ACNP 64th Annual Meeting: Poster Abstracts P584-P872

January 12–15, 2026

Nassau, Bahamas

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

Only disclosures for presenting authors are listed. Underlined names 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.

P584. Connectivity- versus scalp-based targeting of accelerated TMS for depression: a randomized trial

Joseph Taylor, Marina Kare, Dania Haj-Darwish, Emma Jones, Lauren Sanderson, Sanaz Khosravani, Jessica Leach, Leanna Bomer, Nicole Chiulli, Elizabeth Steuber, Christopher Lin, William Drew, Stephan Palm, Anjali Chandra, Summer Frandsen, Anastasia Below, Tracy Barbour, Sheena Baratono, Irene Gonsalvez, Stanley Lyndon, Samuel Snider, David Silbersweig, Shan Siddiqi, Michael Fox

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

Background: Neuroimaging is widely used in psychiatric research, but there is limited evidence that it improves clinical outcomes. Recently, functional connectivity neuroimaging was used to target brain circuits with accelerated transcranial magnetic stimulation (aTMS), leading to rapid antidepressant effects and FDA clearance. However, the importance of connectivity-based targeting remains unclear.

Methods: In this triple-blind trial, we randomized adults with treatment-resistant depression (TRD) (n = 40) to connectivity- versus scalp-based targeting of aTMS (10 treatments per day for 5 days). Our primary outcome was the effect size of connectivity-based targeting using Montgomery–Åsberg Depression Rating Scale (MADRS) change one month after treatment.

Results: MADRS improved by 83% with connectivity-based targeting versus 62% with scalp-based targeting. Connectivity-based targeting had a large effect size (1.09, 95% CI 0.06–4.05), significantly improved MADRS (p = 0.01), and was reproducible within an individual (p < 0.001).

Conclusions: These results suggest that individualized connectivity-based targeting has a large effect size that may be used to improve the antidepressant effects of aTMS.

Keywords: transcranial magnetic stimulation (TMS), functional neuroimaging, Clinical trial, fMRI-guided targeting, TMS targeting.

Disclosure: Nothing to disclose.

P585. Modeling transcranial magnetic stimulation (TMS) treatment response in veterans

Yosef Berlow, Michelle Madore, F. Andrew Kozel, Noah Philip

Alpert Medical School, Brown University, Barrington, Rhode Island, United States

Background: Transcranial magnetic stimulation (TMS) is safe and effective for treatment-resistant depression and has been implemented at more than 35 VA facilities through the VA National TMS Program. In a previous study of 97 Veterans (Scientific Reports 13.1 (2023): 7138), we observed a characteristic nonlinear response pattern in which the largest improvements in depressive symptoms occurred in the first two weeks, followed by smaller, continued gains that approached a plateau. This trajectory was well modeled by an exponential decay function. The present study aimed to replicate and validate this modeling approach in a substantially larger sample of Veterans (n = 1529) receiving TMS for depression.

Methods: We analysed longitudinal symptom data from 1529 Veterans who received TMS within the VA National TMS Program. Depression symptoms were assessed using the Patient Health Questionnaire 9-item (PHQ-9) at baseline and at weekly intervals during treatment. We fit a nonlinear mixed-effects (NLME) model based on an exponential decay function to characterise symptom reduction over the treatment course and compared this with a corresponding linear mixed-effects (LME) model. Model performance was evaluated using Agog information criterion (AIC), Bayesian information criterion (BIC) and a likelihood ratio test (LRT).

Results: The mixed-effects models incorporated 8945 longitudinal PHQ-9 ratings from 1529 Veterans undergoing TMS. The cohort was mostly male (22% female) with a mean age of 51 years. Treatment response (≥ 50% PHQ-9 reduction from baseline) was achieved by 38% of the sample, and remission (PHQ-9 < 5) by 24%. Symptom reduction during TMS was well captured by the exponential NLME model, yielding significant estimates for all parameters (all p < 0.0001). On average, PHQ-9 scores declined by 7.5 points, with a time constant of 2.1 weeks, approaching a plateau of 10.6 points. When compared to the corresponding linear model (LME), the NLME model yielded a better fit, with lower AIC and BIC values, and a highly significant likelihood ratio (LRT 758.8, p < 0.0001).

Conclusions: In a large, multistage VA cohort, the antidepressant response to TMS follows a nonlinear trajectory that is well described by an exponential decay function, replicating prior observations and demonstrating that the greatest improvements typically occur early in the TMS treatment course. This modelling framework provides a concise conceptual and mathematical summary of the expected symptom change during TMS, informing clinical decision-making, including early course monitoring and treatment planning.

Keywords: transcranial magnetic stimulation (TMS), Major depression, statistical methods, veterans.

Disclosure: Nothing to disclose.

P586. Neuromodulation of the extended emotional face perception network reduces amygdala BOLD signal and self-reported anxiety in anxious young adults

E. Kale Edmiston, Jay Fourier, Connor Bell, Tiffany Nan, Shaken Zheng, Anthony Rothschild, Mary Phillips

University of Massachusetts Chan Medical School, Worcester, Massachusetts, United States

Background: Aberrant processing of emotional faces has been linked to elevated threat sensitivity, and the extended face processing system, comprised of visual cortices and the amygdala, is implicated in anxiety disorders. In this randomised, sham-controlled, single-blind study, we targeted the extended face processing system in anxious young adults using transcranial direct current stimulation (tDCS). We hypothesised that active tDCS would be associated with alterations in fusiform (FG) and amygdala blood oxygen level dependent (BOLD) signal, as well as reduced anxiety, relative to sham stimulation.

Methods: Twenty-six unmedicated young adults (18–30 years of age, 15 female) ranging in trait anxiety underwent sham and active tDCS targeting the extended face processing network. Order of active vies sham tDCS was randomized. Participants received 1 mA current or sham for 12 minutes while completing an emotional face morph task with happy and fear faces during functional MRI. Participants completed the Spielberger Trait Anxiety Inventory at an earlier baseline visit and the Spielberger State Anxiety Inventory (STAI) immediately pre/post scan. Mixed linear models included extracted unilateral amygdala or FG BOLD as the dependent variable, tDCS condition (active vs. sham), STAI trait scores, STAI trait by tDCS condition interaction, and demographics as independent variables. For our model of symptom change, STAI state scores (post scan-pre scan baseline) were included as the dependent variable and tDCS condition and demographic variables as independent variables.

Results: There was a significant effect of both trait anxiety (F(1,19) = 8.64, p = 0.008) and tDCS condition (F(1,23) = 7.24, p = 0.013) on left amygdala BOLD signal during fear face perception, such that active tDCS and lower trait anxiety were associated with lower BOLD signal. There was a significant effect of tDCS condition on the left FG such that BOLD increased with active stimulation (F(1,23) = 8.37, p = 0.008) and a significant trait anxiety by tDCS condition effect in the R FG (F(1,22) = 4.46, p = 0.046) during fear vs happy face perception. No other neuroimaging models demonstrated significant effects of tDCS condition, trait anxiety, or their interaction (all ps > 0.10). There was a significant reduction in STAI state scores following active stimulation relative to the sham condition (F = 5.43, p = 0.027).

Conclusions: The extended face processing system is a promising novel target in anxiety. Our montage was associated with a reduction in anxiety symptoms, as well as modulation of the amygdala and FG BOLD response to social threat. Trait anxiety moderated this effect in the FG, such that active stimulation caused the FG response in those with high trait anxiety to be akin to that of individuals without anxiety. Dynamic causal modeling will help to determine if modulation of FG signal mediates the effect in amygdala or vice versa. Longitudinal studies will clarify the duration of observed effects on anxiety symptoms and an appropriate dose for this intervention.

Keywords: face perception, Anxiety, Neuromodulation, Visual threat detection, amygdala-based networks.

Disclosure: Nothing to disclose.

P587. Vagus nerve stimulation reduces suicidal ideation in markedly treatment-resistant depression: recover findings

Vasilis Hristidis, Harold Sackeim, A. John Rush, Scott Aaronson, David Dunner, Victoria de Leon, Andrew van der Vaart, Britt Gott, Ying Chieh (Lisa) Lee, Mark Bunker, Charles Gordon, Olivia Shy, Quyen Tran, Yvette Sheline, Rebecca M Allen, Ziad Nahas, Charles Conway

Washington University School of Medicine, Saint Louis, Missouri, United States

Background: Individuals with markedly treatment-resistant depression (TRD) have substantial risk for suicide attempt and completion. To date, few antidepressant treatments offer benefit for suicide, and it is critical that novel treatments not increase/worsen suicidal ideation or behavior. RECOVER is the largest, triple-blinded, device-based, randomized controlled trial (RCT) ever conducted in patients with markedly TRD, with 84 sites throughout the continental United States. A large subset (~40%) of RECOVER participants with major depressive disorder had a previous history of suicide attempt, and an even larger subset (~70%) entered the trial with ongoing suicidal ideation.

Methods: Participants (N = 493) in this 12-month, triple-blinded, randomized sham-controlled trial were implanted with a vagus nerve stimulation (VNS) device and received treatment as usual throughout. The device was activated or not based on 1:1 randomization, and patients received active or sham stimulation for the duration of the trial. After 2 months of upward device electrical parameter titration, with pseudo-titration for sham, suicide assessments were performed monthly through month 12 of the RCT phase of the trial. Patients with active suicidal ideation or a suicide attempt in the previous 6 months were excluded from participation. The sample was profoundly treatment resistant, chronically ill, and disabled, with an average of more than 13 failed lifetime antidepressant treatments and a current major depressive episode duration of nearly 18 years. A total of 75% were unemployed. The suicide outcome measure was a composite combining the suicide items on the blinded, off-site, clinician-rated Montgomery-Åsberg Depression Rating Scale (MADRS) and Quick Inventory of Depressive Symptomatology–Clinician (QIDS-C) and the patient-rated Quick Inventory of Depressive Symptomatology–Self-Report (QIDS-SR). The composite summed the suicide questions from QIDS-C and QIDS-SR, as well as the equivalent 0–3 scaling of the MADRS suicide question, for a total composite score ranging from 0 to 9. The number of participants with and without meaningful change in this composite was determined at each assessment occasion. The analysis comparing the randomized groups used a repeated measures generalized linear mixed model on whether the suicidal ideation score was changed (reduced or increased) by at least 3 points at each follow-up visit. This threshold of 3 or more points was selected to denote clinically meaningful improvement/worsening, as it corresponded to a change in 1 level on each of the suicide scales. Remission of suicidal ideation was considered to be a reduction of 3 or more points and a composite score of 2 or less. Two windows of time were used for these analyses: the entire 10 months of active/sham stimulation and a single measure averaging the final 3 months of stimulation (months 10–12). This delayed window for evaluating efficacy was selected because previous RECOVER analyses demonstrated it to be the time period of maximal benefit. Finally, using the same suicide composite scale, assessment of emergence of clinically meaningful suicidal ideation (defined as an increase of at least 3 points on the scale) was performed in those with and without suicidal ideation at baseline over the course of the entire trial.

Results: Of the 493 patients entering the trial, 463 had at least 1 suicidal assessment after implantation (modified intent-to-test sample). Among those with significant suicidal ideation at baseline (N = 286), as compared to sham VNS, those receiving active VNS had significantly greater clinically meaningful reduction in suicidal ideation over the duration of the trial (odds ratio [OR], 1.43; 95% CI, 1.004–2.021). The greater reduction in the active versus sham VNS group was observed throughout the trial, from the first assessment onward. Additionally, across the last 3 months of the trial (months 10–12), active VNS showed statistically superior reduction in likelihood of remission of suicidal ideation compared to sham VNS (OR, 1.67; 95% CI, 1.007–2.763). Across all subjects, emergence of suicidal ideation was rare, and active VNS (as compared to sham) was not associated with emergence of greater suicidal ideation in either those with (1.4% of visits for active VNS, 1.6% for sham) or without (4.4% of visits for active VNS, 3.6% for sham) clinically meaningful suicidal ideation at baseline.

Conclusions: During the blinded and randomized phase of the study, active VNS resulted in greater reductions in suicidal ideation than sham VNS. Further, as compared to sham VNS, active VNS did not result in worsening suicidal ideation. VNS is safe in patients with markedly TRD with suicidal ideation and could potentially contribute to reduction in suicidal ideation.

Keywords: Treatment-resistant depression, Major Depressive Disorder, Vagus Nerve Stimulation, Suicidal ideation, Neurostimulation.

Disclosure: LivaNova, PLC, Contracted Research, Self, LivaNova, PLC, Honoraria, Self.

P588. The pregenual anterior cingulate cortex: a novel target for intermittent theta-burst stimulation in depression?

Paulo Suen, Beatriz Cavendish, Pedro Henrique Silva, Andre Brunoni

UTSW Medical Center Dallas, Dallas, Texas, United States

Background: Intermittent theta-burst stimulation (iTBS) to the left dorsolateral prefrontal cortex (DLPFC) is effective for major depressive disorder (MDD), yet the neural mechanisms that mediate clinical improvement remain incompletely defined. Converging work implicates fronto-cingulate circuitry in antidepressant response, but most TMS studies emphasize the subgenual ACC. The pregenual ACC (pgACC; BA24/BA32) is a key affect-regulatory hub that integrates top-down control signals from DLPFC and valuation signals from ventromedial networks, shows state-dependent abnormalities in MDD, and is modulated across effective treatments. However, despite this well-described pathophysiology, pgACC has been underexplored in the context of iTBS for depression and has received comparatively less attention than the subgenual ACC in TMS studies. Therefore, we posited that early structural and functional changes within pgACC during an iTBS course would relate to subsequent symptom improvement, motivating an explicit, a priori interrogation of pgACC as a candidate mechanistic target.

Methods: We conducted an open-label iTBS trial in adults with treatment-resistant depression. Participants received 20 weekday iTBS sessions over four weeks (1,800 pulses/session at 100% motor threshold) targeting the left DLPFC. Depression severity (HDRS-17) was assessed weekly. MRI was acquired at baseline and after 10 treatment days. Regions of interest (ROIs) were defined with the Glasser atlas. The primary, a priori ROI was the left pgACC (regions a24, p24, d32, p32, and s32). Primary neural outcomes were (i) gray matter volume (GMV) change within pgACC and (ii) resting-state functional connectivity (rsFC) between pgACC and DLPFC parcels from baseline to Day-10. Paired t-tests quantified within-subject neural change; linear mixed models (LMM) related neural measures (main effects and time interactions) to HDRS-17 trajectories across the treatment course. (clinicaltrials.gov preregistration: NCT04969549).

Results: Out of 59 enrolled patients, 50 completed all MRI and clinical assessments and were included in the analyses (mean age = 39.89 ± 10.79; 44% female). Participants showed a significant reduction in symptom severity from baseline (M = 18.2 ± 2.7) to post-treatment (M = 8.0 ± 5.1). Neuroimaging analyses revealed early changes after 10 days of iTBS. Structurally, we observed a significant increase in GMV within the left pgACC from baseline to early treatment (p = .006, CI = [0.01, 0.10], ES = 0.22). Importantly, this volumetric increase was positively associated with greater symptom reduction after treatment (p < .005, CI = [0.93, 2.67], ES = 1.81). In terms of functional connectivity, target engagement and validation were centered in the left s32 region of the pgACC. Specifically, FC changed from baseline to day 10 (t = −2.02, p = 0.048), and we observed a significant correlation between changes in connectivity and early HDRS-17 changes (r = 0.28, p = 0.047). Finally, early FC changes were associated with greater improvement from baseline until week 6 - the interaction factor between connectivity and changes was significant (p = 0.0066) with a slope of 2.44.

Conclusions: This study provides converging structural and functional evidence of early fronto-cingulate modulation during iTBS treatment for depression. Specifically, the pgACC showed both volumetric increases and connectivity changes after 10 days of stimulation, and changes were associated with later clinical improvement. These findings support the role of this region as a key target in iTBS.

Keywords: Major Depressive Disorder (MDD), Theta-burst stimulation, Functional MRI (fMRI), Novel Targets, Anterior Cingulate Cortex (ACC).

Disclosure: Nothing to disclose.

P589. Repetitive transcranial magnetic stimulation (rTMS) targeting frontostriatal circuitry to reduce irritability

Wan-Ling Tseng, Dongyu Kang, Sofia Trujillo, Michal Assaf, Vaughn Steele

Yale Child Study Center, Yale University School of Medicine, New Haven, Connecticut, United States

Background: Irritability, defined as proneness to anger and frustration that may reach an impairing extent, is a transdiagnostic symptom cutting across many psychiatric disorders including unipolar depression, bipolar disorder, generalized anxiety disorder, posttraumatic stress disorder, and substance use disorders. Irritability in young people is among the top reasons for referrals to mental health services, accounting for up to 75% of inpatient admissions and ~60% of outpatient visits. In addition to high rates of hospitalization and service use, irritability is associated with risks for multiple poor outcomes including anxiety and depressive disorders, low income and educational attainment, and high suicidality, even after controlling for previous levels of irritability and concurrent psychiatric diagnoses. Indeed, irritability is a stronger predictor of suicide-related outcomes than overall depressive symptoms and, in patients with depression, the presence of irritability is associated with more impaired functioning and worse treatment outcomes. Despite the high public and personal health impact, very few evidence-based treatments exist for irritability, and there are no brain-based treatments. Frustrative non-reward (FNR), defined as the psychological state induced when an expected reward is withheld, plays an important mechanistic role in the pathophysiology of irritability. Research in rodents, non-human primates, and humans shows that FNR is associated with increased aggression and motor activity, paralleling the temper outbursts observed in individuals with irritability. Neuroimaging paradigms inducing frustration in humans thus provide an effective means to probe neural mechanisms underlying irritability. We conducted this pilot study using repetitive Transcranial Magnetic Stimulation (rTMS) as a non-invasive circuit-based neural modulation to target frontostriatal circuitry to reduce FNR responses central to the symptomatology and pathophysiology of irritability.

Methods: This is a within-subject active vs. sham pilot study aimed to recruit 20 young adults aged 18–25 years with elevated irritability. Data collection is ongoing. Here, we present data from 5 participants (3 males; mean age = 20.6 years, range = 18–23 years), totaling 20 fMRI scans. If accepted for presentation, more data will be added. Participants received acute single-session rTMS (excitatory or sham; counterbalanced) applied to left dorsolateral prefrontal cortex. fMRI data were collected during a well-validated frustration task (i.e., the Affective Posner) to examine changes in frontostriatal connectivity (between dorsal anterior cingulate cortex [dACC] and dorsal striatum) from before (PRE) to after (POST) each rTMS. Motor activity, a cross-species translational indicator of frustration response (indexed by grip force) and self-reported feelings of unhappiness during the task are used to examine PRE-POST rTMS changes in behavioral and affective responses to FNR, respectively. We conducted one-tailed paired t-tests comparing PRE and POST changes between active and sham rTMS in the neural (frontostriatal connectivity), behavioral (motor activity indexed by grip force), and affective (self-reported feelings of unhappiness) FNR responses. Our power analyses indicated that at one-tailed alpha of 0.05, with 80% power, N = 62 is needed to detect effects as small as d = 0.32 and N = 5 is needed to detect effects as large as d = 1.43 when examining PRE to POST changes in neural, behavioral, and affective FNR responses.

Results: Results showed that active rTMS decreased grip force and feelings of unhappiness (Cohen’s |d|s = 0.87 and 1.43, respectively) during frustration from PRE to POST, while sham did not (t = 2.37 and 4.07, p = 0.013 and 0.038, respectively). Although the result also showed that active rTMS increased frontostriatal connectivity (Cohen’s |d| = 0.32) during frustration from PRE to POST while sham did not, this is not statistically significant (t = 0.50, p = 0.321).

Conclusions: Data suggest the feasibility and promise of a single active rTMS session to modulate neural, behavioral, and affective measures of frustration. This project will provide critical pilot data to support frontostriatal circuitry as a neural circuitry mediating frustration processing and regulation that could be modulated using rTMS as a potential transdiagnostic treatment for irritability. Results will have broad impact, informing development of evidence-based, neurobiologically-informed interventions for irritability in the context of multiple psychiatric disorders.

Keywords: irritability, frustration, repetitive transcranial magnetic stimulation (rTMS).

Disclosure: Nothing to disclose.

P590. Personalized fMRI-guided TMS targeting the threat neurocircuitry in PTSD: a randomized clinical trial

Sanne van Rooij, Ryan Langhinrichsen-Rohling, Sean Minton, Cecilia Hinojosa, Joshua Lukemire, Rebecca Hinrichs, Natalie Merrill, Timothy Ely, Kristina Dahlgren, Patlapa Sompolpong, Patricio Riva Posse, Paul Holtzheimer, Vince Calhoun, Joan Camprodon, Sheila Rauch, Nadine Kaslow, Kerry Ressler, Tanja Jovanovic, William McDonald

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

Background: There is increasing interest in using Transcranial Magnetic Stimulation (TMS) to specifically target the functional neurocircuitry of the individual. Posttraumatic stress disorder (PTSD) neurobiology is well understood with current models rooted in translational neuroscience. Dysregulation within the threat neurocircuitry, particularly right amygdala hyperreactivity, is highlighted as a central mechanism. TMS has shown promise in reducing PTSD with variable clinical outcomes. However, a mechanistic understanding is needed to optimize treatment parameters and identify patients who may benefit most. The current neuroscience-informed trial aimed to modulate the threat neurocircuitry using fMRI-guided TMS to treat PTSD.

Methods: Sixty-three men and women (18–65 years of age) with PTSD symptoms were consented into the TMS study and n = 50 were randomized to 10 twice-daily sessions of 1Hz TMS (1800 pulses/session, total 36,000) or sham. TMS was delivered to a personalized fMRI-guided target within the right dorsolateral prefrontal cortex with strongest functional connections with the right amygdala. Right amygdala threat reactivity was measured pre- and post-TMS when participants viewed fearful and neutral faces in the MRI scanner. Skin conductance response (SCR) during trauma recall was measured with eSense. PTSD symptoms were measured with the PTSD checklist for DSM-5 (PCL-5) pre- and post-TMS. The primary outcomes of this randomized clinical trial were PTSD neuroimaging and psychophysiological markers as described in clinicaltrials.gov. The secondary goal of this research was to inform on clinical outcomes and learn if this treatment could be most beneficial for patients with a threat reactive neurobiological profile.

Results: Forty-seven (94%) participants completed the TMS visits. Active TMS compared to sham lowered right amygdala threat reactivity. A significant group by time interaction was observed (Beta = −0.09, 95% CI (−0.18, −0.01); p = 0.035). Specifically, the active group showed a significant reduction in right amygdala response to fearful faces from pre- to post-TMS (estimate post-pre = −0.110, 95% CI = −0.201, −0.020, p = 0.018) whereas the sham group did not (estimate post-pre = −0.005, 95% CI = −0.105, 0.095, p = 0.924). No significant effect of TMS was observed for SCR to trauma recall. Secondary analyses showed an overall reduction in hyperarousal (t(43.3) = 2.17, p = 0.036) and PTSD symptoms (t(43.3) = 3.68, p = 0.0006), but no interaction with group. However, following our neurobiological rationale, we performed exploratory analyses including only individuals exhibiting baseline threat reactivity (baseline Beta value fearful faces > =0.1; n = 34 out of 44).

Conclusions: This clinical trial is the first to demonstrate the effect of 2 weeks of personalized neuroscience-informed TMS on threat neurocircuitry biomarkers and PTSD symptoms. This approach supports the development of more precise, circuit-based interventions for psychiatric disorders.

Keywords: PTSD, transcranial magnetic stimulation (TMS), Threat neurocircuitry, personalized medicine, Amygdala.

Disclosure: Nothing to disclose.

P591. Tuning out the amygdala in depressed patients with low intensity focused ultrasound

Amanda Arulpragasam, Mascha van 't Wout-Frank, Yosef Berlow, Emily Aiken, Julia Gillotti, Alison Gorbatov, Nicole McLaughlin, Ryan Van Patten, Yosef Berlow, Israel Liberzon, Jennifer Barredo, Stephen Correia, Benjamin Greenberg, Noah Philip

Brown University and VA Providence, Providence, Rhode Island, United States

Background: Low intensity focused ultrasound (FUS) delivers acoustic energy to non-invasively modulate deep brain regions. In September 2021, we launched a first-in-human clinical trial of the controlled use of FUS to modulate the amygdala in patients with depression (NCT 05147142). Our selection of the amygdala was motivated by its centrality to many neuropsychiatric conditions. This study tested whether FUS could safely modulate the amygdala, and whether this occurred with a high degree of anatomical precision.

Methods: Study procedures were performed under FDA IDE G200146 and VA Providence IRB approval. This describes the first n = 10 a priori defined cohort. The study used a within-subject crossover design, where participants with depression (with or without anxiety) were randomized to FUS to right amygdala (target) or left primary somatosensory cortex (S1; active control). Safety assessments occurred immediately via MRI, at 24 hours (clinical rating scales and neuropsychological testing), and 1 week after FUS (clinical rating scales and neuropsychological testing). FUS was delivered under MRI guidance in two 10-minute applications, using a 10Hz pulse repetition frequency at the upper limit of FDA-permitted intensity. Assuming safe completion of FUS to the first randomized region, the subsequent one was sonicated with identical procedures. Primary outcomes were safety (MRI; side effects via systematic inquiry and spontaneous report) and target engagement (via change in arterial spin labeling (ASL)).

Results: Ten participants (2/10 women, 9/10 with comorbid anxiety) completed the study; of these, 9/10 completed both FUS sessions. There was no evidence of physical injury across all MRI scans (acute, 24 hours, and 1 week) in all participants. Neuropsychological performance was also not significantly impacted after FUS. There were no serious adverse events; common adverse effects were sedation and headache (p < .022 indicating more during amygdala vs. S1 FUS). One participant demonstrated acute worsening (~24 hours) after amygdala FUS and required a study-defined enhanced monitoring protocol over 1 month; no long-term sequelae were observed and he returned to near baseline symptoms after approximately two weeks without further intervention. Regarding neuroimaging, there was significant ASL change in the targeted amygdala vs. S1 (t = 6.02, p < .001; primary outcome measure). To add additional rigor we compared ASL change in the amygdala vs. ipsilateral right hippocampal head, and still observed significant changes (t = 5.62, p < .001). Spontaneous reports of benefit heavily favored the amygdala target (p < .001).

Conclusions: This study provides the first evidence that non-invasive MRI-guided FUS to the amygdala can provide highly spatially specific neuromodulation. While no physical or neuropsychological injury was observed, one participant exhibited psychiatric deterioration with a somewhat delayed onset that resolved over time without intervention. These findings demonstrate the spatial specificity of this technology and can be used to design future precision neuropsychiatric studies and cautious investigation into therapeutic use.

Keywords: focused ultrasound, Depression and anxiety, amygdala.

Disclosure: Pulvinar Neuro, Advisory Board, Self, Motif, Consultant, Self.

P592. Defining and engaging a novel rTMS target for nicotine craving in psychotic disorders

Heather Ward, Sophia Blyth, Simon Vandekar, Baxter Rogers, Gulcan Yildiz, Brett Clementz, Elliot Gershon, Matcheri Keshavan, Shashwath Meda, 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 remains unknown. In schizophrenia, small studies have linked default mode network (DMN) organization to tobacco use and showed that nicotine normalizes DMN disorganization. We sought to 1) validate the relationship between DMN organization and tobacco use using a large psychosis-spectrum sample (Bipolar-Schizophrenia Network on Intermediate Phenotypes 2, B-SNIP2); and 2) test if targeting this network with single and multiple sessions of repetitive transcranial magnetic stimulation (rTMS) affects craving.

Methods: In B-SNIP2, we tested associations between DMN connectivity and tobacco use. In the Single Session DMN-targeted rTMS study, individuals received single rTMS sessions (intermittent theta burst stimulation, iTBS; continuous theta burst stimulation, cTBS; sham) with pre-/post-neuroimaging and craving assessment. In the Accelerated, Multi-Session DMN-targeted cTBS study, individuals received 5 sessions of cTBS with pre-/post-neuroimaging and craving assessment.

Results: In B-SNIP2 (n = 596), current smokers had lower DMN connectivity than former (p = .017) and never smokers (p = .021). These differences were also observed in the psychosis group (current vs. former p = .044; current vs. never p = .011). In the Single Session DMN-targeted rTMS study (n = 11), there was a nonsignificant treatment*time interaction (p = .059) where iTBS increased craving (padj = .015) compared to cTBS and sham. In the Accelerated, Multi-Session DMN-targeted cTBS study (n = 12), DMN-targeted cTBS reduced craving after each session (p < .001) and reduced DMN connectivity (p = .052).

Conclusions: We identified a mechanism of nicotine use in psychosis and demonstrated that engaging this target reduces craving, suggesting a novel target for nicotine interventions in psychosis.

Keywords: psychosis, nicotine dependence, default mode network, resting-state functional connectivity, transcranial magnetic stimulation.

Disclosure: Nothing to disclose.

P593. Integrating virtual reality and E-field modelling for interventional psychiatry training: development of a mixed reality neuromodulation simulation lab

Tarik Hawkins, Jasmine Chen

Thomas Jefferson University, Philadelphia, Pennsylvania, United States

Background: Training in interventional psychiatry is inconsistent across programs, with limited exposure to electroconvulsive therapy (ECT) and transcranial magnetic stimulation (TMS). Lecture based didactics struggle to fully convey principles of coil/electrode placement, electric field (E-field) distribution, or functional neuroanatomy. Mixed reality (MR) offers unique advantages over virtual reality (VR) or augmented reality (AR), enabling integration with real-life instructors and engagement with props such as coils or mannequins for anatomical reference. This interactive format has the potential to enhance realism and skill transfer while standardizing training across diverse settings.

Methods: We developed a MR training platform incorporating 3D designed neuromodulation devices, parcellated brain models, and SimNIBS-generated E-field simulations for both ECT and TMS. The simulator allows visualization of coil/electrode placement, cortical E-field distributions, and target localization. Psychiatry residents (N = 20 across PGY-1–4) and faculty attendings (N = 5) at Jefferson Philadelphia Hospital were enrolled. Five participants from each class were selected for assessment. Outcomes included usability, perceived educational value, confidence in coil/electrode placement, and knowledge in physics and neuromodulation protocols.

Results: Preliminary data indicate that attending physicians demonstrated the largest increases in physics-based knowledge domains (waveforms, Ohm’s law) but modest gains in other areas. Residents showed greater improvements in understanding neuromodulation protocols, particularly TMS, along with a notable increase in interest in pursuing further neuromodulation training. Both groups rated the platform as highly usable and valuable for bridging conceptual and procedural gaps.

Conclusions: A mixed reality neuromodulation training platform with integrated E-field modeling is feasible, well-accepted, and may address gaps in interventional psychiatry education. Preliminary findings suggest differential benefits across training levels, with attendings advancing in physics concepts and residents gaining confidence in TMS protocols and interest in neuromodulation careers. Findings to be presented include ECT simulation and effects of E-field modelling on education. This immersive approach highlights the potential of MR to standardize and enhance interventional psychiatry training.

Keywords: virtual reality, Education and Training, Neuromodulation.

Disclosure: Nothing to disclose.

P594. Elunetirom, a first-in-class, brain-targeting, antidepressant candidate in Ph2 development, triggers robust hippocampal synaptogenesis and cognitive benefits in preclinical studies

Jason Harris, Jill Baccei, Bridgette Franey, Jeffrey A. Vivian, Deidre MacKenna, Will Stratton, Thomas S. Scanlan, Gudarz Davar

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

Background: Impaired neuroplasticity, synaptic loss, and dysfunctional mitochondria in the cortical and limbic regions of the brain are hallmarks of many neuropsychiatric disorders. Psychoplastogens, such as ketamine and psilocybin, have shown efficacy in treating a variety of human neuropsychiatric disorders. These clinical compounds are hypothesized to elicit their benefit through the promotion and support of neuroplasticity, but the accompanying psychotomimetic experiences have limited their widespread clinical use. The thyroid hormone T3 is an established driver of neuroplasticity during pre- and postnatal development via its upregulation of both brain-derived neurotrophic factor (BDNF) expression and mitochondrial bioenergetics. T3 acts on nuclear hormone receptors to directly stimulate neuritogenesis and synaptogenesis and supports the increased energy demands of the growing neurons. While T3 has shown benefit in the adjunctive treatment of major depressive disorder, it is often underutilized by clinicians due to the risk of adverse cardiac effects. We designed the brain-targeting thyroid hormone receptor agonist prodrug, elunetirom, to achieve similar efficacy to the psychoplastogens and T3 without their undesirable properties. In a phase 1 study of elunetirom, no CNS dissociative or clinically meaningful adverse cardiac effects were observed at any of the dose levels explored.

Methods: The effects of elunetirom were examined in a series of in vitro and in vivo assays relevant to neuroplasticity and the treatment of major depressive disorder and bipolar depression. Cortical and hippocampal neurons were isolated from Wistar rat embryos; male C57BL/6J mice were used in the aged-mouse and scopolamine studies. All experimental protocols in animal studies were approved by the appropriate 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. Analysis of variance (ANOVA) was performed on the resulting data; Fisher’s test was used for pairwise comparisons, and p value ≤ 0.05 was considered significant.

Results: Elunetirom’s active metabolite demonstrated significant effects on measures of neuroplasticity across in vitro and in vivo experiments. In a primary culture of cortical neurons, the active metabolite showed significant increases in number of neurons, neurite length, number of roots, and number of neurite extremities. In a primary culture of mature hippocampal neurons, elunetirom’s active metabolite also significantly increased markers of synaptogenesis. In both experiments, the active metabolite’s effects were comparable to those achieved with the positive control BDNF (50 ng/mL). Elunetirom significantly improved spontaneous alternation in aged mice (12-mo) in a T-maze after seven days of treatment, indicating an improvement in cognition presumably stemming from improvements in neuronal plasticity. Similarly, seven days of elunetirom treatment significantly reversed scopolamine-induced cognitive disruption in the mouse T-maze assay.

Conclusions: Elunetirom is a brain-targeting thyroid hormone receptor agonist prodrug that is in Phase 2 clinical development for MDD and bipolar depression. Elunetirom drives robust neuroplastic effects in brain neurons. Preclinical and clinical data suggest that elunetirom has the potential to have a rapid and enduring antidepressant effect in humans without cardiac or psychotomimetic side effects. These potential therapeutic effects are the result of direct actions on the genes that drive restorative neuroplasticity in brain (e.g., BDNF).

Keywords: Depression, Neuroplasticity, precision psychiatry, Synaptogenesis, neuritogenesis.

Disclosure: Autobahn Therapeutics, Inc., Employee, Self.

P595. Discovery and development of novel drug candidates for the treatment of neurological and psychiatric disorders through NIH blueprint neurotherapeutics network (BPN)

Pascal Laeng, Mohamed Hachicha, Shamsi Raeissi, Jamie Driscoll, Oreisa O'neil, Ranga Rangarajan, Charles Cywin

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

Background: Although the success rate of novel treatments for nervous system diseases is low, the perceived difficulties have not limited the enthusiasm of academic and industry scientists to undertake novel drug discovery approaches for the treatment of neurological and psychiatric disorders. The findings in basic research are often not translated to new neurotherapeutic drugs and clinical testing due to lack of relevant infrastructure, and resources. To boost and de-risk drug discovery and development in the neuroscience field, the Division of Translational Research (DTR) within NINDS, and in collaboration with NIMH and other NIH-institutes through the NIH Blueprint for Neuroscience Research, introduced a series of translational programs to promote neuroscience drug discovery and development efforts to mitigate the current pipeline gaps.

Methods: The NIH Blueprint Neurotherapeutics (BPN) Network provides non-dilutive funding for small molecule or biologics drug discovery and development, from exploratory/hit-to-lead through phase I clinical testing. This is accomplished through a combination of grant funding and access to a full range of BPN-sponsored contractors (medicinal chemistry, pharmacokinetics/ADME, toxicology, drug manufacturing/formulation and phase I clinical trials) and to NIH funded subject matter experts with extensive pharma experience. Projects with promising starting molecules and structure activity driving assays can enter the BPN program during the Discovery phase or more advanced lead molecules can enter at the Development stage with the same goal of filing an IND and achieving first in human testing while protecting investigators intellectual property. Discovery involves medicinal chemistry optimization to improve the potency and ADMET properties of compounds in order to identify a development candidate (DC) and seamlessly progress into the development phase. More advanced projects with a well characterized DC can move directly to the Development stage to undergo non-GLP and GLP preclinical toxicology, targeted phase I formulation and manufacturing testing required for an IND. Ultimately, if approved by the FDA, the BPN-program supports advancing into phase I/first-in-human testing.

Results: In this presentation, we will briefly describe the overall organization of the BPN program and how BPN team projects are organized as virtual “biotech” groups gathering bi-weekly. We will show examples and lessons learned demonstrating this funding mechanism, resources and know-how have contributed to translate academic and industry discoveries in basic disease biology into development of novel drug candidates for IND approval and advancement into clinical testing.

Conclusions: Several new molecules entered trials in Phase 1 utilizing the BPN mechanism and those projects have garnered outside funding and continued clinical advancement with several now in Phase 3 including two with psychiatric indications. This clinical progression represents a validation of a non-dilutive alternative funding model that should be highly relevant to the ACNP’s mission and audience to develop new treatments for psychiatric indications.

Keywords: Drug Discovery - new approaches, CNS drugs, Therapeutic Development, mental illness, Neurological and Psychiatric Disorders.

Disclosure: Nothing to disclose.

P596. Modulating the endocannabinoid system in alcohol use disorder: a translational systematic review and meta-analysis of preclinical and human studies

Gabriel Costa, Mayte Cerezo-Matias, Melissa Funaro, Deniz Bagdas, Alfred Kaye, John Krystal, Ismene Petrakis, Joao De Aquino

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

Background: Alcohol use disorder (AUD) affects 28.9 million adults in the United States, with only three Food and Drug Administration-approved medications showing limited effectiveness. The endocannabinoid system (ECS)—including cannabinoid receptors (primarily CB-1 and CB-2), endogenous ligands such as 2-arachidonoylglycerol (2-AG) and anandamide, and metabolic enzymes including fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL)—represents a promising therapeutic target. CB-1 receptors, the most abundant G-protein-coupled receptors in the brain, function as presynaptic inhibitory receptors that dynamically regulate neurotransmitter release across limbic and cortical circuits. Convergent evidence links ECS dysregulation to AUD: chronic alcohol exposure induces adaptive changes, including reduced CB-1 receptor availability and lower FAAH levels correlating with consumption severity. Cannabinoid Receptor 1 (CNR1)/FAAH polymorphisms are associated with AUD risk. This systematic review synthesizes evidence from experimental models and human studies to evaluate the ECS as a potential therapeutic target for AUD.

Methods: We conducted a PROSPERO-registered systematic review (CRD42024593040) adhering to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. We searched Ovid MEDLINE, Embase, PsycINFO, Web of Science Core Collection, Scopus, and Cochrane Library through July 2025, yielding 62 eligible studies (44 preclinical, 18 human). Preclinical studies were controlled laboratory experiments using rodent models with prior alcohol exposure, assessing alcohol preference or self-administration. Human studies included randomized controlled trials and experimental studies involving adults (healthy volunteers to individuals diagnosed with AUD). Meta-analyses calculated standardized mean differences (SMD) using random-effects models for preclinical data grouped by mechanism: CB-1 receptor antagonists (23 studies), CB-1 receptor agonists (12 studies), cannabidiol/CBD (9 studies). Dose-response relationships were assessed using linear and non-linear regression. Human studies were narratively synthesized due to methodological heterogeneity. Both sexes were included where reported.

Results: Meta-analyses of 37 preclinical experiments demonstrated CB-1 receptor antagonists (primarily SR141716A/Rimonabant) significantly reduced alcohol intake (SMD = −1.21, 95% confidence interval[−1.59, −0.83]). Non-linear dose-response analysis indicated biphasic effects with optimal efficacy at intermediate doses (2–5mg/kg, p = 0.002). Newer neutral CB-1 receptor antagonists (AM4113, AM6527) showed comparable efficacy while potentially avoiding adverse effects associated with inverse agonism. CB-1 receptor agonists increased alcohol consumption (9 experiments; SMD = 0.66, 95%CI[0.30,1.02]). CBD demonstrated significant reductions (20 experiments; SMD = −0.70, 95%CI[−1.07, −0.34]) with a potential non-linear dose-response pattern and sex-specific effects (males responding at 30–60mg/kg, females requiring 90mg/kg). FAAH inhibitors reduced alcohol intake by 40–60% through CB-1 receptor-dependent mechanisms. Unlike preclinical experiments, human studies have shown inconsistent and generally null effects. Two rimonabant trials (n = 297) demonstrated minimal efficacy despite achieving plasma drug concentrations, with non-significant reductions in relapse (41.5% versus 47.7%, p = 0.13). Recent CBD trials showed more promise: 800mg significantly reduced nucleus accumbens activation to alcohol cues (p < 0.001) and attenuated craving. Full-spectrum CBD containing minimal delta-9-tetrahydrocannabinol reduced craving at weeks 8 and 16 (p < 0.001). CBD-dominant cannabis significantly reduced drinks per day by 40% in naturalistic studies.

Conclusions: Our systematic review and meta-analysis reveal a translational gap between robust preclinical ECS findings and limited clinical success. While preclinical evidence supports ECS modulation, particularly CB-1 receptor antagonists and CBD, as promising interventions for reducing alcohol use behaviors, clinical translation has been limited by safety concerns, methodological inconsistencies, and under-investigation of novel compounds. Rimonabant's psychiatric adverse effects ultimately led to discontinuation despite efficacy. Aside from CBD’s multi-targeted effects, promising alternatives include peripherally restricted CB-1 receptor antagonists, neutral antagonists (AM4113, AM6527), and FAAH/MAGL inhibitors. Mechanistically informed trials of novel compounds, including next-generation CB-1 receptor antagonists and CBD, are needed to bridge this translational gap and yield new treatments for AUD.

