ACNP 63rd Annual Meeting: Poster Abstracts P609-P914

December 8-11, 2024

Phoenix, Arizona

Sponsorship Statement: Publication of this supplement is sponsored by the ACNP.

Only disclosures for presenting authors are listed. Asterisks in the author lists indicate presenter of the abstract at the annual meeting.

Abstract numbers do not correlate to poster number assigned for presentation at the Annual Meeting.

P609. Determining Optimal Parameters for Noninvasive Thalamic Neuromodulation With Trancranial Focused Ultrasound

Ryan Ash*, Patricia Limon, Martin Scott, Morteza Mohammadjavadi, Kim Butts Pauly, Anthony Norcia

Stanford University School of Medicine, Stanford, California, United States

Background: Clinical neuroscience has led to a revolution in our understanding of how different neural circuits contribute to phenotypes of neuropsychiatric conditions. Unfortunately, the tools to modulate these circuits in the human brain and thereby ameliorate symptoms are limited by poor focality and depth penetration. Transcranial ultrasound stimulation (TUS) is an emerging tool to achieve noninvasive focal brain-wide neuromodulation with high focality ( < 1cm) and the ability to achieve high intensities in-depth. This technology is at an early stage of development, and many key optimizations are needed to accelerate clinical trials in psychiatry. One key need is to better understand how different sonication parameters relate to neuromodulatory effects. Key parameters include the pulse repetition frequency (PRF), intensity, and duty cycle (DC). Importantly, by varying these parameters it may be possible to have either predominantly suppressive/inhibitory effects or facilitatory/excitatory effects on neural activity and synaptic strength. Previous work by our group and others in large animal models (e.g. Fry et al Science 1958, Mohammadjavadi, Ash et al., Scientific Reports 2022) have demonstrated a suppression of EEG visual-evoked potentials (VEPs) when TUS is targeted to the visual thalamus (lateral geniculate nucleus). We are therefore adapting this paradigm into human, as an efficient testbed to evaluate TUS effects.

Methods: We implemented steady-state visual evoked potential (ssVEP) measures of contrast-response (Ash, Norcia Psychophysiology 2023) and contrast increment detection psychophysics as robust neural and behavioral readouts of subcortical visual pathway function. We developed a robust neuroimaging and simulation pipeline to target LGN, and we are using a neuronavigated depth-steerable 4-element TUS transducer.

Results: Our preliminary data suggests that VEPs can be reversibly suppressed with TUS to the human LGN. We are parametrically varying PRF, intensity, and DC to determine which parameters lead to the most robust suppressive and/or facilitatory online and offline effects.

Conclusions: This work provides the foundation for a dissection of the roles of subcortical and deep cortical nuclei in cognition, emotion, and sensory processing in health and disease.

Keywords: focused ultrasound, Electroencephalography (EEG), Thalamus, noninvasive brain stimulation, Visual plasticity

Disclosure: Nothing to disclose.

P610. Neurocomputational Mechanisms of Cognitive Flexibility and its Modulation via Deep Brain Stimulation of the Ventral Capsule/Ventral Striatum

Jaejoong Kim*, Alik Widge

The University of Minnesota, St Paul, Minnesota, United States

Background: Flexible decision-making requires efficiently integrating new, relevant information into ongoing mental processes to enable context-appropriate behavior. Impaired cognitive flexibility during decision-making is a transdiagnostic symptom across multiple mental illnesses, including OCD, depression, and autism spectrum disorder, and is associated with poor treatment outcomes. However, the mechanisms behind cognitive flexibility, especially how new information is integrated with ongoing contextual information to enable adaptive behavior, remain unclear. Given this incomplete understanding, an effective treatment strategy to enhance cognitive flexibility is still a long way off. Here, we investigated a neurocomputational mechanism for encoding both new task-related and old contextual information in high-frequency neural oscillations and how this information is integrated through cross-frequency coupling between multiple regions to support flexible decision-making. Furthermore, we demonstrated how these local and global neural mechanisms supporting cognitive flexibility can be effectively modulated through deep brain stimulation of the ventral capsule/ventral striatum (VCVS DBS), which has been shown to enhance decision efficiency in previous studies.

Methods: Twenty-one epilepsy patients without psychiatric diagnoses (study 1) and fourteen patients with MDD + /- OCD (study 2) performed the Multi-Source Interference Task with and without VCVS stimulation while undergoing intracranial EEG (study 1) and scalp EEG (study 2) recordings. A novel drift-diffusion model decomposing cognitive control into proactive control (context computation) and flexible control (new information computation) was applied to behavioral data. To investigate local mechanisms of encoding and integrating new and contextual information in the non-DBS state, we performed model-based local field potential (LFP) analyses on high-frequency oscillations (β-band: 15-30 Hz, low γ-band: 30-50 Hz, high γ-band: 70-150 Hz) in brain regions including the bilateral dorsolateral prefrontal cortex (dlPFC), lateral orbitofrontal cortex (lOFC), hippocampus, rostral anterior cingulate cortex (rACC), and dorsal anterior cingulate cortex (dACC) using cluster-based permutation tests (CBPT). To investigate network-level mechanisms of cognitive flexibility, which were hypothesized to support the integration of new and ongoing contextual information, we computed phase-amplitude couplings (PACs) between brain regions and assessed their significance using permutation distributions. Model-based PAC analyses were then conducted using non-parametric linear mixed-effect regression (LMER) analyses. Finally, we tested how VCVS DBS alters these local and global mechanisms of cognitive flexibility using LMER analyses and mediation analyses.

Results: In behavioral analyses, we demonstrated that participants utilized both new task-related and ongoing contextual information, balancing reactive/flexible and proactive cognitive control. VCVS DBS enhanced flexible cognitive control by increasing the general drift rate (p < 0.001) and decreasing the learning rate (p < 0.001). LFP analyses regarding local information processing showed that fast oscillation power in the bilateral dlPFC encoded only task-related new information (p < 0.05 in CBPT), correlating with faster response times (RT; p < 0.001 in LMER), while other regions (bilateral lOFC, rACC, dACC, hippocampus) encoded both new and contextual information (p < 0.05 in CBPT). VCVS DBS enhanced new information encoding in the dlPFC (p < 0.05 in LMER) and decreased old context encoding in the lOFC (p < 0.05 in LMER). However, local information processing mechanisms were not related to the integration process. PAC analyses revealed significant coupling between the phase of θ oscillations in the hippocampus and rACC and the amplitude of high-frequency (β/γ) oscillations in multiple regions (all p < 0.05), indicating their role as phase-modulating hubs in the cognitive control network. rACC-centered PACs were significantly associated with the integration of new and old context-related information (p < 0.05 in LMER), with stronger PACs linked to the suppression of old context information during integration. VCVS DBS accelerated RT in response to new, unexpected contexts by enhancing rACC-centered PACs in that context (p < 0.05 in mediation analysis), indicating increased cognitive flexibility.

Conclusions: We demonstrated that local fast oscillations encode both new and old context-related information. The rACC-centered PACs coordinate these local computations across various brain regions to support information integration. Furthermore, VCVS DBS enhanced cognitive flexibility by optimizing local information processing—specifically, new information processing in the dlPFC and old context-related information processing in the lOFC—and by improving overall integration through the modulation of global computations via the rACC-centered PACs. This provides new insights into how both local and global computations contribute to cognitive flexibility and how neural circuits can be adjusted through neuromodulation to enhance cognitive flexibility.

Keywords: deep brain stimulation, cognitive flexibility, local field potentials, Obsessive Compulsive Disorder, computational psychiatry

Disclosure: Nothing to disclose.

P611. Trajectories of Irritability Improvement in Depressed Adolescents Treated With 1 Hz and 10 Hz Transcranial Magnetic Stimulation

Karina Delaney*, Paul Nakonezny, Arjun Athreya, Jennifer Vande Voort, Magdalena Romanowicz, Can Ozger, Paul Croarkin

Mayo Clinic, Rochester, Minnesota, United States

Background: Prior research with antidepressants suggests that early improvement in irritability may presage response in depressed adolescents. However, this relationship has not been examined in TMS studies of adolescents. This study sought to examine the trajectory of improvement in irritability for depressed adolescents undergoing therapeutic transcranial magnetic stimulation (TMS) and the interaction of these changes in irritability with treatment response.

Methods: Participants aged 12 to 18 years (N = 41) underwent 6 weeks of treatment (30 sessions) in a double-blind, randomized trial of 1 Hz vs. 10 Hz TMS. A logistic regression was implemented to examine the relationship between Clinical Global Impressions-Improvement (CGI-I) treatment response at week 6 (defined as a CGI-I score of 2 or less) and the change in Children’s Depression Rating Scale-Revised (CDRS-R) irritability symptoms from baseline to week 4.

Results: The results from the logistic regression revealed a significant negative (inverse) relationship between the change in irritability symptoms and CGI-I response for the 10 Hz TMS group (b log odds =  -1.5474, SE  =  0.7343, p  =  0.0351) and for the 1 Hz TMS group (b log odds = -1.2852, SE  =  0.5656, p  =  0.0231). In other words, as CDRS-R irritability symptoms improved over the first 4 weeks, the probability of CGI-I response at week 6 increased significantly for both the 1 Hz and 10 Hz TMS groups. However, there was no significant interaction between TMS treatment group and change in irritability symptoms on CGI-I response (p  =  0.7323). Thus, TMS stimulus frequency (1 Hz vs. 10 Hz) did not moderate the relationship between the change in irritability symptoms and CGI-I response. Although this suggests that an improvement in irritability symptoms by week 4 predicts subsequent response to TMS treatment irrespective of TMS frequency, a greater probability of treatment response was observed at lower levels of change in irritability symptoms (e.g., -1 to 1) for the 1 Hz TMS frequency vs. the 10 Hz TMS frequency.

Conclusions: An early reduction in irritability may be a useful predictor of response to TMS in the treatment of adolescent depression. These findings will inform clinical decision making in treating adolescents with depression as well as intervention development for irritability in youth. Future studies should examine the potential differential benefit of 1 Hz TMS for irritability.

Keywords: Adolescent, Irritability, Repetitive transcranial magnetic stimulation, Depression

Disclosure: Nothing to disclose.

P612. The Impact of Accelerated iTBS and cTBS on White Matter Microstructure in OCD - A Pilot Study

Bora Kim*, Andrew Geoly, Flint Espil, Jennifer Lissemore, Wiebke Struckmann, Nolan Williams

Stanford University School of Medicine, Palo Alto, California, United States

Background: Theta burst stimulation (TBS) is a novel repetitive Transcranial Magnetic Stimulation paradigm that employs theta burst stimulation to enhance or dampen cortical excitability by mimicking cortical theta rhythms, allowing for accelerated stimulation in short periods. Despite its potent clinical effects, the mechanisms of TBS, particularly its impact on microstructural integrity, remain largely unknown. This pilot study utilized diffusion tensor imaging (DTI) to examine the relationship between accelerated continuous TBS (acTBS) and intermittent TBS (aiTBS) and white matter microstructure, aiming to further understand the neural mechanisms of action of TBS.

Methods: DTI data were acquired from 21 individuals with Obsessive-Compulsive Disorder (OCD) who participated in an ongoing open-label accelerated TBS study for this pilot analysis [NCT04286126]. The diagnosis of OCD was determined by a structured clinical interview for DSM disorders, and symptom severity was measured using the Yale-Brown Obsessive-Compulsive Scale (YBOCS). All participants received either 18,000 pulses per day of aiTBS to the dorsomedial prefrontal cortex (DMPFC) or acTBS to the right orbitofrontal cortex (rOFC) over 5-10 days at 100% or 110% of the resting motor threshold, respectively. MRI data were acquired at baseline (1 week before treatment) and post-treatment.

Multishell diffusion images (1.5mm isotropic, TR/TE = 5650/65 ms, b = 1500, 3000s/mm2, 65 directions) and Anatomical images (3D, T1-weighted, FOV = 256x256mm, 0.9mm isotropic, TR/TE = 6.39/2.62ms) were acquired using a 3T GE Discovery MR750 scanner with a 32-channel head-neck imaging coil at the Center for Cognitive and Neurobiological Imaging at Stanford University. Participants were screened for MRI safety before scanning procedures. We chose to examine fractional anisotropy (FA) changes in five regions of interest (ROI) from a recent ENIGMA paper on OCD. Tract-based spatial statistics (TBSS) were employed to analyze skeletonized microstructural estimates for each ROI. Repeated-measures ANOVA was performed with covariates, including stimulation targets of DMPFC and rOFC, follow-up time between visits (days), sex, age, and baseline OCD symptom severity as measured by the YBOCS.

Results: Demographics included an age range of 18-57 years old (mean 37.3 years old), with 45% male. The mean (SD) baseline YBOCS score was 30.62 (3.92). Thirteen individuals received aiTBS over DMPFC, and eight individuals received cTBS over rOFC. Post-treatment follow-up time ranged from 20 days to 524 days (median 42 days). A repeated-measures ANOVA showed a main effect of study visit on FA in the Uncinate fasciculus after including covariates of age, sex, treatment duration and follow-up time (F = 4.736, p = 0.049, η²p = 0.267), but post-hoc testing comparing FA between Pre- and Post-treatment did not substantiate the main effect (p > 0.05). No significant differences were detected for radial diffusivity (RD), axial diffusivity (AD), or mean diffusivity (MD) in any ROI. ROI-based TBSS analysis did not demonstrate significant white-matter microstructural changes from baseline to post-treatment in specific brain regions.

Conclusions: Our pilot data (N = 21) analysis suggests the need for a larger sample to thoroughly investigate the effects of acTBS and aiTBS on white matter microstructure with adequate statistical power.

Keywords: TMS, Diffusion Tensor Imaging (DTI), Obsessive-Compulsive Eisorder (OCD)

Disclosure: Nothing to disclose.

P613. Effects of Individualized Transcranial Magnetic Stimulation on Social Cognitive Network Functional Connectivity in Schizophrenia Spectrum Disorders: A Target Engagement Study

Lindsay Oliver*, Daniel M. Blumberger, Colin Hawco, Erin Dickie, Julia Gallucci, Jerrold Jeyachandra, Salim Mansour, Zhi-De Deng, Stephanie Hare, James Gold, George Foussias, Miklos Argyelan, Zafiris Daskalakis, Robert Buchanan, Anil Malhotra, Aristotle Voineskos

Centre for Addiction and Mental Health; University of Toronto, Toronto, Canada

Background: Individuals with schizophrenia spectrum disorders (SSDs) often experience social cognitive impairments, which are associated with poor functional outcomes. There are currently no approved treatment options for these debilitating deficits in SSDs or other brain disorders. Social processes affected in SSDs are thought to rely on a frontoparietal ‘simulation network’ and a cortical midline/lateral temporal ‘mentalizing network’, which may be amenable to modulation using neurostimulation. This target engagement study examined, for the first time, the effect of individually-targeted repetitive transcranial magnetic stimulation (rTMS) to the dorsomedial prefrontal cortex (DMPFC) on social cognitive network functional connectivity in individuals with SSDs.

Methods: We conducted a two-week, double-blind, randomized clinical trial comparing conventional 10 Hz rTMS and intermittent theta burst stimulation (iTBS), a more time-efficient rTMS paradigm, to sham stimulation of the DMPFC in clinically stable adults with SSDs across sexes at three sites. Our primary outcome measure was change in within mentalizing network connectivity. Our secondary outcome measure was tolerability. We also explored connectivity changes between the mentalizing and simulation networks. Participants completed pre- and post-treatment magnetic resonance imaging (MRI) scans, including the social processing Empathic Accuracy task. Neurostimulation (ten 10 Hz rTMS, iTBS, or sham sessions) was individually targeted by combining functional connectivity mapping and electric field modeling, which account for functional and anatomical brain variability respectively, using our group’s Bayesian Optimization of Neuro-Stimulation (BOONStim) pipeline. Following MRI preprocessing, within-mentalizing and between mentalizing-simulation network functional connectivity metrics during the Empathic Accuracy task were calculated both pre- and post-rTMS. Given that this was a feasibility study designed to determine effect sizes and tolerability, Cohen’s d effect sizes were used to compare the magnitude of changes after 10 Hz rTMS and iTBS versus sham stimulation.

Results: Of the 51 randomized participants, 46 completed treatment and 43 (10 Hz rTMS N = 15, iTBS N = 15, sham N = 13) were included in the imaging analyses following quality control. There was a greater increase in mentalizing network connectivity following iTBS compared to sham (d = 0.53, 95% CI [-0.26, 1.32]). Negligible differences were observed in changes in mentalizing network connectivity following 10 Hz rTMS when compared to sham (d = 0.07, 95% CI [-0.85, 0.71]). Little change was seen in mentalizing-simulation between-network connectivity when comparing iTBS to sham and 10 Hz rTMS to sham. Pain scores were low in all groups (mean < 5/10) and there was no difference in drop-out rate among the three groups, suggesting good tolerability.

Conclusions: This study marks the first investigation of rTMS effects on social cognitive circuitry in SSDs, incorporating individualized targeting approaches and inclusion of both a 10 Hz rTMS and a more time-efficient iTBS arm. Our results suggest that iTBS can change mentalizing network connectivity with medium effect size. We are now embarking on a larger clinical trial using iTBS versus sham stimulation to assess changes in social cognitive performance.

Keywords: schizophrenia spectrum disorders, repetitive transcranial magnetic stimulation (rTMS), social cognition, fMRI Functional Connectivity, personalized treatments

Disclosure: Nothing to disclose.

P614. Independent Validation and Network Engagement of a Psychosis-Specific, Personalized rTMS Target for Nicotine Dependence

Heather Ward*, Sophia Blyth, Gulcan Yildiz, Baxter Rogers, Simon Vandekar, Brett Clementz, Elliot Gershon, Matcheri Keshavan, Godfrey Pearlson, Carol Tamminga, Mark Halko, Roscoe Brady

Vanderbilt University Medical Center, Nashville, Tennessee, United States

Background: Tobacco use is the top preventable cause of early mortality in schizophrenia, but the underlying pathophysiology explaining the high prevalence of tobacco use remains unknown. Previous research in small samples has linked organization of the default mode network (DMN) to tobacco use in schizophrenia and observed that acute nicotine administration normalizes DMN hyperconnectivity in schizophrenia. We therefore sought to 1) replicate the relationship between DMN organization and tobacco use using a large, multisite sample of individuals with psychotic disorders and healthy controls (Bipolar-Schizophrenia Network on Intermediate Phenotypes 2, B-SNIP2); and 2) to test if modifying this network with repetitive transcranial magnetic stimulation (rTMS) would affect craving.

Methods: Using data from B-SNIP2 (n = 596, 320 psychosis, 276 healthy), we calculated average connectivity of the DMN and tested the associations between DMN connectivity and tobacco use. We then conducted a randomized, controlled, crossover study of single rTMS sessions (intermittent theta burst stimulation, iTBS; continuous theta burst stimulation, cTBS; and sham) applied to an individualized left lateral parietal DMN target with fMRI and craving assessment immediately before and after rTMS in people with schizophrenia and nicotine dependence (n = 10). Connectivity change from the rTMS target was analyzed with longitudinal repeated measures analysis. We used the SPM Sandwich Estimator Toolbox for Longitudinal and Repeated Measures Data (SwE) to identify patterns of connectivity change from the rTMS target. We first generated connectivity maps from the left lateral parietal DMN rTMS target for each MRI scan. We then created difference maps for each rTMS session by subtracting the pre-rTMS map from the post-rTMS map. We used the SwE toolbox to model 1) the effect of rTMS session (iTBS, cTBS, sham) on connectivity change and 2) the effect of craving change on connectivity change using 20,000 bootstraps. We performed voxel-wise tests with FDR correction. Mean connectivity values were extracted from the largest clusters and used to visualize the results. We compared craving change between each rTMS type using a mixed ANOVA.

Results: In B-SNIP2, we observed associations between history of tobacco use and DMN connectivity in the entire sample (p = 0.028) such that current smokers had lower DMN connectivity than former and never smokers. These differences were driven by the psychosis group (p = 0.023) and were not evident in controls (p = 0.66). We also observed a dose-response relationship where higher cigarettes smoked per day was associated with lower DMN connectivity in the entire sample (r = -0.21, p = 0.0095), which was driven by the psychosis group (r = -0.22, p = 0.013) and was not present in controls (r = -0.026, p = 0.915).

In our study of DMN-targeted rTMS, we observed a treatment x time relationship (p = 0.017) where craving was increased by iTBS but not by cTBS or sham. We also observed a treatment x time relationship (mean pFDR < 0.011) where connectivity between the DMN rTMS target and dorsal anterior cingulate cortex (a key region in the salience network) increased with iTBS and decreased with cTBS and sham. Craving change was related to connectivity change within the DMN (mean pFDR < 0.012), where cTBS reduced within-DMN connectivity. Increased craving was associated with increased connectivity within the DMN (r = 0.60, p =.003).

Conclusions: Using a large, multisite dataset of individuals with psychotic disorders, we observed psychosis-specific relationships where lower DMN connectivity was associated with higher daily cigarette use and recent tobacco use. We then showed that DMN-targeted rTMS modulated DMN-salience network connectivity and modulated craving via within-DMN connectivity. Our findings suggest a psychosis-specific mechanism for the high prevalence of nicotine use in this population and a novel target for intervention.

Keywords: nicotine dependence, repetitive transcranial magnetic stimulation (rTMS), Default mode network (DMN), resting-state functional connectivity, psychosis

Disclosure: Nothing to disclose.

P615. Toward Individualizing TMS Treatment Targets for PTSD Using Neuroimaging

Sanne van Rooij*, Cecilia Hinojosa, Malin Au, Lois Teye-Botchway, Sean Minton, Ryan Langhinrichsen-Rohling, Timothy Ely, Gregory Job, Patricio Riva-Posse, Kerry Ressler, Nadine Kaslow, Sheila Rauch, Tanja Jovanovic, Paul Holtzheimer, Vince Calhoun, Joan Camprodon, William McDonald

Emory University School of Medicine, Atlanta, Georgia, United States

Background: About 30-50% of patients with posttraumatic stress disorder (PTSD) do not respond to, or are not able to engage in, currently recommended trauma-focused therapies for PTSD, signifying the need for alternative treatment approaches. Neuroimaging studies report a neurobiological profile of PTSD treatment response that most consistently points to over engagement of the (right) amygdala as the main driver of non-response. Transcranial Magnetic Stimulation (TMS) is a non-invasive neuromodulation treatment, and prior studies have shown promising effects of 1Hz TMS to the right dorsolateral prefrontal cortex (DLPFC) on PTSD symptoms. However, the exact mechanisms of TMS are unclear and there is no method to select the best TMS target for an individual patient. Here we investigate pre-TMS neuroimaging to define the target within the rDLPFC most strongly positively connected with the right amygdala as a method to individualize TMS targets for PTSD.

Methods: Fifty participants with PTSD symptoms were enrolled and resting state (RS) functional neuroimaging data was collected for n = 39 (n = 4 dropout; n = 7 discontinued per protocol). RS functional connectivity (FC) was used to define the area within the rDLPFC most strongly positively connected with the right amygdala. The resting state functional MRI scans were collected on a 3T Siemens scanner and were pre-processed using afni_proc.py. A correlation map between the extracted timecourse of the r amygdala ROI (CIT168) and the rDLPFC ROI (BA8,9,19,46; WFU_PickAtlas) was generated and unwarped into native space. Brainsight neuronavigation was used to identify the TMS target using three criteria: 1. Largest positive RSFC peak, 2. Peak within 1.0 cm of the cortex, 3. Location tolerable for 1Hz 20 TMS sessions as part of a clinical trial. To determine TMS target variability and stability, X, Y, Z MNI target coordinates were extracted and exported to SPSS v.29. Repeated measures general linear models were conducted for three analyses: 1) Variability (between subjects): to compare the positively defined target between subjects, 2) Variability (between targeting method): to compare targets defined using positive versus anti-correlation, 3) Stability (within subjects): to compare positively defined targets before and after engagement of the fear neurocircuitry during the MRI protocol. The fear neurocircuitry was engaged with three functional MRI tasks, i.e., social threat processing using fearful faces, contextual response inhibition, and fear conditioning.

Results: Using motion cutoff < 0.8mm (framewise displacement), RSFC data for n = 35 (83% female, 29% Black, 6% Hispanic) were used for functional targeting and analyses. Individualized TMS targeting for PTSD showed significant variability between participants (F(1,34) = 5,929, p < 0.001, ηp2 = 0.99). Significant variability between targeting methods was observed such that targets defined using positive versus anti-correlation RFSC were significantly different (F(1,34) = 29.17, p < 0.001, ηp2 = 0.46). Finally, stability of the target over time before and after engagement of the fear neurocircuitry was observed (F(1,33) = 0.14, p = 0.712, ηp2 = 0.004).

Conclusions: This novel functional targeting approach is an important step towards individualized TMS treatments for patients with PTSD symptoms. Here we address three important questions regarding individualized targeting and show significant variability in targets between participants and for different targeting methods, suggesting the relevance of RSFC-based individualized targeting for TMS. Importantly, the target location did not change after engaging the fear neurocircuitry, providing important data for the stability of this TMS target. This data therefore supports the potential of fMRI-based individualized TMS targeting in the first clinical trial to use RSFC-based targeting in PTSD.

Keywords: transcranial magnetic stimulation, Posttraumatic stress disorder, personalized treatments, Resting State Functional Connectivity, Amygdala

Disclosure: Nothing to disclose.

P616. Strength of tDCS-Induced Electrical Field is Associated With Psychophysiological Habituation and Predicts Symptom Reduction in Response to tDCS+virtual Reality Exposure for PTSD

Mascha van ‘t Wout-Frank*, David Sorensen, Lucas Parra, Yu Huang, Alexis Harle, Benjamin Greenberg, Amanda Arulpragasam, Noah Philip

Warren Alpert Medical School of Brown University;Center for Neurorestoration and Neurotechnology, Providence VA Healthcare System; COBRE Center for Neuromodulation, Butler Hospital, Providence, Rhode Island, United States

Background: In a recent sham-controlled double-blind trial, we replicated and extended pilot results that 2mA transcranial direct current stimulation (tDCS) targeting ventromedial prefrontal cortex combined with virtual reality exposure (tDCS+VR) vs. sham+VR, improved PTSD symptoms and functioning in warzone-exposed veterans, with effects lasting up to three months. tDCS+VR, vs. sham+VR, also augmented psychophysiological habitation to VR warzone stimuli. While these results suggest tDCS+VR engaged brain circuitry implicated in PTSD, an important hypothesis to test is that greater current delivered to PTSD-relevant neural regions yielded greater clinical benefit and greater psychophysiological habituation. If confirmed, this would fit with the theory that tDCS augments learning-related neuroplasticity for clinical benefit and provide an important demonstration that electrical field modeling can be used to predict therapeutic success. We present planned secondary analyses testing the prediction that e-field strength in PTSD-relevant neural regions correlates with greater psychophysiological habituation during VR and reduction in PTSD symptom severity, and that this is specific to participants randomized to active stimulation+VR.

Methods: Forty-four of the 54 veteran participants in our clinical trial (NCT03372460), had T1-weighted MRI; half (n = 22) were randomized to six sessions of active tDCS+VR and half to sham+VR. Active tDCS involved 2mA stimulation, 3x3cm electrodes (2.22 A/m2), the anode over EEG coordinate Fp1/AF3 and the cathode over P8/PO8. We modeled electric field strength delivered to PTSD- and threat-processing implicated regions using open-source “Realistic vOlumetric Approach to Simulate Transcranial Electric Stimulation” (ROAST) software. A priori regions chosen were: ventromedial prefrontal cortex (including Brodmann Area (BA) 10, orbitofrontal (BA 11), subgenual (BA 25) and anterior cingulate cortex (BA 32)), insula, hippocampus, amygdala, and dorsolateral prefrontal cortex (DLPFC (BA 9 and 46)). Individual e-fields were then correlated with 1) the degree of psychophysiological habituation, quantified as the change in skin conductance reactivity to VR events from the first to the last completed VR session, and 2) reduction in self-reported PTSD symptoms from baseline to the end of the tDCS+VR intervention. We also tested the association between PTSD symptom reduction the degree of habituation during VR. Correlations were tested with Pearson r or Kendall’s Tau-b (τb), depending on whether normality was met based on Shapiro-Wilk test for small sample size. Benjamini-Hochberg correction was applied to adjust for multiple comparisons. Two-tailed α = 0.05 was used for significance.

Results: In the active tDCS+VR group, e-field strength in insula (r = 0.63, p = 0.003, adj-p = 0.01), amygdala (r = 0.58, p = 0.007, adj-p = 0.01), hippocampus (r = 0.62, p = 0.004, adj-p = 0.01), and DLPFC (τb = 0.51, p = 0.002, adj-p = 0.01) correlated significantly with greater habituation. The correlation between anterior cingulate cortex (BA 32) modeled e-field and habituation was significant but did not survive correction (r = 0.34, p = 0.04, adj-p = 0.06). E-field strength in the insula (r = 0.43, p = 0.049, adj-p = 0.13), DLPFC (τb = 0.39, p = 0.01, adj-p = 0.14) and ventromedial prefrontal cortex (BA10; τb = 0.32, p = 0.04, adj-p = 0.1) also correlated significantly with greater PTSD symptom reduction, but these did not survived correction. As expected, there were no significant correlations between e-field strength, the degree of habituation, and PTSD symptom reduction in the sham+VR group (all p > .05). Fisher r-to-z transformation revealed that the correlation between e-field strength in the insula and habituation success was significantly greater in the active tDCS vs. the sham group (z = 2.03, p = 0.04). Finally, in the active tDCS+VR group only, degree of habituation correlated significantly with PTSD symptom reduction (r = 0.41, p = 0.049). Groups did not differ in age (t(42) = -.59, p = 0.55) or sex (Χ2(1) = 0.36, p = 0.55).

Conclusions: The modeling results suggest that active tDCS, when applied during VR exposure, provides current to cortical and subcortical regions implicated in PTSD and threat processing (i.e., insula, amygdala, hippocampus, and DLFPC) at intensities sufficient for neuromodulation to boost learning-related habituation to improve clinical response. Although it should be noted that e-field strength in the ventromedial prefrontal cortex (BA 10), our targeted region, was not associated with either habituation or clinical improvement after correction. Still, these findings expand our prior observations by pointing towards region-specific tDCS action. This approach may inform a precision medicine approach where pre-treatment individual e-field modeling might predict outcomes of tDCS combined with therapeutically relevant exposure.

Keywords: Posttraumatic stress disorder, virtual reality, electrical field modeling, transcranial direct current stimulation, habituation

Disclosure: Nothing to disclose.

P617. Repetitive Transcranial Focused Ultrasound Amygdala Neuromodulation for Transdiagnostic Mood, Anxiety, and Trauma-Related Symptoms

Bryan Barksdale*, Lauren Enten, Annamarie DeMarco, Rachel Kline, Manoj Doss, Charles Nemeroff, Gregory Fonzo

Dell Medical School, Austin, Texas, United States

Background: Mood, anxiety, and trauma-related disorders (MATRDs) are characterized by prolonged states of negative affect triggered by reactivity to emotionally provocative stimuli. Hyperactivity of the amygdala, a subcortical limbic brain structure underlying the detection of salient environmental stimuli, is thought to underlie this common exaggerated emotional reactivity phenotype. Currently available non-invasive neuromodulatory therapies, such as transcranial electrical (TES) and transcranial magnetic stimulation (TMS), are limited by their capacity to modulate the function of only superficial cortical regions. Transcranial focused ultrasound (tFUS) is a novel, safe, non-invasive neuromodulation technique that can target and modulate the function of deep subcortical brain structures, such as the amygdala. In-house preliminary and published data have identified a tFUS neuromodulation protocol that attenuates human amygdala functional magnetic resonance imaging (fMRI) blood oxygenation level dependent (BOLD) signal. In a neuroimaging-coupled open-label trial, we tested if a 3-week daily application of this inhibitory protocol, i.e. repetitive transcranial focused ultrasound (rtFUS), targeted to the left amygdala would attenuate symptoms and amygdala reactivity in a transdiagnostic sample of patients with MATRDs. We also tested if amygdala responses to naturalistic probes (emotional faces) would show a lasting attenuation from pre- to post-treatment.

Methods: To first establish target engagement, we used MRI-guided targeting of the left amygdala and applied active vs. sham tFUS (a 10-min protocol) during functional magnetic resonance imaging (fMRI) in a double-blind, repeated-measures, within-subject design. Participants also underwent task-based fMRI with a popular emotional face matching task prior to and after each stimulation session. We first tested if active vs. sham tFUS modulated left amygdala BOLD signal in MATRD patients (N = 29) and healthy controls (HCs)(N = 23) and examined group differences in degree and magnitude of evoked effects. Using neuronavigation and individualized targeting, MATRDs (N = 24) then underwent open-label (OL) once-daily rtFUS treatment targeting the left amygdala for 3 weeks with the same tFUS inhibitory protocol to assess possible therapeutic potential of repetitive administration. In both in-scanner administration sessions and open-label treatment sessions, participants received the same rtFUS stimulation protocol. This protocol consisted of the following parameters: 10 Hz pulse repetition frequency, 5% duty cycle, 5ms pulse width, derated spatial peak pulse average intensity of 14.4 watts/cm2, a derated spatial peak temporal average intensity of 719.91 milliwatts/cm2, and a derated instantaneous peak pressure of 0.64 megapascals over 10 minutes in 10 sets of 30 second on/off blocks. Target engagement rtFUS effects on blood oxygenation level dependent (BOLD) signal were quantified with a block regressor (modeling rtFUS on vs. off) in separate sessions with visually identical active and sham (FUS-blocking) transducer pads. Voxelwise linear mixed models with probabilistic threshold free cluster enhancement (pTFCE) multiple comparison correction identified main effects of active vs. sham tFUS and interactions with group (patients vs. healthy controls). Voxelwise linear mixed models were also used to assess pre- to post-treatment changes in amygdala responses in patients to emotional stimuli and pre- to post-treatment changes on the primary outcome symptom outcome measure, which was the 30-item Mood and Anxiety Symptom Questionnaire General Distress (MASQ-GD) subscale.

Results: A wide spectrum of MATRD diagnoses were represented in the patient sample, including major depression, bipolar disorder, several anxiety disorders, and posttraumatic stress disorder. Analysis of concurrent tFUS/fMRI demonstrated target engagement, i.e. a significant reduction in left amygdala BOLD signal when tFUS was on vs. off for active vs. sham administrations in patients and healthy controls. Analyses of patient vs. healthy differences revealed a failure in patients to inhibit the contralateral (right) anterior hippocampus, which was deactivated by active tFUS to the left amygdala in healthy controls. In the open-label trial, rtFUS was well-tolerated (no serious adverse events) with high completion rates (N = 21 completing). A significant reduction was observed on our primary outcome (Mood and Anxiety Symptom Questionnaire General Distress subscale; F = 12.89, p = 0.001; Cohen’s d = 0.77), several secondary symptom outcomes (Cohen’s d of 0.51-1.50), and amygdala blood oxygenation level-dependent (BOLD) functional magnetic resonance imaging (fMRI) responses to naturalistic emotional stimuli.

Conclusions: These findings provide compelling initial evidence of safety, tolerability and potential efficacy of rtFUS, a novel non-invasive subcortical neuromodulation technique, and motivate conduction of double-blind randomized trials to confirm safety/efficacy.

Keywords: focused ultrasound, affective disorders, Amygdala, Human Clinical trial, Open Label Trials

Disclosure: Nothing to disclose.

P618. Inhibition of the Left Amygdala via Low-Intensity Focused Ultrasound Enhances the Encoding of Emotional and Neutral Episodic Memories

Manoj Doss*, Sydney Lambert, Charles Nemeroff, Gregory Fonzo, Joseph Dunsmoor

The University of Texas at Austin Dell Medical School, Austin, Texas, United States

Background: The amygdala is thought to be crucial to the formation (i.e., encoding) of emotional episodic memories by enhancing the salience of survival-related stimuli and post-encoding processing of these stimuli. However, causal evidence for this assertion is limited. In typical emotional episodic memory paradigms, memory for negative emotional information, specifically, is typically > memory for neutral and positive. Although patients with amygdala lesions exhibit less preferential memory for negative stimuli, there may be compensatory processes that develop over time, obfuscating the role of the amygdala in intact brains.

Methods: In an ongoing double-blind, sham-controlled, repeated measures study, we non-invasively sonicated the left amygdala with low-intensity focused ultrasound using a standard inhibitory protocol during the encoding of emotional episodic memories. Following localization of the amygdala with structural magnetic resonance imaging, participants (N = 18) received active or sham sonication of their left amygdala. Immediately after, participants completed the encoding phase of an emotional episodic memory task in which emotionally negative, neutral, and positive pictures were presented. The following day, participants completed the retrieval phase in which memory for pictures was tested on a cued recollection test and picture recognition test. We hypothesized that active vs. sham sonication would selectively attenuate the typical memory enhancement for negative vs. neutral and positive stimuli.

Results: Surprisingly, active vs. sham sonication of the left amygdala during encoding enhanced memory for negative, neutral, and positive stimuli (main effect of sonication on cued recollection memory accuracy: F(1,17) = 4.35, p = 0.005, η_p^2 = 0.20), without differential effects of sonication on emotional stimuli (sonication × emotion interaction on cued recollection memory accuracy: (F(2, 34) = 0.86, p > .250). Additionally, memory enhancements were found across multiple other measures on both memory tests and came from an increase in hit rates (rather than a decrease in false alarm rates).

Conclusions: Contrary to expectations, inhibition of the left amygdala enhanced episodic memory in an emotionally non-specific manner. These findings motivate a novel hypothesis for the role of the amygdala in emotional episodic memory. Rather than the amygdala enhancing memory via amplification of salient (e.g., emotional) stimuli, it may instead act as a filter that attenuates the maintenance of non-salient stimuli in long-term memory. In patients with amygdala lesions, there are potentially compensatory processes that prevent global episodic memory enhancements. However, with temporary inhibition, these findings that suggest memory can be acutely enhanced, thereby opening avenues for transient memory enhancements in disorders exhibiting mnemonic impairments.

Keywords: focused ultrasound, episodic memory, emotional memory, Amygdala, encoding

Disclosure: Nothing to disclose.

P619. rtfMRI-Neurofeedback Training to Increase the Amygdala’s Response to Positive Memories in Adults With Borderline Personality Disorder

Laurie Compere, Scott Barb, Carolyn Webb, Emily Riley, Emily Leiker, Greg Siegle, Kymberly Young*

University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, United States

Background: The primary aim of this study is to adapt an intervention for major depressive disorder (MDD) to a population with borderline personality disorder (BPD), based on similar hypothesized mechanisms across these disorders. Patients with BPD have several symptoms in common with patients with MDD, including the same profile of reduced positive memory recall and decreased activity in regions important for salience and self-referential processing during processing of positive stimuli, including the amygdala. It is therefore likely that patients with BPD can also benefit from our neurofeedback intervention targeted at increasing the amygdala response during positive memory recall, which has been shown to be effective at reducing symptoms of MDD.

Methods: 10 participants diagnosed with BPD completed two rtfMRI-nf sessions in which they were trained to increase their amygdala response during positive autobiographical memory recall. Symptoms of BPD and MDD were assessed at baseline, immediately following the intervention, and 12-weeks following the intervention via the BEST and BDI-II.

Results: Participants were able to significantly increase their amygdala response from the initial baseline while engaging in positive memory recall. Symptoms of both BPD and MDD significantly decreased from baseline to the 12-week follow up at p < 0.05. There was an average decrease of 8 points (SD 4.98) on the BEST and an average decrease of 6 points (SD 3.15) on the BDI-II.

Conclusions: This small proof-of-concept study suggests that amygdala rtfMRI-nf may be a promising intervention for BPD. Symptoms of both BPD and MDD significantly decreased following the neurofeedback intervention. Follow-up randomized controlled trials are warranted.

Keywords: Borderline Personality Disorder, real-time fMRI neurofeedback, Amygdala, autobiographical memory

Disclosure: Nothing to disclose.

P620. Electroconvulsive Therapy Generates a Hidden Wave of Activity After Seizure in Mice and Humans

Zach Rosenthal*, Joseph Majeski, Ala Somarowthu, Mary Putt, Antoneta Karaj, Chris Favilla, Golkoo Hosseini, Jenny Rodriguez, Mario Cristancho, Yvette Sheline, Arjun Yodh, Ethan Goldberg

Department of Psychiatry, University of Pennsylvania, Philadelphia, Pennsylvania, United States

Background: Electroconvulsive therapy (ECT) is a life-saving intervention for medication-resistant depression, achieving rapid remission in 75% of patients and reducing suicide risk by 50%. Despite its safety and efficacy, ECT has been historically understudied, underutilized, and stigmatized. Much remains unclear about the physiology underlying ECT’s therapeutic mechanism.

No study to date has recorded brain activity in human subjects or animal models during ECT pulse delivery. Typically, 2-channel electroencephalography (EEG) monitors seizure duration, but there is a need for more robust biomarkers that measure therapeutic plasticity, detect side effects, or rationalize stimulation parameters. The choice of parameters like pulse width and electrode configuration is known to modulate ECT efficacy and side effects, but the remaining stimulation parameter space is vast and largely unexplored.

To address these gaps, we use optical neuroimaging to record brain activity during ECT in mice and humans. We reveal that ECT reliably induces a second, larger-amplitude event immediately after seizure that has gone undetected for 86 years.

Methods: Mouse studies were approved by the Institutional Animal Care and Use Committee at the Children’s Hospital of Philadelphia; human studies were approved by the Institutional Review Board of the University of Pennsylvania.

In mice, we implanted 5 electrodes and a clear polymer cranial window for optical access to the cortex in transgenic Thy1-jRGeCO1a animals (n = 10; 10-16 week old; mixed sex) that express a fluorescent reporter of neuronal calcium dynamics. We recorded 159 ECT treatments, surveying the stimulation parameter space (10 electrode configurations, pulse frequencies: 5-100 Hz, current: 1-25 mA) while monitoring cortical activity using widefield imaging of neuronal fluorescence and hemoglobin optical intrinsic signal. Data were processed in MatLab and statistically analyzed with linear mixed effects modeling.

Inspired by these preclinical results, we developed a prospective single-center observational cohort study using non-invasive optical monitoring of brain oxygenation and blood flow in human patients receiving ECT. A total of 19 recordings from n = 11 patients have been collected to date.

Results: In mice, we find that with increasing pulse current, all ECT treatments reached a threshold where seizure was followed by a second, larger traveling wave of activity called spreading depolarization (SD). Also known as a “brain tsunami”, a SD is a high amplitude, slowly propagating wave of depolarization, that is nearly invisible on routine electroencephalography (EEG) due to highpass filters and volume conduction.

Electrode configuration shapes the spatial topography of both the initial evoked seizure and subsequent SD. Seizures spread bilaterally, with spatial biasing depending on electrode position: bilateral electrode placement produced homotopically symmetric seizures, while unilateral placement resulted in asymmetric seizures (Kruskal-Wallis p = 0.0026). SDs occurred in an all-or-none fashion, recruiting one or both entire hemispheres. Unilateral stimulation tended to generate single-hemisphere SDs, while bilateral stimulation elicited bilateral SDs (χ² test, p < 0.0001). SD foci originated near electrodes (permutation test, p < 0.0001) and regions of highest seizure amplitude (p < 0.0001).

Pulse train parameters also influence seizure and SD. Increasing current resulted in higher amplitude (p < 0.001, mixed effects model) and longer duration seizures (p < 0.001). Higher amplitude seizures were associated with SDs, but seizure duration was not. The same pulses produced distinct seizure and SD patterns depending on inter-pulse spacing. Low-frequency pulses (5 Hz) generated longer, lower amplitude seizures, requiring more current to reach the SD threshold. High-frequency pulses (100 Hz) induced briefer, higher amplitude seizures, needing less current for SD.

We hypothesized SDs occur in humans receiving ECT. Since EEG cannot reliably detect SD waves, we used diffuse optical monitoring to detect hemodynamic signatures of SD. Pilot data from n = 11 patients revealed transient blood flow and oxygenation changes during seizure, followed by a secondary hemodynamic waves lasting several minutes, consistent with SD. Similar to mice, electrode placement in humans appears to spatially bias the amplitude of hemodynamic changes during seizure and putative SD, which were nearly twice as large in the right hemisphere in right unilateral ECT. Ongoing data collection will explore the relationship between SD waves, stimulation parameters, seizure characteristics, and treatment outcomes.

Conclusions: It has long been assumed that seizure is the central outcome of ECT. This investigation reveals that in both mice and humans, seizure is a means to another end: spreading depolarization (SD). We further highlight opportunities for refining stimulation parameters in ECT using the first real-time measurement of brain activity during ECT pulse delivery. We find in mice that high-frequency pulses generate briefer, higher amplitude seizures, and require less current to achieve post-ictal SD. In both mice and humans, we find that electrode configuration recruits brain activity regionally during seizure and SD. This framework may enable future studies on ECT stimulation parameter optimization, circuit-level mechanisms of plasticity after ECT, and novel neuromodulatory therapies.

Keywords: electroconvulsive therapy, optical imaging, individualized Low Amplitude Seizure Therapy (iLAST), Seizure, electric field modeling

Disclosure: Nothing to disclose.

P621. Dose-Related Effects of Smoked Cannabis on Simulated Driving Performance: A Randomized, Controlled Crossover Trial

Bernard Le Foll*, Adam Zaweel, Madison Wright, Justin Matheson, Patricia Di Ciano, Bruna Brands

University of Toronto / CAMH / Waypoint, Toronto, Canada

Background: Δ9-tetrahydrocannabinol (THC) impairs driving-related skills and increases the risk of collision involvement. However, in contrast with alcohol, the relationship between THC dose and driving outcomes remains unclear. Driving studies have not yet tested high doses of THC, despite the increased availability of high-potency THC products on the market. Using objective driving simulator parameters, we sought to examine the relationships between driving performance and THC dose.

Methods: In this randomized, double-blind, placebo-controlled, within-subjects trial, adult cannabis users aged 19-36 years, consuming cannabis 1-5 days/week, attended four driving simulation sessions at the Center for Addiction and Mental Health. Sessions were spaced at least 72 hours apart and were identical except for the potency of cannabis administered that day. Participants received placebo, low-dose (6.25%/47 mg THC), medium-dose (12.5%/94 mg THC) and high-dose (22%/165 mg THC) in a randomized and counterbalanced order and drove the simulator at baseline, and 30 and 90 minutes after smoking. At different intervals, blood, oral fluid, and vital signs were collected and participants completed self-perceived driving questionnaires. The primary outcome measure was mean speed (km/h). Secondary outcomes included: maximum speed, standard deviation of lateral position (SDLP, meters), reaction time (RT, seconds), subjective measures of driving abilities and intoxication. Linear mixed modelling was used to test our primary aim, with driving outcomes as dependent variables. Condition (Placebo, Low-dose, Medium-dose, High-dose), and condition interacting with time (30 minutes, 90 minutes) were the predictors of interest. The same model was used for estimating the association between THC concentrations in blood and driving outcomes.

Results: Of the 36 adults (19 females; 17 males) randomized, one was excluded from analysis due to high baseline THC concentrations. No difference in mean speed was detected. Maximum speed was elevated in the high-dose (Δmean=2.4 km/h, p = 0.013) and medium-dose (Δmean=2.7 km/h, p = 0.006) conditions compared to placebo. SDLP was significantly higher at all active THC conditions compared to placebo (p < 0.001) and the high-dose condition resulted in significantly higher SDLP than both the medium-dose (Δmean=0.013m, p = 0.040) and low-dose conditions (Δmean=0.013m, p = 0.032). For RT, the high and medium-doses significantly increased RT compared to the low-dose (Δmean=0.038s, p < 0.001; Δmean=0.031s, p = 0.006, respectively) and placebo (P < 0.001) conditions. SDLP (slope=0.01, SE = 0.001, p < 0.001) and RT (slope=0.01, SE = 0.003, p < 0.001) showed significant positive associations with levels of THC in blood. This means that approximately for every 1 ng/mL increase in blood THC, SDLP increased by approximately 1 cm, and RT increased by 10 milliseconds (0.01 seconds). We observed high rates of adverse events with high THC doses. Subjects reported worse driving performance and lower skill and willingness to drive at higher THC doses.

Conclusions: Our results suggest dose-related effects of THC on measures of driving performance. While minimal differences were observed between the medium and low doses, there was a consistent and significant difference in driving outcomes associated with the high-dose. Driving impairment effects of cannabis appear to be more prominent at higher doses, which presents a significant public health concern considering the prevalence of high-THC cannabis products currently available in the market.

Keywords: cannabis use, Simulated Driving, Clinical trial, cannabis

Disclosure: Pfizer: Grant (Self). Shinogi: Consultant (Self). Indivor: Grant (Self). Brainsway: Other Financial or Material Support (Self). Canopy Growth Corporation: Grant (Self). Center for Addiction and Mental Health: Employee (Self). Waypoint Centre for Mental Health Care: Employee (Self).

P622. Characterization of Distinct NMDA Receptor Modulator Approaches to Evaluate as Potential Therapeutics

Tanya Wallace*, Liam Scott, Tarek Deeb, Megan Fitzgerald, Ryan Hudson, Nicholas Brandon

Neumora Therapeutics, Watertown, Massachusetts, United States

Background: Hypofunction of the N-methyl-d-aspartate receptor (NMDAR) is a leading hypothesis underlying the etiology of schizophrenia. NMDARs are ligand-gated ion channels that assemble as heterotetramers comprising two GluN1 and GluN2 (A–D) subunits. These receptors are distributed widely throughout the brain and are involved in synaptic plasticity. Common and rare variants in the GRIN2A gene, which encode the GluN2A NMDAR subtype, have been associated with an increased risk of schizophrenia. Therefore, modulation of the NMDA receptor may provide an opportunity to enhance physiological NMDA receptor function, restore dysregulated signaling, and improve symptoms in patients with schizophrenia. To evaluate the potential of NMDAR positive allosteric modulators (PAMs), we investigated two distinct approaches with novel compounds that act non-selectively across NMDAR subunits GluN2A-D (NMRA-1) or selectively at the GluN2A receptor subunit (NMRA-2). The objective of the present studies was to evaluate the pharmacological, physiological, and behavioral properties of two distinct NMDAR PAM profiles.

Methods: The in vitro properties of NMDAR PAMs were determined in cell lines stably expressing human GluN1/GluN2 heterodimers. Potency at human GluN2A-D subunits was assessed separately in stable HEK293 cells using a Ca2+ flux FLIPR assay, reported here as EC50 and EC2X at each subunit. Site-directed mutagenesis studies to determine compound binding sites were conducted in HEK293 cells using sub-saturating glutamate (20 μM) concentrations. Electrophysiological characterization of NMRA compounds to determine the PAM concentration response curve and effects on receptor kinetics in human GluN1/GluN2A stable HEK293 cells using the whole-cell patch clamp technique at physiological temperatures (34 °C) with a CsCl-based internal solution, a Mg2 + -free bath saline.

The in vivo effects of NMRA-1 (1-100 mg/kg, po) and NMRA-2 (1-10 mg/kg, po) on EEG band powers were assessed utilizing pharmaco-electroencephalography (EEG) in freely moving male, Sprague Dawley rats (n = 7-9/group). Effects of NMRA-1 (1-10 mg/kg, po) and GNE-5729 (GluN2A selective PAM, 1-10 mg/kg, po) on recognition memory were determined following acute administration in the novel object recognition assay in male, Lister Hooded rats that had undergone sub-chronic phencyclidine (PCP, 2 mg/kg, intraperitoneally x 7-days) induced deficit model with a 7-day washout period. Tolerability was assessed in the spontaneous locomotor assay, with total distance traveled determined in a novel open field for 120 minutes following administration of NMRA-1 (1-300 mg/kg, po) and NMRA-2 (0.1-20 mg/kg, po), in C57/Bl6 mice (n = 10/group) and Sprague Dawley rats (n = 6-10/group). Data were analyzed using one-way or two-way ANOVA with repeated measures when appropriate followed by post hoc analyses. Statistical significance was determined as p < 0.05.

Results: NMRA-1 demonstrated approximately equipotent properties across NMDA subunits (GluN2A/B/C/D EC50 = 86/101/138/66 nM), while NMRA-2 demonstrated selective GluN2A receptor PAM properties (GluN2A/B/C/D EC50 = 78/8242/3852/7343 nM). Site-directed mutagenesis studies confirmed NMRA-1 (1microM) and NMRA-2 (1microM) bind to distinct sites on the NMDAR as measured by potentiation of glutamate-activated currents. Whole-cell patch clamp recordings showed unique electrophysiological properties across the two compounds with the nonselective NMDAR PAM significantly blocking macroscopic desensitization and slowing decay rates by ~3-fold more than the GluN2A selective compound, while having no effect on the peak activation of 1mM glutamate current (p < 0.05). When tested in rat pharmaco-EEG studies, NMRA-1 (minimally active dose =10 mg/kg, po; p < 0.05) reduced cortical gamma power and had no effect on locomotor activity at doses up to 300 mg/kg, po, whereas NMRA-2 showed smaller margins for tolerability between reduced gamma power (minimally active dose = 3 mg/kg, po; p < 0.05) and locomotor suppression (20 mg/kg, po; p < 0.05). Additionally, in a sub-chronic PCP-induced recognition memory deficit paradigm in rat, NMRA-1 (minimally effective dose = 10 mg/kg, po; p < 0.05) significantly blocked a PCP-induced deficit whereas the GluN2A selective GNE-5729 had no effect in this model.

Conclusions: In these studies, we characterized distinct pharmacological approaches to NMDAR potentiation at the level of subunit specificity and binding site. Our results support that the nonselective NMDAR PAM interacts with and potentiates human NMDARs significantly differently to the GluN2A selective PAM and exhibits a more favorable in vivo profile supportive of additional development.

Keywords: NMDA modulators, Schizophrenia (SCZ), Schizophrenia novel treatment, Electroencephalography (EEG)

Disclosure: Nothing to disclose.

P623. The Novel Anti-Epilepsy Therapeutic BHV-7000 Demonstrates Anti-Depressant and Impulse Control Properties in Preclinical Operant Models

Steven Dworetzky*, Cam MacMillan, Leo Silenieks, Guy Higgins

Biohaven, Pittsburgh, Pennsylvania, United States

Background: BHV-7000 is a selective Kv7.2/7.3 potassium channel activator, potent and well-tolerated in antiseizure models, that is in clinical development for multiple CNS indications. Comorbidities associated with epilepsy include depression, anxiety, and impulse control deficiencies. Additionally, compelling evidence exists for targeting Kv7 activation in bipolar disorder and MDD. Complex clinical conditions such as depression, ADHD, and bipolar can be fractionated into discrete symptom clusters such as low motivation, inattention, and low impulse control, which may be more amenable to translational research and treatment. Accordingly, BHV-7000 was investigated for effects on motivation in the progressive ratio task, and on attention and impulse control in the 5-choice task. BHV-7000 was evaluated in these operant experiments at doses that corresponded to efficacious exposures in seizure-based models.

Methods: Male Long Evans rats were trained to either a progressive schedule (PR) of food reinforcement (N = 40) – a test of motivation, or the serial reaction time task (5-CSRTT) – a test of attention and response control (N = 32). BHV-7000 was tested at oral doses of 0.3-10 mg/kg, which covered a dose range effective against MES-induced seizures in rats. All testing was conducted using a repeated measures design once test subjects had achieved stable baseline performance. In addition to studying treatment effects in all animals, a subgroup analysis was also conducted on the low performance tertile subgroups.

Results: In the PR task, BHV-7000 (1-10 mg/kg) produced a non-significant trend to increase lever press and break point. However restricting analysis to the low lever press/break point subgroup (N = 14) revealed a significant main effect of treatment on lever press (F4,52 = 3.2; P = 0.01), with this measure significantly increased at the 3 and 6 mg/kg doses relative to vehicle (Veh: 161 ± 13; 3 mg/kg: 270 ± 25; 6 mg/kg: 293 ± 53). In rats trained to the 5-choice task, under standard test conditions of 5s inter-trial interval (ITI), 0.75s visual stimulus duration (SD), 100 trials, BHV-7000 (1-10 mg/kg) did not detrimentally affect attentional accuracy as measured by % correct (F3,90 = 0.5, NS) or total trials. However, a main effect of treatment was noted on premature responses (PREM; F3,93 = 7.0; P < 0.01), with BHV-7000 reducing this measure relative to vehicle (Veh: 23 ± 4; 3 mg/kg: 12 ± 4; P < 0.01). To investigate further, the ITI was lengthened to 10s to increase the baseline level of PREM responses. Under this test condition, BHV-7000 (1-3 mg/kg) again significantly reduced this measure relative to vehicle pretreatment (Veh: 61 ± 6; 1 mg/kg: 36 ± 14; P < 0.05), without detrimentally affecting trial number.

Conclusions: BHV-7000 shows evidence to increase a measure of motivation, in low performing rats, similar to ketamine responses. BHV-7000 also reduced premature responses in high-impulsive rats without affecting attentional accuracy or trial number. These effects were detected at doses that overlap with efficacious doses in seizure-based models and were well tolerated in that no detrimental effects on task performance were noted, i.e., similar response speed, trials completed, and choice accuracy. These results support the potential for BHV-7000 to treat certain comorbidities associated with epilepsy such as amotivation, a feature of depression, and deficient impulse control, a feature of ADHD and bipolar.

Keywords: Operant behavior, Ion channel activator, Bipolar, MDD, BHV-7000

Disclosure: Biohaven: Employee (Self).

P624. Investigating Signaling Mechanisms of PDGFRb-Mediated Opioid Analgesic Tolerance

Matthew Fanelli, Jorge Pineda, Anastasia Yokum, Ryan Logan, TuKiet Lam, Stephanie Puig*

UMass Chan Medical School, Worcester, Massachusetts, United States

Background: Opioids are the most effective treatments for severe pain. However, the rewarding effects of these drugs can lead to severe addiction and lethal overdose, which hinders their therapeutic efficacy. The risk for these side-effects arises as opioid-mediated analgesia decreases over time, due to tolerance, and requiring dose-escalation. Selective inhibition of tolerance without affecting analgesia has been a major challenge because both, positive opioid analgesia and side effects, are convergently mediated via the mu-opioid receptor (MOPr). However, recent data showed that opioid-activation of MOPr also transactivates the platelet-derived growth factor receptor beta (PDGFRb). Importantly, inhibition of PDGFRb using Imatinib completely prevents analgesic tolerance. While this data strongly suggests PDGFRb signaling downstream of MOPr mediates tolerance, the precise underlying mechanisms have yet to be elucidated.

Methods: In this study, we used mass-spectrometry based proteomics to explore underlying pathways of PDGFRb-mediated tolerance in lumbar dorsal root ganglia (DRG) and spinal cord (SC), critical structures involved in opioid tolerance. Male and female rats received morphine intrathecally for 5 days to induce tolerance, or with morphine and imatinib to prevent tolerance. Vehicle and imatinib alone were used as controls. A tail-flick assay was used daily to measure analgesia and follow the development tolerance. DRG and SC protein lysates were processed for Mass-Spectrometry to analyze differentially expressed (DE, from flow-through fractions) and differentially phosphorylated (DP, from phosphor-enriched fractions) proteins and pathways between tolerant and non-tolerant groups. Differential expression and pathways were analyzed using Limma Voom and GSEA. We also ran pairwise comparisons between treatment groups using the Rank Rank Hypergeometric Overlap (RRHO) algorithm.

Results: Limma Voom analysis showed that most DE and DP proteins between morphine (tolerant) and morphine +imatinib (non-tolerant) groups were observed in DRG samples rather than SC. Strikingly, RRHO highlighted that DE and DP proteins upregulated by morphine were conversely downregulated in the morphine + imatinib group. Similarly, DE and DP proteins downregulated by morphine were subsequently upregulated by morphine + imatinib. These results show that imatinib reverses morphine-mediated effects and highlight a critical role for PDGFRb signaling in tolerance.

Next, GSEA analysis showed that pathways downregulated by morphine tolerance were re-upregulated by morphine +Imatinib. In DRG, these involved synaptic function and signaling, such as presynaptic endocytosis, (Ctbp1, Ap2m1), vesicular exocytosis (Syt1, Syp, and Vsnl1), and small molecule catabolic processes (Ahcy, Ech1, Ldha, Gapdh).

Conclusions: Identified protein candidates and signaling pathways highlight that PDGFRb signaling downstream of opioid-activated MOPr involve synaptic and molecular trafficking mechanisms. Identified proteins will be the subject of further behavioral and biochemical validation for their role in tolerance.

Keywords: Opioid signaling, Opioid side-effects, Proteomics

Disclosure: Nothing to disclose.

P625. Novel Brain Penetrant Ion Channel Inhibitors to Treat Long-Lasting Social Stress Susceptibility

Carole Morel*, Emily Teichman, Sarah Montgomery, Lyonna Parise, Lin Hsiao-yun, Nikolaos Tzavaras, Allyson Friedman, Jianping Hu, Husnu Umit Kaniskan, Jian Jin, Scott Russo, Ming-Hu Han, Jun Wang

Icahn School of Medicine At Mount Sinai, New York, New York, United States

Background: Background: Establishing rapid and long-lasting therapeutic effects is the main challenge of drug discovery efforts for neuropsychiatric symptoms. Anhedonia, apathy, and anxiety are widespread symptoms in human neuropsychiatric profiles, often associated with complex symptomatology, greater treatment resistance, chronicity, and profound impact on aging patients. Clinical and preclinical studies implicate the ventral tegmental area (VTA) dopamine circuits in the emergence of anhedonia, apathy, and depressive phenotypes. However, how specific dopamine circuits may be selectively targeted to rescue motivational, cognitive, and affective deficits is still unknown. Here, we aim to design novel compounds to inhibit brain hyperpolarization-activated cyclic nucleotide-gated ion channels (HCN) and alleviate long-lasting stress-induced social, reward, and cognitive deficits - hallmarks of major depressive disorders.

Methods: Methods: We use a repeated social defeat stress paradigm, which induces long-lasting individual and complex behavioral stress outcomes, to capture motivational, cognitive, and affective deficits in mice. Following social trauma, we perform a longitudinal study of reward sensitivity/processing and approach/avoidance behaviors in female and male mice. Using circuit-probing techniques, in vitro and in vivo electrophysiological, multi-circuit fiber photometry, and operant behavioral methods, we then investigate the dopamine activity circuit adaptations associated with short- and long-term stress-induced behavioral outcomes. We also leverage medicinal chemistry approaches to design novel brain penetrant ion channel inhibitors to treat long-lasting social stress susceptibility.

Results: Results: We first observed that social trauma induces long-lasting social, reward, and cognitive deficits in male and female mice – an impact more profound in aging mice (Kruskal-Wallis, p < 0.01, n = 9-12 mice per group). Our circuit-probing approaches confirm that VTA dopamine projections to the striatum and the amygdala emerge from two distinct neuronal populations. In line with previous studies, we show that social and reward processing deficits are associated with VTA-striatum dopamine neuron increased firing rate activity and bursting activity in both male and female stressed mice (Kruskal-Wallis, p < 0.01, n = 6-11 mice per group). We then highlight HCN channels as critical regulators of dopamine neuronal functions (Kruskal-Wallis, H9/60 = 42.94, p < 0.0001; p = 0.0049; p = 0.0012; p = 0.0002; p < 0.0001, N = 4-9 neurons, n = 3-6 mice). Our multilevel electrophysiological approach shows that the new compounds rapidly reduce VTA dopamine neuron hyperactivity induced by stress (Kruskal-Wallis, H3/59 = 7.752, p = 0.0207; p = 0.0494; p = 0.0364, N = 18-15 neurons, n = 3-7 mice). We further established that the top two new compounds have improved brain bioavailability compared to the parent compound, Cilobradine. Lastly, we establish that a single systemic injection of our novel ion channel inhibitor holds a prolonged therapeutic efficacy on stress-induced social (paired t-test, p = 0.0238, n = 11-15), reward, and cognitive impairments (one-way ANOVA: F2,39 = 4.539, p = 0.02, n = 11-15).

Conclusions: Conclusion: Stressful and traumatic social experiences in humans are critical environmental factors triggering neuropsychiatric profiles across the lifespan. Beyond social domains, social stress can severely impair biological rhythms, motivation, and cognitive and executive functions. Here, our results show that social trauma 1- alters social, reward, and cognitive deficits in male and female mice – an effect more profound in aging mice; and 2- disrupts the coordinated dopamine activity in the striatum and the amygdala. We also identify novel ion channel inhibitors that rescue dopamine activity and thereby reduce anhedonia and social avoidance behaviors. Our results provide a roadmap of how to investigate and exploit the influence of ion channels on complex brain disorders, as well as an HCN inhibitor compound series for future studies.

Keywords: Depression, Dopamine circuit, Social Trauma, Ion channel, Novel treatment

Disclosure: Nothing to disclose.

P626. Comparative Effectiveness of Long-Acting Injectable Vs. Oral Antipsychotic Medication After a First Episode of Psychosis: A Target Trial Emulation in the FEP-Causal Collaboration

Alejandro Szmulewicz*, Gonzalo Martinez-Ales, Dost Ongur, Celso Arango, Jai Shah, Covadonga M.Diaz-Caneja, Vinod Srihari, Lakshmi Yatham, Deepak Sarpal, Ann Shinn

Harvard School of Public Health, Boston, Massachusetts, United States

Background: Evidence about a potentially beneficial effect of long-acting antipsychotic injectable (LAI) therapy initiation compared with oral continuation in first episode psychosis (FEP) is conflicting. While some observational studies and randomized trials showed a lower risk of relapse and hospitalization for LAI initiation, other trials and network meta-analyses reported no differences. The largest individually-randomized trial among FEP patients, EULAST, compared initiation of LAI paliperidone, risperidone, or aripiprazole vs. continuation of oral therapy. The trial found similar risks of therapy discontinuation and hospitalization in both groups, but was not large enough to obtain precise estimates in subgroups, had a follow-up of 18 months, and only considered hospitalizations before non-adherence to the assigned treatment.

Methods: Here we compare initiation of LAI therapy vs. continuation of oral therapy using observational data from a consortium of cohorts of FEP patients, the FEP-CAUSAL Collaboration, including monthly information on over 2,200 individuals with FEP. Clinics included in the present report are: OnTrack at McLean Hospital (Massachusetts), STEP at Connecticut Mental Health Center/Yale University-Yale (Connecticut), RAISE-ETP (US), NYU Langone Health (New York), Pittsburgh-STEP at the University of Pittsburgh/UPMC Western Psychiatric Hospital, University of British Columbia (British Columbia, Canada), 7 hospitals participating in AGES-CM (Madrid, Spain), and PEPP-Montréal at McGill University (Montreal, Canada).

One way to ensure that observational analyses preserve the features of randomized trials is to design them to explicitly emulate the target trial. We first emulate, as closely as possible, the protocol of the EULAST trial (i.e., “benchmarking analysis”). For this, we applied the same eligibility criteria (age older than 18, diagnosis of psychotic disorder, between 6 months and 7 years since FEP clinical onset; no use of LAI in the past month; and no intolerance to study drugs), the same treatment strategies (LAI vs. oral aripiprazole, risperidone, or paliperidone), outcomes (psychiatric hospitalizations), and follow-up period (18-months) in our observational data. We used the results of this target trial emulation as a benchmark.

Next, we use the observational data to emulate a randomized trial that extends the findings from EULAST (i.e., “extension analysis”) to study the 3-year risk of psychotic relapse, overall and in subgroups defined by number of prior relapses, documented non-adherence to antipsychotics, and a diagnosis of substance use disorder to identify subgroups of FEP individuals for whom the LAI treatment initiation may be most beneficial.

Results: Of 2,228 individuals with FEP, 1,067 were eligible for the benchmarking analyses. Both our target trial emulation and EULAST showed little effect of LAI therapy initiation on the 18-month hospitalization risk. Our estimated 18-month hospitalization risks in LAI therapy initiation and oral therapy continuation were 18.5% (95% CI: 12.2, 23.2) and 19.3% (95% CI: 16.3, 22.6), respectively (as compared to 22.0 and 19.2, respectively in EULAST).

In the extended analysis, with 1,193 eligible individuals, the 3-year risk difference of psychotic relapse comparing LAI therapy initiation with oral continuation was -7.0% (95% CI:-12.1, -0.7). The 3-year risk was even lower for LAI therapy initiation in individuals with a prior relapse or prior non-adherence.

After adjusting by non-adherence in per-protocol analyses, the risk of psychotic relapse was identical in LAI therapy initiation and oral continuation suggesting the protective effect of LAI therapy is mediated by improved adherence.

Conclusions: We estimated that, compared with oral therapy continuation, LAI therapy initiation reduced psychotic relapses over 3 years. LAI therapy initiation may be beneficial in subgroups at particularly higher risk of relapse, such as those with prior relapses or documented non-adherence. These findings demonstrate the value of emulating a target trial in providing additional insights for decision-making in complex clinical situations.

Benchmarking against existing trial results can be used to increase confidence in the results. In benchmarking analyses, the investigator first aims to replicate findings from a previously published randomized trial before extending the observational analysis to answer a broader question not originally considered by trial investigators (e.g., due to sample size limitations or limited follow-up periods). If findings from the randomized trial can be accurately replicated in the observational data, confidence is gained that the data sources and analyses were adequate to adjust for confounding.

Keywords: first episode psychosis, causal reasoning, Long-acting injectable antipsychotics, Pharmacoepidemiology

Disclosure: Nothing to disclose.

P627. Discovery of Biased Agonists for Orphan Receptor GPR52 to Oppose Dopamine D2 Receptor Signaling

Ryan Murphy, Sagir Ali, Nolan M. Dvorak, Fernanda Laezza, Jia Zhou, John Allen*

University of Texas Medical Branch, Galveston, Texas, United States

Background: GPR52 is an orphan G protein-coupled receptor (GPCR) and recently identified GWAS schizophrenia risk gene. GPR52 is a Gs/olf class receptor that activates cAMP signaling and is expressed primarily in the ventral striatum of the human brain in dopamine D2 pathway medium spiny neurons (D2 MSNs). The unique expression profile of GPR52 has distinguished this orphan receptor as a promising drug target for psychiatric disorders including schizophrenia and substance use disorders. Here, we describe the creation and evaluation of a new series of novel aniline-carboxamide GPR52 agonists with G protein biased activity and characterize GPR52 agonist opposition of D2 receptor neuromodulation in the nucleus accumbens (NAc).

Methods: Human GPR52 receptor was expressed in HEK293 cells and balanced or biased signaling and pharmacology of ligands was assessed using the Glosensor cAMP assay or Beta-arrestin TANGO assay. Medicinal chemistry synthesis and pharmacology were used in a small molecule discovery campaign to create novel GPR52 balanced or G protein biased activators. To evaluate potential desensitization of GPR52, cells were treated for 2h with vehicle, balanced or G protein biased agonists and receptor desensitization determined using the Glosensor cAMP assay. Molecular docking of novel ligands into a solved GPR52 crystal structure (PBD:6LI0) was done to explore receptor-ligand binding modes. To determine if GPR52 agonist activation modulates NAc MSN excitability and D2R signaling responses, a whole cell patch clamp study was performed measuring evoked action potentials (APs) in D1 or D2 MSNs using mouse NAc brain slices (n = 6-12 cells analyzed per treatment group).

Results: Medicinal chemistry and pharmacology were used to modify an indoline carboxamide ligand and determine compound features crucial for GPR52 activation. Surprisingly, when the nitrogen containing ring of the indoline system was broken into more flexible variants, this greatly increased compound potency (EC50: ~40 nM) and efficacy. Unexpectedly conversion of the indoline to analine derivatives greatly reduced GPR52 activation of the Beta-arrestin pathway, resulting in the creation of several G protein biased activators. Molecular docking of the balanced agonist PW0787 or highly G protein biased agonist PW0866 into the GPR52 crystal structure suggested compound binding with extracellular loop 2 (ECL2) and an allosteric site of action. Notably the highly G protein-biased agonists showed significantly less desensitization than parent compound PW0667/4a, indicating reduced agonist activation of the GPR52 Beta-arrestin pathway reduces receptor desensitization. Treatment of striatal slices with balanced agonist PW0787 increased the frequency and number of evoked action potentials in D2, but not D1, medium spiny neurons of the NAc. In addition, PW0787 reversed the inhibitory effect of the D2 agonist quinpirole on evoked APs, indicating GPR52 activation functionally opposes D2R signaling and D2R neuromodulation in NAc MSN circuitry.

Conclusions: These findings indicate orphan GPR52, via activating Gs/Golf cAMP signaling, is an excitatory GPCR selectively expressed in D2 MSNs. GPR52 functionally opposes D2R signaling via the cAMP pathway and reverses D2 MSN inhibitory neuromodulation by the D2R in the NAc. Our drug discovery effort has resulted in novel GPR52 activators with optimized potency and efficacy, including both balanced and functionally selective G protein biased activators. The novel G protein biased GPR52 activator PW0866 has greatly reduced activity for receptor desensitization. These balanced and biased GPR52 agonists provide important pharmacological probes to determine the impact of biased agonism on GPR52 function. In addition, these compounds provide an innovative conceptual framework for evaluating reduced β-arrestin-mediated desensitization of GPR52, which may provide sustained agonist efficacy for therapeutic applications. Taken together, the findings further support that GPR52 is a druggable target with therapeutic potential for treating neuropsychiatric disorders involving dysregulated NAc signaling, including psychosis and substance use disorders.

Keywords: orphan receptors, substance use disorders, Nucleus Accumbens, Drug Discovery - new approaches, D2 medium spiny neuron

Disclosure: Maplight Therapeutics: Contracted Research (Self).

P628. Adjunctive Administration of an NMDAR Antagonist and 5-HT4R Agonist Exerts Sex- and Region-Specific Effects on C-Fos and PV Networks

Briana Chen*, Michelle Jin, Indira Mendez-David, Denis David, Alain Gardier, Christine Ann Denny

Columbia University, New York, New York, United States

Background: N-methyl-D aspartate receptor (NMDAR) antagonists and serotonin receptor type IV (5-HT4R) agonists exhibit rapid-acting antidepressant effects. We have previously shown that simultaneously targeting both receptors via a combined administration of (R,S)-ketamine (K), an NMDAR antagonist, and prucalopride (P), a 5-HT4R agonist, can exert extended protection against stress-induced anxiety-like behaviors in comparison to administration of either drug alone. Here, we investigated whether combined administration of (R,S)-ketamine and prucalopride exhibits distinct effects on c-fos, a marker of neural activity, and parvalbumin (PV), a marker of inhibitory interneurons, in the medial prefrontal cortex (mPFC) and hippocampus (HPC).

Methods: A single injection of saline, (R,S)-ketamine, prucalopride, or a combined dose of (R,S)-ketamine + prucalopride (K + P) was administered one week prior to contextual fear conditioning (CFC) stress in male and female 129S6/SvEv mice. Drug efficacy was assessed using CFC re-exposure, the forced swim test (FST), and novelty-suppressed feeding (NSF). c-fos and PV expression in the mPFC and HPC was assayed using immunohistochemistry and hand-counts. All experiments were approved by the Institutional Animal Care and Use Committee (IACUC) at the New York Psychiatric Institute (NYSPI). For behavioral analysis and cell counts, data were analyzed using a one- or two-way analysis of variance (ANOVA), using repeated measures and post-hoc Tukey tests where appropriate. c-fos and PV networks correlations and analysis were calculated with Pearson correlations between regions using the Hmisc, igraph, and tidygraph packages in R, with thresholding set to p < 0.05. Networks and summary statistics were visualized using ggraph and Prism 10.

Results: In comparison to saline controls, K, P, and K + P attenuated learned fear in male mice (n = 6-10 male mice per group; FDrug(11,86) = 3.069, p = 0.0016; pSal vs. K = 0.02; pSal vs. P = 0.03; pSal vs. K + P = 0.002) and reduced behavioral despair in both sexes (n = 6-10 male mice per group, FDrug(11,81) = 6.336, p < 0.0001; pSal vs. K = 0.04; pSal vs. P = 0.02; pSal vs. K + P < 0.0001; n = 6-12 female mice per group, FDrug(7,47) = 6.138, p < 0.0001; pSal vs. K = 0.0008; pSal vs. P = 0.01; pSal vs. K + P = 0.004). K + P, but not either drug alone, suppressed hyponeophagia in both sexes (pmale < 0.0001; pfemale = 0.0002). In male mice, K + P, but not K or P alone, enhanced c-fos network connectivity within the mPFC and dorsal HPC while also reducing correlated c-fos expression in the ventral HPC. In female mice, K + P, but not either drug alone, enhanced correlated c-fos expression between the mPFC and dorsal HPC while simultaneously enriching c-fos network connectivity in the ventral HPC. These findings suggest that K + P exhibits distinct sex-specific effects on neural activity in brain regions critical for learning, memory, and affective behaviors. In male mice, K + P significantly enhanced PV network correlations, particularly in ventral HPC, an effect that was not observed in female mice. K + P, in comparison to all other drug groups, significantly increased PV and c-fos overlap in the mPFC and ventral CA3 subregion of the HPC in male, but not female mice (n = 7-9 male mice per group; mPFC FDrug(3,74) = 21.04, p < 0.0001; vHPC FDrug(3,84) = 4.41, p = 0.006; pvCA3 Sal vs. K + P = 0.0013, pvCA3 K vs. K + P = 0.0007, pvCA3 P vs. K + P = 0.01; n = 6-7 female mice per group; mPFC FDrug(3,63) = 0.2948, p = 0.82; vHPC FDrug(3,63) = 2.178, p = 0.09).

Conclusions: Overall, our results indicate that combined K + P exerts extended protection against anxiety-like behavior in both male and female mice, a result that is not observed in either drug alone. c-fos and PV network analysis indicate that adjunctive K + P administration results in a refinement in neural signaling, particularly in ventral HPC, in a sex-specific manner that may contribute to enhancing stress resilience in both sexes. Together, our findings demonstrate the potential of leveraging combinatorial pharmacological treatment to advance targeted therapies for stress-induced psychiatric disorders.

Keywords: NMDA Antagonists, 5HT4R agonist, network analysis, Sex-specific effects

Disclosure: Intellectual Property: Patent (Self).

P629. Didesmethyl-Cariprazine, the Major Active Human Metabolite of Cariprazine with Greater Dopamine D3 Vs. D2 Receptor Selectivity, Increases Dopamine Release in the Rat Medial Prefrontal Cortex

Jan Kehr, Bence Farkas, Viktor Román, Atilla Horváth, Balázs Lendvai, Nika Adham*

Abbvie, Englewood, New Jersey, United States

Background: Cariprazine (CAR) is a dopamine D3-preferring D3/D2 and serotonin 5-HT1A receptor partial agonist approved by the US Food and Drug Administration to treat adults with schizophrenia, manic, mixed, and depressive episodes associated with bipolar I disorder as well as adjunctive treatment for major depressive disorder. CAR is metabolized to desmethyl-CAR (DCAR) and ultimately to didesmethyl-CAR (DDCAR) with DDCAR being the prominent moiety ( ~ 70% of total) at steady state in humans (Periclou et al., 2021). Although in general DDCAR displays comparable in vitro binding receptor profile and has a similar antipsychotic-like activity as CAR, this metabolite has greater D3 receptor (D3R) affinity and selectivity compared to the parent (Kiss et al., 2019). The objective of this study was to evaluate, whether there are any differences in the neurochemical effects of CAR in comparison with DDCAR, which is more D3 receptor-preferring, to provide potential mechanism(s) that can account for the unique clinical profile of CAR. Neurochemical effects of these agents were evaluated in rats using microdialysis probing in the medial prefrontal cortex (mPFC), a brain area of importance with dysregulated function in various psychiatric and cognitive disorders. The negative and/or cognitive symptoms of schizophrenia may be due to reduced dopaminergic and/or cholinergic tone in cortical brain areas (McCutcheon et al., 2023; 2020). Moreover, alteration in the function of dopamine D3 receptors may play a role in this cortical hypofunctionality and underlie the deficits in social behaviors and cognitive functions in various mood disorders (Gross et al., 2012; Leggio et al., 2016).

Methods: Male Sprague-Dawley rats (weighing 300–320 g, total number of 49 rats) were used in the study. Seven days before the microdialysis experiment, the rats were pre-treated with buprenorphine and carprofen, anesthetized with isoflurane and the guide cannulas were stereotaxically implanted into the mPFC. On the day of experiment, the separate groups of awake rats were treated with CAR and DDCAR and the samples were collected in 20-min intervals for the following 4 h. Concentrations of neurotransmitters (dopamine (DA), norepinephrine (NE), serotonin (5-HT), acetylcholine (ACh), histamine (HA), glutamate (Glu), GABA, glycine (Gly)) as well as, the levels of D-serine (D-Ser) and the monoamine metabolites (MHPG, DOPAC, HVA and 5-HIAA) in the microdialysis samples were measured UHPLC-MS/MS following their derivatization with benzoyl chloride (except of ACh). Concentrations of D-Ser were measured by UHPLC-MS/MS using a chiral separation column. Concentrations of CAR and DDCAR in the remaining volumes of the microdialysates were measured by UHPLC-MS/MS.

Results: The major findings of the study are: 1) DDCAR but not CAR significantly increased the AUC(0-240min) levels of DA and its metabolites DOPAC and HVA in the mPFC. 2) CAR caused significant decreases in ACh, HA, and monoamine metabolites MHPG and 5-HIAA levels in the mPFC. The levels of other analytes were only marginally decreased or not affected by any treatment. 3) The extracellular levels of CAR and DCAR gradually increased over a period of 140 min in the mPFC and remained increased over the entire 240-min sampling period. In the mPFC, the AUC(0-240min) of DDCAR was 41% of that obtained for CAR.

Conclusions: DDCAR produced a significant increase in the level of DA and DA metabolites in the mPFC, whereas CAR did not show any significant effects on these parameters with even a trend towards a decrease at lower dose tested. These data suggest that DDCAR’s greater D3 receptor affinity and selectivity over D2 receptor compared to CAR may contribute to the increase in DA neurotransmission in the mPFC, a brain region markedly associated with negative symptoms and cognitive deficits in schizophrenia. These results confirm the idea that the increasing D3 vs D2 selectivity of compounds produces more benefit for ameliorating these negative and cognitive symptoms. Moreover, DDCAR treatment did not have any impact on ACh and HA levels, whereas CAR resulted in a significant decrease in these neurotransmitters, which may be due to the latter’s greater D2 receptor activity compared to DDCAR. Collectively, the results of this study demonstrate a differential neurochemical profile of DDCAR in comparison to CAR, indicating that the predominance of this metabolite following stable dosing of CAR may contribute to the unique clinical profile of CAR therapy in various psychiatric symptoms.

Keywords: cariprazine, dopamine D3 receptors, Microdialysis, Medial Prefrontal Cortex (mPFC), rats

Disclosure: AbbVie: Employee (Self).

P630. Epigenome-Wide Association Study of Antenatal Corticosteroid Exposure in Preterm Infants: Potential Impacts on Neonatal Neurodevelopment

Bingbing Wu, Zhaohui Qin, Yun Wang*

Emory University, Decatur, Georgia, United States

Background: Antenatal corticosteroids (ACS) are synthetic glucocorticoids given to women at risk of preterm birth to improve neonatal survival and reduce morbidity. However, children exposed to ACS may be at increased risk of developing neurobehavioral disorders later in life, with the underlying mechanisms still unclear. A recent Epigenome-Wide Association Study (EWAS) has shown that ACS induces DNA methylation changes in neonatal blood, particularly in genes associated with neurological function. Interestingly, previous research suggests that DNA methylation patterns in saliva may more closely resemble those in the brain than those in blood. This prompts the question: Can we detect the impact of ACS on neonatal saliva DNA methylation by comparing exposed and non-exposed preterm infants?

Methods: We conducted an EWAS to identify differentially methylated CpG sites associated with antenatal corticosteroid (ACS) exposure. Saliva samples from a birth cohort within the Developmental Human Connectome Project were analyzed using the HumanMethylationEPIC BeadChip. Quality control included filtering out low-quality samples and probes with detection p-values > 0.01, low bead counts, non-cg probes, SNP-related probes, multi-aligned probes, and sex chromosome probes. After filtering, 728,165 probes and 143 samples remained for analysis. The study included 143 preterm infants (gestational age at birth 32 ± 3.4 weeks; 82 males), with 95 exposed to ACS and 48 not exposed. Additional potential confounders considered in the analysis were maternal age, maternal education, maternal BMI, infant sex, delivery method, birth weight, and head circumference.

Results: In total, 778 differentially methylated CpG sites (DMCs) were identified between ACS-exposed and non-exposed groups at a significance threshold of p < 10-6. Among these, 756 were hypermethylated and 22 were hypomethylated in the ACS-exposed group compared to non-exposed group. Key genes identified include LMX1B, ECE1, GLI3, PRDM14, GABBR1, GDNF, SIM2, and HIC1, which are associated with neurodevelopmental processes. The differential methylation patterns suggest that ACS exposure may influence specific molecular pathways that could have long-term implications for brain development.

Conclusions: This study identified significant differentially methylated CpG sites associated with antenatal corticosteroid exposure, with many top annotated genes playing crucial roles in brain and neurological functions. The identification of these epigenetic changes suggests that ACS exposure may have lasting impacts on neurodevelopment through specific molecular pathways.

Moving forward, we will examine longitudinal DNA methylation patterns to assess the persistence of these epigenetic alterations over time. If consistent patterns are observed, it could indicate a stable epigenetic marker of ACS exposure with potential implications for predicting neurodevelopmental outcomes. Furthermore, this research may open avenues for linking these epigenetic changes directly to brain structure and function, providing a deeper understanding of how early life interventions can shape long-term brain health.

Keywords: EWAS, Glucocorticoids, DNA Methylation

Disclosure: Nothing to disclose.

P631. G Protein Signaling-Biased Kappa Opioid Receptor Agonists Produce Anxiolytic-Like Effects in Mice

Laura Bohn*

UF Scripps Biomedical Research, Jupiter, Florida, United States

Background: Kappa opioid receptor (KOR) agonists are clinically used to treat pruritis yet their wide-spread use for treating pain remains limited by their side-effect profile (dysphoria, sedation and diuresis). KOR antagonists are emerging as promising anxiolytics and mood modulators. We have previously described a G protein signaling-biased KOR agonist, triazole 1.1, that is efficacious in suppressing itch; yet it is non-sedating in mice, rats, and non-human primates. However, it is less potent than currently available KOR agonists; therefore we sought to make more potent biased KOR agonists. In the process we found that we could separate the anxiolytic-like properties of KOR agonists from their sedating properties.

Methods: Using an array of signaling assays in human KOR-expressing stable cell lines, we have characterized the pharmacological profile of several new triazole-derived agonists. These assays include: 35S-GTPgammaS binding; radioligand binding; cAMP accumulation assays (Revvity HTRF); beta- arrestin2 recruitment assays (Millipore DiscoveRx); and ERK1/2 phosphorylation assays (all data presented is n > 3 experimental replicates). The compounds were then characterized in mouse behavioral assays which include: chloroquine-induced pruritis; locomotor activity in open field; nesting; elevated plus maze; climbing/exploratory behavior; and diuresis studies. Both male and female adult mice were studied; dose response studies are presented (n = 8-10 in most studies). One-way ordinary ANOVA is the most prevalent test used to compare endpoint measures. Two-way ANOVA is used to repeated measures over time.

Results: We present studies of Triazole 187 performed in comparison to Triazole 1.1 and U50,488H. Triazole 187 is ~5 fold more potent than Triazole 1.1 yet retains signaling bias for promoting GTPgammaS binding over other measures. Triazole 187 and U50,488H are efficacious at 0.3 mg/kg in the pruritis assay while Triazole 1.1 requires higher doses. U50,488H induces sedation at 5 mg/kg ip while triazole 1.1 and triazole 187 do not produce sedation at doses up to 30 mg/kg ip. Triazole 187 and U50488H produce an increase in center time spent in open field and promote climbing/exploration (although the curve is U-shaped for U50,488H as high doses are sedating). Triazole 187 increases time spent in the open arms of the elevated plus maze at 5 mg/kg while U50,488H (3 mg/kg) and Triazole 1.1 (15 mg/kg) do not.

Conclusions: Like KOR antagonists, G protein signaling-biased KOR agonists may be promising for the development of anxiolytic therapeutics. Considering the emotional component of pain perception; this may be an added benefit of a KOR agonist for treating pain and pruritus.

Keywords: kappa opioid receptors, Kappa Opioid Receptor Antagonist, kappa agonist, New anxiolytics, Biased signaling

Disclosure: Nothing to disclose.

P632. Pharmacological Characterization of a Selective Brain Penetrant Agonist of Mouse D5 Receptor

Daming Chen, Gangling Liao, Frank Piscotta, Glen Ernst, Ingrid Buchler, Gregory Carr, James Barrow, Evgeny Shlevkov*

Johns Hopkins University School of Medicine, Baltimore, Maryland, United States

Background: The D5 receptor (D5) is part of the D1-type dopamine receptor family, coupled via Gαs/olf and initiating cAMP/PKA signaling. Compared to the D1 receptor (D1), D5 exhibits a ~ 10-fold higher affinity for dopamine and greater constitutive activity. However, the lack of selective D5 ligands has hindered our understanding of the relative function of D5 and D1 in intact animals given that only non-selective compounds have been available.

Methods: Tool compound was made in one step from commercially available reagents via nucleophilic aromatic substitution SNAr. Identity and purity were confirmed via LCMS and 1H NMR. All pharmacology experiments were conducted in transiently transfected HEK293 cells. cAMP assays were performed using cAMP-Gs Dynamic Kit (Revvity) following manufacturer’s instructions. Except when indicated, concentration-response curves (CRC) were constructed with data obtained from three independent transfections. EC50 and Emax values were obtained using Dose-Response - Special, X is log (concentration) equation in GraphPadPrism.

Results: In this study we synthetize and characterize a brain-penetrant, non-catechol D1/D5 agonist with ~100-fold potency shift toward mouse D5 (mD5) over mouse D1 (mD1) (mD5 EC50 20nM, Emax 90%; mD1 EC50 1400nM, Emax 56%). Pharmacological assays demonstrated that the compound behaves as an affinity-biased full agonist of mD5 with no detectable allosterism. Sequence alignment, structure-activity relationship (SAR) and site-directed mutagenesis led to the identification of two divergent residues in the ECL2 domain of mD1 and mD5 site partially responsible for selectivity shift toward mD5. Interestingly, the mapped residues are divergent between D5 and D1 exclusively within the rodent family. Consistently, the presented compound was equipotent toward human D5 and D1 (D5 EC50 1400 nM, Emax 50%; D1 EC50 1000nM, Emax 45%), where it acted as a partial agonist.

Conclusions: Our study presents a selective agonist of mD5. The pharmacokinetic properties of the compound make it a suitable tool for in vivo studies. Our study supports the hypothesis that ECL2 domain of mD5 contributes to the selectivity of the presented non-catechol orthosteric agonist. We propose that further exploration of divergent residues in the vicinity of the orthosteric site of mD5 and mD1 can be a fruitful strategy for the discovery of rationally designed selective tools.

Keywords: D1 dopamine receptors, Pharmacology, drug discovery/development

Disclosure: Biogen: Stock / Equity - Publicly Traded Company (Self). Myrobalan Therapeutics: Stock / Equity - Privately Held Company (Self). Intellia Therapeutics: Employee (Spouse/Partner).

P633. Cannabinoid Type 1 Receptors Control Recognition Memory via Gaba Release in the Prefrontal Cortex of Mice

Tomohiro Tokutake*, Takashi Asano, Jun Yokose, Yusuke Yano, Shin-Ichi Muramatsu, Shin-Ichi Muramatsu, Atsumi Nitta

University of Toyama, Japan, Toyama, Japan

Background: Cannabis is a regulated substance used thorough the world. Negative psychiatric symptoms by cannabis have an important issue. Memory impairments are the best known as the acute effects of cannabis in animal and humans. CB1R was identified as binding site for delta9-tetrahydrocannabinol (THC). CB1R is mainly expressed in the brain and regulates the release of neurotransmitter such as GABA and glutamate. The PFC is important role for memory. CB1R in the PFC might be associated with memory. However, the relationship between CB1R in the PFC and the memory has not yet been clarified. Elucidation of the mechanism is needed to develop for medical treatments of cannabis-induced memory impairment. The knowledge in each brain region of CB1Rs is insufficient has not yet been not enough. Therefore, we focused on CB1Rs in PFC and attempted to explore these functions for the memory. In addition, CB1Rs are highly expressed in the GABAergic neurons in the PFC. The activation of CB1Rs is reported to reduce the release of GABA. Therefore, we examined effects of activating GABAergic neurons for CB1R-related memory impairment.

Methods: We implanted the guide cannula into the PFC of male C57BL/6J mice. Mice were administered arachidonylcyclopropylamide (ACPA), CB1R agonist, into the PFC via the guide cannula. The memory tests involving novel object and object location tests were performed. The stages of memory can be divided acquisition, consolidation and retrieval. We defined administration before the pre-training as acquisition, administration immediately after training as consolidation, and administration before pre-trial as retrieval. ACPA was administered into the PFC at different stages. We measured GABA levels changes in the PFC by intra-PFC administration of ACPA using in vivo microdialysis. We constructed the adeno-associated virus (AAV) vector with glutamate decarboxylase promotor upstream of the hM3D (Gq) gene and injected it into mice PFC. We measured GABA levels in the PFC of mice injected AAV vector by the i.p. administration of deschloroclozapine (DCZ), hM3D agonist, using in vivo microdialysis. In the object location test, we i.p. administered DCZ before intre-PFC administration of ACPA.

Results: In object location tests, intra-PFC administration of ACPA impaired the acquisition of memory. However, the consolidation and retrieval of memory were not impaired. In the novel object tests, intra-PFC administration of ACPA didn’t have an influence on the acquisition, consolidation and retrieval. In in vivo microdialysis, ACPA significantly reduced GABA levels in the PFC. In addition, the administration of DCZ to mice injected AAV vector significantly increased GABA levels in the PFC. In the object location test, the activation of GABAergic neurons in the PFC by hM3D rescued the memory impairment by intra-PFC administration of ACPA.

Conclusions: These results suggested that CB1R in the PFC is involved in the acquisition of memory in the object location memory via controls of GABA release. Our findings should be useful for the functional analysis of CB1R in the PFC and distribute studies for medication tools for cannabis use disorder.

Keywords: cannabinoid receptor type 1, recognition memory, Cortical GABA

Disclosure: Nothing to disclose.

P634. Preclinical Models of SSRI Exposure on Gut Motility Translate to Human Constipation Phenotypes in Early Life

James Woodruff*, Virginie Bouchard, Virginie Gillet, Jialiang Hua, Ying Liu, Jonathan Posner, Jay Gingrich, Larissa Takser, Kara Margolis, Ardesheer Talati*

College of Physicians and Surgeons, Columbia University/New York State Psychiatric Institute, New York, New York, United States

Background: Serotonin (5-HT) plays a crucial role in the development of both the central (CNS) and enteric (ENS) nervous systems during gestation. While most human studies have concentrated on CNS outcomes, preclinical research indicates that gestational changes to serotonin, such as through exposure to selective serotonin reuptake inhibitors (SSRIs), reduces colonic innervation and motility, consistent with a constipation phenotype. However, whether these associations persist in human populations has not been prospectively studied. Here, we leverage a new birth cohort, Gestational SSRI Exposures in the Development of Functional Gastrointestinal Disorders (GETGAS), to explore associations between prenatal SSRI exposure and the development of disorders of gut-brain interaction (DGBIs) in the first two years of life.

Methods: Mother-child dyads were followed longitudinally from early pregnancy until the child reached 24 months. Data on SSRI exposure were gathered through maternal self-reports and validated by pharmacy records; data on mental health disorders were collected via the SCID-5, and monthly depression symptoms via the PHQ-9. Mother-child dyads were classified in one of three groups based on the child’s exposure to maternal medication and depressive symptoms in pregnancy: healthy control, self-reported prenatal maternal depression symptoms without SSRI exposure, and SSRI exposed group. Final classifications were made only after conclusion of pregnancy, using all available clinical information to minimize misclassification. DGBI were assessed longitudinally using gold-standard ROME-IV criteria, and toddler cognitive and socioemotional development were evaluated using the Bayley’s Scales for Infant Development (BSID-IV) and the Brief Infant-Toddler Social and Emotional Assessment (BITSEA).

Results: After adjusting for infant birthweight and length, sex, gestational age, and self-reported maternal depression symptoms across pregnancy (using the average PHQ-9), prenatal SSRI exposure was associated with increased risk for developing DGBI compared to exposure to depressive symptoms alone, or to neither. Findings were most significant for functional constipation, where SSRIs were associated with a more than three-fold increased risk (OR = 3.587, p = 0.008), consistent with the preclinical findings suggestive of a constipation phenotype. In contrast, infant colic was more prevalent among offspring exposed to prenatal depression symptoms but not SSRIs.

Conclusions: Our findings translate previous pre-clinical work showing reduced colonic motility following SSRI exposure to a human population framework, while rigorously controlling for prenatal depression symptoms and other confounders. The findings also underscore the differential impact that depression versus antidepressant exposures can have on earliest gut development, and suggest that tailored prenatal care strategies that consider maternal psychiatric well-being and medication use in pregnancy.

Keywords: Antidepressant, Prenatal SSRI Exposure, functional constipation, disorders of gut-brain interaction

Disclosure: Nothing to disclose.

P635. Differential Efficacy of Selective Full and Partial M4 Machr Agonists Across Preclinical Rodent Models

Steven Leiser*, Polina Stolyar, Julianna Locantore, Sokhom Pin, Georgette Suidan, Srinivas Chakilam, Hanh Nguyen, Philip Iredale

Cerevel Therapeutics, Cambridge, Massachusetts, United States

Background: The therapeutic potential of targeting M4 muscarinic acetylcholine receptors (mAChR) for the treatment of neuropsychiatric indications is rapidly emerging through preclinical animal research as well as in the clinic. Here we propose an in vivo cascade to screen putative candidate M4 activators.

Methods: Amphetamine-induced locomotor activity (aLMA) evaluates the effects of compounds on ameliorating an induced hyperdopaminergic state. Such a hyperdopaminergic state, which has been postulated to manifest during psychosis and is relevant to schizophrenia, bipolar mania, and potentially other neurobehavioral symptoms, provides construct validity to aLMA. In fact, for evaluation of the effects of M4 activators, aLMA is preferred over MK-801 (glutamatergic) modulation of circuits, as we demonstrate efficacy of full and partial M4 agonists in aLMA with clear PK/PD responses yet show these effects are notably reduced in the MK-801-induced LMA. As a translational biomarker, EEG was used as it holds face or content validity as whether in rodent, NHP or human, EEG is representative of what it aims to measure, which is the underlying neuronal network changes in response to drug.

Results: As expected, partial agonists require a greater projected change in cAMP to demonstrate equivalent reduction in aLMA compared to full agonists. Similarly, in rat pharmaco-EEG, clear PK/PD responses were observed on the power spectra following full and partial M4 agonists. Most strikingly were the increased power in Delta (0.5-4 Hz) and suppression of Sigma (12-16 Hz) from frontoparietal EEG during waking states.

Conclusions: We demonstrate acute administration of full and partial M4 agonists yield distinguishing behavioral profiles and varying degrees of efficacy across assays. The behavioral and neurophysiological endpoints measured from the selected preclinical assays are associated with dopamine function, serve as a metric of M4 activation, and provide utility in screening candidate molecules for neuropsychiatric indications. Accordingly, these mechanistically-defined classes of selective M4 mAChR agonists may provide multifaceted therapeutic utility due to the complex pathophysiology of psychotic disorders.

Keywords: EEG, Muscarinic M4, Behavioral Pharmacology

Disclosure: Cerevel Therapeutics: Employee (Self).

P636. 5-HT2C Receptors Contribute to Viloxazine’s Effects on Impulsivity and Attention in Rats on the 5-Choice Serial Reaction Time Task at Therapeutically Relevant Concentrations

Brittney Yegla*, Jennie Garcia-Olivares, Alvin Terry, Chungping Yu, Jonathan Rubin

Supernus Pharmaceuticals Inc, Rockville, Maryland, United States

Background: Viloxazine ER (extended release capsule) is a nonstimulant medication FDA-approved for Attention-Deficit/Hyperactivity Disorder (ADHD). Its effects in ADHD are thought to be due to norepinephrine (NE) transporter (NET) inhibition, which increases NE and dopamine in select brain regions. However, in preclinical experiments, viloxazine also shows pharmacological activity at serotonin receptors, including partial agonism at 5HT2C and antagonism at 5HT2B and 5HT7. In cynomolgus monkeys, viloxazine was shown to bind to 5HT2C receptors, with an EC50 value (4.1 μM) within the unbound plasma concentration range observed in patients with ADHD taking viloxazine ER (0.4-3.6 μM) and thus is likely therapeutically relevant. Therefore, the question remains as to the role of 5HT2C pharmacology in the efficacy of viloxazine ER for ADHD. Employing the 5-choice serial reaction time task (5-CSRTT), we evaluated ADHD-relevant measures of attention and impulsivity in rats treated with viloxazine alone or viloxazine with SB 242084 (selective 5HT2C antagonist) to assess the potential contribution of 5HT2C activity to viloxazine’s mechanism of action.

Methods: Adult, male Wistar rats (N = 48) were food restricted (85% bodyweight) and trained on the 5-CSRTT. In this task, rats attended to 5 ports and nose poked into a port illuminated for 1s to receive a food reward. No response (omission) or incorrect responses resulted in a timeout period before starting a new trial. Attention (%Hit) was calculated as the percent of correct responses over the sum of correct and incorrect responses, including omissions. Nose pokes during the intertrial interval (ITI) (i.e., premature responding) or timeout periods (i.e., timeout responses) served as measures of impulsivity and cognitive inflexibility, respectively. To utilize a preclinical model exhibiting a relative ADHD phenotype of increased impulsivity, rats were ranked on premature responses for high (N = 12) or low (N = 12) trait impulsivity during sessions with a 10s ITI and were used for further testing. Rats were treated with nicotine (0.2 mg/kg, sc; positive control), viloxazine (1, 3, 10, 20, 30 mg/kg, ip), and SB 242084 (0.025, 0.05, 0.1 mg/kg, ip) under standard conditions (stimulus duration: 1s, ITI: 5s) and variable ITI (vITI; 2.5, 5, 10 s ITI) conditions. Additionally, SB 242084 (0.05, 0.1 mg/kg) was paired with viloxazine (10, 20 mg/kg) to generate 4 groups examining the effects of 5HT2C blockade on viloxazine’s behavioral profile. Repeated-measures ANOVA was conducted, followed by post-hoc tests for specific comparisons (Holm-Sidak), with significance set to p < 0.05. A separate cohort of viloxazine-treated rats had blood collected one hour after treatment, to align with the onset of behavioral testing; plasma was isolated and analyzed using LC-MS/MS to quantify concentrations of viloxazine. Animals were cared for in accordance with the Guide for the Care and Use of Laboratory Animals (U.S. National Institutes of Health) and the experimental protocols were approved by the Institutional Animal Care and Use Committee at Augusta University.

Results: As expected, nicotine improved attention but increased impulsivity. SB 242084 (0.05, 0.1mg/kg) exhibited no behavioral effects during the vITI condition and was subsequently paired with viloxazine. Viloxazine (3 mg/kg and higher) dose-dependently decreased impulsivity (premature responses, p < 0.001) in high and low trait impulsivity rats under standard and vITI conditions. SB 242084 (0.05 and 0.1 mg/kg) reversed viloxazine’s effects at 10 mg/kg on premature and timeout responses under standard and vITI conditions, suggesting a potential role for 5HT2C in viloxazine’s pharmacologic effect, as well as a contributing factor to viloxazine’s efficacy on behavioral inhibition. High doses of viloxazine (20, 30 mg/kg) reduced attention (i.e., decreased %Hit and increased omissions) under standard conditions, with low impulsive rats exhibiting greater sensitivity to viloxazine’s attention-impairing effects than high impulsive rats. Under conditions of high attentional load (short ITI trials), 10 and 20 mg/kg of viloxazine impaired %Hit (p < 0.001); however, 5HT2C antagonism abolished this impairing effect. The doses of viloxazine that reduced impulsivity and were sensitive to 5HT2C blockade produced unbound plasma concentrations of viloxazine akin to those observed in patients administered viloxazine ER (human Cavg with Viloxazine ER: 0.4-3.6 μM; rats 1 h post-dosing with viloxazine: 0.13 µM following 3 mg/kg; 0.48 µM following 10 mg/kg; 1.58 µM following 30 mg/kg), suggesting the likely relevance of 5HT2C to viloxazine’s therapeutic effects.

Conclusions: This preclinical model supports the involvement of 5HT2C in viloxazine’s pharmacologic and therapeutic effects in ADHD treatment.

Keywords: ADHD, Serotonin, Serotonin 5-HT2C Receptor, Non-stimulant, impulsivity

Disclosure: Supernus Pharmaceuticals Inc: Employee (Self).

P637. Pharmacological Characterization of Centanafadine – Potential Implications for Efficacy and Safety in ADHD and Comorbid Psychiatric Disorders

David Heal*, Sharon Smith, Jane Gosden, Reem Elbekai

DevelRx Ltd., Nottingham, United Kingdom

Background: Though current drugs for attention-deficit/hyperactivity disorder (ADHD) are clinically effective, their pharmacological actions may produce burdensome treatment-emergent adverse events (TEAEs) including appetite suppression, anxiety, insomnia, and abuse potential.[1,2] Furthermore, individuals with ADHD are at high risk of developing other serious psychiatric disorders (eg, anxiety, affective, conduct, and substance use disorders) which limit pharmacological options available.[3,4] Centanafadine (CTN) is a norepinephrine (NE), dopamine (DA), and serotonin (5 HT) reuptake inhibitor (NDSRI)[5] under investigation for the treatment of ADHD and showed positive efficacy and safety results in Phase 2/3 trials in adults [6,7] and children/adolescents. CTN is pharmacologically unique compared to existing ADHD [1] drugs by virtue of inhibiting 5-HT reuptake with 5-HT playing a key role in emotion, mood, sleep-wake cycle, and appetite.[8] Per in vitro binding assays, CTN has high affinity with rank order of potency NE > DA > 5-HT.[5]

Intracerebral microdialysis can quantify the pharmacodynamic effects of drugs on multiple CNS monoamines simultaneously. We used microdialysis to measure CTN’s effects on extracellular monoamines in the prefrontal cortex (PFC) and ventral striatum (vSTR) – two key brain regions mediating therapeutic actions of ADHD drugs.[6]

Methods: Microdialysis probes (CMA; 2 mm tips) were implanted into PFC and vSTR of male Sprague-Dawley rats (n = 7–8/group)[9] and perfused with artificial cerebrospinal fluid (1.2 µL/min) throughout ~16-h recovery and experiment. After 80 min baseline sampling, CTN (3, 10, 30 mg/kg p.o.) or vehicle was administered, with dialysate samples collected in 0.1 M perchloric acid at 20 min intervals for 4 h. Monoamines were measured by HPLC with electrochemical detection. Locomotor activity and microdialysis were performed simultaneously using the Culex Bambino/Raturn system. Microdialysis data were log transformed and adjusted for differences between treatment groups at baseline. Data were analysed (ANCOVA) with log(baseline) as a covariate (5-HT) and robust regression (M-estimation with Huber weighting) using the default parameter c = 1.345 with treatment as a factor and log(baseline) as a covariate (NE and DA). Locomotor data were sqrt transformed and analysed (ANCOVA) with sqrt(baseline) as a covariate. All treatments were compared to vehicle by Williams’ test with significance p < 0.05.

Results: CTN dose-dependently increased extracellular NE, DA, and 5-HT in PFC, and DA and 5-HT in vSTR (NE was below quantification level). Monoamine increases were gradual in onset, reaching a plateau ~60 min after CTN dosing, and were broadly maintained through 4 h. Peak monoamine increases (% baseline) produced by CTN 30 mg/kg in PFC were: 5-HT: 1638% @ 80 min; NE: 739% @ 120 min; and DA: 302% @ 180 min (all p < 0.001); and in vSTR were: 5-HT, 1590% @ 80 min and DA: 219% @ 160 min (both p < 0.001). Over 0–4 h, average monoamine increases in PFC were: 5-HT: 151% (NS, 3 mg/kg), 284% (p < 0.001, 10 mg/kg), 1184% (p < 0.001, 30 mg/kg); NE: 151% (p < 0.001, 3 mg/kg), 244% (p < 0.001, 10 mg/kg), 604% (p < 0.001, 30 mg/kg); and DA: 103% (NS, 3 mg/kg), 143% (p < 0.001, 10 mg/kg), 281% (p < 0.001, 30 mg/kg); and in vSTR were: 5-HT: 133% (NS, 3 mg/kg), 294% (p < 0.02, 10 mg/kg), 1119% (p < 0.001, 30 mg/kg); and DA: 122% (p < 0.02, 3 mg/kg), 121% (p < 0.02, 10 mg/kg), 186% (p < 0.001, 30 mg/kg). CTN had no effect on locomotor activity vs vehicle.

Conclusions: The findings show CTN functions as a NDSRI in vivo with a rank order for increasing extracellular monoamine levels of 5 HT > NE > DA in PFC, and 5-HT > DA in vSTR. The finding that CTN substantially increased catecholaminergic neurotransmission in PFC is consistent with its clinically proven efficacy in ADHD. CTN’s ability to increase DA ergic drive in vSTR - an important contributor to efficacy [4] - may explain its efficaciousness within a week of treatment initiation in adult trials.[6,7] Although CTN enhanced DA signaling in vSTR, it did not increase the rats’ locomotor activity predicting a low risk for undesirable psychomotor activation in patients. CTN’s pharmacological uniqueness in treating ADHD is its ability to enhance 5-HT signaling in cortical and subcortical regions, which may contribute to its favourable TEAE profile. Data in adults with ADHD showed a low incidence of TEAEs, with the frequency of the most common potentially related TEAEs (decreased appetite, headache, dry mouth, and nausea) increasing slightly with increasing dose. Indirect comparisons predicted reduced insomnia, anxiety, and feeling jittery with CTN vs lisdexamfetamine (LDX), and less appetite suppression, dry mouth, and nausea than LDX, atomoxetine, or viloxazine.[2] Moreover, the substantial enhancing effect of CTN on 5-HT-ergic and NE-ergic neurotransmission in cortical and subcortical areas suggests CTN has the potential to have beneficial effects on anxiety and mood disorders that are frequently comorbid in adults with ADHD.

References:

1. Heal D et al. Curr Top Behav Neurosci. 2022;57:79-126

2. Schein J et al. JMCP. 2024;30:528-40

3. Kessler RC et al. Am J Psychiatry. 2006;163:716-23

4. Anker E et al. Brit Med J. 2018;8:e019700

5. Bymaster F et al. Synapse. 2012;66:522-32

6. Wigal SB et al. Neuropsychiatr Dis Treat. 2020;16:1411-26

7. Adler LA et al. J Clin Psychopharmacol. 2022;42:429-39

8. Kitson SL. Curr Pharm Des. 2007;13(25):2621-37

9. Paxinos G and Watson C. A Stereotaxic Atlas of the Rat Brain. AP, 1986

Keywords: Centanafadine, Intracerebral microdialysis, Serotonin, Catecholamines, ADHD

Disclosure: Otsuka: Consultant (Self). DevelRx: Founder(Self).

P638. In Vivo Effects of Psilocybin and Lisuride on Serotonergic, Dopaminergic, and Prefrontal Cortex Neuronal Activity and Antidepressant-Like Paradigms

Brandon Richardson, Michael Pileggi, Thomas Prud’homme, Tadhg Strand, Derek Wan-Yan-Chan, Sofia Nasini, Marco Leyton, Nahum Sonenberg, Danilo De Gregorio, Jeffrey Sprouse, Francis Bambico, Gabriella Gobbi*

McGill University, Montreal, Canada

Background: Psilocybin is a 5-HT2A receptor agonist psychedelic drug that has been recently demonstrated to treat depression and drug-resistant depression after one or two administrations with psychotherapy. However, psilocybin provokes hallucinations, giving rise to growing concerns for its large-scale use. Lisuride is a 5-HT2A receptor agonist structurally analogous to LSD, but without its hallucinogenic effects. The goal of this research was to understand their different mechanisms using in vivo electrophysiology and behavioral pharmacology.

Methods: We compared effects of psilocybin and lisuride on behavior and on serotonin (5-HT) neuronal firing activity within the dorsal raphe nucleus (DRN), dopamine (DA) neuronal firing within the substantia nigra (SN) focusing on the pars compacta, as well as neuronal firing activity within the infralimbic region of the medial prefrontal cortex (mPFC) in adult male C57BL6/N mice (n = 5-23). In vivo extracellular single-unit electrophysiology recorded the effects of psilocybin (0.3, 1, 3, and 6 mg/kg) and lisuride (0.1, 0.2, 0.3, 0.4, and 0.5 mg/kg) on these three different neuronal firing activities. MDL100907 (0.2 mg/kg), a selective 5-HT2A receptor antagonist, was used to block the 5-HT effects of psilocybin and lisuride (n = 5 - 8). Behavioral tests included head twitch response (HTR) assessment and forced swim test (FST). HTR was scored during the first 10 minutes after acute intraperitoneal injections of psilocybin, lisuride or vehicle. Forced swim test (FST) was conducted 20 min after injection of psilocybin, lisuride or vehicle. Behavioral measures were analyzed using one-way analyses of variance (Regular, Welch’s, or Kruskal-Wallis) followed by their respective corrected post hoc analysis (Dunnett’s, Dunnett’s T3, or Dunn’s). Electrophysiological data were analyzed using repeated measures one-way ANOVA (RM ANOVA) or mixed-model ANOVA (if missing values were present) followed by Bonferroni post-hoc for multiple comparisons.

Results: Psilocybin, but not lisuride, elicited HTRs (p < 0.0001), as previously reported. FST immobility was decreased by the highest dose of lisuride (0.5 mg/kg, p = 0.0072), but not by psilocybin (0.3, 1, 3 mg/kg); however, when we increased the dose of psilocybin up to 6mg/kg, a lower immobility time was detected (p = 0.0034). 5-HT cell firing was decreased by both psilocybin (p = 0.0023) and lisuride (p = 0.0011). The maximal 5-HT firing inhibition was reached at the dose of 3mg/kg for psilocybin and 0.5mg/kg for lisuride. Only psilocybin inhibition was prevented by MDL100907. Using the same doses, DA cell firing activity in the SN pars compacta was decreased by psilocybin at 2 mg/kg (p = 0.0081, -68%) and 3 mg/kg (p = 0.0022, -79%), but only moderately at the highest dose of lisuride (0.5 mg/kg; p = 0.0459, -59%). Psilocybin decreased the firing rate of 8/11 (73%) mPFC neurons at the dose of 6 mg/kg (p = 0.0353). Similarly, lisuride decreased the firing rate of 4/6 (66%) of PFC neurons at the dose of 0.4 (p = 0.0272) and 0.5 mg/kg (p = 0.0293). MDL100907 reversed psilocybin-induced decrease of mPFC neuron firing, but did not reverse the effects of lisuride.

Conclusions: Together, these results confirm that HTR is evoked only by psilocybin and not by lisuride. Both of them display antidepressant-like effects as detected in the FST. Psilocybin decreases 5-HT firing activity in a 5-HT2A receptor-dependent manner, but lisuride does so independently of this receptor. Similarly, both psilocybin and lisuride decrease the firing activity of the mPFC at the highest doses tested, and this effect is partially mediated by the 5-HT2A receptor for psilocybin, but independent from 5-HT2A receptor for lisuride. While psilocybin decreases the DA firing activity almost completely, lisuride only partially decreases DA activity at the doses tested. Thus, lisuride and psilocybin show distinct activities in vivo beyond that previously documented by HTR. Further research is ongoing to better understand the intrinsic activity of these compounds on 5-HT, DA and glutamatergic receptor subtypes in different brain areas.

Keywords: Psychedelics, Psilocybin, Psilocybin analog, Single-unit electrophysiology in vivo, Novel Rapid Antidepressant

Disclosure: Diamond Therapeutics: Contracted Research (Self). Cosmas Therapeutics Development: Founder (Self).

P639. Pharmacological Reversal of Opioid-Induced Ventilatory Depression in Rats: Differences Among Opioid Agonists

Shawn Flynn*, Charles France

University of Texas Health Science Center at San Antonio, San Antonio, Texas, United States

Background: The opioid epidemic remains a major public health challenge, with over 83,000 opioid overdose deaths reported in 2022. Over 75,000 of these deaths involved synthetic opioids, primarily fentanyl and fentanyl analogs such as carfentanil, which is ~100-fold more potent than fentanyl. Some clinical reports suggest that larger doses of naloxone are necessary to reverse overdoses involving highly potent synthetic opioids such as fentanyl and carfentanil. The opioid receptor antagonist naloxone is the commonly used drug to reverse opioid overdose in humans and diprenorphine is an opioid receptor agonist used to reverse the effects of superpotent opioid agonists in veterinary medicine. The aims of this study were: 1) to characterize the ventilatory depressant effects of carfentanil, fentanyl, and heroin and 2) compare the potencies of the naloxone and diprenorphine to reverse opioid-induced ventilatory depression in rats, testing the hypothesis that opioid receptor agonists differ only in their potency to produce ventilatory depression, and that the potency of opioid receptor antagonists to reverse the effects of an opioid agonist will not differ across the agonists studied.

Methods: Ventilation was measured in male and female Sprague-Dawley rats (n = 15-16/sex) using whole-body plethysmography in 90-minute sessions. Following a 30-minute baseline period, rats received an i.v. infusion of carfentanil, fentanyl, or heroin. The primary dependent variables were minute volume, the volume of air ventilated on average per minute, and apnea duration. Dose-response curves were determined for carfentanil, fentanyl, and heroin, and the potency of each drug to reduce minute volume by 25% was compared across drugs and sex. Functionally equivalent doses of each drug were determined by estimating the dose of each drug (within each sex) to reduce minute volume by 50%. These doses were then used for the reversal experiments.

For reversal experiments the same procedure was used, with the addition of an infusion of naloxone or diprenorphine 5 minutes following the infusion of an opioid receptor agonist. Rats were assigned to one of the three opioid receptor agonists (n = 6/sex/group; 36 total). The doses of each agonist administered to each sex were functionally equivalent as determined by the first experiment to allow for accurate comparison of antagonist potencies. Dose-response curves for naloxone and diprenorphine were determined in each animal and the potency of each drug to restore minute volume to 85% of baseline was calculated. The potency of each antagonist was compared across agonists and sex. The relative potency of naloxone and diprenorphine to reverse the effects of the same agonist was also compared across groups.

All procedures were approved by the University of Texas Health Science Center at San Antonio Institutional Animal Care and Use Committee.

Results: Carfentanil, fentanyl, and heroin dose-dependently reduced minute volume in male and female rats. Fentanyl was approximately 50-fold more potent than heroin in female rats, and 10-fold more potent than heroin in male rats. Carfentanil was ~300-fold more potent than heroin in both sexes. All three opioid receptor agonists produced a similar profile of effects on ventilatory parameters, including long periods of apnea following administration.

Naloxone and diprenorphine did not differ in potency, and dose-dependently reversed the ventilatory depressant effects of all three opioid receptor agonists. Interestingly, the potencies of both antagonists to reverse the effects of carfentanil were significantly lower compared with the potencies to reverse the effects of heroin. That is, larger doses of opioid receptor antagonists were required to reverse the effect of carfentanil compared with an equi-effective dose of heroin. Ongoing analysis of the potencies of naloxone and diprenorphine to reverse the ventilatory depressant effects of fentanyl will determine whether this difference is unique to carfentanil or perhaps a property of fentanyls in general.

Conclusions: These findings are consistent with existing literature regarding the relative potencies of carfentanil, fentanyl, and heroin, and demonstrate that the profile of the ventilatory depressant effects of mu opioid receptor agonists in awake and behaving rats is not different among agonists. The finding that naloxone was less potent at reversing the effects of carfentanil compared with heroin is consistent with some existing literature, and ongoing experiments will determine whether this difference is unique to carfentanil or if it is true for fentanyl as well. These findings provide important information relevant to both the treatment of opioid use disorder and opioid overdose, as well as the design of improved therapeutics for these conditions. Further study of potential differences in the pharmacodynamics of carfentanil compared to other mu opioid receptor agonists is warranted.

Keywords: opioids, overdose, substance use disorder

Disclosure: Nothing to disclose.

P640. Drug Repositioning for Bipolar Disorder: Leveraging Lithium Effects on Peripheral Blood Gene Expression to Discover Novel Therapeutic Targets

Avery Martin, Bo Hu, Amit Anand*

Harvard Medical School and Mass General Brigham, Boston, Massachusetts, United States

Background: Bipolar Disorder (BD) is a prevalent psychiatric disorder, affecting about 1 in 100 people. Although there are many antipsychotic treatments for BD, lithium remains the gold standard. Despite lithium’s effectiveness, its true mechanism of action remains unclear, and it is unknown if there are any other therapeutics which can closely mimic lithium’s effect. To address the question of whether any existing drugs may be able to fill this role, we explored the possibility of drug repurposing for the treatment of BD.

Methods: Gene expression data was collected by blood sample from a cohort of patients (N = 25) undergoing lithium treatment at Indiana University Hospital (1). They were recruited through the Indiana University Hospital outpatient psychiatry clinic and community advertisement. 52% participants were female, the mean age was 33.9 (SD = 7.7) and the range spanned 19-59. The mean Hamilton Depression score was 11.8. Gene expression at baseline and after 8-week’s of treatment was compared, and the results were adjusted for multiplicity by the Benjamini-Hochberg procedure. Using gene-level results, pathway analysis was then performed using four different methods and three different biological term databases (GO, KEGG, and Reactome) to identify pathways implicated after lithium treatment. The genes comprising significant pathways were used to identify repurposing drugs using the NeDRex platform (2). The ROBUST (robust disease module mining via prize enumerating Steiner Trees) algorithm was used to expand the seed genes into disease modules, or networks of closely related genes involved in disease pathways. The Closeness Centrality algorithm, which considers protein-protein and protein-drug associations within disease modules, was used to identify potential targets for drug repurposing of BD.

Results: Six genes were found to be significantly differentially expressed: ADORA 3 (padj =0.02, log fold change (lfc) = 0.38), ILR5A (padj =0.02, lfc = 0.69) P2RY14 (padj =0.03, lfc = 0.67), TPPP3 (padj =0.04, lfc = 0.43), IDO1 (padj =0.04, lfc = 0.54), and CPT1A (padj =0.04, lfc = -0.48). Among the 109 pathways used, the most notable pathway found was the purinergic signaling pathway, with p adj =0.011 in over-representation analysis and p = 0.002 in active subnetwork enrichment analysis. This pathway is thought to be relevant to neurotransmission and neuroinflammation involved in BD mood regulation (3).

Using NeDRex platform, several drug classes were identified as acting on similar pathways as lithium. Various known antipsychotics which have been used to treat BD, such as Chlorpromazine and Thioridazine, were identified. Antifungals such as Miconazole and Econazole were highly ranked and have evidence suggesting that their anti-cortisol effects or effects on melastatin 2, a calcium (Ca2 + ) channel may be a mechanism of their beneficial effects in BD.. Bipolar disorder is known to have disrupted Ca2+ signaling, and l-type calcium channel antagonists have been used in BD treatment in the past (4). Notably, Tamoxifen, a selective estrogen receptor modifier which has been shown to be a promising BD treatment in prospective clinical trials (5) was identified. It has been established that the Protein Kinase C inhibition is a mechanism of both lithium and Tamoxifen, which may also explain the presence of many Tyrosine Kinase Inhibitors in the final list of drug repurposing candidates in our analysis.

Conclusions: It is important to note that the results of the NeDRex analysis do not necessarily suggest that the candidates will have a positive effect on BD treatment, as it only ranks drugs based on association strength between drugs and proteins. Several drugs, such as Sunitinib, are the subjects of case reports in which patients with existing BD experienced worsening symptoms after drug administration (6). Thus, it is important that the results of this analysis are verified through literature review, further research, and clinical trials to determine truly effective BD targets.

Multiple drugs were identified to be potential therapeutics for bipolar disorder, presumably by imitating the mechanisms of lithium. These may act through a variety of pathways, including kinase inhibition, lowering cortisol levels, or affecting calcium signaling. More research must be conducted to confirm whether these targets may be beneficial in treatment.

Funding: NIH R01MH075025 and R01 MH113256-05 (P.I. Amit Anand MD)

References:

1. Anand A, Nakamura K, Spielberg JM, Cha J, Karne H, Hu B. Integrative analysis of lithium treatment associated effects on brain structure and peripheral gene expression reveals novel molecular insights into mechanism of action. Transl Psychiatry 2020; 10(1): 103.

2. Sadegh S, Skelton J, Anastasi E, et al. Network medicine for disease module identification and drug repurposing with the NeDRex platform. Nat Commun 2021; 12(1): 6848.

3. Gonçalves MCB, Andrejew R, Gubert C. The Purinergic System as a Target for the Development of Treatments for Bipolar Disorder. CNS Drugs 2022; 36(8): 787-801.

4. Cipriani A, Saunders K, Attenburrow MJ, et al. A systematic review of calcium channel antagonists in bipolar disorder.. Mol Psychiatry 2016; 21(10): 1324-32.

5. Palacios J, Yildiz A, Young AH, Taylor MJ. Tamoxifen for bipolar disorder: Systematic review and meta-analysis. J Psychopharmacol 2019; 33(2): 177-84.

6. Kunene V, Porfiri E. Sunitinib-induced acute psychosis: case report. Clin Genitourin Cancer 2011; 9(1): 70-2.

Keywords: gene expression, Drug discovery, lithium, bipolar disorder, mood disorders

Disclosure: Nothing to disclose.

P641. Cognitive and Neuronal Benefits of Targeting the α5-GABAA Receptor in Animal Models of Brain Disorders

Thomas Prevot*, Ashley Bernardo, Michael Marcotte, Kayla Wong, Cassandra Marceau-Linhares, Celeste Pina-Leblanc, Adrien Bouchet, Prithu Mondal, James J. Cook, Etienne Sibille

University of Toronto, Centre for Addiction and Mental Health, Toronto, Canada

Background: Reduced GABA/somatostatin (SST) signaling is reported in psychiatric, stress-related and neurodegenerative disorders. Cortical SST+ interneurons inhibit the dendrites of excitatory neurons, largely through α5-containing GABAA receptors (α5-GABAAR). We showed that an α5-positive allosteric modulator (α5-PAM) alleviates working memory deficits and reverses neuronal atrophy in old mice. We then started to investigate the behavioral and neurotrophic effects of this α5-PAM in animal models of aging, chronic stress, β-amyloid load, and tauopathy.

Methods: Four studies are presented ( ~ 12 mice/group, 50% female): 1) Young C57BL6 subjected to chronic stress. 2) 22-month-old C57BL6 developing an age-related cognitive decline. 3) 5xFAD transgenic mice with progressive amyloid load. 4) PS19 transgenic mice developing tauopathy. Efficacy of 4 weeks of GL-II-73 (30mg/kg, p.o) at rescuing cognitive deficits was assessed for working memory (alternation task), spatial memory (water maze), and cognitive flexibility (set-shifting). Brains were then stained (Golgi-Cox; n=4brain/group; 8cell/brain) for spine density and maturation quantification in the prefrontal cortex and hippocampus (NeuroLucida).

Results: Chronic treatment in all models reversed cognitive deficits (ps < 0.01), with a strong effect on working memory. Chronic treatment reversed UCMS-, age-, amyloid- or Tau-protein-induced reduction of spine density at all maturation steps (p < 0.001 in PFC and CA1).

Conclusions: Results support that selective α5 targeting of GABAA receptors overcomes chronic stress-, aging-, amyloid- or tau-related cognitive deficits and detriments in neuronal morphology at all maturation steps. For the first time, a GABAergic intervention shows a symptomatic and disease-modifying therapeutic potential and could represent a major avenue for clinical development for patients suffering from cognitive deficits across brain disorders.

Keywords: GABA, GABA-A, positive allosteric modulators, Cognition, Neuronal morphology, Mouse models

Disclosure: DAMONA Pharma: Consultant (Self).

P642. Does Delta-9-Tetrahydrocannabinol Exposure Impact Acute Pain Sensitivity?

Elisa Pabon*, Anna Hilger, Conor Murray, Stephanie Lake, Timothy Fong, Ziva Cooper

UCLA Center for Cannabis and Cannabinoids, Jane and Terry Semel Institute for Neuroscience and Human Behavior, Department of Psychiatry and Biobehavioral Sciences, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, California, United States

Background: Chronic pain affects billions of people worldwide, leading to decreased quality of life, increased disability, and higher mortality rates, making it a significant public health priority. Pain management is one of the most common indications for medical cannabis use. While acute administration of cannabis and cannabinoids has been shown to reduce pain responses in both healthy individuals and certain patient populations—with notable sex differences—the impact of repeated exposure to delta-9-tetrahydrocannabinol (THC), the primary psychoactive component of cannabis, on pain sensitivity during abstinence remains underexplored, particularly with respect to sex differences. Limited clinical evidence suggests increased THC exposure is associated with increased pain sensitivity during abstinence, potentially due to the desensitization of the endocannabinoid system. Circulating cannabinoid levels may play a critical role, as there is evidence suggesting that cannabinoid exposure influences these levels, which are inversely correlated with pain sensitivity. The present analysis investigates the relationship between weekly THC exposure, sex, and their effects on acute pain response and circulating cannabinoid levels in a sample of healthy adults who regularly use inhaled cannabis.

Methods: Healthy adults (21-55 years) who reported inhaled cannabis use within the past month were recruited for the study. Exclusion criteria included a current chronic pain, use of prescription medications, regular use of over-the-counter pain relievers, major psychiatric disorders (i.e., mood, anxiety, psychotic disorders), and women who were pregnant, lactating, or using any form of hormonal contraceptive. Eligible participants self-reported their recent cannabis use, including the average days per week of cannabis use over the past month, average daily grams of cannabis consumed, and the estimated THC concentration of the cannabis used. From these data, estimated weekly THC exposure was calculated. After confirmation of cannabis abstinence for at least twelve hours via urine and saliva, participants had their blood drawn and underwent a Cold Pressor Test (CPT), a validated experimental pain procedure. During the CPT, participants immersed their hand in 4°C water, and both pain threshold (time to first report of pain) and pain tolerance (time to hand withdrawal) were recorded. Subjective pain ratings were assessed using the Short-Form McGill Pain Questionnaire (SF-MPQ) and the ‘Painfulness’ and ‘Bothersomeness’ scales. Mixed effects models, accounting for individual variability, were used to test the effects of weekly THC exposure and sex on all outcome measures (α = 0.05).

Results: Healthy male (N = 43; 32.0 ± 6.9 years) and female (N = 27; 28.7 ± 8.0 years) adults who reported inhaled cannabis use in the past month (M: 4.6 ± 2.2 days of cannabis use/week, F: 4.5 ± 2.3 days of cannabis use/week) completed the study. There was a significant effect of sex on pain tolerance, males exhibited greater pain tolerance than females (p < 0.05). However, average weekly THC exposure in these healthy, recently abstinent individuals was not significantly related to pain threshold, tolerance, or subjective ratings of pain (all p > 0.05). The ongoing analysis of plasma cannabinoid levels aims to further elucidate their relationship to THC exposure and acute pain sensitivity.

Conclusions: These findings indicate that, in healthy adults who report current inhaled cannabis use, average weekly THC exposure did not significantly alter sensitivity to pain, suggesting that repeated cannabis use may not lead to increased pain sensitivity in this group during acute abstinence. These conclusions are specific to the studied population and this abbreviated period of abstinence; greater THC exposure and longer durations of abstinence may reveal increased pain sensitivity. Further analysis of plasma cannabinoid levels is ongoing to better understand their relationship to THC exposure and pain sensitivity. These findings contribute to the ongoing discussion regarding the clinical use of cannabinoids as potential analgesic therapies, but further research is needed to explore these effects in different contexts and populations.

Keywords: cannabis, THC, Pain sensitivity, abstinence, cannabinoids

Disclosure: Nothing to disclose.

P643. Evaluation of Synthetic Cannabidiol (CBD) Within the Preclinical Screening Platform for Pain

Smriti Iyengar*, Mark Urban, Mark Varney, David Budac, Elizabeth Dugan, Taleen Hanania, Sarah A. Woller

NINDS, NIH, Rockville, Maryland, United States

Background: The NIH HEAL Initiative Preclinical Screening Platform for Pain (PSPP) program aims to accelerate the preclinical development of non-opioid, non-addictive pain therapeutics. Towards this goal, the program accepts small molecules, biologics, natural products, and devices from industry, academic, or government asset owners. In collaboration with PsychoGenics Inc., the program evaluates in vitro functional properties and protein binding, and in vivo pharmacokinetics, side effect profile, abuse liability, and efficacy in preclinical pain models, including plantar incision, L5/L6 spinal nerve ligation (SNL), monosodium iodoacetate (MIA), and chemotherapy-induced peripheral neuropathy (CIPN) models. As part of the validation within the program, clinically used drugs as well as CNS active compounds and negative controls have been evaluated, to better understand their efficacy profile in preclinical pain related models. Within this presentation we will describe ongoing efforts to systematically profile synthetic cannabidiol (CBD) in rats within the PSPP program.

Methods: CBD was evaluated for in vitro functional activity (0.001-100uM) at known targets of drugs of abuse and key safety targets in human receptor preparations. CBD (PO, 5% DMAC, 30% Solutol HS-15, 65% saline) was then evaluated in vivo, including a dose-response assessment of pharmacokinetics (1 and 10 mg/kg, N = 3-4), side effect profile assays (3-300 mg/kg, N = 4-10), and efficacy (1-100 mg/kg, N = 10) in pain related models, using two endpoints each, in male and female Sprague Dawley rats (Envigo). All experimenters were blinded to treatment. Each study included vehicle and positive control groups. Pre-determined inclusion criteria were applied for each endpoint assessed. Data were analyzed using two-way repeated measures ANOVA with Bonferroni’s or Dunnett’s post hoc test where appropriate.

Results: Partial activation of the cannabinoid 1 receptor (CNR1) with a potency of 50nM was noted while higher concentrations inhibited CNR1 in functional in vitro screens. Following PO administration (10 mg/kg) maximal CBD levels in plasma were observed at 1 hour which decreased over 2 hours in both male and female rats. Higher Cmax was noted in female compared to male rats (4.28 and 1.64 µM respectively) and the AUCs over the 24 hour period were greater in female than male rats. In the accelerated rotarod assay (10, 30, 100, 300 mg/kg, PO) and the modified Irwin assay (3, 10, 30, 100 mg/kg PO), CBD did not cause any significant deficits in either male or female rats. CBD (1, 3, 10, 30 mg/kg PO) did not significantly reduce guarding behavior or mechanical allodynia behavior in the plantar incision model and did not affect either mechanical allodynia behavior or cold allodynia in the L5/L6 SNL model. 7 days of administration in the SNL model also did not have any effects on either endpoint. In the MIA model of osteoarthritis pain, however, CBD (3, 10, 30, 100 mg/kg PO) administered BID for 4 days starting at Day 5 after MIA administration, showed a statistically significant and dose dependent reversal of mechanical allodynia behavior in both sexes after repeated administration. CBD had no effect on dynamic weight bearing. In the paclitaxel CIPN model, CBD (3, 10, 30, 100 mg/kg PO), showed a significant reversal of mechanical allodynia behaviorin paclitaxel-treated male rats, (a) after single administration and (b) after 6 days repeated administration. In female rats, no significant reversal was noted either after single or repeated administration, although female rats were shown to have higher levels of CBD compared to male rats, at doses tested, as noted above. CBD (3, 10, 30, 100 mg/kg PO) did not reverse mechanical allodynia behavior in the oxaliplatin CIPN model and did not reverse cold allodynia in either the paclitaxel or oxaliplatin CIPN models, at any dose tested

Conclusions: A rigorous evaluation of synthetic CBD is being performed within the PSPP program. Evolving data suggest that CBD did not have any pain-relevant effects in the plantar incision or L5/L6 SNL models but showed statistically significant effects on mechanical allodynia behavior in the MIA model of osteoarthritis pain at doses assessed in male and at 30 and 100 mg/kg in female rats but had no effect on dynamic weight-bearing in this model. CBD reversed mechanical allodynia behavior in the paclitaxel CIPN model at 100 mg/kg, in male rats, but not in female rats, and had no effect on mechanical allodynia behavior in the oxaliplatin CIPN model and no effect on cold allodynia in either CIPN model.

Keywords: Cannabidiol, Pain, Therapeutic Development, NIH HEAL Initiative

Disclosure: Nothing to disclose.

P644. Prevalence Estimates of Cytochrome P450 Phenoconversion in Youth Receiving Pharmacotherapy for Mental Health Conditions

Samuel Gerlach, Abdullah Al Maruf, Sarker Shaheen, Ryden McCloud, Madison Heintz, Laina McAusland, Paul D. Arnold, Chad Bousman*

University of Calgary, Calgary, Canada

Background: Pharmacogenetics-predicted drug metabolism may not match clinically observed metabolism due to a phenomenon known as phenoconversion. Phenoconversion can occur when an inhibitor or inducer of a drug metabolizing enzyme is present. Although estimates of phenoconversion in adult populations are available, prevalence estimates in youth populations are limited. To address this gap, we estimated the prevalence of phenoconversion in youth (6-24 years) receiving pharmacotherapy for mental health conditions and who had underwent pharmacogenetics testing.

Methods: 1285 participants (49% female) enrolled in the Pharmacogenetic-Supported Prescribing in Kids (PGx-SParK) study and completed their baseline visit before May 1, 2024 were included. PGx-SParK is an ongoing trial examining pharmacogenetic testing implementation in Western Canada (NCT04797364). The Research Electronic Data Capture (REDCap) application was used to collect participant’s self-reported demographic information (age, sex, biogeographical group), as well as current prescribed and over the counter medications. Current cannabis and cannabidiol use was collected via binary (yes/no) response. Genomic DNA was isolated from saliva samples collected using Oragene OG-600 kits (DNA Genotek, Ottawa). DNA was genotyped for CYP2B6, CYP2C19, CYP2D6, and CYP3A4 using an Agena VeriDose custom assay and the CYP2D6 VeriDose copy number variation assay. These genes were selected as they are the only CYP genes with published pharmacogenetics-based prescribing guidelines for mental health medications. For participants taking concomitant inhibitors or inducers of the CYP enzymes under study, published phenoconversion calculations were performed. For each of the four CYP genes, we estimated the proportion of participants that where phenoconverted (i.e., change in metabolizer phenotype) as well as the proportion with an actionable phenoconversion, defined as a change in metabolizer phenotype that would result in a recommended dose adjustment or drug avoidance action for one or more of their current mental health medications. Differences in the distribution of CYP2B6, CYP2C19, CYP2D6, and CYP3A4 metabolizer phenotype groups before and after adjustment for phenoconversion were assessed using a two-proportions Z test. An alpha threshold of 0.05 was used for all analyses.

Results: Nearly half (46%) of the cohort was estimated to be phenoconverted for one of the four genes examined. Comparison of metabolizer phenotype frequencies before and after adjustment for phenoconversion, showed significantly more youth had actionable phenotypes for CYP2C19 (60.3% vs. 69.2%; p = < 0.001), CYP2D6 (49.3% vs. 63.0%; p = < 0.001), and CYP3A4 (8.5% vs.12.3%; p = 0.003) after phenoconversion adjustment. Of youth that were phenoconverted, 24% had a change in their metabolizer phenotype that current PGx-based prescribing guidelines would recommend a change to standard prescribing (dose adjustment, alternative medication).

Conclusions: We found a high prevalence of phenoconversion among youth receiving pharmacotherapy for mental health conditions. The most pronounced impact was on CYP2C19 and CYP2D6, where we found 7.3-fold and 4.5-fold increases in the proportion of poor metabolizers after adjustment for phenoconversion. Given the increased adoption of PGx testing in child and adolescent mental health settings, the clinical implications of phenoconversion are non-trivial and should be considered when implementing PGx testing in practice.

Keywords: pharmacogenetics, Children and Adolescents, Drug metabolism, drug-drug interaction

Disclosure: Sequence2Script Inc: Founder (Self).

P645. New Insights From Electrophysiological and Immunofluorescence Experiments on the Integration of Kappa and Mu Opioid Receptor Signalling in the Paraventricular Thalamus

Eloise Kuijer*, Chris Bailey, Sarah Bailey, Sue Wonnacott, Sharon Smith, David Heal

NIH/NIAAA, Bethesda, Maryland, United States

Background: Kappa and mu opioid receptors (KOPrs and MOPrs) respectively contribute to stress and reward processing. Both receptor subtypes are highly expressed in the paraventricular thalamus (PVT), which is a central node for integrating stress- and reward-related information, and in addition, the PVT may be a pivotal site for opioid receptors to exert their physiological effects. That said, how aversive and rewarding opioidergic signalling converge to dictate PVT activity is not known. First, we investigated cFOS expression in the PVT after systemic pharmacological KOPr activation. In subsequent whole-cell patch-clamp experiments in single PVT neurons, we analysed the functional implications of KOPr and MOPr co-expression. We describe how KOPr and MOPr currents on single PVT neurons overlap and where they critically differ. Collectively, these new findings provide important insights on how these opioidergic receptors shape PVT functioning.

Methods: cFOS immunohistochemistry was performed on 8 weeks old, C57BL/6J mice (18 males, 18 females) after a single IP injection of KOPr agonist (±)U50,488 (5mg/kg or 20mg/kg) or saline. Sections were incubated at 4°C with primary antibody for c-FOS (9F6) rabbit mAb #2250 (1:500) and secondary goat anti-rabbit IgG antibody (H + L) Alexa Fluor 568 (1:250). Microscope images were analysed for cFOS+ neurons using ImageJ.

For electrophysiological experiments, brains of young C57BL/6J mice (27 males; 29 females) were removed under anaesthesia. Coronal sections (300μm) containing PVT were prepared and equilibrated for 30 mins in warm, 95% O2/5% CO2 saturated aCSF. Anatomical locations determined according to Paxinos and Watson Mouse Brain Atlas.

Whole-cell, voltage-clamp recordings of PVT neurons were made using pipettes with electrode resistance 4-6 MΩ and a potassium gluconate-based intracellular solution. Neurons were voltage clamped at -60mV and perfused with KOPr agonist spiradoline (30 μM), KOPr antagonist norBNI (1 μM), MOPr agonist met enkephalin (10 μM), and MOPr antagonist CTAP (1 μM).

Reversal potentials were calculated by measuring current during a ramp from a holding voltage of -140mV to -60mV over 8s. Reversal potentials were compared with the calculated Nernst Equation value in these experimental conditions -101mV.

As no sex-differences were detected, results from both sexes were pooled. Immunofluorescence counts analysed by 1-way ANOVA followed by Tukey’s multiple comparisons test, co-expression of KOPr and MOPr currents by Χ2 test, apparent negative opioid currents by 1-sample t-test against 0, and effects across anterior-posterior axis (mm from bregma) by regression analysis. Data shown as mean ± SEM; significance p < 0.05 (2 tailed).

Results: cFOS expression across the PVT was significantly increased after systemic U50,488 in both anterior PVT (aPVT, -0.70mm from bregma) (1-way ANOVA (treatment) F(2,33) = 6.835, p = 0.0033, n = 12/group) and medial PVT (mPVT, -1.82mm from bregma) (F(2,32) = 13.06, p < 0.0001, n = 11-12/group). Specifically, 20mg/kg U50,488 increased cFOS counts compared to saline in the aPVT by 188% [144-233%] (p = 0.003) and in the mPVT by 247% [200-293%] (p = 0.0001).

About 72% of PVT neurons examined produced a G protein-gated inwardly rectifying K+ (GIRK) mediated current in response to spiradoline (30 μM) of 14.1 ± 1.8pA (n = 28). Of the KOPr current-producing neurons, 95% also exhibited a MOPr agonist current (χ2 = 30.7, df=3, p < 0.0001) indicating that most PVT neurons functionally co-express both opioid receptor subtypes.

Analysis of reversal potentials revealed that the MOPr current reversed around the theoretical K+ equilibrium potential of –101mV (n = 11) while the KOPr current did not reverse at any holding voltage between -140 and -60mV (n = 12). The KOPr current was confirmed to be due to K+ conductance, as BaCl2 (2mM) completely blocked KOPr currents (n = 6).

Persistent tonic activation of MOPrs was inferred from the observation that MOPr antagonist CTAP itself produced a negative current (t(7) = 2.50, p = 0.041, n = 8), whereas KOPr antagonist norBNI did not (t(6) = 1.0, p = 0.36, n = 7). The negative CTAP current significantly differed over the anterior-posterior axis of the PVT (comparison to whether slope is significantly non-zero, F(1,6) = 11.66, p = 0.014) with a slope of 4.86 ± 1.42pA per mm from anterior to posterior (R2 = 0.66, Sy.x = 2.11).

Conclusions: The PVT is an important hub in the integration of stress and reward, our data suggest that the balance of activation of KOPrs and MOPrs in the PVT may play a critical role in integrating stress- and reward-related signals. We show that: (1) the PVT is sensitive to systemic pharmacological KOPr activation, (2) most PVT neurons functionally co-express KOPr and MOPr, (3) even though both KOPr and MOPr activation produces similar effects (i.e. outward hyperpolarising currents), there are key differences: MOPrs are constitutively active, especially in the aPVT. Furthermore, the finding that MOPr, but not KOPr, currents reversed at close to the K+ equilibrium potential suggests that MOPrs are located proximal to the cell body, whilst KOPrs are distal. Taken together, our data give insight in how opioidergic signalling shapes PVT neural activity with subtle differences across the PVT.