Keywords: endocannabinoid system, Alcohol Use Disorder - Treatment, Translational research, systematic review, meta-analysis.

Disclosure: Nothing to disclose.

P597. Brain structural effects of dopaminergic antagonism vs. partial agonism in first-episode psychosis: a normative modeling approach

Anna Feveile, Karen Ambrosen, Ketevan Shalikashvili, Jayachandra Raghava, Mette Nielsen, Kirsten Bojesen, Barbora Bučková, André Marquand, Birte Glenthøj, Warda Syeda, Bjørn Ebdrup

Center for Neuropsychiatric Schizophrenia Research, Mental Health Center Glostrup, University of Copenhagen, Glostrup, Denmark

Background: Schizophrenia is associated with subtle brain structural alterations but separating disease from medication effects is challenging. Antipsychotic dopamine D2 receptor (D2R) antagonism has been associated with striatal volume increases, but effects of partial D2R agonism by newer antipsychotics are largely unexplored. We applied a normative model to existing data from two cohorts of initially antipsychotic-naïve patients with first-episode psychosis to compare short-term brain changes associated with either D2R antagonism or partial D2R agonism.

We first aimed to confirm previous conventional analyses showing volumetric striatal increases after six weeks of amisulpride treatment. Second, we hypothesized that patients treated with aripiprazole for six weeks would display less pronounced volumetric increases in the striatum compared to patients treated with amisulpride. Finally, we expected that patients at long-term follow-up timepoints would conform to a more chronic stage of psychosis by displaying cortical thinning of the frontal and temporal lobes.

Methods: Eighty-six antipsychotic-naïve patients with first-episode psychosis (mean age 23.4 +/−4.8 years, 47 females; 39 males) received 6 weeks of monotherapy with either amisulpride (D2R antagonist) (N = 41) or aripiprazole (partial D2R agonist) (N = 45). All patients underwent 3 Tesla structural magnetic resonance imaging before and after 6 weeks of treatment. A subset was re-scanned at 6 months, 1 year, and 2 years of naturalistic antipsychotic treatment.

A pre-trained normative model was applied to 186 subcortical and cortical regions. We used Wilcoxon signed-rank tests to identify longitudinal regional brain structural deviations.

Results: Both antipsychotics were associated with striatal volume increases after six weeks (amisulpride left and right caudate and right putamen; aripiprazole left and right caudate, corrected p-values < .001). No cortical effects were observed with amisulpride. Thinning of 11 temporal lobe regions were observed with aripiprazole (corrected p-values < .05).

After 6 months of naturalistic treatment, and 1 and 2 years, the striatal changes abated but cortical thinning in 11 regions in the frontal lobes emerged (corrected p-values < .05).

Conclusions: Our analyses confirmed transient striatal volume increases in initially AP-naïve FEP patients treated with amisulpride and identified similar increases in patients treated with aripiprazole. In addition, short-term aripiprazole monotherapy was found to induce temporal thinning. The identified cortical thinning in the frontal lobes of patients after years of naturalistic treatment, likely reflect the combined structural effects of pharmacological treatment and illness progression in patients with psychosis.

Conclusively, interpretation of structural brain changes in patients with psychosis require consideration of short-term pharmacological effects as well as factors related to illness progression.

Keywords: Antipsychotic-naïve first-episode psychosis, Antipsychotic monotherapy, brain structure, Normative Modeling.

Disclosure: Boehringer Ingelheim, Advisory Board, Self, Lundbeck Pharma A/S, Advisory Board, Self, Orion Pharma A/S, Advisory Board, Self, Boehringer Ingelheim, Honoraria, Self, Otsuka Pharma Scandinavia AB, Honoraria, Self, Lundbeck Pharma A/S, Honoraria, Self.

P598. Effect of antipsychotics on brain insulin action in relation to glucose metabolism in healthy volunteers: a pilot study

Emily Smith, Laurie Hamel, Nicolette Stogios, Marcos Sanches, Gary Remington, Aristotle Voineskos, Ariel Graff-Guerrero, Sri Mahavir Agarwal, Satya Dash, Margaret Hahn

Centre for Addiction and Mental Health, Toronto, Canada

Background: Antipsychotics are currently the cornerstone of treatment for schizophrenia spectrum disorders. However, despite their efficacy, their use is associated with serious metabolic side effects, including dysglycemia. Previous work from our group has demonstrated that these effects can occur independent of weight gain and in the absence of illness-related risk. Moreover, studies in rodents suggest that this process may in part be mediated by brain insulin; however, this has yet to be explicitly tested in humans. Therefore, the primary objective of the current proof-of-concept study was to examine whether antipsychotics can impact brain insulin action in healthy humans.

Methods: We conducted a single-blind, randomized, placebo-controlled crossover study (NCT03741478) to test if an acute dose (10 mg) of the antipsychotic olanzapine can directly inhibit the brain’s response to insulin in healthy, non-obese, antipsychotic-naïve participants. Each participant completed three different conditions in a randomized order: placebo spray/oral placebo (PLA/PBO), intranasal insulin/oral placebo (INI/PBO), intranasal insulin/olanzapine (INI/OLA). Our primary dependent variable was endogenous glucose production (EGP), which is known to be regulated by brain insulin activity. We measured EGP using a basal pancreatic euglycemic clamp, which involved using somatostatin to suppress pancreatic hormone production, which we then replaced at basal levels via an intravenous infusion. Brain insulin receptors were directly stimulated using 40 IU insulin lispro administered via a metered nasal spray. Data were analyzed using a linear mixed-effects model with condition, sex, visit number, and timepoint specified as fixed effects and subject included as a random intercept. All values were normalized to t = 0 (time of intervention administration) prior to analysis.

Results: In total, five males and two females completed the full study (age: 30.4 ± 11.2 years). Analyses to identify levels of EGP and other metabolic parameters are still ongoing. Therefore, preliminary findings regarding changes in blood glucose levels throughout the clamp are presented below. Compared to the control condition, intranasal insulin produced a modest but measurable decrease in blood glucose levels during the last 180 minutes of the clamp (i.e., t = 180 to 360 min; INI/PBO vs. PLA/PBO: −0.032 mmol/L, 95% CI: −0.051, −0.013; p = 0.001) whereas co-administration of olanzapine appeared to attenuate this effect (INI/OLA vs. PLA/PBO: 0.021 mmol/L, 95% CI: 0.002, 0.040; p = 0.029). Accordingly, glucose levels were significantly higher in the INI/OLA condition compared to INI alone (difference: 0.053 mmol/L, 95% CI: 0.034, 0.072; p < 0.001). No sex differences were detected; however, the small sample limits interpretation.

Conclusions: Although blood glucose levels during a pancreatic euglycemic clamp do not necessarily track with changes in EGP, they can still provide important insight into how antipsychotics interact with brain insulin. Considering this, our preliminary results are consistent with the hypothesis that olanzapine may interfere with brain insulin action. Ongoing analyses of EGP will clarify these effects. If confirmed, these findings could guide future studies investigating strategies to mitigate antipsychotic-induced metabolic side effects, which could ultimately support better treatment outcomes and quality of life for patients.

Keywords: Antipsychotics, Metabolic Side Effects, Glucose Metabolism, Pancreatic Euglycemic Clamp, Pilot Study.

Disclosure: Nothing to disclose.

P599. Symptom-driven identification of neurotransmitter correlates in early psychosis

Haley Wang, Zhen-Qi Liu, Charles Schleifer, Carrie Bearden, Bratislav Misic, Katherine Karlsgodt

University of California, Los Angeles, Los Angeles, California, United States

Background: Current antipsychotic medications, while able to reduce positive symptoms of psychosis, such as delusions and hallucinations, often fail to adequately address other disabling aspects of the illness, such as negative symptoms. This is in part due to a lack of granularity in the development of pharmaceuticals, with standard antipsychotics typically indicated for psychosis as a whole and not specific symptom domains. Recent advances enabling spatial correlation of in-vivo neurotransmitter receptor maps with other neuroimaging modalities offer a new path forward, to identify potential symptom-specific treatment targets. Here, we employ a novel, symptom-dimension-driven approach to bridge this gap. We first identify robust, multivariate resting-state functional connectivity (RSFC) patterns for specific symptom dimensions. These symptom-specific neural signatures are then spatially correlated with a comprehensive atlas of neurotransmitter receptor maps to identify their underlying molecular basis, thereby revealing potential molecular targets for mechanistically-informed pharmaceutical intervention.

Methods: We analyzed resting-state fMRI data from 124 early psychosis patients in the Human Connectome Project for Early Psychosis (HCP-EP) cohort (aged 16–35, 38.7% female). Patients were diagnosed with schizophrenia, schizoaffective disorder, and/or psychotic mood disorder and were within the first three years of onset. Four separate Partial Least Squares (PLS) correlation analyses identified multivariate relationships linking whole-brain RSFC (among 216 cortical and subcortical regions) with specific symptom dimensions: positive, negative, general psychopathology, and mania. Significant latent components (LCs) were determined via permutation tests (5,000 iterations, FDR correction q < .05). Sensitivity and post-hoc analyses for the RSFC strength signatures of each symptom domain examined potential confounds (e.g., age, sex, IQ, head motion, diagnoses, duration of illness, substance use, antipsychotic medication dosage). The resulting significant LCs, representing the spatial maps of symptom-specific functional connectivity, were then correlated with a comprehensive atlas of 19 neurotransmitter receptor and transporter distributions (e.g., 5-HT_1a/_1e/_2a/₄/₆, D₁/₂, NET, CB₁, NMDA, mGluR₅, GABA_a, H₃, M₁, etc) using Spearman’s correlation. To assess statistical significance, we used a spatial autocorrelation-preserving null model (“spin test”, Pspin < .05) for all tested correlation models.

Results: PLS correlation analyses identified significant multivariate relationships between whole-brain RSFC and three of the four symptom dimensions tested. Specifically, the positive symptom model revealed one significant LC (LCpos) explaining 54.8% of the shared covariance (p = .01). The negative symptom model also yielded a single significant LC (LCneg), accounting for 49.7% of the covariance (p = .002). Similarly, the general psychopathology model produced one significant LC (LCgps) that explained 44.2% of the covariance (p = .003). No significant LCs were identified for the mania symptom dimension (p > .05). The composite scores were not correlated with any confounds (q > .05). Spatially correlating these RSFC signatures with neurotransmitter receptor maps revealed distinct molecular associations (pspin < .05). The RSFC signature for LCgps was significantly positively correlated with the distribution of multiple serotonin receptors (.30 < r < .42), as well as cannabinoid (r = .39), mu-opioid (MOR, r = .41), and mGluR₅ (r = .45) receptors; it was negatively correlated with the GABA_a receptor (r = −.17). The RSFC signature for LCneg was positively associated with norepinephrine transporter (NET; r = .37) and vesicular acetylcholine transporter (VAChT; r = .38) distributions. Finally, the RSFC signature for LCpos was positively correlated with the serotonin transporter (5-HTT; r = .23) and NET (r = .24), but negatively correlated with α₄β₂ nicotinic acetylcholine (r = −.37), CB₁ (r = −.41), and mGluR₅ (r = −.26) receptors.

Conclusions: Our findings show that symptom-specific brain connectivity in early psychosis maps onto distinct neurotransmitter receptor systems, highlighting potentially biologically separable mechanisms for different symptom domains. General psychopathology is aligned with serotonin, cannabinoid, opioid, and glutamatergic systems, consistent with dysregulated affective and stress circuits. Negative symptoms were associated with noradrenergic and cholinergic transporters, pointing to impaired arousal and motivational drive. Positive symptoms mapped to serotonergic and noradrenergic systems but were inversely related to nicotinic, glutamatergic, and endocannabinoid signaling, consistent with the hypotheses that decreased nicotinic function is associated with psychosis symptoms and cognition, as well as the notion that glutamate hypofunction contributes to psychotic symptomatology. These results from a symptom-driven multimodal framework demonstrate that psychosis psychopathology might be supported by distinct molecular mechanisms, moving beyond categorical diagnoses and offering a path toward more targeted therapeutic development.

Keywords: Receptor maps, Drug Discovery - new approaches, Psychosis spectrum symptoms, Multivariate analysis.

Disclosure: Nothing to disclose.

P600. Topiramate in clozapine-treated patients with schizophrenia spectrum disorders: effects on weight and psychopathology

Margaret Hahn, Emily Smith, Kristoffer Panganiban, Kateryna Maksyutynska, Jiwon Lee, Kenya Costa-Dookhan, Gary Remington, Mahavir Agarwal

Centre for Addiction and Mental Health, Toronto, Canada

Background: Antipsychotics (APs) are currently the cornerstone of treatment for schizophrenia spectrum disorders; however, their use is associated with serious metabolic consequences including weight gain, resulting in an increased risk of diabetes and cardiovascular disease. In parallel, roughly 30% of patients demonstrate an inadequate response to AP treatment. Although clozapine remains the only medication indicated for treatment-refractory schizophrenia, it is also the AP with the greatest metabolic liability. Previous work suggests that topiramate (an anticonvulsant) may have beneficial effects on both weight and psychopathology when used adjunctively with APs in patients with schizophrenia. Thus, it is possible that topiramate may pose a unique dual benefit compared to other weight-reducing agents like metformin; however, evidence for its use with clozapine specifically is limited. In addition, topiramate has been shown to worsen cognition in epilepsy populations, which has not been adequately examined in schizophrenia.

Methods: The primary objective of this pilot triple-blinded (participant, primary investigator, and outcome assessor) randomized control trial was to examine the effect of topiramate on weight in clozapine-treated schizophrenia patients. As a secondary objective, we explored the effects of topiramate on other metabolic parameters, psychopathology, and cognition. Stable, overweight/obese (BMI ≥25) clozapine-treated adults with schizophrenia or schizoaffective disorder were randomized (2:1) to 16 weeks of topiramate or placebo up to a maximum tolerated dose of 100 mg BID. Anthropometric measures including weight and BMI were collected at each visit; clinical symptom scores and laboratory parameters were measured at baseline (week 1), midpoint (week 8), and endpoint (week 16). Results were analyzed using a linear mixed model with body weight as the primary outcome and group, time, and the interaction between group and time as predictor variables. Baseline weight and other demographic variables that may be associated with weight were included as covariates.

Results: In total, 36 participants (26 topiramate, 10 placebo) completed at least one follow-up visit and were included in the analysis. Demographic characteristics were comparable between groups except for the percentage of tobacco users, which was higher in the placebo group (27% vs. 70%, p = 0.026). In the primary model controlling for baseline weight, participants in the topiramate group lost 2.54 kg after 16 weeks of treatment (p < 0.0001) whereas those in the placebo group gained 1.09 kg (p = 0.26). Comparing the groups, participants in the topiramate group weighed significantly less at endpoint than those in the placebo group (97.39 vs. 100.94 kg, p < 0.0001) and showed significantly greater percent weight loss (difference: −3.54%, p = 0.007). These differences persisted after controlling for age, sex, and tobacco use. We did not find significant between-group differences in other metabolic parameters. For clinical symptom scale scores, after controlling for age, sex, and tobacco use, the topiramate group showed significant decreases from baseline to endpoint in the positive and negative syndrome scale (PANSS) positive subscale (13.12 vs. 11.68, p = 0.034) and PANSS-General subscale (29.62 vs. 26.10, p = 0.005). No significant differences were observed in the placebo group. There were also no significant differences in cognition between the groups and over time.

Conclusions: In this triple-blinded randomized placebo-controlled trial examining topiramate’s effects on metabolic parameters and psychopathology, we found that topiramate significantly lowered body weight after 16 weeks of treatment compared to placebo. We also found evidence for a significant reduction in positive and general symptom scores, suggesting a beneficial effect of topiramate on psychopathology. Future studies should aim to extend this work by increasing the sample size and/or extending the follow-up time.

Keywords: Schizophrenia (SCZ), clozapine, Obesity, Psychopthology, Topiramate.

Disclosure: Alkermes, Advisory Board, Self, Merck, Consultant, Self, Eli Lilly, Consultant, Self.

P601. Efficacy and safety of an NR2B NAM, BI 1569912, as monotherapy in people living with major depressive disorder: results from a phase II clinical trial

Gerard Sanacora, Franco De Crescenzo, Alla Feldberg, J. Corey Fowler, Rainer-Georg Goeldner, Brittney Starling

Yale University, New Haven, Connecticut, United States

Background: Many patients living with major depressive disorder (MDD), a common, severe, and recurrent mental illness, do not respond sufficiently to antidepressive treatment. An emerging treatment class, the NMDA subunit 2B-selective negative allosteric modulators (NR2B NAMs), is hypothesized to enhance synaptic plasticity in the prefrontal cortex and thereby improve mood in MDD. BI 1569912, an oral NR2B NAM, was investigated for the treatment of MDD. Here, we describe the outcome of a Phase II, multicenter, randomized, double-blind, placebo-controlled trial in patients living with MDD.

Methods: 1447-0012 (NCT06558344) was a Phase II dose-finding monotherapy trial that enrolled adults (18–65 years old) with an established MDD diagnosis by Diagnostic and Statistical Manual of Mental Disorders Fifth Edition (DSM-5) with 0 to 1 previous antidepressant treatment failures in the current episode. Participants were randomized to receive 6 weeks of treatment with either placebo or 1 of 3 doses of BI 1569912 (5 mg, 10 mg, or 20 mg) once daily (randomization ratio 2:1:1:2). Assessments included the Montgomery-Åsberg Depression Rating Scale (MADRS), Symptoms of Major Depressive Disorder Scale (SMDDS), safety, and tolerability. The primary efficacy endpoint was change from baseline in MADRS total score at Week 6. Exploratory endpoints included: change from baseline in MADRS total score at Day 8; change from baseline in Generalized Anxiety Disorder-7 (GAD-7) at individual time points and change from baseline in SMDDS total score at individual timepoints. Dissociation was measured using the Bowdle-Visual Analogue Scale (B-VAS).

Results: Participants (N = 225) were randomized to placebo (n = 76), BI 1569912 5 mg (n = 35), 10 mg (n = 41) or 20 mg (n = 73). Median age (range) was 41 (18–65) years; 54.2% of subjects were female, and mean (SD) MADRS total score at baseline was 31.9 (4.4) points. Mean (SE) change from baseline to Week 6 in MADRS total score was −13.6 (2.1) for BI 1569912 5 mg, −10.6 (2.0) for BI 1569912 10 mg, −10.0 (1.4) for BI 1569912 20 mg and −10.3 (1.4) for Placebo. At Day 8, no relevant differences in MADRS total score were observed between treatment arms with adjusted mean differences (SE) for BI 1569912 vs placebo of −0.4 (1.8) for 5 mg, 1.7 (1.6) for 10 mg and −1.0 (1.4) for 20 mg. Mean (SE) change from baseline to Week 6 in SMDDS total score was −12.3 (2.7) for BI 1569912 5 mg, −11.4 (2.2) for BI 1569912 10 mg, −11.1 (1.7) for BI 1569912 20 mg and −12.2 (1.6) for Placebo. No relevant separation from placebo on the SMDDS was observed at any of the other timepoints assessed either. Adjusted mean (SE) change from baseline to Week 6 in GAD-7 total score was −4.2 (0.9) for BI 1569912 5 mg, −3.8 (0.8) for BI 1569912 10 mg, −3.5 (0.6) for BI 1569912 20 mg and −2.9 (0.6) for Placebo. No relevant separation from placebo on the GAD-7 was observed at any of the other timepoints assessed either. The most common adverse events (with incidence > 2% and higher than in placebo) in BI 1569912 total treatment group were headache (12.1% vs 7.9% in placebo), dizziness (4% vs 1.3% in placebo), diarrhea (4% vs 1.3% in placebo), constipation (2% and 0% in placebo), nasopharyngitis (6% vs 3.9% in placebo). No dose dependent trend was observed. B-VAS scores of internal and external perception showed a difference of < 10 mm (100 mm scale) for BI 1569912 compared to placebo which is considered as not clinically relevant.

Conclusions: BI 1569912, an oral NR2B NAM investigated for the treatment of MDD as monotherapy, failed to meet the primary and exploratory efficacy endpoints in this Phase II trial. BI 1569912 was well tolerated with minimal adverse events and no signs of dissociation observed in any BI drug treated group.

Keywords: CNS Clinical Trials, NMDA glutamate receptors, Phase II clinical trial, Major depressive disorder.

Disclosure: Actinogen Medical, Consultant, Self, Alto Neuroscience, Consultant, Self, Atai, Consultant, Self, Axsome Therapeutics, Consultant, Self, Biogen, Consultant, Self, Biohaven Pharmaceuticals, Boehringer Ingelheim International GmbH, Bristol-Myers Squibb, Clexio, Daiichi Sankyo, Denovo Biopharma, EMA Wellness, Douglas Pharmaceuticals, Embark, Engrail Therapeutics, Freedom Biosciences, Gilgamesh, Holmusk, Intra-Cellular Therapies, Janssen, Levo therapeutics, Lundbeck, Merck, MiCure, Navitor Pharmaceuticals, Neumora Therapeutics, Neurocrine, Newleos Therapeutics, Novartis, Noven Pharmaceuticals, Otsuka, Perception Neuroscience, Praxis Therapeutics, Relmada Therapeutics, Seaport Therapeutics, Sage Pharmaceuticals, Seelos Pharmaceuticals, Supernus, Taisho Pharmaceuticals, Tetricus, Transcend Therapeutics, Usona Institute, Valeant, Vistagen Therapeutics, and XW, Advisory Board, Self, Biohaven Pharmaceuticals, Stock/Equity—Publicly Traded Company, Self, Boehringer Ingelheim International GmbH, Advisory Board, Self, Bristol-Myers Squibb, Consultant, Self, Clexio, Consultant, Self, Daiichi Sankyo, Consultant, Self, Embark, Advisory Board, Self, Engrail Therapeutics, Consultant, Self, Freedom Biosciences, Advisory Board, Self, Freedom Biosciences, Stock/Equity—Privately Held Company, Self, Gilgamesh, Consultant, Self, Holmusk, Advisory Board, Self, Intra-Cellular Therapies, Consultant, Self, Merck, Consultant, Self, Merck, Contracted Research, Self, Freedom Biosciences, Patent, Self, Neumora Therapeutics, Consultant, Self, Neurocrine, Consultant, Self, Newleos Therapeutics, Advisory Board, Self, Novartis, Consultant, Self, Otsuka, Consultant, Self, Relmada Therapeutics, Advisory Board, Self, Seaport Therapeutics, Advisory Board, Self, Sage Pharmaceuticals, Consultant, Self, Seelos Pharmaceuticals, Consultant, Self, Supernus, Consultant, Self, Tetricus, Advisory Board, Self, Transcend Therapeutics, Advisory Board, Self, Usona Institute, Consultant, Self, Usona Institute, Contracted Research, Self, Tetricus, Stock/Equity—Privately Held Company, Self, Newleos Therapeutics, Stock/Equity—Privately Held Company, Self, Lilly, Consultant, Self, Abbvie, Consultant, Self.

P602. Cannabidiol enhances conditioned pain modulation at opioid trough but blunts early post-dose analgesia in adults with opioid use disorder and chronic pain: a randomized crossover trial

Joao De Aquino, Gabriel Costa, Brian Pittman, Peggy Compton, Mohini Ranganathan, Mehmet Sofuoglu

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

Background: Chronic pain affects over half of individuals receiving medications for opioid use disorder (OUD), contributing to poorer treatment outcomes and increased risk of return to non-medical opioid use. Opioid agonist therapy (OAT), while effective for OUD, is associated with neuroadaptive changes in pain processing, including opioid-induced hyperalgesia and impaired conditioned pain modulation (CPM)—a psychophysical measure of top-down pain modulatory control. Psychophysical evidence indicates that higher opioid agonist doses are associated with dose-dependent impairments in pain modulation, compounding deficits already present in individuals with OUD and co-occurring chronic pain. Cannabidiol (CBD), a non-intoxicating phytocannabinoid with activity at multiple nociceptive-relevant targets, represents a promising adjunctive therapy for OUD. However, CBD's acute effects on pain modulation among those with OUD and co-occurring chronic pain remain unexplored. We hypothesized that CBD would enhance top-down pain modulation at opioid trough—when opioid tone is lowest and pain sensitivity highest—and tested whether CBD modulates acute responses to routine opioid dosing.

Methods: We conducted a randomized, double-blind, placebo-controlled, crossover human-laboratory study in 23 adults (mean age 44.0 ± 8.4 years, 48% female) receiving stable OAT (mean dose of methadone 85.7 ± 29.7 mg) for OUD with co-occurring chronic pain (≥6 months high-intensity, non-cancer low back pain). Participants attended four 6-hour sessions separated by ≥72 hours. Each session began at trough opioid levels (24 hours post-opioid dose) and followed this sequence: baseline pain assessments, oral CBD administration (400, 800, or 1200 mg “Epidiolex” or placebo in counterbalanced order), repeated assessments over 210 minutes, opioid (methadone) dose administration, and final assessment at 240 minutes. We performed quantitative sensory testing (QST) using thermode-based thermal stimuli (Medoc TSA-II). Our primary endpoint was change in top-down pain modulation (CPM) at 120 min post-CBD, quantified as the difference between unconditioned heat pain (46.5 °C/115.7 °F) and pain during concurrent cold conditioning (12 °C/53.6 °F). Secondary outcomes included heat pain threshold/tolerance (ramped from 32 °C/89.6 °F at 0.5 °C/s) and bottom-up pain facilitation (temporal summation of pain, TSP; area under curve of 10 sequential heat pulses at 46.5 °C/115.7 °F, 0.5s duration, 2.5s inter-stimulus interval). To explore CBD-opioid interactions, we compared QST measures from immediately pre-OAT dose (120–180 min post-CBD) to 30 min post-opioid dose (240 min). Linear mixed-effects models accounted for crossover design with subject random effects; effect sizes reported as Cohen's d. With our sample (n = 23), a within-subject design at α = 0.05 achieves 80% power to detect an effect of approximately d’ ≈ 0.60 for pairwise comparisons.

Results: During the pre-opioid dose period and peak effects of CBD (120 min), CBD produced dose-dependent enhancement of top-down pain modulation (CPM) relative to placebo: 400 mg d’=0.21 (p = 0.54), 800 mg d’=0.66 (p = 0.061), 1200 mg d’=0.77 (p = 0.063). This improvement occurred without consistent changes in heat pain threshold (all doses p > 0.16) or tolerance (all doses p > 0.27), and minimal effects on bottom-up pain facilitation (TSP) (all doses p > 0.43), supporting specificity to top-down modulatory mechanisms rather than bottom-up antinociception. Conversely, at 30 min post-opioid dose (240 min post-CBD), CBD 1200 mg was associated with worsened heat pain threshold relative to placebo (d’=−1.25, p = 0.006), with CBD 400 mg showing a similar trend (d’=−0.79, p = 0.066). CBD 1200 mg also showed a trend toward increased bottom-up pain facilitation (TSP) post-opioid dose (d’=0.78, p = 0.087), while top-down pain modulation (CPM) remained unchanged. Notably, this early post-opioid dose window occurs during CBD's absorptive phase (tmax ~4 hours), suggesting potential pharmacokinetic interactions. Tolerability was excellent with no serious adverse events, stable vital signs, preserved cognitive performance, and minimal hepatic enzyme changes (mean ALT change < 2 U/L).

Conclusions: Among persons with OUD and co-occurring chronic pain receiving OAT, acute CBD selectively enhanced top-down pain modulation (CPM) at trough opioid levels. The dose-dependent improvement in top-down pain modulation suggests CBD may help normalize specific pain modulation deficits associated with chronic pain and long-term opioid exposure. The transient increase in pain sensitivity during the early post-opioid window warrants further investigation as it they may reflect timing-dependent pharmacokinetic interactions, given that both CBD and methadone are CYP3A4 substrates and CBD affects gastric motility, potentially altering opioid absorption. These findings highlight timing relative to opioid dosing as a determinant of CBD's net analgesic effect and provide mechanistic support for CBD's therapeutic potential when appropriately scheduled. Future studies should extend post-opioid dose assessment window, incorporate serial pharmacokinetic sampling, and evaluate the effects of long-term CBD administration. Collectively, these findings support CBD’s analgesic potential in OUD with chronic pain while underscoring that when and how it is given may matter most—priorities for PK-informed confirmatory studies.

Keywords: cannabinoids, Methadone, psychophysics, chronic pain, inflammation, biomarker, Opioid addiction.

Disclosure: Jazz Pharmaceuticals, Other Financial or Material Support, Self, Ananda Scientific, Other Financial or Material Support, Self, Boehringer Ingelheim, Consultant, Self.

P603. Real-world use of the novel antipsychotic xanomeline/trospium chloride in a large academic medical center

Halide Bilge Turkozer, Ann K. Shinn, Dost Ongur

McLean Hospital, Harvard Medical School, Belmont, Massachusetts, United States

Background: Xanomeline/trospium chloride is a novel antipsychotic, approved by the FDA in September 2024. Xanomeline, a muscarinic receptor agonist, has a unique mechanism of action distinct from that of currently available antipsychotics which target dopamine D₂ receptors. This novel mechanism offers a promising alternative for patients with psychotic disorders who do not respond to or tolerate existing therapies. Given its recent FDA approval, real-world data remain limited. This study examines xanomeline/trospium chloride use during the first ten months after the FDA approval within the Mass General Brigham healthcare system, providing insights into its clinical use beyond controlled trial settings.

Methods: The Mass General Brigham (MGB) Research Patient Data Registry (RPDR) was used to conduct a retrospective analysis of adult patients prescribed xanomeline/trospium chloride from September 2024 to early August 2025. Patient demographics, total and mean prescription counts, prescription trends, and dosing patterns were assessed.

Results: 104 unique patients (42 female) were prescribed xanomeline/trospium chloride between September 2024 and early August 2025. There were 97 patients with outpatient prescriptions and 39 with inpatient prescriptions, with 7 patients receiving only inpatient prescriptions. The mean patient age was 40.8 (SD = 18.1, median age = 37), ranging from 18 to 94 years. There were no age or sex differences between patients receiving inpatient or outpatient prescriptions. The first prescription was about a month after FDA approval. The average number of prescriptions per patient was 2.40 (SD = 2.32), ranging from 1 to 12. Among patients receiving outpatient prescriptions, 55 patients received a single prescription and 42 received multiple prescriptions, with no significant differences in age or sex between the groups. After excluding patients who received their first outpatient prescription within the past month, 27 (12 female, mean age = 45.8, SD = 24.3) of the 79 patients with outpatient prescriptions were prescribed the starting dose (50 mg/20 mg) (n = 24) or the Starter Pack (n = 3) and did not receive a repeat prescription. 47 unique outpatients (18 female, mean age = 37.5, SD = 13.2) were prescribed the 100 mg/20 mg (n = 34) or 125 mg/30 mg (n = 25) capsules. Patients with a one-time starting dose outpatient prescription (50 mg/20 mg or the Starter Pack, n = 27, mean age = 45.8, SD = 24.3) were significantly older than those prescribed the highest dose (125 mg/30 mg, n = 25, mean age = 32.6, SD = 8.19) (p = 0.01), with no differences in sex.

Conclusions: In one of the largest academic healthcare centers in the US, xanomeline/trospium chloride adoption has been gradual but spanned a wide age range, with use most common among younger adults. About one-third of patients received only a one-time starting dose without a repeat prescription. Outpatients prescribed the highest dose regimen were younger than those received a one-time only starting dose prescription, suggesting better tolerability in younger patients. Limitations include the retrospective nature of the analysis limited to MGB records, which may miss care received elsewhere. Future work will include detailed chart reviews to examine if xanomeline-trospium chloride was used as a monotherapy or adjunctively, reasons for discontinuation, side-effect profiles, and real-world effectiveness.

Keywords: Antipsychotic Treatment, Schizophrenia, Antipsychotics, xanomeline.

Disclosure: Nothing to disclose.

P604. Ketamine for depression, but at what cost? a review of ketamine’s neurotoxic effects from preclinical and human studies

Samuel Wilkinson, Songjun Li, Kristina Kumpf, Julian Ripoll, John Krystal, Gerard Sanacora

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

Background: Ketamine, an N-methyl-D-aspartate antagonist, has emerged as an effective, off-label interventional treatment for a range of psychiatric conditions including treatment resistant depression. Despite robust evidence for efficacy, regulatory entities remain concerned about the neurotoxic risk of exposures in excess of 8 lifetime doses of 60mg for research studies in adults. However, as ketamine is used off-label for psychiatric applications, there is a high degree of variability in the dose and frequency of ketamine treatment in clinical settings, raising questions regarding the risk of neurotoxicity.

Methods: Here we examine ketamine’s neurotoxic potential across preclinical and clinical studies to inform clinicians and advise researchers of where critical knowledge gaps exist. We synthesized data from preclinical models, then integrated findings from human clinical trials of esketamine and observational studies in recreational users.

Results: Animal studies indicate that repeated or high dose subanesthetic ketamine resulted in consistent excitotoxic neuronal damage and lasting cognitive deficits, especially in younger subjects. Moderate, infrequent subanesthetic doses do not yield overt histopathology in animal models. In humans, observational studies in frequent high-dose (>1g/day) ketamine users show memory and executive-function impairments. In contrast, a large clinical trial of intranasal esketamine at doses up to 84mg, administered weekly or biweekly for several years, is associated with maintained or slightly improved higher cognitive function. Lower cognitive function (attention, processing speed) showed some worsening, especially in elderly patients; the significance of this finding is unknown. Direct comparisons of esketamine and off-label racemic ketamine at higher doses have not been done.

Conclusions: These evidence underscores the potential for neurotoxic effects when ketamine is used at doses or frequencies beyond those utilized in clinical trials, highlighting a critical need for robust, longitudinal research. Clinicians are advised to exercise caution, particularly when prescribing racemic ketamine off-label at doses significantly higher than those used in clinical trials. When deviating from this in clinical practice, strong consideration should be given to conducting repeated cognitive assessments. Funding agencies should incentivize preclinical researchers to conduct studies that further elucidate the threshold of ketamine’s neurotoxicity.

Keywords: Ketamine, Esketamine, Neurotoxicants, Cognitive impairments.

Disclosure: LivaNova, Consultant, Self, Mind Medicine, Consultant, Self.

P605. Neurotherapeutic target GPR52 constitutively activates multiple G proteins and beta-arrestin2: insights for orphan receptor and GPR52 signaling

Jones Addiah-NIckson, Ryan Murphy, John Allen

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

Background: GPR52 is an orphan class-A G protein-coupled receptor and recently identified schizophrenia risk gene. GPR52 is exclusively expressed in the human brain, primarily in the striatum in dopamine D2 expressing medium spiny neurons. This unique expression profile has distinguished GPR52 as a promising drug target for both neurological and psychiatric disorders. GPR52 is a current CNS target of interest at several pharmaceutical companies (see Ali et al Drug Discov Today. 2024 Apr;29(4):103922). We have previously created novel GPR52 selective agonists which attenuate psychostimulant-induced behavior in rodents, and some of these agonists have reduced activity to engage GPR52-Beta-arrestin signaling to reduce receptor desensitization (see Murphy et al, J Med Chem. 2024 Jun 13;67(11):9709–9730). Previous studies by our group, and others, have also shown that GPR52 has unusually high constitutive activity, which activates the Gs/adenylyl cyclase/cAMP pathway. However whether additional GPR52 signaling occurs is not known and has not been investigated. In this project, we use three robust assays to broadly elucidate GPR52 signaling mechanisms.

Methods: GPR52 mediated elevations in cAMP levels in HEK293 cells were assessed using the GloSensor assay. The TRUPATH BRET assay was also performed using HEK293 cells to assess potential human GPR52 G protein signaling through fourteen different heterotrimeric G protein alpha pairs. The Tango Beta-arrestin assay using HEK293-HTLA cells was performed to assess human GPR52 activity for Beta-Arrestin2. CRISPR/Cas9 knockout of Gs and Golf proteins was performed to create HEK293 cells lacking protein expression of these G proteins, and GPR52 signaling was compared using parental HEK293 or the Gs/Golf knockout cells. Site directed mutagenesis of extracellular loop 2 domains of GPR52 was done to create several receptor mutants and the constitutive signaling of mutants versus wildtype human GPR52 were assessed.

Results: Transfection of GPR52 into HEK293 cells profoundly elevated cAMP levels proportionally with increasing receptor plasmid expression. This increase of cAMP was eliminated by CRISPR/Cas9 knockout of Gs and Golf proteins, or by mutating residues in the GPR52 extracellular loop 2 to alanine. Re-expression of Gs or Golf into Gs/olf knockout cells showed that GPR52 primarily uses Gs to elevate cellular cAMP. Using the TRUPATH BRET-based assay, GPR52 transfection resulted in significantly reduced alpha/gamma BRET ratios for select G proteins, including Gs long, Gs short, gustducin, G12, and G15, indicating GPR52 basally activates these G proteins. GPR52 did not robustly activate other G proteins including Gi1, Gi2, Gi3, G11 or G13. Interestingly, GPR52 expression versus an empty vector control, significantly increased the BRET ratio for Gz and Gq, suggesting GPR52 signaling may basally inactivate these G proteins. Additionally, GPR52 profoundly elevated Beta-Arrestin2 recruitment and signaling using the Tango assay.

Conclusions: Taken together, these studies reveal that GPR52 has high constitutive activity and simply introducing this orphan receptor into cells activates multiple G proteins. GPR52 readily activates Beta-Arrestin2, Gs, gustducin, G12 and G15 and may inactivate Gz and Gq. These studies point to complex GPR52 constitutive signaling for both heterotrimeric G protein and Beta-Arrestin pathways. Ongoing studies are examining these GPR52 signaling mechanisms in response to agonists and in striatal neurons. The complex constitutive signaling by GPR52, including its high elevation of cAMP, also provides a mechanistic basis to further study GPR52 and its association with schizophrenia.

Keywords: orphan receptors, drug target, GPCR, cAMP signalling, Schizophrenia novel treatment.

Disclosure: Maplight Therapeutics, Contracted Research, Self.

P606. Emergent properties of metabotropic glutamate receptor heterodimerization confers unique pharmacology

Justin Steinfeld, Madeline Laramee, Jonathan Javitch

Columbia University and New York State Psychiatric Institute, New York, New York, United States

Background: Pharmacological interventions for childhood-onset schizophrenia and neurodevelopmental conditions, such as autism, have notable side effects and limited efficacy. Growing evidence from neuroimaging and genetic studies has implicated the metabotropic glutamate receptor (mGluR) as a promising pharmacological target for these illnesses. mGluRs appear to likely form both homodimers and heterodimers in vivo, creating the possibility for modulation of specific glutamatergic circuits. In this study, we use a technique developed in the Jonathan Javitch lab called CODA-RET (Complemented Donor Acceptor-Resonance Energy Transfer) to examine properties of Group I receptors (mGlu1 and mGlu5) homo- and heterodimers with agonists and antagonists in order to understand and develop new pharmacological targets for future treatments of psychiatric disorders.

Methods: CODA-RET places a split luciferase(HiBit and LgBit) on the specific mGluRs, creating a luminescent signal only when the two mGluRs pair together. mGluR activation is measured when a fluorescently tagged-G alpha is recruited to the receptor yielding a BRET (bioluminescence resonance energy transfer) signal. This assay is performed by transfecting HEK293T cells with these modified mGluRs and G alpha and measuring BRET in 96-well plates on a Lumistar plate reader. We fit the varied drug concentrations to a sigmoidal dose-response relationship by nonlinear regression using Prism 10. At least 3 independent experiments will be performed with triplicate samples to produce error bars represented as standard error of the mean. Half-maximum effective concentration (EC50) values derived from the non-linear regression will be compared via unpaired t-test.

Results: We were able to create functional mGlu1 and mGlu5 protomers that were modified to have HiBit and LgBit yielding delta BRET ratios of 0.08 to 0.12. Here we show that Group I receptors are poor at both cis- and transactivation, however the heterodimer confers unique properties on the system showing asymmetric Gq activation via mGlu1 leading to unique properties in pharmacology. This bias towards mGlu1 appears to be mediated by the transmembrane domain. Additionally, we find specific point mutations in mGlu1 that prevent PAM, 4578, or NAM, FITM, binding and that these drugs work in cis.

Conclusions: Previous studies have attempts to look at the properties of Group I receptors have relied on downstream calcium mobilization assays which are more sensitive to expression levels and overall less sensitive due to amplification effects. Our system has a large dynamic range that specifically measures G protein recruitment allow for very sensitive measure and differences in the dimers. We are able to show that Group I dimers act in unique ways compared to previously studies mGluR's. These findings serve as a primer to work under way showing that understanding these properties are in fact critical for preventing toxicity in preclinical stages of these drugs.

Keywords: Neuropharmocology, metabotropic glutamate receptor 5 (mGluR5), Autism, Schizophrenia- Novel Treatment.

Disclosure: Nothing to disclose.

P607. Computational and molecular characteristics of DRD2 mutations on the risperidone-DRD2 receptor dynamics

Ranran French, Akshi Pant, Ryan Gumpper

UNC at Chapel Hill, Dept of Psychiatry, Chapel Hill, North Carolina, United States

Background: Schizophrenia is among the top 20 causes of disability worldwide, affecting approximately 3.7 million adults in the United States alone, creating over $300 billion in economic costs. Despite advancements in treatment, schizophrenia remains challenging to manage, with a recovery rate as low as 13%. A significant barrier to achieving remission is the variability in individual responses to antipsychotic medications, which necessitates a trial-and-error approach in drug selection that can last months. During this period, patients continue to experience poorly controlled psychotic symptoms. Risperidone is a second-generation antipsychotic and is often used as a first-line treatment for schizophrenia due to its comparatively good efficacy, lower risk for extrapyramidal symptoms, and availability in multiple formulations. Like most antipsychotics, risperidone exerts its antipsychotic effects by working as a DRD2 and 5-HT2A antagonist. In this study, we combine both in silico and in vitro techniques to examine 20 frequent mutations of the human DRD2 receptor and their effects on risperidone-DRD2 receptor dynamics, aiming to inform drug selection based on pharmacogenetic information.

Methods: The crystal structure of the DRD2-risperidone complex has been previously solved by Wang et al. and exhibits specific molecular interactions between the ligand and the receptor. In this study, we investigated the potential effects of 20 frequently observed mutations in the human DRD2 receptor on the protein dynamics as well as the risperidone binding mode. Since most of these mutations reside on experimentally unresolved disordered loops, we utilized AlphaFold2 to predict their structure and then performed Molecular Dynamics (MD) simulations. Simultaneously, we conducted functional cell assays on these mutations. Using Bioluminence Resonance Energy Transfer (BRET), we quantitatively measured the functional activation of these DRD2 mutants to dopamine in living cells.

Results: In the MD simulation of the wild-type DRD2 receptor, we found three distinct risperidone poses. Among the 20 mutations we investigated, we observed varied distributions of these three poses. We found that some mutations are more stabilizing for the risperidone-DRD2 receptor dynamics while others are destabilizing. Within these 20 mutations, we have identified two mutations in vitro, namely V200M and R145H, that completely attenuate dopamine’s ability to activate DRD2.

Conclusions: DRD2 receptor genetic variances have shown a distinct impact on risperidone-DRD2 receptor interaction.

Keywords: precision psychiatry, Schizophrenia, Dopamine, Antipsychotics, antipsychotic drug.

Disclosure: Vivid Vivion, Inc, Consultant, Self.

P608. Differential antagonism of Mu opioid receptor agonists in vitro

Shawn Flynn, John Traynor

University of Michigan, San Antonio, Texas, United States

Background: The opioid epidemic remains a major public health challenge, with ~80,000 opioid overdose deaths reported in 2023. Over 93% of these deaths involved a synthetic opioid, and some clinical reports suggest that larger doses of opioid antagonists might be needed to reverse overdoses involving these drugs. While several factors could result in this apparent difference in humans, some recent preclinical studies suggest that larger doses of antagonists are in fact needed to reverse the effects of some “superpotent” synthetic opioids (e.g., carfentanil). It has been proposed that this is related to the binding kinetics of the agonist involved, but this unexpected difference in antagonist effect across different agonists is not fully understood. Additionally, the role of the pharmacological and physiochemical properties of the antagonist in this phenomenon remains unexplored. This study characterized interactions between multiple opioid receptor agonists and antagonists with varying physiochemical properties in two different in vitro assays to test the hypothesis that the apparent affinities of antagonists will be lower when determined using highly potent, slowly dissociating agonists, independent of structural class.

Methods: Two events following agonist occupation of the mu opioid receptor namely inhibition of forskolin-stimulated cAMP accumulation and recruitment of β-arrestin were measured in Chinese hamster ovary cells stably expressing the human mu opioid receptor. In each assay, concentration response curves for the opioid receptor agonists DAMGO, met-enkephalin, morphine, fentanyl, as well as the superpotent agonists carfentanil, BU72, and etonitazene, were determined in the presence of varying concentrations of the opioid receptor antagonists naloxone, naltrexone, CTOP, methylnaltrexone, and naloxone methiodide. Rightward shifts in agonist concentration response curves were quantified and plotted as a function of antagonist concentration to determine a pA2 value, the apparent affinity, for each antagonist in the presence of each agonist. Apparent affinities for each antagonist were then compared across agonists.

Results: All mu opioid receptor agonists studied concentration-dependently reduced forskolin-stimulated cAMP accumulation with the following rank order of potency: carfentanil > BU72 > etonitazene > fentanyl > met-enkephalin > DAMGO > morphine. All agonists tested increased β-arrestin recruitment with a similar rank order of potency, but differences in maximum effect. For β-arrestin recruitment, DAMGO, met-enkephalin, carfentanil, and BU72 were full agonists while fentanyl and morphine were partial agonists, producing 75 and 35% of the maximal effect, respectively. As expected, the addition of each opioid antagonist resulted in concentration-dependent rightward shifts in agonist concentration response curves. However, the magnitude of these shifts varied across agonists and thus, the pA2 values for each antagonist differed based on the agonist used. In general, apparent potencies of antagonists were lower when determined using carfentanil, BU72, and etonitazene than other agonists, but the magnitude of this difference varied across antagonists. For example, in β-arrestin recruitment pA2 values for naloxone were approximately 50-fold lower for the superpotent agonists compared to other agonists, while the pA2s for methylnaltrexone differed by 100-fold. Interestingly, there was greater variance in pA2 values determined in the cAMP assay compared with β-arrestin, though antagonist pA2 values determined using the superpotent agonists were still lower compared with other agonists like fentanyl or morphine.

Conclusions: These findings are consistent with the hypothesis that the apparent affinities of antagonists would be lower when determined using highly potent, slowly dissociating agonists and that this effect is independent of structural class. This supports the theory that agonist binding kinetics underly this difference in agonist effects. Ongoing expansion of this project will determine the role of antagonist binding kinetics to determine which properties might make one antagonist more effective than others. As a difference was observed across assays, the role of intracellular receptors/receptor internalization in these observed differences will also be evaluated. Further understanding of these interactions will allow for targeted design of improved therapeutics for opioid overdose and likely improve understanding of drug-drug-receptor interactions at GPCRs in general.

Keywords: Opioids, Fentanyl, In vitro.

Disclosure: Nothing to disclose.

P609. Cross-disease transcriptomic comparison of opioid and Alzheimer’s mouse models reveals disease-dependent effects of (2R,6R)-hydroxynorketamine on pathological circuits

Anna Onisiforou, Panos Zanos, Polymnia Georgiou

University of Cyprus, Nicosia, Cyprus

Background: Opioid use disorder (OUD) and Alzheimer’s disease (AD) are distinct brain disorders that share overlapping mechanisms, including synaptic dysfunction and neuroinflammation. (2R,6R)-Hydroxynorketamine (HNK), a ketamine metabolite with rapid-acting antidepressant properties, has shown promise in treating affective and cognitive symptoms in mouse models; however, its molecular actions in different disease states remain unclear.

Methods: We compared hippocampal transcriptomic responses to HNK in two preclinical models: chronic morphine exposure with prolonged abstinence (OUD) and APPswe/PS1ΔE9 transgenic mice (AD). In the OUD model, adult male C57BL/6J mice (n = 8 HNK-treated vs. 8 saline controls) underwent a 5-day escalating-dose morphine paradigm followed by 28 days of drug-free abstinence; mice then received a single intraperitoneal injection of HNK (10 mg/kg) or saline, and ventral hippocampi were collected 24 h later for RNA sequencing. In the AD model, adult male and female APPswe/PS1ΔE9 mice and wild-type littermates received daily HNK (0.5 mg/kg, i.p.) or saline for 14 days before whole hippocampi were collected for RNA sequencing. Differential gene expression was analyzed using DESeq2, with reversal defined as genes upregulated in one contrast and downregulated in another. Functional enrichment was performed using clusterProfiler with Gene Ontology (GO) and KEGG annotations (adjusted p < 0.05).

Results: Enrichment analyses confirmed that hippocampal upregulated genes in APP/PS1 mice compared to controls were strongly associated with immune activation and Toll-like receptor signaling (adjusted p < 0.05), consistent with microglia-driven neuroinflammation. HNK selectively normalized a subset of these immune–vascular processes and reversed disease-associated dysregulation of a vascular inflammation gene module, including Angptl4 (log₂FC = −1.53), Selp (−7.13), Apob (−7.37), and Ccr9 (+3.47), which are functionally connected through protein–protein interaction networks. Their coordinated reversal suggests that HNK restores neurovascular homeostasis by modulating an integrated inflammatory–vascular network rather than isolated targets. Cross-disease comparisons showed no overlap between HNK-responsive genes in the AD (n = 89 reversal DEGs) and OUD (n = 55 reversal DEGs) models, indicating strong disease-specificity. In healthy controls, HNK elicited minimal and inconsistent transcriptomic changes, with only four shared downregulated genes across studies.

Conclusions: HNK exerts disease-restricted actions, reversing AD-specific neurovascular dysfunction, engaging distinct pathways in OUD, and showing negligible effects in healthy brains. These findings provide the first cross-disease comparison of HNK’s transcriptomic effects and highlight the importance of evaluating neurotherapeutics in disease-relevant models to uncover clinically meaningful mechanisms of action.

Keywords: (2R,6R)-hydroxynorketamine, Alzheimer’s Disease, opioid use disorder, brain transcriptomics, Hippocampus.

Disclosure: Nothing to disclose.

P610. Selective D4R ligands display varying effects on ADHD-like behaviors in a rat model

Comfort Boateng, Tian Li, Alkhatib Mohammad, Samantha Andino, Ashley Amis, Thomas Keck, Ike de la Peña

High Point University, Fred Wilson School of Pharmacy, High Point, North Carolina, United States

Background: The dopamine D4 receptor (D4R), a G protein-coupled receptor, is predominantly expressed in the prefrontal cortex which plays an important role in the brain such as cognition, attention, and reward. These features make D4Rs attractive therapeutic targets for neuropsychiatric disorders such as substance use disorders, schizophrenia and attention-deficit/hyperactivity disorder (ADHD). However, therapeutic development has been hampered by the limited availability of selective and efficacious D4R ligands.

Methods: A series of new D4R compounds were recently synthesized, purified and analytically characterized including CHN combustion elemental analysis. In vitro binding affinities were assessed using [3H]N-methylspiperone radioligand binding on membranes from HEK293 cells expressing dopamine D2-like receptors (D2R, D3R, D4R), along with functional and pharmacokinetic studies. Selected compounds were investigated for their behavioral effects in adolescent (4–6 weeks) spontaneously hypertensive (SHR/NCrl) rats, a validated model of ADHD, and Wistar rats as controls. SHR/NCrl show typical ADHD-like behaviors, including hyperactivity (elevated locomotor activity in open field tests), inattention (reduced spontaneous alternation in the Y-maze), impulsivity (delay discounting), and cognitive dysfunction (decreased novel object preference).

Results: We have identified several compounds with nanomolar D4R binding affinity and excellent D2-like subtype selectivity with displaying antagonist and partial agonist profiles with desirable results in plasma half-life and brain exposure measures. High-efficacy partial agonist FMJ-01-38 (5–10 mg/kg) dose-dependently reduced hyperactivity, improved attention and cognition, and decreased impulsive choice in SHR/NCrl rats. In Wistar rats, FMJ-01-38 improved attention at a dose of 10 mg/kg and cognition at 5 mg/kg. Antagonist FMJ-01-54 did not reduce hyperactivity but enhanced attention and cognition in SHR/NCrl rats, with no effects in Wistar rats.

Conclusions: The newly developed compounds displayed binding and selective affinity for D4R, with favorable pharmacokinetics in rats. Selected D4R compounds improved core ADHD-like behaviors in SHR/NCrl, with FMJ-01-38 also enhancing Wistar rats’ cognitive functions. The differing behavioral results may stem from distinct partial agonist and antagonist actions at D4Rs, involving varied signaling pathways. These new compounds could contribute to research in targeted drug discovery and support further investigation into the role of D4Rs in neuropsychiatric disorders. Future studies aim to elucidate these mechanisms and evaluate their therapeutic potential.

Keywords: Triazole, ADHD, Dopamine D4 receptor, Partial agonist Ligands, Antagonist Ligands.

Disclosure: Nothing to disclose.

P611. Exploring the effects of calpain and SARM1 inhibition in preventing neurodegeneration in human and rodent neuropathic pain models

Cameron Pernia, Geraldine Edu, Greg Joseph, Kymmy Lorrain, Mike Poon, Dan Lorrain

Contineum Therapeutics, San Diego, California, United States

Background: The prevalence of neuropathic pain in adults is reported to be as high as 10% in the US, with certain demographics having higher risk based on age, diabetes diagnosis, cancer history, and those with physical trauma or nerve injury. Chemotherapy induced neuropathy (CIPN) is a subtype associated with axonal degeneration, metabolic dysfunction, and ion-channel dysregulation; characteristics that are shared across other pain subtypes. Chemotherapies like vincristine depolymerize microtubules and cause the degeneration of C-fibers in the epidermis that pathogenically manifest neuropathic pain. Mechanistically, vincristine causes an elevation in intracellular calcium that activates the protease calpain, leading to impaired axonal transport and cytoskeletal destruction. As part of our efforts to explore other calpain dependent mechanisms of axonal protection, we wanted to investigate whether an upstream regulator, Sterile Alpha and TIR Motif Containing 1 (SARM1) may be as efficacious a target.

SARM1 is a metabolic sensor that is triggered by an increase in NMN:NAD+ as seen, for example, in CIPN. As a result, SARM1 further hydrolyzes NAD+ in a feed-forward manner, leading to metabolic failure. SARM1 also converts NAD+ into cyclic ADP-ribose (cADPR), which can trigger calcium release from intracellular stores, leading to calpain activation and axon degeneration. The objective of this study was to explore the activity and efficacy of SARM1 and calpain inhibition across in vitro and in vivo models of neurodegeneration and pain.

Methods: SARM1 hydrolase activity was measured via a fluorogenic assay that measured the conversion of etheno-NAD, a substrate of SARM1, into etheno-ADPR in the presence of SARM1 protein and inhibitors. Fluorescence was measured via a plate reader, as a direct quantification of SARM1 activity in the presence of inhibitors.

Human induced pluripotent stem cell (iPSC) derived sensory neuron cultures were generated following Deng et als procedure, relying on a combination of small molecules (CHIR98014, A 83-01, DBZ, PD173074, PD0332991), supplements (N2 and B27), and recombinant proteins (BDNF, GDNF, NGF, NT-3) over the course of 60 days and were characterized by BRN3A, PRPH, NF200, and TUJ1 protein staining.

Cultures were pre-treated with SARM1 or calpain inhibitors then treated with vincristine for 24h to induce neurite degradation. The SARM1 specific neurotoxin vacor was also tested in the presence of inhibitors. After either insult, cells were fixed and stained for two cytoskeletal proteins, TUJ1 and PRPH, to assess neurite degeneration via confocal microscopy. Neurite morphology was measured using NIS Elements software based off TUJ1/PRPH fluorescence (n = 15 per group) or total fluorescence was measured on a plate reader (n = 3/group).

SARM1 inhibition was tested in mouse models for neurodegeneration. In a sciatic nerve transection (SNT) model, mice were dosed (n = 4/group, IP, BID) with inhibitor 1h before surgically removing ~1mm of nerve. Plasma was collected 48h after SNT.

Vincristine-induced CIPN was tested by exposing mice (n = 8/group) to two 1.5 mpk IP doses of vincristine over 3 days to induce degeneration. Mice were then dosed with SARM1 or calpain inhibitors for 20 days. Axonal protection was evaluated histologically by counting PGP9.5+ intraepidermal nerve fibers (IENFs) in the footpads. For SARM1, thermal pain sensitivity was also evaluated with a Hargreaves instrument.

Vacor induced neurodegeneration was also tested by dosing mice (n = 4/group) with SARM1 inhibitor 1h before and 7h after a 30mpk vacor (PO) exposure. Plasma was collected after 24h.

Plasma from all experiments were analyzed for neurofilament-light (NF-l) protein, a clinically translatable endpoint for neurodegeneration, via ELISA. Differences across groups were tested with ANOVA.

Results: We validated that our human iPSC derived sensory neuron cultures expressed PRPH, TUJ1, MAP2, BRN3A, and NF200 via immunofluorescence microscopy. Using either vincristine (CIPN insult) or vacor (SARM1 activator) we observed that calpain or SARM1 inhibition decreased neurodegeneration.

In our mouse SNT model, while SARM1 inhibition at the highest dose tested (30 mpk) prevented elevated NF-l plasma levels, NF-l levels at lower doses were surprisingly elevated. Similarly, mice treated with vacor and inhibitor showed complete rescue of plasma NF-l levels at higher concentrations, with lower inhibitor doses increasing NF-l plasma levels.

Calpain and SARM1 inhibition were able to rescue nerve fibers in our vincristine induced CIPN mouse model. Further, SARM1 inhibition also showed efficacy in the Hargreaves test, with vincristine treated mice displaying increased paw withdrawal latency, which could be blocked by an inhibitor.

Conclusions: Prior studies have shown that SARM1 activates calpain-dependent neurodegeneration, which we explored via in vitro and in vivo pain models. Calpain and SARM1 inhibition decreased vincristine induced degeneration in our human sensory neurons. Both calpain and SARM1 inhibition could rescue the effects of neuron degeneration in our in vivo mouse models. However, low doses of SARM1 inhibitor increased plasma NF-l levels, implying SARM1 activation, which matches previous SARM1 inhibitor reports. These findings further validate calpain as a target for pain-associated disorders and highlight SARM1 as a possible target, provided inhibitors can be designed without the adverse low dose effects.

Keywords: neuropathic pain, Neuropharmocology, Sensory neurobiology.

Disclosure: Contineum Therapeutics, Employee, Self.

P612. Inhibition of cea-dynorphin neurons is aversive in mice with chronic pain

Merel Dagher, Catherine Cahill

UCLA, Los Angeles, California, United States

Background: Chronic pain is a debilitating condition that affects more than 50 million Americans annually. The underlying neurobiological circuits that mediate chronic pain states are not fully elucidated, contributing to the limited treatment options. The dynorphin neurons in the central amygdala (CeA) are known to contribute to pain-induced aversion and the CeA is a central hub for integrating pain and stress states. We investigated if inhibiting or activating the CeA dynorphin neurons was sufficient to produce negative reinforcement using chemogenetics and conditioned place preference. We hypothesized that mice in a chronic pain state will prefer the chamber associated with clozapine-N-oxide (CNO), a designer drug used to inhibit dynorphin neurons expressing an inhibitory designer receptor.

Methods: We transfected male and female transgenic mice (dyn-cre+ and dyn-cre-) with AAV-hSyn–DIO-hM4D-mCherry to express the inhibitory designer receptor in dynorphin neurons or AAV-hSyn-DIO-mCherry as a control; the same mice were transfected with AAV-hSyn-DIO-ChR-yfp to express channelrhodopsin in dynorphin neurons or AAV-hSyn-DIO-yfp as a control. For the dyn-cre- mice, we transfected non-cre dependent AAVs of the same constructs. Two weeks after the viral injection, we conducted either a chronic constriction injury or sham surgery to induce neuropathic pain. Two weeks following nerve injury, mice were conditioned with either CNO or saline using a three-chamber conditioned place preference unbiased paradigm counter-balanced for day and chamber. Mice were conditioned for six days followed by a post condition test (drug free) and a state-dependent test (drug onboard). Following conditioned place preference, mice were run through a real-time place preference test using optogenetic stimulation.

Results: We found chemogenetic inhibition of CeA neurons in chronic pain mice produced a place preference, whereas inhibition of CeA-dynorphin neurons produced a place aversion. This suggests that inhibiting the CeA is negatively reinforcing and alleviates the negative affect induced by chronic pain state while dynorphin neuron inhibition in the CeA is aversive. Moreover, we found that mechanical allodynia was not changed, as CNO did not alter mechanical threshold compared to baseline values using von frey. Complementary studies reveal that activating these dynorphin neurons in the CeA using optogenetics produces a real-time place preference.

Conclusions: Central amygdala dynorphin neurons contribute to pain aversion but not the sensory component of pain. This study adds to the knowledge of the underlying circuits of chronic pain and the potential for the kappa/dynorphin system to be a neurobiological target for its treatment.

Keywords: chronic pain, neuropathic pain, central amygdala, Dynorphin, kappa opioid receptor.

Disclosure: Nothing to disclose.

P613. Poster Withdrawn

P614. Chronic exposure to combined delta-9 tetrahydrocannabinol and cannabidiol disrupted sensorimotor gating and dysregulated transcriptomic profiles in heterogeneous stock rats

Samantha M. Ayoub, Daniel De Siqueira Lima, Sunitha Vemuri, Michael N. Noback, Wei Ling Lim, Neal A. Jha, M. Melissa Flesher, Abraham A. Palmer, Arpi Minassian, Maria Cecilia G. Marcondes, Jared W. Young

University of California San Diego, San Diego, California, United States

Background: Chronic cannabis exposure disrupts sensorimotor gating in humans, a key sensory information-processing function that reflects the brain’s ability to sort and inhibit irrelevant information. While the cannabis plant contains hundreds of phytocannabinoids, the relative contribution of each of these compounds and the mechanisms leading to sensorimotor disruptions have not been defined. Here, we focused on the two primary and most abundant cannabis constituents, delta-9 tetrahydrocannabinol [THC] and cannabidiol [CBD], to test their independent and combined effects on sensorimotor gating in outbred genetically diverse rats, as measured by prepulse inhibition (PPI) of the acoustic startle response, and in relation to relevant transcriptional patterns in the brain.

Methods: Heterogeneous stock (HS) rats (n = 72; 50% female) were baseline-matched on percentage (%) PPI into one of four treatment groups: 1) Vehicle (VEH); 2) THC (3 mg/kg); 3) CBD (3 mg/kg), and 4) THC+CBD (3 mg/kg each), administered once daily via intraperitoneal injection. The PPI task exposed animals to trial types enabling the calculation of %PPI, including a loud noise stimulus (120 dB) alone and softer prepulse stimuli (68, 71, and 77 dB) that preceded the 120 dB noise. PPI was re-tested 30 min after acute (1 day) and chronic (16 day) treatment exposures, then again 24-hours after treatment withdrawal. Behavioral data were assessed via 3-way analyses of variance (ANOVA) using sex and drug as between-subjects’ factors and prepulse intensity as a within-subjects factor. Sphericity violations detected (Mauchly’s) were corrected with Greenhouse-Giesser. A subset of rats were perfused at the end of testing and the brains were harvested and flash-frozen for RNA extraction and transcriptome analysis using bulk RNAseq. Samples were subjected to poly(A)+ selection, cDNA synthesis and deep sequencing on an Illumina NovaSeq followed by normalization by counts for sequencing depth and clustering analysis. Systems biology and pathway enrichment were performed in iPathwayGuide and Cytoscape softwares.

Results: Acute THC mildly decreased %PPI (drug: [F(3,62) = 2.264, p = 0.090]) vs. VEH and isolated CBD (ps < 0.05). On the other hand, chronic THC+CBD produced robust deficits in %PPI (drug: [F(3,62) = 5.4, p = 0.002]) relative to all other treatment groups (ps < 0.01). These effects of combined THC+CBD were observed alongside numerous transcriptomic alterations linked to neuroactive ligand-receptors (e.g., upregulated Npy2r, Oxtr, Calcrl; downregulated Lpar3) relative to VEH. Moreover, transcriptional differences in cell adhesion processes (e.g., upregulated Cd22, Icam2, Slitrk6, Rt1-Db1), arachidonic acid metabolism (upregulated Pla2g3, Akr1c15), and metabolic pathways (upregulated Pla2g3, Acer2, Elovl7, Akr1c15, Ears2, Ak8, Rdh16, Crppa; downregulated Hsd17b3, Cyp11a1, Tha1, Nmnat1, Acsm5) relative to VEH were also observed. Interestingly, these gene changes were observed only when THC+CBD were administered in combination, but not when either compound was administered alone. There were no behavioral differences observed between groups at the 24-hr withdrawal test, though significant differences occurred in genes involved in inflammation and its regulation (upregulated Nlrc4, Tlr4, BMpr1b, Tr2; downregulated Gnat2, Dnase1, Fut7, Alox15, Ptger2, Nfkbia, Nr1d1).

Conclusions: Acute, but not chronic, THC produced a mild PPI impairment suggesting a short-lived THC effect on sensorimotor gating. In contrast, chronic THC+CBD produced robust PPI deficits, not observed with the administration of THC or CBD alone. Chronic THC+CBD exposure altered neuroactive ligand-receptors previously associated with sensorimotor gating processes (e.g., neuropeptide Y, oxytocin receptor and lysophosphatidic acid receptors) which may explain poorer PPI in these animals. While the impact of chronic THC+CBD on behavior was short-lived, and not observed at acute withdrawal, withdrawal produced important changes consistent with enhanced inflammatory profiles, independent from PPI behavior. Our data suggest that combined THC and CBD may drive poorer PPI observed in chronic cannabis users, potentially explained by changes in transcriptional patterns in the brain relevant to behavior and inflammation.

This work was funded by: P30DA060810, 1R01DA059344, 1R01DA051295.

Keywords: cannabinoid, translational psychiatry, neurotranscriptomics, cognition.

Disclosure: Nothing to disclose.

P615. Discontinuation of selective serotonin reuptake inhibition produces a rebound of the suppressed firing activity of rat norepinephrine neurons

Naomichi Okamoto, Mostafa El Mansari, Pierre Blier

University of Ottawa Institute of Mental Health Research, Ottawa, Canada

Background: A selective serotonin (5-HT) reuptake inhibitors (SSRI) discontinuation syndrome has been well documented following their abrupt cessation, especially with medications that have a shorter half-life than fluoxetine [1]. These symptoms may include worsened mood, irritability, agitation, dizziness, confusion headaches, nervousness, and sensation of electrical shocks. Prolonged administration of SSRIs markedly decrease in a sustained fashion the firing rate of norepinephrine (NE) neurons in the rat locus coeruleus (LC) by about 50% [2]. In patients with major depressive disorder, there is a corresponding 50% decrease of the main metabolite of NE, MHPG, in their cerebrospinal fluid following a 6 to 8 week treatment with the SSRI fluoxetine or fluvoxamine, which is correlated with a decrease in 5-HIAA, the main metabolite of 5-HT [3]. It was thus postulated that the SSRI discontinuation syndrome could be due, at least in part, by the rapid release of the serotonin inhibitory tone on NE neurons.

Methods: Experiments were performed in male Sprague-Dawley rats anesthetized with chloral hydrate. In vivo electrophysiological recordings of NE neuronal activity were carried out in the LC in rats implanted with an osmotic minipump delivering either saline or the SSRI escitalopram (10 mg/kg/day, subcutaneously) for 14 days and recordings were carried out with the minipump delivering the drug or 48 hours after its removal to allow drug elimination. Neuronal firing was analyzed for changes in spike frequency and burst activity in groups of 5 to 7 rats, recording at least 15 neurons per rat.

Results: The NE neuronal firing rate differed significantly among groups (control: 1.87 ± 0.16, escitalopram: 1.14 ± 0.21, washout: 2.47 ± 0.18 Hz). Pairwise comparisons remained significant after Holm’s correction (control vs. escitalopram, p = 0.024; control vs. washout, p = 0.029; escitalopram vs. washout, p < 0.001). The percentage of neurons exhibiting burst activity only showed a nominally significant increase of the control–washout difference (control: 7%, escitalopram: 5%, washout: 14%, p = 0.04), however not after Holm’s correction (p = 0.13).

Conclusions: A robust decrease in the spontaneous firing rate of NE neurons following sustained administration of escitalopram was replicated for a fifth time in our laboratory. Strikingly, the firing rate of NE following the elimination of the drug was more than doubled that observed in rats still receiving the SSRI. This phenomenon appears to be akin to the rebound activity of NE neurons during opiate withdrawal [4]. It is concluded that an abrupt lifting of the inhibitory influence of 5-HT on NE neurons could account for some of the well-documented SSRI discontinuation symptoms.

References:

[1] Rosenbaum JF et al. Selective serotonin reuptake inhibitor discontinuation syndrome: a randomized clinical trial. Biol Psychiat 44:77-87, 1998.

[2] El Mansari et al. Long-term administration of cariprazine increases locus coeruleus noradrenergic activity and serotonin1A receptor neurotransmission in the hippocampus. J Psychopharmacol 34:1143-1154, 2020.

[3] Sheline et al. Correlated reductions in cerebrospinal fluid 5-HIAA and MHPG concentrations after treatment with selective serotonin reuptake inhibitors. J Clin Psychopharmacol 17:11-14, 1997.

[4] Rasmussen et al. Opiate withdrawal and the rat locus coeruleus: behavioral, electrophysiological, and biochemical correlates. J Neurosci 10:2308-2317, 1990.

Keywords: SSRI treatment, discontinuation, norepinephrine, locus coeruleus, firing rate.

Disclosure: Janssen, Honoraria, Self, Otsuka, Grant, Self, Allergan, Grant, Self, Eisai, Advisory Board, Self, Idorsia, Advisory Board, Self, Lundbeck, Advisory Board, Self.

P616. Potent protein-protein interaction modulators targeting reward-encoding medium spiny neuron firing rates

Aditya K. Singh, Cun Zhang, Zahra Haghighijoo, Hamid Teimouri, Jully Singh, Haiying Chen, Marijn Lijffijt, Jia Zhou, Fernanda Laezza

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

Background: The nexus of protein:protein interactions (PPI) organized around the voltage-gated Na+ (Nav) channels is the molecular determinant of neuronal excitability and a converging node of neuroplasticity associated with neuropsychiatric disorders. Recent studies have determined that dynamic regulation of nucleus accumbens (NAc) medium spiny neuron (MSN) firing rates in response to rewards is precisely modulated by the PPI complex between fibroblast growth factor 14 (FGF14) and the Nav1.6 channel. Compound 1028, a highly selective, brain-permeant, modulator of the FGF14/Nav1.6 complex, shifts the voltage sensitivity of FGF14/Nav1.6-mediated Na+ currents and selectively enhances reward-encoding MSN firing rates, effectively restoring motivation to obtain rewards in operant behavioral testing of low-motivation states. This work positions the FGF14/Nav1.6 complex as a validated target for first-in-class therapeutic development aimed at correcting aberrant MSN firing, a key mechanism underlying deficient reward processing and anhedonia in major depressive disorder (MDD).

Methods: We employed medicinal chemistry, chemoinformatics, the split-luciferase complementation assay (LCA) and patch-clamp electrophysiology to validate the activity of optimized leads in in vitro, in-cell and ex vivo preparations.

Results: A series of approximately 50 non-commercially available chemical analogues of compound 1028 and its original scaffold 7605 were synthesized and screened using a multimodal pipeline that integrated biological assays and chemoinformatics. Initial screening was conducted against the FGF14/Nav1.6 complex reconstituted in cells using the LCA. Compounds were ranked based on their percentage of maximal luminescence during triplicate validation, along with counter-screening against full-length luciferase in transiently transfected HEK293 cells. Docking analysis revealed that active compounds clustered into two distinct groups based on hydrogen bond interactions with either R117 or E152 on FGF14—two critical hotspots at the PPI interface. Compounds resembling the 1028 scaffold that interacted with R117 were predicted to modulate intermolecular interactions within the Nav1.6 C-terminal tail and the inactivation gate, acting as PAMs. In contrast, compounds interacting with E152 resemble the original 7605 scaffold and were predicted to function as negative allosteric modulators (NAMs). Among the predicted NAM group, compound CZ0251 significantly inhibited FGF14/Nav1.6 complex formation 100 nM, 27.86 %, n = 8, p < 0.0001, one-way ANOVA) compared to control (0.5% DMSO vehicle) and was selected for electrophysiological testing. Patch-clamp recordings showed that CZ0251 profoundly suppressed FGF14/Nav1.6 currents, with an estimated IC₅₀ of 166 pM and efficacy approaching 83.8%. CZ0251 is currently being evaluated in MSN NAc slice preparations, with future plans for pharmacokinetic studies and in vivo testing in reward-related operant behavior assays.

Conclusions: This approach provides a paradigm shift and an innovative strategy in CNS drug discovery that could lead to first-in-class medications for anhedonia in MDD and other psychiatric disorders.

Keywords: Early Phase Drug Development, ion channel, anhedonia.

Disclosure: IonTX Inc., Founder, Self.

P617. A novel cannabinoid receptor 1 (CB1) pam-antagonist dampens amphetamine-induced dopamine levels; evidence of CB1-mediated indirect modulation of high da states

Catharine Mielnik, Brandon Oliver, Alexandra Arcenas, Natalie Zlebnik, Ruth Ross

University of Toronto, Toronto, Canada

Background: Several serious, debilitating, and lifelong conditions, including psychosis in schizophrenia, mania in bipolar disorder, and attention deficit hyperactivity disorder are believed to be related to a dysregulation in dopamine (DA) signalling (DA pathologies). However, effective and patient-compliant treatment remains stagnant, with D2-targeting compounds the continued gold standard for “anti-psychotic” treatment. In search of novel therapeutic avenues, we have turned our focus on the indirect modulation of DA via endocannabinoid (EC) signaling. Cannabinoid Receptor 1 (CB1) is the most highly expressed G-protein coupled receptor in the brain and represents a potential therapeutic target in numerous conditions, including DA-related ones. ECs interact with CB receptors to modulate several biological processes, including mood, appetite, and metabolism. The EC system acts as a homeostatic filter that integrates several signaling pathways, including DA signaling. Within the context of psychiatry, and based on previous pitfalls of targeting DA directly, indirectly targeting dysregulated DA signaling may be more therapeutically advantageous. Targets such as CB1 offer this indirect modality.

Our recent studies have achieved experimental proof-of-principle for this hypothesis via a novel class of drug-like allosteric modulators that target CB1; CB1 positive allosteric modulator (PAM)-antagonists. CB1 PAM-antagonists increase agonist affinity while simultaneously blocking signalling and are highly effective in mouse models of hyperDA, without evidence of comorbid anxiety effects that have marred CB1 orthosteric inverse agonists in clinical trials. High DA pathologies are likely exacerbated by a positive feedback loop between high DA and high 2-arachidonoylglycerol (2-AG), the most abundant EC in the brain. Known overactive 2-AG signaling is found in several psychiatric conditions, in which high levels of 2-AG and CB1 signalling appear to drive the pathology. Here, we look to directly measure DA levels in vivo to establish a mechanism of action for our novel CB1 PAM-antagonists and their indirect modulation of DA signaling levels via CB1.

Methods: Using a pharmacological model of hyperDA, we measured changes in acute amphetamine-induced DA release following the administration of our novel CB1 PAM-antagonist (ABM300, 10 mg/kg i.p. 30-min pre-treatment) in both male and female adult C57Bl/6J mice. Using the fluorescent GRABDA sensor to record DA transients with millisecond resolution, we simultaneously measured both DLS and NAc DA release during open-field exploration. Briefly, adult mice were anesthetized with isoflurane and GRABDA2m AAV was injected (500nl total/hemisphere) in both the DLS and NAc via unilateral injection of AAV9-hSyn-GrabDA-4.4(2m) (gift from Yulong Li, Addgene viral prep #140553-AAV988) at the following coordinates: antero-posterior [NAc]1.1 mm, [DLS] −0.4 mm; medio-lateral [NAc] 1 mm, [DLS] 2.5 mm; dorsoventral [NAc] −3.8 mm, [DLS] 3.2 mm. GRABDA sensor was infused in the target brain areas, and immediately after, an optical fiber was embedded in the same target brain area, 0.1mm above the previously mentioned coordinates. Optical fibres were secured with resin cement. Mice were allowed to recover for 4-weeks while awaiting viral expression. Fiber photometry of DA signals was conducted during an open field test (2-hour assay/recording) following administration of our novel CB1 PAM-antagonist, ABM300. Samples were collected at 1017 Hz (1017 samples/sec).

Results: Pre-treatment with ABM300 dampened amphetamine-induced DA release, decreasing the area under the curve (AUC) in a time-course of DA signal, over a 120-min recording. This was true for both the DLS (AUC: p = 0.016) and the NAc (AUC: p = 0.0676). Interestingly, DA transient peak width and peak prominence were unchanged following administration of ABM300 in both the DLS and NAc. Preliminarily, we see influences of sex on the magnitude of DA signal decrease in both brain regions. These DA dynamics mirrored the behavioural effects we previously published, showing correlations between simultaneous open-field and fiber photometry recordings; ABM300 dampened amphetamine-induced psychomotor activity.

Conclusions: The present study confirms our hypothesis and provides mechanistic insight that indirect modulation of DA occurs in the striatum via the targeting of CB1 with our novel CB1 PAM-antagonist ABM300. ABM300 decreases psychomotor hyperactivity, which occurs via a dampening of DA signals in both the DLS and NAc. This work provides substantiative evidence for a novel therapeutic strategy in the treatment of hyperDA psychiatric pathologies: indirect modulation of DA via CB1. In contrast with typical orthosteric and allosteric antagonists, PAM-antagonists have a distinctive advantage in reversing receptor overactivation in the persistent presence of high levels of endogenous agonist. PAM-antagonists could be a “seek and destroy” that selectively targets agonist-activated GPCRs, offering a novel approach to generate “antipsychotic”-like molecules to treat hyperDA psychiatric conditions.

Keywords: Endocannabinoids, Dopamine, Antipsychotics, CB1 receptor, drug discovery/development.

Disclosure: Nothing to disclose.

P618. Pharmacological GLP1 receptor modulation impacts decision making under risk of punishment and corresponding nucleus accumbens dopamine signals in rats

Barry Setlow, Zachary Krumm, Wonn Pyon, Jennifer Bizon

University of Florida, Gainesville, Florida, United States

Background: Predictions of expected rewards vs. costs are instrumental in several types of behaviors. Risk-based decision making relies on accurate and dynamic predictions of these variables to guide adaptive choice behavior, while extreme weighting of risk or cost is a feature of neuropsychiatric conditions such as substance use disorders (risk seeking) and anorexia nervosa (risk avoidance). Dopamine signaling in the nucleus accumbens (NAc) plays a fundamental role in mediating both expected and experienced rewards and costs, making it a compelling target for therapeutic modulation in conditions characterized by dysregulated risk-taking behaviors. Pre-clinical evidence has identified the glucagon-like peptide 1 receptor (GLP1R) agonist family of medications as modulators of NAc dopamine signaling, and initial clinical studies indicate that they may impact behaviors ranging from gambling to alcohol consumption.

Methods: To determine if GLP1R activation can influence risk-based decision making and associated dopamine signals, we evaluated the effects of acute administration of the GLP1R agonist Exendin-4 (Ex4) on performance of a risky decision-making task in rats (n = 8M, 7F). In this task, rats make discrete choices between two response levers, one that yields a small reward (1 food pellet) with no risk of punishment and another that yields a large reward (2 food pellets) accompanied by a mild footshock that ranges in probability from 0 to 100%. Ex4 (0, 0.1, 0.3, 1.0, 3.0 µg/kg in 0.9% saline vehicle, s.c.) was administered 1 hour prior to test sessions in the task, using a randomized, within-subjects design. In another set of rats (n = 3M, 2F), effects of Ex4 on NAc dopamine signals were evaluated during task performance using fiber photometry with GRAB-DA2m.

Results: Ex4 reduced choice of the large, risky reward (reduced risk taking) in a dose-dependent manner (two-way ANOVA, dose x shock probability, main effect of dose, F(16,175) = 2.27, p < 0.01 males, F(16,150) = 2.03, p < 0.01 females). The effects of Ex4 on risk taking were mimicked by acute administration of 10 µg/kg semaglutide (F(4, 70) = 2.12, p = 0.02 males, F(4, 60) = 2.81, p < 0.01 females). Trial omissions induced by Ex4 were significantly increased only at the highest dose, which was selected based on its ability to reduce appetite. Co-administration of the GLP1R antagonist Exendin-9 (Ex9) attenuated effects of Ex4 on risk taking in a dose dependent manner (F(8, 105) = 2.03, p < 0.01 males, F(8, 90) = 2.89, p < 0.001 females), supporting the GLP1R specificity of the Ex4 effects. Ex9 alone also increased risk taking (F(8, 105) = 1.98, p < 0.05 males, F(8, 90) = 2.70, p < 0.001 females), suggesting a role for endogenous GLP1R signaling in modulating decision making. Fiber photometry data showed prediction error-like dopamine signals in NAc upon receipt of the large reward; these signals increased when the absence of punishment accompanying the large reward was unexpected (i.e., “wins”), and decreased when punishment accompanying the large reward was unexpected (i.e., “losses”). Ex4 attenuated the magnitude of dopamine signals in response to wins and increased the magnitude of these signals in response to losses (p’s < .02), providing a potential mechanism for the effects of GLP1R activation on decision making.

Conclusions: Collectively, these data begin to inform the mechanisms by which GLP1R-targeting medications can influence decision making guided by both reward and punishment.

Keywords: Decision Making, Nucleus Accumbens, GLP-1 receptor agonist, rats.

Disclosure: Nothing to disclose.

P619. Selective GluN2D inhibition recapitulates non-selective NMDAR antagonists effects on synaptic plasticity in rodents

Bartosz Balana, Vivien Zell, Jeffrey Schoellerman, Ian Fraser, Rory Pritchard, Heidi Huysmans, Abdellah Ahnaou, Juan-Pita Almenar, Michele Rizzolio, Jose Cid, Ryan Wyatt, Wayne Drevets, Pascal Bonaventure

Johnson and Johnson, San Diego, California, United States

Background: The proposed mechanisms behind rapid antidepressant effects shown by several selective and non-selective NMDAR antagonists include the inhibition of interneurons, which leads to a surge in neurotransmitter release, and results in enhanced synaptic plasticity. Given that the GluN2D subunit of NMDA receptors is selectively expressed in interneurons, we hypothesized that inhibiting GluN2D-containing receptors would produce similar effects on synaptic plasticity.

Methods: JNJ-GluN2D NAM is a novel selective negative allosteric modulator (NAM) of GluN2D-containing receptors. Here, we investigated the in vitro and in vivo effects of JNJ-GluN2D, on interneuron function, neurotransmitter release, and synapse density. This assessment was conducted using EEG, microdialysis, and ex vivo brain slice electrophysiology in rodents. Additionally, we evaluated potential psychomimetic effects through pre-pulse inhibition of startle reflex testing in mice.