Keywords: kappa opioid receptor, Mu opioid receptor, slice electrophysiology, reward and aversion, paraventricular thalamus

Disclosure: Nothing to disclose.

P646. Oxycodone Dependence Alters Mu and Delta Opioid Receptor Regulation of PFC Inhibitory Transmission in a Cell-Type Specific Manner

Rebecca Cole, Vaishnavi Sankaran, Corinna Oswell, Gregory Corder, Max Joffe*

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Opioid use disorder (OUD) is a chronic and relapsing brain disorder that is characterized by an inability to control drug use and an intense withdrawal syndrome upon cessation. While Mu opioid receptor (MOR)-based therapies alleviate somatic discomfort and attenuate craving in OUD, these treatments are less effective at addressing long-lasting psychological symptoms that maintain drug use, necessitating a more complete understanding of changes to opioid signaling that occur during dependence. The delta opioid receptor (DOR) system has been consistently linked to hedonic deficits in rodent models of affective and reward behavior, supporting a role for dysregulated DOR signaling in these symptoms of OUD. The prefrontal cortex (PFC) is heavily implicated in OUD, and a substantial preclinical literature has linked PFC function with affective and motivational behavior. Though existing studies have focused mainly on opioid actions at excitatory synapses, the PFC endogenous opioid system is strongly localized to GABAergic interneurons (INs) which tightly regulate PFC circuitry and affective behavior.

Methods: We use genetically engineered female and male mice, viral-assisted optogenetics, whole-cell patch-clamp electrophysiology and slice pharmacology to determine how mu (MOR) and delta opioid receptors (DOR) regulated several functions of inhibitory neurons in PFC.

Results: We show that show that PFC MOR and DOR signaling suppresses spontaneous and evoked PFC inhibitory transmission through dissociable mechanisms. Furthermore, cell type-specific optogenetics revealed that PFC MOR and DOR signaling is synapse-specific and differentially expressed by PV, SST, and CCK-INs. An escalating dose regimen of oxycodone that produces physical dependence and motivational and affective behavior alters SST- and PV-IN MOR and DOR signaling in a cell type-specific manner

Conclusions: These findings suggest that cell- and receptor-specific adaptations to PFC GABAergic transmission may promote different facets of opioid-related behavior following dependence. Ongoing work uses cell type-specific calcium imaging and chemogenetics to monitor PFC IN activity during the development of opioid dependence and withdrawal and probe their role in behavior.

Keywords: Opioid signaling, prefrontal cortex, opioid dependence

Disclosure: Nothing to disclose.

P647. Effects of Novel β-Lactam, MC-100093, on Ethanol Consumption and Glutamate Uptake in the Nucleus Accumbens of Female and Male Alcohol-Preferring Rats

Ahmed Alotaibi, Magid Abou-Gharbia, Wayne Childers, Youssef Sari*

University of Toledo, College of Pharmacy and Pharmaceutical Sciences, Toledo, Ohio, United States

Background: Glutamate is the predominant excitatory neurotransmitter in the brain. The uptake of extracellular glutamate is regulated in majority by the astrocytic glutamate transporters. Therefore, astrocytic glutamate transporter 1 (GLT-1), a Na + -dependent excitatory amino acid transporter, is responsible for the uptake of more than 90% of total extracellular glutamate. In addition, other transporters such as cystine-glutamate exchanger (xCT), a Na + -independent transporters, controls < 10% of glutamate uptake. Ethanol consumption alters glutamate homeostasis in several brain regions, including the nucleus accumbens (NAc). Chronic ethanol-drinking downregulated GLT-1 in the NAc of male P rats. Studies from our lab showed that ceftriaxone, a beta-lactam antibiotic, reduced ethanol drinking and normalized extracellular glutamate concentrations in the NAc, and this effect was associated with upregulation of GLT-1. Besides, we have recently reported that a novel beta-lactam, MC-100093 which doesn’t have any antibiotic action, attenuated ethanol consumption and normalized the expression of GLT-1 in the NAc. The current study aimed to investigate whether the attenuating effect of MC-100093 on ethanol consumption was associated with restoring glutamate uptake. Thus, we focused on measuring glutamate uptake in the NAc of male and female P rats.

Methods: Male and female rats were divided into three groups; 1) Water control group treated with saline, 2) Ethanol group treated with saline, and 3) Ethanol group treated with MC-100093. Following five weeks of continuous exposure to a free access to ethanol (15% v/v) and (30% v/v) and/or water, rats were treated for five consecutive days on the beginning of week 6 with saline or MC-100093 at a dose of 100 mg/kg/day (i.p.). Glutamate uptake was determined through [3H]Glutamate uptake assay in crude membrane fractionation of the NAc.

Results: The results revealed that treatment with MC-100093 attenuated ethanol intake as well as ethanol preference in both male and female P rats. Importantly, MC-100093 increased the Na + -dependent glutamate uptake in the NAc of both female and male P rats. Furthermore, Western blot analysis showed a normalization of GLT-1 expression in the NAc following treatment with MC-100093. In male P rats, ethanol group treated with saline showed a decrease in the Na + -dependent glutamate uptake. We further analyzed the effect of ethanol and MC-100093 on Na + -independent glutamate uptake. We found that treatment with MC-100093 increased glutamate uptake as compared to ethanol saline group in the NAc of both male and female P rats. The apparent effect of MC-100093 on Na + -independent glutamate uptake was associated with a normalization of xCT expression in both female and male P rats.

Conclusions: These findings suggest that MC-100093 effect in attenuating ethanol consumption and reducing the excitatory effect of glutamate imbalance is mainly through the normalization of Na + -dependent excitatory amino acid transporter GLT-1. Thus, MC-100093 may have potential therapeutic benefits in treating alcohol dependence.

Keywords: alcohol dependence, GLT-1, xCT, glutamate

Disclosure: Nothing to disclose.

P648. MK-1942, a Negative Allosteric Modulator of Metabotropic Glutamate Receptor 2, for Treatment-Resistant Depression

Aristide Merola, Yuki Mukai, Tjerk Bueters, Paulette Ceesay, Nicole Dupre, Christine Furtek, Gerard Sanacora, Michael Thase, Steven D Targum, Zoi Alexopoulou, Chantal Mahon, Wenping Li, Callie Miller, Kelly Yee, Sean M Smith, Jason Uslaner, Aubrey Stoch, Michael Egan*

Merck and Co., Inc., Rahway, NJ, USA, North Wales, Pennsylvania, United States

Background: MK-1942 is a negative allosteric modulator of the metabotropic glutamate receptor 2 (mGluR2) that was being studied for treatment-resistant depression. mGluR2 antagonism has been shown in some preclinical models to mimic and mediate the rapid antidepressant effects of ketamine. Preclinical data with MK-1942 suggest that receptor occupancies ≥40% are associated with maximal increase in glutamate metabolism. In Phase 1 studies in healthy participants, MK-1942 was generally well tolerated at single doses up to 12 mg and up to 50 mg bid with titration. The most frequently reported adverse events were dizziness, headache, and nausea, which showed a dose-response relationship and tended to decrease with repeated dosing. Based on the plasma concentration-receptor occupancy relationship determined from a Phase 1 PET imaging study in healthy participants, single MK-1942 doses of 10 mg are anticipated to achieve a peak receptor occupancy of 41%. We report here on results from a Phase 2 proof-of-concept study of MK-1942 for treatment-resistant depression.

Methods: The Phase 2 trial was a randomized, double-blind, 4-week study of MK-1942 given orally added to stable antidepressant therapy in adults 18-65y with treatment-resistant depression (NCT 04663321). Participants were randomized in a 2:1:2 ratio to either: (1) daily treatment with MK-1942 titrated from 5 mg bid to 20 mg bid; (2) intermittent, biweekly, treatment with MK-1942 10 mg; or (3) placebo. Efficacy was assessed using the Montgomery Asberg Depression Rating Scale (MADRS) at 3 weeks for the daily MK-1942 group (titrated dose of 20 mg) and 1 week for the biweekly MK-1942 group (10 mg). Safety and tolerability were assessed by adverse events and routine laboratory tests over the duration of the trial.

Results: The study was terminated on Aug 10 2023, following a safety review triggered by findings from liver function laboratory tests in this study and another Phase 2 study in people with Alzheimer’s disease. The identified cases were asymptomatic and did not meet criteria for serious adverse events. At the time of termination, the number of randomized participants was 40 for the daily MK-1942 group, 19 for the biweekly MK-1942 group and 40 for the placebo group, constituting about 70% of the planned enrollment. No significant differences from placebo in efficacy were seen for either MK-1942 group. The difference in least squares mean change-from-baseline in MADRS total score between the daily MK-1942 group and the placebo group directionally favored placebo at Week 3 (-8.1 vs -11.4, difference = 3.3 [95% CI: -1.4, 8.0], p = 0.168) and at all other time points. The difference between the biweekly MK-1942 group and the placebo group directionally favored placebo at Week 1 (-6.3 vs -8.1, difference = 1.9 [95% CI: -2.7, 6.4], p = 0.417) but not at later time points (e.g. Week 4, -14.3 vs -12.7, difference = -1.5 [95% CI: -8.0, 4.9], p = 0.636). At every timepoint, the difference between the biweekly MK-1942 group and the daily MK-1942 group directionally favored the biweekly group. There were no serious adverse events in the MK-1942 groups. Discontinuations of treatment due to an adverse event were more common in the MK-1942 groups than in the placebo group (daily MK-1942 = 10.0%, biweekly MK-1942 = 15.8%, placebo = 2.5%). The most frequently reported adverse events ( > 10%) in either of the MK-1942 groups were dizziness, nausea, diarrhea, and headache. The proportions of participants with dizziness were higher in the MK-1942 groups than in the placebo group (daily MK-1942 = 25.0%, biweekly MK-1942 = 31.6%, placebo = 7.5%).

Conclusions: MK-1942 administered daily for 3 weeks and biweekly for 1 week did not show significant differences compared with placebo in change-from-baseline MADRS total score. The numerically worse scores for the daily MK-1942 treatment group compared to the other two groups suggests the possibility that glutamate focused antidepressant treatments may require intermittent dosing strategies. Interpretation of the findings is limited by the early termination of the trial.

Keywords: Treatment Resistant Depression, Metabotropic glutamate receptor 2 (mGluR2), randomized clinical trial

Disclosure: Merck and Co., Inc. : Employee (Self). Merck and Co., Inc.: Stock / Equity - Privately Held Company (Self).

P649. Effects of COMT Inhibition on the Neural Correlates of Inhibitory Processing Among Individuals With Alcohol Use Disorder

Drew Winters*, Joseph Schacht

University of Colorado Anschutz Medical Campus, Children’S Hospital, Centennial, Colorado, United States

Background: Dysregulation of inhibitory control is a core feature of Alcohol Use Disorder (AUD). In the brain, inhibitory control is mediated, in part, by dopamine signaling in the prefrontal cortex (PFC). In the PFC, the principal method of dopamine inactivation is degradation by the enzyme catechol-O-methyltransferase (COMT). Tolcapone, a brain-penetrant, reversible COMT inhibitor that is FDA-approved for the treatment of Parkinson’s disease, potentiates extracellular dopamine release elicited by exogenous factors. Tasks that require inhibitory control also elicit cortical dopamine release, and drugs that increase extracellular dopamine concentrations enhance performance on such tasks. Thus, tolcapone might potentiate control-associated cortical dopamine release, and thereby improve inhibitory control in AUD.

Methods: We conducted a randomized, placebo-controlled trial of tolcapone (Schacht et al., 2022) in which non-treatment-seeking participants with AUD (n = 65 with usable neuroimaging data) received tolcapone (titrated to 200 mg t.i.d.) or placebo for 8 days. Participants completed an fMRI stop-signal task (SST; Li et al., 2006) at baseline and following 7 days of study medication ingestion. We first tested the interaction between treatment group and time to identify brain regions in which activation for the contrast of successful vs. unsuccessful stop trials differed between groups on Day 7 relative to baseline. We then tested whether activation to this contrast was associated with changes in drinking during the medication period or behavioral performance on the SST.

Results: The interaction between treatment group and time was significant for several regions, including the right dorsolateral prefrontal cortex (dlPFC), inferior frontal gyrus, and primary motor cortex and left occipital cortex, posterior cingulate, and precuneus. Of these regions, the greatest difference between groups was in the right dlPFC, in which activation for the successful > unsuccessful stop contrast increased relative to baseline in the tolcapone group but not the placebo group. Greater right dlPFC activation for this contrast was associated with reduced drinking during the medication period (β = –0.970, p = 0.011) and improvements in inhibitory control (i.e., decreased stop-signal reaction time) (β = –0.370, p = 0.005) relative to baseline. Additionally, reduced drinking was associated with improved inhibitory control (β = –0.872, p = 0.002).

Conclusions: In this human laboratory study of non-treatment-seeking individuals with AUD, the COMT inhibitor tolcapone increased activation of the right dlPFC, a brain area canonically associated with inhibitory control. This result is consistent with data from healthy controls suggesting that tolcapone improves a variety of cognitive functions, including executive functioning, working memory, and set-shifting (Apud et al., 2007; Giakoumaki et al., 2008; Roussos et al., 2009; Farrell et al., 2012). The associations between increased dlPFC activation, improved inhibitory control, and reduced drinking suggest that pharmacological interventions that increase cortical dopamine may yield promise for rescuing dysregulated inhibitory control among individuals with AUD.

Keywords: alcohol use disorder, Inhibitory control, tolcapone, dorsolateral prefrontal cortex (DLPFC), Clinical trial

Disclosure: Nothing to disclose.

P650. Characterization of Zincergic and Dopaminergic Dynamics in the Striatum During Locomotion and Cocaine Exposure

Oscar Solís*, Fallon Curry, Will Dune, Hao Zhang, Wenyuang Huang, Huiwang Ai, Michael Michaelides

National Institute on Drug Abuse, NIH, Baltimore, Maryland, United States

Background: The striatum integrates dopaminergic and glutamatergic inputs from the ventral tegmental area/substantia nigra pars compacta and the cortex, respectively. Dopamine (DA) in the striatum is critical for behaviors such as movement control, reward, and addiction. Glutamate is critical for the modulation of synaptic plasticity and regulating the activity of striatal neurons. Importantly, a subset of cortical glutamatergic projections to the striatum co-release Zn2 + . Synaptic zinc (Zn2 + ) is an essential element necessary for maintaining neurophysiological homeostasis. Previous studies from our lab demonstrated that synaptic Zn2+ in the striatum modulates cocaine-induced behaviors and dopaminergic transmission. Little is known about the molecular profile of Zn2 + -releasing (zincergic) neurons in circuits related to rewards, as well as what brain regions project zincergic neurons to the striatum. In addition, we investigated the changes in synaptic Zn2+ release and its relationship to the striatal dynamics of other neurotransmitters, such as dopamine, in both normal locomotor behavior or the behavioral effects of cocaine.

Methods: First, we combined retrobeads and fluorescent in-situ hybridization (RNAscope), this approach allowed us to trace the brain regions that send zincergic projections to the nucleus accumbens (NAcc) and the dorsal striatum (DS) of mice (n = 4). Additionally, to characterize zincergic neurons, we performed RNAscope in situ hybridization (ISH) in reward related circuits (prelimbic cortex, cingulate cortex, and amygdala) of mice (n = 5).

We next determined the dynamics of Zn2 + , and dopamine in the NAcc and the DS. C57BL/6 mice were injected with a Zn2+ sensor (AAV-hSyn-GRISZ, green fluorescent indicator for synaptic Zn2 + ; n = 5-6) and a DA sensor (AAV-hSyn-GRABDAm1; n = 7) in the NAcc or DS. Three weeks after AAV injection, we performed fiber photometry to record DA and synaptic Zn2+ dynamics during locomotion (mouse exceeded 5 cm/s), periods of immobility (below 0.25 cm/s), and following an injection of cocaine (10 mg/kg; i.p.).

Results: Our in-situ hybridization results showed that both the DS and NAcc received zincergic projections from the prelimbic cortex, cingulate cortex and amygdala. Additionally, the molecular profile analysis showed that in the prelimbic and the cingulate cortex, 44-55% of glutamatergic neurons are zincergic and 2-5% of gabaergic neurons are zincergic. In the amygdala, 51% of glutamatergic neurons are zincergic and 2% of gabaergic are zincergic neurons.

Our fiber photometry experiments revealed that synaptic Zn2+ and dopamine release increased in the DS (t = 3.08, p < 0.01, for Zn2 + ; t = 4.98, p = 0.001, for DA) and the NAcc (t = 4.06, p < 0.01, Zn2 + ; t = 8.49, p = 0.001, DA) during locomotion and remained low when the mouse was immobile. Additionally, following cocaine administration, there was an increase in the frequency (t = 4.05, p < 0.05, for Zn2 + ; t = 4.23, p = 0.01, for DA) and duration (t = 2.60, p < 0.05, for Zn2 + ; t = 4.94, p < 0.05, for DA) of Zn2+ and DA transients in the DS. In the NAcc, cocaine administration led to an increase in the frequency (t = 2.76, p < 0.05) and duration (t = 3.18, p < 0.05) of DA transients, and an increase in the duration of Zn2+ transients (t = 5.2, p < 0.05).

Conclusions: Our study provides new insights into the dynamics of synaptic Zn2+ and DA in the striatum, highlighting their significant roles in modulating locomotor behavior and cocaine response. The identification of zincergic projections from the prelimbic cortex, cingulate cortex, and amygdala to both the dorsal striatum (DS) and nucleus accumbens (NAcc) shows the importance of these brain regions in Zn2 + -mediated neurotransmission. Molecular profile analysis revealed that a significant proportion of glutamatergic neurons and a small portion of GABAergic neurons in these regions are zincergic, suggesting different roles for glutamatergic neurons. Our fiber photometry experiments further demonstrate that Zn2+ and DA release is activity-dependent, increasing during locomotion and following cocaine administration. These findings enhance our understanding of the neurophysiological mechanisms underlying movement and reward-related behaviors.

Keywords: Dopamine, cocaine, Zinc

Disclosure: Nothing to disclose.

P651. Cue-Evoked Nucleus Accumbens Acetylcholine Tracks Reward Value and Delay Cost in Impulsive Choice

Julianna Cavallaro, William Mannheim, Ronald Salazar, Czarina Maysonet, Ruby Setara, Dustin Zuelke, Eduardo Gallo*

Fordham University, Bronx, New York, United States

Background: Impulsive choice, often characterized by excessive preference for small, short-term rewards over larger, long-term rewards, is a prominent feature of substance use and other neuropsychiatric disorders. Nucleus accumbens (NAc) dopamine and its actions on D2 receptors (D2Rs) have been implicated in impulse control but defining the cellular site(s) of action has been difficult given that different NAc cell types and their afferents express D2Rs. One of these cell types, the cholinergic interneurons (CINs) not only regulate NAc dopamine release, but also respond to reward-associated cues with multiphasic alterations in their firing activity, including a pause in firing. Our recent work showed that D2Rs selectively expressed in NAc CINs promote impulsive choice. Here, to begin to unravel the underlying neural mechanisms, we used in vivo fiber photometry to investigate ACh signals in the mouse NAc during impulsive choice behavior, focusing on phasic responses to reward-predictive cues. We further determined whether CIN-selective D2R genetic ablation contributes to these cue-evoked phasic ACh responses.

Methods: We delivered adeno-associated viruses expressing a GPCR-Activation Based ACh sensor (GRABACh3.0) followed by unilateral implantation of an optic fiber into the NAc of adult ChAT-Cre mice and, in a separate set of experiments, in Drd2loxP/loxP and ChAT-IRES-Cre x Drd2loxP/loxP (CIN-D2KO) mice. Mice performed a delay discounting task that measures the choice between lever pressing for a small, immediate reward or for a larger reward presented after increasing delays. Each session began with forced choice trials, in which one of two light cues was presented for 5 s before the extension of its associated lever. Forced choice trials were followed by free choice trials, in which both light cues were presented for 5 s prior to the extension of both levers. Pressing either lever would result in the corresponding small or large reward. Photometry GRABACh3.0 signals were analyzed as time-locked events aligned to light cue onset and lever extension.

Results: In the absence of delays to the large or small rewards, GRABACh3.0-expressing ChAT-Cre mice increased their preference for the large reward across training days. As delays to the large reward were increased, mice showed significant discounting, as measured by decreased choice of the large reward. To determine whether “small” and “large” light cues would elicit contrasting ACh responses, we analyzed ACh signals during small and large forced trials of the initial no-delays training phase. Light cue onset elicited a transient increase followed by a sustained reduction in ACh below baseline that persisted through the end of the 5-s cue period. The ACh reduction evoked by the small light cue remained stable across training, while it became significantly more pronounced for the large light cue by the last day. Similarly, lever extension evoked a brief “dip” in ACh signal, which became significantly larger across training days, particularly in response to large lever extension. Increasing delays to the large reward resulted in progressive attenuation of the ACh responses to the large cues, reducing the contrast between large and small responses at delays longer than 2 s. Furthermore, we found a significant correlation between the size of the cue-evoked ACh reductions/dips in the forced trials and the large option preference in the subsequent free choice trials. In free choice trials, simultaneous presentation of both small and large cues (lights or lever extensions) led to ACh reductions and dips that became progressively larger with training in the 0-s delay sessions, but which were significantly diminished by increasing delays to the large reward. In preliminary findings, CIN-D2KO were not only less impulsive than control Drd2loxP/loxP mice, but they also showed significantly decreased cue-evoked ACh dips following simultaneous presentation of both small and large cues during choice trials.

Conclusions: Our data suggest that cue-evoked NAc ACh tracks the value of the rewards of different magnitude, is modulated by delay to reinforcement, and can track subsequent choice behavior. In addition, our preliminary data suggest a role for CIN D2Rs in enabling cue-evoked reductions in ACh signals, an effect that was associated with decreased impulsive choice. Together, these findings suggest a novel role for dynamic changes in NAc ACh activity in the neural mechanisms of impulsive choice.

Keywords: Dopamine, Acetylcholine, impulsivity, cholinergic interneuron, D2 dopamine receptor, In vivo fiber photometry

Disclosure: Nothing to disclose.

P652. The Sense and Nonsense of Antipsychotic Combinations: A Model for Dopamine D2/3 Receptor Occupancy

Moritz Spangemacher*, Christian Schmitz, Paul Cumming, Luca Färber, Xenia Hart, Hiroyuki Uchida, Gerhard Gründer

Central Institute of Mental Health, University of Heidelberg, Mannheim, Germany

Background: Approximately 20 – 30 % of patients treated for schizophrenia concomitantly take two or more antipsychotic substances, despite the limited evidence that antipsychotic combination treatment is superior to monotherapy. The common neurobiological rationale that combination therapy engages a broader range of pharmacodynamic targets may fail to consider the need for dose adjustment, which could become necessary due to additive occupancy at dopaminergic D2-receptors (D2Rs) in the brain.

Positron emission tomography (PET) studies can reveal the relationship between plasma levels of an antipsychotic medication and occupancy at striatal dopamine D2/3 receptors (D2R).

It is now generally accepted that striatal D2 occupancy of 65-80% is associated with effective treatment of most antipsychotics. This means that a threshold of at least 65% should be reached in order to achieve a therapeutic effect at the D2 receptor, while the risk for extrapyramidal side effects increases above the 80% threshold

But there is almost no consideration in the literature of the net occupancy obtained with antipsychotic combination treatment. In this report, we introduce a novel model for predicting net D2R occupancy in antipsychotic polypharmacy (APP), taking as illustrative examples five commonly prescribed antipsychotic medications.

Methods: In an extension of the law of mass action for predicting receptor occupancy from the plasma concentration of a single psychopharmacological agent, we test a model for inferring the net striatal D2R occupancy in APP from the individual Michaelis-Menten kinetics of two (or more) antipsychotic medications. Striatal D2-dopamine receptor occupancy of antipsychotic combination treatment can be deduced based on formula:

occupancy [%] = 100 – 100/(1 + C1/EC501 + C2/EC502)

We derived EC50 values for the five antipsychotic medications from an exhaustive, systematic search of published PET and SPECT studies reporting D2R occupancy as a function of plasma drug concentrations.

Results: Based on literature PET findings for striatal D2R occupancy in monotherapy, our model predicts that widely used antipsychotic medication combinations may exceed the optimal therapeutic window of 65-80% occupancy. Our extended model accurately predicted occupancy for the only APP combination documented by PET, the augmentation of clozapine with haloperidol.

Given EC50 values of 0,32 ng/mL for haloperidol and 483 ng/mL for clozapine, our model predicts that combination treatment with these drugs at plasma concentrations at the low ends of the respective recommended therapeutic reference ranges (haloperidol 3 ng/mL; clozapine 400 ng/mL) will lead to 92% D2R occupancy.

Our model predicts a mean ± standard deviation (SD) absolute difference of 7.8 ± 1.6% (range 5 - 9 %) between the estimated D2R occupancy at the indicated plasma concentrations and the receptor occupancy measured by PET. There was a high correlation between the observed and predicted values of D2R occupancy (r = 0.98, p < 0.005). The magnitude of the mean (SD) absolute difference closely matches those reported for the individual substances, namely 12.1 ± 9.1 % for clozapine and 7.0 ± 7.1 % for haloperidol.

In accord with the available clinical data, our model predicts the attainment of an effective receptor net occupancy even with low doses.

Conclusions: Present results call for caution in the design of antipsychotic medication combination therapy, aiming to avoid excessive occupancy by adjusting drug concentrations and doses. This scenario calls for confirmation through follow-up PET studies aiming to assess the separate and combined occupancies obtained with combination therapy. Our analysis shows that combining two antipsychotic substances could easily result in excessive net blockade of striatal D2Rs, resulting in extrapyramidal side effects and other deleterious side effects.

Keywords: Schizophrenia, Antipsychotics, Positron Emission Tomography (PET) Imaging, Dopamine (D2, D3) receptors, Antipsychotic Treatment Practice

Disclosure: Nothing to disclose.

P653. Accumbal Serotonin Signaling Tracks Both Appetitive and Aversive Stimuli

Ruixiang Wang*, Nagalakshmi Balasubramanian, Thomas James, Zeid Aboushaar, Kanza Khan, Daniel Canales, Catherine Marcinkiewcz

The University of Iowa, Iowa City, Iowa, United States

Background: The nucleus accumbens (NAcc) is arguably the most important node in modulation of appetitive and aversive information processing. While dopaminergic signaling in the NAcc responds to both appetitive (rewarding) and aversive stimuli by encoding incentive salience, the function of serotonergic (5-HT) transmission in this brain region is not well understood. Recent developments in fiber photometry and genetically encoded fluorescent indicators/biosensors have made it easier to monitor 5-HT neural activity and transmission in awake, freely behaving animals with high spatiotemporal resolution. In the present study, we used a wireless fiber photometry system to track accumbal 5-HT release in response to different stimuli.

Methods: First, two-photon imaging was employed to validate the 5-HT biosensor (GRAB5-HT3.5) we would use. Then, male C57BL/6J mice were transduced with GRAB5-HT3.5 in the right NAcc (ML = 0.70, AP = 0.98, DV = -4.55 mm), and fiber optic cannulae were implanted afterwards, with the fiber terminal reaching the viral injection site. Two weeks later, wireless fiber photometry recordings were performed using the Amuza TeleFiPho system, while mice were engaging in various behaviors upon presentation of different appetitive and aversive stimuli. To quantify 5-HT transients induced by these stimuli, we computed peri-event fluorescent changes versus the baseline (ΔF/F) and compared area under the curve (AUC) above the x-axis in the peri-event plot.

Results: In the two-photon imaging experiment, application of 5-HT (20 µM) to acute slices containing the dorsal raphe nucleus from mice transduced with GRAB5-HT3.5 resulted in an immediate increase in fluorescent intensity, confirming that this biosensor can swiftly detect extracellular 5-HT and emit green fluorescence signals accordingly.

In fiber photometry experiments, appetitive stimuli, like palatable food (Froot Loops) and sucrose water, led to increases in NAcc 5-HT transients (for AUC during 2 min post-event: empty petri dish vs. Froot Loops, p < 0.001, Froot Loop fragrance oil drops vs. Froot Loops, p = 0.07; in the sucrose water experiments, post- vs. pre-sucrose, p < 0.05). Interestingly, Froot Loops-induced rise in 5-HT transients was associated with “foraging,” or appetitive behavior, but not consummatory behavior, as 5-HT transients started to drop when food consumption began.

Aversive stimuli could also trigger NAcc 5-HT responses, whether the stimuli were internal, e.g., visceral malaise caused by lithium chloride (LiCl) injections (for AUC during 20 min post-injection: 125 mg/kg LiCl vs. saline: p < 0.05), or external, e.g., a predator odor from red fox urine (for AUC within 2 min post-event: empty petri dish vs. fox urine, p = 0.078). Interestingly, sucrose and quinine, stimuli with opposite valences, enhanced NAcc 5-HT transients in a similar pattern (post- vs. pre-quinine, p < 0.05; post-sucrose vs. post-quinine, p = 0.813).

In addition, repeated footshocks eventually triggered a reduction in NAcc 5-HT transients (for the 5th footshock: 2 s before shock vs. 2 s after shock, p < 0.05) perhaps because repeated footshocks constituted inescapable and uncontrollable aversive stimuli, leading to learned helplessness, a model for depression.

Conclusions: Accumbal 5-HT signaling tracks both appetitive and aversive stimuli. Local serotonergic transmission is increased in response to both types of stimuli except when animals have learned helplessness facing repeated footshocks, after which the decreased serotonin tone aligns with the notion that depression is associated with reduced 5-HT activity.

Keywords: Serotonin, Nucleus Accumbens, fiber photometry, Reward, aversion

Disclosure: Nothing to disclose.

P654. Preliminary Evidence for Contributions of Outcome Insensitive Associative Learning to Compulsivity in Obsessive Compulsive Disorder

Amy Rapp*, Anne G. E. Collins, H. Blair Simpson

Icahn School of Medicine At Mount Sinai, New York, New York, United States

Background: Compulsivity is a maladaptive trait in which a behavior persists despite resulting in an unfavorable consequence or no longer being rewarded. It is a core characteristic of Obsessive Compulsive Disorder (OCD), considered to be the prototypical compulsive disorder. Compulsivity in OCD has been theorized to result from an imbalance between deliberate, adaptable goal-directed control and more automatic, inflexible habits. Two distinct cognitive systems putatively underlie goal-directed and habitual behaviors: a “model-based” cognitive system that uses a model/map of the environment to make goal-directed decisions and “model-free” reinforcement learning (RL) that updates the expected value of stimuli through reward prediction errors (RPEs) to make habitual decisions. As such compulsivity in OCD is thought to occur due to a deficit in model-based learning and an excess of model-free RL.

Recent evidence has challenged assumptions about model-free RL contributions to reward-based learning. Re-analysis of several studies that used the same simple experimental working-memory task showed that an associative process that ignored outcome (a “Hebbian” or “habit-like” process referred to as the H-agent) accounted for reward-based learning behavior better than model-free RL. As such, the present study aimed to replicate and extend these findings by: (1) testing if a habit-like associative process that fails to integrate outcome accounted for reward-based learning better than model-free RL in data from adults diagnosed with OCD, and (2) evaluating if a computational parameter representing this outcome insensitive associative learning correlates with severity of compulsivity.

Methods: Participants were unmedicated adults diagnosed with OCD of at least moderate severity (n = 20) and healthy controls (n = 20) who completed the RLWM paradigm, a computerized reward learning task designed to examine the dynamic interaction of working memory (an executive function that supports model-based learning) and model-free RL. In this task, participants determined through trial and error which of three button presses (left, up, or right arrows on a standard keyboard) was associated with each novel visual stimulus presented in a given block. Working memory demand was manipulated by varying the number of visual stimuli over 12 blocks, ranging in set size (ns) from two to six visual stimuli. Feedback was provided indicating if the participant selected the correct corresponding action. Participants also completed a battery of clinician-administered and self-report measures, including a transdiagnostic measure of compulsivity.

Six computational models were fit to behavioral data and included in model comparison. Model variants included six or seven free parameters: working memory capacity, working memory decay, working memory reliance undirected noise, sticky choice, and H or RL learning rate. In the H-agent model, subjective outcome was set to one, meaning that both correct and incorrect feedback were treated the same. As such, the H-agent model tracked association strength, neglecting valence of outcome. Model fitting was conducted in Matlab2022a using fmincon optimization. The winning model was the model with the lowest AIC.

Results: Replicating past studies using the RLWM task, behavior was nearly optimal in small set sizes with an increasingly incremental learning curve as set size increased. The H-agent model demonstrated the best model fit compared to other models that included RL modules. This model also accounted for qualitative response patterns seen in behavioral data. On each trial, participants select from one of three button presses, meaning that there are two possible errors that could be made. If guided by model-free RL, participants would demonstrate a reduced frequency of repeating mistakes, suggesting integration of error feedback (i.e., the error that was made in a given trial would been made less frequently than the other error option). As set size increased, the chosen and unchosen error were selected with increasingly similar frequency. In simulations, only the H-agent model was able to predict this pattern. All other models that included RL predicted that a difference in frequency of selection for the two errors persisted as set size increased, inconsistent with the behavioral data obtained.

Across, the full sample, compulsivity was not associated with the Hebbian bias term, r = 0.17, p = 0.30, 95% CI = -0.15—0.45. However, within the OCD subgroup, compulsivity demonstrated a large correlation with the Hebbian bias term, r = 0.53, p = 0.017, 95% CI = 0.11—0.78, and the sticky choice parameter representing choice perseveration, r = 0.49, p = 0.03, 95% CI = 0.06—0.76.

Conclusions: This study provided preliminary evidence challenging longstanding assumptions regarding the cognitive processes involved in reward learning and conceptualizations of neurocognitive mechanisms of compulsivity. Contrary to understanding of the contributions of model-free RL to reward learning, a Hebbian or habit-like process that did not rely on RPEs or integrate outcome valance was found to integrate with working memory to account for behavior on a simple experimental task. A computational parameter quantifying this Hebbian process was associated with compulsivity within the OCD subgroup. Findings highlight outcome insensitive associative learning as a potential novel mechanism of compulsivity, offering inroads for treatment personalization if replicated.

Keywords: OCD, Computational Methods, Compulsivitiy

Disclosure: Nothing to disclose.

P655. Relationships Between Kappa Opioid Receptor Availability, Hoarding Behaviors, and Suicide Ideation in Psychiatric Subjects: Results From an in vivo Pilot Study Using [11C]EKAP PET

Emily Weiss*, Victoria Hart-Derrick, Christopher Pittenger, Mia Weed, Nabeel Nabulsi, David Matuskey, Margaret Davis

Yale University School of Medicine, New Haven, Connecticut, United States

Background: Hoarding disorder (HD) is a debilitating psychiatric condition characterized by difficulties discarding possessions and excessive acquisition of new objects, resulting in extreme clutter. Recent studies have highlighted risks of impulsive suicide attempts (SA) in HD, reporting SA rates of 24%, and an emerging neurobiological model of hoarding points to the critical role of emotion dysregulation in hoarding pathology. The endogenous opioid system has been virtually unexplored in HD, despite its relevance to emotion dysregulation, obsessive compulsive disorder, and compulsive behavior. Additionally, pre-clinical work has demonstrated that opioid antagonists decrease food hoarding behaviors. Similarly, the opioid system is implicated in impulsivity and suicide risk. Specifically, kappa opioid receptors (KOR) are implicated in stress, depressed mood, and cognition. HD is characterized by impulsivity, self-control difficulties, co-morbid depression, exposure to early life stress, and executive dysfunction. Using [11C]EKAP positron emission tomography (PET), this study presents pilot data from an in-vivo exploration of KOR availability in psychiatric subjects with hoarding behaviors.

Methods: Secondary data analyses were performed using a sample of 12 psychiatric subjects (primary diagnoses of borderline personality disorder or posttraumatic stress disorder) with and without hoarding behaviors. Participants were 30.08 years old on average (SD = 9.09) and 83.30% were biologically female. Subjects participated in an [11C]EKAP PET scan and comprehensive clinical assessment. Volume of distribution (VT: ratio of parent radioligand concentration in tissue relative to that in blood) in grey matter regions was the outcome measure; VT values were available for 9 subjects. The regions of interest (ROI) included the cingulo-opercular network (anterior-cingulate cortex [ACC] and insula), involved in the affective aspects of decision-making and implicated in HD. Hoarding behaviors were measured by the Saving Inventory-Revised (SI-R), a scale assessing three core HD symptoms: difficulties discarding, clutter, and excessive acquisition. Current suicide ideation (SI) was assessed by the suicide item on the Beck Depression Inventory (BDI-II). Analyses included independent samples t-tests (group differences) and Fisher’s exact test (FET; sex differences).

Results: In total, 9% of the sample scored within the sub-clinical range for the full SI-R ( > 24). Additionally, 67%, 42%, and 25% scored in subclinical range for the Discarding, Acquiring, and Clutter subscales, respectively. KOR availability in the cingulo-opercular regions differed between subjects with and without hoarding behaviors. Specifically, KOR availability was 38.9% lower in subjects scoring within the sub-clinical range on the SI-R (n = 3) relative to those scoring in the non-clinical range (t(5.34) = -5.00, p = 0.003, d = 2.50). There were no significant age (p = 0.18) or biological sex differences (FET = 0.11) between subjects with sub-clinical SI-R scores relative to those with non-clinical scores. Similarly, 32.8% lower KOR availability was observed for subjects scoring above the clinical cut-off on the Discarding subscale (n = 2) relative to those scoring below the cut-off (t(6.02) = -3.58, p = 0.012, d = 1.60). No significant age (p = 0.35) or sex differences (FET = 0.49) in difficulties discarding were observed. Finally, we examined hoarding behaviors in subjects with and without current SI. Independent samples t-tests demonstrated that subjects reporting current thoughts of suicide scored significantly higher on the Discarding (t(10) = 2.40, p = 0.038, d = 1.60) and Acquiring (t(10) = 2.40, p =.041, d = 1.44) subscales, with large effect sizes. Effect sizes for differences on the Clutter subscale were also large, however, differences did not reach statistical significance (t(7.89) = 2.35, p = 0.130, d = 2.42). No significant age (p = 0.34) or sex differences (FET = 0.46) were observed for subjects reporting SI relative to those without SI.

Conclusions: The findings show preliminary evidence for the role of KOR dysfunction in hoarding behaviors. Specifically, in cingulo-opercular regions implicated in HD, we found 33-39% lower KOR availability in psychiatric subjects with hoarding symptoms relative to psychiatric subjects without these symptoms. Relatedly, we observed preliminary evidence for relationships between hoarding behaviors and suicide ideation. Subjects reporting current SI endorsed significantly more difficulties with discarding and excessive acquisition, and effect sizes were large. Although differences in clutter levels between subjects with and without SI did not reach statistical significance, observed effect sizes were large. It is possible that these differences would reach significance with a larger sample. Certain limitations must be considered; subjects were not formally diagnosed with DSM-5 HD. Additionally, due to the small sample size and the pilot nature of this data, results must be regarded as preliminary. Finally, although we did not observe sex or age differences, other demographic variables may have affected the findings and should be considered in future work. Overall, the preliminary results highlight the relevance of the endogenous opioid system in hoarding behaviors, as well as links between hoarding and thoughts of suicide, providing a foundation for further exploration of HD’s pathophysiology.

Keywords: Hoarding Disorder, kappa opioid receptor, Suicide risk factors

Disclosure: Nothing to disclose.

P656. Acute Stress Effects on Repetitive Behaviours and Neural Mechanisms

Aidan Price, Isabel Chew, Jacqueline Iredale, Chris Dayas, Erin Campbell, Elizabeth Manning*

The University of Newcastle, Australia, Callaghan, Australia

Background: Acute stress is known to exacerbate the symptoms of neuropsychiatric disorders associated with uncontrollable repetitive behaviours, including Tourette Syndrome (TS) and obsessive compulsive disorder (OCD). Much research around the role of stress in these disorders has focussed on stress hormones, specifically those involved in the hypothalamic pituitary adrenal (HPA) axis such as cortisol. However stress hormone changes fail to explain many aspects of stress effects on repetitive behaviours. Recently, a new circuit was identified linking hypothalamic stress sensitive neurons that initiate the HPA axis, and a nucleus in the indirect pathway of the basal ganglia, which is involved in suppression of actions, however the functional role of this circuit is unknown. We hypothesize that synaptic activity in this pathway may mediate stress-induced exacerbation of inflexible behaviour in psychiatric disorders, and reflect a new treatment target for TS and OCD.

Methods: To examine this circuit, cell-type and pathway specific optogenetic activation and inhibition was performed during baseline and stress sessions. A mix of male and female transgenic mice expressing cre-recombinase in corticotrophin releasing hormone (CRH) neurons were used (n = 24). Optogenetics was used to examine the circuit connecting these neurons in the paraventricular nucleus (PVN) of the hypothalamus to the globus pallidus externa (GPe) in the indirect pathway, by expressing either the excitatory opsin ChR2 or inhibitory opsin eOPN3 in PVN-CRH neurons and implanting bilateral fiber optic probes above the GPe. Changes in active stress coping (rearing) and passive stress coping (grooming) were manually scored during optogenetic stimulation.

In separate studies aimed at examining the effects of stress in a mouse model relevant to OCD and TS, male and female Sapap3 knockout (KO) mice (n = 13) and wildtype littermate controls (n = 14) were exposed to an acute acoustic stress (5 minutes). Grooming was manually scored before, during and after the acoustic startle stress.

Results: Optogenetic activation of the PVN-CRH- > GPe pathway induces grooming in mice, which is similar to what is observed following an acute stress (laser x virus interaction p = 0.01, posthoc comparison with Šidák correction during laser period Chr2 vs control p = 0.001). Preliminary findings suggest that inhibition of this PVN-CRH- > GPe pathway suppresses repetitive behavioural responses to stress (grooming; 0.076), and instead mice shift to a more active stress coping response with increased rearing in eOPN3 expressing mice.

When comparing grooming in Sapap3-KO mice and controls before and during startle stress, as expected there was a main effect of genotype with increased grooming in KO mice (p = 0.009). There was also a trend for a stress effect (p = 0.10). However, there was significant heterogeneity in the severity of grooming phenotype in Sapap3-KO mice at baseline, consistent with our previous studies. Taking a median split of the data, and comparing stress effects in high vs low baseline grooming Sapap3-KO mice, stress increased grooming in Sapap3-KO mice with low grooming at baseline, but decreased grooming in Sapap3 KO mice with high baseline grooming, whereas stress had no significant effects on control mice (1-WAY ANOVA, p = 0.01, low vs high baseline grooming Sapap3-KO, change in grooming between baseline and during startle stress, p = 0.0078).

Conclusions: This work adds to a growing literature demonstrating the functions of synaptic projections of PVN-CRH neurons beyond their role in the HPA axis, that have important implications for how stress contributes to the pathophysiology and symptoms of neuropsychiatric disorders. Further examination of this population in the Sapap3-KO model may support a role in stress-induced repetitive behaviours relevant to TS and OCD. Future work identifying strategies to target these pathways may help guide the development of new treatment strategies that enhance control over symptoms when patients are exposed to stress.

Keywords: Obsessive Compulsive Disorder, Tourette syndrome, Acute Stress, Corticotropin-Releasing Hormone, Basal Ganglia

Disclosure: Nothing to disclose.

P657. The Wicked and Combined Role of Infections, Inflammation, and Vitamins in OCD

Donatella Marazziti*

Donatella Marazziti, Pisa, Italy

Background: Recent evidence highlights that different agents may trigger immunemediated processes involved in the pathophysiology of different neuropsychiatric conditions. Given the limited information on obsessive-compulsive disorder (OCD), the present study aimed at assessing current/past infections and plasma evels of vitamin D, vitamin B12, folic acid, homocysteine, antistreptoslysin (ASO) titer and common peripheral inflammatory markers in a group of OCD outpatient

Methods: The sample included 255 adult outpatients of both sexes (120 men and 130 women, mean age:) with an OCD diagnosis according to the DSM-5 criteria. The Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) was used to assess the clinical phenotype and symptom severity. Laboratory blood tests measured levels of vitamin D, vitamin B12, folic acid, homocysteine, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), blood count and antibodies titers for cytomegalovirus (CMV), Epstein Barr virus (EBV), Toxoplasma gondii and ASO titer.

Results: Eighty-five patients had a previous EBV infection, 55 were seropositive for CMV IgG, 25 showed positive antistreptolysin titer, 14 were seropositive for Toxoplasma gondii IgG, and four for CMV IgM. More than a half of patients showed vitamin D insufficiency. Seventy-five patients resulted ASO positive with no recall of a previous streptococcal infection Compared to seronegative patients, patients with a past EBV infection displayed significantly higher scores on the Y-BOCS total score and compulsion subscale, and other symptoms. Vitamin D was negatively correlated with both the Y-BOCS total (r = -0.254, p = 0.002) and the compulsions (r = -0.210, p = 0.012) subscale scores. Folic acid was negatively correlated with the Y-BOCS total (r = -0.230, p = 0.046) and obsessions (r = -0.226, p = 0.049) subscale scores. In the ASO titer-positive patients, the white cell count was negatively related to “avoidance” (r = -.394, p = 0.026) and “overvalued sense of responsibility” (r = -.371, p = 0.037) item scores, and directly to neutrophil counts (Rho = 0.485, p = 0.007). Forty patients of this group who were on medication for OCD showed significantly lower ASO titers than the other (Z = -2.359, p = 0.018).

Conclusions: The findings of our study show an association between Epstein Barr infection, hypovitaminosis D, decreased levels of folic acid and the overall severity and specific symptom patterns of OCD. The laboratory measures used in this study are useful, cheap and easy parameters that should be routinely assessed in patients with OCD. Together with the results in OCD subjects with ASO titer-positivity, they suggest the pesence a chronic inflammatory state, in spite of no symptoms or recall of previous infections, that might be involved in both the onset and the maintenance of OCD, with immunological alterations being related to symptom dimension to be identified. They also support the notion of possible anti-inflammatory effects of some psychotropic compounds and the potential therapeutic impact of vitamins and antibiotics/immunomodulatoryagents in OCD.

Keywords: OCD, infection, vitamin D, Folic acid

Disclosure: Nothing to disclose.

P658. Habenula Connectivity Related to Loss and Avoidance in Obsessive-Compulsive Disorder

Qian Liu*, Feng Gao, Xiang Wang, Lucas Trambaiolli, Hayden Peel, Changlian Tan, Chuting Li, Jie Fan, Xiongzhao Zhu, Jamie Feusner

Medical Psychological Center, The Second Xiangya Hospital, Central South University, Changsha, Hunan, China; Medical Psychological Institute of Central South University, Changsha, Hunan, China ; National Clinical Research Center for Mental Disorders (Xiangya), Changsha, Hunan, China; National Center for Mental Disorder, Changsha, China, Toronto, Canada

Background: High sensitivity to potential loss is a common characteristic of those with obsessive-compulsive disorder (OCD) and may be of special importance to understanding the clinical symptoms. Convergent human neuroimaging and animal studies suggest that the habenula is a key center for encoding the aversive experience of loss. The habenula and its functional connections may be involved in different loss experience phases; specifically, there is evidence that the right habenula-right hippocampus and the right habenula-subcallosal cingulate connections are associated with loss anticipation and that the striatum-pallidum-habenula pathway is associated with monetary loss. White matter tract tracing data in macaques, and previous human studies, indicate that the habenula receives major afferent input from the pallidum. Further, from animal studies, the habenula modulates the ventral tegmental area (VTA), controlling dopamine release, which promotes motivated behavior to avoid loss. This evidence of the habenula’s involvement with loss, and the phenomenology of high sensitivity to loss in those with OCD, suggests the possibility that individuals with OCD may exhibit abnormal habenula neural connectivity during loss experiences; however, this has yet to be tested. Accordingly, this study examined functional and effective connectivity in the habenula during loss experiences in those with OCD compared with healthy controls (HC). We hypothesized that: 1) individuals with OCD would show differences from HC in functional connectivity of the right habenula-right hippocampus and the right habenula-subcallosal cingulate for the contrast of the loss anticipation vs. neutral anticipation; 2) individuals with OCD would demonstrate differences in functional connectivity of the right habenula-right striatum during the contrast of monetary loss vs. neutral outcomes compared to HC; 3) individuals with OCD would show differences in effective connectivity from the right pallidum to the right habenula during the contrast of monetary loss vs. neutral outcomes compared to HC; 4) individuals with OCD would show greater negative effective connectivity from the right habenula to the right VTA during loss avoidance compared to HC; 5) individuals with OCD with more negative habenula-VTA effective connectivity during loss avoidance would demonstrate faster response time, greater experiential avoidance, and higher sensitivity to punishment. Correlations between effective connectivity of the habenula-VTA and Y-BOCS compulsion scores were also explored. Hypotheses were preregistered: https://aspredicted.org/CHD_PSH.

Methods: Forty HCs and 36 participants with OCD were recruited. All participants completed clinical assessments, and functional magnetic resonance imaging data was acquired during the Monetary Incentive Delay (MID) task. Behavioral performance during different conditions was compared between the HC and OCD groups. Psychophysiological interaction analyses and dynamic effective connectivity were applied to test functional connectivity and effective connectivity hypotheses, respectively. Pearson correlation was used to investigate relationships among behavioral performance, functional connectivity of Hb, effective connectivity of habenula, and obsessive-compulsive symptoms in OCD.

Results: Individuals with OCD had greater experiential avoidance than HC on the Acceptance and Action Questionnaire II. There were no significant group differences in response times during loss trials. Those with OCD compared with HC showed decreased functional connectivity between right habenula and left hippocampus during loss anticipation and increased functional connectivity of right habenula-bilateral pallidum and right habenula-left caudate for monetary loss vs. neutral outcomes. There were no significant differences between groups in mean dynamic effective connectivity from right pallidum to the right habenula. There was lower mean negative dynamic effective connectivity (less inhibition) from right habenula to right VTA (F = 4.40, p = 0.04, η²=0.058) and from right habenula to left VTA (F = 20.16, p < 0.001, η²=0.221) during loss avoidance in the OCD group compared to HCs. In the OCD group, there was a negative correlation between experiential avoidance and negative effective connectivity from the right habenula to the left VTA during loss avoidance (r = -0.383, p = 0.021).

Conclusions: During monetary loss, individuals with OCD exhibited higher functional connectivity within the striatum-pallidum-habenula circuit compared to HCs. Although there was abnormal connectivity of the right habenula with ipsilateral VTA during loss avoidance in OCD, group differences were more pronounced between the right habenula and contralateral VTA. Further, while we predicted greater negative connectivity (greater inhibition), results revealed less inhibition from right habenula to the contralateral VTA in OCD, which was associated with experiential avoidance. Less inhibition of the VTA during loss avoidance could potentially lead to increased dopamine release, reinforcing avoidance behaviors in OCD. In summary, the habenula may be a key node associated with loss experience in OCD. This study suggests that dysfunction within the striatum-pallidum-habenula circuit, and habenula-VTA connectivity, may be contributing pathophysiological mechanistic features underlying loss experiences in those with OCD.

Keywords: Obsessive-Compulsive Eisorder (OCD), dynamic effective connectivity, monetary incentive delay task, habenula, Functional MRI (fMRI)

Disclosure: Nothing to disclose.

P659. Medial Orbitofrontal Cortex Representation of Active Avoidance and Refinement Over Learning

Brittany Chamberlain*, Matthew Geramita, Elizabeth Crummy, Vijay Mohan K Namboodiri, Susanne Ahmari

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Negative reinforcement-based theories of obsessive compulsive disorder (OCD) suggest that the manifestation and maintenance of compulsions is driven by temporary relief from obsession-related anxiety, but the underlying mechanisms are largely unknown. Medial orbitofrontal cortex (mOFC) has emerged as a critical region underlying excessive avoidance in OCD, as patients with OCD exhibit heightened mOFC activity when avoiding compulsion-related images. Importantly, decreasing mOFC activity using transcranial magnetic stimulation reduces avoidance behavior and compulsive urges, highlighting its promise as a therapeutic target. However, how mOFC representations of avoidance emerge and are refined over learning remains unknown. To investigate the cellular-level mechanisms underlying active avoidance learning, we measured single-cell mOFC activity via single-photon calcium imaging in mice during a novel instrumental avoidance task.

Methods: C57Bl6/J mice (n = 18; 10M, 8F) were trained to avoid foot shocks predicted by a cue light by lever-pressing within a 20sec cued period (completing an ‘avoid response’). If mice did not press during the cued period, the cue light co-terminated with delivery of the first in a series of up to five foot shocks. Mice could escape remaining foot shocks by lever pressing during this period (completing an ‘escape response’). Sessions consisted of 50 trials/day, and successful acquisition was defined by ≥80% avoid responses by session 7. Prior to training, mOFC was injected with AAV5-CaMKII-GCaMP6f and implanted with 500µm diameter GRIN lens to visualize activity using miniature microscopes (Inscopix). Single-cell calcium fluorescence was extracted (CNMFe) and time-locked to task-relevant events. Individual neurons were tracked across sessions (CellReg).

Results: Mice learned to avoid foot shocks over 7 sessions with 14/18 meeting acquisition criterion by Session 3, suggesting that most learning occurs within the first 150 trials. We observed increased mOFC representation of threat cue over learning [Session 1 threat cue onset mean AUC = 0.2258 vs. Session 3 threat cue onset mean AUC = 0.4033; unpaired t-test, t(2465) = 2.318, p = 0.0205] that was driven by an increase in the proportion of neurons selectively activated by threat cue onset [RM one-way ANOVA Sessions 1-3: main effect of session, F(1.899, 32.28) = 3.781, p = 0.0355; Dunnett’s multiple comparisons test: Session 1 mean = 8% (106/1183) vs. Session 3 mean = 12% (165/1284), p = 0.0175].

Given that mice transition from escape to avoid responses over learning, we next investigated whether discrete or overlapping populations of mOFC neurons are modulated by these behaviors. We found that only 3% of neurons were activated by both avoid and escape responses, while 12% and 14% of neurons were selectively activated by only avoid or escape responses, respectively, indicating discrete mOFC representation of these defensive behaviors. Additionally, we investigated how mOFC encoding of avoidance evolves over each subsequent training day, and we found a refinement of the avoidance ensemble, with a progressive increase in both the proportion of neurons that are avoid-activated and avoid-inhibited over time [Avoid-activated: mixed-effects analysis: main effect of session, F(3.726, 59) = 4.807, p = 0.0025; Dunnett’s multiple comparisons test: Session 1 mean = 12.9% vs. Session 7 mean = 21.99%, p = 0.0073; Avoid-inhibited: mixed-effects analysis: main effect of session, F(4.121, 65.25) = 5.00, p = 0.0013; Dunnett’s multiple comparisons test: Session 1 mean = 3.4% vs. Session 7 mean 10%. p = 0.0059]. Furthermore, longitudinal tracking of single neurons revealed increased stability of the avoidance ensemble over time, with increasing likelihood that avoid-activated neurons are maintained from session to session [Mixed-effects analysis, main effect of session, F(3.357, 44.98) = 2.915, p = 0.0392; Dunnett’s multiple comparisons test: Session 2 percent maintained = 25.37% vs. Session 7 percent maintained = 48.23%, p = 0.0182].

Finally, we interrogated the necessity of mOFC activity for active avoidance learning by lesioning mOFC bilaterally. mOFC-lesioned mice showed impaired avoidance learning, with an overall acquisition rate of 25% in the NMDA group compared to 83% in the saline control group (paired t-test of acquisition rate across sessions, t(6) = 4.307, p = 0.0068), supporting a causal role for the mOFC in active avoidance learning in this novel task.

Conclusions: These data show robust representation of threat-predictive cues in mOFC that evolve over avoidance learning. Initial representation of escape and avoid behaviors is largely separable, indicating discrete neural representation of these context-dependent defensive behaviors. We observe increased recruitment of avoid-modulated neurons and increased stability of avoidance encoding in late – relative to early – sessions that may reflect consolidation of a well-learned behavior following initial learning. Finally, lesion data support a previously undescribed causal role for the mOFC in active avoidance learning.

Keywords: Active Avoidance, OCD, Aversive Learning

Disclosure: Nothing to disclose.

P660. A Novel, Non-Hallucinogenic Psychedelic for the Treatment of Obsessive-Compulsive Disorder

Bernard Lerer*, Peretz Golding, Michal Brownstien, Alan Kozikowski, Tzuri Lifschytz

Hadassa Medical Center, Hebrew University, Jerusalem, Israel, Jerusalem, Israel

Background: OCD is a common disorder in which the sufferer experiences repeated obsessive thoughts and the need to perform compulsive behaviors. Self-grooming disorders such as skin picking and trichotillomania are closely related to OCD. The most widely accepted treatments are cognitive-behavioral therapy and selective serotonin reuptake inhibitors (SSRIs). At least a third of OCD patients do not have an adequate response to these treatments. In an open-label study (Moreno et al. 2006) and several case reports, the serotonergic psychedelic psilocybin was reported to alleviate OCD. We have designed and synthesized a novel psychedelic compound (HBL20017) that is characterized by high functional efficacy at the 5-HT2A and 5-HT1A receptors and moderate efficacy at the 5-HT2B receptor. We assessed its hallucinogenic properties using the mouse head twitch response (HTR) and tested its potential as an anti-OCD treatment on the marble burying test (MBT). We further tested its therapeutic potential for OCD in mice homozygous for deletion of the Synapse Associated Protein 3 (SAPAP3) gene. These mice manifest a phenotype that is characterized by severe, potentially lethal, obsessive self-grooming and also head body twitches, which are regarded as a model for tic disorders in humans. This same gene is associated with OCD in humans. The obsessive grooming of SAPAP3-/- mice is alleviated by fluoxetine and, in a recent study from our group, by the serotonergic psychedelic psilocybin (Brownstien et al, 2024).

Methods: HBL compounds were tested for functional activity at 5-HT1A, 5-HT2A, 5-HT2B and 5-HT2C receptors by calcium mobilization in HEK cells expressing the receptors. Male C57bl/6j mice were used for HTR studies, and ICR mice were used for the MBT. Mice that carry the SAPAP3 deletion were provided by Dr. G. Feng (MIT), and a breeding colony was established. C57bl/6j and ICR mice were 10-11 weeks of age at the time of behavioral testing. SAPAP3 KO mice were at least 6 months old since the over-grooming phenotype manifests most strongly from this age. HBL20017 and its hallucinogenic analogue, HBL20016, were administered at doses of 1, 6 or 10 mg/kg by i.p. injection. HTR was determined in a magnetometer for 30 minutes immediately after drug injection using C57bl/6j mice with ear-tag magnets. MBT was performed 30 minutes after drug administration to ICR mice, and the number of marbles buried over a further 30 minutes was determined. Self-grooming and head-body twitches were determined in SAPAP3 mice over one hour by an observer blind to the treatment status of the mice at baseline and then 2, 12, 21, 28, and 42 days after treatment.

Results: HBL20017 showed potencies (ED50) of 1.84nM at the 5-HT1A receptor; 7.95nM at the 5-HT2A receptor; 17.1 nM at the 5-HT2B receptor; and, 0.80 nM at the 5-HT2C receptor on calcium mobilization assays in HEK cells. The potencies of HBL20016 were 645nM at 5-HT1A; 1.64 nM at 5-HT2A; 3.42 nM at 5-HT2B; and, 8.37 nM at 5-HT2C. In vitro biological assays showed no indication of cytotoxicity or hERG inhibition. Unlike HBL20016, which induced significant HTR in C57Bl/6j mice, HBL20017 did not induce HTR, suggesting it is non-hallucinogenic. On the marble-burying test, a putative screening test for anti-OCD activity, HBL20017 significantly reduced the number of marbles buried over 30 minutes compared to vehicle control. When administered to 6-month-old SAPAP3-/- mice that manifested severe obsessional grooming, a single dose of HBL20017 significantly attenuated the increase in self-grooming observed in SAPAP3-/- mice treated with vehicle. The ability of a single dose of HBL20017 to alleviate obsessive self-grooming was seen 48 hours after treatment and persisted at 12, 21, and 28 days after treatment. The effect still persisted 42 days after a single dose. No mice in the HBL20017 group developed grooming-associated skin lesions during the post-dose period of observation. The same dose of HBL20016 also attenuated excessive self-grooming but to a lesser extent than HBL20017. Head body twitches were also alleviated by the two HBL compounds, to a greater extent by HBL20017.

Conclusions: HBL20017, which has shown a positive safety profile on initial evaluation and did not induce HTR, yielded positive results on two tests for anti-obsessional potential – the marble burying test and the SAPAP3 obsessional self-grooming model. These findings suggest that HBL20017 has the potential to be a non-hallucinogenic psychedelic treatment for OCD. Reduction of excessive self-grooming was already evident 48 hours after administration of HBL20017 to SAPAP3 KO mice and persisted up to 42 days after a single dose. Comparison with the results of our previous research (Brownstien et al, 2024) suggests that HBL20017 may be more effective than psilocybin in reducing excessive self-grooming over the long-term. HBL20017 also alleviated head-body twitches suggesting it may have potential in the treatment of tic disorders such as Tourette’s Syndrome. The pharmacological properties of HBL20016, which induced HTR but also attenuated obsessional self-grooming in SAPAP3 KO mice, are of considerable interest in further understanding the therapeutic mechanism of HBL20017.

This work was supported in part by Negev Labs. SAPAP3-/- model development was supported in part by ParowBio.

Keywords: Obsessive Compulsive Disorder, Tic disorders, Psychedelics, Head twitch response, SAPAP3 KO mice

Disclosure: Yes, I (or my spouse/partner) do have a financial relationship to disclose.

Financial Relationships Details Parow Entheobiosciences, Contracted Research, Self, Negev Labs, Grant, Self

P661. Event-Related Potentials Correlate with Magnetic Resonance Spectroscopy Measures of Excitatory Neurotransmission and Symptom Severity in Obsessive-Compulsive Disorder

Ana Maria Frota Lisboa Pereira de Souza*, Marjan Biria, Paula Banca, Naomi Fineberg, Trevor Robbins

University of Hertfordshire, UK, Hatfield, United Kingdom

Background: Obsessive-Compulsive Disorder (OCD), a highly debilitating condition affecting circa 1-3% of the population, is linked to dysfunction in the cortico-striato-thalamo-cortical (CSTC) circuitry. Two brain regions, the anterior cingulate cortex (ACC) and the supplementary motor area (SMA), have been strongly implicated in key OCD-related cognitive and motor deficits. Brain imaging and electroencephalographic (EEG) findings converge in revealing an overactivation of these regions in OCD. For example, magnetic resonance spectroscopy (MRS) shows increased levels of the excitatory neurotransmitter glutamate in these regions, which correlate with OCD symptomatology and overreliance on habitual behaviour (Biria et al., 2023). Enhanced amplitudes of event-related potentials (ERPs), including the error-related negativity (ERN), representing error-monitoring and generated in the ACC, and the readiness potential (RP), representing motor preparedness and generated in the SMA, have also been found in OCD (Pereira de Souza, 2020). Nevertheless, the precise relationship between changes in excitatory and inhibitory neurotransmission (glutamate-Gamma-Amino-Butyric Acid (GABA) ratio) and EEG markers in OCD is not yet known. We aimed to investigate the relationship between increased glutamate levels in the ACC and SMA with the amplitudes of the ERN and the RP, respectively.

Methods: Twenty-one individuals with DSM-5 OCD and 22 matched healthy volunteers, male and female, selected from a larger sample reported by Biria et al (2023) underwent 128-channel EEG alongside 7-Tesla proton MRS (measuring glutamate and GABA) of the ACC and SMA. During the EEG, participants performed a combined Stop-Signal/Go-No Go task, to measure the amplitudes of the ERN and the RP. Participants also completed a battery of clinical and self-report questionnaires that investigated OCD symptomatology and habitual behaviour. Between-group t tests and correlational analyses were employed to assess the differences between patients and controls and the associations between variables.

Results: Independent-sample t tests indicated ERN amplitudes were increased in OCD vs controls (p = 0.013, rrb=0.442), but RP amplitudes were not (p = 0.388, rrb=0.156). Correlational analyses indicated significant associations between the amplitudes of the ERN and the RP and the glutamate-GABA ratio in the ACC (r = -0.32, p = 0.03, z = -0.33), and SMA (r = 0.34, p = 0.027, z = 0.36), respectively. Furthermore, higher self-reported symptomatology as measured by the Obsessive-Compulsive Inventory (OCI) correlated significantly with both the increased ERN amplitudes (r = -0.3, p = 0.04, z = -0.33), and the glutamate-GABA ratio in the ACC (r = 0.33, p = 0.03, z = 0.35).

Conclusions: Perhaps the first of its kind, this study utilised two high resolution brain imaging techniques, and showed an association between the amplitudes of the ERN and the RP and the glutamate-GABA ratios in the ACC and SMA, respectively, representing increased excitatory over inhibitory neurotransmission in these critical OCD-related brain regions. Moreover, the amplitude of the ERN correlated with OCD symptom severity. These findings suggest that specific EEG markers involving ERPs, namely the ERN and possibly the RP, hold potential as markers of increased excitatory neurotransmission in OCD-related corticostriatal nodes. Furthermore, ERPs may have utility as affordable and well tolerated proxy biomarkers of MRS neurotransmitter measures of OCD symptomatology. Further research is indicated to corroborate and strengthen these findings in larger datasets.

Keywords: obsessive-compulsive disorder (OCD), Biomarker, Human Neuroimaging

Disclosure: Nothing to disclose.

P662. Characterizing Insula Functional Alterations in Obsessive-Compulsive Disorder From Parcellation to Gradient Model

Xiaoqi Huang*, Lingxiao Cao, Hailong Li, Jiaxin Jiang, Bin Li, Qiyong Gong

West China Hospital of Sichuan University, Chengdu, China

Background: The insula is a heterogeneously connected brain region that typically acts as an integration hub involved in a range of functions. Prior studies on insula functional connectivity (FC) in obsessive-compulsive disorder (OCD) have typically treated it as a unitary structure, disregarding any diversity in the connectional architecture across the topography of the insula.Moreover, past neuroimaging studies in OCD have always been limited by small sample sizes and potential confounding factors such as medication effects and comorbidities. In current study, we examined insula functional connectivity architecture in a relatively large and well-characterized sample of medication-free, non-comorbid OCD patients, relative to matched control subjects.

Given previous reports of subregion-specific functionality and connectivity, we speculated that exaggerated anterior insula connectivity could account for heightened cognitive and emotional responses, whereas the posterior insula hypoconnectivity for difficulties in sensorimotor integration in OCD patients. Considering the known association between insula connectional diversity and cognitive-affective function, we cautiously hypothesized that OCD patients would show reduced insula connectional diversity. Additionally, we expected more perturbed insula functional connectivity to correlate with more severe clinical symptoms in OCD patients.

Methods: The study was approved by the Research Ethics Committee of West China Hospital, Sichuan University. All participants provided written informed consent after receiving a complete description of the study. We recruited 93 medication-free OCD patients without any comorbid psychiatric disorders. The diagnosis of OCD was confirmed by consensus between two licensed psychiatrists using the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID). We also recruited 93 healthy controls (HC), matched for average age and sex, from the local community through advertising posters. They were screened with the SCID (non-patient edition) to confirm the absence of a current or history of psychiatric disorders, and reported no known history of psychiatric disorders in their first-degree relatives. All participants were native Chinese speakers and right-handed.

Anatomical and resting-state functional brain images were acquired for each subject on a 3-Tesla GE Signa EXCITE MRI scanner with an 8-channel phase-array head coil. We application data-driven method to cluster insula voxels into subregions based on similarity in their functional connectivity profiles, as well as to model continuous variation in connectivity among insula voxels using the recently proposed concept of the gradient.

Group analyses were fulfilled using SPM12 that treated subjects as a random variable in a two-by-two flexible factorial ANOVA, with insula subregion as a within-subject factor and diagnosis as a between-subject factor. Age, sex and mean FD value were entered as covariates to control for their possible confounding effects. The global-level gradient distribution of the insula was compared between OCD patients and HC using two-sample Kolmogorov-Smirnov tests to examine the cumulative distribution rather than the mean value of groups. Voxel-level between-group differences in the gradient scores were determined by a two-sample t-test after controlling for age, sex and FD, with the threshold set at p < 0.05 with FDR correction.

Results: We first parcellated the insula into two subregions corresponding to anterior and posterior insula using a data driven clustering approach, and evaluated group differences with a diagnosis-by-subregion flexible factorial ANOVA. Compared to controls, OCD patients exhibited unbalanced insula subregional FC alterations with several brain regions, including the supplementary motor area, precentral gyrus and dorsolateral prefrontal cortex, demonstrated as increased FC strength in the anterior subregion while decreased FC strength in the posterior subregion. We then modeled connectivity variation among insula voxels continuously and generated the principal anterior-posterior gradient. We found global topographic alterations in this gradient in OCD patients, including reduced explanation ratio, gradient range and gradient variation.

Conclusions: We found (1)unbalanced insula subregional FC alterations in OCD patients, with increased FC strength in the anterior subregion while decreased FC strength in the posterior subregion which is more prominent in right insular; (2)OCD patients exhibited less variance in global topographic alterations in the principal anterior-posterior gradient, specifically in the left insula. These findings highlight the specific role of left insular and anterior subdivision underline the neuropathology of OCD.

Keywords: Resting State Functional Connectivity, Connectivity gradients, Obsessive Compulsive Disorder, insular cortex

Disclosure: Nothing to disclose.

P663. Update on Health and Psychosocial Outcomes With Co-Occurring Disorders

Tanya Alim*, Amina Richardson, Narayan Rai, Maria Hipolito, Najiy Shabazz, Janet Bolomope, Alicia Anderson, Michael Serlin, Flora Terrell-Hamilton, William Lawson, Robert Drake, Richard Schottenfeld, Evaristus Nwulia

Howard University, White Plains, Maryland, United States

Background: The Howard University Minority AIDS Network Effort (HUMANE) Program (initially presented 2021) aims to address major gaps in services provided by peer support, case management, and HIV/Hepatitis prevention at three sites – an outpatient mental health services clinic, a primary care clinic, and an infectious disease clinic. During the pandemic a homeless shelter community program was added. The HUMANE program’s overarching goal is to improve the health and well-being of persons with serious mental illness (SMI) and co-occurring disorders (COD), who are living with or at high risk for HIV and/or Hepatitis. One of the aims of the program is to deliver Integrated Dual Disorders Treatment (IDDT) to clients. The IDDT model is an evidenced-based practice that improves the quality of life for people with co-occurring SMI and COD by combining substance abuse services with mental health services.

Methods: Between July 2019 and November 2022, a sample of 184 participants, aged 18–75-year-old, were enrolled in the HUMANE Program and given access to peer support specialists trained to conduct IDDT weekly. Those enrolled completed the Center for Mental Health (CMHS) National Outcome Measures (NOMs) Client-Level Measures for Discretionary Programs Providing Direct Services service tool at enrollment and at a 6-month reassessment.

Results: Out of the 184 clients, 98 were women and 171 were African American. Among the enrollees, 53 (28.80%) had a primary diagnosis of Major Depressive Disorder, 28 (15.22%) with bipolar disorder, 13 (7.07%) with Schizophrenia, 14 (7.61%) with anxiety, 35 (19.02%) with Opioid Use Disorder, and 6 (3.26%) with Alcohol Use Disorder. During enrollment, 184 (92.93%) participants reported having HIV test and 40 (23.39%) were HIV positive. Forty HIV+ participants were already prescribed antiretroviral therapy (ART). 164 (89.13%) participants reported having HBV test during enrollment and 2 of them were HBV+ and were connected to treatment service. Likewise, 166 (90.22%) participants had HCV test during enrollment and 8 of them were HCV+ and all were connected to treatment services. During 6-month reassessment, 99.23 % of the participants reported of having HIV test. Likewise, 98.81% participants reported of having HBV and HCV tests during their 6 months reassessment.

Of the 184 clients due for a 6-month reassessment, 130 completed the NOMs survey. All received IDDT interventions provided by trained peer support specialists. 38 (26.65%) unemployed at baseline and 21 (15.38%) were unemployed at 6 months.

Conclusions: Overall, program participants positively viewed the IDDT evidenced-based model, run by peer support specialists. At the 6-month assessment, improvements occurred in how participants viewed their overall health and quality of life. There were also decreased rates of unemployment.

Keywords: Dual Diagnosis, Integrated Dual Disorder Treatment, Alcohol and substance use disorders

Disclosure: Nothing to disclose.

P664. The Effect of High-Frequency Repetitive Olfactory Stimulation on Pain Severity in Chronic Back Pain

Muyiwa Ogunsola*, Abisola Asante, Kambi Ebo, Onyeka Nwulia, Narayan Rai, Maria Hipolito, Evaristus Nwulia

Evon Medics, Elicott City, Maryland, United States

Background: The medial temporal lobe (MTL) regions are reliably activated by high frequency olfactory stimulation. On the other hand, clinical neuroscience studies indicate that reduced MTL resting state functional connectivity (rsFC) to other cortical regions and smaller MTL volume predict transition from acute pain to chronic pain. We investigated if repetitive olfactory stimulation will reduce pain severity in patients with chronic back pain. We hypothesized that daily olfactory activations of the MTL would lead to reduced pain intensity through treatment-induced increase in the resting-state functional connectivity of the MTL regions.

Methods: Twenty-eight adult subjects (age, 20-68 years; 60% females; and 82% African Americans) with chronic back pain were randomly assigned to receive either daily low frequency intranasal stimulation with 10 odorants (N = 14) over 14 days or daily high frequency intranasal stimulation with similar odorants (N = 14) over 14 days. All participants received MRI studies using 3Tesla Siemens platform at baseline and at the 7th day of treatment. MRI studies included structural MRI and resting-state functional MRI. Pain intensity and pain interference measures were also acquired at baseline, 7th day and 14th day of treatment using validated PROMIS scales for intensity and interference. Linear mixed effect models were used to examine the effects of treatment-associated changes in rsFC on trajectory in pain intensity.

Results: Both activation and the strength of functional connectivity between the olfactory cortex and anterior hippocampus parts of the MTL and the orbitofrontal cortex were inversely related to pain intensity. High frequency repetitive olfactory stimulation paradigm results in more significant (P < 0.05) change in functional connectivity between the MTL affective and pain-related networks. Finally, high-frequency stimulation led to quicker and greater improvements in pain intensity. At day 7, 7/14 (64.3%) of high-frequency treatment recipients vs. 6/14 (42.9%) of low-frequency recipients had experienced 30% reduction in pain; and at day 14, 11/14 (78.6%) of short-burst vs. 9/14 (64.3%) of long-burst recipients had experienced 30% reduction in pain.

Conclusions: Neuroplasticity in the MTL regions relevant to pain and affect can be achieved non-invasively through targeted daily olfactory stimulation. Our study provides the first pilot evidence for use of repetitive olfactory stimulation for treatment of chronic pain in patients with chronic back pain.

Keywords: Medial temporal lobe, olfactory cortex, orbitofrontal cortex (OFC), chronic pain, MRI

Disclosure: Nothing to disclose.

P665. Unbiased Phenotyping and Systematic Analysis of Social Attachment Behaviors in Prairie Voles

Shuyu Wang*, Kara Quine, Audrey Jordan, Christoph Kirst, Devanand Manoli

UCSF, San Francisco, California, United States

Background: Throughout the animal kingdom, organisms build short- and long-lasting bonds that enable social cooperation to facilitate reproduction and survival. In humans, impairments in forming or maintaining long-lasting social relationships play a key role in many neuropsychiatric conditions. The challenge of understanding the biology behind deficits in social attachment, such as those seen in autism, schizophrenia, or trauma, stems from the lack of common model organisms that demonstrate meaningful attachment behaviors. Prairie voles are socially monogamous and form robust, life-long attachments, thus providing an invaluable opportunity to study how the brain orchestrates social attachment behavior.

Methods: Of particular interest is how a prairie vole without a preexisting mate can form a new social attachment, whereas a bonded vole does not. This suggests that prairie vole social circuitry begins in a plastic state that is sensitive to social cues and then transitions to a stable state that is optimized for long-term attachment. We are interested in using unbiased high-throughput sequencing to identify molecular factors that drive attachment.

To comprehend the broader behavioral strategies that underlie social attachment, we developed novel methods to measure social pose and to describe social dynamics. Recent advances in computational neuroethology have enabled high-fidelity, high-throughput tracking of animal pose as well as detection of recurrent behavior modules. However, existing methods are geared toward the study of individual, rather than social, behaviors, and no known platform exists to investigate the social attachment behaviors unique to prairie voles.

Results: Through single-nucleus sequencing of the medial amygdala, we identified pair bond-dependent molecular changes in multiple neuropeptide pathways that correlate with plasticity restriction. Local pharmacologic manipulation of these neuropeptide pathways altered pair bonding behavior. We are now performing molecular perturbations and circuit recordings to assess if these neuropeptide pathways are causally involved in governing social plasticity.

Our multifaceted pipeline tracks masks and user-defined skeletons of markerless voles without loss of identity over hours. Our supervised classifiers enable unbiased detection of known behaviors and our unsupervised machine learning models are uncovering dynamically coupled social behavior patterns. We generated fingerprints of recurring behavior patterns that distinguish male from female voles, pair-bonded from not pair-bonded voles, wild-type from mutant voles, and socially monogamous from socially non-monogamous vole species. By revealing such behaviors across multiple relevant timescales, we hope to lay a foundation for interpreting how neural activity drives social attachment.

Conclusions: Our experimental findings provide a molecular entry point for dissecting neural circuits relevant for social attachment behaviors. Using the computational methods we are developing, we will soon be able to compare, for the first time, reciprocal behavioral patterns mediating social attachment as well as their disruption in the context of mutations linked to neuropsychiatric illness such as autism spectrum disorder and schizophrenia. Likewise, we will be able to rigorously assess the impact of environmental disruptions on distinct components of social attachment behaviors. Our high-throughput behavior phenotyping platform may facilitate the screening of social behavior modulators and nominate treatments that reverse pathological behavior patterns. In the future, we also plan to pursue the joint analysis of behavior syllables and neural recording data, yielding hypotheses about the neural mechanisms that underlie social behavior.

Keywords: Social Behavior, social attachment, prairie voles, Computational Neuroscience, neuromodulators

Disclosure: Nothing to disclose.

P666. A Review of Evidence-Base for Nutritional Supplementation in Correctional Settings

Paige Harris, Lori Lai, Fikayomi Kumapayi, Yedishtra Naidoo, Britta Ostermeyer, Gary Chaimowitz, Andrew Olagunju*

McMaster University, Hamilton, Canada

Background: While there is evidence supporting the mental and physical health benefits of nutrition-based interventions for the general population, little is known about the impacts of such interventions in correctional settings. Individuals in such settings are at a higher risk of chronic mental and physical health conditions, which may be impacted by inadequate nutrition. This review aims to synthesize evidence from current literature to better understand nutritional needs in correctional settings and summarize any beneficial effects of dietary supplements for improving the wellbeing of individuals experiencing incarceration.

Methods: This review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline. Major databases (PubMed/MEDLINE, PsychINFO, Web of Science, and Cochrane Register) were searched using strategies developed in-line with the study objectives to identify eligible reports. Article screening and data extraction were completed by at least two independent investigators. Conflicts were resolved by discussion between investigators and, where necessary, by consulting a senior investigator. Risk of bias assessment will be conducted using the Joanna Briggs Institute (JBI) critical appraisal tool.

Results: A total of 1286 articles were identified from the search and 57 reports were selected for full-text review following title and abstract screening. Thirty-one reports were included in the final review for data extraction. The included reports were derived from different study designs, including clinical trials (n = 7). The major themes addressed in the included reports are assessments of specific nutrient levels amongst incarcerated individuals (n = 10 reports), analyses of foods served in correctional facilities (n = 14), and randomized control trials involving dietary supplements (n = 7). Deficiency studies and menu analyses were largely conducted in correctional facilities for adult males in the United States and Europe. Though menus appear to meet caloric requirements, inadequate provisions of several nutrients, such as fat-soluble vitamins, magnesium, calcium, and folate were identified. Sodium and saturated fats were often in excess, especially in foods available for purchase. Vitamin D was both the most commonly investigated and most commonly identified nutritional insufficiency amongst incarcerated individuals, and may be associated with length of prison stay, race, season, and security level. Blood levels of Omega-3 fatty acids were negatively associated with aggressive and attention deficit behaviors in incarcerated adults. Several articles noted that the lack of menu standardization across facilities limits generalization, and that nutritional inadequacies identified through menu analysis may be worse than reported, as analyses assume that all foods served are consumed. Long and short trials (duration ranged from 2 weeks to 9 months) largely examined the effect of Omega-3 and 6 fatty acids and/or multivitamins on aggression and antisocial behavior in incarcerated adult males, albeit the majority examined effects after ~3 months of supplementation. The findings showed that Omega fatty acids have a degree of benefits in mitigating aggression, violence, and antisocial behavior. The majority of results from multivitamin interventions failed to reach significance; however, a common theme was that more severely deficient individuals displayed the greatest behavioral changes.

Conclusions: Findings from this review highlights some promising benefits of Omega fatty acid supplementation in the population in incarceration. However, there is a need for further well-designed, sufficiently-powered randomized control trials to investigate the impacts of dietary supplements in the correctional setting. Given the nutritional inadequacies reported by nutrition analyses and serum studies, supplementing prison menus with vitamins, minerals, and essential fatty acids would likely improve the health of individuals in such settings. Nutrition education may also be a crucial intervention for enhancing the autonomy and wellbeing of incarcerated individuals. The findings from this review will be presented to highlight the current evidence for nutrition-based interventions in correctional settings and include specific recommendations to support the equitable and inclusive provision of adequate nutrition to improve the wellbeing of individuals in correctional settings.

Keywords: Correctional Psychiatry, nutrition, Wellbeing

Disclosure: Nothing to disclose.

P667. Mapping the Human Amygdala Through Multiscale Spatial Transcriptomics

Michael Totty*, Svitlana Bach, Madeline Valentine, Madhavi Tippani, Sarah Maguire, Ryan Miller, Ishbel Rosario, Kristen Maynard, Thomas Hyde, Stephanie Cereceo Page, Stephanie Hicks, Keri Martinowich

Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United States

Background: The amygdala (AMY) is a complex brain structure critical for emotional learning and memory. AMY dysfunction is implicated in stress-related disorders including depression, anxiety, and post-traumatic stress disorder, as well as substance use disorders. In humans, the AMY is made of numerous subnuclei, including the cortical-like basolateral complex composed of the lateral (LA), basal (BA), and accessory basal (aBA) nuclei, in addition to the striatal-like central and medial nuclei. Studies in both rodents and non-human primates demonstrate that these subregions display diverse molecular and functional specializations. While single nucleus RNA sequencing (snRNA-seq) has begun to broadly unveil the cell types found within the human AMY, spatial transcriptomic investigations are needed to link these cell types to their anatomical subregion, and by extension, their potential functional roles.

Methods: To address this gap, we employed multiple spatial transcriptomic technologies to map gene expression patterns across the entire coronal cross-section of the human AMY from both male and female neurotypical brain donors (n = 10; 5M and 5F). Tissue blocks containing the AMY were dissected from fresh-frozen coronal hemislabs of postmortem human tissue utilizing established neuroanatomical landmarks. Inclusion of the AMY was confirmed by visualization of histological stains and single molecule fluorescent in situ hybridization (smFISH). Because there is a lack of established marker genes for the human AMY, the anatomical structure of the AMY was visualized using marker genes for excitatory neurons (SLC17A7) and white matter (MBP). The Visium Spatial Gene Expression Platform (10x Genomics) was first used to comprehensively profile the AMY with full-transcriptome coverage at 55 µm resolution. AMY tissue blocks were scored into strips and 10 um cryosections were mounted onto Visium slides (6-8 capture areas per donor; n = 71 total), yielding > 300,000 total Visium spots. Spatially-informed, unsupervised clustering methods were used to detect spatial domains. Differential gene expression analyses were then used to discover novel marker genes for AMY spatial domains.

We next used the Xenium In Situ Platform (10x Genomics) to gain single cell resolution in n = 4 of the original ten human donors (2M and 2F). We mounted the entire extent of the AMY within a single Xenium capture area for each donor (n = 4 total capture areas). Using the spatial domain-specific marker genes derived from Visium, we designed a custom gene expression panel (100 gene targets), which we used in addition to the pre-designed Human Brain gene expression panel (266 gene targets; 10x Genomics) to target a total of 366 genes. In total, this resulted in > 1,000,000 cell segmentations and > 300,000,000 highly quality transcripts.

Results: Spatially-informed, unsupervised clustering recapitulated AMY subnuclei consistent with the Paxinos atlas of the human brain (Mai, et al. 2015), including further anatomical subdivisions such as the ventromedial subdivision of the BA (vmBA) and dorsal and lateral subdivisions of the LA. Pseudobulk differential expression analyses revealed novel marker genes for each subregion, including CYP26B1 (LA), RGS4 (BA), LAMP5 (vmBA), and PDYN (aBA). Integrating our previously generated snRNA-seq data, we successfully registered distinct human AMY cell types to precise anatomical subregions, which were further validated using the Xenium In Situ platform. Finally, while Visium’s low-resolution has thus far proved insufficient to reliably identify intercalated cells unique to the AMY, we identified two classes of intercalated cells using single cell segmentations in the Xenium platform. Future work will expand these analyses towards the final development of an atlas of the human AMY, focusing on both comprehensive mapping of single cell types and enrichment for gene sets associated with psychiatric disease.

Conclusions: We generated a comprehensive multiscale spatial transcriptomic atlas of the human AMY. Utilizing both Visium and Xenium platforms, our preliminary analyses revealed the molecular landscape of AMY subregions, both identifying novel marker genes and mapping single cell types to precise anatomical locations. These discoveries will prove useful for all researchers investigating the postmortem human AMY, and will likely be useful for researchers using non-human primates. The finalized atlas will be made publicly available upon publication and will be useful for both the discovery of molecular targets within defined AMY subregions, as well as cross-species comparisons for preclinical researchers. For example, researchers investigating transcriptional changes in AMY cell types in the context of human psychiatric disease can use our atlas to register their snRNA-seq datasets to determine the anatomical localization of impacted cell types. Similarly, preclinical researchers will be able to determine the cross-species conservation or divergence in the AMY subregions and cell types of their interest.

Keywords: Amygdala, spatial transcriptomics, Postmortem Human Brain Tissue, single cell omics, Atlasing

Disclosure: Nothing to disclose.

P668. A Three-Factor Diathesis-Stress Model of Commonly Comorbid Early Onset Psychiatric Disorders: Replication in a Large Sample With fMRI Data

Charlotte Caswell, Niki Hosseini-Kamkar, Sylvia Cox, Tobias Banaschewski, Gareth Barker, Arun Bokde, Rüdiger Brühl, Sylvane Desrivières, Herta Flor, Hugh Garavan, Penny Gowland, Antoine Grigis, Andreas Heinz, Hervé Lemaitre, Jean-Luc Martinot, Marie-Laure Paillère, Eric Artiges, Frauke Nees, Dimitri Papadopoulos Orfanos, Luise Poustka, Michael Smolka, Nilakshi Vaidya, Henrik Walter, Tomas Paus, Marco Leyton*

McGill University, Montreal, Canada

Background: We recently reported evidence that commonly comorbid early onset mental health outcomes were predicted with high accuracy ( > 90%) and statistical robustness (p = 2.4 x 10ˆ5) by a three-factor model composed of adolescent externalizing (EXT) traits, early life adversity, and midbrain dopamine autoreceptor availability (n = 52). Here, we investigated whether this positron emission tomography (PET) based model could be reproduced in a much larger sample using functional magnetic resonance imaging (fMRI).

Methods: IMAGEN database members with Childhood Trauma Questionnaire (CTQ) data were selected for analyses (n = 1338). Externalizing scores were calculated at ages 14 and 16 based on the Strengths and Difficulties Questionnaire. fMRI reward anticipation responses (large minus no reward, Lg-No; large minus small reward, Lg-Sm) at ages 14 and 19 were assessed using a monetary incentive delay (MID) task. Regions of interest were the ventral tegmental area (VTA), substantia nigra (SN), ventral striatum (VS), caudate, putamen, and anterior cingulate cortex (ACC). DSM-IV diagnoses were made using the Development and Wellbeing Assessment at ages 14, 16, and 19. Binary logistic regression models were constructed to determine whether a combination of EXT traits, childhood trauma, and MID fMRI contrasts at either 14 or 19 years of age identified participants with a DSM-IV disorder by age 19.

Results: With all possible fMRI contrast parameters, the three-factor models were highly significant (p < 1.0 x 10ˆ21). In each of these models, EXT and CTQ scores were significant individual predictors (p < 0.001). At age 14, reward anticipation responses in the SN (Lg-Sm contrast), VS (Lg-No and Lg-Sm), caudate (Lg-No and Lg-Sm), putamen (Lg-Sm), and ACC (Lg-Sm) were significant predictors in their respective models (p ≤ 0.05). At age 19, reward anticipation responses in the VTA (Lg-Sm) and VS (Lg-No) were significant predictors within their models (p ≤ 0.029). The models had an overall accuracy of nearly 75% and accounted for ≥ 11% (Nagelkerke R2) of the variance in psychiatric disorders. Adding sex as a fourth factor improved each model’s significance. At age 14, significant interactions with CTQ were found for the ACC (Lg-Sm, p = 0.003) and putamen (Lg-Sm, p = 0.015).

Conclusions: The results replicate and extend our previous findings in a large sample, increasing confidence in our novel diathesis-stress model of commonly comorbid early onset mental health problems. The results have profound implications for diagnostic classification schemes (dimensions vs categories) and pleiotropic views of psychiatric disorder etiology (shared vs discrete).

Acknowledgements: Thanks also Robert Whelan, Gunter Schumann and other members of the IMAGEN consortium, too many to include on this website.

Keywords: neuroimaging biomarkers, Vulnerability, Stress

Disclosure: Nothing to disclose.

P669. Impact of Comprehensive Dental Care on Substance Use Disorder Treatment Outcomes

Annette Fleckenstein*, Glen Hanson

University of Utah, Salt Lake City, Utah, United States

Background: In a study published in the Journal of the American Dental Association, we reported the effects of integrated comprehensive oral health care for patients with substance use disorder (SUD) on SUD therapeutic outcomes. Briefly, after 1 - 2 months of SUD treatment, ~300 subjects (men and women, aged 20 through 50 years) received integrated comprehensive dental treatment. Their treatment outcomes were compared with those of sex matched patients who were likewise treated for SUD but did not receive comprehensive oral health care. Results revealed that patients that received dental care had substantially elevated improvements including: 1) an increase in rates of SUD treatment completion; 2) an increase in drug abstinence; 3) an increase in employment; and 4) a reduction in homelessness. The purpose of the present study was to extend these findings by investigating if comprehensive dental care improves quality of life in individuals receiving Targeted Adult Medicaid (TAM), which includes adult SUD patients receiving dental care, in a manner that correlates with the improved SUD treatment outcomes.

Methods: Self-selected Medicaid patients in SUD treatment completed comprehensive dental care, as well as demographic questionnaires, the oral health impact profile-14 (OHIP-14) and the Outcome Questionnaire-45.2 (OQ-45.2).

Results: Preliminary results demonstrate improvement in OHIP-14 and OQ45.2 scores, as well as improvement in employment status, among those that receive comprehensive dental care.

Conclusions: This is among the first studies to demonstrate the value of comprehensive dental care in Medicaid SUD patients. The present study confirms the findings that addressing oral health substantially enhances SUD treatment outcomes in a manner that appears to correlate with associated increases in quality of life elements.

Keywords: dental pain, substance use disorders, employment

Disclosure: Nothing to disclose.

P670. Central and Peripheral Signatures of Insulin Resistance as Biomarkers of Accelerated Cognitive Aging

Audrey Evers*, Fahim Abbasi, Katie Watson Lin, Alison Myoraku, Natalie Rasgon

Stanford University School of Medicine, Stanford, California, United States

Background: The trajectory of pathological brain aging is not well defined; however, the detrimental effects of insulin resistance (IR) on brain function have been demonstrated with direct and indirect measures. We previously reported a significant correlation between peripheral IR and performance on selective cognitive tasks and hippocampal structure and function. Specifically, IR was linked to disruptions in memory and executive function, metabolic decline in the medial prefrontal cortex, reduction in hippocampal volume, and aberrant intrinsic connectivity between the hippocampus and medial prefrontal cortex. In the current study, we aimed to compare the effects of concurrently obtained central and peripheral signatures of IR as mediators of premature cognitive aging in young and middle-aged adults with metabolic dysfunction.

Methods: We recruited 104 overweight, non-diabetic, non-depressed adults ages 25-50 years old who completed an insulin suppression test for direct measurement of insulin resistance, a battery of cognitive and neuropsychiatric measures, and gave samples for total and phosphorylated insulin receptor substrate (IRS) measures in L1CAM enriched exosomes. PEARSON correlation coefficients were calculated to examine the relationship between peripheral insulin resistance, metabolic variables, and neuronally derived IRS1 exosomes.

Results: The sample was 62% female, with a mean age of 40 years and a mean body mass index (BMI) of 28.9 kg/m2. We observed a significant negative correlation between peripheral (leptin, BMI) and central markers of metabolism (IRS measures in L1CAM enriched exosomes). Pearson correlation analyses revealed that phosphorylated IRS1 was significantly negatively correlated with BMI (r = -.23, p = 0.012) and leptin (r = -0.25, p = 0.006). Total IRS1 was significantly negatively correlated with steady-state plasma glucose (SSPG; r = -0.20, p = 0.039), fasting plasma glucose (r = -0.25, p = 0.006), BMI (r = -0.19, p = 0.041), and c-peptide (r = -0.21, p = 0.021).

Conclusions: To our knowledge, this is the first assessment of central and peripheral biomarkers of metabolism in nondiabetic and non-depressed subjects. We postulate a novel crosstalk in insulin signaling between the CNS and the periphery. These findings will be discussed as mediators of premature cognitive aging. Future research should focus on expanding these findings to various phenotypes of both metabolic and cognitive dysfunction.

Keywords: metabolic function, insulin resistance, cognitive aging

Disclosure: Nothing to disclose.

P671. Development of an Economic Impact and Monetization Framework for Mental Health and Substance Use Disorders: Changing the Dialogue

Mark Rapaport*, Fernando Wilson, Janis Dubno, Valentin Olivry, Gabe Freeman, Alan Lo

Huntsman Mental Health Institute, Salt Lake City, Utah, United States

Background: Mental health and substance use (MH/SU) services are underfunded nationwide. The creation of a monetization framework and socio-economic analysis that identifies the social and economic impact of these services presents one approach to more accurately value and compensate MH/SU providers for the value they create, not only for their patients, but for insurance companies, governments (ranging from municipal to federal level), and society at large. A monetization framework can also support implementation of novel funding mechanisms such as social impact investing. In 2023, the Huntsman Mental Health Institute (HMHI) engaged the Sorenson Impact Institute (SII) and the University of Utah’s Health Economics Core (HEC) to identify the social/economic impact generated from provision of MH/SU healthcare services. The SII/HEC team is leveraging the findings to create a monetization framework that is intended to capture the fiscal and societal benefits associated with improvement of patient and community-level outcomes.

Methods: The SII/HEC conducted an inventory of HMHI services, consisting of 26 programs and initiatives that included clinical and residential programs, crisis care and response programs including receiving centers and crisis hotlines, academic and workforce training programs, and strategic initiatives such as Stop Stigma Together. To inform initial identification of outcomes, an evidence-based review of prior literature relevant to HMHI services was conducted. This review provided associations of HMHI programs with specific social or economic outcomes, leading to creation of an impact indicator map. In addition, this information was supplemented with findings from over 100 stakeholder interviews that included clinicians, program administrators, and policymakers. Afterward, potential financial, social and/or economic outcomes of each programmatic area were identified and categorized into impact indicators.

Based on the impact indicator map, data requirements were identified to construct the monetization framework and socio-economic impact analysis (SEIA). Subsequently, the team collected key programmatic, demographic, and outcome data, as well as information to identify the key beneficiaries of fiscal savings associated with improved mental health outcomes. This information allowed the team to identify programs having the greatest potential for a feasible SEIA and monetization of social and economic impact and also informed our activities to quantify that impact.

Results: Analysis of programs and outcomes resulted in identification of direct and indirect impact indicators, linked to social and economic impact, that may be monetized based on the cost savings produced by these programs. Overall, the SII/HEC team identified 23 direct outcomes and 35 indirect outcomes that could be potentially monetized. These direct and indirect indicators included patient measures such as improved access to care, mental health status, and lives saved, financial measures including diversion from emergency department/inpatient utilization and cost savings, societal measures such as diversion from the criminal justice system and improved labor market productivity, and scholarly and health professions impact such as research funding generated, publications, and MH/SU workforce trained. Interviews with stakeholders identified several contributions of HMHI programs and challenges to the MH/SU field. For example, ongoing challenges reported by stakeholders included low reimbursement rates (or no reimbursement) from 3rd party payers, positive but unquantified externalities and cost savings from MH/SU services to the wider academic health system and society, and high need for diversion of MH/SU patients from the ED setting. As part of its review of programs and activities, the team identified significant data gaps notably with respect to patient-level outcomes data. As a result, additional literature reviews of similar programs were conducted to supplement actual programmatic data in order to find suitable impact rates from MH/SU treatment and the economic value of those potential benefits.

Conclusions: The SII/HEC team’s work thus far has found that the provision of MH/SU services creates substantial and diverse social and economic impact for patients and society. Through its diverse program offering, HMHI generates significant societal impact resulting in substantial financial savings. However, there is a need for an outcome measurement and data framework and infrastructure that supports regular and consistent patient- and program-level data collection in order to accurately quantify and value this impact.

Keywords: Health Economics, Social Impact, Population Health

Disclosure: Nothing to disclose.

P672. NIH Heal Pain Therapeutics Development Program (PTDP): A Funding Mechanism that Supports Development of a Non-Opioid Analgesics

Mohamed Hachicha*, Jim Pomonis, Matthew Rice, Mary Pellymounter

National Institute of Neurological Disorders and Stroke (NINDS), Rockville, Maryland, United States

Background: The Pain Therapeutics Development Program (PTDP) is part of the NIH Helping to End Addiction Long-term® (HEAL) Initiative, a trans-NIH research effort focused on improving prevention and treatment for opioid misuse and addiction as well as enhancing pain management. To bolster drug discovery and development activities in pain treatment, the division of translational research (DTR) within NINDS, and in collaboration with other NIH-institutes, launched several HEAL programs to promote and facilitate the discovery and development of non-opioid analgesics for the treatment of pain.

Methods: PTDP is a milestone-driven funding mechanism evaluated on an annual basis throughout the 5-year funding period. PTDP provides access to consultants with extensive biopharma experience and CRO resources for a variety of drug development activities including medicinal chemistry, drug manufacturing and formulation, DMPK, pre-IND enabling studies and Phase I clinical testing to accelerate drug discoveries into new therapies to treat pain.

Results: PTDP has funded a portfolio of 18 therapeutic projects representing multiple modalities (small molecules and biologics) and innovative targets originating from industry and academic institutions. Since its inception in 2019, two projects have reached IND submission status under the PTDP funding mechanisms.

Conclusions: PTDP has been a successful program focusing on novel molecular targets which are considered high-risk/ high-reward projects with an opportunity to test for their analgesic and non-abuse potential preclinically, and clinically for their safety and tolerability. In addition to grant support, the program provides access to consultants with extensive biopharma experience and CRO resources for a variety of drug development activities and no claim of sponsor’s intellectual property.

Keywords: Drug development, Pain, Preclinical, Clinical, Drug discovery

Disclosure: Nothing to disclose.

P673. Characterization of the Rat Spinal Nerve Ligation (SNL) Model of Neuropathic Pain in Support of the NIH Heal Initiative Preclinical Screening Platform for Pain (PSPP)

Mark Urban*, Elizabeth Dugan, Katelyn Buban, Jennifer Hagedorn, Latha Devi, Stephen Morairty, Mei Kwan, Afshin Ghavami, Sarah A. Woller, Smriti Iyengar, Taleen Hanania

PsychoGenics, Inc., Paramus, New Jersey, United States

Background: The National Institutes of Health Helping to End Addiction Long-term Initiative, or NIH HEAL Initiative, Preclinical Screening Platform for Pain (PSPP) program aims to accelerate the discovery and development of novel non-opioid, non-addictive pain therapeutics. Under the HEAL initiative, the NIH is collaborating with PsychoGenics, Inc. to screen and profile novel assets in vivo in rat models of pain to determine efficacy, and in additional tests to examine potential adverse effects and abuse liability. To systematically profile novel assets in rat models of pain, efforts in the PSPP program are also focused on the characterization and optimized use of these models. Here, we describe one such effort to characterize the rat spinal nerve ligation (SNL) model of neuropathic pain in terms of optimal evoked and non-evoked behavioral pain endpoints, non-behavioral pain biomarkers (i.e. EEG), and use of the model to determine target engagement for select compounds to establish occupancy/efficacy relationships.

Methods: Adult male and female Sprague Dawley rats were used in these studies, and all housing and testing of animals were in accordance with the Principles of Laboratory Animal Care and the approval of the PsychoGenics Inc., Institutional Animal Care and Use Committee in AAALAC-accredited facilities. For the rat SNL model, rats received a surgical procedure where the left L5 and L6 spinal nerves were tightly ligated distal to the dorsal root ganglia using 4-0 silk suture. For the EEG studies, rats were implanted with screw electrodes at S1 (primary somatosensory cortex) and PFC (prefrontal cortex) using a standard rat head mount (Pinnacle 8239 2-EEG/1-EMG). For the target engagement studies, serotonin (SERT) and norepinephrine (NET) transporter occupancy was evaluated using [3H]-citalopram and [3H]-nisoxetine, respectively, in brain and spinal cord slices ex vivo following dosing (PO) with duloxetine. Behavioral pain endpoints that were assessed included evoked hind paw mechanical (von Frey filaments) and cold allodynia (acetone test), and non-evoked hind limb weight bearing deficits and wheel running behaviors. Non-evoked pain endpoints involved measurement of EEG activity by determining changes in spectral power within standard EEG bands (i.e. delta, theta, alpha, beta, low gamma, high gamma).

Statistical analysis, effect size, and power analysis: Data were analyzed using two-way repeated measures ANOVA with Bonferroni’s or Dunnett’s post hoc test when appropriate. Effects p < 0.05 were considered to be statistically significant. Power analysis and effect size were determined using SAS/STAT, and appropriate sample size was based on a power value of 0.8 to ensure adequate power for F-tests for two-way interactions.

Results: In the rat SNL model, tight ligation of the left L5 and L6 spinal nerves produced unilateral hind paw mechanical and cold allodynia in male and female rats which was maximal at Week 2 and persisted for 12 weeks relative to sham-operated rats. Additionally, reduced dynamic weight bearing was observed on the left hind limb in male and female SNL rats which was maximal at Week 1 and persisted throughout the 6-week testing period relative to sham-operated rats. Potential changes in general activity were measured using the wheel running test during the dark and light cycles, and no significant differences in activity were observed between SNL and sham-operated male or female rats during the 4-week testing period. In the EEG studies, no significant changes in spectral power were observed between SNL and sham-operated male or female rats during the 6-week testing period. For the target engagement studies, administration of duloxetine (6, 30, 60 mg/kg, PO) resulted in dose-dependent SERT occupancy in spinal cord (maximum ~50% occupancy), but did not result in detectable NET occupancy in the spinal cord. In brain sections, SERT occupancy (maximum ~70%) and NET occupancy (maximum 50%) were detected following duloxetine administration (6, 30, 60 mg/kg, PO); however, SERT and NET occupancy was not observed to be dose-dependent.

Conclusions: The results from these studies demonstrate that hind paw mechanical and cold allodynia represent a reliable, evoked pain endpoint in the SNL model which may be used to routinely evaluate efficacy of novel compounds for neuropathic pain. Reduced hind limb weight bearing may represent a non-evoked pain endpoint in the SNL model to further examine efficacy of novel compounds, and additional pharmacological data with reference analgesic compounds may support the use of this endpoint for routine screening. EEG activity was examined in SNL, sham, and naïve rats to identify a potential non-behavioral pain biomarker; however, no clear changes in EEG spectral power were identified in SNL rats relative to control animals. SERT and NET target engagement following administration of duloxetine (PO) was observed in brain (SERT, NET) and spinal cord (SERT) at efficacious doses. These efforts represent an example of the strategy used within the PSPP program to comprehensively characterize the utility of rat pain models to accelerate the development of novel therapeutics for the treatment of pain.

Keywords: neuropathic pain, Behavioral Model, Behavioral Pharmacology

Disclosure: Nothing to disclose.

P674. NIMH IRP Translational Neuropsychopharmacology Initiative

Janet Clark*, Carlos Zarate, Susan Amara

Division Intramural Research Program, NIMH, Bethesda, Maryland, United States

Background: The need for novel therapies for central nervous system (CNS) disorders with improved efficacy, safety, and tolerability is unquestionably high, as it is widely recognized that, despite currently recognized treatments, CNS disorders are major contributors to the global burden of illness and incur high economic costs. Unfortunately, over the past two decades, CNS drug discovery has been, with a few exceptions, relatively unsuccessful in delivering new chemical entities especially for the treatment of psychiatric disorders. The National Institute of Mental Health (NIMH) has become increasingly aware of the reduced investment by pharmaceutical companies in the development of therapeutics for treating psychiatric disorders despite the unmet medical need. In response to the reduced efforts in the pharmaceutical industry towards psychiatric drug discovery the NIMH Intramural Research Program (IRP) is proposing to re-invigorate psychiatric drug discovery by facilitating and de-risking the discovery and development of novel treatments. Support for the discovery and development of new treatments for psychiatric disorders including target validation, biomarker development, IND enabling studies, and Phase I safety / tolerability and Phase II proof of concept studies are all in scope for this important NIMH IRP initiative.

Methods: After completing a confidentiality agreement, interested parties including academic groups, small and large pharma or biotech may apply by submitting a proposal including detailed information on: the psychiatric indication, therapeutic rationale, target engagement, current stage of development, a development plan, project milestones, intellectual property and patent landscape, future objectives and anticipated outcomes. In addition, a compound ‘report card’ with details on the chemical properties of the compound of interest is completed and submitted with each proposal.

Results: To steer this initiative the NIMH IRP has established the NIMH Translational Neuropsychopharmacology Task Force (TNTF). The TNTF is comprised of a panel of neuroscience drug discovery/development experts from the pharmaceutical industry, academia and the NIH who prioritize, critically review, and recommend new proposals for support to the NIMH IRP Faculty and Leadership. Supported proposals will either be sourced by NIMH IRP Principal Investigators and Staff or the desired work may be supported by a collaborative agreement between the interested party and the NIMH IRP. The NIMH IRP has experience in the clinical testing of novel therapies for the treatment of neuropsychiatric diseases including generalized anxiety disorder (GAD), treatment resistant depression (TRD), suicide, bipolar disorder (BPD), perimenopausal depression (PMD), irritability in children, autism spectrum disorder, and neuroimmune disorders. In addition, the NIMH IRP has extensive CNS imaging capabilities including PET, MEG, EEG and fMRI that are readily integrated into clinical studies to answer questions regarding target engagement or better understand disease pathology.

Conclusions: The NIMH is committed to engaging with the scientific community and reinvigorating the development of new therapies for the treatment of psychiatric disorders. The current reduction in psychiatric drug discovery and development in the pharmaceutical industry provides an opportunity for the NIMH IRP to contribute in an important way to de-risking novel therapeutics and engage industry in an effort to identify and develop more effective treatments for patients.

Keywords: Neuropsychopharmacology (NPP), drug discovery/development, Neuropsychiatric Disorders, NIMH

Disclosure: Nothing to disclose.

P675. Targeting Glymphatic Pathways to Ameliorate Cognitive Impairments in an Aged Mouse Model of Postoperative Delirium

Yuto Hasegawa*, Robyn Wiseman, Feiyi Xiong, Yannan Li, Xiaolei Zhu, Barbara Slusher, Atsushi Kamiya

Johns Hopkins University School of Medicine, Baltimore, Maryland, United States

Background: Postoperative delirium is a common cognitive complication after major surgery in older adults. However, the molecular and cellular mechanisms underlying its pathophysiology are poorly understood, delaying advances in prevention and treatment. Systemic inflammation triggers neuroinflammation, primarily regulated by brain resident immune cells, leading to acute cognitive dysfunction. Recent studies suggest that dysfunction in the glymphatic waste clearance system may contribute to cognitive decline in the elderly. Interestingly, glutamate can induce astrocyte swelling by altering the expression of aquaporin-4 water channel (AQP4) in perivascular astrocytes, which is a critical regulator of the glymphatic system. In our current study using an aged mouse model of abdominal surgery, we investigate whether abdominal surgery-induced systemic inflammation increases glutamate levels in the central nervous system, which may affect AQP4 signaling in perivascular astrocytes, leading to aberrant glymphatic waste clearance system and cognitive impairments. Furthermore, we explore whether pharmacological inhibition of specific enzymes involved in glutamate regulation could normalize glial cell and glymphatic phenotypes, thereby improving cognitive deficits.

Methods: 18-20 months old C57BL/6 mice (both sexes) were used in this study. They were divided into the following three groups; The first control group was naïve mice without anesthesia and surgical intervention. The second control group (sham) received 2-2.5 % isoflurane in 0.5-1 L/min O2 flow anesthesia for 15 minutes. The surgery group received the same anesthesia setting as the sham group during abdominal surgery which is a 1-minute intestine manipulation. After 2 days of recovery from surgery, we performed histochemical, biochemical, and behavioral assays to identify surgery-induced glutamate expression, astrocyte morphology, and AQP4 signaling changes in perivascular astrocytes, resulting in disturbance of cognition, especially spatial memory and selective attention in aged mice. We also examined long-lasting cognitive effects three weeks after 1set of behavioral assessment (i.e., 25 days after abdominal surgery) to determine whether the phenotype observed are transient, which may recapitulate delirious phenotypes in aged humans.

Two-tailed analysis of unpaired Student’s t tests were used for data analysis in comparing control and surgery group. This statistical analysis was conducted by using GraphPad Prism software. The sample size was calculated based on our pilot results. We set 1-beta=0.80, alpha=0.05 in power analyses. A value of p < 0.05 was considered statistically significant.

Results: Inhalation anesthesia followed by abdominal surgery produced male-specific impairments in object and spatial recognition memory, as assessed by the novel object and location recognition tests. Abdominal surgery also induced deficits in spatial learning and memory, as well as selective attention in aged mice, as confirmed by the Barnes maze test and Object-based attention test. These cognitive deficits were not observed 3-weeks post-surgery. To determine whether abdominal surgery affects glymphatic function, we injected fluorescent tracers (BSA-647 and 3kDA-Dextran-FITC) into the subarachnoid cerebrospinal fluid (CSF) of the cisterna magna after abdominal surgery in aged mice. We observed a significant reduction in the tracer penetration in the brains of mice subjected to surgery compared to the control (sham) group (p = 0.0343, t = 2.728, df=6). Abdominal surgery induced excess glutamate production (p = 0.0008, t = 4.749, df=10), morphological changes of astrocytes (p = 0.0002, t = 5.570, df=10), and altered expression of AQP4 signaling in the perivascular astrocytes (p = 0.0002, t = 5.570, df=10). We are currently investigating whether pharmacological inhibition of excess glutamate production can ameliorate these phenotypes.

Conclusions: We identified that the inhalation anesthesia followed by abdominal surgery in aged mice leads to transient male-specific cognitive impairments. These behavioral outcomes are linked to a disturbance in glymphatic function, potentially caused by excess glutamate production, which induces astrocyte morphological changes and altered AQP4 signaling. Our research findings highlight disturbances in glymphatic waste clearance system as potential pathological mechanisms underlying postoperative delirium. We hope our studies contribute to develop mechanism-based pharmacological intervention for the prevention and treatment of postoperative delirium in the elderly.

Keywords: postoperative cognitive dysfunction, Glymphatic clearance, glutamate

Disclosure: Nothing to disclose.

P676. The Impact of Estrogen-Suppressing Contraceptives on Mental Health Outcomes in Borderline Personality Disorder

Seyma Katrinli*, Alex Rothbaum, Raneeka Foster, Chase Turner, Ben Hunter, Abigail Powers, Vasiliki Michopoulos, Alicia Smith

Emory University School of Medicine, Atlanta, Georgia, United States

Background: Borderline Personality Disorder (BPD) is a chronic mental health disorder characterized by rapidly shifting emotional, interpersonal, and behavioral symptoms, and often co-morbid with mood and anxiety disorders. Females are more likely to be diagnosed with BPD than males and exhibit greater functional impairment. Hormonal fluctuations, particularly estrogen, may influence the manifestation of BPD symptoms. Here we investigated the influence of estrogen-suppressing contraceptives on mental health outcomes in those with and without a BPD diagnosis.

Methods: The study includes 348 females ages 18-50 undergoing residential treatment for psychiatric disorders, with 131 having a BPD diagnosis. Patients were categorized based on their contraceptive method: 1) Estrogen-suppressing contraceptives (e.g., oral contraceptives, Depo-Provera; N = 145) and 2) Naturally cycling (e.g., no contraceptives; N = 203). Mixed effect interaction models with a random intercept for subjects tested the impact of estrogen-suppressing contraceptives on the relationship between BPD diagnosis and mental health outcomes at admission and discharge ( ~ 90 days treatment), assessed by the four Behavior and Symptom Identification Scale (BASIS-32) domains: difficulties in relationships; daily living; depression/anxiety; and impulsivity.

Results: Females with a BPD diagnosis were more likely to use estrogen-suppressing contraceptives (OR[95% CI] = 1.60[1.01, 2.55], p = 0.045). However, estrogen-suppressing contraceptive use was not associated with mental health outcomes at admission and discharge or changes in symptom severity pre- to post-treatment. Estrogen-suppressing contraceptive use moderated the association of BPD diagnosis with difficulties in relationships (Beta[SE] = -0.56[0.20], p = 0.004), depression and anxiety symptoms (Beta[SE] = -0.59[0.20], p = 0.004), and difficulties in daily living (Beta[SE] = -0.50[0/20], p = 0.01). Patients with a BPD diagnosis expressed increased difficulties in relationships (Beta[SE] = 0.51[0.14], p = 2.4e-4), depression and anxiety symptoms (Beta[SE] = 0.46[0.14], p = 8.8e-4), and difficulties in daily living (Beta[SE] = 0.42[0.14], p = 0.003) only if they were naturally cycling.

Conclusions: Our findings suggest that estrogen-suppressing contraceptives may help to regulate the rapidly shifting emotional, interpersonal, and behavioral symptoms in those with BPD by stabilizing estrogen levels.

Keywords: Borderline Personality Disorder, Contraceptive use, Estrogen, Mental health outcomes

Disclosure: Nothing to disclose.

P677. 56 Symptoms and Counting: Item Content Overlap of the Pediatric Catatonia Rating Scale With Other Measures of Catatonia

Musa Yilanli*, Colleen Waickman, Zachery Mondlak, Eric Youngstrom

Nationwide Children’s Hospital, Columbus, Ohio, United States

Background: Catatonia is a neuropsychiatric syndrome associated with mood, psychotic, substance-induced, and medical disorders in children and adults. It is challenging to differentiate catatonia from other comorbid disorders in the pediatric population, and most catatonia research has centered on adults. The Pediatric Catatonia Rating Scale (PCRS) is the only scale that evaluates children with catatonia and was adapted from the Bush Francis Catatonia Rating Scale (BFCRS) to use in child and adolescent inpatients (Benarous et al., 2016). It was developed within a prospective study of catatonia and had moderate internal consistency. Other scales exist to evaluate catatonia in adults but have not been explicitly compared to the PCRS. It is not known how much this scale can be differentiated from pre-existing scales that evaluate catatonia in adults. This study compared eight different scales and the DSM-5-TR criteria for catatonia to investigate the degree of overlap between these separate measures, with the goal of guiding future research on catatonia rating scales for children and adolescents.

Methods: The PCRS, DSM-5-TR criteria for catatonia, BFCRS, BFCRS-Revised, Kanner Scale, Rogers Catatonia Scale (RCS), Modified Rogers Scale (MRS), Northoff Catatonia Rating Scale (NCRS), and Braunig Catatonia Rating Scale (BCRS) were evaluated and compared. A clinical expert in catatonia rated symptoms found in each measure as either a 0 or 1 (0 = No, not on scale; 1 = symptom featured on scale). We used the Jaccard Index in R Studio to determine the content overlap for each scale. Symptoms for each scale were visualized represented using an item content overlap map.

Results: 56 symptoms of catatonia were identified in the content overlap analysis. A Jaccard Index Scale Analysis indicated that the highest overlap between scales was a 0.88 (BFCRS and BFCRS-R), while the lowest was 0.28 (DSM-5-TR criteria and NCRS). The most represented symptoms were waxy flexibility, stupor/immobility, stereotypy, posturing, negativism, mutism, mannerisms, and echolalia/echopraxia, found across all nine measures. 23 symptoms were represented only once, with schizophasia, acrocyanosis, and automatic compulsive movements unique to the PCRS. The PCRS had the greatest overlap with the BFCRS (0.53) and the least overlap with the RCS (0.33).

Conclusions: The primary goal of this study was to examine the overlap in symptom content between the PCRS and eight other catatonia measures, including DSM-5-TR criteria. As anticipated, the PCRS exhibited the greatest overlap with the BFCRS, given its derivation from this scale with specific pediatric additions. The NCRS, as the longest scale, was the most distinct. The RCS, originally developed to differentiate catatonic depression from Parkinson’s disease, primarily focuses on motoric features of catatonia. This exclusive focus limits its generalizability to other populations, particularly pediatrics. The broad spectrum of symptoms identified in catatonia highlights the need to determine which symptoms are most predictive of both pediatric and general catatonia. Limitations of this study include the use of a single coder and the absence of symptom validity assessment. Future research could evaluate the predictive value of symptoms across catatonia rating scales to inform clinical practice.

Keywords: catatonia, Rating scales, DSM-5

Disclosure: Nothing to disclose.

P678. Exposure to Antipsychotic Medication is Associated with Less Days Alive and Free From Catatonia in Critically Ill Patients: Results From the Delirium and Catatonia Prospective Cohort Study

Gloria Mina, Trey McGonigle, Jinyuan Liu, Nathan Brummel, Mayur Patel, Joshua Smith, Stephan Heckers, Pratik Pandharipande, Robert Dittus, Eugene “Wes” Ely, Jo Wilson*

Vanderbilt University Medical Center, Nashville, Tennessee, United States

Background: Catatonia, a neuropsychiatric syndrome characterized by changes in motor behavior (increased, decreased or abnormal), volitional and affective changes, once linked to schizophrenia, is increasingly recognized in mood disorders and in the medically ill, typically co-occurring with delirium. The relationship between catatonia development and antipsychotic medication exposure, frequently initiated in intensive care units to manage psychotic phenomena and agitation in the context of delirium has not been previously described. As dopaminergic blockade is one mechanistic theory underlying catatonia, we sought to understand if exposure to antipsychotic medication is associated with an increased risk of developing catatonia in critically ill adults.

Methods: The Delirium and Catatonia (DeCat) Prospective Cohort Investigation is a convenience sample, prospective cohort study nested within two longitudinal cohort studies. Data was collected between December 2018 – July 2021 in medical, surgical and trauma intensive care units at Vanderbilt University Medical Center, Nashville, TN. Prospective participants were eligible if > 18 years old on mechanical ventilation and/or vasopressors who did not have a major neurocognitive disorder at baseline. We excluded patients with severe psychiatric illness at baseline or who could not participate in follow-up assessments. Our primary exposure was antipsychotic medication administration (ever / never exposed) and cumulative dosage during the first 14 days of enrollment, compared to no exposure; antipsychotics were converted to haloperidol equivalents (mg). Catatonia was assessed using the Bush-Francis Catatonia Rating Scale and mapped to DSM-5 criteria. The primary outcome was Catatonia Free Days (CFD), defined as number of days the patient was alive and free from catatonia during first 14 days of study enrollment in the main model, and during the first 5 days in a sensitivity analysis. Proportional odds logistic regression was used to estimate the odds ratio (OR) of outcome events, with robust covariance estimates clustered by cohort study. Models were adjusted for age, sex, mechanical ventilation at enrollment, Sequential Organ Failure Assessment (SOFA) score (severity of illness measure), Charlson comorbidity index, and short form of the Informant Questionnaire on Cognitive Decline in the Elderly (SF-IQCODE) (assessing cognitive decline) total score at enrollment. ORs were reported and interpreted as an increase or decrease in the odds having more CFD (OR < 1 indicating a lower odds of more days alive and free from catatonia). Multiple imputation (40 replications) was used to account for missing data. The Vanderbilt Institutional Review Board approved this study. All analyses were performed in R studio (R Foundation for Statistical Computing, Vienna, Austria; http://www.R-project.org/).

Results: 270 patients met inclusion criteria. The median (IQR) age of patients was 54.5 (36.7, 67.2) years. 168 (62%) patients were men. 39 patients (14%) experienced catatonia during the study period, 27 (26%) of whom were exposed to antipsychotic medication and 12 (7.1%) of whom were not exposed (p < 0.001). Only 5 patients were catatonic at the time of study enrollment. Compared to patients who were never exposed to antipsychotics, those exposed had a lower chance of more CFD (OR, 0.4939 [95% CI, 0.3857-0.6325]) in the 14-day model and an even lower odds of more CFD (OR, 0.2568 [95% CI, 0.158-0.4173]) in the 5-day model. Furthermore, those exposed to higher cumulative dosages had a lower chance of more CFD compared to those exposed to lower dosages (OR, 0.0335 [95% CI, 0.0166-0.0673]) in the 14-day model, and 0.0281 (0.0142-0.0556) in the 5-day model.

Conclusions: In this cohort of critically ill patients, exposure to antipsychotic medication and higher cumulative dosage was associated with less days alive and free from catatonia compared to no exposure, with the effect strengthening in the acute period for the binary exposure group. This study may influence how intensivists approach the use of antipsychotic medications in this population, as well as highlight underlying potential biologic mechanism, further strengthening the dopaminergic blockade hypothesis underlying catatonia.

Keywords: Antipsychotic drugs, catatonia, epidemiology

Disclosure: Nothing to disclose.

P679. The Phenomenology of Postpartum Psychosis: Preliminary Findings From the Massachusetts General Hospital Postpartum Psychosis Project

Lee Cohen*, Miranda Arakelian, Taylor Church, Madison Dunk, Margaret Gaw, Hannah Yoon, Lauren Kobylski, Rachel Vanderkruik, Marlene Freeman

Ammon-Pinizzotto Center for Women’s Mental Health, Massachusetts General Hospital, Boston, MA, Boston, Massachusetts, United States

Background: Postpartum psychosis (PP) is a severe psychiatric disorder–with limited data or consensus on diagnostic criteria and clinical presentation–that affects thousands of people each year. The Massachusetts General Hospital Postpartum Psychosis Project (MGHP3) was established to: 1) describe the phenomenology of PP, and to 2) identify genomic and clinical predictors in a large cohort. Results thus far point to a richer understanding of the heterogeneity and complexity of this often-misunderstood illness and its nature over time.

Methods: Data are collected from those who experienced PP within 6 months postpartum of delivery and within the ten years prior to the MGHP3 interview. Participants provide information via the Mini International Neuropsychiatric Interview for Psychotic Disorders Studies (MINI-PDS), MGHP3© Questionnaire (including assessment of episode onset, duration, symptoms, and treatment received) and other relevant history. This retrospective study uses validated diagnostic tools to evaluate psychiatric history across participants’ lifetime. Descriptive statistics (e.g., median values, frequencies) were conducted to describe the phenomenology of PP.

Results: As of November 3, 2022, 248 participants with histories of at least one episode of PP completed the MGHP3 interview. Most participants met criteria for Bipolar I Disorder with psychotic features (71.8%). During PP episode(s), participants reported odd beliefs or delusions (87.6%), persecutory delusions (75.2%), ideas of reference (55.8%), visual (52.3%) and/or auditory (48.1%) hallucinations. The median time between delivery and symptom onset was 10 days (SD = 43.72). Most participants reported receiving medication (93.0%) and/or psychotherapy (65.9%).

Conclusions: This report describes findings regarding the phenomenology of postpartum psychosis among the MGHP3 cohort, the largest cohort with validated PP studied to date. This ongoing effort to refine the phenotype of PP and to delineate underlying genetic determinants of the disorder will contribute to an enhanced understanding of this serious illness. It also underscores areas for further rigorous assessment using other research methods and sets the stage for translational reproductive neuroscience – including ongoing analyses of neuroimaging and genetic data from the MGHP3 cohort.

Keywords: postpartum psychosis, reproductive psychiatry, Bipolar Disorder, postpartum, mania

Disclosure: Alkermes, Inc.: Other Financial or Material Support (Self). Dr. Reddy’s Laboratories, Inc.: Other Financial or Material Support (Self). Eisai Inc.: Other Financial or Material Support (Self). Otsuka America Pharmaceutical, Inc.: Other Financial or Material Support (Self). Supernus Pharmaceuticals: Other Financial or Material Support (Self) Teva Pharmaceutical Industries Ltd.: Other Financial or Material Support (Self). SAGE Therapeutics: Other Financial or Material Support (Self).

P680. Assessing the Potential Cardiovascular Risk of Microdosing Lysergic Acid Diethylamide in Mice

Devin Effinger*, Jillian King, Joselynn Calderon, Janae Strong, Scott Thompson

University of Colorado School of Medicine, Aurora, Colorado, United States

Background: Psychedelics are hallucinogenic compounds that have received renewed interest due to observed therapeutic effects for multiple psychiatric disorders. Microdosing, the prolonged ingestion of a sub-hallucinogenic dose of a psychedelic, has gained popularity in the public due to anecdotes from users suggesting potential cognitive and emotional benefits. Classical psychedelics, such as lysergic acid diethylamide (LSD) and psilocybin, act on a myriad of different serotonin (5-HT) receptors, with a high affinity for both the 5-HT2A and 5-HT2B receptors. A strong body of evidence has linked activation of 5-HT2B receptors to the development of valvulopathy and drug-induced valvular heart disease, resulting in several 5-HT2B agonist drugs being pulled off the market. Given the recent explosion in popularity of psychedelic use, changes in state legislation surrounding the use of these drugs, and the known activity at the 5-HT2B receptor, investigating the effects of microdosing psychedelics on cardiovascular health is imperative. We hypothesized that chronic administration of low dose LSD would induce valvulopathy in mice.

Methods: Male and female C57/B6J mice were chronically administered either 5-HT (positive control, 40 mg/kg, i.p., n = 6) or LSD (0.01 mg/kg, i.p., n = 6). Dosing followed a typical human microdosing schedule: 5 days on 2 days off. Echocardiograms were performed at baseline, 4-weeks, and 8-weeks to monitor changes in heart structure and function. Histological analyses are being performed to determine changes in collagen ratio, valve thickness, etc. Statistical analyses were performed using 2-way analysis of variance (ANOVA) to compare differences between groups and across the different time points.

Results: Following the 4- and 8-week scans, the 5-HT group showed decreased left ventricle inner diameter at end systolic (Ftreatment(1,10) = 11.90, p = 0.0062; Finteraction(2,20) = 5.487, p = 0.0126) and at end diastolic (Finteraction(2,20) = 4.913, p = 0.0184; Finteraction(2,20) = 5.487, p = 0.0126). Left ventricle volume at end systolic (Ftreatment(1,10) = 9.569, p = 0.0114; Finteraction(2,20) = 5.028, p = 0.0170) and at end diastolic (Finteraction(1,10) = 9.569, p = 0.0114) were also decreased compared to baseline, indicative of left ventricle valvulopathy. No changes were seen in the LSD microdose group. Further follow up scans will be conducted at 12 weeks, along with the addition of a 0.03 mg/kg LSD group to determine if cardiovascular risks arise at higher doses.

Conclusions: Here we show that chronic, low-dose LSD did not produce any pathological changes to heart structure as compared to a 5-HT positive control group. Ongoing work will examine a higher dose of LSD (0.03mg/kg, i.p.) to determine if cardiovascular effects are dose responsive. Insight into the effects of LSD and other psychedelics on cardiovascular structure and function will be critical in established safe dosing regimen for prolonged, low-dose application of psychedelic and psychedelic-derivative drugs.

Keywords: Microdosing with psychedelics, Psychedelics, cardiovascular function

Disclosure: Nothing to disclose.

P681. Characterizing the Effects of Serotonergic Psychedelics on Interval Timing in Mice

Jared Kopelman*, Landon Klein, Susan Powell, Richard Sharp, Jared Young, Adam Halberstadt

UC San Diego, San Diego, California, United States

Background: Deficits in temporal processing occur in multiple psychiatric disorders and may be a key feature of schizophrenia. Certain drugs also have the capacity to alter temporal processing. For instance, serotonergic psychedelics such as lysergic acid diethylamide (LSD) and psilocybin have profound effects on the subjective experience of time. Previous studies have shown that phenethylamine psychedelics, which are relatively selective for the 5-HT2A receptor (the primary CNS target for psychedelics in the brain), can alter performance on rodent timing tasks. However, tryptamine psychedelics such as LSD and psilocybin, which are less selective pharmacologically, have not been evaluated in these paradigms. For this study, we tested whether psilocin, the O-dephosphorylated active metabolite of psilocybin, alters rodent performance on an interval timing task and examined whether 5-HT1A and 5-HT2A receptors play a role in the effect of psilocin. We also further probed the role of 5-HT2A receptors in task performance by testing the effect of several antagonists with varying degrees of selectivity. Lastly, given the evidence that the medial prefrontal cortex (mPFC) plays important roles in both interval timing and psychedelic drug effects, we examined whether that brain region is involved in the effects of psychedelics on interval timing in rodents.

Methods: A series of in vivo experiments were performed in male C57BL/6J mice to assess the effect of psychedelics on interval timing using the discrete trials task (DTT). In the DTT, mice are presented with two levers after a variable interval; responding on lever A is reinforced if the interval was < 6.5 s, and responding on lever B is reinforced if the interval was > 6.5 s. A two parameter logistic function is then fitted to the proportional choice for lever B (%B responding), yielding estimates of the indifference point (T50, a measure of timing accuracy), as well as the slope, Weber Fraction (WF) and Difference Limen (DL), which are measures of timing precision. We tested the effect of psilocin (0.4 mg/kg and 0.8 mg/kg) compared to saline vehicle on DTT performance. Next, we pretreated mice with the 5-HT1A antagonist WAY-100,635 (0.5 mg/kg) prior to psilocin (0.8 mg/kg) to examine the role of the 5-HT1A receptor in the effect of psilocin on DTT performance. We performed a similar experiment with M100907 (0.03 mg/kg), a highly selective 5-HT2A antagonist, to examine the role of this receptor in the effect of psilocin. We also assessed the effects of the selective 5-HT2A antagonists M100907 and MDL 11,939 and the mixed 5-HT2A/2C antagonist ketanserin on DTT performance to study the role of the 5-HT2A receptor in interval timing more broadly. Lastly, we examined the effect of microinfusion of the psychedelic 5-HT2A agonist 2,5-dimethoxy-4-iodoamphetamine (DOI) into the mPFC and compared it to the effects of systemic (IP) DOI administration.

Results: There was a main effect of psilocin on %B responding (F(2,30) = 7.19, p = 0.0028), T50 (F(2,30) = 9.29, p = 0.0007), and DL (F(2,30) = 5.07, p = 0.0127). The highest dose of psilocin (0.8 mg/kg) significantly reduced %B responding at longer time intervals (6.9, 7.8, 8.7, 9.6, and 10.5 s), increased T50 from 6.73 ± 0.16 s to 7.69 ± 0.26 s, and increased the DL from 1.19 ± 0.07 to 1.70 ± 0.23. Pretreatment with WAY-100,635 did not significantly alter the effect of psilocin on %B responding, and there was no interaction between psilocin and WAY-100,635 for T50, curve slope, DL, or WF, indicating that the effect of psilocin on DTT is not dependent on activation of 5-HT1A receptors. In contrast, there was a significant interaction between psilocin and M100907 (F(1,28) = 6.25, p = 0.0185) for %B responding, such that psilocin had no effect on %B responding in mice pretreated with M100907, indicating that the 5-HT2A receptor mediates the effect of psilocin on DTT performance. Blockade of the 5-HT2A receptor with M100907, MDL 11,939, or ketanserin produced significant changes in %B responding and shifted the psychometric curve to the right, increasing T50. Although ketanserin significantly increased the WF (F(3,33) = 3.19, p = 0.0361) and DL (F(3,33) = 9.26, p = 0.0001), neither M100907 nor MDL 11,939 altered timing precision. Lastly, microinfusion of DOI bilaterally into the mPFC (2 µg/side) mimicked the effect produced by IP administration of DOI, producing a significant alteration of %B responding (F(1,31) = 8.84, p = 0.0057) and increasing T50 from 6.57 ± 0.11 s to 7.20 ± 0.15 s (F(1,31) = 10.75, p = 0.0026).

Conclusions: Similar to phenylalkylamine psychedelics, the tryptamine psilocin altered DTT performance in mice, shifting the response curve to the right and reducing the precision of timing. The effect of psilocin on %B responding was blocked by pretreatment with M100907 but not WAY-100,635, indicating that the effect of psilocin on DTT performance is mediated by the 5-HT2A receptor but not by the 5-HT1A receptor. Interestingly, 5-HT2A antagonists also affected DTT performance, indicating timing can be altered by both increases or reductions in 5-HT2A activity. In addition, we found that microinfusion of a psychedelic drug directly into the mPFC had the same effect on DTT performance as systemic administration, indicating that this region may be an important site of action for psychedelic drug effects on interval timing in rodents.

Keywords: Psychedelics, 5-HT2A receptors, hallucinogens, Psilocybin, timing

Disclosure: Nothing to disclose.

P682. Acute and Lasting Changes in Cortical Dynamics After Ibogaine Treatment in Veterans With Traumatic Brain Injury

Jennifer Lissemore*, Anna Chaiken, Kirsten Cherian, Derrick Buchanan, Flint Espil, Jackob Keynan, Malvika Sridhar, Camarin Rolle, Manish Saggar, Corey Keller, Nolan Williams

Stanford University School of Medicine, Stanford, California, United States

Background: Recent evidence has shown striking improvements in the sequelae of traumatic brain injury (TBI) after treatment with magnesium-ibogaine, including remission of post-traumatic stress disorder (PTSD), depression, and anxiety symptoms. The neural mechanisms underlying this promising rapid treatment option, however, are not yet clear. In particular, TBI can lead to widespread disruptions to the cortical dynamics that underlie healthy human brain function, yet the neurophysiological effects of ibogaine on cortical dynamics in humans, and how these effects relate to therapeutic outcomes after ibogaine, have yet to be investigated.

Methods: Cortical dynamics were measured using resting-state electroencephalography (EEG) before, immediately after, and one month after ibogaine treatment in 30 male veterans with a history of TBI who showed a robust response to ibogaine treatment. EEG measures used to index cortical dynamics included spectral measures of band power, theta/beta ratio and peak alpha frequency. We also assessed the relationship between these EEG measures and neurocognitive, PTSD, depressive, and anxiety symptom improvements after ibogaine treatment.

Results: We found significant acute changes in cortical dynamics after ibogaine treatment, including enhanced power in slower cortical rhythms (4-13 Hz), diminished power in faster rhythms (13-50Hz), increased theta/beta ratio and lowered peak alpha frequency. Some neurophysiological changes were sustained for at least one month after ibogaine, including lowered peak alpha frequency. Neurophysiological markers that were significantly modulated after ibogaine treatment also correlated significantly with improvements in executive function and PTSD symptoms post-treatment. Additionally, EEG measures of cortical dynamics at rest before treatment were significantly correlated with improvements in executive function, PTSD, and anxiety after ibogaine treatment.

Conclusions: We identified neurophysiological markers that are modulated by ibogaine both acutely and for at least one month after treatment, which may be associated with psychiatric and neurocognitive treatment outcomes. We also identified pre-treatment neurophysiological markers that may be useful as predictors of ibogaine treatment outcomes. By adding novel insight into the neurophysiological processes that underlie the therapeutic effects of ibogaine treatment, we improve the understanding of the therapeutic potential of ibogaine, such that this promising treatment can be optimized for the treatment of TBI and related psychiatric disorders.

Keywords: ibogaine, EEG biomarkers, Traumatic Brain Injury

Disclosure: Nothing to disclose.

P683. Functional and Sexual Disability and Quality of Life After One Dose of MM120 (lysergide) in Adults With Generalized Anxiety Disorder

Paula Jacobsen*, Rob Barrow, Craig Conant, Eric Foster, Jamie Freedman, Jamileh Jemison, Sarah Karas, Daniel Karlin, Todd Solomon, Miri Halperin Wernli, Reid Robison

MindMed, New York, New York, United States

Background: Generalized Anxiety Disorder (GAD) is a common psychiatric disorder associated with reduced quality of life (QoL), psychosocial functioning, work productivity, and a wide range of chronic and episodic psychiatric and somatic symptoms. Current treatments ineffectively alleviate symptoms that impact QoL, including those associated with sexual dysfunction. We evaluated whether a single-dose administration of MM120 (lysergide D-tartrate) resulted in durable improvements in functional disability, sexual dysfunction, and QoL measures in participants with GAD.

Methods: This phase 2b (NCT05407064) multicenter, randomized, double-blind, placebo-controlled, dose-finding study enrolled adults aged 18 to 74 years diagnosed with GAD as defined by a Hamilton Anxiety Scale (HAM-A) of ≥20. Participants were randomized equally across 5 arms to receive a single administration of MM120 (lysergide D-tartrate) at a dose of 25µg, 50µg, 100µg, or 200µg, or placebo. The effects of MM120 on functional disability, QoL, and sexual dysfunction were assessed throughout the trial (weeks [w] 1, 2, 4, 8, and 12) by the Sheehan Disability Scale (SDS); the EQ-5D-5L and the Pittsburgh Sleep Quality Index (PSQI); and the Arizona Sexual Experiences Questionnaire (ASEX), respectively. Changes from baseline for each of these measures were analyzed descriptively.

Results: In total, 198 participants were enrolled. MM120 100µg and 200µg demonstrated consistent improvements in functional disability. These improvements were evident at w1 and continued through w12. The MM120 100µg dose demonstrated an optimal level of clinical activity on the primary endpoint, the Hamilton Anxiety Scale (HAM-A), with placebo-adjusted improvements of 7.17 (-14.6 MM120 vs -7.4 placebo) and 6.9 (-15.1 MM120 vs -8.2 placebo) points at w1 and w12, respectively. Similar improvements were observed with MM120 200µg. On the EQ-5D-5L, MM120 100µg demonstrated placebo-adjusted improvements in utility index of 0.111, 0.081, and 0.116 points at w4, w8, and w12, respectively. On the visual analog scale (VAS) of the EQ5, a measure of the participant’s self-rated health, MM120 100µg demonstrated placebo-adjusted improvements of 4.02, 5.41, and 6.04 points at w4, w8 and w12, respectively. Mean PSQI improved at wQ4, w8, and w12 across all groups, including placebo. At w12, there was a considerable decrease from baseline in the proportion of male participants who reported sexual dysfunction with MM120 100µg (29.2% at baseline vs 10% at w12) and 200µg (41.7% at baseline vs 0 at w12) vs placebo (15.4% at baseline vs 12.5% at w12). Similar decreases from baseline were observed in the proportion of female participants who reported sexual dysfunction at w12 in the MM120 100µg (75% at baseline vs 46.2% at w12) and 200µg (57.1% at baseline vs 33.3% at w12) groups vs placebo (50% at baseline vs 33.3% at w12).

Conclusions: Overall, QoL measures demonstrated a clinically significant dose-dependent and durable change in response to a single-dose treatment of MM120. Since current medications often negatively impact QoL, these data represent a significant distinction from existing standard-of-care treatments.

Keywords: GAD, Psychedelics, generalized anxiety disorder, Quality of LIfe (QoL), Anxiety and Depression

Disclosure: MindMed: Employee (Self).

P684. Results of a Phase 1 Clinical Trial to Determine the Safety and Tolerability of the Psychedelic Drug Mebufotenin Administered via Intravenous Injection (GH002) in Healthy Volunteers

Velichka Valcheva*, Theis H. Terwey, Emese Gulyás-Hudák, Viktoria McDonald, Ronan Kelly, Peter Dogterom, Khalid S. Abd-Elaziz

GH Research, Dublin 2, Ireland

Background: Mebufotenin, also known as 5-methoxy-N,N-dimethyltryptamine (5-MeO-DMT), is a potent psychedelic drug belonging to a group of psychoactive indole-alkylamine drugs. It acts as a non-selective serotonin (5-HT) agonist with highest affinity for the 5-HT1A receptor subtype. Early-phase clinical trials of mebufotenin administered via pulmonary inhalation (GH001) in healthy volunteers and patients with treatment-resistant depression demonstrated that GH001 administered as single doses (dose range 2 − 18 mg) or as an individualized dosing regimen (IDR) of up to three escalating doses (6, 12, and 18 mg) is well tolerated, with an ultra-rapid onset (generally within seconds) and a short duration of psychoactive effects. The unique IDR ensures maximum opportunity for response in each individual. An intravenous (i.v.) formulation of mebufotenin (GH002) is also being developed. This Phase 1 clinical trial investigated the safety and tolerability of GH002 in healthy volunteers.

Methods: This Phase 1 clinical trial enrolled 64 healthy volunteers aged 18 − 45 years who were in good physical and mental health without previously diagnosed psychiatric disorders (clinical trial identifiers: ClinicalTrials.gov, NCT05753956; EudraCT, 2022-002620-13). This trial consisted of two parts. Part A had a single-dose, double-blind, placebo-controlled, randomized design. Single ascending doses of GH002 in the dose range of 0.25–10 mg or placebo (randomized 3:1) were administered via i.v. bolus injection to seven cohorts with eight healthy volunteers in each cohort. Part B had a multiple-dose, open-label, nonrandomized design in which GH002 was administered via i.v. bolus injection as an IDR of up to three escalating doses (2, 4, and 6 mg) to one cohort of eight healthy volunteers. The IDR was administered on a single day with a scheduled 1-hour interval between doses. GH002 was administered without additional mandated visits for structured psychological support before or after dosing. The primary endpoint of this trial was to determine the safety and tolerability of GH002 administered as a single dose or as an IDR.

Results: GH002 was well tolerated in healthy volunteers at all doses tested, without occurrences of severe or serious adverse events. In Part A, treatment-emergent adverse events (TEAEs) were observed in 22 of the 42 participants (52.4%) who received GH002 and in 5 of the 14 participants (35.7%) who received placebo. In Part B, TEAEs were observed in 4 of 8 participants (50%). The incidence of TEAEs did not appear to be dose-dependent in either part of the trial. Of the observed TEAEs in this trial, 98.2% were of mild severity and 1.8% were of moderate severity. In Part A, of the 42 participants who received GH002, the TEAEs reported in more than one participant were fatigue (six participants), nausea (five participants), dizziness (four participants), abdominal pain and vomiting (each in three participants), emotional distress, head discomfort, headache, pain in extremity, muscle spasms, and grunting (each in two participants). In Part B, the only TEAE reported in more than one participant was headache (two participants). No noteworthy changes in vital signs were observed except for temporary, non–clinically relevant increases in heart rate and blood pressure shortly after GH002 administration. No clinically relevant changes were observed for any cohort in the safety laboratory analyses, cardiac monitoring (electrocardiogram and continuous Holter), or any measures of cognitive function. With the exception of changes associated with the TEAEs of emotional distress and poor quality sleep, no clinically relevant changes in psychiatric symptoms scales were observed.

Conclusions: Overall, this trial demonstrated that GH002 was well tolerated in healthy volunteers at the investigated single-dose levels and in the IDR. These results warrant further investigation in patient populations.

Keywords: mebufotenin, healthy volunteers, Safety, 5-methoxy-N,N-dimethyltryptamine, psychedelic drugs

Disclosure: GH Research: Employee (Self).

P685. Systematic Review of Genetic Markers of Serotonergic Psychedelics Exposure

Andreea Chiorean*, Helena K. Kim, Brett D. M. Jones, Stanley Wong, Ishrat Husain

University of Toronto, Centre for Addiction and Mental Health, Toronto, Canada

Background: Serotonergic psychedelics are serotonin 2A receptor (5HT2A) agonists or partial agonists that induce cognitive, perceptual, and emotional changes. There is increasing evidence supporting serotonergic psychedelics as potential therapeutics for a range of neuropsychiatric disorders, including mood and anxiety disorders. Emerging evidence suggests that the mechanisms of action of these compounds likely involve multiple pathways, contributing to variability in their hallucinogenic and therapeutic effects. Individual variability in genetic expression may in part explain different treatment outcomes in psychiatric disorders. Understanding the genetic pathways implicated in psychedelic treatment will further enhance our understanding of the mechanisms underlying therapeutic effects and support the identification of markers that can be explored as therapeutic targets or predictors of response. The goal of the current systematic review is to identify and synthesize existing literature evaluating (1) genetic markers as predictors of psychedelic treatment response or (2) change in genetic markers after psychedelic exposure or treatment.