Results: JNJ-GluN2D demonstrated high selectivity for recombinant GluN2D receptors as measured by automated patch-clamp technique (IC50 = 0.11 µM for human GluN2D; n = 9 experiments; IC50 > 11 µM for GluN2A/B/C). It reduced inhibitory tone on primary pyramidal cells in hippocampal slices, leading to a significant decrease in sIPSC frequency (−18% vs. control; p = 0.03; n = 8 slices). The compound showed good brain permeability (mouse Kpuu = 0.23; rat Kpuu = 0.3), making it a suitable in vivo tool. After subcutaneous administration, JNJ-GluN2D exhibited several effects shown previously by selective and unselective NMDAR antagonists. These included an enhanced number of synapses in rat prefrontal layer 5 pyramidal neurons after 24 hours following a 60 mg/kg dose, as evidenced by mEPSC frequency (5.34 Hz for JNJ-GluN2D vs. 3.25 Hz for vehicle; n = 15–16, p = 0.003). Furthermore, JNJ-GluN2D at 120 mg/kg sc dose significantly increased dopamine, serotonin and acetylcholine release in the rat cortex (p < 0.05; n = 4–5 animals). Similar to non-selective NMDAR antagonist, the compound enhanced EEG gamma band power compared to vehicle control (p < 0.05; n = 8 rats). Importantly, unlike NMDAR blockers JNJ-GluN2D did not disrupt pre-pulse inhibition (89% of control after 60 mg/kg sc dose; p = 0.25; n = 18 mice), suggesting a lack of dissociative side effects.

Conclusions: Since interneuron firing is dependent on activation of NMDA receptors, we have hypothesized that selective inhibition of interneuron-specific GluN2D receptor subtype would shift exhibition/inhibition balance toward increased excitation. Disinhibition of primary neurons would lead to a glutamate burst and subsequently increased synaptogenesis. Our results confirm that GluN2D-selective antagonist decreases inhibitory tone onto primary cells which resulted in increased levels of neurotransmitters in vivo (a proxy for glutamate burst) and lead to increased number of synapses 24h after dosing.

Gamma EEG band is highly dependent on activity of cortical PV interneurons. Non-selective NMDAR antagonists increase gamma band power, and this effect is absent in GluN2D KO mice. Selective inhibition of GluN2D-containing receptors recapitulates those effect and further supports proposed mechanism of antidepressant action via inhibition of interneurons. Absence of any effect of JNJ-GluN2D on PPI suggests that the selective inhibition of GluN2D offers a more precise control of interneurons activity and downstream effects vs. unselective NMDA receptor blockers.

In summary, our preclinical findings indicate that the inhibition of GluN2D-containing receptors could serve as a promising mechanism to induce synaptic plasticity and exert antidepressant effects.

Keywords: NMDA glutamate receptors, Fast-acting Antidepressant, Synaptic Plasticity, GRIN2D, GluN2D.

Disclosure: Johnson and Johnson, Employee, Self, Johnson and Johnson, Stock / Equity - Publicly Traded Company, Self, Johnson and Johnson, Employee, Spouse/Partner, Johnson and Johnson, Stock / Equity - Publicly Traded Company, Spouse/Partner.

P620. Effects of phosphodiesterase-4 B inhibition on inflammation and social behavior in a model of chronic social defeat stress (CSDS) in male mice

Lauren Seabrook, Robert Aukema, Kyongnyon Nam, Thomas Wassmer, Kerry Ressler, Steven Leiser, Stefka Gyoneva, Sophia Bardehle, William Carlezon

Harvard Medical School McLean Hospital, Belmont, Massachusetts, United States

Background: Stress is one of the strongest risk factors for the onset and maintenance of depression and anxiety disorders, and a large body of evidence links stress exposure to alterations in immune function. Elevated inflammatory cytokines have been reported in subsets of individuals with depression, suggesting that inflammation may contribute to symptom expression and treatment resistance. However, the role of inflammation in the pathogenesis of stress disorders remains incomplete. Phosphodiesterase-4 (PDE4) is a family of intracellular enzymes comprising several isoforms. PDE4 regulates inflammation and neurotransmission through cyclic adenosine monophosphate (cAMP) degradation. Targeting this pathway via PDE4 inhibitors has been an approved treatment for individuals with inflammatory conditions (such as chronic obstructive pulmonary disease [COPD]) since 2011 in the USA, and there is a growing interest in the use of PDE4 inhibitors for the treatment of depression in humans. Because the mechanisms by which PDE4B influences stress responses and social behavior are not well defined, we investigated the functional role of PDE4B using both pharmacological inhibition and genetic deletion strategies in a novel accelerated chronic social defeat stress (CSDS) model.

Methods: Adult male C57BL/6J mice were subjected to accelerated CSDS, consisting of exposure to an aggressive CFW intruder twice daily for four consecutive days. Control mice were exposed to novel, non-aggressive conspecifics on an identical schedule. To assess the contribution of PDE4 activity, mice received rolipram (a non-selective PDE4 inhibitor; 0.3 mg/kg), ketamine (positive control for antidepressant efficacy; 20 mg/kg), or vehicle on the final day of defeat stress. Social behavior was evaluated ~24 hours later using the open field social interaction (OFSI) test. Two days following the OFSI test blood, brain, thymus, and adrenal glands were collected for profiling of peripheral immune activation and brain transcriptomic analyses. Plasma corticosterone was quantified by ELISA, and inflammatory cytokines were measured using MSD multiplex assays. Whole-brain RNA sequencing was conducted to identify transcriptomic alterations associated with social defeat and drug treatment. In a parallel cohort, male pde4b knockout (KO) mice (C57BL/6 background) and wild-type littermates were subjected to the same CSDS protocol to test whether genetic deletion of PDE4B recapitulates the effects of pharmacological inhibition.

Results: Both rolipram and ketamine increased social interaction following CSDS. Specifically, rolipram (P < 0.05) and ketamine (P < 0.05) treated mice spent more time with a novel conspecific in the OFSI compared to vehicle-treated defeated mice. Defeat stress induced robust immune changes, including reductions in thymus size (involution; P < 0.05) and elevated plasma cytokines (TNFα, IL-1β, IL-10, IL-6; P’s < 0.05), all of which are consistent with systemic inflammatory activation. However, these peripheral immune alterations were not significantly modulated by rolipram or ketamine. Transcriptomic analyses revealed the upregulation of immune pathways indicative of stress-induced neuroinflammation in defeated versus control mice, supporting a central immune component to stress-induced behavioral changes. Importantly, pde4b KO mice also displayed increased social interaction scores compared to wild-type controls following CSDS (P < 0.05).

Conclusions: Both pharmacological inhibition and genetic deletion of PDE4B enhanced social behavior following defeat stress, pointing to PDE4B as an important regulator of stress-induced social withdrawal. Interestingly, these behavioral effects were dissociable from canonical peripheral inflammatory measures such as thymic involution, adrenal hypertrophy, and circulating cytokines. Ultimately, delineating the interactions between PDE4B signaling, neuroinflammation, and stress resilience may enable the development of novel therapeutic approaches for mood disorders.

Keywords: Depression Inflammation Cytokine, Phosphodiesterase-4B (PDE4B), Social defeat stress.

Disclosure: Nothing to disclose.

P621. An anterior cingulate cortex neuronal ensemble controls contextual opioid analgesic tolerance

Rafael Perez, Yann Mineur, Cheryl Chen, Wenliang Zhou, Julian Thompson, Saagar Motupally, Angelica Minier-Toribio, Marina Picciotto

Yale University, New Haven, Connecticut, United States

Background: Opioids are first-line treatments for pain, but rapid tolerance results in reduced efficacy and increased dosing, heightening the risk of opioid use disorder. Remarkably, after repeated administration, tolerance can be quickly reversed when the drug is given in an unexpected context.

This phenomenon has significant implications for the long-term use of opioids in therapy and addiction treatment. However, the brain mechanisms underlying the contextual control of tolerance remain poorly understood. This study aimed to identify the neuronal populations that mediate this environmental tolerance control.

Methods: We first developed a model of contextual analgesic tolerance. Male and female mice (n = 17/sex) received daily injections of fentanyl (Fen) (25 ug/kg) in one context, alternating with saline (Sal) in a different context over 14 days. Tolerance was evaluated after each conditioning session using the hotplate test. We then identified neuronal ensembles activated by tolerance and contextual reversal using whole brain clearing with ArcCreER:Ai14 (ArcTRAP) mice (n = 4/group) to tag active neurons in the fen context through Arc- dependent tdTomato expression, followed by cFos labeling of active cells after exposure to the fen- (tolerance) or sal- paired (reversal) context. Next, we used in vivo fiber photometry to measure ACC activity across contextual tolerance conditioning and during reversal (n = 8). Finally, we employed chemogenetic tools to inhibit the ACC or stimulate tolerance-active ensembles in ArcTRAP mice to modulate tolerance (n = 5–7/group).

Results: Mice exhibited tolerance, showing a decrease of about 50% in analgesic responses from the first to the last Fen session (p < .0001), which was reversed in the Sal- paired context (p < .01). Contextual tolerance significantly increased active cells in the ACC, an effect specific to the co- presentation of fen and the fen- paired context (p < .01). In vivo ACC activity in the context correlates with analgesic tolerance (p < .01). ACC silencing blocked tolerance without affecting basal nociception (p < .0001), while stimulating ACC tolerance- active ensembles prevented tolerance reversal in the sal- paired context (p < .05).

Conclusions: We identify ACC neuronal ensembles as orchestrators of contextual opioid analgesic tolerance without impacting basal nociception.

Keywords: opioid tolerance, Context, Neuronal ensembles, Anterior Cingulate Cortex (ACC), analgesia.

Disclosure: Nothing to disclose.

P622. Metabotropic mGlu1 receptor regulation of cortical inhibition and cognitive function: implications in adolescent cocaine exposure

Deborah Luessen, Isabel Gallinger, Mac Meadows, Brenna Wolfe, P. Jeffrey Conn

Emory University School of Medicine, Nashville, Tennessee, United States

Background: Exposure to psychostimulants, such as cocaine, during adolescence produces persistent changes in the prefrontal cortex (PFC) which parallel cognitive deficits seen in adulthood. Further, adolescent exposure to psychostimulants impairs inhibitory transmission in the PFC in adulthood, suggesting that enhancing PFC inhibitory transmission may be a promising strategy to reverse drug-induced cognitive deficits. Activation of the mGlu1 subtype of metabotropic glutamate receptor increases inhibitory transmission in the PFC and working memory by selective excitation of somatostatin-expressing GABA interneurons (SST-INs). Therefore, we hypothesize that repeated exposure to cocaine during a critical developmental period in adolescence disrupts PFC inhibition via SST-INs and drives working memory impairments in adulthood which can be mitigated by activation of mGlu1.

Methods: Male and female SST- and PV-Ai9 tdTomato mice were injected once daily with cocaine (20 mg/kg, i.p.) for 7 days (postnatal day 35–42). Whole-cell patch-clamp electrophysiological recordings from interneuron populations within the PFC were conducted between 10–12 weeks of age. Additionally, touchscreen-based automated cognition testing was used to determine working memory performance in adult mice. Novel mGlu1 positive allosteric modulators (PAMs) were leveraged in behavioral and electrophysiology studies to determine their procognitive efficacy and mechanism of action within the PFC.

Results: We found that repeated administration of cocaine during a critical adolescent period impaired PFC SST-IN, but not parvalbumin-expressing interneuron (PV-IN), firing compared to saline-treated mice. Adolescent cocaine exposure significantly decreased the frequency of spontaneous excitatory postsynaptic currents onto SST-INs but not PV-INs. These findings were paralleled by adolescent cocaine-induced impairments in spatial working memory in adulthood.

Importantly, these physiological and behavioral effects of adolescent cocaine exposure were reversed by selective mGlu1 activation. Lastly, repeated amphetamine administration during the same adolescent critical period did not result in impaired SST-IN function or spatial working memory in adulthood.

Conclusions: These studies show that: 1) cocaine, but not amphetamine, exposure during an adolescent critical period induces persistent and selective deficits in PFC SST-IN function and cognition in adulthood and 2) selective activation of mGlu1 with PAMs represents a novel strategy for reversing cocaine-induced cognitive impairments.

Keywords: Cognition, Addiction, Neurodevelopment, Electrophysiology, Prefrontal Cortex.

Disclosure: Nothing to disclose.

P623. Prefrontal cortex vasoactive intestinal peptide neurons regulate motivation to drink alcohol

Dakota Brockway, Samuel Boehm, Nilah Jordan, Lars Nelson, Max Joffe

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Alcohol use disorder (AUD) remains one of the most prevalent and costly psychiatric conditions: with nearly 30 million individuals affected in the United States alone. Despite its profound societal impact, existing pharmacological treatments for AUD are limited and offer only modest efficacy. The prefrontal cortex (PFC) plays a critical role in executive function, decision-making, and regulation of motivation, all processes that are disrupted in AUD. Clinical imaging studies and population-level analyses consistently link aberrant PFC activity with heavy drinking, alcohol craving, and impaired self-control over intake. Within the PFC, vasoactive intestinal peptide-expressing interneurons (VIP-INs) represent a sparse but powerful population that primarily target other inhibitory neurons to enhance pyramidal cell output, and prior data from our lab has shown that alcohol bidirectional affects the excitability of these neurons, but the relationship with behavior and role of neuropeptide signaling have not been established.

Methods: We interrogated PFC IN physiology using a combination of ex vivo whole-cell electrophysiology and biosensor imaging with a suite of genetically engineered mouse lines. We employed selective pharmacological tools to modulate VIPR1 signaling. Furthermore, we evaluated causative relationships between IN physiology and function using chemogenetics.

Results: VIP selectively depolarizes VIP neurons in the PFC. VIP depolaritzes resting membrane potential of VIP neurons (one-sample t test; t14 = 4.384, p = 0.0006)n = 15 cells from 7♀, 4♂ while having no effect on other populations such as SST (one-sample t test; t10 = 0.8579, p = 0.411) n = 11 cells from 4♀, 6♂, or pyramidal neurons (one-sample t test; t6 = 0.6841, p = 0.5195) n = 7 cells from 3♀, 4♂. In chemogenetic studies, activation of hM3Dq signaling in PFC VIP cells increased alcohol-seeking behavior (paired t test; t8 = 3.158, p = 0.0134) and alcohol intake (paired t test; t8 = 2.433, p = 0.216) in the progressive ratio (PR) self-administration task N = 9 mice (3♀, 6♂).

Conclusions: Our findings suggest that alcohol acutely increases VIP-IN excitability, resulting in increased VIP release. VIP then acts in an autocrine fashion via VIPR1 to further depolarize VIP-INs and elevate intracellular cAMP. This feedback loop increases activity of pyramidal neurons through disinihibition of interneurons targeting pyramidal cells (SST interneurons), driving maladaptive cortical excitation that promotes alcohol seeking. We propose that this loop can be attenuated by targeted inhibition of VIP-INs or pharmacological antagonism of VIPR1.

Keywords: GABAergic interneurons, prefrontal cortex, vasoactive intestinal peptide, chemogenetics, alcohol use disorder.

Disclosure: Nothing to disclose.

P624. Lithium as a therapeutic strategy in a SNAP25 L50S mouse model of developmental and epileptic encephalopathy

Burak Uzay, Lisa Monteggia, Ege Kavalali

Icahn School of Medicine at Mount Sinai- Mount Sinai Hospital, New York City, New York, United States

Background: Synaptic vesicle fusion relies on soluble N-ethylmaleimide–sensitive factor attachment protein receptor (SNARE) proteins, including SNAP25, synaptobrevin-2, and syntaxin-1. Pathogenic variants in the SNAP25 gene result in a group of neurodevelopmental disorders called SNAP25 encephalopathies. Among these pathological variants, we previously identified SNAP25 L50S as a mutation that exclusively disrupts spontaneous glutamate release. Patients carrying this mutation present with treatment-resistant seizures and a range of abnormal neurobehavioral phenotypes. Using in vitro models, we previously demonstrated that chronic lithium treatment can restore abnormal network activity and action potential kinetics by inducing homeostatic downscaling of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors. Here, we tested whether lithium could serve as a therapeutic strategy in vivo using a CRISPR knock-in (KI) mouse model of SNAP25 encephalopathy.

Methods: We generated SNAP25 L50S CRISPR-KI mice and performed behavioral assays to characterize their neurobehavioral phenotype. We performed slice patch-clamp recordings to assess abnormal spontaneous glutamate release. We then treated these mice with lithium for 16–17 days, followed by repeated behavioral testing to evaluate lithium’s potential therapeutic effects.

Results: SNAP25 L50S mice showed infertility, increased neonatal mortality, in addition to frequent spontaneous tonic–clonic seizures. Behavioral assays revealed reduced muscle strength, significantly increased anxiety-like behavior, impaired social interaction, hypomobility, and deficits in learning and memory, including impaired long-term memory and abnormal cue-based fear conditioning. Chronic lithium treatment significantly alleviated hypomobility, long-term memory deficits, and seizure frequency.

Conclusions: Patients with SNAP25 encephalopathy often experience treatment-resistant epilepsies that do not respond to conventional antiepileptic drugs. Our findings suggest that lithium may represent a viable therapeutic strategy in the context of abnormal spontaneous glutamate release, capable of ameliorating key behavioral phenotypes including hypomobility, memory deficits, and seizure burden. Building on our previous in vitro work, we propose that lithium exerts these effects through modulation of homeostatic AMPA receptor downscaling, highlighting a potential novel treatment approach for SNAP25 encephalopathies.

Keywords: SNAP25, Lithium, epilepsy, Mouse models.

Disclosure: Nothing to disclose.

P625. A non-canonical role for dopamine pauses in driving behavioral persistence

Rene Carter, Sasha Burwell, Haidun Yan, Shaun Lim, Brenda Shields, Mike Tadross

Duke University, Durham, North Carolina, United States

Background: Animals adapt to changing environments by gradually unlearning behaviors that no longer yield expected rewards—a process known as extinction learning. This adaptation is thought to rely on transient pauses in ventral tegmental area dopamine (VTADA) neuron firing, which have been proposed to enhance behavioral flexibility and accelerate extinction. However, directly testing the causal role of these pauses has been hindered by a lack of tools to selectively eliminate them without broader circuit disruption.

Methods: We developed gabazineDART, a cell- and receptor-specific pharmacological tool that acutely blocks GABAa receptor–mediated pauses in VTADA neurons while maintaining normal tonic and burst firing. DAT::Cre mice of both sexes were injected with AAV vectors expressing HaloTag tethering proteins, permitting selective Gabazine capture to VTADA neurons. In vivo electrophysiology (n = 39 cells from 5 mice vs 18 cells from 3 controls) confirmed selective disruption of pause dynamics. Experimental paradigms include a Pavlovian reward conditioning and extinction assay (n = 12 gabazineDART, n = 12 control), fiber photometry of VTADA and lateral nucleus accumbens dopamine signals (n = 6 vs 6; n = 7 vs 7), and auditory fear extinction with freezing readouts (n = 4 vs 4). Data were analyzed using permutation tests, regression analyses, and bootstrap methods.

Results: Contrary to canonical models, gabazineDART-treated mice exhibited significantly faster extinction of conditioned licking responses compared to controls (AUC extinction metric, P = 0.0043). Fiber photometry revealed that gabazineDART abolished omission-evoked dopamine pauses in both VTADA neurons (P = 0.001, early extinction) and lateral nucleus accumbens (P = 0.002, early extinction), without significantly affecting cue-evoked bursts. Trial-level analyses revealed that larger dopamine pauses were associated with greater behavioral perseveration (R² = 0.65, P = 0.00005), and extinction learning emerged only after the pauses subsided. In fear extinction, gabazineDART robustly suppressed freezing across three days (day 2 P = 0.0029, day 3 P = 0.0173, day 4 P = 0.0098).

Conclusions: These findings overturn the long-standing view that dopamine pauses facilitate extinction learning. Instead, we found that GABAa-mediated pauses in VTADA neurons appear to promote behavioral persistence, constraining flexibility in both appetitive and aversive contexts. By selectively eliminating pauses, gabazineDART accelerated reward extinction and reduced conditioned fear expression, highlighting dopamine pauses as a novel therapeutic target for disorders characterized by pathological perseveration, such as OCD, PTSD, and Parkinson’s disease.

Keywords: Dopamine, Behavioral Pharmacology, cell-type specific, Extinction learning, Fear conditioning and extinction.

Disclosure: Nothing to disclose.

P626. Preclinical behavioral and pharmacological characteristics of TN-001, a novel, non-hallucinogenic neuroplastogen for the treatment of major depressive disorder

Samuel Slocum, Nicholas Anas, Elizabeth Cogan, Mark Demitrack

Transneural Therapeutics, Issaquah, Washington, United States

Background: Psychedelics have recently demonstrated significant clinical promise for the treatment of psychiatric illnesses including major depressive disorder (MDD) and post-traumatic stress disorder (PTSD). However, the hallucinogenic and cardiotoxic properties of psychedelic compounds present unique challenges in the clinical setting that may limit their widespread use. Neuroplastogens are an emerging class of compounds that promote neuroplasticity, similar to psychedelics, without inducing hallucinations. Here we present data on TN-001, a novel neuroplastogen for the treatment of MDD.

Methods: Agonist and antagonist activity of TN-001 was assessed using PI hydrolysis and PathHunter® β-arrestin2 assays, and responses were normalized to the efficacy of serotonin in these assays. For receptor occupancy studies, C57BL/6J mice received intraperitoneal (i.p.) injections of either vehicle, psilocybin, or TN-001 and cortical membranes were prepared after 60 minutes. Using these membranes, 5-HT2A occupancy was determined by assessment of specific binding of [3H]Cimbi-36 and nonspecific binding of 25CN-NBOH. Using competition binding methods, Ki values were determined using cortical membranes from untreated mice. Neural plasticity was assessed using primary culture of mature cortical neurons from Wistar rat embryos. Neuronal cell cultures were incubated with vehicle, psilocin, BDNF, or TN-001 on Day 11 of culture for three days. Cells were then fixed on Day 14 and immunostaining for synaptophysin (SYN, a pre-synaptic marker), postsynaptic density protein 95 (PSD-95), a post-synaptic marker), and microtubule-associated protein 2 (MAP-2, binds to neurons and neurites) was performed. Images were captured using Operetta® CLS (PerkinElmer) and analyzed using Harmony™ software (PerkinElmer).

All behavioral assays were performed in C57BL/6J mice with free access to food and water. For head twitch response (HTR) and locomotor activity assays, mice were placed into a clean cage after i.p. injection with vehicle, psilocybin, or TN-001. HTRs were scored by a blinded experimenter and locomotor activity was recorded as distance moved in cm using Ethovision® (Noldus). The forced swim test (FST) was used to assess antidepressant activity. Mice were placed into a cylinder of water 1 hour following i.p. injection with vehicle, fluoxetine, or TN-001. Immobility time was measured using SMART 3.0 video tracking software (Panlab). Experimental protocols were approved by the 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.

Results: TN-001 is a Gq-biased partial agonist of human 5-HT2A receptor with an EC50 of 68 nM and Emax of 67% (relative to serotonin response) and an antagonist of human 5-HT2B receptor with an IC50 of 142nM and Imax of 97%. Similar potency was observed for mouse 5-HT2A receptor (106 nM, 45%). In ex vivo occupancy studies, TN-001 showed robust occupancy of cortical 5-HT2A receptors which exceeded that of psilocin at the same doses. In radioligand binding studies with cortex homogenates, TN-001 showed comparable binding affinities for both mouse and human cortical 5-HT2A receptors (55 nM and 79 nM, respectively). In models of neuronal plasticity using mature rat cortical neurons, TN-001 induced significant dose-dependent increases in the number of MAP-2 positive neurons, number of MAP-2 positive neurites in µm, and overlap in SYN/PSD-95/MAP-2 staining in µm2. These observations represent increases in the number of neurons, total neurite network, and number/area of synapses, respectively. In vivo, TN-001 does not induce HTR and does not impair locomotor activity at relevant doses. In FST, TN-001 significantly reduced immobility time compared to vehicle.

Conclusions: TN-001 is a potent dual 5-HT2A partial agonist and 5-HT2B antagonist that demonstrates structural neuroplasticity ex vivo and antidepressant activity in mice without inducing HTR or locomotor impairments. While agonism of 5-HT2B is a common off-target activity among many 5-HT2A agonists, and confers significant risk of 5-HT2B agonist-induced cardiac valvular hypertrophy, TN-001 is an antagonist at 5-HT2B receptors, and therefore is unlikely to carry this risk. These results provide a promising avenue toward the safe application of 5-HT2A-mediated neuroplasticity to treat MDD and PTSD.

TN-001 is an investigational compound and not approved by the Food and Drug Administration for human use.

Keywords: neuropharmacology, neuroplasticity, serotonin 2A receptor.

Disclosure: Transneural Therapeutics, Consultant, Self, Transneural Therapeutics, Consultant, Spouse/Partner.

P627. ENX-205, a novel neuroplastogen with 5-HT2A agonism and D2/D3 antagonism: a promising therapeutic approach for PTSD

Krishna C. Vadodaria, Dave S. Garvey, Brian Kangas, William Brubaker, John McCorvy, Diego Pizzagalli, Vikram Sudarsan, Kimberly Vanover

Engrail Therapeutics, San Diego, California, United States

Background: Posttraumatic stress disorder (PTSD) is a debilitating condition with limited effective treatments. 5-HT2A receptor agonists, including next-generation neuroplastogens show promise for enhancing cortical plasticity and facilitating trauma processing. ENX-205 is a novel neuroplastogen that combines potent 5-HT2A agonism (partially Gq-driven) and D2/D3 antagonism. This dual mechanism is hypothesized to promote neuroplasticity, enhance social interaction, and increase dopamine signaling in reward and cognitive circuits, thereby targeting core PTSD symptoms such as avoidance and mood/cognitive disturbances. Despite potent Gq signaling, its unique polypharmacology is predicted to reduce hallucinogenic liability.

This novel pharmacological profile offers a potential multimodal approach to treating PTSD, addressing limitations of existing treatments and advancing the development of safe, effective therapeutics.

Methods: The pharmacological profile of ENX-205 was evaluated using in vitro assays with CHO cells expressing human dopaminergic and serotonergic receptors. Pharmacokinetics (PK) were evaluated in male rats (n = 3), orally dosed with ENX-205 (5 mg/kg PO). Neuroplasticity was assessed in both rat and human neuronal cultures following treatment with ENX-205 for 48 hours and Sholl analyses for neurite outgrowth.

Receptor occupancy for D2/D3 and 5-HT2A receptors was assessed in male rats (n = 5) using ex vivo radioligand binding assays. Dopamine release in the nucleus accumbens and prefrontal cortex was measured via microdialysis in freely moving male rats (n = 8). Reward processing behavior was assessed using the probabilistic reward task (PRT) in male Sprague Dawley rats (n = 8), trained on a touchscreen-based paradigm; ENX-205 (1, 2.5, 5 mg/kg PO) was dosed 4 hours prior to testing in a Latin square design. Social behavior was tested using the social interaction assay in male rats (n = 10/group) treated with vehicle or ENX-205 (0.5–2.5 mg/kg PO). Extrapyramidal symptoms and hallucinogenic liability were evaluated in the catalepsy and head twitch response assays, respectively, in adult male rats.

Results: ENX-205 exhibited a distinctive polypharmacological profile, demonstrating potent 5-HT₂A receptor agonism (Ki = 9.3 nM; EC₅₀ = 5 nM) and D₂L receptor antagonism (Ki = 0.04 nM; IC₅₀ = 0.5 nM).

In vitro, ENX-205 significantly enhanced neuroplasticity. In rat cortical neurons, 1 µM ENX-205 increased neurite outgrowth (p < 0.001, ANOVA). In human iPSC-derived neurons, ENX-205 improved both cell survival and total neurite length at low nanomolar concentrations (0.1–0.3 nM; p < 0.05, SSMD analysis), suggesting robust neurotrophic effects across species.

In pharmacokinetic studies, ENX-205 showed rapid oral absorption (Tmax < 1 hour) and strong brain penetration (>10× brain/plasma ratio at peak). Receptor occupancy studies confirmed sustained D₂ (~66%) and 5-HT₂A (~20%) receptor engagement following a single 2.5 mg/kg oral dose, with peak occupancy observed at 4 hours post-dose.

Microdialysis experiments revealed that ENX-205 (2.5 mg/kg) significantly increased extracellular dopamine in the nucleus accumbens (p < 0.001 at peak) and prefrontal cortex (p < 0.01 at peak), with effects sustained over the 8-hour sampling window. These findings support a mechanism involving presynaptic D₂/D₃ receptor antagonism and disinhibition of dopaminergic transmission within the reward circuitry and the frontal cortex.

In the probabilistic reward task (PRT), at 1 and 2.5 mg/kg, ENX-205 significantly increased response bias (log b: p < 0.05; Cohen’s d = 1.2 and 1.4, respectively), without affecting discriminability (log d) at any of the doses tested, consistent with anti-anhedonic antidepressant-like effects. No effect was observed at 5 mg/kg.

In the social interaction test, ENX-205 enhanced, in a dose-dependent manner, active social exploration with a significant increase at 2.5 mg/kg (p < 0.01), indicating anxiolytic and pro-social properties.

ENX-205 did not induce head twitch responses at doses up to 10 mg/kg, indicating a lack of hallucinogenic-like effects typically associated with 5-HT2A agonism. Additionally, catalepsy, a marker of extrapyramidal motor side effects linked to postsynaptic D₂ antagonism at high D2 RO, was observed only at supratherapeutic doses (5 and 10 mg/kg), with no motor effects at behaviorally active doses (1–2.5 mg/kg). This suggests a favorable therapeutic window with a low risk of hallucinations or motor side effects at clinically relevant exposures.

Together, these findings indicate that ENX-205 promotes neuroplasticity and dopaminergic neurotransmission in limbic and cortical brain regions, producing behavioral effects consistent with alleviation of symptoms of PTSD, including the potential to reduce social withdrawal and avoidance and improve cognition, mood, and emotional regulation, while demonstrating a strong safety margin with minimal hallucinogenic or motor liabilities.

Conclusions: ENX-205 is a novel, orally bioavailable neuroplastogen with dual 5-HT2A agonism and D2/D3 antagonism, promoting cortical plasticity and dopaminergic tone. It demonstrates preclinical efficacy in translational models of social behavior and reward, with a favorable safety margin. These findings support its potential as a next-generation treatment for PTSD.

Keywords: 5-HT2A Agonist, neuroplastogen, Non-hallucinogenic, PTSD.

Disclosure: Engrail Therapeutics, Employee, Self, Zoll Medical Devices, Employee, Spouse/Partner.

P628. Characterization of a novel nitazene with impaired brain permeability and high analgesic potency

Juan Gomez, Zachary Frangos, Anna Tischer, Agnieszka Sulima, Michael Baumann, Kenner Rice, Mike Michaelides

NIH / NIDA, Baltimore, Maryland, United States

Background: Mu opioid receptor (MOR) agonist medications are the most effective analgesics available. However, such medications have known adverse effects including respiratory depression and abuse liability. As such, there is a concerted effort to develop alternative medications via partial agonism of MOR or through different pathways. However, most alternative medications are not as effective as currently available MOR agonists. Nitazenes are synthetic opioids that show high selectivity and efficacy at the MOR. Historically they have been deemed unsafe due to their strong potency and adverse effects, including active and potent metabolites. Nitazenes have been found in toxicology reports from overdose fatalities, and it has been suggested that they may contribute to the deleterious effects found in the poly-drug mixtures of street drugs. However, not all nitazenes are alike, and research is too limited to simply classify all nitazenes as harmful substances which merit no usage. In addition to optimizing MOR potency, efficacy, and signaling properties to decrease adverse effects of MOR agonists, we have been working on the development of novel fluorinated nitazene analogs with impaired brain permeability. Here we characterize the novel nitazene FDNZ (2-fluoroethyl-benzimidazole hydrochloride).

Methods: The affinity of FDNZ was assessed via competitive binding assays against [3H]DAMGO using rat membrane preparations. Increasing concentrations of unlabeled DAMGO and FDNZ were tested, and one-site competition curves were fitted for Ki values calculated using the Cheng-Prusoff equation. FDNZ efficacy was determined using [35S]GTPγS autoradiography in rat brain tissue. The behavioral and physiological effects of both compounds were assessed using intravenous self-administration (IVSA), analgesia, catalepsy, body temperature, and locomotion. We radiolabeled [18F]FDNZ to assess the in vivo kinetics and measure its bioavailability across time.

Results: Competitive binding assays against [3H]DAMGO in whole rat brain tissue membranes showed that FDNZ had a Ki ≈ 2 nM. FDNZ produced supra-maximal [35S]GTPγS binding compared to DAMGO. [18F]FDNZ was injected i.v. and rats were dynamically scanned for 60-minutes. [18F]FDNZ showed no brain accumulation. Analgesia was observed in a rat hot plate task with EC50 = 100 µg/kg. Catalepsy and hypothermia were observed at higher doses with only the highest dose (1 mg/kg) inducing a significant drop in temperature. Rats self-administered FDNZ and showed peak responding for 3 µg/kg/inf during a dose response test. However, rats immediately extinguished responding for FDNZ on the first session the active lever was replaced with saline and showed no drug-primed reinstatement of self-administration.

Conclusions: We show how a novel drug can be characterized by using several in vivo and in vitro radiometric and behavioral assays to gain an understanding of its function as a potential analgesic. Direct evidence of FDNZ in vivo kinetics with PET suggest that brain penetrance is poor, which suggests peripheral mediation for analgesic effects and may underlie FDNZ’s weak reinforcement. Future studies will focus on the effects of FDNZ on respiratory depression and tolerance.

Keywords: Nitazene, Mu opioid receptors, analgesia.

Disclosure: Nothing to disclose.

P629. Predictive genomic and electrophysiological insights into cocaine addiction susceptibility in rats

Ran Qiao, Michelle Doyle, Benjamin Johnson, Giordano de Guglielmo, Abraham Palmer, Marsida Kallupi

UCSD, La Jolla, California, United States

Background: Cocaine Use Disorder (CUD) manifests as persistent drug-seeking behavior with profound social and psychological consequences. Rodent models are instrumental in uncovering the neurobiological underpinnings of CUD, yet predicting addiction vulnerability in naive animals remains complex. This study leverages genomic and electrophysiological approaches to identify predispositions to cocaine addiction-like behaviors.

Methods: The RATTACA approach integrates genome-wide association study (GWAS) data, analyzing ~7.3 million SNPs from a 556-rat cohort to assess polygenic risk for cocaine addiction (heritability ~0.14). Using an 80:20 training-testing split, we identified the top and bottom 5% of rats as high- and low-vulnerability groups. Whole-cell patch-clamp recordings were performed on brain slices from the central amygdala (CeA) and nucleus accumbens core (NaCC) to evaluate baseline and cocaine-induced electrophysiological changes.

Results: In the CeA, high-vulnerability rats exhibited elevated baseline spontaneous inhibitory postsynaptic current (sIPSC) frequencies, suggesting enhanced GABAergic activity. Application of 1 µM cocaine ex vivo selectively reduced sIPSC frequency in the high-vulnerability cohort, indicating distinct GABAergic modulation. No differences were observed in sIPSC amplitude, kinetics, or glutamatergic transmission. In contrast, NaCC recordings showed no baseline GABAergic differences between groups, with both cohorts displaying comparable sIPSC reductions following cocaine exposure, highlighting region-specific neural adaptations.

Conclusions: This study reveals pre-existing GABAergic dysregulation in the CeA as a key factor in cocaine addiction susceptibility. By combining the RATTACA methodology with electrophysiological analyses, we uncover neurobiological markers of CUD vulnerability, offering insights into targeted therapeutic strategies and advancing our understanding of neurotransmitter dynamics in addiction risk.

Keywords: Addiction phenotypes, GABA and Glutamate, Animal Models.

Disclosure: Nothing to disclose.

P630. Targeting the right fibers: precision mapping of internal capsule pathways for personalized psychiatric neuromodulation

Marina Celestine, Chiara Maffei, Nicole McLaughlin, Steven Rasmussen, Anastasia Yendiki, Suzanne Haber

Harvard Medical School McLean Hospital, Belmont, Massachusetts, United States

Background: Precision targeting is the frontier of next-generation deep brain stimulation (DBS) and lesions for psychiatric illness. While motor pathways have been well characterized, the intricate trajectories of prefrontal and anterior cingulate (PFC/ACC) fibers—key to emotion and cognitive regulation—remain poorly resolved, particularly within the internal capsule (IC), a core target for neuromodulation in OCD and depression. These fibers follow complex, tortuous paths, posing a significant challenge for accurate visualization and therapeutic modulation. Building on our prior work identifying topographic PFC/ACC fiber organization in the anterior limb of the IC (ALIC), we now expand this atlas, introduce a probabilistic model to predict fiber positions, and link specific circuit disruptions to clinical outcomes in a case of psychiatric capsulotomy. This approach redefines our capacity to personalize DBS based on individual circuit anatomy.

Methods: We constructed a cross-species, multi-modal atlas of IC organization using non-human primate (NHP) tract-tracing and high-resolution diffusion MRI (dMRI). Histological tracer data were registered to dMRI space to validate the anatomical accuracy of tractography. Human homologs were identified using matched cortical seed regions. We then developed a probabilistic Gaussian-mixture model of fiber positioning within the IC, enabling prediction of fiber trajectories even where direct anatomical data are lacking. Finally, we applied this framework to a clinical case of OCD treated with capsulotomy, identifying which cortical circuits were disrupted and linking them to therapeutic effects.

Results: Across species and modalities, cortical fibers showed a conserved and predictable topographic organization entering, within and leaving the IC. Our probabilistic model captured this organization with high fidelity and revealed distinct patterns for fibers originating from functionally divergent prefrontal regions. In the OCD case, lesion-induced disruption of ventromedial and dorsal PFC pathways corresponded to improved affective symptoms, underscoring the translational power of anatomically informed targeting.

Conclusions: We demonstrate that cortical projections through the IC are not only systematically organized but can be accurately predicted using combined anatomical and imaging-based approaches. This work directly informs DBS and lesion targeting by identifying which fiber pathways are most likely to mediate clinical outcomes. By bridging basic neuroanatomy with human neuroimaging and intervention, this platform paves the way for precision neuromodulation in psychiatric illness.

Keywords: internal capsule, diffusion tractography, precision psychiatry, Obsessive Compulsive Disorder, capsulotomy.

Disclosure: Nothing to disclose.

P631. Obsessive-compulsive traits show dimension-specific associations with serotonin 4 receptor binding

Kristian Larsen, Paula Banca, Christelle Langley, Barbara Sahakian, Trevor Robbins, Gitte Knudsen

Copenhagen University Hospital, Rigshospitalet, Copenhagen, DK-2100 Denmark

Background: Obsessive-compulsive disorder (OCD) affects 1–2% of the population. It is characterised by persistent, intrusive thoughts (obsessions) and repetitive behaviours or mental rituals (compulsions) aimed at reducing distress or preventing feared outcomes. The serotonergic system is consistently implicated in the pathophysiology of OCD, but it is based mainly on the clinical efficacy of selective serotonin reuptake inhibitors (SSRIs), which remain the first-line treatment. Direct evidence for serotonergic dysfunction in OCD is limited. Molecular imaging studies are few, have produced inconclusive findings, and to date, no study has examined serotonin 4 receptor (5-HT4R) binding in relation to OCD. The 5-HT4R is a promising target because its binding potential is sensitive to changes in central serotonin levels in the brain and has been shown to decrease following prolonged SSRI treatment. In a randomised, double-blind clinical trial, we examined the association between serotonergic and clinical outcome measures in compulsive behaviour. We hypothesised that in individuals with obsessive-compulsive traits, 5-HT4R binding would be associated with symptom severity and distinct from that in non-compulsive individuals.

Methods: We preregistered our primary analysis as part of a clinical trial (ClinicalTrials.gov ID: NCT04336228). We included 23 unmedicated individuals with elevated obsessive-compulsive traits and 23 healthy controls, matched for age, sex, and education. Participants completed clinical assessments, including the Padua Inventory and the Obsessive-Compulsive Inventory (OCI) and underwent [¹¹C]SB207145 positron emission tomography (PET). The PET data were modelled with a single reference tissue model (SRTM), using cerebellum as the reference region, to estimate regional binding potential (BPND). Our primary model tested whether a latent factor of BPND across prefrontal (ventrolateral/dorsolateral prefrontal cortex, anterior cingulate) and striatal (caudate, putamen) regions was associated with obsessive-compulsive trait severity. We included age and injected radioligand mass as covariates. In a follow-up exploratory analysis, we applied principal component analysis (PCA) to Padua and OCI subscales to extract latent symptom dimensions and tested their associations with 5-HT4R binding.

Results: Regional analyses showed no significant group differences in 5-HT4R BPND between individuals with obsessive-compulsive traits and controls. In the preregistered latent variable model, higher Padua Inventory scores were significantly associated with higher 5-HT4R BPND (r = 0.56, p = 0.006), whereas OCI total scores were not associated with 5-HT4R BPND (r = −0.24, p = 0.30). Our PCA revealed three components of interest. The first principal component reflected overall obsessive–compulsive severity, with positive loadings across all but one subscale; this component was not associated with 5-HT4R binding (PC1: r = 0.19, p = 0.46). The second component contrasted contamination/washing against checking and doubt, such that higher 5-HT4R BPND was observed in individuals with greater contamination/washing symptoms (PC2: r = 0.51, p = 0.02). The third component contrasted obsessional harm-related thoughts with ritualistic behaviours (neutralising, checking, grooming), indicating that higher 5-HT4R BPND was associated with stronger harm-related obsessions (PC3: r = –0.73, p = 0.0001).

Conclusions: This study provides the first direct human evidence linking obsessive-compulsive symptom dimensions to serotonergic function in vivo. Elevated 5-HT4R BPND was associated with obsessive-compulsive traits measured by the Padua Inventory, but not with OCI scores or general severity, which suggests that these instruments capture different aspects of symptom expression. These findings lend support to dimensional frameworks of OCD neurobiology and suggest that serotonin-sensitive PET imaging could help identify biologically distinct subtypes. By refining our understanding of how low serotonin levels map onto discrete symptom domains, this approach may ultimately inform precision medicine strategies for OCD.

Keywords: obsessive-compulsive disorder (OCD), PET Imaging, Clinical Neurobiology, Clinical psychiatry.

Disclosure: Nothing to disclose.

P632. Response inhibition and error monitoring in adults with obsessive-compulsive disorder: preliminary reports from the global OCD initiative

Jônatas Santos, Marcelo Batistuzzo, Eyal Kalanthroff, Himani Kashyap, Niels de Joode, Benjamin Huber, Dianne Hezel, Madhuri Narayan, Clara Marincowitz, Christine Lochner, Dan Stein, Janardhanan Narayanaswamy, Helen Simpson, Odile A. van den Heuvel, Melanie Wall, Roseli Shavitt, Marcelo Hoexter

University of Sao Paulo, São Paulo, Brazil

Background: Response inhibition, the ability to suppress prepotent actions, is a key component of the broader executive function of inhibitory control and has been a primary focus in the search for neurocognitive deficits in Obsessive-Compulsive Disorder (OCD). Research suggests that circuits involving the prefrontal cortex and basal ganglia, particularly the inferior frontal gyrus (IFG) and pre-supplementary motor area (pre-SMA), are crucial for successful inhibition. Despite some insights, studies in OCD remain limited, especially those using multimodal imaging to understand brain mechanisms of inhibition. Such an approach could provide deeper insights into the neurocircuitry underpinning response inhibition in OCD patients and healthy controls. The Global OCD initiative was designed to address this gap by using a large, cross-cultural sample to identify reproducible brain signatures related to cognitive domains in OCD. Our proposed study explores primarily the response inhibition impairment and post-error slowing in adults with obsessive-compulsive disorder.

Methods: Behavioral data were analyzed from a large, pre-registered (https://osf.io/zsa6g/), cross-cultural study involving five international sites: Brazil, India, South Africa, the Netherlands, and the USA. A total of 268 unmedicated adults with OCD and 256 healthy controls (HC) completed the stop-signal task (SST). The primary outcomes were the stop-signal reaction time (SSRT), a measure of the latency of the stopping process, and post-error slowing (PES), an index of behavioral adjustment after mistakes. The secondary outcomes were mean reaction times in non-stop trials (nsRT), no-signal accuracy rates (nsACC), stop signal delay (SSD), mean RT of non-stop trials after correctly responding to non-stop trials (pNS), mean RT of non-stop trials after successful stop trials (pSS), mean RT of non-stop trials after failed stop trials (pSE). To test for group differences, we used generalized linear models and assessed three different models: an unadjusted model, a partially adjusted model controlling for site, sex, and age, and a final, fully adjusted model that also controlled for IQ, socioeconomic status (SES), and years of education. Furthermore, analyses searching for associations between SST measures of executive function and clinical variables - Yale-Brown Obsessive-Compulsive Scale (Y-BOCS), Hamilton Depression Scale (HAM-D), Hamilton Anxiety Scale (HAM-A), age of onset, treatment history, impulsivity scale - were run using data from the OCD group. To control for multiple testing type 1 errors we used the Bonferroni correction and adjusted the significance level by the number of variables separately for primary and secondary outcomes.

Results: No significant group differences were found in the primary outcomes of the SST: for the SSRT, the OCD group (Mean = 207.4 ms, SD = 71.6) did not differ significantly from HC (Mean = 198.6 ms, SD = 64.2; p = 0.140), with a negligible effect size (Cohen's d = 0.140). Similarly, for PES, no significant difference was observed between the OCD group (Mean = 9.59 ms, SD = 70.2) and HC (Mean = 12.9 ms, SD = 76.1; p = 0.273). Furthermore, no significant group differences were found across any of the six secondary outcomes after correcting for multiple comparisons (all p > 0.0083). Within the OCD patient group, task performance was not significantly associated with clinical characteristics. Specifically, SSRT showed no correlation with the severity of obsessive-compulsive symptoms (Y-BOCS score, p = 0.666), depression (HAM-D score, p = 0.422), or anxiety (HAM-A score, p = 0.953). Looking deeper, task performance also showed no significant association with specific OCD symptom dimensions, including hoarding, contamination, symmetry, or aggressive/taboo thoughts.

Conclusions: The absence of response inhibition deficits in this large, carefully controlled, cross-cultural study challenges the widely held view that impaired inhibitory control is a core feature or robust endophenotype of OCD. However, several factors should be considered when interpreting these null findings. It is possible that the specific visual modality of the stop-signal task used, which consisted of 300 sequential trials, may not have been sensitive enough to detect subtle deficits that could emerge with an auditory version of the task or with different methodological parameters. Given our large sample size, any medium to large effect size would have been detected, in order that the absence of a signal suggests that if a true deficit exists, it is likely small. Despite the null behavioral results, the lack of a group-level behavioral difference does not preclude the existence of subtle, underlying neurobiological variations. The critical next step of this project is to leverage the collected multimodal neuroimaging data - structural (sMRI), diffusion-weighted (DWI), and resting-state functional MRI (rsfMRI) - to conduct a detailed investigation of the structural and functional brain correlates of response inhibition across the entire sample, as outlined in our analysis plan. We will test specific hypotheses regarding the relationship between SST performance and gray matter volume in the IFG and pre-SMA, white matter integrity in fronto-basal-ganglia pathways, and resting-state functional connectivity between key inhibitory control regions, providing a more nuanced understanding of the neurocircuitry supporting response inhibition and its potential alterations in OCD.

Keywords: Obsessive Compulsive Disorder, Cognitive Functioning, Response inhibition, Error-monitoring, Cross-cultural.

Disclosure: Nothing to disclose.

P633. Multimodal brain imaging and blood-based markers of inflammation as predictors of response to celecoxib in treatment-resistant obsessive-compulsive disorder: a preliminary study

Gabriella Restifo-Bernstein, Francesca Zanderigo, Gabrielle Messner, Arturo Sanchez-Lacay, R. Todd Ogden, Dianne Hezel, Tyler Cutforth, Dritan Agalliu, H. Blair Simpson, Jeffrey Miller

Columbia University / New York State Psychiatric Institute, New York, New York, United States

Background: First-line pharmacological treatments for obsessive-compulsive disorder (OCD), serotonin reuptake inhibitors (SRIs), fail to achieve remission in the majority of patients. This may be related to the heterogeneity of biological pathways associated with OCD, which may not all be effectively targeted with the same treatment. Emerging evidence suggests that neuroinflammation may play a role in treatment resistant OCD (TR-OCD). In a pilot trial, we examined whether multimodal brain imaging- and blood-based inflammatory markers at baseline were associated with OCD severity, whether they predicted treatment outcome from an anti-inflammatory medication as augmentation to SRI treatment in individuals with TR-OCD, and whether treatment-associated changes in a peripheral inflammatory biomarker were associated with symptom change.

Methods: Participants included 12 adults (9 female and 3 male) with moderate to severe OCD (Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) score ≥16) despite current SRI treatment. Blood sampling was performed at baseline and following treatment (described below) to quantify C-reactive protein (CRP), a peripheral inflammatory marker. Baseline positron emission tomography (PET) imaging with 11C-ER176 and concurrent arterial blood sampling was performed to quantify binding to the 18 kilodalton translocator protein (TSPO) (radiotracer total volume of distribution, VT), indexing density of inflammatory cell types including microglia and astrocytes, within the cortico-striato-thalamo-cortical circuit (CSTCC) (orbitofrontal cortex (OFC), anterior cingulate cortex (ACC), caudate, putamen, and thalamus) implicated in OCD pathophysiology. T1-weighted structural magnetic resonance imaging (MRI) was acquired for coregistration to PET images and identification of regions of interest (ROIs). Given high inter-regional correlation of 11C-ER176 VT across ROIs, a scalar summary of 11C-ER176 VT within CSTCC was quantified and adjusted for TSPO genotype (TSPO-CSTCC). Baseline diffusion-weighted MRI was acquired to quantify extracellular water (dMRI-FW), a putative marker of neuroinflammation, within CSTCC. Participants then received 8 weeks of standardized open-label treatment with the anti-inflammatory medication celecoxib as augmentation to ongoing SRI treatment, at a dose of 100mg twice daily for one week, and then 200mg twice daily for seven weeks if well-tolerated. OCD severity was assessed every two weeks with the Y-BOCS.

Clinical improvement from celecoxib was calculated as the slope of all available Y-BOCS scores during the trial using linear regression. Linear regression was used to assess relationships between baseline inflammatory indices (CRP, TSPO-CSTCC, dMRI-FW) and clinical outcomes (baseline Y-BOCS and clinical improvement), as well as between longitudinal change in CRP and clinical improvement. Multiple linear regression was used to explore prediction models incorporating multiple biomarkers. Prediction of clinical outcome was assessed in study completers (n = 10, with PET data available in n = 7). Given this sample size, we report relationships of imaging and blood markers to clinical outcome using effect size (Pearson r values).

Results: OCD symptom severity improved significantly following celecoxib treatment in the full sample (n = 12; Week 0 = 24.6 ± 5.5; final Y-BOCS = 19.1 ± 5.9; p < 0.001) and in study completers (n = 10; Week 0 = 24.2 ± 5.5; Week 8 = 17.5 ± 4.4; p < 0.001). We did not observe a relationship between sex and any clinical outcome or biomarker (all p > 0.5) except for TSPO-CSTCC, which was higher in males (p = 0.012). Baseline dMRI-FW in OFC and ACC was strongly associated with clinical improvement (OFC: r = 0.72; ACC: r = 0.75) but only weakly correlated with baseline OCD severity (OFC: r = 0.27; ACC: r = 0.10), suggesting specificity to anti-inflammatory treatment outcome. Baseline CRP was correlated with symptom improvement (r = 0.52) and also with baseline OCD symptom severity (r = 0.68); furthermore, post-treatment reductions in CRP were associated with clinical improvement (r = 0.38). Baseline TSPO-CSTCC was correlated with symptom improvement (r = 0.41) and was associated with baseline OCD severity (r = 0.30). Combining all biomarkers in a multiple linear regression model to predict clinical improvement yielded the highest adjusted r value among all models.

Conclusions: Celecoxib augmentation significantly improved OCD symptoms in adults with TR-OCD, although treatment outcomes varied across participants. Baseline dMRI-FW in cortical regions of the CSTCC, CRP, and TSPO-CSTCC predicted clinical response to celecoxib treatment, such that higher pre-treatment levels of these inflammatory biomarkers were all associated with greater response to celecoxib; combining biomarkers improved prediction. These preliminary findings support the hypothesis that inflammation is present in a subset of individuals with TR-OCD, and that these individuals may benefit from anti-inflammatory treatments. If replicated in larger samples, this finding may help advance a precision medicine approach to TR-OCD.

Keywords: obsessive-compulsive disorder, inflammation, positron emission tomography, C-reactive protein, diffusion-weighted magnetic resonance imaging.

Disclosure: Abbvie, Grant, Self.

P634. Learning and extinction of repetitive avoidance behavior: neural correlates and effects of trait obsessive-compulsivity

Hannah Berg, Riley Rozniarek, Rayus Kuplicki, Helen B. Simpson, Martin Paulus, Robin Aupperle

Laureate Institute for Brain Research, Tulsa, Oklahoma, United States

Background: Compulsions are a hallmark of obsessive-compulsive disorder (OCD), but the neural mechanisms underlying compulsions remain sparsely understood. In the present study, we employed a novel behavioral paradigm during functional magnetic resonance imaging (fMRI) to elicit compulsion-like repetitive avoidance behavior and to study its neural correlates and investigate obsessive-compulsive related individual differences.

Methods: Adults (n = 30, mean age = 32.35, 20 female), including those with anxiety- and OCD-related diagnoses (n = 12), were recruited to represent a range of obsessive-compulsive traits. Participants completed the Dimensional Obsessive-Compulsive Scale (DOCS; mean = 11.28), followed by the Tap-To-Safety Task, an experimental paradigm designed to elicit compulsion-like behavior, during fMRI. The task began with Pavlovian conditioning to a threat cue (CS+) paired with mild electric shock, and safety cues never paired with shock (CS-). In the behavioral test phase, participants could repeatedly tap a button to reduce the risk of shock, at the cost of reduced point accrual. Next, during an extinction phase, no shocks were delivered regardless of tapping. Regions of interest were identified as areas with activity to CS+> CS- during tap-preparation (“You may soon tap the button”). Effects of DOCS scores and neural activity on tapping behavior were examined.

Results: During the behavioral test, heightened activity to CS+ (>CS-) during tap-preparation was identified in dorsal anterior cingulate cortex and left and right dorsal striatum (voxelwise p < .001). Greater activity in these regions was associated with more tapping behavior across threat and safety cues (ps < .002, ηp² > .05). During extinction, higher DOCS scores were associated with less-steep declines across trials in a) tapping behavior to the former threat cue (F([1,143] = 4.61, p = .034, ηp² = .04) and b) right dorsal striatal activity to the former threat cue (F[1,113] = 4.74, p = .031, ηp² = .04).

Conclusions: These preliminary findings provide evidence for a cortico-striatal mechanism underlying the persistence of repetitive avoidance behavior. Findings suggest that obsessive-compulsive traits may stem from dysfunctional safety learning, involving excessive persistence of dorsal striatal threat representations in safe situations, providing a neurobehavioral target for future translational research and intervention.

Keywords: Anxiety, fMRI, obsessive-compulsive disorder (OCD), Fear conditioning and extinction, avoidance.

Disclosure: Nothing to disclose.

P635. Brain metabolic correlates of SAPAP3 gene deletion: effect of treatment with psilocybin and psychedelic mushroom extract

Michal Brownstien Zerem, Tareq Eweiwi, Alexander Botvinnik, Bernard Lerer

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

Background: SAPAP3 is a scaffolding protein enriched in the postsynaptic density of excitatory synapses. Mice homozygous (HOM) for SAPAP3 deletion (SAPAP3-KO) manifest compulsive self-grooming, tic-like head-body twitches, and anxiety. SAPAP3-KO is an increasingly accepted model for OCD. We implemented a randomised, controlled trial of psilocybin (PSIL) and psychedelic mushroom extract (PME) in HOM SAPAP3-KO mice. A subset of mice underwent brain 18FDG-PET/MRI before and after treatment to determine brain metabolic correlates of SAPAP3 deletion and treatment.

Methods: HOM SAPAP3-KO mice of both sexes were randomly assigned to a single i.p. injection of PSIL 4.4 mg/kg, PME (psilocybin 4.4 mg/kg), or vehicle (VEH) and blindly evaluated up to 42 days. 4 (2 F) VEH-treated, 6(4 F) PSIL-treated, and 6(3 F) PME-treated HOM mice underwent FDG-PET before treatment (and 7(4 F) wild-type [WT] mice), as well as 21 days after. Images were analyzed using PMOD. Brain ROIs in the striatum, cortex, anterior cingulate, and thalamus, based on the PNROD mouse atlas, and normalized to cerebellum, were used to calculate 18FDG uptake.

Results: PSIL and PME eliminated the >100% increase in self-grooming and head-body twitches observed in VEH-treated mice (Brownstien et al., 2024). Across genotypes, two-way ANOVA showed a significant effect of brain area (F3,77 = 9.54, p < 0.0001; Striatum p < 0.001 vs. cortex and anterior cingulate). Uptake was lower in HOM than WT across the 4 brain areas (Genotype F1,77 = 4.75, p = 0.03). Following treatment, uptake of HOM mice administered VEH increased (striatum) or remained unchanged (cortex); uptake of PSIL or PME treated mice was lower (time × treatment interaction in the striatum [F2,10 = 4.92, p = 0.032; PSIL, Pre vs. Post, p = 0.027], and treatment effect in the cortex (F2,11 = 5.22, p = 0.025; PME vs. VEH, p = 0.02).

Conclusions: (1) Brain metabolic activity is highest in the striatum of SAPAP3 KO mice, irrespective of genotype. (2) HOM mice are characterized by lower activity than WT mice. (3) Following VEH treatment, activity increases further in the striatum; it remains unchanged in the cortex, while PSIL and PME reduce uptake. These observations accord with imaging findings in OCD and with SAPAP3 KO mouse studies using other modalities. Analyses of sub-regions of these and other brain areas are underway.

Keywords: Psilocybin, Obsessive-Compulsive Eisorder (OCD), MR-PET, Psychedelics.

Disclosure: Negev Labs, Consultant, Self, DioTree, Founder, Self, ParowBio, Consultant, Self.

P636. Continuous performance improvement for the administration of long-acting medication in psychiatric and substance use disorders

Tanya Alim, Michelle Brooks, Sanaa Belrhiti, Walter Bland

Howard University, White Plains, Maryland, United States

Background: Mental health disorders significantly impair daily functioning for many individuals. In the United States, schizophrenia affects approximately 1% of the population, and bipolar disorder affects approximately 2.6% of adults 18 years or older. While oral antipsychotic medications are commonly prescribed, relapse is common. Long-acting medications (LAMs) offer important benefits by improving adherence and reducing relapse risk. Evidence also supports their utility among patients with co-occurring substance use disorders. The advantages of LAM include improved adherence, greater clarity on treatment discontinuation, and less frequent dosing (Wang et al, 2024). Patients treated with LAMs have better outcomes, including reduced mortality and lower risk of re-hospitalization (Sajatovic et al, 2018).

Despite these benefits, LAMs remain underutilized, with barriers to widespread use at the patient, provider, and system levels (Lindenmayer et al., 2020). Importantly, racial disparities have been documented: Black patients are more likely to be prescribed LAMs compared to White patients, often in the context of coercive or mandated treatment settings, raising concerns about equity and patient autonomy (Bareis et al, 2020). The National Council for Mental Wellbeing recommends that organizations use a continuous quality improvement (CQI) process to improve patient understanding of and access to LAMs.

Methods: At Howard University Hospital (HUH), we implemented a LAMs program in December 2023 to improve access for patients with psychiatric and substance use disorders. Partnerships with pharmaceutical collaborations enabled uninsured patients to access free or reduced-cost medications, including newer agents, through a patient assistance program. Inpatients at HUH were systematically educated about and offered LAMs. Acceptance rates were tracked from electronic health record reviews and presented monthly at the Department of Psychiatry CQI committee. Barriers to acceptance, including patient preference and clinician factors, were recorded. Improvement efforts followed the Plan, Do, Study, and Act (PDSA) model of performance improvement. IRB exemption is being sought to expand outpatient monitoring of LAMs and evaluate additional barriers.

Results: Between 2023 and 2025, approximately 200 patients with a psychotic disorder diagnosis and 100 patients with alcohol use disorder were admitted to the inpatient service. Forty-three patients accepted and received LAMs: 33 patients received antipsychotic LAM, and 10 patients received extended-release naltrexone. The most common reason for refusal was patient preference for oral medications. Educational information regarding the benefits of LAMs was provided, though variability in provider delivery was noted. Patients were offered the LAMs even if they did not have a history of non-adherence.

Conclusions: Implementation of an LAM program at an urban academic hospital expanded access through pharmaceutical partnerships and routine inpatient education. While more patients were offered the LAMs as compared to previous years, the frequency of acceptance by patients is still low. Strategies to improve LAM administration will include enhancing education of patients and providers on the inpatient and outpatient teams. The importance of offering LAMs regardless of adherence is particularly important.

Keywords: Long-acting antipsychotics, long-acting injectable, psychiatric and substance use disorders.

Disclosure: Nothing to disclose.

P637. Variation in M1 and M4 muscarinic cholinergic receptor genes and human striatal dopaminergic function

Daniel Eisenberg, Philip Kohn, Michael Gregory, Jasmin Bettina, Bhaskar Kolachana, Karen Berman

Clinical and Translational Neuroscience Branch, National Institute of Mental Health, Bethesda, Maryland, United States

Background: Dopamine systems dysfunction in schizophrenia spectrum illness has been a central principle guiding current understanding of the biology of psychosis. Molecular neuroimaging experiments have repeatedly identified schizophrenia-associated elevations in striatal dopamine synthesis capacity, a phenotype that may be relevant to therapeutic response. Furthermore, nearly all pharmacological treatments for schizophrenia involve D2 dopamine receptor blockade. However, additional neurochemical systems have also been implicated in psychosis, including the acetylcholine system. An agent targeting M1 and M4 muscarinic cholinergic receptors has recently been approved for the treatment of schizophrenia in the US, and clozapine, a unique antipsychotic medication with efficacy in treatment resistance, also has affinity at these receptors. Preclinical data have identified acetylcholine-dopamine intersections in the striatum, where both neurotransmitters abound, and suggested the possibility of meaningful muscarinic effects on D1 dopamine receptor operations. How alterations in muscarinic cholinergic function might influence striatal dopaminergic systems may be relevant to antipsychotic mechanisms, though in vivo data on potential acetylcholine-dopamine interactions in the human brain are limited.

Methods: One hundred and fifteen individuals free of neuropsychiatric illness (age 38+/−11 years, 58 female) participated in neuroimaging and provided blood for genotyping at the NIH Clinical Center in Bethesda, Maryland. PET scans were performed using [11C]-NNC112 (N = 115) to measure D1 dopamine receptor availability. A separately acquired T1-weighted MRI scan was obtained for anatomical segmentation and spatial normalization of PET images. Participants abstained from caffeine and nicotine for four hours prior to neuroimaging. PET emission data were collected over 90 minutes following tracer injection. Attenuation-corrected, dynamically binned frames were normalized to MNI space using ANTS software and smoothed to improve signal-to-noise ratios. These data were then subjected to modeling using the simplified reference tissue model with a cerebellar reference region as implemented in PMOD to estimate BPnd. Genetic data were obtained from Illumina SNP chips, and genetic scores predicting CHRM1 and CHRM4 striatal expression were calculated for each individual based on cis-eQTL data from the GTEx project. Voxelwise general linear modeling controlling for sex, age, and ancestry-genetic components tested for associations between gene scores and striatal BPnd.

Results: Individuals with genetic variation predicting greater M1 receptor gene expression showed greater striatal D1 receptor availability (p < 0.005, uncorrected). Genetic variation predicting greater M4 receptor gene expression was associated with less striatal D1 receptor availability (p < 0.005, uncorrected).

Conclusions: These results echo prior preclinical data and are consistent with the possibility that genetic biases associated with muscarinic receptor gene expression are relevant to striatal dopamine systems in the living human brain. If replicated, the divergent muscarinic receptor subtype genetic effects on D1 dopamine operations may merit further study. Additional work in clinical populations is needed to better understand whether these possible cholinergic-dopaminergic interactions relate to muscarinic therapeutic action.

Keywords: Acetylcholine, D1 Dopamine Receptors, Dopamine, PET, M1 and M4 Muscarinic Receptors.

Disclosure: Nothing to disclose.

P638. Using large language models to model semantic predictive processing in the brain and alterations in schizophrenia

Asieh Zadbood, Guillermo Horga

Columbia University, New York, New York, United States

Background: Predictive processing models are an influential mechanistic framework for psychosis, proposing that altered prediction-error signals underlie psychosis. However, work in this area falls short of explaining the type and content of psychotic symptoms, especially the characteristic semantic content, which may require an explanation at the level of semantic predictions. Semantic predictions and prediction errors occur during natural language comprehension, where the brain continuously anticipates the next word and content. Therefore, natural language comprehension serves as a useful test for studying semantic predictive processing in the brain. Recently, large language models (LLM), built on next-word prediction, have been effectively used to extract neural signals relevant to semantic prediction during natural language understanding. Additionally, their generative nature in predicting the next word offers a valuable tool for quantifying prediction errors in speech processing. We apply LLM-driven semantic prediction errors to model brain activity during natural language comprehension in healthy individuals and those with schizophrenia. We hypothesize that semantic prediction errors during speech comprehension are reflected in neural fMRI responses of brain regions involved in language processing, and that this relationship is weakened in individuals with psychosis.

Methods: We used a publicly available fMRI dataset with 80 healthy participants (age 18–45 years, 45 females) who listened to a 7-minute English narrative (Pieman from the Moth series). The fMRI data were collected on a 3-T Skyra scanner (2 mm isotropic voxels, TE = 0.028 s, TR = 1.5 s). Participants were instructed to passively listen to the story without any specific guidance aside from paying close attention to the narrative. Data were preprocessed using fMRIprep. We performed detrending, high-pass filtering (128 s cutoff), and noise removal (including 6 head motion regressors and their derivatives, and the first 5 CompCors for white matter and CSF) using Nilearn. GPT-2 was employed to extract model variables relevant to semantic predictive processing. Semantic prediction error was calculated as cross-entropy for each token. These cross-entropy values were averaged across each TR to produce a vector of values equal in length to the number of TRs. We then correlated this vector with neural signals in predefined regions of interest (ROIs). For the schizophrenia dataset, we accessed a publicly available sample of 25 healthy controls (HC; age 20–64 years, 7 females), 23 individuals with schizophrenia experiencing auditory hallucinations (AVH+; age 19–66 years, 3 females), and 23 individuals with schizophrenia without auditory hallucinations (AVH-; age 31–61 years, 7 females). The speech task included passive listening, although the stimulus was not a continuous narrative; it consisted of 48 semantically unrelated sentences presented across six blocks (8 sentences per block). The speech was in Spanish, and all participants were native Spanish speakers. Other task blocks included words, reversed speech, and white noise. Preprocessing and the main analysis (semantic prediction error measure from GPT-2 correlated with brain signal) were performed using the same methods as stated previously. Additionally, we conducted a GLM analysis to assess the neural response during speech perception using a main contrast [speech > white noise].

Results: During narrative comprehension in healthy participants (Pieman set), we found a significant correlation between semantic prediction error (cross-entropy) and neural signal in a set of ROIs overlapping with the auditory cortex, superior temporal sulcus, lateral frontal regions, parahippocampal cortex, and retrosplenial region, after correction for multiple comparisons. In our a priori ROI analysis of the superior temporal gyrus, we observed a mean correlation value of r = 0.02 (t(79) = 2.78, p = 0.04). In the schizophrenia dataset, there was no significant difference in responses to speech (vs. noise) between groups in this ROI. In the prediction error analysis, however, a one-way ANOVA revealed a main effect of group (F(2,68) = 3.92, p = 0.02), driven by a significantly weaker relationship between semantic prediction error and neural signals in AVH+ compared to AVH- (t(44) = −2.66, p = 0.01).

Conclusions: Our findings in a large group of healthy participants suggest that LLMs provide a reliable measure for assessing the semantic prediction error signal in neural fMRI recordings, consistent with previous work with intracranial recordings. Additionally, our preliminary results suggest alterations in semantic predictive processing in psychotic individuals with AVH, alterations that were not paralleled by general changes in speech responses using traditional analyses. Our schizophrenia findings are limited because the task used a set of semantically unrelated sentences instead of a continuous narrative; this prevented us from fully leveraging the LLM models, which tend to perform better with longer contexts. Future studies can address this limitation by collecting data during narrative comprehension in individuals with schizophrenia. These results nevertheless support the utility of LLM-based modeling of speech data in examining alterations in semantic predictive processing in psychotic and other clinical populations, offering a novel approach to probe the predictive processes of interest with simple passive listening paradigms.

Keywords: Large Language Models, psychosis, Semantic prediction error, Predictive coding, speech processing.

Disclosure: Nothing to disclose.

P639. Identifying trajectories of depressive symptoms in early psychosis

Lindsay Oliver, Julia Gallucci, Sierra Vaillancourt, Alyssa Qian, Katie Lavigne, Martin Lepage, Aristotle Voineskos, Colin Hawco

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

Background: Depressive symptoms are common, poorly understood, and associated with negative outcomes in individuals with psychosis. The overlap between depressive and negative symptoms interferes with determining diagnostic specificity, underlying neurobiology, and targeted treatments for depression in psychosis. Previous work from our group suggests that divergent neurobiological pathways underlie depressive versus negative symptoms, and that neurostimulation may be an effective treatment for general depressive symptoms in schizophrenia spectrum disorders. Heterogeneity in clinical and functional outcomes across the course of psychosis and between individuals also highlights the need for individualized treatment approaches. Thus, early characterization of symptom progression and identification of individuals who will experience persistent depressive symptoms is crucial. Our aim was to identify trajectories of depressive symptoms in individuals with first episode psychosis across two years of an early intervention service. The overall goal is to identify predictors of symptom course for early identification of individuals likely to develop persistent depressive symptoms.

Methods: The study utilized data from individuals diagnosed with first episode psychosis across sexes, aged 14–35 years old, who were part of The Prevention and Early Intervention Program for Psychoses (PEPP-Montreal), a two-year early intervention service in Montreal, Canada. Baseline sociodemographic, non-social and social cognitive, and a subsample of structural magnetic resonance imaging data were collected, as well as longitudinal functioning and clinical data. Depressive symptoms were assessed using the Calgary Depression Scale for Schizophrenia (CDSS) over a two-year follow-up period, with the current analysis including eight timepoints (baseline, 1, 2, 3, 6, 9, 12, and 18 months). Growth mixture modelling was used to parse heterogeneity and identify trajectories of depressive symptoms. Model selection was determined based on fit statistics, percentage of participants per class, and clinical interpretability.

Results: Among the 611 participants included in trajectory modelling (185 female; mean age = 24 years, SD = 4.7), a range of baseline diagnoses were represented, including schizophrenia, bipolar I disorder with psychotic features, depression with psychosis, psychosis not otherwise specified, and other first episode psychosis diagnoses. Trajectories of depressive symptoms were best modelled using a three-class growth mixture model based on the combination of the Bayesian information criterion, scaled entropy, average posterior probabilities, bootstrap likelihood ratio tests (3-class > 2-class solution: p < .001), and interpretability. The three identified trajectories were 1) persistently low depressive symptoms (N = 411, 67%), characterized by low baseline symptom severity that remained relatively stable over time; 2) decreasing depressive symptoms (N = 140, 23%), beginning with elevated symptom severity and showing a marked decline over the follow-up period; and 3) persistently high depressive symptoms (N = 60, 10%), which showed high baseline symptom severity that remained elevated relative to other classes throughout the follow-up period. Trajectory classes did not differ in age at entry, gender, or IQ, but differences were seen in depressive symptoms at baseline and 18 months, negative symptoms at 18 months, and diagnoses represented (all p < .01).

Conclusions: The findings revealed three distinct trajectories of depressive symptoms. Effective treatments for depressive symptoms are a major unmet need in people with psychosis, given their impact on functional outcomes. Understanding trajectories of depressive symptoms in the early stages of psychosis, as well as how these relate to negative symptoms, is critical for the development of personalized treatment options, such as neurostimulation. This work will also examine overlap with negative symptoms and predictive models for early identification of patients with a high likelihood of developing persistent depressive symptoms and poor outcomes, to inform more precise and earlier interventions for those who will need it most.

Keywords: early psychosis, depressive symptoms, Negative Symptoms, symptom trajectory.

Disclosure: Nothing to disclose.

P640. Neural correlates of stress and alcohol cue reactivity in chronic pain

Milena Radoman, Cheryl Lacadie, Rajita Sinha

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

Background: Chronic pain, which is defined as pain that lasts or recurs for more than three to six months beyond the normal healing time, is a significant public health issue affecting nearly 116 million Americans and costing the United States an estimated $635 billion annually. One in four individuals with chronic pain use alcohol to alleviate suffering, and those who do report greater incidence of alcohol use disorder. Growing research suggests that chronic pain may disrupt stress and reward processing, and that such dysfunction may increase stress- and alcohol cue-related craving, a well-established risk factor for the development of alcohol use disorder. Stress and alcohol cues have been shown to robustly engage overlapping cortico-striatal-limbic networks, including the insula, dorsal anterior cingulate cortex, amygdala, striatum, and ventromedial prefrontal cortex, all of which are also central to pain processing. However, it is not well understood how chronic pain may alter neural responses to these cues. To address this gap, the present study examined the subjective craving and neural responses to stress and alcohol cues relative to neutral cues in individuals with chronic pain compared with healthy controls.

Methods: Twenty-four treatment-entering men and women (ages 18–60) with chronic pain (CP) and 23 healthy controls (HC) underwent fMRI scanning during which they participated in a well-validated cue provocation task exposing them to standardized and matched stress (S), alcohol (A) and neutral control (N) cues over six successive runs in a randomized block design per condition with repeated subjective alcohol craving assessments. Linear mixed effects (LME) models with a random intercept were conducted to assess main and interaction effects for group (CP and HC) and condition (S, A and N) on alcohol craving during fMRI. A whole brain, voxel-based second level 3dLME (AFNI) analysis (p < .001, whole brain cluster correction at α < .05) examined the condition (S, A, N) x group (CP vs HC) interaction, with participant as a random effect and age and sex as covariates.

Results: Across alcohol, stress, and neutral cue conditions, no significant group differences in subjective alcohol craving were observed (p > .05), indicating that subjective craving did not distinguish participants with chronic pain and healthy controls. In fMRI, following a significant Task x Group interaction, patients with chronic pain demonstrated heightened activation in the right putamen and rostral anterior cingulate cortex (rACC) during exposure to neutral-relaxing cues relative to healthy controls. In contrast, during stress versus neutral cues, chronic pain was associated with a predominantly blunted striatal-limbic response, including reduced activation in the right amygdala, anterior insula, putamen, bed nucleus of the stria terminalis (BNST), and dorsal anterior cingulate cortex. Finally, during alcohol versus neutral cues, patients with chronic pain exhibited hyperactivation in the BNST, nucleus accumbens, dorsal caudate, and posterior thalamus, but blunted responses in the orbitofrontal cortex and rACC, relative to healthy controls.

Conclusions: Despite no group differences in subjective alcohol craving, chronic pain was associated with distinct neural alterations across cue conditions. Heightened responsivity to neutral-relaxing cues, blunted striatal-limbic engagement during stress, and a mixed pattern of hyperactivation in subcortical salience/reward regions with reduced prefrontal control during alcohol cues suggest that chronic pain shifts the balance of neural processing across affective and motivational circuits. These preliminary findings highlight that pain-related changes in brain function may reveal hidden markers of vulnerability not captured by self-report, providing mechanistic insight into pathways linking chronic pain, stress dysregulation, and risk for maladaptive coping.

Keywords: chronic pain, Cue Reactivity, craving, fMRI.

Disclosure: Nothing to disclose.

P641. Peak alpha frequency slowing predicts pain after traumatic injury hospitalization

Heather Webber, Amélie Grandjean, Ali Mazaheri, Bruno Kluwe-Schiavon, Joy Schmitz, Jin Yoon, John Harvin, Preethi Gunaratne, Julia Simon, Kendall Farrell, Kandice Motley, Erin Fox, Jessica Vincent, Darrion Mouton, Consuelo Walss-Bass

The University of Texas Health Science Center At Houston, Houston, Texas, United States

Background: Non-medical use of opioids and opioid overdoses remain a prominent public health concern. Those experiencing continued pain following a traumatic injury or surgery are at risk of persistent prescription opioid use. Discovery of biomarkers that are predictive of continued pain could be useful in identifying groups of people for targeted early intervention and prevention strategies. Slowing of the peak alpha frequency (PAF) measured with electroencephalogram (EEG) has been linked to chronic pain and is predictive of postoperative pain sensitivity in lung cancer patients. The current study aimed to expand these findings by assessing if PAF measured 1-week post-trauma injury hospitalization is predictive of continued pain 1- and 3-months post-hospitalization.

Methods: A prospective cohort observational design was used to assess associations between PAF and pain outcomes. Participants were recruited and consented following admission to the UTHealth Red Duke Trauma Institute for a traumatic injury or trauma surgery due to injury. Five minutes of eyes-closed resting-state EEG was collected at 7 days post-discharge. Self-reported pain (pain catastrophizing scale, brief pain inventory) were collected 7 days (V1: n = 38), 1 month (V2: n = 33), and 3 months (V3: n = 25) post-discharge. EEG data were segmented into 5-second non-overlapping epochs. Frequency decomposition was performed on cleaned EEG data using FieldTrip. Peak frequency per person in the alpha range (7–14 Hz) was calculated at each epoch and then averaged across epochs. PAF was extracted using the center of mass method and averaged over the central electrodes (C3, Cz, C4). We assessed the association between PAF and the pain scales at each timepoint using Pearson’s correlations.

Results: Overall, pain catastrophizing (V1 = 11.79; V2 = 9.36; V3 = 5.40), pain interference (V1 = 5.10; V2 = 3.86; V3 = 1.78), and pain severity (V1 = 3.70; V2 = 2.48; V3 = 1.63) decreased with each visit. PAF collected at V1 was associated with pain catastrophizing at V2 (r = −.347, p = .048) and V3 (r = −.470, p = .018). PAF was lower for those in the upper quartile of pain catastrophizing (M = 9.48, SD = .62) compared to the lower quartile (M = 10.21, SD = .31) at V2 as well as at V3 (M = 9.72, SD = .66; M = 10.26, SD = .41). PAF collected at V1 was also associated with pain interference (r = −.453, p = .023) and severity (r = −.535, p = .006) at V3 only.

Conclusions: Slowing of PAF (i.e., lower peak frequencies) measured at 7 days following traumatic injury hospitalization was predictive of continued post-discharge pain. Specifically, lower PAF was predictive of higher pain catastrophizing at 1- and 3- months post-discharge and higher pain interference and severity at 3 months post-discharge. Therefore, while pain measures went down with time overall, those with a lower PAF might be more susceptible to persistent pain that continues for months after hospitalization. These preliminary results highlight the utility of PAF as a potential biomarker of risk for persistent pain and pain interference. If replicated, these results have implications for individualized treatments for those at risk, with the goal of lowering the impacts of post-discharge pain and, ultimately, lowering continued prescription opioid use.

Keywords: EEG biomarkers, Pain, prescription opioids.

Disclosure: Nothing to disclose.

P642. Delta-9-tetrahydrocannabinol-sparing effects of cannabigerol: a placebo-controlled human laboratory study

Elisa Pabon, Stephanie Lake, Conor H. Murray, Alisha Eversole, Katherine Hampilos, Samantha Baglot, Timothy Fong, Ziva Cooper

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

Background: Pain remains a major public health challenge, with limited treatments that are both effective and well-tolerated. Appetite loss frequently accompanies pain, and greater appetite impairment correlates with higher pain intensity. Delta-9-tetrahydrocannabinol (THC), the principal psychoactive compound in cannabis, can reduce pain and stimulate appetite and is FDA-approved for anorexia. Yet, its therapeutic potential is constrained by intoxicating and abuse-related effects. Cannabigerol (CBG), a non-intoxicating minor cannabinoid, has emerged from preclinical studies as a promising candidate with both analgesic and orexigenic properties and potentially anxiolytic effects, without THC’s psychoactive profile. To our knowledge, this is the first human study to directly compare the effects of vaporized THC, CBG, and their combinations across varying ratios on pain, appetite, and subjective experience in healthy volunteers.

Methods: Healthy adults (21–55 years) who reported cannabis use within the past month were recruited for a placebo-controlled, within-subject study. Exclusion criteria included current chronic pain, use of prescription medications, major psychiatric disorders and women who were pregnant or lactating. After biochemical confirmation of cannabis abstinence for at least twelve hours, participants inhaled vaporized THC (0 mg, 5 mg, 15 mg) and CBG (0, 5 mg, 15 mg) alone and co-administered (5 mg THC + 5 mg CBG, 15 mg THC + 5 mg CBG, 5 mg THC + 15 mg CBG, 15 mg THC + 15 mg CBG) in a randomized order across nine outpatient sessions. Subjective effects were assessed via Mood and Physical Symptoms Visual Analog Scales (MPS-VAS) and the Vaporized Cannabis Rating Form (V-CRF). Analgesic effects were assessed via the 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. Mixed effects models and post-hoc pairwise comparisons tested the effects of THC and CBG dose on all outcome measures (α = 0.05).

Results: Fourteen healthy adults (28 ± 8 years, 42.9% female) who reported occasional cannabis use (1.6 ± 0.9 days/week) completed the study. There were significant main effects of dose on ratings of ‘Hungry’ and ‘Want food’ (p < 0.001). ‘Hungry’ ratings increased with 5 mg THC, 15 mg CBG, and THC-CBG combinations (15 mg THC + 5 mg CBG; 15 mg THC + 15 mg CBG) relative to placebo (p ≤ 0.01). ‘Want food’ ratings increased with 5 mg THC, 5 mg CBG, and THC-CBG combinations (5 mg THC + 5 mg CBG; 15 mg THC + 5 mg CBG; 15 mg THC + 15 mg CBG) (p ≤ 0.05). There was a significant main effect of dose on ratings of ‘Anxious’ (p < 0.001), with 15 mg THC and 15 mg THC + 5 mg CBG increasing anxiety (p ≤ 0.01) and 15 mg CBG decreasing anxiety (p < 0.01) relative to placebo. ‘High’ and ‘Good effect’ ratings increased with all doses of THC (p < 0.001); 15 mg CBG also increased these ratings (p ≤ 0.05) relative to placebo. In THC-CBG combinations, CBG did not significantly increase ratings of ‘High’ or ‘Good Effect’ relative to THC alone. Ratings of ‘Bad effect’ increased with all THC doses (ps < 0.001) relative to placebo. CBG alone did not impact these ratings and in all THC-CBG combinations, CBG did not increase these ratings relative to THC alone. Dose effects on pain threshold (p < 0.01) and tolerance (p < 0.001) were modest; 5 mg THC + 5 mg CBG decreased pain threshold (p < 0.05), whereas 5 mg CBG increased pain tolerance (p < 0.05) relative to placebo.