Methods: A literature search was performed on PubMed/Medline, Embase, PsychINFO, and Web of Science from inception to November 2023 (PROSPERO CRD42023483144). Peer-reviewed publications in English investigating genetic markers in relation to serotonergic psychedelic exposure or response prediction were included for review. Given the relative novelty of contemporary psychedelic trials, we hypothesized there will be limited clinical genetic studies on human participants. Thus, we extended our inclusion criteria to in vitro, animal, and human studies. A qualitative synthesis of the sample models studied, psychedelics examined, outcomes examined, reported genetic changes and genetic predictors of psychedelic treatment outcomes will be performed. For papers published prior to the completion of the Human Genome Project (2003), a narrative summary will be completed as a historical framework of genetic studies in psychedelic research prior to the use of more modern molecular techniques. If there are five or more papers examining the same genetic marker, same psychedelic type, and the same sample, a meta-analysis will be performed using a random-effects meta-analysis with the Comprehensive Meta-Analysis software. Heterogeneity will be assessed by the χ2 test and the I² statistic; an I² > 50% will be considered as significant heterogeneity. The literature search will be re-run before the final analyses to retrieve the most recent studies.

Results: In total, 2714 unique articles were identified, of which 293 were included for full-text review. Of these papers, 138 papers met inclusion criteria to date, with 67 of 138 being published prior to 2003 and included in the narrative summary only. Of the remaining papers, 23 had examined genetic predictors, 35 examined changes in genetic expression and 13 had data available for both.

Summary statistics for papers published after 2003 are described and include 7 human studies, 48 animal studies, and 24 cell line studies, where several papers included more than one sample type (N = 6) or psychedelic type (N = 13). Gene knockout studies (N = 17) were considered predictor studies and included: 5HT2A, SIGMAR1, GNAQ, GRM2, PPP1R1B, ARRB2, CREB. The included studies investigated the following psychedelics using modern molecular techniques: LSD (N = 30), psilocybin (N = 14), DOI (N = 15), 5-MeO-DMT (N = 12), DMT (N = 10), mescaline (N = 4), ayahuasca (N = 3), 25B-NBOMe (N = 2), 4-OH-DiPT (N = 1), 5-MeO-AMT (N = 1), peyote extract (N = 1), harmine (N = 1) and cacti (active component mescaline; N = 1). Genetic variation and gene expression levels of genes associated with psychedelic treatment were examined and identified as psychedelic receptor targets (5HT1A, 5HT2A, 5HT2B, 5HT2C, SIGMAR1, GRM2, NTRK2), genes associated with neuroplasticity (NPTN, BDNF, NEGR1, PSD95, CFOS, CEBPB, ARC, EGR1, EGR2, SYN2, SYN3), downstream inflammatory gene expression (CXCL8, ICAM-1, VCAM-1, CSF2, IL1B, IL6, IL5, IL13, IL10, TNF), cellular transport (SLC6A2, SLC6A3, SLC6A4, SLC22A1, SLC22A3), and metabolism (CYP2D6, CYP2C19) of psychedelics.

There was one genetic marker that was examined by at least five studies using serotonergic psychedelics in the same sample. Five rodent studies (n = 5) all reported increased expression of c-fos mRNA in rat brain tissue following the administration of psychedelics. This marker will be further considered for meta-analysis. Data extraction and synthesis of the psychedelic genetic marker data are ongoing with more detailed synthesis to be completed. When available, final results and curated genotypic database will be published to contribute to collaborative data sharing efforts.

Conclusions: To our knowledge, this is the first review to synthesize all available genetic marker data of psychedelic exposure and genetic predictors of psychedelic treatment response. By extending study samples to include human, animal, and in vitro models, we are able to offer a comprehensive examination of both upstream and downstream genetic markers associated with psychedelic exposure including multiple levels of evidence. The findings of this review will inform genetic targets for future studies and identify predictive markers of psychedelic treatment outcomes to advance precision psychiatry.

Keywords: Psychedelics, Genetic biomarker, gene expression, systematic review

Disclosure: Nothing to disclose.

P686. Psilocybin for Treatment-Resistant Major Depression – Results From an Academic Phase 2b Randomized-Controlled Trial

Lea Julia Mertens*, Michael Koslowski, Manuela Brand, Ricarda Evens, Laura Kärtner, Tomislav Majić, Dennis Scharf, Moritz Spangemacher, Max Wolff, Gerhard Gründer

Central Institute of Mental Health, University of Heidelberg, Mannheim, Germany

Background: ilocybin is intensively researched in the treatment of depressive disorders with promising results (Goodwin et al., 2022), although methodological limitations do not allow firm conclusions and warrant further research.

Methods: Aim of this bicentric, double-blind, phase 2b, active placebo-controlled trial (EudraCT: 2019-003984-24; NCT04670081) was to investigate the efficacy and safety of a high-dose oral psilocybin (25 mg) administered in a psychotherapeutic context in 144 adults of both sexes with treatment-resistant depression (TRD) (Mertens et al., 2022). The trial was conducted from July 2021 until February 2024 at two German trial sites (follow-up still ongoing). Patients were randomly assigned to one of four treatment arms, receiving 1) placebo (100 mg nicotinamide) first and a high-dose psilocybin (25 mg) second; 2) a low-dose psilocybin (5 mg) first and a high-dose second; 3a) a high-dose psilocybin first and a low-dose second; 3b) a high-dose psilocybin at both sessions, over a period of 13 weeks. The two dosing sessions were separated by 6 weeks and embedded in 7 psychotherapeutic preparatory and integration sessions (lasting up to 2 hours each). The primary endpoint was treatment response, defined as a minimum of 50% reduction in symptoms, measured with the Hamilton Rating Scale for Depression (HAMD-17; range: 0-52), six weeks after the first dose (i.e. one day before the second dose). Secondary endpoints were at 1 week after the first dose, 1 and six weeks after the second dose as well as long-term follow-ups after six and 12 months; secondary efficacy parameters were change from baseline on the HAMD-17 and rate of remission (Mertens et al., 2022).

Results: 144 patients (59% male; mean age: 42.6 years, SD = 10.83; 98% white) were randomized into the trial, of which 143 received the first dose, 137 completed the primary endpoint by protocol after six weeks and 137 received both dosings. The mean HAMD-17 scores ranged from 21.46 to 22.40 in all treatment groups at baseline, indicating a similar and moderate depression level. Of those patients who received 25 mg psilocybin first, 8 (17 %) classified as responders at six weeks, while in the 5 mg psilocybin arm 6 (13 %) and in the active placebo arm 4 patients (9 %) met the response criterion. A logistic regression model on response rates after six weeks, tested at one-sided significance levels of 0.025, revealed no significant differences between nicotinamide and the 25 mg dose (z = 0.99, p = 0.162) and the 5 mg dose (z = 0.41, p = 0.342); however, one week after the first dose, there were significantly more responder in the 25 mg group (34%) when compared to nicotinamide (6%; z = 3.02, p = 0.001, Odds Ratio = 7.60 [95% Confidence Interval: 2.29 – 34.75] and the 5 mg group (10%; X² = 7.08, p = 0.004).

Mean changes from baseline on the HAMD-17 after one week and six weeks were -7.15 and -6.06 for 25 mg, -3.81 and -3.55 for 5 mg and -2.30 and -1.12 for the nicotinamide group respectively. An Analysis of Covariance (ANCOVA) was calculated on the HAMD-17 change from baseline scores one and six weeks after the first dose, with HAMD17 baseline score and trial centre added as covariates and tested at one-sided significance levels of 0.025. The results showed that the 25 mg group had a significantly larger decrease in depression scores after one and six weeks than the nicotinamide (1 week: b = -4.87, p < 0.001, 6 weeks: b = -4.83, p < 0.001) and the 5 mg group (1 week: b = -3,54, p = 0.002, 6 weeks: b = -2.65, p = 0.018) with the latter not differing significantly from nicotinamide at either timepoint (1 week: b = -1.33, p = 0.138, 6 weeks: b = -2.18, p = 0.047).

(Temporary) worsening of suicidal ideations after baseline until six weeks after the first dose (intervention phase 1) and until six weeks after the second dose (intervention phase 2) occurred in all treatment groups with no significant group differences (X² = 0.007, p = 0.997 and X² = 0.82, p = 0.85 for intervention phases 1 and 2 respectively). Suicidal behavior in form of preparatory actions occurred in 1 person in the nicotinamide arm during intervention phase 1 (i.e. before the second dose). After the second dose, 2 people performed preparatory actions and one aborted suicide attempt occurred.

Serious Adverse Events (SAEs) rated as unrelated to the study drug occurred in 2 patients in form of a colon cancer diagnosis and worsening of depression requiring hospitalization in the weeks after dose 2; SAEs evaluated as related to the study drug occurred in 2 patients in form of a hypertensive urgency during the dosing session and worsening of anxiety/panic paired with dissociative symptoms 2 days after dosing.

Conclusions: Psilocybin when administered in a psychotherapeutic setting appears to be a safe treatment option, although dynamic changes in depression severity and suicidality should be carefully monitored. While the primary endpoint in terms of categorical response was not met, the results still provide strong evidence for the efficacy of a high-dose psilocybin (25 mg) in reducing depressive symptoms in patients with TRD when administered in a psychotherapeutic setting with the strongest effect after one week.

Keywords: Psilocybin, TRD, Psychedelic therapy, Human Clinical trial, Major Depression Disorder

Disclosure: Nothing to disclose.

P687. LSD Base and Tartrate Bioequivalence and Determination of Absolute Bioavailability in Healthy Participants

Denis Arikci, Friederike Holze*, Lorenz Müller, Patrick Vizeli, Dino Luethi, Deborah Rudin, Matthias Liechti

University Hospital Basel, University of Basel, Switzerland, Basel, Switzerland

Background: Lysergic acid diethylamide (LSD) is currently being investigated as novel treatment for various psychiatric and neurological illnesses and is likely to enter phase 3 trials for the treatment of generalized anxiety disorders at the end of 2024. LSD occurs in different crystalline drug forms and different LSD formulations (base or tartrate, oral or intravenous) that are currently being used in research and by recreational drug users. It is still unclear if LSD base and tartrate formulations are equivalent. Additionally, the absolute oral bioavailability of LSD is unknown. Furthermore, it has previously been hypothesized to have slow receptor kinetics. The present study aimed to determine the bioequivalence of oral LSD base and tartrate formulations and define LSD’s absolute oral bioavailability. The study was registered at clinicaltrials.gov (NCT04865653).

Methods: We conducted a randomized, double-blind, placebo-controlled, 5-period cross-over study in 20 healthy participants (10 female, 10 male) investigating three different oral formulations of LSD at equivalent LSD base doses to probe bioequivalence and an intravenous formulation to determine the absolute oral bioavailability (BA). Conditions included (1) a drinking solution of LSD base (“oral base”, 83.11 µg) in 96% ethanol, (2) a watery drinking solution containing LSD tartrate (“oral tartrate”, equivalent to 81.0 µg LSD base), (3) a novel solid oral formulation in the form of a rapid dissolving tablet containing LSD base (“RDT”, 80.5 µg), (4) an intravenous formulation of LSD tartrate (“IV”, equivalent to 81 µg LSD base) and (5) corresponding placebos. We assessed pharmacokinetic parameters, acute subjective effects using visual analog scales (VAS) and the 5 Dimension of Altered State of Consciousness scale (5D-ASC), autonomic effects and adverse effects up to 24 h. For the bioequivalence analysis maximal concentrations (Cmax) and areas under the concentration-time curves (AUCs) were determined using non-compartmental analysis according to regulatory guidelines (within an acceptance interval of 80-125%) in Phoenix WinNonlin. Peak values (Emax) were calculated for repeated pharmacodynamic measures and were analyzed using repeated-measures of variance analysis (ANOVA) with drug as the within-subjects factor, followed by Tukey’s post hoc test. The criterion for significance was p < 0.05.

Results: All oral formulations were bioequivalent (AUC, base/tartrate: 90%CI 92-110%, base/RDT: 90%CI 92-111%; Cmax, base/tartrate: 90%CI 99-120%, base/RDT: 90%CI 94-114%). BA was ~80% of the IV formulation (AUC oral base/IV: 79.7%, oral tartrate/IV: 80.5%, RDT/IV: 80.7%). In line with the bioequivalence, acute subjective effects did not differ statistically between oral formulations. In contrast to oral formulations, the IV formulation produced higher “any drug effect” (all p < 0.01), “good drug effect” (all p < 0.05), “ego dissolution” (all p < 0.05), “experience of unity” (all p < 0.5) and “insightfulness” (all p < 0.01). Furthermore, the IV formulation produced greater “anxiety” when compared with oral formulations (oral tartrate and RDT p < 0.001, oral base p < 0.01), “nausea”, and “bad drug effect” when compared with oral base (p < 0.01 and p < 0.05, respectively) and tartrate (both p < 0.01). All oral formulations produced small to moderate increases in cardiovascular stimulation compared with placebo. The IV formulation moderately increased heart rate compared with oral formulations (oral base p < 0.05, RDT p < 0.01, oral tartrate p < 0.001) and placebo (p < 0.001). All oral formulations produced comparable alterations in state of consciousness reflected by the 5D-ASC total score, while LSD i.v. produced significantly higher responses compared with the RDT and oral tartrate (p < 0.05 and p < 0.01, respectively). The mean ( ± SD) lag between plasma LSD peak and peak “any drug effect” after i.v. administration was 53 ± 43 minutes. Mild acute and subacute adverse effects were comparable for all four LSD conditions.

Conclusions: LSD base and tartrate are bioequivalent when dosed orally at equivalent base doses and share a comparable pharmacodynamic profile and the three oral formulations share comparable pharmacokinetic parameters. The oral BA of LSD is 80% of an IV administration. Dosing with LSD in research with LSD base and LSD tartrate can therefore be considered equivalent. The lag between plasma peak and peak drug effects is in line with LSD’s slow receptor kinetics. These findings are important for the further development of LSD into a medication.

Keywords: Psychedelics, LSD, Pharmacokinetics, Pharmacodynamics, Bioequivalence and Bioavailybility

Disclosure: Nothing to disclose.

P688. Development of Non-Hallucinogenic, Neuroplasticity-Inducing Psychedelic Treatments: Pharmacological Approaches

Orr Shahar*, Alexander Botvinnik, Tzuri Lifschytz, Bernard Lerer

Hebrew University of Jerusalem, Department of Medical Neurobiology, Jerusalem, Israel

Background: Psychedelic treatment requires controlled conditions including hours of specialized supervision. Current research explores harnessing long-term neurobiological effects of psychedelics while reducing short-term psychoactive effects. We focus on modulating the short-term effects of psychedelics (psilocybin (PSIL), 5-MeO-DMT, and DMT) by administering serotoninergic receptor agonists and antagonists. We investigate combined treatment effects using the head-twitch-response (HTR), a rodent correlate of psychedelic activity, the marble-burying test (MBT), and neuroplasticity-related synaptic proteins to assess long-term neurobiological effects.

Methods: Animals: Male C57BL/6J mice (11 weeks old, ~30g) were used for HTR and synaptic protein experiments, and male ICR mice ( ~ 30g) for MBT. Treatments were administered by intraperitoneal injection. Experiments were approved by the Authority for Biological and Biomedical Models, Hebrew University.

HTR Dose Response: Mice implanted with mini-ear magnets were treated with PSIL (0.1-25.6 mg/kg), DMT (1.25-80 mg/kg), and 5-MeO-DMT (1.25-80 mg/kg) immediately before HTR was measured for 20 minutes using a magnetometer.

Pharmacological Modulation of Psychedelic-Induced HTR: Mice were injected with PSIL (4.4 mg/kg), DMT (5 mg/kg), or 5-MeO-DMT (10 mg/kg). Doses were based on human clinical trial doses and our dose-response data. The injections were immediately preceded by administration of serotonin receptor modulators including the 5-HT2A antagonist, M107900 (0.5 mg/kg), the 5-HT1A agonist, 8OH-DPAT (1 mg/kg), the 5-HT1A antagonist, NAD-299 (3 mg/kg), the 5-HT2C antagonist, RS-102221 (4 mg/kg) and the 5-HT2C agonist, WAY-161503 (3 mg/kg).

MBT: Mice were individually placed in a cage containing 20 evenly-spaced marbles after being injected with PSIL (4.4 mg/kg), DMT (5 mg/kg), or 5-MeO-DMT (5 mg/kg), alone or immediately preceded by a serotonin receptor modulator that reduced HTR (5-HT2A antagonist, 5-HT1A agonist). Number of marbles buried after 30 minutes was counted.

Neuroplasticity Markers: Mice were sacrificed 12 days following a single injection of PSIL (4.4 mg/kg), DMT (5 mg/kg), and 5-MeO-DMT (5 and 10 mg/kg). Western blot analysis was performed on the frontal cortex, amygdala, and hippocampus to assess levels of PSD-95, GAP43, synaptophysin, and SV2A.

Results: Dose-Response and Modulation: PSIL, DMT, and 5-MeO-DMT induced a dose-dependent increase in HTR that peaked at 4 min and subsided to baseline by 20 minutes. Significant attenuation of HTR was observed with the 5-HT2A antagonist (p < 0.0001) and the 5-HT1A agonist (PSIL and DMT p < 0.0001, 5-MeO-DMT p = 0.03). The 5-HT1A antagonist significantly increased HTR induced by DMT (p = 0.002) and 5-MeO-DMT (p < 0.0001) but not by PSIL. In contrast, the 5-HT2C antagonist significantly increased HTR induced by PSIL (p = 0.03) but not DMT and 5-MeO-DMT.

MBT: A significant reduction of marbles buried was observed with single administration of PSIL 4.4 mg/kg (p = 0.0017), DMT 5 mg/kg (p < 0.0001), and 5-MeO-DMT 5 mg/kg (p = 0.018). The 5-HT2A antagonist did not itself affect marbles buried; however, the 5-HT1A agonist did reduce the number of marbles buried (p < 0.0001). A significantly greater reduction in marbles buried as compared to the psychedelic alone was observed after combined treatment with DMT and a 5-HT2A antagonist (p = 0.002), DMT and a 5-HT1A agonist (p < 0.0001), 5-MeO-DMT and a 5-HT2A antagonist/5-HT1A agonist (p < 0.0001), and PSIL + 5-HT1A agonist (p = 0.009). Pretreatment with a 5-HT2A antagonist did not alter the reduction of marbles buried induced by PSIL.

Neuroplasticity Markers: Significant increases in synaptic proteins were observed with all the psychedelics over all three brain areas. Nested one-way ANOVA revealed increase in GAP43 (5-MeO-DMT 10mg/kg p = 0.003), PSD95 (PSIL 4.4 mg/kg p = 0.01, 5-MeO-DMT 10 mg/kg p = 0.0002, DMT 5 mg/kg p = 0.02), synaptophysin (5-MeO-DMT 5 mg/kg p = 0.09, 5-MeO-DMT 10 mg/kg p = 0.02, DMT 5 mg/kg p = 0.012), and SV2A (PSIL 4.4mg/kg p = 0.019, DMT 5 mg/kg p = 0.02).

Conclusions: Our study shows that the tryptaminergic psychedelics PSIL, DMT, and 5-MeO-DMT dose-dependently induce HTR, which can be modulated by targeting specific serotonin receptors. Attenuation of HTR by a 5-HT2A antagonist and a 5-HT1A agonist was expected but not interesting discrepancies such as increased HTR induced by a 5-HT1A antagonist in DMT/5-MeO-DMT treated mice but not in PSIL-treated mice, and increased HTR induced by a 5-HT2C antagonist in PSIL-treated mice and not DMT/5-MeO-DMT-treated mice.

The significant reduction in marble-burying observed with all three tryptaminergic psychedelics indicates their potential in treating OCD. Modulators that reduced HTR (5-HT2A antagonist/5-HT1A agonist) and are known to reduce trip effects in humans did not block the reduction in marble-burying but enhanced it. This highlights the therapeutic potential of psychedelics co-administered with a trip-reducing 5-HT2A antagonist or 5-HT1A agonist. Moreover, this finding suggests that a psychedelic trip may not be required for the efficacy of serotonergic psychedelics in OCD.

Increases in synaptic proteins observed 12 days post-injection suggest long-term neurobiological benefits from a single dose of psychedelics and support the potential of psychedelics to promote synaptic plasticity.

Supported in part by Negev Labs and Parow Bio

Keywords: Psychedelics, Neuroplasticity, serotonin receptors, Obsessive Compulsive Disorder, Behavioral Pharmacology

Disclosure: Nothing to disclose.

P689. Single-Dose Psilocybin for U.S. Military Veterans With Severe Treatment-Resistant Depression: Short and Long-Term Follow-Up Outcomes

Mathieu Fradet, Sara Ellis, Anna Donnelly, Michael Ostacher, Scott Aaronson, Trisha Suppes*

Stanford University School of Medicine and VA Palo Alto Health Care System, Palo Alto, California, United States

Background: Major depressive disorder (MDD) and treatment resistant depression (TRD) are among the most burdensome diseases in the general population and among Veterans. A large proportion of patients suffering from MDD are not relieved by conventional treatment options, especially when they suffer from comorbid conditions such as PTSD. Psilocybin has recently emerged as a promising treatment for TRD. We aimed to evaluate the efficacy and safety of psilocybin for the treatment of severe TRD among US military Veterans at short-term (3 and 12 weeks) and long-term follow-up (6, 9 and 12 months).

Methods: Fifteen US military Veterans with severe TRD were recruited in an open-label pilot study at the VA Palo Alto Healthcare System. Each participant received 25 mg of synthetic psilocybin (COMP360) and psychological support. The primary outcome was change in Montgomery-Asberg Depression Rating Scale (MADRS) at 3 weeks post dosing. Secondary outcome on the MADRS included response (50% decrease from baseline) and remission rates ( ≤ 10) as well as self-reported depression symptoms (QIDS), and functional disability (SDS). During short-term follow-up, changes in MADRS, QIDS and SDS were assessed for 14 participants at 3 and 12 weeks.

10 participants agreed to engage in a longer term follow up, for whom MADRS, QIDS and SDS were reassessed at 6, 9 and 12 months post-treatment. Adverse events were recorded during the treatment period and long-term follow-up.

Results: 15 participants were enrolled from March 2022 to February 2023. Total MADRS scores markedly decreased between baseline and week 3 (F (2, 23) = 50.75, p < .0001), with a mean change of -22.93 (SD = 13.47; SE = 2.81; 95% CI, -28.75 to -17.10; Cohen’s d = 0.42). Co-morbid PTSD or gender did not influence the change in MADRS scores (p > 0.05). At week 3, 60% (n = 9) of participants met response criteria and 53% (n = 8) met remission criteria. At 12 weeks, 47% of participants maintained a response (n = 7), and 40% (n = 6) maintained remission.

Among the 10 individuals who participated in long-term follow-up, 80% of participants met response criteria (n = 8) at 6 months, and 50% (n = 5) met remission criteria. At month 12, 50% (n = 5) met response criteria, and 40% (n = 4) met response criteria. As observed for short-term outcomes, co-morbid PTSD or gender did not significantly affect MADRS scores at long-term follow-up (p > 0.05). MADRS scores were notably lower (p < 0.0001) at long-term endpoints than at baseline, including month 6 (M = -26.37, SD = 10.58), month 9 (M = -22.62, SD = 10.61), and month 12 (M = -18.50, SD = 10.61).

There was a notable decrease in QIDS scores from baseline to week 3 and week 12 (F(3, 30) = 38.94, p < 0.0001). For long-term follow-up participants, QIDS score was significantly lower than baseline at month 6 (M = -12.50, SD = 5.67), month 9 (M = -11.13, SD = 5.87), and month 12 (M = -9.88, SD = 5.89). There was also a small but significant decrease in total SDS from baseline to week 3 and week 12 post-treatment (F(3, 26) = 13.65, p < .0001). In our cohort, SDS scores then significantly increased during long-term follow-up from week 3 to month 12 (p = 0.01).

In terms of adverse events, during or immediately after psilocybin intake, 47% of participants (n = 7) experienced mild-to-moderate headaches, 20% (n = 3) experienced nausea and 13% (n = 2) experienced mild back pain. Two participants reported increased psychological distress during dosing, with no increase in suicidal ideation or behavior. No serious adverse event was observed, and no additional adverse event occurred during long-term follow-up.

Conclusions: Surprisingly high rates of response and remission were observed in this first-in-kind open-label study of psilocybin for Veterans with long standing treatment resistant depression. Many showed persistent benefits at long-term follow-up. The study intervention was well tolerated by most participants, with adverse events being transient. While blinded and controlled trials are needed to further assess the efficacy and safety of this intervention, psilocybin therapy holds promise for Veterans suffering from TRD.

Keywords: Depression, Treatment resistant depression, Psilocybin, Veterans, Psychedelics

Disclosure: psilotec: Stock / Equity - Privately Held Company (Self). MindMed: Advisory Board (Self). Compass Pathways: Grant (Self).

P690. Effects of Psilocybin on Brain Circuits Underlying Immediate Improvement in Treatment-Resistant Depression

Xue Zhang*, Emily Zhai, Claire Bertrand, Sara Ellis, Trisha Suppes, Leanne Williams

Stanford University School of Medicine, Palo Alto, California, United States

Background: Recent clinical trials have demonstrated the potential of psilocybin as a fast-acting novel therapeutic for treatment-resistant depression (TRD). However, the precise neural mechanisms underlying its efficacy remain elusive. This study aimed to identify intrinsic brain circuits acutely modulated by psilocybin in TRD and examine its association with treatment outcomes, to provide critical insights for personalized interventions.

Methods: In an open-label trial, 15 adult participants (ages 18-65; 2 female) with TRD were recruited to receive a single administration of psilocybin (25 mg) under supportive conditions. The trial is registered on ClinicalTrials.gov under the identifier NCT04433858. Depression symptom severity was measured using the Montgomery-Åsberg Depression Rating Scale (MADRS) at multiple timepoints, including baseline and Day 1 post-treatment. Intrinsic brain circuits were assessed using resting-state functional magnetic resonance imaging (fMRI) at the same timepoints. Resting-state fMRI data were preprocessed using an established pipeline, and were then used to estimate intrinsic activities and functional connectivities (FCs) using the Brainnetome atlas, which consists of 246 cortical and subcortical regions. Each region was labeled according to Yeo’s 7 networks parcellation. The intrinsic activity of each region was quantified by the amplitude of low-frequency fluctuations (ALFF). ROI-to-ROI pairwise FCs (246 x 246) were evaluated and transformed into Z-scores using Fisher’s z transformation. A paired t-test was conducted on the ALFF of each region and on each pairwise FC between baseline and Day 1 post-treatment. Results for ALFF were corrected for multiple comparisons using false discovery rate (FDR) correction at p < 0.05, while results for FC were thresholded at p < 0.001. Intra-subject correlation was calculated between the brain circuit activity and connectivity that survived correction and the MADRS scores.

Results: Among subjects with quality neuroimaging data at both timepoints (N = 7), psilocybin induced a significant decrease in intrinsic activity in the bilateral mediodorsal thalamus nucleus (MDT; left MDT Z = -4.44, right MDT Z = -4.19), at Day 1 post-treatment relative to pre-treatment baseline. Psilocybin also induced a widespread decrease in functional connectivity (FC), including connectivity within the bilateral MDT (Z = -4.42), between the MDT and the dorsal attention circuit (Z = -3.64), between MDT and the frontoparietal circuit (Z = -3.12), and within the somatomotor circuit (Z = -3.70). All Ps < 0.001. Intra-subject correlation analysis showed that the decrease in activity in MDT at Day 1, as well as the average decrease in FC, was associated with reductions in MADRS (P < 0.05).

Conclusions: These findings suggest that psilocybin’s immediate therapeutic effects may arise from its ability to target and modulate rigid, maladaptive brain circuits. Our finding highlights the need for a precision approach, in which psilocybin will only be applied to individuals with identified circuit deficits, for a better treatment outcome.

Keywords: Treatment-resistant depression, Psilocybin, Brain Circuits, Mechanistic Clinical Trial

Disclosure: Nothing to disclose.

P691. The Non-Hallucinogenic Serotonin 1B Receptor is Necessary for the Persisting Behavioral Effects of Psilocybin in Mice

Sixtine Fleury, Kate Nautiyal*

Dartmouth College, Hanover, New Hampshire, United States

Background: Recent studies have demonstrated the potential of psychedelic therapies for the treatment of psychiatric disorders, and there are a number of clinical trials aimed at investigating the effect of psilocybin on major depressive disorder. The persisting behavioral effects of psychedelic therapies are most commonly attributed to activation of the hallucinatory serotonin 2A receptor (5-HT2A R). However, psilocin, the active metabolite of psilocybin, binds to many serotonin receptor subtypes. Recent studies in rodents suggest some therapeutic effects may not require activation of the 5-HT2AR. Investigating the role of other serotonin receptors is important for understanding psilocybin’s lasting clinical effects. We hypothesize that psilocybin may influence depressive-like behaviors via the serotonin 1B receptor (5-HT1BR), a non-hallucinatory subtype of serotonin receptors previously implicated in mediating depressive phenotypes and neural plasticity.

Methods: We used corticosterone or forced swim stress to induce a depressive-like phenotype in female and male mice, respectively. Psilocybin (5mg/kg) or saline was administered i.p, and head twitch and locomotor responses were measured during the acute drug response (N = 5/group). In the 1-3 days following drug administration, we measured anhedonia in a gustometer assay as well as anxiety-like behavior in the novelty-suppressed feeding (NSF) assay and elevated plus maze (EPM). We performed these experiments in mice lacking the 5-HT1BR and controls to assess the role of 5-HT1B in the effects of psilocybin (N = 11-14/group). In a follow up experiment, we also measured the effect of psilocybin following administration of a 5-HT1BR antagonist (GR-127935 at 10mg/kg; N = 8-11/group). We also assessed neural activity following psilocybin using whole-brain c-fos labeling in mice lacking the 5-HT1B R and controls (N = 7/group).

Results: Using genetic and pharmacology loss-of-function approaches, we find that the 5-HT1BR is required for the behavioral and neural effects of psilocybin in mice. The acute response to psilocybin was measured through increased head twitches and decreased locomotion. We found no sex differences in these acute responses, and no effect of 5-HT1B KO on the head twitch response (main effect of drug: F(1,15) = 88.0, p < 0.001; main effect of genotype: F(1,15) = 0.2, p = 0.637). However, mice lacking 5-HT1BR had a diminished hypolomotor response to psilocybin, suggesting that 5-HT1BR is involved in mediating the acute locomotion effect of psilocybin (drug x genotype interaction: F(1,15) = 3.3, p = 0.002). Interestingly, the persisting behavioral effects of psilocybin were also mediated by 5-HT1BR. Specifically, we found that psilocybin was able to rescue corticosterone-induced anhedonia and anxiety-like behavior in females, but not males. Control females showed increased hedonic responding in the gustometer (main effect of drug: F(1, 160) = 6.6, p = 0.002) and decreased anxiety following psilocybin in the NSF (Chi2 = 11.4, p < 0.005) and EPM (t(29) = 2.23, p = 0.03). These effects of psilocybin were absent in mice lacking 5-HT1BR (main effect of drug in gustometer: F(1, 115) = 0.7, p = 0.3; NSF: Chi2 = 1.9, p = 0.17; EPM: t22 = 0.1, p = 0.9). We saw this same pattern of results when a 5-HT1BR antagonist was administered prior to psilocybin. Next, using a forced swim stressor, we found that we could replicate this effect in males. Finally, using c-fos labeling to assess whole-brain neural activity patterns following psilocybin administration, we saw differential neural activity in mice lacking 5-HT1BR compared to controls in a number of brain regions including the striatum, amygdala, and PFC.

Conclusions: Current work is aimed at identifying the 5-HT1B-responsive circuits through which psilocybin has persisting behavioral effects. Overall, our work suggests that psilocybin induces antidepressant-like effects in mice that are dependent on the non-hallucinatory serotonin 1B receptor.

Keywords: Psilocybin, Psychedelics, 5-ht1b, Anxiety and Depression

Disclosure: Nothing to disclose.

P692. An Open-Label Study of Single-Dose COMP360 Psilocybin for Post-Traumatic Stress Disorder: Safety, Tolerability, and Secondary Efficacy Outcomes

Niall McGowan, James Rucker, Rachel Yehuda, Manish Agrawal, Hollie Simmons, Shriya Das, Guy Goodwin*

Compass Pathfinder Ltd. (a subsidiary of Compass Pathways plc), London, United Kingdom (UK)

Background: Post-traumatic stress disorder (PTSD) is a common psychiatric disorder emerging after exposure to a traumatic event and is associated with severe and debilitating symptoms impacting quality of life. Many patients are not helped by pharmacologic treatments, with SSRI treatment response rates scarcely exceeding 60% and < 30% of patients achieving full remission (1). This unmet treatment need underscores the importance of novel treatment development for a regrettably underserved patient group.

Methods: A 12-week Phase II, multicenter, open-label trial conducted at sites in the UK and US investigated the safety and tolerability of a single dose of 25 mg COMP360 psilocybin with psychological support in adult outpatients with moderate-to-severe PTSD (ClinicalTrials.gov ID: NCT05312151). Participants were recruited from the community and referring healthcare providers. PTSD with at least moderate symptom severity was confirmed using the Clinician Administered PTSD Scale for DSM-5 (CAPS-5; scale range 0-80, score ≥25 required) at Baseline. Participants taking antidepressant or antipsychotic medications at Screening completed medication washout at least two weeks before Baseline. The trial primary outcome was the safety and tolerability of treatment. Improvement in PTSD symptom severity was assessed via change from Baseline on the clinician-rated CAPS-5 at Week 4 and Week 12, and the self-reported PTSD Checklist for DSM-5 (PCL-5) at each study visit. Treatment response (defined by a reduction of ≥15 points in CAPS-5 score) and remission (defined as a CAPS-5 score ≤20) were also assessed. Improvements in functional impairment and quality of life were assessed at Week 4 and Week 12 using the Sheehan Disability Scale (SDS; change from Baseline) and the EQ-5D 5-levels instrument (EQ-5D-5L; index score), respectively.

Results: Twenty-two participants (63.6% female) were enrolled in total, and all completed the 12-week trial. The mean and median [standard deviation (SD)] CAPS-5 score at Baseline was 47.5 [9.78] indicating that most participants had severe symptoms of PTSD; and the mean duration of PTSD amongst participants was 88.2 [68.65] months. There were no reported treatment-emergent serious adverse events. The most common treatment-emergent adverse events (TEAEs) included headache (n = 11; 50.0%), nausea (n = 8; 36.4%), crying (n = 6; 27.3%), and fatigue (n = 6; 27.3%). There were two TEAEs of suicidal ideation that resolved during the study. The first was a moderate and transient event which resolved on administration day in a patient who went on to be a responder, and it was deemed to be related to study drug. The second event was mild and occurred at Week 7 in a non-responder, resolved during the study, and it was deemed to be possibly related to study drug. Both participants had previous history of suicidality as measured by the Columbia-Suicide Severity Rating Scale. Treatment was associated with a clinically meaningful reduction in CAPS-5 score from Baseline at Week 4 (-29.9 [14.06]) and Week 12 (-29.5 [15.43]). This translated to an 81.8% response rate and a 63.6% remission rate at Week 4. Response and remission rates at Week 12 were 77.3% and 54.5%, respectively. PTSD symptom reduction was rapid, and durable indicated by a change from Baseline PCL-5 score (52.1 [13.68]) at Day 2 (-33.5 [14.32]) which was sustained to Week 12 (-34.3 [18.13]). Participants showed an improvement in functional impairment over the 12 weeks of the study; from a mean SDS total score of 22.7 [5.38] at Baseline, there was a -11.7 [8.41] point reduction at Week 4 and a -14.4 [8.21] reduction at Week 12. Quality of life scores improved throughout the study, indicated by an EQ-5D-5L index score of 0.51 [0.287] at Baseline increasing to 0.73 [0.272] at Week 4 and 0.78 [0.269] at Week 12.

Conclusions: Treatment with 25 mg COMP360 was generally well-tolerated and no serious safety concerns were observed. Participants experienced a clinically meaningful and durable reduction in clinician-rated PTSD symptoms, and rapid self-reported improvement by Day 2. Response and remission rates were high and sustained at the end of the trial. Functional impairment was reduced and quality of life scores improved across the 12 week study period. Results should be interpreted within the context of the modest sample size and the open-label design with no control comparator. Whilst the results are promising, larger well-controlled studies are required to inform the viability of COMP360 psilocybin as a potential efficacious treatment for PTSD.

Reference:

1. Kelmendi et al. (2016). Eur J Psychotraumatol, 7:31858

Keywords: PTSD, Psilocybin, Open Label Trials, Clinical trial, Safety/tolerability

Disclosure: Compass Pathways plc: Stock / Equity - Publicly Traded Company (Self). Takeda Pharmaceutical Company: Consultant (Self).

P693. The Feasibility, Safety, and Potential Efficacy of Psilocybin With Psychological Support as an Adjunct Treatment for Anorexia Nervosa in Adult Females

Hannah Douglass*, Meg Spriggs, Kate Godfrey, Jennifer Danby, Frederico Magalhaes, Kirsty Alderton, Lauren Macdonald, Stephanie Archer, Kirran Ahmad, Tim Read, Fernando Rosas, Allan Blemings, Adele Lafrance, Dasha Nicholls, David Erritzoe, Rebecca Park, David Nutt, Robin Carhart-Harris

Imperial College London, London, United Kingdom

Background: Anorexia nervosa (AN) is a life-threatening eating disorder (ED) with a mortality of approximately 10%, the highest of all psychiatric conditions. It is characterised by severely low body weight resulting from stringent restriction of food intake due to a pathological fear of weight gain, and the overvaluation of body weight and shape and its relevance to self-worth. Although there are several approved psychological therapies for the treatment of adults with AN, with pharmacological interventions used to treat comorbid conditions, their effectiveness is reported to be limited in maintaining remission. Psilocybin with psychological support has shown promise as an alternative treatment for a range of psychiatric conditions, including those comorbid with AN, with indications of safety and potential efficacy demonstrated by existing literature. Here, we present data collected from the “Psilocybin as a Treatment for Anorexia Nervosa” clinical trial (NCT04505189), which had a recruitment target of 20 participants and aimed to investigate the preliminary safety, feasibility, and potential efficacy of this intervention.

Methods: Eligibility criteria included: female (at birth), 21-65 years old, primary diagnosis of AN present for ≥3 years, body mass index (BMI) ≥14kg/m2 and have found previous or current treatments ineffective at maintaining remission. Over a 6-week period, participants were administered three doses (1 mg, 25 mg, 25 mg) of the investigational drug COMP360 (COMPASS Pathways’ proprietary synthetic psilocybin formulation) in a fixed-order, single-blind, within-subject design. The 1 mg dose acted as a sub-perceptual “placebo”. The pharmacological intervention was conducted in a therapeutic setting, with music-listening, and was accompanied by psychological support provided by two therapists. Each dosing session was encompassed by a preparation session the day prior, and an integration session the next day. Baseline assessments were conducted one day prior to the first dosing session, with the final in-person visit two weeks after the third dosing session acting as a key endpoint. Feasibility, recruitment, and retention rates of participants were assessed. Safety was indexed through the monitoring of adverse events (AEs) throughout the 6-week active period of the study, and subsequent remote 12-month follow-up which comprised questionnaires, interviews, and behavioural tasks. It was hypothesized a] that the severity of global ED psychopathology would be improved long-term at the final visit, 3-month follow-up and 6-month follow-up, as assessed by the Eating Disorder Examination interview (EDE); b] these improvements would be greater following the two 25 mg doses of psilocybin, than following the 1 mg session, assessed using the self-report EDE questionnaire (EDE-Q); c] that the Readiness and Motivation Questionnaire (RMQ) would demonstrate reduced “precontemplation” from baseline to final visit, which would be reduced to a greater degree by the 25 mg doses.

Results: Of 161 individuals that were sent a participant information sheet, 100 (62.1%) attended a remote video screening call and 21 (13.0%) were recruited. There was a retention rate of 95.2%, with twenty participants completing the study up to the final visit. The majority of these participants were white British (76.2%) and had an undergraduate degree or higher (90.5%). Of those recruited, 6 were assessed as having comorbid general anxiety disorder and 5 with major depressive disorder at in-person screening. AEs were reported by all participants, with 107 documented in total, headaches (37.3%) and nausea (25.2%) being the most common. Two suicide attempts (at seven and nine months) by one participant were documented as serious AEs and were assessed as unlikely to be related to the drug. A linear mixed-effect model demonstrated enduring reductions in mean global EDE scores between baseline and the final visit (β = -1.00, p < 0.0001, d = 1.14), 3-month follow-up (β = -1.26, p < 0.0001, d = 1.15) and 6-month follow-up (β = -1.15, p < 0.0001, d = 1.02). Compared to baseline, mean RMQ precontemplation scores were seen to be reduced from two weeks after the second dosing session (β = -9.80, p = 0.02, d = 0.879), and were sustained at most timepoints up to the 12-month follow-up (β = -10.57, p = 0.009, d = 0.610). This indicates that participants, on average, had increased desire to undertake actions to recover from AN. Repeated measures ANOVAs demonstrated that there were no significant differences in the incremental change in global EDE-Q and RMQ precontemplation scores between doses, indicating that all doses likely contributed to the reported effects.

Conclusions: This study provides preliminary evidence of the feasibility, safety, and potential efficacy of psilocybin with psychological support as an adjunct treatment in adults with AN. Despite two serious AEs, the AEs documented here are in-line with those previously reported in clinical psilocybin research, indicating that those with AN are not at elevated psychological or physical risk. Potential efficacy was seen for the combined effect of psilocybin and integrated psychological support in reducing severity of ED psychopathology and improving motivation to engage in recovery. Further higher-powered and extended randomised-controlled trials are warranted to elucidate the individual impact of the pharmacological intervention and accompanying psychological support on the reported therapeutic effect.

Keywords: Psilocybin therapy, Psychedelics, anorexia nervosa, Eating disorders, clinical psychopharmacology

Disclosure: Nothing to disclose.

P694. Preliminary Feasibility of a Novel Medication-Assisted Psychotherapy: Ketamine Enhanced Radically Open Dialectical Behavior Therapy (RO DBT) for Treatment Resistant Depression

Kirsten Gilbert*, Jason Xie, Molly Steinhoff, Nuri B. Farber

Washington University in St. Louis, St. Louis, Missouri, United States

Background: Ketamine is an effective means of treating treatment resistant depression (TRD), a disorder with high psychiatric comorbidity, premature mortality, and high societal burden. However, ketamine is fast-acting and transient, which has led to growing interest in combining it with psychotherapy to enhance its’ long-term benefits. A novel psychotherapy demonstrating effectiveness in TRD is Radically Open Dialectical Behavior Therapy (RO DBT). RO DBT targets the underlying mechanism of overcontrol, a phenotype characterized by cognitive inflexibility/rigidity, extreme concern for errors, checking behaviors, and perfectionism by increasing openness, flexibility, and reward experiences in social connections. Given ketamine improves social functioning and theoretically increases openness, ketamine may provide fast-acting change to catalyze longer-term change in RO DBT for individuals with overcontrol. The current study tests preliminary feasibility of ketamine-enhanced RODBT in targeting mechanistic processes of overcontrol, flexibility, reward processing and social functioning in adults with TRD.

Methods: The pilot study (Clinicaltrials.gov ID# NCT06138691) is ongoing with 9 overcontrolled adults (n = 6 biological female) with TRD enrolled. Participants complete a behavioral and electroencephalogram (EEG) assessment at baseline, followed by four weeks of twice-weekly intravenous ketamine infusions and four months of RO DBT (individual and group therapy, each once weekly). Ketamine sessions occur one day prior to individual and group therapy sessions. Participants also complete a mid-treatment and post-treatment EEG and behavioral assessment.

Results: Three participants dropped out of treatment due to time constraints (n = 2) and disliking ketamine (n = 1). Six participants have completed the ketamine portion and mid-treatment assessment and are continuing RO DBT therapy. Preliminary findings demonstrate that following individual ketamine infusions, in addition to expected decreases in sadness and anxiety (p < .001), participants also experienced momentary decreases in social isolation (p = 0.002) and increases in openness, flexibility, and a willingness to try new things (p = 0.046). After one month of ketamine-enhanced RO DBT, symptoms of depression (p = 0.008) and anxiety (p = 0.047) significantly decreased, with mechanistic change demonstrated via significant increases in self-reported psychological flexibility (p = 0.023) and anticipatory and consummatory reward responding (p = 0.001). One month of ketamine-enhanced RO DBT also demonstrated nonsignificant increases in social connectedness (p = 0.18) and quality of life (p = 0.19), nonsignificant decreases in loneliness (p = 0.19) and maladaptive perfectionism (p = 0.07) as well as a nonsignificant increase in the ERN, a neural marker of error and performance monitoring (p = 0.051).

Conclusions: Preliminary findings demonstrate ketamine-enhanced RO DBT to be feasible for overcontrolled individuals with TRD, but time burden might necessitate treatment modification. In addition to decreasing psychiatric symptoms, ketamine-enhanced RO DBT demonstrates potential at targeting mechanistic processes characteristic of overcontrol, including psychological inflexibility, reward responding, social connectedness, perfectionism, and neural performance monitoring. The sample size is small, but data collection is ongoing. Ketamine-enhanced RO DBT shows promise as a personalized treatment for overcontrolled individuals with TRD that could provide fast-acting benefit that catalyzes durable and longer-term mechanistic change.

Keywords: IV- Ketamine, Treatment Resistant Depression, Medication Assisted Treatment, Psychotherapy, Psychological inflexibility

Disclosure: Upaya Counseling: Speakers Bureau (Self).

P695. Developing Animal Models of Psychedelic Therapy: Methods and Mechanisms for Reversal of Trauma-Induced Behavior in Rats

Kate Lawson*, Hannah Vu, Naomi Cole, Christina Ruiz, Stephen V. Mahler

University of California, Irvine, Irvine, California, United States

Background: Post traumatic stress disorder (PTSD) is a common mental health disorder that affects millions of Americans each year. The current few pharmacological treatments available for individuals suffering from PTSD are often insufficient or ineffective. Therefore, it’s necessary to develop new approaches that are persistent and effective for broad patient populations. In the last few years, psychedelic drugs like LSD, psilocybin, and DMT have re-emerged in psychiatry, with apparent potential for treating PTSD and other psychiatric disorders. Clinical studies have reported that these drugs reduce psychiatric symptoms almost immediately and can remain effective for weeks to months after acute administration. Yet mechanistic studies cannot be conducted in humans, and double blind dosing is also unlikely given the profound subjective effects of these drugs.

Therefore, despite their obvious limitations, translationally relevant animal models of “psychedelic therapy” are required. Here we validate such a model in rats using the serotonin 2A (5HT-2A) receptor agonist 2,5-dimethoxy-4-iodoamphetamine (DOI), testing its behavioral and neural consequences in a rat model of traumatic learning.

Methods: We used stress-induced enhanced fear learning (SEFL), a well-validated rat model of adulthood trauma, to investigate the neural mechanisms underlying stress-induced excessive fear and potential reversal by psychedelics. In our SEFL procedure, male and female adult Long Evans rats undergo 15 repeated, uncued, unpredictable shocks on day 1 in Context A (1 mA, 1 s for males, 0.75 mA, 1 s for females). Control rats are exposed to Context A in the same way, but without shocks. On days 2-5, all rats were returned to Context A for 30 min each day, without shocks (context extinction). The SEFL test occurred on day 9 in a distinct Context B, where 210s after placement in the chamber, rats were given one shock (1 mA, 1 s for males, 0.75 mA, 1 s for females). Freezing and darting behavior was scored manually for the first 3 mins in Context B, and the 3 mins following the single shock that was delivered.

We sought to determine whether a single, acute dose of the 5-HT2A agonist DOI was capable of persistently decreasing the impact of prior trauma upon subsequent fear in response to a modest threat. We also examine whether the environment in which DOI was administered was relevant to this “therapeutic effect.” We do so with a variety of behavioral tests at various time points, both before and after acute DOI treatment, including locomotion, head shakes, and ultrasonic vocalizations. We also measure brain activity elicited by modest threat, comparing rats that previously received DOI to those that did not.

Results: We find that a single dose of the psychedelic 5-HT2A agonist DOI persistently reverses trauma-induced fear hypersensitivity in both sexes (DOI lowers freezing relative to saline control; p = 0.00683). We also find evidence for “set and setting” influencing the therapeutic and neural activity effects of the psychedelic, supporting claims stemming from human studies, but which were previously untested in animals.

Conclusions: Results suggest that rodent models of “psychedelic therapy” can be improved, and that the details of these experiments can markedly influence the behavioral and neural outcomes measured, and thus influence their translational relevance.

Keywords: Psychedelics, Acute Stress, stress-enhanced fear learning (SEFL)

Disclosure: Nothing to disclose.

P696. Informed Consent in Psychedelic-Assisted Therapy

Mazdak Bradberry*, Paul Appelbaum, Natalie Gukasyan

Columbia University, New York, New York, United States

Background: Evidence continues to accumulate for the efficacy of therapies involving psilocybin and MDMA, which we describe collectively as “psychedelic therapies,” across a range of clinical indications. The adaptation of these therapies into clinical practice is complicated by several cultural, ethical, and clinical factors, such as widespread extant “underground” psychedelic therapy, diverse cultural frameworks and long-standing indigenous practices in psychedelic therapies, significant public optimism about the effectiveness of these therapies, drug-induced changes in interpersonal boundaries, and uncertain risks of psychedelic experiences in some individuals. Rigorous standards of informed consent will help to mitigate the potential harms arising from these issues and thus maximize the safety and efficacy of psychedelic treatments in clinical settings.

Methods: We reviewed recent and historical studies on the medical and psychiatric risks and benefits of psychedelic therapies in clinical as well as in naturalistic, non-clinical settings. We explicitly considered the diverse settings, regulatory frameworks, and systems of knowledge in which psychedelic therapies may be expected to occur in the coming years, especially in settings outside of clinical research. We reviewed the fundamentals of informed consent, including the capacity to consent to treatment, and adapted existing informed consent frameworks to define a set of essential elements in informed consent disclosures for psychedelic therapies.

Results: Psychedelic use occurs in a variety of settings including professional therapeutic, supervised non-professional, and recreational settings. Use in each of these settings is associated with varying degrees of therapeutic efficacy and risk of adverse outcomes. Emerging state legal frameworks exist for psilocybin outside traditional clinical settings. Across studies in clinical and non-clinical settings, we found that the risk of ineffective treatment or clinical worsening after a psychedelic treatment or experience was approximately 10-15%. In some of these individuals, symptoms such as worsened or less stable mood, sleep difficulties, and existential distress persisted for weeks, months, or years. The risk of prolonged psychosis or mania after classical psychedelic use was estimated at 0.2-0.5%, though with considerable uncertainty in the amount of drug exposure per person. Insufficient data were found to estimate the rate of prolonged adverse effects from MDMA. Insufficient data were found to estimate the rate of increased suicidal ideation following any psychedelic therapy. Risks of adverse cardiovascular effects and seizure were identified but could not be estimated quantitatively. Little evidence exists to guide the use of touch, a highly sensitive and high-risk element of psychedelic therapy.

Conclusions: While the principles of informed consent usually apply to licensed clinical professions, we determined that supervised psychedelic use outside the clinical professions should still adhere to the principles of informed consent. Informed consent for psychedelic therapy should include a clear disclosure of the psychological and somatic risks and potential benefits, and an individual’s understanding of this information should be rigorously assessed before proceeding. Further research on treatment processes and outcomes across clinical and non-clinical settings will help to better determine the risks, benefits, and safety profile of psychedelic therapies.

Keywords: Psychedelics, Psilocybin, MDMA, informed consent

Disclosure: Nothing to disclose.

P697. RE104: A Novel, Shorter-Acting Psychedelic for Post Partum Depression

Mark Pollack*, Jasna Hocevar-Trnka, Nathan Bryson, Beatrix Taylor, Matthew Johnson, Robert Alexander

Reunion Neuroscience, Sacramento, California, United States

Background: RE104, a unique 4-OH DiPT prodrug, is a proprietary, novel serotonergic psychedelic compound in clinical-stage development for the treatment of postpartum depression (PPD) and other mental health conditions. Preclinical and clinical characterization confirmed similar pharmacology of 4-OH DiPT to the well-characterized psychedelic active form of psilocybin (4 OH-DMT), while in vivo studies demonstrated a significantly shorter and reproducible psychedelic experience. We chose PPD as our lead indication due to the unmet medical need for a rapid acting, highly effective treatment requiring only a limited interruption of breast feeding. Here we present the results of the first-in-human (FIH) phase 1 study characterizing the safety, pharmacokinetics (PK) and pharmacodynamics (PD) of REi04 as well as our design for the recently initiated study randomized active-dose controlled Phase 2 trial in women with moderate to severe PPD.

Methods: A phase 1, FIH, double-blind, parallel group trial was conducted with 6 ascending dose cohorts of 8 psychedelic experienced health volunteers (randomized 6 active, 2 placebo). Predefined dose escalation ranged from 5 mg to 47.9 mg RE104 administered subcutaneously as a single injection,. Adequate set and setting included one preparatory session followed by a dosing session with a qualified and trained session monitor,. Follow-up study visits occurred on days 2 and 10. Study objectives included assessing safety and tolerability, PK and PD (Drug Effect Questionnaire (DEQ) and Mystical Effect Questionnaire 30 (MEQ)). A “complete” mystical experience (CME) was defined as > /=60% max value in total MEQ score. DEQ < /= 1 represented a subjective end of the psychoactive experience.

Results: A total of 48 subjects with a mean age of 36 years, 27% female and 88% white were enrolled across 6 cohorts. There were no serious AEs and no clinically significant vital signs, clinical laboratory, or electrocardiogram findings at doses up to and including 40 mg RE104. PK demonstrated dose-proportionality. At 30 mg RE104, mean experience duration was 3.7 hours, with all participants having a score off < /=1 at 5 hours post-dose. 66.7% of participants in the RE104 30 mg treatment group had a CME predictive of clinical efficacy. This data informed the dose selection of RE104 at 30 mg for the phase 2 trial.

The PPD trial is a multi-center, randomized, double-blind parallel group, active dose-controlled study evaluating the safety and efficacy of a single dose of RE104 in participants aged 18-45 with PPD. The primary endpoint is MADRS at day 7.

Conclusions: A single dose of RE104 was found to be safe and generally well-tolerated with robust PD effects and a short induced psychoactive state (approximately 4 hours). RE104 has the potential to be an accessible, fast-acting, single dose treatment for PPD.

Keywords: Postpartum mental health, Psilocybin analog, psychedelic

Disclosure: Reunion Neuroscience: Employee (Self).

P698. A Novel Class of Non-Hallucinogenic Neuroplastogens with a Unique Combination of Serotonin 5-HT1A/2A Receptor Agonism and Dopamine D2/3 Receptor Antagonism

Krishna C. Vadodaria*, Brian D. Kangas, Jordi Serrats, William Brubaker, Diego A. Pizzagalli, Vikram Sudarsan, Kimberly E. Vanover

Engrail Therapeutics, Inc., San Diego, California, United States

Background: A novel class of neuroplastogens has been discovered, uniquely combining 5-HT1A/2A receptor agonism with D2/3 receptor antagonism. We describe here for the first time the pharmacological profile of a representative prototype of the class, ENX-105. For this unique combination of pharmacologies, we hypothesized that (a) 5-HT1A and 5-HT2A receptor agonism would elicit anxiolytic, antidepressant, and pro-cognitive effects, downstream of enhanced neuroplasticity, (b) at low receptor occupancy (RO), preferential antagonism of presynaptic inhibitory D2/3 receptors would boost dopamine levels in brain reward circuitry, producing anti-anhedonic effects, and (c) at higher RO, antagonism of postsynaptic D2/3 receptors would produce antipsychotic effects. We predicted ENX-105 to be non-hallucinogenic, as D2/3 receptor antagonism may counteract potential 5-HT2A agonism-driven hallucinations. The pharmacokinetic (PK) profile, corresponding receptor occupancies, and behavioral impact of high dose-induced catalepsy were also characterized.

Methods: ENX-105 pharmacology was characterized using binding and functional assays in vitro with mammalian cell lines overexpressing recombinant human receptors. For PK characterization, 3 male rats were orally dosed, plasma collected at 8 timepoints post dose (5 min-24 h), with brain collection in a subgroup at 4 timepoints.

For D2/3 RO time course in rats (n = 5 male rats), oral ENX-105 was administered and plasma and brains were collected at 1, 2, 4, 8, and 24 h post dose. In an ex-vivo competitive tracer binding assay, striatal tissue was incubated with radiolabeled [3H]raclopride, to calculate inversely proportional %RO (Dunnetts’s test).

Therapeutic efficacy of oral ENX-105 and dose-response were tested in various rodent behavioral assays.

Psychogenics SmartCube® (n = 12 male mice) measured spontaneous and evoked behaviors and compared to a reference data set of several drug classes. Classifier algorithms established a drug class and subclass signature.

Probabilistic reward task (PRT): rats were trained on a touchscreen based PRT analog using parameters based on a human task to determine bias for a more richly-rewarded stimulus (log b) and discriminability (log d) as a pro-cognitive measure. Conditioned avoidance response (CAR) was used to examine antipsychotic-like efficacy where avoidance response and escape failures were measured in rats. Head-twitch response was used to examine hallucinogen-like effects. Catalepsy was scored over a period of 4 h post dose in male rats. These behavioral assays were performed in adult male rats (n = 5-8/group).

For microdialysis experiments, probes were surgically implanted in the nucleus accumbens and prefrontal cortex (n = 8 male rats). Microdialysate samples were collected 30 min intervals from 2 h before drug/vehicle to 8 h post-dose (1 mg/kg and 2.5 mg/kg). Data were log transformed and statistical comparisons were made using the Williams’ test and multiple t-tests.

Results: ENX-105 is an agonist at 5-HT1A (EC50 = 16 nM) and 5-HT2A (EC50 = 5 nM) receptors, and an antagonist at D2S (IC50 = 0.8 nM), D2L (IC50-0.08nM), and D3 (IC50 = 3.8 nM) receptors. ENX-105 exhibits significant brain enrichment and retention over the course of 24 hours despite a short plasma half-life.

Given the novel combination of pharmacologies, the SmartCube® assay platform was utilized to examine broad behavioral signatures of ENX-105 and revealed a partial anxiolytic ( < 20%) and novel unclassified ( > 40%) signature at lower doses, antipsychotic signature at higher doses ( ~ 70%), without any hallucinogenic-signature. Consistent with a non-hallucinogenic profile, ENX-105 did not induce head twitches at any dose tested (1-10 mg/kg; p > 0.05).

In the PRT, a translational model predictive of anti-anhedonic efficacy, low dose (1 mg/kg; p = 0.03; d = 0.784) ENX-105 significantly enhanced reward responsiveness (log b). Discriminability (log d) also was significantly enhanced by ENX-105 treatment [ENX-105 vs. vehicle, 0.5 (p = 0.02), trend at 1 mg/kg (p = 0.06), and 2.5 (p = 0.02)], consistent with a pro-cognitive profile.

In CAR, only higher doses of ENX-105 showed positive effects (2.5 mg/kg, p < 0.01; 5 mg/kg, p < 0.001), consistent with antipsychotic-like efficacy mediated by postsynaptic D2/3 receptor antagonism.

Microdialysis experiments in freely moving rats revealed a sustained increase in dopamine in the nucleus accumbens [1 mg/kg (0-4h, p < 0.05), 2.5 mg/kg (0-4h, p < 0.001; 0-8h, p < 0.01)], reflecting dopaminergic mechanisms underlying the anti-anhedonic and antipsychotic efficacy. Additionally, ENX-105 induced a mild and transient decrease in serotonin levels in the prefrontal cortex (4-5h, p < 0.01).

Notably, ENX-105 did not induce catalepsy at anti-anhedonic (1.0 mg/kg/ p > 0.05) nor antipsychotic (2.5 mg/kg; p > 0.05) doses.

Conclusions: Anxiolytic and pro-cognitive effects of ENX-105 are highly consistent with its 5-HT1A/2A receptor agonism whereas anti-anhedonic and antipsychotic effects are highly consistent with its D2/3 receptor antagonism, without the liability of hallucinations that typically accompany psychedelic 5-HT1A/2A receptor agonists. These data support further investigation of ENX-105 and related non-hallucinogenic neuroplastogens for clinical use in the treatment of psychiatric disorders.

Keywords: neuroplastogen, Non-hallucinogenic, anhedonia, 5-HT2A receptor, Depression

Disclosure: Engrail Therapeutics: Employee (Self). Zoll Medical: Employee (Spouse/Partner).

P699. Evaluating the Helioscope Effect: Factor Analysis and Validation of a Novel Instrument for Psychedelic Experiences

Gregor Hasler*, Vincent Diehl, Abigail Calder

University of Fribourg, Villars-sur-Glâne, Switzerland

Background: In recent years, there has been a notable surge in scientific inquiry into psychedelic compounds such as lysergic acid diethylamide (LSD), psilocybin, and 3,4-methylenedioxymethamphetamine (MDMA). These substances have demonstrated potential therapeutic effects in treating various mental health disorders. Especially MDMA-assisted therapy has shown remarkable efficacy in treating post-traumatic stress disorder (PTSD), but the underlying mechanisms facilitating these effects largely remain unclear. This study introduces the Helioscope Questionnaire (HQ), a novel tool designed to measure the processing of difficult or traumatic memories during psychedelic experiences. The ‘helioscope effect’ metaphorically describes an important potential effect of psychedelic drugs, wherein they offer a protective filter against the overwhelming intensity of distressing memories, allowing for more detailed observation and subsequent processing.

Methods: The HQ was developed through extensive clinical observations and expert consultations, resulting in a 25-item scale assessed on a 5-point Likert scale. For this study we recruited 320 participants who had a psychedelic experience within the past year to participate in an online survey. Participants were recruited through online advertisements and word of mouth. Data collected included demographics, drug use history, “set and setting” variables, and personality traits for exploratory analysis of their effects on the HQ using multiple regression.

Data were analyzed using exploratory factor analysis (EFA) and Spearman correlation to determine the scale’s structure and validity. Convergent validity was assessed by correlating the HQ with the Psychological Insight Questionnaire (PIQ) and the relationship between negative HQ items and the Challenging Experience Questionnaire (CEQ). Discriminant validity was evaluated using the Mystical Experience Questionnaire (MEQ) and the CEQ. Predictive validity was tested using the Persisting Effects Questionnaire (PEQ).

Results: Factor analysis revealed three distinct subscales within the HQ: the Helioscope Protection Factor (HPF), the Helioscope Exposure Factor (HEF), and the Anti-Helioscope Factor (AHF). The HPF measures the feeling of safety and protection provided during memory processing in the psychedelic experience, the HEF assesses the emergence of new content and the confrontation with normally avoided topics, and the AHF evaluates negative impacts on processing difficult experiences. The DEPS demonstrated excellent internal consistency with McDonald’s Omega (ω) values of 0.91 for the total scale, 0.91 for the HPF, 0.91 for the HEF, and 0.82 for the AHF. Substantial correlations between the DEPS and PIQ (r = 0.64) supported its convergent validity, while moderate correlations with the MEQ (r = 0.42) and nil correlations between the HQ and the CEQ (r =- 0.005) confirmed discriminant validity. The DEPS total score significantly predicted positive long-term changes as assessed by the PEQ, demonstrating its predictive validity (β = 1.09, p < .001). Additionally, the AHF subscale significantly predicted negative long-term changes (β = 1.09, p < .001), indicating its relevance in identifying potential risks associated with psychedelic experiences.

Conclusions: The HQ is a robust tool that provides novel insights into the processing of difficult experiences during psychedelic states. Its predictive validity for both positive and negative long-term changes underscores its potential utility in psychedelic-assisted psychotherapy. The HQ offers a specific measure that can facilitate understanding of how psychedelics enable the processing of traumatic memories, which is crucial for optimizing therapeutic interventions. Future research should explore the contextual and intrapersonal factors influencing the helioscope effect and assess the HQ’s applicability in diverse clinical settings. The scale’s ability to predict positive life changes beyond established measures like the MEQ highlights its unique contribution to understanding the therapeutic mechanisms of psychedelics. The HQ can serve as a valuable clinical tool for assessing how psychedelics facilitate the confrontation and processing of difficult memories, contributing to the overall well-being of individuals undergoing psychedelic-assisted psychotherapy.

Keywords: drug side effects, Psychotherapy, Trauma exposure

Disclosure: Nothing to disclose.

P700. Interneuron Cell Class Specific Circuit Influences on the Action of Psilocybin

Pasha Davoudian*, Neil Savalia, Clara Liao, Ling-Xiao Shao, Jack Nothnagel, Takeshi Sakurai, Alex Kwan

Yale School of Medicine, New Haven, Connecticut, United States

Background: Psilocybin is a serotonergic psychedelic with emerging therapeutic potential for

treating psychiatric diseases. In preclinical studies to-date, psychedelics have been

shown to induce rapid and robust structural plasticity of cortical pyramidal

neurons. However, pyramidal neurons are embedded in local microcircuit networks which

also contain GABAergic neurons that can dramatically shape and modulate neural

activity. Here we use a variety of methods to understand and investigate how GABAergic interneurons are influenced by psilocybin and how such physiological changes may contribute to psilocybin’s behavioral effects.

Methods: We use in vivo cell-type specific electrophysiology (N = 22 mice) to examine the acute effect of psilocybin and saline control on the firing rate of SST and PV interneurons. We used in vivo two-photon optical imaging (N = 12 mice) to examine the acute effect of psilocybin and saline control on SST and PV interneuron calcium dynamics. We additionally used pharmacological blockade (N = 3 mice) and cell-type specific receptor genetic deletion (N = 3 mice) concurrently with in vivo cell-type specific electrophysiology to examine the role of specific serotonin receptors in governing psilocybin’s actions. Lastly, we used cell-type specific gene deletion of specific receptors to examine in both tail suspension (N = 41 mice) and learned helplessness (N = 50 mice). Both male and female mice were used for all experiments.

Results: To determine serotonin receptor expression, we analyzed publicly available RNA-seq of the mouse frontal cortex to examine expression of serotonin receptors in SST, PV, and VIP interneurons. For SST interneurons there was a greater share of 5-HT1A receptor expression, PV interneurons expressed greater level of 5-HT2A, and VIP interneurons had prominent expression of 5-HT2C and 5-HT3A.

To determine if the firing patterns changes generally follow from the RNA expression data, we used cell-class specific electrophysiology to record from PV and SST interneurons in awake, head-fixed mice. Comparing the firing rates of opto-tagged SST interneurons before and after psilocybin administration showed a significant decrease in spiking activity (P = 0.043, bonferoni corrected). Conversely, PV interneurons demonstrated a delayed but significant increase in firing rate following psilocybin administration (P = 1.4*10-4). To fully understand psilocybin’s effect on interneurons, we used two-photon microscopy to image SST and PV interneurons in the frontal cortex of awake, head-fixed mice. In line with our electrophysiological results, we observed a decrease in calcium events of SST interneurons following psilocybin administration (P = 0.17). However, we did not see an increase in calcium event rate dynamics in PV interneurons following psilocybin administration (P > 0.999).

To investigate these different mechanisms, we first globally blocked 5-HT1A receptors in the brain before administering psilocybin by using the 5-HT1A antagonist, WAY-100635. Consistent with 5-HT1A receptors playing a key role in the inhibitory effect of SST interneurons, we did not observe a decrease or increase in SST firing rate with WAY-100635 pretreatment before psilocybin administration (P = 0.44). We thus sought to specifically knockdown 5-HT1A receptor functionality in SST interneurons. We generated a cell class specific knockout to specifically knockout the 5-HT1A receptor in SST interneurons. Psilocybin failed to induce a similar decrease in firing rate of SST neurons with genetic knockdown of the 5-HT1A receptor (P = 0.09).

Given psilocybin’s inhibitory effect on SST interneurons, we wanted to understand how relevant this physiological effect was for psilocybin’s behavioral effects. To answer this question, we took cell class specific SST-5-HT1A knockout mice and their control littermates (Fig 4.4A) and tested two behavioral assays. Tail-suspension, a long-standing preclinical assay for putative anti-depressive treatments as well as learned helplessness. A single dose of psilocybin reduced immobility time in control (P = 0.043), but not SST-5HT1A knockout mice (P = 0.778). A single dose of psilocybin did not reduce escape failures in either control (P = 0.531) or knockout animals (P = 0.86), suggesting that the 5-HT1A was not relevant for this behavioral-paradigm.

Conclusions: The current study reveals that psilocybin acutely decreases the spiking frequency and calcium dynamics of SST interneurons in the frontal cortex and that these physiological changes may modulate psilocybin’s behavioral effects. Psilocybin’s disinhibitory effect via SST interneurons is strikingly reminiscent of the leading hypothesis of how subanesthetic ketamine exerts its rapid antidepressant. However, work from our group and others has demonstrated that while ketamine and psilocybin have certain shared brain targets, they also exhibit distinct pharmacological action. How then could two disparate drugs exert similar physiological and behavioral effects? Dendritic disinhibition as gate toward neuronal plasticity extends beyond preclinical depression research with examples in motor and task learning. Removal of dendritic inhibition could thus provide a general mechanism by which neocortical microcircuits could reshape to integrate additional local and long range inputs in an adaptive manner. Such circuit mechanisms are likely key in mediating psilocybin’s long-term plasticity and behavioral effects and should be tested in future studies.

Keywords: Psychedelics, GABAergic interneurons, Cortical microcircuit processing, Somatostatin

Disclosure: Nothing to disclose.

P701. Assessing the Relationship Between PTSD and Type 2 Diabetes in a Large-Scale Analysis of Veteran Health Records

Katharine Liang*, Abigail Schindler, Rebecca Hendrickson

VA Puget Sound MIRECC/University of Washington School of Medicine, Seattle, Washington, United States

Background: Post-traumatic stress disorder (PTSD) affects up to 15% of Veterans following deployment and is associated with high rates of medical comorbidities. Many comorbidities are linked to insulin resistance (IR), which is present in over half of people with PTSD. PTSD severity is prospectively associated with increased risk of both worsening IR severity and new-onset type 2 diabetes mellitus (DM2), leading to premature mortality. Mounting evidence suggests that trauma-related mechanisms leading to IR may be biologically distinct from that of IR in the general population. However, it remains unclear which trauma-exposed populations are more susceptible to distinct mechanisms, and whether other factors such as medications or obesity may be at play. Assessing the influence of these factors on the relationships between PTSD and IR is one approach to inform future research objectives and mechanistic study design addressing this important physiological relationship that has potential significant clinical impact in the mental health population. Here, we present preliminary results from an analysis of diagnostic frequency and laboratory results in a sample of post-9/11 US military Veterans that examines the influence of PTSD diagnosis and symptom burden on the likelihood of a DM2 diagnosis and/or a laboratory measure of recent blood glucose (hemoglobin A1c [HbA1c]).

Methods: Data were extracted from the Department of Veterans Affairs (VA) Corporate Data Warehouse clinical database using the VA Informatics and Computing Infrastructure (VINCI) as database/registry coded values. Veterans were included who served at least two years of post-9/11 military service, exited the military prior to 30 years of age, and had completed the traumatic brain injury screening tool. Diagnostic history was identified by ICD diagnosis codes. HbA1c and independent variables of body mass index (BMI), race, sex and PTSD Checklist (PCL) responses were extracted from medical records. PTSD symptom burden was quantified using PCL total scores (PCL5-weekly, PCL5-monthly, PCL-civilian and PCL-military, each normalized by type then combined into one variable of PCL total score). Diagnoses of type 1 diabetes, thyroid or other non-diabetic endocrine disorder, bipolar or primary psychotic disorder were excluded. For the regression analyses, Veterans were excluded during and following any year in which they received any directly anti-glycemic or corticosteroid medication. Multivariable linear mixed effects models were implemented in using the nlme package in R and carried out using results binned and averaged by calendar year, with HbA1c as the dependent variable and independent variables of PTSD total score, age, BMI, and same calendar year prescription for α1 adrenoceptor (AR) antagonist, β AR antagonist, antipsychotic, or antidepressant medication, and run separately for male (M) and female (F) Veterans. Mediation analyses were implemented using the psych package in R.

Results: Over 4.9 million Veterans ages 22-45 were identified for analysis. The full sample prevalence of PTSD in the sample was 28.1%, and of DM2 was 1.2%; the conditional prevalence of DM2 given a diagnosis of PTSD was 1.8% and the conditional prevalence of PTSD given a DM2 diagnosis was 1.8%. The prevalence of DM2 in the population increased with age, such that by age 45 the overall prevalence of DM2 was 7.5%, but the conditional prevalence of DM2 given a diagnosis of PTSD was 9.1%. The relative risk of DM2 given a diagnosis of PTSD was approximately 1.2-1.5 across age, and remained similarly elevated when individuals were excluded during and following either any year in which an antipsychotic was prescribed, or any year in which an antipsychotic or antidepressant was prescribed.

In multivariable mixed effects models, PTSD symptom burden was a significant predictor of HbA1c (normalized β and 95% CI = 0.015 [.013-.017] for M and β = 0.010 [.004-.015] for F, both p < .001), as compared with age (β = 0.014 [0.012-0.016] for M and β = 0.009 [0.003-0.015] for F, both p < .001), BMI (β = 0.094 [0.092-0.097] for M and β = 0.108 [0.102-0.114] for F, both p < .001), and same-year use of an α1 AR antagonist (β = -0.003 [-0.005,-0.001] for M and β = -0.000 [-0.007,0.006] for F, p < .01 for M, N.S. for F). In secondary analyses, the effect of PTSD symptom burden on HbA1c was greater in males, at younger ages, and at lower BMI. In mediation analysis, BMI was not found to be a mediator contributing to the relationship between PTSD symptom severity and HbA1c.

Conclusions: In this analysis of medical records from nearly 5 million Veterans we found that PTSD diagnosis increased risk of DM2, and that this increased risk was not accounted for by antipsychotic or antidepressant use. Furthermore, increased PTSD severity was correlated with worsened glycemic control, a relationship that was not mediated by BMI. This correlation was stronger in males, at younger ages, and at lower BMI, suggesting a relationship between PTSD and glucose metabolism in this population that may potentially be distinct from canonical mechanisms of metabolic dysfunction. Future research exploring mechanisms that explain the link between PTSD and IR has the potential to influence existing treatment algorithms and lead to new treatments for IR-related comorbidities that comprise a large source of morbidity and mortality in Veterans with PTSD. Additionally, such progress may also expand our understanding of PTSD biology and possible treatment targets.

Keywords: PTSD, Brain regulation of metabolism, insulin resistance, Type-2 Diabetes, trauma and stress disorders

Disclosure: Nothing to disclose.

P702. Sexually Divergent Effects of Activity in the Ventral Hippocampus to Nucleus Accumbens Projection on Fear Learning in Rats

Francesca Czesak, Cristina Maria-Rios, Jacklyn Staffeld, Jonathan Morrow*

University of Michigan, Ann Arbor, Michigan, United States

Background: Cues in the environment can gain intrinsic motivational value when paired with rewarding substances. This relationship can become maladaptive when these cues are associated with harmful or addictive substances. Pavlovian conditioned approach training is used to assess individual variation in attributing incentive value to reward cues, which is associated with a proclivity towards addiction and relapse. When the presentation of a lever precedes a food reward, sign-tracking rats will approach and manipulate lever itself, while goal-tracking rats treat the lever merely as a predictor and approach the location of impending food reward. Sign-tracking rats have also been found more likely to freeze to a fear-associated cue, but not a fear-associated context, showing that their increased attention towards cues may apply to cues for both appetitive and aversive stimuli. Neuronal activity in the nucleus accumbens (NAc) – a critical region of the motive circuit – seems to be necessary for the development of sign-tracking behavior and the attribution of incentive salience. However, little is known about the sources of afferent glutamatergic neurons that regulate NAc activity and may contribute to individual variation in the associative learning process. Recent work in our lab has indicated that, in male rats, inhibition of projections from the ventral hippocampus (vHPC) to the NAc enhances the acquisition of sign-tracking, whereas excitation of that same projection shifts behavior toward goal-tracking. However, the vHPC-NAc projection does not seem to influence the sign- and goal-tracking behavior of female rats. For this study, we explored the hypothesis that the vHPC-NAc projection plays a functional role in determining Pavlovian conditioned fear responses, paralleling its effects on Pavlovian conditioned approach behaviors.

Methods: We used an in vivo dual-vector approach to bilaterally inject viruses expressing Cre recombinase into the NAc and either a virus expressing a Cre-dependent, inhibitory Gi-coupled or an excitatory Gq-coupled designer receptor exclusively activated by designer drugs (DREADD) into the vHPC to selectively target the vHPC-NAc projection. These viral injections were performed via stereotaxic surgery on male and female Sprague Dawley rats (7-8 weeks old). Rats then underwent an aversive conditioning task five weeks after surgery. The fear conditioning paradigm consisting of a habituation session, a trace conditioning session pairing tone and foot shock, and a cued fear extinction test in a novel context. All of these sessions occurred with either clozapine-N-oxide (CNO; 3mg/kg, i.p) or vehicle (6% DMSO) on board. Time spent freezing throughout each session was recorded.

Results: Among males in the inhibition (n = 6), excitation (n = 6), and vehicle (n = 6) groups during the fear conditioning task, we found a main effect of treatment on percent freezing during each tone, or CS period (F(1,8) = 5.551, p < 0.05). Specifically, there was a significant difference between inhibition and excitation groups, where excitation increased cued freezing while inhibition decreased cued freezing (p = 0.03, 95% CI = [-72.30 to -5.962]). Additionally, during the cued fear extinction test, we found a main effect of treatment on percent freezing during each tone (F(2,15) = 4.519, p < 0.01). Specifically, there was a significant difference between inhibition and excitation groups, where excitation increased cued freezing while inhibition decreased cued freezing (p = 0.01, CI = [-86.52 to -14.38]). Despite these findings, we found no significant differences in freezing across tasks when comparing female inhibition (n = 4), excitation (n = 6), and vehicle (n = 9) groups suggesting the vHPC-NAc projection does not affect cued fear learning or expression in female rats.

Conclusions: Our data suggest that the vHPC-NAc projection may not be necessary for the acquisition of sign- and goal-tracking behavior nor for cued fear learning or expression in females. However, it may have a more complex and selective role in the performance of sign-tracking and cued fear learning behaviors in males, as chronic vHPC-NAc inhibition increased while excitation decreased the expression/performance of both sign-tracking and cued fear learning in male rats. These experiments add to our growing understanding of the neurocircuitry responsible for biasing sign- and goal-tracking and fear behavior. Ultimately, this will lead to a better understanding of increased susceptibility to cue-driven psychopathologies such as addiction and PTSD.

Keywords: Addiction comorbidity, individual differences, Pavlovian conditioning, Fear learning, Vulnerability traits

Disclosure: Nothing to disclose.

P703. Functional and Structural Connectivity of the Nucleus Accumbens, Amygdala, and Prefrontal Cortex in Relation to Social Anhedonia and Engagement in Trauma-Exposed Young Adults

Steven J. Granger*, Habiballah Rahimi Eichi, Sylvie J. Weinstein, Isabelle R. Vratimos, Justin T. Baker, Isabelle Rosso

Harvard Medical School, McLean Hospital, Belmont, Massachusetts, United States

Background: The transition from adolescence to young adulthood marks a period of rapid normative growth in social networks; however, individuals experiencing symptoms from trauma exposure often show lower social engagement. Both interpersonal trauma exposure and lower social engagement during this developmental stage have been associated with higher risk of poor outcomes, including posttraumatic stress disorder. Despite these associations, the biological correlates of low social engagement following interpersonal trauma exposure remain poorly understood in adolescents and young adults.

Extensive translational research suggests that the basolateral amygdala (BLA), ventromedial prefrontal cortex (vmPFC), and nucleus accumbens (NAcc) are involved in approach and avoidance behaviors and are altered in trauma-exposed adults with anhedonia symptoms. Our prior research in trauma-exposed adults indicates that social anhedonia is associated with social network properties, including lower social network diversity (SND). It remains unknown whether brain connectivity is associated with social anhedonia and SND in trauma-exposed adolescents and young adults.

This investigation utilized resting state functional connectivity (rsFC) and diffusion-weighted magnetic resonance imaging (dwMRI) to determine whether functional and structural connectivity among the BLA, NAcc, and vmPFC are associated with self-report measures of social anhedonia and social network properties in symptomatic trauma-exposed older adolescents and young adults. In addition, we used ecological momentary assessment (EMA) and passive digital phenotyping (PDP) over one year to capture objective real-world social engagement. We hypothesized that altered functional and structural connectivity of the BLA, vmPFC, and NAcc would be associated with greater social anhedonia and lower social engagement.

Methods: We recruited sixty-seven (67) individuals (ages 17-20; 57 female) exposed to an interpersonal DSM-5 Criterion A trauma, with baseline posttraumatic and/or depressive symptoms scores over 28 on the Posttraumatic Stress Disorder Checklist for DSM-5 (PCL-5) and/or over 22 on the Center for Epidemiologic Studies Depression Scale (CESD). Self-report instruments, including the Revised Social Anhedonia Scale (RSAS) and Social Network Index (SNI), were administered at an initial study visit. Total score on the RSAS was used to measure social anhedonia. The SNI was used to derive an SND score, quantified as the number of social roles participants regularly engaged in (e.g., spouse, friend, colleague). During a second study visit, participants completed a multi-modal MRI scan on a 3 Tesla Siemens Prisma scanner. Pearson correlation coefficients were calculated between the NAcc and BLA, NAcc and vmPFC, and BLA and vmPFC from rsFC data. dwMRI outcomes included fractional anisotropy (FA) averaged across hemispheres of the uncinate fasciculus (UF, connecting the BLA and vmPFC) and anterior corticostriatal tract (aCT, connecting NAcc to vmPFC), which connect the regions of interest and are implicated in social affect and behavior.

EMA and PDP began after the MRI scan. EMA consisted of mobile-device surveys querying instances of in-person communication with others, from which we derived the proportion of time spent alone and the average number of in-person interactions per survey over the year-long study period (78.66 ± 54 surveys). A subset of participants (n = 45) had PDP data through the Beiwe smartphone app. Using the Forest package to process the mobile GPS data, we derived the average amount of time spent at home per day and the average number of significant GPS locations visited per day over the study period (201 ± 107 days of data). Statistical analyses consisted of partial correlations between connectivity values and the self-report and objective social engagement variables, controlling for age, sex, scan motion, PCL-5 total score, and CESD total score. We corrected for multiple comparisons using the Hommel method within imaging modality, and comparisons were made separately within self-report (RSAS/SND), EMA, and PDP measures.

Results: Greater functional connectivity between the NAcc and vmPFC was associated with greater social anhedonia (r = 0.40, p = 0.0017, q = 0.010). Higher FA of the UF was associated with a greater average number of in-person interactions (r = 0.27, p = 0.046, q = 0.14) and a lower proportion of time spent alone (r = -0.28, p = 0.038, q = 0.11). Higher FA of the aCT was associated with a larger average number of significant GPS locations visited per day (r = 0.32, p = 0.045, q = 0.18). However, the associations involving FA did not persist after multiple comparisons correction.

Conclusions: This investigation provides preliminary evidence that functional connectivity between the NAcc and vmPFC is associated with greater self-reported social anhedonia but not with objective measures of social engagement in a symptomatic and predominantly female sample of young adults with interpersonal trauma. This suggests that self-report and objective measures of social affect and engagement in this demographic may have different neurobiological signatures. Future work will use repeated EMA and PDP measures over time, given evidence that progressive social withdrawal may better predict adverse outcomes than chronically low social engagement.

Keywords: anhedonia, Social Behavior, Trauma exposure, Digital phenotyping, Resting State Functional Connectivity

Disclosure: Nothing to disclose.

P704. PACAP in Fear Extinction and PTSD

Shane Adams, Thomas Neylan, Victor May, Kerry Ressler, Sabra Inslicht*

University of California, San Francisco, San Francisco, California, United States

Background: Pituitary adenylate cyclase-activating polypeptide (PACAP38), a neuropeptide involved in homeostasis and stress adaptation, has been associated with PTSD and fear conditioning, particularly in women. PACAP38 acts as a primary neurotransmitter and activates the sympathetic nervous system, including the hypothalamic-pituitary-adrenal (HPA) axis, resulting in glucocorticoid release following psychogenic stress (i.e., cognition-related stress), and is required for sustained release of catecholamines (i.e., dopamine, epinephrine, norepinephrine) during periods of chronic psychogenic and physiological stress (i.e., survival-related stress). PACAP, PACR1, and its genetic components (e.g., ADCYAP1R1) may modulate fear-conditioned responses of PTSD through neural mechanisms of synaptic plasticity (e.g., brain-derived neurotrophic factor), critical to the formation and consolidation of new memories that promote fear extinction and extinction retention necessary for adaptation to traumatic stress. Some evidence suggests that individuals with PTSD have more difficulty extinguishing conditioned fear responses, which may prevent improvements in first-line fear extinction-related treatments like prolonged exposure therapy. While PACAP38 was found to be associated with greater fear-conditioned startle responses in women with PTSD, suggesting increased fear acquisition, no studies have examined the association between PTSD symptoms, PACAP38, and fear extinction in humans. This study aims to evaluate the association of PACAP38 with fear extinction and retention.

Methods: We conducted a 2x2 (sex-by-PTSD status) between-subjects experiment to evaluate the association of PACAP38 with fear extinction and retention using a fear conditioning paradigm. Participants included 123 trauma-exposed individuals, including representative proportions of women (48.8%), participants with ≥subthreshold PTSD (39.8%), veterans (33.3%), and participants of color (59.5%). All participants were evaluated for PTSD symptom severity and status using the Clinician Administered PTSD Scale (CAPS-IV) and for the presence of current and lifetime DSM-IV diagnoses using the Structured Clinical Interview for DSM-IV (SCID). Data were collected at 3 visits, including: 1) habituation and fear conditioning (Day 1); 2) fear extinction (Day 4); and 3) extinction retention (Day 11). The conditioned stimuli (paired = CS + ; unpaired = CS–) were two different colored circles, randomly selected for each participant and presented for 8 seconds on a computer monitor positioned in front of the participant. The US was a 500-msec electric pulse, ranging from 0.5 to 5.0 mA, as self-selected by the participant to be “highly annoying but not painful,” and delivered through electrodes attached to the second and third fingers of the dominant hand. The primary psychophysiology outcome variables were differential skin conductance responses (SCR) to CS + /CS- during fear acquisition, extinction, and retention. Blood samples were drawn 1-hour prior to each experimental phase (Time 1, 2, and 3) to measure serum PACAP levels.

Results: Neither PACAP38 nor differential SCR at fear acquisition, extinction, or retention were significantly correlated with CAPS-IV scores for the total sample or when examined by sex/PTSD status. Examinations of potential subsamples employing the Middle-Out Approach (MOA) and latent class analysis indicated three classes: Low PTSD symptoms (45.5%), High PTSD symptoms (40.7%), and Intrusive-Hypervigilant symptoms (13.8%; i.e., intrusive memories, flashbacks, avoidance, social distancing, hypervigilance). Intrusive-Hypervigilant participants were more likely to be women of color, had lower fear extinction retention, and PACAP38 levels at least 2.22 times higher (ηp2 = 0.52-.56) than other participants.

Conclusions: Although PACAP38 was higher in women and PTSD+ participants, PACAP38 was not significantly associated with PTSD symptom severity or fear conditioning/extinction for the total sample. Secondary examinations using MOA provided preliminary evidence of a homogeneous subgroup for whom peripheral PACAP38 levels were more than double that of other participants and associated with Intrusive-Hypervigilant PTSD symptoms and impairments in fear extinction retention. Participants in this subgroup were more likely to be women of color who may share an increased burden of chronic psychosocial stressors. Findings suggest specificity in the association between PACAP38, PTSD symptoms, and fear extinction retention supported by behavioral, biological, and physiological indicators.

Keywords: PTSD, PACAP, Sex Differences

Disclosure: Nothing to disclose.

P705. High Inflammation Marked by C-Reactive Protein is Associated With Dropout in Prolonged Exposure

Alex Rothbaum*, Tanja Jovanovic, Alexandra Klein, Alexander Kline, Barbara Rothbaum, Kerry Ressler, Vasiliki Michopoulos, Norah Feeny

Skyland Trail, Atlanta, Georgia, United States

Background: While most people exposed to trauma follow a natural trajectory of recovery, a significant minority will develop posttraumatic stress disorder (PTSD). Effective therapies such as prolonged exposure (PE) have robust scientific support. However, there remain patients who do not respond to current treatments. Understanding underlying biological mechanisms of disease maintenance and treatment response provides promise for personalizing psychological treatment and improving outcomes. Inflammation is a mechanism of interest, which can be measured by C-reactive protein (CRP). Inflammation has been associated with diagnosis and severity in multiple psychiatric disorders, including depression severity and PTSD. CRP, which is already widely used in medical practice (e.g., cardiology, endocrinology) as a broader marker of systemic inflammation, is a good candidate biomarker as it has a longer half-life. Existing literature has found CRP to be associated with PTSD risk in retrospective and prospective studies. The impact of pro-inflammatory state on PTSD treatment prognosis and response remains to be determined, however, studying the association between CRP and PE outcomes may provide insight into pathways that can be engaged to better personalize and augment treatment for PTSD. This study aimed to examine the relationship between pre-treatment CRP and PTSD symptom change and treatment dropout.

Methods: Participants (N = 68) were a subset from a single site of a larger randomized controlled trial of PE vs. PE + sertraline. Participants (72.1% female, ages 18-65, 56% African American) were recruited from the community, met DSM-IV criteria for PTSD, and received 10 90-minute sessions of PE. Symptoms were assessed at baseline, each session, post-treatment, and at 3-, 6-, and 9-month follow-up using the PTSD Symptom Scale Self-Report (PSS-SR). Blood samples were collected at the baseline visit and frozen as whole blood. Aliquots of 0.2 ml of participant blood was diluted with 0.8 ml of saline, allowing blood to sieve through the saline layer. The aliquot was centrifuged for 30 min at 12,000 RPM with the resulting 0.6ml of supernate removed to a new cryovial. Of this, 0.1 ml was analyzed for hsCRP with a x5 dilution factor. All samples and fluids were kept on ice throughout. CRP was dichotomized in line with previously published cutoffs (high: ≥ 3 mg/L; low: < 3 mg/L). As CRP was derived from whole blood and not plasma, concentrations obtained are considered underestimates of true circulating CRP.

Results: Nine-months post-treatment, those with high baseline CRP had greater mean PSS-SR change scores from baseline than those with low baseline CRP when controlling for age, race, and sex (19.72 vs. 25.59; F[4,46] = 4.38, p = 0.04, d = 0.45). There was no main effect of treatment condition between PE and PE + Sertraline on this finding (p = 0.80, R2 = 0.17). Those with high baseline inflammation were more likely to drop out of treatment after session four of PE compared to those with low baseline inflammation (OR = 3.25, 95% CI [1.05-10.07], p = 0.04).

Conclusions: High baseline CRP was associated with greater reduction in PTSD severity following PE and PE + sertraline. However, high baseline inflammation was also associated with a three-fold increase in likelihood of premature treatment dropout. As such, high inflammation may be a risk factor for early treatment dropout, precluding the opportunity to see symptom change as a measure of treatment response. Considering a more comprehensive approach to treatment response that includes treatment dropout as lack of response, these results are consistent with literature from depression and schizophrenia showing high inflammation is predictive of poorer treatment response. Interestingly, the current dropout finding is after two sessions of imaginal exposure, which should be explored further as it does not appear to be an immediate reaction to imaginal exposure. One possible explanation is that inflammation-driven reward deficits may blunt the perceived benefits of early treatment engagement, including the goal of building confidence to handle difficult situations. CRP is already a widely used marker in medicine, thus additional research should be done to determine if it can be a prognostic marker for PE dropout. Identifying those who may be at risk for dropout will provide opportunities to implement preventative interventions, such referral to intensive or massed treatment options. Additionally, these and future related findings may be broadly applicable to exposure-based treatment for fear-based disorders, not just PTSD.

Keywords: inflammation, Psychotherapy, Translational Neuroscience, PTSD, Treatment-Response

Disclosure: Nothing to disclose.

P706. Identifying the Neural Ensembles Mediating Fear Generalization During Adolescence

Alessia Mastrodonato*, Juliana Tapia, Christine Ann Denny

Columbia University, New York State Psychiatric Institute, New York, New York, United States

Background: Fear generalization is defined as when a conditioned, fear-inducing stimulus is generalized to neutral stimuli leading to maladaptive fear responses in a safe environment. It is a core symptom observed in anxiety disorders, which typically begin in childhood or adolescence. We previously reported that adolescent male, but not adolescent female mice overgeneralize fear, while adult female, but not adult male mice overgeneralize fear. Therefore, there is clearly an age- and sex-dependent influence on fear generalization. However, the neural ensembles mediating fear generalization across the lifespan have yet to be identified.

Methods: ArcCreERT2 x enhanced yellow fluorescent protein (EYFP) male and female mice (5-week-old and 6-month-old, n = 10 per group) were injected with 4-hydroxytamoxifen (4-OHT) in order to open up a window to tag fear encoding neural ensembles. Five hours after 4-OHT, mice were administered a 1-shock contextual fear conditioning (CFC) protocol. Five days later, mice were administered the contextual fear discrimination (CFD) protocol in which they received 1-shock CFC in context A and then a similar, neutral context exposure in context B each day. Following 8 days of CFD, mice were either euthanized following exposure to the aversive context A or the neutral context B. Brain tissue were processed for EYFP and Arc immunolabeling, imaged, and analyzed to identify whole-brain neural ensembles mediating fear overgeneralization. Activity patterns were correlated between brain regions to identify how age and sex change fear generalization neural activity patterns.

Results: We found that 5-week-old female, but not male mice discriminate between the aversive context A and the neutral context B by day 6 (p < 0.01). Six-month-old male, but not female mice discriminate by day 7 (p < 0.01). These behavioral changes are paralleled by changes in the neural ensembles’ reactivation. Specifically, the 5-week-old male, but not female mice and the 6-month-old female, but not male mice show decreased reactivation of the aversive fear ensembles in the hippocampus (HPC) regions dentate gyrus (DG), CA3 and CA1 (p < 0.05 for all groups) suggesting that they might not be able to discriminate between the aversive and the neutral context because of improper activation of the aversive neural ensemble in the neutral environment, which might result in exaggerated fear.

Conclusions: Our data show that fear generalization behavior is associated with altered HPC fear ensembles in 5-week-old and 6-month-old mice and sex greatly impacts fear generalization behavior. Future studies will build whole-brain networks at different ages to determine how Age, Sex, and Context impact the fear ensembles mediating fear generalization. These data will allow us to find new nodes in the network mediating fear generalization so that we can target with pharmacological manipulations to alleviate fear overgeneralization.

Keywords: Fear generalization, Memory engram cell, Adolescence

Disclosure: Nothing to disclose.

P707. The Short-Acting Psychedelic 5-MeO-DMT Produces Opposing Effects on Fear Extinction Depending on the Timing of Drug Application Relative to Extinction Learning

Sarah Jefferson*, Naishaliz N. Lorenzo, Patrick H. Wehrle, Samuel C. Woodburn, Alex C. Kwan, Alfred P. Kaye

Yale University School of Medicine, New Haven, Connecticut, United States

Background: Serotonergic psychedelics have been proposed as potential therapeutics for a range of mental illnesses. Drugs with short-lived subjective effects may be clinically useful because dosing time would be reduced, improving patient access. The short-acting psychedelic is 5-MeO-DMT has been associated with improvement in depression and anxiety symptoms in early clinical studies and we have previously shown that it increases structural plasticity in the frontal cortex for > 30 days after a single dose. However, the effects of 5-MeO-DMT-induced structural plasticity on cognitive and emotional processing have not been fully characterized. Here we assessed the effect of various doses of 5-MeO-DMT on fear extinction learning, a process that is relevant for treatment of PTSD, and determined whether the timing of drug application relative to learning affected ability to extinguish a learned fear response.

Methods: We subjected adult male and female mice (n = 7-11 per group) to cued a fear conditioning protocol consisting of a single tone-shock pair in Context A. The following day, 5-MeO-DMT (10 mg/kg or 20 mg/kg) or vehicle were injected intraperitoneally at either 30 min or 24 hours prior to extinction learning. Extinction consisted of 6 tones with no associated shock in Context B. Extinction retention was tested 24 hours later and also consisted of 6 tones in Context B. Freezing responses were assessed by two-way ANOVA with drug and cue or drug and sex as conditions.

Results: Surprisingly, we found that 20 mg/kg 5-MeO-DMT impaired fear extinction retention when applied 30 min prior to extinction learning (p = 0.0018 for drug condition, two-way ANOVA). However, when drug was injected 24 hours prior to extinction learning, fear responses were suppressed on extinction learning (p = 0.0016 for drug, two-way ANOVA) and extinction retention (p = 0.0195 for drug, two-way ANOVA) days. This reduction in freezing was consistent with a suppression of fear expression rather than enhanced learning as there was no effect of cue on freezing (p > 0.05 for cue, two-way ANOVA). 10 mg/kg 5-MeO-DMT had no significant effect on freezing when delivered 24 hours prior to extinction learning (p > 0.05 for drug, two-way ANOVA), suggesting that suppression of fear expression is dose-dependent. There were no differences in freezing response between male and female mice (p > 0.05 for sex, two-way ANOVA).

Conclusions: 5-MeO-DMT can produce opposing effects on learned fear responses depending on the timing of drug delivery relative to fear extinction learning and it does so in a dose-dependent fashion. This has relevance when considering the potential clinical application of this drug in conjunction with psychotherapy.

Keywords: Psychedelic effects, 5-MeO-DMT, Fear conditioning and extinction, structural neuroplasticity

Disclosure: Nothing to disclose.

P708. Hostile Interpretation Bias and Avoidance Secondary to Fear of Reactive Aggression in a PTSD Pharmacotherapy Treatment Trial

Rebecca Hendrickson*, Millie Wijenaike-Bogle, Marlene Tai

University of Washington, Seattle, Washington, United States

Background: Avoidance of situations where there is a risk of harm to self or others is a core feature of PTSD, and patients with PTSD expressing concern about the potential for them to engage in violent behavior, generally out of proportion to objective risk of this occurring, is a known component. One relatively under-explored potential contributor to this potential risk is that of hostile interpretation bias (HIB), where an encounter that others might consider neutral is experienced as conveying hostility or threat. Here, we describe preliminary results addressing baseline predictors and treatment response of HIB in a treatment trial of prazosin for PTSD in treatment-seeking Veterans (N = 54).

Methods: Participants were treatment-seeking Veterans with PTSD participating in a randomized placebo controlled trial of prazosin for PTSD, in which all participants spent time on open label prazosin, blinded prazosin, and placebo; the results presented here are from the initial 8 week open label portion. Concerns about the risk of being harmed by others versus the risk of harming others, associated avoidance behaviors, and the self-perceived role of reactive aggression and HIB were assessed using a custom questionnaire. HIB was assessed using the Word Sentence Association Paradigm – Hostility Scale (WSAP-H). PTSD diagnosis and symptoms were assessed using the Clinician Administered PTSD Scale for DSM5 (CAPS-5) and the PTSD Symptom Checklist for DSM5 (PCL-5). Functional impact was assessed using the Inventory of Interpersonal Functioning (IPF) and the Social Roles and Activities Short Form 8A PROMISE measure.

Results: A majority of participants expressed at least come concern about the risk of harming others based on their perception of hostility or threat, with > 25% expressing significant or extreme concern about this risk. Concern about risk to others/HIB was significantly related to PTSD symptom severity (CAPS-5 total score, R = 0.42, p < .0001). Concern about HIB was also significantly related to degree of functional impairment on the Inventory of Psychosocial Functioning (IPF, R = 0.46, p < .001). By comparison, concern about the risk of harm from others was similarly related to PTSD symptom severity (R = 0.48, p < .0001), but more weakly related to interpersonal functioning (R = 0.28, p < .05).

Overall, response of PTSD symptoms to treatment was significant across the first 8 weeks of the trial (open label portion), with significant improvement in PTSD symptom severity (CAPS-5 total score, p < .0001), problems with psychosocial functioning (NIH PROMIS Social Roles and Activities 8-item short form, p < .0001), and concern about the risk of threat from others (p < .05). However, there was not a significant decrease in concern about the risk of harming others (p = 0.32) or in HIB as measured by the WSAPH-H (p = 0.15). The degree of improvement in concern for HIB was significantly related to the degree of improvement in psychosocial functioning (R = 0.39, p = 0.019) and total PTSD symptom severity on the self-report PCL5 (R = 0.6, p < .001). In contrast, the degree of improvement in concern for risk from others was not significantly related to improvement in psychosocial functioning (R = 0.17, p = 0.33) or PLC5 (R = 0.27, p = 0.12).

Conclusions: In this treatment-seeking US Veteran population with PTSD, concern regarding HIB and the risk of causing harm to others was common, and associated with significantly worse interpersonal and workplace functioning on self-report. Treatment with open label prazosin resulted in significant improvement in total symptom burden and functioning, as well as self-reported concern about the risk of harm from others, but did not result in a significant decrease in concern about HIB and the risk of causing harm to others. However, those who did report decreased concern about HIB and the risk of harm to others experienced significantly greater decreases in total PTSD symptom burden and improvement in interpersonal functioning. These findings suggest HIB/self-perceived risk of harm to others represents an important symptom domain, which is highly related to functional outcomes, but which may have distinct patterns of treatment response.

Keywords: PTSD, Hostile Interpretation Bias, Irritability/Aggression, prazosin

Disclosure: Nothing to disclose.

P709. Acid-Sensing Ion Channels Mediate Distinct Neural and Behavioral Responses to Interoceptive Carbon Dioxide

Aubrey Chan*, Rebecca Taugher-Hebl, Youngcho Kim, Nandakumar Narayanan, John Wemmie

University of Iowa, Iowa City, Iowa, United States

Background: Respiration is strongly tied to emotion. Controlled breathing has a long tradition in diverse cultures of promoting relaxation. Conversely, respiratory interoceptive threats, such as the threat of suffocation signaled by rising carbon dioxide (CO2) levels, provoke prominent anxiety in humans and defensive behaviors in animals. Similarly, people with respiratory disorders have a higher prevalence of anxiety. Thus, understanding the neural mechanisms underlying interoception of respiratory threats may lead to improved understanding and treatment of anxiety disorders. Regulation of anxiety has been associated with the amygdala, and recently, the amygdala has also been implicated in both interoceptive fear evoked by CO2 as well as in respiratory control. Interestingly, CO2 evokes exaggerated panic attacks in people with amygdala lesions, suggesting a critical role for the amygdala in processing respiratory interoceptive threat. In mice, CO2 inhalation evokes a spectrum of defensive behaviors ranging from freezing to fight-or-flight, and mice with amygdala lesions also show increased flight responses to CO2. Because CO2 inhalation rapidly lowers brain pH, we hypothesized these different CO2-evoked behaviors are mediated by acid sensing ion channels (ASICs) in the amygdala.

Methods: To test this hypothesis, we exposed wild-type and Asic1a-/- mice to various CO2 concentrations and quantified freezing and flight-related jumping behaviors. We also delivered acidic pH directly into the amygdala of awake, behaving mice. Further, we examined CO2 induction of neural activity by using c-Fos staining and by recording unit activity and local field potentials (LFPs) from amygdala in awake, behaving mice.

Results: Both CO2 and acidic pH elicited freezing and jumping responses. In Asic1a-/- mice, freezing responses to both CO2 and acidic pH were impaired, whereas jumping responses to both stimuli were exaggerated. Furthermore, specifically disrupting Asic1a in amygdala reduced CO2-evoked freezing but did not increase CO2-evoked jumping, suggesting effects of ASIC1A on these two behaviors involve distinct cellular sites of action.

Surprisingly, CO2 induced more c-Fos in Asic1a-/- mice, indicating behavioral effects of ASIC1A disruption were due to more than silencing the amygdala. CO2 suppressed firing of some units and stimulated others; strikingly, proportions of both activated and suppressed neurons were increased in Asic1a-/- mice. LFPs also revealed interesting effects: CO2 increased theta power but suppressed other frequencies, especially gamma. In Asic1a-/- mice, the magnitudes of both effects were significantly enhanced.

Conclusions: These data suggest ASIC1A in the amygdala plays a central role regulating responses to respiratory threats and begins to delineate distinct circuit activity patterns corresponding with respiratory interoception. The changes to theta and gamma LFP power suggest that ASIC1A may also affect amygdala communication with other brain sites that regulate defensive behaviors.

Keywords: defensive and motivated behaviors, interoception, local field potentials, Acid-sensing, Amygdala

Disclosure: Nothing to disclose.

P710. Unbiased Identification of the Anterior Hypothalamic Nucleus as a Novel Regulator of Stress Vulnerability

Zachary Pennington*, Alexa LaBanca, Shereen Abdel-Raheim, Madeline Bacon, Afra Mahmud, Patlapa Sompolpong, Bumjin Ko, Zhe Dong, Yu Feng, Alexander Smith, Paul Kenny, Denise Cai

Icahn School of Medicine At Mount Sinai, New York, New York, United States

Background: Prior adversity is a major factor governing vulnerability to stressful life events, predisposing individuals to numerous psychiatric illnesses. Although much is known about neuronal stress circuits, how these circuits are modified by experience to increase stress vulnerability remains unknown. Here, we identify a novel locus regulating stress vulnerability.

Methods: Immediate early gene imaging of cleared, intact, mouse brains was first conducted to identify how prior adversity (10 foot-shocks) influences brain-wide activity levels in response to a subsequent stressful experience days later (an auditory startle stimulus). After identifying the anterior hypothalamic nucleus (AHN) as a possible regulator of stress vulnerability, in vivo calcium imaging was performed to assess how AHN neurons dynamically respond to physical as well auditory stressors. To determine the AHN’s causal role in stress vulnerability, we optogenetically silenced AHN neurons either during an initial stressor, or during subsequent defensive behaviors (associative recall of the initial stressor, anxiety-related behavior in the light-dark test, and exposure to a novel auditory stressor in a distinct context). Chemogenetic activation of AHN neurons was performed to test if increasing their activity is sufficient to augment defensive behavioral responding to threats. Lastly, projection-specific chemogenetic inhibition of AHN inputs from the lateral septum, amygdala, and ventral hippocampus was performed to determine where the AHN receives information about stress from. N = 10-20/group. Both sexes included. Negative-binomial regression and hierarchical clustering was used for the analysis brain-wide data. General linear model was used to assess all behavior.

Results: Harnessing an unbiased brain-wide approach in mice, we found that the AHN – a region that to date has received little attention – displays potentiated responses to future stressors in previously stressed mice. Moreover, the AHN shows increased functional connectivity with a limbic defensive network in mice with a history of prior stress. Using in vivo calcium imaging, we found that AHN activity linearly tracks the magnitude of stress severity. Addressing the causal role of the AHN in stress vulnerability, optogenetic inhibition of AHN neurons during an initial stressor blunted responses to future stressors. Similarly, inhibition of AHN neurons after stress was able to mitigate the observed increase in subsequent defensive behaviors. Demonstrating a bidirectional influence of these neurons on defensive behavior, stimulation of AHN neurons was able to facilitate defensive behavioral responses to stress. Lastly, after tracing limbic inputs to the AHN, we utilized a projection-specific chemogenetic approach and found that inputs from the amygdala to the AHN facilitate sensitized responding to future stressful events, whereas inputs from the lateral septum exert an inhibitory influence on defensive behavior.

Conclusions: Collectively, this work highlights the AHN as a novel regulator of stress vulnerability and demonstrates its ability to gait limbic inputs to regulate defensive behaviors. Given the AHN’s rich cellular diversity, it represents a compelling target for pharmacological intervention.

Keywords: Fear regulation, hypothalamus, Acute Traumatic Stress

Disclosure: Nothing to disclose.

P711. Decoding Threat Neurocircuitry Representations During Traumatic Memory Recall in PTSD

Joshua Cisler*, Kierra Morris, Joseph Dunsmoor

University of Texas at Austin, Austin, Texas, United States

Background: Heightened negative emotional responding upon exposure to reminders of the traumatic memory is a core feature of PTSD1–3. Understanding the basic mechanisms engaged when individuals with PTSD recall traumatic memories also necessarily has implications for understanding the basic mechanisms underlying exposure-based treatment approaches, where individuals are asked to repeatedly recall traumatic memories4. Whether trauma-related stimuli and trauma memory recall engage more general cognitive and neurocircuitry mechanisms of threat memory and threat stimulus responding, or if trauma-related memories and responding are unique, is an important and unresolved line of inquiry5–9. This study tested the hypothesis that traumatic memory recall engages neurocircuitry representations that mirror activity patterns engaged during generalized threat stimulus processing in women with Post Traumatic Stress Disorder (PTSD).

Methods: We used multivariate pattern analysis to train 3 decoders. A “trauma” decoder was trained on fMRI patterns during idiographic trauma vs neutral narratives in a sample of 73 adult women with PTSD. A separate cohort of 125 adult participants from the community completed a reward and threat learning task during fMRI, from which “shock” and “reward loss” decoders were trained on neural activity at the time of threat or reward outcome delivery, respectively. These decoders were then cross-tested on the alternative datasets, allowing analyses of the degree to which traumatic memory recall engaged neurocircuitry representations that overlap with more general aversive stimuli. For example, the shock decoder was applied to the trauma memory recall data, thereby making predictions about engagement of shock neurocircuitry representations during trauma memory recall. Decoders were trained and tested in four networks related to salience processing as well bilateral amygdala and hippocampal masks.

Results: Shock neural representation engagement was higher during trauma vs neutral memory recall in the medial PFC t(1,73) = 2.71, p = 0.0068, and mid-cingulate cortex (MCC) networks, t(1,71) = 2.87, p = 0.004, with no significant interactions with script repetition. Reward loss neural representations were not significantly higher during trauma vs neutral memory recall for any network.

The traumatic memory decoders were then applied to neural activity patterns during shock (vs no shock), providing predictions about engagement of neural patterns associated with traumatic memory during shock (vs no shock) delivery. Trauma predictions were significantly higher for shock vs no shock trials for decoders trained in the dorsal ACC, t(1,132) = 7.31, p < .001, inferior frontal / anterior insula, t(1,132) = 3.14, p = 0.002, medial PFC, t(1,126) = 7.99, p < .001, and MCC networks, t(1,130) = 7.8, p < .001, but not amygdala or hippocampal masks.

The traumatic memory decoders were similarly applied to neural activity patterns during reward loss (vs reward wins), providing predictions about engagement of neural patterns associated with traumatic memory during reward losses (vs wins). Neural pattern engagement for traumatic memory recall was not significantly elevated during reward loss for any network decoder tested.

We also observed a significant interaction between shock decoder predictions during trauma memory recall and PTSD re-experiencing symptoms, t(1,72) = 2.27, p = 0.024. Post-hoc tests using a median split of CAPS-5 re-experiencing symptoms demonstrated this interaction was due to greater shock neural pattern engagement for trauma vs neutral memories in individuals with higher re-experiencing symptoms, t(1,39) = 3.23, p = 0.001, compared to those with less severe re-experiencing symptoms.

Conclusions: These results suggest that trauma memory recall engages neurocircuitry representations that overlap with threatening, and specifically painful, stimulus delivery. The results also encouragingly support the clinical validity of laboratory-based models of fear learning and extinction that use painful stimuli as the unconditioned stimulus10, such that neurocircuitry representations for this unconditioned stimulus appear to at least partially overlap with those of clinically-defined trauma memories among a PTSD sample.

Keywords: PTSD, multivariate pattern analysis, trauma memory

Disclosure: Nothing to disclose.

P712. Cannabinoid Receptor 1 Availability in Trauma-Exposed Individuals With Childhood Versus Adulthood Onset: A Positron Emission Tomography Study

Ardavan Mohammad Aghaei*, Ansel T. Hillmer, Nachshon Korem, Brian Pittman, Henry Huang, Nabeel Nabulsi, Ilan Harpaz-Rotem, Anahita Bassir Nia

Yale University School of Medicine, New Haven, Connecticut, United States

Background: Over the past two decades, the endocannabinoid (eCB) system has emerged as a potential target for the treatment of post-traumatic stress disorder (PTSD). Despite extensive evidence from animal studies suggesting the efficacy of eCB modulators in PTSD models, results from human clinical trials are mixed. Studies on animal models have shown heterogenicity in the trauma-induced eCB system alterations based on trauma characteristics, which could partially explain the mixed results in clinical trials. One factor that is proposed as the potential driver of mixed results is overlooking the differential impact of trauma exposure onset on the eCB system. Given the natural development of the eCB system during childhood, childhood-onset trauma exposure may disrupt the natural course of the eCB system pruning during adolescence, especially when trauma is severe enough to result in PTSD diagnosis. To address this gap in the literature, we investigated the availability of the Cannabinoid Receptor Type 1 (CB1R) in individuals exposed to trauma stratified by the age of trauma onset.

Methods: We recruited individuals with PTSD (N = 13, 7 females, 8 with childhood-onset trauma) and Trauma Controls (TC, N = 23, 2 females, 14 with childhood-onset trauma) who were meeting criterion A for PTSD but not the full criteria. The number of experienced traumas and the age of the subject at the time of each traumatic event were recorded using the Trauma History Questionnaire. Subjects were further categorized with respect to the age of first exposure as either childhood (= < 18 years) or adulthood ( > 18 years). Volume of distribution (VT) of [11C]OMAR, an antagonist radioligand specific for CB1R, was measured as a proxy of CB1R availability. Participants were scanned for 120 minutes after injection of 481.2 ± 151.5 MBq of [11C]OMAR, in a High Resolution Research Tomograph (HRRT) Positron Emission Tomography (PET) scanner. Arterial blood samples were collected while scanning to measure the metabolite-corrected input function. [11C]OMAR VT was calculated using multilinear analysis (t*=30 min). Linear regression models were used to assess group (PTSD vs TC), trauma onset, and their interaction in CB1R availability within each Region of Interest (i.e., amygdala and hippocampus). All models were adjusted for sex and age.

Results: There were no significant group differences in age, body mass index, and distribution of trauma onset. Of note, the number of traumatic events was significantly higher in the PTSD group compared to TC. However, within each group, there was no significant difference in the number of traumatic events between childhood and adulthood onset trauma sub-groups. In models excluding the group x onset interaction effect, we failed to observe a group effect in amygdala (ß=0.10, d = 0.21, p = 0.62) or hippocampus (ß=-0.02, d = 0.04, p = 0.91). However adding the interaction term to models revealed a significant group x onset interaction effect with high effect sizes (amygdala: ß=-0.69, d = 1.54, p = 0.049; hippocampus: ß=-0.76, d = 1.81, p = 0.02) and a main group effect with trend-level significance (amygdala: ß=0.47, d = 1.06, p = 0.07; hippocampus ß=0.39, d = 0.94, p = 0.12). Exploring the posthoc contrasts stratified based on the trauma onset, we observed lower CB1R availability in individuals with PTSD compared to TC, when the onset was in adulthood (amygdala:-16.2%, d = 1.01, p = 0.07; hippocampus:-12.6%, d = 0.91, p = 0.12), but higher when the onset was in childhood (amygdala: 8.9%, d = 0.31, p = 0.38; hippocampus:13.8%, d = 0.70, p = 0.12).

Conclusions: Consistent with some preclinical literature, we observed a divergent pattern of association of CB1R availability with PTSD diagnosis based on the onset of trauma. These findings support the differential impact of childhood trauma onset on the eCB system, given the developmental changes of this system during childhood. Randomizing participants in clinical trials without considering these variations may mask the potential therapeutic effects of eCB modulators in a subgroup of individuals with PTSD. Future larger cross-sectional and longitudinal studies should gather more detailed data regarding trauma characteristics to understand better the alterations of the eCB system following trauma exposure to inform clinical trials.

Keywords: PTSD, endocannabinoid system, PET Imaging

Disclosure: Nothing to disclose.

P713. Latent State Fear Learning in Ambiguous Threat Contexts is Differentially Modulated by Interpersonal Violence Exposure, Anxiety Disorders, and Post-Traumatic Stress Disorder

John Leri*, Kevin Crombie, Joseph Dunsmoor, Joshua Cisler

The University of Texas at Austin, Austin, Texas, United States

Background: Anxiety disorders are characterized by biased inferences about safety and threat. One prominent antecedent to anxiety disorders is exposure to interpersonal violence (IPV), which has been linked to deficits in threat processing. Given the role of IPV in the etiology of anxiety disorders, and the common link to perturbations in threat processing, this study sought to determine if anxiety disorders and IPV have shared or distinct impacts on learning and memory processes involved in a contextual fear laboratory task.

Methods: Participants were 116 adults (ages 21-50; M = 30.11, SD = 8.13) that underwent clinical interviews to identify IPV (National Trauma Assessment), post-traumatic stress disorder (PTSD; Clinical Administered PTSD Scale for DSM-5), and non-PTSD anxiety diagnoses (Diagnostic Interview for Anxiety, Mood, and Related Neuropsychiatric Disorders) and completed a two-day contextual fear learning task. On day-1 participants learned the threat contingencies of three conditioned stimuli (faces were the conditioned stimuli, electrical shock was the unconditioned stimulus) in four unique contexts (two high threat and two low threat) which were explicitly labeled (with colors and images). On day-2, participants entered ambiguous contexts (ie., without context labels), were exposed to the conditioned stimuli, and were asked to estimate the likelihood that they were in each of the previously learned contexts based on the observed pairings of the conditioned stimuli and shock. Nested linear mixed effects models were used to determine if experimental or clinical characteristics were associated with day-2 likelihood estimates. The baseline model predicted likelihood estimates based on experimental characteristics (ie., if the likelihood estimate was about the context that participant was in, if the participant was in one of the high/low threat contexts, if the participant was making an estimate about a high/low threat context, and if the participant was making an estimate about a high threat context while being in a high threat context or vice versa). The full model predicted likelihood estimates based on experimental characteristics, clinical characteristics (IPV exposure, PTSD diagnosis, non-PTSD anxiety diagnosis), and the interactions between experimental and clinical characteristics.

Results: Experimental characteristics significantly predicted day-2 likelihood estimates (χ2(4) = 353.61, p < .001) such that participants provided higher likelihood estimates if they were rating the context that they were actually in (b = 0.95, p < .001) or if they were rating a context which corresponded to the observed threat level (b = 1.80, p < .001; ie., rating a low threat context when in a low threat context). Nested model comparisons indicated that the clinical characteristics also accounted for a significant amount of variance in participant’s day-2 likelihood estimates (χ2(15) = 40.96, p < .001). Participants with IPV exposure were less accurately able to differentiate what context they were in (b = -0.91, p = 0.015). Participants with anxiety disorders provided higher likelihood ratings for contexts which corresponded to the observed threat level (b = 1.24, p < .001; ie., higher likelihood estimates of being in either high threat room when in one of the high threat rooms), participants with PTSD rated low and high threat contexts similarly, irrespective of what context they were in (b = -1.03, p = 0.017).

Conclusions: These results indicate that IPV exposure, PTSD, and non-PTSD anxiety disorders impact the recognition of threat contexts under ambiguous circumstances in distinct ways. Participants who had been exposed to IPV provided lower likelihood estimates for the true threat context than non-IPV controls, which indicates IPV exposure is associated with less accurate identification of learned threat contexts under conditions of ambiguity. In contrast, anxiety disorders (PTSD and non-PTSD anxiety disorders) were associated with participant’s ability to differentiate between contexts with differing degrees of threat. Specifically, participants with PTSD provided more similar ratings for high and low threat contexts whereas non-PTSD anxiety diagnoses were associated with a greater difference between ratings for the high and low threat contexts. These findings suggest that IPV exposure impacts distinct mechanisms of threat learning and memory relative to anxiety disorders and that specific types of anxiety disorders correspond to different biases in the recognition of latent threat contexts.

Keywords: Fear learning, Anxiety and PTSD, interpersonal violence, Behavior

Disclosure: Nothing to disclose.

P714. Maternal PTSD and Neural Reactivity to Infant Cues in Postpartum Women

Rebecca Lipschutz*, Kristina Dahlgren, Ellie McAfee, Justin Santos, James Rilling, Vasiliki Michopoulos, Abigail Powers, Jennifer Stevens

Emory University School of Medicine, Atlanta, Georgia, United States

Background: Posttraumatic stress disorder (PTSD) in women has been linked with dysfunction in neural circuits that support fear and emotion regulation. Functional magnetic resonance imaging (fMRI) studies have shown increased amygdala reactivity to threat and decreased activity in regions underlying emotion regulation (e.g. vmPFC, anterior cingulate cortex[ACC]) in women with PTSD compared to trauma-exposed controls. In the context of motherhood, these neural mechanisms may also contribute to dysregulation of arousal responses to infant cues of distress such as crying. Prior studies in healthy postpartum women have found that the circuits involved in emotion regulation and threat sensitivity (i.e. amygdala, insula, ACC) are also activated in response to infant cries and are further associated with sensitive parenting behaviors. Emerging evidence has shown links between maternal stress and trauma with neural reactivity to infant cry sounds, however only one small study (n = 11) has examined associations between maternal PTSD and brain activity to child-related stimuli. Thus, little is known about the impact of maternal PTSD on the maternal brain response to infant cues in the postpartum period. Understanding brain activation patterns related to maternal PTSD symptoms will help elucidate underlying mechanisms that may contribute to negative parenting behaviors and adverse outcomes for mothers and children. This study examined maternal PTSD symptoms and brain activation in response to social threat (fearful faces) and infant distress (infant cry sounds) in postpartum women.

Methods: The current study examined how maternal PTSD symptoms related to maternal neural reactivity in an ongoing study of n = 31 mother-infant dyads. All mothers endorsed at least one Criterion A trauma to be eligible for the study (M(SD) = 5.8(3.4), range = 1-15). PTSD symptoms were assessed with the PTSD Checklist for DSM-5 (PCL-5; M(SD) = 13.9(12.6), range = 0 – 51). Mothers attended a fMRI session between 8-12 weeks postpartum and completed a social threat task (adult fearful and neutral face stimuli) and infant cry task that included recordings of their own infant’s cries along with an unknown infant’s cries, and control stimuli matched for auditory features. Data were preprocessed using fMRIprep 20.2.3 and analyzed in SPM12.

Whole-brain analyses corrected for multiple comparisons were used to detect neural activity during the infant cry task. Region-of-interest (ROI) analyses tested correlations between PTSD symptoms and activation in the amygdala, insula, and anterior cingulate cortex (ACC). Correlation analyses examined associations between PTSD symptoms and whole-brain activation during the infant cry task (Own infant cry > matched control sound).

Results: Consistent with prior work, PTSD symptoms correlated with higher right amygdala reactivity to fearful > neutral faces, r = 0.44, p = 0.011. In whole-brain voxel-wise analyses of the cry task, mothers showed significantly greater activation to own infant cry sounds (compared to matched control sounds) in the bilateral primary auditory cortex, hypothalamus, bilateral anterior insula, and dmPFC (pFDR < .05). ROI analyses revealed that higher maternal PTSD symptoms were associated with decreased neural reactivity to own infant cry (compared to matched control sounds) in the right amygdala (r = -.37, p = 0.04) and bilateral subgenual ACC (r = -.45, p = 0.01). Beyond the a priori ROIs, whole-brain analyses of maternal PTSD symptoms showed significant negative associations with activation in a midbrain cluster overlapping the raphe nucleus, the bilateral frontal pole, and bilateral nucleus accumbens to own infant cry sounds (pFDR < 0.05). No positive associations between brain activation and PTSD symptoms reached significance in whole-brain analyses.

Conclusions: Similar to prior work, we observed expected associations between PTSD and right amygdala hyperreactivity to a social threat task in trauma-exposed mothers. However, in contrast, mothers with higher PTSD symptoms showed lower amygdala reactivity to infant cries. This suggests that mothers may respond differently to infant cries than other threat cues in the context of posttraumatic stress. Maternal PTSD symptoms were also related to lower activation around the raphe nucleus, frontal pole, and nucleus accumbens, suggesting that mothers with higher PTSD symptoms are less responsive to infant cries in regions related to affective responding and orienting. Overall, associations between maternal PTSD and neural activity were observed in regions that are important for mothers’ threat detection, processing of infant social and emotional cues, and responsive and sensitive caregiving behaviors. These results provide novel evidence for neural mechanisms underlying maternal caregiving in postpartum women exposed to trauma.

Keywords: Postpartum mental health, PTSD, Intergenerational transmission of trauma, maternal caregiving, fMRI

Disclosure: Nothing to disclose.

P715. Compression of the Principal Gradient of Functional Connectivity Spanning the Sensory Cortex and Default Mode Network Tracks the Vividness and Intrusiveness of Trauma Memories

Kevin Clancy*, Yara Pollmann, Boyu Ren, Poornima Kumar, Isabelle Rosso

McLean Hospital - Harvard Medical School, Belmont, Massachusetts, United States

Background: Trauma-related intrusive memories (TR-IMs) are prevalent among individuals exposed to trauma and are leading contributors to post-traumatic psychiatric sequelae. Clinical presentations of TR-IMs highlight their vivid, sensory details that emerge seemingly “out of the blue” to intrude on ongoing thought. These features may distinguish TR-IMs from general episodic memories and thus represent critical therapeutic targets.

There is accruing evidence that TR-IMs involve interactions between different cortical networks linked to sensory processing (sensory cortex; SC), threat detection (salience network; SN), and autobiographical memory (default mode network; DMN). The organization of these cortical interactions in the human brain falls along a principal gradient of functional connectivity, which situates externally oriented, unimodal sensory systems on one end and internally oriented, transmodal cognitive networks on the opposite end. Within this organizing pattern, the SC and DMN are maximally separated. This is believed to protect higher-order cognitive processes from spontaneous and unfiltered sensory intrusions, which could spontaneously activate sensory-based representations of autobiographical memories. Therefore, a collapse of this organizing pattern, or principal gradient, of functional connectivity could be a mechanism driving the sensory vividness and intrusiveness of TR-IMs.

In this study, we combined daily ecological assessments (EMAs) of TR-IMs with functional neuroimaging in trauma-exposed adults to examine how the macroscale organization of cortical networks is related to TR-IM properties. Specifically, using diffusion map embedding techniques to estimate the principal gradient of functional connectivity, we tested the hypothesis that a shorter functional distance between the SC and the DMN would be associated with more vivid and intrusive TR-IMs.

Methods: 109 trauma-exposed adults (89 female) were assessed using the Clinician-Administered PTSD Scale and completed a two-week EMA protocol assessing TR-IM properties, followed by resting-state functional magnetic resonance imaging (rs-fMRI). Cleaned rs-fMRI timeseries were extracted from 400 cortical parcels using the Schaefer atlas to generate a 400x400 functional connectivity matrix. Connectivity matrices were then submitted to diffusion map embedding to compute participant-level principal gradients. These gradients reflect a one-dimensional pattern that explains the maximal amount of variance in functional connectivity patterns across cortical regions, with regions showing similar patterns of functional connectivity falling closer to each other within the principal gradient. Principal gradient values were averaged across parcels within the SC (visual and somatosensory) and the DMN, as defined by the Yeo 7-network atlas. These gradient values index the position of these networks within the principal gradient space. The position of the SN was examined in a similar manner as a control network, given its intermediate position within the principal gradient, in order to demonstrate if the gradient differences were specific to the SC and DMN or generalized to other networks. The SC-DMN difference was then computed to index their separation along this gradient. Pearson correlations were used to examine the relationships between these gradient values and averaged EMA ratings of TR-IM intrusiveness and vividness.

Results: A shorter SC-DMN distance along the principal gradient was associated with more vivid (r = -0.26, p = 0.006) and more intrusive TR-IMs (r = -0.23, p = 0.017). Follow-up analyses demonstrated that vividness was associated with a compression of both the DMN (r = -0.27, p = 0.006) and the SC (r = -0.25, p = 0.009) along the principal gradient. In contrast, intrusiveness was associated with the compression of the SC (r = -0.29, p = 0.002) but not the DMN (r = -0.16, p = 0.09). Control analyses examining the SN’s position within the principal gradient revealed no associations with TR-IM features (p’s > 0.317). The associations between SC-DMN separation and TR-IM features were independent of total PTSD symptom severity (p’s > 0.894) or PTSD diagnosis (p’s > 0.121). Additionally, there was no influence of Age or Sex (p’s > 0.503).

Conclusions: Consistent with our hypotheses, the vividness and intrusiveness of TR-IMs in trauma-exposed adults were associated with a shorter distance between the SC and DMN within the principal gradient. This suggests that less functional separation, or more shared patterns of functional connectivity, between these intrinsically divergent systems may contribute to the spontaneous and involuntary activation of sensory-based representations of autobiographical trauma memories. These findings align with existing neurobiological and cognitive models of TR-IMs and trauma-related disorders, which implicate hyperactivity of sensory systems and dysfunction of autobiographical memory systems like the DMN. Moreover, these findings offer a parsimonious account for the nuanced presentations of TR-IMs, which involve both sensory-perceptual and cognitive-affective processes. Restoring the functional organization of these cortical systems could be a critical target for neurotherapeutics.

Keywords: Posttraumatic stress disorder, ecological momentary assessment, autobiographical memory, Default mode network (DMN), Connectivity gradients

Disclosure: Nothing to disclose.

P716. Associations Between Menopause Symptoms, PTSD, and Physiological Hyperarousal in Perimenopausal Black Women With a Trauma History

Amanda Arnold*, Megan Huibregtse, Linzie Taylor, Amy Murphy, Donia Hassan, Leigha Lee, Esther Jung, Rebecca Hinrichs, Britton Chahine, Abigail Powers, Alicia Smith, Jennifer Stevens, Vasiliki Michopoulos

Emory University School of Medicine, Atlanta, Georgia, United States

Background: Perimenopause is defined as the transitional period preceding menopause and is characterized by irregular menstrual cycles and increasing cycle length. This menopause transition is marked by significant hormonal fluctuations, particularly in estradiol levels, and often accompanied by various physical, cognitive, and psychological symptoms that can impact quality of life. Importantly, our lab has shown that perimenopausal women from minoritized backgrounds experience more post-traumatic stress disorder (PTSD) symptoms compared to pre- and postmenopausal women, suggesting that the perimenopause may serve as a period of increased risk for PTSD symptoms. Furthermore, Black women often enter perimenopause earlier, report longer durations of menopause-related symptoms, and exhibit more severe PTSD symptoms than White individuals. However, the relationship between menopause-related symptoms and PTSD remains poorly understood. Hyperarousal symptoms (e.g., irritability, hypervigilance), a core PTSD symptom cluster, significantly impacts daily functioning, contributes to overall PTSD severity, and is driven by dysfunction of the autonomic nervous system, which has also been implicated in some menopause-related symptoms. Measures like skin conductance response (SCR) offer objective indicators of this psychophysiological hyperarousal. The current study aimed to understand how PTSD symptoms and subjective and objective markers of hyperarousal relate to self-reported menopause symptoms in perimenopausal Black women with a trauma history.

Methods: Black women (N = 50) aged 40-55 with a trauma history were recruited in the Atlanta metropolitan area and provided informed consent. Participants completed validated self-report measures: Menopause Rating Scale (MRS), PTSD Checklist for DSM-5 (PCL-5), Patient Health Questionnaire (PHQ-9), and Generalized Anxiety Disorder-7 (GAD-7). Skin conductance was measured with electrodes on the index and middle fingers. Baseline skin conductance using eSense (N = 35) was recorded during a post-MRI questionnaire, followed by a skin conductance measurement during a PCL-5 trauma interview. SCR was calculated by subtracting average baseline skin conductance from maximum skin conductance during trauma recall. Pearson correlations examined associations between SCR and self-report measures, with 95% confidence intervals and significance set at p < 0.05.

Results: Menopause symptoms were significantly positively correlated with PTSD symptoms (R = 0.637 [95%CI 0.433, 0.778], p < .001). SCR was significantly positively correlated with menopause symptoms (R = 0.414 [95%CI 0.098, 0.653], p = 0.012), PTSD symptoms (R = 0.446 [95%CI 0.138, 0.676], p = 0.006), and PTSD hyperarousal cluster symptoms (R = 0.408 [95%CI 0.091, 0.649], p = 0.014). No significant correlations were found between SCR and general anxiety symptoms (R = 0.137, p = 0.425) or depression symptoms (R = 0.206, p = 0.229).

Conclusions: The current study reveals significant associations between physiological arousal (measured by SCR), menopause-related symptoms, and PTSD symptoms, particularly hyperarousal, in trauma-exposed perimenopausal Black women. These findings suggest that psychophysiological hyperarousal and PTSD symptoms may contribute to menopause symptoms or vice versa in trauma-exposed women. The lack of significant associations with general anxiety and depressive symptoms underscores the specificity of this relationship to PTSD. The current results highlight the interconnected nature of menopause and PTSD symptoms in perimenopausal Black women with a trauma history. The findings suggest that addressing either menopause-related symptoms or PTSD symptoms may have beneficial effects on the other. This relationship also underscores the importance of considering both hormonal and trauma-related factors in interventions for perimenopausal women for informing more effective and comprehensive treatment strategies during the menopausal transition. Future research should characterize the underlying mechanisms of these associations, particularly hormonal influences on psychophysiological hyperarousal and fear neurocircuitry during the menopause transition.

Keywords: Perimenopause, PTSD, hyperarousal, Skin conductance responses

Disclosure: Nothing to disclose.

P717. Effects of Intraparietal Sulcus Stimulation on Extinction Learning in PTSD

Sonalee Joshi*, Milan Patel, Nicholas Balderston, Kevin Lynch, Barbara Fureman, Desmond Oathes, Yvette Sheline, Mingcong Tang, Lily Brown

University of Pennsylvania, Philadelphia, Pennsylvania, United States

Background: Posttraumatic stress disorder (PTSD) is associated with deficits in extinction learning following trauma exposure. Gold standard treatments such as prolonged exposure improve extinction learning. However, these treatments are not effective in approximately 40-67% of patients, and further research is needed to increase their efficacy. While a degree of arousal is necessary during extinction learning, excessive arousal (i.e., over-engagement) may prohibit learning. The intraparietal sulcus (IPS) has been implicated as a center mediating arousal. Continuous theta burst stimulation (cTBS), a form of transcranial magnetic stimulation (TMS), may attenuate arousal. Prior studies have examined the impact of prefrontal TMS on PTSD symptoms, but no prior studies have targeted the IPS with cTBS. This is the first study to examine the effect of IPS on extinction learning in individuals with PTSD.

Methods: Adult participants with PTSD were recruited to complete six separate visits including clinical interviews, functional magnetic resonance imaging (fMRI), and cTBS. Participants underwent a diagnostic interview for PTSD, suicide risk, and other possible comorbidities and completed self-report measures of PTSD symptoms (i.e., Posttraumatic Diagnostic Scale; PDS), depression (Patient Health Questionnaire-9; PHQ-9), and anxiety (General Anxiety Disorder-7; GAD-7). They generated scripts of a traumatic and neutral event included in a script-driven imagery task and reported their expected outcome for listening to the trauma script (Visit 1). Participants completed a baseline fMRI scan including Threat of Shock Task (TST) and Stroop Task (Visit 2). Participants received active or sham cTBS to the IPS and completed the script driven imagery task (Visit 3). Twenty-four hours later, they underwent fMRI scanning and completed the TST and Stroop Task (Visits 4). Thirty days later, participants completed two additional visits identical to Visit 3 and 4 (Visits 5 and 6). In the present analysis, we examined descriptive statistics for baseline clinical data (n = 10) and conducted preliminary analyses on behavioral task data from the Visit 2 scan (n = 4). We compared average anxiety ratings during threat vs. safe trials on the TST and examined differences in response times by accuracy on congruent vs. incongruent trials on the Stroop Task.

Results: At baseline, participants reported moderate depression (PHQ-9; M = 10.90, SD = 5.09) and anxiety (GAD-7; M = 10.90; SD = 4.82), and moderate to severe PTSD symptoms (PDS; M = 33.90, SD = 16.82). Participants who completed the TST during the Visit 2 fMRI scan rated their anxiety on a scale from 0 (no anxiety) to 255 (highest anxiety) and reported greater anxiety for threat trials when the shock was delivered (M = 134.69, SD = 69.79) compared to trials where no shock was delivered (M = 98.63, SD = 72.35). During the Stroop Task, mean reaction time for congruent trials was 1046.87ms (SD = 440.70) compared to 1187.34ms (SD = 484.44) on incongruent trials. We anticipate an additional 20 new enrolled participants by December 2024, at which point statistical comparisons will be reported.

Conclusions: These preliminary results lay the groundwork for future analyses in a larger sample examining the impact of cTBS in the IPS in combination with extinction training on behavioral task performance, physiological indicators of arousal, and neural function. Findings from this novel study will inform treatment approaches augmenting extinction training for PTSD with targeted TMS.

Keywords: PTSD, TMS, Human Neuroimaging, Extinction learning

Disclosure: Nothing to disclose.

P718. TSND-201 (methylone), a Rapid-Acting Neuroplastogen that Stimulates Neurite Outgrowth

Jennifer Warner-Schmidt*, Martin Stogniew, Amanda Jones, Blake Mandell, Benjamin Kelmendi

Transcend Therapeutics, New York, New York, United States

Background: Post-traumatic stress disorder (PTSD), depression, and anxiety show high rates of comorbidity and are all currently treated with selective serotonin reuptake inhibitors (SSRIs). However, SSRIs require weeks of daily dosing to take effect and have limited efficacy. There is a pressing clinical need for acute, episodic pharmacotherapies that are rapid-acting and long-lasting. To address this need, TSND-201 (methylone) is in clinical development to treat PTSD where it has significantly reduced symptoms of PTSD, depression, and anxiety in an open-label Phase 2 study of PTSD patients. The current study was undertaken to explore TSND-201’s underlying mechanism of action.

TSND-201 is the beta ketone analog of MDMA. Despite their structural similarity, TSND-201 shows distinct pharmacological and subjective effects. It is a non-hallucinogenic, triple (serotonin, norepinephrine, and dopamine) reuptake inhibitor and releaser with no known off-target activity. In preclinical studies, TSND-201 has rapid, robust and long-lasting anxiolytic and antidepressant-like effects and enhances fear extinction learning and retrieval. A role for neuroplasticity in its mechanism of action was evidenced by regulation of neuroplasticity-related gene expression in the frontal cortex and amygdala. Altered neuroplasticity in these two brain areas is a common mechanism underlying the pathophysiology and treatment of all three disorders. Therefore, the current studies were undertaken to examine direct effects of TSND-201 on neuroplasticity, specifically on neurite outgrowth, and to identify pathways that mediate its neuroplastogenic effects.

Methods: E18 rat cortical cultures were treated with TSND-201 (10 uM) or vehicle on day in vitro (DIV) 1 and were fixed on DIV3. MDMA (10 uM) was tested as a comparator, and brain-derived neurotrophic factor (BDNF) was used as a positive control. The total neurite length per cell, longest neurite(s), and number of branches per cell were measured following immunohistochemistry using beta-III tubulin antibodies. To determine whether TSND-201’s effect on neurite outgrowth is mediated by common pathways activated by other rapid-acting antidepressants, cells were pretreated with inhibitors of trkB (Ana-12) or mTor (rapamycin) before TSND-201 or vehicle. To further evaluate the pathways involved in TSND-201’s neurite outgrowth effect, cortical cultures (N = 6 per group) treated with TSND-201 or vehicle were harvested 2 or 24 hours after drug treatment for RNA sequencing. RNA extraction, sequencing, and analysis methods were exactly as described previously (Warner-Schmidt et al., 2024).

Results: Total neurite length per cell was significantly increased by methylone (62%, p < 0.01), MDMA (50%, p < 0.05) and BDNF (37%, p < 0.05) compared to vehicle. However, only TSND-201 increased the length of the longest neurites by 2-fold compared to vehicle (p < 0.001). MDMA and BDNF did not have any effect on longest neurites. All three treatments significantly increased the number of branches per cell. Pretreatment with Ana-12 or rapamycin showed different effects on neurite outgrowth and branching. RNAseq results identified pathways regulated at short time points following TSND-201’s treatment that may underlie its effects on neurite outgrowth

Conclusions: TSND-201 shows rapid and direct effects on neurite outgrowth, specifically neurite length, which is distinct from MDMA. Results identify a novel mechanism by which TSND-201 may exert its rapid, long-lasting beneficial effects in preclinical and clinical studies related to PTSD, depression and anxiety.

Keywords: PTSD, Depression, Anxiety, neuroplastogen, Fear extinction

Disclosure: Transcend Therapeutics: Employee (Self). Transcend Therapeutics: Stock / Equity - Privately Held Company (Self).

P719. Single-Cell Atlas Reveals Chromatin and Gene Transcription Dynamics in Human PTSD Brain

Matthew Girgenti*, Ahyeon Hwang, Mario Skarica, Jing Zhang, CheYu Lee, Rosemarie Terwilliger, Lin Lin, Alexa-Nicole Sliby, Jiawei Wang, Dianne Cruz, Douglas Williamson, Alicia Che, Ke Xu, Keith Young, John Krystal

Yale University School of Medicine, New Haven, Connecticut, United States

Background: Post-traumatic stress disorder is a multigenic disorder occurring in the aftermath of severe trauma exposure. Recent studies have begun to detail the molecular biology of the postmortem PTSD brain using bulk-tissue transcriptomic and epigenetic analyses. However, given the array of PTSD-perturbed molecular pathways identified thus far, it is unlikely that a single cell type is responsible. It is therefore necessary to uncover the individual cell type contributions to the molecular pathology of PTSD.

Methods: We isolated ~2M nuclei from human postmortem dorsolateral prefrontal cortex (BA 9/46) from cases and controls for single nucleus RNA sequencing across three diagnostic cohorts: PTSD, MDD (Psychiatric control), and neurotypical controls to identify neuronal and non-neuronal cell type clusters and cell type-specific gene expression changes. We then performed paired sequencing of the same samples for ATAC-sequencing, to measure differential chromatin accessibility. In addition, we validated our transcriptomic findings by performing spatial transcriptomics on a subset of donors using the Xenium platform. We identified open chromatin regions harboring risk alleles for PTSD and through integration of snRNA and snATAC we identified disease specific cis-regulatory elements. We used the most current and largest PTSD and MDD GWAS to fine map risk variants within specific cell types.

Results: We identified 14 distinct cell type clusters including neuronal and non-neuronal cell types. We identified over 1142 FDR significant differentially expressed genes across many cell types and confirmed expression changes of several genes implicated in PTSD pathophysiology by spatial transcriptomics. We found PTSD specific cis-regulatory elements for several genes including ELFN1, FKBP5, and KCNIP4. We show disease-specific receptor-ligand communication patterns between cell types. We constructed a gene expression regulatory landscape of PTSD by integrating RNA and ATAC modalities to define cis-regulatory elements (CREs) and link them to specific genes. We also assemble PTSD-specific TF regulatory networks and link these to DEGs. This resource enabled us to fine-map all of the PTSD GWAS genes from the Million Veteran Program into specific cell types.

Conclusions: We discovered selective changes in the glucocorticoid system (long implicated in PTSD pathology), that were surprisingly most pronounced in endothelial cells and to a lesser extent other non-neuronal cells. In addition, we identify vulnerability of somatostatin (SST) interneurons in PTSD and global shifts in the transcriptome reflecting decreases in SST signaling and neurotransmission. These changes are accompanied by decreased output of microglia signaling suggesting suppression of neuroinflammatory mechanisms in the PTSD PFC. Overall, this work enabled characterization of gene pathways and their dynamics in diverse cortical cell types and prediction of cis-regulatory logic and associated factors underpinning the molecular etiology of PTSD.

Keywords: PTSD, Major Depression Disorder, Postmortem Brain Tissue, single-cell genomics, dorsolateral prefrontal cortex (DLPFC)

Disclosure: Nothing to disclose.

P720. Psychological Resilience and Brain Aging in Cognitively Healthy World Trade Center Responders

Saren Seeley*, Rachel Fremont, Zoe Schreiber, Laurel Morris, Leah Cahn, James Murrough, Daniela Schiller, Dennis Charney, Robert Pietrzak, M. Mercedes Perez-Rodriguez, Adriana Feder

Icahn School of Medicine At Mount Sinai, New York, New York, United States

Background: World Trade Center (WTC) responders show elevated rates of psychiatric and neurodegenerative disorders, greater biological aging, and cognitive impairment, related to exposure to psychological stressors and environmental hazards from the September 11, 2001, attacks. We hypothesized that responders who remained psychologically healthy despite severe WTC-related trauma exposure (i.e., who were highly resilient) would show differences in brain aging vs. responders with PTSD and vs. a control group of responders who experienced few WTC-related trauma exposures. We also examined relationships between brain aging, cognitive measures, and resilience-linked psychological factors in the entire sample.

Methods: The present study analyzed structural MRI data from N = 97 (82% male, M age = 54) cognitively healthy WTC responders who participated in a parent CDC/NIOSH-funded neuroimaging study (#U01-OH01147301), 2018-2023. All responders were previously assessed for 10 WTC exposures (e.g., exposed to human remains; injured onsite) yielding a 0-10 count for each participant. Responders were recruited into three demographically matched groups, WTC-related PTSD (n = 32, M exposures = 5.4), Highly Resilient (n = 34, M exposures=5.5), and Lower WTC-exposed (n = 31, M exposures = 2.4). All in the PTSD group met full DSM-5 criteria for lifetime WTC-related PTSD, and 75% met full past-month PTSD. Responders in the other two groups had no lifetime DSM-5 psychiatric disorders. All participants had normal hearing and vision, no history of head trauma or stroke, normal full-scale IQ estimated by the Wechsler Test of Adult Reading (WTAR), and no cognitive impairment, based on a Montreal Cognitive Assessment (MoCA) score ≥23. We applied BrainStructuresAge, a novel pipeline for T1-weighted MPRAGE data using a deep learning algorithm to estimate brain structure-specific ages (Nguyen et al., 2024, Hum. Brain Map.). Age gap per structure was calculated as the residual of [estimated brain age] ~ [participant’s chronological age], such that a positive age gap indicates accelerated brain aging.