Conclusions: In healthy adults who use cannabis, vaporized CBG selectively enhanced appetite and increased pain tolerance and, at 15 mg, reduced anxiety and elicited abuse-related subjective effects. THC robustly increased abuse-related subjective effects and co-administration of CBG did not alter these effects. These findings suggest CBG may offer orexigenic, anxiolytic, and analgesic benefits with a more favorable side-effect profile than THC, warranting further investigation in clinical populations experiencing pain and appetite loss.

Keywords: tetrahydrocannabinol, cannabigerol, cannabinoid, Acute effects.

Disclosure: Nothing to disclose.

P643. Lifetime correlates and consequences of chronic antipsychotic exposure in the veterans health administration

Tim Bigdeli, Yash Joshi, Yuli Li, David Graham, Georgios Voloudakis, Gregory Light, Alan C. Swann, David L. Braff, Thomas R. Kosten, Cooperative Studies Program #572, Million Veteran Program, Sumitra Muralidhar, Grant D. Huang, J Michael Gaziano, Mihaela Aslan, Panos Roussos, Philip D. Harvey

VA New York Harbor Healthcare System, Brooklyn, New York, United States

Background: The Veterans Health Administration (VHA) is the largest integrated healthcare system in the United States, with 170 medical centers and 1,193 outpatient clinics serving 9.1 million veterans. Patients utilizing VHA services are more likely to suffer from multiple chronic health conditions than the general population. These risks are even more profound among Veterans treated with antipsychotic medications.

Methods: We surveyed prescription records in the electronic health records (EHRs) of >700,000 participants in Cooperative Studies Program #572 and the Million Veteran Program, calculating total standardized daily doses (TSDD) and anticholinergic burden (AChB). We performed phenome-wide association studies (PheWAS) of ICD-9/10 billing codes and lab values, and Cox Hazards modeling of incident co-morbidities and all-cause mortality over a 14 year period. Joint models incorporated inherited risk factors including schizophrenia and Alzheimer's disease polygenic risk scores (PRS).

Results: Between 2000 and 2025, we identified 12M antipsychotic prescriptions for 177,749 individuals; 382,386 were for injectables (3%) and 5,363,051 (44.6%) were inpatient. PheWAS of medication burden highlighted significant (p < 1e-20) associations across neurological, circulatory, musculoskeletal, digestive, and endocrine domains, and with non-specific symptoms and adverse drug events. Comparing 19,035 SCZ cases with 5-fold matched controls, cases had significantly higher Charlson comorbidity index (CCI; p < 1e-300). Among cases, cumulative treatment days were associated with CCI (OR = 1.06, 95% CI:[1.04,1.07]; p = 1.4e-10) after adjusting for SCZ PRS (OR = 0.95, 95% CI:[0.93,0.96]), as well as diabetes (OR = 1.18, 95% CI:[1.13,1.23]; p = 8.13e-19), pulmonary disease (OR = 1.07, 95% CI:[1.04,1.1]; p = 4.85e-11), and dementia (OR = 1.2, 95% CI:[1.10,1.30]; p = 1.15e-5). AChB predicted sooner time to dementia diagnosis (b = .23, 95%CI: [0.17,0.29], p = 3e-6) after adjusting for measured cognition and PRS.

Conclusions: We demonstrate that chronic antipsychotics exposure among patients diagnosed with SCZ, are associated with a range of physical health conditions, independently of genetic risk for SCZ. Ongoing analyses extend these findings to a broader range of psychotic and affective disorder patients treated with antipsychotic medications.

Keywords: Anticholinergic medication burden, veterans, electronic health record (EHR), Antipsychotic Treatment, Dementia.

Disclosure: Nothing to disclose.

P644. Socio-cultural perspective of mental illness among the shona-speaking people in mutare district, zimbabwe: towards a decolonial approach

Sifikile Songo

Manicaland State University of Applied Sciences, Mutare, Zimbabwe

Background: An appreciation of indigeneity serves to recognize and respect a group’s identity and cultural practices. This study sought to unravel the socio-cultural perspective of mental illness among the Shona-speaking people in Zimbabwe, in an effort to resist cultural hegemony in mental health practice.

Methods: To lay the groundwork for future research in this area, this research was informed by exploratory epistemology. Purposively sampling was done from a population of Lay community health workers and semi-structured interviews were administered to fifteen men and women aged between twenty-five and forty-five. Data was thematically analysed.

Results: The findings revealed that mental illness is understood as a range of mental conditions which affect an individual’s thinking, feelings, behaviour and their overall well-being. Findings further indicated that mental illness is believed to be a Eurocentric phenomenon whose existence is questioned in the Shona culture, mental illness is instead attributed to supernatural forces such as ancestral possession, avenging spirits and witchcraft.

Conclusions: The influence of the United States of America in psychiatry has inhibited the dissemination and articulation of alternative ideas in psychiatric nosology and practice. Beliefs, interpretations, explanations and perceptions in mental health world over have been dominated by the United states. Future research may consider exploring the expression and interpretation of cultural concepts of distress from varied geopolitical regions and cultures.

Keywords: Culture, Hegemonies, Lay Community Health Workers.

Disclosure: Nothing to disclose.

P645. Alterations in endocannabinoid system gene expression in the amygdala following inflammatory and neuropathic pain

Haley Vecchiarelli, Bartosz Lelas, Ana L. Simal, Jaime Tuling, Giannina Descalzi

University of Guelph, Guelph, Canada

Background: Chronic pain is a debilitating condition affecting almost 1 out of every 4 adults in the United States and approximately 1.5 billion adults worldwide (>20% of adults). Despite its widespread prevalence and position as a major global health concern, current treatments for chronic pain do not alleviate it in a majority of patients. Additionally, chronic pain is a strong risk factor for affective conditions, such as anxiety and depression, and is associated with increased risk of death by suicide. Pain is important as it allows individuals to avoid harm and survive through the ability to detect and respond to harmful and potentially harmful stimuli. Activation of peripheral nociceptors by noxious stimuli leads to the transmission of signals via ascending spinal pathways to the brain, including activating limbic and emotion-related brain structures. Experiences, including exposure to noxious stimuli and pain, lead to neuroplasticity, wherein neurons and glia change their function, transcriptomic signature and structure. People experiencing chronic pain show altered brain activity, including in the anterior cingulate cortex (ACC), nucleus accumbens (NAc) and basolateral amygdala (BLA).

We previously demonstrated that increased astrocyte-neuron lactate shuttling (ANLS) in the ACC is a fundamental mechanism for pain-dependent neuroplasticity. However, the cellular signaling cascades that influence lactate production in astrocytes, particularly in pain chronification, are not entirely understood. An intriguing target is the endocannabinoid system (the system cannabis acts on). Mounting reports from people suffering from chronic pain indicate that cannabis may reduce neuropathic pain symptoms, particularly those impacting quality of life and anxiodepressive states. Accordingly, studies using animal models have observed that cannabis vapor reduces pain hypersensitivity in rats experiencing chronic inflammatory pain, and that disruption of cannabinoid type-1 receptors (CB1) blocks this effect. However, how the ANLS and the endocannabinoid system is altered throughout regions involved in pain chronification, particularly the ACC and BLA, is not yet well understood. In particular, the ANLS in the BLA has not been assessed under chronic pain conditions.

Methods: Young adult (8–10 weeks old) female and male C57BL/6CR mice were used for all experiments. Inflammatory pain was modeled using Complete Freund’s Adjuvant (CFA) (10 μL of 50 % CFA in PBS) or vehicle (PBS) injected subcutaneously on the dorsal side of the left hindpaw. Mice were assessed for mechanical nociceptive thresholds using von Frey filaments before the procedure to determine baseline scores and at 24 and 72 hours. Brain regions were collected at 72 hours post injection. Neuropathic pain was modeled using the spared nerve injury (SNI) model (transection of the common peroneal and tibial sciatic nerve branches), and sham treatment (incision but no ligation) was used as control. Mice were assessed for mechanical nociceptive thresholds using von Frey filaments at 3 and 7 days, in addition to baseline scores. Brain regions were collected at 7 days post injury.

Brain regions of interest were excised via dissection/brain punch before freezing on dry ice and stored at −80 °C. RNA was extracted from the BLA using the RNeasy Micro Kit and converted to cDNA using the QuantiTect Reverse Transcription Kit (both Qiagen) according to the manufacturer’s instructions. Quantitative (q)PCR was performed in the BLA using TaqMan Gene Expression Assays for the following targets: Mct4, CB1 (Crn1), CB2 (Crn2), Trpv1, Ppara, Pparg, Grm5, Dagla, Daglb, with Gapdh as a control. Relative gene expression was determined. Data were compared using a two-way ANOVA using sex and pain as factors, with a Tukey’s post-hoc correction where appropriate.

Results: As previously reported, there were reduced mechanical nociceptive thresholds in the injected (but not contralateral) hind-paw following CFA at 24 and 72 hours in both males and females. Preliminary results indicate that MCT4 gene expression was slightly, albeit non-significantly, increased in the BLA with CFA, across sexes.

Conclusions: We here demonstrate that MCT4 is not significantly altered in the BLA following CFA and ongoing work is assessing the endocannabinoid system in the BLA. We will expand this further to assess in other relevant brain regions such as the ACC and NAc. A thorough understanding of the mechanisms promoting chronic pain development and co-morbid anxiodepressive symptoms (e.g., ANLS, the endocannabinoid system and their interactions) may identify novel treatment targets, providing therapeutic breakthroughs.

Keywords: Pain, Amygdala, Endocannabinoids, lactate shuttle.

Disclosure: Nothing to disclose.

P646. Genetic versus environmental risk for serious mental illness (SMI) and youth outcomes in a transdiagnostic high-risk cohort

Jennifer Forsyth, Mahnoor Hyat, Jinhan Zhu, Katherine Scheuer, Carrie Bearden, Carlos Lopez-Jaramillo, Nelson Freimer, Loes Olde Loohuis, Johanna Valencia-Echeverry, Juan David Palacio-Ortiz

University of Washington, Seattle, Washington, United States

Background: Relationships have been identified between clinical and neurobehavioral phenotypes in childhood and genetic and environmental risk for later-onset serious mental illnesses (SMI), such as schizophrenia (SZ), bipolar disorder (BP), and severe-depression (DEP). However, few studies compare associations across environmental risk scores and genetic risk scores for multiple disorders simultaneously, or across multiple classes of genetic risk. Furthermore, most genetic studies have been conducted in European ancestry individuals.

Methods: In a currently-enrolling study of Admixed American youth at elevated transdiagnostic risk for SMI in Colombia (n = 471; mean age = 10.83; 49.3% female), we examined associations between clinical diagnoses and cognitive and motor functioning in childhood and adolescence versus polygenic risk scores (PRS) for SZ, BP, and DEP; copy number variants associated with neurodevelopmental disorders (NDD-CNVs); and a cumulative adverse childhood exposure (ACE) score.

Results: SZ-PRS (R2 = 1.42%) and DEP-PRS (R2 = 1.75%) were associated with poorer cognitive function (FDR p < .05). DEP-PRS (R2 = 2.71%) and ACE score (R2 = 1.60%) were associated with attention deficit hyperactivity disorder (ADHD; FDR p < .05). SCZ-PRS (R2 = 3.17%), DEP-PRS (R2 = 4.22%), NDD-CNVs (R2 = 2.76%), and ACE score (R2 = 5.41%) were associated with conduct disorder. Only ACE score was associated with depressive and anxiety disorders, following multiple testing correction (R2 = 3.17% and R2 = 2.08%, respectively, FDR p < .05).

Conclusions: Results suggest that environmental risk scores and PRS for multiple disorders are associated with clinical and neurobehavioral signs and symptoms during childhood and adolescence, with stronger associations between internalizing disorders and ACE scores than psychiatric PRS at this age. Analyses will be updated as the cohort expands to clarify general vs specific effects of risk scores.

Keywords: adverse childhood experiences (ACE), Cognitive Functioning, polygenic risk scores, major psychiatric disorders.

Disclosure: Nothing to disclose.

P647. Effects of stress- and drug-cue reactivity on craving and cognition in opioid use disorder: preliminary evidence

Benjamin Varnas, Alexis Koslofsky, Shayne Thomas, Greg Perlman, Jodi Weinstein, Richard Rosenthal, Anissa Abi-Dargham, Helen Fox, Scott Moeller

Stony Brook University, Seaford, New York, United States

Background: The opioid epidemic remains a pressing public health crisis. A main driver of compulsive drug use is psychosocial stress, which can trigger drug craving and relapse even after longer-term abstinence. Yet, stress reactivity and its underlying biology in opioid use disorder (OUD) are not well-understood. Here, we used a validated laboratory stress challenge, which used personalized scripts along with guided imagery, to transiently induce stress and (separately) drug craving in individuals with OUD compared with healthy controls (HC). We examined the effects of the personalized stress- and drug script conditions (versus personalized neutral script condition) on self-reported drug craving and cognition (Stroop task). We hypothesized that individuals with OUD would self-report more craving and show more Stroop interference than HC during both the stress and drug imagery conditions.

Methods: We acquired 3-condition, within-person data in 25 OUD participants maintained on opioid medications and 24 HC (data collection is still ongoing). During screening, participants responded to clinically-oriented prompts that were used to write and then audio-record three vividly described vignettes drawn directly from their own experience. We developed one vignette per experimental condition: a recent stressful event (stress condition), a recent tranquil event (neutral condition), and a recent event that culminated in using opioids (in OUD) or alcohol (in HC) (drug condition) (note that 3 HC did not drink alcohol, and instead scripts were produced on coffee or food). During the procedure, each vignette was played to participants through headphones in counterbalanced order. For each condition, participants self-reported the intensity of their craving on a 10-point scale, at the following intervals: 5 minutes before listening to the respective script (−5 min), immediately after listening (0 min), and then 5, 15, 30, and 45 min after listening (+5, +15, +30, and +45, respectively). Participants also performed a classical color-word Stroop task, acquired only at the +5-min interval. After the +45-min interval, participants were given a 30-min break before the next condition, to return to their baseline state. In addition to the craving and Stroop data, participants produced saliva samples for analysis of cortisol and alpha amylase (saliva data not yet available). To safeguard data integrity, (1) stress labs began before 10:30 A.M. (for consistent hormone capture); (2) participants needed to produce negative urine toxicology for illicit drugs on study day (except opioids in OUD); and (3) participants were requested to abstain from phone use, food, and all beverages except water throughout the procedure.

Results: To analyze the Stroop data, we performed a linear mixed model (LMM) with the following predictors: Diagnosis (OUD, HC), Condition (neutral, stress, drug), and their interaction. There were no main effects (both p > 0.42), but the interaction was significant [χ2(2) = 8.99, p = 0.011]. HC had more interference during the stress Condition than the drug Condition, whereas OUD had more interference during the drug Condition than the stress Condition.

To analyze the craving data, we performed generalized estimating equations specifying a negative binomial distribution and log link function (given many zeros in the data). The predictors were Diagnosis (OUD, HC), Condition (neutral, stress, drug), Time (−5, 0, +5, +15, +30, +45), and all their 2-way and 3-way interactions. Results revealed main effects of Diagnosis [χ²(1) = 15.93, p < 0.001] (OUD > HC), Condition [χ²(2) = 7.69, p = 0.021] (both stress and drug > neutral), and Time [χ²(5) = 33.90, p < 0.001] (0, +5, and +15 min > −5 min). Significant 2-way interactions were observed for Diagnosis × Time [χ²(5) = 16.27, p = 0.006] and Condition × Time [χ²(10) = 80.89, p < 0.001]. All of these effects were qualified by the significant 3-way interaction [χ²(10) = 126.26, p < 0.001]. In both OUD and HC, the drug Condition increased drug craving at Times 0 and +5 (all p < 0.01) (returning to normal by +10). Uniquely in OUD, however, the stress Condition also increased craving at Times 0 and +5 (both p < 0.01) (whereas this increase following the stress Condition did not occur in HC).

Conclusions: These initial results provide a rigorous laboratory model for how stress reactivity may ultimately translate into drug craving (and perhaps drug use/relapse). Whereas personalized drug imagery increased drug craving in both OUD and (somewhat surprisingly) HC, stress uniquely increased drug craving in OUD. This finding of stress augmenting craving is consistent with results from a prior study from members of our study team, but extends that study by (1) nearly doubling the number of OUD participants, (2) including HC for comparison, and (3) showing the trajectory of craving over 45 min rather than at a single post-imagery timepoint. We also observed an effect of stress- and drug-cue reactivity on cognition. HC showed more Stroop interference following stress imagery, whereas OUD showed more interference following drug imagery. Increased drug craving and impaired cognition may together create vulnerabilities for drug use in OUD. As data collection is still ongoing and the saliva data remain to be analyzed, these results are preliminary. Nevertheless, if confirmed, this study will enhance basic knowledge about stress processing and underlying biology in OUD, which may lead to new treatment and preventative strategies.

Keywords: Opioid addiction, cue-induced craving, Stress reactivity, Cognition, Cue Reactivity.

Disclosure: Nothing to disclose.

P648. The effect of cognitive profile compared with healthy controls on patterns of frontostriatal brain wiring in early psychosis non-affective and affective subjects: an MRI diffusion imaging tractography study

James Levitt, Mark Vangel, Fan Zhang, Marek Kubicki, Martha Shenton, Lauren O'Donnell, Kathryn Lewandowski

VA Boston Healthcare System, Harvard Medical School, Psychiatry Neuroimaging Laboratory, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, United States, Brockton, Massachusetts, United States

Background: Previously, we examined the impact of clinical subgrouping, based on DSM-V diagnostic criteria, on brain connectivity between the prefrontal cortex and the caudate nucleus (Levitt, et al., 2023, 2025) in early affective psychoses. Here, using the same methodology, but dividing early affective and non-affective psychosis patients based on cognitive, not diagnostic, criteria, we assessed the frontostriatal wiring organization of these subjects compared to HCs. We used diffusion MRI tractography from the Human Connectome Project in 193 subjects for whom we had cognitive measures and divided early psychosis subjects into 3 k-means cognitive clusters (C1 (n = 47), C2 (n = 48), C3 (n = 42)). There were 56 healthy controls (HCs), 97 Early Psychosis non-affective patients and 40 affective patients; Mean age: 24.1. The sex distribution was: 70 male non-affective patients; 27 non-affective female patients; 18 male affective patients and 22 female affective patients. We used our novel method of fiber cluster analysis of whole brain diffusion Magnetic Resonance Imaging (dMRI) tractography to assess brain wiring. This method allows us to quantify the degree of deviation from a topographic, parallel, arrangement in brain wiring connectivity between the frontal cortex (FCtx) and the caudate (Cd), a component of the associative striatum.

Methods: Data used in this study come from the Human Connectome Project for Early Psychosis (HCP-EP) study (MPI: Shenton, Breier). Diffusion MRI Data from 3 HCP sites (University of Indiana, Massachusetts General Hospital and McLean Hospital) were harmonized (Cetin-Karayumak S et al, 2019). From this harmonized data set we generated whole brain tractography using our unscented Kalman filter (UKF) 2-tensor tractography methodology (Malcolm JG et al, 2010). We used a data-driven fiber clustering atlas that allows for a whole brain tractography parcellation into 2000 white fiber clusters according to white matter (WM) fiber geometric trajectory (Zhang et al., 2018). Then, fiber clusters of interest (i.e., from frontal cortex (FCtx) to caudate (Cd)) based on FreeSurfer parcellation from the whole brain white matter (WM) were identified for each subject. We identified 17 WM fiber clusters (F1–17)) connecting FCtx and Cd in both each hemisphere in each subject group. Cognitive subgroups were identified using data-driven K-means clustering in early psychosis subjects. Cognition was assessed using the NIH Toolbox Cognition battery, which includes 5 measures of fluid cognition (card sort, list sort, picture sequencing, flanker, pattern completion) and two measures of crystallized cognition (oral reading, picture vocabulary). For the clustering, the NIH Toolbox subtests were entered into the model. A 3-cluster solution best fit the data. Cluster 1 was characterized by impairment in fluid cognition but above average crystallized cognition, (i.e., mixed). Cluster 2 was characterized by deficits in both fluid and crystallized cognition (i.e., impaired). Cluster 3 was characterized by average performance on most measures (i.e., unimpaired). We assessed the between-group difference for each fiber cluster pair in each fiber cluster in the degree of convergence, reflected by a convergence quotient (CQ). Our CQ was calculated as: (Cortex Distance - Caudate Distance)/ (Cortex Distance + Caudate Distance). For each fiber cluster, we used a mixed model regression analysis of CQ in each hemisphere, separately. For each of 17 fiber clusters in each hemisphere, we assessed the between-group difference in CQ for the 16 pairings with all other fiber clusters in the same hemisphere. We fit a mixed-model regression for each of 17 fiber clusters in a hemisphere, with CQ as response, and subject as a random effect. The Fixed effects are all 2 factor interactions including group, pair and sex, where group is one of the following (cognitive clusters C1 vs HC, or C2 vs HC, or C3 vs HCs). The p value is for the group by pair interaction in this model.

Results: First, in the right hemisphere (RH), we found 3 fiber clusters (F5, F11, F15) emanating from subregions of the rostral middle frontal gyrus (rMFG), and frontal pole (FP) that showed a significant cognitive cluster (C2) by fiber cluster pair interaction (adjusted ps < 0.05). Second, in the left hemisphere (LH), we found 1 fiber cluster (F8) emanating from the inferior frontal gyrus (IFG) that showed a significant cognitive cluster (C2) by fiber pair interaction (adjusted p < .05).

Conclusions: Here, we employed a research domain criteria (RDOC) framework approach to subgroup early psychosis patients with quantifiable cognitive measures which may be closer to the underlying neurobiology of patients and, thus, more likely to yield meaningful biological subgroups. Of note, as expected, after adjusting for sex, we found that it was the early psychosis patients who were in the impaired cognitive cluster (C2) group that showed significant brain wiring deviations from HCs in the pattern of frontostriatal connectivity. Our data further support that long-tract brain wiring measures may serve as trait biomarkers for early psychosis subjects, and that heterogeneity in complex symptoms such as cognition can be leveraged to identify subgroups with common underlying biological characteristics.

Keywords: diffusion tractography, early psychosis, brain wiring, Fronto-striatal circuits, Cognition.

Disclosure: Nothing to disclose.

P649. Frontal and amygdalar KOR availability and BPD traits: neurobiological correlates of impulsivity, emotional dysregulation, and suicide risk

Tamiah Lewis, Dayna Freeman, Emily Weiss, Mika Naganawa, Margaret Davis

Yale University School of Medicine, Waterbury, Connecticut, United States

Background: Borderline Personality Disorder (BPD) is a complex psychiatric condition characterized by pervasive emotional dysregulation, impulsivity, recurrent self-harming behaviors, and risk for suicide. Between 70–80% of individuals diagnosed with BPD engage in non-suicidal self-injury (NSSI), and 10% die by suicide, making BPD one of the highest-risk psychiatric disorders. In addition, some BPD traits can function as barriers to treatment (e.g., interpersonal sensitivity, impulsivity, emotion dysregulation), adding to increased provider burnout and burden with downstream effects on quality and availability of care. Despite this, there are currently no effective FDA-approved pharmacological treatments for BPD symptom severity overall, and few capable of affecting the mentioned traits individually. Promisingly, studies show the kappa opioid receptor (KOR) system plays a role in the modulation of stress, emotion, and impulsive behavior; activation of KORs, especially in the amygdala and frontal cortex, has been associated with dysphoria, anhedonia, and cognitive deficits – symptoms associated with the BPD. However, the relationship between KOR, BPD, and BPD traits that represent treatment barriers have not been directly investigated. This study aimed to examine the relationship between KOR availability in the amygdala and frontal cortex, impulsivity, and emotional dysregulation, and suicide history in individuals with BPD.

Methods: Our sample included forty individuals with BPD, with (SA; n = 20), and without (NSA; n = 20) a history of suicide attempt. All participants reported a history of difficulty securing and maintaining care with medical providers. Participants completed 1 MRI scan and 1 [11C]EKAP PET scan and a battery of psychiatric and cognitive assessments. [11C]EKAP was injected as bolus and subjects were scanned for 120 min. Volume of distribution (VT: the ratio of activity in tissue relative to that in blood, corrected for metabolites) was computed using an arterial input function. Analyses focused on the frontal cortex and amygdala. We examined both within and between group differences in KOR availability and suicide attempt history using ANOVA and correlations with effect sizes quantified using percent difference and Cohen’s d. Emotion dysregulation was assessed with the Difficulties in Emotion Regulation Scale (DERS). Impulsivity was assessed with the UPPS Impulsive Behavior Scale (UPPS), and the Barratt Impulsivity Scale (BIS-11). Interpersonal distress was measured by the Interpersonal Reactivity Index (IRI). BPD symptoms were measured by the Borderline Symptom List (BSL-23). Analyses were conducted in SPSS v. 29.

Results: Group comparison showed lower KOR availability in participants with a history of SA compared to NSA in both the amygdala (F(1,38) = 4.39, p = 0.033, 17.44%, d = −0.70) and frontal cortex (F(1,38) = 8.23, p = 0.007, 18.55%, d = −0.92). However, SA participants did not differ from NSA on intensity of emotion dysregulation, impulsivity, or BPD symptom severity. Across all participants, KOR availability in the amygdala was negatively associated with overall emotion dysregulation (DERS total: r = –0.40, p = 0.01), impulsivity (BIS-11 total: r = –0.41, p = 0.007), and antisocial behavior (UPPS negative urgency: r = –0.44, p = 0.004). Greater BPD symptom severity was also associated with lower KOR availability in the amygdala (BSL score: r = –0.412, p = 0.007) and higher interpersonal distress (IRI score r = –0.506, p = 0.023). In contrast, KOR availability in the frontal cortex was only significantly negatively related to interpersonal distress (IRI score r = –0.47, p = 0.04) and exhibited weaker, non-significant associations with other measures mentioned above (r’s = −0.296 – −0.15, p’s = 0.06 – 0.35).

Conclusions: Antagonism of KOR has been shown to reduce negative affect and improve emotional regulation in both preclinical and clinical studies. For instance, buprenorphine (a partial KOR antagonist) and novel selective antagonists such as PF-04455242 and LY2456302 (Aticaprant) have demonstrated potential in reducing depressive symptoms and enhancing affective stability. The amygdala plays a central role in detecting emotion and generating affective responses—often hyperactive in individuals with BPD, contributing to emotion dysregulation. Findings highlight the importance of amygdalar KOR as a potential neurochemical substrate of emotional dysregulation, impulsivity, and suicide in BPD. By identifying neurobiological correlates of risk and treatment interference in BPD, this research may contribute to future targeted treatment development.

Keywords: kappa opioid receptor, PET Imaging, Borderline Personality Disorder, Suicide attempt, Impulsive behavior.

Disclosure: Nothing to disclose.

P650. Hormonal adaptations and clinical outcomes in MDMA-assisted therapy for PTSD: a phase 2 veteran trial

Lauren Lepow, Thomas Hildebrandt, Amy Lehrner, Rachel Yehuda

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

Background: MDMA-assisted therapy has demonstrated promise for Post-Traumatic Stress Disorder (PTSD), offering an opportunity to examine the biological predictor and correlates of durable recovery. The current study investigated long-term clinical outcomes and neuroendocrine markers in veterans in relation to moral injury burden and sustained remission status.

Methods: Military veterans with chronic PTSD (N = 22) completed two or three MDMA-assisted therapy sessions at the James J. Peters VA Medical Center. Clinical variables such as moral injury as assessed by the 9-item Moral Injury Events Scale (MIES) and blood samples for brain-derived neurotrophic factor (BDNF), cortisol, catecholamines, and neuropeptide Y (NPY) were collected at baseline and follow-up. PTSD severity was assessed with the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5) at baseline and 12-month long-term follow-up. Outcomes included categorical response, remission, and sustained remission (CAPS < 12 and no PTSD diagnosis at 12 months). Analyses explored clinical change across moral injury groups and hormonal adaptations linked to remission.

Results: Participants demonstrated robust and durable clinical improvement. Mean baseline CAPS severity was 40, with a mean reduction of 22.8 points at 12-month follow-up. By categorical outcome, 21/22 (95%) achieved response, 16/22 (73%) met stringent responder criteria, 18/22 (82%) lost PTSD diagnosis, and 11/22 (50%) achieved sustained remission at 12 months.

Moral injury did not prevent clinical benefit. Veterans with high moral injury (MIES > 16/27) (n = 7) achieved a mean CAPS reduction of 43.6%; while those with lower moral injury (n = 15) improved by 64.1%. The greater reduction in the lower moral injury group was not statistically significant (p = 0.21).

The most robust biological signal was observed when outcomes were defined by sustained remission (assessed a year following study completion).

Both remission and non-remission groups improved clinically, but only those in sustained remission (n = 11) exhibited significant hormonal adaptations. Remission was associated with change in cortisol (β = –0.394, p < 0.01), catecholamines (β = 0.265, p < 0.05), and NPY at the trend level (β = –0.202, p < 0.10). Non-remitters (n = 11) showed no comparable hormonal changes. Across the full cohort, percent change in cortisol correlated positively with percent change in CAPS severity (r = 0.24, ns), suggesting a link between HPA-axis recalibration and clinical recovery. BDNF did not significantly differentiate groups.

Conclusions: In this Phase 2 veteran trial, MDMA-assisted therapy for PTSD produced substantial and durable reductions in PTSD symptoms. Clinical gains were observed in both high and low moral injury groups, though there was a non-significant trend toward greater reductions in those with lower moral injury. Sustained remission was uniquely marked by hormonal adaptations in cortisol, catecholamines, and NPY, pointing to possible biological mechanisms of enduring recovery. These preliminary findings in this sample underscore the robust clinical potential of MDMA-assisted therapy and highlight candidate biomarkers of sustained response that warrant testing in larger trials.

Keywords: Psychedelics, PTSD, veterans, Cortisol.

Disclosure: Nothing to disclose.

P651. Stratifying MDMA response using negative affect circuit biotypes with clinical validation

Xue Zhang, Laura Hack, Claire Bertrand, Rachel Hilton, Nancy Gray, Leyla Boyar, Jessica Laudie, Boris Heifets, Trisha Suppes, Peter van Roessel, Carolyn I. Rodriguez, Karl Deisseroth, Brian Knutson, Leanne M. Williams

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

Background: MDMA-assisted therapy shows promise for treating posttraumatic stress disorder (PTSD), yet clinical response varies substantially across individuals. Identifying who is most likely to benefit is essential for advancing precision psychiatry. This work applies a circuit-based stratification framework focused on the negative affect circuit to identify biotypes predictive of MDMA response. Findings from a mechanistic trial are further prospectively validated in an ongoing clinical trial.

Methods: In a mechanistic trial, N = 16 participants of both sexes with subthreshold PTSD symptoms completed one baseline and three double-blinded drug sessions (placebo, 80 mg, 120 mg MDMA) in randomized order. Task and resting-state fMRI were acquired 90 minutes post-administration, along with behavioral readouts from a computerized test battery. Task-based and resting-state activity and connectivity with a focus on the amygdala, subgenual anterior cingulate cortex (sgACC), and thalamus within the negative affect circuit were examined for all four sessions. Participants were stratified into heightened (n = 8) or intact (n = 8) negative affect subgroups based on baseline amygdala activity in response to nonconscious threat stimuli. Baseline neural, behavioral, and demographic subgroup differences were tested using t-tests or χ² tests. Subgroup differences in MDMA-induced effects were analyzed using linear mixed-effects models with subgroup, session, and their interaction as fixed effects. Random intercepts were included. Covariates, including age and sex were added if associated with outcomes. All p-values were FDR-corrected (q < 0.05).

In a clinical trial (N = 30, enrolling; both sexes), participants with PTSD are randomized to MDMA-assisted cognitive processing therapy or therapy alone. The same imaging and behavioral measures are collected pre-, during, and post-therapy.

Results: In the mechanistic trial, the heightened subgroup showed greater amygdala and mediodorsal thalamus (MDT) activity to nonconscious threat at baseline, along with elevated MDT resting-state activity and exaggerated behavioral responses. Following MDMA, this subgroup showed reduced amygdala activity and increased sgACC–amygdala connectivity during threat processing, decreased MDT and enhanced sgACC–amygdala connectivity and global functional connectivity during resting-state, compared to the intact group. All p-values < 0.05, corrected. The ongoing clinical trial applies this stratification prospectively to validate these findings case by case.

Conclusions: Stratifying individuals based on negative affect circuit function identifies distinct neurobehavioral profiles predictive of MDMA response. This approach links brain mechanisms to clinical application and offers a promising path toward personalized MDMA-assisted therapy for PTSD.

Keywords: MDMA, Precision Psychiatry, Functional MRI (fMRI), Neurocircuitry, Stratification.

Disclosure: Nothing to disclose.

P652. Patient narratives in psilocybin therapy for mood dysfunction in Parkinson’s disease: a computational analysis of language

Sachin Pendse, Ellen Bradley, Balázs Szigeti, Gisele Fernandes, Kim Sakai, Sophie Bax, Joshua Woolley

University of California San Francisco, San Francisco, California, United States

Background: Psilocybin clinical trials thus far have involved a substantial psychotherapy component, though there has been little systematic analysis of these sessions. Natural language processing (NLP) offers an accessible, scalable way to examine how patients’ expression over the course of a study intervention, potentially shedding light on mechanisms of clinical improvement.

Methods: We conducted an exploratory analysis of language from psychotherapy sessions in a non-randomized pilot trial (NCT04932434) evaluating the feasibility of psilocybin therapy for mood dysfunction in Parkinson’s disease (PD). Data from 10 people (5 women) who completed treatment involving two administrations of psilocybin (a 10 mg safety dose followed by a 25 mg treatment dose) paired with eight psychotherapy sessions (two before drug exposure, two between the 10 mg and 25 mg doses, and four after the 25 mg dose) were used. Sessions were audio recorded, transcribed via the Whisper Application Programming Interface, and diarized. We examined empirically-validated measures of emotional and cognitive experience using 15 metrics in the Linguistic Inquiry and Word Count (LIWC) tool linked to treatment experiences in previous research on mood states in PD; these included measures of language related to cognitive processing, health, affect, self, and community affiliation. We also assessed metrics associated with PD in prior research, including: usage of action verbs, identified via counting all non-auxiliary verbs using the spaCy part of speech tagger; semantic variability measured via average semantic distance between words within a given turn using FastText word embeddings; and concreteness of a turn measured via Brysbaert et al.’s validated dictionary of word concreteness scores. We tested for differences across (1) preparatory sessions versus all sessions post-psilocybin exposure, (2) preparatory versus post-treatment dose (25 mg) sessions, and (3) post-safety dose (10 mg) versus post-treatment dose sessions. For all comparisons, we used Welch’s t-test and adjusted p-values for the False Discovery Rate via the statsmodels Python package. To understand themes discussed during sessions, we used BERTopic topic modeling with speaker turns as our unit of analysis. We converted turns to embeddings via the ModernBERT Embed Base language model and created a model across all transcript text, using each turn as a single document. To finetune clustering, we reduced the embedding space via a Uniform Manifold Approximation and Projection dimensionality reduction approach and then applied Hierarchical Density-Based Spatial Clustering of Applications with Noise with a minimum cluster size of 15. This achieved a silhouette score of 0.53, Cv coherence score of 0.38, and a topic diversity score of 0.70, indicating that about 70% of the top-10 words across topics were unique. Our model identified 131 distinct topics across 14,505 participant speaker turns, with 45.46% of documents being outliers. The research team manually reviewed topics to identify those most relevant given study context.

Results: From preparatory sessions to post-psilocybin exposure sessions, we found increases in language related to the self (dz = 2.6, padj < 0.001), past experiences (dz = 2.3; padj < 0.001), and cognitive processing (dz = 1.3, padj = 0.006). There were concurrent decreases in language related to the future (dz = 1.9, padj = 0.001), religion (dz = 1.8, padj = 0.001), and group affiliation (dz = 1.37, padj = 0.006). Between preparatory sessions and post-25 mg sessions, these same features shifted significantly in the same directions, though there was a greater decrease in language related to the future (dz = 2.1, padj = 0.001). We did not see significant differences between sessions after the 10 mg safety dose versus after the 25 mg treatment dose, or in the use of action verbs, semantic variability, or concreteness (padj > 0.05 for all features). Our topic model showed that across sessions, patients tended to discuss people in their lives and their social connections (7.7% of all speaker turns), including discussion of how their experience of PD is perceived and understood by their family members, the impact of PD (4.0%), as well as how physical activities and art/creativity (1.8%) shape their daily lives.

Conclusions: Following exposure to psilocybin, patients with PD engaged in more cognitive, self-referential verbal processing, shifted focus from the future to the past, and made fewer references to religion and connection with others. These changes were apparent relatively early in the protocol, after an initial test dose of 10 mg. NLP may be a useful tool for future randomized controlled trials to examine how linguistic changes relate to clinical outcomes.

Keywords: Parkinson's disease, Psilocybin, Language.

Disclosure: Nothing to disclose.

P653. The safety and preliminary effects of psilocybin for phantom limb pain: a feasibility pilot study

Adam Halberstadt, Jon Dean, Ethan Hurwitz, Timothy Furnish, Joel Castellanos, Cassandra Vieten, Albert Lin, Mark Geyer, Fadel Zeidan

University of California, San Diego, La Jolla, California, United States

Background: Phantom limb pain (PLP) is a refractory condition defined by pain in a missing limb after amputation. Evidence is accumulating that classical psychedelic drugs such as lysergic acid diethylamide (LSD) and psilocybin can alleviate treatment-resistant disorders. Notably, case reports indicate that LSD and psilocybin can produce persistent reductions of PLP. However, controlled data about the safety of administering psychedelic drugs to amputees is lacking, and randomized, controlled trials are ultimately required to determine whether psychedelics can relieve PLP and other types of chronic pain.

Methods: We conducted a placebo-controlled, double-blind, feasibility pilot study (NCT05224336) to test the preliminary safety and efficacy of psilocybin for PLP. Nine participants (mean±SD 37 ± 13 years) with one amputation and PLP were randomized to receive a single 25-mg oral dose of psilocybin (n = 5; 2 female) or a 100-mg oral dose of niacin (n = 4; 2 female). Two monitors met with the participants on three occasions pre-dosing to prepare them for drug administration and again one-day post-dosing for an integration session. Participants provided visual analog scale ratings for past-week PLP and residual limb pain (RLP; i.e., pain at the amputation-site) intensity at baseline and two- and four-weeks post-dosing. Suicidality was assessed at all visits. Blood pressure (BP), heart rate (HR), and adverse drug effects were assessed during the dosing session.

Results: While the group × time interaction for the primary outcome of PLP was not statistically significant (p > 0.05), psilocybin was nonetheless associated with clinically-meaningful reductions of PLP two weeks (−65% [Cohen’s d = 1.34]) and four weeks (−32% [Cohen’s d = 0.56]) post-dosing compared to baseline (niacin group: −3% at two weeks [Cohen’s d = 0.30]; −11% at four weeks [Cohen’s d = 0.42]). There was, however, a significant group × time interaction [χ2(1) = 9.31, p = 0.002, np2 = 0.41; 95% CI 0.06 to 0.66] for RLP. Psilocybin was associated with significant RLP decreases two-weeks (−78% [p < 0.001, Cohen’s d = 3.21, 95% CI 1.00 to 5.41]) and four-weeks (−62% [p < 0.001, Cohen’s d = 1.73, 95% CI 0.02 to 3.44]) post-dosing compared to baseline. Conversely, the niacin group exhibited non-significant increases of RLP at those time points (two weeks: +18% [Cohen’s d = −0.22]; four weeks: +50% [Cohen’s d = −0.52]). No significant increases in suicidality, serious adverse events, group differences for adverse drug effects, or group × time interactions for BP or HR (p > 0.05) were noted during the study.

Conclusions: This is the first placebo-controlled study to show that a 25 mg dose of psilocybin is safe and potentially efficacious for chronic pain. Addition clinical trials are required to confirm the therapeutic efficacy of psilocybin for PLP and RLP and to understand the time-course of its effects and how best to employ psychedelics as treatments for chronic pain.

Keywords: Psychedelic therapy, psilocybin, phantom pain.

Disclosure: Compass Pathways, Contracted Research, Self, atai, Contracted Research, Self.

P654. Pharmacokinetics and pharmacodynamics of oral 4-bromo-2,5-dimethoxyphenethylamine (2C-B) administration in healthy participants

Friederike Holze, Denis Arikci, Jan Thomann, Dino Luethi, Matthias E. Liechti

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

Background: Psychedelics have recently gained interest as a treatment for psychiatry and neurological disorders. 4-Bromo-2,5-dimethoxyphenethylamine (2C-B) is widely used recreationally and similarly to LSD and psilocybin its primary effects are thought to be mediated by serotonin 2A (5-HT2A) receptors. Scientific studies describing its dose-effect relationship and its pharmacokinetic properties are largely lacking. The present study therefore aimed to describe the pharmacokinetics of 2C-B and its main metabolites, BDMPAA and B-2-HMPAA, as well as the pharmacodynamics and dose-response relationships of three different oral doses of 2C-B.

Methods: We used a double-blind placebo-controlled cross-over design in 24 healthy participants (12 male, 12 female) investigating three different oral doses of 2C-B. Conditions included 1. 10 mg of 2C-B, 2. 20 mg of 2C-B, 3. 30 mg of 2C-B and 4. placebo. We assessed pharmacokinetic parameters and acute subjective effects using visual analog scales (VAS). Plasma concentrations were determined using a validated LC-MS/MS method. Pharmacokinetic parameters and effect durations were calculated using non-compartmental methods 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. The study was registered at clinicaltrials.gov (NCT05523401).

Results: Mean (95% confidence interval) maximal 2C-B concentrations were 2.5 (2.2–2.8) ng/mL, 4.6 (4.1–5.1) ng/mL, and 6.4 (5.8–7.0) ng/mL after administration of 10, 20, and 30 mg of 2C-B, respectively. Maximal concentrations were reached on average after 2.0 h, with a mean elimination half-life of 1.3 h.