Results: On average across both cortical and subcortical structures, we observed accelerated brain aging (i.e., a positive age gap) in the PTSD group, and decelerated brain aging (i.e., a negative age gap) in the Highly Resilient group. The Lower WTC-exposed group’s estimated brain age generally approximated their chronological age (i.e., aging-as-expected). Aging was significantly more decelerated in the Highly Resilient group vs. PTSD, age gap t(94) = -2.34, p = 0.021, Cohen’s d = 0.58. K-means clustering (k = 3) identified clusters corresponding to positive age gap, little to no age gap, and negative age gap across brain structures. Proportions in each cluster differed by group, (χ² = 11.98, p = 0.016). The decelerated aging cluster had a majority of Highly Resilient responders (58%, 21% each PTSD and Lower WTC-exposed). The accelerated aging cluster had a majority of responders with PTSD (52%; 16% Highly Resilient, 32% Lower WTC-exposed). The aging-as-expected cluster contained similar proportions of Highly Resilient (38%) and Lower WTC-exposed (36%) responders.

Brain aging in the PTSD group was most prominent in frontal and left temporal structures involved in language, memory, and emotional processing. These included Wernicke’s and Broca’s areas, temporal pole, middle temporal gyrus, insula, orbital gyrus, as well as subcortical regions (hippocampus, amygdala, cerebellum). These structures showed the opposite pattern in Highly Resilient responders, having some of the most decelerated brain aging. In contrast, in the Lower WTC-exposed group, accelerated aging was most prominent in somatomotor regions, e.g., occipital lobe; precentral and postcentral gyri.

Across the entire sample, decelerated brain aging was associated with less use of emotional suppression (r = 0.24, p = 0.020), more dispositional positive affect (r = -.21, p = 0.043) and more optimism (r = -.26, p = 0.009). Age gap (positive or negative) was not significantly correlated with scores on a brief cognitive screener (MoCA) in any group (p > .29), but accelerated brain aging in the PTSD group was associated with lower reading (WTAR) score, r = -.35, p = 0.051. In the 85 participants who completed a computer-based cognitive testing battery (CogState), accelerated brain aging was associated with lower episodic verbal memory encoding in the PTSD group only (CogState ISL score r = -.46, p = 0.014).

Conclusions: Highly resilient WTC responders showed decelerated brain aging, both overall and in structures involved in emotion processing, language, and memory. This is notable given the group’s exposure to a high level of traumatic stressors and environmental hazards (e.g., via working in search and rescue) in the aftermath of 9/11. Further, decelerated brain aging was related to higher positive emotionality and optimism, and less maladaptive emotion regulation. Accelerated brain aging was associated with poorer verbal encoding in the PTSD, despite cognitive impairment being exclusionary for this study. The Lower WTC-exposed control group showed a different pattern of aging altogether, primarily in somatomotor regions. Though we cannot infer directionality from these cross-sectional data, our findings add to emerging knowledge of relationships between psychological resilience and biological aging.

Keywords: stress resilience, World Trade Center responders, brain aging, Structural MRI, PTSD

Disclosure: Nothing to disclose.

P721. Reductions in Sensorimotor-Default Mode Network Connectivity During Trauma Symptom Provocation After Deep Brain Reorienting Treatment

Breanne Kearney*, Ruth Lanius

University of Western Ontario, London, Canada

Background: Despite a growing interest in mind-body treatments for trauma-related conditions, their neurobiological mechanisms of action remain unclear. This presentation will build on recent findings of hyperconnectivity patterns between the sensorimotor (SMN) and posterior default mode (DMN) networks involved in somatomotor perception and self-referential memory processes, respectively, during traumatic memory retrieval in PTSD by showing a reversal of these patterns after Deep Brain Reorienting (DBR) treatment. DBR is a mind-body treatment targeting lower-level sensorimotor mechanisms during trauma processing.

Methods: Individuals were assessed by the Clinician Administered PTSD Scale (CAPS-5) and underwent 7T fMRI scans before and after the intervention phase. Participants with PTSD were randomized to either eight sessions of DBR (n = 20) or waitlist control (WL; n = 22). During fMRI scans, participants were visually presented with personalized neutral and trauma-related words in separate six-minute runs. Subjective re-experiencing intensity scores were collected in situ. This randomized clinical trial is registered with the U.S. National Institute of Health (NCT04317820).

Results: Using whole-brain voxel-based analyses (voxelwise threshold p < 0.001; cluster corr pFDR < 0.05), seed-based approaches show significant reductions in functional connectivity (FC) between precuneus/posterior cingulate cortex and right precentral gyrus (t = 6.2; pFDR=0.003) as well as decreased FC between sensorimotor network and precuneus (t = 7.3; pFDR=0.008) in those who received DBR treatment; no significant FC changes emerge in WL. Data-driven group independent component analysis (ICA) also shows significantly reduced recruitment of PCC/precuneus by the sensorimotor network (t = 5.7; pFDR=0.04) in those who received DBR with no FC changes seen in WL. These DBR-related findings occur alongside significant differences between DBR and WL in CAPS-5 total (pre p = 0.71; post p < 0.001), CAPS-5 re-experiencing (pre p = 0.80; post p = 0.002), and in situ reliving intensity during trauma provocation (pre p = 0.46; post p = 0.001).

Conclusions: These findings further support the indexing of SMN-DMN hyperconnectivity patterns in trauma re-experiencing symptomatology by showing a reduction in these patterns after DBR and strengthen the rationale for sensorimotor-based treatments.

Keywords: PTSD, Human Clinical trial, Psychotherapy

Disclosure: Nothing to disclose.

P722. Brain Aging and PTSD in a Community Sample of South African Women: Exploring Protective Factors and Implications of Model Bias

Kate Webb*, Sheri-Michelle Koopowitz, Jonathan Ipser, Kerry Ressler, Nathaniel Harnett, Dan Stein

Harvard Medical School, McLean Hospital, Belmont, Massachusetts, United States

Background: Trauma exposure is linked to accelerated aging across various brain regions. Protective factors are theorized to mitigate this neurobiological consequence of trauma and help prevent posttraumatic stress disorder (PTSD). However, recent work suggests protective factors may also carry a biological cost. For example, individuals with PTSD who have higher psychological resilience, appear biologically older than individuals with similar clinical profiles and lower psychological resilience. Therefore, the current study investigated whether psychological resilience moderated the relationship between PTSD and accelerated aging in a community sample of South African women.

Methods: Women (N = 189; Mean age = 30.67 ± 6.65) were recruited as part of the Drakenstein Child Health Study. Participants underwent a structural MRI and completed the Connor-Davidson Resilience Scale (CD-RISC) and a PTSD measure. A predicted brain age gap (BAG) was derived using a pre-trained model. Linear models tested a CD-RISC x Group (no trauma exposure (NTE), trauma-exposed [TE], and PTSD) interaction on BAG after adjusting for relevant covariates. Follow-up models probed regional BAG (e.g., sub-cortical) and sub-cortical volumes (amygdala and hippocampus).

Results: On average, the predicted brain age was 10 years older than the participants’ chronological age. There were no group differences on chronological age (F(2,186) = 0.33, p = 0.722), predicted whole brain age (F(2,186) = 0.47, p = 0.629), or whole brain BAG (F(2,186) = 0.99, p = 0.374). There was a significant CD-RISC x Group interaction, such that higher CD-RISC scores were associated with accelerated aging in those with suspected PTSD but not in NTE or TE individuals (t(177) = 2.49, p = 0.014). Follow-up tests showed the CD-RISC x Group interaction was only significant for sub-cortical BAG (t(177) = 3.11, p = 0.014) and hippocampus volume (t(174) = -2.01, p = 0.046).

Conclusions: Pre-trained models may overestimate BAG among individuals from lower-middle-income countries and results emphasize the need for diversity in training data. Still, findings indicate greater psychological resilience may be associated with a biological cost in women with PTSD. Future longitudinal work should examine other culturally relevant factors and consider the generalizability of brain age models across sociodemographic groups.

Keywords: PTSD, Risk and Resilience, algorithmic bias and fairness

Disclosure: Nothing to disclose.

P723. A Retrospective Analysis of Ketamine Intravenous Therapy for PTSD in Real-World Care Settings

Lynne McInnes*, Mark Berman, Matthew Worley

Osmind, San Francisco, California, United States

Background: There remains tremendous unmet need for the treatment of PTSD, as no new treatments have been approved since 2001 and response rates to current treatments are low. Approximately 60% of patients with PTSD respond to selective serotonin reuptake inhibitors but only 20% − 30% achieve complete remission following pharmacotherapy (Stein et al. 2002; Zohar et al. 2002) resulting in polypharmacy in clinical practice (Krystal et al. 2017). While some forms of psychotherapy can be effective for PTSD, elevated symptoms often prevent patients from completing a course of treatment. MDMA holds great promise for PTSD however its approval may well be delayed and even if it is approved in 2024 it is widely acknowledged that the initial rollout will be very slow. Meanwhile, it has been suggested that ketamine intravenous therapy (KIT) may have some efficacy for PTSD. Two small randomized controlled trials (RCTs) of off-label KIT for PTSD by Feder and colleagues (2014, 2021) showed positive results using 0.5 mg/kg however one additional study by Abdallah and colleagues yielded equivocal results using doses of 0.2 and 0.5mg/kg. However, there is at least one paper showing that doses up to 1.2 mg/kg may be more effective for depression in real-world patients than the 0.5 mg/kg dose used in RCTs (Cusin et al. 2017). Therefore we chose to explore available real-world data on the use of KIT in the treatment of PTSD leveraging our Osmind EHR-derived real-world database of over 8,000 patients with a diagnosis of PTSD receiving at least one treatment with KIT.

We report outcomes in a sample of 1309 real-world patients with a diagnosis of PTSD who received at least one KIT treatment and completed a PCL-5 both before and after the index treatment.

Methods: We used data from a large sample of patients who initiated KIT within Osmind’s electronic health record-derived de-identified database. The EHR is used by mental health clinicians in community practice. Patients and providers consented to the use of anonymized data from the EHR for research purposes. The study protocol was deemed exempt by the WCG institutional Review Board. The outcomes measure utilized by clinicians was the PTSD Checklist for DSM-5 (PCL-5) which is a self-report measure with a look back period of one month and scores ranging from 0-80. Patients scoring above a cut-point score of 31-33 are considered to be in need of treatment. We included patients ≥18 years old as of the index date who had at least one KIT administration note and a diagnosis of PTSD. Of these 1,306 had both a baseline PCL-5 (M = 46.6, SD = 15.8) and at least one post-index PCL-5 completed within 30 days after any infusion, with mean age of 42.3 (SD = 12.1), 94% Caucasian, and 66% Female. The most common psychiatric comorbidities were MDD (90%), GAD (72%), ADHD (33%), and panic disorder (19%). We estimated changes in 7,928 PCL-5 scores as a function of the number of infusions received with mixed-effects models, with the best-fitting model (linear, nonlinear, log) selected according to BIC. Covariates included age, sex, race, and BMI. We also report response rates with a threshold of 30% reduction from baseline PCL-5, similar to prior research (Hamner et al. 2004, Feder et al. 2014, 2021).

Results: The best-fitting model using log-transformed infusions (BIC = 62,384) found a statistically-significant reduction in PCL-5 scores, b = -11.56 (0.22), z = -51.73, p < 0.001. Model-estimated changes in PCL-5 closely approximated observed means, with an estimated 18.6-point reduction in PCL-5 scores from baseline (46.16) to after the 4th infusion (27.56), which 82% of the initial sample reached within 28 days after the first infusion. Observed response rates increased from 27% after the 1st infusion to 65% after the 4th infusion. Mean dose at the final infusion during the induction phase, defined as up to 6 infusions within 28 days, was 1.02 mg/kg (SD = 0.39).

Conclusions: We observed statistically significant reductions in PCL-5 scores after KIT administration in a population of adult predominantly caucasian female middle-aged patients with a diagnosis of PTSD and significant comorbidity. Response rates continued to increase with each treatment from the first to beyond the 4th infusion.

Keywords: PTSD, (R,S)-ketamine infusion therapy, Real-world clinical outcomes

Disclosure: Nothing to disclose.

P724. The Influence of Childhood Physical Abuse on Heart Rate Variability Over Pregnancy

Mariana Rocha*, Sean Minton, Shimarith Wallace, Sriya Karra, Meghna Ravi, Rebecca Hinrichs, Nicole Nugent, Suchitra Chandrasekaran, Alicia Smith, Abigail Powers, Vasiliki Michopoulos

Emory University School of Medicine, Atlanta, Georgia, United States

Background: Black pregnant individuals are at disproportionate risk for trauma exposure and trauma-related psychopathology, including post-traumatic stress disorder (PTSD) and major depressive disorder (MDD). Although trauma-related psychopathology is associated with increased risk for pregnancy-related metabolic morbidity and mortality (MM and M), the physiological mechanisms underlying this relationship are poorly understood. Trauma exposure, PTSD and depression are associated with low heart rate variability (HRV), an indicator of a dysregulated autonomic nervous system (ANS). In non-pregnant people, low HRV is associated with adverse behavioral and metabolic health outcomes. Because pregnancy is accompanied with significant cardiometabolic adaptations to meet the energy demands of the developing fetus, it is necessary to understand how HRV changes over the course of pregnancy, and whether specific trauma histories impact the trajectory of HRV over the prenatal period. Childhood physical abuse (CPA), over other types of traumas, has been shown to be associated with ANS dysfunction. Thus, we assessed how HRV changes over pregnancy and how CPA impacts initial HRV and the trajectory of HRV over pregnancy in a sample of Black pregnant individuals. We hypothesized that HRV would decrease over pregnancy and that CPA would lead to lower initial HRV and a greater decrease in HRV over pregnancy.

Methods: Black pregnant persons (N = 127, mean age = 27.5) seeking prenatal care from a large publicly funded hospital in Atlanta, GA, who endorsed at least one criterion A trauma were invited to participate in this study. Participants completed an interview involving collection of demographic information and trauma history. Self-reported CPA and adulthood physical assault were obtained from participant interviews using the Childhood Trauma Questionnaire (CTQ) and the Traumatic Events Inventory (TEI) measures, respectively. Resting respiratory sinus arrhythmia (RSA), a measure of HRV in synchrony with respiration, was collected via electrocardiogram (ECG) sensors three times, once during each pregnancy trimester. To evaluate the influence of CPA on RSA over the course of pregnancy, growth curve models were conducted in MPlus and the alpha value for significance was set at p < .05. An unconditional latent growth curve model was used to assess overall changes in RSA over pregnancy in our sample. A second latent growth curve model was conducted to assess the effect of CPA on initial RSA and change in RSA over pregnancy, while controlling for adulthood physical assault and employment. Covariates for the CPA model were included if they were correlated with RSA at any timepoint (p < .1). In both models, the intercept was set to conception, and timepoints were defined by gestational age at RSA assessment, calculated by the participant’s estimated due date.

Results: In our sample, 61.6% of participants endorsed having experienced CPA. The unconditional model had a significant intercept (β = 5.607, S.E. = 0.481, p < .001) but not slope (β = -0.008, S.E. = 0.018, p = 0.650). In our CPA model, CPA was significantly associated with the model intercept (β = -.348, S.E. = 0.165, p = 0.035) but not slope (β = 0.012, S.E. = 0.006, p = 0.05).

Conclusions: The current findings indicate that RSA does not change over pregnancy in our sample of trauma-exposed Black pregnant persons. However, individuals who have experienced CPA showed lower RSA early on in pregnancy. Future directions include assessing whether prenatal trauma-related psychopathology symptoms predict early prenatal RSA and change in RSA over pregnancy, and whether early prenatal RSA and change in RSA over pregnancy are related to the development of pregnancy-related MM and M. Uncovering a modifiable biological mechanism of risk, like HRV, may inform preventions and interventions that reduce pregnancy-related health disparities in those that have experienced trauma.

Keywords: autonomic nervous system, pregnancy, heart rate variability, Childhood trauma

Disclosure: Nothing to disclose.

P725. Temporal Dynamics of Dissociation Across Diagnostic Categories: Insights From Everyday Life and Laboratory Data in Patients With Posttraumatic Stress Disorder, Borderline Personality Disorder, and Dissociative Disorders

Johannes Heekerens*, James Gross, Sylvia Kreibig, Katja Wingenfeld, Stefan Roepke

Charite University Medical Center Berlin, Berlin, Germany

Background: Dissociation is a ubiquitous mental state characterized by sudden, involuntary disruptions in how a person experiences and integrates different aspects of their life. For example, this can impact their perception of their surroundings and body, and include a subjective sense of gaps in one’s awareness. Dissociation has been proposed to be accompanied by specific alterations in affective experience and physiology. Specifically, dissociation was conceptualized as a coping response and, as such, is assumed a role in regulating unpleasant affective and high-arousal physiological states. In the current research project, we investigated changes in affective and physiological states before, during and after dissociative episodes.

Methods: We collected dense experience sampling and electrocardiogram data over one week from 88 participants with recurrent dissociative symptoms associated with borderline personality disorder, posttraumatic stress disorder, and/or dissociative disorder. Participants reported momentary depersonalization/derealization, gaps in awareness, and affective arousal and valence twelve times per day. The prompts were presented in three blocks, each containing four prompts separated by 15-minute intervals. The same individuals participated in a laboratory stress induction, the Trier Social Stress Test, during which we repeatedly assessed the aforementioned self-reports, along with electrocardiogram and electrodermal data, blood pressure, and salivary cortisol levels. Data were analyzed using dynamic structural equation models and latent change score models to explore the temporal dynamics between dissociative, affective, and physiological states.

Results: Results from everyday life indicate strong contemporaneous relationships between dissociative states and high unpleasant arousal. Increases in arousal and decreases in valence both predicted higher dissociation approximately 15 minutes later. Standardized cross-lagged effects were estimated at 0.09, 95% CI [0.06, 0.12], for the effect of arousal on later depersonalization/derealization, and -0.07, 95% CI [-0.10, -0.04], for the effect of valence on later depersonalization/derealization. For gaps in awareness the estimates were 0.07, 95% CI [0.03, 0.09], and -0.08, 95% CI [-0.10, -0.05], respectively. Higher dissociation was associated with prolonged episodes of unpleasant arousal. Standardized cross-lagged effects were estimated at 0.03, 95% CI [0.01, 0.06], for the effect of depersonalization/derealization on later arousal, and 0.01, 95% CI [-0.02, 0.03], for the effect of depersonalization/derealization on later valence. For gaps in awareness the estimates were 0.07, 95% CI [0.03, 0.09], and -0.03, 95% CI [-0.06, -0.01], respectively. Increases in heart rate predicted the onset of dissociative episodes in an additional analysis, with a positive linear trend between heart rate and a negative linear trend between high-frequency heart rate variability and dissociation. Standardized cross-lagged effects were estimated at 0.02, 95% CI [-0.01, 0.04], for the effect of heart rate on later depersonalization/derealization, and 0.02, 95% CI [-0.01, 0.04], for the effect of depersonalization/derealization on later heart rate. For gaps in awareness the estimates were 0.01, 95% CI [-0.01, 0.04], and 0.01, 95% CI [-0.01, 0.04], respectively. During the laboratory stress induction, we observed marked increases in dissociation, which ran parallel to increases in unpleasant arousal, heart rate, mean arterial blood pressure, and electrodermal conductance. However, we found no temporal relations between dissociation and affective or physiological states during the laboratory session. For example, higher dissociation at the beginning of the stress induction did not predict changes in affective states 5-10 minutes later during the stress induction. Specifically, the unstandardized average effect of latent change in depersonalization/derealization was estimated at -0.27, p = 0.485, on later arousal and 0.16, p = 0.698, on later valence. For gaps in awareness the estimates were 0.00, p = 0.999, and 0.19, p = 0.772, respectively.

Conclusions: Our results support close contemporaneous relationships between unpleasant arousal and dissociation. In everyday life, dissociation appears to be triggered by increased unpleasant arousal and prolongs such affective states. This aligns with the conceptualization of dissociation as a maladaptive emotion regulation strategy or general psychopathological mechanism. We find no evidence that dissociation effectively regulates or blunts unpleasant affective or high-arousal physiological states.

Keywords: Stress and Trauma, Dissociation, Borderline Personality Disorder, Affect Regulation

Disclosure: Nothing to disclose.

P726. TSND-201 (Methylone) for the Treatment of PTSD: Improvements Across Each CAPS-5 Cluster and Anxiety Symptoms From the Open-Label Portion of the IMPACT-1 Study

Amanda Jones*, Jennifer Warner-Schmidt, Martin Stogniew, Blake Mandell, Hannah Kwak, Paul Miller, Iain Jordan, Terence Ching, Benjamin Kelmendi

Transcend Therapeutics, New York, New York, United States

Background: Post-traumatic stress disorder (PTSD) is a debilitating psychiatric illness affecting approximately 13 million adults in the US annually, yet only 2 FDA-approved (Paxil and Zoloft) pharmacological treatments are available. Both treatment options have limited effectiveness, delayed efficacy onset, and the potential for significant side effects. Novel compounds with acute and episodic dosing paradigms may be clinically useful as they produce rapid relief with minimal, transient side effects. TSND-201 is a non-hallucinogenic, rapid-acting neuroplastogen and the beta-ketone analog of MDMA; its distinctive pharmacological properties and profile, affinity to monoamine transporter affinity, and lack of off-target effects produce subjective drug effects that offer notable clinically practical advantages over MDMA.

Preclinical studies have demonstrated that TSND-201 produces fast-acting, robust, and long-lasting anxiolytic and antidepressant-like activity, uniquely positioning TSND-201 as a potential treatment for PTSD.

Methods: The IMPACT-1 study is a multi-center, two-part clinical trial. Part A has completed and was an open-label evaluation of 14 participants with PTSD. Eligible participants were adults (male and females) with severe PTSD (CAPS-5 ≥ 35) who failed at least 1 prior treatment (pharmacotherapy and/or psychotherapy) for PTSD. Participants were treated with 4 doses of TSND-201 (methylone) given once a week for 4 weeks. Throughout each dosing session, participants were provided non-directive psychological support by a trained mental health practitioner. Following the 4-week treatment period, participants were followed for an additional 6 weeks to evaluate the durability of the therapeutic effect. PTSD symptom improvement was evaluated on the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5; including symptom clusters: [Intrusion, Avoidance, Cognition and Mood, Arousal and Reactivity] and PTSD Checklist for DSM-5 (PCL-5), and anxiety symptoms were assessed using 4 items of the Montgomery-Åsberg Depression Scale (MADRS; inner tension, reduced sleep, reduced appetite, and concentration difficulties). Safety was assessed by monitoring adverse events, vital signs, and C-SSRS.

Results: Treatment with TSND-201 resulted in significant improvements on PTSD and anxiety symptoms. On the CAPS-5, treatment with TSND-201 resulted in a mean change from baseline (47.8 points) of -23.3 points (p < 0.001) on Day 10 (2 days after the second dose); the durability of the therapeutic effect is demonstrated by a decrease of 36.2 points (p < 0.001) on Day 64 (6 weeks after the last dose). This trend was mirrored by a 23.5-point decrease in PCL-5 score (p = 0.003) after the first dose and 35.8 points (p < 0.001) at the end of the study (Day 64). The efficacy of TSND-201 on PTSD symptoms was consistent across each of the four CAPS-5 symptom clusters. Mean end of study improvements from baseline were 79.4% for Intrusion (-9.8 points), 79.9% for Avoidance (-4.8 points), 66.1% for Cognition and Mood (-13.2 points), and 75% for Arousal and Reactivity (-8.54 points). Anxiety symptoms, measured by the four MADRS items (range 0 to 24), were also significantly improved after treatment with TSND-201. Mean improvements from baseline (12.9 points) were 51.9% (-6.7 points, p = 0.004) points on Day 10 (2 days after the second dose) and were durable with a mean improvement of 65.9% (-8.5 points, p < 0.001) on Day 64 (6 weeks after the last dose). TSND-201 demonstrated a favorable safety profile and was well tolerated with transient adverse events. The common adverse events were headache, fatigue, and decreased appetite.

Conclusions: TSND-201 demonstrated rapid, robust, and durable effects on each of the PTSD symptom clusters and anxiety symptoms. Given the urgent need for rapid-acting, safe, effective treatments for PTSD, this study has generated encouraging and promising evidence in support of further development of TSND-201 as a treatment for PTSD. The ongoing Part B study of IMPACT-1, a randomized, placebo-controlled study, will provide additional insights into TSND-201’s potential as a groundbreaking treatment for PTSD.

Keywords: PTSD, Neuroplastogen, TSND-201, Clinical trial

Disclosure: Transcend Therapeutics: Employee (Self). Transcend Therapeutics: Stock / Equity - Privately Held Company (Self).

P727. Maternal Posttraumatic Symptoms are Related to Negative Parenting Behavior and Lower Infant Respiratory Sinus Arrythmia in Early Postpartum

Abigail Powers*, Rebecca Lipschutz, Elizabeth McAfee, Catherine Abrams, David O’Banion, Vasiliki Michopoulos, Patricia A. Brennan, Jennifer Stevens

Emory University School of Medicine, Atlanta, Georgia, United States

Background: Women are at substantially heighted risk of posttraumatic symptoms (PTS) compared to their male counterparts and PTS can be carried across generations, making the early postpartum period a critical time to characterize early mechanisms through which maternal PTS leads to adverse child outcomes. Maternal PTS is related to lower levels of maternal sensitivity and higher levels of disengagement, intrusiveness, or both in older infants and toddlers. Infant autonomic nervous system (ANS) function is directly impacted by early parenting behaviors and low respiratory sinus arrythmia (RSA, i.e., heart rate variability in synchrony with respiration) is a biomarker of risk in children that may be a critical mechanism to consider in the intergenerational transmission of PTS-related risk but has not yet been examined. Additionally, the relation between maternal PTS and both parenting behaviors and RSA have not been studied in the early postpartum period.

Methods: Data was collected as part of an ongoing study of intergenerational PTS-related risk in the postpartum period with trauma-exposed mothers and their infants recruited from an urban public hospital obstetrics clinic and the greater Atlanta community. Mother-infant dyads (N = 51; maternal mean age = 28.5; 90% Black) attended a 6-week postpartum visit and engaged in a caregiving stressor task. Specifically, during the task, mothers unswaddled infants, removed clothes, changed diaper, and applied lotion before reclothing infants. Maternal PTS were measured using the PTSD Symptom Checklist for DSM-5 (mean (SD) = 19.30 (16.53), range 0-57). Maternal behavior during a caregiving stressor task was coded by blinded coders based on a coding scheme adapted from the National Institute of Child Health and Human Development (NICHD) Study of Early Child Care; codes for maternal sensitivity, intrusiveness, and detachment were used in analyses. Infant RSA data was collected within the study visit on a subsample of the mother-infant dyads who completed the study once additional funding was obtained (n = 28). RSA was captured at baseline and during the caregiving stressor task using electrocardiogram (ECG) sensors; segments were averaged within tasks. First, we evaluated the differential association between overall PTS and PTS clusters with maternal parenting behavior during the caregiving task using bivariate correlation analysis. Differences in patterns based on infant sex were examined based on prior research showing relevant infant sex differences in parenting behavior. Then, correlations between PTS and infant RSA levels at baseline and during stressor task were examined within the subsample. There were no significant differences in means levels of PTS or parental codes between mothers with or without RSA data.

Results: Higher levels of maternal re-experiencing and avoidance PTS were significantly associated with lower maternal sensitivity (r = -.29, p = 0.035 and r = -.38, p = 0.005, respectively), and higher intrusiveness (r = 0.44 p < .001 and r = 0.36, p = 0.009, respectively) but not detachment. Overall PTS was correlated with lower maternal sensitivity (r = -.27, p = 0.054). Correlation analyses split by sex showed that associations were stronger and only remained significant in male infants (n = 26) across maternal sensitivity (PTSD re-experiencing: r = -.39 p = 0.043; PTSD avoidance: r = -.52, p = 0.005), intrusiveness (PTSD re-experiencing: r = 0.62, p < .001; PTSD avoidance: r = 0.54, p = 0.003), and detachment (PTSD re-experiencing: r = 0.39, p = 0.038; PTSD avoidance: r = 0.45, p = 0.015). Within the subsample with RSA data, higher levels of negative cognition and mood PTS (r = -.37, p = 0.043) and hyperarousal PTS (r = -.43, p = 0.017) were significantly correlated with lower infant RSA during the caregiving stressor task. Higher overall PTS was negatively correlated with lower baseline infant RSA (r = -.36, p = 0.053) and lower infant RSA during the caregiving stressor task (r = -.36, p = 0.053).

Conclusions: Maternal PTS early in postpartum are related to lower maternal sensitivity and higher levels of intrusiveness during a lab-based caregiving task. PTS may impact parenting behavior during caregiving more strongly for male infants than female infants, although larger samples are necessary to formally evaluate sex differences in these patterns. Maternal PTS are also related to lower infant RSA during the caregiving task, suggesting a potential link between maternal PTS and an early biomarker of health risk in infants. Given that differential associations were found by PTS clusters, it may be that certain PTS profiles affect early parenting behavior and infant autonomic response in varied ways, and continued longitudinal research in this area with a larger sample is warranted to further clarify pathways of intergenerational PTS-related risk.

Keywords: Posttraumatic stress disorder, Postpartum Period, psychophysiology, Intergenerational transmission of trauma

Disclosure: Nothing to disclose.

P728. Stress and Cannabis Use Change Stress Responses by Altering Perineuronal Nets in the Ventral Pallidum

Ritchy Hodebourg*, Peter Kalivas

Medical University of South Carolina, Charleston, South Carolina, United States

Background: The increasing legalization of cannabis and its high comorbidity with post-traumatic stress disorder (PTSD) necessitates a better understanding of the interaction between stress and cannabis in the brain. Using acute restraint stress combined with a rat cannabis self-administration paradigm, I recently found that cannabis use promotes two primary PTSD-like symptoms: avoidance coping behaviors and the generalization of stress-coping responses to a neutral stimulus (NS) not previously associated with stress exposure. However, the neuroadaptations underlying these changes are poorly understood. Recent studies demonstrated that perineuronal nets (PNNs), the highly condensed form of the extracellular matrix, are dysregulated in several brain regions after stressful events. Although the ventral pallidum (VP) is a brain region well characterized for its role in aversive behaviors, the effects of stress and cannabis on PNNs in this structure are unknown. Here, we aimed to investigate the effects of acute stress, stress-conditioned stimulus (stress-CS), and cannabis use on PNNs in VP

Methods: In the first experiment, rats were restraint stressed for 2h and simultaneously exposed to an odor that became the stress-CS. Control rats were exposed to the same odor in the home cage. 3 weeks after the stress, the effect of the CS or an NS was tested in a defensive burying task (DBT) for 15 min. The active (buying behavior) and avoidant coping strategies (immobility and escape behaviors) and the number of neurons surrounded by PNNs were quantified. Then, to directly assess the impact of PNNs on stress responses, chondroitinase ABC was used to enzymatically degrade PNNs before the DBT. Finally, to evaluate the effect of stress and cannabis on PNNs, 3 weeks after the acute stress, another cohort of rats self-administered cannabinoids (delta9-tetrahydrocannabinol+cannabidiol; THC + CBD) or vehicle for 10 days. After 10 days of withdrawal, coping strategies were evaluated in the DBT.

Results: The acute stress induced an enduring increase in the number of neurons surrounded by PNNs in the VP. Moreover, exposure to a stress-CS led to a decrease in PNNs, which was negatively correlated with the burying behavior, supporting an important role of PNNs in coping strategies. In addition, PNN removal in the VP mimics the effect of cannabis use by generalizing stress responses to the NS. This suggests that cannabis use may worsen PTSD symptoms by reducing PNNs in the VP. Finally, as expected, cannabis withdrawal reduced the number of PNNs in the VP and generalized stress responses even in non-stressed rats.

Conclusions: These data strongly point to the need to investigate how stress and cannabis affect PNNs, potentially contributing to the comorbidity between PTSD and cannabis use disorder.

Keywords: Perineuronal nets, PTSD, cannabis use disorder, Ventral Pallidum

Disclosure: Nothing to disclose.

P729. Obsessive Compulsive Personality Disorder Traits are Associated With Increased Anger and Lower Quality of Life in Trauma-Exposed Veterans

Meghan Kulak*, M. Tracie Shea, Timothy Mariano, Benjamin Greenberg

Alpert Medical School, Brown University, Providence, Rhode Island, United States

Background: Trauma exposure and post-traumatic stress disorder (PTSD) are associated with high rates of co-occurring axis II pathology, including Obsessive Compulsive Personality Disorder (OCPD). OCPD is characterized by a maladaptive level of rigidity and perfectionism that results in internalized distress, interpersonal difficulties, and lower quality of life (QOL). A diagnosis of OCPD may increase the risk of developing PTSD following trauma-exposure and may negatively impact treatment course and outcomes. Like PTSD, individuals with OCPD report high levels of anger which may mediate internal distress and interpersonal difficulties. Despite this, the relationship between OCPD, anger, and QOL in trauma-exposed veterans remains understudied.

Methods: A subsample of 92 veterans with a history of warzone trauma-exposure and at least 3 hyperarousal symptoms and moderate to severe problems with anger were recruited from the VA Medical Center in Providence, RI as part of a larger study on personality disorders in veterans. Standardized interviews and self-reports assessed demographics, medical/psychiatric history and trauma history. A total score for OCPD traits was calculated by summing items that assess the DSM-5 criteria for OCPD using the Schedule for Nonadaptive and Adaptive Personality-2 (SNAP-2). PTSD severity was assessed via the Clinician-Administered PTSD Scale (CAPS). Anger dimensions were assessed via the State-Trait Anger Expression Inventory (STAXI-2), and quality of life was assessed using the global item and two subscales (psychological and social relationships domains) from the World Health Organization Quality of Life Brief Form (WHOQOL-BREF).

Results: SNAP-2 OCPD scores were significantly correlated with state anger (r =.30, p < .01), trait anger (r = 0.34, p < .01), anger expression out (r = 0.25, p < .05), and anger expression in (r = 0.23, p < .05). Of the OCPD criteria, inflexible morality and ethics, reluctance to delegate, and rigidity/stubbornness were most strongly associated with STAXI-2 anger measures. OCPD total scores on the STAXI-2 did not predict PTSD severity on the CAPS, however was significantly associated with the CAPS anger criterion (-.31, p < 0.01). The OCPD total score on the SNAP-2 was not significantly associated with lower global quality of life but was associated with significant impairment in the social relationships domain (r =.38, p < .001). This association remained significant (p = 0.003) in linear regression analyses controlling for PTSD severity.

Conclusions: In a sample already selected for having anger problems, OCPD traits in trauma-exposed veterans are associated with significantly higher scores on anger scales and more impairment in social quality of life. This highlights the importance of screening for OCPD traits in this population to inform treatment strategies and improve outcomes.

Keywords: Combat PTSD, Obsessive-compulsive personality traits, anger

Disclosure: Nothing to disclose.

P730. Endocannabinoid Dysregulation and PTSD in Urban Adolescents: Associations With Anandamide Concentrations and FAAH Genotype

Hilary Marusak*, Samantha Ely, Clara Zundel, Leah Gowatch, MacKenna Shampine, Carmen Carpenter, Reem Tamimi, Alaina Jaster, Tehmina Shakir, Len May, Terri deRoon-Cassini, Cecilia Hillard

Wayne State University, Detroit, Michigan, United States

Background: The endocannabinoid system plays a critical role in regulating fear- and anxiety-related behaviors. Our recent meta-analysis revealed higher circulating concentrations of the endocannabinoids anandamide (AEA) and 2-arachidonoylglycerol (2-AG) in adults with posttraumatic stress disorder (PTSD). This suggests that dysregulation of the endocannabinoid system may contribute to PTSD. Additionally, a common functional polymorphism (C385A, rs324420) in the gene encoding fatty acid amide hydrolase (FAAH)—the enzyme responsible for catabolizing AEA—is associated with higher circulating AEA concentrations, altered fear extinction, and elevated PTSD symptoms in adults. However, there is less understanding of whether the endocannabinoid system is associated with PTSD symptoms in youth. This is critical, as adolescence is a high-risk period for trauma exposure and the emergence of psychiatric disorders and is characterized by dynamic fluctuations in endocannabinoid signaling. Furthermore, PTSD rates among trauma-exposed youth are as high as 15%, but may be even higher in urban areas with high rates of individual and community level trauma exposure. In this study, we examined the effects of trauma exposure and PTSD symptoms on peripheral endocannabinoid concentrations and FAAH genetic variation in a sample of urban youth.

Methods: Detroit-area adolescents (n = 102, 54.9% female, M ± SD = 13.33 ± 2.21 years) and a parent/guardian participated in a cross-sectional community-based study. The sample included 40.2% White Non-Hispanic, 34.3% Black Non-Hispanic, 6.9% White Hispanic, 4.9% Asian/Pacific Islander, and 12.7% Biracial participants (1% not reported). Lifetime exposure to trauma and past-month PTSD symptoms were assessed using the adolescent self-report version of the UCLA PTSD Reaction Index for DSM-5. Plasma concentrations of AEA and 2-AG were quantified using liquid chromatography with tandem mass spectrometry and log-transformed for analysis. Saliva samples were collected for genotyping at rs324420 in FAAH using EndoDNA test kits and the Illumina Infinium high-throughput screening assay.

Results: Ninety percent of youth reported exposure to one or more traumas (M ± SD = 4.02 ± 2.85 categories), and 20% exceeded the threshold for detecting PTSD ( ≥ 35). Youth with higher PTSD symptoms showed higher AEA concentrations (r = 0.23, p = 0.034). This result remained significant after adjusting for age, sex, and body mass index (F = 4.491, p = 0.003), and was similar when using PTSD cutoff scores (p = 0.036). When split by PTSD subtype, the effect of PTSD on AEA was significant for hyperarousal (F = 6.611, p < 0.001) but not for other PTSD subtypes (p’s > 0.05). 2-AG concentrations were not associated with PTSD symptoms or diagnosis (p’s > 0.05). For the FAAH (rs324420) genotype, the A-allele (AA, AC) was associated with significantly higher PTSD symptoms (t = 2.25, p = 0.027) and circulating AEA concentrations (t = 2.27, p = 0.026) compared to CC homozygotes.

Conclusions: This study suggests that elevated AEA concentrations are associated with PTSD symptoms in urban adolescents, particularly within the hyperarousal subtype. These findings extend prior research in adults, indicating that endocannabinoid dysregulation may play a role in PTSD among youth. Additionally, the FAAH (rs324420) A-allele is linked to both higher PTSD symptoms and higher AEA levels, pointing to a potential genetic factor in endocannabinoid system involvement in PTSD. These results underscore the need for further research into the role of the endocannabinoid system in adolescent PTSD and the exploration of targeted interventions.

Keywords: Endocannabinoids, Childhood trauma, Adolescent PTSD, FAAH, Anandamide

Disclosure: Nothing to disclose.

P731. Neurophysiological Correlates of Emotional Numbing During Mindfulness Training in Dissociative Trauma-Exposed Adults: An Electroencephalography Study

Timothy McDermott*, Amanda Johnston, Rebecca Krawczak, Jacob Semerod, Divya Jagadeesh, Negar Fani, Greg Siegle

Emory University School of Medicine, Atlanta, Georgia, United States

Background: Trauma can contribute to dampened emotional processing and is often co-occurring with dissociation, which is an involuntary psychological detachment from one’s present experience (bodily sensations, emotions, attention). The neurophysiological underpinnings of emotional numbing symptoms have previously been examined in cross-sectional neuroimaging studies, implicating frontal regions such as the left dorsolateral prefrontal cortex (dlPFC) and dorsomedial prefrontal cortex (dmPFC) as either being over- or under-modulated depending on the task context and response of interest. Examination of individual variability in emotional numbing over time could further elucidate the neurophysiological correlates of emotional numbing and strengthen the reliability of findings by identifying temporally stable relationships. Our ongoing clinical trial (NCT04670640) compares four mindfulness training interventions with a cross-design of interoceptive (breath focus) and exteroceptive (vibration feedback) awareness seeking to reduce dissociation in trauma-exposed adults. Here, we examined if self-reported ratings of emotional numbing during mindfulness training related to neurophysiological activity recorded by electroencephalography (EEG). As this is an ongoing trial, we did not test for differences between mindfulness training conditions.

Methods: 94 EEG recordings were collected from 19 trauma-exposed adults (mean age=34.11 years; SD = 10.87 years; 47.4% identified as men) who volunteered to participate in the mindfulness clinical trial. EEG data were collected using 64-channel HydroCel Geodesic Sensor Net 130 system [Electrical Geodesics Inc. (EGI)]. EEG recordings consisted of six 3-minute blocks of mindfulness practice (18 minutes per visit), and visual analog ratings of feeling “Numb” were collected before and after each block (7 ratings per visit). EEG recordings underwent standard data cleaning procedures and visual quality checking, and activity was extracted and averaged across all six blocks for canonical frequency bands (theta = 4-7 Hz; alpha = 8-12 Hz; beta = 15-25 Hz; gamma = 35-45 Hz). Electrodes of interest were located over lateral frontal (EGI 64-Channel Montage: electrodes 10-14; included F3 and Fp1) and medial frontal (EGI 64-Channel Montage: electrodes: 3, 6, 8-9; included Fz) areas. Activity was averaged across electrodes of interest for each frequency. “Numb” ratings were averaged for each visit, and this average was the criterion in a linear mixed-effects regression model with left frontal and medial frontal activity for each frequency band as predictors (total of 8). Training visit was included as a fixed effect, and participant ID was included as a random effect with an auto-correlation function to account for within-subject residuals (order 1 structure). The model used maximum likelihood estimation to provide estimations for missing data points and outliers. An α threshold of p < .05 indicated statistical significance.

Results: In the linear mixed-effect model, there were significant effects of medial frontal theta activity [F = 5.97; p = 0.020], medial frontal alpha activity [F = 5.08; p = 0.043], medial frontal beta activity [F = 7.21; p = 0.009], and training visit [F = 14.8; p < .001] on “Numb” ratings. These effects were such that theta and beta activity were inversely related to “Numb” ratings (theta: r = -.31; beta: r = -.34), while alpha activity was positively related to “Numb” ratings (r = 0.29). Training visit was also inversely related to “Numb” ratings (r = -.45), which demonstrated that increased mindfulness practice led to decreased emotional numbing over time. Interactions between training visit and each frequency for lateral and medial frontal electrodes of interest predicting “Numb” ratings were not significant (p’s > .084); as some interactions were marginally significant, this will be tested again when the study has a larger sample.

Conclusions: Our findings from this ongoing clinical trial of mindfulness training for dissociative trauma-exposed adults demonstrate that practicing mindfulness improves symptoms of emotional numbing, likely by increasing neurophysiological activity correlated with attentional control. This includes increased medial frontal theta activity, which has been shown to reflect increased engagement of attentional control, increased medial frontal beta activity, which reflects increased sensorimotor control, and decreased medial frontal alpha activity, which is inversely associated with attentional control. This study provides support for the relevance of neurophysiological mechanisms of attentional control in decreasing subjective feelings of numbness, potentially reflecting increased engagement with present experience. Future work from the completed clinical can test for differences between mindfulness training conditions on emotional numbing symptoms and examine for differential changes in neurophysiological mechanisms.

Keywords: Electroencephalography (EEG), Dissociation, Trauma exposure, Emotion Circuitry, Mindfulness Meditation

Disclosure: Nothing to disclose.

P732. Traumatic Stress Acutely Modulates Neural Reactivity to Visual Stimulation

Nathaniel Harnett*, Grace Rowland, Kate Webb, Kerry Ressler, Isabelle Rosso

Harvard Medical School, McLean Hospital, Belmont, Massachusetts, United States

Background: Threat neurocircuitry plays a demonstrable role in the development of posttraumatic stress disorder. However, contributions of sensory circuits – notably visual circuits – are integral to appropriate threat processing. Here, we investigated the influence of traumatic stress on reactivity to visual stimuli and subsequent relationships with threat processing and PTSD symptom development.

Methods: A total of 24 recent trauma-exposed (TE) participants and 15 non-trauma exposed (NTE) controls were recruited from emergency departments and the community. Participants completed assessments of PTSD symptoms (using the PCL-5) and prior trauma (using the LEC-5 and CTQ). Participants completed a visual evoked response procedure consisting of an alternating contrast flickering checkerboard (8Hz) block with an attentional check block (15 seconds per block). We used standard general linear models to generate neural reactivity maps to the blocks and used the maps in group-level analyses with cluster correction approaches to maintain a = 0.05.

Results: Robust visually evoked responses were observed during stimulation within visual cortex (pcluster-corrected < 0.001). TE participants showed greater deactivation within the visual cortex compared to NTE participants [t(37) = 5.74, pcluster-corrected < 0.05]. Further, we observed an interaction with the dorsomedial prefrontal cortex such that TE participants showed greater deactivation to checkerboard stimulation compared to NTE participants but no difference was observed during the attentional check [F(1,37) = 40.46, pcluster-corrected < 0.05]. Groups did not differ in LEC-5 [t(37) = 1.99, p = 0.054] or CTQ [t(36) = 1.65, p = 0.109], but differential reactivity within visual cortex was positively associated with PCL-5 scores in the TE group [r(21) = 0.47, p = 0.028].

Conclusions: The present findings suggest traumatic stress induces alterations in visual processing of stimuli. Specifically, trauma may potentiate visual reactivity to stimulation and be associated with altered attentional processing of stimuli, as indicated by the interaction within the dorsomedial prefrontal cortex. Further, the altered reactivity may be related to future PTSD development, although additional longitudinal research is needed.

Keywords: Trauma exposure, PTSD, fMRI, visual cortex

Disclosure: Nothing to disclose.

P733. Trauma Produces HDAC6 Suppression in the Amygdala of Humans and Rodents With PTSD

Robin Bonomi*, Mika Naganawa, Delaney McRiley, Takuya Toyonaga, David Matuskey, Matthew Girgenti, Kelly Cosgrove

Yale University, Department of Psychiatry, New Haven, Connecticut, United States

Background: Histone deacetylase 6 (HDAC6), one of the 18 HDACs, is a primary regulator of gluocorticoid receptor (GR) nuclear translocation and highly expressed in serotonergic neurons of the brain. Working as a primary regulator for GR chaperone complex, HDAC6 is key in the development of symptoms following trauma. Using 18F-Bavarostat, we are able to effectively image this enzyme in the brain of living individuals and animals for the first time.

Methods: 27 individuals were recruited for this study including non-trauma exposed (N = 8), trauma -exposed (experienced criterion A trauma but without meeting PTSD criteria) (N = 9), and individuals with criterion A trauma who meet clinical diagnostic PTSD symptoms based on CAPS-5, PLC-5, and SCID (N = 9). Individuals underwent 120min [18F]-Bavarostat PET scan (bolus injection of 146□48 MBq) on Siemens mCT. PTSD symptomatology was characterized by Clinician Administered PTSD Scale for DSM-5 (CAPS-5). Individuals with significant medical or psychiatric conditions were excluded. Arterial blood sampling allows measurement of the metabolite-corrected input function. MA1 analysis using plasma input function was used to estimate volume of distribution (VT) for regions of interest determined using an anatomic labeling (AAL) template co-registered to each subject’s T1-weighted structural MRI (MPRAGE; isotropic voxel: 1mm3).

Male Sprague Dawley rats underwent single prolonged stress model and were imaged with 18F-Bavarostat (0.3 + /- 0.1 mCi) for visualization of HDAC6 both before and after stress (N = 12) Behavior analysis confirmed PTSD- like phenotype in stressed animals at 7 days post stress using open field testing. Similar studies were done at 48- hours post stress to examine the acute stress period.

Results: Demographically, the groups were well matched with no significant differences in gender or race (p = 0.96, p = 0.12, respectively). PTSD individuals had significantly higher scores on the PCL-5 and CAPS-5 scoring panels than the trauma-exposed control group p < 0.05 using multiple comparisons ANOVA for each of the questionnaire). HDAC6 expression in the PTSD group was significantly lower than the healthy control group in the amygdala (p = 0.008). Clinical correlations were examined but significant correlations with PET imaging findings were not realized for this group.

The open field test (41) confirmed multiple parameters consistent with a PTSD-like phenotype including: decreased center time (27.9 sec control vs 18.3 sec SPS, p = 0.78) and number of center entries (15.5 control vs 8.72 min SPS, p = 0.013), increased latency to accessing the center (55.4 seconds control vs 88.8 seconds SPS, p = 0.047), decreased total distance traveled (2.39 meters control vs 1.38 meters SPS, p = 0.001). A separate acute stress group underwent testing at 48 hours post SPS, no significant differences in behaviors were demonstrated. Significant reductions in the SPS group (N = 10) were seen in the hippocampus/amygdala (p = 0.0028), limbic cortex (p = 0.0123), brainstem (p = 0.0036), and striatum (p = 0.0032). Notably, the whole brain and cerebellum did not show significant group differences. To further confirm the reduction in HDAC6 following SPS, within subject comparison demonstrated, robust reductions in HDAC6, where the stressed animals had an average SUV of 2.33 in the striatum and 2.21 in the amygdala /hippocampus (striatum, p = 0.066 and amygdala/hippocampus p = 0.086). Cross group comparison for animals at 48- hours post stress was also performed with comparable group injection parameters at time of scan (Supplemental table 4).

The acute stress group at 48 hours post SPS also demonstrated lower HDAC6 in these regions with a mean SUV in the amygdala/hippocampus of 2.21 (p = 0.086), brainstem 2.51 (p = 0.113), striatum 2.33 (p = 0.066). at the acute post-trauma period, prior to behavior differences Injection parameters across all groups were well matched.

Conclusions: This study marks the first translational study of HDAC6 using [18F]-Bavarostat and the first to study it in a clinical population. This work begins to probe the understanding of changes in the glucocorticoid system following trauma via HDAC6 chaperone regulation. This allows for novel pharmacologic tools to be developed targeting individuals with significant trauma exposure but before they demonstrate symptoms of PTSD. A larger sample size and more research may further elucidate important clinical correlations between HDAC6 and PTSD symptoms.

Keywords: F-18 PET Imaging, HDAC, PTSD, Translational research, Clinical Neurobiology

Disclosure: Nothing to disclose.

P734. One Year Outcomes from a Randomized Controlled Trial of Transcranial Direct Current Stimulation and Virtual Reality for PTSD

Noah Philip*, Kyra Brettler, Benjamin Greenberg, Amanda Arulpragasam, Samantha Cilli, Emily Aiken, Mascha van ‘t Wout

Brown University and VA Providence, Providence, Rhode Island, United States

Background: Background: Posttraumatic Stress Disorder (PTSD) is a debilitating psychiatric condition with a profound impact on quality of life. Our group conducted a randomized controlled tiral that demonstrated transcranial direct current stimulation (tDCS) combined with virtual reality improved PTSD psychophysiological arousal, with effects lasting up to three months. However, longer-term outcomes are critical when considering whether and how and when to apply this intervention in the real-world.

Methods: This study reviewed naturalistic outcomes up to 1-year from study endpoint, with the hypothesis that those in the active group would have superior outcomes. The primary outcome was psychiatric relapse, defined as suicide attempt, inpatient or PTSD residential program admission, use of rescue transcranial magnetic stimulation, or emergency/urgent care visits. Changes in medications were examined, and whether psychophysiological habituation during tDCS+VR predicted survival.

Results: All (n = 54; 100%) of the original sample was included. A total of 31% (n = 17) participants relapsed; of these, the majority (n = 13/17, 76.5%) were in the sham group; mean survival time for the sham group was 231 ± 28 days, compared to the active group which was 315 ± 22 days (p = 0.019; odds ratio=3.02 CI 1.1-8.1). Reduction in medication was more frequent in the active group (p = 0.01), and habituation correlated with survival time (p = 0.04) in the sham group.

Conclusions: Our findings indicate tDCS+VR holds promise as an effective intervention for PTSD, with durable long-term clinical outcomes and reduction in medications. Psychophysiological habituation was associated with superior long-term outcomes. These observations support the utility of this novel therapeutic approach in PTSD management and can inform future mechanistic inquiry and treatment development.

Keywords: PTSD, TDCS, veterans

Disclosure: Nothing to disclose.

P735. Dopamine in the Tail of the Striatum Manages Potential Threats

Iku Tsutsui-Kimura*

Hokkaido University, Sapporo, Japan

Background: Canonical dopamine signals reward value and reinforce rewarding action. Conversely, we recently found that dopamine in the tail of the striatum (TS) signals stimulus salience and promote avoidance to a novel object. However, the roles of TS dopamine in more naturalistic environment with a potential threat was still largely unrevealed.

Methods: We used a semi-naturalistic foraging paradigm to model avoidance of potential threats. In this paradigm, a thirsty mouse freely goes out of a homing area and enters a foraging arena to obtain a water reward. In test sessions, a predator-like object (“monster”) surges when a mouse approaches a reward. In the presence of the monster, mice failed to retrieve reward and fled into the homing area. Across trials, mice returned home earlier and earlier before the monster charged, indicating that mice learned to predict a potential threat.

Results: Over multiple monster sessions, mice gradually succeeded in acquiring reward. This task, thus, allows us to study various aspects of threat managements under threat-reward conflicts, such as threat avoidance, threat prediction and eventual overcoming. We found that ablation of dopamine neurons that project to the TS impaired threat avoidance and prediction. Mice with ablation of dopamine receptor type-1-expressing medium spiny neurons (D1-MSNs) decreased avoidance of the monster, while mice with ablation of D2-MSNs exhibited intact avoidance but failed to improve success rate. D1-MSNs were more activated in avoidance trials (failure of reward acquisition) while D2-MSNs were more activated in successful trials, especially in later sessions.

Conclusions: These results indicate differential roles of D1- and D2-MSNs in threat management: avoidance and overcoming, respectively. These findings provide a basic principle of how D1- and D2-MSNs dynamically regulate different aspects of behavior under the control of dopamine.

Keywords: Dopamine, Medium spiny neurons, approach-avoidance conflict

Disclosure: Nothing to disclose.

P736. Neurobehavioral Effects of Aversive-To-Appetitive Counterconditioning in Posttraumatic Stress Disorder

Samuel Cooper*, Nicole Keller, Ameera Azar, Sophia Bibb, Sydney Lambert, Augustin Hennings, Elizabeth Bauer, Emily Leiker, Jarrod Lewis-Peacock, Joseph Dunsmoor

The University of Texas At Austin, Austin, Texas, United States

Background: Background: Exposure therapy for PTSD is based on prolific threat extinction work. In extinction paradigms, a danger cue (CS + ) is no longer paired with an aversive outcome (US), forming a new inhibitory safety memory that outcompetes the established fear memory. Similarly, exposure therapies for PTSD provide patients with trauma-relevant situations that cannot actually result in the expected negative outcome. Unfortunately, some patients fail to respond or experience symptom relapse. One issue is that although fear might habituate during exposure, a robust safety memory does not reliably emerge and fear returns. Thus, effective and noninvasive techniques to augment extinction and enhance safety learning are vital for PTSD treatment research. Human neuroimaging work indicates that rewarded extinction strengthens safety learning relative to standard extinction, but this has not yet been tested in PTSD. We address this gap by testing if rewarded extinction (i.e., aversive-to-appetitive counterconditioning) outperforms standard extinction in PTSD, and approximates successful extinction as seen in non-PTSD comparisons.

Methods: Participants (25 comparison, 32 PTSD, 73% women, 27% Hispanic [representative of recruitment population]) completed a 2-day fMRI experiment. Two CS + ’s were paired with an aversive US on Day 1, followed by one CS+ being paired with a pleasant US (CS+Rew) and the other undergoing standard extinction (CS+Ext). Neural responding to both CS + ’s was measured 24-hours later.

Results: Univariate fMRI extinction analyses found relatively lower CS+Rew neural activity in threat regions (insula, dorsal anterior cingulate [dACC], periaqueductal gray) compared with CS+Ext for all participants, indicating rewarded extinction facilitated increased safety learning over standard extinction. However, safety learning during extinction, as characterized in ventral medial prefrontal cortex (vmPFC), was improved in comparison participants, but not PTSD, for the CS+Rew, β = 0.48 95%CI = [0.09, 0.87], t = 2.41, p = 0.01. Multivariate pattern analysis in the vmPFC further showed a lack of a CS+Rew pattern reinstatement in the PTSD, but not comparison, group. Additionally, the CS+Rew and CS+Ext patterns were significantly reinstated in the dACC for PTSD, but not comparison participants, at follow-up.

Conclusions: Our results offer some evidence for rewarded extinction’s efficacy in PTSD and largely accord with increasing focus on enhancing inhibitory learning during exposure therapy. Group differences in putative threat vs. safety regions indicate rewarded extinction might operate as fundementally different neural process than standard extinction. Limitations include: sample size precluded analyses of individual PTSD symptom dimensions in relation to rewarded extinction, and durability of learning over longer time periods (e.g., months) was not tested. We discuss next steps and neuroscience-informed considerations in implementing rewarded extinction within exposure therapy.

Keywords: PTSD, fMRI, MVPA, Fear conditioning and extinction, Reward

Disclosure: Nothing to disclose.

P737. Exploring the Association Between Self-Reported Sleep Quality and Prefrontal Glutamate and GABA in Individuals With Posttraumatic Stress Disorder

Meredith Reid*, Sarah Comer, Faith Miller

Auburn University, Auburn, Alabama, United States

Background: Posttraumatic stress disorder (PTSD) is often accompanied by poor sleep quality. Converging lines of evidence point to glutamatergic and GABAergic dysfunction in PTSD. Previous studies using magnetic resonance spectroscopy (MRS) have reported an association between insomnia severity and glutamate and GABA in the parieto-occipital cortex of individuals with PTSD. The primary purpose of this study was to measure glutamate concentrations in the prefrontal cortex of participants with PTSD, trauma-exposed participants without PTSD, and participants without trauma exposure. In an exploratory analysis, we examined the correlation between self-reported sleep quality and prefrontal glutamate and GABA. We hypothesized that individuals with PTSD would have lower glutamate levels and worse sleep quality compared to trauma-exposed individuals without PTSD and individuals without trauma exposure. We further hypothesized that sleep quality would be correlated with glutamate and GABA levels.

Methods: The sample consisted of 27 participants with PTSD, 27 trauma-exposed participants without PTSD (TE), and 26 participants without trauma exposure (NT). PTSD Criterion A trauma exposure was determined from the Life Events Checklist for DSM-5 extended version based on participants’ self-reported worst event (i.e., index event). PTSD symptom severity was measured using the PTSD Checklist for DSM-5. Sleep quality was assessed in a subset of participants (26 PTSD, 24 TE, 21 NT) using the Pittsburgh Sleep Quality Index (PSQI). MRS data were acquired on a 7T Siemens MAGNETOM scanner. Three-dimensional structural images were acquired for anatomical reference and segmentation. Spectra were acquired from a voxel in the left dorsolateral prefrontal cortex (25 x 25 x 25 mm) using an ultra-short TE STEAM sequence (TE/TR/TM = 5/10,000/45 ms, 32 averages, 4 kHz spectral bandwidth, 2048 points). Spectra without water suppression (4 averages) were acquired for pre-processing and quantification. Spectra were processed and fit with Osprey and LCModel. Tissue- and relaxation-corrected molal concentration estimates (mmol/kg of tissue water) of glutamate were calculated. Metabolites were compared between groups using ANCOVA controlling for age and sex followed by post hoc pairwise t-tests with Tukey-adjusted p-values. Pearson correlation coefficients were used to explore the association between PSQI scores and glutamate and GABA.

Results: The groups were comparable in age and sex. There was a significant group difference in glutamate (F(2,71) = 6.17, p = 0.003). Glutamate was significantly lower in the PTSD group compared to the NT group (p(Tukey) = 0.005, Cohen’s d = 0.92) and significantly lower in the TE group compared to the NT group (p(Tukey) = 0.02, Cohen’s d = 0.80). GABA was not significantly different between the groups (F(2,62) = 0.66, p = 0.52). There was a significant group difference in PSQI scores (F(2,68) = 16.18, p < 0.001). PSQI was higher in the PTSD group compared to the TE group (p(Tukey) = 0.007, Cohen’s d = 0.88) and the NT group (p(Tukey) < 0.001, Cohen’s d = 1.66) and higher in the TE group compared to the NT group (p(Tukey) = 0.03, Cohen’s d = 0.78). There was a trend-level association between glutamate and PSQI in the combined sample (r(67) = -0.21, p = 0.08), but the correlations within each group were not significant (PTSD: r(24) = -0.25, p = 0.22; TE: r(22) = -0.01, p = 0.97; NT: r(17) = 0.28, p = 0.24). The correlation between GABA and PSQI was not significant (combined sample: r(59) = -0.12, p = 0.37; PTSD: r(22) = -0.09, r = 0.67; TE: r(17) = -0.20, p = 0.42; NT: r(16) = 0.07, p = 0.77).

Conclusions: We observed that individuals with PTSD had lower glutamate and worse sleep quality compared to individuals without PTSD. The results of our study add to the growing evidence of glutamatergic dysfunction in individuals with PTSD. We provide some preliminary evidence of an association between prefrontal glutamate and self-reported sleep quality, but more studies with larger samples are needed to validate this finding. Further research is needed to unravel the precise mechanisms underlying glutamate alterations in individuals with PTSD and the implications for the pharmacological and behavioral treatment of the disorder. Interventions that enhance glutamatergic function are promising targets for drug development.

Keywords: PTSD, sleep, magnetic resonance spectroscopy (MRS), glutamate, dorsolateral prefrontal cortex (DLPFC)

Disclosure: Nothing to disclose.

P738. Nicotinamide Supplementation Prevents Gestational Stress-Induced Deficits in Maternal Care and Metabolism

Fiona Hollis*, Breanna Robertson, Gabrielle Wood, Archana Ramesh, Hannah Brennen, Jiya Desai, Baruk Srivastava, Alexia Crockett, Erin Gorman-Sandler

University of South Carolina School of Medicine, Columbia, South Carolina, United States

Background: Postpartum depression (PPD) is an understudied psychiatric complication of childbirth that afflicts up to 20% of women and can induce lasting consequences for their children. Mitochondria are dynamic organelles that are crucial for cell homeostasis and share a link with many of the proposed mechanisms underlying PPD etiology. Neurons rely on mitochondrial energy production to function, and stress, a major risk factor for PPD, increases brain energy demands. In turn, brain mitochondrial function is also affected by stress and linked to avoidance and social behaviors. We previously showed that gestational stress in rats decreased mitochondrial complex I respiration in the prefrontal cortex (PFC) in association with behaviors relevant to PPD. However, whether PFC mitochondrial respiration had functional consequences for behavior was unknown. We hypothesized that enhancing complex I respiration during stress exposure would prevent these stress-induced PPD-relevant behaviors.

Methods: Nulliparous and time-mated adult female Wistar rats (n = 8-10/group) received nicotinamide (NAM, a NAD+ precursor that stimulates complex I respiration, 3mM) or vehicle (VEH) in the drinking water from gestational day 8-21. On gestational day 10, females were exposed to 10 consecutive days of chronic mild unpredictable stress or handling. PPD-relevant behaviors including body weight gain, maternal care, and elevated plus maze were assessed from postpartum day 2 (PD2) – PD10. Ex vivo mitochondrial respiration, protein expression, and plasma corticosterone were measured in PFC of females at PD11-12 using high resolution respirometry, immunoblots, and ELISA, respectively. Data were analyzed by 3-way ANOVA, followed by post-hoc tests where appropriate. Associations between mitochondrial measures and behavior were tested with Pearson correlations.

Results: Gestational stress significantly decreased body weight gain (p = 0.005), maternal care (p = 0.04), and increased avoidance behavior in the elevated plus maze (p = 0.03). Ex vivo analysis replicated previous findings of reduced mitochondrial respiration (p = 0.02) and complex I protein (p = 0.04) in the PFC. Furthermore, preliminary analysis suggests that stress blunted resting plasma corticosterone levels, relative to postpartum controls (p = 0.03). NAM treatment did not prevent stress effects on body weight (p = 0.85) or avoidance behavior (p = 0.94), but mitigated decreases in maternal care (p = 0.001), mitochondrial respiration (p = 0.04), and potentially plasma corticosterone levels (p = 0.08). Moreover, we observed a significant correlation between maternal care and PFC mitochondrial respiration (R2 = 0.38, p = 0.03). While NAM had no effect on complex I protein level, complex V expression was significantly increased (p = 0.02), highlighting a potential mechanism of enhanced mitochondrial respiration.

Conclusions: Here we showed that brain mitochondrial respiration can functionally modulate specific postpartum behaviors. Moreover, our effects in vehicle-treated animals further highlight stress as a key mediator of postpartum health. Finally, we show proof-of-principle evidence that targeting mitochondrial function during stress can be a viable strategy to mitigate specific aspects of stress-induced behaviors and endocrine changes relevant to PPD.

Keywords: postpartum depression, Mitochondrial Respiration, prefrontal cortex, Gestational stress

Disclosure: Nothing to disclose.

P739. Demographic and Clinical Characteristics Associated With Postpartum Depression Risk in Women With a History of Depressive Disorders in a Community-Based Health Care System: A Retrospective Observational Study

Kathryn Ridout*, Kathleen Branhan, Chethana Eswarrapa, Mubarika Alavi, Kelli Peterman, Brooke Harris, Lyndsay Avalos, Samuel Ridout

The Permanente Medical Group, Santa Rosa, California, United States

Background: Postpartum depression (PPD) is the most common psychiatric condition after childbirth, affecting 1 in 7 patients in the U.S., and is a leading contributor to maternal mortality by increasing suicide risk. Patients with PPD have poorer postpartum outcomes, decreased maternal-infant bonding, and show impaired infant cognitive-emotional development. Identifying characteristics of patients at risk for PPD can facilitate early identification and personalized intervention. Understanding risk factors in patients with a history of depression is particularly important given the strong relation between PPD risk and a personal history of depression. Previous literature has identified maternal characteristics associated with PPD risk in the overall population, but not in patients with a history of depression. This study aimed to close this literature gap by identifying demographic and clinical characteristics associated with PPD risk in patients with a history of depressive disorders.

Methods: Retrospective cohort study of electronic health record data in a large, diverse, integrated, community-based health care system with universal PPD screening at six weeks postpartum. Inclusion criteria included an age ≥18-years-old with a live birth delivery and a history of depression as defined by International Classification of Diseases (ICD)-9 or -10 codes in the year prior to their last menstrual period between January 1st, 2010, to December 31st, 2019. Exclusion criteria included one or more previous primary diagnoses of schizophrenia, schizoaffective disorder, psychotic disorders, or bipolar disorders. Using modified Poisson regression, demographic and clinical characteristics were modeled as exposures and evaluated for association with the outcome of PPD. PPD was defined as an international classification of diseases (ICD)-9 or -10 diagnostic code or a patient health questionnaire (PHQ)-9 score of ≥10 up to a year after delivery. A PHQ-9 score of ≥20 defined severe PPD.

Results: Of the 6,552 patients in the cohort, 2,469 (37.7%) developed PPD. Patients identifying as Asian/Pacific Islander, Black, and LatinX had significantly higher risk of PPD compared to patients identifying as White (aRR=1.54; 95% CI 1.18-2.01; aRR=1.65; 95% CI 1.25-2.17 and aRR=1.29; 95% CI 1.06-1.58, respectively for severe symptoms). Patients with median neighborhood incomes of < $60,000 or $60,000-120,000 had higher risks of severe PPD compared to a median income of > $120,000 (aRR=1.33; 95% CI 1.02-1.73 and arr=1.31; 95%CI 1.06-1.62 for severe symptoms) as did patients receiving Medicaid insurance (aRR=1.09; 95% CI 1.02-1.28). Supratherapeutic antidepressant dosing prior to pregnancy was associated with a significantly higher risk of severe PPD (aRR=1.34; 95%CI 1.02-1.76) as was a history of PPD (aRR=1.31; 95% CI 1.21-1.41). Patients with any depressive symptoms early in pregnancy had a higher risk of PPD compared to none, with patients having severe depressive symptoms early in pregnancy showing the highest risk of PPD (aRR=4.74; 95% CI 3.54-6.35 for severe postpartum symptoms). Higher parity was associated with a lower risk of PPD (aRR=0.90; 95%CI 0.84-0.98 for 1 and aRR=0.90; 95%CI 0.83-0.99 for 2 + ) while age, carrying multiple gestations, or starting prenatal care after the first trimester were not associated with PPD risk.

Conclusions: Women with a history of depression have unique predictors of PPD compared to the general pregnant population. Previous literature examining PPD risk by race and ethnicity in the general population reported a lower risk for patients identifying as non-White; our findings of higher risk in non-White patients with a history of depression may reflect trends towards later depression detection and care engagement and suggest a care implementation gap. We did not find a significant association between age and PPD; previous literature reporting a higher risk of PPD in older patients with a history of depression were in nulliparous women, suggesting that risk for PPD by age may be specific to patients in their first pregnancy. We show that depression outside of pregnancy, depression during pregnancy, and depressive symptom severity all increase PPD risk, which may warrant additional surveillance, care, or support for these patients. Previous work examining parity or multiple gestations reported mixed findings related to PPD risk. In this cohort of patients with a history of depression, we found higher parity was associated with lower PPD risk, while carrying multiple gestations was not associated with PPD risk. Literature examining prenatal care start date in the first trimester versus later and PPD risk is sparse; our non-significant findings suggest that factors inherent to the process of care do not significantly contribute to PPD risk in this cohort. Future work should investigate the effectiveness of targeted interventions for these risks factors and can inform innovations in screening and personalized care delivery.

Keywords: postpartum depression, Risk factors, Real World Data

Disclosure: Nothing to disclose.

P740. Differences in Circulating Immune Markers in Pre- and Post-Menopausal Women With Major Depressive Disorder and Their Association With Symptoms of Anhedonia

Nabila Haque*, Snoben Kuruvila, Abu Minhajuddin, Cherise Chin-Fatt, Manish Jha, Madhukar Trivedi

The University of Texas Southwestern Medical Center, Department of Psychiatry, Dallas, Texas, United States

Background: Subgroups of individuals with depression have elevated levels of circulating markers of inflammation, such as c-reactive protein. Based on literature linking menopausal status to inflammation, this report evaluated whether menopause in women with depression is associated with elevated levels of circulating immune markers.

Methods: Pre- (n = 49) and postmenopausal (n = 45) women participants of Combining Medications to Enhance Depression Outcomes trial (NCT00590863) with immune biomarker data were included. Circulating levels of immune markers were assayed with Bioplex Pro™ human cytokine 27-plex kit. Differences in levels of immune markers based on menopausal status were compared with unpaired t-tests after log transformation due to skewed distribution. Spearman correlation coefficients were used to estimate the association of immune markers with symptoms of anhedonia.

Results: Postmenopausal women had higher mean level of Eotaxin [mean difference (95% Confidence Limit; CL) = 0.32 (0.18, 0.46)], interleukin (IL) 13 [mean difference (95% CL) = 0.40 (0.11. 0.69)], IL 5 [mean difference (95% CL) = 0.12 (0.01. 0.23)], and tumor necrosis factor alpha [mean difference (95% CL) = 0.10 (0.01. 0.19)] compared to premenopausal women. Only Eotaxin levels were significantly different after Bonferroni adjustment for multiple comparisons. Higher levels of Eotaxin were associated with higher levels of anhedonia (rspearman = 0.37, p = 0.0137) in postmenopausal women but not in premenopausal women (rspearman = -0.11, p = 0.47).

Conclusions: Menopause in women with depression may be associated with distinct patterns of abnormalities in immune system. In this exploratory analysis, we found evidence for increased levels of Eotaxin in postmenopausal women, which in turn was associated with higher levels of anhedonia. Prior research has linked Eotaxin to decreased neurogenesis, neurodegeneration and aging. Future research is needed to elucidate the potential mechanisms underlying this link between menopausal status, levels of Eotaxin and symptoms of anhedonia and to determine whether Eotaxin can serve as a biomarker to predict treatment response or as a therapeutic target.

Keywords: menopause, inflammation, anhedonia, Depression, Adaptive immunity

Disclosure: Nothing to disclose.

P741. Cellular Mechanisms Mediating the Long-Term Effects of Pregnancy, Postpartum and Stress Experiences in the Maternal Brain

Jennifer Chan*, Giuseppina di Salvo, Ashley Cunningham, Sohini Dutta, Elizabeth Brindley, Ian Maze

Icahn School of Medicine At Mount Sinai, New York, New York, United States

Background: Pregnancy and postpartum periods represent incredible physiological stressors affecting over 100 million people who give birth each year globally. Yet, while the effects of environmental and psychosocial stresses are well-studied, how these reproductive experiences persistently impact the maternal brain remain unknown. Moreover, parity (previously carrying one or more pregnancies to term) is a complex risk factor. In general, parity promotes parenting adaptations. However, in some individuals, parity is associated with increased susceptibility for perinatal or postpartum mood and affective disorders. Our understanding as to why some (and not all) women become more susceptible to brain disorder risk following birth is unclear. This gap in knowledge necessitates clarification of the additional risk factors that interact with parity, as well as a need to understand the mechanisms that both orchestrate and disrupt the adaptive long-term effects of reproductive experiences in the maternal brain.

Methods: To determine how reproductive experiences influence brain plasticity long-term, we used bulk RNA-seq to identify 1) brain regions with high vs low transcriptional sensitivity to parity, 2) the timing when changes emerge across pregnancy and postpartum windows, and 3) the contribution of individual reproductive events on brain region programming (N = 4-8). To examine the functional impact of these gene expression changes, we used behavioral assays directed specifically towards brain regions of interest, including contextual fear conditioning, object location task, and pup retrieval (N = 6-12). Next, we assessed the impact of maternal stress during postpartum - a period of dynamic neuroplasticity which we observed to be crucial for programming of parity-sensitive brain regions - on long-term behavioral changes. To do this, we used a maternal separation with limited nesting paradigm, in which stressed dams were separated from pups for 3 hours per day from postpartum days 10-20. To detect cell type-specific changes contributing to parity- and postpartum stress-induced brain alterations, we used single nuclei RNA-seq (N = 4-6). Finally, integrating our bulk and single cell RNA-seq findings, we used a projection-specific chemogenetics approach to test the mechanism by which postpartum exposures impart persistent cellular changes contributing to behavioral plasticity (N = 8-15).

Results: In our comparison of mouse dams one month after offspring weaning vs age-matched nulliparous females, we identified the dorsal hippocampus (dHpc) as the region most transcriptionally sensitive to parity (502 differential genes, adj. p < 0.05). Analyses of gene co-expression networks and predicted upstream regulators suggest robust transcriptional changes by parity may be mediated by sensitivity to hormone receptors, oxytocin and dopamine signaling. These gene expression changes in dHpc corresponded to enhanced performance in contextual fear conditioning by mouse dams, as indicated by significantly heightened freezing during acquisition and context retrieval (p < 0.05). Comparing dHpc changes over the pregnancy and postpartum periods, we observed that persistently altered gene expression networks (one month after postpartum end) begin to arise during pregnancy and become more robustly expressed during postpartum. To tease apart the specific contributions of pregnancy, parturition, and pup interactions on dHpc changes by parity, we compared the transcriptome of pregnancy-only vs pup sensitized female mice. We observed that while pregnancy is the major contributor of dHpc gene expression changes, complete dHpc programming involves additional key postpartum experiences, such as pup interactions. Thus, we next tested whether adverse experiences over the postpartum period could disrupt the long-term effects of parity on the dHpc. Using a mouse model of maternal separation with limited nesting, we show that postpartum stress inhibited both the transcriptional and behavioral adaptations observed in control dams. We used single nuclei RNA-sequencing to determine potential cellular origins of persistent dHpc changes. Interestingly, we identified populations of dopamine receptor-expressing neurons altered both by parity alone and in combination with postpartum stress, suggesting changes in dopamine regulation may also underlie dHpc behavioral changes. Indeed, we observe that dHpc dopamine signaling from the ventral tegmental area (VTA), which is highly responsive during normal postpartum interactions, is dysregulated by chronic postpartum stress. Finally, using a chemogenetics approach to inhibit projection neurons from the VTA to the dHpc, we demonstrate the important role of dopaminergic signaling in mediating the long-term, adaptive effects of parity in dHpc cellular and behavioral plasticity.

Conclusions: This research provides novel insight into the persistent effects of pregnancy and postpartum exposures in the maternal mouse brain. We demonstrate transcriptional and cellular mechanisms involved in promoting long-term behavioral adaptations, which are vulnerable to interference by chronic postpartum stress. Altogether, this work provides a novel framework for the term “mommy brain”, which has long implied negative cognitive changes after birth; instead emphasizing the importance of stress reduction during these dynamic windows of brain plasticity to support maternal adaptations.

Keywords: Postpartum mental health, maternal separation, Dopaminergic system, Dorsal Hippocampus, Single-cell RNA sequencing

Disclosure: Nothing to disclose.

P742. Differential Central Monoamine Metabolite Concentrations in Primiparous and Multiparous Rhesus Macaques (Macaca mulatta) Across the Early Postpartum Period

Elizabeth Wood*, Natalia Gabrielle, Stephen G. Lindell, Melanie Schwandt, Christina Barr, Stephen Suomi, J. Dee Higley

Oregon Health and Science University, Tigard, Oregon, United States

Background: The perinatal period is a time of heightened neuropsychological plasticity and vulnerability, particularly for first-time mothers (primiparas). Postpartum depression (PPD) and anxiety (PPA) affect a substantial portion of mothers (approximately 10-20%), with studies suggesting that first-time mothers have a higher risk of PPD and PPA, compared to multiparous mothers. Despite this well-replicated finding, etiological markers which may underlie PPD and PPA are unclear. Among the neurobiological changes of the postpartum period are the monoamine systems, which play a crucial role in mood and behavior, as well as psychopathological states, such as depression and anxiety. Surprisingly, comparisons of the monoamines between primiparous and multiparous mothers are limited, generally using peripheral measures and lacking follow-up measurements. Longitudinal measurements of the central monoamine metabolites are difficult in postpartum human populations and comparisons of primiparous and multiparous humans are confounded by contextual factors such as socioeconomic status, race, ethnicity, and education. Nonhuman primate (NHP) models partially overcome these limitations due to their controlled environments, known life histories, and increased capacity to obtain central measures, such as cisternal cerebrospinal fluid (CSF) to measure neurotransmitter functioning. Like humans, nonhuman primates show evidence of postpartum affective dysfunction, and their central nervous systems (CNS) is organized similarly to the human CNS. Using a NHP model, this study investigates differences in central monoamine metabolite concentrations between primiparous and multiparous NHP mothers during the early postpartum period.

Methods: Sixty-nine rhesus macaque (Macaca mulatta) mothers were included in the study. All mothers were reared in identical conditions that approximate rhesus monkey normative social life to control for environmental variables. Cisternal CSF was repeatedly obtained from mothers following 105 live infant births (n = 46 births were primiparous and n = 60 births were multiparous). Samples were collected under highly controlled conditions at Days 7, 14, 21, 30, and 60 postpartum (approximating the neonatal and early infancy periods) and assayed for levels of 3-methoxy-4-hydroxyphenylglycol (MHPG), homovanillic acid (HVA), and 5-hydroxyindoleacetic acid (5-HIAA). Linear mixed models were used assess the effects of parity on MHPG, HVA, and 5-HIAA concentrations over the five time points. A random intercept was incorporated into the models to account for within-subject correlations across the repeated measures.

Results: Results showed a significant effect of parity on MHPG concentrations (Estimate = -15.75, SE = 7.36, p = 0.04, 95% CI[-30.41, -1.09]). Primiparas exhibited higher CSF MHPG levels (M = 110.17 □ 4.00), when compared to multiparas (M = 99.34 □ 4.36).There was also an effect of time on CSF HVA levels (Estimate = -196.21, SE = 40.16, p < 0.001, 95% CI [-275.98, -116.429]). Day 7 HVA concentrations were lower than all other timepoints (p < 0.008), while Day 60 HVA concentrations were higher than all other timepoints. (p < 0.003). There were no parity or time effects on cisternal CSF 5-HIAA.

Conclusions: Changes in neurobiological functioning of the postpartum period may be modulated, in part, by parity. Higher CSF MHPG concentrations among primiparas may reflect greater neurochemical adjustments consequential to the stress associated with first-time motherhood. Although the effects of stress alone cannot be ruled out, it is noteworthy that elevated platelet MHPG levels are also observed in human women experiencing the postpartum blues, indicating that MHPG concentration changes could be part of the underlying pathways of postpartum mood disorders. The absence of differences in CSF HVA and 5-HIAA levels suggests that dopamine and serotonin metabolism may be less affected by parity. A major strength of this study is the repeated capture of cisternal CSF monoamine metabolites, allowing for monoamine level comparisons between the very early postpartum periods and beyond. On average, the primiparous monoamine system does not appear to adapt to the novel experience of caring for an infant over the early postpartum period. Findings also indicate a potential pathway for heightened perinatal maternal neuroplasticity. Further research should explore the underlying mechanisms and long-term implications of these findings for maternal and offspring health. Planned future studies will assess parity differences in maternal behavior, particularly in mothers that show extremes in monoamine responsiveness, which may be an early indicator of risk for post-partum mood disorders. Pregnancy measures of the monoamines would also be of utility to examine whether parity differences in monoamine concentrations are extant pre-delivery, as well as to assess whether aberrant pregnancy monoamine levels predict postpartum affective behavioral dysfunction, particularly in primiparas. Such insights could inform the development of targeted interventions and preventive PPD and PPA measures. This study contributes to the understanding of the neurobiological impacts of the perinatal period, particularly the influence of parity, and provides a foundation for future research on the role of monoamine systems in early maternal behavior and mood regulation.

Keywords: perinatal mental health, monoamines, rhesus macaques

Disclosure: Nothing to disclose.

P743. Impact of Prenatal Cannabis Exposure on the Placental Transcriptome: A Proof-of-Concept Weighted Co-Expression Gene Network Analysis (WCGNA)

Laura Stroud*, Janet Joseph, Todd Everson, Carmen Marsit

Psychiatry and Human Behavior, Warren Alpert Medical School, Brown University, Providence, Rhode Island, United States

Background: Prenatal cannabis use is common and increasing, with rates as high as 1 in 4 in poor, young, underserved pregnant people. Expanding legalization of cannabis has led to increasing availability, use, perceptions of safety as well as increased potency. Although emerging data has highlighted the impact of prenatal cannabis exposure on offspring health and neurodevelopment, neurobiological mechanisms remain elusive. The placenta is a unique organ that regulates the intrauterine environment and plays a critical role in supporting fetal growth and neurodevelopment. Yet, little is known regarding the impact of prenatal cannabis exposure on placental biology. This proof-of-concept study aims to investigate the impact of prenatal cannabis exposure on placental gene expression.

Methods: We conducted transcriptome-wide bulk RNAseq analysis of 67 placenta specimens (n = 12 cannabis-exposed) from a larger study of prenatal tobacco exposure and infant development. Cannabis exposure was determined through synthesizing maternal report (Timeline Followback interview) and meconium biomarker data. We conducted preliminary analyses of expression of the impact of prenatal cannabis exposure on specific genes involved in endocannabinoid, opioid, and cortisol signaling and metabolism. We then performed preliminary weighted co-expression gene network analysis (WCGNA).

Results: Placental gene expression analyses highlighted dysregulation of endocannabinoid and stress-related genes in cannabis-exposed placentas, including downregulation of genes involved in endocannabinoid signaling (FAAH and FAAH2; ps < .05), and upregulation of stress-responsive genes (NR3C1, FKBP5; p < 0.1). Results from the WCGNA analysis revealed 25 co-expression modules; average eigengene values for two modules were greater among cannabis-exposed vs. unexposed placentas (p < 0.1); these modules were overrepresented for genes involved in cell-cycle regulation, cell to cell interaction, extracellular matrix organization, and G-protein coupled receptors, suggesting potential mechanisms underlying placental dysfunction and fetal development.

Conclusions: This proof-of-concept study highlights a significant neurobiological impact of prenatal cannabis exposure on placental gene expression. Results from the co-expression analyses revealed intriguing gene sets since alterations to the placental extracellular matrix can affect barrier functions and are associated with intrauterine growth restriction. Additionally, findings from the candidate gene analyses highlight the importance of stress-related and endocannabinoid signaling pathways in determining the impact of prenatal cannabis exposure. Pinpointing genomic pathways will allow us to identify biomarkers of fetal harm and develop novel screening tools and intervention targets to protect and treat exposed offspring. Overall, results emphasize the importance of understanding the molecular mechanisms underlying prenatal CB exposure to inform targeted interventions and policies aimed at mitigating its impact on maternal and child health.

Keywords: Prenatal Cannabis Exposure, Transcriptome, placenta

Disclosure: Nothing to disclose.

P744. Trajectories of Prenatal Depressive Symptoms in Relation to Infant Brain Connectivity

Venus Mahmoodi*, Marisa Spann

College of Physicians and Surgeons, Columbia University, New York, New York, United States

Background: Prenatal depression is a significant risk factor for postpartum depression, medical complications, and adverse neonatal brain development. While existing studies link prenatal depression with functional connectivity in the neonate’s amygdala and hippocampus, they often assess maternal depression at a single time point or as an average over pregnancy. This study aimed to characterize variations in maternal depressive symptoms across gestation and explore their association with infant amygdala and hippocampus functional connectivity.

Methods: As part of a larger study, healthy pregnant participants completed the Patient Health Questionnaire-9 (PHQ-9) at three time points during pregnancy, approximately once per trimester. Out of 109 recruited women, 82 completed at least two sessions. We calculated individual mean and standard deviation of the PHQ-9 scores. Imaging data were collected for twenty infants scanned between 0-6 months. Standard fMRI preprocessing was performed. For each infant, we constructed functional connectivity maps using a 90-node infant atlas and extracted connectivity values for the hippocampus and amygdala.

Results: Few mothers demonstrated a change in their depression score across pregnancy. However, 9% of the pregnant participants were characterized as depressed. Values for the infant’s left and right hippocampus and amygdala were combined, resulting in 89 hippocampal connections and 89 amygdala connections to associate with maternal PHQ-9 scores. Five hippocampal connections and one amygdala connection significantly correlated (p < 0.05) with the mean PHQ-9 scores. In contrast, one hippocampal connection and seven amygdala connections significantly correlated (p < 0.05) with the standard deviation of the PHQ-9 scores.

Conclusions: Our findings showed that prenatal depressive symptoms mapped to different aspects of the infant’s amygdala and hippocampus over time. Infant amygdala functional connectivity mapped better on the variation of PHQ-9 scores over pregnancy, while the neonate’s hippocampal functional connectivity mapped better on the average PHQ-9 scores. These results suggest that the infant amygdala’s connectivity is more related to how the prenatal depressive symptoms vary over time, likely due to changes in emotion regulation, and the infant hippocampus’s connectivity relates more to the overall level of prenatal depressive symptoms, which is likely linked to severity of symptoms. These nuanced findings of the infant’s amygdala and hippocampus connections can help in the development of targeted interventions for prenatal depression.

Keywords: prenatal depression, Infant fMRI, functional connectivity

Disclosure: Nothing to disclose.

P745. An Examination of SARS-CoV-2 Infection, Gene Expression in the Placenta, and Postpartum Depression With a Validation Analysis

Kendall Moore, Bushra Amreen, Juliana Camacho Castro, Tess Rooney, Frederieke Gigase, Rebecca Jessel, Emma Smith, Alan Adler, Bethany Dubois, Alexandra Mills, Elianna Kaplowitz, Leah Habersham, Noel Strong, Sophie Ohrn, Mara Graziani, Rachel Brody, Yula Ma, Whitney Lieb, Teresa Janevic, Veerle Bergink, Lotje De Witte, Anna-Sophie Rommel, Jia Chen, M. Mercedes Perez-Rodriguez*, Corina Lesseur

Icahn School of Medicine at Mount Sinai, New York, New York, United States

Background: The COVID-19 pandemic has had a significant negative impact on the mental health of pregnant women with increased risk of anxiety and depression during pregnancy and perinatal period. The immune system and inflammatory processes play critical roles in psychiatric health and the placenta plays an integral role in the maternal immune system. As a master regulator of both the in-utero environment and multi-organ pregnancy-related processes in the mother, the placenta orchestrates maternal-fetal interactions and neuroendocrine changes during pregnancy including modulation of the maternal immune system. Our research investigates the impact of gestational SARS-CoV-2 infection on immune and inflammation gene expression in the placenta and maternal postpartum depression (Edinburgh Postnatal Depression Scale, EPDS).

Methods: Study 1: In 141 pregnant participants from the Mount Sinai Hospital we used placenta RNA-sequencing data to investigate associations between 347 immune and inflammation-related genes, gestational SARS-CoV-2 infection, and maternal mental health. First, we conducted differential gene expression analyses of placental immune genes by gestational SARS-CoV-2 status using multivariable linear regression models adjusted for maternal age, gestational age at delivery, and infant sex. Next, we interrogated the relationship between differentially-expressed genes (unadjusted p-value < 0.05) from the previous analysis (predictors) and log transformed EPDS scores around 6-weeks postpartum (continuous outcome) with linear regression adjusted for gestational age, Distressed Communities Index scores (DCI; a measure of economic wellbeing of communities) based on zip code, and pre-pregnancy BMI. A nominal p-value of < 0.10 was used to establish genes of interest for further analysis.

Study 2: In 167 additional placental samples, we generated gene expression data (qPCR) for the top 5 genes of interest from Study 1. We conducted differential gene expression analysis of these 5 genes by gestational SARS-CoV-2 status and then interrogated the relationship between these genes and log transformed postpartum EPDS scores with linear regression models adjusted for the same covariates as in Study 1. Finally, we repeated these investigations in stratified analysis by fetal sex.

Results: In Study 1, SARS-CoV-2 infection was associated with downregulation of PTGER2 expression in the placenta, (logFC = -0.50, adjusted p-value = 0.022) and dysregulation of 37 additional genes (p-value < 0.05). Placental expression of four (CXCR6, BANK1, KCNJ2, and MET) of these 38 SARS-CoV-2 associated immune genes was also inversely associated with maternal postpartum EPDS scores (p-value < 0.10).

In Study 2 we observed no significant effects of SARS-CoV-2 infection on placental gene expression; nor effects of expression on postpartum EPDS scores in the whole sample. Yet, in fetal sex-stratified analyses, we observed a decrease in the expression of CXCR6 amongst male placentas from individuals who had SARS-CoV-2 infection during pregnancy (β= -0.87, p-value = 0.03). We also observed an inverse association between placental MET expression and postpartum EPDS scores in individuals with female placentas (β= -0.10, p-value = 0.067).

Conclusions: Overall, this study suggests that in placentas from individuals carrying male fetuses, gestational SARS-CoV-2 infection was associated with lower CXCR6 expression. Additionally, lower MET expression in the placentas from individuals carrying female fetuses may be associated with more severe postpartum depression symptoms at 6 weeks postpartum. However, this study does not fully validate either of these associations in two separate samples, and further research should be done to elucidate this connection.

Keywords: Novel coronavirus (SARS-CoV-2), postpartum depression, placenta, gene expression, pregnancy

Disclosure: Neurocrine Biosciences, Inc: Consultant (Self).

P746. Brain Functional Connectivity Predicts Hormone-Sensitive Mood Dysregulation in a Scaled-Down Experimental Model of the Perinatal Transition

Melissa Walsh*, Megan Hynd, Lis Bernhardt, Gabriel Dichter, David Rubinow, Crystal Schiller

Univ. of North Carolina at Chapel Hill, School of Medicine, Carrboro, North Carolina, United States

Background: The hormonal fluctuations of pregnancy and childbirth can significantly impact mood, especially in individuals with a history of postpartum depression. Our previous hypothesis-driven work suggests fMRI estimates of reward function do not distinguish hormone-sensitive (HS + ) vs. hormone-insensitive (HS − ) groups during hormone manipulation. This study used a data-driven, hypothesis-agnostic approach to characterize atypical brain functional connectivity (FC) changes associated with hormone sensitivity in a scaled-down experimental model of pregnancy. We further explored neurotransmitter systems correlated with atypical FC changes using normative atlases.

Methods: Participants were non-pregnant, euthymic women with a history of major depressive episodes with peripartum onset (n = 15) and healthy parous controls without psychiatric history (n = 15). Perinatal hormonal changes were simulated by inducing hypogonadism, administering estradiol and progesterone to first-trimester levels for 8 weeks, and then withdrawing both hormones. Individuals who experienced a 30% or greater increase in affective symptoms during hormone addback or withdrawal were classified as HS+ and all others were classified as HS-. T1 and resting-state fMRI scans (5 min) were acquired on a 3T Siemens Magnetom scanner at baseline and following hormone withdrawal. After excluding motion outliers (n = 5), the final sample comprised 24 participants (HS + : n = 9; HS-: n = 15). FC maps were generated using the CONN Toolbox and multivariate voxel pattern analysis (MVPA). An F-test was performed at each voxel to identify clusters showing significant group-by-time differences in multivariate FC (height threshold: p = 0.001; cluster threshold: p = 0.05, FWE-corrected). FC changes were characterized using post-hoc seed-to-voxel analysis (MVPA-derived seeds) and the Neurosynth decoder. To determine molecular correlates of group-by-time FC effects, unthresholded statistical maps were spatially correlated with nineteen normative PET receptor atlases using neuromaps (Hansen et al., 2022). For these exploratory analyses, significance was defined at p = 0.05, uncorrected.

Results: The voxel-wise MVPA analysis revealed significant group-by-time effects in clusters within 1) the right putamen/amygdala and 2) the right frontal pole. In HS+ individuals, right putamen FC atypically increased with the somatomotor cortex and decreased with parietal regions associated with working memory; right frontal pole FC atypically increased with the default mode network and decreased with inferior frontal regions associated with language processing. Atypical right putamen FC changes aligned spatially with the regional distribution of serotonin 5-HT2A (r = -.39, p = 0.01) and serotonin transporter (r = 0.48, p = 0.01), derived from PET receptor atlases. Atypical right frontal pole FC changes corresponded with GABA-A (r = 0.33, p = 0.02) and NMDA (r = 0.31, p = 0.02) receptor distributions.

Conclusions: In a scaled-down experimental model of pregnancy, hormone-sensitive mood dysregulation was associated with atypical FC changes in the right putamen/amygdala and right frontal pole. Atypical FC changes in the right putamen aligned with PET atlases for serotonin 5-HT2A and serotonin transporter; the latter was recently linked to premenstrual mood dysregulation (Sacher et al., 2023). Changes in the right frontal pole FC aligned with GABA-A and NMDA PET atlases; the former is the target of the drug brexanolone, approved for treatment of postpartum depression. These findings stand to be replicated in larger samples and molecular targets validated in PET studies.

Keywords: perinatal depression, experimental design, Resting State Functional Connectivity, neurotransmitters, brain connectome

Disclosure: Nothing to disclose.

P747. Evidence for the Biological Embedding of Stress During Pregnancy in Free Living Conditions

Cassandra Hendrix*, Clare A. McCormack, Katharina Schultebraucks, Moriah Thomason

New York University Langone Health, New York, New York, United States

Background: Research has repeatedly isolated stress exposure as a potent risk factor for psychiatric illness. Altered autonomic nervous system functioning is a key candidate mechanism by which stress is understood to become biologically embedded and linked to negative psychological outcomes. Reduced heart rate variability (HRV), a widely used metric of autonomic functioning, is observed in the context of chronic stress, is linked to poorer emotion regulation, and is a consistent feature of numerous psychiatric illnesses. A critical gap in understanding is whether embedding of stress occurs via similar biological mechanisms during pregnancy. Pregnancy is characterized by immense hormonal, neural, metabolic, and immune changes that may influence bodily sensitivity to stress to differing extents depending on gestational timing. At least one experimental study in humans found that autonomic reactivity to stress is higher at 24 compared to 36 gestational weeks, suggesting reduced stress sensitivity late in pregnancy. Yet whether reduced stress sensitivity is present outside of carefully controlled lab conditions, and when it may emerge during pregnancy, remains unknown. The present pilot study leveraged daily ratings of stress and directly measured maternal HRV each night across 5 months of pregnancy to test the hypotheses that (1) daily ratings of stress will correspond to lower heart rate variability, and (2) stress-HRV associations will be stronger earlier in gestation.

Methods: Analyses are based on 867 observations from 16 unique pregnant women. Pregnant women enrolled in a fully remote, longitudinal study that assessed stress and maternal physiology daily between 16 and 40 weeks of pregnancy. At a random time between 8am and 9pm each day, participants were prompted to complete 5 questions about their current stress, mood, and social context via SMS message. Perceived stress was assessed via a single question (“How stressed are you right now?”) rated on a 0 (not stressed) to 100 (extremely stressed) scale. Compliance with daily stress assessments was high, with a median of 77% of pushes receiving a valid response (interquartile range: 65-93%). Participants were also provided with an Oūra ring to wear continuously throughout the study period, which provided daily assessments of HRV. HRV was calculated as the root mean square of successive differences during nightly sleep. Gestational age at each assessment was calculated based on estimated due date. A linear mixed effects model was used to determine effect of daily stress ratings, gestational age, and their interaction on nightly maternal HRV. Participant was a random factor, and slope of HRV across gestation was allowed to have individually varying slopes. Restricted maximum likelihood was used to account for missing data. Maternal HRV data were log transformed to correct for positive skew. Posthoc simulation estimated this analysis was powered to detect a minimum effect size of f2 = 0.013 with 80% power. Data collection is ongoing so the present analyses should be considered preliminary.

Results: There was no linear main effect of gestational age (b = -0.017, 95%CI[-0.181, 0.149], se=0.082, p = 0.842) or daily stress ratings (b = 0.012, 95%CI[-0.039, 0.065], se=0.026, p = 0.640) on nightly maternal HRV. Instead, the influence of stress on nightly HRV was moderated by gestational age at assessment (b = 0.052, 95%CI[0.001, 0.104], se=0.026, p = 0.048). Simple slopes analysis revealed that higher maternal-rated stress predicted lower HRV that night, but only during gestational weeks 16 to 27 (p’s < 0.05, FDR-corrected). The association between stress and HRV was not significant after 27 gestational weeks (p’s > .0.214). Mean stress levels were equivalent before and after 27 weeks (t = -1.058, se=0.055, p = 0.293).

Conclusions: These preliminary analyses provide novel evidence that maternal HRV is sensitive to psychosocial stress in free living conditions through the second trimester of pregnancy. Consistent with experimental evidence, autonomic sensitivity to stress was reduced later in gestation, and the present analyses pinpoint 27 gestational weeks as a potential switch point after which autonomic sensitivity may change. Additional research in larger samples is required to test whether changes in stress sensitivity correspond to differences in psychiatric risk across gestation. The present study is among the first to demonstrate feasibility of long-term, daily monitoring of maternal physiology and psychosocial experiences across pregnancy, which holds potential to provide unparalleled insight into stress-related programming of psychiatric risk.

Keywords: pregnancy, autonomic nervous system, wearables, Perceived stress, developmental origins

Disclosure: Nothing to disclose.

P748. Association Between Menopause and Fear Learning Among Trauma-Exposed Women

Jennifer Stevens*, Linzie Taylor, Megan Huibregtse, Amanda Arnold, Amy Murphy, Donia Hassan, Leigha Lee, Esther Jung, Rebecca Hinrichs, Britton Chahine, Abigail Powers, Alicia Smith, Vasiliki Michopoulos

Emory University School of Medicine, Atlanta, Georgia, United States

Background: Black women experience a disparate burden of exposure to traumatic experiences over the lifespan, which is associated with increased risk for trauma-related hyperarousal (TRH) and dysregulated fear responses that are dependent upon the ventromedial prefrontal cortex (vmPFC) and its regulation of amygdala responses to threat. Natural hormonal variation across the female lifespan is linked to increased risk for depressive and anxiety symptoms. However, factors that increase vulnerability to the psychological consequences of hormonal variation remain unclear. This is especially true for women during the menopausal transition, who have received minimal focus in trauma and mental health research to date. Here we investigate the TRH and changes in the engagement of the vmPFC and amygdala during fear learning, over the perimenopausal transition, among Black women with experiences of trauma.

Methods: N = 40 Black and African American women ages 40-55 completed the study procedures. Inclusion criteria required participants to have an intact uterus and at least 1 ovary, not be taking hormonal or psychiatric medications, and to report experiencing at least 1 potentially traumatic event. Participants completed self-report assessments of PTSD symptoms (PCL-5), depression symptoms (PHQ-9). Suicide risk assessment (CSSRS) and PTSD diagnosis (CAPS-5) were conducted via structured interviews. Participants endorsed a wide range of PTSD symptoms (PCL-5 range 0-49; n = 6 met for PTSD on CAPS-5). Baseline menopausal status was assessed using STRAW + 10 criteria, and tracked prospectively using weekly mobile menstrual cycle surveys over 18 months of follow-up. Lab visits with fMRI neuroimaging and psychophysiology were collected at baseline and at 6 and 18 months. Participants completed a visual fear conditioning paradigm during fMRI, in which categories of visual stimuli (CS cues; animals and tools) were paired with either a loud aversive sound (the US) or never paired with sound. Initial analysis of the baseline data for fear conditioning is presented here, with regions of interest (ROIs) for the vmPFC, central amygdala, and early visual cortex (V1/V2).

Results: Women in peri- and early post-menopause, relative to premenopausal women, reported greater PTSD symptoms, F1,35 = 4.20, p = 0.05 (Pre: M + -SE = 12.00 + -2.44, Peri/Post: 21.1 + -3.62) and greater TRH, p = 0.03. Moreover, menopausal symptom severity was strongly associated with PTSD symptoms, F = 21.55, p = 0.00005, and TRH, F = 28.20, p = 0.000007, after controlling for BMI and trauma load. In the fear conditioning task, the comparison of CS + > CS- produced significant activation in the right central amygdala, dACC, bilateral insula, ventral visual cortex, and auditory cortex, pFWE < 0.05. Peri/postmenopausal women showed greater engagement of the vmPFC ROI during fear learning (CS + > CS-), compared to premenopausal women, t = 2.30, p = 0.02. VmPFC engagement during fear learning was not associated with PTSD symptoms or TRH, p’s > 0.05, but did predict lower suicide risk, r = -0.43, p = 0.008. Additionally, peri/postmenopausal women showed less activation of the early visual cortex ROI, relative to premenopausal women, t = 2.47, p = 0.02, but the engagement of this region was not related to PTSD symptoms, TRH, or suicide risk.

Conclusions: These initial analyses suggest that women in peri-menopause and early menopause experience elevated TRH and overall PTSD symptoms relative to pre-menopausal women, consistent with prior cross-sectional work in larger samples. The greater engagement of the vmPFC during fear learning, appeared to be an adaptive response during peri/post-menopause, associated with lower suicide risk scores. These findings provide novel insight on the effects of the menopausal transition on fear learning, as well as relevance for trauma-related symptoms, neither of which have been explored in humans. However, further work is needed in order to identify neural mediators of the perimenopausal elevation in TRH, including in a larger sample size and longitudinal change over time.

Keywords: PTSD, Fear conditioning, fMRI

Disclosure: Nothing to disclose.

P749. Pharmacologic and Non-Pharmacologic Treatment for Alcohol Use Disorder in Pregnancy and Postpartum: An Analysis of Unmet Needs Using Multi-State Administrative Claims Data

Caitlin Martin*, Bridget Galati, Joanna Buss, Jennifer Bello-Kottenstette, Mishka Terplan, Hendree Jones, Richard Grucza, Kevin Xu

Virginia Commonwealth University, Richmond, Virginia, United States

Background: Alcohol use disorder (AUD) is the most prevalent substance use disorder among reproductive age people, including during the pregnancy and postpartum periods. Prenatal AUD is associated with increased odds of severe maternal morbidity, and alcohol is the leading cause of preventable disability in the United States. AUD treatments, including medications for alcohol use disorder (MAUD) and psychosocial interventions, are effective across populations studied, yet many people with AUD do not receive treatment. In pregnancy, AUD treatment effectiveness and safety studies are lacking, especially for MAUD. To date, there are no published real-world data on rates of AUD treatment receipt in birthing people. We report on the prevalence of receipt of MAUD and psychosocial treatments among a multi-state AUD sample through the perinatal period (preconception, pregnancy, year postpartum).

Methods: We used the Merative MarketScan databases (2015-2019) to conduct a retrospective cohort study of pregnant individuals with AUD ( > 1-2 claims with AUD diagnosis in the 1-year preceding conception) who had received either MAUD or AUD-related psychotherapy within the year before pregnancy. The primary outcome, filled MAUD prescriptions (naltrexone PO = daily tablets; naltrexone ER = monthly injection; acamprosate), were identified via pharmacy claims. We also evaluated claims for psychosocial services associated with at least one AUD diagnosis code based on ICD codes for counseling, psychotherapy, case management, and behavioral health services. We computed rates of MAUD and psychotherapy receipt in the perinatal period, stratifying by five 12-week observation windows (12 weeks preceding conception; first trimester; second trimester; third trimester; 12 weeks after delivery [postpartum]).

Results: Our sample included 1,840 pregnant people with AUD, encompassing 1,874 deliveries. With a median age of 27 years, 60.6% identified as non-Hispanic White, 30.6% non-Hispanic Black, 2.4% Hispanic, and 2.1% ‘other’ race/ethnicity. The majority of our sample had diagnoses for co-morbid mood (65.9%) or anxiety disorders (62.7%).

In the baseline 12 weeks preceding conception, 15.3% filled MAUD prescriptions (8.5% naltrexone PO, 7.8% naltrexone ER, 1.0% acamprosate). During the first trimester, rates of MAUD prescriptions decreased to 9.9% (4.2% naltrexone PO, 5.7% naltrexone ER, 0.5% acamprosate). In the second and third trimesters, only 1.3% and 0.4% of our sample filled a MAUD prescription, respectively. After delivery, MAUD prescriptions remained low at 2.0%.

For AUD psychosocial treatments, 47.0% had > 1 claim at baseline; by the first trimester, psychosocial claims decreased to 25.2%. By the third trimester, psychosocial treatment claims were one-fourth (12.2%) of baseline rates. Postpartum rates of AUD psychosocial treatment remained low at 10.9%.

Conclusions: In this multi-state sample of birthing people with AUD, using administrative claims data, we found prescribing rates for MAUD and claims for AUD-related psychosocial treatment to be low pre-pregnancy. AUD treatment rates decreased further during pregnancy and remained very low after delivery, with approximately 2 in 100 of our sample with postpartum claims for MAUD and 1 in 10 with psychosocial treatment receipt. The perinatal period typically offers increased opportunity to provide evidence-based treatments for chronic health conditions. Our very low rates of AUD treatment during pregnancy and postpartum warrant further investigation.

Keywords: pregnancy/postpartum, Alcohol Use Disorder - Treatment, acamprosate, Naltrexone

Disclosure: Nothing to disclose.

P750. Peripartum Allopregnanolone Blood Concentrations and Depressive Symptoms: A Systematic Review and Individual Participant Data Meta-Analysis

Georgios Schoretsanitis, Lauren M. Osborne, Inger Sundstrom-Poromaa, Elizabeth S. Wenzel, Jennifer L. Payne, Corrado Barbui, Chiara Gastaldon, Kristina Deligiannidis*

Psychiatry Research, Zucker Hillside Hospital; Feinstein Institute for Medical Research, Glen Oaks, New York, United States

Background: Neuroactive steroids including allopregnanolone (ALLO) are implicated in the pathophysiology of peripartum depressive symptoms (PDS) yet a consistent association between neuroactive steroid blood levels and PDS has been elusive (1-6) despite robust data on their pathophysiological role (7, 8). Our aim was to systematically review and subsequently meta-analyze observational data on peripartum blood ALLO concentrations in females with and without PDS and assess moderators of ALLO patterns.

Methods: We performed a systematic review searching PubMed/Embase/PsychInfo/Cinhail through 08/2023 (updated in 07/2024) and conducted a random-effects meta-analysis of studies comparing ALLO blood concentrations in females with versus without PDS at timepoints during the 2nd and 3rd trimester and the postpartum period, calculating standardized mean differences (SMDs) and 95% confidence intervals (CIs). Meta-regression and subgroup analyses included age, current antidepressant treatment, and analytical method. Study quality was assessed using the Newcastle-Ottawa-scale. The study protocol was registered on PROSPERO (registration number CRD42022354495). We retrieved 13 studies with 2,509 women (n = 849 with PDS). Females with versus without PDS during pregnancy and up to one year postpartum were included. Participants included antidepressant-naïve and antidepressant-free females as well as those taking psychotropics. We used a random-effects model for our primary outcome, considering the large heterogeneity related to study cohorts, analytical methods, and the inherently essential related variability. Results for each timepoint of ALLO assessments were summarized using the SMD and 95% CI presented in forest plots; further, we assessed the predictive role of gestational ALLO concentrations’ differences between individuals with versus without PDS at postpartum follow-ups. The heterogeneity variance parameter (τ2) was calculated using the DerSimonian-Laird estimator (9). For longitudinal studies with ALLO assessments at multiple timepoints, cohorts from the same study were considered separately for different meta-analyses in a cross-sectional manner, based on the eight predefined timepoints. We also calculated the I-square (I2) statistic as a measure of the proportion of variability that can be attributed to heterogeneity (10). Thereafter, the effect of maternal age was assessed in a meta-regression analysis (11). As appropriate, subgroup or sensitivity analyses examined effects of current antidepressant use and the use of different ALLO analytical methods. We also performed a sensitivity analysis excluding poor-quality studies. Meta-regression and subgroup analyses were performed when data for a minimum of three studies were available. Analyses were performed using the meta package in R (12).

Results: We retrieved 13 studies with 2,509 women (n = 849 with PDS). ALLO concentrations did not differ between females with versus without PDS at any timepoint (p > 0.05). Pregnancy ALLO concentrations did not differ for females with versus without PDS at postpartum follow-up (p > 0.05). Sensitivity analyses indicated, that in women without current antidepressant use, higher ALLO concentrations in women with versus without PDS at gestational weeks 21-24 and 25-28 (SMD = 1.07, 95%CI = 0.04, 2.11 and SMD = 0.92, 95%CI = 0.26, 1.59 respectively). We observed effects for maternal age, i.e. larger ALLO concentration differences in samples of younger ages, at gestational weeks 29-33 (estimated co-efficient -0.42, 95%CI = -0.72, -0.11, p = 0.008). Moreover, there were differences between studies using mass-spectrometry combined with chromatography versus immunoassays at gestational weeks 25-28 (p = 0.01). Study quality was rated as poor, good, and fair for two, one and ten studies respectively.

Conclusions: To our knowledge, this is the first meta-analysis to assess ALLO blood concentrations in females with versus without PDS. Despite the lack of differences in our main analysis, in our subgroup and sensitivity analyses we reported higher ALLO concentrations in women with vs. without PDS at gestational weeks 21-24 and 25-28 when only including studies with individuals not receiving antidepressants. Moreover, at gestational weeks 25-28 studies using chromatography/mass spectrometry reported higher ALLO concentrations in females with versus without PDS, whereas immunoassays did not. The use of heterogenous peripartum time points, study cohorts, depression symptom measures and analytical methods has hampered progress in elucidating neuroactive steroid signaling linked to PDS.

Keywords: allopregnanolone, perinatal, depression, meta-analysis

Disclosure: Sage Therapeutics: Consultant (Self). Sage Therapeutics: Contracted Research (Self). Brii Biosciences: Consultant (Self). Gerbera Therapeutics: Consultant (Self). Gerbera Therapeutics: Contracted Research (Self). GH Research: Consultant (Self). Neuroscience Software: Consultant (Self). NextSense: Consultant (Self). Woebot Health: Contracted Research (Self). Premier Healthcare: Contracted Research (Self). Biogen: Consultant (Self).

P751. Investigating Opioid and Nicotine Effects on Perineuronal Net Structures and Activity-Regulated Cytoskeleton Associated Protein (ARC/ARG3.1) in Wistar Rats

Benjamin Schwartz, Mariana Dejeux, Sarah Jewanee, Blake Reeves, Andriy Fomin, Aliyah Ogden, Jacques Nguyen*

Baylor University, Waco, Texas, United States

Background: The nonmedical use of opioids is a significant health problem, and co-occurring use of opioids and nicotine may exacerbate the risk of developing polysubstance use disorder/dependence. Specialized extracellular matrix proteins, perineuronal nets (PNNs), and the activity-regulated cytoskeleton associated protein (ARC/ARG3.1) are implicated in metaplasticity and drug-related conditioning. We hypothesized that subjects repeatedly administered oxycodone, fentanyl, and/or nicotine would exhibit tolerance and dependence-like effects and alter expression of PNNs and ARC, further enhancing addiction vulnerability.

Methods: Male Wistar rats ( ~ 10 weeks old) were randomly assigned to groups that received acute or repeated injections (twice daily for 7 days) of oxycodone (2-4 mg/kg, i.p.), fentanyl (4-125 ug/kg, s.c.), nicotine (0.4-2mg/kg, s.c.) or 0.9% saline vehicle. Thermal antinociception (i.e. tail withdrawal test) was used to validate acute drug intoxication, tolerance, and spontaneous or precipitated withdrawal, following a naloxone (1 mg/kg, i.p.) injection and after an acute or extended (1 day vs. 4 week) abstinence interval. A separate cohort of rats were implanted with jugular vein catheters and trained to self-administer intravenous fentanyl (0.625-10 ug/kg/inf) during 8-hour sessions. Following acquisition rats were tested under a progressive ratio (PR) reinforcement schedule. Serial collection of whole-brain coronal slices and immunohistochemistry were used to identify Wisteria Floribunda Agglutinin (WFA)-positive perineuronal net structures and ARC-positive brain regions.

Results: Oxycodone, fentanyl, and nicotine injections produced significant antinociception (increased latency in the tail withdrawal test), as well as tolerance following repeated injection (p < 0.05). Increases in perineuronal net count and intensity and ARC density were observed in the orbitofrontal cortex and CA1 hippocampus following abstinence to both opioid and nicotine administration. Rats trained to self-administer fentanyl escalated responding across 21 sessions and showed increased motivation for 2.5 ug/kg/inf dose (higher breakpoint) under PR reinforcement schedule. Escalation of fentanyl self-administration increased PNN expression (quantity and intensity) in the orbitofrontal cortex.

Conclusions: Overall, these data confirm that opioid and nicotine administration and abstinence in rats may alter extracellular matrix structures and immediate early genes integral to neuroplasticity-related learning and memory. This may indicate a key mechanism underlying polydrug-use vulnerability.

Keywords: opioid, extended-access self-administration, Perineuronal nets, Activity-regulated cytoskeletal associated protein (Arc)

Disclosure: Nothing to disclose.

P752. Fronto-Cortical Gene-Regulatory and Cell-Cell Interaction Changes in Mice Susceptible to Motivational Deficits of Chronic Psychosocial Stress

Puja Parekh*, Devin Rocks, Margaux Kenwood, Jacob Roshgadol, Conor Liston

The University of Texas at Dallas, Richardson, Texas, United States

Background: Chronic stress is a significant risk factor for depression in patients and elicits anhedonia-like behavior and amotivation in animal models. Understanding the circuit and molecular mechanisms that contribute to these states is integral to the development of more efficacious therapeutic strategies. Importantly, individual variation in stress response has been observed in several models and can shed light on mechanisms promoting resilience. We hypothesized that vulnerability to stress-induced motivational deficits correlates with genetic and molecular changes within the anterior cingulate cortex (ACC), a region implicated in depression-like behaviors and antidepressant response.

Methods: We developed and employed a head-restrained Pavlovian effortful reward-seeking task to measure motivated behavior in adult male and female C57BL/6J mice (8-12 weeks) and investigated the effect of chronic nondiscriminatory social defeat stress (CNSDS) on sensitivity to reward availability and physical effort requirement. We previously validated this task in a large number of animals in which we also observed individual variation in the response to stress (n = 56). To investigate potential contributing mechanisms to the differences in post-stress reward-seeking behavior, we classified a smaller cohort of animals as ‘control’, ‘resilient’ and ‘susceptible’ and performed 10X Visium slide-based spatial transcriptome sequencing on micro-dissected ACC tissue (n = 9 mice; 3/condition; 2 replicates). Clusters were identified corresponding to major cell types within the ACC. Spot-level data were deconvolved for single-cell analysis of differentially expressed genes (DEGs) across experimental groups. Gene ontology (GO) analysis revealed genes enriched for a number of pathways and biological processes associated with synaptic function and neurotransmission. Finally, the spatial component allowed us to evaluate communication between cell groups expressing a given ligand and its receptor. We compared the interaction strength between specific sender-receiver cell type pairs across experimental conditions.

Results: Mice (n = 9) learned to discriminate cues in the reward task and showed sensitivity to rewarded and effortful trials as indexed by averaged baseline-subtracted anticipatory and consummatory lick responses (F3,32 = 6.038, p = 0.0022; F3,32 = 22.06; p < 0.0001). Compared with control animals, stressed mice had a significant reduction in anticipatory responding for reward (2-way ANOVA; effect of treatment; F1,20 = 17.35; p = 0.0009). Resilience and susceptibility were determined by the change in effortful reward seeking behavior from baseline. Transcriptomic analyses show that expected cell types are identified in ACC sections (7 unique clusters) and the greatest number of differentially expressed genes are found in layer 2/3, 5 and 6 excitatory neurons and inhibitory interneurons in ACC tissue from susceptible versus resilient animals. Furthermore, transcriptional states differ significantly between control, resilient and susceptible groups across cell types. The number of cells of each type was relatively consistent across groups, replicates and slides. In our analysis of cell-cell communication, we found that interaction strength between inhibitory and excitatory neurons was significantly greater in samples from resilient animals compared with control and susceptible (p < 0.5). Significantly modulated ligand-receptor cell pairs in the susceptible group include corticotropin releasing hormone (CRH)-expressing inhibitory interneurons and proximal Crhr1 + L5 excitatory neurons as well as nociceptin-expressing inhibitory neurons and proximal Oprl1 + L5 neurons. Additionally, we found increased astrocyte signaling to L2/3 excitatory neurons in the susceptible group. These signaling pathways may be particularly relevant for stress response and regulation of behavior as well as synapse development and function.

Conclusions: We leveraged spatial sequencing technology to investigate gene-regulatory and cell-cell communication related changes in the ACC in animals classified as resilient or susceptible to stress-induced effortful reward seeking deficits. Ongoing and future studies are aimed at understanding the functional consequences of diminished interaction strength in key signaling pathways as well as exploring transcriptomic changes in distinct projection-targeted neural populations within the ACC and striatal regions as a result of chronic stress exposure.

Keywords: Chronic social stress, Reward, motivation, and anhedonia, individual variability, gene expression, anterior prefrontal cortex

Disclosure: Nothing to disclose.

P753. NR4A2 in Cholinergic Interneurons of Nucleus Accumbens Regulates Cocaine-Seeking Behavior and Synaptic Plasticity in Male and Female Mice

Jessica Childs*, Tracy Fetterly, Gerardo Sandoval, Eni Kramar, Jacob Rounds, Vanessa Alizo Vera, Natalie Thatch, Sham Labanieh, Bailee Valenzuela, Dina Matheos, Javier Diaz Alonso, Marcelo Wood

University of California, Irvine, Irvine, California, United States

Background: The unnaturally strong and persistent cue/reward associations that are made during drug use are thought to underly the lifelong risk of relapse associated with substance use disorder (SUD). There is evidence for epigenetic regulation of drug memory, including effects of the HDAC3-target epigenetic effector gene nuclear orphan receptor subfamily4 groupA member2 (Nr4a2), which the Wood Lab has studied for it’s role in the medial habenula’s regulation of relapse to cocaine-seeking. NR4A2 is a transcription factor important for the development and maintenance of dopamine neurons, and cocaine exposure alters Nr4a2 expression in the nucleus accumbens (NAc). The D1 and D2 medium spiny neurons (MSNs) of the NAc have been extensively linked NAc activity and subsequent cocaine-associated behaviors. NAc MSNs are regulated by a sparse population of cholinergic interneurons (ChIns) within the NAc that express Nr4a2. We hypothesized that reducing NR4A2 function by expressing it’s dominant negative (NURR2C) in NAc ChIns would enhance cocaine-associated memory formation and related cocaine-seeking in mice, suggesting NR4A2 normally limits drug-induced behaviors.

Methods: These studies were approved by the UCI Institutional Animal Care and Use Committee and were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The transcriptionally inactive dominant negative of NR4A2, NURR2C, or GFP was expressed in NAc ChIns using cre-dependent AAV and male and female ChAT-Cre mice. Cocaine conditioned place preference (CPP) and intravenous self-administration (IVSA) were used to study cocaine-associated memory formation and behavior. For CPP experiments, mice in both groups received either a subthreshold dose of cocaine (7mg/kg/I.P.) or saline in one of two distinct sides of a three-compartment apparatus. After four days of conditioning, mice were assessed for their preference for the cocaine-paired side. In IVSA experiments, mice learned to press a cocaine-rewarded lever (8.5µl/kg/infusion) in operant conditioning chambers. Mice in both groups were allowed to self-administer cocaine during one-hour sessions for 12 days.

Electrophysiological approaches were used to study to overall activity of the NAc (local field potential recordings, LFP), the function of MSNs (whole-cell patch-clamp recordings), and the function of ChIns (loose cell attached and whole-cell current-clamp recordings).

Sex as a Biological Variable: Sex and estrous cycle are both known to affect drug-associated behaviors. Therefore, where possible, the data described in this abstract come from male and female mice and ongoing work includes both males and females. We refer to the NIH’s guidelines on sex as biological variable and monitor the effects of the estrous cycle on our findings.

Results: Expressing NURR2C in NAc ChIns increased cocaine CPP for a subthreshold dose of cocaine. After four days of conditioning, NURR2C mice had increased preference for the cocaine paired side compared GFP controls (males: t(48) = 2.061, p = 0.048, females: t(16) = 2.906, p = 0.010). There were no differences in locomotor behavior. NURR2C expression in ChIns also increased volitional cocaine taking. Over the 12 days of IVSA cocaine self-administration, NURR2C mice had increased active lever pressing compared to GFP controls (F(1, 20) = 5.086, p = 0.0355). To study the effect of NURR2C in ChIns on NAc activity, LFP recordings were made before and after a high frequency stimulation was used to induce LTP. NURR2C was found to increase NAc long-term potentiation (LTP) in naïve males (t(17) = 3.6, p = 0.0021) and females (t(13) = 3.6, p = 0.003). This study was repeated in males sacrificed 1-hour after the end of cocaine CPP experiments, and here NURR2C mice also exhibited enhanced LTP (males, t(12) = 4.9, p = 0.0003. Consistently, in whole-cell patch-clamp recordings from MSNs in naïve male mice, ChIn NURR2C expression was found to increase AMPA/NMDA ratios (t(24) = 3.42, p = 0.0022).

Conclusions: These findings contribute to the growing body of literature demonstrating the role of NAc ChIns in drug-associated behaviors, and identify ChIn NR4A2 function as a key negative regulator of overall NAc activity and related cocaine-seeking behavior. NR4A2 is a nuclear orphan receptor and thus highly druggable, and these findings further suggest that Nr4a2 is a desirable target for pharmacological therapeutic approaches for SUD treatment.

Keywords: Nucelus accumbens, cholinergic interneuron, cocaine self-administration, slice electrophysiology

Disclosure: Nothing to disclose.

P754. Neural Dynamics in the Ventromedial Prefrontal Cortex Underlying Approach-Avoidance Conflict Resolution in Mice

Rodolfo Flores Garcia*, Yuan Zhao, Nathanyal Ross, Miguel Arenivar, Monique Awanyai, Huikun Wang, Francisco Pereira, Bruno Averbeck, Hugo Tejeda

NIH/NIMH, El Paso, Texas, United States

Background: Approach-avoidance conflicts, where actions promising potential rewards also carry risks of adverse consequences, are common in everyday life. In psychiatric disorders such as depression and addiction, patients display deficits in resolving approach-avoidance conflict, leading to maladaptive behaviors. Adaptive approach-avoidance conflict resolution involves the interplay between the brain’s reward and threat systems. The ventromedial prefrontal cortex (vmPFC) is of particular interest, given its role in suppressing fear states and attributing value to rewards. In this study, we investigated the role of vmPFC neurons during the platform-mediated avoidance (PMA) task in mice to expand our understanding of how this brain region mediates adaptive behaviors during motivational conflict.

Methods: Eight-week-old C57BL/6J mice were surgically prepared with a viral infusion of the genetically encoded calcium indicator (AAV-Syn-GCaMP8f) targeting the vmPFC, followed by GRIN lens implantation above the infusion site. Following recovery, mice underwent six daily sessions to learn how to nose-poke for Ensure reward. During this training phase, mice were presented with a 30-second light cue, during which nose-pokes were rewarded on a fixed ratio 2 schedule. Each session consisted of 30 trials, with an inter-trial interval averaging 60 seconds. Subsequently, mice were trained in conflict sessions over 10 daily sessions using the PMA task. During these sessions, a tone co-terminating with a shock was introduced. The tone was presented with full, partial, or no overlap with the light cue, creating a set of 5 distinct trial types: ‘Light only’ trials, ‘Tone only’ trials, ‘Co-presentation’ trials (simultaneous tone and light), ‘Light then Tone’ trials, and ‘Tone then Light’ trials. Mice were exposed to multiple instances of each trial type during each session. Calcium activity in the vmPFC was recorded on every test session using in vivo, single-cell calcium imaging. We employed reinforcement learning to model the mice’s learning and a multiple linear regression encoding model to examine the encoding of behavior and task-related variables in vmPFC neurons.

Results: The results revealed that mice balanced approach and avoidance behaviors to resolve conflict by adapting their actions in response to changing reward and threat contingencies. Additionally, a reinforcement learning model successfully captured the behavior of mice during both reward and conflict learning phases. The neural encoding results indicated that, across the training sessions, more neurons in the vmPFC increasingly encoded task- and behavior-related variables. Notably, as mice trained in the PMA task, vmPFC neurons exhibited changing representations of motivationally relevant cues. During reward training, more vmPFC neurons encoded the light cue, and their activity in response to this cue increased in a training-dependent manner. However, during conflict training, the representation of the reward cue in vmPFC neurons remained stable, while the representation of conflict and threat-predicting cues decreased over training days. In addition to encoding cues, vmPFC neurons also encoded decision-related variables and action values during both reward and conflict training. This suggests that vmPFC neurons are involved not only in processing motivationally relevant cues but also in integrating decision-related information and action value representations to guide adaptive behavior.

Conclusions: Together, these results highlight the functional diversity of vmPFC neurons, suggesting a role for these neurons in integrating information about learned associations, decision-related variables, and action values for adaptive decision-making. Our findings contribute to the understanding of motivational and emotional neural circuits, which is relevant to psychiatric disorders characterized by maladaptive conflict resolution, such as addiction and depression.

Keywords: Approach-Avoidance Conflict, Medial Prefrontal Cortex, In Vivo Calcium Imaging, Threat, Reward

Disclosure: Nothing to disclose.

P755. Investigating the Impact of Predator Odor Exposure on Natural Rewards: Uncovering the Role of L1

Amelia Cuarenta*, Margie Stricklin, Kevin Mesape, Moushmi Bhat, Debra Bangasser

Georgia State University, Atlanta, Georgia, United States

Background: Adversity experienced early in life can impact reward-related behaviors and increase the risk for the development of psychiatric disorders later in life. Much of the current research has focused on drugs of abuse, but the effect of early life adversity on natural rewards, including food and social rewards, remains understudied. We use a predator odor exposure (POE) model in which neonatal rat pups are exposed to a predator odor for 5 minutes/day for the first three days of life. This fear of harm model has previously shown an increase in anxiety-like behavior and a reduction in juvenile social play behavior. This reduction in social play is accompanied by an increase in copy number in the juvenile amygdala of the retrotransposon, L1. This increase in L1 was inversely correlated with juvenile social play behavior. L1 is an autonomous mobile element that can copy and paste itself elsewhere within the genome. L1 has also been associated with several psychiatric disorders including major depression and substance use disorder. Therefore, it is crucial to understand the role L1 may play in social behaviors relevant to some psychiatric disorders. We are now extending our research to investigate whether changes in social behavior extend to adulthood. We are using social self-administration in adulthood to assess whether POE alters social behavior beyond the adolescent period. We are also investigating whether L1 in the amygdala is a regulator of social behavior via L1 transient knockdown in the amygdala.

Methods: In the POE model, rat pups are exposed to cat, rat, and ferret odor for 5 minutes/day on postnatal days (PND) 1, 2, and 3, respectively. Control pups are handled in the exact same manner but exposed to clean bedding. Prior work from our lab has found that this model increases anxiety-like behavior and reduces social play during the juvenile stage of development.

Experiment 1: We conducted oral sucrose self-administration and social self-administration using MedPC Operant Chambers to assess how POE affects the motivation for sucrose and social interactions. Male and female adult rats (n = 6-10/group) were placed in operant chambers and permitted to lever press on a fixed ratio 1 (FR1) schedule for sucrose for four days, followed by 2 days of fixed ratio 2 (FR2), fixed ratio 4 (FR4), and fixed ratio 8 (FR8) each. They were then tested on progressive ratio testing to assess their motivation for sucrose. All rats had one day off before following the same schedule for social self-administration. All sessions began 2-3 hours after the start of the animals’ dark cycle and lasted for 1 hour.

Experiment 2: In vivo knockdown of L1 using antisense oligonucleotides (ASO) will be used to transiently knockdown L1 in the amygdala. Stereotaxic surgery will be used for site-specific injections into the neonatal rat amygdala (coordinate □ 1 mm lateral, 2 mm caudal, and 5.5 mm ventral from bregma. After cold anesthetization, rats will be bilaterally infused in the amygdala with 1 µl (100 nmol) of either L1 FANA ASO or scrambled control ASO. Confirmation of knockdown will be validated using qPCR. For behavioral analysis, 30 rats (15 L1 FANA ASO, 15 scrambled ASO) will undergo juvenile social play behavior testing in the home cage on PND 25 – 30.

Results: Our preliminary behavioral analysis demonstrates that male rats exposed to POE during the neonatal period show a reduction in the motivation for social reward relative to controls [t(10) = 3.481, p = 0.006, η2 = 0.55]. Our POE-exposed male rats also show a reduction in active lever presses on the social progressive ratio task [t(10) = 2.641, p = 0.025, η2 = 0.41]. This behavioral change is reward specific as animals exposed to POE show no difference from controls in their motivation for sucrose rewards [t(17) = 0.486, p = 0.633, η2 = 0.02]. Ongoing studies are aimed at linking this change in social motivation to L1 copy number in the amygdala. We anticipate data revealing a role of L1 in mediating these effects.

Conclusions: Prior work showed POE reduced juvenile social play. Here we found that POE reduced motivation for a social reinforcer, but not a sucrose reinforcer. This suggests that POE induces selective impairments in social behavior, underscoring the utility of this model to assess the link between early adversity and social deficits that characterize many psychiatric disorders. Our ongoing study is investigating L1 as a potential mechanism that regulates social behavior. If our results are consistent with our hypothesis this would be an interesting finding as there is very little research linking L1 genomic changes with social behavior. Altogether we are developing a nuanced understanding of how early life adversity alters the genome to affect motivation for reward.

Keywords: Social reward, Early life stress, L1 retrotransposons

Disclosure: Nothing to disclose.

P756. Shared Immune-Related Pathology in Response to Chronic Stress or Alcohol Exposure

Lyonna Parise*, Flurin Cathomas, Kenny Chan, Romain Durand-de Cuttoli, Carole Morel, Antonio Aubry, Arman Tavallaei, Johana Alvarez, Rachel Fisher-Foye, Tory Drescher, Long Li, Hsiao-Yun Lin, Scott Russo

Icahn School of Medicine At Mount Sinai, New York, New York, United States

Background: Alcohol use disorder is often comorbid with major depression and manifests similar behavioral alterations including deficits in reward valuation and heightened stress reactivity. Biologically, the peripheral immune system has been shown to contribute to these deficits with the blood-brain barrier emerging (BBB) as a crucial structure for maintaining appropriate behavioral responding. Both chronic alcohol use and stress have shown to facilitate a pro-inflammatory profile however, how these factors interact with the BBB remains unclear. Interestingly, using the intermittent access two-bottle choice paradigm and chronic social defeat stress model, we find that neutrophil elastase (NE), a secreted neutrophil serine protease, is upregulated in stress-susceptible mice and after ethanol exposure. Given the proteolytic nature of NE, we hypothesize that aberrant expression after chronic insults may be contributing to maladaptive behavioral reactivity.

Methods: To address this phenomena, adult male mice were exposed to a drinking paradigm that models moderate binge drinking and is thus relevant to both recreational drinking and the progression to pathological alcohol drinking habits. After 4 weeks, alcohol was discontinued, and mice were left undisturbed for 2 weeks. At this time, all mice were exposed to a subthreshold stressor and behavioral reactivity was assessed. For chronic stress, Mice were expose to 10 days of chronic social defeat stress and then stratified by their interaction ratio ( > 1=resilient (RES), < 1=susceptible (SUS)). Mice were testing in a behavioral battery to assess stress reactivity and separate cohorts of mice were trained in reward based operant learning as an additional assessment of reward-related deficits. Blood was collected to be analyzed by flowcytometry and plasma was obtained to determine the expression of pro-inflammatory mediators. To assess BBB integrity we calculated whole brain expression of peripherally injected dyes as well as expression of the tight junction protein Claudin 5, within the nucleus accumbens. Recombinant NE administration was used to precipitate stress- and alcohol-induced dysregulation and pharmacological inhibition of NE was used to attenuate stress reactivity. In a separate group of mice Claudin 5, was virally knocked-down to evaluate the direct contribution of BBB breakdown.

Results: Stress and alcohol similarly promote expression of pro-inflammatory cytokines and increase BBB permeability via Claudin 5 downregulation and is recapitulated by endogenous NE administration. Attenuating NE activity promoted adaptive behavioral responses to stress-eliciting stimuli normalized reward-related deficits.

Conclusions: NE is a common pathological factor between stress-susceptibility and chronic alcohol use and could be a promising target for affective dysregulation.

Keywords: Chronic social and non-social stress, Alcohol, Reward, motivation and Anhedonia

Disclosure: Nothing to disclose.

P757. Obesogenic State is Associated With Changes in Striatal Insulin Receptor Levels and Functioning in Mice

Miriam Bocarsly*

Rutgers NJMS, Newark, New Jersey, United States

Background: Understanding the neuronal circuitry underlying feeding behaviors is necessary as current estimates suggest that 40% of adults, globally, are overweight. Historically, excessive weight gain and obesity were considered lifestyle or behavioral disorders. However, recent understanding points to the complexity of the disease, which relies heavily on the actions of the central nervous system and the brain’s control of food intake. While neuronal circuitry underlying food intake is traditionally studied in the context of maintaining homeostatic control of energy balance, amassing evidence supports the idea that control of caloric intake also involves calculations of hedonic value, reward and motivation. Our current research focuses on the role of the striatum, a brain area implicated in reward and motivation, in regulating feeding behaviors.

An interesting observation has linked the actions of insulin in the striatum to local dopamine release. Here, in a preclinical model, we demonstrate the effects of a high-fat diet on insulin-mediated striatal dopamine release, identify an interaction between striatal insulin receptors and obesity-related alterations in dopamine D2-receptors, and propose a circuit-level mechanism.

Methods: Utilizing transgenic technology, we selectively knockdown the dopamine D2 receptor in the striatum, and examined the effects on behavioral measures, such as food intake, locomotion, and willingness to work for a preferred food, peripheral measures such as body weight, glucose tolerance and insulin sensitivity, and central measures of insulin receptor availability and dopamine release in response to insulin.

Results: Using transgenic methods, we designed a mouse model with low striatal D2Rs. These mice gained more weight on a standard rodent chow compared to littermate controls, and indicated that these mice were willing to work for food regardless of internal hunger state. Interestingly, qPCR results showed these mice have increased levels of striatal insulin receptor mRNA, and they showed elevated dopamine release in response to insulin using FSCV.

Conclusions: We demonstrate a compelling, interaction between striatal dopamine D2 receptors and insulin receptors. Further, we demonstrate that this system is dysregulated in an model of obesity, suggesting a novel mechanism by which central reward processing may be perturbed in instances of obesity.

Keywords: Obesity, Natural rewards, Striatal Dopamine signaling, D2 dopamine receptor

Disclosure: Nothing to disclose.

P758. Temporal Dynamics of Mesocorticolimbic Dopaminergic Modulation of Risk/Reward Decision Making

Jackson Schumacher*, Mudi Zhao, Stan Floresco

The University of British Columbia, Vancouver, Canada

Background: Optimal risk/reward decision making is integral to human mental and physical health and relies on dopaminergic (DA) signalling. Risk/reward decision making is often studied using operant tasks such a probabilistic discounting which requires rats to choose between a small, certain and a large, risky reward whose probability varies systematically across blocks of trials. Prior work has shown that nucleus accumbens (NAc) DA, via D1 receptors, promotes preference for the large/risky reward whereas DA in the medial prefrontal cortex (mPFC) refines choice via opposing action on D1 and D2 receptors. Microdialysis studies demonstrate that DA efflux in the NAc and mPFC tracked reward uncertainty and overall reward rates, respectively. However, DA release manifests as discrete phasic events and pharmacology and microdialysis lack the temporal resolution to disentangle these temporally specific effects.

Methods: To address this gap in understanding we performed 2 experiments. First, we selectively silence DA neurons in the VTA of male and female TH-cre long evans rats (n = 11 total) using the inhibitory opsin Arch3.0. Silencing occurs at specific timepoints in probabilistic discounting behavior (prior to choice, following small/certain rewards or following large/risky wins or losses). Performance during silencing sessions is analyzed compared to baseline days and a separate EYFP control group undergoes the same testing regiment. Data are analyzed using ANOVA including sex and task variant (ascending or descending probability across blocks) as a between subjects factor and probability block and treatments as within subjects factors. Second, fiber photometry experiments use the dopamine sensor GRABDA3h to measure DA in the mPFC, NAc shell or NAc core during probabilistic discounting.

Results: Data using optogenetic silencing of dopaminergic VTA neurons has demonstrated that silencing during receipt of either the large/risky or small/certain reward biases choice towards the other option. These changes in choice behavior are associated with changes to the propensity to follow an unrewarded large/risky choice with a small/certain choice (lose-shift behavior). Specifically, inhibition on large/risky and small/certain rewards increased and decreased lose-shift behavior, respectively. Preliminary data from fiber photometry studies have uncovered transient fluctuations in DA release prior to choice and after various outcomes with the type of choice, outcome and probability block modulating the direction and magnitude of the fluctuation. Specifically, in the NAc shell, transient increases in DA release prior to choice are greater preceding large/risky choices than those preceding small/certain choices. For outcomes, there is a robust increase in DA signal after large/risky wins, a decrease following losses and no change following receipt of the small/certain reward. Finally, the magnitude of win and loss related fluctuations is modulated by block with the largest post-win DA increases being seen in the lowest probability blocks and the largest post-loss DA decreases being seen in the highest probability blocks.

Conclusions: These findings illuminate that DA plays a complex role in guiding action based on prior outcomes in situations where reward contingencies change across time. Follow up studies using fiber photomerty will expand these results, examining how mPFC and NAC (shell and core) DA encode reward magnitudes and probabilities of different outcomes. Follow up studies with optogenetics will examine the role of specific VTA pathways (VTA- > NAc vs VTA - > mPFC) in probabilistic discounting behavior. Together, these findings will help extend our theoretical understanding of dopamine signalling to complex scenarios that require animals to adapt to changing risk/reward contingencies.

Keywords: Dopamine, Decision Making, Nucleus Accumbens, Medial Prefrontal Cortex, risk-reward paradigm

Disclosure: Nothing to disclose.

P759. Validation of Assessments for Reward Processing, Positive Emotion, and Anhedonia for Use in a Phase 2 Study of a Novel Histamine H3 Inverse Agonist, ALTO-203, in Major Depression

Nicholas Cooper*, Josh Jordan, Quentin Huys, Greg Hajcak, Mike Avissar, Jessica Powell, Amit Etkin, Adam Savitz

Alto Neuroscience, Mountain View, California, United States

Background: Increasing evidence suggests decreased reward system dopaminergic signaling is a key factor in some patients with depression, yet there are no treatments directly addressing this issue. Anhedonia, the absence of pleasure or motivation in depression, likely results from this dopaminergic deficit. ALTO-203, a novel Histamine 3 receptor (H3R) inverse agonist, has shown distinct pharmacological effects, including increased dopamine release in the nucleus accumbens in animal studies. In a Phase 1 trial, ALTO-203 significantly enhanced positive subjective emotion as measured by the Bond-Lader Visual Analogue Scale (BL-VAS), on par with modafinil, which is known to boost dopamine release in the reward system and has antidepressant effects. Therefore, ALTO-203 shows potential as an antidepressant targeting reward system dopamine dysfunction, with significant single-dose impacts. To better understand the best way to measure treatment effects with ALTO-203, we conducted a study on reward tasks and clinical measures in individuals with a range of anhedonia.

Methods: We tested three novel behavioral tasks designed to assess different aspects of reward processing and motivation: Effort Tapping (sensitivity to high or low reward), the Probabilistic Instrumental Learning Task (PILT, learning which symbol gives more points), and Buckets (reward threshold at which to leave the current activity patch in favor of a new one). These tests were studied alongside established reward tasks from the literature: the Probability Reward Task (PRT) and the Effort Expenditure for Rewards Task (EEfRT). Additionally, we included survey measures of anhedonia and affect, including the Dimensional Anhedonia Rating Scale (DARS), the Positive Valence Systems Scale (PVSS), and the Snaith-Hamilton Pleasure Scale (SHAPS) for anhedonia, as well as the Behavioral Inhibition and Behavioral Activation System (BIS/BAS) and the Positive and Negative Affect Schedule (PANAS) for affect. Lower scores for the DARS and PVSS, and higher scores for the SHAPS correspond to increasing anhedonia. We also collected depression symptomatology via the Patient Health Questionnaire (PHQ-8, minus the suicidality question).

Results: We found significant correlations between performance in the new reward tasks and survey measures of anhedonia and positive emotion (N = 168), as well as BIS/BAS ‘Behavioral Approach’ subscales (N = 114). The Effort Tapping task showed a positive correlation between the difference in tapping rates for high and low reward conditions and the Dimensional Anhedonia Rating Scale (r = 0.16, p < .05) and Positive Valence Systems Scale (r = 0.18, p < .05), indicating increased reward sensitivity in anhedonic subjects. The Buckets task also showed a positive correlation between the slope of reward gained and the DARS (r = 0.16, p < .05) and PVSS (r = 0.22, p < .01), with less anhedonia relating to a tendency to stay longer in the current activity over the course of the task. The Buckets slope of reward gained was also associated with lower depressive symptoms (PHQ-8, r = -0.31, p < .001). The Probabilistic Instrumental Learning Task’s beta coefficient, derived from a reinforcement learning model, correlated positively with the BIS/BAS Drive (r = 0.27, p < .01) and Reward Responsiveness measures (r = 0.23, p < .05), linking higher motivation and reward responsiveness to improved task performance. In contrast, PRT and EEfRT did not show significant correlations with the survey measures, suggesting they assess different reward dimensions or are limited in their sensitivity.

Conclusions: Our results provide support for the use of our novel behavioral tasks and measures of positive emotion to understand putative reward system dopaminergic effects. As such, they are included in an ongoing Phase 2 proof-of-concept trial of ALTO-203 in patients with major depression and anhedonia. The study is designed as a single-dose, three-way, placebo-controlled crossover between two doses of the drug and placebo, with the primary outcome being a change in acute subjective positive emotion on the BL-VAS (NCT06391593), with change in these new reward tasks as exploratory endpoints.

Keywords: Dopaminergic system, Reward, motivation, and anhedonia, Major Depressive Disorder (MDD), Cognitive test, biomarker development

Disclosure: Alto Neuroscience: Stock / Equity - Publicly Traded Company (Self).

P760. Opioid-Induced Hypofrontality Serves as Computational Landscape to Prioritize Relapse

Elizabeth Doncheck*, Roger Grant, Lisa Green, Elizaveta Romanova, Josh Boquiren, Aryan Parmar, Sophie Buchmaier, Jacqueline Paniccia, Rachel Clarke, Jade Baek, Michael Scofield, James Otis

Medical University of South Carolina, Charleston, South Carolina, United States

Background: Paradoxical prefrontal cortical activity patterns are hallmarks of substance use disorders. While phasic prefrontal hyperactivation is predictive of drug cravings, relapse, and is necessary for drug seeking, chronic reductions in prefrontal activity, volume, and function are common in patients with a history of drug dependence and such hypofrontality is broadly believed to impair decision making. The computational underpinnings of these divergent activity dynamics are unknown, in part due to challenges in measuring single-cell computations as drug use develops.

Methods: To address this, we developed a head-fixed heroin self-administration procedure to allow for longitudinal tracking of prelimbic prefrontal cortex (PrL) neuronal activity in mice during behavior. Chemogenetic approaches were used to assess the necessity for PrL pyramidal neuron activation for heroin-seeking reinstatement in head-fixed mice (AAV9-CaMKIIa-hM4Di n = 10, -eYFP = 14 mice; mixed-sex). To visualize their activity, PrL pyramidal neurons were labelled for calcium imaging (AAVdj-CaMKIIa-GCaMP6s) and a GRIN lens implanted dorsal to PrL. In vivo two-photon recordings of PrL pyramidal neuron calcium dynamics (n = 9 male, 7 female mice; n > 2000 cells/day) were taken from the onset of intravenous heroin self-administration through extinction and reinstatement. Principal component analysis-based unsupervised spectral clustering was used to identify uniquely responding neuronal ensembles among heterogeneous population-level activity. Single-cell tracking and deep learning-based spike inference were used to dissect the longitudinal evolution and stability of activity dynamics within individual neurons.