Mean (95% confidence interval) maximal BDMPAA and B-2-HMPAA concentrations were 610 (554–672) and 28 (25–31) ng/mL, 1072 (995–1156) and 55 (49–63) ng/mL, and 1294 (1168–1433) and 86 (76–97) ng/mL after administration of 10, 20, and 30 mg of 2C-B, respectively.

Maximal “any drug effects” (mean ± SEM) were 27 ± 4 (p < 0.01), 56 ± 5, and 72 ± 5 (both p < 0.001) for the 10, 20, and 30 mg doses compared with placebo, respectively. Maximal “good drug effects” and “bad drug effects” (mean ± SEM) were 29 ± 5 (p < 0.05) and 2.0 ± 0.6 (p = 0.84), 60 ± 6 (p < 0.001) and 7.4 ± 2.4 (p < 0.05), and 68 ± 5 (p < 0.001) and 9.8 ± 2.5 (p < 0.001) for the 10, 20, and 30 mg doses compared with placebo, respectively.

Drug effects showed a dose-dependent onset (mean ± SD) of 0.82 ± 0.46 h, 0.60 ± 0.29 h, and 0.50 ± 0.28 h for the 10, 20, and 30 mg doses, respectively, and lasted for 3.1 ± 1.5 h, 4.6 ± 1.4 h, and 4.9 ± 1.5 h, respectively.

Conclusions: Here, we show that both the pharmacokinetics and pharmacodynamics of 2C-B increase with increasing doses. Plasma concentrations of 2C-B and its metabolites increased with higher doses, however dose-proportional increase in maximal concentrations was only observed for B-2-HMPAA. Subjective drug effects (“any,” “good,” and “bad” drug effects) also increased with dose with a relatively rapid onset of effects (<1 h) lasting between 3–5 h depending on the administered dose. “Bad drug effects” were generally mild at low doses but became more pronounced at higher doses, similar to classical psychedelics such as LSD and psilocybin.

Keywords: Psychedelics, Pharmacokinetics, Pharmacodynamics.

Disclosure: Nothing to disclose.

P655. Psilocybin modulates plasticity and metaplasticity in human visual cortex 24 H post-administration

Nathan Heller, Gabi Lofland, Frederick Barrett

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

Background: Synaptic plasticity mechanisms govern stable transformations in neural response patterns. Metaplasticity is a persistent change in the rules governing those mechanisms. Preclinical studies show serotonergic psychedelics (SPs) modulate both plasticity and metaplasticity, but these effects have not yet been demonstrated in humans. Because SPs target visual processes, we use a sensory plasticity assay measuring visual evoked potential (VEP) changes after repetitive visual stimulation. The ‘visual tetanus’ assay we employ has revealed plasticity-related alteration in major depression and enhancements by SSRIs (Normann et al., 2007). Related protocols have been investigated extensively in rodent models, where they have been shown to reflect LTP-like mechanisms (Clapp et al., 2006) and recruit specific subpopulations of inhibitory interneurons to potentiate stimulus-selective visual encoding (Montgomery et al., 2022). This cross-species compatibility positions this approach well for testing SP-induced modulation of plasticity and metaplasticity in humans, where direct evidence is lacking.

Methods: Healthy volunteers were recruited for a within-subjects, counterbalanced crossover, double-blind, placebo-controlled study (N = 18; F = 7). 24 h after dosing with placebo or oral psilocybin (10 mg), we administered a 10-min, 2 Hz checkerboard stimulation protocal that induces LTP-like VEP modulation. VEPs were compared at baseline (pre-checkerboard) and 13-min post-checkerboard stimulation at a posterior ROI (electrodes: PO7, PO3, POz, PO4, PO8, O1, Oz, O2, and Iz). A manipulation check (cluster-based permutation testing, two-sided sign-flip) was applied to post-placebo data (averaged-ROI) to determine whether VEP component (C1/P1/N1) modulation was consistent with prior implementations of this protocol (Normann et al., 2007). Linear mixed-effects (LME) was applied to pre-specified window-defined VEP-peaks, in order to test for drug effects on VEP amplitude (Amp) as a function of Condition (placebo vs psilocybin) and Time (preCheckerboard vs postCheckerboard). Model: Amp ~ Condition*Time + (1 + Time | Subj) + (1 | Electrode).

Results: Cluster testing (placebo-only, ROI-averaged; two-sided sign-flip) confirmed C1 attenuation (82–96 ms; mean Δ = 1.03 µV; 95% CI [0.51, 1.54]; p_cluster = 0.002; dz = 0.99) and P1 potentiation (~127–141 ms; mean Δ = +0.96 µV; 95% CI [0.44, 1.48]; p_cluster = 0.002; dz = 0.92), but no N1 potentiation was detected. In crossed-effects LME models on pre-specified window-defined VEP-peaks, only P1 amplitude revealed effects of psilocybin administration. Two drug effects were observed. First, preCheckerboard (i.e., baseline) P1 amplitude was higher following psilocybin administration compared to placebo administration (+0.51 µV, 95% CI [0.18, 0.84], p = 0.003; d = 0.24). Second, P1 amplitude was subject to a highly significant Condition-by-Time interaction (ΔΔ = −1.06 µV, 95% CI [−1.53, −0.59], p ≈ 1 × 10⁻⁵; d = 0.35): following checkerboard stimulation, placebo administration led to canonical P1 potentiation, whereas psilocybin administration led to P1 attenuation. The interaction was spatially extensive: all 9/9 electrodes remained significant after BH–FDR (q = 0.05); per-electrode ΔΔ ranged −1.32 to −0.77 µV (mean ≈ −1.06 µV), matching the model-based estimate.

Conclusions: Here we observe two dissociable effects of psilocybin on human visual-cortical plasticity 24 h post-dose: elevated baseline P1 amplitude and a reversal of the checkerboard-induced P1 potentiation. The first psilocybin-induced effect suggests that visual cortex underwent stable synaptic changes to become more responsive to perturbation: a plasticity effect. The second suggests a persistent change in the rules governing LTP-like visual plasticity: a metaplastic effect. Together, these data provide the first direct human evidence that psilocybin shifts the visual-cortical set point and alters subsequent LTP-like plasticity 24 h post-dose.

Keywords: Psilocybin, 5-HT2A receptor, Visual plasticity, Metaplasticity.

Disclosure: Nothing to disclose.

P656. Post-acute effects of lysergic acid diethylamide on motor learning in healthy volunteers

Abigail Calder, Vincent Diehl, Morten Lietz, Parsa Yousefi, Kristian Beichmann, Antonin Rouaud, Gregor Hasler

University of Fribourg, Villars-sur-Glâne, Switzerland

Background: Preclinical studies demonstrate that lysergic acid diethylamide (LSD) can enhance cortical synaptic plasticity beginning within hours and lasting for at least several days. This suggests that LSD could positively impact learning ability. Improvements in motor learning, in particular, would have important clinical applications in neurorehabilitation and stroke recovery. Motor learning involves processes of “online” learning during active practice and “offline” learning during waking rest. Additionally, LSD may improve perceived stress and cognitive flexibility.

Methods: Forty-three healthy subjects (24 women) received 100 µg LSD and a placebo in randomized, counterbalanced order (ClinicalTrials.gov: NCT05177419). All participants completed a baseline visit, dosing days, and follow-up visits one and seven days after dosing. Participants rated subjective positive and negative drug effects during dosing days. One day after dosing, online and offline motor learning were assessed with a sequence typing task. Participants typed a 9-digit sequence with the left hand as quickly and accurately as possible. In the online learning phase, they practiced for 30 minutes. In the offline phase, they completed two test blocks after resting for 10 and 80 minutes. Additionally, at baseline and one week after dosing, participants completed the Cognitive Flexibility Inventory (CFI) and Perceived Stress Scale (PSS).

Results: LSD was well-tolerated and produced mostly positive subjective effects. One day after LSD, offline motor learning significantly improved compared to placebo. LSD did not affect accuracy, online learning, or initial speed in the task. Additionally, one week after dosing, stress was reduced and specific aspects of cognitive flexibility were increased.

Conclusions: Improvements in motor learning, stress and cognitive flexibility support the hypothesis that LSD enhances neuroplasticity. If replicated, these findings may have important implications for the use of psychedelics in neurorehabilitation.

Keywords: Psychedelic effects, motor learning, Neurological and Psychiatric Disorders, Perceived stress.

Disclosure: Nothing to disclose.

P657. MM120 (Lysergide d-tartrate; LSD) phase 1 and phase 2 studies: a detailed safety analysis

Sarah Karas, Paula Jacobsen, Jamileh Jemison, Nithya Srinivas, Daniel Karlin

Mind Medicine, Bend, Oregon, United States

Background: MM120 (lysergide d-tartrate; LSD) is a compound under development by Mind Medicine, Inc. for the treatment of generalized anxiety disorder (GAD) and major depressive disorder (MDD). Unlike the current standard of care for GAD and MDD, which requires daily medication dosing with resultant persistent adverse event (AE) burden, the intermittent single-dose paradigm for MM120 is being clinically evaluated as a potential treatment option with improved rapidity of onset, magnitude of efficacy, prolonged durability, and a favorable tolerability profile.

Methods: These safety data were derived from three studies that included 171 adult participants: two Phase 1 open-label, randomized, 2-period, 2-sequence, crossover studies evaluating MM120 100 µg in 27 and 64 participants, and a multicenter, randomized, double-blind, placebo-controlled Phase 2b study including 40 participants in each cohort of 100 and 200 µg. Treatment-emergent adverse events (TEAEs), vital signs, and suicidality (Columbia–Suicide Severity Rating Scale) were monitored and reported. TEAEs were coded using MedDRA (v26.0) and categorized as occurring either within 24 hours of dosing or more than 24 hours after dosing. For each AE, severity and relatedness to treatment were determined.

Results: In the Phase 1 studies, >75% of participants experienced TEAEs. Most TEAEs occurred on dosing day and were consistent with the known acute pharmacodynamic profile of MM120. The most frequently reported TEAEs across the Phase 1 studies were mydriasis, visual hallucinations, euphoric mood, and illusions at similar rates across the cohorts.

In the Phase 2b study, TEAEs were reported in 97.5% of participants with MM120 100 µg and 92.5% with 200 µg. TEAEs occurring on the dosing day were mild to moderate, and aligned with expected psychoactive and physiological effects, including hallucinations, illusions, euphoric mood, anxiety, mydriasis, nausea, not clinically significant increases in blood pressure. Post-treatment TEAEs occurred in approximately 56% of participants in both groups; the most common were headache, fatigue, low mood, and insomnia. These were comparable to placebo at 38.5%.

Related serious adverse events (SAEs), hallucinogen persisting perception disorder, or deaths were not observed. No participants in any of the studies exhibited active suicidal ideation or behavior during their sessions or post-treatment. Vital signs changes did not reach clinical significance or require intervention.

Conclusions: In the three studies that included 171 eligible participants, a single administration of MM120 100 or 200 µg was characterized by transient sensory-perceptual and physiological effects, largely confined to the dosing day. The majority of TEAEs on dosing day were rated mild to moderate. Across studies, post-treatment TEAEs were uncommon and mild. No suicidality signal was observed. No cumulative or study-wide safety trends emerged. These findings demonstrate a consistent safety profile across studies, supporting further clinical evaluation of MM120.

Keywords: Psychedelics, Safety profile, CNS Clinical Trials, LSD, generalized anxiety disorder.

Disclosure: MindMed, Employee, Self.

P658. Two patients with stroke treated with psilocybin

Gregor Hasler, Vincent Diehl, Josua Zimmermann, Meret Branscheidt

University of Fribourg, Villars-sur-Glâne, Switzerland

Background: Psychedelics such as psilocybin act primarily at serotonin 5-HT2A receptors and have been shown to enhance neuroplasticity, learning, and emotional processing.1 Recent preclinical and clinical data suggest that psychedelics may facilitate functional recovery after neurological injury.2 In a recent study, Calder and Hasler showed that LSD has the potential to enhance motor learning in humnas (see 2nd ACNP 2026 Poster by Gregor Hasler). However, evidence for their use in stroke rehabilitation remains scarce. We report two patients with ischemic stroke who received psilocybin under medical supervision to support neurorehabilitation.

Methods: Both patients were treated in an inpatient neurorehabilitation setting at Cereneo, Hertenstein, Switzerland. Psilocybin was administered under Swiss federal exemption permits with psychological support before, during, and after dosing. Standardized assessments included the Montgomery-Åsberg Depression Rating Scale (MADRS) and language/cognitive testing where applicable.

Results: Case 1: A 72-year-old woman with ischemic stroke of the right middle and anterior cerebral arteries and comorbid recurrent depression underwent five psilocybin sessions (15–20 mg). Sessions were well tolerated without serious adverse events. Transient improvements were observed after each session in mood, motivation, and functional engagement in neurorehabilitation therapies, accompanied by subjective experiences of emotional relief and openness. Functional recovery of motor skills, as estimated by the time the patient was able to stand, progressed beyond the expected trajectory after the third session. Apart from mild confusion on the day after the fourth session, the patient and her relatives reported no side effects.

Case 2: A 44-year old male patient with aphasia following ischemic stroke received two psilocybin sessions (15 mg and 25 mg). The first session elicited only mild psychedelic effects (subjective peak intensity 5/10). Minor side effects included transient dizziness and a three-day depressive episode that resolved spontaneously. Post-session testing showed improvements in reading speed (10→13 words) and object naming (8→10 correct). The second session (25 mg) produced somewhat stronger subjective effects (peak 8/10) with mild somatic sensations and emotional changes, but again limited visual or perceptual alterations. Naming tests revealed mild gains across frequency categories. By follow-up, MADRS scores had normalized (0 points). Overall, incremental improvements in expressive language ability were documented across sessions. The patient did not report any adverse effects of the treatment.

Conclusions: These two case reports suggest that psilocybin can be safely administered in stroke patients under controlled medical conditions, with potential benefits for neurorehabilitation and emotional well-being. The antidepressant and neuroplasticity-promoting effects of psilocybin may extend to patients with neurological injury, offering a novel therapeutic adjunct to conventional stroke rehabilitation. Controlled trials are warranted to systematically evaluate safety, efficacy, and optimal dosing strategies in this population.

Keywords: Stroke, Psychedelics, motor learning, Neuroplasticity, speech production.

Disclosure: Nothing to disclose.

P659. RE104: a novel serotonergic psychedelic 4-OH-DiPT prodrug for the treatment of postpartum depression

Anita Clayton, M. Camille Hoffman, Peter Hendricks, Daniel Fallon, Matthew Johnson, Beatrix Taylor, Jasna Hocevar-Trnka, Nathan Bryson, Joe Hirman, Robert Alexander, Stephen Ross, Mark Pollack

Reunion Neuroscience, Sacramento, California, United States

Background: RE104, a unique, proprietary 4-OH-DiPT prodrug, is a novel psychedelic 5HT2A agonist investigational compound being developed for the treatment of postpartum depression (PPD) and other mental health conditions. Preclinical and clinical characterization confirmed similar pharmacology of RE104 to other well-characterized classical psychedelics such as psilocybin, while demonstrating a significantly shorter and reproducible acute psychedelic experience in the clinic.

PPD is a serious disorder occurring in approximately 15% of pregnancies, and is associated with significant distress, dysfunction and disability in affected women as well as adverse effects on infants.

Herein we present new results from a recently completed Phase 2, double-blind active-controlled trial of RE104 for the treatment of PPD.

Methods: Participants (N = 84) were randomized to receive a single, subcutaneous dose of 30 mg or 1.5 mg (active control) RE104 for injection at 28 sites across the United States and were subsequently followed for 4 weeks post-treatment. While session monitors were involved in the preparation, dosing session monitoring and follow-up, the study treatment did not include formal psychotherapy.

Eligible participants were women aged 18–45, ≤15 months post-partum, not breastfeeding, with moderate to severe depression based on an entry HAMD-17 score of ≥ 24 at screening and baseline. In addition, if applicable, eligible participants could remain on a pre-existing stable regimen of SSRIs or therapy during the study.

The primary endpoint was change from baseline at Day 7 in the Montgomery-Åsberg Depression Rating Scale (MADRS) total score. Additional endpoints included changes in the MADRS over the course of the study, rates of response and remission, and changes in maternal function including care and relationship with the infant as assessed by the Barkin Index of Maternal Function (BIMF). Analysis of the primary outcome was via MMRM including treatment, visit and visit-by treatment interaction, and baseline MADRS; an unstructured covariance structure. Analysis of secondary endpoints was performed using descriptive statistics.

Results: A single 30 mg dose of RE104 resulted in a statistically and clinically significant reduction from baseline of 23.0 points in the MADRS total score on Day 7, as compared to a reduction of 17.2 points in the active control arm (difference of 5.80; one-sided p = 0.0094). Clinically meaningful reductions in the MADRS were also observed for RE104 30 mg on the first day following administration and maintained through the Day 28 follow-up. 77.1% of participants treated with RE104 30 mg were responders (with at least 50% improvement from baseline MADRS) to therapy at Day 7 compared to 61.6% of participants on active control. 71.4% of participants on RE104 30 mg achieved remission (with a MADRS score of ≤10) at Day 7 compared to 41.0% on active control. Both response and remission improvements were maintained through the Day 28 follow-up period. Maternal well-being and function, including care of the infant as assessed by the BIMF, also demonstrated substantial improvement.

RE104 was observed to be well-tolerated, during the trial with the majority of adverse events being mild to moderate in intensity, resolving spontaneously and consistent with the pharmacology observed with other agents in the class. Consistent with its pharmacokinetics, over 90% of participants on RE104 30mg showed no signs or symptoms posing a risk for discharge at the first measure of discharge readiness, taken 4 hours following treatment.

Preliminary results from a contemporaneous lactation study suggest that the total amount of metabolites observed in the breastmilk represents less than 0.1% of the 30 mg RE104 administered to the mother representing a safety margin that is an order of magnitude below levels that might be considered of potential concern relative to the infant. Pending review by the FDA, these results suggest that breastfeeding following RE104 treatment may be able to resume with limited interruption.

Conclusions: In this first sizable randomized controlled trial of a psychedelic agent for the treatment of Postpartum Depression, administration of 30 mg of RE104 resulted in rapid, significant and durable improvement, and was safe and well-tolerated.

Keywords: Psychedelic, Postpartum depression, Antidepressant.

Disclosure: Reunion Neuroscience, Employee, Self, Argus Cognitive, Founder, Self.

P660. Long-term efficacy of psilocybin in treatment-resistant major depression (EPIsoDE): 6- and 12-month observational follow-up of a phase 2b trial

Lea Julia Mertens, Felix Betzler, Manuela Brand, Ricarda Evens, Laura Kärtner, Andrea Jungaberle, Henrik Jungaberle, Tomislav Majić, Christian N. Schmitz, Andreas Ströhle, Dennis Scharf, Moritz Spangemacher, Max Wolff, Zahra Assadi, Scharif Bahri, Lilith Becher, Luca V. Färber, Henry Harder, Eugenia Kulakova, Linda Kunz, Andy Meijer, Barbara Rohrmoser, Moritz M. Berger, Michael Koslowski, Gerhard Gründer

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

Background: Psilocybin has shown promising short-term efficacy for depressive disorders, including treatment-resistant depression (TRD). Long-term follow-up data, however, are scarce and limited by small sample sizes. The aim of this study was to investigate the antidepressant effect of one or two doses of psilocybin with adjunct psychotherapy in TRD until twelve months.

Methods: This is an observational 12-months follow-up of a phase 2b, randomized, active placebo-controlled trial in male and female patients with TRD of moderate to severe degree. Patients were recruited predominantly from out-patient settings at two university hospitals in Germany. Eligibility criteria included 25–65 years of age, no somatic or psychiatric contraindication and withdrawal from antidepressant medications. Participants were randomly assigned to receive two doses six weeks apart, embedded in seven psychotherapeutic sessions: (i) placebo (100 mg nicotinamide) followed by 25 mg psilocybin, (ii) 5 mg followed by 25 mg psilocybin, or (iii) 25 mg psilocybin followed by either 5 mg or 25 mg psilocybin. The primary endpoint was treatment response (≥50% reduction on the Hamilton Rating Scale for Depression [HAMD17]) six weeks after the first dose, which has been reported elsewhere. The end-of-study of the main trial was 12 weeks after the first dose, after which participants were free to use other forms of treatments. Follow-up assessments were scheduled after six- and twelve-months. All participants who completed at least one follow-up assessment were included in the follow-up analysis set. Primary efficacy endpoint was the Hamilton Rating Scale for Depression (HAMD17), evaluated in terms of change from baseline, response, remission (HAMD17 score < 8) as well as sustained response and remission since the end-of-study. Change from baseline (primary follow-up endpoint) was evaluated using a generalized additive mixed regression model, including treatment, time, treatment x time interactions, center and center x time interactions as covariates (fixed effects) and participant-specific intercepts (random effects). Time was treated on a continuous scale (days since first dose) with the effects modeled by smooth functions based on penalized B-splines. Response and sustained response were analyzed by logistic regression models comparing treatment groups, including center and baseline HAMD17 as covariates; remission and sustained remission were each analyzed descriptively. The trial is registered at ClinicalTrials.gov, NCT04670081.

Results: Between June 10, 2021 and December 5, 2023, 144 patients (59 females [41%], 85 males [59%]) were enrolled and randomly assigned across treatment groups. Of these, 126 participants (51 females [40%], 75 males [60%]) completed at least one follow-up visit, with similar drop-out rates across groups. Mean baseline HAMD17 scores ranged from 21.6–22.2. A generalized additive mixed model for change in HAMD17 scores showed a significant time effect across all groups for both follow-up timepoints, with estimated average changes from baseline of −7.93 (95% CI: −9.17, −6.70, adj. p < 0.0001) at six months and −7.74 (95% CI: −9.04, −6.43, adj. p < 0.0001) at twelve months. No significant differences between treatment groups were apparent at either timepoint. Results were consistent when controlling for antidepressant pharmacotherapy and psychedelic use during the follow-up period. They indicated a strong positive association of antidepressant pharmacotherapy with HAMD17 scores (β = 3.79, 95% CI: 1.98 to 5.60), whereas no association for psychedelic use (β = −0.27, 95% CI: −2.23 to 1.69) was found.

Response rates did not differ significantly across groups: They ranged from 29.5% to 50.0% at six-months (p = 0.65) and from 34.1% to 50.0% at twelve-months (p = 0.71). Sustained response rates (since end-of-study) were also similar across groups and ranged from 20.0% to 25.0% at six-months (p = 0.94) and from 17.5% to 25.0% at twelve-months (p = 0.45).

Conclusions: This is by far the largest systematic follow-up of patients treated with a psychedelic drug for any psychiatric disorder. The findings demonstrate a stable and clinically meaningful antidepressant effect of one or two doses of psilocybin with adjunct psychotherapy lasting up to twelve months in TRD. These results provide strong evidence for the treatment’s long-term efficacy. While this study did not find an added benefit of a repeated 25 mg psilocybin dose, future studies should investigate different treatment regimens, optimally in randomized-controlled study designs.

Funding: Funded by the German Federal Ministry of Education and Research (BMBF, grant number: 01EN2006 A/B).

Keywords: Psilocybin therapy, Treatment-resistant depression, Clinical trial, long-term results.

Disclosure: Nothing to disclose.

P661. Stronger desire for and perceived helpfulness of psychedelic assisted therapy in adults vs. adolescents in preliminary study of caregivers of adolescents formerly in intensive psychiatric treatment

Alex Rothbaum, Jessica L Maples-Keller, Rory O'Brien, Michelle Tobar, Raneeka DeMoss, Noah Marchuck, Molly Mercer-Deadman, Ben Hunter, Alicia K Smith, Seyma Katrinli

Skyland Trail, Atlanta, Georgia, United States

Background: Psychiatric illness in adolescence has been on the rise, especially since the COVID-19 pandemic, with more patients needing treatment than there is capacity in the existing care delivery system. Adolescents who are treatment refractory and/or significantly impaired may end up in intensive psychiatric services, such as residential or partial-hospitalization treatment. In such cases, caregivers have had to put forth significant effort and incur large expenses to find appropriate treatment. While data in adolescents is extremely limited, recent studies have laid the translational building blocks to support the potential efficacy of psychedelic assisted therapy (PAT) for treatment-refractory adults in depression and posttraumatic stress disorder. Extant literature has shown the treatment choice and preference is an important, yet oft-ignored predictor of treatment response and cost-effectiveness. Given psychedelics such as MDMA and psilocybin are currently Schedule 1 drugs, thus are illegal for use outside of clinical research, it is vital to understand how caregivers may view a typically illegal substance within their child’s mental health treatment. Community-focused participatory research is proven method to derive hypothesis-generating insights with input from stakeholders to develop and implement new interventions. As such, the current project aimed to engage caregivers of adolescents discharged from intensive psychiatric services to assess attitudes and beliefs surrounding PAT, especially related to potential adolescent treatment with the goal of tailoring future research and implementation.

Methods: Participants (N = 52) were caregivers of adolescents (13–18 years old) who had previously received intensive psychiatric services at a non-emergent non-profit treatment center who consented to participate in survey research post-discharge. Caregivers were recruited via email from a list of addresses exported from an electronic medical record; demographic information was not collected to reduce potentially identifiable information; links were unique limiting one response to specified addresses. Attitude toward psychedelics was assessed using methods adapted from the psychiatric treatment preference literature. Participants were presented with a brief definition of psychedelics and psychedelic-assisted therapy (PAT) followed by questions assessing familiarity, desire for availability of PAT for adults and adolescents in the United States (US), willingness to seek out PAT for self and child if it were available, perceived helpfulness of PAT, and perceived benefit of PAT on a 10-item Likert scale where 1 = “Not at all” and 10 = “Extremely”.

Results: Caregivers had a stronger desired for PAT availability in the US for adults (mean = 6.56 ± 2.75, mode = 5) than adolescents (mean = 5.29 ± 2.74, mode = 5; t(51) = 4.12, p < .001, d = .57). There were no significant differences in caregiver willingness to seek PAT if it were available for themselves (mean = 4.63 ± 3.12, mode = 2) or their dependent adolescent (mean = 4.79 ± 3.22, mode = 1; t(51) = −.54, p = .59). However, caregiver perceived helpfulness of PAT was stronger for adults (mean = 7.06 ± 2.40, mode = 7) than for adolescents (mean = 5.88 ± 2.75, mode = 5; t(49) = 4.57, p < .001, d = .65). Mean familiarity with PAT prior to the survey was 5.4 ± 3.23 with modal answer of 1. Mean extent of belief that psychedelics would improve psychotherapy gains was 6.12 ± 2.51, mode of 5.

Conclusions: Overall, there is a lack of familiarity and hesitancy for PAT in caregivers of adolescents who have undergone intensive psychiatric services. While caregivers have significantly stronger desire and perceived helpfulness of PAT for adults than adolescents, the Likert ratings were not overwhelming strong. Relatedly, the belief the psychedelics would improve psychotherapy outcomes was relatively neutral, reflected in the similar rankings of willingness to seek PAT if it were available. We believe these are the first quantitative data to assess caregiver perceptions of PAT for dependent adolescents. Importantly, these adolescents have needed residential and/or partial-hospitalization treatment suggesting that they have a history of treatment resistance and/or significant functional impairment, as such may be the likely population focused on for these treatments in the future. As such, these caregivers may be more open to novel treatments given lack of quick success with existing evidence-based treatment. It may be that caregivers with treatment-naïve dependent adolescents may be less open than the current sample, with a desire to try established treatment options first, however, this is a testable hypothesis in future research. As PAT methods are developed, it will be important to educate the public so that they are familiar and understand the data in terms of potential benefit as an adjunctive to psychotherapy, should the data in adolescents indicate such. Further, recruiting adolescents for such studies may be difficult without significant caregiver education and engagement.

Keywords: Psychedelics, Adolescent psychiatry, Emerging Therapies.

Disclosure: Nothing to disclose.

P662. An open-label study of single-dose psilocybin for borderline personality disorder with co-occurring major depressive disorder

Jon Grant, Dustin Ehsan, Margaret O’Brien, Laurie Avila, Megha Neelapu, Sophia Boutouis

University of Chicago, Chicago, Illinois, United States

Background: Borderline personality disorder (BPD) is characterized by mood instability, cognitive symptoms, impulsive behavior, and disturbed relationships. A variety of psychotherapies have been developed, and while research on the use of medication is ongoing, no drug has been approved in the United States or elsewhere for its treatment. Second generation antipsychotics have been the most extensively studied, but current treatments for BPD are often inadequate.

Methods: We conducted an open-label pilot study of a single dose of psilocybin (25mg) for borderline personality disorder with co-occurring major depressive disorder. Baseline assessments included demographics, psychiatric and medical evaluations, vital signs, a physical examination, an electrocardiogram, a urine drug screen, liver function testing, and pregnancy testing for women of childbearing potential. Each participant completed the BPD Symptom Assessment Scale (BPD-SAS), a 12-item, self-report scale concerning BPD symptoms during the past 2 weeks. In addition, each participant was evaluated with the Montgomery-Asberg Depression Rating Scale (MADRS), the Columbia-Suicide Severity Rating Scale (C-SSRS), and the Clinical Global Impression – Severity (CGI-S) scale at Baseline and at all subsequent visits. Two trained medical professionals sat with the participants and provided supportive techniques during the 8-hour dosing day. Follow-up assessments were conducted at 1-, 2-, and 4-weeks post-dosing. The primary outcome measures were changes in BPD-SAS and MADRS scores from dosing to 4 weeks post-treatment. We used a paired-samples t-test to examine changes in BPD and depressive symptoms from Baseline to 4 weeks post-dosing.

Results: 9 participants (4 male, 5 female; mean age = 31.33 [SD = 13.45] years; 7 Caucasian, 2 multiracial) with BPD and major depressive disorder were enrolled. All 9 completed each study visit.

BPD scores did not significantly change from the Baseline visit to the Week 4 follow-up visit: Baseline (M = 32.67, SD = 4.18) to the final visit (M = 28.11, SD = 8.98); t(8) = 1.70, p = .129. The effect size for this change was moderate (Cohen’s d = .565, 95% CI [−.158, 1.26]; Hedges’ g = .510, 95% CI [−.142, 1.14]).

Depression symptoms significantly changed from the Baseline visit to the Week 4 follow-up visit: Baseline (M = 28.56, SD = 4.53) to the final visit (M = 17.22, SD = 10.39); t(8) = 4.217, p = .003. The effect size for this change was strong (Cohen’s d = 1.41, 95% CI [.445, 2.33]; Hedges’ g = 1.27, 95% CI [.402, 2.10]).

Conclusions: This open-label pilot study of 9 participants showed that a single dose of psilocybin resulted in improvement in depressive symptoms for a period of 4 weeks. There was no statistically significant change in BPD symptoms during the trial. While statistical significance was not reached regarding BPD symptoms, likely due to the small sample size, the observed effect size and directional trend suggest potential clinical benefit and support the need for further study in a larger, controlled trial. BPD symptoms did not prevent subjects from experiencing improvement in their depressive symptoms. Importantly, no adverse outcomes were reported during the study period, suggesting the intervention was well-tolerated.

Keywords: Borderline Personality Disorder, Psilocybin, Major Depressive Disorder (MDD).

Disclosure: Biohaven, Contracted Research, Self, Janssen, Contracted Research, Self.

P663. Psilocybin enhances familiarity-based encoding: evidence from memory errors and electroencephalography

Manoj Doss, Nathan Heller, Gabriella Lofland, Audrey Duarte, Joshua Koen, Frederick Barrett

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

Background: The acute effects of psychedelics (hallucinogenic 5-HT2A agonists such as psilocybin) are rarely found to enhance cognitive processes, but we have recently found an idiosyncratic enhancement in episodic memory. Although psychedelics impair the encoding (i.e., formation) of hippocampal-dependent recollection, they can enhance the encoding of cortical-dependent familiarity (e.g., recognizing a face but not recalling where or when this individual was met). Recollection is typically accurate and accompanied by high-confidence memory decisions, as it involves remembering specific details such as where or when an event took place. In contrast, familiarity is accompanied by more variable memory confidence and can be readily misattributed resulting in memory errors, especially when recollection fails. Indeed, in one study we found that participants falsely remembered a picture of labels due to a misattribution of heightened familiarity produced by encoding these labels under the effects of psychedelics. These effects were largest for emotional compared to neutral stimuli. Here, we aimed to replicate these effects of psilocybin and examine their neural mechanisms with electroencephalography (EEG). We focused on the FN400, a neural correlate of familiarity in which reduced negative deflections at 300–500 ms in frontal-central electrodes during memory retrieval are associated with greater familiarity.

Methods: In an ongoing double-blind, placebo-controlled, repeated measures study (N = 22), we tested the effects of 10 mg of psilocybin during the encoding phase of an emotional episodic memory task with memory tested 24 hours later under drug-free conditions. During the encoding phase, participants viewed labels describing emotionally negative, neutral, and positive scenes, half of which were presented with their corresponding picture. The following day EEG was collected and memory for these pictures was tested. Participants were presented with the labels from the encoding phase and had to decide whether each label had been presented alongside its corresponding picture.

Results: Compared to placebo, psilocybin during encoding impaired hit rates (F(1, 21) = 7.88, p = .010, η_P^2 = .27) and memory accuracy (F(1, 21) = 16.04, p < .001, η_P^2 = .43) made with high-confidence responses, consistent with an impairment in recollection. Psilocybin during encoding also tended to increase false alarm rates (F(1, 21) = 3.40, p = .079, η_P^2 = .14) but not high-confidence false alarm rates (F(1, 21) = 1.67, p = .21), especially for emotional stimuli (drug × emotion interaction: F(1, 21) = 3.25, p = .049, η_P^2 = .13), consistent with a misattribution of heightened familiarity. Further evidence for enhanced encoding of familiarity under psilocybin came from smaller FN400s across frontal-central electrodes to labels that had been previously seen under psilocybin compared to those seen under placebo.

Conclusions: In the fourth study, we provide evidence that psychedelics can enhance the encoding of cortical-dependent familiarity despite impairing hippocampal-dependent recollection. This drug effect is not found with GABAA sedatives (e.g., alcohol), NMDA dissociatives (e.g., ketamine), or Δ9-tetrahydrocannabinol, which impair the encoding of both recollection and familiarity, and therefore, such findings may shed light on unique phenomena and therapeutic mechanisms of psychedelics. For example, under drug-free conditions, when recollection fails and familiarity is high, cognitive illusions can occur such as déjà vu and presque vu (illusory insight), phenomena also experienced under psychedelics. Finally, enhanced cortical-dependent memory encoding may provide an opportunity to fast-track cortical learning that could disrupt maladaptive, rigid memories that have become instantiated over time in the cortex.

Keywords: episodic memory, Psychedelics, Psilocybin, EEG, emotional memory.

Disclosure: VCENNA, Advisory Board, Self.

P664. Sustained 5-HT2AR activation engendered by select serotonergic psychedelics in vitro

Joshua Zamora, Kathryn Cunningham

Center for Addiction Sciences and Therapeutics, John Sealy School of Medicine, University of Texas Medical Branch, Galveston, Texas, United States

Background: The ‘psychedelic renaissance’ is fueled by the exciting clinical potential of hallucinogenic serotonin 5-HT2A receptor (5-HT2AR) agonists for the treatment of neuropsychiatric diseases. Through actions at centrally expressed 5-HT2ARs, psychedelics evoke fervent hallucinations and catalyze neuroplastic changes with enduring therapeutic effects. Additionally, psychedelic stimulation of the 5-HT2AR can alter the functional status of the receptor. Specifically, psychedelics desensitize and downregulate the 5-HT2AR leading to subsequent reduction in 5-HT2AR-directed signaling. Classically, receptor internalization is a key driver of desensitization and downregulation of G-protein coupled receptors (GPCRs) such as the 5-HT2AR. Internalization serves as a mechanism to halt heterotrimeric G-protein-dependent signaling and adjust agonist-responsiveness of a given receptor pool. However, this classical model of receptor regulation is challenged by growing reports of GPCR-governed signaling that persistently continues within intracellular compartments following receptor internalization. While induction of neuroplasticity is postulated to be the therapeutic mechanism of action of psychedelics, how psychedelic challenge spatiotemporally reorganizes 5-HT2AR signaling to bring about ameliorative outcomes is less understood. We utilized human embryonic kidney 293 cells (HEK293) transiently expressing the human 5-HT2AR (h5-HT2AR) as a cellular model to further interrogate how treatment with hallucinogenic 5-HT2AR agonists perturbs 5-HT2AR function.

Methods: Accumulation of inositol monophosphate (IP1) was measured with the Cisbio IP-One Homogeneous Time Resolved Fluorescence (IP1-HTRF) assay to index 5-HT2AR activity in vitro following exposure to 5-HT, (-)-2,5-dimethoxy-4-iodoamphetamine [(-)-DOI], (+)-lysergic acid diethylamide [(+)-LSD], dimethyltryptamine (DMT) or lisuride (LIS). HEK293 cells transiently expressing the h5-HT2AR were pretreated with various concentrations of 5-HT, (-)-DOI, (+)-LSD, DMT or LIS for 60 min followed by extensive drug washout to discern sustained, residual 5-HT2AR signaling. Residual agonist responses were compared to functional responses produced by an acute 60 min agonist treatment without drug washout for each respective ligand. To examine alterations in 5-HT2AR functionality following extended agonist pretreatment, h5-HT2AR-expressing HEK293 cells were treated with vehicle (0.1% DMSO) or (-)-DOI (10 µM) for 24 hours followed by removal of ligand from the extracellular media prior to addition of 5-HT. Cell permeable 5-HT2AR antagonists (M100,907, pimavanserin, ketanserin) or a cell-impermeable 5-HT2AR antagonist (methylated-ketanserin) were employed to assess the contributions of intracellular or plasma membrane pools of 5-HT2AR to residual IP1 accumulation manifested post (-)-DOI washout. Each experiment was performed in technical triplicate at minimum with four to eight biological replicates. Where appropriate, data were statistically analyzed using an unpaired Student’s t-test with an experimental error rate of α = 0.05.

Results: (-)-DOI pretreatment for 60 min sustained 5-HT2AR signaling following washout with significantly reduced potency (p < 0.05) relative to an acute (-)-DOI treatment. (+)-LSD pretreatment for 60 min sustained 5-HT2AR signaling following washout with significantly lower maximal efficacy (p < 0.05) versus an acute (+)-LSD treatment. For the non-hallucinogenic LIS, 60 min pretreatment sustained 5-HT2AR signaling following washout with significantly greater potency (p < 0.05) relative to an acute LIS response. In contrast, neither 5-HT nor DMT pretreatments elicited persistent 5-HT2AR signaling following ligand washout. Surprisingly, 24-hour (-)-DOI pretreatment significantly augmented the basal IP1 response (p < 0.05) despite drug washout. The addition of escalating concentrations of 5-HT did not further enhance production of IP1. The persistent elevation in basal IP1 accumulation following 24-hour (-)-DOI pretreatment was completely reversed by M100,907, pimavanserin or ketanserin. Strikingly, the cell impermeant 5-HT2AR antagonist methylated-ketanserin only partially reversed this sustained, wash-resistant IP1 response induced by 24-hour (-)-DOI pretreatment.

Conclusions: Our data demonstrate that the 5-HT2AR is persistently activated in vitro despite the elimination of excess agonist from the extracellular space. Sustained 5-HT2AR activation is a ligand-dependent phenomenon that can be differentially coaxed by hallucinogenic [(-)-DOI, (+-LSD)] or non-hallucinogenic (LIS) 5-HT2AR agonists. Residual 5-HT2AR activation present following prolonged (-)-DOI exposure is sensitive to complete reversal only by permeant antagonists suggesting that persistent 5-HT2AR signaling following extended (-)-DOI treatment is driven, in part, by an intracellular pool of 5-HT2AR. Enduring 5-HT2AR signaling produced by select 5-HT2AR agonists may contribute to their neuroplastic or psychopharmacological actions. Ongoing studies will glean the mechanisms that dictate wash-resistant 5-HT2AR signaling and evaluate if persistent 5-HT2AR signaling from intracellular compartments initiates or maintains neuroplasticity in vivo.

Keywords: Psychedelics, 5-HT2A receptor, GPCR, Molecular Pharmacology, Intracellular Signaling.

Disclosure: Nothing to disclose.

P665. Modeling MDMA’s therapeutic effects on plasticity and fear memories in mice

Nitzan Geva, Sarah Jefferson, Anne George, Axel Rosado, John Krystal, Christopher Pittenger, Alfred Kaye

Yale University, New Haven, Connecticut, United States

Background: MDMA-assisted psychotherapy recently showed efficacy in a Phase 3 clinical trial for PTSD, and enhances fear extinction in rodents. The precise mechanisms of MDMA role in altering emotional learning and expression are thus a vital question. Here, we utilized cutting edge optical dissection of neuronal activity and structure using in vivo microscopy to identify circuit and subcellular mechanisms of MDMA effects on plasticity in medial prefrontal cortex (mPFC) in the context of fear extinction learning.

Methods: Two-photon imaging of dendrites in Thy1-GFP mice was over the course of 37 days to assess frontal cortex spine density changes following MDMA (vs saline control). Miniscope calcium imaging of mPFC neurons was conducted for 3 days over the course of fear learning and extinction (day 1–30s tone with 1mA 1s shock Context A, day 2 - MDMA 7.8 mg/kg or saline 30 min prior to 6 tones in a Context B, day 3–6 tones in context B).

Results: MDMA induced dendritic spine density increases on days 1 to 7 relative to saline control (p < 0.001); effects were shared across frontal cortex regions (Cg1 vs PL/IL). Administration of MDMA enhanced fear cue response at recall, while also increasing trial-to-trial changes in activity (>1000 neurons per condition). Fear expression was attenuated in a timing specific manner when MDMA was administered with fear cues, demonstrating precise memory valence effects of MDMA fear extinction.