Results: Akin to observations in freely moving drug self-administration assays, chemogenetic manipulations demonstrate PrL pyramidal neuron activity is necessary for drug-seeking reinstatement in head-fixed mice. Two-photon calcium imaging highlights the activity of individual PrL pyramidal neurons is, however, highly heterogeneous during heroin seeking. Within the greater population, spectral clustering uncovers eight distinct PrL neuronal ensembles (n = 4 excitatory, n = 4 inhibitory), with phasic activity patterns that differentially stabilize and computationally predict heroin seeking, that emerge from the onset of heroin use through relapse. These analyses also uncover that not all PrL neurons encode heroin seeking ( ≥ 53% neurons/day). Deep learning-based spike inference reveals global and persistent prefrontal hypoactivation emerges following heroin use in terms of both basal and mean firing rates, burst frequency and firing rate, and in signal-to-noise ratios. Strikingly, while the activity of neurons that show little encoding of heroin seeking becomes profoundly hypoactive, neuronal ensembles with activity that stably predicts future heroin seeking are spared (encoding neurons = 67% excitatory). Finally, ensembles showing both the most profound excitatory responses and best predictive validity of lever press rates are consistently the most prevalent during seeking sessions.

Conclusions: Overall, these findings reveal that opioid use does not merely drive hypofrontality, but rather provides a biased landscape upon which discrete neuronal ensembles can prioritize drug seeking. Single-cell holographic optogenetics studies are currently underway to dissect whether relapse-encoding ensembles in PrL (n = 6 mice) are functionally relevant for behavior.

Keywords: Substance abuse disorders, Medial Prefrontal Cortex, Two-photon calcium imaging, intravenous drug self-administration

Disclosure: Nothing to disclose.

P761. Chemogenetic Stimulation of BNST Corticotropin-Releasing Factor Cells Perturbs Effort Related Choice

Mudi Zhao*, Jackson Schumacher, Payton Mackwood, Stan Floresco

The University of British Columbia, Vancouver, Canada

Background: Human and animal research highlight the contributions of the neuropeptide, corticotropin-releasing factor (CRF), in physiological and behavioural responses during stress. In humans, elevated plasma CRF levels are observed in stress-related disorders, characterized by cognitive impairments and motivational deficits, including a reduced willingness to pursue rewarding stimuli. In rats, acute restraint stress reduces preference for larger rewards associated with greater effort cost, without affecting the incentive value of large rewards. These effects are driven in part by increased CRF transmission, as they are blocked by CRF antagonists and recapitulated by treatment with exogenous CRF. Yet, the endogenous CRF pathways that may perturb effort-related decision and motivation remain unclear. One possible candidate is the bed nucleus striatum terminalis (BNST), a region of the extended amygdala. The BNST is activated under stressful conditions and BNST CRF cells project to key nodes within the mesocorticolimbic reward system. Here, we examined whether chemogenetic excitation of BNST CRF cells is sufficient to drive motivational deficits observed during stress.

Methods: Adult male and female Long Evans rats were trained on an effort discounting task where they choose between a low effort/small reward option that requires 1 lever press for the reception of 1-sugar pellet reward, or a high effort/large reward option that requires a set number of presses (1, 5, 10, 20) to obtain a larger, 4-sugar pellet reward, with the effort requirement increasing across blocks within a session. BNST CRF cells were targeted selectively by infecting them with two viral vectors: (1) AAV8-CRF-Cre encoding the non-endogenous Cre-recombinase under the control of the CRF promoter, and (2) AAV8-hSyn-DIO-hM3Dq-mCherry (or AAV8-hSyn-DIO-mCherry in control animals) inducing expression of the excitatory designer receptor, hM3Dq, only in the presence of Cre-recombinase. Stimulation of BNST CRF cells was achieved by systemic administration of clozapine-N-oxide (CNO; 1mg/kg and 3mg/kg) 10 minutes prior to animals performing on the effort discounting task.

Results: In hM3Dq-mCherry animals, excitation of BNST CRF cells with systemic CNO administration (1mg/kg and 3mg/kg) reduced preference for the high effort/large reward option. This was accompanied by a trend increase in choice latency, which may reflect decision hesitancy. Shift in choice was not primarily driven by a loss of task engagement as measured by trial omissions or incomplete trials where rats did not complete the required number of presses on the high effort/large reward option. High dose of CNO (3mg/kg) also reduced rates of pressing on the high effort lever, indicating reduced response vigor. In animals that expressed mCherry only, CNO administration had no effect. Together, these findings demonstrate that heightened excitability of BNST CRF cells mirrors the effects of acute restraint stress and treatment of exogenous CRF on effort-related choice, suggesting a potential role of BNST CRF cells in how stress affects motivation.

Conclusions: Reduced willingness to expend effort for rewarding stimuli is a core symptom in many stress-related disorders. Hyperexcitability of BNST CRF neurons may contribute to the pathogenesis of stress-related impairments in motivation and serve as a potential therapeutic target.

Keywords: Reward, motivation and Anhedonia, Acute Stress, Corticotropin-releasing factor (CRF), Effort Discounting

Disclosure: Nothing to disclose.

P762. Inflammation in Combination With Metabolic Deficits Underly Anhedonia-Like Behavior in Treatment Resistant Male Rats

Roger Varela*, Heather Macpherson, Tristan Houghton, Clarissa Yates, Nathanael Yates, Dara Daygon, Susannah Tye

Queensland Brain Institute, The University of Queensland, Brisbane, Australia

Background: Anhedonia is defined as either a reduced ability to experience pleasure or a diminished interest in engaging in pleasurable activities, also known as motivational anhedonia. It is a transdiagnostic psychiatric feature and a negative prognostic factor for depression, predicting poor response to antidepressant therapies. The role of mesocorticolimbic dopamine (DA) circuitry underlying anhedonia in humans is well established, however the precise biological triggers for this DA dysfunction are not fully understood. According to the immunometabolic theory of depression, the increased peripheral metabolic demands induced by chronic low-grade inflammation signals the brain to induce a sustained reduction of striatal DA resulting in impaired motivation. The present study aims to model these central-peripheral crosstalk in animals to better elucidate the biological mechanisms of DA-immune axis disruption underlying anhedonia-like behaviors in chronic stress state.

Methods: All experimental protocols were approved by the Institutional Animal Care and Use Committee and were conducted in accordance with the National Institutes of Health and National Health and Medical Research Council Guide for the Care and Use of Laboratory Animals. Motivation was assessed through the Effort Related Choice Task (ERCT). For this, adult male Wistar rats were trained in operant chambers to press a lever to acquire a sucrose pellet (reward) in a progressive-ratio paradigm. Free lab chow was concurrently available in the chamber during the test. The number of lever pressing, rewards retrieved, maximum ratio achieved, and chow consumed were recorded by the end of each session. This DA-dependent behavioural paradigm examines motivation and preferential bias between differently valued food sources based upon associated effort costs. After reaching a stable baseline in the ERCT, animals received daily intraperitoneal (i.p.) injections of adrenocorticotrophic hormone (ACTH) or saline (n = 8) for 24 days in order to mimic a chronic stress state. Antidepressant response was then measured by injecting bupropion (DA transporter inhibitor) at 20mg/kg and 40mg/kg i.p. prior to the last two ERCT sessions. After the last session animals were euthanized, and serum, peripheral blood mononuclear cells (PBMCs), and brain were isolated. Peripheral and central markers of inflammation; glucose metabolism; and dopamine cell integrity were quantified using immunohistochemistry, Western blot, ELISA and metabolomics approaches. Normality of data were determined using Shapiro-Wilk and D’Agostino-Pearson tests, and variance by calculating F-ratios. Mean differences between groups were examined using unpaired one-tailed t-tests; repeated-measures one-way ANOVAs with Dunnett’s post hoc, Mann-Whitney one-tailed U-tests or Friedman tests with Dunn’s post hoc; or two-way ANOVA with Tukey’s post hoc. Correlations analysis was performed with Pearson’s or Spearman’s test and tested using linear regression, followed by Holm-Šídák post hoc. A p-value of < 0.05 was considered statistically significant for all results. Supervised analysis using partial least squares-discriminant analysis (PLSDA) was performed. Permutation analysis was also performed to cross-validate the PLSDA model.

Results: Chronic ACTH treatment induced an anhedonia-like phonotype in animals observed by reduced lever pressing (p = 0.0208) and rewards (p = 0.0213) in the ERCT when compared to controls. Bupropion challenge was not able to reverse this alteration. Increased Interleukin(IL)-6 levels in brain (p = 0.0228) and serum (p = 0.0203), decreased glucose uptake (p = 0.0080) and metabolism (p < 0.05 for CMM metabolites), and decreased Tyrosine Hydroxylase (TH) levels in the nucleus accumbens (NAc) shell (p = 0.0047) and infralimbic cortex (p = 0.0134) were also observed in ACTH animals compared to saline-treated controls. Multivariate analysis revealed that peripheral IL-6 levels, PBMCs glucose uptake and impaired nucleotide metabolism were important variables to the anhedonia-like phenotype model (VIP scores > 1). Finally, correlation analyses also confirmed that NAc Shell TH levels were positively correlated with rewards acquired in ERCT (p = 0.0479), and PBMCs glucose uptake was positively correlated with rewards (p = 0.0462) and number of lever pressing (p = 0.0113). TH levels were also correlated with local IBA1 density (a marker of ramified microglia) in infralimbic cortex (p = 0.0380) and prelimbic cortex (p = 0.0444).

Conclusions: These data suggest that stress hormone-induced upregulation of inflammation concurrent with the impairment of insulin-mediated glucose uptake into immune cells is associated with disruption of nucleotide metabolism essential for TH synthesis with impaired central dopamine synthesis contributing to the behavioral expression of treatment-resistant anhedonia. These results suggest that immunometabolic perturbations concomitant with impaired insulin action at the level of the immune cell result in a metabolically deficient state that directly impacts nucleotide precursors essential for dopamine synthesis and effortful behavior. These results highlight the potential for immune and metabolic markers for individualized treatment of refractory depression and anhedonia.

Keywords: anhedonia, Immunometabolism, inflammation, animal model, Dopamine

Disclosure: Nothing to disclose.

P763. Dynorphin Neurons in the Dorsal Nucleus Accumbens Shell Block Fentanyl- and Natural-Reward Seeking

Sineadh Conway, Nathanyl Ross, Blanca Perez-Palomar, Enze Hao, Ream Al-Hasani*

Department of Anesthesiology, Center for Clinical Pharmacology, Washington University in St. Louis, St. Louis, Missouri, United States

Background: The dynorphin and kappa opioid receptor (KOR) system has been identified as critical in the homeostatic regulation of natural rewards. This system in also known to modulate negative affect associated with motivated states, however, we previously showed that photostimulation of dynorphin neurons in the dorsal nucleus accumbens shell (dNAcSh) but not the ventral nucleus accumbens shell (vNAcSh) surprisingly drives preference-, reward-like behaviors. We also showed that this photostimulation is accompanied by a real-time increase in both dynorphin and dopamine. This work highlights that the dynorphin/KOR system may play a more nuanced role in reward seeking behaviors. Therefore, the goal of the this work is to better understand the functional role of dynorphin neurons in the dNAcSh in natural and fentanyl reward seeking behaviors.

Methods: Sucrose self-administration:

Preprodynorphin-Cre (pDyn) mice were injected with either AAV-DIO-ChR2-eYFP or AAV-DIO-eYFP virus into the dNAcSh (AP + 1.3, ML + /- 0.5, DV -4.25). After 30 days to allow for viral expression, mice first underwent magazine training overnight, followed by fixed ratio 1 (FR1) training in daily 1-hour sessions. A test session was implemented after mice reached FR1 criterion ( < 20% variability in nosepokes and pellets consumed across 3 days). Then mice received FR1 test session plus 20 Hz stimulation. Test session performance was compared to the average performance across the 3 previous days on which FR1 criteria was met.

Food-restriction sucrose self-administration:

The same protocol as above was followed but mice underwent overnight food restriction and an additional test session the following day.

Fentanyl self-administration:

pDyn-Cre mice were injected with Cre-dependent excitatory DREADD (hM3Dq) or control (mCherry) virus into the dNAcSh (AP + 1.3, ML + /- 0.5, DV -4.25). After 30 days to allow for viral expression, mice underwent fentanyl self-administration training for 5 days. Mice were given the option to drink either water or fentanyl (10ug/ml) for 2 hours for 5 days. Test session 1 followed during which mice received an injection of saline or DCZ (DREADD ligand) prior to water and fentanyl access. Mice received test session 2 so that each mouse received both a saline and DCZ test. Mice underwent extinction (2 sippers containing water) followed by reinstatement test in which mice were given an injection of either saline or DCZ prior to the reintroduction of fentanyl and water.

Social Interaction:

On day 1 test mouse allowed to explore an arena containing two mesh cups to habituate to the novel environment for 10 minutes. On day 3 and 5 the same procedure was repeated but with a stranger mouse in the mesh cups plus either no stimulation or 20Hz stimulation. Time spent in a defined interaction zone were scored using ethovision XT.

In all the experiments, within subject statistical comparison were made using a one-way or two -way ANOVA and post-hoc comparisons, n = 38 including male and female mice.

Results: Sucrose self-administration:

In both male and females mice activation of dNAcSh dynorphin neurons significantly decreased operant responding for sucrose pellets and pellets consumed compared to baseline. Additionally, ChR2 mice consumed significantly less sucrose pellets during the 20 Hz test compared to eYFP controls.

Food-restriction sucrose self-administration:

Food restricted mice showed significantly less nosepokes during the 20 Hz stimulation test compared to the 0 Hz test. Food restricted mice showed no significant difference in pellets consumed during the 20 Hz stimulation test. In a food restricted state with no stimulation females consume significantly more pellets than males.

Fentanyl self-administration:

In male and female mice chemogenetic activation of dNAcSh dynorphin neurons significantly decreased fentanyl consumption and preference, compared to controls. We also observed a significant decrease in fentanyl reinstatement.

Social Interaction:

Photostimulation of dNAcSh dynorphin neurons significantly decreased interaction time as compared to no stimulation. This effect seemed to be driven by male mice who showed a significant decrease in interaction time during photostimulation.

Conclusions: We observed a decrease in operant responding for a sucrose reward in both males and females. We found that following a food restricted state, activation of dynorphin neurons blocked the food restriction-induced increases in sucrose seeking and consumption observed in controls. To further determine if this dynorphin-mediated effect is limited to natural rewards we investigated whether activation of dynorphin neurons decreased fentanyl seeking and reinstatement. We observed that both fentanyl seeking and reinstatement were blocked following stimulation of dynorphin neurons in the dNAcSh. Finally, we wanted to determine whether this dynorphin-mediated effect was limited to consummatory behaviors so we investigated social interaction and found that activation of dynorphin neurons in the dNAcSh decreased social interaction. Together we show that stimulation of dynorphin neurons in the dNAcSh using both optogenetics are chemogenetics blocked reward in a variety of natural reward paradigms and fentanyl seeking and reinstatement behaviors. We previously showed that activation of these dynorphin neurons drives reward-like behavior so it seems likely that during activation of these neurons mice are experiencing a hedonic state that outcompetes the other rewarding stimuli.

Keywords: Dynorphin, Fentanyl, Reward seeking, Natural rewards

Disclosure: Nothing to disclose.

P764. Sex-Specific Effects of Dextroamphetamine on Choice and Affective Response to Reward in a Counterfactual Gambling Task

Jessica Cooper*, Jonathan Ryan, Heather Webber, Paula Lopez-Gamundi, Jennifer Hoots, Cecilia Nunez, Scott Lane, Margaret Wardle, Michael Treadway

Emory University School of Medicine, Atlanta, Georgia, United States

Background: While reward processing is generally considered to be dopamine-dependent, research on the effects of dopaminergic manipulation on affective responses to rewards has been limited. Moreover, there is growing interest in how biological sex might influence these effects. To address this gap, we examined the impact of dextroamphetamine on choice and affective responses to outcomes during a counterfactual gambling task. This work aims to explore how dopaminergic manipulation affects reward processing differently across sexes, potentially revealing sex-specific patterns in affective responses to rewards.

Methods: Over three sessions, 30 participants (50% female, ages 18-35) each received placebo,10mg, or 20mg of dextroamphetamine under counterbalanced double-blind conditions. Following drug or placebo administration, participants completed an 80-trial counterfactual gambling task where they made choices between two wheels, each with two potential reward or loss outcomes of varying probability of receipt. Affective responses were collected following receipt of the outcome of the selected wheel (partial feedback) and after viewing the outcome of the unchosen wheel (complete feedback). Mixed effects models were used to examine effects of d-amphetamine, biological sex, and the interaction between sex and drug condition on choice and affective ratings.

Results: Choices in both drug conditions (10mg, 20mg) were characterized by reduced impact of expected value (β = -.323, z = -3.01, p = 0.002; β = -.249, z = -2.27, p = 0.023) and reduced sensitivity to potential risk (β = 0.170, z = 2.45, p = 0.014; β = 0.168, z = 2.41, p = 0.016) relative to choices made in the placebo condition. Biological sex did not moderate the effect of d-amphetamine on decision variables (i.e. expected value, risk, ps > .05). For affective rating following the outcome of the chosen wheel, a significant sex x drug x outcome 3-way interaction was observed (β = -.018, t = 3.66; p < .001) for the 20mg condition relative to placebo, where male participants showed increased affective reactivity to outcomes under the drug condition compared to placebo.

Conclusions: Here, we explore a previously unexamined effect of d-amphetamine on decision making and affective reactivity. We find that administration of d-amphetamine affects decision making behavior by reducing the impact of both expected value and potential risk on choice. We additionally find that affective reactivity to outcomes is increased by d-amphetamine in male participants. These findings highlight the need for larger studies on sex-specific effects of d-amphetamine on affective response and sensitivity to reinforcement, and suggest new directions for research on stimulant abuse risk.

Keywords: Reward-based decision-making, Risky decision-making, Amphetamine, Sex-specific effects

Disclosure: Nothing to disclose.

P765. Endogenous Opioid Peptide in Dorsal Raphe Nucleus Modulates Pain and Motivation

Kathryn Braden*, Marcela Arguello, Diego De Gregorio, Aidan Evans-Strong, Daniel Castro

Washington University in St. Louis, St. Louis, Missouri, United States

Background: Chronic pain is a complex disease state associated with comorbidities such as depression, anxiety, and increased risk of suicide. Historically, the majority of analgesic research focused on peripheral and spinal sensory mechanisms despite emotional unpleasantness being a key feature of the human pain experience. It is important to understand the basic neurobiology behind the motivational dysregulation of pain to effectively treat it. The endogenous opioid system can powerfully modulate both analgesia and motivational neural circuits and is therefore well-suited to modulate pain-induced affect. Dorsal midbrain nuclei such as periaqueductal grey (PAG) and the adjacent dorsal raphe nucleus (DRN) have been shown to be important sites of opioid action. The PAG is canonically associated with opioid-mediated descending pain inhibition and is desensitized during chronic pain states. The specific role of the DRN during chronic pain has not been characterized despite studies showing opioid activity here also modulating pain and motivated behaviors.

Methods: To investigate the functional significance of DRN opioid signaling we are using CRISPR-Cas9 mediated knockdown to disrupt enkephalin production and measuring consequent changes in a variety of appetitive and aversive behaviors (expected n = 12 mice/sex). These behaviors include: thermal and mechanical allodynia, odor preference/avoidance, social interaction, food intake, and naloxone-precipitated withdrawal. All behavioral tests will be counterbalanced when possible, with several days between each test. When possible, mice are tested within subjects. We have also begun to characterize the overlap of DRN enkephalin with other neuronal cell types using Hiplex Fluorescent In Situ Hybridization (FISH). For all studies, male and female mice are being used. All animal procedures were approved by the Washington University IACUC and were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Results: We found that Cre-dependent CRISPR mediated knockdown of enkephalin peptide (Penk) in the DRN produced a hyperalgesic response to low-dose carrageenan injection compared to controls (Two-Way ANOVA, between subjects, Sidak’s multiple comparison; F (1,14) = 9.482, p = 0.0082; t(8,8) = 4.002, p = 0.0008; Cre- males = 4, Cre- females = 4, Cre+ males = 5, Cre+ females = 3). Loss of enkephalin in the DRN also enhanced avoidance of an aversive odor compared to controls (Welch’s t-test, between subjects, t(4,8) = 2.903, p = 0.0247; Cre- males =6, Cre- females = 2, Cre+ females = 4). However, there were no differences between Penk-Cre+ knockdowns and littermate Penk-Cre- controls in appetitive behaviors such as social interaction or food intake.

Preliminary FISH studies indicate that Penk gene expressing cells compose 16.47% ( ± 4.11) of the DRN. Of those Penk-expressing cells 49.93% ( ± 1.88) also express vGlut2 and 39.79% ( ± 1.15) express vGlut3 and are therefore glutamatergic while 58.76% ( ± 4.41) co-express vGat and would be considered GABAergic (Data are Mean ± SEM; C57Bl6/J male= 1, female = 1). Since opioid peptides are typically co-released with other neurotransmitters to affect behavior, we also investigated whether these behavioral effects were being driven by either a primarily glutamatergic or GABAergic cell population. To accomplish this, we used vGat-, vGlut2-, and vGlut3-Cre mice in conjunction with our CRISPR mediated Penk knockdown and found that the hyperalgesic response to low-dose carrageenan injection was primarily driven by enkephalin from vGlut2-expressing neurons as the vGlut-2-Cre mice with Penk CRISPR were the only group to have an enhanced mechanical allodynia (Two-Way ANOVA, between subjects (Penk-Cre- vs. vGlut2-Cre + ), Dunnett’s multiple comparison; F (4,32) = 5.959, p = 0.0011; q(8,7) = 2.803, p = 0.0238; Penk-Cre- males = 4, Penk-Cre- females = 4, vGlut2-Cre+ males = 3, vGlut2-Cre+ females = 4). Ongoing work will continue to characterize the effect of DRN enkephalin knockdown within GABAergic and glutamatergic cell types on aversive behaviors.

Conclusions: These results suggest that DRN enkephalin peptide acts to buffer responses to aversive stimuli during certain aversive states. With the loss of enkephalin from the DRN these responses are exaggerated (e.g. increased allodynia following mild inflammatory state, decreased interaction with aversive odor). Preliminary results suggest this effect may be driven by glutamatergic enkephalin cells. Future work will further characterize the role of DRN enkephalin within various neuronal subtypes as well as its role during other aversive states such as naloxone-precipitated withdrawal.

Keywords: endogenous opioids, dorsal raphe, Pain, aversion, CRISPR/Cas9

Disclosure: Nothing to disclose.

P766. Thalamic Neurotensin Projections Regulate Behavioral Selection Under Motivational Conflict

Hao Li*, Valentina Olivera-Pasilio, Reesha Patel, Wanyi Guo, Sirin Jitklongsub, Callista Polasek

Feinberg School of Medicine, Northwestern University, Chicago, Chicago, Illinois, United States

Background: A mild noise can become remarkably alert and trigger defensive reactions when encountered in a fearful state, and food would taste particularly delicious when hungry. To survive in this ever-changing environment, we must adaptively determine the perceived value (valence) of stimuli based on internal and external contexts and adapt our behavior accordingly. The impairments in such flexible processing of valence are reflected in a number of mental health disorders, including anxiety, depression, and schizophrenia1. However, it remains unclear how valence is dynamically encoded in neural circuits to guide behaviors in response to changes in contexts. Compared to fast-acting neurotransmissions, neuropeptides are known to produce more sustained and state-dependent modulatory effects on synaptic activity and behavior. My previous work revealed that neurotensin (NT), a 13 amino acid neuropeptide, plays a critical role in assigning positive or negative valence during associative learning via differential NT release from the paraventricular nucleus of the thalamus (PVT).

Methods: The study used both male and female mice. To model context-dependent behavioral selection, we adopted a progressively punished reward-seeking (PPRS) task, in which mice nose poke for sucrose solution delivery that coincides with a progressively increased footshock. As the shock intensity increased, mice switched from reward-seeking to punishment-avoidance.

Results: Using optogenetics, in vivo electrophysiology, and neurotensin-specific manipulation and recording, we found that PVT NT neurons encode valence (P < 0.001), which is context-dependent and mediated by inputs from the prelimbic cortex (PL). Furthermore, we found that PVT NT neurons can mediate approach and avoidance behavior during the PPRS task via their projections to the central nucleus of the amygdala (CeA) and the nucleus accumbens (NAc), respectively (P < 0.05).

Conclusions: These data place PVT NT in an anatomically and functionally central position in integrating context-related inputs and orchestrating appropriate behavioral selections during valence processing.

Keywords: paraventricular nucleus of the thalamus, neurotensin, Amygdala, Nucleus Accumbens, Neural Reward Circuitry

Disclosure: Nothing to disclose.

P767. Early -Life Exposure to Fluoxetine Impacts Endogenous Opioid Modulation of Motivated Behavior

Emily Cambre, Misbah Sheikh, Elena Christenfeld, Arturo Herraez Torres, Sarah Canetta*

Columbia University, New York, New York, United States

Background: Brief exposure to the SSRI, fluoxetine (FLX), during early life leads to a variety of affective behavioral changes in adulthood, including increased avoidance and despair as well as impaired motivation. We were interested in refining the reward-related behavioral changes following early-life FLX exposure as well as identifying therapeutic strategies for reward-related deficits. To that end, we focused on the opioid system given our prior finding that chronic administration of the mu opioid receptor (MOR) agonist, tianeptine (TIA), improved avoidant behavior in these early FLX-exposed animals.

Methods: Mice were exposed to FLX (10 mg/kg) in early-life (P2-11, EL FLX) or vehicle (EL VEH) and in adulthood ( > 90) were tested in the progressive ratio, lickometer and pavlovian-to-instrument transfer tasks. We chronically administered the MOR agonist, TIA (30 mg/kg s.c., 2x/day for 14 days), or the MOR antagonist, methocinnamox (MCAM, 10 mg/kg i.p., 1x/day every 3 days for 3 weeks), in EL FLX, EL VEH or no early exposure control mice prior to behavioral testing. We also tested acute MCAM (10 mg/kg i.p.) in these tasks. Sample sizes ranged from 11 to 29 mice and both sexes were included.

Results: Adult EL FLX mice showed impaired motivation (e.g. decreased breakpoint: p = 0.0148) while hedonic response to reward and pavlovian-to-instrumental transfer was intact. TIA had no effect on motivation but surprisingly chronic MCAM improved motivation in EL FLX mice (p = 0.0059) while trending towards decreasing hedonic response to reward. Chronic MCAM did not affect motivation in control mice but significantly decreased hedonic response to reward (p = 0.0006). Acute MCAM increased motivation in EL FLX (p = 0.0058), but not in EL VEH, mice.

Conclusions: Brief, early-life exposure to FLX impairs adult motivation, while leaving hedonic response to reward and cue-reward learning intact. Surprisingly, the MOR antagonist, MCAM, improves motivation in EL FLX mice. MCAM does not impact motivation in control animals and decreases hedonic response to reward in both control and EL FLX mice. Together, this work suggests that briefly elevating serotonin signaling in early life affects the development of the endogenous opioid system such that it suppresses motivation, in adulthood. Ongoing experiments aim to identify the mechanism underlying this effect.

Keywords: Selective Serotonin Reuptake Inhibitors (SSRIs), Brain Development, Motivation, Mu opioid receptor, Reward Seeking

Disclosure: Nothing to disclose.

P768. Reward Encoding in Prefrontal Somatostatin-Expressing Interneuron Ensembles

Hunter Franks, Julia Lopes Gonҫalez, Rodrigo Campos-Cardoso, Kirstie Cummings*

University of Alabama at Birmingham, Birmingham, Alabama, United States

Background: Substance use disorder (SUD) is a debilitating disorder characterized by severe alterations in the neural circuits of reward. Among others, the gamma aminobutyric acid (GABA) system is a critical component of the pathology of SUDs. While historically studied in the context of alcohol use, the GABAergic system is now gaining more attention for its roles in mediating the rewarding properties of other drug types (such as opioids) and natural rewards. However, in comparison to other systems like dopamine, our understanding of the GABAergic mechanisms of reward seeking is far less complete. Here, we seek to understand how prefrontal GABAergic interneurons, in particular somatostatin-expressing interneurons (SST-INs), function to shape and encode reward.

Methods: To study the mechanisms underlying reward encoding in prefrontal SST-INs, we used a combination of cFos immunohistochemistry (n = 3-5 mice per group (naive and reward-trained)), whole-cell brain slice electrophysiology (n = 7-12 cells per group from 3-4 mice per group (naive and reward-trained)), fiber photometry calcium imaging (n = 4-5 mice per group (reward trained; PV-Cre and SST-Cre)), and in vivo optogenetic manipulations (n = 6-8 mice per group (eYFP, channelrhodopsin, and archaerhodopsin)). To model reward memory encoding, we used sucrose operant reward conditioning. Data were tested for normality and equal variance prior to statistical testing. If assumptions were met, we performed unpaired t-test or one-way ANOVA with or without repeated measures, where appropriate. If assumptions for parametric tests were violated, non-parametric equivalents were used. Mice of both sexes were used in all experiments in equal numbers.

Results: We observed that prefrontal SST-INs exhibit several correlates consistent with a potential role in reward memory encoding. First, cFos analysis revealed layer and cell type-specific recruitment of prefrontal SST-INs compared to control groups (t(7) = 3.85, P = 0.002, unpaired t-test). Second, we observed that prefrontal SST-INs exhibit physiological correlates related to increased intrinsic excitability as a result of reward learning, including decreased rheobase (t(17) = 2.75, P = 0.015, unpaired t-test), firing threshold (t(17) = 2.74, P = 0.018, unpaired t-test), latency to fire (t(17) = 4.12, P = 0.001, unpaired t-test), as well as significant differences in action potential input-output (F (1, 16) = 6.658, P = 0.02, two-way repeated measures ANOVA). Finally, in vivo fiber photometry calcium imaging revealed that prefrontal SST-INs are recruited across sessions during reward encoding and track reward-related cues (numerous tests; reported on poster).

Conclusions: Overall, we are beginning to understand when and how prefrontal SST-INs are recruited during operant sucrose reward seeking. Aside from affording novel fundamental insight regarding mechanisms of sucrose reward memory encoding in prefrontal SST-INs, our lay the foundation for understanding how the function of these GABAergic ensembles go awry in SUDs, where GABAergic dysfunction is prevalent.

Keywords: Reward memory, Somatostatin-expressing interneurons, Medial Prefrontal Cortex (mPFC), Plasticity

Disclosure: Nothing to disclose.

P769. Cholinergic Interneurons Gate Dopamine Release to Drive Effortful Behavior

Gavin Touponse, Matthew Pomrenze, Teema Yassine, Zihui Zhang, Robert Malenka*, Neir Eshel

Stanford University, Stanford, California, United States

Background: Effort is costly, but it can also be valued: across the animal kingdom, cases abound in which individuals value rewards more when they have worked harder for them. Decades of work in psychology, economics, and neuroscience have revealed intricate mechanisms behind effort’s costliness. Much less is understood about effort’s value. We recently found that dopamine (DA) release in the core of the nucleus accumbens (NAc) tracked the effort required for reward, even when the reward was kept constant. When mice worked harder for reward, more DA was released in response to reward delivery. Surprisingly, we observed the same effect even when animals worked for optogenetic stimulation of DA axon terminals in the NAc. Despite identical optogenetic stimulation parameters, more effort led to greater DA release. Our findings provide a plausible neural basis for how effort can augment a reward’s value. The source of this effect, however, remains mysterious. What is modulating DA release according to effort? Here we focused on the potential role of cholinergic interneurons (CINs) in the NAc. Although they make up a small fraction of the cells in the NAc, CINs are thought to play an outsized role in the modulation of motivated behavior. In vitro studies have long demonstrated that acetylcholine (ACh) can directly stimulate DA release from striatal axon terminals. Several recent studies, however, have questioned whether CINs modulate DA release in vivo. We hypothesized that CINs might modulate DA release specifically in high-effort conditions.

Methods: We established a simple operant task that varies the number of nosepokes mice are required to perform to receive sucrose reward. During a single session, the fixed ratio (FR, or effort requirement) fluctuated in 10-minute bins over 50 minutes. While mice (N = 14, 7 female) performed the task, we measured DA release in the NAc using the genetically-encoded sensor GRAB-DA, allowing us to test how DA reward responses reflected the amount of effort the mice had expended. We then observed how these DA responses changed when we manipulated cholinergic signaling in one of two ways: either local infusion of a nicotinic receptor antagonist (DHβE) vs saline into the NAc (N = 8 mice, 4 female), or optogenetic inhibition of CINs using halorhodopsin (N = 6 mice, 3 female). In a separate cohort of mice (N = 7, 4 female), we then used the genetically-encoded sensor GRAB-ACh to measure acetylcholine release in our task, allowing us to determine if effort modulates cholinergic signaling. In a final cohort of mice (N = 9, 4 female), we infused DHβE or saline bilaterally into the NAc and measured the number of rewards the mice earned at low- or high-effort conditions. We used Friedman’s tests to compare DA or ACh responses across effort requirements and Wilcoxon signed rank tests to compare matched DA responses to DHβE (vs saline) and halorhodopsin (light ON vs OFF). We used a mixed effects model to examine the interaction between effort requirement and drug administration (DHβE vs saline) on the number of rewards earned, accounting for the random effect of subject.

Results: Upon reward delivery, both DA and ACh release were modulated by the amount of preceding effort, with higher effort evoking greater release (Friedman’s test, P < 0.001 for both DA and ACh). When we locally infused DHβE to block the nicotinic receptors expressed on DA axon terminals, DA release was no longer modulated by effort (Friedman’s test, P = 0.58); instead, DA release after high effort remained at the low level usually induced by low-effort trials. Optogenetic inhibition of CINs at the time of reward delivery had the same flattening effect on DA release, such that high-effort DA release was significantly reduced on light ON compared to light OFF trials (Wilcoxon signed-rank, P = 0.01) while low-effort DA release remained the same (Wilcoxon signed-rank, P = 0.42). Finally, we found an interaction between bilateral drug infusion (DHβE vs saline) and effort requirement (FR 1, 5, 10, 21, or 46) on the number of rewards the mice earned on each block (mixed-effects model, P = 0.002), such that DHβE specifically decreased the number of high-effort rewards while leaving low-effort rewards unchanged. In all experiments, we did not observe any behavioral or neural differences between sexes.

Conclusions: By pairing optogenetic and pharmacological manipulations with neural recordings while mice performed an operant reward-seeking task, we found that local ACh facilitates DA release in high-effort conditions, thereby promoting effortful behavior. These results reconcile in vitro studies, which imply strong cholinergic regulation of DA release, with recent in vivo studies, which did not detect such regulation, but also did not modulate effort requirements. Our findings imply that interactions between neuromodulators are context-dependent; in this case, we observed the interaction only by including an effort component. Our results help explain a longstanding observation across species that individuals value rewards more when they have worked harder for them. And they speak to an ongoing controversy about whether the activity of DA cell bodies can be dissociated from the release of DA in striatal target regions. We show that local ACh can expand the range of DA release, with an important function in valuing (and motivating) effort.

Keywords: Dopamine, Acetylcholine, Effort, Nucleus Accumbens, Neuromodulation, Reward seeking

Disclosure: Boehringer Ingelheim: Consultant (Self).

P770. Effects of State vs. Trait Affect on Exploration and Exploitation: A 28-day Reinforcement Learning Study

Aliona Tsypes*, Michael Hallquist, Eran Eldar, Alexandre Dombrovski

University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, United States

Background: Extreme affective states shape people’s decision-making and can trigger substance use, interpersonal conflict, and even suicidal acts. One of the mechanisms through which affect may shape behavior is by altering the balance between choosing familiar good options (exploitation) and pursuing more uncertain options in hopes of uncovering superior alternatives (exploration). Failure to balance exploration vs. exploitation can undermine one’s problem-solving and has been implicated in maladaptive behaviors from anxious avoidance to attempted suicide (Brown et al., 2023; Tsypes et al., 2024). It is unclear, however, to what extent such failures reflect traits vs. are induced by extreme affects typical, for example, of a suicidal crisis.

Normatively, negative affect may promote exploration through threat vigilance or suppress it due to uncertainty avoidance or disrupted uncertainty estimation. Positive affect may increase exploration by inflating long-term reward estimates or decrease it by amplifying short-term opportunity cost estimates. Empirical findings on the effects of negative affect on explore-exploit decision-making are mixed. This is in part due to the failure of existing studies to distinguish between person-level trait affect and within-person affective fluctuations. Even less is known about the effects of positive affect. This intensive longitudinal study sought to disambiguate state vs. trait affective influences on exploration-exploitation.

Methods: We tested competing hypotheses about mood and decision-making in 226 participants (76% female; 70% Caucasian; Mean/SD age: 30.5/7.2) with varying levels of affective instability. Participants completed a validated reinforcement learning task (Eldar et al., 2018) twice daily, resulting in 3, 905 trials with the feedback about winning, losing, or neither winning nor losing. On each trial, they chose between two images with experimentally manipulated reward and punishment probabilities. New images were tested alongside familiar ones to independently manipulate the values and uncertainties of the stimuli and assess exploratory behavior. Computational modeling was used to estimate the expected values of the images on a given trial. Participants reported their positive (elation and energy) and negative mood (anxiety, sadness, and irritability) four times daily. To assess the effects of personality traits characterized by high levels of positive and negative affectivity, participants completed the Personality Inventory for DSM-5, the Big Five Inventory, and the UPPS-P Impulsive Behavior Scale. To test our hypotheses, we conducted a series of multilevel logistic regression models examining the effects of mood on the link between stimulus uncertainty and choice, controlling for the expected values of the stimuli.

Results: On average, stimuli with higher values (exploitation) and higher uncertainty (exploration) were preferred (all ps < .001). Between-person analyses revealed that trait positive affectivity increased exploration, while negative affectivity decreased exploration (all ps < .001). Further, trait positive affectivity predicted optimal choices and trait negative affectivity predicted suboptimal choices. Within-person analyses showed a different pattern: positive mood (energetic, elated) decreased exploration (both ps < .001), while effects of negative mood were null.

Conclusions: State vs. trait positive affects have diverging effects on exploration. While trait positive affectivity increased exploration, aligning with the hypothesis of optimistic long-term reward estimates, momentary positive mood decreased exploration, consistent with upward-biased estimates of short-term opportunity cost. Conversely, trait negative affectivity predicted a general failure of both exploitation and exploration, suggesting broad disruptions in both value and uncertainty learning. Clinically, our observations support interventions focused on increasing general levels of positive affect to promote optimal exploration rather than primarily addressing in-the-moment negative affect regulation.

Keywords: Reinforcement learning, affective instability, exploration-exploitation tradeoff, ecological momentary assessment, personality

Disclosure: Nothing to disclose.

P771. Directional and Temporal Dissociations in VTA Dopamine and GABA Responses to Changes in Reward and Punishment Contingencies

Merridee Lefner*, Bita Moghaddam

Oregon Health and Sciences University, Portland, Oregon, United States

Background: In order to efficiently and safely traverse an environment, one must learn to distinguish between cues that predict outcomes of different modalities. In a dynamic environment where cues predicting rewarding or aversive outcomes may unexpectedly change, it is critical to flexibly update behavioral responding while retaining the ability to recall previous associations. Impairments in the ability to flexibly respond when cue contingencies change are evident in a number of psychiatric disorders. Dopamine and γ-aminobutyric acid (GABA) neurons in the ventral tegmental area (VTA) are involved in appetitive and aversive learning, yet little is known about how they encode changes in competing contingencies. In the current study, we determined how VTA dopamine and GABA population dynamics update following the reversal of previously learned appetitive and aversive associations.

Methods: All procedures were approved by the Oregon Health and Science University Institutional Animal Care and Use Committee and were conducted in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals. Adult male and female TH::Cre (n = 8; 5 females) and GAD::Cre (n = 9; 5 females) Long-Evans rats were trained on 24 sessions of a flexible contingency learning (FCL) paradigm developed in our laboratory to assess responses to reward, punishment, and neutral contingencies experienced concurrently. During FCL, subjects were exposed to three distinct auditory cues (tone, whitenoise, and click) that precede either an appetitive outcome (45 mg sucrose pellet reward), an aversive outcome (0.2 mA, 180 ms foot shock), or no outcome within the same session. The cue-outcome pairings were counterbalanced across subjects. After 8 sessions of initial exposure, the appetitive and aversive outcomes reverse such that the cue previously paired with a shock instead precedes reward delivery and vice versa for 8 more sessions. The cues then reversed back to the initial associations for the final 8 sessions.

Fiber photometry recordings were used to measure changes in calcium activity from VTA dopamine or GABA neuron populations during learning and the two reversal phases of the FCL paradigm. We infused a calcium indicator (AAV1-Syn-Flex-GCaMP6f-WPRE-SV40) into the VTA of TH::Cre or GAD::Cre rats, followed by an optic fiber implanted above the infusion. Behavioral responding and quantification of neural data were analyzed using a 2-way mixed-effects model fit (restricted maximum likelihood method), repeated measures where appropriate, followed by post hoc Tukey’s or Sidak’s tests. The Geisser–Greenhouse correction was applied to address unequal variances between groups.

Results: Rats discriminated between the 3 different cues across initial learning sessions in the FCL task, demonstrating increased conditioned responding to the reward-paired cue (cue effect p < 0.0001; session x cue interaction p < 0.0001). Conditioned responding updated accordingly following the reversal of the appetitive and aversive associations (cue effect p = 0.002; session x cue interaction effect p = 0.0003). There was no effect of sex on behavioral responding during initial learning (sex effect on appetitive cue p = 0.11, aversive cue p = 0.55, neutral cue p = 0.52) or reversal sessions (sex effect on appetitive cue p = 0.12, aversive cue p = 0.40, neutral cue p = 0.37).

Fiber photometry recordings indicate dissociable responses between VTA cell groups throughout the initial acquisition of contingencies during the first phase of the FCL paradigm. The VTA dopamine cell population expectedly increased activity in response to the appetitive association and decreased activity in response to the aversive association (cue effect p < 0.0001; session x cue interaction effect p = 0.005). In contrast, the VTA GABA population indiscriminately increased activity in response to all sensory events regardless of valence, including the neutral stimulus (cue effect p = 0.07; session x cue interaction effect p = 0.13).

The reversal phase selectively influenced GABA, and not dopamine, responses to the reward-predictive cue. Across reversal sessions, GABA activity discriminated the new appetitive association via increased cue-evoked activity (cue effect p = 0.02; session x cue interaction effect p = 0.003). Furthermore, we found that within-session GABA activity also increased in response to the appetitive cue during reversal, and this response profile was sustained across both early and late sessions of the reversal phase (GABA early sessions reward cue trial effect p = 0.002; FCL phase x reward cue trial interaction effect p = 0.01; late sessions reward cue trial effect p = 0.0006).

Conclusions: These findings provide evidence for dissociations in both directionality and temporal response profiles between VTA dopamine and GABA neuron activity during initial learning as well as the reversal of competing appetitive and aversive stimulus-outcome contingencies. Our results implicate the VTA GABA population as a potential modulator of flexible associative learning.

Keywords: Dopamine, GABA, behavioral flexibility, VTA, reward

Disclosure: Nothing to disclose.

P772. Neural Response to Reward and Loss Following Basic Combat Training

Clara Freeman*, Collin Teich, Eric Rawls, Scott Sponheim, Melissa Polusny, Craig Marquardt

Yale University School of Medicine, New Haven, Connecticut, United States

Background: The brain’s responsiveness to rewarding stimuli is essential for adaptive functioning while deficits in neural reward processing have been linked to the transdiagnostic symptom of anhedonia. Acute or prolonged stress exposure may lead to a significant reduction in neural reward responses, possibly contributing to further vulnerability to psychopathology. Conversely, greater baseline levels of reward responsivity may buffer against the impacts of stress exposure. However, most studies that have examined links between reward processing and stress have used laboratory paradigms in which stressors may be tightly controlled but artificial or have used cross-sectional designs in which it is impossible to infer directionality. The extent to which real-world stressors significantly impact how the human brain processes rewards remains unclear. In order to address this gap, the present study examined a sample of Army National Guard recruits who underwent Basic Combat Training (BCT) to observe whether electrophysiological measures of neural response to reward would decrease following this stressor.

Methods: The present pre-registered analysis (https://osf.io/f6e8w) used data from the Advancing Research on Mechanisms of Resilience (ARMOR) project, a longitudinal study that follows Army National Guard recruits before and after their completion of basic combat training (BCT), a physically and mentally demanding 10-week military training program. Participants (N = 69, 73.9% male) completed a virtual gambling task while electroencephalography (EEG) was recorded both before and after completing BCT. This gambling task elicits the Reward Positivity (RewP), an early frontocentral event-related potential that is enhanced to win relative to loss feedback and indexes neural responsiveness to reward. Mean EEG activity in the time window of the RewP (175 – 325 ms post-feedback at FCz) was averaged separately for the win and loss condition, as were delta- and theta-band time-frequency power. Repeated-measures ANOVA was used to assess change in the RewP following BCT and mediation models were used to assess associations between neural responses and self-reported BCT-related stress exposure.

Results: We found that neural response to both win and loss feedback in the time-window of the RewP was significantly reduced following BCT relative to baseline (F(1,68) = 8.02, p = 0.006, ηp2 = 0.105). Furthermore, greater perceived stress exposure during basic combat training was associated with a smaller post-BCT RewP in the win condition after adjusting for baseline reward response (β = -.16 p = 0.030); this was not replicated in the loss condition. Lastly, although delta- and theta-band power did not significantly decrease from pre- to post-BCT, greater delta-band activation at baseline predicted lower reports of perceived stress exposure during basic training (β = -.25, p = 0.041).

Conclusions: These findings support the hypothesis that experiencing real-world stressors can lead to a significant reduction in neural responses to reward. Decrease in response to reward but not loss was larger in those who reported greater perceived basic training stress exposure. Given the link between stress and disorders such as depression, stress-induced decreases in reward responsivity may be a possible mechanism. Furthermore, those with larger baseline activation to reward in the delta-band frequency reported less perceived stress exposure during basic combat training, suggesting a possible indicator of resilience to stressors and their consequences. Future studies are warranted to understand whether these stressor-related decreases in neural reward response increase risk for future psychopathology and whether delta-band responses to reward reliably increase resilience and buffer against future stressors.

Keywords: Reward, Stress, Electroencephalography (EEG), Military

Disclosure: Nothing to disclose.

P773. Prenatal Circadian Rhythm Disruption Induces Sex-Specific Substance Use and Mood-Related Phenotypes in Mice

Lauren DePoy*, Colleen McClung

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: 20% of Americans are at risk for environmental circadian rhythm disruptions (CRD) due to shift work. Shift workers experience substantial negative health outcomes, but females are especially affected with greater vulnerability for substance use (SU) and adverse outcomes associated with pregnancy. These outcomes not only occur during pregnancy, but offspring are affected at birth and later in life. In mice, prenatal CRD (pCRD) recapitulates these risks, increasing adverse pregnancy outcomes and altering behavior in adult offspring. However, it is unknown whether pCRD affects SU in mature offspring.

Methods: C57Bl/6J dams were sham handled or disrupted, by reversing the light/dark cycle 4 times during gestation. Reward- and mood-related behaviors were measured in adult offspring. Cocaine reward was measured using conditioned place preference (5 mg/kg cocaine). Contingency degradation was used to measure decision making. Mice were trained to respond on two levers for food, then the likelihood that one of those levers will be reinforced was degraded. Another cohort was trained to respond for food before jugular catheterization. After recovery, mice were trained to respond on a different lever for cocaine. Acquisition, the reinforcing and motivational properties of cocaine, extinction and cue-induced reinstatement were measured. A separate cohort of mice were tested in the light/dark box, open field and elevated plus maze. Throughout, 2-way ANOVAs were used for data analysis (disruption group by sex), with additional repeated measures as appropriate (i.e. session, lever, dose). Significant interactions were followed up with Sidak posthoc tests. p < 0.05 was considered significant and p < 0.1 trending. Samples sizes of 6-14, depending on the experiment. In order to investigate possible underlying mechanisms, we investigated how gene expression rhythms (RNAsequencing) change in the nucleus accumbens of adult mice with a history of pCRD. Tissue was collected at 6 times of day (ZT2/6/10/14/18/22)(zeitgeber time-time from lights onset) in male and female mice. Procedures were approved by the University of Pittsburgh IACUC.

Results: Interestingly, females exposed to pCRD developed an anhedonic-like phenotype with decreased food self-administration (interaction, p = 0.007), cocaine intake (disruption, p = 0.008) and reinforcing properties of cocaine (interaction, p = 0.01). On the other hand, pCRD males showed a SU-like phenotype with increased cocaine preference (interaction, p = 0.08), higher order food self-administration (disruption, p = 0.099) and cocaine reinforcement (interaction, p = 0.08). Furthermore, while male pCRD mice maintained goal-directed decision making, responding more on a reinforced lever, female pCRD mice did not, indicating habit formation. Together, these results suggest that male and female mice exposed to pCRD respond differently for rewarding outcomes. In order to determine whether these divergent behavioral outcomes are unique to reward I next measured anxiety-like behavior. As expected, preliminary results in the open field test and elevated plus maze (interaction, p = 0.005) paralleled reward-related behavior. Anhedonic-like female pCRD mice showed increased anxiety-like behavior and pCRD males showed decreased anxiety/increased risk-taking behavior. Gene expression rhythms were changed in the nucleus accumbens after pCRD.

Conclusions: These results suggest that pCRD may predispose individuals to distinct psychiatric disorders based on sex, mood disorders in females and SU disorders in males. By understanding how disrupted rhythms during pregnancy affect behavior in adulthood, we can develop novel therapeutic approaches for SU and mood disorders in adults.

Keywords: Reward, circadian rhythms, pregnancy, sex differences

Disclosure: Nothing to disclose.

P774. Single-Nucleus Transcriptome Analysis Reveals a Conserved Ventral Pallidal Neural Landscape Across Mammals

Lite Yang, Lisa Fang, Michelle Lynch, Vincent Costa, Vijay Samineni, Meaghan Creed*

Washington University School of Medicine, St. Louis, Missouri, United States

Background: The ventral pallidum (VP) plays a critical role in regulating emotional processing and motivated behaviors by orchestrating activity throughout the ventral striatum, habenula and midbrain. While contemporary work has begun to appreciate the functional diversity of VP with respect to its neurochemically-defined neural subpopulations, the molecular heterogeneity underlying its functional diversity remains incompletely understood.

Methods: We used snRNA-seq and multiplexed fluorescence in situ hybridization to define the transcriptional taxonomy of distinct cell types in the VP in mice (Mus musculus, n = 10M/10F), macaques (Macaca mulatta n = 3F, 1M) and baboons (Papio anubis n = 2M). The VP was dissected, single nuclei (44,518 in mice, 49,802 in macaque and 57,724 in baboon) were isolated and sequenced using 10X Genomics Chromium Single Cell Gene Expression 3’ V3.1 Assay. Libraires were sequenced on a NovaSeq6000 with 150 cycles each for Read1 and Read2. Raw sequencing data from individual libraires were processed and mapped to the reference genome and the gene-cell counts matrices from all snRNA-seq libraries of the same species were concatenated using cellranger-6.1.2.

Gene names in the primate datasets were converted to mouse genes using biomaRt, and datasets from all species were joint clustered using Seurat V3 to identify conserved and species-specific clusters.

Multiplexed fluorescent in situ hybridization was used to detect 24 key marker genes for each transcriptional cluster in mouse. Tissue was hybridized, and successive rounds of imaging was used to detect 3 mRNA probes. Images were registered using the DAPI signal, and mRNA for each marker gene was quantified within each DAPI-positive nucleus within the VP in QuPath (3.2). Custom python scripts were used to assign each VP neuron to a cluster based on molecular identity, and to visualize its location within the VP.

Finally, gene ontology analysis identified key neurotransmitter, neuropeptide and neurotransmitter-receptor genes enriched in each integrated cluster, identifying possible neuromodulatory targets for these molecular sub-populations.

Results: Molecularly-defined neuronal cell types displayed a high degree of transcriptional similarity between all three species, with the majority of molecularly-defined clusters conserved across all three species. Initial clustering resulted in 20 neuronal clusters in mice and macaque and 24 clusters in baboon. Each cluster was assigned a neurochemical identity based on expression of canonical marker genes for glutamatergic, gabaminergic and cholinergic populations: Vesicular glutamate transporter 1 and 2, (Slc17a6, Slc17a7 mouse; SLC17A6, SLC17A7 primate), glutamic acid decarboxylase; vesicular gaba transporter (Gad2, Slc32a1 mouse, GAD2, SLC32A1 primate), or choline acetyltransferase, high-affinity choline transporter (ChAT, Slc5a7 mouse; CHAT, SLC5A7 primate). Cross-species conserved clusters represented cholinergic, glutamatergic and gabaminergic populations previously reported to play unique roles in motivated behavior in rodent circuit dissection studies. Within the two primate species, unique dopaminoceptive and cholinergic subclusters were identified and conserved across both primate species but had no homolog in mice.

Conclusions: Collectively, our results provide a transcriptional, cross-species atlas for understanding the structure and function of the VP, and identifies key neuropeptides, neurotransmitters and neurotransmitter receptors that could be targeted within specific VP cell types. Our results also emphasize the well conserved nature of VP cell types across mammalian species and provide a critical resource for future studies.

Keywords: Ventral Pallidum, opioids, Non-human primates, mouse, RNA-sequencing

Disclosure: Nothing to disclose.

P775. Endocannabinoid Tuning of Behavioral Engagement via an Anterior Paraventricular Thalamus-Nucleus Accumbens Circuit

David Marcus*, Anthony English, Gunn Chun, Rachel Oomen, Emma Seth, Sabrina Hwang, Christian Pedersen, Avery Hunker, Azra Suko, Yulong Li, Nephi Stella, Larry Zweifel, Michael Bruchas

University of Washington, Seattle, Washington, United States

Background: A key clinical feature of addiction is compulsive engagement with rewarding stimuli, such as drugs of abuse, despite negative outcomes. However, we lack a detailed understanding of the neural mechanisms that regulate compulsive behavioral engagement. The Nucleus Accumbens (NAc) represents an integral functional component of the mesocorticolimbic pathway, canonically known as the “reward” pathway. The NAc receives dense projections from the Paraventricular Thalamus (PVT), and these projections have been shown to regulate the behavioral effects of opiate withdrawal, sucrose seeking/consumption, and behavioral responses to painful stimuli. The PVT is a highly heterogenous structure, and recent studies examining the PVT-NAc circuit have generated contradictory results, partially driven by a lack of genetic and anatomical targeting within the PVT.

Methods: All studies were conducted in equal numbers of male and female mice in accordance with NIH guidelines, approved by the IACUC at the University of Washington. No sex-dependent effects were observed. Below were the methods used:

• 1. RNAscope: 4 mice total, 2 anterior and 2 posterior PVT slices per mouse were used for quantification of mRNA transcript numbers and colocalization using HALO software

• 2. Ex vivo electrophysiology: Acute coronal slices containing the NAc were prepared from NTS-Cre x D1-tdTomato mice injected with DIO-ChR2-eYFP. Minimum of 4 mice per group, 2 slices per mouse, and a maximum of 4 cells per group per slice.

• 3. GCaMP fiber photometry: 10 NTS-Cre mice were injected with DIO-GCaMP6s in the aPVT and implanted with a unilateral fiberoptic in the NAc. 6 weeks after surgery, mice were trained in Pavlovian reward and Pavlovian fear conditioning.

• 4. In vivo optogenetics: Activation: 12 NTS-Cre mice were injected with DIO-ChR2 in the aPVT and 11 were injected with DIO-eYFP control virus. Inhibition: 7 NTS-Cre mice were injected with DIO-PPO and 6 were injected with DIO-eYFP control virus. Both groups were implanted with bilateral fiberoptics above the NAc. Mice were trained in Pavlovian reward and Pavlovian fear conditioning.

• 5. GRABeCB fiber photometry: 7 WT were injected with hSyn-GRABeCB in the aPVT and implanted with a unilateral fiberoptic above the NAc. Mice were trained in Pavlovian reward and Pavlovian fear conditioning.

• 6. CRISPR/Cas9 deletion of the CB1 receptor w/ fiber photometry. NTS-Cre mice were injected with a 5:3 mixture of DIO-saCas9-sgCNR1 and DIO-GCaMP6s in the aPVT and were implanted with a unilateral fiberoptic above the NAc. Mice were trained in Pavlovian reward and Pavlovian fear conditioning.

• 7. 1 photon imaging of NAcaPVT PENK neurons. PENK-Cre mice were injected with a transsynaptic AAV1-EF1a-DIO-FLP in the aPVT and an fDIO-GCaMP6s in the NAc, followed by implantation with a GRIN lens in the NAc. Mice were trained in Pavlovian reward and Pavlovian fear conditioning.

Results: Here, we identify the neuropeptide neurotensin (NTS) as a novel marker for a select population of anterior PVT neurons. Furthermore, we demonstrate that these NTS neurons send strong excitatory projections to the NAc, and that this input is biased toward Dopamine Receptor 1 (D1R) expressing medium-spiny neurons. This input bias is mediated through tonic endocannabinoid (eCB) suppression of excitatory input onto Dopamine Receptor 2 (D2R) expressing neurons. Using fiber photometry in mice expressing DIO-GCaMP in PVT NTS neurons, we observed that aPVT-NAc projections are inhibited during engagement in both sucrose seeking/consumption (R2 = 0.2553, p = 0.0002) and conditioned freezing behaviors (R2 = 0.1300, p = 0.0068), and activated upon behavioral disengagement (R2 = 0.3299, p < 0.0001). Ectopic activation and inhibition of this circuit using optogenetics was sufficient to attenuate (p = 0.0413) and prolong behavioral engagement (p = 0.0389), respectively. To assess the role of eCBs in modulating aPVT-NAc circuit activity and behavioral engagement, we used GRAB-eCB, a fluorescent sensor for eCB release, to demonstrate that eCBs are released in the NAc and bind to impinging aPVT terminals during engagement in reward-seeking behaviors (p = 0.0009). We augmented this approach by demonstrating that CRISPR/Cas9 deletion of the Cannabinoid 1 Receptor (CB1) from aPVT NTS neurons was sufficient to reduce engagement in reward-seeking behaviors (p = 0.0481) and concomitant aPVT-NAc terminal inhibition (p = 0.0136). Lastly, using an anterograde transsynaptic labeling technique combined with 1-photon miniscope imaging, we observed that D2R/PENK neural activity in the NAc precedes aPVT terminal inhibition, ostensibly linking NAc neural activity with resultant eCB release and retrograde action at aPVT terminals in the NAc.

Conclusions: Taken together, these results implicate a novel eCB mechanism for regulation of behavioral engagement through modulation of an aPVT-NAc circuit. Our data suggest that eCB inhibition of this excitatory circuit plays a critical role in regulating engagement with both rewarding and aversive stimuli. These data lay the groundwork for investigating how drugs of abuse elicit maladaptive changes in the innate function of this circuit to drive compulsive drug seeking behaviors.

Keywords: Endocannabinoids, Nucleus Accumbens, paraventricular thalamus, Reward seeking, aversion

Disclosure: Nothing to disclose.

P776. Sex-Specific Contrasting Role of BECLIN-1 Protein in Pain Hypersensitivity and Anxiety-Like Behaviors

Fariya Zaheer, Gabriel J. Levine, Ana L. Simal, Seyed R.F. Tabatabaei, Tami A. Martino, Giannina Descalzi*

University of Guelph, Guelph, Canada

Background: Chronic pain is a debilitative disease affecting 1 in 5 adults globally, and is a major risk factor for anxiety. Given the current dearth of available treatments for both individuals living with chronic pain and mental illnesses, there is a critical need for research into the molecular mechanisms involved in order to discover novel treatment targets. Cellular homeostasis is crucial for normal bodily functions and investigation at the cellular level may reveal a better understanding of the processes that occur leading to the development of chronic pain. A critical mechanism involved in homeostasis is the process of autophagy, which ensures appropriate disposal of damaged intracellular organelles to maintain its equilibrium. BECLIN-1 plays a major role in the process of macro-autophagy (herein autophagy) through its interaction with PI3K complexes and is responsible for the delivery of autophagosomes to lysosomes for intracellular degradation, promoting the restoration of cellular homeostasis. In this study, we used a transgenic mouse model to investigate the role of autophagy in chronic pain, and comorbid anxiety-like behaviors. We found contrasting roles for BECLIN-1 in the development of pain hypersensitivity and anxiety-like behaviors in a sex-dependent manner.

Methods: We exposed female and male adult mice to a model of chronic neuropathic pain, the spared nerve injury (SNI) model. Sham surgeries were used as control. We assessed mechanical and thermal pain thresholds at baseline (prior to injury), and up to 28 days, post injury. At 30 days, mice were assessed for anxiety-like behaviors using the elevated plus maze and open field tests.

Results: We found that male SNI mice with impaired BECLIN-1 function demonstrated heightened mechanical and thermal hypersensitivity compared to male wildtype SNI mice, while female SNI mice with impaired BECLIN-1 function demonstrated similar thresholds to the female wildtype SNI mice. We also found that disruptions of BECLIN-1 prevented SNI induced increases in anxiety-like behaviors male mice but did not affect SNI induced increases in anxiety-like behavior in female mice. Our data thus indicate that BECLIN-1 is differentially involved in the nociceptive and emotion components of chronic pain in male but not female mice.

Conclusions: Our data indicate a novel role for BECLIN-1 in neuropathic pain, and comorbid anxiety-like behaviors in mice. We found that deletion of Beclin-1 reduces nociceptive hypersensitivity whilst preventing pain-induced increases in anxiety-like behaviors. Notably, these effects were sex-dependent, where only males, but not females, showed BECLIN-1 mediated effects. Our data thus indicates that macroautophagy is differentially involved in nociception and anxiety, in male, but not female mice.

Keywords: chronic pain, anxiety, mouse models, autophagy

Disclosure: Nothing to disclose.

P777. Escalation of Cocaine Self-Administration Induces Sex-Dependent Dysregulation of Estrogen and Nicotinic Receptors in Key Brain Regions of Heterogeneous Stock Rats

Elizabeth Sneddon-Yepez*, Supakorn Chonwattanagul, Kokila Shankar, Sonja Plasil, Michaela Cullum-Doyle, Benjamin Sichel, Dyar Othman, Narayan Pokhrel, Molly Brennan, Lisa Maturin, Lieselot Carrette, Francesca Telese, Abraham Palmer, Olivier George

University of California - San Diego, La Jolla, California, United States

Background: In the past ten years, cocaine use among men and women has risen in the USA. Yet, women are more likely to progress from initial use to dependence. There is a clear consensus that sex hormones influence cocaine-related behaviors. Estrogens increase while progesterone and testosterone decrease cocaine-related behaviors. In addition, we previously identified the nicotinic acetylcholine receptor (nAChR) cluster as a risk factor for cocaine use in our genome-wide association study. However, the association between these mechanisms to cocaine behaviors is unknown. Heterogenous stock rats allow for the investigation of individual differences allowing for a translationally relevant and powerful model to investigate cocaine-related behaviors. This study aimed to identify the effects of a history of cocaine self-administration on sex steroids and nAChRs expression in the hippocampus, prelimbic/infralimbic cortices, and the ventral midbrain. We hypothesized that different histories of cocaine self-administration will lead to sex-dependent alterations in sex steroid and nAChR mRNA expression.

Methods: Male and female heterogeneous stock rats (n = 57) underwent intrajugular catheterization surgeries and operant tasks (short- and long-access self-administration, progressive ratio, and aversion-resistant responding). Rats were grouped into addiction-like groups (resistant vs. severe) based on an addiction index determined by behavioral measures. Punches of the hippocampus, prelimbic/infralimbic cortices, and ventral midbrain were collected for RT-qPCR. Genes of interest included, Esr1, Esr2, Ar, Pgr, Chrna3, Chrb4, Chrna5, and Chrna7. All experimental protocols in animal studies were approved by the Institutional Animal Care and Use Committee at UCSD and were conducted following the National Institutes of Health Guide for the Care and Use of Laboratory Animals. A 2-Way ANOVA was used with sex and group (addiction index) as the between-subjects factors. Post hoc Holm Sidak’s tests were used where appropriate.

Results: For Ar in the hippocampus, the interaction almost met the threshold of significance (F(2, 56) = 3.011, p = 0.057, η^2 = 0.09, 2-Way ANOVA). Post hoc Holm Sidak’s tests revealed Ar mRNA was upregulated in severe females vs. resistant females, (p = 0.026), naïve females, (p = 0.023), and severe males (p = 0.009). For Esr1 in the hippocampus, a 2-Way ANOVA found a main effect of sex (F(1, 46) = 7.189, p = 0.01, η^2 = 0.1), a main effect of group (F(2,46) = 3.973, p = 0.025, η^2 = 0.11), and the interaction almost met the threshold of significance(F(2,46) = 2.907, p = 0.065, η^2 = 0.09). Post hoc Holm Sidak’s tests revealed Esr1 mRNA was upregulated in resistant females vs. severe (p = 0.023) and naïve (p = 0.003) females and resistant males (p = 0.007). For Esr2 in the prelimbic/infralimbic cortices, there was a main effect of group (F(2,29) = 4.563, p = 0.019, η^2 = 0.2, 2-Way ANOVA). For Ar in the ventral midbrain, there was a main effect of sex (F(2, 37) = 8.322, p = 0.007, η^2 = 0.17, 2-Way ANOVA). For Esr1 in the ventral midbrain, there was a significant interaction (F(2, 28) = 3.522, p = 0.043, η^2 = 0.17) and a main effect of sex (F(1, 28) = 4.718, p = 0.039, η^2 = 0.11).). Post hoc Holm Sidak’s tests revealed Esr1 mRNA was upregulated in severe males vs. resistant males, (p = 0.036), naïve males, (p = 0.036), and severe females (p = 0.007). For Pgr in the ventral midbrain, a significant interaction was revealed (F(2, 29) = 3.807, p = 0.034, η^2 = 0.2, 2-Way ANOVA). Pgr was downregulated in severe females vs. males (p = 0.46, Holm Sidak’s). For Chrna3 in the ventral midbrain, there was a main effect of sex (F(1,31) = 5.076, p = 0.032, η^2 = 0.12, 2-Way ANOVA). For Chrna5 in the ventral midbrain, there was a main effect of sex (F(1, 30) = 4.482, p = 0.043, η^2 = 0.12, 2-Way ANOVA).

Conclusions: These data show sex differences in sex steroid receptors and nAChR mRNA expression in addiction-relevant brain regions following a history of cocaine self-administration. We found unique patterns of dysregulation in these regions. For the hippocampus, we see dysregulation in sex steroid receptors. While we see dysregulation in nAChR mRNA expression in the prelimbic/infralimbic cortices. Interestingly, there was dysregulation of both sex and nAChR mRNA expression in the ventral midbrain. These findings underscore the importance of considering sex-specific molecular changes in the development of cocaine use disorder and have identified specific targets we can further explore in future studies.

Keywords: cocaine sex differences, sex steroid hormone receptors, nicotinic acetylcholine receptors, cocaine self-administration

Disclosure: Nothing to disclose.

P778. Sex Differences in the Locus Coeruleus Drive Cocaine-Seeking Behavior During Extinction and Withdrawal

Jon Person, Eric Vallender, Amy Kohtz*

University of Mississippi Medical Center, Jackson, New Jersey, United States

Background: The inability to maintain abstinence is a trademark of substance use disorders (SUDs) yet effective maintenance therapies remain elusive. Furthermore, SUD treatment may be particularly complex in women; as psychological and biological responses to drugs of abuse differ in women compared to men. Several measures of cocaine use disorder are greater in women, and can be paralleled in female rodents, yet the biological mechanisms for these sex differences remain unclear. Thus, understanding circuit and molecular signatures that drive increased drug-seeking among females is critical to the development of effective SUDs therapies.

Methods: Herein, we tested sex differences in the role of locus coeruleus (LC) noradrenergic signaling to the dorsal hippocampus (dHPC) in driving operant cocaine memories cocaine-seeking persistence during extinction from self-administration (CSP) in male and female adult Sprague Dawley rats (n = 6-8/group). Rats were implanted with PRSx8-HA-hM4Di or hM3Dq DREADDs to the locus coeruleus and guide cannulae were inserted to the dHPC. On extinction day 1, rats were infused with CNO and tested for cocaine seeking behavior. In a separate group, male and female rats (n = 6-8/group) were similarly trained to self-administer cocaine, and euthanized naïve to cocaine, in 24 hr withdrawal, or in 24 hour withdrawal with cocaine-seeking. LC was fresh dissected and processed for RNA-Seq analyses.

Results: Inhibition of LC-NE signaling to the dHPC using DREADDs attenuated CSP in females only, whereas excitation of LC-NE signaling to the dHPC increased CSP in males only (F(1,24) = 26.28, p < 0.0001). RNA-Seq analysis observations revealed notable sex-specific genetic changes in noradrenergic signaling pathways of the LC to the dHPC in rats. Also, transcriptomic alterations were discernible in gene expression profiles among control, extinction, and withdrawal rats. These changes were modulated by sex, with both universal and sex-specific changes observed.

Conclusions: These studies show the importance of the LC in SUDs, shedding light on potential sex-specific responses during the distinct behavioral phases of the addiction cycle. Furthermore, these effects may be driven by sex differences in adrenergic tone between the LC and dHPC. Thus, the substantial sex differences in the retrieval and retention of cocaine memories may be driven in part by the locus coeruleus.

Keywords: cocaine sex differences, Locus coeruleus, Dorsal Hippocampus, DREADDs, Extinction and Reinstatement

Disclosure: Nothing to disclose.

P779. Decoding the Neural Representation of Value in Schizophrenia With Negative Symptoms

Reihaneh Forouhandehpour, Guillermo Horga, Clifford Cassidy*

Stony Brook University, Stony Brook, New York, United States

Background: Deficits in pleasure and motivation are a common and burdensome aspects of several neuropsychiatric conditions, including schizophrenia with negative symptoms. Negative symptoms are not effectively treated with standard pharmacotherapies and the mechanisms underlying them remain poorly understood. Here we designed a novel fMRI paradigm to test a leading theory that value representation in the brain is deficient in schizophrenia with negative symptoms.

Methods: We developed an engaging fMRI paradigm, the Movie Shopping Task (MST) to study the neural mechanisms of value-based decision-making in schizophrenia. This task incorporates both naturalistic stimuli fMRI and task-based fMRI. During each trial of the MST, participants watch a movie trailer (ranging between 25 and 90 seconds in length), then rate enjoyment of the trailer (experience of pleasure), bid money in an auction to win the movie (willingness-to-pay), and exert key-press effort to win it (willingness-to-work for reward). There are 90 trials over 10 runs divided into 2 sessions (total task duration is around 120 minutes). After each MRI session participants can win a DVD copy of the film advertised in a randomly selected trial based on the amount money they bid on that trial (using the Becker-Degroot-Marschak procedure to assess willingness-to-pay) and have another chance to win a DVD based on the number of key presses on a different randomly selected trial. A complete dataset was available from 19 healthy individuals and 22 schizophrenia patients with negative symptoms (item score on the Scale for the Assessment of Negative Symptoms ≥3). Behavioral data was analyzed with two-sample t-test. fMRI data was subjected to standard preprocessing with fMRIPrep, followed by first-level analysis in FSL to generate beta maps for each dollar bid value. A mask of value representation regions in the ventromedial prefrontal cortex and ventral striatum was applied and multivariate pattern analysis (MVPA) with support vector regression was performed on each participant’s data to predict financial decisions using 5-fold cross validation. For each participant, the Pearson correlation coefficient between predict and observed bid responses was retained and used in group-level 1-sample and 2-sample t-tests.

Results: Some schizophrenia patients made higher bids to win movies in the auction procedure (t36 = 2.51, p = 0.02). The groups did not differ on liking scores, effort scores, or the agreement between liking scores and bids. Brain activity patterns predicted willingness-to-pay in healthy individuals (t18 = 4.21, p = 0.0003, 1-sample t-test of correlation between predicted and true bids per participant) but not in schizophrenia patients (t21 = 1.01, p = 0.16). Predictions of bids were significantly worse in the schizophrenia group compared to the healthy group (t39 = 3.04, p = 0.004, 2-sample t-test).

Conclusions: These preliminary results suggest the presence of deficits in value representation in prefrontal cortex and striatum in schizophrenia but no deficit in the experience of pleasure. These results are consistent with long-established models of mechanisms underlying negative symptoms, and could provide insight into the origin of social and occupational deficits in schizophrenia. Future work on this dataset will further investigate whether deficits in value-based decision making are independent of hedonic experience, which may be intact in schizophrenia.

Keywords: Schizophrenia (SCZ), Negative Symptoms, fMRI, Value-based Decision-Making, MVPA

Disclosure: Nothing to disclose.

P780. Synaptic Density in Early Stages of Psychosis and Clinical High Risk: A Link With Negative Symptoms

M. Belen Blasco*, Kankana N Aji, Christian Ramos Jiménez, Ilana Ruth Leppert, Christine Lucas Tardif, Johan Cohen, Pablo Rusjan, Romina Mizrahi

McGill University, Montreal, Canada

Background: Synaptic dysfunction is involved in schizophrenia pathophysiology. However, whether in vivo synaptic density is reduced in early stages of psychosis, including its high-risk states remains unclear.

We aim to investigate whether synaptic density (SV2A BPND) is reduced in first-episode psychosis (FEP) and in clinical high-risk (CHR). Moreover, we aim to investigate the effect of cannabis use on synaptic density and examine its relationship with psychotic symptoms and grey matter microstructure across groups.

Methods: This is a cross-sectional study performed in a tertiary care psychiatric hospital from July 2021 to October 2023. The study analyzed a total of 49 subjects, including 16 antipsychotic-free/minimally exposed FEP, 12 CHR and 16 healthy controls with clean urine drug screen (except cannabis).

Synaptic density (SV2A BPND) was quantified with dynamic 90 minutes [18F]SynVesT-1 PET scans across prioritized brain regions of interest (ROIs) delineated in individual Magnetic Resonance Images (MRIs). Cannabis use was confirmed with urine drug screens. Grey matter microstructure was assessed using diffusion-weighted MRI to estimate neurite density.

Results: 16 FEP patients (mean[SD] age, 26.1[4.6] years; 7 females, 17 CHR (mean[SD] age, 21.2[3.5] years; 9 females), and 16 healthy controls (mean[SD] age, 23.4[3.6] years; 9 females) were studied. Synaptic density was significantly different between groups (F(2,273) = 4.02, p = 0.02, Cohen’s F = 0.17; ROI: F(5,273) = 360.18, p < 0.01, Cohen’s F = 2.55) with a group by ROI interaction (F(10,273) = 2.67, p < 0.01 Cohen’s F = 0.32). Synaptic density was lower in cannabis users (F(1,272) = 5.31, p = 0.02 Cohen’s F = 0.14). Lower synaptic density (SV2A BPND) across groups was associated with more negative symptoms (PANSS Negative Scores: F(1,81) = 4.31, p = 0.04, Cohen’s F = 0.23; SOPS negative scores: F(1,90) = 4.12, p = 0.04, Cohen’s F = 0.21). SV2A BPND was significantly associated to neurite density index (NDI) (F(1,138) = 6.76, p = 0.01, Cohen’s F = 0.22).

Conclusions: Synaptic density reductions are present during early stages of psychosis and its risk states and associated with negative symptoms. The implications of SV2A for negative symptoms in psychosis and CHR warrant further investigation. Future studies should investigate the impact of cannabis use on synaptic density in CHR longitudinally.

Keywords: Schizophrenia, synaptic aberrations, Schizophrenia, negative symptoms, Neurodevelopment, Cannabis

Disclosure: Nothing to disclose.

P781. A Neuron-Glia Axis of Transcriptomic Dysregulation in the Subgenual Anterior Cingulate Cortex in Schizophrenia

Rachel Smith*, Agoston Mihalik, Nirmala Akula, Pavan Auluck, Stefano Marenco, Petra E. Vértes, Armin Raznahan, Francis J. McMahon

National Institute of Mental Health, NIH, Bethesda, Maryland, United States

Background: Classic approaches to identifying changes in transcriptomic expression patterns in post-mortem brain tissue from donors with psychiatric disorders do not align well with risk genes identified through exome-wide and genome-wide association studies (GWAS). These analyses are challenged by genetic complexity and polygenicity, regional and cell-type variability, and environmental confounds such as recreational and medicinal drug use. Here, we characterize transcriptomic differences in the relatively under-studied subgenual anterior cingulate cortex (sgACC) in bipolar disorder (BD), major depression (MDD) and schizophrenia (SCZ), using a novel multivariate regression technique that accounts for complex interactions among features, including environmental confounds.

Methods: We investigated the association between post-mortem sgACC transcriptomic expression patterns and neuropsychiatric disorder diagnosis using traditional univariate analyses (differential gene expression analysis, DGE, and weighted gene co-expression network analysis, WGCNA) and canonical correlation analysis (CCA). CCA is a multivariate regression technique that quantifies the linear association between two multivariate data matrices (in this case, gene expression-by-sample and covariates-by-sample). In this work, we consider both regularized CCA (RCCA), which treats all features as independent, and group RCCA (GRCCA), which accounts for underlying data structure using a grouping vector (here, gene co-expression module determined using WGCNA).

Bulk RNA sequencing was performed on 185 post-mortem sgACC samples obtained from individuals with BD (n = 35), MDD (n = 51), SCZ (n = 44), and no known psychiatric diagnosis (n = 55) (sex at birth: male = 117, female = 68; self-reported race: Black or African American = 68, Caucasian = 117). Recreational drug and medication use across 17 known drugs was reported for each sample; we used multiple correspondence analysis to identify dimensions of drug use and included the first 8 dimensions ( > 75% variance explained) as covariates in each model.

Diagnosis-associated gene lists identified using each analysis method were characterized functionally with gene ontology (GO) pathway analysis and gene set enrichment analysis (GSEA) across each GO ontology. We benchmarked the genes identified in each analysis against genes identified in the literature, including DEGs from a previous study in this dataset (Akula et al, 2021), a list of consensus DEGs (Merikangas et al, 2022), common variants (Trubetskoy et al, 2022), and rare variants (Singh et al, 2022).

Finally, to consider the role of alternative splicing, we ran the GRCCA analysis on transcript expression data, and provide the results as a resource.

Results: GRCCA identified a linear combination of gene expression across samples that was significantly associated with SCZ diagnosis, but not with MDD, BD, or variation due to recreational or medicinal drug use. Genes that were significantly associated with SCZ (abs(z-score) > 2; FDR < 0.05; n = 2026) demonstrated strong alignment with known risk genes in SCZ. These genes were enriched for SCZ common variant-associated genes (GSEA p = 0.0007), especially among upregulated genes (normalized enrichment score, NES = 1.43). Significant genes also showed overlap with SCZ rare variant-associated genes (hypergeometric p = 0.06). Furthermore, GRCCA genes exhibited substantial overlap with DEGs from this dataset (Akula et al 2021; hypergeometric p < < 0.001), but not with a consensus list of SCZ DEGs (p = 0.53).

Overall, significant GRCCA genes were enriched for glia-related immune response, cilium organization and motility, and GTPase activity pathways. GSEA revealed that the immune and ciliary enrichments were strongest in genes that were downregulated in SCZ, specifically lymphocyte activation and proliferation. Genes that were upregulated in SCZ were enriched for neuronal processes, especially synaptic signaling, and ATP-dependent activity as it pertains to DNA and RNA (all FDR < 0.05).

In contrast, genes identified through DGE (n = 1747; p < 0.05) and WGCNA (n = 2533 across 3 modules) did not overlap with genes implicated by common or rare variants (hypergeometric p > > 0.05 in all cases). Significant RCCA genes (n = 1194) did overlap with genes implicate by association studies (common p = 0.0007 and rare p = 0.016); however, these genes did not exhibit strong functional enrichments (all FDR > 0.05).

Conclusions: GRCCA identified a SCZ-specific pattern of gene expression that aligns well with known SCZ risk genes and implicates a distinct axis of glia-neuron functional pathways. This multivariate analysis method uniquely linked dysregulation in transcriptomic expression to genetic risk for SCZ, providing a basis for aligning across genomic levels to describe a comprehensive, mechanistic framework for molecular dysregulation in SCZ.

Keywords: Schizophrenia, Bipolar disorder, Depression, Transcriptomics, human post-mortem brain, subgenual anterior cingulate cortex, Canonical Correlation Analysis (CCA)

Disclosure: Nothing to disclose.

P782. Clozapine as a Long-Term Therapeutic Choice: Longitudinal Analysis of Schizophrenia Symptoms in a Naturalistic Setting

Rachel Scheinberg, Zhirui Fu, Laura Scott, Krista Baker, Arlene Cuerdo, Lillian Zhong, Chloe Bethany, Malaka Harper, Leslie Nucifora, Allison Brandt, Russell Margolis, Gayane Yenokyan, Frederick Nucifora*

Johns Hopkins University School of Medicine, Baltimore, Maryland, United States

Background: Treatment resistant schizophrenia (TRS) occurs in up to 35% of patients and leads to a substantial burden on patients, families and society. The only medication approved by the FDA for TRS is clozapine. Unfortunately, clozapine is dramatically underutilized due to rare but potentially severe side effects, and FDA REMS requirements that result in the need for regular blood draws and enrollment of patients, prescribers, and pharmacies in an associated registry. Increasing clozapine use could decrease symptoms and improve the quality of life for patients with TRS. However, while the value of clozapine for TRS has been well-established, the pattern of improvement on clozapine is less-well established, particularly in naturalistic clinical settings. Establishing this pattern could strengthen the case for clozapine utilization and provide practical guidance for prescribers. The objective of the current study is therefore to assess the symptom trajectories on clozapine using longitudinal data from a clinical setting with the goal of determining the time to improvement, the duration of improvement, the extent of improvement, and the particular symptom domains that improve.

Methods: The study population consisted of patients treated with clozapine at a specialty clozapine clinic of the Johns Hopkins Bayview Community Psychiatry Program. Each patient is assigned to a psychiatrist and a therapist, one therapist oversees clozapine REMS issues, and site nurses provide some assistance with managing side effects. Approximately 100 patients are treated in the clinic at any one point in time. This study focused on 28 patients started on clozapine after enrolling in the clinic (New Starts). These patients had been designated treatment resistant based on standard criteria, and were at least moderately ill as defined by Positive and Negative Symptoms Scale (PANSS) scores of > 58 or by overall clinical presentation. PANSS symptoms were assessed approximately every 6 months while patients were enrolled in the clinic and taking clozapine, including prior to the onset of clozapine treatment. Patients were considered clozapine-responders if PANSS score declined to ≤58 or improved by 20%.

Results: PANSS scores of patients started on clozapine significantly improved during the first year of treatment, with significant improvements in both positive and general symptoms and a trend for improvements in negative symptoms. 29.4% of New Starts had a PANSS score of ≤58 by 12 months on clozapine. Of the 70.6% of patients with a PANSS score above 58, 50% improved to a PANSS score of ≤58 after an average of 22.6 months on clozapine. Overall, 64.7% of New Starts improved to a mild level of symptoms or better during treatment with clozapine. At 12 months after clozapine initiation, 70.6% of New Starts had not improved by 20%. However, 33% of those patients reached > 20% improvement by 28.4 months. Additionally, 20% of New Start patients with a PANSS score of ≤58 at 12 months improved by 20% after 12 months on clozapine.

Conclusions: These results suggest that treatment with clozapine is associated with improvement of symptoms across positive, negative and general domains of the PANSS in a naturalistic setting. Notably, in some individuals, improvement may take a year or more, justifying prolonged treatment with clozapine, particularly in the absence of established alternatives. The results further support the use of clozapine for patients with TRS.

Keywords: schizophrenia, Clozapine, Treatment resistant schizophrenia

Disclosure: Nothing to disclose.

P783. Intolerance of Deliberation and Suspicious Decision-Making in Psychosis and Other Persecuted Individuals

Angus MacDonald*, Rebecca Kazinka, Christopher McCain

University of Minnesota, Minneapolis, Minnesota, United States

Background: Work on delusions has focused on “cold” cognitive processes, or the mechanisms underlying delusion formation, such as a tendency to use less information when making decisions known as a jumping to conclusions (JTC) bias (Garety and Freeman, 1999). However, such theories largely fail to incorporate the content of positive symptoms. Importantly, this content is mainly “hot” -- emotion-laden and largely self-referential. For example, work using the Minnesota Trust Game (MTG) has shown that persecuted patients and community members are more likely to ascribe malevolent motivations, not simply selfish motivations, to anonymous partners. This work begins to bridge this gap, focusing on persecutory ideation as a model symptom and spite sensitivity -- the concern about a partner’s malevolence -- as a dispositional factor.

Methods: 48 psychosis patients and 62 community-dwelling twins (including 23 persecution-discordant MZ pairs) completed the Minnesota Trust Game (MTG) during MRI scanning. which is an adaptation of other trust games in which in one condition the partner is incentivized to be untrustworthy (rational mistrust condition), while in another condition, the partner would also lose money if the player does and therefore is disincentivized to be untrustworthy (suspiciousness condition). MTG decisions were fit to an econometric model to quantify spite sensitivity. Participants also completed the Boxes task variant of the beads task (Moritz, et al., 2017) that measured JTC. Participants are shown a sequence of balls and asked when they are ready to guess which of two options is more likely to be drawn next. Primary outcome measures are the number of balls drawn before making a decision (draws) and the number of times they switch which ball they believe to be more likely as they are completing the task (switches).

Results: 1) In both patients and community twins, MTG analyses replicated the link between decisions in the suspiciousness and persecutory ideation, and computational analyses related these differences to the expectation of a spiteful partner. 2) Confirmatory MTG imaging demonstrated that dysconnectivity between the frontoparietal/central executive network connectivity and the valuation/salience networks (lateral OFC/anterior insula, in particular) during the decision-making phase of the suspiciousness condition, but not the rational mistrust condition. JTC biases, as reflected in fewer draws and fewer switches, predicted 3) patients and community twins with more persecutory ideation, and 4) those with more suspicious decisions and greater expectations of a spiteful partner. Furthermore, 5) reduced switching predicted reduced frontoparietal-valuation/salience dysconnectivity during the suspiciousness condition of MTG.

Conclusions: These results are broadly consistent with theories that link persecutory ideation to JTC bias. They advance our understanding of JTC by suggesting that dysregulation or discomfort with the deliberative processes required to interpret others’ intentions lead to the adoption of easier decision-making systems, such as habit or instinct, that opt out of potential social threats. If so, it remains to be determined whether spite sensitivity is a predisposing factor for opting out of social threats, or whether it is a post-hoc rationalization. Either way, these findings provide strategies and metrics for CBT for psychosis and other interventions.

Keywords: Schizophrenia Spectrum Illness, Paranoia, jumping to conclusions

Disclosure: Nothing to disclose.

P784. Baseline and 8-week Cognitive Assessment of Individuals at Clinical High Risk (CHR) for Psychosis in the AMP SCZ Initiative: A Model for Future Interventions?

William Stone*, Walid Yassin, Kelly Allott, Cassandra Wannan, Setari Parsa, Kathryn E. Lewandowski, Matcheri Keshavan, Rene Kahn, Patrick McGorry, Carrie Bearden, Barnaby Nelson, Martha Shenton, Scott Woods, Ruben Gur, Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ)

Harvard Medical School/BIDMC, Boston, Massachusetts, United States

Background: Cognitive impairments are core features of schizophrenia that also occur at higher rates in individuals at clinical high risk (CHR) for psychosis compared to community controls (CON). These impairments uniquely contribute to predictive models of psychosis transition, making cognitive assessment critical in the Accelerating Medicines Partnership® Program for Schizophrenia (AMP SCZ). Here we present preliminary data on baseline cognitive performances in CHR and CON and changes at an 8-week follow-up to assess the suitability of the battery for the current observational study and for future efforts to prevent CHR transitions to psychosis.

Methods: The AMP SCZ cognitive battery integrates tests of general intellect (Wechsler Abbreviated Scale of Intelligence II; WASI-II) and specific cognitive domains, including eight tests from the Penn Computerized Neurocognitive Battery (PennCNB). These tests assess verbal learning (Short Penn List Learning Test; SPLLT), visual learning (Short Visual Object Learning Test; SVOLT); attention (Short Penn Continuous Performance Test; SPCPT), emotion recognition (Penn Emotion Recognition Test; ER40); working memory (Short Fractal N-Back Test; SFNB2), processing speed (Digit-Symbol Test; DIGSYM), and motor speed (Short Computerized Finger Tapping Test, dominant hand; (SCTAP-DH). A composite measure was also constructed. Baseline group differences and reliability of measures at an 8-week follow-up were assessed, as were relationships to age, gender, participant education, parental education and income.

Results: Individuals with CHR (n = 768) performed significantly lower than CON (n = 98) in WASI II IQ (p < 0.05), ER40 (p < 0.05), DIGSYM (p < 0.001), SCTAP-DH (p < 0.01) and the composite score (p < 0.001) at baseline. Longitudinally, CHR showed significant but modest performance increases at 8 weeks in SPCPT (p < 0.001), ER40 (p < 0.001), SFNB2 (p < 0.05), DIGSYM (p < 0.01) and the composite score (p < 0.01). By contrast, performances in SCTAP-DH (p < 0.001), SVOLT (p < 0.001), and SPLLT (p < 0.001) declined. Age showed a positive association with SCTAP-DH (p < 0.001), and SPCPT (p < 0.001), and a negative relationship with SPLLT (p < 0.001), DIGSYM (p < 0.01), and SVOLT, (p < 0.001). Male gender was related to higher scores on SCTAP-DH (p < 0.01). Education of the participants indicated a positive relationship with SVOLT (p < 0.01), SCTAP-DH (p < 0.05), SPCPT (p < 0.05), DIGSYM (p < 0.001), and SPLLT (p < 0.001), while education of the parents was positively associated with ER40 (p < 0.05), SVOLT (p < 0.05), SCTAP-DH (p < 0.05), and SFNB2 (p < 0.05). Income had no association with any of the cognitive variables. Reliability analysis showed good reliability for DIGSYM (ICC = 0.89), SPCPT (ICC = 0.79), ER40 (ICC = 0.75), and the composite score (ICC = 0.78). Moderate reliability was found for SPCPT (ICC = 0.69) and SCTAP-DH (ICC = 0.50), with lower reliability for SFNB (ICC = 0.47) and SVOLT (ICC = 0.30).

Conclusions: Initial observations of baseline and 8-week data showed that CHR performed lower than CON in multiple measures, consistent with expectation. Notably, CHR showed only modest practice effects on several PennCNB measures re-administered at 8 weeks, and prominent declines in performance associated with alternate forms of the verbal and visual learning tests. These performances suggest there is ample room for cognitive improvement with potential therapeutic interventions. These findings suggest further that demographic factors such as participant education and parental education were associated with positive moderating effects on learning tests showing lower performances at 8 weeks than at baseline.

Keywords: Schizophrenia (SCZ), Clinical High Risk State for Psychosis, Cognition

Disclosure: Nothing to disclose.

P785. Stress Gene Expression and Social Cognition and Motivation in Schizophrenia

Yvonne Yang*, Jonathan K. Wynn, Steve Cole, Michael F. Green

VA Greater Los Angeles Healthcare System, UCLA, Los Angeles, California, United States

Background: Individuals with schizophrenia are often socially isolated and show impairments on tests of social cognition and social motivation. Isolation and loneliness are a global concern, but how do they affect patients with schizophrenia compared to socially isolated and socially connected individuals from the community? The conserved transcriptional response to adversity (CTRA) is a well-characterized stress gene expression program that is activated in conditions of social isolation and loneliness in healthy individuals, and is associated with downstream negative health effects. We evaluated associations among the CTRA and social variables including social cognition and social motivation in individuals with schizophrenia, socially connected healthy volunteers, and socially isolated healthy volunteers, to evaluate physiological correlates of social stress across the three groups.

Methods: Socially connected and self-reported isolated healthy volunteers (n = 36, 66, respectively) were recruited using two different online advertisements. For the patient group (n = 47), diagnosis of schizophrenia was confirmed with the Structured Clinical interview for DSM-5 (SCID-5). Subjects were between age 18 and 65 and excluded if they met criteria for moderate or severe substance use disorder in the prior three months. Assessments for isolation, loneliness, social motivation, and social cognition were conducted by trained staff in person. Blood samples for mRNA CTRA analysis were obtained and stored at -80 °C in PAXGene RNA tubes; RNA sequencing was batch processed after all samples were collected. The CTRA was defined as a composite score derived from expression of 53 genes (19 pro-inflammatory, 34 antiviral), as per previously published work. Linear mixed models were performed to characterize relationships among CTRA and social variables.

Results: Preliminary analyses revealed between-group differences, within group differences, and interaction effects for CTRA expression and social variables. For group comparisons, compared to socially-connected healthy volunteers, the patient group was found to have lower inflammatory gene expression and a trend towards higher antiviral gene expression. For within group comparisons, in the patient group, CTRA expression was unchanged with higher loneliness, and lower with higher isolation. In socially-connected healthy individuals, CTRA was elevated with both loneliness and isolation, consistent with previous work on the CTRA. Socially-isolated healthy volunteers presented an intermediate phenotype, with CTRA higher in the context of isolation and lower in the context of loneliness. With regards to interactions with social motivation and cognition, CTRA activation in patients was intermediate between socially-connected community members and socially isolated community members. Socially-connected community members exhibited higher CTRA levels with social avoidance and lower CTRA with social approach motivation, as might be predicted. Individuals with psychosis exhibited the same pattern but to a lesser degree, and socially isolated healthy volunteers showed little change with relation to social avoidance and social approach. For social cognition, socially-connected individuals from the community showed higher CTRA with poorer social cognition performance, socially-isolated individuals showed no change in CTRA, and individuals with schizophrenia had more CTRA activation with greater social cognition performance.

Conclusions: In contrast to socially-connected individuals, patients with schizophrenia showed lower physiological stress when more socially isolated. Higher social avoidance and better social cognition were also associated with elevated CTRA expression in patients. These findings raise the possibility that, for patients with schizophrenia, increased social activity could be associated with higher levels of stress, as opposed to healthy individuals, in whom social activity is associated with decreased physiological stress. These results call into question prior assumptions that attempting to increase social interactions, a key functional outcome, has a universally positive impact on patients. Next steps could include examining the relationships among negative symptoms, social variables, and the CTRA in schizophrenia.

Keywords: Schizophrenia (SCZ), social stress, social isolation, genetics, immune mechanisms

Disclosure: Nothing to disclose.

P786. Thalamic D2/3 Dopamine Receptor Systems and Antipsychotic Treatment Resistance Genetic Variation

Daniel Eisenberg*, Rachael Blackman, Maria Tietcheu, Philip Kohn, Jasmin Bettina, Bhaskar Kolachana, Michael Gregory, Karen Berman

Clinical and Translational Neuroscience Branch, National Institute of Mental Health, Bethesda, Maryland, United States

Background: Pharmacologic blockade of the dopamine D2 receptor is a common feature of medication treatments for psychosis. However, for many individuals with schizophrenia spectrum illness, mainstay antipsychotic drugs do not provide adequate relief from symptoms, a clinical dilemma for which clozapine, a unique atypical agent, has utility. Such treatment resistance remains a major source of morbidity and mortality among those with schizophrenia spectrum illness, and its molecular and neurochemical underpinnings are incompletely understood. Positron emission tomography has allowed for in vivo characterization of dopamine systems neurobiology in schizophrenia, and recent work using high-affinity dopamine D2/3 receptor ligands has suggested that lower D2/3 receptor availability in the thalamus, a complex structure housing an array of functionally diverse nuclei, may be an important feature of this illness, though whether this phenotype is at all linked to the neurobiology of treatment response is uncertain. Genetic association studies of treatment resistance in schizophrenia have proposed a broad network of single nucleotide polymorphisms that might differentially contribute to illness in people with treatment resistance though its biological mechanistic throughlines continue to require investigation. Whether cumulative genetic burden at these markers has implications for the distribution of in vivo D2/3 dopamine receptor availability across the thalamus is not known.

Methods: One-hundred and four healthy people without psychiatric illness (mean age 38 + /-11 years, 47 women) were studied at the NIH Clinical Center in Bethesda, Maryland, where they provided blood for genotyping and had [18F]-fallypride PET scans to measure dopamine D2/3 receptor availability. For each person, genotyping using genome-wide Illumina SNP chips was performed, and polygenic scores representing cumulative genetic risk burden were generated for treatment-resistant schizophrenia based on genome-wide association study literature summary statistics. For neuroimaging, a 4-hour abstinence for caffeine and nicotine were required prior to tracer injection. Each individual’s separately-collected T1-weighted anatomical MRI scan was segmented for a cerebellar reference region and was co-registered to each participant’s native space, attenuation-corrected, and motion-corrected PET data. After extraction of the reference region’s time activity curve, PET data were normalized to MNI-space using ANTS software. PMOD software was used to implement simplified tissue reference modeling to estimate BPND, a measure of D2/3 receptor availability. Principal components analysis implemented with AFNI software was applied to voxelwise calculations of tracer binding potentials across the thalamus, and resultant components were tested for association with polygenic scores for treatment-resistant schizophrenia employing general linear models that included covariates for age, sex, and ancestry associated genetic principal components.

Results: Greater polygenic risk burden for treatment resistant schizophrenia was associated with patterns of [18F]-fallypride tracer binding suggesting both widespread relative reductions in receptor availability (PC1, p < 0.003) as well as a centroposterior-anterior gradient (PC2, p < 0.05).

Conclusions: In healthy individuals not exposed to D2 modulating treatments, D2/3 receptor availability across the thalamus may be influenced by the genetic liability to treatment resistant schizophrenia. Thus, at least cumulatively, common genetic determinants of treatment resistance are not independent of dopaminergic systems biology, though whether these associations represent downstream or more fundamental mechanisms requires further investigation. To the extent that the identified associated spatial patterns within the thalamus are relevant to particular thalamo-corticostriatal circuits implicated in models of schizophrenia, additional work to determine network-level functional sequalae of this variability may be valuable. Future longitudinal studies of antipsychotic naïve individuals with schizophrenia spectrum illness may ultimately be critical to provide a clinical context for these observations.

Keywords: D2 dopamine receptor, Schizophrenia, Genetics

Disclosure: Nothing to disclose.

P787. Biomacromolcular Atlas of the Dorsal Anterior Cingulate Cortex Reveals Dysregulated Circadian Rhythm Function and Metabolism in Schizophrenia

Andrew DeMarco*, Ryan Salisbury, Kevin Xu, Jordan Gilardi, Akayla Lewin, Lambertus Klei, Anastasia Yocum, Steven Mullet, Robert Sweet, David Lewis, Berine Delvin, Stacy Gelhaus, Matthew MacDonald

University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, United States

Background: Next generation sequencing has nominated specific cell types, states, and signaling pathways in schizophrenia etiology, including multiple metabolic processes, neuro-inflammation pathways, and kinases with critical roles in synaptic signaling cascades. However, the specific mechanisms by which these signaling pathways and biochemical processes interact to drive synapse and circuit pathology in schizophrenia are difficult to predict from mRNA and protein expression data, as they weave an interconnected and dynamic biomacromolecular matrix. We posited that mapping and integrating multiple tranches of biomacromolecules would illuminate the interaction dynamics across macromolecular subtypes to provide a more comprehensive neurophysiological window into schizophrenia pathology. To accomplish this, we characterized the proteome, phosphoproteome, lipidome, and metabolome in dorsal postmortem anterior cingular cortex (dACC) grey mater from 56 schizophrenia patients and matched controls to provide novel insights into the disease.

Methods: Postmortem dACC grey matter was dissected from 56 schizophrenia subjects from the University Common Mind cohort, each paired with an unaffected control, matched for sex, age, and postmortem interval (PMI). The tissue was homogenized in S-Trap buffer (5% SDS, 50 mM TEAB) and digested on S-Traps. Samples were randomized into 8 groups containing 14 samples and two pooled controls and labeled with TMTpro. Phosphopeptides were enriched using an AssayMAP (Agilent), and both fractions were analyzed by LC-MS/MS with additional offline fractionation, facilitating deep protein and phosphopeptide quantification. Lipids and metabolites were extracted by methanol chloroform extraction. Untargeted quantification of both classes was performed by LC-MS/MS. Peptides and phosphopeptides were identified and quantified in Proteome Discoverer (v2.5), and lipids and small molecules were identified and quantified in Lipid Search (v5.1) and Compound Discoverer (v3.3). Data from each tranche were individually imputed using a variable autoencoder, and case-control statistics were performed with mixed multi-level modeling accounting for clinical covariates using in-house R-scripts. Enrichment analyses were conducted on altered features in each tranche (q < 0.05) with KEGG and Reactome databases, Motif-enrichment analysis, Lipid Ontology, and Metabolanalyist. Multiomic analyses were performed in KEGG using OmicsNet.

Results: Following quality control, statistical analyses were performed on 4,282 proteins, 11,177 phosphorylation sites mapping to 3,631 proteins, 8,347 lipid features, and 1,472 metabolite features. In line with prior studies, the 389 upregulated proteins were enriched for complement cascade (p = 2.6E-7), Toll-like receptor signaling (p = 1.3E-3), and interleukins 4 and 13 signaling (p = 1.3E-5), the 656 downregulated proteins were enriched for mitochondria translation (p = 2.2E-32) and mitochondrial proteostasis (p = 6.7E-8). Proteins with increased phosphorylation (n = 1371): guanosine triphosphatase signaling molecules (RAC1 p = 1.1E-5 and CDC42 p = 1.1E-4). Down-regulated phosphoproteins (n = 1888): insulin signaling (p = 8.7E-3). Proline-directed basophilic kinases (p = 3.0E-50) were predicted to be hyperactive, while acidophilic kinases (p = 1E-36) were hypoactive. Sphingolipids (p = 1.9E-13) and ceramides (p = 5.2E-21) were the two major classes of the 135 upregulated lipid features, while fatty acids comprised most of the 336 down-regulated lipid features (p = 3.9E-9). The 624 upregulated metabolite features carboxylic acid derivatives (p = 6.4E-28) and pyrimidine nucleosides (p = 3.0E-4). 458 downregulated metabolite features: alanine/aspartate/glutamate metabolism (p = 9.3E-7) and histidine metabolism (p = 1.1E-3). Multiomics analysis utilizing all dysregulated features mapped in KEGG from all tranches yielded multiple hits to metabolism, including alanine, aspartate and glutamate metabolism (p = 5.5E-10), glycine, serine and threonine metabolism (9.1E-22) and glycerophospholipid metabolism (p = 1.8E-8)), as well as circadian rhythm (p = 1.1E-3) cAMP(p = 1.1E-3) and AMPK (p = 2.8E-3) signaling pathways.

Conclusions: Circadian rhythm and expression of circadian-related genes are inextricably linked with glutamate, aspartate metabolism, and fatty acid homeostasis. Interestingly, glutamate and aspartate activate key receptors in circadian rhythm signaling, which involves cAMP and AMPK signaling inputs. Alterations in normal sleep patterns are clinically associated with schizophrenia. Our data lead us towards one of two models. The first suggests that alterations in metabolic homeostasis in schizophrenia patients may be the primary driver of circadian dysregulation affecting cAMP and AMPK signaling. Alternatively, dysregulated cAMP and AMPK could be upstream of the loss of metabolic homeostasis, which in turn leads to circadian dysfunction. Our findings merit further investigation of kinase and metabolic functions in a circadian paradigm to define key biological regulatory nodes that could contribute to disease symptoms and serve as targets for novel therapeutics.

Keywords: systems neurobiology, Schizophrenia (SCZ), signal transduction, Postmortem Brain Tissue, liquid chromatography/mass spectrometry

Disclosure: Nothing to disclose.

P788. AMP SCZ Observational Study: Severity of Illness at Baseline and Over Two Months Follow-Up in Clinical High Risk for Psychosis as a Function of Ascertainment Method

Catalina Mourgues*, Nora Penzel, Arielle Ered, Alyssa Bathery, Vijay Mittal, Rene Kahn, Patrick McGorry, John Kane, Carrie Bearden, Martha Shenton, Monica Calkins, Barnaby Nelson, Alison Yung, Ofer Pasternak, Jean Addington, Scott Woods, Accelerating Medicines Partnership® Schizophrenia AMP® SCZ

Yale University School of Medicine, New Haven, Connecticut, United States

Background: Drug development for early intervention in psychosis depends on the ascertainment of patients with the Clinical High-Risk (CHR) for Psychosis syndrome and on assessing severity of illness over time. Two interviews, the Structured Interview for Psychosis-risk Syndromes (SIPS) and the Comprehensive Assessment of At-Risk Mental States (CAARMS), have been extensively used for both purposes. While these tools both evaluate attenuated positive symptoms, they differ in their approach, potentially leading to distinct profiles at baseline and variations over time. The extent of these differences remains unclear, as few studies have conducted both SIPS and CAARMS interviews on the same participants. The Positive Symptoms and Diagnostic Criteria for the CAARMS Harmonized with the SIPS (PSYCHS, https://doi.org/10.1111/eip.13457) offers a unified approach to CHR assessment, providing a standardized severity scoring system compatible with both CAARMS and SIPS frameworks. This study explores positive symptom and other severity measure scores at baseline and over two-month follow-up from both vantages.

Methods: The ongoing Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ) Observational Study (https://doi.org/10.1093/schbul/sbae011) second public data release includes complete longitudinal data from baseline (BL) to month 2 (M2) for 150 CHR participants. Participants were divided into two overlapping groups: Group 1 (G1, n = 150, all participants in this subsample) consisted of individuals meeting CAARMS CHR criteria at the screening visit, and Group 2 (G2, n = 68) comprised individuals who fulfilled SIPS CHR progression criteria. A battery of clinical assessments was administered, including the PSYCHS for attenuated positive symptoms, Brief Psychiatric Rating Scale (BPRS) for positive and negative symptoms, Negative Symptom Inventory-Prodromal Risk (NSI-PR), Calgary Depression Scale for Schizophrenia (CDSS), Overall Anxiety Severity and Impairment Scale (OASIS), Social and Occupational Functioning Assessment Scale (SOFAS), Global Functioning: Social (GFS), and Global Functioning: Role (GFR). Since G1 and G2 groups are overlapping, no formal between-group statistical comparisons were conducted.

Results: At BL, G2 showed numerically higher severity scores for positive symptoms than G1 (PSYCHS G1 19.16 ± 8.66, G2 20.79 ± 9.75; SIPS positive G1 11.47 ± 4.12, G2 11.93 ± 4.41; CAARMS positive G1 38.21 ± 17.26, G2 40.32 ± 18.18; BPRS positive G1 4.87 ± 3.22, G2 5.19 ± 3.42). G1 showed numerically higher severity scores for both negative symptom measures (NSI-PR G1 15.43 ± 7.69, G2 14.72 ± 7.17; BPRS negative G1 1.73 ± 2.06, G2 1.20 ± 1.67). Other measures showed no consistent pattern (CDSS G1 5.47 ± 4.13, G2 5.88 ± 4.23; OASIS G1 8.47 ± 4.08, G2 8.35 ± 4.41; SOFAS G1 68.96 ± 13.69, G2 68.62 ± 14.19; GFS G1 7.33 ± 1.25, G2 7.22 ± 1.20; GFR G1 7.40 ± 7.58, G2 7.68 ± 1.65). From BL to M2, both groups showed a significant decrease in all measures except for social and role functioning (SOFAS) and negative symptoms (BPRS subscale and NSI-PR). Numerically larger within-subject improvement effect sizes across visit were observed in G2 for three of the four positive symptom measures (PSYCHS G1 d = 0.574, 95%CI:0.400-0.745, G2 d = 0.644, 95%CI:0.380-0.903; SIPS positive G1 d = 0.580, 95%CI: 0.406-0.752, G2 d = 0.539, 95%CI:0.283-0.792; CAARMS positive G1 d = 0.567, 95%CI: 0.393-0.738, G2 d = 0.612, 95%CI:0.350-0.869; BPRS positive G1 d = 0.441, 95%CI: 0.268-0.613, G2 d = 0.528, 95%CI:0.264-0.788), one of two negative symptom measures (NSI-PR G1 d = 0.073, 95%CI:-0.099-0.244, G2 d = 0.126, 95%CI: -0.133-0.383; BPRS negative G1 d = 0.070, 95%CI:-0.095-0.235, G2 d = -0.020, 95%CI:-0.265-0.225), and four of five of the remaining measures (CDSS G1 d = 0.333, 95%CI:0.168-0.497, G2 d = 0.317, 95%CI:0.072-0.560; OASIS G1 d = 0.450, 95%CI:0.279-0.620, G2 d = 0.509, 95%CI:0.250-0.764; SOFAS G1 d = -0.126, 95%CI:-0.290-0.040, G2 d = -0.154, 95%CI:-0.398-0.091; GFS G1 d = -0.101, 95%CI:-0.263-0.061, G2 d = -0.188, 95%CI:-0.429-0.054; GFR G1 d = -0.062, 95%CI:-0.224-0.100, G2 d = -0.099, 95%CI:-0.339-0.141). Effect sizes were larger for PSYCHS, SIPS, and CAARMS attenuated positive symptoms than for BPRS positive symptoms in both groups.

Conclusions: In the available dataset, severity scores evaluated by PSYCHS, SIPS, and CAARMS attenuated positive symptoms were higher in participants meeting SIPS CHR criteria than in the larger group meeting CAARMS criteria. Over time, both groups showed, on average, significant symptom improvement, with larger effect sizes usually observed in participants meeting SIPS CHR criteria. Notably, for positive symptom measures, the scales designed for CHR (PSYCHS, CAARMS, and SIPS) demonstrated larger within-subject change effect sizes than the positive symptom scale of the BPRS, for participants in both ascertainment groups.

These results, pending completion of the full sample, can inform strategies for ascertainment and outcome measure selection in future CHR clinical trials. Requiring SIPS progression for ascertainment appears to produce a more severely ill sample than CAARMS for positive symptom measures, which could potentially enhance statistical power for treatment effects, although at a cost of a higher screen failure rate than with CAARMS ascertainment. The PSYCHS and/or its derived SIPS and CAARMS attenuated positive symptoms scales may be more sensitive to change over time, given an effective treatment, in CHR participants than the BPRS.

Keywords: Clinical high-risk for psychosis, Clinical outcome assessments, Positive symptoms

Disclosure: Nothing to disclose.

P789. Characterizing Reproducible Brain-Behavioral Association Models in Clinical High Risk for Psychosis

Clara Fonteneau*, Zailyn Tamayo, Jie Lisa Ji, Samuel A. Brege, Kangjoo Lee, Michael P. Harms, Ofer Pasternak, Rene Kahn, Patrick McGorry, John Kane, Carrie E. Bearden, Martha E. Shenton, Barnaby Nelson, Scott Woods, Tyrone Cannon, Alan Anticevic, Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ)

Yale University, New Haven, Connecticut, United States

Background: There have been increasing efforts to generate a robust and reproducible mapping between symptoms and neural features in psychosis spectrum illness. Recent findings in chronic schizophrenia patients suggest that identifying neural effects benefits from data reduction of the symptom variance into a low dimensional ‘geometry’ (e.g., via Non-negative Matrix Factorization or Principal Component Analysis). Mapping these low dimensional symptom axes to neural effects has proven to be more robust and reproducible than using the average of a symptom score on a standard clinical scale. This approach has yet to be tested in clinical high risk (CHR) patients to identify neural markers of psychosis risk. Here we propose to use a discovery-replication approach by combining efforts from the North American Prodrome Longitudinal Study (NAPLS3) and the Accelerating Medicines Partnership Schizophrenia (AMP SCZ) projects. We hypothesize that a data reduction procedure of CHR symptoms will produce a robust low-rank solution in NAPLS3 that will be associated with specific neural circuits. We hypothesize that this brain-behavioral CHR ‘geometry’ will replicate in AMP SCZ. Finally, we hypothesize that the neural maps associated with the low rank symptom solution will produce more robust neural maps than using average scores from full clinical scales alone.

Methods: We studied how behavior relates to resting state functional MRI in control and CHR participants in both NAPLS3 and AMP SCZ datasets. Behavior includes clinical and cognitive domains present in both datasets: Positive, Negative, Disorganized, General functional impairment symptoms, Cognition and Medication status. Notably, the symptom scores collected between NAPLS3 and AMP SCZ datasets do not completely overlap. Therefore, we used the following harmonization approaches: 1) Analysis with exactly overlapping item scores; 2) Analysis with all available items and cross-correlation from Approach 1. The neuroimaging data was processed with the Quantitative Neuroimaging Environment and ToolboX (Qu|Nex), which integrates the Human Connectome Project Minimal Preprocessing Pipelines. To examine the network-level connectivity for each individual, a neural global brain connectivity (GBC) map was computed using a multimodal whole-brain CAB-NP brain parcellation. We performed a principal component analysis (PCA) on the behavioral data to generate the low-rank symptom solution. The PC score for each subject was regressed onto the single-subject connectivity maps using a mass-univariate approach. The discovery analysis was performed on the NAPLS3 dataset. The PCA was performed at baseline using 22 symptom measures (specifically 19 clinical items from the Scale of Prodromal Symptoms (SOPS), 2 cognitive items from the MATRICS Consensus Cognitive Battery (MCCB) and 1 item reflecting the medication dosage). The replication analysis focuses on the baseline data from the second AMP SCZ public data release (anticipated N = 720 CHR participants).

Results: Behavioral analyses from the NAPLS Consortium converged on a stable solution after 10,000 permutations with N = 670 CHR patients and N = 93 controls evaluated at baseline. The PCA solution yielded 4 components (PCs) after permutation testing (p < 0.05, 10,000 permutations), explaining 43% of the clinical symptom variance. We evaluated each PC according to the pattern of the 22 symptom loadings: PC1 - Global Dysfunction, PC2 – Thought Disorder, PC 3 - Disorganized and Cognitive Deficits, PC4 - Positive symptoms. Each PC revealed distinct neural patterns of GBC covariance in NAPLS3.

Conclusions: We identified stable data-driven clinical dimensions specific to baseline CHR patients in the discovery dataset (NAPLS3), which we mapped to neural GBC effects. By leveraging both NAPLS3 and AMP SCZ cohorts we advance the largest discovery-replication effort of CHR symptom-to-neural mapping to date. These results have the potential to inform reproducible neural biomarker development for the CHR syndrome with the ultimate goal of informing personalized treatment decisions that map onto specific neural targets associated with psychosis risk.

Keywords: Clinical high-risk for psychosis, symptom mapping, functional neuroimaging, data reduction

Disclosure: Nothing to disclose.

P790. Non-Specific Markers of Inflammation and the Association With Psychopathology and Functioning Across the Schizophrenia-Spectrum: Findings From a Lower-Middle Income Country

Muhammad Husain*, Ameer Bukhsh Khoso, Suleman Shakoor, Brett Jones, Ishrat Husain, Inti Qurashi, Nusrat Husain, George Foussias, Imran Chaudhry

Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada

Background: Inflammatory mechanisms are thought to contribute to the pathophysiology of schizophrenia spectrum disorders (SSD). We investigated the differences in non-specific markers of inflammation between individuals with at-risk mental state (ARMS), first episode psychosis (FEP) and multi-episode schizophrenia. We also examined the associations between markers of inflammation, psychopathology and functioning within these three groups.

Methods: This study is a cross-sectional secondary analysis of pooled data gathered from five clinical trials evaluating novel anti-inflammatory treatments as adjuncts to standard care for individuals with SSD. The primary studies study took place in Pakistan between 2010 to 2022. For the purposes of this study we analysed demographic, clinical, functional and complete blood count (CBC), erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) data from three groups: ARMS, FEP, and multi-episode schizophrenia. Neutrophil to lymphocyte ratio (NLR), monocyte to lymphocyte ratio (MLR), platelet to lymphocyte ratio (PLR), ESR and CRP were compared between ARMS, FEP and multi-episode schizophrenia. Associations between markers of inflammation and psychotic symptoms and functioning were also evaluated. Analysis of variance (ANOVA) was used to examine the differences between groups with continuous variables when data was normally distributed, and group differences were analysed using post hoc tests to identify significant differences between the groups.

Results: A total of 921 participants, comprising 550 males and 371 females, were included in this analysis (ARMS: n = 326; FEP: n = 274; multi-episode schizophrenia: n = 321). Significant mean differences (p < 0.001) were found between NLR and CRP between the three groups, with multi-episode schizophrenia having higher levels (NLR: Mean ± SD = 2.55 ± 0.98; CRP: Mean ± SD = 6.43 ± 2.06), followed by FEP (NLR: Mean ± SD = 2.27 ± 0.92; CRP: Mean ± SD = 5.23 ± 1.50) and lowest levels were observed in the ARMS group (NLR: Mean ± SD = 1.78 ± 0.72; CRP: Mean ± SD = 4.57 ± 1.15). A statistically significant difference (p < 0.05) in ESR was observed between ARMS (ESR: Mean ± SD = 24.53 ± 17.62) and chronic schizophrenia (ESR: Mean ± SD = 28.53 ± 20.60), while no significant differences were found between ARMS and FEP or between FEP and multi-episode schizophrenia. No statistically significant differences were found in PLR between the three groups. MLR values were also significantly different (P < 0.001) between the three groups with multi-episode schizophrenia having lowest levels (MLR: Mean ± SD = 0.82 ± 0.48), followed by FEP (MLR: Mean ± SD = 0.91 ± 0.68) and highest levels were observed in the ARMS group (MLR: Mean ± SD = 1.06 ± 0.72). We found weak correlations between markers of inflammation, psychopathology and functioning.

Conclusions: We found evidence that non-specific markers of inflammation differ between individuals across SSD from a lower middle-income country. NLR and CRP may increase with disease progression and higher levels were observed in individuals with multi-episode schizophrenia compared to individuals with ARMS and FEP. We acknowledge that the observed differences in inflammatory markers may have been driven by other factors including iatrogenic (e.g., pharmacotherapy, psychotherapy, neurostimulation), lifestyle (e.g., smoking, exercise, diet) or other environmental factors (e.g., stress, medical co-morbidity). Nonetheless, the results suggest that inflammatory mechanisms may contribute to the causation and persistence of psychosis and/or its association with medical comorbidity and shortened life span.

Keywords: Schizophrenia Spectrum Illness, systemic inflammation, neuroinflammation, lower-middle income countries

Disclosure: Nothing to disclose.

P791. Pathway to Care and Baseline Characteristics of Hispanic Individuals Enrolled in Coordinated Specialty Care in the ESPRITO Hub of the EPINET Initiative

Juan Gallego*, Nina Schooler, John Kane, Delbert Robinson

FIMR, Zucker Hillside Hospital, Hofstra North Shore-LIJ, Glen Oaks, New York, United States

Background: Given the complexities in the diagnosis and treatment of a first episode of psychosis (FEP), an evidence-based program called Coordinated Special Care (CSC) has been developed to improve outcomes in individuals suffering from a FEP. On the other hand, limited work has been conducted to understand the pathways to CSC care and the ethno-racial differences related to CSC. To our knowledge, and despite the fact that Hispanics comprise 18% of the total US population, very few articles have been published describing the characteristics of Hispanic individuals that participate in CSC programs. To help decrease this gap in knowledge, we will present data from the ESPRITO hub of the Early Psychosis Network (EPINET). EPINET initiative is a NIMH-sponsored initiative that uses a learning health system model to integrate CSC programs across the United States. The EPINET initiative comprises 8 regional centers or “hubs” and each hub coordinates several CSC programs. For this analysis, we will use baseline data obtained at the time of enrollment of participants at the 13 clinics that are members of the ESPRITO hub of EPINET.

Methods: Statistical analyses were conducted stratified by ethnicity (Hispanic vs. Non-Hispanic). Baseline characteristics were compared between Hispanics and non-Hispanics using either t-tests for normally distributed variables or Wilcoxon rank-sum tests for non-normally distributed variables. Multivariable regression analyses will be conducted over the next weeks to determine the association between baseline and demographic variables and duration of untreated psychosis, defined as the period of time between onset of psychosis and enrollment in CSC.

Results: Seven hundred and sixty-four participants were included in the analysis. Of those, 162 (21.2%) were Hispanics and 602 (78.8%) were non-Hispanic. In the whole sample, 535(71%) were male and their mean age was 22.7 years (SD = 4.4). Hispanics, compared to Non-Hispanics, reported less commonly English as their preferred language (99.2% vs. 82.5, p < 0.001), had lower rates of employment in the past 6 months (43.1% vs. 55.8%, p = 0.005), lived more commonly with family (93.7% vs. 79.8, p < 0.001), had lower scores in symptoms based on the PANSS 6 scale (median=9, IQR = 7-14 vs. median=11, IQR = 7-17, p = 0.02) and were less commonly in foster care (0.7% vs. 4.8%, p = 0.02). There were no significant differences in marital status, education, diagnosis, and neurocognition. In terms of referral source, Hispanics were more commonly referred by crisis stabilization programs (40.7% vs. 8.3%, p < 0.001), whereas non-Hispanics, compared to Hispanics, were more commonly referred by family (12.3% vs. 5.3%, p = 0.036) or by themselves (10.4% vs. 2.7%, p = 0.010). Additionally, Hispanics were more commonly supported by parents (39.5% vs. 27.7%, p = 0.004), were more commonly uninsured (30.8% vs. 16.1%, p < 0.001) and were less often covered by Medicaid (30.8 vs. 46.5%, p < 0.001). Hispanics also had a higher preference of family involvement (p = 0.04). There were no significant differences in the number of psychiatric hospitalizations between Hispanics and non-Hispanics before enrollment in the CSC. Results of multivariate analysis not yet available but to be presented at the meeting.

Conclusions: Several baseline characteristics differ between Hispanic and Non-Hispanic individuals at the time of enrollment into a CSC program. Several of those characteristics, such as less Medicaid coverage, higher rates of no insurance, lower rates of employment, reflect some of the obstacles that Hispanics need to overcome to obtain specialized care. Importantly, Hispanics were more commonly referred to CSC by crisis clinics, in contrast to non-Hispanics who were most commonly referred by either family members or by themselves. There are several potential explanations for this, such as stigma, or decreased access to specialized outpatient care. Expectedly, Hispanics had more family support and preferred a higher degree of family involvement in their care. Despite the overall richness of the database, few details were available for prior health care contacts before enrollment into CSC. Efforts to obtain more details at the individual and family level (stigma, self-reliance) and more information related to prior contacts with primary care physicians and community psychiatrists would help understand the pathways to care as well as the barriers and obstacles faced by Hispanic individuals in their goal of obtaining treatment and recovering from a psychotic episode.

Keywords: coordinated specialty care, Hispanic/Latinos, early intervention services

Disclosure: Nothing to disclose.

P792. Pathway-Partitioned Polygenic Risk Scores for Schizophrenia in Two Patient Cohorts Suggest Novel Mechanistic Insights Into Clinical Variability

Jinhan Zhu, Toni Boltz, Keith Nuechterlein, Robert Asarnow, Michael Green, Katherine Karlsgodt, Jean Addington, Tyrone Cannon, Kristin Cadenhead, Barbara Cornblatt, Matcheri Keshavan, Daniel Mathalon, Diana Perkins, William Stone, Scott Woods, Elaine Walker, Matthew Conomos, Roel Ophoff, Carrie Bearden, Jennifer Forsyth*

University of Washington, Seattle, Washington, United States

Background: Schizophrenia (SCZ) is a neurodevelopmental disorder involving genetic and clinical heterogeneity, including heterogeneity in age of psychosis onset (AOO) that is associated with differences in course of illness. Determining whether AOO differences can be explained by genome-wide polygenic risk for SCZ or levels of polygenic risk partitioned within distinct neurodevelopmental pathways is a key question that may elucidate mechanisms underlying clinical variability. The current study therefore examined relationships between AOO, genome-wide polygenic risk scores for SCZ (PRS-SCZ), and pathway-partitioned PRS-SCZ in two datasets with distinct AOO distribution, namely, in UK Biobank, where SCZ patients show a later-than-average AOO distribution (SCZ-UKBB), and in a cohort of predominantly recent-onset SCZ patients with an earlier-than average AOO distribution, harmonized across multiple studies in North America (SCZ-NA). Since the majority of SCZ-associated single nucleotide polymorphism (SNPs) are located in non-protein coding regions of the genome and the optimal mapping of non-coding SNPs to genes to capture regulatory elements is unknown, pathway PRS-SCZ were investigated using multiple SNP-to-gene mapping definitions.

Methods: Genome-wide PRS-SCZ and pathway-PRS-SCZ were generated for a set of 18 mutually exclusive and previously validated neurodevelopmental gene-sets using PRS-CS for European-like ancestry subjects in SCZ-UKBB (n patients = 1,259, n controls = 19,915) and SCZ-NA (n patients = 235, n controls = 290). PRS-SCZ was first tested for association with SCZ case status. PRS-SCZ associations with AOO were then examined among SCZ cases with AOO information in each cohort (SCZ-UKBB n = 668, mean AOO (SD) = 31.85 (12.96) years; SCZ-NA n = 231, mean AOO (SD) = 19.16 (5.69) years) using linear mixed models, controlling for sex, ancestry principal components, and genetic relatedness. Pathway-PRS-SCZ associations with AOO were tested using various non-coding SNP-to-gene mapping definitions, including a commonly used distanced-based definition of 35 kilobases (kb) upstream and 10 kb downstream and leveraging more specific experimentally-derived annotations including adult brain quantitative trait loci (QTL) in GTEx and PSYCH Encode, fetal brain 3D chromatin contacts from H-MAGMA, and/or cross-tissue annotations of gene enhancers and promoters from GeneHancer.

Results: Genome-wide PRS-SCZ was significantly associated with SCZ case status in both SCZ-UKBB (p = 2.14E-129, R2 = 7.4%) and SCZ-NA (p = 1.29E-07, R2 = 7.1%). Genome-wide PRS-SCZ was not associated with AOO in SCZ-UKBB (p = 0.41) or SCZ-NA (p =.88); however, nominal associations were observed for multiple pathway-PRS-SCZ, correcting for number of SNPs within each gene-set. Most notably, pathway PRS-SCZ for three neurodevelopmental gene-sets were associated with AOO in both SCZ-UKBB and SCZ-NA across multiple non-coding SNP-to-gene mapping definitions: 1) a postnatal synaptic signaling and plasticity gene-set previously found to be strongly enriched for common SNPs associated with SCZ (SCZ-NA maximum R2 = 1.0%, p = 0.0098; SCZ-UKBB maximum R2 = 0.4%, p = 0.0311); 2) a sensory perception gene-set (SCZ-NA maximum R2 = 3.6%, p = 0.0005; SCZ-UKBB maximum R2 = 0.2%, p = 0.0344); and 3) a fetal angiogenesis gene-set (SCZ-NA maximum R2 = 1.2%, p = 0.0195; SCZ-UKBB maximum R2 = 0.8%, p = 0.0458). Higher postnatal synaptic signaling-PRS-SCZ was associated with later AOO in the SCZ-NA cohort that was over-sampled for early onset cases, and with earlier AOO in the SCZ-UKBB cohort that had a relatively late mean AOO. Conversely, higher sensory perception-PRS-SCZ was associated with earlier AOO in the SCZ-NA cohort and later AOO in the SCZ-UKBB cohort. Angiogenesis-PRS-SCZ was associated with earlier AOO in both cohorts.

Conclusions: Genome-wide polygenic risk for SCZ was not associated with AOO in the current study; however, neurodevelopmental pathway-PRS-SCZ results suggest novel insights into mechanisms underlying AOO differences. In particular, associations between PRS-SCZ loading within a postnatal synaptic signaling and plasticity gene-set and AOO in the SCZ-NA and SCZ-UKBB cohorts in a centralizing direction suggests that higher polygenic risk within this canonical SCZ-associated pathway yields psychosis onset during the late adolescence-early adulthood age that is most common for the overall SCZ population. Conversely, PRS-SCZ loading within a fetal angiogenesis pathway was associated with earlier onset in both cohorts. Although speculative, this may suggest that SCZ risk variants that alter integrity of the blood-brain barrier increase risk for early symptom onset by increasing vulnerability of the brain to external pathogens. Associations between sensory perception-PRS-SCZ and AOO in dispersing directions in the SCZ-NA versus SCZ-UKBB cohorts are less intuitive but may suggest that differences in neural processing of environmental stimuli contributes to differences in AOO. Overall, results suggest that pathway-specific PRS-SCZ may provide novel mechanistic insights for understanding clinical variability in SCZ. Future research will explore whether optimizing SNP-to-gene definitions at the individual gene level increases explanatory power of pathway-PRS for parsing clinical heterogeneity.

Keywords: Schizophrenia (SCZ), Clinical heterogeneity, polygenic risk scores, neurodevelopment

Disclosure: Nothing to disclose.

P793. Effects of KarXT (Xanomeline and Trospium Chloride) on Ambulatory Blood Pressure: Results From an Open-Label, Multicenter Trial in Participants With Schizophrenia

Alexander Vandell*, Colin Sauder

Bristol Myers-Squibb, Boston, Massachusetts, United States

Background: KarXT is an investigational treatment for schizophrenia comprising the dual M1/M4 muscarinic receptor agonist xanomeline and peripherally restricted muscarinic receptor antagonist trospium. In three 5-week, randomized, double-blind, placebo-controlled EMERGENT trials, KarXT improved symptoms and was generally well tolerated in participants with schizophrenia experiencing acute psychosis. In pooled analyses of data from the acute EMERGENT trials, KarXT was associated with higher rates of hypertension treatment-emergent adverse events (TEAEs) and slight elevations in blood pressure after 5 weeks of treatment. This trial was conducted to determine whether treatment with KarXT is associated with clinically significant increases in systolic blood pressure (SBP), which may confer long-term risk.

Methods: An open-label, inpatient ambulatory blood pressure monitoring (ABPM) trial was conducted in stable adults (aged 30-65 years) with schizophrenia to assess the effects of KarXT on 24-hour ambulatory SBP. Diastolic blood pressure (DBP) and heart rate (HR) were assessed as secondary outcomes. Participants with current or a history of significant cardiovascular disease, orthostatic hypotension at screening or baseline, currently receiving medication(s) for hypertension, or significant electrocardiogram (ECG) abnormalities were excluded. Participants received twice-daily oral doses of KarXT beginning at xanomeline 50 mg/trospium chloride 20 mg and titrated to a maximum dose of xanomeline 125 mg/trospium chloride 30 mg for 8 weeks. 24-hour ABPM was conducted via an automated ABP recording device, which measured BP every 30 minutes for 24 hours at baseline and week 8. On days 1, 3, 8, 21, 28, 35, and 42, seated vital signs were collected (in triplicate) within 30 minutes before and 2 hours after the morning KarXT dose. The primary endpoint was change from baseline to week 8 in 24-hour average ambulatory SBP. Secondary endpoints included change from baseline to week 8 in average 24-hour DBP, 24-hour HR, and daytime and nighttime ambulatory SBP and DBP. Safety endpoints included TEAEs, clinical laboratory evaluations, and ECG monitoring. Safety analyses were performed in the safety population, defined as all participants who received > /=1 KarXT dose. All ABPM analyses were performed in the modified intent-to-treat (mITT) population, defined as all participants who received > /=1 dose of KarXT, had valid baseline and week 8 24-hour ABPM assessments, and had KarXT compliance between 80% and 120%.

Results: A total of 128 participants were included in the safety population; 95 participants who completed the 8-week trial were included in the mITT population. The majority of participants were male (79.7%) and Black or African American (73.4%) with a mean (SD) age of 44.8 + /-9.31 years and baseline Positive and Negative Syndrome Scale total score of 59.3 + /-8.55. KarXT was not associated with an increase in 24-hour ambulatory SBP. At week 8, 24-hour ABPM demonstrated a mean change in 24-hour ambulatory SBP of -0.59 mmHg + /-1.10. KarXT was associated with a small increase in 24-hour ambulatory DBP of 1.89 mmHg + /-0.78 and a mean change from baseline to week 8 in 24-hour ambulatory HR of 9.82 + /-1.17 beats/min. The effects of KarXT on ambulatory SBP, DBP, and HR were similar for both daytime and nighttime periods. KarXT was generally well tolerated; 75.0% of participants had > /=1 TEAE, and 5.5% had a TEAE leading to discontinuation. The most common TEAEs were dyspepsia (26.6%), headache (18.8%), nausea (16.4%), constipation (14.8%), dizziness (11.7%), dry mouth (9.4%), and vomiting (7.8%). Consistent with the pivotal EMERGENT trials, most TEAEs were related to the activity of xanomeline and trospium at muscarinic receptors and mild or moderate in severity.

Conclusions: KarXT is unlikely to be associated with a clinically meaningful increased risk of hypertension and related cardiovascular events. The results of this open-label 24-hour ABPM trial demonstrate that there is no clinically significant effect of KarXT on ambulatory SBP in adults with schizophrenia. Nominal increases in HR and DBP were observed but not associated with meaningful adverse outcomes. Together with results from the pivotal EMERGENT trials, these results support the potential of KarXT, if approved, to be the first in a new class of medications based on muscarinic receptor agonism and a well-tolerated alternative to currently available antipsychotics.

Keywords: KarXT, Safety, Xanomeline, Blood Pressure, Antipsychotic

Disclosure: Yes, I (or my spouse/partner) do have a financial relationship to disclose.

Financial Relationships Details Bristol Myers Squibb, Employee, Self

P794. Electroencephalography-Based Event-Related Potential and Oscillation Measures Deployed in the Accelerating Medicines Partnership Schizophrenia Study: Interim Analysis Assessing Test-Retest Stability/Reliability

Daniel Mathalon*, Gregory Light, Spero Nicholas, Tashrif Billah, Suzie Lavoie, Thomas Whitford, Brian Roach, Sylvain Bouix, Holly Hamilton, Oliver Pogarell, Andrey Anhokin, Raoul Jenni, Sarah Kerins, Clarice Lee, Ilvana Dzafic, Aysenil Belger, Stephen Wood, Rene Kahn, Patrick McGorry, John Kane, Carrie Bearden, Barnaby Nelson, Scott Woods, Martha Shenton, Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ)

University of California, San Francisco, San Francisco, California, United States

Background: Based on current clinical criteria, only 15-20% of young people at clinical high risk for psychosis (CHR) convert to full psychosis over a follow-up period of 2-3 years, with subsequent conversion rates decreasing over time. Moreover, CHR non-converters are heterogeneous, with 20-30% remitting from the CHR syndrome over similar follow-up periods. These variable outcomes and low conversion/remission rates limit the development of novel treatments and personalized medicine approaches to treatment. Moreover, given side effects and long-term risks of antipsychotic medication, there is a critical need for effective and safe early-stage treatments for CHR individuals.

This clinical and scientific context motivated the launch of the Accelerating Medicines Partnership (AMP) Schizophrenia (SCZ) program in 2020. AMP SCZ is a public-private partnership between the U.S. NIH, the U.S. FDA, the European Medicines Agency, pharmaceutical and life science companies, non-profit and other organizations. AMP SCZ is the largest multi-site prospective longitudinal study of CHR individuals undertaken to date. It comprises two data collection networks, ProNET and PRESCIENT, spanning 43 sites across 5 continents. It is also supported by a Data Processing and Coordinating Center (PREDICT DPACC).

Building on prior CHR studies, AMP SCZ aims to develop tools for predicting transition to psychosis and other outcomes in CHR individuals. By adopting a biomarker-informed framework, which optimizes clinical trial design and provides more precise measures of target engagement and treatment response, AMP SCZ also sets the stage for more rapid and efficient testing of new treatments targeting CHR symptoms and prevention of psychosis. Finally, a major deliverable of AMP SCZ is the creation of a data repository accessible by the broader scientific community through the NIMH Data Archive.

Among the biomarker domains considered for inclusion in AMP SCZ, a set of electroencephalography(EEG)-based event-related potential (ERP) and event-related oscillation measures were nominated based on prior studies showing them to 1) predict CHR clinical outcomes, 2) have sensitivity to schizophrenia, 3) have good test-retest reliability, 4) have theoretical/empirical links to pathophysiological mechanisms (e.g., NMDA receptor hypofunction). Based on these considerations, AMP SCZ converged on a set of EEG measures and paradigms, including mismatch negativity (MMN), auditory and visual P300 to target (P3b) and novel (P3a) stimuli, and 40 Hz auditory steady state response (ASSR). The AMP SCZ study design includes biomarker assessments at baseline and 2-months in order to capture biomarker change trajectories as possible predictors of clinical outcomes. Here, in an interim analysis of the AMP SCZ EEG data, we capitalize on this design to assess the test-retest stability of the EEG-based measures. Since true change can occur over a 2-month retest interval, these stability coefficients can be interpreted as conservative estimates of the true test-retest reliability of the measures.

Methods: Participants who completed baseline and 2-month follow-up EEG sessions were included in this interim analysis, including 654 CHR individuals who met SIPS or CAARMS criteria based on the PSYCHS interview, and 87 community controls (CON). EEG data were recorded from AMP SCZ sites using identical recording systems and were regularly processed and rated for data quality on weekly quality control (QC) calls with sites. AMP SCZ EEG experts reviewed processed data and several QC metrics (e.g., impedances, bridging, task performance) and assigned QC ratings of 4 (excellent), 3 (good), 2 (some usable), or 1 (fail). These metrics were displayed on web-based dashboard that was used by sites to review their processed data. Test-retest stability/reliability coefficients were calculated as generalizability (g)-coefficients (intraclass correlations based on generalizability theory).

Results: Test-retest g-coefficients for CHR and CON groups for each of the principal EEG/ERP measures were generally moderately high for data that received QC ratings of 3 or 4 (CON n = 84; CHR n = 580), n. The g-coefficients for each ERP/ERO measure’s peak amplitude, averaged over a cluster of electrodes where the peak was prominent, were as follows: MMN (gCON = 0.72; gCHR = 0.78), Visual Target P3b (gCON = 0.83; gCHR = 0.79),Visual Novelty P3a (gCON = 0.76; gCHR = 0.74), Auditory Target P3b (gCON = 0.66; gCHR = 0.71), Auditory Novelty P3a (gCON = 0.74; gCHR = 0.69), 40 Hz ASSR Total Power (gCON = 0.67; gCHR = 0.65), 40 Hz ASSR Evoked Power (gCON = 0.64; gCHR = 0.75), 40 Hz ASSR Inter-trial Coherence (gCON = 0.68; gCHR = 0.68). When participant sessions with QC ratings of 2 were added to the groups (3 additional CON and 74 CHR), g-coefficients were modestly reduced, supporting the validity of our QC ratings.

Conclusions: These analyses show that the EEG methods and paradigms implemented in AMP SCZ are yielding measures with moderately high test-retest stability, suggesting that the test-retest reliability of the measures, if assessed over a shorter interval when only random sources of measurement error operate, are likely even higher. Thus, the EEG measures assessed in AMP SCZ are demonstrably reliable, maximizing their potential to be used to select CHR participants for treatment studies, or to serve as target engagement measures or secondary outcome measures in clinical trials

Keywords: Clinical high-risk for psychosis, EEG/ERP electrophysiology, P300, Auditory Mismatch Negativity, Gamma-band auditory steady-state response (ASSR)

Disclosure: Nothing to disclose.

P795. Automatic Assessment of the Scale of Prodromal Symptoms (SOPS) Using Large Language Models

Carla Agurto*, Eduardo Castro, Jenna Reinen, Dheshan Mohandass, Agrima Srivastava, Nora Penzel, Pablo Polosecki, Zarina Bilgrami, Einat Liebenthal, Scott Woods, Martha Shenton, Rene Kahn, Patrick McGorry, John Kane, Carrie E. Bearden, Ofer Pasternak, Guillermo Cecchi, Phillip Wolff, Romina Mizrahi, Barnaby Nelson, Cheryl Corcoran

IBM, Ossining, New York, United States

Background: Large Language Models (LLMs) have become valuable tools in the field of text analysis due to their advanced natural language processing capabilities. Their ability to comprehend language context makes them essential for extracting insights from large volumes of text data. In psychiatry, LLMs are particularly beneficial as they can analyze patient narratives with high accuracy and efficiency. Due to the limited number of available mental health professionals and the need for continuous monitoring of patients, these models could be leveraged to provide insights into patients’ mental health and aid in their remote assessment, especially for those who are at risk of psychosis. With this in mind, we propose an approach to analyze patients’ transcribed speech, infer their Scale of Prodromal Symptoms (SOPS) scores, and evaluate their accuracy compared to the scores assigned by a trained professional in two studies.

Methods: We have tested our approach using data from two studies of English-speaking cohorts:

a) SPEAK study (grants # R01MH115322 and R01MH107558), which contains data from 221 participants from 3 cohorts: community controls (CON: 88), clinical high risk (CHR: 79), and individuals with psychosis (PSY: 54), acquired in 4 sites (2 US, 1 Australia, 1 Canada). Participants have mean age and education years of (23.6 + /- 4.5, 14.4 + /- 2.9), (21.5 + /- 4.6, 12.4 + /- 2.7) and (24.0 + /- 66.1, 5.5 + /- 2.5) for CON, CHR and PSY, respectively. The cohort was ethnically diverse, including 109 Caucasian, 61 Asian, 32 African-American and 19 individuals who identified as ‘other’. From the 221, we obtained the SOPS scores for 101 participants (CON:48, CHR: 47, PSY: 6); and b) Accelerating Medicines Partnership® in Schizophrenia (AMP SCZ study), which contains data from 337 participants (CON:56, CHR:281) acquired in 21 sites. Participants have a mean age of 21.5 + /- 3.9 and 21.4 + /- 4.2 years for CON and CHR, respectively. For this study, the positive and negative symptoms were assessed with the PSYCHS and Negative Symptoms Inventory-Psychosis Risk (NSI-PR) tests.

All participants had an open-ended interview in English. These interviews were recorded and transcribed using TranscribeME. Then, we instructed Llama 3.1 405B model from Meta, to analyze the content of the participants’ transcripts and provide scores the 4 domains (positive, negative, disorganization and general) evaluated in the SOPS test. The parameters of the model were set to greedy sampling, max number of token= 4095, repetition penalty=1, and temperature=0 to ensure consistent and precise responses for this type of analysis. Additionally, we instructed the LLM to analyze the content using sentiment analysis and emotion recognition, along with the definition of each of the 19 items in the SOPS test.

Below, we show an example of the provided LLM scores for 3 positive symptoms.

Unusual thought content: 4 (patient reports experiencing static behind everything, seeing auras around objects, and having memories that may not be real)

Perceptual Abnormalities: 4 (patient reports experiencing dissociation, feeling like she’s not real, and having trouble with her sense of self)

Disorganized Communication: 3 (patient’s speech is mostly coherent, but she jumps between topics and has trouble articulating her thoughts)

Results: The number of words (mean + /- standard deviation) in the responses of the participants during the open-ended interviews were CON: 4209 + /- 1150, CHR: 3611 + /- 1467, and PSY: 3040 + /- 1087 for the SPEAK study; and CON: 2068 + /- 927, CHR: 1829 + /-876 for the AMP-SCZ study. For our first analysis, we used the automatic SOPS total score calculated by Llama 3.1 in the SPEAK study to discriminate CON, CHR, and PSY using a Kruskal-Wallis test, obtaining H = 59.7 (p = 1.1e-13). For the AMP SCZ study, we used a Mann-Whitney U test to discriminate CON vs CHR, achieving U = 12436 (p = 2.3e-12). Then, we computed the correlations between the Llama-scored SOPS and the actual SOPS sub- and total scores for the SPEAK study, achieving Pearson correlation values of r = 0.47 (p = 9.3e-07) for positive symptoms, r = 0.54 (p = 8.2e-09) for negative symptoms, r = 0.41 (p = 2.7e-05) for disorganization symptoms, 0.52 (p = 1.2e-08) for general symptoms, and r = 0.53 (p = 1.5e-08) for the total score. In the case of the AMP SCZ study, the correlation between the positive symptoms (Llama-scored SOPS vs. PSYCHS) was r = 0.32 (p = 3.2e-09), and for the negative symptoms (Llama-scored SOPS vs. NSI-PR) was r = 0.29 (p = 1.5e-07). Of note, the range of predicted negative symptoms scores (0-18) in the SPEAK study was narrower than that of observed negative symptom scores (0-28).

Conclusions: We found statistically significant correlations between the SOPS predicted scores and those assigned by clinical experts. In fact, the results were still significant when comparing the predicted negative symptoms of the SOPS with the NSI-PR scores in the AMP SCZ study, demonstrating the potential of using LLMs in psychiatry. Reasons that LLM’s did not identify more severe negative symptoms may be due to its reliance on language content, which is more limited in the context of severe negative symptoms, as well as the reliance on clinician observation for negative symptom ratings, including not only language content, but also speech acoustics, pauses, face expression and general motor activity. Given the promising results obtained in this preliminary analysis, we plan to expand this approach by using retrieval-augmented generation to make these assessments grounded in expert documentation.

Keywords: Clinical high-risk for psychosis, Large Language Models, Scoring Accuracy, Automated natural speech analysis

Disclosure: Nothing to disclose.

P796. Cardiovascular Safety and Tolerability of KarXT (Xanomeline and Trospium Chloride) in People With Schizophrenia: Pooled Results From the 5-Week, Randomized, Double-Blind, Placebo-Controlled Emergent Trials

Colin Sauder*, Alexander G Vandell, Ian Ramsey, Amy Claxton, Shilling Ruan, Ronald Marcus, Inder Kaul

Bristol-Myers Squibb, Boston, Massachusetts, United States

Background: Currently approved antipsychotics are associated with a range of problematic adverse events (AEs) such as movement disorders, sedation, sexual dysfunction, and effects on cardiovascular health, including weight gain, dyslipidemia, and metabolic syndrome. A significant need remains for more effective, better-tolerated treatments with novel mechanisms of action for people with schizophrenia. The investigational treatment for schizophrenia KarXT combines the dual M1/M4 preferring muscarinic receptor agonist xanomeline and peripherally restricted muscarinic receptor antagonist trospium. In the 5-week, randomized, double-blind, placebo-controlled EMERGENT-1 (NCT03697252), EMERGENT-2 (NCT04659161), and EMERGENT-3 (NCT04738123) trials, KarXT improved symptoms with a safety profile that is differentiated from current standard of care. Overall, KarXT was generally well tolerated in people with schizophrenia experiencing acute psychosis. The most common AEs associated with KarXT were gastrointestinal in nature (eg, nausea, constipation), but also included AEs associated with increases in heart rate and blood pressure

Methods: Data from the acute EMERGENT trials were pooled to further characterize the cardiovascular safety and tolerability of KarXT. The acute EMERGENT trials enrolled adults with a DSM-5 diagnosis of schizophrenia, a Positive and Negative Syndrome Scale total score 80-120, a Clinical Global Impression-Severity score > /=4, and recent worsening of psychosis warranting hospitalization. Eligible participants were randomized 1:1 to twice-daily oral KarXT or matched placebo for 5 weeks. KarXT dosing started at xanomeline 50 mg/trospium 20 mg and was titrated to a maximum dose of xanomeline 125 mg/trospium 30 mg. Safety analyses included the incidence of treatment-emergent AEs (TEAEs), serious TEAEs, and TEAEs leading to discontinuation. Orthostatic hypotension was evaluated as an AE of special interest (AESI). Vital signs were recorded 2 hours after the morning dose at each postbaseline visit to correspond with maximum concentration (Cmax) and included both seated and standing (after 2 minutes) measurements to assess orthostatic change. A 12-lead electrocardiogram (ECG) was obtained on day 1 and at weeks 4 and 5. Analyses were performed in the safety population, defined as all participants who received > /=1 dose of trial medication.

Results: A total of 683 participants (KarXT, n = 340; placebo, n = 343) were included in the safety population for analysis. Baseline demographics and clinical characteristics were similar between the KarXT and placebo groups. The most common cardiovascular TEAEs in the KarXT vs placebo group were systemic hypertension (including hypertension, blood pressure increased, labile hypertension, and orthostatic hypertension; 8.5% vs 1.7%) and tachycardia (5.0% vs 2.3%). Hypertension and tachycardia TEAEs were all mild or moderate in severity; the majority resolved with continued dosing, and none led to treatment discontinuation. One TEAE each of increased heart rate (0.3%), palpitations (0.3%), and abnormal ECG (0.3%) led to treatment discontinuation in the KarXT group compared with none in the placebo group. There were no cases of orthostatic hypotension meeting the criteria for AESI in either treatment group. KarXT was associated with small increases in observed blood pressure (BP); placebo-adjusted mean systolic BP increased 0.7 mmHg and mean diastolic BP increased 1.5 mmHg at week 5 vs baseline. KarXT was also associated with a placebo-adjusted mean increase from baseline to week 5 in observed heart rate (HR) of 6 bpm. Except for HR, which was higher in the KarXT vs placebo group, there was no difference in ECG parameters between the KarXT and placebo groups.

Conclusions: In the acute EMERGENT trials, KarXT was associated with a low risk for cardiovascular TEAEs and small, generally transient increases in BP and HR. The recently completed 52-week, open-label, outpatient EMERGENT-4 and EMERGENT-5 trials as well as an 8-week ambulatory BP monitoring trial will provide additional information on the cardiovascular safety and tolerability of KarXT. If approved, KarXT has the potential to be the first in a new class of medications based on muscarinic receptor antagonism and a well-tolerated alternative to currently available antipsychotics.

Keywords: KarXT, Cardiovascular, Xanomeline, Antipsychotic, Safety

Disclosure: Bristol-Myers Squibb: Employee (Self).

P797. Leveraging Generative AI to Identify Linguistic Biomarkers of Psychosis Risk

Zarina Bilgrami, Cheryl Corcoran, Barnaby Nelson, Scott Woods, Martha Shenton, Guillermo Cecchi, Gahan Pandina, Phillip Wolff*, Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ)

Emory University, Atlanta, Georgia, United States

Background: Psychosis and its risk syndromes are often characterized by disturbances in spoken language, suggesting illogical thinking, loosening of associations, and poverty of content (Andreasen, 1979a, 1979b; Bearden et al., 2011). Recent studies have shown that linguistic features predictive of psychosis onset can be detected in transcripts using hand-engineered algorithms, such as discourse coherence and syntactic complexity (Corcoran et al., 2018; Bedi et al., 2015), semantic density and content (Rezaii et al., 2019), and speech connectedness (Spencer et al., 2021). With the advent of Generative AI, a novel assessment approach has emerged, synthesizing abstract features for theory-agnostic machine learning classification, potentially leading to more robust and unbiased assessments.

Methods: Language samples were collected from 286 individuals identified as Clinical High Risk (CHR) for psychosis and 65 Community Control (CON) participants enrolled in the Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ) project (Release 2.0). Samples were obtained from PSYCHS clinical assessments and open-ended interviews. The PSYCHS clinical interviews followed a semi-structured format, assessing 15 distinct positive symptom categories, while the open-ended interviews were participant-directed, promoting natural and expansive conversations.

We utilized the LLaMA 3 Large Language Model (LLM) from Meta (2024) based on 70 billion parameters. LLaMA 3 was queried about the normality of the interviewees’ responses. Since the concept of normality is not specific to a particular mental health disorder, this approach allows for unbiased feature elicitation. For instance, in response to this question during the PSYCHS, “Have you ever felt that some person or force has been controlling your thoughts?” an interviewee might reply, “Sometimes if I’m on a strange train of thought, I hear ringing in my ear.” LLaMA 3 would determine if the response was normal and, if not, describe the nature of its abnormality. In this example, LLaMA 3 might respond, “No, the interviewee’s responses are not normal. They describe experiencing physical sensations (ringing in the ear) in response to certain thoughts.” When asked to explain why the response was not normal, LLaMA 3 might elaborate, “The interviewee’s responses are abnormal because they describe an unconventional connection between a physical sensation and their thoughts.”

Results: A Naïve Bayes (NB) classifier trained on words and phrases generated by LLaMA 3’s initial and follow-up responses successfully distinguished CHR individuals from CON individuals across 10 test sets. For the PSYCHS interviews, the classifier achieved an AUC of 0.682, a precision of 0.603, and a recall (sensitivity) of 0.680. In the open-ended interviews, the classifier achieved an AUC of 0.585, a precision of 0.581, and a recall of 0.585. This difference between modes of speech elicitation is expected, as PSYCHS specifically queries symptoms. The modest performance in discriminating CHR from CON (compared to psychosis prediction) aligns with, and slightly exceeds, previous manual linguistic studies (Bearden et al., 2011). In open-ended interviews, CHR patients may not discuss diagnostic symptoms unless specifically prompted. Conversely, the NB classifiers struggled to distinguish CHR and CON individuals based solely on the participants’ direct responses. For the PSYCHS interviews, the classifier yielded an AUC of 0.543, a precision of 0.598, and a recall of 0.543. For the open-ended interviews, the results were an AUC of 0.539, a precision of 0.562, and a recall of 0.538. This suggests that AI enrichment might be necessary for effectively differentiating CHR and CON individuals based on language alone.

Interestingly, PSYCHS interviews were linked to features related to positive symptoms, such as “unconventional beliefs,” “disorganized thinking,” and “perceptual disturbances.” In contrast, open-ended interviews yielded responses typical of negative symptoms, such as “unclear responses,” “avoidance,” and “unenthusiastic responses.” To explicitly test this observation, four clinically trained raters evaluated the top 100 CHR-indicative features for association with positive or negative symptoms. A mixed ANOVA revealed that interview type interacted with symptom type, F(1,198) = 23.864, p < .001, η^2 = 0.108, with post hoc analyses indicating that PSYCHS interviews elicited more positive symptoms than negative symptoms, t(99) = 4.389, p < .001, d = 0.439, and vice versa for open-ended interviews, t(99) = 2.244, p = 0.027, d = 0.224.

Conclusions: Abnormal spoken language can differentiate CHR from the norm, through clinical ratings, manual linguistic analyses, hand-engineered algorithms, and now using large language models and generative AI. Interestingly, positive symptoms were elicited more through structured clinical assessment whereas negative symptoms were elicited more through open-ended interviews. Notably, language data effectively differentiated CHR and CON individuals only after being re-expressed and enriched with meta-level comments from an AI, a result highlighting the potential benefits of integrating advanced AI methodologies with clinical data to enhance early detection and intervention strategies for psychotic disorders.

Keywords: psychosis, generative artificial intelligence, language, clinical assessment, early detection

Disclosure: Nothing to disclose.

P798. Patterns of Differential Gene Expression and Co-Expression in Layer 3 Pyramidal Neurons Across 3 Cortical Regions of the Human Visuospatial Working Memory Network in Schizophrenia

Kevin Dowling*, John Enwright, Dominique Arion, David Lewis

University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, United States

Background: Visuospatial working memory (VSWM) depends on coordinated transfer of information across a hierarchical network of functionally distinct cortical regions, including primary visual (V1) and association cortices (posterior parietal cortex, PPC; dorsolateral prefrontal cortex, DLPFC). Deficits in VSWM constitute a core feature of schizophrenia (SZ) and are associated with altered patterns of activity that differ across cortical regions during VSWM tasks. Regional differences in layer 3 pyramidal neuron (L3PN) properties (e.g., morphology, excitability) support distinct patterns of activity during VSWM tasks, therefore we quantified L3PN gene expression in V1, PPC and DLPFC from SZ and matched unaffected comparators (UC).

Methods: For each region, samples of 100 L3PNs were collected from SZ (n = 39, 11 female) and gender- and age-matched UC by laser capture microdissection. Samples, processed in pairs to reduce technical variance, were sequenced to an average depth of 50 million 100-nt paired end reads and aligned to the genome using STAR. Pair-wise differential gene-level mRNA expression (DGE) analyses were performed within region using edgeR, voomwithQualityWeights, and limma. All analyses were corrected for batch effects using COMBAT, and included age, RNA integrity number, and postmortem interval as covariates. P-value threshold-free rank-rank hypergeometric overlap (RRHO) analyses were used to compare concordance/discordance of gene-level mRNA alterations in SZ across regions. LeafCutter was used to analyze differences in local splice variation (LSV), an orthogonal index of the transcriptome relative to gene-level expression. Gene ontology (GO) analyses were performed using PANTHER with Revigo to consolidate redundant terms. To identify groupings of co-expressed genes (modules), weighted gene correlation network analyses (WGCNA) were performed on all genes with quantifiable mRNA levels. High quality modules were identified using WGCNA:ModulePreservation. To assess the plausible biological basis for WGCNA module organization, ChEA3 was used to identify transcription factors (TFs) known to regulate the expression of genes in each module. Cell-type enrichment analysis of WGCNA modules were performed with Seurat using publicly available snRNA-seq data from total tissue homogenates from each of the 3 regions. All analyses were FDR corrected for multiple comparisons (□ = 0.05) and a ≥ 10% difference was considered meaningful for both DGE and LSV analyses.

Results: The number of differentially expressed genes (DEGs) between SZ and UCs were lowest in DLPFC L3PNs (338 upregulated in SZ, 319 downregulated in SZ), intermediate in PPC (557 up, 705 down), and greatest in V1 (1909 up, 1666 down) L3PNs. There were few significant LSV differences between L3PNs in the DLPFC (n = 1), PPC (n = 3), or V1 (n = 16), even when statistical thresholds were relaxed, suggesting differences in gene-level mRNA expression were a primary driver of L3PN transcriptome alterations in SZ. In each region the levels of activity-related transcripts (e.g., BDNF) and genes in mitochondrial bioenergetic GOs (e.g., respiratory chain complex) were lower in SZ. Perhaps consistent with prior reports that measures of dendritic structure (e.g., spine density, length) are less affected in V1 than DLFPC L3PNs in SZ, DEGs upregulated in SZ were overrepresented among GO terms related to regulation of dendritic morphology and stability with the degree of overrepresentation following this regional pattern: V1 > PPC > DLPFC. However, despite differences in the number of DEGs, p-value threshold-free RRHO analyses indicated the overall pattern of SZ-UC gene expression alterations was highly similar across each of the 3 regions, with the greatest similarity in alterations between PPC and V1 L3PNs. To further investigate the basis for these alterations, WGCNA analyses were performed, which identified a greater number of co-expression modules in DLPFC L3PNs (n = 15) than in PPC (n = 5) or V1 (n = 1) L3PNs. Moreover, the size of modules significantly enriched for DEGs had this regional pattern: DLPFC < < PPC < < V1 L3PNs. Genes in each module were predicted to be regulated by largely distinct TFs and were not enriched for markers of distinct excitatory neuron subtypes. Consistent with the possibility that L3PN module complexity may reflect transcriptional regulation, within DLPFC or PPC L3PNs similarity between modules (eigengene correlation) was positively correlated with the number of shared TFs (both r > 0.6, p < 0.001).

Conclusions: L3PN transcriptome alterations in SZ were similar across regions, but DEGs were most numerous in V1, intermediate in PPC, and lowest in DLPFC L3PNs. One explanation for these findings is that a higher number of robust co-expression modules in DLPFC than PPC or V1 L3PNs may reflect greater transcriptional complexity in DLPFC L3PNs, which could buffer against transcriptome alterations in SZ. However, it remains unclear whether the observed transcriptome alterations reflect causal and/or compensatory processes. As L3PN morphology was previously reported to be more altered in DLPFC L3PNs than V1 L3PNs in SZ, our findings that genes associated with L3PN morphology were upregulated in SZ, despite consistently lower levels of activity-related and metabolic transcripts in these neurons, suggests that V1 L3PNs better compensate for SZ-related alterations than do DLPFC L3PNs.

Keywords: brain transcriptomics, pyramidal neuron, Schizophrenia (SCZ), visuospatial working memory

Disclosure: Nothing to disclose.

P799. Multivariate Resting-State Functional Connectivity Features Linked to Transdiagnostic Psychopathology in Early Psychosis

Haley Wang*, Zhen-Qi Liu, Jason Nomi, Taylor Bolt, Charles Schleifer, Carrie Bearden, Bratislav Misic, Lucina Uddin, Katherine Karlsgodt

University of California, Los Angeles, Los Angeles, California, United States

Background: Aberrant synchronization of brain activity between regions, indicative of resting-state functional connectivity (RSFC) alterations, putatively reflects disruptions in neurotransmitter signaling and synaptic plasticity, crucial for higher-order cognitive functions. Meta-analyses reveal widespread RSFC reductions in psychosis, especially in the default mode, salience, and central executive networks. However, the relationship between RSFC and various dimensions of psychopathology remains unclear due to heterogeneity in symptoms and variable findings in RSFC studies. Univariate approaches and small sample sizes contribute to this variability, thus, multivariate, data-driven methods may offer more comprehensive insights. In addition, early psychosis (EP) is marked by diagnostic instability and variable prognosis, making categorical diagnoses less useful. However, despite overlapping symptoms and shared etiologies, EP research often focuses on diagnostic categories. Thus, investigating brain-symptom relationships transdiagnostically, especially in adolescence and EP, may provide vital insight. For instance, identifying shared neurobiological underpinnings of symptom profiles, independent of diagnosis, is essential for understanding core neuropathological mechanisms and developing targeted treatments for youth.

Methods: Partial least squares (PLS) correlation, a statistical learning method was employed to derive latent components (LCs) that optimize covarying patterns between brain and clinical features. Significant LCs were determined via permutation tests (5,000 iterations, FDR correction q < .05). Projecting raw data onto LCs yielded composite scores for RSFC and clinical features. Loadings, as Pearson’s correlations, reflected variable contributions to LCs. Confidence intervals were determined through cross-validation and bootstrapping (10,000 repetitions). Data from the Human Connectome Project-Early Psychosis Release 1.1 (HCP-EP) included 124 patients aged 16-35 years (38.7% female). Patients were diagnosed with schizophrenia, schizoaffective disorder, and psychotic mood disorders and were within the first three years of onset. Sensitivity and post-hoc analyses examined potential confounds (e.g., age, sex, IQ, head motion, diagnoses, duration of illness, substance use, antipsychotic medication dosage).

Results: A single significant LC (p < .001) captured 41.8% of the covariance between RSFC variables and clinical symptoms (r = 0.43, p < .001), indicating its potential relevance to EP pathophysiology. Examination of the LC loadings revealed a pattern of hyperconnectivity in RSFC associated with specific symptomatology across early psychosis diagnoses. Greater RSFC composite scores were linked to loadings featuring increased between-network RSFC across the brain. Sensory-motor (visual and somatomotor) and default mode (A and B) networks exhibited greater RSFC with the control B network and with subcortical regions, including the amygdala, nucleus accumbens, posterior thalamus, and putamen. The RSFC signature was particularly linked to a unique symptom profile, which may broadly represent an inflexible cognitive pattern, with symptoms including elevated rigidity, anxiety, hostility, and tension. No significant differences in RSFC or clinical composite scores were observed across diagnostic groups after false discovery rate (FDR) correction (q > .05), suggesting a transdiagnostic pattern of brain-behavior associations. The composite scores were not correlated with any confounds (q > .05).

Conclusions: This data-driven transdiagnostic approach identified a neurobiological signature (LC) of functional brain alterations across EP diagnoses, showing strong covariance with a unique symptom profile. This LC implicated hyperconnectivity of the somatomotor and default networks with subcortical and control networks. The identified multivariate brain-behavior association might index intermediate neuro-psychological processes and potentially serve as transdiagnostic phenotypes, providing a more comprehensive characterization of individuals’ clinical variability at the early stage of psychotic illness. These findings highlight the utility of data-driven methods to reveal common symptom domains that share RSFC-related neurobiological underpinnings, beyond traditional symptom groupings. This approach could potentially guide the refinement of existing interventions and the development of new treatment targets related to specific symptom clusters or neurobiological profiles.

Keywords: Resting State Functional Connectivity, early psychosis, multivariate analysis, brain-behaviour relationships

Disclosure: Nothing to disclose.

P800. Beta-Band Hyperconnectivity in Early-Course Schizophrenia Revealed by TMS-EEG of the Dorsolateral Prefrontal Cortex

Francesco Donati*, Bruno Couto, Ahmadreza Keihani, Caitlin Moore, Allison Kim, Sabine Janssen, Chloe Huston, Ahmad Mayeli, Adenauer Casali, Fabio Ferrarelli

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Accumulating evidence from functional neuroimaging and neurophysiology studies has recently shown altered connectivity in large-scale brain networks - including the frontoparietal executive network - in schizophrenia, to the extent that neural dysconnectivity is thought to represent a key pathophysiological mechanism implicated in the neurobiology of this disorder. Single-pulse transcranial magnetic stimulation (TMS) coupled with simultaneous electroencephalography (TMS-EEG) can non-invasively and directly investigate the activity and connectivity patterns of virtually any cortical area. Previous TMS-EEG studies in schizophrenia showed a slowing in the EEG response to TMS in the prefrontal cortex of chronic and early-course patients, relative to healthy subjects, with the local prefrontal EEG activity showing a shift from predominantly gamma-frequency responses to predominantly beta-band activity. However, it is unclear whether patterns of long-distance, beta-band-dysconnectivity, which may contribute to the slowing of the prefrontal TMS-EEG responses, are also present in patients with schizophrenia. To answer this question, here we aimed to quantify the connectivity of TMS-EEG responses in early-course schizophrenia (ECSCZ) compared to healthy control (HC) subjects.

Methods: We recruited 47 subjects with ECSCZ, defined as up to 5 years from their first psychotic episode, and 27 HC participants. Age (ECSCZ: 25.48 ± 5.22; HC: 23.21 ± 4.12) and gender (ECSCZ: M = 32, F = 15; HC: M = 17, F = 10) did not significantly differ between groups. Each subject underwent TMS-EEG recordings of the dorsolateral prefrontal cortex (DLPFC). In each session, at least 150 trials were acquired. The accurate targeting of the DLPFC was ensured by neuronavigation using the subject’s individual brain magnetic resonance imaging (MRI). The intensity of TMS was set at 120% of the resting motor threshold. To reduce EEG artifacts and maximize the signal-to-noise ratio, a real-time graphic user interface was used to monitor the quality of the TMS-EEG response. Auditory confounders due to the TMS ‘click’ sound were reduced using an adaptive noise masking tool. After preprocessing using a semi-automated artifact rejection procedure, EEG signals were bandpass filtered in the beta (13-25 Hz) frequency band. To estimate the strength of the TMS-induced EEG connectivity recorded at the sensor level, we calculated the weighted phase lag index (wPLI) in an early (20-120ms) and a late (120-300ms) time window following the TMS pulse. The wPLI quantifies the degree to which phase differences between paired EEG electrodes (or regions of interest, ROIs, of electrodes) exhibit a consistent directional bias, rather than being uniformly distributed, suggesting a non-random relationship. Group differences in wPLI were investigated between frontal, central, and parietal EEG ROIs within and across both the left and right hemispheres using Wilcoxon rank-sum tests.

Results: In the early time window, we found no differences in beta-band connectivity between ECSCZ and HC subjects. However, ECSCZ patients exhibited a significant increase in beta-band wPLI in the late time window between 1) left-frontal and left-parietal ROIs (p = 0.0438); 2) left-central and left-parietal ROIs (p = 0.0475); 3) left-central and right-frontal ROIs (p = 0.0076); 4) left-central and right-central ROIs (p = 0.0054); and 5) left-parietal and right-frontal ROIs (p = 0.0093).

Conclusions: These initial findings showed widespread patterns of late, beta-band-specific hyperconnectivity in ECSCZ, relative to HC participants, after TMS of the DLPFC. These large-scale alterations in connectivity downstream to the prefrontal cortex may represent a novel biomarker of network dysfunction in schizophrenia, particularly within the frontoparietal executive network, with important implications for the understanding of the pathophysiology of this disorder.

Keywords: Schizophrenia (SCZ), TMS EEG, Connectivity

Disclosure: Nothing to disclose.

P801. Combining Large Language Models and Feedback From Clinical High Risk Individuals to Identify Features of Lived Experience Narratives

Jenna Reinen*, Carla Agurto, Carlos Larrauri, Dheshan Mohandass, Cheryl Corcoran, Lana Kambeitz-Ilankovic, Avi Reichenberg, Kathryn E. Lewandowski, Walid Yassin, Phillip Wolff, Tina Kapur, Sylvain Bouix, Rene Kahn, Patrick D. McGorry, John M. Kane, Carrie Bearden, Barnaby Nelson, Scott W. Woods, Martha E. Shenton, Ofer Pasternak, Nora Penzel, Eduardo Castro, Pablo Polosecki, Guillermo Cecchi

IBM Research, Rye Brook, New York, United States

Background: Narrative can provide nuanced insight about lived experience and symptom evolution that can be difficult to capture in standardized assessments. While narrative is historically difficult to assess and quantify, recent advances in artificial intelligence (AI), and specifically using natural language processing (NLP) and large language models (LLMs), have expanded the possibilities to quantitatively evaluate more abstract linguistic concepts. Here, we describe an approach derived from a focused effort to combine input from Clinical High Risk (CHR) individuals, literature review, and technological methodology to develop specific features that can be captured in digital data, such as speech, audio, and visual inputs. We will also provide information about feature development and extraction, and how they differed in community controls (CON) versus at-risk individuals (CHR). These current results can inform the capacity of this approach to digitally capture and quantify aspects of human narrative.

Methods: Part I: We first intended to find important symptom-related aspects in those previously identified as high-risk for schizophrenia. Following an extensive literature review, a group of researchers, clinicians, and a National Alliance on Mental Illness (NAMI) representative conducted a two-day discussion at the Accelerating Medicines Partnership® in Schizophrenia (AMP® SCZ) hackathon, aimed at identifying key aspects and specific examples of lived CHR experience. For each topic, examples were added to further describe each experience, and to inform the NLP lexical similarity / LLM analysis (for example, “changes in school performance” was added as an example under “cognitive changes”). Once these topics were identified, a secondary discussion occurred with a team of audio-visual signal processing research experts to determine how these features could be extracted from interviews using NLP and LLMs. Statements related to each topic were developed for future use with the NLP lexical similarity analysis. Part II: Our computational approach aimed to identify and quantify expressions in the original CHR individual’s response that are related to the key aspects identified in Part I, which can help us to profile the CHR population. We transcribed open-ended clinical interviews performed with participants (CON and CHR) recruited for the AMP SCZ study. Then, we processed those transcriptions using robust BERT Transformers (RoBERTA) to compute lexical similarity to each of the key aspects identified in Part I by tokenizing the content in each sentence using Natural Language Toolkit (NLTK). Cosine similarity was computed between the interview sentence and the topic-specific statements designed to reflect the key aspects from Part I. For each topic, a mean score across the interview was then computed for each individual. In addition to mean score, other metrics were explored and will be discussed. Following this, group differences and significance tests were calculated.

Results: In this analysis, we used data from 297 CHR individuals and 60 community controls across 21 international sites derived from a baseline visit (Mean age CHR 21.3 + /- 4.2 years; CON 21.5 + /- 3.9 years; CHR 153F/144M; CON 36F/24M). We first finalized the list of key topics identified in the focused meeting. This included changes in identity, feelings of isolation, feelings of suspiciousness and changes in salience processing, trauma/stress, fear/anxiety, appreciation for help and connection, cognitive changes, and future thoughts and planning. Additional experience-based examples of each of these topics will be outlined. The lexical similarity results indicated that feelings of isolation (p = 0.02), including reports of engagement with family and friends (p < 0.01), suspiciousness (p < 0.01), anxiety (p = 0.01), fear (p < 0.001), and cognitive properties, including ability to think (p < 0.001), and school performance (p < 0.01) were significantly different between CHR individuals and community control individuals.

Conclusions: We have described a method and implementation by which lived experience narrative and CHR individual feedback may be translated into features used by NLP and LLMs to identify symptoms and group assignment. In some cases, this approach adds dimensionality beyond standard clinical assessments of CHR symptoms, in particular as it applies to affective experience. It should be taken into consideration that these findings are part of an ongoing study, and that additional data will be available for future analysis. Additionally, those tested represent a cohort early in at-risk identification, and findings may be characteristically different from a longitudinal lens. Here, we have presented a feasibility experiment that lays the grounds for future work, thereby enabling CHR individuals, researchers, and engineers to combine experience and expertise intended to guide machine learning models in identifying symptom changes and disease progression in psychosis.

Keywords: Clinical high risk state for psychosis, Prodromal Psychosis, Large Language Models, natural language processing (NLP)

Disclosure: Nothing to disclose.

P802. Mono-Clonal Antibodies for Schizophrenia: Preliminary Results of an Individual Patient Data Meta-Analysis

Mark Weiser, Ragy Girgis, Brian Miller, Thomas Weickert, Jin-Young Park, Linda Levi, Fuller Torrey, John Davis*

University of Illinois at Chicago, Chicago, Illinois, United States

Background: One of the hypotheses regarding the etiology of schizophrenia is that inflammatory mechanisms are involved in its pathophysiology. This is supported by studies showing that patients with schizophrenia have increased brain, plasma and serum levels of Interleukin-6 (IL-6) and Interleukin 1-beta (IL-1β). The Stanley Medical Research Institute funded three RCTs administering monoclonal antibodies for the treatment of schizophrenia: canakinumab: a monoclonal antibody that interferes with the bioactivity of IL-1β; tocilizumab: a monoclonal antibody against the IL-6 receptor; and siltuximab, a monoclonal antibody that binds IL-6. This study presents an individual patient data meta-analysis of these three studies together.

Methods: In our preliminary analysis, we performed an individual patient data meta-analysis combing the data of three RCTs. We examine the end points of the Positive and Negative Syndrome Scale (PANSS) Total, Positive, Negative, and General symptoms scores by ANCOVA, with baseline score as a covariant, and monoclonal antibody and study as factors. We will also evaluate the early time point, and a mid-time point, and explore other variables that might alter psychopathology.

Results: The monoclonal antibodies non-significantly decreased PANSS total, by -2.5 points (se=1.9, t = -1.3, p = 0.19 ns), positive symptoms by -.84 points (se = 0.90, t = -.94, p = 35), and negative symptoms by -.42 points (se = 0.80, t = -.52, p = 0.60 ns), but non-significantly increased general symptoms, by 0.42 points (se=1.4, t = 0.31, p = 76 ns).

Conclusions: The results of this preliminary meta-analysis did not find efficacy of monoclonal antibodies in the treatment of the symptoms of schizophrenia. Whereas some of the blood-based markers were changed by these monoclonal antibodies, this did not translate to reduction in clinical symptomatology. One potential reason for these results might be that these drugs do not cross the blood-brain barrier, hence while they have an effect on peripheral blood-based immune markers, they do not lead to change in symptoms. These results do not encourage further studies on monoclonal antibodies in reducing the symptoms of schizophrenia

Keywords: Schizophrenia (SCZ), monoclonal antibodies, Interleukin 1beta, Interleukin-6

Disclosure: Nothing to disclose.

P803. Comprehensive Isoform Analysis Reveals Differential Expression in Schizophrenia Dorsolateral Prefrontal Cortex

Anthony S. Abrantes, Lisa Bast, Fatima Memic, Shuyang Yao, NaEshia Ancalade, Anke Dijkstra, Shadia Sekle, Elizabeth Tseng, Craig Stockmeier, Gloria Sheynkman, Jens Hjerling-Leffler, Guus Smit, Patrick F. Sullivan, Paola Giusti-Rodriguez*

University of Florida, College of Medicine, Gainesville, Florida, United States

Background: Genetic studies have provided important insight regarding the architecture of multiple psychiatric disorders, including schizophrenia. It is now widely acknowledged that schizophrenia is highly polygenic, consisting of both common variants of unknown significance (identified through genome-wide association studies [GWAS]) and rare coding variants (identified through exome and copy number variant studies). Despite these advances, our comprehension of the implication of these rare, coding variants and of the single nucleotide polymorphisms that reside in the non-coding genome, which are not easily interpretable.

Multiple lines of evidence implicate alternative splicing in the pathophysiology of schizophrenia. Common risk variants associated with schizophrenia and major depressive disorder reside within the gene locus of the RBFOX1 RNA binding protein, and mutations of RBFOX1 have been linked to autism and epilepsy. This is further supported by gene-set analysis of schizophrenia GWAS findings which are enriched for genes that harbor binding sites for RBFOX1 and RBFOX2, another important splicing factor. Long-read RNA-sequencing technologies have made dramatic improvements in throughput, accuracy, access, and cost. This study aimed to generate a comprehensive isoform survey of human dorsolateral prefrontal cortex (DLPFC) and to identify potential isoform level differences in schizophrenia cases and controls.

Methods: To achieve this, we generated long-read RNA-sequencing data from post-mortem human DLPFC from schizophrenia cases (n = 10) and controls (n = 15) using the PacBio Sequel II system. We also generated complementary short-read RNA-seq data to examine isoform abundances. We used published and in-house pipelines to generate a comprehensive isoform survey of human DLPFC and to compare differential isoform usage in DLPFC from schizophrenia cases vs healthy controls.

Results: We identified differential isoform usage (DIU) in DLPFC from schizophrenia cases compared to healthy controls using established and in-house pipelines (IsoSeq, Cupcake, SQANTI3, IsoAnnot, and DiffSplice) and contextualized these findings at the transcript- and gene-level using pathway and network analyses.

Conclusions: Our findings support DIU as a regulatory mechanism at play in the pathophysiology of schizophrenia. DIU may reveal schizophrenia-associated isoforms that impact protein-protein interactions (PPIs) and could nominate actionable targets.

Keywords: Genomics, RNA splicing, transcript isoforms, Schizophrenia (SCZ)

Disclosure: Nothing to disclose.

P804. Personalized High-Definition MR-Guided tDCS Improves Face Emotion Recognition in Schizophrenia

Antigona Martinez*, Maria Belen Aburto-Ponce, Pejman Sehatpour, Odeta Beggel, Kristin Micceri, Daniel Javitt

Nathan S Kline Institute for Psychiatric Research, Orangeburg, New York, United States

Background: Impaired face emotion recognition (FER) is a key feature of schizophrenia (SCZ) and is closely linked to functional outcomes. FER depends on the coordinated function of subcortical and cortical regions within the primate-specific ‘third visual pathway’ (TVP), which is specialized for processing social information. This pathway preferentially processes motion inputs relayed through the primary visual cortex (V1) and the motion-sensitive middle temporal cortex (MT + ), extending into posterior (pSTS) and anterior (aSTS) regions of the superior temporal cortex. Previous functional MRI (fMRI) studies in SCZ have shown deficits in V1 and MT+ activation in response to simple visual motion stimuli, which correlate with reduced FER. Patients with SCZ also show impairments in V1, MT + , and pSTS activation when processing dynamic emotional faces. This study uses EEG and fMRI to localize regions involved in FER deficits in SCZ and combines these with personalized, high-definition, MR-guided transcranial direct current stimulation (pHD-tDCS) targeting MT + , to explore the causal mechanisms underlying FER dysfunction in SCZ.

Methods: Deficits in MT+ activation were identified in 22 individuals with SCZ and 27 neurotypical (NT) controls using a random dot kinematogram (RDK) task. FER accuracy and TVP activation was assessed using dynamic emotional faces. A subset of 18 SCZ participants underwent three sessions of pHD-tDCS concurrent with EEG acquisition and one session concurrent with fMRI. The tDCS/EEG sessions consisted of either sham (30-second ramp-up/ramp-down), cathodal, or anodal stimulation, while the tDCS/fMRI session involved sequential sham and cathodal stimulation. Realistic head modeling was employed to generate personalized models that maximized current flow in MT + . Behavioral, EEG, and fMRI data were collected pre-, during, and post-tDCS. Statistical analyses were conducted using univariate repeated measures ANOVA and Pearson correlations to explore relationships among measures, corrected for multiple comparisons. EEG data were analyzed using single-trial time-frequency decomposition of event-related activity. fMRI activations were mapped onto individual cortical surfaces, clustered, and thresholded at p < .05.

Results: SCZ participants demonstrated significant deficits in MT+ activation (p < .01) and higher-order TVP regions (p < .001) in response to RDK stimuli, as well as reduced amplitude of the motion-specific N2m component (p = 0.032) which maps primarily within the theta-frequency range. Cathodal pHD-tDCS significantly reversed MT+ deficits (p < .01) and increased theta evoked power in response to RDK stimuli. During the FER task, cathodal tDCS improved both FER accuracy (p = 0.037) and activation of TVP regions, including pSTS (p = 0.047). Behavioral improvements in motion perception and FER were correlated (p = 0.051) with one another and linked with activation of MT+ (p = 0.005) and pSTS (p = 0.023), respectively. Finally, cathodal pHD-tDCS during the motion task significantly increased source coherence between MT+ and nodes of the TVP (p = 0.033).

Conclusions: These findings highlight the role of impaired motion processing in FER deficits in SCZ and demonstrate the feasibility of concurrent pHD-tDCS/EEG and pHD-tDCS/fMRI studies. Additionally, they establish a causal relationship between impaired MT+ activation and social cognitive impairments in SCZ, suggesting that future studies with repeated pHD-tDCS sessions targeting MT+ may significantly alleviate FER deficits in this population.

Keywords: transcranial direct current stimulation (tDCS), face emotion processing, EEG, fMRI, motion processing

Disclosure: Nothing to disclose.

P805. Hepatic Safety of Karxt (Xanomeline and Trospium Chloride): Pooled Results From the Randomized, Double-Blind, Placebo-Controlled Emergent Trials

Amy Claxton*, Ronald Marcus, Shiling Ruan, Inder Kaul, Colin Sauder

Bristol Myers Squibb, Bellevue, Washington, United States

Background: KarXT is an investigational treatment for schizophrenia comprising the dual M1/M4 preferring muscarinic receptor agonist xanomeline and peripherally restricted muscarinic receptor antagonist trospium chloride. KarXT is designed to preserve xanomeline’s beneficial central nervous system effects while mitigating adverse events (AEs) due to peripheral muscarinic receptor activation. The efficacy and safety of KarXT in people with schizophrenia experiencing acute psychosis were demonstrated in the three 5-week, randomized, double-blind, placebo-controlled EMERGENT-1 (NCT03697252), EMERGENT-2 (NCT04659161), and EMERGENT-3 (NCT04738123) trials. Long-term safety and tolerability of KarXT in adults living with schizophrenia were assessed in two 52-week, outpatient, open-label trials: EMERGENT-4 (NCT04659174) and EMERGENT-5 (NCT04820309).

Methods: Data from the completed acute EMERGENT trials and interim data from the long-term EMERGENT-4 and EMERGENT-5 trials as of April 17, 2023, were pooled separately to characterize the hepatic safety of KarXT. The acute EMERGENT trials enrolled people with a recent worsening of psychosis warranting hospitalization, Positive and Negative Syndrome Scale (PANSS) total score 80-120, and Clinical Global Impression–Severity (CGI-S) score ≥4. EMERGENT-4 enrolled participants who completed EMERGENT-2 or EMERGENT-3. EMERGENT-5 enrolled adults with schizophrenia with no prior exposure to KarXT, a PANSS total score ≤80, and a CGI-S score ≤4. Oral KarXT was dosed twice daily at xanomeline 50 mg/trospium 20 mg and titrated to a maximum dose of xanomeline 125 mg/trospium 30 mg for either 5 weeks (acute trials) or 52 weeks (long-term trials). Any liver transaminase elevations ≥3× upper limit of normal (ULN) for alanine transaminase (ALT) or aspartate transaminase (AST) were monitored as AEs of special interest (AESIs). All safety analyses were conducted in the safety population, defined as all participants who received ≥1 dose of trial drug.

Results: A total of 683 participants (KarXT, n = 340; placebo, n = 343) in the acute EMERGENT trials and 674 participants in the long-term EMERGENT trials were included in the safety analyses.

In the acute trials, 1.8% (n = 6) in the KarXT group experienced liver transaminase elevations meeting criteria for an AESI. All but 1 participant receiving KarXT (5 of 6) had transaminase elevations resolve with continued KarXT treatment; 1 participant had transaminase normalize shortly after treatment discontinuation.

In the long-term trials, 1.9% of participants (n = 13) experienced liver transaminase elevations meeting criteria for an AESI. Seven of 13 continued KarXT treatment, and their transaminase levels normalized. Three participants had transaminase levels return to normal after brief treatment interruptions, two had transaminase levels normalize following discontinuation of KarXT, and 1 participant was lost to follow-up.

The majority of liver transaminase elevations occurred within the first month of treatment and resolved rapidly with continued KarXT treatment, consistent with hepatic adaptation to xanomeline.

Conclusions: In the EMERGENT trials, treatment with KarXT up to 52 weeks was generally well tolerated in people with schizophrenia. The incidence of liver transaminase elevation meeting AESI criteria was low and the pattern was similar in both acute and long-term trials. Transaminase elevations occurred early during the course of KarXT treatment, with evidence of hepatic adaptation as events resolved over time, resulting in few discontinuations or interruptions in treatment. These findings, together with the efficacy results showing a clinically meaningful reduction in the symptoms of schizophrenia, support the potential of KarXT, if approved, to be the first in a new class of medications based on muscarinic receptor agonism and a well-tolerated alternative to currently available antipsychotics.

Keywords: KarXT, liver transaminase, Safety, Xanomeline, Antipsychotic

Disclosure: Yes, I (or my spouse/partner) do have a financial relationship to disclose.

Financial Relationships Details Bristol Myers Squibb, Employee, Self

P806. An Integrative Model of Reduced Recurrent Excitation in Schizophrenia

Kenneth Wengler*, Jordan Hamm, Yuriy Shymkiv, Ken Miller, Guillermo Horga, Rafael Yuste

Icahn School of Medicine At Mount Sinai, New York, New York, United States

Background: Schizophrenia is a serious mental illness that affects ~1% of the global population. Despite the development of antipsychotics in the 1960’s for treating the positive symptoms of schizophrenia (i.e., hallucinations and delusions), treatment development for the other symptom domains (i.e., negative symptoms and cognitive deficits) have had limited success. While substantial efforts have been made, there remains a translational gap in the treatment development process due to the unclear pathophysiology of the disease and the corresponding limited construct, face, and predictive validity of preclinical models for schizophrenia. Here, we investigate intrinsic neural timescales (INT) and spontaneous coactivation-reliability in both preclinical imaging data (2P-Ca2 + ) and human neuroimaging data (resting-state functional magnetic resonance imaging [rs-fMRI]) and use a biophysical model to link these two imaging measures and bridge the translational gap.

Methods: Preclinical data consisted of 2P-Ca2+ from layer 2/3 pyramidal cells in the primary visual cortex (V1) of three mouse models of schizophrenia-relevant disease mechanisms: (1) mice undergoing chronic ketamine administration (n = 6; n = 5 controls) that model NMDA-receptor hypofunction; (2) Df16 + /- mice (n = 7; n = 7 controls) that model the 22q11.2 microdeletion copy number variant; and (3) Setd1a + /- mice (n = 9; n = 8 controls) that model the Setd1a loss-of-function variant. Clinical data consisted of rs-fMRI from individuals with schizophrenia (n = 127) and matched healthy controls (n = 152). Analogous measures were derived at the microscopic (i.e., 2P-Ca2+ data) and macroscopic (i.e., rs-fMRI data) scales to maximize translational capacity. For INT, given that previous studies have calculated INT from calcium imaging data and rs-fMRI data (Wengler et al., eLife, 2020), standard analysis methods were used such that INT represents the decay rate of the autocorrelation across individual neurons (microscopic scale) or across voxels within a brain region (macroscopic scale). For spontaneous coactivation-reliability, standard analysis methods (Hamm et al., Biological Psychiatry, 2020) were used for estimating reliability across neuronal ensembles (microscopic scale) while analogous methods were used for estimating reliability across voxel ensembles within a brain region (macroscopic scale). Multiple linear regression was used for all statistical analyses.

Results: Given several reports of shorter INT in patients with schizophrenia, we first estimated INT in the preclinical data. We found that V1 INT was significantly shorter in the mouse models of schizophrenia-relevant disease mechanisms compared to the control mice (t = -2.10, p = 0.043). To demonstrate construct validity, we also analyzed rs-fMRI from patients with schizophrenia (n = 127) and healthy controls (n = 152) and found (as expected; Wengler et al., 2020) shorter whole-brain averaged INT (t = -2.34, p = 0.020). Next, to further explore construct validity, we estimated the spontaneous coactivation-reliability using the rs-fMRI data from the patients and controls and found reduced whole-cortex averaged spontaneous coactivation-reliability in the patients (t = -2.62, p = 0.009). Similarly, when combining all of the mouse data, we found (as expected; Hamm et al., 2020) reduced V1 spontaneous coactivation-reliability in the mouse models (t = -3.70, p = 0.001).

To explore the circuit basis of the INT deficits, we then set out to develop a biophysical model— a network of excitatory (E) and inhibitory (I) neurons with biologically realistic properties—that was able to recapitulate the previous experimental observation of reduced spontaneous coactivation-reliability across three mouse models of schizophrenia-relevant disease mechanisms. Following Hamm et al.’s suggestion of altered attractor dynamics as a potential explanation for their experimental findings, we utilized a ring-attractor network. Spontaneous coactivation-reliability was only sensitive to varying E- > E connections such that weaker connection strengths produced less reliable ensembles. In line with previous modeling work demonstrating a role of recurrent excitatory connections in INT, reduced E- > E connection strengths also produced shorter INT. This is consistent with the role of recurrent excitatory connections in attractor dynamics and supports Hamm et al.’s hypothesis of altered attractor dynamics in their preclinical models.

Finally, to explore the relevance of the biophysical model and to assess predictive validity, we tested for relationships between spontaneous coactivation-reliability and INT in the human and mouse data, as predicted by the biophysical model. Importantly, spontaneous coactivation-reliability and INT were positively correlated both for V1 in mice (t = 2.16, p = 0.037) and for whole-brain averages in humans (t = 2.25, p = 0.025).

Conclusions: Here we report that individuals with schizophrenia and three different mouse models of schizophrenia-relevant disease mechanisms display similar shortenings of their neural circuit time scales. Using biophysical modeling, we demonstrate that these deficits can be explained by reduced recurrent excitatory connectivity that creates fewer circuit attractors. Our results indicate that schizophrenia is a specific pathology of cortical connectivity, demonstrate the utility of circuit-based bioassays of the disease and point at novel therapeutic entry points.

Keywords: Translational Neuroscience, schizophrenia, intrinsic neural timescale, coactivation pattern analysis, biophysical modeling

Disclosure: Terran Biosciences: Patent (Self).

P807. Inflammation and Aging in Psychosis – a Transdiagnostic Proteomics Study Using the Human Connectome Project for Early Psychosis (HCP-EP)

Johanna Seitz-Holland*, Moritz Haaf, Ana Paula Mendes-Silva, Nora Penzel, Cornelius Berberich, Lauren Breithaupt, Sylvain Bouix, Michael J. Coleman, Elana Kotler, Nicholas Kim, Kathryn E. Lewandowski, Daphne Holt, Matcheri Keshavan, Dost Öngür, Alan Breier, Martha E. Shenton, Breno S. Diniz, Steven E. Arnold, Marek Kubicki

Brigham and Women’s Hospital, Harvard Medical School, Sommerville, Massachusetts, United States

Background: Given that protein expression mediates genetic vulnerability, proteomics—the comprehensive study of proteins—is regaining prominence in neuropsychiatric research. This resurgence is driven by novel technical advancements that enable the simultaneous examination of multiple proteins. Numerous studies have demonstrated the potential of proteomics in elucidating pathological mechanisms, identifying clinically relevant biomarkers, and revealing new treatment options for neurodegenerative disorders. Although fewer studies have employed these novel techniques in psychiatry, proteins have been identified as promising transdiagnostic markers for understanding and addressing disease heterogeneity. Consistent with research from other fields, protein studies support the concept of an inflammatory subtype underlying serious mental illnesses, including psychosis. This aligns with the “inflamm-aging“ hypothesis, which posits that chronic low-grade inflammation may contribute to increased vulnerability to premature aging in these populations.

The present study investigates inflammation and aging-associated protein markers in the Human Connectome Project for Early Psychosis (HCP-EP). Specifically, we will examine 374 peripherally circulating plasma proteins associated with inflammation, cardiometabolic health, cell communication, and aging. Additionally, we will calculate specific indices related to cell aging based on senescence-associated secretory phenotype (SASP) proteins.

Methods: The HCP-EP study was conducted across four recruiting sites, collecting data from participants aged 16 to 35 years. The original cohort included 223 individuals within five years of the onset of a DSM-5 diagnosis of psychosis (schizophrenia, schizophreniform disorder, schizoaffective disorder, psychosis not otherwise specified, delusional disorder, brief psychotic disorder, major depression with psychosis, or bipolar disorder with psychosis), as well as 80 matched healthy controls. Participants underwent extensive testing, including the collection of demographic, clinical, and cognitive data.

In the present study, out of the 303 individuals, we included the 120 who had plasma stored in EDTA tubes. Proteomic plasma analyses were conducted using the Olink platform, selecting four validated panels relevant to “inflamm-aging” (Target 96 Cardiometabolic, Target 96 Inflammation, Target 96 Cardiovascular II, and Target 96 Cardiovascular III). The sample comprised 35 healthy individuals, 63 diagnosed with non-affective psychosis, and 22 with affective psychosis, matched by age and sex.

For comparisons between all individuals with psychosis and healthy individuals, as well as between non-affective and affective psychosis groups, we conducted three sets of analyses. First, we performed data-driven analyses using all proteins. Using the results from the data-driven analyses, we then applied gene set enrichment analysis. To identify proteins related to “inflamm-aging” specifically, we then utilized principal component analyses to calculate a SASP protein index based on a selected set of proteins previously studied in individuals with Alzheimer’s disease and depression. Last we examined the association between all protein indices and clinical and cognitive variables in individuals with psychosis only.

Results: The data-driven analyses revealed significant differences between individuals with psychosis and healthy individuals in several proteins related to inflammation, cell communication, blood pressure, and cardiometabolic regulation (e.g., Renin, Tumor necrosis factor receptor superfamily member 9, Leptin, Chemokine ligand 20, Tyrosine-protein kinase Mer; pFDR-corrected < 0.1). We observed more pronounced group differences in the non-affective versus healthy comparison than in the affective versus healthy comparison.

Gene set enrichment analysis indicated significant enrichment of the tumor necrosis factor and monocyte chemotaxis pathways (pFDR-corrected < 0.1) in individuals with psychosis compared to healthy individuals.

When calculating SASP indices, we found a high correlation between individual proteins, consistent with previous studies. Individuals with psychosis exhibited a significantly higher SASP index than healthy individuals (p < 0.05). Additionally, the SASP indices in individuals with psychosis were significantly associated with age, sex, body mass index, and alcohol consumption (all p < 0.05), while associations with cognition were only trend-level significant.

Conclusions: This study represents an initial step toward developing novel transdiagnostic protein-based biomarkers for psychosis. Consistent with previous research on serious mental illness, we observed an increase in protein-levels related to inflammation and aging in individuals with psychosis, particularly pronounced in those with non-affective psychosis. Gene set enrichment and principal component analyses underscored the advantages of studying protein indices over individual proteins. The associations of these variables with physical health parameters further emphasizes the importance of considering psychosis as a whole-body disorder, which may aid in understanding comorbidities. However, future larger and longitudinal studies are needed to replicate our findings.

Keywords: early psychosis, Proteomics, inflammation, Aging

Disclosure: Nothing to disclose.

P808. Neuromelanin Accumulation and Treatment Responsiveness in Patients With Schizophrenia: A Cross-Sectional Study With Neuromelanin-Sensitive MRI

Fumihiko Ueno*, Yusuke Iwata, Shiori Honda, Guillermo Horga, Clifford Cassidy, Edgardo Torres, Jianmeng Song, Vincenzo De Luca, Sakiko Tsugawa, Yoshihiro Noda, Mahavir Agarwal, Gary Remington, Philip Gerretsen, Shinichiro Nakajima, Ariel Graff-Guerrero

Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada

Background: Neuromelanin (NM) is a product of monoamine metabolism including dopamine (DA). NM-sensitive MRI sequences allow in vivo quantification of NM levels in the substantia nigra (SN). NM-MRI signal is thought to serve as a biomarker for SN dopamine neuron integrity, and in turn, striatal DA functioning. Increased striatal DA synthesis has been associated with schizophrenia responding to first-line antipsychotics (first-line responders [FLR]), while normal striatal DA synthesis has been associated with treatment-resistant schizophrenia (TRS). However, although clozapine is the only approved drug for TRS, no studies have investigated the relationship between striatal DA activities, as measured by NM-MRI, and clozapine response in patients with TRS.

Methods: This study enrolled patients with TRS who did not respond to clozapine (ultra-resistant schizophrenia [URS]) and who responded to clozapine (non-URS), patients with FLR, and healthy controls (HCs). We measured NM signals in the substantia nigra (SN) using NM-MRI and compared these signals, as assessed with the contrast ratio (CR), among the groups. The associations between SN-NM signals and participants’ characteristics were also explored.

Results: 78 participants (URS, n = 16; non-URS, n = 16; FLR, n = 20; HCs, n = 26) completed the study. Overall group differences were found in NM-MRI signals after controlling for age (F(3,72) = 6.30, η2 = 0.21, p = 0.001). Specifically, patients with URS (Cohen’s d = 0.86, p = 0.03) and FLR (Cohen’s d = 1.27, p < .0001) exhibited higher CR values compared to HCs. NM signals showed no associations with any of the characteristics or symptom severity scores within each group or the entire patient group.

Conclusions: Our results suggest that SN-NM signals are elevated in FLR patients and similar to controls in TRS patients (i.e., URS and non-URS). Longitudinal studies are required to establish if SN-NM signals are a suitable biomarker to predict treatment response in schizophrenia.

Keywords: treatment-resistant schizophrenia, Dopamine, Neuromelanin-sensitive MRI

Disclosure: Nothing to disclose.

P809. Hospitalization in Early Phase Psychosis: Observations From a Learning Health Network

Jenifer Vohs*, Vinod Srihari, John Cahill, Stephan Taylor, Ashley Weiss, Chaudry Serena, Stephan Heckers, Steven Silverstein, Ivy Tso, Nicholas Breitborde, Alan Breier

Indiana University School of Medicine, Indianapolis, Indiana, United States

Background: Psychiatric hospitalization during the early phase of psychotic illness (EPP) has been linked to significant functional disruption, magnified personal suffering, and substantial economic burden. We examined hospitalizations among EPP participants enrolled in the Academic Community Early Psychosis Intervention Network (AC-EPINET) Learning Health Network (LHN) – a NIMH funded regional hub consisting of six academically-affiliate EPP clinics with nearly 700 patients currently enrolled.

Methods: All AC-EPINET participants with 12-months of complete hospitalization data were included in the present analysis (N = 234). Hospitalization was characterized in three ways (occurrence of an overnight stay in a hospital for a psychiatric reason, number of hospitalization episodes, and length of each hospital stay), and at three time points, six-month prior to program enrollment and at 6- and 12-months post enrollment. Beyond only hospitalization, participants were further assessed using the EPINET common core assessment battery (CAB), which includes a range of validated clinical instruments and hospitalization assessment (https://nationalepinet.org/core-assessment-battery-cab/).

Results: The preponderance of participants was male (70.4) and with at least a high school level education (71.7%). Fifty percent identified as White and 36.1% identified as Black. The primary diagnoses of participants were schizophrenia (42.7%), other unspecified psychotic disorder (25.4%), and schizoaffective disorder (12.5%). Duration of untreated psychosis, defined as time from the onset of the first psychotic symptom to first prescribed antipsychotic medication, was a Median duration of 5.4 months (IQR = 8.5). The number of hospitalizations were significantly reduced from 36.3% at baseline, to 13.7% at 6 months (OR 0.3 (95% CI 0.2-.4), p < .0001) and 12.4% at 12 months (OR 0.3 (95% CI 0.1-.4), p < .0001). Of those who were hospitalized, 74.1% at baseline, 68.8% at 6 months, and 65.6% at 12 months had only 1 hospitalization. The days spent hospitalized decreased slightly from baseline (M = 15.9, SD = 14.8) to 6 (M = 13.2, SD = 16.9) and 12 (M = 13.2, SD = 9.9) months. Symptoms, as measured by the Modified Colorado Symptom Index similarly reduced from baseline (M = 19.3, SD = 14.7) to 6 (M = 14.8, SD = 12.6) and 12 (M = 14.5, SD = 13.2)(p < .001).

Conclusions: Hospitalizations were significantly reduced following enrollment in the LHN. The shared assessment battery utilized across the network also enabled careful examination of demographics, symptoms and other variables that will be used to better understand what factors contribute to hospitalization, ultimately informing interventions to be deployed network wide. Consideration of confounds and examination of potential predictors of hospitalization, including symptoms, substance misuse, DUP, and medication adherence will be discussed.

Keywords: early psychosis, hospitalization, learning health network

Disclosure: Nothing to disclose.

P810. The Role of Cholinergic Interneurons in Amotivation in the 22q11DS Model of Schizophrenia

Mary Patton*, Brett Teubner, Kristen Thomas, Alexandra Trevisan, Sharon Freshour, Cody Ramirez, Jay Bikoff, Stanislav Zakharenko

St Jude Children’s Research Hospital, Memphis, Tennessee, United States

Background: Schizophrenia is a neurodevelopmental disorder partially characterized by notoriously intractable negative symptom categories. The severity of the negative symptom amotivation is negatively correlated with social and occupational functioning and is not currently met with effective treatment options. Advances in treatment are hampered by a lack of mechanistic understanding of the neural circuits underlying this symptom. Despite being implicated in motivational states, the dorsal striatum is largely overlooked when studying motivated behavior. Moreover, disruptions in acetylcholine signaling are associated with schizophrenia symptoms, but the role that cholinergic neurons play in amotivation remains unknown. Striatal cholinergic interneurons (CHIs) modulate glutamatergic signaling onto the principal medium spiny neurons of the dorsomedial striatum (DMS) and provide a basal cholinergic tone in the DMS by firing tonically. Here, we tested the involvement of the DMS in motivated behavior in a mouse model of 22q11 deletion syndrome (22q11DS), a genetic disorder in which 30% of individuals develop symptoms that are indistinguishable from idiopathic schizophrenia.

Methods: To test motivation in female and male 22q11DS mice and wildtype littermates we use the Progressive Ratio behavior task (on average N = 12/genotype). To investigate the circuits underlying amotivation we use ex vivo optogenetics with whole-cell patch-clamp electrophysiology (n = 16 cells/5 mice of both sexes per group) and manipulate these circuits in vivo using chemogenetics (N = 7/sex). To identify candidates underlying disruptions in dorsal striatal neuron firing properties, we use single nucleus RNA-sequencing (N = 4/genotype).

Results: We find that 22q11DS mice of both sexes display amotivation (Unpaired t test, **p = 0.0053) and exhibit weakened glutamatergic inputs to DMS from the parafascicular nucleus of the thalamus (Pf) (Unpaired t test, *p = 0.013). Mimicking the weakened Pf-DMS with chemogenetics in wildtype mice produces the amotivation phenotype (Paired t test, **p = 0.0075), causally implicating this circuit in amotivational states. Further, we show that there are substantially more spontaneously active CHIs in 22q11DS mice (Chi-square test, *p = 0.03) and blocking signaling through the M2 muscarinic acetylcholine receptor rescues the weakened Pf-DMS drive (Unpaired t test, **p = 0.0012). Finally, single nucleus RNA-sequencing identifies overexpression of the nonselective cation channel, Trpc3, in CHIs from 22q11DS mice (g:Profiler, adjusted *p = 0.014), pointing to a possible mechanism underlying hyperactive CHIs.

Conclusions: These findings are the first to implicate disruptions in dorsal striatal functioning, specifically the Pf drive of medium spiny neurons in amotivation symptoms arising in 22q11DS and point to intra-striatal acetylcholine signaling as a possible mechanistic linchpin in the etiology of amotivation.

Keywords: Reward, motivation, and anhedonia, schizophrenia negative symptoms, Dorsal striatum, cholinergic interneuron

Disclosure: Nothing to disclose.

P811. Antipsychotic Co-Initiation Vs. Monotherapy for Schizophrenia and Related Psychotic Disorders: A Systematic Review and Meta-Analysis

Emma A. van Reekum*, Kyle Fediuk, Marco Solmi, Jianping Zhang, Christoph U. Correll

McMaster University, Population Health Research Institute, Hamilton, Canada

Background: Current treatment approaches do not lead to meaningful recovery for most people with schizophrenia. Antipsychotic polypharmacy (APP) is frequently used as an augmentation strategy when monotherapy has failed, but past reviews have yielded conflicting results whether this strategy improves efficacy or increases adverse effects. No prior review has examined the benefits and harms of antipsychotic polypharmacy co-initiation (APP-CI), a different strategy aiming to expedite efficacy and recovery. By reducing the duration of continued psychosis, APP-CI could improve outcomes for people with schizophrenia, though this is not presently known.

Methods: We conducted a systematic review and meta-analysis that was pre-registered on Open Science Framework (osf.io/4rp2j). We searched Ovid Medline, Embase, PsychINFO, and Cochrane Central, from database inceptions to 07/31/2024 and without language restrictions. Grey literature was also searched. We included randomized trials that assessed any combination of APP-CI in relapsed people with schizophrenia and related psychotic disorders, compared to antipsychotic monotherapy initiation (AP-MI). Trials were ineligible if they initiated APP > 1 week apart (i.e., augmentation). Screening and data extraction were done in duplicate, as were risk of bias assessments, using the Cochrane risk-of-bias tool 2.0. Relative risks, mean differences (MDs), and 95% confidence intervals (95%CIs) were calculated using random effects pairwise meta-analyses. We then calculated absolute effects using control group event rates for baseline risk and presented the number-needed-to-treat if results were clinically significant. Clinical significance was determined by a difference of at least 5 people per 100 for binary data, 11 points on the Positive and Negative Syndrome Scale (PANSS) and 3 points on its subscales. Evidence certainty was assessed using the GRADE framework.

Results: Fifteen trials randomizing 1966 inpatients (mean age=32.3 ± 5.6 years, males=60.8%, schizophrenia diagnosis=99.4%, illness duration=6.8 ± 4.3 years, baseline PANSS Score=83.6 ± 4.7, baseline Clinical Global Impression-Severity Score=5.1 ± 0.1) to 14 different APP-CI combinations vs AP-MI during a mean 9.9 (median=8.0) weeks of follow-up were identified. Clozapine plus sulpiride (k = 3) and clozapine plus risperidone (k = 2) were the most used APP-CI combinations. All but one study used antipsychotics orally.

Meta-analytically, APP-CI may improve study-defined treatment response compared to AP-MI (k = 13, n = 1608, 56.0% vs 46.4%, 13 more per 100 people, 95%CI 5 to 23 more, number-needed-to-treat=10, p < 0.01, low certainty). However, AAP-CI was unlikely superior to AP-MI in PANSS symptom improvements (k = 5, n = 871), including total symptoms (MD = -1.9 point, 95%CI -4.7 to 0.9, p = 0.18), positive symptoms (MD = -0.58, 95%CI -1.32 to 0.17, p = 0.13), and negative symptoms (MD = -0.8, 95%CI -1.66 to 0.06, p = 0.07), all moderate certainty. APP-CI may not differ from AP-MI in the proportion of people requiring benzodiazepines (k = 4, n = 644, 46.2% vs. 45.6%, 2 more per 100 people, 95%CI 2 fewer to 7 more, p = 0.34) or discontinuing antipsychotics (k = 3, n = 603, 8.6% vs 10.1%, 2 fewer to 100 people, 95%CI 5 fewer to 5 more, p = 0.54), both low certainty. APP-CI also likely results in no difference in all-cause study dropout compared to AP-MI (k = 7, n = 1407, 24.2% vs 31.1%, 1 fewer per 100 people, 95%CI 6 fewer to 5 more, p = 0.72; moderate certainty).

We found no difference in the proportion of people with ≥1 adverse effect between APP-CI and AP-MI, however, the evidence was very uncertain (k = 8, n = 1376, 45.3% vs 51.4%, 2 fewer per 100 people, 95%CI 7 fewer to 5 more, p = 0.68). For specific adverse effects, we found with very low to low certainty that compared to AP-MI, APP-CI slightly increases the proportion of people reporting weight gain (k = 3, n = 501, 35.1% vs 27.4%, 8 more per 100 people, 95%CI 1 fewer to 18 more, p = 0.07), while reducing restlessness (k = 4, n = 813, 10.1% vs 16.8%, 5 fewer per 100 people, 95%CI 10 fewer to 4 more, p = 0.22), and may result in little to no difference in akathisia, dystonia, cardiac problems, sleep disturbance, sialorrhea, sedation, constipation, and extrapyramidal symptoms.

Conclusions: APP-CI may increase the number of relapsed people with schizophrenia who respond to treatment, without increasing antipsychotic discontinuation or most adverse effects, and without confounding by differences in benzodiazepine co-prescribing. However, there was very low to moderate certainty in these findings, so additional high-quality trials are needed. Although APP-CI may not affect study dropout, approximately 1 in 3 people dropped out over a mean of 10 weeks, leading to high risk of bias from missing outcome data in most trials. Effective strategies are needed to improve retention in schizophrenia trials.

Keywords: Schizophrenia Spectrum Illness, Schizophrenia, Antipsychotics, advantages of the combination of therapies, meta-analysis, randomized controlled trial

Disclosure: Nothing to disclose.

P812. Efficacy of Oral and Lai Aripiprazole Vs Paliperidone in Patients With Early-Phase Schizophrenia: Secondary Analysis From a Large-Scale, Open-Label, Randomized Trial (EULAST)

Sofia Pappa, Tal Elhasid, Jinyoung Park, Stuart Seidman*, Inge Winter-van Rossum, Rene Kahn, Wolfgang Fleischhacker, Linda Levi, Michael Davidson, John M. Davis, Mark Weiser

Chaim Sheba Medical Center, Ramat Gan, Israel

Background: Aripiprazole is the oldest member of a new generation of antipsychotics. Some, but not all, studies have found that partial dopamine agonists may be less efficacious than the classic, second-generation D2 dopamine antagonists. We analyzed data from a recently published open-label, pragmatic RCT in patients with early phase schizophrenia and compared time to all-cause discontinuation, time to hospitalization, and symptom reduction between aripiprazole and paliperidone.

Methods: This is a secondary analysis from the EULAST trial, a multicenter, RCT conducted across 15 European countries and Israel, on patients with up to 7 years of illness duration and randomized 1:1:1:1 to LAI paliperidone, LAI aripiprazole, or their corresponding oral formulations. The primary results of the EULAST study indicated no difference in time to all-cause-discontinuation (ACD) between long-acting and oral antipsychotics. This current study compares the oral and long-acting formulation of aripiprazole with paliperidone, alongside separate analyses for the four groups of LAI paliperidone, oral paliperidone, LAI aripiprazole, and oral aripiprazole.

Results: In total, 523 patients were randomly allocated, 255 to aripiprazole group (131 LAI and 124 oral) and 268 to the paliperidone group (141 LAI and 127 oral). Overall, both time to ACD (HR = ?, p =??) and time to hospitalization (HR = ?, p =??) did not significantly differ between the combined aripiprazole and paliperidone groups. However, comparisons of the four groups found that patients who took oral paliperidone had shorter times to ACD (Paliperidone oral vs. LAI (HR = 1.49, p = 0.05), aripiprazole oral (HR = 1.43, p = 0.09), and aripiprazole LAI (HR = 1.52, p = 0.05)) and trends to shorter times to hospitalization compared to the other groups (Paliperidone oral vs. paliperidone LAI (HR = 1.79, p = 0.08), aripiprazole oral (HR = 1.69, p = 0.12), and aripiprazole LAI (HR = 1.23, p = 0.49)). Similarly, there was no significant difference in changes in total PANSS scores between aripiprazole and paliperidone, however, patients treated with oral paliperidone showed less improvement in PANSS positive score compared to the other three groups (Paliperidone oral vs. paliperidone LAI (coef = -0.11, p = 0.02), aripiprazole oral (coef = -0.07, p = 0.13), and aripiprazole LAI (coef = -0.10, p = 0.03)). The trajectory of negative, general, and total scores did not show significant differences by group over time. When analyzing results for only participants who completed the entire study, similar results were found. There were no differences in time to hospitalisation or PANSS scores between the two combined groups or time to hospitalisation for the four groups; oral Paliperidone showed less improvement in PANSS positive scores (Paliperidone oral vs paliperidone LAI (coef = -0.12, p = 0.01), aripiprazole oral (coef = -0.07, p = 0.15), and aripiprazole LAI (coef = -0.09, p = 0.06)).

Conclusions: This is the first head-to-head RCT directly comparing oral and long-acting injectable aripiprazole and paliperidone. Overall, the real-world effectiveness of the combined aripiprazole and paliperidone groups was comparable in this study although, of the four groups, patients on oral paliperidone were more likely to stop the medication early and require hospitalization.

Keywords: Antipsychotic drugs, D2 dopamine antagonists, dopamine partial agonist, efficacy

Disclosure: Yes, I (or my spouse/partner) do have a financial relationship to disclose.

Financial Relationships Details Negev Capital, Founder, Self

P813. Altered Basal Ganglia Neurocircuitry in Unmedicated Patients With Schizophrenia Revealed During Working Memory Task Performance: An fMRI Back-Translation of Findings From the D2 Receptor Overexpressing Mouse Model

Philip Tubiolo, John Williams, Roberto Gil, Natalka Haubold, Yash Patel, Sameera Abeykoon, Zu Jie Zheng, Dathy Pham, Najate Ojeil, Kelly Bobchin, Eilon Silver-Frankel, Greg Perlman, Jodi Weinstein, Mark Slifstein, Christoph Kellendonk, Guillermo Horga, Anissa Abi-Dargham, Jared Van Snellenberg*

Stony Brook University Renaissance School of Medicine, Stony Brook, New York, United States

Background: Clinical neuroimaging studies of people with schizophrenia (PSZ) have produced extensive evidence of excess dopaminergic activity in the striatum, leading to the development of animal models intended to reproduce this neurobiological abnormality. One such animal model, the dopamine D2 receptor overexpressing (D2R-OE) mouse, which specifically overexpresses D2Rs in the striatum, exhibits a highly plastic change in basal ganglia neurocircuitry. These transgenic mice show a dramatic increase in bridging collaterals from dopamine D1 receptor (D1R) expressing medium spiny neurons (MSNs) to the globus pallidus externus (GPe), in which axons that ordinarily project exclusively to the globus pallidus internus (GPi) bifurcate and send a projection to GPe as well. This phenotype can be reversed within a few weeks, either by turning off the transgene or via treatment with haloperidol, while D2R knockdown mice show a decrease in these bridging collaterals.

To determine whether this alteration in basal ganglia neurocircuitry could be occurring in PSZ, we used functional magnetic resonance imaging (fMRI) to examine functional connectivity between the dorsal (pre-commissural) caudate (DCa) and the GPe in a sample of unmedicated PSZ, both during rest and while performing a working memory (WM) task. The WM task was employed as a condition in which excitatory inputs from the cortex were driving activity in DCa MSNs at the time that DCa-GPe connectivity was measured.

Methods: A sample of 37 PSZ and 30 matched healthy controls (HC) completed 30 minutes of resting state fMRI, with a subset of 29 PSZ and 29 HC also performing a self-ordered WM task during fMRI scanning with a multiband factor of 6. Following standard data preprocessing, task events were modeled and the estimated neural response to task was subtracted from the WM task data. High-motion volumes were identified using the Multiband Censoring Optimization Tool and censored from the dataset after mode 1000 normalization and linear detrending, and before bandpass filtering (0.08-0.009 Hz). For the WM data, “task-on” volumes were defined as all volumes acquired two or more seconds after onset and within four seconds after the end of a trial (to allow for the rise and fall of the hemodynamic response). Only “task-on” volumes were retained for analysis. Regions of interest (ROIs) for DCa, GPe, and GPi were obtained from individual subject T1s using automated segmentation: striatal ROIs were identified by a convolutional neural network model trained to reproduce five manually-drawn striatal subregions, and GPe and GPi ROIs were obtained from DBSegment. Connectivity between ROI pairs was calculated for each subject as the partial correlation between the mean signal in all voxels within each ROI, controlling for 24 motion parameters and white matter, cerebrospinal fluid, and global signals. The primary analysis of interest was the difference between PSZ and HCs in average ipsilateral DCa-GPe connectivity, for which we had 80% power to detect a Cohen’s d of 0.75. We also tested the relationship between DCa-GPe connectivity and WM task performance in a linear regression model controlling for age, self-reported gender, diagnosis, and diagnosis-by-connectivity interaction, for which we had 80% power to detect a Cohen’s f2 of 0.14. As a secondary analysis, all previous tests were repeated using DCa-GPe connectivity recalculated while additionally controlling for GPi activity.

Results: PSZ showed increased DCa-GPe connectivity compared to HCs during WM (d = 0.64, P = 0.024), but not at rest (d = -0.26, P = 0.835). The WM result remained significant in a multiple regression model controlling for age and gender (β* = 0.286, P = 0.025; notably, female participants also showed increased DCa-GPe connectivity: β* = -0.278, P = 0.034). Post-hoc tests of all other basal ganglia ROI pairs revealed no significant differences, with or without correction for multiple comparisons. Finally, across the entire sample, higher DCa-GPe connectivity during WM was associated with poorer performance on the WM task (β* = -0.30, P = 0.048). Increased age was also associated with poorer task performance (β* = -0.29, P = 0.04).

When estimating DCa-GPe connectivity while controlling for GPi activity, PSZ still showed increased DCa-GPe connectivity compared to HCs in both the t-test (d = 0.66, P = 0.006) and multiple regression model (β* = 0.307, P = 0.016; no effect of gender) during the WM task, and not during rest (d = -0.31, P = 0.69). Additionally, higher DCa-GPe connectivity during WM remained associated with poorer WM task performance (β* = -0.313, P = 0.036), as did increased age (β* = -0.279, P = 0.047).

Conclusions: These results provide first-in-human evidence that increased bridging collaterals from DCa to GPe could be a neuropathological feature of unmedicated PSZ, which is associated with impaired cognitive task performance. Although current technology for in vivo imaging of human subjects cannot directly measure the density of these bridging collaterals, our test of DCa-GPe connectivity was directly motivated by the finding of increased bridging collaterals in the D2R-OE mouse, which was created to model the excess D2R stimulation known to occur in schizophrenia. Thus, this was a strong a priori test of the hypothesis that excess bridging collaterals are present in PSZ, which warrants further investigation in larger samples, association with measures of dopaminergic function, and examination of whether this phenotype normalizes with antipsychotic treatment.

Keywords: Schizophrenia, dopamine, functional MRI, task connectivity, basal ganglia

Disclosure: Nothing to disclose.

P814. Lesions That Cause Psychosis Map to a Common Brain Circuit in the Hippocampus

Andrew Pines*, Summer Frandsen, William Drew, Garance Meyer, Calvin Howard, Stephan Palm, Frederic Schaper, Christopher Lin, Konstantin Butenko, Michael Ferguson, Maximilian Friedrich, Jordan Grafman, Ari Kappel, Clemens Neudorfer, Natalia Rost, Lauren Sanderson, Ona Wu, Isaiah Kletenik, Jacob Vogel, Alexander Cohen, Andreas Horn, Michael Fox, David Silbersweig, Shan Siddiqi

Brigham and Women’s Hospital, Boston, Massachusetts, United States

Background: Identifying the anatomy causally involved in psychosis could inform therapeutic neuromodulation targets for schizophrenia.

Methods: This case-control study used lesion network mapping to identify functional connections of published cases of lesions causing psychosis compared with control lesions unassociated with psychosis. 153 cases of lesion-induced psychosis were analyzed. Included subjects had documented brain lesions associated with new-onset psychotic symptoms without prior history of psychosis. Control cases included 1291 patients with lesions not associated with psychosis. Generalizability across lesional datasets was assessed using an independent cohort of 181 patients with brain lesions who subsequently underwent neurobehavioral testing. We then conducted a multiple linear regression analysis on outcomes in an RCT for clinical high-risk patients, evaluating for an interaction effect between the TMS site - Psychosis Circuit connectivity and conversion outcomes while controlling for age, education, and sex.

Results: 153 lesions from published cases were determined to be causal of psychosis, 42 of which were described as “schizophrenia” or “schizophrenia-like”. Lesions that caused psychosis mapped to a common brain circuit defined by functional connectivity to the subiculum of the hippocampus (84% functional overlap, pFWE < 5 x 10-5). At a lower statistical threshold (75% > overlap, pFWE < 5 x 10-4), this circuit included the ventral tegmental area, retrosplenial cortex, lobule IX and dentate nucleus of the cerebellum, and the mediodorsal and midline nuclei of the thalamus. This circuit was consistent when derived from only “schizophrenia-like” cases (spatial r = 0.98). We repeated these analyses after excluding lesions directly intersecting the hippocampus (n = 47) and found a consistent functional connectivity profile with the subiculum remaining the center of connectivity, demonstrating a circuit-level effect. In an independent observational cohort of patients with penetrating head trauma (n = 181), lesions associated with symptoms of psychosis exhibited significantly similar connectivity profiles with the lesion-derived psychosis circuit. Voxels in the rostromedial prefrontal cortex (rmPFC) are highly correlated with this psychosis circuit, suggesting the rmPFC as a promising TMS target for schizophrenia. Results from the linear regression analysis showed that patient outcomes (psychotic episode or none at one year follow up) depends on TMS site - Psychosis Circuit connectivity (β = -1.6129, SE = 0.48598, p = 0.0017).

Conclusions: Lesions that cause psychosis affect a common brain circuit in the hippocampus. These results can help inform therapeutic neuromodulation targeting.

Keywords: Schizophrenia (SCZ), TMS, neuroanatomy

Disclosure: Nothing to disclose.

P815. Understanding Interoception and Emotion in Schizophrenia Using Psychophysiology and Heartbeat-Evoked Potential

Beier Yao*, K. Eve Lewandowski, Mei-Hua Hall

Harvard Medical School, McLean Hospital, Belmont, Massachusetts, United States

Background: Interoception refers to the processing, integration, interpretation, and regulation of bodily signals by the brain. It is crucial to motivational and affective functioning such as subjective experience of arousal (i.e., intensity of emotion). Previous studies found that people with schizophrenia reported higher arousal than healthy controls in response to neutral stimuli, and higher levels of aversive emotion in response to positive and neutral stimuli. However, it’s unclear if this difference in subjective experience of emotion is due to altered interoceptive processing of normative bodily responses, or normative interoceptive processing of elevated baseline arousal state/negative emotion. This study aimed to investigate interoceptive processing and emotional experience in schizophrenia by examining heart rate changes and heartbeat-evoked potential (HEP) during varying emotional states. HEP is an event-related potential (ERP) that is time-locked to heartbeats and reflects cortical processing of heartbeats. HEP can be modulated by attention, interoceptive ability, psychiatric conditions, and arousal state. We hypothesized a potential disconnect between physiological signals, HEP, and subjective emotional experience in schizophrenia.

Methods: Data collection and analysis are ongoing. For the preliminary sample, 19 participants with schizophrenia or schizoaffective disorder (SZ; age: 34.47 ± 9.55; 14M/5F) and 17 healthy controls (HC; age: 32.18 ± 8.73; 10M/7F) have completed the study. Participants viewed 96 images while their electroencephalogram (EEG) and electrocardiogram (ECG) were being recorded. They also rated each image on subjective feelings of valence (unpleasant, neutral, pleasant) and arousal after each image. To examine heart rate changes during image viewing, R peaks were identified and used to derive inter-beat interval (IBI) series. Heart rate analysis included the IBI concurrent with the onset of image presentation (IBI0), 2 IBIs before image presentation (IBI-2, IBI-1), and 4 IBIs after image onset and during image viewing (IBI + 1, IBI + 2, IBI + 3, IBI + 4). The second IBI before image onset (i.e., IBI-2) was used as baseline heart rate to derive changes in IBIs associated with image viewing. HEP was identified by segmenting EEG data into 125-600 ms epochs after each R peak during image viewing. Group comparisons of subjective rating and heart rate changes were conducted on trial-level data using multilevel modeling. HEP were averaged within different stimuli conditions (negative, neutral, positive) per participant for group comparisons.

Results: HC and SZ were matched on age and sex. For subjective rating during image viewing, there were significant group by condition interaction effects on both valence, F(2, 3416) = 12.53, p < .001, and arousal, F(2, 3416) = 3.45, p = 0.03. Specifically, SZ rated the negative images as less unpleasant, t(3416) = 3.21, p = 0.001, and less intense than HC, t(3416) = -2.62, p = 0.009. For heart rate changes, there was a significant increase in the duration of IBI + 3, F(2, 3405) = 7.21, p = 0.0008, and IBI + 4, F(2, 3405) = 7.61, p = 0.0005, for negative images. In other words, participants had a transient decrease in heart rate only when viewing negative images. There was no group or group by condition interaction effects. When looking at the association between subjective rating and heart rate change, we found a significant effect of valence rating, F(1, 3422) = 9.89, p = 0.002, and a significant group by valence rating interaction effect, F(1, 3422) = 5.05, p = 0.02, on the duration of IBI + 4. Specifically, as valence rating decreases (i.e., becomes more negative), the IBI + 4 duration increases in HC, t(3421) = -3.73, p = 0.0002, but this association is not significant in SZ, t(3423) = -0.65, p = 0.52. Similarly, we found a significant effect of arousal rating on the duration of IBI + 4, F(1, 3266) = 9.17, p = 0.002, but no group by arousal rating interaction effect, F(1, 3266) = 0.82, p = 0.37. Lastly, for HEP amplitude, we found a positive shift in SZ relative to HC when viewing neutral images across many frontal-central channels, p = 0.001 - 0.02.

Conclusions: The finding of a transient decrease in heart rate in response to negative images is consistent with previous literature. Though there was no group difference in heart rate changes, people with SZ rated the negative images differently from HC. Moreover, HC’s valence rating of each image was predictive of their heart rate change while viewing the corresponding image, but this relationship was not significant in SZ. In other words, SZ exhibited normative heart rate changes in response to negative images, but such physiological signals were less informative when reporting their subjective emotional experience, resulting in different ratings. These findings suggest an alteration in how people with SZ use heartbeat signals to inform their emotional experience. Further supporting this interpretation is the positive shift in HEP amplitude when SZ viewed neutral images, despite no group differences in either subjective rating or heart rate changes in response to neutral images. In sum, preliminary results from the current study suggest that the differences in emotional experience in people with SZ may be a result of altered cortical processing instead of abnormal physiological responses. Studying interoception using a multimodal approach may reveal important insights into the underlying mechanisms of alterations in emotional experience in schizophrenia.

Keywords: Schizophrenia (SCZ), interoception, EEG, psychophysiology, emotion

Disclosure: Nothing to disclose.

P816. Auditory Thalamocortical Hyperconnectivity is Associated With Positive Symptom Severity in Unmedicated People With Schizophrenia

John Williams*, Philip Tubiolo, Roberto Gil, Natalka Haubold, Sameera Abeykoon, Zu Jie Zheng, Dathy Pham, Najate Ojeil, Kelly Bobchin, Eilon Silver-Frankel, Greg Perlman, Jodi Weinstein, Mark Slifstein, Guillermo Horga, Anissa Abi-Dargham, Jared Van Snellenberg

Stony Brook University Renaissance School of Medicine, Stony Brook, New York, United States

Background: Converging evidence from both animal models and functional neuroimaging studies has strongly implicated dysfunction of the thalamus and associated circuits in the pathophysiology of schizophrenia (SCZ). While positive symptoms of SCZ are potentially devastating for patient outcomes, the role of thalamic circuit dysfunction in their etiology is unclear. In an animal model of SCZ, the 22q11.2 microdeletion syndrome mouse, synaptic transmission from auditory thalamus to primary auditory cortex (AC) was specifically impaired, with stimulation of the medial geniculate nucleus (MGN) resulting in weaker excitatory post-synaptic currents. In contrast, human resting-state (RS) functional magnetic resonance imaging (fMRI) studies of people with SCZ (PSZ) have found a generalized increase in resting-state functional connectivity (RSFC) between the thalamus and sensorimotor cortex, which may reflect dysfunction of cortico–basal ganglia–thalamo–cortical (CBGTC) loops.

Here, we aimed to determine whether alterations in RSFC between auditory thalamus and AC are specifically associated with the severity of positive symptoms. We conducted an fMRI study of unmedicated PSZ and healthy controls (HC) who completed RS fMRI and a Sensory Thalamic Localizer (TL) fMRI task, which identifies individualized auditory thalamic functionally-defined ROIs (fROIs), including MGN. Additionally, we explored whether auditory corticothalamic connectivity disruptions could constitute more generalized CBGTC loop abnormalities by assessing RSFC between cortex and dorsal caudate (DCa).

Methods: Unmedicated PSZ (n = 82) and matched HC (n = 55), ages 18-60, completed RS fMRI (RS only, N = 137), across two sites: New York State Psychiatric Institute (NYSPI) and Stony Brook University (SBU). MGN fROIs were additionally available for 55 PSZ and 46 HC who additionally completed the TL fMRI task (RS + TL, N = 101). Symptom severity was assessed for both PSZ and HC using the Positive and Negative Syndrome Scale (PANSS) Positive (PS), Negative (NS) and General (GS) subscales. PANSS scores were available for 78 PSZ and 51 HC from the RS only sample (N = 129), and 52 PSZ and 42 HC from the RS + TL sample (N = 94).

Data at both sites were acquired using a 3 T MR scanner, 2 mm isotropic voxels, and a multiband factor of 6, with a TR of 850 ms at NYSPI and 800 ms at SBU. Participants at both sites were asked to complete 4 runs each of RS (NYSPI: 7 min 34 s, SBU: 7 minutes 38 s) and TL (3 min 46 s) fMRI. DCa ROIs were produced from each participant’s T1-weighted MR image using a convolutional neural network model.

The TL task is publicly available and has been described previously in detail. In summary, participants were presented with alternating visual or auditory stimuli. Individual participant fROIs were produced using TL fMRI data by extracting regions of posterior thalamus showing modality-specific connectivity to AC. For each participant, MGN and DCa seed RSFC images were generated from RS data as Pearson’s partial correlations, controlling for head motion (Friston-24 model) and mean white matter, cerebro-spinal, and global signals, after censoring volumes acquired during excess motion.

Group-level seed RSFC analyses assessing for associations with SCZ diagnosis and positive symptom severity (PANSS PS) were performed using generalized linear models (GLMs) in Permutation Analysis of Linear Models using 10,000 permutations and threshold-free cluster enhancement; results were family-wise error rate corrected and thresholded at α = 0.05. MGN seed RSFC associations were assessed within an AC mask generated by selecting all Cole-Anticevic Brain Network Partition “auditory” parcels dilating by 4 mm; DCa seed RSFC associations were assessed in whole cortex. SCZ associations were assessed using a GLM with a regressor for diagnosis. PANSS PS associations assessed using a GLM with regressors for diagnosis, PS, NS, and GS. As a confirmatory analysis, PANSS PS associations were also assessed separately in the PSZ subsample using a GLM with regressors for PS, NS, and GS. All GLMs additionally included covariates for age, sex, and handedness.

Results: Seed RSFC analyses localized a region of bilateral superior temporal gyrus (STG) showing an association between MGN hyperconnectivity and PANSS PS subscores (40 grayordinates; N = 94); in PSZ separately (n = 52), this association remained significant in 14 grayordinate subregion of the full sample result in left STG. No associations were found between MGN seed RSFC and SCZ (N = 101). DCa seed RSFC localized regions showing associations between DCa hyperconnectivity and SCZ, including STG, superior temporal sulcus, central sulcus, temporal pole, inferior parietal lobule, postcentral gyrus, and right insula (3,975 grayordinates; N = 137); no associations were found between DCa seed RSFC and PANSS PS (N = 129). 28 STG grayordinates showed both associations between MGN RSFC and PANSS PS and associations between DCa seed RSFC and SCZ.

Conclusions: These results suggest that the severity of positive symptoms in PSZ is related to thalamocortical hyperconnectivity between MGN and associative AC, but not hypoconnectivity between MGN and primary AC. Associative AC also showed substantial regions wherein hyperconnectivity with DCa was predictive of SCZ diagnosis, but not positive symptom severity. These results reinforce existing evidence of CBGTC loop dysfunction in SCZ, and thalamocortical hyperconnectivity as a potential mechanism underlying positive symptoms more specifically.

Keywords: Schizophrenia, Unmedicated, Resting-state fMRI, Positive Symptoms, Auditory Thalamus

Disclosure: Nothing to disclose.

P817. Long-Term Potentiation (LTP)-Like Cortical Plasticity as a Predictor of Response to Cognitive Training in Schizophrenia

Holly Hamilton*, Renée Dembo, Cassandra Marzke, Spero Nicholas, Brian Roach, Judith Ford, Daniel Mathalon

University of Minnesota and Minneapolis VA Health Care System, Minneapolis, Minnesota, United States

Background: Converging evidence suggests that impaired cortical plasticity may be a key pathophysiological mechanism in schizophrenia that contributes to cognitive dysfunction, which is a core feature of the illness and major determinant of functional disability. In particular, long-term potentiation (LTP), a basic mechanism of experience-dependent synaptic plasticity that depends on glutamatergic transmission at N-methyl-D-aspartate receptors (NMDARs) that is widely considered a primary substrate of learning and memory, is predicted to be compromised in schizophrenia based on NMDAR hypofunction models of the illness. Accordingly, impaired LTP may limit the potential benefit of interventions targeting poor cognition that rely on new learning, such as cognitive training. Using an electroencephalography (EEG)-based LTP-like cortical plasticity paradigm in which repetitive visual stimulation produces lasting enhancement of visual evoked potentials (VEPs), we examined whether the integrity of these plasticity mechanisms predicts gains following cognitive training.

Methods: Schizophrenia and healthy comparison participants (men and women) underwent EEG recording during a visual cortical plasticity paradigm. VEPs were elicited by checkerboard stimulus reversals at baseline and after exposure to repetitive visual stimulation (2 Hz, 10 minutes) designed to induce VEP potentiation. Changes in P1 and N1b VEP component amplitudes from baseline to post-visual stimulation were compared between groups. Cognitive function was also assessed before and after participants with schizophrenia completed 10 sessions of visual computerized cognitive training.

Results: Preliminary analyses of data from 22 participants with schizophrenia (86% men) and 24 healthy individuals (75% men) indicated that visual stimulation produced potentiation of P1 and N1b in both groups, reflected by main effects of time (p = 0.03 and p = 0.002, respectively) that did not interact with group (ps > .09). Greater N1b potentiation at baseline predicted gains in cognitive function in participants with schizophrenia (n = 14) who completed 10 sessions of cognitive training (β = -.584, p = 0.02).

Conclusions: Results demonstrate LTP-like plasticity among individuals with schizophrenia and may provide a mechanistic account for variability in response to cognitive training interventions. If replicated, results could facilitate development in personalized treatment approaches that prescribe cognitive training to those who may benefit most. Moreover, results could help identify new neurophysiological targets for the development of treatments aimed at improving plasticity, thereby increasing the potential to remediate cognitive deficits.

Keywords: schizophrenia, cortical plasticity, electroencephalography

Disclosure: Nothing to disclose.

P818. The Distinct Pharmacological Profile of Emraclidine, a Selective Positive Allosteric Modulator of the M4 Muscarinic Acetylcholine Receptor

Ellie Klein*, Angela Quinn, May Fern Toh, Leonard Te, Gwyneth Catipay, Deborah L Smith, Kushali Gupta, Jamie Hatch, Hanh Nguyen, Philip Iredale, Sokhom Pin

Cerevel Therapeutics, Cambridge, Massachusetts, United States

Background: The M4 muscarinic acetylcholine receptor (mAChR) is one of 5 G-protein coupled acetylcholine receptor subtypes and is highly expressed in brain regions implicated in the symptoms of schizophrenia, where it may act to reduce dopamine release. The present study aimed to understand the unique binding and functional profile of the highly selective M4 positive allosteric modulator (PAM) emraclidine using multiple probes and evaluating various signaling events and pathways at steady state, as well as multiple time points. Functional profiling of emraclidine at the M4 receptor confirmed selectivity for M4 over other mAChRs and signaling bias to the G-protein pathway.

Methods: The binding profile and kinetics of emraclidine at M4 was characterized alongside multiple additional M4 binding compounds using radioligand competitive binding assays with multiple probes. Competitive binding assays were performed on CHO-K1 M4 overexpressing cells in the presence of the M4 PAM [3H]MK-6884 (Ki [SE], 17.0 [3.3] nM) and the nonselective muscarinic orthosteric antagonist [3H]N-methylscopolamine (Ki > 1000 nM). Direct agonist activation of G-protein was investigated using the [35S]GTPɣS assay in the scintillation proximity assay (SPA) format, a homogenous assay measuring the proximity of radioligand to the receptor bound to SPA beads. HTRF assays were used to measure the second messenger molecules cAMP (M2 or M4 receptors) or IP1 (M1, M3, or M5 receptors). Activation of specific G-protein subtypes was analyzed by Domain Therapeutics, Inc. (Montreal, Canada) using bioluminescence resonance energy transfer (BRET) assays. The structure of emraclidine bound to M4 (emraclidine-M4) was determined by single particle cryo-EM to a global resolution of 3.3Å.

Results: Functional profiling of emraclidine at the M4 receptor confirmed selectivity for M4 over other mAChRs and signaling bias to the G-protein pathway. The binding profile and kinetics of emraclidine at M4 were characterized alongside multiple additional M4 binding compounds using radioligand competitive binding assays. Binding affinities (Ki) of emraclidine in competitive binding assays performed in the presence of the M4 PAM [3H]MK-6884 or the nonselective muscarinic orthosteric antagonist [3H]N-methylscopolamine confirmed emraclidine primarily behaves as a PAM at M4. Additionally, scopolamine completely inhibited the functional activation of M4 by 80% effective concentrations of both emraclidine and ACh with similar potency, confirming that M4 receptor activation by emraclidine is mediated via the orthosteric binding site and can occur independently of ACh at high emraclidine concentrations. Cryo-electron microscopy 3D structure confirmed that emraclidine binds the M4 receptor at the common extracellular vestibule allosteric binding pocket in a similar manner to other published structures of allosteric binders.

Conclusions: Overall, this study confirms the mechanism by which emraclidine activates the M4 receptor and highlights various mechanisms by which M4-targeting compounds, including partial agonists, full agonists, and PAMs, interact with M4. These data provide a foundational understanding for how various targeting approaches at M4 may activate different signaling pathways, resulting in different downstream effects that may potentially be suited for specific therapeutic applications.

Keywords: Schizophrenia (SCZ), Muscarinic M4, allosteric modulator, acetylcholine, receptor binding

Disclosure: Nothing to disclose.

P819. Inflammatory Cytokine Profiles in Gluten-Sensitive Individuals With Schizophrenia

Valerie Harrington*, Donna Dadkhoo, William W. Eaton, Daniela Cihakova, Monica V. Talor, Bruce Patterson, Matthew Glassman, Deanna L. Kelly

University of Maryland School of Medicine, Baltimore, Maryland, United States

Background: Schizophrenia (SZ) is a prevalent psychiatric disorder that remains incompletely understood. Recent studies indicate stratifying patients with SZ based on IgG antibodies against the wheat gluten protein gliadin. Roughly 1 in 3 non-celiac SZ patients are positive for these anti-gliadin antibodies (AGA); however, only about 10% of controls demonstrate this gluten sensitivity (GS), suggesting a potential role of GS in these symptomatic AGA + SZ individuals. Additionally, these AGA IgG have been found in the central nervous system of AGA+ individuals with SZ but not in AGA+ controls, suggesting a potential role for neuroinflammation in this SZ subgroup. Despite this possible connection, inflammatory profiles within this subset have not been previously described in depth.

Methods: Outpatient human subjects with SZ (n = 631) and controls (n = 70) were recruited and consented to blood draw and serum analysis at Maryland Psychiatric Research Center in Catonsville, MD. All participants were grouped into subsets of 18-34 (n = 190 for SZ, n = 30 for controls), 35-49 (n = 191 for SZ, n = 17 for controls), and 50-65 (n = 196 for SZ, n = 23 for controls) years of age. Participants included both males (n = 404 for SZ, n = 32 for controls) and females (n = 172 for SZ, n = 38 for controls). Antibodies and cytokines were measured via enzyme-linked immunosorbent assays. AGA positivity was defined as serum AGA IgG > 20 IU/mL. Celiac disease was defined as TtG IgG > 20 IU/mL. Individuals with SZ and controls with celiac disease (n = 11 and 2, respectively) were excluded from further analysis. Significance between groups was determined using two-way ANOVA.

Results: Serum inflammatory profiles significantly differed by diagnosis, age, and AGA positivity. AGA IgG levels were significantly higher in participants with SZ under 35 years old compared to individuals 35-49 (p = 0.0004) and 50-65 (p < 0.0001) years old, but there were no AGA IgG differences associated with age in controls. AGA IgG and AGA IgA were significantly correlated in both SZ (p < 0.0001, r = 0.2968) and controls (p < 0.0001, r = 0.5158), and control and SZ AGA+ participants (n = 22 and 219, respectively) had significantly greater AGA IgA compared to AGA- participants (n = 48 and 358, respectively; p < 0.0001). Several inflammatory cytokines were significantly elevated in AGA IgG+ participants with SZ, including GM-CSF (p = 0.0009), IFNγ (p < 0.0001), IL-17A (p = 0.0023), IL-1β (p < 0.0001), IL6 (p < 0.0001), and TNF (p < 0.0001); however, AGA+ controls showed no significant cytokine difference from AGA- controls. Serum cytokine and AGA levels did not significantly differ by sex in controls or individuals with SZ.

Conclusions: These findings not only reveal an inflammatory profile in AGA+ individuals with SZ distinct from AGA- SZ or controls but also suggest potential variations by age. The correlation between AGA IgG and IgA hints at systemic and mucosal inflammation, as IgA is the primary antibody class secreted at mucosal surfaces like the gut. Given the established link between neuroinflammation and SZ pathogenesis, understanding the mechanisms driving increased inflammation in this subgroup could pave the way for more effective future therapeutic interventions, offering the potential for improved patient outcomes.

Keywords: gluten, inflammation in schizophrenia, age differences, dietary interventions, Pro-inflammatory cytokines

Disclosure: Nothing to disclose.

P820. An Aberrant Relationship Between Striatal Dopamine Decarboxylation Rate and Prefrontal Gaba Levels is Related to Poor Treatment Outcome in initially Antipsychotic-Naïve First-Episode Psychosis Patients: A Combined PET-MRS Study

Kirsten Bojesen*, Karen Marie Sandø Ambrosen, Anne korning Sigvard, Mette Ødegaard Nielsen, Albert Gjedde, Yoshitaka Kumakura, Lars Thorbjørn Jensen, Dan Fuglø, Bjørn H. Ebdrup, Egill Rostrup, Birte Glenthoj

Center for Neuropsychiatric Schizophrenia Research, Glostrup, Denmark

Background: Preclinical models of schizophrenia suggest that aberrations in the relationship between the dopaminergic, glutamatergic and GABAergic neurotransmitter systems are more central to illness pathophysiology than single neurotransmitter disturbances. In support, we recently reported that a combination of striatal dopamine, prefrontal GABA and thalamic glutamate brain levels can classify antipsychotic-naïve first-episode patients with psychosis (FEP) from healthy controls (HC) with an accuracy of 83.7%. In this model, we found that especially an interaction between measures of striatal dopamine and prefrontal GABAergic activity contributed to characterize FEP from HC, and that an association between striatal dopamine and thalamic glutamate levels contributed to characterize FEP as well. However, it is unknown how the relationship between dopamine synthesis in nucleus accumbens (NAcc) and prefrontal GABA levels differs between FEP and HC, and if an aberrant relationship between neurotransmitters are related to treatment outcome. Last, the direction of the relationship between striatal dopamine synthesis and thalamic glutamate levels has not been investigated. Therefore, we tested if the relationship between striatal dopaminergic activity and prefrontal GABA levels was aberrant in FEP compared with HC, and if the aberrant relationship was related to treatment outcome. Last, we explored the correlation between striatal dopaminergic synthesis and thalamic glutamate levels.

Methods: We recruited 29 antipsychotic-naïve patients with FEP, and 31 HC matched on age, sex, and parental socioeconomic status and followed participants up after six weeks. Patients were treated with aripiprazole as monotherapy (9.4 ± 3.7mg). Striatal dopaminergic decarboxylation rate (k3) in NAcc was estimated with [18F]-FDOPA PET, and brain levels of GABA in anterior cingulate cortex (ACC) and glutamate levels in left thalamus were measured with magnetic resonance spectroscopy (MRS) on a 3T scanner. Psychopathology was evaluated using the PANSS score, and treatment outcome was estimated as change in the PANSS positive score between baseline and the six weeks follow-up. Multiple linear regression was used to test interactions and associations between striatal k3 (dependent variable) and GABA levels in dACC (independent variable) or glutamate levels in left thalamus (independent variable), as well as the association between treatment outcome (dependent variable) and the striatal k3*prefrontal GABA interaction (independent variable).

Results: The association between k3 in NAcc and prefrontal GABA levels:

The relationship between k3 in NAcc and prefrontal GABA levels differed between FEP and HC (group*GABA: p = 0.002) due to a negative relationship in HC (β = -0.15, p = 0.03) but a positive relationship in FEP (β = 0.15, p = 0.04).

The association between k3*GABA and treatment outcome:

Treatment outcome was associated with the interaction between k3 in NAcc and GABA levels in ACC (k3*GABA: p = 0.02) due to a higher degree of aberration between the neurotransmitters being related to less improvement in PANSS positive score after six weeks.

The association between k3 in NAcc and thalamic glutamate levels:

The interaction between k3 in NAcc and thalamic glutamate did not differ between groups (group*thalamic glutamate insignificant) but there was a positive association between k3 in NAcc and thalamic glutamate levels at baseline in both FEP and HC (thalamic glutamate: p = 0.04).

Conclusions: The data reveal that higher prefrontal GABA levels seem to inhibit striatal dopaminergic activity in HC whereas the reverse relationship is seen in antipsychotic-naïve FEP. Moreover, degree of aberration between prefrontal GABA and striatal dopaminergic function is related to treatment outcome in FEP. Last, higher striatal dopaminergic activity seems related to higher levels of thalamic glutamate in both antipsychotic-naïve FEP and HC. The findings support that combined neurotransmitter disturbances are central to the pathophysiology of psychotic disorders. This implies that development of novel therapeutics may benefit from shifting focus from single to combined neurotransmitter disturbances.

Keywords: Dopamine, Cortical GABA, Antipsychotic-naïve first-episode psychosis, F-18 PET Imaging, 1H MRS

Disclosure: Lundbeck Pharma AS: Honoraria (Self).

P821. Habenular White Matter is Altered in Schizophrenia

John Hassmann, Elisa Ambrosi, Katrina Rufino, Fabrizio Piras, Gabriele Sani, Gianfranco Spalletta, Ramiro Salas*

Baylor College of Medicine, Houston, Texas, United States

Background: Schizophrenia (SZ) is a debilitating psychiatric disorder characterized by symptoms such as disordered thought, delusions, reduced motor responsiveness, and cognitive impairment. The habenula is a small, phylogenetically-conserved, midbrain structure important for avoidance behavior and for encoding negative emotional outcomes. There is evidence of a possible role of the habenula in SZ, but the role of habenular white matter has not been studied.

Methods: Twenty subjects with diagnosed SZ and 23 healthy controls (HC) were studied using diffusion tensor imaging (DTI). DTI was analyzed using a small region of interest approach. Participants were administered the Positive and Negative Syndrome Scale for Schizophrenia (PANSS) and the Rey-Osterrieth Complex Figure (ROCF) test. Analysis of covariance (ANCOVA) was used to compare habenula fractional anisotropy of participants with SC to HC subjects while controlling for age, sex, and tobacco use. Pearson correlations and moderation analyses were conducted.

Results: ANCOVA revealed a significant difference in habenula fractional anisotropy (FA) between SZ and HC groups after controlling for age, gender, and tobacco use [F(1,43) = 17.75, p < .001, partial η2 = 0.32]. SZ had lower average habenula fractional anisotropy (M = 0.29, SD = 0.04), compared to healthy controls (M = 0.34, SD = 0.03). Rey cognitive scores correlated with habenula FA in patients (r = 0.53, p = 0.019) but not in controls.

Conclusions: The habenula may play a role in SZ. FA of habenular streamlines was smaller in SZ than HC, and correlated with SZ symptoms. Thus, the habenula may provide a druggable target for the treatment of SZ symptoms.

Keywords: Schizophrenia (SCZ), habenula, Diffusion Tensor Imaging (DTI)

Disclosure: Nothing to disclose.

P822. Examining the Mediating Role of Inflammation on the Relationships Between Sleep and Cognitive Functioning Among People Living With Schizophrenia and Non-Psychiatric Comparison Participants

Ellen Lee*, Sonia Ancoli-Israel, Atul Malhotra, Xin Tu, Steve Cole, Michael Irwin, Lisa Eyler

University of California, San Diego, La Jolla, California, United States

Background: Sleep disturbances affect up to 80% of people living with schizophrenia (PLWS), and impact cognitive aging in this highly vulnerable group. Inflammation is a key biological mechanism that links sleep with poor cognitive outcomes. While poor subjective sleep quality has been linked to elevated levels of pro-inflammatory biomarkers among PLWS, few studies have examined the associations of objective sleep measures with inflammatory biomarkers among PLWS.

We present cross-sectional data from a longitudinal study of PLWS and non-psychiatric comparison participants (NCs). Our hypotheses were: 1) Objective sleep measures would be worse in PLWS than in NCs. 2) Worse sleep measures would be related to elevated inflammatory biomarker levels in both groups. 3) Worse sleep measures would be related to worse cognitive outcomes in both groups. We explored whether inflammatory biomarkers mediated the effects of sleep disturbances on cognitive functioning.

Methods: The sample included 41 PLWS (DSM-5 criteria) and 59 NCs (total sample: mean age 58.0 years, SD 8.9, range 35 to 74 years). People with dementia or other major neuropsychiatric disorders were excluded.

Cognitive measures included executive functioning (composite) and speed of processing from the Delis-Kaplan Executive Function System.

Participants wore Fitbit fitness trackers (Fitbit Enterprise, San Francisco, CA) for 5 consecutive days and nights, a wrist-worn tri-axial accelerometer that yields results comparable to actigraphy for sleep measures, including total sleep time (TST), percent sleep (percentage of time spent asleep vs. total time in bed), wake after sleep onset (WASO, total duration of overnight awakenings), bed-time, and wake-time. Standard deviations (SD) were used for intra-individual variability of sleep measures.

Inflammation was assessed using blood-based assays for high-sensitivity C-reactive protein (hs-CRP).

Independent sample t-tests and Chi-square tests were used to assess differences between PLWS and NC groups. General Linear Models were performed to assess the relationships of sleep measures with inflammatory biomarker levels and cognitive functioning, covarying for age, sex, race, and diagnostic group.

To investigate whether inflammation mediated the relationship between sleep and cognitive functioning, we conducted causal mediation analyses when we found significant relationships between sleep and cognitive functioning as well as between sleep and inflammatory biomarkers. Using two linear regression models, we estimated the average causal mediation effect (ACME), average direct effect (ADE), the total effect of sleep on cognitive functioning, and the proportion of the total effect mediated by inflammatory biomarker levels. Nonparametric bootstrap procedures with 1000 simulations were used to compute 95% confidence intervals for the ACME, ADE, total effect, and proportion mediated. Statistical significance was evaluated using p-values.

Results: The PLWS and NC groups were comparable on sex and BMI. PLWS were younger (d = -.58), had fewer years of education, and had more non-white participants, and smoked more than NCs. PLWS had similar mean sleep measures, but greater variability of TST (p < 0.001, d = 0.91), percent sleep (p = 0.05, d = 0.40), and waketime (p = 0.002, d = 0.64) than NCs. PLWS had higher hs-CRP levels (p = 0.01, d = 0.53), worse executive functioning (p < 0.001, d = -2.28), and worse speed of processing (p < 0.001, d = -2.1) than NCs.

In the whole sample, greater mean and variability of TST were associated with higher hs-CRP levels (B = -.006, SE 0.003, p = 0.02, partial-eta2 = 0.063; B = 0.002, SE 0.0009, p = 0.04, partial-eta2 = 0.045; respectively). Earlier mean bedtime and longer mean TST were associated with worse processing speed (B = 0.003, SE 0.0001, p = 0.003, partial-eta2 = 0.10; B = -.016, SE 0.008, p = 0.037, partial-eta2 = 0.05; respectively). There were trend relationships between less variability of percent sleep and longer mean TST with worse executive functioning (B = 0.021, SE 0.011, p = 0.06, partial-eta2 = 0.04; B = -.0013, SE 0.00065, p = 0.051, partial-eta2 = 0.043; respectively).

We found a non-significant, small mediation effect of hs-CRP levels on the relationship between TST and speed of processing (ACME: -0.004, 95% CI: [-0.01, 0.01], p = 0.150) and a trend for a direct effect of mean TST on processing speed (ADE: -0.020, 95% CI: [-0.04, 0.01], p = 0.07). The total effect was significant (Total Effect: -0.024, 95% CI: [-0.04, 0.01], p = 0.02), indicating that TST negatively impacts speed of processing. Approximately 17.8% of the total effect was mediated by hs-CRP levels (p = 0.17).

Conclusions: Greater intra-variability and worse mean objective sleep measures were associated with greater inflammation and cognitive outcomes. The preliminary analyses do not support inflammatory levels as a mediator of sleep’s effects on cognitive functioning. This study is continuing and updated results will be presented at the meeting.

Keywords: Psychosis spectrum symptoms, Wearable Sensing, Sleep disturbances, Inflammatory Markers

Disclosure: Nothing to disclose.

P823. Transcriptomic Analysis of the Human Habenula in Schizophrenia

Ege A. Yalcinbas, Bukola Ajanaku, Erik D. Nelson, Renee Garcia-Flores, Nicholas J. Eagles, Kelsey D. Montgomery, Joshua M. Stolz, Joshua Wu, Heena R. Divecha, Atharv Chandra, Rahul A. Bharadwaj, Svitlana Bach, Anandita Rajpurohit, Ran Tao, Geo Pertea, Joo-Heon Shin, Joel Kleinman, Thomas Hyde, Daniel R. Weinberger, Louise A. Huuki-Myers, Leonardo Collado-Torres, Kristen Maynard*

Lieber Institute for Brain Development, Johns Hopkins University School of Medicine, Baltimore, Maryland, United States

Background: Pathophysiology of many neuropsychiatric disorders, including schizophrenia (SCZD), is linked to habenula (Hb) function. While pharmacotherapies and deep brain stimulation targeting the Hb are emerging as promising therapeutic treatments, little is known about the cell type-specific transcriptomic organization of the human Hb or how it is altered in SCZD. Here we define the molecular neuroanatomy of the human Hb and identify transcriptomic changes in individuals with SCZD compared to neurotypical controls.

Methods: Using the 10x Genomics 3’ single cell gene expression platform, we performed single nucleus RNA-sequencing (snRNA-seq) from Hb-enriched postmortem human brain tissue from 7 adult neurotypical control donors. We identified molecularly distinct cell types and validated their spatial location using single molecule fluorescent in situ hybridization (smFISH) with RNAscope technology (n = 3 control donors). We then performed bulk RNA-sequencing and cell type deconvolution in Hb-enriched tissue from 35 individuals with SCZD and 33 neurotypical controls. We identified gene expression changes associated with SCZD in the Hb and compared our findings to previously published results in the dorsolateral prefrontal cortex (DLPFC), hippocampus, and caudate. Finally, leveraging DNA genotyping data, we conducted expression quantitative trait loci (eQTL) and SCZD risk colocalization analyses.

Results: Analysis of snRNA-seq data identified 17 Hb cell type clusters across 16,437 nuclei, including 3 medial and 7 lateral Hb populations, several of which were conserved between rodents and humans. smFISH for cell type marker genes validated snRNA-seq Hb cell types and depicted their unique spatial organization. Bulk RNA-seq analyses yielded 45 SCZD-associated differentially expressed genes (DEGs, FDR < 0.05), with 32 (71%) unique to Hb-enriched tissue. eQTL analysis identified 717 independent SNP-gene pairs (FDR < 0.05). Of these pairs, 16 pairs included a SNP that is a SCZD risk variant and 7 different pairs included a SCZD DEG. eQTL and SCZD risk colocalization analysis identified 16 colocalized genes. Of these colocalized genes, 9 have not been previously identified, suggesting that genome wide association (GWAS) signals associated with SCZD may include unique causative genes in Hb.

Conclusions: These results identify topographically organized cell types with distinct molecular signatures in the human Hb and demonstrate unique genetic changes associated with SCZD, thereby providing novel molecular insights into the role of Hb in neuropsychiatric disorders.

Keywords: Schizophrenia (SCZ), Single nucleus RNA sequencing, Transcriptomics, expression quantitative trait loci (eQTLs)

Disclosure: Nothing to disclose.

P824. Social Engagement Moderates the Relationship Between CRP and Negative Symptoms in Individuals at Clinical High Risk of Psychosis

David Goldsmith*, Emerald Yuan, Jean Addington, Carrie Bearden, Kristin Cadenhead, Tyrone Cannon, Barbara Cornblatt, Matcheri Keshavan, Daniel Mathalon, Diana Perkins, William Stone, Scott Woods, Elaine Walker, Benson Ku

Emory University, Atlanta, Georgia, United States

Background: Negative symptoms are common in individuals at clinical high risk of psychosis (CHR-P), are debilitating, and may predict conversion to psychosis. There are few treatment options for negative symptoms, and as such, discovering novel mechanisms in young CHR-P individuals is of great significance. One potential mechanism that may contribute to the development of negative symptoms is inflammation. Inflammatory markers have been shown to be elevated in CHR-P individuals and may be associated with negative symptoms. Deficits in social engagement are negative symptoms and play an important role in outcomes for CHR-P individuals. In fact, social engagement may buffer the effects of risk factors on psychosis. Social engagement in early developmental periods may decrease stress and interact with downstream processes, such as inflammation. Herein, we hypothesized that lifetime social engagement may moderate the association between C-Reactive Protein (CRP), a marker of inflammation, and negative symptoms in CHR-P young adults such that this association would be significant only among those with lower, but not greater, social engagement.

Methods: 48 individuals (30 CHR-P and 18 healthy controls; HC), which comprised the entire subgroup with data on CRP, negative symptoms, and social engagement from the North American Prodromal Longitudinal Study (NAPLS)-2 cohort, were included in this analysis. Negative symptoms were assessed using the Scale of Prodromal Symptoms (SOPS), and social engagement was calculated using the sum of five items from the Life Events Stress scale: (1) involvement in church or synagogue, club, neighborhood, or other organization; (2) took a vacation; (3) took up a new hobby, sport, craft, or recreational activity; (4) acquired a pet; and (5) made new friends. A generalized linear model with robust estimation was used to test the association of CRP, diagnosis, and social engagement (and their interactions) with negative symptoms, adjusting for age, sex, ethnicity, poverty, and depressive symptoms. Simple slopes for the association between negative symptoms and CRP moderated by social engagement were calculated and stratified by CHR-P groups.

Results: The mean age of the cohort was 22.6 and 21.9 for the CHR-P and HCs, respectively. Eight subjects in each group were female. CHR-P subjects had significantly greater negative symptoms than HC subjects (p < 0.001), though there was no significant difference in CRP or social engagement. In the generalized linear models, negative symptoms were significantly associated with CRP (beta=1.34, SE = 1.35, 95%CI -1.31 to 4.00, p = 0.035) as well as CHR-P (beta=8.16, SE = 1.71, 95%CI 4.80 to 11.52, p < 0.001). There was a significant association between negative symptoms and the interaction of CRP-by-social engagement (beta=0.37, SE = 0.56, 95%CI -0.74 to 1.47, p = 0.008), but not the interaction of CRP-by-CHR-P or CHR-P-by-social engagement (both p > 0.25). There was a significant association between negative symptoms and the three-way interaction CRP-by-CHR-P-by-social engagement (beta = -5.27, SE = 1.70, 95%CI -8.60 to -1.94, p = 0.002). Based on the simple slopes analysis, we observed a significant positive association between negative symptoms and CRP amongst the CHR-P group at low (-1SD; p = 0.02) and mean levels of social engagement (p = 0.04) but not in the individuals with high social engagement ( + 1SD; p = 0.34) or in any of the HC social engagement levels (p all > 0.2).

Conclusions: In this sample of CHR-P individuals, there was an association between negative symptoms and the interaction between diagnosis, inflammation, and social engagement, adjusting for relevant clinical and demographic covariates. Greater engagement in social activities appeared to buffer the relationship between inflammation, as measured by CRP, and negative symptoms. We have previously shown that inflammation is associated with negative symptoms related to motivation and pleasure, including asociality, in patients with schizophrenia, which is consistent with the preclinical and clinical literature of exogenous administration of inflammatory stimuli leading to avolition and anhedonia as well as studies of endogenous concentrations of inflammatory markers in patients with depression being associated with similar phenotypes. The data herein suggests that these associations in young individuals at risk for psychosis may be buffered by social engagement, perhaps by limiting stress and its downstream impacts on the brain and behavior. Though this study is limited by a small sample size, future studies should seek to replicate these findings in a larger sample, and future studies with interventions that target social engagement may be important for limiting negative symptom burden in CHR-P individuals with increased inflammation.

Keywords: Clinical high-risk for psychosis, Social buffering, inflammation, CRP, Negative Symptoms

Disclosure: Nothing to disclose.

P825. Enhanced Theta Oscillatory Responses to Standard Tones Drive Auditory Deviance Mismatch in 22q11.2 Deletion Syndrome

David Parker*, Gabrielle Ruban, Sidney Imes, Opal Ousley, Bruce Cuthbert, Brett Henshey, Nicholas Massa, Brian Roach, Daniel Mathalon, Elaine Walker, Joseph Cubells, Erica Duncan

Emory University, Athens, Georgia, United States

Background: 22q11.2 Deletion Syndrome (22q11DS) substantially increases risk for psychosis and other neurodevelopmental psychiatric conditions. Auditory deviance detection responses index NMDA and GABA receptor dependent predictive coding and stimulus adaptation, and they are reduced in psychosis syndromes. We examined evoked oscillatory responses in individuals with 22q11DS to determine the relationships among deviance mismatch (DMM) responses and their clinical and cognitive profiles.

Methods: Our cohort included subjects with 22q11DS (n = 38) and healthy comparison subjects (HC; n = 33). Event-related potentials to standard, pitch-deviant, duration-deviant, and double-deviant (pitch and duration) tones were measured by EEG sensors at Fz and Cz. A spatial PCA component reduced the two sensors to one virtual sensor (73% of variance explained). Signals were converted to the time-frequency domain (3-15 Hz, -50-400 ms) with a modified Morlet wavelet procedure. Time-frequency responses to the standard tone and each deviant tone were generated. DMM time-frequency responses were generated by subtracting the standard from the deviant response. Permutation tests (10,000) identified 14 regions of interest (threshold p ≤ .0071), which were submitted to Group x Sex ANOVAs (covariates: age and hearing acuity). The 14 significant variables from the auditory responses and deviance mismatch responses were reduced to three principal components to summarize the results efficiently and account for overlapping variance. Each PCA component was submitted to a Group x Sex ANOVA (covariates: age and hearing acuity). Exploratory analyses were performed among EEG responses, prodromal symptoms, autism spectrum traits, and cognitive profiles.

Results: Basic Auditory Response: Individuals with 22q11DS had a significantly higher response to standard tones. In the theta frequencies (4-8 Hz), 22q11DS responses were > HC responses from 125-225 ms (p-fdr=0.0067, ηp²=0.114) and from 275-325 ms (p-fdr=0.0036, ηp²=0.139). 22q11DS responses were > HC responses at 12 Hz from 250-300 ms (p-fdr=0.0067, ηp²=0.112) and from 7-14 Hz from 375-450 ms (p-fdr=0.0133, ηp²=0.092). The 22q11DS group did not differ in their responses to the pitch- or duration-deviant stimuli. The 22q11DS group had significantly lower responses to the double-deviant stimuli than the HC group at: 8 Hz from 250-275 ms (p-fdr=0.0079, ηp²=0.107), 9-10 Hz from 300-325 ms (p-fdr=0.014, ηp²=0.09), and at 7 Hz from 325-350 ms (p-fdr=0.0039, ηp²=0.135).

Deviance Mismatch Response: The 22q11DS group exhibited significantly lower deviance detection responses to the pitch-deviant signals than the HC group at two time-frequency clusters. 22q11DS responses were lower during the “typical” MMN time window at 4-8 Hz from 125-225 ms (p-fdr=0.0055, ηp²=0.121) and at 8-14 Hz from 375-450 ms (p-fdr=0.0014, ηp²=0.175).

Subjects with 22q11DS showed significantly lower deviance detection response to the duration-deviant signals than HC participants at two time-frequency clusters. 22q11DS responses were lower at the typical duration MMN time window at 4-7 Hz from 125-225 ms (p-fdr=0.0024, ηp²=0.156) and at 10 Hz from 275-325 ms (p-fdr=0.0014, ηp²=0.175).

Participants with 22q11DS had a significantly lower deviance detection response to the double-deviant stimuli than HC at three large time-frequency clusters. 22q11DS responses were lower at the typical double-deviant MMN time window at 4-8 Hz from 150-250 ms (p-fdr=0.002, ηp²=0.164), at 9-14 Hz from 100-200 ms (p-fdr=0.0051, ηp²=0.125), and at the typical P3a ERP component window at 5-14 Hz from 250-450 ms (p-fdr=0.0011, ηp²=0.195).

PCA Results: PCA component 1 (PCA-Theta-Alpha) was associated with the cortical response in the 4-15 Hz and 250-450 ms ranges, corresponding to the ERP P3a component time window. Higher PCA-Theta-Alpha values were associated with lower standard responses and higher DMM scores. The 22q11DS response was significantly lower than the HC response (p-fdr=0.0011, ηp²=0.192). PCA component 2 (PCA-DMM Theta) was associated with the theta cortical response (4-8 Hz) from 120-225 ms, corresponding to traditional MMN time windows. The 22q11DS response was significantly lower than the HC response (p-fdr=0.0011, ηp²=0.208). PCA component 3 (PCA-Double Deviant) was associated with the cortical response, primarily to the double-deviant tone in the 7-10 Hz and 250-350 ms ranges. Higher PCA-DD scores were associated with higher cortical responses to double-deviant tones (p-fdr=0.0028, ηp²=0.148).

Relationship to Clinical Symptoms and Cognitive Behaviors: There was a significant association between the PCA-Theta-Alpha (250-450 ms) component and prodromal positive symptoms (uncorrected p = 0.0098, Spearman’s rho = -.41) and disorganized symptoms (uncorrected p = 0.03, Spearman’s rho = -.35). No other significant relationships between EEG PCA components and Negative or General prodromal symptoms were observed in participants with 22q11DS. No significant relationships between ASD traits, Cognition, or Verbal IQ and EEG PCA components were identified after adjusting for group.

Conclusions: Individuals with 22q11DS show abnormal evoked oscillatory responses to basic auditory stimuli and during deviance detection tasks. This implicates NMDA and GABA dysregulation as mechanisms of cortical dysfunction related to psychosis in 22q11DS.

Keywords: Auditory Mismatch Negativity, 22q11.2 deletion syndrome, copy number variant, theta band oscillatory measures, Electroencephalography (EEG)

Disclosure: Nothing to disclose.

P826. Regional Activity and Excitatory/Inhibitory Biomarkers in the Anterior Limbic System: Effects of Aging vs. Disease Course in Schizophrenia

Elena Ivleva*, Jayme Palka, Monserrat Feria-Vargas, Cassidy Soto, Ehsan Shokri-Kojori, Jinsoo Uh, Sina Aslan, Uma Yezhuvath, Changho Choi, Vanessa Le, Kristen Kennedy, Michael Rugg, Carol Tamminga

University of Texas Southwestern Medical Center, Dallas, Texas, United States

Background: The anterior limbic system—hippocampus (Hipp) and its cortical projection network, including medial prefrontal cortex (mPFC)—is the key circuit implicated in neurobiology of schizophrenia (SZ). However, the dynamic evolvement of the Hipp-mPFC alterations, their specific trajectories, and relationships with clinical manifestations over the SZ course are unknown. We aim to disentangle the effects of aging vs. disease course on the Hipp-mPFC circuit alterations captured with multimodal high-resolution brain imaging [Vascular Space Occupancy (VASO); 1H-MRS] and examine their predictive value for cognitive dysfunction and symptoms over the adult lifespan in SZ.

Methods: The study sample constitutes a nearly complete baseline cohort (n = 139: 70 SZ, 69 healthy controls (CON), aged 18-75 years, both sexes) from an ongoing study (Aging and Disease Course: Contributions to Lifespan Neurobiology of Schizophrenia (COURSE), R01MH127317, Ivleva). 3 Tesla multimodal brain imaging data were acquired from bilateral Hipp and mPFC: VASO to estimate relative cerebral blood volume (rCBV) as proxy for regional activity in the Hipp subfields and mPFC (the results are available from n = 52: 23 SZ, 29 CON analyzed to-date); 1H-MRS with optimized inter-RF time delays to estimate regional concentrations of Glu and GABA in whole Hipp and mPFC (n = 109: 55 SZ, 54 CON). Hipp subfield masks [Dentate Gyrus/CA3 (DG/CA3), CA1, subiculum (SUB)] were developed via manual tracings and incorporated into the Automatic Segmentation of Hippocampal Subfields pipeline. mPFC gray matter mask was developed using a customized in-house protocol incorporating FreeSurfer. Statistical analyses: Hierarchical Generalized Additive Modeling (HGAM) with a 2nd order polynomial basis was used to examine (i) age-dependent (SZ vs. CON) and (ii) age- vs. disease course- (i.e., illness duration, ID) dependent (SZ only) effects in MRI and cognition outcomes. For all tests, α1/2 = 0.05.

Results: Hipp-mPFC regional activity (VASO): (i) Age-based HGAMs showed elevated rCBV in left CA1 (p = 0.004) and SUB (p = 0.013) in SZ vs. CON. Age2 (“accelerated aging”) effect (p = 0.026) and group x age2 interaction (p = 0.022) emerged in left CA1, driven by persistently elevated rCBV in SZ vs. CON. (ii) In SZ, age- and ID-based HGAMs revealed effects of ID in left CA1 and bilateral DG/CA3 (p = < .001-.016); and of ID2, in right CA1 (p = < .001), left DG/CA3 and bilateral SUB (p = < .001-.025). Age2 x ID2 interactions emerged in left DG/CA3 (p = 0.004) and SUB (p = 0.01), driven by more robust declines in rCBV with ID relative to those observed with age.

Hipp-mPFC excitatory/inhibitory biomarkers (1H-MRS): (i) Age-based HGAMs showed elevated Glu in bilateral Hipp (p = 0.025-.043) but reduced Glu in mPFC (p = 0.049) in SZ vs. CON. Age2 effect (p = 0.013) emerged in left Hipp Glu, and group x age2 interaction (p = 0.035), in mPFC. The shapes of the trajectories revealed distinct patterns in the Hipp and mPFC. In the Hipp, we observed elevated and rather stable over the lifespan Glu and GABA in SZ, and age-dependent declines in CON. In contrast, in the mPFC, Glu and GABA declined with age in both groups, with steeper declines in SZ. (ii) Age- and ID-based HGAMs in SZ revealed effects of age in mPFC Glu (p = 0.004); ID in left Hipp Glu (p = 0.048); and ID2 in left Hipp Glu, mPFC Glu and right Hipp GABA (p = 0.001-.039). Age x ID2 interactions emerged in Hipp Glu (p < .001) and GABA (p = < .001-.002) bilaterally; and age2 x ID2 interactions, in left Hipp Glu, mPFC Glu and left Hipp GABA (p = 0.005-.048), due to earlier and more robust declines (especially in Glu) with ID than with age.

Cognition (the Brief Assessment of Cognition in Schizophrenia, BACS): (i) Age-based HGAMs indicated lower BACS composite and all subscale scores in SZ vs. CON (all p < .001). Effect of age emerged in verbal fluency (p = 0.033); of age2, in BACS composite and several subtests (verbal fluency, digit sequencing and tower of London) (p = < .001-.033); and group x age2 interaction, in verbal memory (p = 0.011), indicating earlier and more robust declines in cognitive performance with increasing age in SZ than in CON. (ii) Age- and ID-based HGAMs in SZ showed effects of age in verbal memory and verbal fluency (p = 0.015-.02); ID in tower of London (p = 0.032); and ID2 in symbol coding and BACS composite (p = < .001-.035). Age2 x ID2 interactions emerged in symbol coding, tower of London and BACS composite (p = 0.011-.050), driven by earlier and more robust declines in cognitive function with ID relative to the declines observed with age.

Conclusions: Our findings demonstrate significant alterations in the anterior limbic system in SZ vs. CON, with the regionally-specific effects: elevated activity in the Hipp CA1 and SUB; elevated Hipp Glu but reduced mPFC Glu. The age- and illness duration-based HGAMs and trajectories indicate differential sensitivities among the VASO, 1H-MRS and cognitive biomarkers, with more robust disease course-dependent effects, relative to aging, in SZ. Future analyses will be focused on additional MRI biomarkers (i.e., structural MRI, task-based fMRI), predictive utility of the MRI-based biomarkers on cognition and symptoms, and longitudinal (2-year follow-up) data analyses). Ultimately, our study will provide novel, granular, biomarker-informed evidence on the specific contributions of aging and disease course into the lifespan neurobiology of SZ. Future work may inform SZ stage-specific interventions, e.g., reducing limbic hyperactivity in early-course SZ vs. enhancing this same circuit’s function in the advanced stages of SZ.

Keywords: schizophrenia, aging, disease course, biomarkers, neuroimaging (VASO, 1H-MRS), hippocampus, medial PFC

Disclosure: Nothing to disclose.

P827. Activation of Metabotropic Glutamate Receptor 3 (mGlu3) Corrects Schizophrenia-Like Cortico-Accumbal Abnormalities and Cognitive Deficits

Shalini Dogra*, Arisa Timoll, Niki Harris, P. Jeffrey Conn, Colleen M. Niswender

Vanderbilt University, Nashville, Tennessee, United States

Background: Accumulating clinical evidence suggests that imbalance of excitatory-inhibitory neurotransmission in the prefrontal cortex (PFC) may lead to cognitive deficits in schizophrenia and the agents that restore this balance may provide symptomatic relief in schizophrenia patients. In this context, metabotropic glutamate receptor 3 (mGlu3) has emerged as a compelling target for modulating excitatory signaling in the PFC. Polymorphisms in GRM3 are linked with impaired cortical activation and reduced performance in cognitive tasks in schizophrenia patients. These human genetic studies, along with the preclinical studies showing cognition-enhancing effects of mGlu3 activation in rodents expressing schizophrenia-like physiological and behavioral deficits, suggest that there is therapeutic potential in increasing mGlu3 activity to enhance cognition in schizophrenia.

Methods: Acute treatment with MK801 (0.18 mg/kg; s.c.) was used to induce schizophrenia-like physiological deficits in both male and female C57BL6j mice (5-6 mice/group). Cortical pyramidal neurons projecting to the nucleus accumbens (NAc) were traced by injecting retrograde travelling, adeno-associated particles encoding for green fluorescent protein (GFP) in the NAc. Acute brain slices were prepared 30 minutes after saline or MK801 treatment and medial PFC (mPFC) pyramidal neurons projecting to the NAc were identified as the cells expressing GFP. To test the effects of acute MK801 treatment on excitatory (E)-inhibitory (I) balance in mPFC neurons innervating the NAc, GFP-expressing neurons were voltage clamped at -65 mV and +10 mV to record spontaneous excitatory postsynaptic currents (sEPSCs) and spontaneous inhibitory postsynaptic currents (sIPSCs), respectively. The effects of mGlu3 activation on MK801-induced changes in E/I balance were evaluated by bath application of the mGlu2/3 receptor agonist, LY379268 (100 nM), in the presence of the mGlu2 negative allosteric modulator (NAM), VU6001966 (10 µM), to block the mGlu2-mediated effects. Pre-treatment of the mGlu3 NAM, VU0650786 (20 µM), was performed to confirm the specificity of mGlu3 effects. Slice optogenetics, paired with whole-cell patch-clamp electrophysiology, was used to evaluate the effect of MK801 treatment on glutamatergic neurotransmission at synapses made from mPFC inputs in the NAc (5-6 mice/group). Behavioral correlates of the cognitive impairments were accessed using a Y-maze test of spontaneous alternations in C57BL6/j mice (15-20 mice/ group). The effects of various treatments on the excitatory transmission and behavioral measures were evaluated using one-way ANOVA followed by the Newman-Keul’s post hoc test.

Results: We found that administration of acute MK801 enhanced the excitatory-inhibitory (E/I) ratio in mPFC pyramidal neurons projecting to the NAc. Bath application of the mGlu2/3 agonist, LY379268 (100 nM) reduced MK801-induced increases in E/I ratio in mPFC pyramidal neurons projecting to the NAc. Interestingly, the effects of LY379268 were blocked by pre-treatment with the mGlu3 NAM, VU0650786, suggesting that activation of mGlu3 mediates the efficacy of the mGlu2/3 agonist, LY379268 in restoring E/I balance in MK801-treated mice. We also observed that activation of mGlu3 corrected MK801-induced increases in glutamatergic neurotransmission onto D1-MSNs in the NAc, suggesting that mGlu3 can correct schizophrenia-associated glutamatergic abnormalities in the NAc. We next evaluated the ability of mGlu3 to reverse MK801-induced deficits in spontaneous alternations in a Y-maze test of cognition. We observed a significant deficit in spontaneous alternation in mice treated with MK-801 (0.18 mg/kg). Importantly, MK801-induced reductions in spontaneous alternations were not accompanied by a reduction in the total number of entries, demonstrating that reduced alternations observed in MK801-treated mice do not result from reduced locomotion. Interestingly, pretreatment with the mGlu2/3 agonist LY379268 (3 mg/kg, i.p., 30 minutes prior to MK801 treatment) corrected MK801-induced deficits in spontaneous alternations. Further, the effects of LY379268 were blocked by pretreatment with a selective mGlu3 NAM, VU6010572 (10 mg/kg, i.p., 20 minutes prior to LY379268 treatment), suggesting that activation of mGlu3 can rescue schizophrenia-like cognitive deficits in mice.

Conclusions: These data suggest that activation of mGlu3 can 1) correct schizophrenia-like glutamatergic abnormalities in cortico-accumbal circuitry, and 2) rescue behavioral correlates of cognitive impairments in mice modeling schizophrenia-like symptoms. Together, these studies provide important information about the circuit-specific mechanisms by which gene polymorphisms in GRM3 may affect cognition in schizophrenia and may provide future directions to develop highly selective mGlu3 agonists/modulators as novel therapeutics for schizophrenia.

Keywords: Metabotropic glutamate receptor 3 (mGluR3), Medial Prefrontal Cortex (mPFC), retrograde tracing, slice electrophysiology, Cognition

Disclosure: Nothing to disclose.

P828. Natural Language Processing-Based Part of Speech Tags Characteristic of Childhood Onset Psychosis

Anthony Deo*, Cynthia Lado, Andrew Carolan, Johanne Solis, Yuli Fradkin, Thanharat Silamongkol, Chloe Rosenkranz, Benjamin Herrera, Walter Barr, Emma Deaso, David Glahn, Michele Pato, David Zald, Carlos Pato

Rutgers University, Piscataway, New Jersey, United States

Background: Childhood onset schizophrenia is associated with both language delays and persistent subtle changes in language structure. Recent advances in natural language processing (NLP) have revealed consistent quantifiable changes in language structure predictive of diagnosis in clinical high risk for psychosis (CHR), schizophrenia and bipolar spectrum disorders in older adolescents and adults. NLP based measures have not been examined in children and young adolescents with a broad psychosis phenotype including affective and nonaffective psychoses. The goal of this preliminary study was to identify the relative frequencies of different parts of speech that are unique to and thus predictive of broadly defined child and early adolescent onset psychosis.

Methods: All protocols were approved by the Institutional Review Board at Rutgers University. Cases were defined as children and adolescents with a diagnosis of psychosis of any cause including affective (36%), schizophrenia spectrum (21%) and other psychosis (43%) (N = 14) and controls defined as having no personal history of psychosis (N = 9). The mean age at the time of evaluation for cases was 12.29 ( + /- 2.05) and for controls was 12.89 ( + /- 2.57) years. Gender assigned at birth was 50% female for cases and 44% female for controls. Clinical assessment for psychosis included the Kiddie Schedule for Affective Disorders and Schizophrenia (KSADS) and the Structured Interview for Psychosis-Risk Syndromes (SIPS). The Story Game was utilized to collect samples specifically for language analysis as it has been validated to elicit expression of thought disorder in children. Interviews were recorded via remote video conferencing and manually transcribed verbatim by a medical transcription service. Part of speech tags were assigned using the Natural Language Toolkit (NLTK) in Python. Analysis included all available transcribed language samples for a participant including KSADS, SIPS and Story Game. A binary logistic regression model was fitted with case status (presence or absence of psychosis) as the dependent variable, using an exploratory forward selection stepwise procedure to determine the predictors in the model. All 35 parts of speech categorized by NLTK were included in the model selection analysis. Additionally, age and gender assigned at birth were included in the model but were found to have no impact on model performance.

Results: The mean length of recordings was 149 ( + /- 78) minutes per participant. The extended length of the transcripts is important as analysis in adolescents in CHR noted limitations in NLP analysis of transcribed speech samples due to limited speech production in a relatively short interview using only the Story Game. A 3-predictor binary logistic regression model resulted in a 78.3% classification accuracy distinguishing cases of childhood onset psychosis from controls. The 3-predictor model included the NLTK tags MD (modal verbs), RBS (adverb, superlative best) and WRB (wh-adverb where, when).

Conclusions: We present the first NLP based analysis of extensive language samples derived from a broad psychosis phenotype in children and adolescents. A binary logistic regression model based on only 3 part of speech tag predictors was able to classify childhood onset psychosis cases from controls with reasonable accuracy. The predictors in the model including modal verbs which convey ambiguity in describing things that may be possible (i.e. can, could, may, might) and adverbs (including where, when and superlatives) have previously been noted to occur at altered frequencies in schizophrenia.

Keywords: natural language processing (NLP), Childhood-onset schizophrenia, diagnosis, psychopathology, dimensions

Disclosure: Nothing to disclose.

P829. A Mechanistic Investigation of Network Etiologies of Auditory Hallucinations in Schizophrenia

Andrew Murphy*, Michael Avissar, Yadi Chen, Pejman Sehatpour, Gaurav Patel, Daniel Javitt

Columbia University and New York State Psychiatric Institute, New York, New York, United States

Background: Many individuals with schizophrenia experience auditory verbal hallucinations (AVH) as a major symptom of their disease. Dysfunction in the posterior auditory and language regions are thought to underpin these symptoms. Non-invansive brain stimulation, including transcranial magnetic stimulation (TMS), has shown promise in treating these symptoms, but consistent results across individuals has been elusive. Individualized targeting of TMS within individual patients using functional MRI-guided computational modeling is one approach to improve symptom reduction. To investigate the efficacy of these computational models, our group has begun collecting pre-TMS individual fMRI data to use as model input. During an open-label study, inhibitory TMS was applied based on model results determining individual perisylvian language area (PSL) location in the temporal lobe, and post-TMS fMRI data was collected to evaluate functional brain changes. Further, pre- and post-TMS task imagining was acquired during which subjects were exposed to a 80Hz tone to evoke auditory cortex activity. Importantly, in addition to regional activation, specific interregional functional connections correlate with symptom severity in schizophrenia. Functional connections emanating from the left temporal cortex have recently been implicated as a marker for symptom severity. In particular, connections between the left temporal cortex and the anterior cingulate bilaterally, amygdala bilaterally, and left dorsolateral prefrontal cortex have all been associated with the magnitude of AVH in subjects with schizophrenia. To further investigate the role of these functional connections in AVH, our group plans to compute and compare individual pre- and post- functional connection strengths between frontal and temporal regions. Lastly, one known EEG correlate of AVH is an elevated response in the gamma range to 40hz stimuli. Here we investigate whether this correlate is affected by TMS.

Methods: Subjects: The treatment arm included 4 subjects diagnosed with schizophrenia with auditory verbal hallucinations who underwent TMS. The control arm included 6 subjects diagnosed with schizophrenia with auditory verbal hallucinations and 6 subjects diagnosed with schizophrenia without auditory verbal hallucinations.

Cortical stimulation point (CSP) selection: Structural MRI images were obtained and registered to The Human Connectome Project Parcellation. The subject-specific coordinates of the perisylvian language area parcel were determined. Neuronavigation was used to apply TMS to the subject-specific CSP.

TMS: Subjects underwent 1hz TMS sessions. Each session lasted 20 minutes, and were conducted over 10 consecutive days.

Task protocol: Each subject underwent 4 blocks of the task. One task block was composed of 12 13.5s stimulus blocks alternating with 12 13.5s rest blocks. Stimulus blocks were randomized to an equal number of either 80hz or 40hz auditory stimuli.

Imaging protocol: Each subject underwent resting state (rs) fMRI before and after the task. For subjects undergoing TMS, rsfMRI was completed before and after the TMS protocol. Each subject underwent task fMRI during the task protocol. EEG was simultaneously obtained during all imaging sessions.

Results: Results 1: We obtained the Auditory Hallucinations Rating Scale (AHRS) on all treatment subjects both pre- and post-TMS. We found that this score decreased following TMS, suggesting a reduction in AVH severity.

Results 2: In the treatment group, we observed lowered task-evoked activity in the auditory cortex following TMS. We also found that pre-TMS treatment group activation was more similar to control group hallucinators, and post-TMS treatment group activation was more similar to control non-hallucinators. Pre-TMS treatment group and control group hallucinators had higher auditory task activation than post-TMS treatment group and control group non-hallucinators.

Results 3: In the treatment group, we observed decreased frontal - temporal functional connection (FC) strength following TMS. We also found that the pre-TMS treatment group FC strength was more similar to control group hallucinators and post-TMS treatment group FC strength was more similar to control group non-hallucinators. Pre-TMS treatment group and control group hallucinators had higher frontal - temporal FC strength than post-TMS treatment group and control group non-hallucinators.

Results 4: We calculated EEG amplitude in the gamma range for all subjects during task performance. We found that following TMS, the amplitude of this AVH correlate to be decreased post-TMS relative to pre-TMS.

Conclusions: Our results demonstrate that TMS to the PSL (1) decreases the severity of AVH, normalizes both (2) task activation and (3) functional connectivity to become less similar to hallucinating and more similar to non-hallucinating control subjects, and (4) decreases gamma-range EEG responses. Taken together, we demonstrate that personalized TMS to PSL normalizes the brain state of hallucinators across several dimensions, with associated symptom reduction. We next hope to expand this analysis by leveraging existing datasets to better characterize baseline brain states of hallucinating and non-hallucinating individuals.

Keywords: Schizophrenia (SCZ), TMS, Functional Connectivity

Disclosure: Nothing to disclose.

P830. White Matter Microstructure and Associations to Schizophrenia and Childhood Trauma: A Nation-Wide Danish Twin Study

Cecilie Lemvigh*, Rachel Brouwer, Jayachandra M. Raghava, Tina Dam Kristensen, Birte Glenthoj, Bjørn H. Ebdrup

Center for Neuropsychiatric Schizophrenia Research (CNSR), Mental Health Center, Glostrup, Copenhagen University Hospital, København V, Denmark

Background: Schizophrenia is a heritable illness associated with disturbances in white matter microstructure. The integrity of the white matter microstructure itself is heritable, and previous evidence has indicated a shared genetic basis with schizophrenia. In addition, the experience of trauma during childhood is a well-established risk factor for schizophrenia that has also been linked to white matter alterations. Delineating the genetic relationships between schizophrenia liability, white matter microstructure and childhood trauma may increase our understanding of the underlying pathological processes.

Methods: Participants were recruited from 2011-2017 in Denmark by linking two nationwide registers, The Danish Twin Registry and The Danish Psychiatric Registry. Monozygotic (MZ) and dizygotic (DZ) proband pairs concordant or discordant for schizophrenia spectrum disorders and healthy controls twin pairs (HC) matched on sex and age were included. Register diagnoses were confirmed by SCAN interviews. Diffusion weighted magnetic resonance imaging scans were conducted on a Philips 3T Achieva whole-body scanner with a 32-channel SENSE head coil. A diffusion tensor model was applied to obtain measures of fractional anisotropy (FA). Experiences of childhood trauma were examined using the self-report Childhood Trauma Questionnaire assessing five domains: physical, emotional, and sexual abuse, and physical and emotional neglect.

In ongoing analyses, we are examining the genetic and environmental influences on white matter microstructure in this cohort enriched for schizophrenia using structural equation modelling. Further, we examine genetic and environmental associations between white matter, childhood trauma and schizophrenia. As an exploratory aim we will assess potential sex differences.

Results: In total, 216 twins participated (age 41.7 (SD 10.3), 49% females), covering 32 complete MZ proband pairs, 24 complete DZ proband pairs, 29 complete MZ HC pairs and 20 complete DZ HC pairs. Further, six individuals from proband pairs participated without their twin. The majority of patients were diagnosed with schizophrenia (N = 39, 61%) with a mean age of diagnosis of 27.4 ( ± 7.4). Proband pairs had lower FA in cingulate gyrus, t(175) = -1.98, p = 0.025, and inferior fronto-occipital fasciculus, t(175) = -2.82, p = 0.003 compared to HC twin pairs, as well as higher CTQ total, U = 2283.5, p < .001. In the whole group, CTQ total was not associated with any of the included FA measures, however when splitting the sample based on diagnosis, in the HCs we observed a significant negative correlation between CTQ total and superior longitudinal fasciculus, p = 0.044, and in the unaffected co-twins a positive correlation between CTQ total and FA in cingulate gyrus, p = 0.045 and uncinate fasciculus, p = 0.008, respectively.

Conclusions: The study aims to provide further evidence that altered white matter microstructure in schizophrenia is heritable and associated with disease liability, indicating overlapping genetic etiology. White matter disturbances may then constitute an endophenotypes and provide a basis to explore common genes. Further, we aim to provide a genetic basis to understand the process involved in the link between white matter alterations, childhood trauma and schizophrenia. Preliminary results suggest a different relationship between childhood trauma and white matter microstructure in patients compared to their unaffected co-twins and healthy controls. We aim to present results from the structural equation modelling analyses at the 2024 ACNP conference.

Keywords: Psychosis spectrum symptoms, DTI, clinical neuroimaging research, Childhood trauma, White matter microstructural alteration

Disclosure: Nothing to disclose.

P831. Merging 7-Tesla Magnetic Resonance Spectroscopy and Magnetoencephalography to Examine Excitatory-Inhibitory Balance in First-Episode Schizophrenia

Alfredo Sklar*, Annie Blazer, Warren Snover, Fran Lopez-Caballero, Mark Curtis, Brian Coffman, Dean Salisbury, Deepak Sarpal

University of Pittsburgh Western Psychiatric Institute and Clinic, Pittsburgh, Pennsylvania, United States

Background: Excitatory/inhibitory (E/I) imbalance is a proposed neural disruption common to various psychiatric disorders. Magnetic resonance spectroscopy (MRS) has provided evidence of altered gamma-aminobutyric acid (GABA) and glutamate (Glu) levels in schizophrenia that appear to fluctuate over disease course. However, there exists a paucity of data linking these abnormalities to impaired in vivo neurophysiology and brain function. The early auditory gamma-band response (EAGBR), a measure of local gamma-burst activity reflecting initial stages of cortical information processing, exhibits a pattern of longitudinal decline over the first year of illness. Interestingly, the EAGBR may be elevated at disease onset, with an association between larger EAGBRs and worse clinical outcomes suggestive of pathologic hyperexcitability. The current study examined associations between GABA/Glu balance and the EAGBR in a first-episode schizophrenia (FESz) sample to address this gap in the literature.

Methods: Twelve FESz were included. Participants received resting-state 7T-MRS and magnetoencephalography (MEG) recorded during an auditory oddball task. MRS spectra and source-localized MEG data were extracted from overlapping regions of left (LH) and right (RH) hemisphere superior temporal sulci. Associations between EAGBR values, extracted from wavelet-transformed stimulus-evoked responses, and GABA/Glu, GABA/Cre, and Glu/Cre ratios were assessed. The Brief Psychiatric Rating Scale (BPRS) and Global Functioning: Role/Social (GF: Role/Social) scales assessed disease symptoms and functional impairments, respectively.

Results: A lower GABA/Glu ratio was correlated with stronger LH (ρ = -.87, p < .001) and RH (ρ = -.85, p < .001) EAGBRs. These relationships were primarily driven by diminished GABA (LH: ρ = -.76, p = 0.007; RH: ρ = -.68, p = 0.02), as opposed to excessive Glu (LH: ρ = -.01, p = 0.99; RH: ρ = 0.53, p = 0.097) levels. Lower LH GABA/Glu ratios (ρ = -.64, p = 0.03) and larger RH EAGBRs (ρ = 0.70, p = 0.01) were also correlated with higher BPRS scores and lower LH GABA/Glu ratios were correlated with lower GF: Role (ρ = 0.62, p = 0.04) and Social scores (ρ = 0.64, p = 0.03).

Conclusions: The present investigation merged 7T-MRS and MEG recordings, leveraging the exquisite spatio-temporal resolution of these measures to isolate a functional measure of E/I imbalance in FESz. Robust associations between the GABA/Glu ratio, clinical ratings, and the EAGBR suggests this evoked gamma-band response reflects cortical hyperexcitability within the auditory system that is closely tied to disease debility at this early illness stage. In light of recent evidence that diminished EAGBRs over the first year of illness may be associated with symptom reduction, future investigations should apply these methodologies to examination of E/I balance across illness stage in schizophrenia. While preliminary, these data highlight the potential of merging high-precision, in vivo neuro-chemical and physiological techniques to validate established cellular models of disease.

Keywords: first episode schizophrenia, excitatory / inhibitory balance, 7T MRS, magnetoencephalography, Evoked gamma-band response

Disclosure: Nothing to disclose.

P832. Machine Learning Reveals a Sparse Set of Cognitive Domains Impaired in Schizophrenia

Robert Chen*, Tiffany Greenwood, Steven Millard, David Braff, Laura Lazzeroni, Neal Swerdlow, Monica Calkins, Robert Freedman, Michael Green, Ruben Gur, Raquel Gur, Caroline Nievergelt, Keith Nuechterlein, Allen D. Radant, Larry Siever, Jeremy Silverman, William Stone, Catherine Sugar, Ming Tsuang, Bruce Turetsky, Gregory Light, Debby Tsuang

University of Washington, Seattle, Washington, United States

Background: Cognitive impairment is a hallmark of schizophrenia. Development of cognitive phenotyping has relied on large test batteries that are infeasible to deploy in clinical settings. Machine learning (ML) has the potential to identify a subset of cognitive domains that accurately distinguishes between patients with schizophrenia (SCZ) and healthy control subjects (HCS), paving the way for simplified cognitive tests that could be easily adopted in clinical practice. The goal of this study was to determine which subset of a diverse battery of cognitive tests can separate SCZ from HCS most efficiently.

Methods: In this case-control study, a total of 1415 stable outpatient individuals with schizophrenia or schizoaffective disorder and 1062 healthy counterparts were recruited by five laboratories distributed geographically across the US as part of the Consortium on the Genetics of Schizophrenia II (COGS-II). Individuals were excluded if they had neurological or additional Axis I psychiatric disorders or substance abuse. Covariates included age, gender, smoking status, site, and the maximum of parents’ education.

Fifteen cognitive and neurophysiologic biomarkers including the Degraded Stimulus and Identical Pairs Continuous Performance Tests, Letter Number Span, Penn Computerized Neurocognitive Battery (CNB), the California Verbal Learning Test II (CVLT-II), and an anti-saccade task were measured. Eight state-of-the-art ML models were trained on 70% of the dataset and tested on the remaining 30% to develop classifiers that separated SCZ from HCS. The primary model performance metric was the area under the receiver operating characteristic (ROC) curve, which balances sensitivity and specificity for distinguishing between SCZ and HCS. Recursive feature elimination (RFE), a ML method that identifies groups of features important for classification performance, was used to identify the most discriminatory features.

Results: Participants with complete data included 599 SCZ (45 ± 12 [mean ± SD] years, 441 male) and 745 HCS (38 ± 13 [mean ± SD] years, 356 male). Of the eight ML models tested, a boosted generalized linear model (BoostedGLM) classifier performed best, achieving an ROC of 0.899 in the held-out dataset. Elastic net regression (ROC = 0.898), partial least squares regression (ROC = 0.896), and a feed-forward neural network (ROC = 0.894) performed similarly to the BoostedGLM (p > 0.05), while the remaining four tree-based and random forest algorithms performed significantly worse (p < 0.05). RFE revealed that the total number of correct items recalled on five trials of the CVLT-II was the most discriminatory feature, followed by an emotional identification task on the Penn CNB. A minimal ML model based on BoostedGLM using just these two neurocognitive tests achieved an ROC of 0.888, similar to the complete BoostedGLM model trained on all 15 biomarkers (p = 0.17).

Conclusions: A subset of cognitive measures—revealed by a variety of machine learning models—related to verbal learning and emotional identification distinguishes SCZ from HCS. These findings support a ‘less-is-more’ approach to efficient cognitive profiling in schizophrenia.

Keywords: Cognitive impairment associated with schizophrenia, machine learning, Cognitive biomarkers

Disclosure: Nothing to disclose.

P833. Rates of Suicidality, Parasuicidal Behavior and Associated Clinical Features in Individuals at Clinical High-Risk for Psychosis in the NAPLS-3 Cohort

Tara Trujillo*, Heline Mirzakhanian, Jean Addington, Carrie Bearden, Tyrone Cannon, Barbara Cornblatt, Matcheri Keshavan, Daniel Mathalon, Diana Perkins, William Stone, Elaine Walker, Scott Woods, Kristin Cadenhead

University of California - San Diego, San Diego, California, United States

Background: Individuals at clinical high risk (CHR) for psychosis have higher rates of depression, anxiety, trauma, and other comorbid disorders that are associated with suicidality. Prior CHR and psychosis studies have reported higher rates of suicidal ideation (SI) and behavior compared to healthy controls but there is limited research examining predictors and protective factors for suicidality in this vulnerable population. The aims of the current project are to 1) investigate rates of suicidality in CHR participants in the third sample of the North American Prodrome Longitudinal Study (NAPLS-3) and 2) determine which demographic, life event, symptom, functional and treatment information is associated with past or present SI or behavior.

Methods: The sample included 710 CHR and 96 healthy control (HC) participants. Past SI, plans, self-harm and attempts were assessed via the clinician-administered Structured Assessment of Violence Risk in Youth (SAVRY) scale and the Calgary Depression Scale for Schizophrenia (CDSS). Clinical data was obtained via the Scale of Prodromal Symptoms (SOPS), the Structured Clinical Interview for DSM-5 Disorders (SCID-5), the Alcohol/Drug Use Scale (AUS/DUS), Global Assessment of Functioning (GAF), the Global Functioning Social and Role Scales (GF:R, GF:S), the Social and Occupational Functioning Scale (SOFAS), and the Premorbid Adjustment Scale (PAS). The Daily Stress Inventory (DSI) and the Life Events questionnaire scales were also used.

Results: Of the 710 CHR participants, 30.7% endorsed a history of SI or plan while 25.6% endorsed a history of self-harm with no intent. The rate of reported serious self-harm or suicide attempts was 12% with 1% reporting suicide attempts designed to end in death. Of the 96 HC participants, 4.2% endorsed a history of self-harm or suicidal gestures with no clear suicidal intent with 0% endorsing a history of suicidal thoughts, plan, or attempt.

A history of SI was associated with higher symptom ratings across all domains on the baseline SOPS (p’s < .03 – .001) as well as lower premorbid, global, social and role functioning (p’s < .01 – .001). Those with a history of SI or parasuicidal behavior were more likely to have a broad range of comorbid diagnoses including depression, substance use disorders, anxiety disorders, attention deficit disorder, and borderline personality disorder (p’s < .03 –  .001). A history of trauma and other stressful life events including bullying, emotional neglect, and abuse was also prevalent in the group with a history of suicidal thoughts or behavior (p’s < .02 – .001). A history of suicidality was also negatively associated with a history of psychotherapy as well as antidepressant and antipsychotic treatment at baseline (p’s < .04 – .001). Personal risk factors included poor coping, impulsivity, low empathy and fewer protective factors were also associated with past SI or behavior (p’s < .05 – .001). Interestingly, past suicidality was not associated with future conversion to psychosis, symptoms or functioning at 24-month follow-up.

Conclusions: Close to one-third of CHR participants in the NAPLS3 sample report a history of suicidal thoughts, plans or self-harm while an additional 12% report a history of serious self-harm or suicide attempts, consistent with prior CHR reports. Although the rate of deaths by suicide in the CHR population has not been reported, our findings are important because suicide is one of the leading causes of death in people with psychotic disorders. This data supports the importance of suicide risk assessments and the development of interventions specifically targeted to CHR youth. Importantly, we found that participants who had been in treatment were less likely to have SI, which demonstrates a need to actively treat those with CHR symptoms. A better understanding of associated risk and protective factors in the CHR population will help in the further development of targeted treatments in this high-risk population. Future directions will include examining long-term clinical outcomes (5-20 years) in CHR participants to determine how parasuicidal behavior is associated with future outcomes including early death and completed suicide in this population.

Keywords: Clinical high-risk for psychosis, Suicide risk factors, suicide prevention

Disclosure: Nothing to disclose.

P834. Peripheral Markers of Blood Brain Barrier Disruption and Immune Cell Transmigration are Related to Asociality and Mesial Temporal Lobe Cortical Thickness Reductions in Neuroinflammatory Schizophrenia

Thomas Weickert*, Roxanne Daniels, John Burns, Yunting Zhu, Cynthia Shannon Weickert

State University of New York Upstate Medical University, Syracuse, New York, United States

Background: Inflammation is linked to schizophrenia pathophysiology and increased CD163+ macrophages are found in multiple brain regions in schizophrenia. Immune cell transmigration into human brain involves blood brain barrier (BBB) alterations resulting in increased Vascular Endothelial Growth Factor (VEGF) and shedding of CD163. Thus, we hypothesized that soluble CD613 (sCD163) and VEGF would be increased in the serum of people with schizophrenia especially in those with other increased inflammation markers. Further, we expected serum levels of VEGF and/or sCD163 would inversely correlate with cortical thickness.

Methods: Serum samples were obtained from the Australian Schizophrenia Research Bank (N = 1,143, males=581, females=562) and VEGF, sCD163 and CRP proteins were measured by ELISA in 499 people with schizophrenia/schizoaffective disorder and 644 healthy controls. All individuals were stratified by normal or elevated CRP levels ( > 3 µg/ml). Negative symptoms were measured via SANS. Structural 1.5T MRIs (n = 280) were used to determine cortical thickness via FreeSurfer. Diagnostic X inflammatory subgroup differences were determined by ANOVA.

Results: Patients with schizophrenia had elevated sCD163 t(1085) = 4.32, p < .0001, and VEGF levels t(1082) = 4.51, p < .0001, relative to healthy controls. Regarding inflammatory subgroups, we detected significant differences in sCD163 and VEGF (both F > 13.00, p < 0.00001). The elevated inflammation schizophrenia subgroup displayed higher VEGF and sCD163 levels than healthy control subgroups (all p’s < 0.0001). sCD163, but not VEGF, correlated positively with age and duration of illness (p < 0.01 and p < 0.02 respectively). We found inverse correlations between VEGF levels and cortical thickness of the parahippocampal gyrus (r = 0.40, p < 0.01) in patients with schizophrenia who were classified as having elevated inflammation and sCD163 was related to asociality scores.

Conclusions: Serum proteins indicative of blood brain barrier alterations and macrophage migration are increased in patients with schizophrenia classified as having elevated peripheral inflammation. Additionally, these alterations do not appear to resolve, but may worsen with time and relate to brain tissue damage of the mesial temporal lobe. Thus, treatments blocking immune cell transmigration may benefit patients with schizophrenia.

Keywords: Schizophrenia (SCZ), inflammation, Blood-Brain-Barrier, Macrophage, cortical thickness

Disclosure: Nothing to disclose.

P835. A Coordinate-Based Meta-Analysis of Neural Contributions to Antipsychotic Treatment Response

Henry Chase, Sidhant Chopra, Michael Pupi, Tate Overbey, Simon Eickhoff, Thomas Nickl-Jockschat, Jose Rubio, Nicolas Crossley, Shaun Eack, Deepak Sarpal*

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Numerous studies have used neuroimaging to examine the relationship between schizophrenia spectrum disorders (SSD) and the effects of antipsychotic drugs. Across this literature, studies have employed a variety of imaging modalities and analytic methods to address antipsychotic response, with disparate results. Here, we employ a multimodal coordinate-based, activation likelihood estimation (ALE) methods to meta-analyze published neuroimaging studies that focus on markers of antipsychotic response, bringing these disparate studies into a common framework. Given that phenotypic similarities may be attributable to the dysfunction of common functional circuits, we also pursued a novel network-based approach, considering enrichment within three primary large-scale association networks implicated in the pathophysiology of SSD.

Methods: Literature search for neuroimaging studies employing various modalities (fMRI, PET, structural MRI, SPECT, ASL) to evaluate antipsychotic treatment response in SSD was conducted using PubMed in December 2023. Whole-brain coordinates from studies meeting criteria were converted into MNI space, and analyzed using the ALE technique. The significance of converging clusters of loci was assessed using family-wise cluster-corrected p-value of p < 0.05 (cluster-forming threshold: p < 0.001). In addition, correlation of ALE scores with ICA/BrainMap database-derived large-scale networks (default mode network: DMN; central executive network: CEN; and salience network: SAN) was conducted using Spearman’s rho with significance assessed using spatial autocorrelation preserving null models at an alpha value of 0.016 given three networks.

Results: A total of 26 studies met our inclusion criteria: average patient sample size of 49 (range: 10-127); 54% focused on first-episode schizophrenia; average age of 27.9 (7.3) years; 73% studies limited to treatment with 2nd generation antipsychotic drugs (two focused on clozapine); an average treatment duration of 12.3 weeks (range: 3-52 weeks); and imaging modalities including resting-state functional connectivity (n = 17), structural imaging (n = 6), task-based fMRI (N = 2), and SPECT (n = 1). No significant convergence of loci was observed using cluster-level statistics. However, loci were found to be significantly enriched in the SAN (rho=0.20, p = 0.008), but not CEN or DMN (p’s > .38).

Conclusions: To our knowledge, this study is the first to perform a coordinate-based, meta-analytic investigation of the literature on the neural correlates of antipsychotic treatment response in SSD. Although no localized region was identified in a whole-brain analysis, network-based findings suggest that loci are significantly enriched across the SAN. These findings further support the SAN as a key network underlying antipsychotic treatment response and in the pathophysiology of SSD.

Keywords: Antipsychotic response, meta-analysis, Human Neuroimaging

Disclosure: Nothing to disclose.

P836. Social Withdrawal in Schizophrenia is Predominantly Associated With Resting-State Intrinsic Hypoconnectivity in the Social Aversion Network

Theo van Erp*, Darius McLeod, Kamalakannan Vijayakumar, Aysenil Belger, Juan Bustillo, Kelvin Lim, Judith Ford, Daniel Mathalon, Adrian Preda, Bryon Mueller, Steven Potkin, Jaylen Lee, Elizabeth Martin, Jason Schiffman, Magenta Simmons, Ivy Tso, Bramsh Chandio, Jessica Turner, Lianne Schmaal, Paul Thompson, Cameron Carter, Diego Pizzagalli, Kevin Bickart, Vince Calhoun

University of California, Irvine, Irvine, California, United States

Background: Social withdrawal - the active avoidance of social interaction - is a complex, transdiagnostic symptom that is often already observed during the early, developmental stages of neuropsychiatric disorders, yet little is known about its underlying neural circuitry. Social withdrawal is an expression of social dysfunction, which can be defined as a lack of the capacity to integrate behavioral, cognitive, and emotional skills to adapt to diverse social contexts and demands. It predicts disorder onset, is associated with a decreased likelihood of short-term remission, and, along with other negative symptoms, remains an important, unmet therapeutic need. It has a significant impact on long-term mental and physical health outcomes, and overall well-being of individuals with neuropsychiatric disorders. In contrast, social connectedness protects against psychotic and mood disorders. This pilot study examines the neural circuitry associated with social withdrawal in schizophrenia, using resting-state functional magnetic resonance imaging, as part of a larger planned meta-analysis. More specifically, it tests the role of social perception, social affiliation, and social aversion networks, posited by an amygdala-centric model of social connectedness, using seed-based, resting-state connectivity to ventrolateral, medial, and dorsal amygdala subregions, respectively. Social withdrawal can be passive, i.e., due to reduced approach behaviors - likely to involve the social affiliation network - or active, i.e., due to increased avoidance behaviors - likely to involve the social aversion network.

Methods: Resting-state functional magnetic resonance imaging data from 180 individuals with schizophrenia [45 women, mean age (SD) = 38.6 (11.5)], recruited by the Function Biomedical Information Network Phase 3 study, were analyzed with seed-based connectivity analyses using HALFpipe [smoothing 6mm FWHM, confound removal with aCompCor (top 5 components), temporal filtering 125s Gaussian, non-linear registration to MNI152NLin2009cAsym using ANTs13 (Advanced Normalization Tools)]. The Positive and Negative Syndrome Scale (PANSS) item G16 was used to define high (G16 > 3; n = 71) and low (G16 < =2; n = 109) active social withdrawal groups. Whole brain Fisher-Z transformed, seed-based correlations were generated for 6 (3 left- and 3 right hemisphere) amygdala seeds based on the amygdala-centric model of social connectedness. Group comparisons were performed using AFNI’s (Analysis of Functional NeuroImages) 3dLME, predicting each of the 6 Fisher-Z images with active social withdrawal (high, low), while statistically controlling for age, sex, and site (scanner); threshold nominal p < 0.05, two-tailed.

Results: Individuals with schizophrenia with high compared to low active social withdrawal showed predominantly lower intrinsic network connectivity in the social aversion network (nominal p < 0.05) relative to the social perception and social affiliation networks (nominal p < 0.05).

Conclusions: This pilot study - conducted as a proof of concept for a larger, planned meta-analytic study on neural circuitry associated with social withdrawal - found that high active social withdrawal in schizophrenia is predominantly associated with intrinsic hypoconnectivity in the social aversion network. These findings suggest that active social withdrawal in schizophrenia may be predominantly associated with abnormalities in brain networks that are associated with avoidance rather than approach behaviors.

Keywords: social withdrawal, Resting State Functional Connectivity, Schizophrenia (SCZ), Negative Symptoms, Resting-state fMRI

Disclosure: Nothing to disclose.

P837. Exploratory Anchor-Based Analyses to Determine Minimally Clinically Important Difference (MCID) for the Brief Assessment of Cognition (BAC) and the Schizophrenia Cognition Rating Scale (SCoRS)

Hans Klein*, Reuben Fan, Tingting Ge, Mahnaz Asgharnejad, Chris Brady, Elizabeth Hanson, Jan Sedway, Venketesha Murthy, Richard Keefe, Jaskaran Singh, Ni Khin

WCG, Princeton, New Jersey, United States

Background: Identifying the clinical impact of any novel treatment is critical to understand the potential benefits it may have for any given patient. Thus, stakeholders (including regulators, sponsors, and clinicians) have expressed interest in identifying a Minimally Clinically Important Difference [MCID; also known as meaningful score difference (MSD)] for clinical outcome assessments. MCID have not been established for assessments of cognitive performance or cognitive functioning in individuals with Cognitive Impairment Associated with Schizophrenia (CIAS). The objective of this study is to provide preliminary estimates for MCID on two cognitive endpoints, Brief Assessment of Cognition (BAC) and the Schizophrenia Cognition Rating Scale (SCoRS) using anchor-based analyses on an enriched sample from the INTERACT study.

Methods: The INTERACT study was a phase 2, 12-week, randomized, double-blind placebo-controlled efficacy, safety, and PK study of three doses (50, 125, and 500mg) of luvadaxistat (NBI-1065844; formerly TAK-831), an investigational product previously developed for the treatment of negative symptoms in schizophrenia. As the trial recruited for negative symptoms of schizophrenia and not cognitive impairment, an enriched sample was utilized for analyses (i.e., participants with a BAC composite T score between 0 and 40 at baseline). Alongside the 20 items that make up the SCoRS interview, the SCoRS Interviewer Change Rating captures a clinical rating of change from baseline in which clinicians rate the participant’s change in cognitive ability on a 7-point Likert scale with 1 being much worse, 4 being no change, and 7 being much improved. Using anchor-based methods on an enriched sample, we estimated a MCID for the BAC and SCoRS Interview Total Rating. Specifically, we used the SCoRS Interviewer Change Rating at End of Treatment that is captured with the SCoRS assessment, but not included in the Total Rating, to divide the enriched sample into those with an observed clinical improvement (scoring 5 or above) versus those with no clinical change (scoring 4). Data from all treatment arms, including placebo, were considered in the analyses to determine the blinded minimum mean score change that is associated with observed clinical improvement. Groups were compared to determine the average change in BAC composite T score and change in SCoRS Interview Total Rating. Using these average changes as a starting point, we then set MCID thresholds. Finally, we report frequencies for participants who met the MCID thresholds within each of the trials’ treatment arms to describe sensitivity to treatment.

Results: The enriched sample (n = 142, treatment arm distributions: Placebo = 46, 50mg = 36, 125mg = 34, 500mg = 31), 77 participants (Placebo = 25, 50mg = 19, 125mg = 18, 500mg = 15) were noted to have observable improvement from baseline based upon the SCoRS Interviewer Change Rating while 60 participants had no observable clinical change. The average change from baseline on BAC Composite T-Score was larger for those with observable clinical improvement (Mean change = 5.47, SD = 6.91) compared to those with no observable clinical improvement (Mean change = 3.50, SD = 6.06). Similarly, the average change from baseline on the SCoRS Total Rating was larger for those with observable clinical improvement (Mean change = -3.70, SD = 5.06) compared to those with no improvement (Mean change = -0.88, SD = 4.27). Although this indicates a potential 2-point MCID for both BAC and SCoRS based upon group differences, frequency tables indicate that a higher threshold for BAC (3-point change from baseline) may be more sensitive to treatment. A higher proportion of participants attained a 3-point change threshold in the 50mg (82.9%) and 125 mg (76.5%) treatment groups relative to placebo (55.6%; p = 0.043). For SCoRS Total Rating, frequency tables show that a higher proportion of patients achieved a 2-point improvement in the 50mg treatment condition (65.7%) relative to the placebo condition (50.0%), however this did not reach significance (p = 0.440).

Conclusions: These results provide a good initial estimate for MCID in assessments of cognitive test performance (BAC) and interview-based assessment (SCoRS). A 3-point improvement on the BAC assessment seems to align with observed clinical improvement as determined by the SCoRS Interviewer Change Rating and is sensitive to treatment as a higher proportion of individuals in the 50mg and 125mg treatment arms achieved this benchmark. For the SCoRS Total Rating MCID, a 2-to-3-point difference may be clinically meaningful as the group with observable improvement based upon the SCoRS Interviewer Change Rating corresponded with a SCoRS Total Rating mean change 2.8-points higher than those without clinical improvement. Further analyses are needed to confirm the stability of these estimates in alternate samples.

Keywords: Cognition, MCID, Cognitive impairment associated with schizophrenia

Disclosure: WCG: Employee, (Self).

P838. Structure-Function Correlates of the Cerebello-Thalamo-Frontal Network in Schizophrenia

Neelabja Roy*, Urvakhsh Mehta, Rakshathi Basavaraju

National Institute of Mental Health and Neuro Sciences (NIMHANS), India, Bengaluru, India

Background: Resting state functional connectivity (rsFC) within the cerebellar-thalamo-cortical (CTC) circuitry is found impaired in schizophrenia (Andrease, Paradiso, and O’Leary., 1998) with posterior cerebellar hypoconnectivity with thalamus, prefrontal cortical areas now deemed as its consistent marker across treatment stages (Mehta, Ithal, Roy et al 2024). However, it is not clear if morphometric aberrations within this critical circuitry are associated with the functional hypoconnectivity. We first compared brain volumes of the CTC-RoI across drug-naive and remitted schizophrenia, and normative groups. We then examined the association between brain volumes and rsFC within the CTC-RoI. Lastly, we examined if brain volumes were associated with cognitive performance.

Methods: We recruited three age and sex matched subject groups: i) 54 drug naïve first-episode schizophrenia (FESdn), ii) 22 FES remitted after 2 years of treatment with risperidone (FESrm), iii) 43 normative controls (NC). We obtained their MRI scans on a 3-Tesla MRI scanner using a 32-channel head coil to acquire (a) high-resolution structural MRI (T1 MPRAGE sequence; 5m) of 1-mm section with no inter-slice gap and (b) rs-fMRI echoplanar imaging of 303 volumes from 10m 6s sequences, processed to yield 25 independent component networks (ICNs). We first preprocessed and standardised the rsFC scans to a common template and used a beta-extraction approach to get the strength of functional connectivity in our CTC-RoI for each of our subjects. For the structural scans, we conducted cortical surface reconstruction on the scans with freesurfer recon-all processes. After validating satisfactory segmentation into cortical, white matter and subcortical structures, we used automatic segmentation (aseg) and automatic parcellation (aparc) to derive volumetric measures of cerebellar cortices, thalamus, posterior and anterior cingulate, DLPFC for subjects. Then we performed independent t-tests and correlational analysis for assessing volumetric differences between NC vs. FESdn and NC vs. FESrm, along with their correlational analysis with factors of the derived functional strength measures and the assessed clinical symptom scores for measuring the associations of our interest.

Results: Comparing FESdn with NC, we found significantly reduced volumes in the bilateral cerebellar cortices (right: p = 0.002, Cohen’s d = 0.667; left: p = 0.017, d = 0.515), no difference in either thalamus and a significantly reduced right posterior cingulate (p = 0.003, Cohen’s d = 0.657). We also observed a similar significantly lesser volume in the same structures in the NC-FESrm comparison, indicating their importance. However in FESrm, Right Anterior Cingulate (p = 0.02, Cohen’s d = 0.638) and Bilateral DLPFC (right: p = 0.019, Cohen’s d = 0.642; left: p = 0.029, Cohen’s d = 0.598) were additionally significantly reduced in volume than NC. For looking at the associations, we aggregated through the CTC-RoI subregions and across all subjects, and we did find the volumetric morphometry to be significantly correlated with the rsFC strength of the critical cerebellar-thalamic-cortical network (r = 0.254, p = 0.005, N = 119). Lastly, even though the rsFC was not associated with the clinical scores of PANSS, it was found significantly correlated to important symptoms, scores of neurocognition (r = 0.371, p < .001, N = 90) and social cognition (r = 0.327, p = 0.002, N = 91).

Conclusions: That we found structural morphometric measures to be indeed associated with the rsFC activity and certain symptom scores, and that volumetric deficiencies in key subregions were consistent across two schizophrenia stages when compared to normative brains, may be indicative of a key interplay of cerebellar morphometry in the disorder. The potential role of medications towards remission (as evidenced by additional morphometric differences in the remitted group), and the progressive nature of aberrations and variations, should be studied more with an approach of integrating functional and associated morphometric markers for more holistic analytical markers.

Keywords: morphometry, Structural and Functional Connectivity, Schizophrenia (SCZ), cerebellum

Disclosure: Nothing to disclose.

P839. Sex-Specific Alterations in Gene Expression Across the Striatum in Psychosis

Kyle Ketchesin*, Megan Perez, RuoFei Yin, Madeline Scott, Wei Zong, Marianne Seney, Xiangning Xue, Mariah Hildebrand, Vaishnavi Shankar, Jill Glausier, David Lewis, George Tseng, Colleen McClung

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Psychosis is a defining feature of schizophrenia and highly prevalent in bipolar disorder. Despite known sex differences in the symptoms and progression of these disorders, little is known about the underlying molecular mechanisms driving these differences. Prior work from our group identified diurnal alterations in gene expression across the human striatum in subjects with psychosis. In this study, we investigated sex differences in the transcriptome within striatal subregions in unaffected and psychosis subjects.

Methods: RNA-sequencing was performed on nucleus accumbens (NAc), caudate, and putamen samples from subjects with psychosis (n = 36) or unaffected subjects (n = 59). For analysis of sex differences, we created a sex-matched cohort (psychosis: n = 10/sex; unaffected: n = 11/sex) and evaluated sex and psychosis effects within each brain region using an ANOVA of expression data. Transcripts were considered differentially expressed if p < 0.01 and a fold change of 1.2 (20% expression change). Metascape was used for pathway enrichment.

Results: Significant sex differences in gene expression were observed across striatal regions. Notably, angiogenesis and immune-related pathways were significantly upregulated in female subjects with psychosis across the striatum, while mitochondrial-related pathways were downregulated in the NAc. Rank-rank hypergeometric overlap analyses revealed a discordant expression pattern with opposite directions of gene expression between male and female subjects with psychosis. Meta-analyses showed that this expression pattern is consistent across regions.

Conclusions: We found significant sex differences in gene expression across the striatum that may underlie sex-specific striatal dysfunction and symptomatology in psychosis. Future studies will focus on the potential mechanisms and functional consequences of these gene expression alterations.

Keywords: Schizophrenia, Psychosis, Postmortem, Striatum, RNA-seq

Disclosure: Nothing to disclose.

P840. Differentially Interacting Effects Between Lifetime and Acute Stress on Striatal Prediction Error Signals and Their Relation to Symptoms in Patients With Schizophrenia

Teresa Katthagen, Jacob Nudelman, Olivia Hutchinson, Florian Schlagenhauf, James Waltz*

University of Maryland School of Medicine, Baltimore, Maryland, United States

Background: Both the emergence of psychotic illness (at first break) and the exacerbation of symptoms in chronic psychotic illness are commonly associated with stressful life events. Adverse childhood events (ACEs) have been linked to an increased risk of conversion to psychotic illness in those at clinical high-risk and has been shown to compound the effects of acute stressors on psychotic symptom severity. There is also evidence that acute stress can exacerbate the negative symptoms of psychotic illness, such as anhedonia and avolition. Our goal was to investigate how neural circuits for stress reactivity, reward processing, and salience signaling interact in mediating the effects of cumulative and acute stress on both the positive and negative symptoms of psychotic illness. We hypothesized that ACEs impact salience attribution and motivation by altering neural mechanisms of learning.

Methods: Participants consisted of a sample of individuals between 18 and 64 years old (inclusive) with diagnosed schizophrenia or schizoaffective disorder (collectively termed SZ; N = 58; mean age = 39.3; 70.7% male) and a healthy volunteer (HV) group comprised of individuals with no diagnosed psychiatric condition (N = 37; mean age = 42.2; 56.8% male). Participants performed a 3-choice reversal learning task twice, once after being administered an acute stressor (the Socially-evaluated Cold Pressor Task/SECPT), and once after not being stressed. The SECPT involved the participant submerging his/her left hand up to the wrist in water just above freezing (1º-4º C) until the pain became unbearable (up to for 3 minutes), while being filmed by an unsympathetic confederate. In the reversal learning task, choices were rewarded probabilistically, with a choice of the optimal deck (i.e., the one with the highest expected value) leading to a 100-point gain on 90% of trials (and a loss of 50 points on 10% of trials). Choices of two non-optimal decks led to 100-point gains on 50% and 10% of trials (and losses of 50 points on 50% and 90% of trials), respectively. Participants were instructed to try to identify the optimal deck as quickly as possible; they were also informed that, occasionally, a new deck would become the optimal one. Participants achieved as many stages as possible in 240 total trials (4 runs of 60 trials). To quantify task performance, we concatenated all trials within subjects and modeled choices with a Hierarchical Gaussian Filter (HGF) with decision noise. Via Bayesian Model Comparison, we tested whether computational parameters remained stable or changed across conditions (stress vs. control). To assess ACEs in participants, we used the 28-item Childhood Trauma Questionnaire (CTQ), which quantifies 3 kinds of abuse and 2 kinds of neglect. To assess anhedonia and avolition in SZ patients, we used the Clinical Assessment Interview for Negative Symptoms (CAINS). We examined brain responses to precision-weighted prediction errors (PEs) at the second level of the learning hierarchy in a priori volumes of interest (VOIs) in the anatomically defined bilateral ventral striatum (VS).

Results: As revealed by model comparison, we observed no effect of the acute stress condition on any model parameter (protected exceedance probability for model with stable parameters = 1). This was mirrored in the brain findings, which showed only weak main effects for acute stress on the precision weighted PE signal across all subjects [left PCC (-8, -34, 36): F = 15.6, p < 0.001 uncorrected, right putamen (24, 8, -6): F = 12.5, p < 0.001]. Across conditions and subjects, precision weighted PEs were accompanied by BOLD response within the salience network (ACC/vmPFC, striatum, and insula) at p(FWE for the whole brain) < 0.05. Overall, we observed a between-group difference (HV > PSZ) in responses to PEs in right [(6, 8, -8), F = 12.48, p = 0.015] and left [(-10, 10, -12), F = 5.53, p = 0.02] VS (small volume corrected). There were no significant interactions between Group and Stress condition within the clusters of the main effect for PEs. However, we observed interacting effects of group, acute stress condition, and the severity of childhood trauma on VS PE signals [for Condition*Group*CTQ total score interaction: β(SE) = -0.3 (0.11), t = -2.63, p = 0.01; for Condition*Group*CTQ emotional neglect interaction: β(SE) = -0.75 (0.25), t = 3, p = 0.003]. In controls, CTQ Total Scores predicted VS PE responses in the stress condition [β(SE) = -0.21 (0.09), t = -2.44, p = 0.02]. In people with SZ, CTQ Emotional Neglect scores predicted VS PE responses in the non-stress condition [β(SE) = -0.24(0.08), t = -2.85, p = 0.008]. Additionally, Motivation and Pleasure (MAP) scores from the CAINS interacted with CTQ Emotional Neglect scores in predicting attenuated VS PE responses in the non-stress condition, in SZ patients [β(SE) = -0.009(0.003), t = 2.69, p = 0.01].

Conclusions: These results replicate prior findings of attenuated reward prediction error signaling in people with schizophrenia, especially those with more severe motivational deficits. In addition, these findings demonstrate differential effects of ACEs on brain responses to reward PEs in people with schizophrenia and healthy volunteers. Further research is required to identify specific pathways from childhood trauma to schizophrenia symptoms, by way of brain mechanisms of learning and motivation.

Keywords: psychosis, trauma, Reward, Reinforcement learning, uncertainty

Disclosure: Nothing to disclose.

P841. Longitudinal Measurement of Brain Glutamate in Psychosis Spectrum Patients Using 7T GluCEST

David Roalf*, Jacquelyn Stifleman, Arianna Mordy, Ally Atkins, Margaret Pecsok, Kosha Ruparel, Monica Calkins, Ravi Prakash Reddy Nanga, Mark Elliott, Fang Liu, Russell Shinohara, Ravinder Reddy, Raquel Gur, Ruben Gur

University of Pennsylvania, Philadelphia, Pennsylvania, United States

Background: Psychosis commonly develops in adolescence or early adulthood. Malfunctioning neurotransmitter systems, such as glutamate, are implicated in the disease progression of psychosis. Monitoring in vivo alterations of the glutamate system within the brain are spatially limited by traditional magnetic resonance techniques. But recently we have shown that brain glutamate levels are lower across the brain in psychosis spectrum (PS) as compared to typically developing (TD) youth using 7 Tesla glutamate chemical exchange saturation transfer (GluCEST). Here, we extend our previous work in a longitudinal framework and measure GluCEST over a two-year period In this study, we seek to 1) measure developmental time course of glutamate across the cortex in a cohort of TD and PS and 2) to establish longitudinal comparison data of glutamate measures at 7T MRI (1HMRS vs. GluCEST).

Methods: Longitudinal GluCEST data was acquired using a 7T Siemens Terra scanner in 33 TD and 49 PS. GluCEST (slice number: 1, slice thickness: 5 mm) data was acquired in a mid-sagittal plane. GluCEST images were first corrected for B0 and B1 inhomogeneity effects using PyGluCEST analytical pipeline. GluCEST effects were tested using linear mixed models with GluCEST level as dependent variable, age as a continuous predictor, group and sex fixed effects and subject as the random effect.

Results: The effect of age on the outcome GluCEST differed between TD and PS patients for the thalamus and lingual gyrus region (p < 0.05). More specifically, as age increases for PS, GluCEST levels decrease. In addition, age and diagnostic group had varied effect of GluCEST that was regionally distinct. 1HMRS data indicated similar trends for glutamate as GluCEST (PS < TD, p = 0.09) for diagnostic group and age (lower Glu with increasing age).

Conclusions: GluCEST levels declined more in PS subjects than TD, but in regionally specific manner. Our findings describe a pattern of abnormal brain neurochemistry early in the course of psychosis that changes in late adolescence and early adulthood. We suggest that neurochemical profiles of GluCEST contrast across cortex and subcortex may be considered markers of early psychosis. GluCEST methodology thus shows promise to further elucidate the progression of the psychosis disease state and dynamic changes to therapeutic interventions.

Keywords: glutamate, schizophrenia, neuroimaging

Disclosure: Nothing to disclose.

P842. Poster Withdrawn

P843. Multimodal Neuroimaging-Based Prediction of Treatment Response in Early Psychosis

Jason Smucny*, Tyler Lesh, Daniel Berge, Joyce Guo, Cameron Carter

University of California - Davis, Sacramento, California, United States

Background: Response to treatment in psychotic disorders is highly variable, and biomarkers of treatment response have been lacking to date. As a result, trial and error remains the basis for care in early psychosis (EP) and poor outcomes continue to be common in individuals. Early identification of patients who are less likely to have good treatment responses would help identify individuals who would benefit from alternative and/or supplemental interventions. Previous work suggests that frontoparietal activation during a proactive cognitive control task can significantly predict treatment response in EP, albeit with modest accuracy. Here we tested the hypothesis that this classification model can be enhanced by including whole brain gray matter free water (FW), a measure possibly related to neuroinflammation, as a predictor of response.

Methods: Patients with either schizophrenia (n = 32, 27M/5F, mean age 20.4 years) or Type I bipolar disorder with psychotic features (n = 15, 8M/7F, mean age 21.7 years) were scanned at 1.5T during the AX-Continuous Performance task and underwent diffusion imaging at the beginning of treatment. Participants were re-evaluated clinically after 12 months of EP specialty care. Patients were classified as Improvers ( > 20% improvement on Brief Psychiatric Rating Scale (BPRS) Total Score at one-year follow-up vs. baseline) or Non-Improvers. Stepwise logistic regression was then performed using baseline BPRS score and frontoparietal activation in an initial model and whole brain gray matter FW included as a predictor in a final model.

Results: 26 of the 47 patients were classified as Improvers. Frontoparietal activation and FW were higher in Improvers vs. Non-Improvers. The initial model that included baseline BPRS and frontoparietal activation was significant (χ2 = 16.35, p < .001, Nagelkerke R2 = 0.39, accuracy for Improvers = 73%, accuracy for Non-Improvers = 71%). Including gray matter FW as a predictor significantly improved the model (χ2 = 24.36, p < .001, Nagelkerke R2 = 0.54, Δ-2 Log Likelihood if FW term removed = 8.25 (p = 0.004), accuracy for Improvers = 85%, accuracy for Non-Improvers = 76%). Frontoparietal activation and FW were not significantly correlated (r = 0.10, p = 0.52).

Conclusions: These results suggest that gray matter FW makes a significant independent contribution to an initial task fMRI-based model for predicting treatment response in EP. These results suggest that multimodal models may be more powerful than unimodal models for predicting clinical improvement in recent onset illness. Future work in larger samples as well as independent replication are needed to confirm these findings.

Keywords: Schizophrenia (SCZ), Treatment-Response, Magnetic Resonance Imaging, Free water imaging, functional magnetic resonance imaging

Disclosure: Nothing to disclose.

P844. Neural Patterns of Social Touch and the Association With Social Deficits in Schizophrenia

Rene Hurlemann*, Danilo Postin

School of Medicine and Health Sciences, University of Oldenburg, Oldenburg, Germany

Background: Even after successful treatment of the psychotic symptoms, most patients with schizophrenia suffer from persistent impairments in cognitive, occupational and social domains. These impairments hamper participation in everyday life and are often accompanied by increasing social isolation and loneliness. Subtle changes in social behaviors, including social withdrawal, can already be observed in the prodromal phase of schizophrenia (at-risk mental state) and play a crucial role in the chronic course of the disease. Patients increasingly lose contact with their social environment and are deprived of the health-promoting effects of interpersonal relationships, including romantic relationships and interpersonal touch. While previous studies have demonstrated behavioral deficits in various social-cognitive domains in patients with schizophrenia, little is known about disease-related changes in the neural processing of social touch and the association with social deficits in patients with schizophrenia.

Methods: This pre-registered clinical trial (see ClinicalTrials.gov Identifier: NCT04968223) aimed to investigate differences in the neural and behavioural correlates of social touch between patients with schizophrenia (n = 40) and a healthy control group (n = 42) without a family history of psychotic illness. Both groups underwent task-based functional magnetic resonance imaging (fMRI) to examine the neural correlates of social touch. To further characterize potential biases related to social touch, additional physical touch “permission maps” (painted areas in human silhouette indicating touch allowance) were measured. Psychometric questionnaires and semi-structured interviews were used to assess symptom burden and levels of psychosocial functioning at baseline and at 3- and 6-months follow-ups. Relapse rate was measured at 3-, 6- and 12-months follow-ups.

Results: Behavioral results revealed an overall more negative attitude towards social touch in patients with schizophrenia compared to the healthy control group. Patients reported higher levels of touch avoidance in the self-report questionnaire and fewer comfort zones of social touch. In line with the behavioral results, these differences were also evident in the processing of touch at the neural level. Compared to the healthy control group, patients exhibited a reduced somatosensory activation in response to touch compared to no touch. In addition, patients displayed an increased reduction in the activation of the bilateral caudate nucleus in response to touch, which was not present in the healthy control group. This particular activity reduction of the caudate nucleus was significantly associated with social deficits in the patient group.

Conclusions: This study sheds light on the role of close physical contact in the context of social functioning deficits in schizophrenia. Moreover, our results highlight the role of the caudate nucleus as part of the neural basis for social functioning deficits in schizophrenia.

Keywords: social touch, Schizophrenia (SCZ), Functional MRI (fMRI), Social Functioning

Disclosure: Nothing to disclose.

P845. Neuroanatomical Subtypes Indicate Treatment Resistance in Schizophrenia

Koki Takahashi*, Nobuaki Hondo, Shiori Honda, Yoshihiro Noda, Tetsuo Ishikawa, Sakiko Tsugawa, Jose Rubio, Yusuke Iwata, Ryosuke Tarumi, Kamiyu Ogyu, Fumihiko Ueno, Akihiro Koreki, Edgardo Torres-Carmona, Wanna Mar, Teruki Koizumi, Hideo Kato, Keisuke Kusudo, Vincenzo De Luca, Philip Gerretsen, Gary Remington, Mitsumoto Onaya, Hiroyuki Uchida, Masaru Mimura, Ariel Graff-Guerrero, Shinichiro Nakajima

Keio University School of Medicine, Shinju-ku, Japan

Background: Approximately one-third of patients with schizophrenia do not respond to standard antipsychotic treatments. This condition, known as treatment-resistant schizophrenia (TRS), is associated with distinct neurobiological abnormalities including pronounced cortical structural aberrations. However, how cortical structure affects the response to antipsychotics is largely unknown, partly because of the heterogeneity in the pathophysiology of schizophrenia and the limited explanatory power of cortical morphology in terms of antipsychotic response.

In this study, we hypothesized that structural abnormalities across the cortex in patients with schizophrenia can be classified into subtypes with differential molecular mechanisms and thus can explain the differences in TRS/nonTRS.

Methods: We included 190 patients with schizophrenia (108 [56.8%] TRS and 82 [43.2%] non-TRS, 42.9 ± 12.4 years old, 77 females [41%]) and 84 healthy controls (HCs) (41.7 ± 12.4 years old, 36 females [43%]) from 2 sites in Japan (Tokyo and Chiba) and 1 site in Canada (Toronto). All participants provided written informed consent, and the study protocol was approved by the ethics committees at each institute. We acquired T1-weighted magnetic resonance imaging (MRI) scans (3T GE scanner for Tokyo and Toronto and 1.5T GE scanner for Chiba) in all three sites. Cortical thickness for 148 regions defined in the Destrieux atlas were calculated for each participant with FreeSurfer 7.2.0. These metrics, controlling for age and sex as covariates, were then transformed into Z-scores using HC data from each site.

Firstly, patterns of cortical morphology were explored. X-means clustering with all 148 Z-scores as input data and cosine distance as the distance metric was performed to automatically identify cortical structure patterns of schizophrenia. Robustness of clustering was evaluated with leave-one-site-out cross-validation. The cluster-level cortical morphology was assessed with one-sample t-tests of Z-scores across regions with Bonferroni corrections.

Then, the relationship between cortical morphology and TRS/nonTRS was tested with a chi-squared test. The effects of the TRS/nonTRS, site, and their interaction on clustering results were tested with logistic regression.

Moreover, to explore the molecular basis underlying cortical structure, the spatial Pearson correlation analysis with FDR correction between Z-score of the centroid of each cluster and publicly available maps of receptors and transporters in healthy brains were carried out. They included 5HT1A, 5HT1B, 5HT2A, 5HT4, 5HT6, CB1, D1, D2, GABAA, H3, M1, mu-opioid, N-methyl-D-aspartate, and mGluR5 receptors and serotonin, noradrenaline and vesicular acetylcholine transporters.

Results: We identified two distinct cortical morphology clusters (mean silhouette Score=0.32) with cluster 1 (n = 145 [76%], 57 females [39%], 42.3 ± 12.2 years old, 625 ± 417 mg chlorpromazine-equivalent (CPZ mg eq) /day) characterized by widespread cortical atrophy (significant atrophy in 137/148 regions, p-Bonferroni < 0.05) and cluster 2 (n = 45 [24%], 20 females [44%], 44.9 ± 13.0 years old, 527 ± 362 CPZ mg eq/day) by localized cortical hypertrophy (hypertrophy in 24/148 regions, p-Bonferroni < 0.05). The identified clusters were highly robust in leave-one-site-out cross-validation (mean adjusted rand index=0.95)

The ratio of TRS was higher in cluster 1 in comparison with cluster 2 (cluster 1: 92 patients with TRS [63%], cluster 2: 16 patients with TRS [36%]) (χ2 = 10.9, p < 0.001). In logistic regression, the overall model was significant (p = 0.001). The main effect of TRS/nonTRS was significant (p = 0.023) but neither main effect of site nor site-by-response interaction was significant.

In cluster 1, Z-scores showed spatial correlation only with serotonin 5HT2A receptor map (r = -0.28, FDR-corrected p = 0.01). Conversely, in cluster 2, Z-scores showed negative spatial correlation with densities of 5HT1A, D1, H3, and mu-opioid receptor and serotonin, dopamine and vesicular acetylcholine transporters. A positive spatial correlation was found between Z-score and noradrenaline transporter in cluster 2 while no such association was found in cluster 1.

Conclusions: We found distinct cortical morphology subtypes of schizophrenia with unsupervised clustering and cross-validation. Also, these subtypes showed differential antipsychotic responses with widespread cortical atrophy associated with TRS. These results confirm the notion that pronounced cortical damage is associated with TRS. Moreover, these patterns exhibited differential neuroreceptor profiles, suggesting differential involvement of neurotransmitter systems in the pathophysiology of TRS/nonTRS.

In cluster 1, serotonin regions with rich 5HT2A receptor density could be more vulnerable to atrophy, suggesting the importance of 5HT2A antagonism of atypical antipsychotics. In cluster 2, regions with rich target molecules for antipsychotics including 5HT1A, D1, and H3 receptor had more normal structure, possibly suggesting that D2 receptor-independent activity in the cerebral cortex could contribute to the effects of antipsychotics.

In future studies, attempts for both data-driven explainable prediction of treatment response to individual drugs and further understanding of heterogeneous molecular mechanisms of schizophrenia would be warranted.

Keywords: Neuroimaging Analysis, treatment resistant schizophrenia, Machine learning clustering, Schizophrenia (SCZ)

Disclosure: Nothing to disclose.

P846. Precision Mapping of Cerebellar Connectivity to Cognition in Psychosis: Converging Evidence From fMRI and TMS

Hengyi Cao*, Miklos Argyelan, Joanna Yan, Halil Aziz Velioglu, Franky Fang, Andrea Joanlanne, Simran Kang, Lara Prizgint, Jenna Schugart, Philip Watson, John Cholewa, Sunny Tang, Ricardo Carrion, Todd Lencz, Anil Malhotra

Feinstein Institute for Medical Research, Queens, New York, United States

Background: Cerebellar dysfunction has been strongly implicated in both cognition and psychosis, and the “cognitive dysmetria” hypothesis posits an intermediate mechanism of cognition in the pathway from cerebellar dysfunction to psychopathology in psychotic disorders. However, the nuanced pattern of cerebellar connectivity contributing to cognitive function and in turn to clinical symptoms remains unclear. Establishing such linkage is important to identify potential neural target for cognitive enhancement in psychosis.

Methods: In the first study, we investigated a total of 100 patients with early-stage psychosis from the Human Connectome Project Early Psychosis (HCP-EP) study (mean age 22.8 years, 64 males). The entire cerebellum was parcellated into 125 fine-grained functional parcels using the CAB-NP atlas. The cerebellar functional connectome measuring connectivity between each cerebellar parcel and the whole brain was computed from the resting-state data. Cognitive function was assessed using the NIH toolbox including six distinct domains (working memory, episodic memory, attention, executive function, verbal processing, and processing speed). We used connectome-based predictive modeling to evaluate the performance of cerebellar connectome in prediction of each cognitive domain, with significance based on cross validation and permutations. Connections most predictive of cognitive scores were further associated with PANSS symptom assessments using canonical correlation analysis (CCA). Mediation analysis was subsequently conducted linking cerebellar connectivity to PANSS symptoms, with cognitive scores as mediator.

In the second study, 12 patients with a schizophrenia spectrum disorder were recruited for a randomized, double-blind, sham-controlled cerebellar TMS treatment trial (6 active TMS, 6 sham). Each patient was treated for two weeks with an iTBS protocol (one session with 1800 pulses per day, five days per week), targeting the posteriormost part of the cerebellum left to the midline under neuronavigation (corresponding to the left crus 1/2 area). Cognitive evaluations were performed using the Brief Assessment of Cognition in Schizophrenia (BACS) at baseline and end of the 2nd week.

Results: In the HCP-EP sample, the cerebellar connectome significantly predicted three cognitive domains in patients, namely, verbal processing (r [predicted vs observed] = 0.47, p < 0.001), working memory (r [predicted vs observed] = 0.42, p = 0.002), and executive function (r [predicted vs observed] = 0.33, p = 0.01). While each cognitive domain was predicted by a distinct cerebellar connectivity pattern, the left crus 1/2 area turned out to be a common region whose connectivity strongly predicted all three domains. The CCA analysis revealed a significant association between cerebellar connectivity and negative symptoms (canonical r = 0.44, p = 0.03); and the mediation analysis further unveiled that verbal processing score was a significant mediator that fully mediated the relationship between cerebellar connectivity and negative symptoms (p = 0.002).

In the TMS sample, despite the small sample size, we observed a significant group by time interaction effect for overall cognitive composite score (p = 0.025, Cohen’s d = 1.52), with score improvement in the active TMS group but not sham group. Domain-specific analysis revealed large interaction effect on working memory (Cohen’s d = 1.51) and medium-to-large interaction effects on verbal memory (Cohen’s d = 0.76) and executive function (Cohen’s d = 0.73), remarkably consistent with the domains identified in the HCP-EP sample.

Conclusions: These findings suggest a potential causal pathway from cerebellar connectivity to domain-specific cognitive deficits and psychopathology in schizophrenia. Moreover, these findings also point to the cerebellum as a potential neural target for treatment of cognitive dysfunction in patients with psychotic disorders.

Keywords: Psychotic Disorders, cerebellum, Cognition, Resting-state fMRI, transcranial magnetic stimulation (TMS)

Disclosure: Alkermes: Grant (Self).

P847. Exploring the Role of Genetic Factors and Cannabis Use on Cognitive Control in Individuals With Schizophrenia

Tyler Lesh*, Anika Minoji, Tracy Warren, Marina Albuquerque, Sarvenaz Pakzad, Joshua Rhilinger, Alexander Nord, Cameron Carter

University of California - Davis, Sacramento, California, United States

Background: Cannabis use has been linked to increased risk for psychosis and continued use is associated with higher relapse rates, longer hospital stays, and greater positive symptom severity in individuals with schizophrenia. Despite these negative outcomes, many studies have found that patients with a history of cannabis use show higher cognitive performance than never-users. Our own group recently published evidence of higher cognitive control performance and frontal brain activity during the AX-CPT in past cannabis users with schizophrenia (Lesh et al., 2024; Schiz Bull Open). A number of theories attempt to address this conundrum, including higher premorbid cognition, social functioning, and even putative anti-inflammatory effects of cannabidiol in past users. These findings could also be potentially explained by lower genetic liability for schizophrenia in past cannabis users with schizophrenia, which may be linked to higher cognition.

Methods: To test this hypothesis, 65 individuals with recent onset schizophrenia, schizoaffective disorder, or schizophreniform disorder (48 past cannabis users and 17 never-users) were recruited from the EDAPT early psychosis clinic at the UC Davis Medical Center. All participants had no history of other non-cannabis substance use disorders. Participant blood samples underwent genotyping to generate a polygenic risk score (PGS) using the Psychiatric Genomics Consortium GWAS summary statistics for schizophrenia. Participants’ cognitive control ability was measured using the AX-Cognitive Performance Task (AX-CPT) performance during fMRI. Diagnosis and cannabis use was evaluated using a structured clinical interview (SCID-IV). General linear models covarying for age, sex, and nicotine use were used to investigate the relationship between cannabis use and PGS with cognitive control performance (d’-context) and brain activity.

Results: As predicted, schizophrenia patients with a history of cannabis use showed lower PGS compared to never-users with schizophrenia (p < .05). Furthermore, past users showed significantly higher dorsolateral prefrontal cortex activity during cognitive control compared to never-users (p < .05). While only at a trend level (p = 0.11), past users showed some preliminary evidence of higher cognitive control performance based on d’-context scores. No significant differences were identified in global functioning or clinical symptomatology (all p > .24).

Conclusions: These data suggest lower genetic risk may be one of the factors that can explain the paradoxically higher cognition in past cannabis users with schizophrenia. Lower vulnerability might be reflected in more preserved cognition and neurobiology, which could be consistent with higher performance and brain activity in individuals with a history of cannabis use. Based on this perspective, individuals in this group might not have developed a psychotic disorder in the absence of cannabis use, although this is speculative. Recent reports have explored the significant genetic overlap between risk for schizophrenia and risk for cannabis use/abuse, which highlight the complex intertwined nature of these disorders. Future and ongoing prospective studies, such as the Adolescent Brain Cognitive Development study, are also likely to offer important new insights into disentangling these factors.

Keywords: cannabis, Cognition, Schizophrenia (SCZ), fMRI

Disclosure: Nothing to disclose.

P848. Theta Oscillations Assessed From a Mismatch Negativity Paradigm in NAPLS2 Individuals at Clinical High Risk for Psychosis and Healthy Controls: Associations With Clinical Outcomes and Cognition

Jessica Hua*, Brian Roach, Holly Hamilton, Peter Bachman, Aysenil Belger, Ricardo Carrion, Erica Duncan, Jason Johannesen, Gregory Light, Margaret Niznikiewicz, Jean Addington, Carrie Bearden, Kristin Cadenhead, Barbara Cornblatt, Diana Perkins, William Stone, Ming Tsuang, Elaine Walker, Scott Woods, Tyrone Cannon, Daniel Mathalon

UCSF and San Francisco VA Medical Center, San Francisco, California, United States

Background: Mismatch negativity (MMN) amplitude reduction is an event-related potential (ERP) candidate biomarker of schizophrenia. This amplitude reduction reflects N-methyl-D-aspartate receptor (NMDAR) hypofunction, a pathophysiological mechanism underlying psychosis and cognitive dysfunction. Accumulating evidence suggests this reduction is present in individuals at clinical high-risk for psychosis (CHR-P) who subsequently convert to psychosis as well as being predictive of psychosis.

Animal and human studies of schizophrenia highlight the unique contributions of specific oscillatory frequency bands, motivating interest in parsing oscillatory frequencies and phases summed together in MMN ERPs. Time-frequency analysis (TFA) can be used to compute frequency-specific inter-trial phase coherence (ITC; event-related phase consistency) and total power (event-related oscillation magnitude changes) from ERPs. Notably, TFA studies suggest that abnormalities in theta oscillations, which are critical to adaptive and cognitive control and facilitate long-range communication among brain regions, underlie MMN schizophrenia deficits. However, whether deviance-related theta deficits are present prior to psychosis onset and whether they are predictive of psychosis is unclear.

This is the first study to examine associations between theta oscillations with CHR-P clinical outcomes, time to psychosis conversion, and cognition. We hypothesized that: 1) CHR-P converters (CHR-C) would show theta oscillatory deficits vs. healthy controls (HC) and CHR-P non-converters (CHR-NC), 2) theta oscillatory deficits would predict shorter time to conversion in CHR-P, and 3) theta oscillatory deficits would be associated with worse cognition in CHR-P and HC.

Methods: CHR-P (n = 580) and HC (n = 241) were recruited as part of the 8-site North American Longitudinal Prodromal Study 2 (NAPLS2) and completed the MMN paradigm. 77 CHR-P converted to psychosis, and 238 were followed clinically for ≥24 months without converting. Theta ITC and total power differences were assessed using 3x4x2 repeated measures analyses of variance with Group (CHR-C, CHR-NC, HC) as a between-subjects factor, and Stimulus Type (standard, duration-deviant, pitch-deviant, double-deviant) and Fronto-Central Lead (frontal, central) as within-subjects factors. Hierarchical Cox regressions modeled associations between theta standard and deviant metrics (averaged across leads) and time to conversion in the full CHR-P group. Follow-up models tested associations between significant theta metrics and time to conversion over and above positive symptoms. Lastly, general linear models examined the Group x Global Cognition interaction in the full CHR-P group and HC. Interactions were dropped when not significant, providing a test of the common slope.

Results: For theta ITC, there was a Group effect (F[2,553] = 3.37, p = 0.035, d = 0.22) that did not interact with Stimulus Type or Fronto-Central Lead. Follow-up tests showed CHR-C had lower ITC vs. CHR-NC and HC (ps < .047), and no group difference between CHR-P NC and HC (p = 1.000). For theta total power, there was a significant Group effect (F[2,553] = 3.37, p = 0.028, d = 0.23) that did not interact with Stimulus Type or Fronto-Central Lead. Follow-up tests showed CHR-C had lower total power vs. HC (p = 0.027). CHR-NC showed an intermediate pattern, with no significant group differences between CHR-NC and CHR-C or HC (ps < .469).

Lower duration-deviant theta ITC was associated with shorter time to conversion (p = 0.039, 95% CI[0.61-0.98]) in the full CHR-P group. Controlling for positive symptoms, duration-deviant theta ITC remained a significant time to conversion predictor (change from previous model: p = 0.025; duration-deviant ITC: p = 0.027).

As for cognition, Group x Global Cognition interaction was not significant and was dropped for theta ITC. Common slope tests showed worse global cognition was associated with lower ITC for duration-deviants (F[1,763] = 12.65, p < .001, d = 0.26), pitch-deviants (F[1,763] = 7.19, p = 0.008, d = 0.19), and double-deviants (F[1,763] = 15.94, p < .001, d = 0.29), but not standards (p = 0.343, d = 0.06). Follow-up common slope tests showed worse cognition across most individual tests was associated with lower theta ITC for deviants. For theta total power, Group x Global Cognition was significant for pitch-deviants (F[1,762] = 4.42, p = 0.036, d = 0.16), with a stronger association between worse global cognition and less theta power for pitch-deviants in HC vs. CHR-P (p = 0.016). Following-up this result, there was a significant Group x Trail Making Test interaction (F[1,760] = 6.01, p = 0.014, d = 0.18), with a stronger association between worse cognition and less theta total power for pitch-deviants in HC vs. CHR-P (p = 0.014). Global cognition was not associated with theta total power for standards and other deviants (ps > .074, ds=0.03 to 0.13).

Conclusions: Current study results provided evidence that theta oscillatory deficits predate the onset of psychosis and predict a shorter time to psychosis conversion. Further, worse cognition was associated with lower theta activity, consistent with research finding that theta as a key factor in cognition. Importantly, these TFA methods allow for further interrogation of local and circuit level abnormalities underlying ERP MMN amplitude reduction, which in turn can lead to the identification of circuits that can be modulated in psychosis.

Keywords: theta band oscillatory measures, Electroencephalography (EEG), Clinical high-risk for psychosis, Auditory Mismatch Negativity, time-frequency

Disclosure: Nothing to disclose.

P849. Risk of Breakthrough Symptoms on Antipsychotic Medication When Remitted Patients With Schizophrenia are Treated With Long-Acting Injectable Medications: Preliminary Results

Robert Zipursky*, Ofer Agid, Gary Remington, Olivia Spandier, Nicole Sung, Sheng Chen, Wei Wang

University of Toronto, Toronto, Canada

Background: The risk that psychotic symptoms will return is known to be very high if individuals with schizophrenia discontinue maintenance antipsychotic treatment. However, the risk of breakthrough psychotic symptoms is less clear when patients receive maintenance antipsychotic treatment. Recurrences may occur despite maintenance treatment for a number of reasons including poor treatment response, suboptimal maintenance dose, and incomplete adherence. This study was initiated to estimate the risk of breakthrough psychotic symptoms when these considerations are minimized by limiting our study to individuals with schizophrenia who: 1) have experienced a remission of psychotic symptoms with antipsychotic treatment, 2) are receiving maintenance treatment with a second-generation long-acting antipsychotic medication (LAI), and 3) have established adherence.

Methods: All participants were recruited from the outpatient services of the Centre for Addiction and Mental Health (CAMH) in Toronto, Canada. Individuals 18 years of age and older who met DSM-5 criteria for schizophrenia were invited to participate if they met the following inclusion criteria: 1) receiving treatment with long-acting paliperidone, risperidone, or aripiprazole, 2) on a stable dose for at least three months, and 2) a history of improvement as defined by a rating of mild or lower on the Clinical Global Impression Schizophrenia (CGI-SCH) – Severity Scale for Positive Symptoms that has been sustained for three months or longer, and 4) demonstrated adherence to LAI treatment defined as not having received any injections more than 7 days past the due date for the past three months. Participants were excluded if, in the previous three months, they had: 1) been treated with oral antipsychotic medication, 2) had a psychiatric hospitalization, or 3) met criteria for a current major depressive or manic episode. After undergoing a screening visit, participants were assessed at the time of their next LAI injection, and then at 12 weeks, 24 weeks, 36 weeks and 48 weeks using the Clinical Global Impression Schizophrenia (CGI-SCH) and the 24-item Brief Psychiatric Rating Scale (BPRS). Breakthrough psychotic symptoms were defined as a rating of “Much Worse” or “Very Much Worse” on the CGI-SCH –Improvement Scale for Positive Symptoms. Psychiatric hospitalizations and emergency room visits were documented at each visit. Medication side effects and measures of functioning including the Personal and Social Performance Scale (PSP) were assessed at baseline, 24 weeks and 48 weeks. Trough antipsychotic levels were assessed at baseline, 24 weeks and 48 weeks.

Results: To date, 37 participants (12F, 25M) have completed a baseline assessment including 19 receiving paliperidone (Invega Sustenna or Invega Trinza), 1 receiving risperidone (Risperdal Consta) and 17 receiving aripiprazole (Abilify Maintena). Mean age of the sample is 37.2 years (s.d. = 11.9). Race/ethnicity was reported by 30% as White, 35% as Black, 24% as Asian and 11% as Other. At the baseline visit, mean BPRS was 30.2 (s.d.=5.4) and mean PSP was 69.9 (s.d.=15.8). Follow-up assessments have been completed for 29 participants at 12 weeks, 19 at 24 weeks, 15 at 36 weeks, and 10 at 48 weeks. To date, one participant was hospitalized due to relapse, one additional participant experienced breakthrough symptoms during the follow-up period, and one participant was hospitalized for reasons unrelated to psychotic symptoms.

Conclusions: In this study, we have been assessing the 1-year risk of breakthrough psychotic symptoms in individuals with DSM-5 schizophrenia who have experienced a remission of psychotic symptoms and have been adherent with treatment with a long-acting injectable formulation of a second-generation antipsychotic medication. Our preliminary results suggest that the risk of breakthrough symptoms is low. Additional data will be presented on antipsychotic plasma levels and medication side effects.

Keywords: Schizophrenia (SCZ), Long-acting injectable antipsychotics, Maintenance treatment, Relapse

Disclosure: Janssen Inc: Grant (Self). Boehringer Ingelheim: Advisory Board (Self).

P850. Effects of 40 Hz tACS Stimulation on Cognition and Symptoms in Patients With Schizophrenia

Robert Smith*, Xinyi Cao, Yong Lui, Yuanyu Lu, Hua Jin, John Davis, Chunbo Li

New York University School of Medicine and NKI, Woodmere, New York, United States

Background: Transcranial alternating current stimulation is a technique of brain stimulation to modify neural activity and plasticity by entraining more specifically defined cortical oscillation frequencies. Recent reviews of tACS effects in schizophrenia suggest that tACS stimulation may affect symptoms and cognition in schizophrenia, mostly from open label trials and case reports, but there have been very few randomized controlled trials, especially with tACS targeting γ band oscillations. There is considerable evidence of disturbances in gamma cortical oscillations in schizophrenia; there are also deficits in the modulation of γ oscillations during working memory tasks in schizophrenics, and lack of an increase in γ oscillations strength with working memory load similar to that seen in healthy controls.

Methods: The current study was a randomized double-blind trial of 10 sessions of active vs. sham tACS 40 Hz gamma band stimulation in 50 patients with schizophrenia evaluating effects changes in cognition and symptoms conducted in Shanghai China. The primary outcome measure was change in the overall composite score on the MATRICS battery. Secondary outcomes were additional cognitive measures. on the MATRICS battery and other neuropsychological tests, and symptoms as measured on the PANSS scale. tACS stimulation was performed using a Starstim stimulator. Placement of stimuli electrodes was: a) active electrode over the left DLPFC (F3), and b) reference electrode over the right parietal region (P2). Comparison to baseline values were done with evaluations 1-2 days after completion of 10 tACS sessions and 2 and 4 weeks later. The main analysis was a mixed model analysis of difference scores from baseline using SAS mixed procedure with three time points (immediately after 10 tACS sessions, 2 weeks later, and 4 weeks later) with baseline scores as covariate. Additional analysis used actual scores at the four time points in a mixed model without baseline covariate. Side effects were evaluated with a modification of a scale developed by Bruononi and associates. Blinding was assessed by questionnaires of subjects and guesses whether they received active or sham stimulation.

Results: The analyses showed no statistically significant (P < .05) effects on improvement in any of the cognitive measures or PANSS rated positive or negative symptoms. There was a trend (P < .06) for the MATRICS domain of verbal learning to show greater improvement compared to sham within 1-2 days after the 10 tACS sessions. tACS was well tolerated and side effects were minimal. Results of guess questionnaire showed effective blinding of subjects.

Conclusions: The results of this RCT do not support the efficacy of 40 Hz γ tACS stimulation to improve cognition or symptoms in patients with chronic schizophrenia. Three RCTs using alpha or theta band tACS stimulation and other reports of theta band stimulation have shown some more positive effects in patients with schizophrenia and should be investigated further to define the optimal stimulation parameters for tACS trials in patients with schizophrenia.

Keywords: Schizophrenia (SCZ), tACS, Gamma synchronisation

Disclosure: Nothing to disclose.

P851. Ascertainment of Patients at Clinical High Risk for Psychosis - Concordance of SIPS and CAARMS Criteria

Alyssa Bathery, Catalina Mourgues, Nora Penzel, Ered Ariel, Jean Addington, Alison Yung, Rene Kahn, Patrick McGorry, John Kane, Carrie E. Bearden, Martha Shenton, Barnaby Nelson, Ofer Pasternak, Scott Woods*, Monica Calkins, Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ)

Yale University School of Medicine, New Haven, Connecticut, United States

Background: The clinical high risk syndrome for psychosis (CHR) offers a paradigm for the secondary prevention of schizophrenia and other psychotic disorders. Two instruments have been extensively used for CHR ascertainment: the Structured Interview for Psychosis-risk Syndromes (SIPS) and the Comprehensive Assessment of At Risk Mental States (CAARMS). Both have been used extensively in the literature and have displayed excellent psychometric properties and good predictive validity for conversion to psychosis in CHR populations. While both have been shown to accurately identify CHR and they have identical psychosis spectrum symptom construct coverage, the two measures have operationalized CHR syndrome criteria differently. Although there is evidence that the SIPS and CAARMS criteria are comparable in their ability to identify psychosis risk, discrepancies in their operational criteria led to questions regarding the ability to generalize findings from one instrument to the other. Very few studies have examined the concordance of the CAARMS and SIPS criteria, but there is some evidence of significant discrepancies between the two. A new instrument (Positive SYmptoms and Diagnostic Criteria for the CAARMS Harmonized with the SIPS, or PSYCHS, https://doi.org/10.1111/eip.13457) deployed in the Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ) observational study (https://doi.org/10.1093/schbul/sbae011) provides the opportunity to investigate for the first time a large cohort ascertained with both sets of criteria. The harmonization process that led to the development of the PSYCHS achieved full SIPS/CAARMS harmonization for individual symptom ratings and threshold psychosis criteria, but modest harmonization for CHR criteria. Evaluation of overall concordance of CAARMS and SIPS criteria within the PSYCHS is essential to assessing the degree of SIPS/CAARMS criteria harmonization and characterizing the phenomenology of clinical high-risk symptoms experienced by AMP SCZ participants.

Methods: The ongoing AMP SCZ observational study is being conducted at 43 international sites. Certified PSYCHS raters conduct semi-structured interviews that determine all CAARMS (Subthreshold Intensity=SubInt, Subthreshold Frequency=Sub Freq, Brief Limited Intermittent Psychotic Symptoms=BLIPS, and Vulnerability=Vuln) and SIPS (Attenuated Positive Symptoms Syndrome=APSS, Brief Intermittent Psychotic Symptoms=BIPS, and Genetic Risk and Deterioration=GRD) CHR syndromes, which are confirmed in one of five weekly diagnostic consensus calls. Four current statuses as assessed for each SIPS syndrome (Progression=Prog, Persistence=Pers, Partial Remission=Part, and Full Remission=Full). Observational study inclusion criteria required participants meet any CAARMS syndrome or any SIPS syndrome in progression.

Results: In the second AMP SCZ public data release, 720 CHR participants (age 12-30) of both sexes provided complete data on all CAARMS and SIPS subgroups. Among these, 287 (40.2%) met criteria for both SIPS Progression and CAARMS, 427 (59.8%) met criteria for CAARMS but not SIPS Progression, 6 (0.8%) met criteria for SIPS Progression but not CAARMS, and zero by design met criteria for neither. Concordance between the criteria sets was low (Cohen’s kappa = -0.017, Phi = -0.111).

All but six cases (0.8%) met criteria for a CAARMS syndrome. Of the remaining 714, nearly all (708, 99.2%) met criteria for CAARMS SubInt. Only seven (1.0%) met criteria for CAARMS SubFreq and five (0.7%) for CAARMS BLIPS, and all were comorbid with CAARMS SubInt. CAARMS Vuln was observed more frequently (58/714, 8.1%) but only 6 cases (0.8%) were not comorbid with SubInt.

Only 293 cases (40.1%) met criteria for SIPS CHR Progression. Of these, nearly all (287, 98.1%) met criteria for SIPS APSS Prog. Only three (1.0%) met criteria for SIPS BIPS Prog (all comorbid with APSS Prog) and 21 (0.7%) for SIPS GRD Prog with only 6 (0.8%) not comorbid with APSS Prog.

Within the most common CAARMS syndrome (SubInt), SIPS APSS criteria were met for Progression in 281 (39.7%), Persistence in 383 (54.1%), Partial Remission in 21 (3.0%), and Full Remission in 2 (0.3%). SIPS-CAARMS concordance among less frequent CHR syndromes will be reported, as will detailed reasons for syndromal differences.

Conclusions: This first large sample ascertaining CHR participants for both CAARMS criteria and SIPS progression demonstrates that the two criterion sets, although similar in a number of respects, identify largely non-overlapping samples. The CAARMS criteria are generally more inclusive and SIPS progression criteria more restrictive, which appears primarily due to the difference between the system requirements for recency of onset or worsening of symptoms, and/or symptom frequency. As previously reported, the less common CHR syndromes occur infrequently, particularly in isolation. Implications for clinical trial inclusion criteria will likely interact with selection of the primary outcome.

Keywords: Early intervention, Clinical high-risk for psychosis, CNS Clinical Trials

Disclosure: NW PharmaTech: Stock / Equity - Privately Held Company (Self).

P852. The Association of the Urban Exposome and Persistent Distressing Psychotic-Like Experiences

Benson Ku*, Emerald Yuan, Grace Christensen, Lina Dimitrov, Benjamin Risk, Anke Huels

Emory University School of Medicine, Atlanta, Georgia, United States

Background: Urban upbringing is a known risk factor for the development of psychotic disorders. Recent research has demonstrated that domains of social determinants of health (SDOH) (e.g., air pollution and social context) may explain this association. However, various SDOHs have often been studied in isolation. This study identifies distinct SDOH exposure profiles and estimates their associations with persistent distressing psychotic-like experiences (PLE).

Methods: This population-based study used data from the Adolescent Brain and Cognitive Development (ABCD) Study. Participants were recruited from 22 US sites between September 2016 and January 15, 2022. Data from the baseline and three follow-up time points were included. Area-level, geocoded variables spanning five domains of SDOH, including socioeconomic status, education, crime, natural environment, social context, and crime, were studied. A self-organizing map (SOM) clustering method was used to identify SDOH patterns. Persistent distressing PLE was derived from the Prodromal Questionnaire-Brief Child Version using four years of data. Generalized linear mixed modeling tested the association between SOM exposure clusters and persistent distressing PLE as well as physical activities (i.e., team and individual sports), adjusting for individual-level covariates including age, sex, race/ethnicity, highest level of parent education, family-relatedness, and study sites.

Results: Among 8,145 participants (baseline mean [SD] age, 9.92 [0.63] years; 3868 (47.5%) females; 5,566 (68.3%) White, 956 (11.7%) Black, 159 (2.0%) Asian, and 1,480 (18.4%) Hispanic participants), five distinct exposure profiles clusters were identified. Compared to the reference Cluster 1 (suburban affluent areas, 2521 children, 30.9%), Cluster 3 (rural areas with low walkability and high ozone; 1459 children, 17.9%; adjusted OR: 1.34, 95% CI: 1.09—1.64) and Cluster 4 (urban areas with high ADI, high crime, and high pollution; 715 children, 8.8%; adjusted OR: 1.40, 95% CI: 1.08—1.81), were associated with persistent distressing PLE. Team sports mediated 6.14% of the association for Cluster 3, but not for Cluster 4.

Conclusions: This study found that neighborhoods characterized by rural areas with low walkability and urban areas with high socioeconomic deprivation, air pollutants, and crime were associated with persistent distressing PLE. Further research is needed to explore the biopsychosocial pathways through which different environmental factors may impact the development of psychosis.

Keywords: psychotic-like experiences, Neighborhood Socioeconomic Deprivation, Air Pollution, exposome

Disclosure: Nothing to disclose.

P853. Neuromelanin Accumulation in Patients With Treatment-Resistant Schizophrenia: A Cross-Sectional NM-MRI Study

Ryosuke Tarumi, Shiori Honda, Clifford Cassidy, Takahide Etani, Saki Homma, Shunya Sekihara, Yuka Kaneko, Sotaro Moriyama, Yui Tobari, Fumihiko Ueno, Sakiko Tsugawa, Hiroyuki Uchida, Ariel Graff-Guerrero, Shinichiro Nakajima*

Keio University, Tokyo, Japan

Background: Approximately 30% of patients with schizophrenia do not respond to antipsychotic treatment while about 10% of patients show sustained remission. While abnormalities of dopamine function are implicated in the pathophysiology of schizophrenia, reports on treatment responsiveness and presynaptic dopamine function in the nigrostriatal pathway are inconsistent.

Neuromelanin (NM)-sensitive magnetic resonance imaging (MRI) allows in vivo quantification of NM, a product of dopamine metabolism, in the substantia nigra (SN). Few studies have examined the relationship between antipsychotic responsiveness and SN NM levels in patients with schizophrenia.

The purpose of this study was to investigate the relationship between SN NM levels and treatment responsiveness in patients with schizophrenia. We compared SN NM levels among patients with TRS, patients with remission in positive symptoms of schizophrenia, and healthy controls (HCs).

Methods: This study was approved by the Ethics Committee of Keio University School of Medicine and Komagino Hospital (approval numbers: 20170313, 20230003). We included age- and sex-matched patients with TRS, patients with remission in positive symptoms of schizophrenia, and HCs. We used a 3T GE MRI with an 8-channel head coil and applied NM-sensitive MRI (2D GRE MT, TR=260ms) to measure SN NM signals. We also evaluated the severity of symptoms using the Positive and Negative Symptom Scale (PANSS). First, we conducted an analysis of covariance to compare SN NM levels among the groups controlling for age and sex as covariates. Subsequently, we performed correlation analyses to explore relationships between the severity of symptoms and SN NM signals.

Results: Age- and sex- matched 72 participants (age [years]: HC: 47.00 ± 12.41, remission: 45.96 ± 10.73, TRS: 47.67 ± 11.51; 9 females [37.5%] for each group) completed the study. There was no group difference in SN NM levels among the three groups. NM levels did not correlate with clinical symptom severity in the patient groups.

Conclusions: Our results suggest that dopamine function in the SN may be similar among patients with TRS, patients with remission of positive symptoms, and HCs. No significant difference in SN NM levels between the TRS and HC groups is keeping in line with previous studies employing positron emission tomography on presynaptic dopamine function. On the other hand, no significant difference between the remission and HC groups was inconsistent with previous PET studies noting higher or lower dopamine synthesis capacity in the striatum in patients with remission of positive symptoms in comparison with HCs. Brugger et al. (2019) demonstrated that individual variability in DA synthesis capacity is similar between patients with schizophrenia and HCs, suggesting no significant difference in presynapcti dopamine function in relation treatment response in this population. The relationship between dopamine activity and treatment responsiveness in schizophrenia may vary depending on brain regions or symptom severity. Further research is required to examine the relationship between treatment responsiveness and dopamine activity in the nigrastriatal pathway of schizophrenia employing multimodal neuroimaging methods.

Keywords: Schizophrenia (SCZ), Schizophrenia, Antipsychotics, Neuromelanin-sensitive MRI, treatment-resistant schizophrenia

Disclosure: Nothing to disclose.

P854. Dysregulated mRNA Translation and Schizophrenia-Relevant Behaviours in Mice

Brandon Rodrigue, Edna Matta-Camacho, Vern Lewis, Argel Aguilar Valles*

Carleton University, Ottawa, Canada

Background: Schizophrenia (SCZ) is a heterogeneous neurodevelopmental disorder and a leading cause of disability worldwide. Several studies have implicated the mammalian target of rapamycin complex 1 (mTORC1) – eukaryotic initiation factor 4E binding protein (4E-BP) translation initiation pathway in the therapeutic effects of antipsychotics. Furthermore, a recent exome study identified a single nucleotide polymorphism in the EIF4EBP2 gene (encoding for 4E-BP2) in an SCZ patient. Still, this pathway’s causal role in this disorder’s development has not been explored. Therefore, this study aims to examine how this mutation in the 4E-BP2 gene leads to a SCZ-like phenotype in mice.

Methods: In vitro phosphorylation analysis of the mutant 4E-BP2 by mTORC1 was performed. Then, to test the effects of this mutation in vivo, mice bearing the single nucleotide mutation in 4E-BP2 were created using CRISPR-Cas9 (4E-BP2 knock-in [KI]). These mice were compared to mice lacking the 4E-BP2 gene (4E-BP2 KO mice). Mice from the wildtype and mutant strains were subjected to behavioural assays, including amphetamine-induced locomotion, pre-pulse inhibition of acoustic startle (PPI) and Y-maze.

Results: In vitro analysis of mutant 4E-BP2 indicated that phosphorylation by mTORC1 was increased compared to WT controls, suggesting a greater rate of inactivation of this protein. In the behavioural tasks, we found that adult male 4E-BP2 KI and 4E-BP2 KO mice had a greater locomotor response to amphetamine compared to WT controls. In contrast, female 4E-BP2 KI mice responded to amphetamine like wildtype mice. Male and female 4E-BP2 KO mice had deficits in the PPI task.

Conclusions: These results suggest the SCZ-associated single nucleotide polymorphism in 4E-BP2 acts as a loss-of-function mutation, leading to an SCZ-like phenotype in male mice.

Keywords: mRNA translation, Schizophrenia (SCZ), mTORC1

Disclosure: Nothing to disclose.

P855. Impact of Homeostatic Sleep Pressure on Hypocretin Neuron Activity

Oscar Gonzalez*, Jorge Morga-Rodriguez, Gordon Wang, Julie Kauer, Luis de Lecea

Stanford University School of Medicine, Stanford, California, United States

Background: Sleep is a universal behavioral state observed across all animals. Its role in learning and memory has been recently established, and disrupted sleep has been associated with various neurological and psychiatric disorders. As the amount of time spent awake increases, there is a buildup of pressure to sleep. Following sleep deprivation, this buildup of sleep pressure leads to homeostatic sleep rebound. How the amount of time spent in awake impacts specific brain circuits, thereby driving it towards sleep, remains to be fully understood. It is likely that changes in activity of wake-promoting hypocretin neurons of the lateral hypothalamus may be impacted by sleep pressure.

Methods: In this study we performed simultaneous in vivo fiber photometry and EEG/EMG recordings in Hcrt-IRES-Cre knock in mice (n = 4; 3-4 month old; both sexes). We recorded GCaMP6s signal from Hcrt neurons transduced with AAV-DJ-EF1a-DIO-GCaMP6s. Mice were maintained on a 12h light-dark cycle. Mice were habituated to recordings cages and cables for 3 days then recordings were done starting at either ZT0, ZT12, ZT6, or ZT6 + 6h sleep deprivation. Offline staging of the EEG was done on Matlab using AccuSleep. Raw photometry signals were corrected for photobleaching by fitting the signal with a double exponential and subjecting the fit. dF/F was computed for each signal and then converted to a z-score. Z-scored time series were the used for subsequent data analysis. In addition to the in vivo recordings, we performed ex vivo whole-cell patch clamp recordings in acute brain slices collected from mice genetically expressing eGFP in hypocretin (Hcrt) neurons (3-4 month old mice; both sexes; n = 3-5 animals per time point). 220-um thick coronal brain slices were collected from Hcrt-eGFP mice at one of 4 time points (ZT0, ZT12, ZT6 or ZT+6h sleep deprivation). Metrics of intrinsic excitability were measures using a K-gluconate internal solution and spontaneous post-synaptic currents were recorded with a CsCl internal solution. All data analysis for both fiber photometry and whole-cell patch clamp was performed in Matlab and Python using standard libraries and modules. Normality of data was tested using the Shapiro test and subsequent analysis for statistical significance was done using a t-test, Mann-Whitney U test, or ANOVA.

Results: We found that hypocretin activity differed during periods of high (ZT0) and low (ZT12) sleep pressures. We analyzed the GCaMP signal from Hcrt neurons at the sleep to wake transition in freely behaving mice using fiber photometry. We found a significantly larger Hcrt GCaMP signal at sleep to wake transitions in recordings starting at ZT0, when mice have highest amounts of sleep pressure, as compared to ZT12 recordings, when sleep pressure is lowest (40.1% +/- 5.5% (ZT0) vs 20.2% +/- 3.9% (ZT12) increase from baseline; Mann-Whitney U test; p < 0.05). The increased signal at sleep to wake transitions during periods of high sleep pressure (ZT0) was also reflected in the normalized area under the curve (1.17 + /- 0.23 (ZT0) vs 1.0 + /- 0.02 (ZT12); t-test; p < 0.05). To test whether this phenomenon was driven by changes in intrinsic excitability or synaptic currents of Hcrt neurons, we performed ex vivo whole-cell patch clamp recordings from Hcrt neurons expressing eGFP in acute brain slices collected at ZT0, ZT12. We no difference in the resting membrane potential, action potential threshold, rheobase, or evoked firing between these two conditions (p > 0.05; n = 15 neurons per condition). Since we observed no difference in intrinsic excitability between the high and low sleep pressure conditions, we next explored the possibility that changes in excitatory synaptic inputs onto Hcrt neurons may underly the observed differences in photometry data at these time points. We recorded spontaneous excitatory post-synaptic currents (sEPSC) from Hcrt neurons expressing eGFP in acute brain slices collected at ZT0, ZT12. Bath application of 30uM Bicuculine + 1uM Strychnine was used to pharmacologically isolate EPSCs in the slices. We found that sEPSCs recorded under high sleep pressure (ZT0) were significantly larger (-24.4 pA +/- 3.36pA) than those recorded at low sleep pressure (ZT12) (-10.8 pA +/- 0.54 pA) (t-test p < 0.01; n = 8 neurons per condition).

Conclusions: We found significant differences in the amplitude of Hcrt responses at sleep to wake transitions under different amounts of sleep pressure. Furthermore, we provide evidence suggesting that the mechanism driving this change may involve accumulated changes of excitatory synaptic currents onto hypocretin neurons across extended periods of sleep or wake. Together, these data provide novel insights into the mechanisms underlying local circuit sleep pressure in wake-promoting hypocretin neurons of the lateral hypothalamus, which may drive wake/sleep transitions.

Keywords: sleep, hypocretin, sleep homeostasis

Disclosure: Nothing to disclose.

P856. Sleep and Social Anhedonia in Adolescents: Combining Experience Sampling Methodology and Smartphone Accelerometry

Jaclyn Kirshenbaum*, Zia Bajwa, David Pagliaccio, Esha Trivedi, Katherine Durham, Nicholas Allen, Stewart Shankman, Randy Auerbach

Columbia University, New York, New York, United States

Background: Anhedonia is characterized by a reduced ability to experience pleasure and is a cardinal symptom of major depressive disorder. The prevalence of anhedonia increases during adolescence and social anhedonia, in particular, is related psychopathology symptoms (e.g., depression, suicidal ideation) relative to general anhedonia in adolescents. At the same time, sleep behavior and patterns (e.g., onset, duration) undergo substantial change. Approximately 70% of adolescents do not receive the recommended amount of daily sleep, which can impair psychological functioning. Given that sleep health is a common target of interventions, it is important to enhance our understanding of how sleep contributes to proximal risk for social anhedonia to inform early intervention and prevention efforts. Smartphone sensors (i.e., mobile accelerometry) provide an unobtrusive and scalable way to monitor daily sleep health in real-time over longer time periods. However, given the nascency of smartphone sensors to measure sleep, it is important to test whether smartphone sleep features relate to other measurements of sleep (e.g., experience sampling methodology [ESM]). Furthermore, it is important to test day-to-day longitudinal associations linking sleep to social anhedonia levels to identify opportunities for intervention. Accordingly, this study tests whether: (1) smartphone sensor sleep features relate to daily subjective reports of sleep and (2) proximal changes in sleep health associate with increases in daily social anhedonia. We hypothesized that diminished sleep health (i.e., shorter sleep duration; later bedtimes) assessed through ESM and smartphone sensors will be associated with daily increases in social anhedonia.

Methods: Participants included adolescents (N = 154, 73% female, 13-18-years-old) recruited from the greater New York City and Chicago areas as part of a larger study examining social processing and depression. Eligible participants participated in a 12-month study and were categorized as healthy (HC, n = 54), remitted depressed (remMDD, n = 66), or currently depressed (MDD, n = 27). At baseline, participants installed the Effortless Assessment Research System (EARS) application on their personal smartphones, allowing access to motion sensor data over 12 months. Activity states were derived from continuous passive monitoring of motion data to quantify sleep periods. The start and end times of the longest phone stationary period within each 24-hour window were used to operationalize bedtimes and risetimes, respectively. Following data quality checks, sleep duration was quantified by the difference between bedtime and risetime. For one week following their baseline assessment and approximately one week prior to a 6- and 12-month follow-up assessment, participants received daily prompts via ESM regarding their sleep quality (i.e., restfulness), sleep duration from the previous night, and three prompts per day regarding their negative affect, positive affect, and social anhedonia. First, analyses examined whether daily subjective ESM sleep items were associated with smartphone sensor sleep features. Second, multilevel linear regressions analyses examined the longitudinal association between daily subjective ESM sleep items (i.e., quality, duration) with social anhedonia. Last, multilevel linear regressions examined longitudinal associations between smartphone sensor sleep metrics and ESM-based anhedonia.

Results: ESM-reported sleep duration and was significantly correlated with smartphone sensor sleep duration (β[95%CI] = 0.36[0.31, 0.40], p

Conclusions: Overall, these effects highlight the possibility of applying scalable tools (i.e., software installed on personal smartphones) to identify short-term increases in social anhedonia among adolescents. In the future, these tools may be leveraged to develop just-in-time interventions, focusing on modifying specific sleep behaviors in order to reduce the risk of anhedonia during adolescence.

Keywords: social anhedonia, Sleep disturbances, experience sampling methodology, smartphone, Adolescence

Disclosure: Nothing to disclose.

P857. Experimentally-Induced Inflammation in the Mouse Hind Paw Promotes and Fragments Active-Phase Sleep

Dominika Burek*, Khairunisa Ibrahim, Andrew Hall, Ashish Sharma, Erik Musiek, Jose Moron-Concepcion, William Carlezon, Jr.

Harvard Medical School McLean Hospital, Belmont, Massachusetts, United States

Background: Chronic pain is prevalent, highly comorbid with mental health disorders, and a persistent public health challenge. As a syndrome, pain comprises physical, emotional, and cognitive symptoms such as disability, negative affect, feelings of stress, and fatigue. Classical behavioral outcomes such as exploratory behavior and stress coping have proven insufficient in aiding the identification of biomarkers, treatment targets, and novel therapeutics, particularly in preclinical models of persistent and chronic pain. Objective and continuous measures, such as sleep telemetry recordings, provide a more complete understanding of how pain may act as a stressor, over periods of time that more accurately reflect chronic conditions as they appear in humans. A rodent model of long-term inflammatory pain, induced by injection of Complete Freund’s Adjuvant (CFA) into a hind paw, has previously been shown to cause anhedonia and dysregulated naturalistic behaviors, in a manner similar to animal models of stress. Here, we examined whether this extended to alterations in circadian rhythms and sleep, such as those induced by chronic social defeat stress.

Methods: For circadian rhythm studies, male and female C57BL/6J mice were individually housed in cages equipped with passive infrared sensor wireless nodes for daily locomotor activity recording in either a 12h light/12h dark cycle or constant darkness condition for 7 days of baseline, 7 days after saline (vehicle, control) injection, and 7 days after CFA injection. For measurement of sleep and wake states, male and female C57BL/6J mice were individually housed in cages on an automated piezoelectric sleep monitoring system for daily breathing and gross body movement recording in a 12h light/12h dark cycle for 2 days of baseline, 7 days after saline injection, and 7 days after CFA injection. For sleep architecture analyses, male and female C57BL/6J mice were implanted with wireless transmitters for EEG, EMG, body temperature, and locomotor activity recording in a 12h light/12h dark cycle for 7 days of baseline, 7 days after isoflurane exposure (anesthetic to enable restraint in implanted animals, control), 7 days after saline injection, and 21 days after CFA injection. Piezosensor and EEG data were separated into light phase (12 hours of lights-on, a period when mice are normally more likely to be sleeping) and dark phase (12 hours of lights-off, when mice are normally more likely to be active) and analyzed with repeated measures 1-way ANOVAs and Tukey’s multiple comparisons tests.

Results: CFA selectively reduced intra-daily variability in locomotor activity, a measure of the fragmentation of a 24-hour rhythm, in standard light cycling conditions only (p < 0.05). Saline and CFA injection both reduced the percent variance, an indirect measure of daily phase stability (p < 0.05) and the circadian period length (p < 0.001) in standard light cycling conditions. Saline and CFA injection also both increased relative amplitude, the ratio of the most active 10 hours to the least active 5 hours, in constant darkness (p < 0.05). Next, using piezosensor tracking, we observed that CFA decreased wake bout length (p < 0.0001) and increased sleep duration (p < 0.001) in the first week after injection. EEG recording indicated this effect was due to increases in both rapid eye movement (REM) and slow-wave sleep (SWS): specifically, CFA increased REM duration (p < 0.05), REM bouts (p < 0.01), SWS duration (p < 0.05), and SWS bouts (p < 0.05), particularly in the dark phase.

Conclusions: CFA-induced inflammatory pain profoundly altered sleep architecture in male and female mice. Injection of the hind paw, whether with CFA or saline, reduced some measures of circadian rhythmicity such as variance, period, and amplitude. CFA increased sleep duration primarily in the dark phase—when mice are more likely to be active—whereas it decreased sleep bout length in the light phase and increased in the dark phase. Additionally, CFA reduced wake bout length, especially during the dark phase. The overall increase in sleep was due to increases in both REM and SWS. Increases in REM and SWS duration and bouts were most significant in the dark phase, regardless of whether CFA had been injected at its onset or 12 hours prior. Taken together, these results indicate that inflammatory pain acutely promotes but also fragments sleep.

Keywords: sleep, pain models, inflammation, circadian, telemetry

Disclosure: Nothing to disclose.

P858. Examining Variants in Mitochondrial DNA and Sleep Measures in Individuals With Depression

Fernanda dos Santos, Natasha Camasho, Scott D. Gordon, Brittany L. Mitchell, Enda M. Byrne, James Kennedy, Nicholas G. Martin, Lindsay Melhuish-Beaupre, Vanessa Gonçalves*

Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada

Background: Major depressive disorder (MDD) is a leading cause of disability worldwide, with a lifetime prevalence ranging up to 21%. Around 75% of the individuals suffering with MDD experience sleep disturbances. Mitochondrial function is intrinsically related to sleep. The circadian rhythm, a mechanism regulating sleep and metabolism, is modulated via mitochondrial dynamics. Moreover, the daytime buildup of reactive oxygen species (ROS), produced during mitochondrial oxidative phosphorylation, fosters the ideal metabolic conditions for a state of sleep. Thus the impairment of mitochondrial function can contribute to sleep disturbances, especially in MDD patients. The aim of this work was to examine whether mitochondrial DNA (mtDNA) variants are directly associated, and/or interact with nuclear DNA, to influence sleep disturbances in MDD.

Methods: To achieve this goal, we obtained individual-level genotype data and insomnia severity scores from 8,403 individuals diagnosed with MDD (75.4% female; aged 45.3 y.o. [s.d.= 14.8]) from two Australian cohorts. Insomnia severity was evaluated using the Insomnia Index (ISI) scale. A quality control pipeline was applied, and only mtDNA variants with overall missingness < 2% were analyzed. mtDNA variants with minor allele frequency (MAF) > 1% were classified as common (n = 38), and those with MAF < 1% were classified as rare (n = 63). Logistic and linear regressions were fitted in R to assess the 1) main effect of common variants and 2) the interaction of them with MDD polygenic risk score (MDD-PRS) in regard to sleep disturbance phenotypes. A kernel-regression-based association with SKAT-O method was used to assess the 3) effect of rare variants. Sex, age and 10 genetic principal components were included as covariates in all analyses.

Results: Results from the common variants analyses were meta-analyzed using METAL. As a sensitivity analysis, all analyses were repeated after filtering out individuals who did not work regular hours. Multiple mtDNA variants were nominally (p < 0.05) associated with sleep disturbances in MDD. After, multiple comparison correction, the mt variants m.16391G > A (MT-DLOOP1), m.10034T > C (MT-TG) and m.13780A > G (MT-ND5) remained significantly associated with problems waking up too early (ISI3) (p < 0.0019). Moreover, the variant m.4917A > G (MT-ND2) seemed to interact with MDD-PRS to modify the risk for difficulties staying asleep (ISI2). No evidence of association of rare variants with sleep disturbances was found.

Conclusions: These results shed light on the complex mechanisms underlying sleep disturbances in MDD, and open new avenues for the study of novel targets and markers.

Keywords: Mitochondrial DNA, Major Depressive Disorder (MDD), Sleep disturbances, insomnia, Mitochondria

Disclosure: Nothing to disclose.

P859. Maternal Immune Activation Alters Circadian Rhythms in Parvalbumin Interneurons

Jennifer Donegan*, Andrew Hynes, Laura Fonken

Dell Medical School at the University of Texas at Austin, Austin, Texas, United States

Background: The circadian system is a 24 h time keeping system that regulates multiple physiological processes throughout the body. Multiple lines of evidence link circadian disruptions to neuropsychiatric disorders, such as schizophrenia. In addition, people with schizophrenia consistently demonstrate a loss of inhibitory interneuron function in key brain regions, including the medial prefrontal cortex (mPFC). Specific deficits occur in parvalbumin (PV) interneurons, fast-spiking GABAergic cells that inhibit pyramidal cell firing and control synchronized network activity. Previous work shows that circadian control of PV interneurons in the visual cortex contributes to plasticity and visual system development; however, the role of circadian rhythms in prefrontal cortical and hippocampal PV interneuron activity and schizophrenia-like behaviors remains unknown. In the current experiments, we will test the hypothesis that prefrontal cortical and hippocampal PV interneurons display circadian rhythms and examine whether maternal immune activation alters these rhythms.

Methods: Maternal immune activation (MIA) was used to alter development and produce schizophrenia-like behavioral deficits. Specifically, MIA was induced in C57BL/6J mice with an injection of Poly I:C (20 mg/kg) on embryonic day 12.5. Adult MIA offspring were euthanized at Zeitgeber Time 10 (ZT10, light phase) and ZT22 (dark phase). RNAscope was then used to measure the expression of circadian genes NR1D1 (Rev-erbα) and PER1 (Period 1) in parvalbumin interneurons of the mPFC.

Results: We observed a significant interaction between prenatal treatment and time of day in both NR1D1 and PER1 expression in PV interneurons of the mPFC. Specifically, in control animals, there was a significant increase in NR1D1 expression between ZT10 and ZT22. This differential expression was absent in MIA offspring. In addition, PER1 expression was significantly increased in MIA offspring compared to controls at ZT10.

Conclusions: Together, these data suggest that clock gene expression in PV interneurons of the mPFC varies across the light cycle, and this gene expression is altered by MIA. Future experiments will manipulate circadian gene expression in PV interneurons to determine the effect on behaviors relevant to schizophrenia. Our results may have important implications for the treatment of schizophrenia and other disorders associated with disrupted circadian rhythms.

Keywords: parvalbumin interneurons, maternal immune activation, Medial Prefrontal Cortex, circadian rhythms

Disclosure: Nothing to disclose.

P860. Modulating Kynurenic Acid and Orexin Activation to Improve Sleep-Wake Quality in Aging

Snezana Milosavljevic, Mindal Reese, Maria Piroli, Andrew Smith, Joseph McQuail, Homayoun Valafar, Jim Fadel, Ana Pocivavsek*

University of South Carolina School of Medicine, Columbia, South Carolina, United States

Background: Advanced age presents challenges that increase the likelihood of poor quality sleep-wake experiences. Disruptions in sleep that come with aging may result in fatigue, irritability, headaches, and deteriorating cognition. Aging pathology points to dysfunctional cholinergic transmission, which may be driven, in part, by altered regulation of hypothalamic orexin neurons, recognized for their role in regulating sleep-wake cycles. We hypothesize that kynurenic acid (KYNA), an endogenous antagonist of cholinergic neurotransmission, plays a mechanistic role in age-related sleep dysfunction and orexin activation. The small molecule KYNA is an astrocyte-derived metabolite of the kynurenine pathway of tryptophan catabolism, synthesized primarily by kynurenine aminotransferase II (KAT II) in the brain. An excess of KYNA in the brain results in dysfunctional sleep architecture and cognitive impairment (Pocivavsek et al. 2017 Sleep). We hypothesize that KYNA contributes to age-related alterations in physiological regulation of orexinergic neurons in the lateral hypothalamus.

Methods: Young (4-6 months) and aged (24-28 months) male and female Fischer 344 x Brown Norway F1 hybrid rats were used in all experiments. In vivo microdialysis was conducted in the lateral hypothalamus in 1 h fractions; extracellular KYNA levels were evaluated at baseline and after injection with kynurenine (5 mg/kg, i.p.), the direct bioprecursor of KYNA, to stimulate de novo KYNA formation (N = 6-13 per group). Separate rats were implanted with telemetry transmitters for continuous electroencephalography (EEG) and electromyography (EMG) data acquisition. EEG/EMG waveforms were classified into rapid eye movement (REM) sleep, non-REM (NREM) sleep and wakefulness by a trained artificial neural network, a supervised predictive machine learning technique with high accuracy, and reviewed by an expert hand-scorer. Sleep recording experiments evaluated basal sleep for 24 h and in subsequent experiments de novo KYNA synthesis was inhibited during the latter half of the light phase with KAT II inhibitor (PF-04859989, 30 mg/kg, s.c.) treatment at Zeitgeber time (ZT) 6 (N = 10 per group). A separate cohort of rats received PF-04859989 (30 mg/kg, s.c) at ZT0 or vehicle, and brains were collected at ZT6 to evaluate neuronal activation in the lateral hypothalamus with immunohistochemistry (IHC) to colocalize cFos and orexin (N = 4-7 per group). Lastly, we conducted additional immunohistochemical analysis of brain sections from male young and aged rats to evaluate orexin, astrocytic marker GFAP and KAT II expression in the lateral hypothalamus.

Results: In vivo microdialysis showed significant increase in basal KYNA in aged compared to young males (*P < 0.05) in the lateral hypothalamus. Formation of KYNA with kynurenine challenge was enhanced in aged compared to young males (*P < 0.05). Both sexes of aged rats exhibited increased NREM sleep duration and reduced wake duration during the latter half of the dark phase (P < 0.01). The higher number of REM bouts (P < 0.001), NREM bouts (P < 0.001) and wake bouts (P < 0.001) in the dark phase in aged rats suggests more fragmented sleep. KAT II inhibitor treatment restored sleep-wake behavior in aged rats such that NREM sleep and wake durations during the dark phase were at the level of young rats. Sleep architecture was also more consolidated in aged rats following KAT II inhibitor treatment with a reduced number of REM sleep, NREM sleep and wake bouts in the dark phase. Neuronal activation studies revealed enhanced activation of orexinergic neurons in aged compared to young males during the light phase (ZT 6)(P < 0.01). Of note, acute PF-04859989 treatment restored orexinergic activation to the levels of young males (P < 0.01). Immunofluorescence in male rats confirmed overlap between GFAP and KAT II within astrocytes and notably aged rats showed reduced orexin expression, but increased KAT II within the lateral hypothalamus.

Conclusions: Our findings place new attention on the interplay between KYNA, aging and orexin neurons in the lateral hypothalamus. Our data suggest that mitigating elevations in KYNA in aged rats effectively treats sleep-wake behavioral abnormalities experienced with aging. Taken together, reducing brain KYNA levels via inhibition of KAT II enzyme could be a potential therapeutic target for improving deteriorating sleep quality associated with aging.

Keywords: Tryptophan catabolites (TRYCAT), Kynurenine pathway, sleep disturbance, orexin/hypocretin

Disclosure: Nothing to disclose.

P861. Seasonality of Dorsolateral Prefrontal Cortex Transcript Expression in Bipolar Disorder, Schizophrenia and Depression

Katherine Lyman*, Madeline R. Scott, Kyle D. Ketchesin, Kaitlyn Petersen, RuoFei Yin, Xiangning Xue, Jill Glausier, David Lewis, Marianne L. Seney, George Tseng, Colleen A. McClung

University of Pittsburgh School of Medicine Pittsburgh, Pennsylvania, United States

Background: Seasonal patterns have been clinically noted in psychiatric illness; for example, mania peaks in the spring for some patients, suicidality has been observed to peak in spring/summer, and psychosis symptoms may become exacerbated during summer months. However, these phenomena have yet to be fully explained at a physiological level. Our lab has previously used time of death to calculate transcript expression rhythms in human postmortem brain tissue, and found that there are large differences in gene expression rhythms associated with psychiatric illnesses. Given these findings, we sought to use date of death in a similar analysis to calculate seasonal rhythms of gene expression. We hypothesized that subjects with psychiatric illnesses have differences in seasonal gene expression rhythms in the brain that may contribute to the seasonal patterns in clinical features observed in these disorders.

Methods: Postmortem brain tissue samples were taken from 33 bipolar disorder (BD) subjects (21 male, 12 female), 50 schizophrenia (SCZ) subjects (47 male, 13 female), 83 major depressive disorder (MDD) subjects (58 male, 26 female), and 83 non-psychiatric comparison (NPC) subjects (116 male, 50 female). Samples were selected from brain banks in Pittsburgh and Nashville; subjects ranged in age from 18-65 and their dates of death were distributed across the calendar year. Tissue samples were extracted from the dorsolateral prefrontal cortex (DLPFC), after which RNA sequencing was performed. RNA-seq data was then processed via a time-series analysis fitting the expression data for each gene to a mathematical model representing seasonal rhythms via a sinusoidal curve, with a period length of 365 days. For each gene’s sinusoidal curve, p-value was calculated, as well as amplitude (A) and peak (i.e., day of the year representing peak gene expression).

Results: Of 16,885 total genes sequenced in the DLPFC, 6227 genes were shown to have a significant seasonal pattern (p < 0.05 and A > 0.15) in the BD group, compared with 433 genes in the MDD group, 360 genes in the SCZ group, and 70 genes in the NPC group. These results indicate a much higher prevalence of seasonality in the BD group, as well as a milder effect of increased seasonality in the MDD and SCZ groups compared to NPCs.

Regarding biological sex, male subjects showed a mild increase in seasonally significant genes compared to females in the MDD group (Male: 785; Female: 380 genes) and in the comparison group (Male: 211; Female: 170 genes). However, when analyzing sex differences in a combined BD/SCZ group, females showed far more of a seasonality effect than males (Male: 889; Female: 2300 genes).

Conclusions: For healthy subjects, gene expression within the DLPFC appears to be stabilized throughout the calendar year, with few genes demonstrating significant vacillation in expression related to the seasons. However, we found a significant seasonal destabilization amongst subjects with BD, for whom approximately 1/3 of the total genes expressed in the DLPFC show a pronounced effect of season. While less pronounced, there is also a seasonal destabilization effect observed in subjects with MDD and SCZ. These findings suggest a physiological seasonal response regulation process which may be overactive in psychiatric illness, and which requires further study to understand how these transcript expression rhythms relate to the clinical patterns observed in these illnesses. Future studies will also require investigation in a sex-specific manner given our findings of differential seasonal effects between females and males.

Keywords: Seasonality, dorsolateral prefrontal cortex (DLPFC), Bipolar Disorder, Postmortem Human Brain Tissue

Disclosure: Nothing to disclose.

P862. The Role of Paraventricular Nucleus of Thalamus in the Sleep Disturbance Induced by Withdrawal From Repeated Ethanol Exposure

Seungwoo Kang*

Augusta University Medical College of Georgia, Augusta, Georgia, United States

Background: Sleep disturbances are a common phenomenon seen in individuals with alcohol use disorder (AUD), both in time of drinking and abstinence. Specially, alcohol withdrawal-induced sleep disturbance adversely affects the condition of AUD patients and has emerged as a major negative reinforcement in relapse drinking. Unfortunately, our understanding of the brain mechanisms underlying the impact of alcohol withdrawal on sleep remains incomplete, posing an urgent need to identify the detailed cellular and molecular target for innovative AUD therapeutic potentials.

A growing body of research demonstrates the importance of the neuronal activities in the paraventricular nucleus of thalamus (PVT) in sleep homeostasis. Histological and functional studies have shown that PVT neurons relay the signals among the brain regions where all have been spotlighted recently because of their critical roles in arousal and the related modulation in sleep/wake cycle. Indeed, selective inhibition of projecting glutamatergic PVT neurons strongly induces sleep behavior, while their activation suppressed sleep. However, gaps remain to be filled with respect to how chronic intermittent ethanol exposure (CIE) impacts maladaptive activities in the PVT and consequent behavioral outcomes in sleep homeostasis.

Methods: To investigate this, we compared the cellular activities in the PVT and behavioral consequences of mice withdrawn from chronic intermittent ethanol exposure (CIE) and ethanol naïve counterparts (CON) using ex vivo and in vivo electrophysiology, in vivo calcium and neuromodulator imaging, virus-based gene transfer, designer receptors exclusively activated by designer drugs (DREADDs)-mediated chemogenetics, and electroencephalogram (EEG)/electromyograpy(EMG)-based sleep cycle testing. Briefly, mice were exposed to air or vaporized ethanol in a vapor inhalation chamber for four weeks. Each daily cycle consisted of ethanol vapor for 16 h followed by 8 h of abstinence in their home cage. This was repeated each day for 4 consecutive days, followed by 3 days of abstinence. The cellular activities and sleep patterns were evaluated after 24 withdrawal from the last exposure of CIE paradigm.

Results: We found that withdrawal from chronic intermittent ethanol exposure (CIE) induced the progress in disturbed sleep patterns by increasing wake time and decreasing non-rapid eye movement (NREM) sleep time, and parallelly increased the levels of the c-Fos expression in the anterior part of the PVT (aPVT). Conversely, chemogenetic inhibition of the projecting neurons in the aPVT ameliorates the effects of repeated ethanol exposure on sleep patterns. Interestingly, when the neuronal activities in the anterior part of PVT (aPVT) were increased after CIE, simultaneously, the major subunit of m-type potassium ion channel in the thalamus, KCNQ2, which is selectively expressed in the aPVT along the anterior-posterior axis of PVT, were decreased. Importantly, the rescue of the KCNQ2 expression in the projecting neurons of aPVT via viral gene transfer ameliorated the CIE-induced sleep disturbance.

Conclusions: These findings reveal a central role of the PVT in sleep disturbance during withdrawal from repeated intermittent ethanol exposure, highlighting this brain area as a potential therapeutic target for sleep disturbance in alcohol use disorder and other psychiatric disorders.

Keywords: Sleep disturbances, paraventricular nucleus of the thalamus, alcohol withdrawal, potassium channel

Disclosure: Nothing to disclose.

P863. Age Dependent Changes in 24 Hour Gene Expression Rhythms Across Cells of the Human Dorsolateral Prefrontal Cortex

Madeline Scott*, Hui Yang, Tereza Clarence, Prashant N.M., Xinyi Wang, Milos Pjanic, Sanan Venkatesh, Aram Hong, Clara Casey, Zhiping Shao, Marcela Alvia, Stathis Argyriou, PsychAD Consortium, Pavan, K. Auluck, Stefano Marenco, Vahram Haroutunian, Georgios Voloudakis, Jaroslav Bendl, Donghoon Lee, John Fullard, Gabriel Hoffman, Kiran Girdhar, Panos Roussos, Colleen McClung

University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, United States

Background: The importance of circadian rhythms in health and disease has become increasingly apparent, especially in psychiatry where sleep and circadian behavior disruptions are a common feature across many disorders. Circadian rhythms, or 24 hour (h) physiological and behavioral cycles, are important in regulating sleep/wake cycles, alertness, cognition, and many other essential processes. Various physiological and activity rhythms deteriorate with aging in both humans and mice, including cognitive function and prefrontal cortex-specific dendritic length and complexity. At the molecular level, gene expression rhythms have also been shown to drastically change with age, suggesting that this may be an important feature of late adulthood. However, to our knowledge, no study has investigated cell type specific gene expression rhythms in the prefrontal cortex, nor their changes in aging.

Methods: These analyses were done using single nucleus RNA-sequencing (snRNA-seq) data from the new PsychAD Consortium’s dorsolateral prefrontal cortex (DLPFC) atlas. For this study, we focused on 192 subjects with no diagnosed psychiatric illness or neurodegenerative disease, and for whom we could identify their time of death (TOD). We then used a joint rhythmicity analysis to determine and compare 24 h gene expression rhythms in Adulthood (21 – 60 years old (yo); n = 116) and Late adulthood ( > 61 yo; n = 76) age groups within 26 cell type subclasses.

Results: 45 significant rhythmic genes (FDR < 0.05) were observed in 9 subclasses of the adulthood group, with 84% (38/45) found in upper layer intratelencephalic excitatory neurons (EN). As expected, many of these genes are associated with the circadian molecular clock. Conversely, Versican (VCAN) was the only significant rhythmic gene (FDR < 0.05) observed in late adulthood. For a deeper analysis, we loosened our rhythmicity significance threshold (p-value < 0.01) and found that, for most subclasses, more rhythmic genes were observed in adulthood (Median Rhythmic Genes per Subclass +/- Standard Deviation (st.dev); All = 129 + /-80.3; ENs = 186 + /-67.2; Inhibitory Neurons (IN) = 155 + /-44.1; Glia (Microglia, Astrocytes, Oligodendrocytes, Oligodendrocyte Progenitor Cells) = 128 + /-80.4; Other (Perivascular Macrophage, Adaptive Immune Cell, Smooth Muscle Cell, Vascular Leptomeningeal Cell, Endothelial Cell) = 24 + /-17.6) than in late adulthood (All = 112 + /-57.9; EN = 111 + /-43.0; IN = 134 + /-26.7; Glia = 152 + /-62.1; Other = 22 + /-17.2). Some notable exceptions were EN_L6_CT, EN_L6B, IN_LAMP5_LHX6, microglia, and oligodendrocytes, which had more rhythmic genes in late adulthood. However, comparative rhythmicity analyses identified very few genes that were rhythmic in both groups within each subclass (2 + /-2.1 median rhythmic genes in both groups per subclass), indicating that the identity of transcripts with rhythmic gene expression differs between adulthood and late adulthood. Additionally, of the few genes that were identified as having rhythms in both age groups, 46% (28/61) had significantly different peak times (FDR < 0.05) of expression, further emphasizing the differences between these age groups. Furthermore, while the expected circadian clock signature was observed in adulthood neuronal subclasses, this signal was lost in late adulthood. While genes that gained rhythmicity in late adulthood were not consistent across subclasses, rhythmic genes in microglia and oligodendrocytes were both enriched for the unfolded protein response.

Conclusions: Overall, our findings suggest that an intact, rhythmic molecular clock, which is known to be important for regulating sleep-wake patterns, cognitive function, and cellular metabolism, is a fundamental component of the DLPFC transcriptomic landscape during adulthood, particularly in neuronal subclasses. However, in late adulthood, rhythmicity of these core circadian clock genes is lost, but different genes gain a 24 h rhythmic expression pattern, consistent with a growing literature that has observed circadian reprogramming within various tissue types and across different species (humans, mice, Drosophila). Microglia and oligodendrocytes in particular had more rhythmic genes in late adulthood, which were enriched for genes associated with the unfolded protein response. Rhythmicity of this pathway suggests increased demand for proteins that regulate quality control of protein synthesis during aging, perhaps in response to higher activity and/or cellular stress within these cells. Investigating how these patterns influence neuropsychiatric and neurodegenerative processes may offer novel strategies for mitigating both age-related and disease-associated cognitive decline and improving overall brain health.

Keywords: dorsolateral prefrontal cortex (DLPFC), Single nucleus RNA sequencing, circadian rhythms, Biology of aging

Disclosure: Nothing to disclose.

P864. Sleep Reactivity as a Vulnerability to PTSD: Identifying High-Risk Individuals for Early Intervention Within the Acute Aftermath of Trauma

Anthony Reffi*, David Moore, Greg Mahr, Lily Jankowiak, Tanja Jovanovic, Christopher Drake

Henry Ford Health System, Novi, Michigan, United States

Background: Research suggests treating acute insomnia shortly after trauma exposure could help prevent posttraumatic stress disorder (PTSD). However, it is unknown who is most at risk of acute insomnia after trauma, thus obstructing the ability to identify high-risk groups in need of early intervention. We tested the hypothesis that individuals with high sleep reactivity—a trait vulnerability to sleep disturbance after stress—are most vulnerable to acute insomnia after trauma that increase their risk of PTSD.

Methods: We recruited 88 patients hospitalized in Detroit, MI following traumatic injury (Mage = 39.53 ± SD 14.31 years, 67.0% male, 67.0% Black). Patients reported their trait sleep reactivity and pre-trauma insomnia within one week of trauma exposure (T1) using the Ford Insomnia Response to Stress Test (FIRST) and the Insomnia Severity Index (ISI), respectively. Patients completed the ISI again one-month post-trauma (T2; n = 61) and the PTSD Checklist for DSM-5 (PCL-5) two months post-trauma (T3; n = 59). We tested whether high trait sleep reactivity (FIRST > 20) moderated the prospective effect of insomnia one month after trauma on PTSD two months after trauma, while covarying for pre-trauma insomnia, baseline PTSD, sex, acute stress, and days elapsed since trauma exposure.

Results: As expected, sleep reactivity exacerbated the prospective effect of insomnia on PTSD, such that the impact of insomnia one month after trauma on future PTSD was strongest for those with high trait sleep reactivity (β = 0.62, SE = 0.56, p = 0.034). More precisely, among patients with insomnia one month after trauma (ISI > 15), those with high trait sleep reactivity went on to report significantly more severe PTSD symptoms two months after trauma (Mean PCL-5 = 32.67) than those with low sleep reactivity (Mean PCL-5 = 14.82; p = 0.016).

Conclusions: These novel findings indicate that reactive sleepers are most vulnerable to acute insomnia within the immediate aftermath of trauma that in turn conveys future risk of clinically significant PTSD symptoms. This might enable the early identification of potentially vulnerable individuals who might develop PTSD, toward whom sleep-focused preventive efforts can be targeted.

Keywords: insomnia, Acute Traumatic Stress, PTSD, sleep

Disclosure: Nothing to disclose.

P865. The Effect of Orexin 2 Receptor (OX2R) Selective Agonist E2086 on the Body Weight of Narcolepsy Model Mice

Reiko Koba, Ken Hatanaka*

Eisai, Co., Ltd., Tsukuba-Shi, Japan

Background: Narcolepsy type 1 (NT1) is a neurological disorder stemming from the loss of orexin neurons and is associated with symptoms such as excessive daytime sleepiness and cataplexy, as well as disrupted sleep. The incidence of overweight or obesity ranges from 25% to 74% in patients with NT1. However, the mechanisms involved in the association of narcolepsy with obesity and precocious puberty have not been fully elucidated yet (Palhano et al., 2018). Previous research has indicated that NT1 patients typically exhibit a higher body mass index compared with population controls (Schuld et al., 2000). Similarly, narcolepsy model mice lacking orexin neurons (orexin/ataxin-3 Tg mice) also develop obesity (Hara et al., 2001). Given the orexin deficiency in NT1, orexin receptor agonists might improve metabolic control in affected individuals. Prior studies, including enhanced OX2R signaling preventing diet-induced obesity and improving leptin sensitivity (Funato et al., 2009), as well as findings that danavorexton, an OX2R selective agonist, reduced weight gain in orexin neuron-deficient mice (Ishikawa et al., 2022), support this hypothesis.

E2086, a novel OX2R selective agonist, has been shown to exhibit wake-promoting and anti-cataplectic-like effects in orexin/ataxin-3 Tg mice. This study aimed to further investigate the impact of E2086 on body weight regulation in the nonclinical narcolepsy model.

Methods: In this study, E2086 (1 or 3 mg/kg) or vehicle was administered orally by gavage to both wild-type (WT) and orexin/ataxin-3 Tg mice at ZT12 daily for 21 days. ZT12 refers to the moment when the lights are turned off in a 12/12 light/dark cycle. Body weight was recorded prior to the first administration of E2086 or vehicle (Day 1) and subsequently on Days 8, 15, and 22 to assess weight changes. Food intake was monitored by measuring the difference in food weight before and after each interval.

Results: Preliminary data revealed that orexin/ataxin-3 Tg mice treated with 3 mg/kg of E2086 displayed a significant reduction in weight gain compared with vehicle-treated mice without alterations in food intake. By contrast, WT mice showed no change in weight gain and food intake between the E2086- and vehicle-treated animals.

Conclusions: These findings suggest that E2086 could offer a therapeutic benefit for managing the metabolic symptoms associated with NT1. Additional experiments are underway to replicate and confirm these initial results.

Sponsor: Eisai Co., Ltd.

Keywords: narcolepsy, orexin, Obesity

Disclosure: Eisai: Employee (Self).

P866. Trazodone in the Off-label Treatment of Insomnia: Abuse and Safety Risks

Margaret Moline*, Jocelyn Y. Cheng, Jack Henningfield, Mark A. Sembower, Steve Pype, Emerson M. Wickwire, Ric Procyshyn

Eisai Inc., Nutley, New Jersey, United States

Background: Trazodone, a serotonin reuptake inhibitor/serotonin-2 receptor antagonist, received initial US FDA approval for the treatment of major depressive disorder in 1981. It is a frequently prescribed medication for insomnia in the US despite not being approved for the treatment of insomnia disorder. As of 2018, zolpidem was dispensed more frequently than trazodone; however, off-label use of trazodone for insomnia has been increasing in the US relative to zolpidem such that it is now prescribed more commonly than zolpidem.

It is well known that GABA-A receptor agonists like benzodiazepines and zolpidem can be associated with tolerance, dependence, abuse, and falls. It is possible that the increased off-label use of trazodone may be due to the perception that its lack of scheduling makes it a safer alternative for the treatment of insomnia. However, an abuse potential study was not required at the time that trazodone was approved for the treatment of major depression, and since that time there has been a lack of systematic clinical or real-world studies to assess its potential for abuse or dependence. Additionally, trazodone has been associated with safety concerns, including cardiovascular adverse events (AEs) such as orthostatic hypotension and arrythmias, as well as priapism.

The objective of the current study was to assess trazodone abuse/dependence potential and other safety risks by collating real-world data from publicly available data sources and published literature in comparison with GABA-A receptor agonists.

Methods: Cases where the drug was suspected of contributing to the event (ie, suspect cases) involving trazodone, zolpidem, or those benzodiazepines most frequently prescribed for insomnia (temazepam, triazolam, estazolam) were extracted from openFDA from the FDA Adverse Events Reporting System (FAERS) from January 1, 2015, to March 31, 2024. Search terms were “zolpidem,” “trazodone,” and the benzodiazepines “temazepam,” “triazolam,” and “estazolam.” AEs were classified as non-serious or serious (death, life-threatening, causing initial or prolonged hospitalization, disabling or incapacitating, congenital abnormality or birth defect, or other important medical event). The abuse-related preferred search terms included drug abuse, drug dependence, drug withdrawal syndrome, and falls.

Results: FAERS included 13,348 trazodone, 19,466 zolpidem, and 6111 benzodiazepine reports. In these reports, rates of drug abuse and drug dependence were lower with trazodone than with zolpidem and the benzodiazepines (drug abuse: 6.2% versus 9.1%, 12.7%, respectively; drug dependence: 1.0% versus 5.2%, 3.8%, respectively). Serious cases (82.9%/75.3%/85.1%) were similar between trazodone, zolpidem, and the benzodiazepines, respectively. Deaths were reported less frequently with zolpidem (23.5%) than trazodone (33.7%) and the benzodiazepines (36.9%). Falls were reported for all the drugs; of the total number of reports, 2.5% were falls-related for trazodone, compared with 3.9% for zolpidem and 2.8% for the benzodiazepines.

Conclusions: Data from this important national source demonstrate that trazodone has abuse/dependence potential and important safety risks, including a fall risk similar to benzodiazepines. Given the limited data from well-controlled studies and its common off-label use, re-evaluation of trazodone prescribing rates for treating patients with insomnia is warranted.

SUPPORT: Eisai Inc.

Keywords: insomnia, trazadone, safety

Disclosure: Eisai Inc.: Employee (Self).

P867. Sex Differences in Mood Improvement After Slow-Wave Sleep Disruption in Major Depressive Disorder

Jennifer Goldschmied*, Philip Gehrman

University of Pennsylvania, Philadelphia, Pennsylvania, United States

Background: Current research suggests that deficits in synaptic plasticity may underlie mood dysregulation in MDD. Sleep slow-wave activity (SWA) has been implicated in the facilitation of homeostatic synaptic downscaling and has been shown to be decreased in males with major depression (MDD), but not females. This may suggest that males with MDD have a sleep-related vulnerability in the regulation of neuroplasticity. The aim of the present study was to examine if there are sex differences in the impact of experimental modulation of SWA on mood in MDD.

Methods: 32 individuals (12 M) with MDD were recruited to sleep for two nights (Baseline, Slow-wave Disruption (SWD)) in the sleep laboratory separated by one week. On the SWD night, slow-wave sleep was experimentally decreased using auditory stimulation. In the morning following each overnight visit, mood was assessed via self-report measures including a VAS scale of both positive and negative mood, and the Positive and Negative Affect Schedule (PANAS). Repeated measures MANOVA was used to examine mood changes for constructs of both positive and negative mood with condition (Baseline, SWD) as the within-subject factor and sex (M,F) as the between-subjects factor.

Results: Results from repeated measures MANOVA revealed a significant condition by sex interaction for the construct of negative mood, F(2,29) = 3.97, p = 0.03, but not for positive mood. Post-hoc analyses demonstrated that following SWD, males exhibited a significant decrease on both the VAS, t(11) = 1.97, p = 0.04, and PANAS scales of negative mood, t(11) = 2.28, p = 0.02. In contrast, females demonstrated a significant increase in PANAS-assessed negative mood following SWD, t(19) = -1.89, p = 0.04. Positive mood was unaffected by SWD.

Conclusions: Our results indicate that SWD is associated with sex-specific improvements in negative mood, with benefits seen only in males with MDDs. These findings suggest that mood dysregulation in males with MDD may be associated with sleep-related impairments in the regulation of homeostatic synaptic plasticity. These data also add to the existing literature demonstrating significant sex differences in the presentation and potentially pathophysiology of MDD and suggest that treatment interventions consider sex.

Keywords: slow-wave activity, Mood, sex differences

Disclosure: Nothing to disclose.

P868. Social Interaction Induces Aberrant Activity in the Nucleus Accumbens of the CNTNAP2 Knockout Mouse

Aishwarya Patwardhan, Katrina Choe*

McMaster University, Hamilton, Canada

Background: Autism Spectrum Disorder (ASD) is a neurodevelopmental condition with a strong genetic etiology; however, a clear link between the gene mutations and behavioral symptoms remains unestablished. Cortical Dysplasia-Focal Epilepsy (CDFE) is a monogenic form of ASD caused by a loss-of-function mutation in the CNTNAP2 gene. Cntnap2 knockout (KO) mice exhibit ASD-like behavioral phenotypes, providing an opportunity to understand the mechanistic link between the genetic mutation and associated behavioral symptoms. Interestingly, Cntnap2 KO mice also exhibit deficiency in brain oxytocin (Penagarikano et al., 2015), a neuropeptide hormone that influences various aspects of social interactions through modulatory effects on downstream target brain regions. Our recent work demonstrated that agonist activation of oxytocin receptors in the nucleus accumbens (NAc) increased social interaction in Cntnap2 KO mice (Choe et al., 2022). Given the oxytocin deficiency in Cntnap2 KO mice, and the established role of NAc in social reward, it is feasible that insufficient oxytocin release during social interaction causes aberrant NAc neuronal activity in these mice, thus leading to their low sociability phenotype. To test this hypothesis, we assessed social interaction-triggered changes in NAc activity in Cntnap2 KO mice and compared to their wildtype (WT) littermates. Furthermore, we examined the effects of stimulating endogenous oxytocin release using chemogenetic tools.

Methods: 1. Homecage reciprocal social interaction and assessment of NAc cellular activity: To compare social interaction-induced changes in NAc cellular activity levels, we performed a social interaction assay where an adult (8-12 weeks old) Cntnap2 KO mouse or a WT littermate freely interacted with a novel, sex-matched juvenile (3-5 weeks old) mouse in a new homecage for 10 minutes (n = 8-10 mice/group). A control group of KO mice and WT littermates spent 10 minutes alone in a new homecage. 90 minutes post behavioral assay, mice were paraformaldehyde-perfused and their brains were processed for c-Fos immunostaining and quantification. 2. DREADD stimulation of endogenous oxytocin release: AAV-mOXT-hM3Dq-mCherry (excitatory DREADD) or AAV-mOXT-Venus (control) were injected into the hypothalamic paraventricular nucleus of KO mice and WT littermates (n = 5-6/group). To stimulate endogenous oxytocin release, clozapine-n-oxide (CNO, 1mg/kg) was administered first via intraperitoneal injection and then through drinking water for 12 hours prior to reciprocal social interaction test. Both sexes were included in the study, and statistical comparisons were performed using one- or two-way ANOVA with FDR post-hoc correction as appropriate. q values of below 0.05 were considered as statistically significant.

Results: In the homecage reciprocal social interaction assay, Cntnap2 KO mice spent a significantly shorter amount of time interacting with the juvenile conspecific mouse compared to their WT littermates. Further analysis revealed a significantly shorter duration, but not the number, of interaction bouts in the KO mice compared to WT. As expected, social interaction significantly increased the average number of c-Fos+ cells in the NAc of WT mice, but strikingly decreased it in KO mice. DREADD stimulation of endogenous oxytocin release in Cntnap2 KO mice lengthened social interaction bouts without affecting the number of interactions. We are currently examining whether this behavioral rescue is associated with a corresponding increase in NAc cellular activity levels by comparing c-Fos+ cell counts between groups.

Conclusions: The lower sociability phenotype of Cntnap2 KO mice is driven by shorter bouts of social interactions compared to WT littermates, suggesting a lack of motivation to continue further social engagement. Social interaction with a novel juvenile mouse suppresses NAc activity in Cntnap2 KO mice, in stark contrast to the stimulatory effect observed in WT littermates. With the NAc being an important region of the mesolimbic reward pathway, these data suggest a lack of NAc-encoded social reward in these mice, with a potentially crucial role of oxytocin.

Keywords: oxytocin, autism spectrum disorder, Nucleus Accumbens, Social Behavior, CNTNAP2

Disclosure: Nothing to disclose.

P869. Mothers’ Confidence in Parenting Predicts Higher Brain-To-Brain Synchrony During Mother-Infant Interaction

Lindsay Taraban*, Theodore Huppert, Hendrik Santosa, Caitlin Aloisio, Alison Hipwell, Judith Morgan

University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, United States

Background: The first year of life is a sensitive period for infants to learn foundational emotional self-regulation skills, setting the stage for their regulatory abilities throughout development. During dynamic, back-and-forth interactions, caregivers can reinforce infants’ expression of positive emotions (e.g., mutual smiling, imitative vocalizations) and down-regulate infant negative emotions (e.g., soothing, gentle touch in response to fussiness). This process of parent-infant co-regulation has been found to involve synchronous responding of parent and infant behavior, as well as physiological signals, including heart rate and hormone release. More recently, mother-infant co-regulation has been found to relate in real time to coordinated neural activation in regions that drive emotional expression and regulation, such as the prefrontal cortex (PFC). Although it remains a new area of research, recent studies have begun to examine maternal factors that may disrupt mother-infant co-regulation, observable at the neural level. Overall, higher levels of maternal stress (economic, interpersonal, parenting) have been found to relate to lower levels of neural co-regulation with mothers and their toddlers and preschoolers. Higher maternal stress may distract mothers from their child’s affective cues and disrupt their ability to perspective-take with their child, leading to a decrease in neural co-regulation. Although not explored in prior research, positive maternal factors may conversely relate to higher levels of neural co-regulation between mothers and their babies. To the extent that mothers feel confident in themselves as parents and the relationship they are building with their baby, they may be more present in their interactions with the infant, and more attuned to their infant’s signals, leading to greater neural co-regulation. The present study takes a strengths-based approach to investigate whether mothers’ confidence in their parenting abilities predicted higher levels of synchronous responding in mothers’ and infants’ PFC.

Methods: Participants were 29 mothers and their 3-month-old infants [mean age (SD) = 3.6(.59) months; range = 2.9-4.8 months; 45% female]. Neural co-regulation was measured in the home during 3 minutes of face-to-face play using Near-Infrared Spectroscopy (NIRS), a safe, painless, and non-restrictive method for assessing real-time neural function. NIRS data were collected using a two NIRSport2 systems (NIRx Medical Technologies, LLC), with each system containing 8 illuminators as sources and 8 detectors. Custom-fitting head caps for mother and infant were each centered on the Fp1 position of the 10-20 system and extended bilaterally over the ventrolateral and dorsolateral PFC and medial PFC. Maternal Confidence was measured using a sum score on the 11-item Competency Subscale from the Barkin Index of Maternal Functioning (BIMF). Mothers used a 7-point scale (0 = strongly disagree to 6 = strongly agree) to respond to items related to their perception of their competence in the maternal role (e.g., “I am a good mother;” “My baby and I understand each other;” “I trust my instincts when it comes to taking care of my baby.”). Data analysis was performed in MatLab using the AnalyzIR toolbox, which utilizes statistical methods based on robust regression and is thus less sensitive to motion-related artifacts.

Results: Overall, mothers experienced moderately high to high confidence in their parenting [mean(SD) = 58.48 (5.86); range = 43-66 (out of possible 66)]. Consistent with expectations, greater maternal confidence was associated with stronger coupling of maternal medial and lateral PFC with infant right lateral PFC (5 associations; t = 4.09, p = 0.0004, q-corrected = 0.005; t = 3.21, p = 0.003, q = 0.029; t = 4.90, p < .0001, q = 0.001; t = 4.04, p = 0.0004, q = 0.005; t = 3.47, p = 0.002, q = 0.012). Counter to our expectations, higher maternal confidence was also associated with weaker synchrony mother medial PFC and infant left lateral PFC and mother medial PFC (t = -3.06; p = 0.005, q = 0.04).

Conclusions: Results overall suggest that greater maternal confidence in parenting is associated with greater neural co-regulation in areas of the brain relevant for emotional expression and regulation. Specifically, as a function of higher levels of maternal confidence, we found greater activation between maternal brain regions implicated in emotional expression (medial PFC) and regulation (lateral PFC) and infant right lateral PFC, implicated in emotional regulation. We also found one unexpected negative association between maternal PFC and infant left lateral PFC. At least with older children, the left PFC has been found to relate more to conflict interactions while the right PFC relates more to cooperation. These exciting preliminary results raise several additional questions, such as whether babies with easy temperament may elicit both greater confidence in caregiving and higher levels of co-regulation. Additional analyses will examine associations with child variables, such as temperament, age, and biological sex. Longer term, we will use longitudinal analysis (3, 6, and 9 months) to test whether confidence results from high levels of co-regulation or vice versa. Clinically, it is possible that increasing mothers’ confidence in their parenting may have a positive cascading effect on mother-infant co-regulation and support the development of babies’ early emotional self-regulation skills.

Keywords: hyperscanning, NIRS, emotion regulation

Disclosure: Nothing to disclose.

P870. Preliminary Evidence of the Unique Contribution of Three Early Relational Health Constructs at 4 Months to Child Neurodevelopmental and Socioemotional Outcomes at 2 to 3 Years of Age

Andreane Lavallee*, Jennifer Warmingham, Imaal Ahmed, Ginger Atwood, Maggie Kyle, Maha Hussain, Elena Arduin, Ruiyang Xu, Nicole Shearman, Morgan Firestein, Marissa Lanoff, Dani Dumitriu

Columbia University Medical Center, New York, New York, United States

Background: The positive and nurturing early-life relationship that forms between infants and their parents/caregivers – hereafter early relational health (ERH) – is heralded as a resilience builder and buffer against childhood adversity. Decades of correlational data support the association between ERH and child neurodevelopment and socio-emotional functioning across the life-course. ERH is generally used as a blanket term for all constructs describing the tie between parents/caregivers and infants, like bonding, attachment, sensitivity, and emotional connection, indiscriminate of their theoretical origins. Likewise, there is also important heterogeneity in patterns of child developmental attainment across neurodevelopmental (e.g., cognitive, motor, language), and socio-emotional domains (e.g., early emerging psychopathology symptoms, prosocial behavior). Critically, the characterization and unique contribution of ERH constructs to the promotion and/or buffering of various child developmental outcome domains remains undefined.

Methods: Leveraging data from the COMBO Initiative at Columbia University Irving Medical Center on a subset of 98 mother-child dyads, we used 2-wave Structural Equation Modeling (SEM) to explore the unique contribution of ERH constructs assessed at 4 months postpartum on prospective child neurodevelopmental and socio-emotional outcome domains at 2 to 3 years of age. ERH constructs considered here include mother-reported bonding on the Postpartum Bonding Questionnaire (PBQ), and video-coded maternal sensitivity with an adaptation of Ainsworth Maternal Sensitivity Scale and dyadic emotional connection with the Welch Emotional Connection Screen. Child neurodevelopmental outcome domains included cognitive, expressive and receptive language, and gross and fine motor development assessed at 2 years through the observer-based Developmental Assessment of Young Children-2nd edition (DAYC-2), and child internalizing and externalizing symptoms reported by mothers on the Child Behavior Checklist (CBCL) at 2 to 3 years. Analyses tested the association among three ERH constructs with Pearson correlations, and their simultaneous contribution to each child developmental domain. Analyses were controlled for maternal race, ethnicity and insurance status, and child sex at birth, and age at neurodevelopmental assessment.

Results: Results report on a subset of 98 mother (mean age at delivery 31□5) and infant (n = 47, 47% female) dyads from the COMBO Initiative who have completed survey and video-based assessments from 4 months to 3 years of age. Of the mothers included here, 72% (n = 71) identify as white and 19% (n = 19) as Hispanic, and 20% (n = 20) receive Medicaid. At 4 months postpartum, there is no association between mother-reported bonding and video-coded maternal sensitivity (r = 0.02, p = 0.81), bonding and video-coded emotional connection (r = -0.004, p = 0.97), as well as maternal sensitivity and emotional connection (r = 0.06, p = 0.57). Preliminary SEM results suggests that video-coded maternal sensitivity at 4 months uniquely significantly predicts more advanced cognitive development on the DAYC-2 at 2 years (Beta=0.247, p = 0.005) and less externalizing symptoms reported by mothers on the CBCL at 2 to 3 years (Beta = -0.221, p = 0.026). Additionally, video-coded dyadic emotional connection at 4 months uniquely significantly predicts expressive language on the DAYC-2 at 2 years (Beta=0.249, p = 0.008). Mother-reported bonding does not predict any neurodevelopmental or socio-emotional outcome domain.

Conclusions: Leveraging multi-informant data on ERH and child development acquired in a subset of 98 mother-child dyads from the COMBO Initiative across the first 3 years of life, we investigated the unique contribution of three ERH constructs as promoters and/or buffers of child neurodevelopment and socioemotional functioning. Data support that bonding, maternal sensitivity and dyadic emotional connection describe unique aspects of the mother-infant relationship. Preliminary results further support the hypothesis that unique ERH constructs predict specific fundamental aspects of child neurodevelopment and socioemotional functioning, therefore contributing to understanding the variations across developmental outcomes throughout childhood.

Keywords: Social Functioning, Social Behavior, neurodevelopment, Social Interactions

Disclosure: Nothing to disclose.

P871. MDMA Increases Preference for Social Rewards in Female Mice

Tyler Dexter*, Richard Sharp, Samuel Barnes, Nicole Satchell, Melissa Flesher, Marco Venniro, Debra Bangasser, Mathieu Wimmer, Susan Powell, Jared Young

University of California - San Diego, San Diego, California, United States

Background: Social decision-making requires the integration of various streams of information. Individuals perceive social information, assess its’ valuation and relationship to prior experience, and have to synthesize this information to guide future behavior. Imaging studies have implicated circuitry between the prefrontal cortex (PFC) and subcortical regions in social decision-making, and PFC damage has been shown to impair indices of social preference across species. Importantly, deficits in social preference and decision-making are debilitating symptoms seen in various mental disorders (e.g., depression, schizophrenia), requiring the development of targeted therapeutic interventions. MDMA is a serotonin and dopamine releaser/reuptake inhibitor that increases sociability in humans and rodents. As a result, researchers have begun investigating how MDMA increases such behavior and whether MDMA has the potential to treat social impairments in psychiatric disorders. Here, we tested whether MDMA (0, 3, and 7mg/kg) could increase social preference and motivation in mice using a novel operant paradigm. We hypothesized that MDMA would: 1) increase the animal’s preference for access to social interaction, and 2) increase willingness of mice to exert effort to gain a social reward.

Methods: Female mice (n = 8) were food restricted and socially isolated for the duration of testing. Mice were first trained to lever press for access to a social partner and a food reward, then tested for social preference on a decision-making paradigm (directly choosing between food and social rewards). Additionally, we assessed social motivation using a progressive ratio breakpoint task (PRBT), where mice had to progressively increase their response rates for access to a social partner until they chose to stop responding (i.e., breakpoint). Following task training, mice received unilateral injections of either the CAMKII-dependent calcium sensor GCaMP8f or dopamine sensor GRABDA3 into the medial PFC (mPFC) to measure mPFC activity and dopamine signals, respectively (n = 4 per group). An optical fiber was implanted above the injection site and in vivo fiber photometry was used to collect fluorescent signals and align activity dynamics with key behavioral events during task performance.

Results: During the social/food preference test, mice initially displayed a strong preference for food over social rewards (F(1, 7) = 19.51, p < 0.01), and higher breakpoints for food rewards on the PRBT (t(7) = 4.952, p < 0.01). Interestingly, MDMA facilitated social decision-making and shifted preference towards social rewards at the high dose (F(2, 14) = 5.243, p < 0.05; p = 0.01). Alternatively, the same dose of MDMA decreased breakpoint for both social (F(2, 14) = 5.678, p < 0.05, p < 0.05) and food rewards (F(1.046, 7.324) = 11.66; p = 0.01), but did not affect ancillary measures of social interest such as anticipatory behavior (F(2, 14) = 0.534, p = 0.598), or time spent interacting with the social partner (F(2, 14) = 0.085, p = 0.919). Moreover, there was no effect of MDMA when animals had to make a single lever press on an FR-1 schedule (F(1.074, 7.521) = 1.844, p = 0.2156), suggesting intact social drive when the effort requirements were kept low. During social self-administration, we observed increased PFC activity time-locked to the time of cue (signaling choice presentation) and during lever responding. Alternatively, PFC dopamine signals increased following lever responding in the time leading up to reward delivery.

Conclusions: Our data demonstrate that MDMA can increase measures of social drive in healthy mice under low work requirements, suggesting a dissociation between reward value and the degree of effort expenditure for their attainment. Contrary to one of our hypotheses, systemic MDMA reduced motivation for social interaction as measured by breakpoint. It is possible that MDMA’s effects on progressive ratio responding may be linked to accelerated reward satiation and a devaluation of rewarding stimuli. Thus, initial task engagement under low work requirements may be aborted as the cost (i.e., effort) outweighed the diminished reward value as animals became satiated in early trials. We did observe, however, that MDMA increased the preference for social rewards under conditions where the animals initially showed an inherently stronger preference for a competing reward. These findings provide support for the use of MDMA as a pro-social compound and highlight the importance of evaluating multiple components of social decision-making for future endeavors testing therapeutic compounds. Our preliminary photometry experiments revealed recruitment of PFC circuitry during the presentation of cues related to social rewards and at the time of responding. Previous studies have demonstrated increased PFC activity during social approach behaviors, while our data support a role for the PFC in processing information related to upcoming social interactions. Future studies will investigate how PFC excitatory activity and dopamine signaling may be modulated by MDMA administration and whether these processes are disrupted in models of psychiatric disorders.

Keywords: Social Behavior, Prefrontal Cortex, MDMA

Disclosure: Nothing to disclose.

P872. Influence of Familiarity on MDMA-Induced Social Connection

Hanna Molla*, Anya Bershad, Emma Hahn, Harriet de Wit

University of Chicago, Chicago, Illinois, United States

Background: MDMA is known to enhance feelings of closeness, which may facilitate a therapeutic alliance in MDMA-assisted therapy. With this therapy, the patient and therapist first engage in preparatory sessions to build a relationship before the MDMA experience, which is thought to enhance the efficacy of the drug. However, it is not known whether familiarity with a conversation partner influences the effects of the drug. This study examined whether the acute effects of MDMA on a social interaction were influenced by the familiarity of the partner.

Methods: Using a randomized, double-blind, crossover design, we compared the effects of MDMA (100 mg or placebo) during interactions with a familiar or unfamiliar conversation partner. Familiarity was established before the drug administration, by engaging participants in a procedure known to induce feelings of closeness. During four sessions, participants (N = 33) had a 15-min conversation with a familiar and a non-familiar partner, after MDMA and placebo. They rated closeness and connection towards the conversation partner after the 15 min conversation.

Results: As expected, MDMA increased ratings of closeness (p = 0.005) and connection (p = 0.02) to conversation partners relative to placebo. Additionally, participants rated liking (p < 0.001) and feeling more connected (p = 0.002) to conversation partners who they were more familiar with compared to unfamiliar partners. There were few interactions between these two variables.

Conclusions: These findings suggest that familiarity with the partner does not significantly enhance the effect of MDMA on social closeness and connection, as defined in this context.

Keywords: MDMA, Social Behavior, Psychedelics

Disclosure: Nothing to disclose.

P873. Single Nucleus RNA-Sequencing Reveals Transcriptional Synchrony Across Different Relationships

Liza Brusman, Allison Fultz, Julie Sadino, Michael Kelberman, Robin Dowell, Mary Allen, Zoe Donaldson*

University of Colorado Boulder, Boulder, Colorado, United States

Background: Relationships are shaped by reciprocal interaction and feedback between individuals. As relationships mature, pairs share common goals, improve their ability to work together, and experience coordinated emotions. However, the neural underpinnings responsible for this unique, pair-specific experience remain largely unexplored.

Methods: We used single nucleus RNA-sequencing to examine the transcriptional landscape of the nucleus accumbens (NAc) in 38 socially monogamous prairie voles in long-term peer or mate relationships (n = 5 - 11 pairs/grp). Voles were sequenced individually, resulting in a robust data set for examining gene-behavior relationships. We further used machine learning and correlational analyses to test the hypothesis that bonding leads to transcriptional convergence.

Results: We developed a high-resolution single cell map of the prairie vole NAc, identifying 15 transcriptionally-defined cell type clusters based on known markers of cell types in this region. We then identified proportional differences in subpopulations of medium spiny neurons; mate-paired females had greater proportion of Drd1-Pdyn and Drd2-Penk medium spiny neurons compared with peer-paired females (GLM with estimated marginal means, FDR corrected: MSN-Drd1Pdyn: T = 2.93, p = 0.024; MSN-Drd2Penk: T = 2.80, p = 0.033). We then used Hotspot, a computational algorithm that groups genes into modules based on similar expression patterns across cells. Using a minimum threshold of 30 genes per module, we identified 23 modules in total that range in size from 31-415 genes. Eight of these modules correlated with stranger-directed behavior while 2 correlated with partner-directed behavior in a relationship-type-specific fashion (Spearman’s correlation, p < 0.05).

To test whether bonded partners are more transcriptionally similar to each other than to other voles, we used an SVM-based decoder to predict the originating vole for cells in our data set. Upon training, the decoder was able to accurately predict vole identity for ~70% of cells in each cell type cluster. This is significantly more than would be expected by chance (2.7%), indicating the validity of using this decoder to infer similarity in cellular transcriptional profiles. When we subsequently trained the decoder in a leave-one-out paradigm where it was forced to choose an originating vole from the other 37 voles in our dataset, we found that cells were predicted to come from the partner at much greater than chance (2.8%) levels, indicating that partners are more transcriptionally similar than would be expected by chance. We then asked whether there was evidence of transcriptional synchrony at the level of gene expression, focusing on the gene modules identified by Hotspot. We found that, regardless of pairing type, two of our Hotspot modules, Module-6 (expressed in oligodendrocyte progenitor cells, or OPCs) and Module-11 (expressed in interneurons), exhibited pairwise expression similarity using multiple metrics. First, Module-6 and Module-11 average gene expression is correlated between partners (Module-6: Rho = 0.61, p = 0.0057; Module-11: Rho = 0.59, p = 0.0075). Second, using the expression of all of the genes within each module to calculate the Euclidean distance between partners, we found that for these two modules, partners were closer together in Euclidean space than non-partners regardless of relationship type (Wilcoxon rank sum tests: Module-6: W = 9781, p = 0.00017; Module-11: W = 8803, p = 0.0083). Finally, when we ranked the animals by their module gene expression and calculated the rank-based distance between partners, we found that partners were closer together in the ranking than would be expected by chance (One-sample t-tests vs. null (expected) value: Module-6: T = -3.46, p = 0.0028; Module-11: T = -2.73, p = 0.014). We further found that Module-6 gene expression was positively correlated with intra-pair interaction levels (Rho = 0.50, p = 0.0013).

Conclusions: While there has been substantial research linking transcriptional state to behavior, here we provide a new layer of understanding by delineating transcriptional changes that may enable relationship-appropriate behaviors as well as the emergence of transcription-based synchrony within pairs. The latter represents a putative biological substrate underlying the neural and behavioral organization of relationships that may help partners understand and infer each other’s mental state and facilitate behavioral coordination.

Keywords: prairie voles, Nucleus Accumbens, social bond, synchrony

Disclosure: Nothing to disclose.

P874. Social Threat Perception is Mediated by Dorsal Striatal Acetylcholine Release

Munir Kutlu*, Oyku Dinckol, Noah Wenger, Taylor Good, Jake Caselli, Zehra Bozdag

Temple University School of Medicine, Philadelphia, Pennsylvania, United States

Background: In the central nervous system, acetylcholine (ACh) is crucial for attention, learning, memory, and mediating physiological and behavioral responses to stress. Previous research suggests ACh may signal negative valence in both social and non-social contexts. The dorsolateral striatum (DLS), which receives substantial ACh input from local and external sources, is key to motivated action. However, the specific role of ACh release in the DLS in regulating social behaviors remains largely unexplored.

Methods: In this study, we investigated the role of ACh release in the dorsolateral striatum (DLS), a brain region heavily involved in motivated action, in threat perception in social settings using fiber photometry recordings combined with a genetically encoded ACh sensor (ACh.Sn.Fr) as well as optogenetics and novel behavioral paradigms.

Results: First, we recorded DLS ACh release during a 3-chamber test, where C57BL/6J mice were allowed to explore either a same-sex or opposite-sex novel conspecific and a familiar conspecific (a cagemate). Our results demonstrated that mice preferred interacting with the opposite-sex conspecific over the cagemate, but did not show a similar preference for the same-sex conspecific. Accompanying this effect, the DLS ACh release was increased at the time of interaction with the novel same-sex conspecific but not during the interaction with the opposite-sex conspecific suggesting that ACh may signal situations with higher threat perception. Next, we examined DLS ACh release patterns during social competition. In a warm spot/cold arena task, where mice competed with a novel conspecific over warmth in a cold arena, we showed that ACh levels increased with distance to the warm spot and decreased with distance to their competitors in the cold arena. When the competitive element was removed (i.e. warm spot or competing conspecific), striatal ACh was no longer affected by distance to the warm spot’s previous location in the arena or distance to conspecific. We also tested our animals in a tube dominance test, where mice competed against a conspecific that was blocking the opposing exit of the tube. In this test, we found that DLS ACh increased when competing against another mouse. This effect was no longer seen if the competing animal was removed. Importantly, we showed that optogenetic stimulation of DLS ACh release resulted in a reversal of preference for the novel opposite-sex conspecific and stronger competition over the warm spot without inducing place preference in the absence of competition.

Conclusions: Together, these results suggest that DLS ACh is critically involved in processing social threats in a time-dependent and causal manner. This has important implications for psychopathologies where social information processing is maladaptive, such as autism spectrum disorders.

Keywords: Dorsal striatum, acetylcholine, In vivo fiber photometry, Social Behavior

Disclosure: Nothing to disclose.

P875. Reward Circuitry Underpinnings for Social Buffering of Stress in Mice

Urszula Skupio*, Giuliana C Bilbao, Djemila Compaore, Nekhama Riznyk, Alexander Harris

Columbia University, New York State Psychiatric Institute, New York, New York, United States

Background: Social buffering, the phenomenon by which the presence of others can reduce the consequences of stressful experiences, has been observed in both humans and rodents, yet the precise mechanism by which social interactions can offset the negative impact of stress remains unknown.

Methods: In this study, we have used a spectrum of tools, including a new behavioral approach, fiber photometry recordings in both dopamine and GABA neurons of the ventral tegmental area (VTA), and optogenetic manipulations, to study neural circuitry of social buffering of stress in both male and female mice.

Results: Utilizing a new behavioral paradigm, we demonstrate that brief social interaction with a familiar, unstressed individual after acute restraint, ameliorates the stress-induced deficits in reward seeking in both male and female mice (F(1,54) = 6.26; p = 0.015; n = 29 males, 28 females, 13-16 mice per group). Previous studies indicate that within the VTA, not only dopamine but also GABA neural activity has been implicated in the stress-induced reductions in reward seeking. Therefore, we aimed to investigate whether social buffering alleviates stress-induced blunting of reward seeking by modulating specific VTA neuronal populations. Using fiber photometry recordings, we observed distinct calcium dynamics of VTA dopamine and GABA neurons during restraint, social interaction, and cued-reward task. Furthermore, we show that brief optogenetic inhibition of these neurons after restraint mimics the effects of social interaction on stress-induced deficits in reward anticipation (F(1.7, 20) = 11.15; p < 0.001; n = 9 males, 5 females), suggesting an important role of VTA GABA neurons in mediating the effects of social buffering.

Conclusions: Together, our data suggests that social communication of affective state engages both GABA and dopamine neurons within the VTA to counteract negative effects of stress. These findings provide new avenues to develop targeted therapeutic interventions for stress-related disorders.

Keywords: Dopamine, GABA, Ventral Tegmental Area (VTA), Social Behavior, Acute Stress

Disclosure: Nothing to disclose.

P876. PHF21B and Social Neuropeptides

Ma-Li Wong*, Qi Ma, Hongyu Ruan, Junchi He, Yike Huang, Julio Licinio

State University of New York, Syracuse, New York, United States

Background: Previous research identified the PHF21B gene in human chromosome 22q13.31. We reported that rats resilient to chronic stress showed reduced hippocampal Phf21b expression. PHF21B, a histone reader in the PHD finger protein family, is expressed in brain regions including the frontal cortex and hippocampus. To better understand its role in the CNS, we developed a Phf21b knockdown mouse model.

Methods: The study, approved by multiple institutions, examined both male and female mice. Key approaches included: 1) Creating Phf21b mutant mice using CRISPR/Cas9 technology; 2) Conducting cognitive and social behavioral tests; 3) Analyzing long-term potentiation in hippocampal pyramidal neurons; 4) Performing hippocampal transcriptome profiling; 5) Using immunoblotting, immunohistochemistry, and Golgi staining; 6) Employing histone peptide array analysis

Results: Phf21b deficient mice showed 60% less PHF21B expression than wild-type mice. Both male and female Phf21b deficient mice exhibited social memory deficits. Phf21b deficiency led to thinner cortices, increased astrocyte numbers, and reduced neurogenesis. Deficient mice showed decreased synaptic protein expression and impaired glutamatergic neurotransmission. RNA sequencing revealed PHF21B’s role in regulating neurotransmission genes. Histone peptide studies indicated PHF21B’s involvement in transcription regulation through specific histone modifications and CREB interaction. Oxytocin and oxytocin receptor expression were reduced in several brain regions in PFH21B deficient mice.

Conclusions: The above results verified that PHF21B is a critical upstream regulator of synaptic plasticity-related genes, and regulates social peptide expressions.

Keywords: Social Behavior, glutamatergic transmission, neuropeptides

Disclosure: Nothing to disclose.

P877. Distinct Effects of Oxytocin Neurons on Social Approach in the Anterior and Posterior Paraventricular Nucleus of California Mice

Brian Trainor*, Hanna Butler-Struben, Alexandra Serna Godoy, Vanessa Minie, Audrey Chrisman, Sophia Wright

UC Davis, Davis, California, United States

Background: Oxytocin is a versatile neuropeptide that can promote or inhibit social approach behaviors. Several lines of evidence indicate that distinct neural circuits mediate the behavioral effects of oxytocin in appetitive or aversive social contexts. For example, oxytocin receptors in the nucleus accumbens promote social approach while oxytocin receptors in the bed nucleus of the stria terminalis promote social avoidance and vigilance. Less clear is whether distinct populations of oxytocin neurons activate these receptor fields in different social contexts. Our prior work shows that in female California mice, anterior but not posterior oxytocin neurons are more reactive in novel social contexts following social defeat.

Methods: Male and female California mice were randomly assigned to be treated with antisense (targeting Oxt) or missense morpholinos (n = 8-10 per group) in the posterior paraventricular nucleus. One week later mice were tested in a social interaction test and the amount of social approach (within 1 body length of a caged target mouse) and social vigilance (more than 1 body length away from target mouse while oriented toward the target) were recorded. A separate set of females were randomly assigned to be treated with antisense or missense morpholinos (n = 14) in the anterior paraventricular nucleus. We also performed single nucleus RNA sequencing on PVN samples of female California mice (n = 8).

Results: In the posterior PVN, antisense treatment reduced social approach (F1,31 = 7.5, p = 0.01) in both females (Cohen’s d = 1.0) and males (d = 0.9). There were no differences in social vigilance, during an acclimation test with an empty cage, or during an open field test. In the anterior PVN there was no effect of antisense treatment on behavior despite a significant decrease in oxytocin immunoreactivity (t18 = 2.32, p = 0.03, d = 1.11). Knock down experiments in anterior PVN of stressed females is in progress. Sequencing data identified two populations of oxytocin neurons. One population showed high expression of Foxp1, which has been identified as a marker for magnocellular oxytocin neurons in mice. The second population expressed Adarb2, which has been put forth as a marker for parvocellular oxytocin neurons.

Conclusions: We found that posterior oxytocin neurons but not anterior oxytocin neurons in the paraventricular nucleus of unstressed California mice. Previous work in male Mus reported that the posterior paraventricular nucleus is populated primarily by parvocellular oxytocin neurons. Intriguingly, these oxytocin neurons played a key role in promoting social learning, similar to our findings. Our sequencing data suggest that it will be possible to use molecular markers to distinguish magnocellular and parvocellular oxytocin neurons in California mice, which is currently under investigation. Overall, our results suggest that distinct populations of oxytocin neurons modulate social behavior in different social contexts.

Keywords: Social defeat stress, sex difference, social salience

Disclosure: Nothing to disclose.

P878. Role of Hypocretin/Orexin Neurons in Social Behavior and Isolation

Matthew Dawson, Dylan J. Terstege, Naila Jamani, Van Anh Lee, Kartikeya Murari, Jonathan R. Epp, Derya Sargin*

Hotchkiss Brain Institute, Alberta Children’s Hospital Research Institute, University of Calgary, Calgary, Canada

Background: Intraspecific social interactions are vital for survival and maintenance of society among all mammalian species. However, our understanding of the brain regions and circuits remain incomplete. Our recent studies (Dawson et al., 2022) demonstrated that the activity of hypothalamic hypocretin/orexin-producing neurons, which govern arousal and motivation, differentially encode social discrimination. We further showed that these neurons are essential for social interaction in a sex-dependent manner. Building on these findings, here we examine the involvement of the hypocretin/orexin system and its downstream pathways in an early-life mouse model of chronic social isolation, which exhibits impaired social behavior during adulthood.

Methods: To recapitulate social deficits, we used a mouse model of chronic social isolation by single-housing female and male mice right after weaning (postnatal day 21) until adulthood. Group-housed littermates served as controls. We infused adeno-associated virus (AAV) encoding GCaMP6s in hcrt-cre mice to express the calcium indicator selective in hypocretin/orexin neurons. We next performed in vivo calcium imaging of hypocretin/orexin neurons in control and isolated mice to investigate the differences in hypocretin activity during social interaction with same-sex conspecifics. We next infused AAV encoding the orexin sensor (GRAB-Ox0.9) into the brain regions that show synchronized activity with the lateral hypothalamus during social behavior. We performed multi-fiber photometry to characterize hypocretin/orexin release during social interaction in control and isolated female and male mice.

Results: We show that hypocretin neuron activity increases in female and male control and isolated mice upon initial interaction with a same-sex stranger conspecific. However, the amplitude of interaction-induced hyporcretin activity is significantly reduced in female and male isolated mice, compared with controls. Our data further reveals the downstream pathways modulated by hypocretin/orexin activity that are disrupted in isolated mice with social deficits.

Conclusions: Together, these findings situate the hypocretin/orexin system within a broader crucial for modulating social behavior. Moreover, we explored the impact of these findings using an animal model of chronic social isolation, which exhibits long-term social impairments. Our data has significant implications for understanding the circuit mechanisms underlying neuropsychiatric diseases associated with social deficits.

Keywords: orexin/hypocretin, hypothalamus, In vivo fiber photometry, social isolation

Disclosure: Nothing to disclose.

P879. What Gets in the Way of Social Support? Functional Connectivity and Social Support Interference in Laboratory Settings

Andrea Coppola*, Erin Maresh, David Sbarra, Jessica Andrews-Hanna

University of California San Diego, San Diego, California, United States

Background: Social support plays a mechanistic role in connecting high-quality relationships to good health. However, the question of what might interfere with establishing and maintaining supportive relationships remains relatively understudied. Self-focused attention is increasingly considered a transdiagnostic risk factor for psychopathology and resting-state functional connectivity (RSFC) can be used as a marker of this construct. Thus, a timely and important question is the degree to which self-focus-related brain activity can extend beyond the individual to degrade vital social support processes in romantic relationships.

Methods: To address this question, one hundred seventy participants (85 dyads) in romantic relationships participated in a two-session study. First, dyads completed a supportive discussion task. Discussions were behaviorally coded for instrumental and emotional support. The second session involved a seven-minute resting state functional MRI scan. RSFC in regions associated with self- and other-focus was computed to yield self- and other-focus scores for each participant. Actor-Partner Interdependence Modeling (APIM) was used to examine dyadic effects of self- and other-focus on provided and perceived support.

Results: APIM models revealed actor (b = -1.06, SE = 0.43, p = 0.01) and partner (b = -1.11, SE = 0.46, p = 0.02) effects of self-focus on instrumental support provided to a partner, suggesting that, to the extent that RSFC reflects greater self-focus, this construct is negatively associated with the provision of instrumental support. Additionally, APIM models revealed an actor effect (b = -1.21, SE = 0.50, p = 0.01) of RSFC in regions associated with other-focus on perceived support such that greater other-focus corresponded with lower ratings of perceived support.

Conclusions: These findings provide evidence for a relationship between RSFC and support provision and perception, and will be discussed in terms of their potential relationship and clinical implications.

Keywords: Social support, Cognition, fMRI, functional connectivity, Brain-Behavior Relationships

Disclosure: Nothing to disclose.

P880. Sex Differences in the Effects of Manipulating the Paraventricular Thalamus to Orbitofrontal Cortex Pathway During Reversal Learning

Kathleen Tuite*, Milena Girotti, David Morilak

The Univ of Texas Health Science Center at San Antonio, San Antonio, Texas, United States

Background: Stress-related psychiatric disorders, such as major depressive disorder and anxiety disorders, have cognitive flexibility deficits that persist even after other symptoms of these disorders go into remission. Reversal learning, a form of cognitive flexibility necessary to adapt to a changing environment, is disrupted in stress-related psychiatric disorders. The orbitofrontal cortex (OFC) mediates reversal learning, and hyperactivity in the OFC is associated with depression and obsessive-compulsive disorder in humans. Using a reward-based discrimination digging task to assess reversal learning in rodents, preliminary data using Fos immunohistochemistry showed a significant decrease in Fos in the lateral OFC following reversal learning. Further, we have previously reported that chronic stress impairs reversal learning and potentiates responses to excitatory input in the OFC, and that inducing long-term depression in the mediodorsal thalamus to OFC pathway reverses these deficits, indicating that increased activity in projections to the OFC is detrimental to reversal learning. However, the circuit-level mechanisms underlying stress-induced reversal learning deficits are not well established. The paraventricular thalamic nucleus (PVT) is highly stress responsive and is known to provide excitatory input to the OFC. Here we test the hypothesis that the PVT to OFC pathway is responsible for the effects of chronic stress on reversal learning.

Methods: All experiments were done in male and female Sprague-Dawley rats. To examine the activation of the PVT to OFC pathway during chronic stress, we utilized ΔFosB immunohistochemistry in the PVT combined with retrograde tracing from the OFC. In a separate cohort of animals, we used an adeno-associated virus to deliver excitatory (Gq) DREADD or GFP control under the CaMKII promoter into the PVT of non-stressed animals, and implanted guide cannulae into the lateral OFC for pathway specific activation. We used the same procedure to deliver an inhibitory (Gi) DREADD or GFP control into the PVT of animals that would undergo chronic unpredictable stress (CUS). Animals received microinjections of the DREADD agonist clozapine-N-oxide (300 µM, i.c. 0.75 μL) directly preceding the reversal learning task into the lateral OFC.

Results: ΔFosB immunohistochemistry revealed that female rats have overall lower expression of ΔFosB when compared to males in the PVT. However, both sexes show increases in ΔFosB after stress as well as similar activation of the PVT to OFC projection. Activating the PVT-OFC pathway with the Gq DREADD significantly impaired reversal learning in non-stressed male rats, while inhibiting the PVT-OFC pathway with the Gi DREADD in chronically stressed animals reversed the stress-induced deficits in reversal learning only in males, with minimal or no effect of these manipulations in females.

Conclusions: These results suggest that the PVT to OFC pathway has no effect on reversal learning in female rats while manipulating this pathway in males produced robust effects.

Keywords: Chronic unpredictable mild stress, lateral orbitofrontal cortex, Cognitive / behavioral flexibility

Disclosure: Nothing to disclose.

P881. Stress-Induced Regulation of Prefrontal Cortical Astrocytes and Behavior in Male and Female Rats

Jayme McReynolds*, Justin Bollinger, Nicolas Wiles, Corey Foltz, A’yanna Nation, Andrew Gaulden, Sierra Rollins, Eric Wohleb

University of Cincinnati, Cincinnati, Ohio, United States

Background: Chronic stress and affective disorders, such as anxiety, depression, and PTSD, share many overlapping behavioral symptomologies including impaired cognition and poor emotional regulation. Additionally, chronic stress can impact the development or severity of these affective disorders. Therefore, understanding how stress impacts neural circuits associated with cognition and affective behavior is a necessary area of study. The medial prefrontal cortex (mPFC) is highly impacted by chronic stress and is critical for executive function, cognition, and emotion regulation. Much of the work to date has focused on stress-induced alterations in mPFC neurons though recent findings have identified glial cells, such as microglia and astrocytes, as having a critical role in the neurobiology of affective disorders. Additionally, there is a growing appreciation for the role of astrocytes in cognitive processing in healthy brains. However, stress effects on astrocyte function in the mPFC remains poorly understood. Therefore, the current study aims to examine how acute or chronic stress regulates behavior and mPFC astrocytes.

Methods: Adult male and female Long-Evans rats underwent 2-hr restraint stress for 0, 1, or 14 days (n = 8/group). Body weight and struggling behavior was measured across the chronic stress paradigm. 24 hrs after the last stressor, adrenal glands and thymus were dissected and weighed, and the mPFC was dissected. Fluorescence-activated cell sorting (FACS) was used to isolate astrocytes from mPFC and gene expression was quantified using qPCR. Current studies are examining the impact of chronic stress on cognitive flexibility, using an operant strategy shift task, anxiety-like behavior, and social behavior. Additionally, ongoing studies are exploring the impact of acute and chronic restraint stress on astrocyte activation using immediate early gene expression.

Results: Both male and female rats that underwent chronic restraint stress (CRS) demonstrated decreased body weight and adrenal hypertrophy compared to the no stress control group. Interestingly, male rats showed overall greater amounts of struggling behavior at both Day 1 and Day 14 of CRS than female rats. This suggests that male rats exhibit delayed stress habituation or higher levels of active coping than female rats. We found that CRS decreased astrocytic expression of Nr3c1, the gene encoding the glucocorticoid receptor (GR), in both male and female rats but acute restraint stress (ARS) only decreased Nr3c1 expression in female rats. Additionally, while ARS and CRS appear to decrease expression of glutamate transporters in male rats, only CRS decreases expression of these transporters in female rats. We are continuing to explore changes in other glucocorticoid-responsive genes and genes related to astrocyte-neuron interaction.

Conclusions: Acute and chronic stress induce sex-specific changes in mPFC astrocyte gene expression in rats. These changes have important implications for the role that astrocyte dysregulation plays in the impact of chronic stress on behavior and neuronal function. Additionally, while higher levels of neuronal GR can be associated with stress-induced behavioral changes, these data suggest that decreased astrocytic GR may also contribute to stress-related behavioral and neuronal changes. Ultimately, understanding the sex-specific impact of stress on glia and their subsequent role in stress-related behavioral changes has important implications for the development of treatments for affective disorders.

Keywords: Acute and Chronic Stress, astrocytes, Medial Prefrontal Cortex (mPFC), glucocorticoid receptor

Disclosure: Nothing to disclose.

P882. Latrepirdine (Dimebon) Modulates Stress-Mediated Behaviors in Rodents

Frank Yocca*, Michael DeVivo, Dinesh Dhull, Subhendu Seth

Bioxcel Therapeutics, BioXcel Corporation, Clinton, Connecticut, United States

Background: Latrepirdine (Dimebon) was under clinical investigation for the treatment of global function and cognition in Alzheimer’s patients but failed to meet the primary clinical endpoints in phase 3 studies. Although primary endpoints were not met, secondary endpoints suggested that latrepirdine might be useful for the treatment of neuropsychiatric symptoms, such as anxiety and agitation, in patients diagnosed with Alzheimer’s Disease. Further development of latrepirdine was impaired by lack of understanding of the mechanism of action for this drug. We find that latrepirdine may be a 5-HT7 receptor antagonist at efficacious doses and that it reduces locus coeruleus mediated behaviors. These findings enable continued clinical development of latrepirdine for chronic treatment of neuropsychiatric symptoms in patients diagnosed with Alzheimer’s Disease.

Methods: Binding potency of latrepirdine was assessed in vitro using membranes from CHO-K1 cells stably transfected with cDNA encoding human 5-HT7 receptors. 50 mcg aliquots of membranes were incubated with 5.5 nM [³H]LSD for 120 minutes at 25°C. Non-specific bind was determined using 10 microM 5-HT. In vitro adenylyl cyclase activity was measured for either 5-HT6 or 5-HT7 receptors. For 5-HT6 activity, HeLa cells were stably transfected with cDNA encoding human 5-HT6 receptors. For 5-HT7 activity, CHO cells stably transfected with cDNA encoding human 5-HT7 receptors. Test compound and/or vehicle were incubated with cells at 37°C for 20 minutes. The reaction was evaluated for cAMP levels by time-resolved fluorescence resonance energy transfer TR-FRET. For 5-HT6, test compound inhibition of 0.3 µM serotonin-induced response by 50 percent or more ( ≥ 50%) indicates receptor antagonist activity. For 5-HT7, Test compound inhibition of 5-carboxamidotryptamine (5-CT)-induced response by 50 percent or more ( ≥ 50%) indicates receptor antagonist activity.

In vivo behavior was measured using yohimbine-induced stress and the elevated plus maze. Male Wistar rats (Janvier; Le Genest St Isle – France) were used for this study. Rats (220 g) were group-housed (3-4 rats per cage) and maintained in a room with controlled temperature (21-22°C) and a reversed light-dark cycle (12h/12h; lights on: 17:30 – 05:30; lights off: 05:30 – 17:30) with food and water available ad libitum. EPM apparatus was a PVC maze covered with Plexiglas and subdivided into four equal exploratory arms (40 x 10 cm), which were all interconnected by a small platform (10 x 10 cm). The apparatus was placed 65 cm above the floor. Two arms were open, and two others were closed with wall (high: 10 cm). The elevated plus maze test was done in lighting conditions of 100 – 150 lux light. All compounds/drugs were prepared in normal saline (0.9% NaCl). After compound/drugs administration, rats were placed on the platform opposite a closed arm. The number of entries and the time spent in each arm were recorded during a 5 min period. The animal was considered as entered in an arm when it placed its four paws in the arm. The apparatus was cleaned using 70% alcohol after each animal tested.

Results: Potency of latrepirdine as assessed by binding was 14 nM and by inhibiting adenylyl cyclase activity was 11 nM at 5-HT7 receptors. Latrepirdine was not potent at 5-HT6 receptors (IC50 > 500 nM). The potency at 5-HT7 receptors is in good agreement with reported peak plasma concentration of latrepirdine at the 20 mg dose used in the Doody study (1), as extrapolated from the 8.16 mg dose used in the Chew pharmacokinetic study (2). Yohimbine is an antagonist of the alpha2-adrenoceptors (alpha2-adrenergic receptors). Because these autoreceptors inhibit the firing of locus coeruleus neurons yohimbine will activate the neurons by blocking the autoreceptors. Increased activity of LC neurons is associated with an increased response to stress. Using yohimbine, stress-induced anxiety was measured with the elevated plus maze. Stress causes the animals to remain in the closed arms and not to explore the open, elevated arms. Latrepirdine dose-dependently increased entries and time spent in the open arms, indicating that latrepirdine affects locus coeruleus mediated stress response. These findings suggest that latrepirdine may work through 5-HT7 receptors to reduce hyperarousal mediated by the locus coeruleus. We propose that this mechanism may account for the latrepirdine-mediated changes in the neuropsychiatric inventory (NPI) measured in clinical studies.

REFERENCES:

1. Doody RS et al. Lancet. 2008 Jul 19;372(9634):207-15.

2. Chew ML et al. Pharm Res. 2016 Aug;33(8):1873-80.

Conclusions: These results indicate that latrepirdine is a potent 5-HT7 receptor antagonist and is highly selective with respect to 5-HT6 receptors. In vitro potency of latrepirdine for 5-HT7 receptors matches the reported plasma concentrations achieved in clinical studies that demonstrated efficacy at secondary endpoints related to neuropsychiatric symptoms. Antagonism of 5-HT7 receptors may result in a reduction in hyperarousal. This is a plausible mechanism to account for previously published clinical findings suggesting that latrepirdine is effective in reducing neuropsychiatric symptoms in patients diagnosed with Alzheimer’s Disease.

Keywords: Acute agitation, Cognition, Neuropsychiatric symptoms (NPS)

Disclosure: Nothing to disclose.

P883. Social Isolation Decreases Hippocampal Mitochondria Respiration in Female, but Not Male, Peromyscus Californicus

Amy Wegener*, Hannah Fulenwider, Hannah Stadtler, Erica Glasper, Gretchen Neigh

Virginia Commonwealth University, School of Medicine, Richmond, Virginia, United States

Background: The neural mechanisms that facilitate the maladaptive consequences of loss, loneliness and grief are not fully understood. Clinical assessments in individuals experiencing prolonged grief suggest a role of synaptic function in grief progressing as a chronic stress disorder. Loss and social isolation are two experiences that are not synonymous in humans and should be approached as such in preclinical models. As mitochondria are implicated in the stress response, we proposed that synaptic mitochondria respiration is altered within the stress sensitive hippocampus following separation from a partner. As such, we investigated hippocampal synaptic mitochondria changes in the monogamous and biparental California mouse (Peromyscus californicus) following separation from an opposite sex partner (partner loss) or a same-sex cage mate (social isolation).

Methods: Two housing paradigms were established in this study. In housing paradigm 1, female mice paired with a male underwent tubal ligation to control for pregnancy and were single housed for two weeks for post-operative recovery. Males were similarly single housed to counterbalance housing experience and handling. Following post-operative recovery, females were paired with a male California mouse for 10 days, or remained single-housed. At the end of the 10-day pairing, mice were separated from their partner for the next 10 days, or remained paired for the same length of time. Three housing groups of males and females were generated: single-housed (males n = 11; females n = 8), 10-day partner separation (males n = 8; females n = 7), or paired (males n = 8; females n = 8). In housing paradigm 2, male and female California mice were placed in same-sex groups of 2-3 at weaning. Mice remained with their weaning cage mates until reaching adulthood (approximately 90 days) and were then separated for social isolation or remained with their cage mates. Three housing groups were investigated for both sexes: same-sex paired housing (males n = 13; females n = 10), 10-day social isolation (males n = 12; females n = 10), or a 30-day social isolation (males n = 13; females n = 12). All mice were tested for anxiety-like behavior in a 20-minute open field test (OFT) at the end of the housing paradigm. The day following the OFT, functional synaptic mitochondria was isolated from the hippocampus and assessed for changes in mitochondrial respiration using Agilent’s Cell Mito Stress Test. Terminal organ weights were recorded to interrogate peripheral markers of stress. Effects of housing were determined using a one-way ANOVA.

Results: In the partner separation paradigm, housing group influenced OFT outcomes such that single-housed males spent the least amount of time in the center (F(2,24) = 4.32, p = 0.025). No effect of housing group was observed on female behavior (p > 0.05). Despite observing behavioral indicators of anxiety-like behavior in males, neither male nor female mice in housing paradigm 1 displayed a significant influence of housing group on synaptic mitochondrial respiration in the hippocampus (p > 0.05). Paired female mice had an increase in terminal normalized spleen weight, compared to other female groups (F(2,20) = 3.95, p = 0.036). There was no peripheral indicator of stress observed in the terminal adrenal weight for either sex (p > 0.05). Interestingly, in the social isolation paradigm, females did exhibit a change in hippocampal synaptic mitochondria respiration, with 10- and 30-day separated females displaying an overall decrease across mitochondrial dynamics: basal respiration (F(2,22) = 6.63, p = 0.0056), maximal respiration (F(2,22) = 6.92, p = 0.0047), spare capacity (F(2,22) = 6.62, p = 0.0056), and proton leak (F(2,22) = 7.15, p = 0.0040). Peripheral metrics were altered by social isolation in males with an increase in normalized terminal adrenal weight (F(2,35) = 3.81, p = 0.032) and spleen weight (F(2,35) = 3.56, p = 0.0390) in isolated males. Female peripheral metrics were not affected by social isolation. Further, there was no impact of social isolation on anxiety-like behavior.

Conclusions: These data indicate that California mice respond to changes in their housing status through peripheral and central mechanisms, but males and females are not mutually affected. Following social isolation, males were more sensitive to the influence of isolation on peripheral metrics, while females showed a robust decrease in hippocampal synaptic respiration, but neither is associated with detectable changes in anxiety-like behavior. Surprisingly, there were no changes induced by partner loss. We anticipated partner loss to be a more salient social stressor than social isolation alone. However, the opposite sex pairs were together for 10 days prior to separation and the same-sex pairs were co-housed since weaning. As California mice function as a biparental family unit, subsequent studies may need to separate males and females following successful rearing of a litter. Pregnancy- and lactation-induced hormonal changes in females would need to be considered as a contributing factor to female outcomes. As we observed limited changes in anxiety-like behavior, future work should determine if the social isolation periods leveraged in this study were perceived as a maladaptive chronic social stressor or a prompt to adapt to new circumstances in females. These data provide insight on approaching loss and loneliness as a synaptic mitochondrial stressor in California mice.

Keywords: Chronic social stress, Mitochondrial Respiration, partner loss

Disclosure: Nothing to disclose.

P884. Adolescent Stress Causes Long-Lasting Behavioral Deficits and Impairs Parvalbumin Interneurons: Prevention by N-Acetylcysteine

Ícaro Freitas, Francisco Guimaraes, Felipe Gomes*

University of Sao Paulo, Ribeirao Preto, Brazil

Background: Adolescent stress is a critical risk factor for the development of psychiatric disorders, including schizophrenia. Previous studies from our group indicate that applying a stress protocol to rats during a period corresponding to adolescence results in schizophrenia-like impairments in adulthood, such as an increase in the activity of the dopamine system in the ventral tegmental area (VTA). These changes were associated with dysfunction in the excitatory-inhibitory (E/I) balance in the ventral hippocampus (vHip), linked to a functional loss of GABAergic interneurons expressing the calcium-binding protein parvalbumin (PV) and their associated perineuronal nets (PNN), with the presence of a redox imbalance in the vHip. Therefore, we propose that antioxidant drugs could potentially “protect” PV interneurons/PNN from adolescent stress-induced damage and, consequently, attenuate behavioral and electrophysiological impairments caused by stress. We evaluated whether the treatment with the antioxidant N-acetylcysteine during the exposure to an adolescent stress protocol would prevent the long-lasting stress-induced behavioral, electrophysiological, and deficits in PV and PNN expression in the vHip in adult rats.

Methods: Male Sprague-Dawley rats were subjected to a stress protocol consisting of a combination of stressors: daily sessions of footshock (25 shocks of 1mA - 20 ± 60s) between postnatal days (PD) 31 to 40 and restraint stress for one hour on PD 31, 32, and 40. Throughout the stress protocol, the experimental groups – stress vs. naïve (non-stressed; n = 10-12/group) – were treated with vehicle or N-acetylcysteine (900mg/L) in drinking water. In adulthood (from PD 60), rats underwent behavioral tests to assess anxiety-like responses (Elevated Plus-Maze - EPM and Light-Dark Box - LDB), sociability (Social Interaction test - SI), and cognitive function (Novel Object Recognition - NOR test). On PD 66, some animals underwent in vivo electrophysiological recordings of VTA dopamine neurons. For other animals, their brains were perfused and collected for immunofluorescence staining of parvalbumin (PV), PNNs using Wisteria floribunda agglutinin (WFA), and 8-Oxo-DG (a marker of DNA damage caused by oxidative stress) in the vHip (CA1-Subiculum).

Results: Adolescent stress resulted in anxiety-like behavior in adulthood in the LDB (F(1,45) = 9.33, p = 0.007), decreased sociability (F(1,45) = 3.74, p = 0.002), and impairments in novel object recognition memory in the NOR test (F(1,45) = 9.32, p = 0.004). The treatment with N-acetylcysteine attenuated these behavioral changes. Additionally, adolescent stress caused the animals to present an increased number of spontaneously active VTA dopamine neurons (F(1,20) = 16.74, p = 0.0006). N-acetylcysteine prevented this change. In the vHip, a reduction in the number of PV+ (F(1,20) = 18.82, p = 0.0003) and PV + /PNN+ cells (F(1,20) = 13.82, p = 0.001), along with an increase in 8-Oxo-DG labeling (F(1,20) = 3.28, p = 0.046), were observed in adult rats subjected to adolescent stress. N-acetylcysteine also prevented these impairments.

Conclusions: Our findings indicate that the treatment with N-acetylcysteine during the stress protocol prevented the long-lasting changes related to schizophrenia in adult rats caused by adolescent stress.

Keywords: adolescent stress, Parvalbumin neurons, N-acetylcysteine, Perineuronal nets

Disclosure: Nothing to disclose.

P885. Maternal Prenatal Stress Association With Placenta Glucocorticoid-Sensitive Polyepigenetic Score

Sameera Abuaish*, Daniel Alschuler, Seonjoo Lee, Benjamin Tycko, Frances A. Champagne, Catherine Monk

Princess Nourah bint Abdulrahman Univeristy, Riyadh, Saudi Arabia

Background: In-utero exposure to stress programs health outcomes of individuals, which may be mediated by epigenetic mechanisms. The placenta is the gateway to the developing fetus, and it regulates fetal exposure to different maternal factors. Maternal stress through glucocorticoid signaling might play a role in affecting the placental epigenome and transcriptome, rendering the fetus more susceptible to the effects of stress. A recently developed glucocorticoid-sensitive polyepigenetic score, derived from 24 CpG sites exhibiting differential methylation in response to glucocorticoid treatment, has been inversely associated with maternal anxiety and depression in cord blood. However, the placental expression of this score in relation to maternal stress remains unexplored.

Methods: A subsample of healthy pregnant women (N = 98) of a larger sample who were recruited at gestation week (GW) T1: 12-22 or at T2: GW 23-28 at Columbia Medical Centre in New York answered the Perceived Stress Scale (PSS) at three different visits during pregnancy T1, T2, and T3: GW 34-36. At birth, the placenta was collected, DNA was extracted and bisulfite converted, and DNA methylation was assessed using the Illumina DNA methylation array. The polyepigenetic scores are methylation levels at 24 CpG sites multiplied by the coefficient of elastic-net regression of dexamethasone changes in DNA methylation reported by Provencal et al., 2020. Linear regression analysis evaluated the association between PSS scores during pregnancy and the polyepigenetic score while controlling for covariates.

Results: Found that maternal PSS scores at T2 and T3 were associated with lower glucocorticoid-sensitive polyepigenetic scores (T2: B = -0.003 (95%CI: -0.005 - -0.001), P = 0.003; T3: B = -0.002 (95%CI: -0.004 - 0.00), P = 0.054). Conversely, no such association was observed in early pregnancy (T1: B = -0.002 (95%CI: -0.005 - 0.001), P = 0.2). Our findings show that higher reported perceived stress during pregnancy was associated with lower polyepigenetic scores.

Conclusions: These results provide evidence for increased demethylation at glucocorticoid-responsive DNA sites in the placentas of women experiencing elevated prenatal stress. These findings align with earlier reports demonstrating differential DNA methylation in glucocorticoid-related genes in the placenta using candidate gene approaches and the polyepigenetic score in the cord blood in relation to maternal distress. Future investigations should focus on elucidating the mediating role of this epigenetic biomarker in the relationship between maternal prenatal stress and adverse birth and infant outcomes.

Keywords: Early life stress (ELS), Epigenetic biomarkers, Placenta, Maternal stress

Disclosure: Nothing to disclose.

P886. Early Life Stress Impacts the Risk for Peripartum Dysfunction From Genes to Behavior

Marissa Nicodemus, Laiklyn Luther, Briana Karem, Kathleen Morrison*

West Virginia University, Morgantown, West Virginia, United States

Background: Women are at increased risk for mood and anxiety disorders, although this risk is heterogeneous and manifests uniquely across the lifespan. Two known risk factors are stress during puberty and later becoming pregnant. We understand little about the mechanisms underlying the enduring risk from either of these factors individually or when they are combined. Pubertal stress is associated with a blunted glucocorticoid response in peripartum humans and mice. Our recently published work shows that pubertally stressed mice respond negatively to the increased allopregnanolone levels of pregnancy, suggesting that the hormonal changes of pregnancy may be detrimental in females with a history of early life adversity. Here, we utilized our mouse model to further understand the effects of pubertal adversity on peripartum behavior, physiology, and gene expression in the brain and other stress axis tissues throughout the perinatal period.

Methods: We performed several studies. In all studies, mice were exposed to chronic variable stress (CVS, n = 6-18) from postnatal day 21-34 using our established paradigm or were left undisturbed during puberty (Controls, n = 8-15). In study 1, adult females were bred with naïve males and were left undisturbed until parturition. Mice were tested for pup retrieval and for their behavioral and hormonal response to a maternal separation stressor. After the separation task, stress axis tissues were collected for gene expression. In study 2, adult females were bred with naïve males. They were repeatedly tested for their behavioral and hormonal response to novel pups at 1.5, 9.5, and 17.5 days post-conception (dpc; early, mid, and late pregnancy).

Finally, in study 3, we collected brains at baseline conditions at one of several time points to examine the molecular underpinnings of the behavioral outcomes: at PN21 (before stress), PN35 (24 hours following the final stressor), or in adulthood (non-pregnant or late pregnancy). Brains were processed either for epigenetic measures via nuclear protein extraction or for in situ hybridization for anatomically-informed gene expression in stress- and maternal-relevant brain regions.

Results: In study 1, pubertal stress led to deficits on pup retrieval task (more failed retrievals, p = 0.006; decreased time spent in nest, p = 0.04, t-test) and prolonged the corticosterone response to a maternal separation. Pubertal stress decreased Crh gene expression (p = 0.01) in the paraventricular nucleus of the hypothalamus, increased Crhr1 expression (p = 0.04) in the pituitary, and decreased Mc2r expression (p = 0.04) in the adrenal (all via t-test). In study 2, we found a significant effect of pubertal stress on the corticosterone response to novel pup retrieval that interacted with stage of pregnancy (3-way ANOVA). Pubertal stress had a main effect of blunting the corticosterone response at 1.5 dpc (F (1, 54) = 4.210, p = 0.045) and 9.5 dpc (F (1, 52) = 5.931, p = 0.018). In late pregnancy, there was an interaction between pregnancy and pubertal stress (F (1, 55) = 4.813, p = 0.035), such that the pubertal stress altered the corticosterone response in opposite directions depending upon pregnancy state. In study 3, we found impacts of pubertal stress and pregnancy on regulators of chromatin acetylation in the paraventricular nucleus of the hypothalamus. Histone acetylation regulator histone acetyltransferase (HAT) activity was significantly altered by pubertal stress throughout adolescence and in adulthood in females, being decreased after two weeks of stress (PN35 t-test, p = 0.0168) and dynamically changed in adulthood by pregnancy.

Conclusions: We found that early life stress that occurred during puberty led to disruption of the developmental trajectory of epigenetic regulation within the hypothalamus, altered the hormonal response to acute stressors during pregnancy, increased anxiety-like postpartum behavior, and altered transcription in hypothalamic brain regions that regulate maternal and stress responses. These results indicate significant alterations in how the stress axis reorganizes during the peripartum period in pubertally stressed females. They provide novel insight into the impact of pubertal stress on risk for negative physiological and behavioral outcomes in the peripartum window. Understanding the nuances of life history that influence risk for peripartum affective dysfunction is key to developing effective treatment strategies.

Keywords: Early life stress, Adolescence, Peripartum, Epigenetic, Hypothalamus

Disclosure: Nothing to disclose.

P887. Childhood Maltreatment is Associated With Stress Reactivity Within Specific Hypothalamic, Bed Nucleus and Brainstem Subregions: A 7-Tesla Study

Layla Banihashemi*, Brandon Sibbach, Helmet Karim, Sara Albert, Tamer Ibrahim

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Dysregulated stress reactivity is an important risk factor for myriad health and disease outcomes. Childhood maltreatment is consistently associated with dysregulated stress reactivity, as well as affective symptoms and disorders. Despite these links between childhood maltreatment, stress reactivity and affective disorders, underlying neural mechanisms are unclear. Examining how childhood maltreatment influences a “central visceral” stress-control circuit also implicated in affective processes may provide insights into mechanisms underlying maltreatment-related health problems. This central visceral circuit includes the apex of the HPA axis, the paraventricular nucleus of the hypothalamus (PVN). The ventral bed nucleus of the stria terminalis (vBNST) directly modulates PVN function and receives dense viscerosensory, noradrenergic signaling from the brainstem nucleus of the solitary tract (NST) that collateralizes to the PVN. The dorsal BNST (dBNST) innervates the vBNST, and each of these regions is preautonomic, innervating the brainstem/NST to control autonomic reactivity. Given the small size and ventral location of these regions, our goal was to use 7-Tesla (7T) imaging to identify stressor-evoked activity within these specific subregions and examine their relationships with childhood maltreatment.

Methods: Participants were young adults (n = 155; mean age=26.30, SD = 4.87) from a transdiagnostic, abuse-enriched sample. 7T images were acquired using Tic-Tac-Toe RF coil systems (Santini Sci Rep 2021). Probabilistic atlases were derived from manual segmentations using MPRAGE and 2D gradient echo sequences (Sibbach BSF 2023). Functional MRI data were acquired using 1.25mm isotropic EPI (TR = 2500ms). Stressor-evoked activity was elicited via a performance-titrated Multisource Interference Task (mild cognitive stress) and parameter estimates were extracted from the control > stress contrast with 20% probabilistic atlases inverse normalized. Hierarchical regression analyses were performed covarying for age, sex, and race, examining childhood maltreatment (MACE multiplicity score) and socioeconomic deprivation (childhood parental education level reverse coded). We also evaluated whether findings survived adulthood variables [traumatic events, socioeconomic status (education level), negative life events]. Curvilinear relationships were also examined using squared terms for maltreatment and deprivation. For significant findings, post-hoc analyses were conducted to examine the timing [prekindergarten (1-5), preadolescent (6-11), adolescent (12-18)] and subtypes (10 subscales) of maltreatment.

Results: Regression analyses revealed a significant relationship between MACE and PVN stressor-evoked activity (β = 0.240, p = 0.015; all reported βs are standardized) in the full model with adulthood variables. There was also a significant relationship between MACE and dBNST stressor-evoked activity (β = 0.298, p < 0.001), which remained significant in the full model (β = 0.322, p < 0.001). Further, there was a significant curvilinear relationship between MACE and NST stressor-evoked activity (β = 0.518, p = 0.014), which remained significant in the full model (β = 0.575, p = 0.008). No significant relationships were found with socioeconomic deprivation or vBNST. Post-hoc analyses indicated significant relationships between MACE and dBNST activity for all three time periods, prekindergarten (β = 0.215, p = 0.009), preadolescent (β = 0.220, p = 0.008), and adolescent (β = 0.206, p = 0.015). A significant curvilinear relationship was found between prekindergarten MACE and NST activity (β = 0.506, p = 0.016). Significant relationships were revealed between several MACE subscales and dBNST activity; the strongest predictors of dBNST stressor-evoked activity were peer physical bullying (β = 0.235, p = 0.005) and peer emotional abuse (β = 0.220, p = 0.008). Significant curvilinear relationships were revealed between MACE subtypes and NST stressor-evoked activity; significant subscales were witnessing violence to siblings (β = 0.552, p = 0.009) and physical neglect (β = 0.461, p = 0.035).

Conclusions: Our findings revealed that childhood maltreatment, not socioeconomic deprivation, was associated with greater PVN, dBNST, and NST stressor-evoked activity. Thus far, relationships between childhood adversity and vBNST stressor-evoked activity were not revealed. This may indicate that childhood maltreatment predominantly shapes preautonomic circuitry with perhaps more secondary effects on modulatory viscerosensory pathways. Maltreatment reported throughout childhood was associated with dBNST stressor-evoked activity, while relationships with NST activity were limited to earlier development. These findings may be related to critical period differences in brainstem vs. forebrain structures. Further, several subscales of maltreatment were related to dBNST stressor-evoked activity, with the most robust subtypes being peer-related, while witnessing sibling violence and physical neglect were the only subtype predictors of NST activity. These findings could suggest regional differences linked to the distance, predictability, controllability and chronicity of maltreatment-related stimuli, with dBNST perhaps being more sensitive to distant stimuli. These findings may provide novel neural insights into childhood maltreatment-related risks for affective disorders and guide novel treatment approaches.

Keywords: Acute Stress, Childhood Adversity, 7 Tesla fMRI, paraventricular nucleus of the hypothalamus, bed nucleus of the stria terminalis

Disclosure: Nothing to disclose.

P888. Testing the Dual-Hormone Hypothesis: The Role of Cortisol Reactivity in Testosterone-Driven Risk-Taking

Félix Duplessis-Marcotte*, Marie-France Marin

Université du Québec à Montréal, Montréal, Canada

Background: Males engage in riskier behaviors than females, an effect observed across sociocultural contexts (Byrnes et al., 1999; Charness and Gneezy, 2012) and species (Orsini and Setlow, 2017). Testosterone has been proposed as a potential mechanism underlying this difference. Relatedly, a review highlighted that endogenous testosterone levels predict economic risk-taking, although findings lacked consistency (Apicella et al., 2015). The dual-hormone hypothesis has been elaborated to explain the inconsistencies in the investigation of testosterone and behavior (Mehta and Josephs, 2010). It states that the link between testosterone and risk-taking is moderated by the stress hormone cortisol. Testosterone and cortisol mutually inhibit their respective secretary axes (Viau, 2002). As such, the hypothesis predicts that testosterone increases risk-taking, but only at low levels of cortisol. This prediction has been confirmed in several studies (Mehta and Prasad, 2015). Yet, most studies have focused on basal hormonal levels, and it has been proposed that the hormonal interaction may be altered in a stressful context eliciting dynamic fluctuations of cortisol (Knight et al., 2020). The objective of this study is to test the role of testosterone, both independently and in conjunction with cortisol (dual-hormone hypothesis), on risk-taking in a basal hormonal state and in a context of stress-induced cortisol fluctuations, with consideration of sex differences.

Methods: A simulation-based power analysis revealed that a minimum of 80 participants were required to reach 80% of statistical power. We recruited 84 healthy young adults (42 women) and extensively screened for psychiatric and medical conditions that could interfere with endocrine systems. Participants were randomly assigned to the Trier Social Stress Test, a validated psychosocial stressor (Kirschbaum et al., 1993), or a control non-stressful condition before completing the Iowa Gambling Task (Bechara et al., 1994). The goal of the task was to maximize monetary profits by drawing 100 cards from four decks: two with large immediate gains and larger long-term losses (risky decks) and two with small immediate gains and smaller long-term losses (safe decks). The dependent variable was a binary variable indicating the selection from the risky decks or the safe decks. At the onset of the experiment, a saliva sample was taken to quantify baseline testosterone. Seven saliva samples were taken throughout the experiment to quantify cortisol reactivity to the social stressor. The area under the curve with respect to the increase (AUCi) was calculated to capture cortisol reactivity in the social stress condition, whereas the area under the curve with respect to ground (AUCg) was calculated to capture basal cortisol output in the control condition (Pruessner et al., 2003). All hormones were quantified using enzyme-linked immunoassays. Continuous variables were centered by sex and z-scored to obtain standardized effect sizes for the logistic multilevel models.

Results: A t-test revealed that the social stressor increased the cortisol AUCi compared to the control condition in men (difference = 1.12 standard deviations (sd), t(35.77) = -4.27, p < .001), but not in women (difference = 0.48 sd, t(32.45) = -1.58, p = 0.124). To compare the effect of testosterone on risk-taking in basal and stressful conditions, logistic multilevel models were fitted by sex to predict risk-taking with an interaction between experimental condition and basal testosterone levels. In men, testosterone was associated with fewer risk-taking only in the control condition (OR = 0.456, p = 0.026). In women, testosterone was associated to greater risk-taking in the control condition (OR = 2.32, p = 0.002) and the social stress condition (OR = 1.87, p = 0.011). Subsequent logistic multilevel models were fitted by sex and condition to test the dual-hormone hypothesis in the control and the social stress condition. In women, testosterone did not interact with AUCg (in the control condition) or AUCi (in the social stress condition). In men assigned to the control condition, no interaction between testosterone and AUCg emerged. A significant interaction between testosterone and AUCi was revealed in men exposed to the social stress (B = 0.47, p < .001). At lower testosterone levels, men’s risk-taking was not affected by AUCi levels. At elevated testosterone levels, AUCi was positively correlated with risk-taking (OR = 4.05, p < .001).

Conclusions: Our findings highlight the importance of considering dynamic cortisol fluctuations in the investigation of the dual-hormone hypothesis. In a non-stressful context, testosterone predicted risk-taking in men and women, irrespective of basal cortisol levels. In the context of a social stressor, only men showed an interaction between testosterone and cortisol reactivity on risk-taking, replicating the “reverse profile” reported by others, where testosterone and cortisol induce a double hit effect. This is consistent with findings showing that both testosterone and cortisol are related to heightened neural responses to reward (Knight et al., 2020), suggesting that cortisol increases reward salience to a greater extent under stress. Given that the social stressor did not significantly elevate cortisol reactivity in women, future studies should examine whether the current findings observed in men could be replicated with women-oriented stressors, such as social rejection (Stroud et al., 2002).

Keywords: Risk-taking, Cortisol response to stress, Testosterone, Sex differences, Acute Stress

Disclosure: Nothing to disclose.

P889. Synergistic Effects of Adrenergic and Glucocorticoid Activity on Spatial Memory and Reward Processing

Kshitij Kumar*, Garima Chauhan, Riddhi Bhalerao, Partha Mohanty, Arjun Ramakrishnan

Indian Institute of Technology Kanpur, Kanpur, India

Background: Noradrenaline and corticosteroid hormone secretion are key components of the stress response. These neuromodulators are known to alter consolidation and retrieval of memory.However, the alteration in spatial memory due to these neuromodulators can affect decision making and molecular dynamics across brain regions.The following study investigates the neurobiological and behavioural mechanism behind this phenomenon.

Methods: Yohimbine (Yoh) / Corticosterone (Cort) or both were administered to Sprague Dawley rats at the dose of 3 mg/kg,i.p; to upsurge nor-adrenergic and glucocorticoid activity, respectively. The behavioral performance was assessed using the eight arm Radial Arm Maze (RAM) task. The animals were divided into 4 cohorts based on the given pharmacological intervention.

The behavioral training protocol consisted of the pre-delay and post-delay phase. In the pre-delay phase, only four randomly selected arms were open and baited, while the other four were blocked. After the animal visited all baited arms, it was kept for a 5-minute delay period in its home cage. In the post-delay phase, previously blocked arms were baited and all arms were open for exploration.

The Performance Index (PI) was calculated as the ratio of baits consumed to total arm entries per trial. Behavioral performance concerning memory was factored using parameters like spatial working memory, reference memory, and performance index. Decision making parameters included reaction time and reward sensitivity (N = 6-18 per group). Bayesian multilevel framework and linear mixed model framework were used to model structure of the data and the relationships between variables across different experimental conditions.

Immunohistochemical (IHC) assessments were performed on 30-μm thick free-floating coronal brain sections to quantify levels of dopamine, p-glucocorticoid, cFos/NeuN and monoamine oxidase in various brain regions (N = 3, every 6th section). The brain sections were obtained immediately after behavioral experiments. Animals were transcardially perfused with ice-cold 4% paraformaldehyde (PFA) for fixation, and their brains were immediately dissected. For IHC analysis, the brain sections were first washed with PBS containing 0.1% Tween-20 and underwent epitope retrieval by heating in a sodium citrate buffer at pH 6 for 10 minutes. They were then blocked for 2 hours at room temperature in a buffer containing 10% goat serum in PBS with 0.03% Triton X-100. The sections were incubated for 24-48 hours at 4°C with primary antibodies targeting c-Fos, tyrosine hydroxylase, phosphorylated glucocorticoid receptor and monoamine oxidase, followed by 2-hour incubation with fluorescent secondary antibodies. Protein expression was quantified by measuring mean fluorescence intensity per microscopic field using a Leica BX51TF microscope. Cresyl violet staining was used to count pyknotic neurons in brain regions like striatum, PFC, amygdala and hippocampus at 400x magnification. Kit based ELISA was used to measure the corticosterone levels in the plasma and tissue.

Results: The co-activation of noradrenergic and glucocorticoid systems significantly impacted spatial memory and reward processing. After training, animals exhibited a significant reduction in working memory errors (p < 0.001) and reference memory errors (p < 0.001). Performance indices, indicating effective learning, increased after training (p < 0.01, RM ANOVA). In the pre-delay phase, Yohimbine (YOH) and Yohimbine + Cortisol (YOH + CORT) groups visited a higher number of arms (p < 0.01) and spent less time per arm, suggesting impaired spatial memory encoding. During spatial memory retrieval, YOH and YOH + CORT groups exhibited increased reference memory errors (p < 0.05) and working memory errors (p < 0.05). The YOH + CORT group showed heightened sensitivity to reward prediction errors, with a significant increase in travel time to the bait during wrong choices (p < 0.001). Corticosterone levels were significantly elevated in the prefrontal cortex and amygdala following YOH + CORT administration, accompanied by increased phosphorylated glucocorticoid receptor levels in the hippocampus. Notably, co-administration induced pyknotic neurons in the amygdala and mPFC and upregulated dopamine levels in the ventral tegmental area and striatum. Further analysis is underway.

Conclusions: This study demonstrates the complex interplay between adrenergic and glucocorticoid systems in modulating spatial memory and reward processing. The synergistic effects of these stress mediators alter neuronal integrity, dopamine levels, and glucocorticoid receptor activation, leading to changes in memory, performance and decision-making strategies. These findings provide insights into the neurobiological mechanisms underlying stress-induced alterations in cognitive flexibility, which may have implications for understanding origin of stress-related disorders.

Keywords: spatial memory, Acute Stress, Cortisol, Yohimbine, Decision Making

Disclosure: Nothing to disclose.

P890. Effects of Unpredictable Stress on Threat Processing and Defensive Learning

Susanna Molas*, Leshia Snively, Elora Williams, Emma Keppler

The University of Colorado Boulder, Boulder, Colorado, United States

Background: Defensive responses are critical to survival. In the face of potential danger, animals instinctively exhibit behaviors like freezing or fleeing to seek shelter. However, with repeated exposures to threatening stimuli not previously associated with a real danger, defensive behaviors decrease as part of an adaptive learning process.

Environmental factors, such as stressors, can potentiate defensive responses and heighten anxiety-like behaviors. Determining whether unpredictable stress disrupts innate threat responses and adaptive learning, as well as understanding the underlying neuronal mechanisms, will help develop better treatments for anxiety disorders.

Methods: We used the overhead dark visual looming stimulus (VLS) paradigm to investigate innate defensive responses and adaptive learning in rodents. The VLS apparatus consisted of a plexiglass box (60x20x30cm) with a rectangular shaped nest (20x12cm) located in one corner and a projector screen (30x20cm) above the arena. A video-camera was used to record and track animals’ behavior. Each VLS involved 15 consecutive dark expansions of 0.5s length and each mouse received 6 looming stimuli per session, repeated over the course of 3 days. To measure the effects of unpredictable stress, a group of experimental animals (N = 9 males and N = 11 females, C57BL6/J, 12-14 weeks old) were subjected to a series of unpredictable stressors for 4 days (1h restrain, 20 min forced swim, 10 min foot shock), prior to measuring defensive responses in the VLS test across 3 days. Control mice (N = 8 males and N = 9 females) remained in the home cage.

Results: Preliminary results indicate that animals subjected to unpredictable stress show impaired VLS-evoked escape responses (latency to the nest) in both males and females (treatment effect, F(1,110) = 9.403, p = 0.03). Additionally, unpredictable stress causes sex-specific effects on avoidance behaviors, measured as time spent inside the nest upon VLS events (sex x treatment effect, F(1,110) = 6.172, p = 0.015). We also found that across 3-days of VLS exposures, male and female mice demonstrate differences in escape (sex effect, F(1,110) = 5.39, p = 0.022) as well as avoidance behaviors (sex effect, F(1,110) = 4.583, p = 0.035). Interestingly, defensive responses to VLS are affected by the female estrous cycle, including freezing (estrous cycle effect, F(3,110) = 3.131, p = 0.039), escape (estrous cycle effect, F(3,110) = 3.587, p = 0.024) and avoidance behaviors (estrous cycle effect, F(3,110) = 8.432, p < 0.001).

Conclusions: The present study demonstrates that unpredictable stress disrupts defensive responses to a visual looming threat. The effects of stress on avoidance behavior are sex-specific, suggesting divergent pathways in males and females. Our findings also indicate that VLS-evoked defensive behaviors differ by sex and that hormonal fluctuations during the female estrous cycle can affect threat and stress responses. This study offers new insights into the neuroethology of stress-induced anxiety and may contribute to improving therapeutic strategies for neuropsychiatric conditions.

Keywords: Chronic unpredictable mild stress, defensive and motivated behaviors, Sex-specific effects

Disclosure: Nothing to disclose.

P891. Prenatal Stress and THC Exposure Impacts Maternal and Adolescent Behaviors

Jimmy Olusakin*, Mahima Dewan, Joseph Cheer, Mary Kay Lobo

University of Maryland, School of Medicine, Baltimore, Maryland, United States

Background: Cannabis is the most used illicit substance worldwide. The commonly cited reasons for continued cannabis use in chronic users is to cope with elevated stress (and/or) anxiety levels, pain and nausea. Pregnant women will often use cannabis to control nausea or stress/anxiety, and the effectiveness of cannabis in controlling these symptoms is likely due to the regulation of dopaminergic signaling in the reward-related brain areas. The main psychoactive component of cannabis, delta-9-tetrahydrocannabinol (THC), can readily cross the placenta during gestation, and is also secreted in the maternal milk during lactation. In addition, both clinical and preclinical studies have shown that chronic exposure to cannabis during gestation and lactation can induce behavioral teratogenic consequences. Little is known about the compound effects of prenatal exposure to stress and cannabis on long-term behaviors and cognition in adolescent offspring.

Methods: In this study, pregnant dams underwent 10 days of chronic witness defeat stress (CWDS) from embryonic day 0 (E0) with concurrent administration of subcutaneous i.p injections of 2mg/kg THC from E0 till birth. Following the 10 days of CWDS, we tested dams for anxiety-like behaviors – open field and elevated plus maze, and social interaction. At adolescence, postnatal day 35, we tested the offspring exposed to prenatal stress and THC on similar anxiety-like and social interaction behaviors. In a different cohort, we tested adolescent mice exposed to prenatal stress and THC on effort-related choice task using the Y-maze barrier test, a read-out for motivation. In the Y-maze barrier test, adolescent mice were presented with the option of either a high effort/high reward or a low effort/low reward food choice. All behaviors were analyzed with 2-way ANOVA (sex/treatment) In a separate cohort, tissue punches from the medial prefrontal cortex (mPFC) and nucleus accumbens (NAc) were obtained from adolescent mice exposed to prenatal THC and stress and using the single molecule detection assay by Nanostring technologies, we assessed the transcriptional substrates driving our observed behavioral deficits. Transcriptome analysis was done using the nCounter downstream analysis package provided by Nanostring Technologies.

Results: Our downstream behavior analysis in the dams showed no interaction effect of THC and stress on anxiety-like behaviors. However, we observed some main effects of THC alone in the social interaction (F1,21 = 4.55, p = 0.04) and a trending main effect in the open field tests suggesting that exposure to THC alone heightens anxiety-like behaviors. In addition, we observed significant decrease in litter sizes with increase in pup mortality in dams exposed to THC or stress alone. In the adolescent mice, we observed significant interaction effects of Prenatal THC and stress in the splash test suggesting a decrease in motivation to groom. Further, we observed sex specific differences in anxiety-like behaviors. Particularly, prenatal THC and stress exposed male adolescent mice showed decreased time spent exploring the open arm and center of the EPM and OFT respectively suggesting that compound exposures to prenatal THC and stress exacerbated anxiety-like behaviors. While we did not observe any interaction effects of prenatal THC and stress in the effort related test for motivation, we did find main effects of THC and stress in females and just main effect of stress in males in the 10cm barrier testing. Our transcriptome analysis revealed significant differentially expressed transcripts in the mPFC compared to the NAc. Further, we observed significant up regulation of genes involved in mitochondrial respiration and synaptic processing in the mPFC of adolescent mice exposed to prenatal stress and combination of prenatal stress and THC. Interestingly, these genes were down regulated in the NAc in both groups of adolescent mice.

Conclusions: Our study highlights some sex-specific behavioral deficits in adolescent mice exposed to prenatal stress and THC. We further observed transcriptional adaptations in these adolescent mice with significant changes differentially expressed genes involved in mitochondrial respiration, synaptic processing.

Keywords: Prenatal stress, THC, Anxiety-related behaviors, Adolescence, Transcriptome

Disclosure: Nothing to disclose.

P892. A Two-Hit Model of Juvenile Immune Activation and Adult Stress Synergistically Alters Cognitive Function in Adult Mice

Rachel Rahn*, Divija Chopra, Gabriela Manzano Nieves, Vivian Block, Christopher N Parkhurst, Conor Liston

Weill Cornell Medical College, New York, New York, United States

Background: Genome-wide association studies indicate that immune dysregulation plays a role in many psychiatric disorders including depression. Epidemiological data suggest a largely unexplored link between childhood infections and the later diagnosis of depression and other stress-related psychiatric disorders. However, how early-life immune activation influences stress susceptibility and behavior in adults remains understudied. We therefore aim to characterize the potential synergistic effects of immune activation during the juvenile period in mice and later stress in adulthood, to determine how inflammation and stress interactions alter the cortical functional connectome, resulting in cognitive behavioral deficits.

Methods: To study immune activation-facilitated cognitive dysregulation, we used a two-hit model of stress, consisting of immune activation during development (intraperitoneal injection of lipopolysaccharide (LPS) at postnatal day 21, 0.1 mg/kg body weight) followed by 7 days of mild unpredictable stress in adulthood. Behavioral studies were conducted using adult wildtype mice (C57BL/6J; n = 9-17 per group). For imaging studies, transgenic mice expressing a calcium indicator were used (Thy1/GCaMP6s; n = 7-9 per group). All groups were balanced for sex. LPS+stress mice were compared to stress-only and saline control mice in preliminary studies, and with LPS-only mice in subsequent cognitive behavioral assays. Comparisons were made to determine if LPS injection during the juvenile period amplified deficits or otherwise affected stress-related cognitive measures. Behavioral assays used included an operant reward conditioning task (tone-reward association; 3 days), and a variable reward task where the reward size depended on the relative difference in licks between the pre-tone and tone period (behavioral inhibition and effort; 4 days). Neural activity changes were assessed by widefield calcium imaging of excitatory pyramidal neurons in the adult mouse cortex and analysis of functional connectivity changes in cortical regions, defined by a previously published 32-seed set as well as a parietal region of interest identified in the preliminary dataset. Widefield calcium imaging data was analyzed in MATLAB, with initial processing utilizing the open-source pipeline of Brier and Culver (2023). Head-fixed resting state datasets were collected during wakefulness using sequential LED illumination with GCaMP6 fluorescence excitation and sCMOS detection, with 15-20 minutes of data per mice after data cleaning. A one-way ANOVA was used for preliminary 3-group analysis, and a 2-way ANOVA for 4-group analysis, followed by Tukey’s post hoc tests for group differences (PRISM Graphpad Software).

Results: In the operant reward conditioning task, where mice learned a tone-reward association, we found that LPS+stress mice completed fewer trials than saline controls (p = 0.0056), amplifying the subthreshold effect of stress only compared to saline controls (p = 0.076). This suggests that immune activation augments stress-induced deficits in reward behavior. Non-specific (inter-trial period, ITI) licking was also reduced in LPS+stress animals compared to saline controls, suggesting that LPS+stress may be decreasing motivation in the mice. This led us to further examine LPS and stress-related changes in reward behavior by using a variable effort task that rewards increased licking to a CS+ tone and inhibition of licking during pre-tone ITI periods. As before, we found that LPS+stress mice displayed decreased ITI lick rates, and this decreased ITI lick rate facilitated the acquisition of larger rewards, allowing the mice to learn the rule (lick less to ITI, more to tones) faster. Interestingly, within this task we did not observe the decrease in trial acquisition observed in the operant reward task. This suggests that the decreased reward motivation may be partly due to reward size, or reward type, thus LPS+stress may be making mice less sensitive to low rewards. When we performed widefield calcium imaging of excitatory pyramidal neurons to examine region-specific changes in activity, we also found that LPS and stress significantly altered adult frontoparietal functional connectivity (one-way ANOVA p = 0.0071). Frontoparietal resting state connectivity was significantly decreased in LPS+stress mice compared to saline controls (p = 0.0054), demonstrating a significant change not observed between stress only and saline control groups (p = 0.14).

Conclusions: Our findings demonstrate that developmental immune activation and adult stress impact brain networks and cognitive domains including reward behavior and motivation. We found that LPS+stress reduced motivated and effortful behavior while not impacting successful task learning. These behavioral deficits suggest that juvenile immune activation and adult stress both contribute to cognitive deficits in effortful reward behavior, and to anhedonia-related deficits similar to those seen in depression in humans. We also identified alterations in the resting-state functional connectome after LPS and stress exposure. The observed deficit in frontoparietal connectivity involves regions implicated in complex reward-related behavior, cognitive control, and depression. Further experiments will directly test how these differences in frontoparietal connectivity relate to our observed effort and motivation deficits in the context of stress and immune activation.

Keywords: Chronic unpredictable mild stress, LPS, Reward, resting-state functional connectivity, Reward, motivation, and anhedonia

Disclosure: Nothing to disclose.

P893. Olfactory Processing During Chronic Social Defeat Stress is Crucial for the Induction of Behavioral Disturbance in Mice

Yuki Okuda, Dongrui Li, Yuzuki Maruyama, Hirokazu Sonobe, Tomoyuki Mano, Kazuki Tainaka, Ryota Shinohara, Tomoyuki Furuyashiki*

Kobe University Graduate School of Medicine, Kobe, Japan

Background: Environmental factors such as trauma and stressful life events increase the risk of mental illnesses, including depression. Rodent studies using chronic stress models have elucidated neural-circuit mechanisms underlying depression-related behaviors. These studies have mainly focused on brain regions associated with emotion and cognition. Chronic stress is known to induce dendritic atrophy, reduce dendritic spines, and decrease excitability in the hippocampus and medial prefrontal cortex (mPFC), leading to emotional and cognitive disturbances. In contrast, acute stress induces effects such as dendritic hypertrophy of mPFC neurons to suppress these disturbances. Since sensory information from external stimuli is initially processed in sensory regions, abnormalities in these regions may contribute to emotional and cognitive disturbances in mental illnesses. Clinical studies have implicated sensory systems in depression and its treatment. However, rodent studies have not yet explored how animals perceive external stimuli as stressors that lead to emotional and cognitive disturbances. It is essential to examine brain-wide neural responses to acute and chronic stress and their relevance to behavior in an unbiased manner.

Methods: Male Arc-dVenus neural response reporter mice were subjected to acute or chronic social defeat stress (SDS). Mice exposed to chronic SDS were categorized into susceptible and resilient groups based on their level of social avoidance, a typical depression-related behavior. We conducted three-dimensional whole-brain imaging and unbiasedly analyzed the fluorescent intensities of each brain region using UMAP clustering and machine learning. We further investigated the behavioral role of brain regions selectively responsive to chronic stress over acute stress using chemogenetic methods.

Results: Brain regions across the brain contributed differently to neural responses to acute and chronic SDS. Machine learning analysis revealed that multiple sensory cortices had a stronger contribution to neural responses to acute and chronic SDS than stress-related association cortices, such as the prelimbic and infralimbic cortices. The contributions of sensory cortices varied depending on the sensory modality. Auditory and visual cortices contributed to both acute and chronic SDS, while olfactory (piriform) and gustatory (insular) cortices were selectively responsive to chronic SDS. The response of the piriform cortex to SDS increased with repeated stress exposure. Chemogenetic manipulation of neuronal activities demonstrated that activation of the piriform cortex to the lateral septum pathway during chronic SDS was crucial for inducing behavioral disturbances.

Conclusions: This study identified distinct patterns of stress-induced neural activity across the brain and highlighted the significance of the olfactory pathway in perceiving chronic stress, leading to behavioral disturbances through its connection to the lateral septum.

Keywords: Chronic social stress, Emotional dysregulation, Whole-Brain Rodent Imaging, sensory processing

Disclosure: Nothing to disclose.

P894. Microglial Glutaminase as a Therapeutic Target in Chronic Stress-Induced Depression

Yannan Li, Meixiang Huang, Ajit G. Thomas, Wathsala Liyanage, Lukas Tenora, Niyada Hin, Mizuho Obayashi, Pavel Majer, Rangaramanujam M Kannan, Takashi Tsukamoto, Rana Rais, Barbara S. Slusher, Xiaolei Zhu*

Johns Hopkins University School of Medicine, Baltimore, Maryland, United States

Background: Major depressive disorder (MDD) is a prevalent, debilitating psychiatric disorder. Despite various treatments, resistance to current antidepressants is common, necessitating novel interventions based on MDD’s pathological mechanisms. Our study focuses on the role of glutaminase (GLS1), an enzyme upregulated in activated microglia, in the pathophysiology of chronic stress-induced depression.

Methods: In this study, we used the Chronic Social Defeat Stress (CSDS) model, a well-established rodent model for both sexes, to study stress-induced psychiatric disorders, including depression. Our approach involved the development of hydroxyl-terminated poly(amidoamine) (PAMAM) dendrimer nanoparticle delivery systems that specifically target microglial GLS1. We synthesized two novel dendrimer-GLS1 inhibitors, Dendrimer-TTM020 (D-TTM020) and Dendrimer-JHU29 (D-JHU29), and evaluated their target engagement effects and efficacy using GLS enzymatic activity assays, immunohistochemistry, histopathology, and a battery of behavioral tests, including social interaction test.

Results: Our dendrimer conjugates showed selective targeting towards activated microglia. Both Dendrimer-TTM020 and Dendrimer-JHU29 significantly inhibited microglial GLS1 activity in two brain regions (PFC: CSDS+Veh vs. CSDS + D-TTM020, P = 0.0213, CSDS+Veh vs. CSDS + D-JHU29, P = 0.0419; HPC: CSDS+Veh vs. CSDS + D-TTM020, P = 0.0004, CSDS+Veh vs. CSDS + D-JHU29, P = 0.0395). Moreover, these conjugates improved sociability in the CSDS murine model (Discrimination index: CSDS+Veh vs. CSDS + D-TTM020, P = 0.0010, CSDS+Veh vs. CSDS + D-JHU29, P = 0.0009) without exhibiting undesirable effects, specifically no GI-related toxicities, indicating their potential as effective treatments for chronic stress-associated depression.

Conclusions: The development of microglial-targeted GLS1 inhibitors, particularly dendrimer-based conjugates, represents a novel therapeutic approach for chronic stress-associated depression. These inhibitors offer a promising alternative to current treatments, setting the stage for future clinical studies.

Keywords: Chronic social defeat stress, Microglia, Glutaminase Inhibitors, Depression, Nanoparticle Delivery Systems

Disclosure: Nothing to disclose.

P895. Testing Optimal iPSC Derived Cell Types and in Vitro Glucocorticoid Conditions for Modeling Stress-Related Psychiatric Disorders’ Multi-Omic Signatures

Cameron Pernia*, Chris Chatzinakos, Aarti Jajoo, Artemis Iatrou, Ilse Kolmans, Clara Snijders, Kerry Ressler, Nikolaos Daskalakis

Harvard Medical School McLean Hospital, Belmont, Massachusetts, United States

Background: Our group published the first in vitro model that recapitulated cell type specific effects of PTSD with iPSC derived neurons exposed to 100nM Dexamethasone (DEX), and the largest multi-omic and multi-brain region investigation of PTSD and MDD to elucidate disease signatures. What combination of in vitro cell types and DEX doses that best models psychiatric stress is still unknown. In this work, we identify the in vitro conditions that best recapitulate stress-related disorders’, such as PTSD and MDD, multi-omic signatures.

Methods: 2 control iPSC lines were differentiated into neural stem cells (NSCs), immature neurons (INs), and neuron cultures. At each stage, cells were treated with DEX (0nM, 5nM, 25nM, 100nM), and harvested for RNAseq and DNA methylation (DNAm) arrays. RNA and DNA underwent RiboZero and EPIC methylation arrays respectively. Differential analysis was performed via voom/limma, gene networks were identified with WGCNA, and relevance to published post mortem brain datasets was tested with RRHO and GSEA.

Results: NSCs showed fewer differentially expressed genes (DEGs) than INs and neurons when exposed to DEX. WCGNA of DEX DEGs in INs and neurons identified gene modules associated with neuronal processes (logFC = -0.26; padj=0.031) and immune signaling (b = 0.005; padj=8.32e-6). RRHO (r = 0.46; padj=2.89e-11) and gene set analysis (NES = 2.57; padj= 1.62e-30) of DEX DEGs with clinical datasets showed distinct enrichments across various in vitro conditions with PTSD and MDD. PTSD and MDD GWAS genes were present unevenly across cell types (n = 0 – 262). Most differentially methylated genes (DMGs) due to DEX across all cell types were associated with stress or psychiatry (n = 111/119). GSEA of DMGs due to DEX in the INs and neurons were enriched for neuronal and synaptic processes (NES = 1.42; padj= 1.33e-8).

Conclusions: In this study, we explore various iPSC derived cells types and DEX dosages for best modeling individual stress associated psychiatric disease based off clinical datasets. We observed GR activation was not consistent across cell types, and cell types’ relevance to the psychiatric datasets varied across analytical approaches. Overall, increasing the maturity of our cultures improved their clinical significance. Our findings can be applied for designing the optimal in vitro drug screen for stress related psychiatric disorders.

Keywords: Glucocorticoids, iPSCs, Post Traumatic Stress Disorder, post-mortem brain, Multiomics

Disclosure: Nothing to disclose.

P896. An Epidaemiological Investigation Into Suicide Death Rates Among Adolescents in Muslim-Majority Countries

Salahudeen Mirza*, Murad Moosa Khan

Yale School of Medicine, New Haven, Connecticut, United States

Background: Suicide is a leading cause of death among youth worldwide. Yet, most of the high-quality epidaemiological data is focused on countries from the Global North (e.g., United States, Canada, United Kingdom). Early on, Durkheim uncovered differences in suicide rates among Christian denominations; following this work, many have assumed that the Islamic prohibition of suicide would lead to a lower rate of suicide deaths in Muslim-majority countries. Though part of the pattern might be attributable to criminalization and under-reporting, it is true that the age-standardised suicide death rate is lower in Muslim-majority as compared to non-Muslim-majority countries. At the same time, suicidal and self-harming behaviours are clearly prevalent among youths in Muslim-majority countries. Though, to our knowledge there has been no focused investigation of suicide death rates among youths specifically in Muslim-majority countries. This information is critical to developing better prevention and interventional strategies.

Methods: Out of 183 countries with available data, 46 countries with at least 50% Muslim population proportion were selected according to World Population Review. Country-by-country suicide death rate estimates for adolescents (ages 15-19 years) were obtained from the World Health Organisation 2019 published data. Weighted average adolescent suicide death rates across all countries, all Muslim-majority countries, and all non-Muslim-majority countries, were taken by multiplying each country’s 15-19 years old suicide death rate by the population ages 15-19 in 2019 (United Nations World Population Prospects), then dividing by the total population ages 15-19 across all countries included in each average. The weighted average adolescent suicide death rate in Muslim-majority countries was compared to the global weighted average adolescent suicide death rate to identify any Muslim-majority countries which showed an above-average adolescent suicide death rate.

Mann-Whitney U-test was implemented to compare the overall adolescent suicide death rate between Muslim-majority and non-Muslim-majority countries, as well as rates within males and females and the suicide death sex ratios. Pearson correlational analysis within Muslim-majority countries considered whether adolescent suicide death rate was significantly associated with scores on the global youth development index, the global peace index, or the Muslim population proportion.

Results: Across all included countries, the weighted average adolescent suicide death rate was 5.93/100,000. Of the Muslim-majority countries, 10 (22%) out of the 46 included were above the global weighted average: Uzbekistan (15.38/100,000), Kazakhstan (15.35/100,000), Kyrgyzstan (10.04/100,000), Pakistan (10.03/100,000), Turkmenistan (7.80/100,000), Iran (6.87/100,000), Morocco (6.75/100,000), Yemen (6.24/100,000), Oman (6.07/100,000), and Albania (5.97/100,000). Uzbekistan and Kazakhstan rank in the top 10 for adolescent suicide death rates across all countries. Weighted average adolescent suicide death rate was 4.91/100,000 in Muslim-majority countries and 6.23/100,000 in non-Muslim-majority countries.

The mean (non-weighted) adolescent suicide death rate was significantly lower (35%) in Muslim-majority countries by Mann-Whitney U-test (P < .05). Follow-up analysis revealed that there was only a significant difference in mean adolescent suicide death rates between Muslim-majority and non-Muslim-majority countries for males (49% lower in Muslim-majority countries, P < .05), but not females. There was no significant difference in mean adolescent suicide death sex ratios between Muslim-majority and non-Muslim-majority countries. In Oman, Kuwait, Malaysia, Qatar, Syria, Maldives, and Bahrain, the adolescent suicide death sex ratio of male-to-female was > 4, suggesting much increased risk for males.

Correlational analysis revealed no significant associations of adolescent suicide death rates with global youth development index, global peace index, or Muslim population proportion.

Conclusions: This preliminary epidaemiological investigation of secondary data from Muslim-majority countries suggests an overall reduced adolescent suicide death rate in Muslim-majority countries. This may not reflect a true reduction in suicide deaths as it may be attributable to under-reporting, issues with estimations, and other artifacts. Some Muslim-majority countries show concerningly high adolescent suicide death rates and may benefit from culturally tailored prevention strategies. Further investigation of associated factors might better inform approaches which could be taken to mitigate adolescent deaths by suicide in Muslim-majority countries.

Keywords: suicide, religion, cultural psychiatry, Adolescence, epidemiology

Disclosure: Nothing to disclose.

P897. The Interplay Between Genetics and Epigenetics Associated With Attempted Suicide

Aysheh Alrfooh, Lucas Casten, Jenny Gringer Richards, John Wemmie, Vincent Magnotta, Jess Fiedorowicz, Jacob Michaelson, Aislinn Williams, Marie Gaine*

University of Iowa, College of Pharmacy, Iowa City, Iowa, United States

Background: Individuals with bipolar disorder are at increased risk for suicide, and this can be influenced by a range of biological, clinical, and environmental risk factors. Biological components associated with suicide include DNA modifications that lead to changes in gene expression. Common genetic variation and DNA methylation changes are some of the most frequent types of DNA changes associated with an increased risk for suicidality. Importantly, the interplay between genetic variation and DNA methylation changes is becoming more prevalent in research.

Methods: We hypothesized that DNA methylation patterns in specific loci already genetically associated with suicide would be altered in individuals with bipolar disorder and a history of suicide attempt. To test this hypothesis, we searched the literature to identify common genetic variants (N = 34) previously associated with suicidal thoughts and behaviors in individuals with bipolar disorder. We then created a customized sequencing panel that covered our chosen genomic loci. We profiled DNA methylation patterns from blood samples collected from bipolar disorder participants of both sexes with suicidal behavior (N = 55) and without suicidal behavior (N = 51). For the primary CpG site analysis a logistic regression model was used where age, sex, race, body mass index (BMI), and smoking history were included as covariates. For the correlation analyses a non-parametric test (Kendall’s test) was used.

Results: We identified seven differentially methylated CpG sites and five differentially methylated regions between the two groups (adjusted p-value < 0.05). Additionally, we found that DNA methylation changes in MIF and CACNA1C were associated with lethality or number of suicide attempts (FDR < 0.05). Finally, we identified three meQTLs in SIRT1, IMPA2, and INPP1 (FDR < 0.05). This study illustrates that DNA methylation is altered in individuals with bipolar disorder and a history of suicide attempts in regions known to harbor suicide-related variants.

Conclusions: This study illustrates that DNA methylation is altered in individuals with bipolar disorder and a history of suicide attempts in regions known to harbor suicide-related variants.

Keywords: QTL, DNA Methylation, Genetic variation, bipolar disorder (BD), Suicide attempt

Disclosure: Nothing to disclose.

P898. Correlation of Quinolinic Acid With Suicidal Ideation and Behavior in Bipolar Disorder

Angelos Halaris*, Anton Shkundin, Stephen Murata

Loyola University School of Medicine, Maywood, Illinois, United States

Background: Depression in patients with bipolar disorder (BD) presents significant clinical challenges, including a high suicide risk. Suicide is a frequent consequence of bipolar depression, with severe depression being the primary predictor of suicide risk in BD. Inflammation has been linked to both depression and suicide risk and is associated with induction of the the kynurenine pathway. This leads to altered levels of quinolinic acid (Quin), a neuroactive metabolite and NMDA receptor agonist, which may contribute to suicidality via modulation of glutamatergic neurotransmission and neuroinflammation. This study explored the potential for using quinolinic acid as a diagnostic predictor of suicidal ideation and behavior in bipolar disorder patients.

Methods: The study included 42 males and females aged 21 to 65 years who met DSM-IV criteria for bipolar disorder (BD I or II) in the depressed phase. Participants had previously failed to respond to at least two adequate antidepressant trials or experienced a breakthrough depressive episode despite being on a mood stabilizer and/or an atypical antipsychotic. After an overnight fast, 20 ml of antecubital venous blood was drawn between 08:00 and 10:00 AM at baseline. Kynurenine pathway metabolites including quinolinic acid were measured using Ultra Performance Liquid Chromatography/Mass Spectrometry (UPLC-MS). Pearson correlation analysis was conducted to explore the relationship between kynurenine pathway metabolites including quinolinic acid (Quin) and the Columbia-Suicide Severity Rating Scale (C-SSRS), including measures of suicidal ideation and behavior.

Results: Quin demonstrated significant correlations with key kynurenic pathway metabolites, including AA (r = 0.733, p < 0.001), Kyn (r = 0.590, p < 0.001), KynA (r = 0.389, p = 0.011), and the Kyn/Trp ratio (r = 0.824, p < 0.001). Quin and its ratios were significantly correlated with C-SSRS measures, including actual attempt with Quin/Trp ratio (r = 0.825, p < 0.001), Quin (r = 0.727, p < 0.001), and Quin/3-HK ratio (r = 0.410, p = 0.018). Active suicidal ideation with intent to act (without a specific plan) correlated with Quin/Trp ratio (r = 0.532, p < 0.001) and Quin (r = 0.503, p = 0.002). Active suicidal ideation with a specific plan and intent correlated with Quin/Trp ratio (r = 0.507, p = 0.002) and Quin (r = 0.461, p = 0.005). Active suicidal ideation (with any method) without intent to act correlated with Quin/Trp ratio (r = 0.488, p = 0.003) and Quin (r = 0.438, p = 0.008). Additionally, Quin correlated with a wish to be dead (r = 0.352, p = 0.026). Interrupted attempt correlated with Quin/Trp ratio (r = 0.472, p = 0.006) and Quin (r = 0.377, p = 0.030). Lastly, actual lethality significantly correlated with Quin (r = 0.469, p = 0.037) and Quin/Trp ratio (r = 0.459, p = 0.041).

Conclusions: These findings demonstrate significant associations between quinolinic acid and measures of suicidal ideation and behavior. They provide important insights into the biological processes that may contribute to suicide risk in individuals with bipolar disorder and could potentially be used as a biomarker for suicide prevention possibly by modulating kynurenine metabolite levels.

Keywords: Bipolar Depression, Quinolinic Acid, Kynurenine Pathway, The Columbia-Suicide Severity Rating Scale (C-SSRS)

Disclosure: Nothing to disclose.

P899. fMRI Biomarkers of Dimensional Personality Functioning as Predictors of Suicide-Related Treatment Outcomes Following Partial Hospitalization

Jenna Traynor*, Aayushi Sangani, Kailyn Fan, Courtney Beard, Roger Pottanat, Kerry Ressler, Lois Choi-Kain

McLean Hospital, Harvard Medical School, Belmont, Massachusetts, United States

Background: Suicide is a public health crisis. Personality dysfunction is a major risk factor for suicide. Several levels of care are available to treat suicidality (e.g., inpatient, partial hospital, outpatient), but we have a poor understanding of which levels of care are best suited to which patients. In part, this is because the DSM-5 diagnoses used to assess patients and inform treatment selection have well-documented psychometric problems, and poor predictive utility. Recently, alternative dimensional models have emerged that quantify personality dysfunction by measuring the severity of impairments in self and interpersonal functioning, but no study has examined whether this approach is a better predictor of treatment response, compared to traditional DSM-5 categories. Using a sample of suicidal patients in a transdiagnostic partial hospital, this study compares the prognostic utility of categorical DSM-5 assessment in a head-to-head fashion against a new self-report measure of dimensional personality functioning. We had four aims: i) characterize changes in suicidality and depression across treatment ii) examine whether categorical vs. dimensional assessment is a better predictor of suicidality and depression at post-treatment, iii) identify associations between DSM-5 categorical and dimensional scores with resting-state fMRI connectivity in major neural networks and iv) use resting-state fMRI profiles to predict the resolution of suicide at post-treatment.

Methods: Data from 28 suicidal patients treated in a CBT partial hospital is included in this abstract. Data collection is ongoing. At baseline, patients completed a 12-min resting-state fMRI scan, the MINI Neuropsychiatric Interview (i.e., categorical DSM-5) and the Levels of Personality Functioning Scale – Brief Form (LPFS-BF; a 12-item self-report measure of dimensional personality functioning). Suicide ideation (the DSI – Suicidality Subscale), and depression severity (the Patient Health Questionnaire – 9 Item) were collected at pre- and post-treatment, and 3-month follow up. Multilevel models explored changes in symptoms from baseline to follow up. Logit and linear regression examined whether number of MINI diagnoses vs. dimensional LPFS-BF scores predicted suicidality (present/absent) and depression severity at post-treatment. fMRI data were analyzed using the CONN Toolbox. Pre-processing included: discarding the first 3 functional volumes; realignment and unwarping; co-registration of functional and structural images; normalization to MNI space; segmentation, 2mm3 resampling; outlier detection; and 6mm smoothing. Participant connectivity maps were produced in first-level analysis after denoising with CompCor. Motion scrubbing, linear detrending, and a 0.008 – 0.09 Hz band pass filter were applied. Regions of interest were set in the default mode network (DMN), salience (SAL), frontolimbic, and dorsal and ventral attention networks (DAN; VAN), using a Yeo-17 parcellation and subcortical atlas. Second-level analysis is currently underway to characterize associations between rest-connectivity in the DMN, SAL, frontolimbic, DAN and VAN with MINI diagnoses and LPFS-BF scores, controlling for age and sex. Results will be considered significant if they survive FDR correction at p < 0.05. fMRI connectivity values will be entered into the logistic regression to examine whether they further improve the prediction of suicidality at post-treatment.

Results: Significant improvements were observed in suicide ideation at post-treatment (β = −2.78, S.E. = 0.35, t[26] = −7.89, p < 0.001) and follow up (β = −2.73, S.E. = 0.47, t[26] = −5.84, p < 0.001), and in depression at post-treatment (β = −7.36, S.E. = 1.37, t[25] =−5.36, p < 0.001) and follow-up (β = −7.68, S.E. = 1.88, t[25] =−4.09, p < 0.01). A trend was observed suggesting that dimensional LPFS-BF scores may predict the resolution of suicide ideation (β = 0.01, SE = 0.01, z = 1.73, p = 0.08) and depression severity (β = 0.05, SE = 0.02, t = 2.19, p = 0.05) at post-treatment, whereas number of DSM-5 MINI diagnoses did not predict the resolution of suicide ideation (β = 0.08, SE = 0.43, z = 0.19, p = 0.85) or depression severity (β = 0.87, SE = 0.96, t = 0.91, p = 0.38) at post-treatment. Associations between DMN, SAL, frontolimbic, DAN and VAN connectivity profiles with dimensional vs. categorical scores are currently being characterized. Subject Fisher-Z values (ROI-to-ROI) will be entered into the logistic regression to predict the resolution of suicidality at post-treatment.

Conclusions: Findings thus far suggest that a dimensional measure of personality dysfunction may be a better prognostic predictor than DSM-5 categorical nosology. The dimensional self-report can be completed in < 10 min, whereas DSM-5 based assessments span at least 60 minutes and require a trained clinician. As such, these preliminary findings are impactful and point to a more reliable and time-saving method of assessment to predict the resolution of suicidality following partial hospitalization. Characterizing the resting-state architecture underlying scores on each diagnostic measure will be a highly novel and potentially informative approach to identify patient profiles that are likely to respond to the partial hospital intervention, ultimately helping to prevent loss of life in patients at the highest risk of suicide. This study is funded by the Brain and Behavior Research Foundation and Families for BPD Research.

Keywords: Suicide prediction, Resting-state fMRI, Transdiagnostic

Disclosure: Nothing to disclose.

P900. Accident-Proneness and Suicide Risk: A Retrospective Chart Review of Trauma Patients in an Urban Emergency Center

Ynhi Thomas*, Nidal Moukaddam, Syed Murtaza, Kelly Keene, Nicholas Murphy, Chad Wilson, Thomas Kosten, Alan C. Swann

Baylor College of Medicine, Henry J.N. Taub Department of Emergency Medicine, Houston, TX, USA, Houston, Texas, United States

Background: Unintentional injuries and suicide are leading causes of death and disability in the United States. Among survivors of medically severe suicide attempts and non-suicidal traumatic injuries, subsequent accidents are a leading cause of premature death. Additionally, individuals who have survived non-suicidal injuries are at a high risk of recurrent severe physical trauma and may also be at an elevated risk of suicide attempts. Notably, approximately 60% of suicide deaths occur during the first attempt without prior mental disorder diagnoses. Understanding the relationships between accident-proneness and suicide risk can provide valuable insights for prevention strategies.

Methods: In this 18-month retrospective chart review (N = 225, ≥18 y/o) study at an urban Emergency Center, our team investigated clinical indicators related to physical trauma, current suicide attempt or self-harm, and past suicide attempts. All 225 subjects were physical trauma patients, and all had a psychiatry consult in the Emergency Center due to concerns for their mental health by the provider. We reviewed demographics, clinical history, and injury mechanisms for trauma patients presenting with versus without past suicide attempts. We compared Emergency Center patients with physical trauma in 2x2 cells: a) current suicide attempt or self-harm versus b) no suicide attempt or self-harm, by c) past suicide attempt versus d) no past suicide attempt.

Results: All 225 participants were included in the data analyses. The mean age was 38 years old; 61% were males and 39% were females. There were 49% whites, 42% blacks, 1% Asian, and 7% other. Mechanisms of injuries for participants with versus without suicide attempt history, respectively, were overdoses (9% vs 4%), penetrating wounds (31% vs 14%), falls (16% vs 19%), and blunt trauma (44% vs 61%) (P < 0.004). Among the 88 participants with past suicide attempts, 64% presented with another suicide attempt versus 21% with a non-suicidal accident (P < 0.001). Participants with previous suicide attempts had higher rates of psychotic, affective, drug use, and post-traumatic stress disorders, as well as more medical and psychiatric hospitalizations, and Emergency Center visits (all P < 0.05). Self-harm rates were comparable (20% vs 17%), but suicide attempt rates were about half (20% vs 50%) in participants without (n = 137) versus with previous suicide attempts (n = 88). Among 137 participants without previous suicide attempts, 50 had suicide attempt or self-harm at Emergency Center presentation and 87 did not. The 87 participants without either previous suicide attempt or self-harm had more past Emergency Center visits (4.61 vs 3.08, P < 0.001) and past medical hospitalizations (0.74 vs 0.24, P < 0.05) than the 50 participants with both previous and current suicide attempt or self-harm. The two groups did not differ in demographics, mental disorder, or psychiatric hospitalization.

Conclusions: Participants with a history of suicide attempts were more likely to present with another suicide attempt and to have past mental disorders compared to those without such a history. However, 37% (50/137) of trauma participants without prior suicide attempts presented with self-harm or suicide attempts at Emergency Center presentation, indicating a high risk for suicide death (SD). These 50 participants may represent part of the 60% of potential SDs occurring without a prior mental disorder diagnosis. Interestingly, participants without previous suicide attempts who did not present with self-harm or suicide attempts had more past Emergency Center visits and medical hospitalizations than those with both previous and current suicide attempts or self-harm. This suggests that frequent Emergency Center visits and hospitalizations may be indicators of underlying issues that increase the risk of severe outcomes, highlighting the need for improved screening and intervention strategies in Emergency Centers. The similarity in demographics, mental disorder rates, and psychiatric hospitalizations between these groups suggests that other factors, such as the nature and context of physical trauma, might play a critical role in suicide risk. More research is needed to identify characteristics of suicide attempts or self-harm in trauma patients, which could represent predictors of potentially lethal first suicide attempts. Understanding these predictors could inform targeted interventions and improve outcomes for high-risk individuals.

Keywords: Suicide risk factors, Trauma, Mental Disorders

Disclosure: Nothing to disclose.

P901. Obsessive-Compulsive Personality Disorder and the Risk of Suicidal Behavior in Individuals With Anxiety and Mood Disorders

Vilma Jakiene*, Naomi Fineberg, Aurelija Podlipskyte, Julija Gecaite-Stonciene, Alicja Juskiene, Vesta Steibliene, Julius Burkauskas

Laboratory of Behavioral Medicine, Neuroscience Institute, Lithuanian University of Health Sciences, Palanga, Lithuania

Background: Recent studies recognize a significant risk of suicidal behavior in patients with obsessive-compulsive and related disorders (Pellegrini et al., 2021; Bowen aet al., 2019). Anxiety and mood disorders are also important risk factors for suicide (Moitra et al., 2021), while the co-presence of personality disorder may further amplify the risk (McClelland et al., 2023). Obsessive-compulsive personality disorder (OCPD) is a common comorbidity in patients with anxiety and mood disorders (Burkauskas et al., 2019) but tends to be overlooked by clinicians (Rizvi et al., 2023). There is limited data analyzing whether the coexistence of OCPD contributes to an increased risk of suicidal behavior in individuals with anxiety and mood disorders.

Methods: A cross-sectional analysis was conducted in a sample of individuals with anxiety and mood disorders. Data were collected on variables such as gender, age, education, history of smoking, current mental disorders, and current medication use. All participants were interviewed for current psychiatric diagnoses and suicidality risk using the Mini International Neuropsychiatric Interview (M.I.N.I. 7.0.2). Suicidality risk was assessed using Section B of the MINI. Based on the sum of the weighted scores of the “yes” items, a total of 1–8 points were classified as ‘low-risk’, 9–16 points as ‘moderate-risk’, and 17 points and higher as ‘high-suicidality risk’. The Compulsive Personality Assessment Scale (CPAS) was used to assess features associated with OCPD. Chi-square tests were used to assess categorical variables, while multivariable logistic regression analyses were used to assess the associations between OCPD and suicidality risk.

Results: We recruited 227 individuals with anxiety and mood disorders (78.1 % women, mean age 40.0 ± 11.6). All study participants had either single or comorbid anxiety and mood disorders diagnoses: 145 (63.9%) major depressive episode, 20 (8.8%) persistent depressive disorder, 10 (4.4%) bipolar disorder, 88 (38.8%) generalized anxiety disorder, 75 (33.0%) panic disorder, 54 (23.8%) agoraphobia, and 46 (20.3%) social anxiety disorder. Forty-nine (21.5%) fulfilled operational criteria for DSM-5 OCPD. Suicidality risk was more prevalent in individuals with anxiety and mood disorders and OCPD in comparison to individuals with anxiety and mood disorders only (59.2% vs. 42.1%, p = 0.034).

A multivariable logistic regression revealed that individuals with anxiety and mood disorders and comorbid OCPD were more likely to have a higher suicidality risk compared to those with anxiety and mood disorders without OCPD (odds ratio = 2.14, 95% CI 1.08–4.24; p = 0.029), after adjusting for age, gender, education, and medication use.

Conclusions: The presence of OCPD is associated with increased suicidality risk in individuals with anxiety and mood disorders. Clinicians should recognize that OCPD is a relevant factor for increasing suicidality risk, just like other forms of personality disorder, and therefore should routinely screen for this in their clinical assessment of patients with anxiety and mood disorders.

Keywords: Mood and anxiety disorders, Suicidality, Obsessive-compulsive personality disorders

Disclosure: Nothing to disclose.

P902. Suicide-Related Language in Smartphone Keyboard Inputs Among High-Risk Adolescents

Paul Bloom*, Julia Greenblatt, Natalia Parjane, Hanga Galfalvy, Karla Joyce, Katherine Durham, Giovanna Porta, Jaclyn S. Kirshenbaum, Esha Trivedi, Lauren S. Chernick, Peter Dayan, David Brent, Nicholas B. Allen, David Pagliaccio, Randy Auerbach

Columbia University, New York, New York, United States

Background: Adolescent suicide is a public health crisis, and thus, there is an urgent need to better understand acute risk factors for suicidal thoughts and behaviors (STB) among adolescents. Natural language processing of typed text, particularly via text messaging and posting on social media, may be one method for identifying risk for STB in real time. Yet, relatively little is known about how young people use suicide-related language in typed text, and prior work has been mostly limited in scope to public conversations within a select set of social media applications. Given these gaps, we sought to characterize patterns of suicide-related language typed by adolescents on smartphones, including text messaging, social media apps, and web browser searches.

Methods: Adolescents ages 13-18-years-old reporting depressive, anxiety, and/or substance use disorders were recruited as part of a larger longitudinal study in the greater New York City and Pittsburgh areas from psychiatric outpatient programs, emergency departments, medical center research registries, and social media. In current preliminary analyses, data from a subsample (N = 48) of a larger cohort (N = 223) were analyzed (the full sample will be included for the final analysis). At baseline, 52.1% of the subsample were experiencing current suicidal ideation, and additionally, 31.3% reported current ideation and a past-year suicide attempt. Over a 6-month period, typed text was passively logged via the EARS smartphone app (n = 180,271 total messages). De-identified messages containing suicide-related tokens (e.g., “suicid*”, “kms”, “kill myself”) were reviewed and coded by research staff for relevance to suicide and content (N = 991, 0.55%). EARS also recorded the app that was open while each entry was typed. Over the same period, youth also completed weekly smartphone surveys on the frequency of suicidal thoughts.

Results: A total of 277 text entries (0.15%) from 23 participants (47.9% of the subsample) were coded as relevant to suicide. The most frequent categories of suicide-related entries included descriptions of current suicidal thoughts (26%), support or advice to others with suicidal thoughts (19%), jokes about suicide (15%), and descriptions of prior suicide attempts (9%). Most suicide-relevant entries were typed in Snapchat (69%), with other entries mostly in text message (19%), Instagram (4%), and web browsing apps (3%). Bayesian linear regression analysis indicated a positive between-participants association between self-reported suicidal ideation severity at baseline and frequency of suicide-relevant entries (β = 0.50, 95% CI [0.23, 0.76]). In addition, Bayesian multilevel linear regression indicated that increased frequency of suicide-relevant entries in a given week was associated with a higher frequency of next-week suicidal thoughts reported via weekly surveys (β = 0.11, 95% CI [0.04, 0.18]).

Conclusions: Preliminary findings suggest that more frequent suicide-relevant entries were associated with increased weekly suicidal ideation. Future analyses will extend the current analyses in the full cohort, and work to identify types of suicide-related language associated with short-term likelihood of suicidal events (e.g. attempts, emergency department visits or hospitalization for suicide-related concerns). More broadly, a better understanding of youth’s typed language around STB may facilitate efforts in identifying risk, and ultimately, preventing suicide.

Keywords: Adolescent, suicide, natural language processing (NLP), smartphone, passive sensing

Disclosure: Nothing to disclose.

P903. Association of Non-Suicidal Self-Injurious Behaviors With Suicide Attempt: Findings From the Texas Youth Depression and Suicide Research Network (TX-YDSRN)

Manish Jha, Abu Minhajuddin, Holli Slater, Lynnel Goodman, Sarah Wakefield, Madhukar Trivedi*

The University of Texas Southwestern Medical Center, Dallas, Texas, United States

Background: Population-based studies suggest high prevalence rates of suicidal ideation (SI) and suicidal behaviors (SB) in youth and young adults in the United States. There is an urgent need to identify clinical features that predict the occurrence of suicide-related outcomes in youth who are receiving clinical care for their depression and/or SI/SB. In this report based on six-month follow-up data from youth receiving care in community, we evaluated the role of non-suicidal self-injurious (NSSI) behaviors in predicting suicide attempts.

Methods: Youth with suicide-attempt data available for six months in the ongoing Texas Youth Depression and Suicide Research Network (TX-YDSRN) were included (N = 1021). The Concise Health Risk Tracking Scale Behavioral Module was used to determine the presence of lifetime and past week SI, suicide attempt, and NSSI. Based on clinician-evaluation, youth were grouped as those with no lifetime NSSI (no NSSI), presence of lifetime NSSI but no past-week NSSI (lifetime NSSI) and presence of NSSI in the past week (past-week NSSI). Logistic regression analyses were used to evaluate if these NSSI-based groups predicted subsequent suicide attempts even after controlling for select baseline clinical and sociodemographic variables.

Results: Of the 1021 individuals, 331 (32.4%), 583 (57.1%) and 103 (10.1%) were in no NSSI, lifetime NSSI and past-week NSSI groups, respectively. Over six months of follow-up, suicide attempt rates were 0.9% (3/331), 6.2% (36/583) and 23.3% (24/103) in no NSSI, lifetime NSSI and past-week NSSI groups, respectively. Odds ratio (95% confidence limit) of suicide attempts in lifetime NSSI and past-week NSSI groups as compared to no NSSI group were 3.69 (1.09, 12.54) and 11.53 (3.09, 43.0), respectively even after controlling for severity of overall depressive and anxiety symptoms, including active suicidal ideation in the past-week at baseline.

Conclusions: In this large cohort of youth receiving clinical care, NSSI behaviors, especially those in the past week, were strongly associated with occurrence of a suicide attempt over a six-month follow-up period.

Keywords: Children and Adolescents, Adolescent Depression, Suicide risk factors, self-harm, Suicide prediction

Disclosure: Acadia Pharmaceuticals, Alkermes Inc., Alto Neuroscience Inc., Axsome Therapeutics, BasePoint Health management LLC, Biogen MA Inc., Cerebral Inc., Circular Genomics Inc., Compass Pathfinder Limited, Daiichi Sankyo Inc., GH Research, GreenLight VitalSign6 Inc., Heading Health, Janssen Pharmaceutical, Legion Health, Merck Sharp and Dohme Corp., Mind Medicine Inc., Myriad Neuroscience, Naki Health Ltd, Neurocrine Biosciences Inc., Noema Pharma AG, Orexo US Inc., Otsuka America Pharmaceutical Inc., Otsuka Europe LTD, Otsuka Pharmaceutical Development and Commercialization Inc., Praxis Precision Medicines Inc, PureTech LYT Inc., Relmada Therapeutics Inc., Sage Therapeutics, Signant Health, Sparian Biosciences, Titan Pharmaceuticals, Takeda Pharmaceuticals Inc.: Consultant (Self).,

P904. Social Neural Circuitry and Emotional Responses to Social Threat Impact Suicidal Thoughts in High-Risk Young Adults

Kristen Eckstrand*, Michele Bertocci, Helmet Karim, Tien Hong Stanley Seah, Carly Lenniger, Chloe Horter, Peter Franzen, Dana Rofey, Jennifer Silk, Erika Forbes

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Young adults are contemplating suicide at crisis levels, particularly among sexual and gender minorities (SGM; i.e., people with non-heterosexual orientations and/or have a gender identity that is incongruent with assigned sex at birth). The Interpersonal Theory of Suicide posits that distal factors, including activity in social brain circuitry, interact with proximal factors such as social threat, to influence suicidal ideation (SI). There is a dire need for biological psychiatry research examining SI in SGM populations. We took a social neuroscience, ecologically informed approach to disparities in SI by examining the combination of neural sensitivity and social threat experiences to SI in a longitudinal study of chronically suicidal, predominately SGM young adults.

Methods: 86 young adults (23.5 ± 3.1yrs; 79% SGM; 67% female; 69% white) reported SI and social threat events via ecological momentary assessment (EMA) 3 times/day over 7 days/month for 6 months. A subset of 39 young adults also completed fMRI during a personalized, social-reward task involving video stimuli from discussions about positive or neutral experiences with a close friend or from an age/race/gender-matched unfamiliar peer (stranger). SI was measured using 5 dichotomous items related to suicidal ideation, intent, and plan, and a single continuous item assessing the intensity of suicidal thoughts on a 0-100 scale. Participants reporting social threat experiences responded whether they had experienced any of four emotional responses to the reported social threat event (i.e., sad, worried, mad, stressed). Mean SI intensity and proportion of negative emotional responses to social threat were summarized across the study. fMRI data were preprocessed in fMRIprep and analyzed in Statistical nonParametric Mapping. Activation in social neural circuitry to Friend Positive > Friend Neutral (FP > FN) and Friend Neutral > Stranger Neutral (FN > SN) was extracted (pFWE < 0.05). Two machine learning models predicting mean SI intensity, one for each contrast, were used for feature selection of neural activation, emotional responses to social threat, and their interaction. Significant social neural circuitry activity and social threat experiences were entered into cross-validated elastic net regression models predicting ST intensity. GLMNET, for variable selection, and post-hoc tests were completed in R. Models were cross-validated with k = 5 folds. Exploratory bivariate associations between ST intensity and emotional responses to social threat were completed in R.

Results: 60.5% (N = 52) of the sample reported 471 instances of SI (M = 9.06, SD = 10.81) ranging in intensity from 0-100 (M = 41.30, SD = 21.80) during the study. 4 regions were active to FP > FN including left caudate, calcarine fissure, lingual gyrus, and middle temporal gyrus. The elastic net regression model predicting SI (R2 = 0.29, p = 0.04) revealed higher SI intensity was predicted by proportion of sad and worry responses to stress, and the combination of higher left caudate activity and higher proportion of negative emotional responses to social threat. 62 regions were active to FN > SN. No neural activity during FN > SN or their interaction with emotional responses to threat predicted SI intensity.

Conclusions: Consistent with conceptual models of SGM health disparities, negative emotional reactions to social threat impact SI in natural settings. However, in this study we specifically demonstrate that the combination of elevated neural sensitivity in reward/salience circuits combined with negative emotional responses to social threat is associated with elevated SI intensity. These results provide preliminary support for the role of social neural circuitry activity, particularly in reward/salience regions, as a distal factor contributing to the emergence of suicidal thoughts in the setting of social stress. These results can inform strategies for detecting and addressing suicide risk in the setting of social stress, particularly among vulnerable populations including SGM. This provides novel insights into the neural mechanisms of suicidality in a vulnerable population for whom social threat is ubiquitous.

Keywords: suicide, young adults, Social Threat, Sexual and gender minority, EMA

Disclosure: Nothing to disclose.

P905. Higher Cardiovascular Risk is Associated With Suicide Attempts Among Youth With Mood Disorders

Simin Jin, Mikaela Dimick, Kody Kennedy, Eric Youngstrom, Benjamin Goldstein*

Centre for Addiction and Mental Health, Toronto, Canada

Background: Self-harm (i.e., suicide attempts (SA), non-suicidal self-injury (NSSI)) can significantly predict future suicide attempts among youth. Cardiovascular disease (CVD) incidence and CVD risk factors are higher in individuals with mood disorders and are associated with SA, NSSI, or suicidal ideation in adults with mood disorders. As the cardiovascular-suicide risk association remains under-studied among youth, the present study examined the association of cardiovascular risk with self-harm in youth with mood disorders.

Methods: Participants were 587 youth (of any sex/gender), ages 13 to 21 years, including youth with mood disorders (i.e., bipolar disorder, major depressive disorder) and a history of SA (n = 84), NSSI (n = 146), or no self-harm (n = 137), and healthy controls (n = 220). A composite cardiovascular risk score included blood pressure, body mass index (BMI), and lifetime cigarette smoking. Multinomial logistic regression investigated the association of composite cardiovascular risk scores and individual cardiovascular risk factors with self-harm. Receiver Operating Characteristic analysis examined the ability of composite cardiovascular risk scores to clinically classify self-harm.

Results: Higher composite cardiovascular risk scores were associated with a history of SA in youth with mood disorders relative to youth with mood disorders but without prior self-harm (OR = 1.37, 95% CI = [1.11, 1.68], p = 0.003). Youth with higher blood pressure, BMI, and smoking risk categories were more likely to have SA. Composite or individual cardiovascular risk was not associated with NSSI. Composite cardiovascular risk scores significantly classified prior SA (Area Under the Curve = 0.65 ± 0.04, 95% CI = [0.57, 0.73], p < 0.001), but the accuracy does not meet the adequate cut-off for clinical utility.

Conclusions: Higher cardiovascular risk is associated with greater odds of having SA in youth with mood disorders. Future longitudinal studies are warranted to examine the predictive effects of cardiovascular risk on self-harm.

Keywords: cardiovascular, Suicide attempt, Suicide risk factors, youth, mood disorders

Disclosure: Nothing to disclose.

P906. Task fMRI Correlates of Irritability and Suicide Risk in People With Depression: A Pilot Study

Jessica Sah, Oghogho Abigail Iyekekpolor, Sakina Rizvi, Sidney Kennedy, Katharine Dunlop*

University of Toronto, Toronto, Canada

Background: Major depressive disorder (MDD) is the most common mental disorder among suicide deaths, the ninth leading cause of death in Canada. Currently, suicidal patients respond less favourably to antidepressant treatment than those without, stressing the urgent need to understand behavioural and biological factors underlying suicidality in the hopes of developing tailored treatments. We propose that one way to better understand these factors is to elucidate the neurobiological link between suicide and comorbid symptoms. Irritability correlates with increased suicidality in MDD, independent of other symptoms. Further, neuroimaging studies assessing irritability or suicidality separately suggest a common fronto-limbic circuit. We aim to identify differences in the behavioural and neural correlates of irritability in individuals with MDD with suicidal ideation and a recent suicide attempt, relative to individuals with MDD with suicidal ideation and no history of attempt, and healthy controls.

Methods: We recruited male and female participants aged 18-65 from the three experimental groups: MDD participants with suicidal ideation and no suicide attempt history (n = 30; MDD-SA), those with suicidal ideation and recent attempt history (n = 30; MDD + SA), and healthy controls (n = 30; HC). Each participant completed a research interview, questionnaires, and two validated paradigms with functional MRI: one will examine negative affect and irritability following social exclusion (Cyberball Task), and the other will enable participants to retaliate against excluder players (Dictator Game). Linear mixed-effects models identified differences in irritability, task behaviour, and brain activity by group, as well as interactions between irritability and group

Results: 26 participants (13 Control, 13 MDD [7 MDD + SA, 6 MDD-SA]) have completed the study to date (mean ± standard deviation age = 33.1 ± 15.0 years, 20 female). Among HC, social exclusion was associated increased irritable mood on a visual analogue scale (VAS; W = 15.00, p = 0.05). Both self-reported irritability on the Brief Irritability Test (BITe) and an increase in irritability during the Cyberball Task correlated with punishment of excluders relative to includer Cyberball players (BITe Spearman’s ρ = 0.68 p = 0.03; VAS Change ρ = 0.66 p = 0.04). Due to the expected small sample at this stage, we collapsed the MDD-SA and MDD + SA groups. Relative to HC, all MDD participants had significantly greater irritability at baseline (BITe U = 3.00, p = 0.001), after inclusion (U = 4.00, p < 0.001) and exclusion (U = 8.00, p = 0.004). Irritability worsened from inclusion to exclusion at trend (W = 25.50, p = 0.06). Those with a higher capacity for suicide punished excluders at higher rates (ρ = 0.76, p = 0.04). We used three two-way continuous covariate interactions (p < 0.05, two-tailed, cluster size > 25 voxels) to test whether the slope of the regression line correlating baseline irritability with task activity differed by diagnosis. During social exclusion on the Cyberball Task, irritability in MDD was associated with increased activity in the ventral/rostral anterior cingulate cortex and dorsomedial prefrontal cortex. During retaliation against excluders, irritability in MDD was negatively correlated with activity in the frontal pole, rostral anterior cingulate cortex, and dorsomedial prefrontal cortex. During the forgiveness of excluders, irritability was positively correlated with task activity in the dorsolateral prefrontal cortex.

Conclusions: Our preliminary results suggest that fronto-limbic hyperactivity during exclusion correlates with irritability in MDD, consistent with previous literature. Highly irritable individuals retaliated more frequently against Cyberball excluders, and in the MDD group, irritability was associated with fronto-limbic hypoactivity during retaliation. Forgiving excluders by opting for a fair coin distribution recruited dorsolateral prefrontal regions implicated with task switching and cognitive control, which correlated increased irritability in the MDD group. The results suggest may indicate a compensatory recruitment of top-down cognitive control and theory of mind regions to regulate behavior. The results of the study may identify brain regions that influence irritability and suicide risk, thereby providing empirically validated targets for transcranial magnetic stimulation interventions, like inhibitory targets related to exaggerated irritability or excitatory targets that help regulate behavioral outbursts.

Keywords: suicide, Irritability/Aggression, fMRI, Social Behavior, Depression

Disclosure: Nothing to disclose.

P907. A Study of Repeated Suicide Attempts in Persons Hospitalized for Depression: The Role of Immune Factors

Faith Dickerson*, Robert Yolken

Sheppard Pratt Health System, Baltimore, Maryland, United States

Background: Suicide is a major cause of death worldwide and is highly prevalent in persons with major depressive disorder (MDD). Individuals with MDD hospitalized for a suicide attempt are at high risk for a repeat suicide attempt in the 6-month period after hospital discharge. Previous studies have identified immune alterations in MDD, but not the prospective association between immunological abnormalities and subsequent suicide behavior.

Methods: The study population consisted of adults with major depressive disorder hospitalized for an acute depressive episode and a recent suicide attempt based on the Columbia Suicide Severity Rating Scale (C-SSRS). Participants were assessed for co-occurring immunological disorders and on clinical measures. Participants had a blood sample drawn from which were measured cytokines, antibodies, and other markers of inflammation. Following hospital discharge, participants were assessed monthly for six months. Cox proportional hazard models examined the relationships between baseline variables and a repeat suicide attempt.

Results: A total of 69 persons were enrolled with an average age of 29.8 years (s.d.11.5); 36 (52%) were White; 44 (64%) female sex; 41 (59%) heterosexual orientation; 6 (9%) transgender or non-binary gender. A total of 62 participants had at least one post-hospital study visit and of these 15 (24%) made a suicide attempt in the follow-up period. These individuals, compared with those who did not make a repeat suicide attempt, had a significant alteration in a combined serological immune marker consisting of the cytokines, IL-1β, TNF-α, and IFN‐γ, as well as IgG antibodies to Epstein Barr Virus (HR = 8.03, 95% CI 1.73, 37.08, p = 0.008). A diagnosis of asthma was also associated with a repeated suicide attempt (HR = 3.10, 95% CI 1.10, 8.79, p = 0.033). The occurrence of a suicide attempt in the follow-up period was also correlated with several clinical measures obtained at baseline. These included the number of psychiatric hospitalizations in the previous two years (coefficient =.3238, 95% CI.168, 0.480, p < .001); the number of lifetime suicide attempts (coefficient = 0.21, 95% CI 0.07, 0.36, p = 0.004); score on the Suicidal Ideation Scale (coefficient = 0.0420, 95% CI.0080, 0.0760, p = 0.015) and on the Social Readjustment Rating Scale (SRRS) (coefficient = 0.0010, 95% CI 0.00004, 0.0020, p = 0.041).

Conclusions: The finding that an active immunological disorder and increased levels of immune activation are predictive factors for a repeated suicide attempt should aid in the identification of individuals who are at particularly high risk. This approach could provide for more personalized methods for the assessment of suicide risk in individuals with serious mental illnesses and the rationale for the study of immune-based interventions.

Keywords: Suicide attempt, Immune Biomarkers, Major Depressive Disorder, Suicide risk factors

Disclosure: Yes, I (or my spouse/partner) do have a financial relationship to disclose.

Financial Relationships Details Boehringer-Ingelheim, Contracted Research, Self, Click Therapeutics, Contracted Research, Self

P908. Exploratory in Vivo Imaging of Blood-Brain Barrier Permeability in Relation to Suicide Risk

Mina Rizk*, Laurel Morris, Flurin Cathomas, Sara Costi, Audrey Evers, Emma Mayer, Sarah Boukezzi, Kenny Chan, Lyonna Parise, Xiang Xu, Cliff Lun, Cheuk Ying Tang, Scott Russo, James Murrough

Icahn School of Medicine At Mount Sinai, New York, New York, United States

Background: The blood-brain barrier (BBB) plays a crucial role in maintaining brain function by shielding against peripheral immune events. Elevated peripheral pro-inflammatory markers are associated with increased suicide risk. Our group’s prior work demonstrated that chronic stress leads to loss of claudin-5, a key tight junction protein in BBB endothelial cells, resulting in cytokine entry, neuroinflammation, and depressive-like behaviors. Recent post-mortem studies have found disrupted BBB in the prefrontal cortex (PFC) and nucleus accumbens (NAc) of suicide decedents. The PFC and NAc are integral to the decision-making and reward-processing domains, respectively, which are implicated in suicide risk. An in vivo examination of BBB integrity in suicidal individuals has not been reported. This exploratory study addresses this knowledge gap by utilizing dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to assess BBB permeability as it relates to peripheral immune activity in suicidal individuals.

Methods: Thirty-seven participants (18-55 years) meeting DSM-5 criteria for major depressive disorder (MDD) were enrolled. Lifetime history of suicide attempts was assessed using the Columbia Suicide Severity Rating Scale (C-SSRS), and depression severity was evaluated using Quick Inventory of Depressive Symptomatology. Exclusions included inflammatory/autoimmune disorders, anti-inflammatory medication use, body mass index (BMI) ≥ 35, and unstable medical conditions. DCE-MRI acquisition protocol with dual-time resolution quantified BBB leakage. An intravenous gadolinium-based contrast agent (0.05 mmol/kg) was administered, followed by coronal T1-weighted DCE-MRI scans, acquired for 16 minutes with 15.4 seconds temporal resolution per image. This protocol captured the first pass of the bolus with high temporal resolution, followed by slower acquisition with higher spatial resolution for interstitial uptake. Arterial input function was computed using an automated algorithm to extract a vascular mask from DCE scans. A two-compartment pharmacokinetic model (Patlak approach) was applied per voxel to produce whole-brain and region-specific estimates of BBB leakage (mean permeability index, Ki). Regions-of-interest (ROIs) were selected based on post-mortem findings and included those implicated in decision-making (dorsolateral PFC, dlPFC; dorsal anterior cingulate cortex, dACC) and reward circuitry (habenula; NAc; dorsomedial PFC, dmPFC; ventromedial PFC, vmPFC) and averaged bilaterally. Circulating immune markers were assessed by the Olink multiplex assay – Inflammatory panel using blood samples collected while participants were fasting, abstinent from smoking for 8 hours and from alcohol for 24 hours. Ki values and immune markers were compared between participants with and without a history of suicide attempts while controlling for age, sex and BMI. Due to the exploratory nature of this analysis, p-values were not corrected for multiple comparisons.

Results: Participants with a history of suicide attempt (N = 7, aged 36.6±11.9, F/M = 5/2) compared to those with MDD but no suicide attempt history (N = 30, aged 32.9±9.0, F/M = 15/15) did not differ in depression severity. Whole-brain Ki values were significantly higher in MDD participants with a history of suicide attempts, compared to MDD non-attempters (p = 0.038). ROIs analyses revealed that compared with MDD non-attempters, MDD suicide attempters had higher Ki in habenula (p = 0.031), dACC (p = 0.021) and dmPFC (p = 0.022), and dlPFC (p = 0.012). No difference between the two groups were found in NAc (p = 0.119) or vmPFC (p = 0.475). Compared with non-attempters, MDD attempters had higher pro-inflammatory cytokines (e.g., Cystatin D, IL18, CXCL10; all ps < 0.05). Further, brain-wide and regional BBB permeability correlated positively with several pro-inflammatory markers (e.g., IL7, FGF23, CXCL11; all ps < 0.05).

Conclusions: This exploratory study, despite its limitations due to small sample size and exploratory nature, sheds light on the association between brain-wide and regional BBB leakage, peripheral inflammation, and suicide risk, underscoring the potential of BBB dysfunction as a biomarker for suicidal behavior. Our findings also demonstrate the effectiveness of non-invasive DCE-MRI in detecting BBB abnormalities associated with suicide risk. The study results have important implications for developing innovative diagnostic and therapeutic strategies targeting BBB dysfunction in individuals at risk for suicide, potentially enabling early intervention and prevention.

Keywords: Blood-Brain Barrier, Major Depressive Disorder (MDD), Suicide attempt, Inflammation, DCE MRI

Disclosure: Nothing to disclose.

P909. Emotional Face Processing Deficits Linked to Cortisol Dysregulation in Individuals With Suicide Risk

Steven Lamontagne*, Jessica Gilbert, Yoojin Lee, Carlos Zarate, Elizabeth Ballard

National Institute of Mental Health, Bethesda, Maryland, United States

Background: Background. Suicide is a leading cause of death, but its neural mechanisms remain poorly understood. Interpersonal deficits, including an impaired ability to interpret emotional social cues, are a major risk factor for a suicide crisis. Cortisol dysregulation, often reported in individuals with suicidal thoughts and behaviors (STBs), can affect perceptions and responses to social cues. Cortisol might therefore impact neural underpinnings of emotional face processing, though this relationship has not yet been studied in the context of suicide. The current study used magnetoencephalography (MEG) to evaluate relationships between tonic cortisol and gamma oscillatory responses (a proxy measure of excitation-inhibition balance) to emotional face stimuli in individuals with varying levels of suicide risk.

Methods: Methods. One hundred eighteen participants (n = 71 female; Mage=40.71, range 19-70) were assigned to one of the following four suicide risk groups: Crisis: those with a suicide attempt and/or ideation with intent in the past two weeks (n = 14), Past Attempt: those with a history of attempt, but no suicidal behavior or ideation with intent in the past year (n = 39), Clinical Control: those with anxiety or mood symptoms, but no suicide history (n = 35), and Healthy Control: those without psychiatric or suicide history (n = 30). Henceforth, the Crisis and Past Attempt groups are collectively referred to as “at-risk”; the Clinical and Healthy Control groups are collectively referred to as “controls”. 24-h urinary free cortisol was analyzed using tandem liquid chromatography and mass spectrometry. A CTF 275-channel whole-head MEG scanner was used to examine electrophysiological correlates of emotional face processing. During MEG scanning, participants completed an emotional evaluation task in which they indicated the sex (male or female) (implicit condition) or emotional valence (positive or negative) (explicit condition) of facial stimuli. MEG data were source-localized using a linearly-constrained minimum-variance beamformer in the gamma (30-58 Hz) frequency from -100 to 1000 ms peristimulus time, and a linear mixed-effects model implemented in AFNI was used to evaluate differences in gamma power between the face stimuli. Hierarchical linear regressions were used to evaluate moderating effects of cortisol on task and MEG data. Dynamic causal modeling (DCM) was used to examine effective connectivity between regions of interest implicated in emotional face processing, namely the early visual cortex (EV), amygdala (AMY), and superior frontal gyrus (SFr). DCMs were fitted using a wide frequency range from 1-40 Hz within 300 ms post-stimulus onset.

Results: Results. Adjusting for biological sex and age, 24-h urinary cortisol levels were blunted in the Past Attempt (ps < 0.05) but not the Crisis (ps > 0.08) group compared to controls. This implicates cortisol dysregulation in those with distant rather than active STBs, which is corroborated by an inverse correlation between cortisol and past (but not current) suicidal ideation across the full sample (p < 0.01). Adjusting for cortisol levels, task-based reaction times (RTs) in the explicit conditions were faster for positive than negative face stimuli in all groups (ps < 0.01) except for the Crisis group, where RTs were similar across stimuli. This suggests that cortisol interacts with acute suicide risk to disrupt emotion recognition and processing. Electrophysiologically, main effects of Group in the EV and SFr (ps < 0.01, cluster corrected) revealed diverging activation patterns by suicide risk. Across face conditions, the at-risk groups showed higher gamma power than the control groups in the EV but lower gamma power in the SFr. Interestingly, cortisol moderated EV gamma power in the Crisis group: adjusting for sex, age, and depression, higher EV gamma power was related to lower cortisol levels. Finally, the DCM findings revealed that, compared to controls, the at-risk groups showed increased effective connectivity from the EV to the AMY across face conditions.

Conclusions: Conclusion. These findings suggest that those with active and past STBs have heightened initial low-level processing but disengaged higher-order processing of emotional faces, with a moderating effect of tonic cortisol in those with recent suicide crisis. At-risk individuals also showed greater feedforward prediction errors propagating from the EV to AMY, further implicating inadequate top-down control of early sensory input from the EV. Disrupted connectivity parameters between these regions might be linked to HPA-axis dysregulation, which could underlie future efforts to identify biomarkers of suicide risk. Collectively, these findings point to emotional processing patterns, as well as their modulation by cortisol, as biomarkers of suicide and could inform targeted interventions for risk mitigation.

Keywords: Suicide Mechanisms, Cortisol, face emotion processing, magnetoencephalography

Disclosure: Nothing to disclose.

P910. Blood MicroRNA as Human Biomarker in Suicide

Thomas Kosten*, Alan C. Swann, Cristian Coarfa, Amrit Koirala, David Nielsen, Preethi Gunaratne

Baylor College of Medicine and Michael E. DeBakey VA Medical Center, Houston, Texas, United States

Background: Suicide’s multiple determinants include genetics and psychiatric and behavioral disorders, and recent reviews have cataloged potential blood biomarkers associated with suicide in major depressive disorder (Mamdani F, et al. Translational Psychiatry (2022) 12:159). In blood, they identified 14 genes which significantly differentiated suicides. The top six genes differentially expressed in blood were: PER3, MTPAP, SLC25A26, CD19, SOX9, and GAR1, and two genes CD19 and TERF1 were increased in blood but decreased in prefrontal cortex. The range of genes clearly reaches well beyond the usual neurotransmitters associated with suicide such serotonergic, noradrenergic and dopaminergic polymorphisms, and most polymorphisms reflect long term risk, not biological markers to predict acute risk or definitive recovery from suicidal ideation (SI) and attempts. We tested plasma microRNAs (miRNA), which enter the blood from brain, as possible epigenetic markers for SI recovery using genome-wide miRNA expression, and we assessed their mRNA-targets using functional annotation analyses.

Methods: We assessed and collected blood samples from 37 SI and 27 non-SI inpatients at admission and 4-6 weeks later in recovery while still inpatients at a private psychiatric hospital. We collected weekly symptom assessments including Columbia-Suicide Severity Rating Scores (C-SSRS). We processed the blood (Qiagen, Germantown, MD) and stored the plasma at -80o C until analysis. We extracted total RNA using the Qiagen miRNeasy serum/plasma kit, and constructed sequencing libraries using Qiagen Qiaseq miRNA library kit and ensured 180 nt miRNA using an Agilent 4200 tape station. Using an Illumina Nextseq500 instrument we sequenced the miRNA.

We used the global miRNA expression profiles to create a list of expressed miRNAs. We removed adapter, primer, and poly-A sequences and discarded reads < 15 nt. We identified known miRNAs using the ENCODE pipeline, performing mapping with STAR and miRNA quantification using miRbase v21 and miRdeep2 v2.0.0.8. We constructed a miRNA expression matrix using the number of reads mapped to each miRNA and determined differentially expressed (DE) miRNAs using the R package DESeq2 (v1.26.0). We considered differences in miRNA expression between SI and non-SI at admission and when SI resolved for SI group significant at Benjamini and Hochberg corrected p < 0.05 and |log2 Fold Change | > 0.25.

We correlated plasma miRNA with C-SSRS, depression and anxiety in these inpatients. The two groups had no differences in sex (54% vs 43% males) and white ethnicity (83% vs 88%), but the SI were older [SD] (29 [12] vs 23 [9] years; P < 0.05). Psychiatric disorder rates did not differ for anxiety (54% vs 70%), depressive (71% vs 76%) and tobacco use (29% vs 46%) disorders, but SI had higher rates of past alcohol use disorder (11% vs 43%; P < 0.01). No patients had withdrawal symptoms nor required alcohol detoxification. The SI group’s C-SSRS score was 14.7 (SD 3.5) (range 9 to 23) with 9 indicating active suicidal ideation with any method. The admission mean PHQ-9 (depression) score was 16.5 with no scores below 10, which represented severe depression. As expected, the 37 SI resolved patients showed reductions in their baseline anxiety of 12.8 (P < 0.001) and depression of 17.2 (p < 0.001), but rates of taking five major medication classes as a group showed no significant changes between admission and discharge (SSRI/SNRI, antipsychotics, anti-anxiety, mood stabilizer, attention deficit).

We also collected an independent replication sample of 30 demographically different patients from Houston County Hospital to compare our top DE miRNA for suicide recovery between this population and our private hospital generated sample.

Results: The 37 SI patients showed down-regulation of about 2% of 2600 miRNAs with the most downregulated miRNA showing a 39-fold higher level at admission compared to after recovery from SI; this miRNA showed no change in non-SI patients. We validated the largest differentially expressed (DE) miRNAs by qRT-PCR, and luciferase assays confirmed miRNA functional activity for miR-10b-5p in the SDC1 gene. Depression and anxiety improved and correlated with changes in miR. The serum protein Clusterin, which the distinguishing miRNA also targeted, significantly increased by 31% in the plasma of the 37 recovered patients. We found 1% of 22,000 mRNA (genes) were targeted by at least 3 DE miRNA. Tissue enrichment using the Human Gene Atlas found the parietal lobe and two parts of cerebellum among the top 20 targets of our DE miRNA. One-third of our top DE miRNA for suicide recovery also replicated in the 30 demographically different patients from Houston County Hospital.

Conclusions: These plasma miRNA epigenetic changes that can each regulate 100 or more genes appear to reflect biomarkers and critical pathways in brain for recovery from suicidality and are not reflecting non-specific changes in miRNA also found among non-SI inpatients. These same biomarkers may provide risk factors for assessing suicide risk among patients presenting to emergency departments with severe physical injuries, but not reporting suicidality.

Keywords: Suicide risk factors, human epigenetics/microRNA, blood biomarker, Epigenetic biomarkers, Suicide Mechanisms

Disclosure: Nothing to disclose.

P911. Identifying Suicide Ideators via Implicit Associations with Death and Entrapment and Corresponding Electrophysiological Features

Michael Bronstein*, Sean Mullen, Blair Brown, Miriam Freedman, Benito Garcia, Melanie Goodman Keiser, Bing Brunton, Alik Widge

University of Minnesota, Minneapolis, Minnesota, United States

Background: The Brief Death Implicit Association Test (BDIAT) represents a significant step toward suicide assessments that are less dependent on patient willingness and capacity for explicit self-report. However, this task alone lacks sufficient sensitivity and specificity to be clinically actionable – perhaps because it neglects essential motivations for suicide. According to the Integrated Motivational-Volitional Model, experiences of defeat/entrapment can provide impetus for suicide ideation and behavior. With this in mind, we developed the Entrapment Go-No-Go Association Test (E-GNAT) – a novel task that captures implicit associations with defeat/entrapment. We hypothesized that the E-GNAT and BDIAT might together provide for better identification of suicide ideators.

Methods: Male and female individuals with severe, treatment-resistant depression and participants from the general population completed the Columbia Suicide Severity Rating Scale (C-SSRS) and one week of 3x/day Ecological Momentary Assessment (EMA) examining suicide ideation. They then completed the E-GNAT and BDIAT with simultaneous electrophysiological recording. We constructed shallow and deep machine learning models to classify participants into ideators and non-ideators using task-related behavior and neural activity.

Results: Analysis of E-GNAT reaction times indicated that suicide ideators, on average, had stronger implicit death-trapped associations (M = 0.91) than non-ideators (M = 0.55). BDIAT reaction times indicated that ideators had stronger implicit death-me associations (M = -0.11) than non-ideators (M = -0.34). We also present results describing our machine learning models’ ability to discriminate between suicide ideators and non-ideators based on the BDIAT alone, the E-GNAT alone, and the combination of both tasks.

Conclusions: The E-GNAT shows significant promise as an additional, theory-informed assessment of suicidality that does not rely on explicit self-report. While suicide-related assessments examining implicit attitudes have generally focused on behavior, electrophysiological responses to these tasks carry additional information about suicidality.

Keywords: Suicide Assessment, Implicit association test, suicide

Disclosure: Nothing to disclose.

P912. Lithium May Reduce Recurrent Suicide Risk by Reducing Impulsive Decision Making: The Case for Lithium in Outpatient Management of Medically Severe Suicide Attempt Survivors

Nicholas Murphy*, Grace Pham, Andreas Weyland, Julia Engelhardt, George Kypriotakis, Ynhi Thomas, Thomas Kosten, Nidal Moukaddam, Sanjay Mathew, Alan C. Swann

Baylor College of Medicine, Houston, Texas, United States

Background: Most suicides are first attempts; however, extant work in behavioral modeling implies that the transition from suicidal ideation to behavior is underscored by a state of high central arousal leading to impulsive behavior. The accumulation of psychological stress lowers the tolerance for physiological arousal, and impulsivity is amplified by this lower tolerance for arousal. Impulsivity then impairs future decision making, leading to further accumulation of stress. Studies attempting to distinguish suicide attempters from ideators have found that impulsivity is consistently associated with attempts across the spectrum of suicidal behavior from self-harm to lethal suicidal behavior. To date the only pharmaceutical that has successfully reduced suicidal behavior is lithium. From studies of conduct disorder, we also know that lithium acts on the central mechanisms of arousal, by reducing norepinephrine and stress-mediated phosphatidylinositol turnover. the purpose was to determine whether lithium affects computational and behavioral processes predisposing to suicidal behavior in people who are at high-risk for suicide.

Methods: Survival of a medically severe suicide attempt (MSSA) was used as a proxy for high suicide risk. This is based on reports of recurrent attempt risk, and increased all-cause mortality, associated with the 10 year period post-attempt. Transdiagnostic MSSA patients were identified by our team based on a recent admission to the Ben Taub General Hospital emergency center or other relevant setting. Study measures were scheduled 6-9 months after MSSA to assure subsidence of nonspecific acute effects of the attempt and time to complete study procedures within 1 year of MSSA. 15 patients (37.5% male, 40.13 ± 13.66 years) received lithium carbonate and matching pill placebo separately in a randomized double-blind crossover design administered six weeks apart. To test the effect of lithium on measures of impulsivity and arousal, participants completed the Immediate Memory Task (IMT), Internal State Scale (ISS) Activation sub-measure, and the Time Perception Task (TPT). We conducted separate analyses for each variable using repeated measures analysis of covariance. Planned contrast effect sizes were estimated using the repeated measures variation of Cohen’s D to account for the correlation between variables. Missing data was assumed to be missing at random and was imputed using the mean within-groups. Violations of sphericity were corrected using the Greenhouse-Geisser method.

Results: Participants (37.5% male) had a mean age of 40.13 years ( ± 13.66 years) and average weight of 198lbs (±63.78lbs). Lithium dosing was associated with increased IMT response latency (p = 0.017, pη2 = 0.23), and decision bias (p = 0.048, pη2 = 0.21). Time perception had a significant main effect of condition (F = 6.41, p = 0.005, pη2 = 0.33) but did not demonstrate significant contrast results. ISS Activation was not significantly associated with the effect of condition (F = 0.23, p = 0.81, pη2 = 0.02).

Conclusions: Our results suggest that lithium may reduce risk in MSSA survivors by increasing the response latency and increasing conservative response bias during decision making. These findings support the need to develop treatments that target systems related to response preparation and impulsivity.

Lithium has a broad mechanism of action that covers circuit properties described in the Research

Domain Criteria description of suicide. However, it is does not target a specific system and requires close clinical monitoring of renal function when prescribed long-term. By studying the effects of lithium on factors predisposing to suicide it will be possible to establish a more firm pharmaceutical management strategy for patients considered high risk. Further, it will help to increase knowledge of clinical targets such that more precisely tailored drugs can be engineered.

Keywords: Suicide Mechanisms, Lithium, impulsivity, clinical psychopharmacology, emergency medicine

Disclosure: Nothing to disclose.

P913. Funding Opportunities Psychiatric Disorders: Molecular and Cellular Neuroscience Research Branch NIMH

Enrique Michelotti*, Jamie Driscoll, Cathleen Hsu, Sofiya Hupalo, Laurie Nadler, Leonardo Tonelli, Yael Mandelblat-Cerf

National Institute of Mental Health, Rockville, Maryland, United States

Background: The NIMH Division of Neuroscience and Basic Behavioral Science (DNBBS) supports research programs in basic neuroscience, genetics, basic behavioral science, research training, resource development, technology development, drug discovery, and research dissemination. The DNBBS Molecular and Cellular Neuroscience Research (MC) Branch focuses on the fundamental understanding of the structure and function of cells and circuits relevant to cognitive, affective, and social domains as well as biomarkers and therapeutic development. Specific research programs within the Branch include neuronal and glial signaling, synaptic plasticity, neuropharmacology, neuroendocrinology, and neuroimmunology.

Methods: The MC Branch provides funding for the following areas of interest:

• Fundamental mechanisms of signal transduction, synaptic transmission, and synaptic plasticity.

• Neuroimmune mechanisms linking inflammatory processes with cells and circuits’ function.

• The transient and sustained neurobiological actions of psychedelics, utilizing a reverse translational approach to test mechanistic hypotheses.

• The biology underlying high-confidence genetic and environmental risk factors for mental health illnesses.

• Pre-clinical research aimed at developing novel approaches for somatic cell gene therapies to treat neurodevelopmental disorders.

• Identifying, validating, and testing novel targets for therapeutic intervention and preclinical biomarkers for the treatment of mental illnesses.

• Development, optimization, and evaluation of novel in vivo measures as potential assays in pre-clinical screening of therapeutic candidates.

• Elucidating the molecular and cellular mechanisms responsible for the differential function, dysfunctional brain processes, and effects of potential therapeutics, by sex, age, and environment.

• Development and validation of screening assays for the discovery of validated hits for drug discovery/development for the treatment of mental illness.

• Development and evaluation of PET and SPECT novel radioligands for imaging in the human brain

• Discovery of in vivo chemical probes for novel brain targets

• Development of high, medium, or low-throughput assays to assess the activity and selectivity of small molecules, biotechnology products, and biologics with potential as candidate therapeutics.

• Accelerate innovative therapies translation from discovery to early human studies.

Approaches may include in vitro and/or in vivo paradigms using vertebrate or invertebrate animal models, human subjects and human cell-based assays, and ML/AI and other computational tools that integrate complex experimental datasets and provide testable predictions.

Results: In this presentation, we illustrate the range of the MC Branch programs and funding mechanisms, grants, cooperative agreements, and contracts, related to the role of structure and function of cells in disease and therapeutic discovery and development.

Conclusions: The Branch has funded multiple successful basic biology projects, neuronal and glial signaling, synaptic plasticity, neuropharmacology, and neuroimmunology, as well as projects in the drug discovery area. Several projects advanced to clinical development.

Keywords: Molecular, Cellular, Neuroscience, Funding

Disclosure: Noting to disclose.