Conclusions: MDMA has precise timing effects on fear memories. Spine plasticity occurs 1 day after MDMA administration, and this time point also corresponds to changes in the neuronal and population-level representations of fear. Links between spine plasticity and network dynamics have not previously been identified, and may further explain MDMA and entactogens could be effective in PTSD.

Keywords: Psychedelics, Brain plasticity, PTSD, systems neuroscience, calcium imaging.

Disclosure: Freedom Biosciences, Contracted Research, Self.

P666. Durability of psychedelic-induced neuroplasticity is associated with distinct changes to the dendritic spine proteome

Sarah Jefferson, Shveta Bathla, Patrick Wehrle, Hanieh Falahati, Pietro Camilli, Angus Nairn, Alfred Kaye

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

Background: Structural neuroplasticity has been proposed as a key mechanism underlying the long-lasting therapeutic effects of psychedelics. A single dose of psilocybin confers antidepressant effects. lasting for weeks to months in humans, and rodent studies have demonstrated that serotonergic psychedelics psilocybin and 5-MeO-DMT enhance structural plasticity in the frontal cortex for >1 month following a single dose (Shao et al 2021, Jefferson et al 2023). In contrast, we have shown that MDMA produces a robust, but shorter-lasting increase in dendritic spine density, indicating that the durability of psychedelic-induced neuroplasticity may differ among drugs with different direct mechanisms of action (Jefferson et al., unpublished). Despite great interest in this area, there continue to be fundamental questions about how newly-formed spines may integrate into circuits to produce therapeutic effects. This study sought to determine whether specific changes to the dendritic spine proteome are associated with the durability of drug-induced structural neuroplasticity.

Methods: We used a novel proximity labeling-based approach with Shank3-BioID2 to label synaptic cleft proteins at early (within 1 week) and extended (over 1 month) timepoints following treatment with psilocybin (1 mg/kg), 5-MeO-DMT (20 mg/kg) and MDMA (7.8 mg/kg) relative to saline vehicle. Drug doses were selected based on prior studies using longitudinal two-photon imaging to track dendritic spine density over the course of one month in the mouse medial frontal cortex. A sample size of n = 6 mice per condition was chosen for 80% power based on a pilot study of Shank3-BioID2 proximity labeling proteomics with 5-MeO-DMT vs saline, in which an effect size of 0.85 was observed. All drug and time conditions utilized male and female mice and sex was considered as a covariate in subsequent analyses. The pAAV-Shank3-BioID2 construct was injected bilaterally into the lateral ventricles of P0-3 neonatal mice. At 6 weeks of age, mice were injected with drug. Biotin was delivered by IP injection for 7 days on either days 0–6 or 30–36 relative to drug injection. The mPFC was isolated and extracts were subjected to affinity purification of biotinylated proteins and LC/MS/MS with label free quantification. Negative controls consisted of mice that did not receive biotin. Only proteins that were upregulated by at least 1.5-fold in the experimental samples vs negative controls were considered for analysis. Proteins that were significantly up- or down-regulated (p < 0.05) for a given drug and timepoint were subjected to pathway analysis with Ingenuity Pathway Analysis.

Results: In the first week following drug injection, we identified differentially expressed proteins (DEPs) for each drug: 71 for psilocybin, 62 for 5-MeO-DMT, and 55 for MDMA. At one month following drug injection, there were still high numbers of differentially expressed proteins for all three drugs: 43 for psilocybin, 84 for 5-MeO-DMT and 92 for MDMA. Relatively few DEPs were shared between drugs (2 for the early timepoint and 5 for the extended timepoint), but on a pathway level, the early changes to the synaptic proteome were more similar for serotonergic psychedelics with long-lasting neuroplastic effects (psilocybin and 5-MeO-DMT) compared to the entactogen MDMA. Many of the shared processes between psilocybin and 5-MeO-DMT were involved in local translation and ROBO receptor signaling, while processes distinct to MDMA included various mitochondrial and metabolic pathways. Interestingly, although MDMA did not significantly alter spine density at 30 days post-drug in our prior unpublished data, it produced sustained changes to the synaptic cleft proteome at this timepoint, with processes including cytoskeletal signaling, GTPase signaling and electron transport.

Conclusions: This study provides evidence that serotonergic psychedelics psilocybin and 5-MeO-DMT, which been shown to induce long-lasting neuroplastic effects in the mouse medial frontal cortex, produce distinct changes to the dendritic spine proteome in the week following drug delivery when compared with MDMA, a drug with shorter-lived neuroplastic effects in this region. Additionally, these data provide mechanistic insight into the early and sustained synaptic effects of psilocybin, 5-MeO-DMT and MDMA, which may be important in understanding their potential therapeutic impacts.

Keywords: Psychedelics, Psilocybin, 5-MeO-DMT, MDMA, Proteomics.

Disclosure: Nothing to disclose.

P667. Sex-specific behavioral and cellular effects of psilocin on context-induced fentanyl seeking in mice

Tiffini Lovell, Ana-Clara Bobadilla

Colorado State University, Fort Collins, Colorado, United States

Background: Psychedelic compounds, including psilocin, are emerging as promising therapeutics for substance use disorders, yet their mechanisms and sex-specific effects remain underexplored. A key feature of substance use disorders is context-induced relapse, where environmental cues previously associated with drug use trigger drug-seeking behavior. Using a Conditioned Place Preference paradigm, we investigated the impact of psilocin on fentanyl-seeking behavior in male and female C57BL/6J mice, with a focus on sex as a biological variable.

Methods: Mice were conditioned over eight days to associate distinct environmental contexts with fentanyl or saline administration. Following conditioning, animals received varying doses of psilocin and were assessed for drug-seeking behavior, quantified as time spent in the fentanyl-paired context. To evaluate whether changes in anxiety-like behavior mediated psilocin’s effects, we conducted Elevated Plus Maze testing.

Results: Psilocin significantly reduced fentanyl seeking in female mice, with no observable effect in males. Psilocin did not alter anxiety-related measures in either sex.

Conclusions: These findings support psilocin as a potential sex-specific therapeutic for opioid use disorder, particularly in females. Ongoing studies focus on identifying the underlying neural mechanisms using 3D brain mapping in Ai14xcFos-TRAP2 transgenic mice to compare neuronal activation associated with psilocin and fentanyl across brain regions and sexes. Preliminary data suggest sex-specific patterns of neural engagement following psilocin treatment.

Keywords: Psilocin, Fentanyl, Neuronal ensembles.

Disclosure: Nothing to disclose.

P668. Acute psilocybin affects spontaneous and context-dependent behaviors in a dose and sex-dependent manner in sprague-dawley rats

Praachi Tiwari, Bryan Jenkins, Catherine Moore, Hudson Zhou, Tylaah George, David Yaden, Elise Weerts

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

Background: Psychedelics are being increasingly investigated for their therapeutic potential in neuropsychiatric disorders. However, their dose-dependent effects remain poorly characterized, which can be parcellated using preclinical models. In rodents, the head twitch response is currently the primary behavioral proxy for “hallucinogenic-like” activity, but it does not directly capture the breadth of psychedelic-induced inherent behavioral changes. Further, it is suggested that psilocybin induced range of behaviors are interdependent in their expression. However, this hypothesis has not been systematically investigated. In the present study, we address this gap by examining spontaneous and context dependent behaviors in Sprague-Dawley rats across multiple psilocybin doses. This approach not only enables identification of systemic behavioral effects beyond the conventionally scored head twitch response but also provides a framework to explore correlations between different behaviors.

Methods: We categorized spontaneous behaviors as all behaviors displayed by the rats (Charles River, Wilmington, MA) in their home cage environment, upon psilocybin administration. These include head twitch response (HTR), as well as rearing, immobility, grooming, sniffing or digging, and circling behavior. Context dependent behaviors putatively measured compulsive and anxiety-like behavior using Marble Burying Test (MBT) ad Open Field Test (OFT) respectively. To identify the effects of different doses of psilocybin on spontaneous and context dependent behaviors, naïve adult rats (13–15 per group) were administered psilocybin at doses of 0, 0.5, 1, or 1.5 mg/kg to determine the acute (0–30 minutes) and post-acute (> 4h) behavioral response. All behaviors were recorded using either a side-view or a top-view camera. Spontaneous behaviors were manually scored by two scorers blind to experimental conditions. Total number of marbles buried on the MBT was scored manually and live by two experimenters. OFT data was analyzed using Ethovision XT 17.0. Data were analyzed on GraphPad Prism (version 10.6.0) using ordinary two-way analysis of variance (ANOVA), considering dose and sex as independent variables, followed by Bonferroni's post-hoc test when the ANOVA indicated a significance. Data were tested for normality of distribution prior to all analyses used to measure statistical significance. For data that did not follow a normal distribution, Dunn’s post-hoc analysis was applied if interaction indicated a significance. Correlation between variables was measured using Pearson correlation coefficient. Data are expressed as mean ± standard error of the mean (S.E.M) and statistical significance was set at p ≤ 0.05.

Results: Our findings revealed a distinct, dose-dependent acute spontaneous and context-specific behavioral response to psilocybin administration in male and female rats. We noted a bell-shaped curve for HTR in response to increasing psilocybin dose, with peak HTRs seen at 1 mg/kg of psilocybin. However, while we did not note an interaction on the two-way ANOVA, there was a main effect of dose. We further noted a significant interaction for immobility (p = 0.002) and circling behavior (p < 0.0001), which increased with increasing psilocybin doses. In contrast, rearing, sniffing and grooming behavior decreased with increasing psilocybin doses, with a main effect of both dose and sex for sniffing response, and a main effect of dose for rearing and grooming behavior. A significant two-way ANOVA interaction was observed for rearing (p = 0.039) as well as sniffing bouts (p < 0.0001). Our correlational analyses for spontaneous behaviors revealed different directionality of relationships between individual behaviors at specific doses. However, there was no significant correlation between behaviors across any dose of psilocybin tested. Furthermore, for context-specific behaviors, psilocybin administration produced a dose-dependent acute and but not post-acute change in compulsive behaviors, as seen with MBT, with minimum burying response noted at 1 mg/kg psilocybin for male rats, and at 1.5 mg/kg psilocybin for female rats, and a significant interaction on two-way ANOVA (p = 0.007) accompanied by a main effect of dose. We also do not note any significant interaction on anxiety-like response on the OFT post-acutely. Ongoing studies aim to elucidate the underlying neurobiological mechanisms through cellular and molecular analyses of neuronal activity patterns and neuroplasticity. We hypothesize that there may be increased cellular plasticity with reduction in PNNs in the target regions involved in compulsive and anxiety-like behaviors, like the basal ganglia, ventral hippocampus and medial prefrontal cortex.

Conclusions: Together, by analyzing behavioral responses across a range of psilocybin doses, our findings help identify patterns that may predict individual behavioral responses and determine optimal doses for studying specific behaviors. Our correlational analyses suggests that spontaneous behaviors may occur independent of each other upon systemic psilocybin administration, pointing to the possibility of distinct neural circuit and/or receptor engagement. Overall, this approach advances our understanding of the dose range relevant for potential therapeutic effects of psilocybin in animal models of neuropsychiatric disorders.

Keywords: Psychedelics, Compulsive behavior, Anxiety-related behaviors, Internal state modulation.

Disclosure: Nothing to disclose.

P669. A preclinical study characterizing the fast-acting and long-lasting antidepressant-like effects of psilocybin in adolescence

Rubén Garcia-Cabrerizo, Itziar Beruete-Fresnillo, Pedro Bergas-Cladera, M. Julia Garcia-Fuster

University of the Balearic Islands, Palma de Mallorca, Spain

Background: Adolescent depression is a significant public health concern, yet treatment options remain limited, particularly due to age- and sex-related differences in antidepressant efficacy. This study explored for the first time the potential antidepressant-like response of psilocybin in adolescence by examining acute, repeated and persistent effects in Sprague-Dawley rats of both sexes without distinction, as measured under the stress of the forced-swim test. As compared to other studies, we relied on a more translational approach by administering psilocybin orally (oral gavage, o.g.), while elucidating its hallucinogenic-like potential through head-twitch responses. Finally, hippocampal neurogenesis markers were evaluated as potential biomarkers of psilocybin’s antidepressant-like responses in adolescence (1- and 16-days post-treatment).

Methods: To do so, we used a total of 118 Sprague-Dawley adolescent rats (58 males and 60 females). Groups of allocated rats were exposed to different testing procedures following adolescent psilocybin (0.3 and 1 mg/kg) treatments: head-twitch responses and locomotor activity; antidepressant-like scoring under the stress of the forced-swim test; hippocampal neurogenesis markers by immunohistochemistry studies. Locomotor activity and head-twitch responses were analyzed through two-way ANOVAs, with treatment (saline vs. psilocybin) and route of administration (i.p. vs. o.g.) as independent variables. Antidepressant-like responses and the regulation of neurogenic biomarkers were analyzed through one-way ANOVAs, followed by Dunnett’s multiple comparisons test when appropriate.

Results: The main results of our study showed that: (1) Acute psilocybin (1 mg/kg, 30 min) induced subjective hallucinogenic effects, as measured by head-twitch responses (F1,19 = 29.96, p < 0.001), independently of the route of administration (i.p. vs. oral gavage, o.g.; F1,19 = 1.58, p = 0.225), and without changing locomotor activity (F1,19 = 0.66, p = 0.425). (2) Acute psilocybin (0.3 and 1 mg/kg, o.g., 30 min) exerted a rapid antidepressant-like response (F2,57 = 18.81, p < 0.001) that paralleled the hallucinogenic effect. Specifically, both doses of psilocybin significantly decreased the time rats spent immobile when compared to saline-treated rats (0.3 mg/kg: −42 s, ***p < 0.001; 1 mg/kg: −57 s, ***p < 0.001). These decreases in immobility were paired with increases in active escape behaviors, such as climbing (F2,57 = 6.69, p = 0.003) and swimming (F2,57 = 17.18, p < 0.001). (3) Repeated psilocybin (0.3 and 1 mg/kg, 7 days, 1 dose/day, o.g.) induced an antidepressant-like response by reducing the time rats spent immobile in the forced-swim test (F2,57 = 22.46, p < 0.001) while increasing escaping behaviors (climbing: F2,57 = 5.95, p = 0.005; swimming: F2,57 = 10.52, p < 0.001). Specifically, both doses of psilocybin significantly decreased the time rats spent immobile when compared to saline-treated rats (0.3 mg/kg: −44 s, ***p < 0.001; 1 mg/kg: −65 s, ***p < 0.001). (4) Repeated psilocybin (0.3 and 1 mg/kg, 7 days, 1 dose/day, o.g.) increased several hippocampal neurogenesis markers (Ki-67, cell proliferation: F2,32 = 6.44, p = 0.005; NeuroD, neural progenitors: F2,32 = 14.51, p < 0.001; and BrdU, cell survival: F2,32 = 9.25, p < 0.001) as measured 1-day post-treatment. (5) Psilocybin induced long-lasting antidepressant-like effects as measured 8- (F2,57 = 17.04, p < 0.001) and up to 15-days (F2,57 = 9.49, p < 0.001) post-treatment in the forced-swim test (i.e., immobility rates). (6) Repeated psilocybin induced long-term neuroplasticity improvements as measured by increased hippocampal NeuroD in adolescent rats of both sexes (F2,36 = 6.48, p = 0.004). (7) A significant positive correlation was observed between the antidepressant-like response (escaping behaviors: climbing + swimming) and hippocampal NeuroD (r = 0.418, n = 39, **p = 0.008) in the same rats.

Conclusions: To the best of our knowledge, these results are the first ones to underscore oral psilocybin’s potential as a fast-acting and long-lasting antidepressant during adolescence, a developmental stage marked by high vulnerability to depression and reduced response to conventional treatments, while also proposing NeuroD as a biomarker of its long-term plasticity.

Keywords: Antidepressant, Adolescence, Psychedelics.

Disclosure: Nothing to disclose.

P670. The psychedelic dopr improves depression-related behavior in mice

Michael Noback, Melissa Flesher, Bruna Cuccurazzu, Adam Halberstadt, Jared Young

University of California - San Diego, San Diego, California, United States

Background: Major depressive disorder (MDD) is strongly associated with impairments in reward learning and effortful motivation. There is evidence that psychedelics can treat depressive symptoms in MDD, but such behavioral/cognitive impairments require investigation. Furthermore, there has been relatively little cross-species preclinical work determining the therapeutic potential of these drugs. We previously showed that the psychedelic compound 2,5-dimethoxy-4-propylamphetamine (DOPR) could elicit head twitch responses (HTR) in mice (analogous to hallucinogenic response in humans) but also increased effortful motivation in mice with low baseline motivation at both HTR-inducing and much lower (micro)dose levels. Here, we specifically tested a targeted model of MDD with these and additional behavioral assays. The acetylcholinesterase inhibitor physostigmine elicited depression symptoms in humans and can induce an MDD-like state in mice. We coadministered DOPR (at micro- and macro-dose levels) and physostigmine to mice and tested them in several cross-species validated touchscreen-based operant tasks sensitive to MDD-relevant impairments, including the progressive ratio breakpoint task (PRBT; effortful motivation), the probabilistic reversal learning task (PRLT; reward learning), and the Iowa gambling task (IGT; reward/punishment risk-taking).

Methods: 60 mice (C57BL/6, 50% female) were trained and baseline tested in touchscreen-based versions of the PRBT and PRLT and counterbalanced into two treatment groups based on performance. Mice were then administered DOPR (0, 0.01, 0.1, or 1 mg/kg; within-subjects design) and either physostigmine (0.03 mg/kg or saline; between-subjects design) and tested in the PRBT and PRLT over four weeks. Mice were then tested in the IGT (0, 0.1, or 1 mg/kg DOPR; within-subjects design), with or without physostigmine (between-subjects). DOPR and physostigmine were administered 30 minutes pre-test.

Results: There was a main effect of DOPR on PRBT performance (F(3,53) = 20.708, p < 0.001), with the lowest dose of DOPR causing a trend increase in PRBT performance (p = 0.064), while the highest dose significantly impaired performance (p < 0.001). After analysis based on a median split (as per prior findings), we found a significant DOPR*median split interaction (F(3,53) = 4.002, p = 0.009), and the breakpoint of low-baseline mice was improved by the lower two doses (p < 0.001, p = 0.055) and impaired by the highest dose (p < 0.001), while high-baseline animals were impaired by the higher two doses (ps < 0.001) and unaffected by the lowest dose (p = 0.935). We observed no effect of physostigmine on breakpoint.

PRLT performance was impaired by DOPR as reflected by the primary outcomes of trials to criterion (TTC; F(3,53) = 5.958, p < 0.001) and switches (F(3,53) = 6.468, p < 0.001), driven by the highest dose of DOPR (TTC: p = 0.005, switches: p = 0.002). As this pattern was also present in total trials, (F(3,53) = 36.380, p < 0.001), it is likely that the effects on TTC and switches are reflective of motor effects of the drug. We observed a DOPR*physostigmine interaction in target touch latency (F(3,53) = 2.376, p = 0.072), in which animals treated with high-dose DOPR showed a significant difference between vehicle- and physostigmine-treated animals, as physostigmine increased latency (p = 0.009).

In the IGT, we observed a significant effect of DOPR on the primary outcome of difference score (F(2,54) = 6.452, p = 0.002), in which high-dose DOPR significantly reduced difference score (p = 0.012). In secondary outcomes, low-dose DOPR was associated with improvements in decision-making behavior in saline-treated mice, specifically safe win-stay (F(2,54) = 2.948, p = 0.061) and risky lose-shift ratios (F(2,54) = 6.552, p = 0.025). Physostigmine-treated mice were insensitive to DOPR in IGT decision-making metrics.

Conclusions: Low doses of DOPR increased effortful motivation in the PRBT and improved reward-related decision making in the IGT. Surprisingly, acute physostigmine did not impair performance in these tasks unlike prior studies and no interaction with DOPR was observed. We replicated prior findings with DOPR and expanded upon them. Importantly, we demonstrated that microdoses of DOPR were effective in improving effortful motivation and reward learning. Given that prior literature primarily focused on the antidepressant effects of macrodoses of related drugs in humans, these data provide important additional evidence for the therapeutic potential of a wider range of doses of psychedelic drugs. Future studies should determine the impact of chronic treatment of such low doses for potential therapeutic potential.

Keywords: Psychedelics, Microdosing with psychedelics, Translational Cognitive Testing.

Disclosure: Nothing to disclose.

P671. Psilocybin differentially recruit GABAergic interneurons in the medial frontal cortex

Pasha Davoudian, Neil Savalia, Ling-Xiao Shao, Joshua Wilson, Amanda Weiner, Quan Jiang, Jack Nothnagel, Takeshi Sakurai, Alex Kwan

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

Background: Psilocybin is a classic psychedelic that may be a promising therapeutic for ameliorating suffering for people living with psychiatric conditions. Most studies of psychedelics in the neocortex have focused on the excitatory pyramidal cells. However, this emphasis is not fully justified, because classic psychedelics are partial or full agonists at numerous serotonin receptors, including the 5-HT2A, 5-HT1A, and 5-HT2C subtypes that are expressed in cortical inhibitory neurons. In agreement, the endogenous neurotransmitter serotonin has been shown to modulate inhibitory synaptic currents and electrophysiological properties of interneurons in brain slices and in vivo.

Methods: We used cell-type specific electrophysiology and two-photon calcium imaging to measure the activity of interneuron subtypes in the mouse medial frontal cortex after the administration of psilocybin. For electrophysiology, we recorded from SST (N = 39 neurons) and PV neurons (N = 53) after saline or psilocybin and compared use bonferroni corrected p-values for multiple comparison. Similarly for calcium imaging we recorded from SST (N = 117 neurons), PV (N = 139 neurons), and VIP neurons (N = 53) and again found a significant decrease in SST calcium effects using mixed effects modeling. We developed and validated a SST-specific 5HT1A knockout. We used the SST-5HT1A KO mice and recorded from N = 15 neurons and saw no change in firing rate after psilocybin. We saw no change in KO mice in terms of head twitch response after psilocybin (N = 18–24 per group, mixed model analysis). We did observe a difference in fear extinction and extinction retention during learned helplessness between (N = 14–20 per group, mixed model analysis).

Results: We find that psilocybin acutely decreases the firing of SST interneurons (P-bonferroni = 8*10-4) but increases spiking of PV interneurons (P-bonferroni = 8*10-5). The suppressive effect of psilocybin on SST interneurons is unlikely due to the disinhibitory circuit, because we did not detect any change in activity in VIP interneurons after psilocybin administration during calcium imaging (P = 0.8 mixed effects model). Instead, 5-HT1A receptor expression in SST interneurons is a key target, supported by experiments where using pharmacological blockade (P = 1.0 bonferroni corrected) and conditional gene knockout (P = 0.3, bonferroni corrected) did not cause a change in firing rate. Deletion of the 5-HT1A receptor in SST interneurons interferes with psilocybin’s long-term effects on stress-related behaviors (mixed effects models P = 0.002 for fear extinction, 0.001 for extinction retention, and 0.004 for fear renewal). Deletion of the 5-HT1A receptor in SST interneurons interferes also interfered with short term stress behavior in the tail suspension test (P = 0.035, mixed effects model) No effect on head twitch response (P = 0.975, mixed effects model).

Conclusions: This study reveals distinct firing patterns among cortical GABAergic cell types in response to psilocybin. In the medial frontal cortex, psilocybin reduced the spiking activity of SST interneurons while increasing the firing of PV interneurons. Pharmacological blockade and conditional knockout experiments indicate that the psilocybin-induced suppression of SST interneurons is mediated by the 5-HT1A receptors expressed on the SST interneurons. This modulation of GABAergic inhibition is functionally important, because deleting the 5-HT1A receptors from SST interneurons impaired the long-term ameliorating effects of psilocybin on stress-related behaviors.

Keywords: Psychedelics, GABAergic interneurons, Psilocybin.

Disclosure: Nothing to disclose.

P672. Effects of psilocybin on behavior and hippocampal genomic response to glucocorticoids in male and female mice

Baila Hammer, Rina Balter, Esther Gross, Deepali Lehri, Ravi Sachidanandam, Jordan Marrocco

Touro University, New York, New York, United States

Background: Psilocybin, as a serotonin receptor agonist, primarily targets the 5-HT2A receptor, inducing neuroplastic changes that may alleviate the effects of chronic stress, such as hippocampal atrophy and reduced neuronal connectivity. Activation of serotonergic receptors regulates the secretion of glucocorticoids (GCs) and the expression of the GC receptor in the hippocampus (HPC) via genomic and epigenomic mechanisms. To test whether the serotonergic agonist psilocybin requires GC signaling to exert anti-stress effects, we used a mouse model of stress susceptibility in which corticosterone (CORT) administered to adult male and female mice in drinking water during the course of several weeks causes disruption of the activity of the hypothalamic-pituitary-adrenal axis, affects emotional behavior, and alters the signaling pathway of the GC receptors.

Methods: CORT (25 ug/ml) or vehicle were administered ad libitum in the drinking water to adult male and female mice over the course of 21 days. The last day of treatment, mice received a single injection of psilocybin (3 mg/kg, i.p.) or saline, and 24 hours later were tested for anxiety-like behavior using the light/dark box test paradigm. On the following day, the ventral HPC (vHPC) was dissected and processed for RNA-sequencing.

Results: We found that male mice treated with vehicle and injected with psilocybin showed increased emotional score compared to male mice treated with vehicle and injected with saline that was similar to male mice treated with CORT and injected with saline. Psilocybin reversed the emotional behavior observed in saline-injected CORT-treated male mice. In addition, psilocybin-injected CORT-treated male mice showed decreased anxiety-like behavior compared to psilocybin-injected vehicle-treated male mice. Treatment with CORT and psilocybin did not affect the behavioral response of females in the light/dark box test. RNA-seq of the vHPC showed differentially expressed genes and pathways involved in GC and glutamate/GABA signaling that were distinct across treatments and were different between males and females.

Conclusions: These findings suggest that GC signaling modulates the molecular and behavioral effects of psilocybin in a mouse model of stress susceptibility.

Keywords: Psilocybin, Glucocorticoids, Ventral Hippocampus.

Disclosure: Nothing to disclose.

P673. Psilocybin (COMP360) induces sustained prohedonic effects in a reverse-translated touchscreen rodent model

Brian Kangas, Kayleigh LaMalfa, Diego Pizzagalli, Jack Bergman, Christopher Thomas

Harvard Medical School, McLean Hospital, Belmont, Massachusetts, United States

Background: In recent years, there has been increased research into the medicinal potential of psychedelics, such as the 5-HT2A receptor agonist psilocybin, for a host of neuropsychiatric disorders. Anhedonia, defined as the loss of interest and motivation from previously reinforcing stimuli and pleasurable activities, is a prominent and often debilitating phenotype in transdiagnostic clinical populations. The aim of the current study was to evaluate the effects of psilocybin (COMP360, COMPASS Pathways’ proprietary formulation of synthetic crystalline psilocybin) and key positive/negative control drugs in the rodent version of the Probabilistic Reward Task (PRT). The PRT is a touchscreen-based assay of reward responsiveness that was reverse-translated from an RDoC-recommended task used to objectively quantify anhedonic phenotypes in psychiatric patient populations.

Methods: Across groups of adult male Long-Evans rats (n = 12 per experiment, n = 72 total), the effects of psilocybin and key comparator drugs were studied for their ability to produce prohedonic phenotypes in the PRT (i.e., increases in adaptive response biases for richly rewarded stimuli). Specifically, psilocybin (0.1–1.0 mg/kg) alone or, in other subjects, following the 5-HT2A antagonist volinanserin (0.03–0.3 mg/kg), the NMDA antagonist ketamine (10 mg/kg), the psychedelic 5-HT2A agonists 2,5-Dimethoxy-4-iodoamphetamine ((±)-DOI; 0.3–3.0 mg/kg) and N,N-dimethyltryptamine (DMT; 1.0–10.0 mg/kg), the non-psychedelic 5-HT2A agonist lisuride (0.1–1.0 mg/kg), and the prototypical SSRI fluoxetine (0.3–3.0 mg/kg) were examined.

Results: Psilocybin and ketamine both produced similarly efficacious and dose-dependent acute increases in reward responsiveness (F[3.1,34] = 5.3, p = 0.004) that persisted 24 hours after dosing (F[2.3,25] = 3.8, p = 0.03) but not at the 7-day time point (F[3.5,38] = 0.47, p = 0.73). Volinanserin blocked this effect in a dose-dependent manner (p > 0.05), suggesting that the mechanism is 5-HT2A mediated. Treatment with (±)-DOI (F[2.6,29] = 4.0, p = 0.02) and DMT (F[2.0,22] = 4.2, p = 0.03) produced significant dose-dependent prohedonic effects acutely, although neither DOI’s (F[2.3,25] = 0.15, p = 0.89) nor DMT’s (F[2.4,27] = 1.2, p = 0.32) effects were present 24 hours post drug administration. Lisuride failed to produce prohedonic phenotypes in the PRT at any time point and, instead, significantly decreased reward responsiveness (F[2.5,27] = 5.2, p = 0.008). Acute doses of fluoxetine did not affect PRT performance (F[1.9,21] = 0.04, p = 0.96), which is consistent with the inability of SSRIs to produce prohedonic effects in humans.

Conclusions: Overall, the present findings demonstrate that psilocybin has rapid and enduring positive effects on reward responsiveness in a translational rodent task. Psilocybin’s prohedonic efficacy appears to similar that of ketamine and may have a more favorable time course than the 5-HT2A psychedelics (±)-DOI and DMT. Studies with volinanserin and lisuride collectively suggested that agonism of the 5-HT2A receptor is necessary, but may not be sufficient, for these effects. Future directions include evaluating the role of relief from anhedonia in mediating the potential therapeutic efficacy of psilocybin and other psychedelics in clinical populations.

Keywords: anhedonia, Touchscreen cognitive testing, Psychedelic medicine.

Disclosure: Nothing to disclose.

P674. Developmental changes in brain structure, function, and sensory perception following exposure to oral psilocybin during adolescence

Craig Ferris, Richard Ortiz, Itishree Sahoo, Sairam Masadi, Ashwath Maheswari, Michael Gitcho, Rachel Utama, Tochi Chukwuemeka, Noah Cavallaro, Eric Brengel, Praveen Kulkarni

Northeastern University, Boston, Massachusetts, United States

Background: Psilocybin is a hallucinogen with complex neurobiological and behavioral effects. Underlying these effects are changes in brain neuroplasticity. We hypothesized psilocybin given during adolescence, a time of heightened neuroplasticity, particularly in the forebrain, would affect emotional behavior and the associated underlying neuroanatomy, neurocircuitry, and epigenetics.

Methods: Female and male mice were given vehicle or 3.0 mg/kg psilocybin every other day by oral gavage from postnatal days 40 – 50 for a total of five exposures. Between postnatal days 90 −120 mice were imaged and tested for affective behavior and perception of rewarding and aversive stimuli. MRI data from voxel-based morphometry, diffusion weighted imaging, and BOLD resting state functional connectivity were registered to a mouse 3D MRI atlas with 139 brain regions providing site-specific differences in global brain structure and functional connectivity between experimental groups. The prefrontal cortex was measured for changes in proteins associated with epigenetics.

Results: Mice showed no significant differences in affective behavior. Mice with early psilocybin exposure showed reduced brain sensitivity to both rewarding and aversive odors during scanning sessions. There were regional reductions in brain volume and alteration in water diffusivity affecting males more than females. Global and regional functional connectivity were increased in both sexes with the prefrontal cortex showing enhanced connections to the hypothalamus, thalamus and midbrain. Males showed reduced levels of epigenetic and neuroplasticity protein markers in the prefrontal cortex. Interestingly, when comparing male and female rats regardless of treatment there were significant sex-dependent changes.

Conclusions: The pronounced changes in brain volume, water diffusivity - a surrogate marker of gray matter microarchitecture, increase in functional connectivity, altered perception of rewarding and aversive stimuli and enhanced levels of protein markers of neuroplasticity provide compelling evidence that exposure to psilocybin during periadolescence has long term developmental consequences particularly in males.

Keywords: Diffusion Tensor Imaging (DTI), Resting State Functional Connectivity, neuroepigenetics.

Disclosure: Ekam Imaging, Founder, Self.

P675. The Pharmacological, pharmacokinetic and potential neuroplasticity-related profile of 5-MeO-DMT (mebufotenin), a 5-HT receptor agonist for the treatment of mood disorders

Michelle Kokkinou, Vaishnavi Madhavan, Tim Mason, Amir Lotfi

Beckley Psytech, Oxford, United Kingdom

Background: Major depressive disorder (MDD) is a leading cause of disability worldwide, impacting over 300 million people, with a substantial proportion of patients not adequately responding to treatment, contributing to the social, economic and personal disease burden. More effective treatment strategies are required. BPL-003 is a benzoate salt formulation of mebufotenin, also known as 5-methoxy-N,N-dimethyltryptamine (5-MeO-DMT), in development for the treatment of mood disorders. 5-MeO-DMT is a naturally occurring psychedelic, serotonin (5-HT) receptor agonist associated with rapid-acting antidepressant effects and has been proposed to promote neuroplasticity. The objectives of this study were to investigate in vitro binding, in vivo pharmacokinetics (PK), efficacy and potential mechanism of action of 5-MeO-DMT preclinically.

Methods: Pharmacology of 5-MeO-DMT (10µM) was evaluated in vitro using 102 receptor binding, and 29 enzyme and uptake assays across a range of targets in human and rat receptor preparations. PK parameters of 5-MeO-DMT were evaluated in male and female rats (n = 3/sex/group) and in male mice (n = 3/group) following intranasal (i.n.) or subcutaneous (s.c.) administration respectively. 5-MeO-DMT (0.5, 1.5, 5mg/kg, s.c.) efficacy was assessed using the forced swim test (FST) in male mice (n = 8/group) with time spent immobile as the outcome measure. In vitro markers of 5-MeO-DMT-induced neuronal plasticity (neuronal number, neurite length, branching points) were analysed in a primary culture of rat cortical neurons, incubated with 5-MeO-DMT (0.1, 0.3, 1, 10µM). Neurogenesis markers Ki-67 and doublecortin (DCX) in the dentate gyrus of the hippocampus were investigated ex vivo using immunohistochemistry, following in vivo acute administration of 5-MeO-DMT (0.5, 5mg/kg, s.c) in mice (n = 4–5/group). Data were analyzed using univariate statistics, one-way ANOVA and post hoc Fisher’s Least Significant Difference tests. Statistical significance was determined as p < 0.05. Protocols in animal studies were conducted in accordance with the OECD Principles of Good Laboratory Practice for the in vivo PK study in rats, in accordance with the principles of the Canadian Council on Animal Care for the FST experiment, or National Institutes of Health Guide for the Care and Use of Laboratory Animals and followed European Union regulations (Directive 2010 / 63 /EU) for neuroplasticity and neurogenesis studies.

Results: 5-MeO-DMT exhibits selective and high affinity for subtypes of 5-HT receptors, as assessed by percent inhibition of the binding of a radioactively labelled ligand specific for each target. Comparable systemic exposures were obtained from s.c. and i.n. routes of administration to support clinical translation of the rodent findings. Where applicable, peak (Cmax) and total systemic exposures (AUC0-Tlast) to 5-MeO-DMT increased approximately in a dose-dependent manner, independent of sex. 5-MeO-DMT relative to vehicle treatment demonstrated reduction in behavioral despair indexed by decrease in immobility time 3 hours following treatment (F4,35 = 9.484, p < 0.001, Saline vs 5-MeO-DMT treatment at 0.5 and 1.5mg/kg, p < 0.05) with no effect when tested at 24 hours and 7 days post administration (p > 0.05). No overt behavioural signs that could have confounded immobility time assessment during the FST were observed at the time of testing (p > 0.05). 5-MeO-DMT significantly increased the number of neurons (at 0.3µM and 1µM, p < 0.05), the neurite network (0.3–10 µM, p < 0.05) and the number of branching points (0.3µM, p < 0.05) relative to control condition. The number of Ki-67 and DCX positive cells in the dentate gyrus of the hippocampus was significantly increased 24h post 5-MeO-DMT vs vehicle treatment (p < 0.05).

Conclusions: In summary, 5-MeO-DMT as a potent and selective serotonergic psychedelic shows a high degree of binding affinity to 5-HT receptor subtypes in vitro and exhibits dose dependent increases in PK parameters. Acute 5-MeO-DMT treatment is associated with reduction in behavioural despair indexed by decrease in immobility time, interpreted as antidepressant-like behavior, at doses and times where no overt behavioural signs are detected. 5-MeO-DMT promotes markers of neuroplasticity in rat cortical cultures and neurogenesis in the hippocampus of mice which help further elucidate the mechanism of action beyond receptor effects. These preclinical findings support effects of 5-MeO-DMT potentially relevant to its therapeutic use for treatment of mood disorders and identify therapeutic approaches for targeting neuroplasticity alterations thought to be implicated in individuals with mood disorders.

Keywords: Psychedelics, 5-MeO-DMT, Antidepressant, Neuroplasticity, Treatment Resistant Depression.

Disclosure: Beckley Psytech, Employee, Self.

P676. Psilocybin triggers an activity-dependent rewiring of large-scale cortical networks

Quan Jiang, Ling-Xiao Shao, Shenqin Yao, Neil Savalia, Amelia Gilbert, Pasha Davoudian, Jack Nothnagel, Guilian Tian, Tin Shing Hung, Hei-Ming Lai, Kevin Beier, Hongkui Zeng, Alex Kwan

Cornell University, Ithaca, New York, United States

Background: The classic psychedelic psilocybin is a 5-HT2A receptor agonist and holds promise as a treatment for mental illnesses. One dose of psilocybin induces structural remodeling of dendritic spines in the medial frontal cortex in mice. As functional synapses, the dendritic spines would be innervated by presynaptic neurons, but the sources of these inputs have not been identified.

Methods: Using monosynaptic rabies viral tracing, we map the brain-wide distribution of inputs to frontal cortical pyramidal neurons after psilocybin or saline administration. Each brain was imaged after whole-tissue clearing and with light sheet fluorescence microscopy. Samples included Fezf2-2A-CreER mice (n = 2680 cells per mouse from 4 male and 5 female animals for psilocybin, 2350 cells per mouse from 4 male and 4 female animals for saline) and PlexinD1-2A-CreER mice (n = 5495 cells per mouse from 4 male and 4 female animals for psilocybin, 4511 cells per mouse from 4 male and 4 female animals for saline). Additional bioinformatics, in vivo electrophysiology, and chemogenetics studies were performed to delineate the mechanisms for the psilocybin-induced effects.

Results: We discover that psilocybin’s effect on connectivity is network-specific: strengthening the routing of inputs from perceptual and medial regions (homolog of default mode network) to subcortical targets, while weakening inputs that are part of cortico-cortical recurrent loops. A major factor that shapes the pattern of synaptic reorganization is the drug-evoked spiking activity acutely after psilocybin administration.

Conclusions: Collectively, the results reveal the impact of psilocybin on the connectivity of large-scale cortical networks and demonstrate neural activity modulation as an approach to sculpt the psychedelic-evoked neural plasticity. Our study hints at an exciting avenue for future research to combine neuromodulation with 5-HT2A receptor agonists to precisely target specific circuits for neural plasticity.

Keywords: Psilocybin, Psychedelics, Rabies tracing, Synaptic Plasticity, Serotonin 5-HT2A Receptor.

Disclosure: Empyrean Neuroscience, Advisory Board, Self, Freedom Biosciences, Advisory Board, Self, Xylo Bio, Advisory Board, Self, Boehringer Ingelheim, Consultant, Self, Eli Lilly, Consultant, Self, Intra-Cellular Therapies, Contracted Research, Self.

P677. Pre-administration of lorazepam negates long-term antidepressant-like effects of psilocybin

Charles Nichols, Sophie Woodruff, Meghan Hibicke

LSU Health Sciences Center, New Orleans, Louisiana, United States

Background: The classic psychedelic psilocybin has been shown to induce rapid-acting antidepressant effects that persist for at least up to 6 months after a single administration. In clinical practice, it is common to prescribe benzodiazepines alongside antidepressant therapy. The effects of benzodiazepines on antidepressant effects elicited by psychedelics have not been explored in either human clinical trials or in preclinical animal models.

Methods: WKY rats were randomly assigned to one of four treatment groups: saline/saline (S/S, n = 8), lorazepam/saline (L/S, n = 8), saline/psilocybin (S/P, n = 8), or lorazepam/psilocybin (L/P, n = 8). Treatments were delivered through two bolus intraperitoneal injections separated by 30-minute intervals, with rats receiving either saline or lorazepam first and either saline or psilocybin second. Antidepressant-like efficacy was assessed using the Forced Swim test (FST) at 3 weeks, 5 weeks, 7 weeks, 9 weeks, and 11 weeks post-treatment. Three subdivisions of the mPFC: the anterior cingulate cortex (ACC), prelimbic cortex (PL), and infralimbic cortex (IL) were extracted to perform quantitative PCR using TaqMan® Gene Expression Assays for 17 targets of interest representing several different types of ion channels.

Results: S/P-treated rats exhibited sustained antidepressant-like effects for up to 9 weeks. Pre-administration of lorazepam 30 minutes prior to psilocybin negated psilocybin’s long-term antidepressant-like effects. Lorazepam was associated with a decrease in Gria3 expression, and psilocybin treatment was associated with an increase in Gria4 expression at 12 weeks post-administration in the PL.

Conclusions: Psilocybin produced long-lasting antidepressant-like effects, and administration of a BZD prior to psilocybin prevented these effects. Two genes encoding for AMPA channel subunits were altered in response to treatment, however their implications in long-term antidepressant-like effects remains to be determined. Our results have important clinical implications on the use of psychedelics to treat depression, and contraindicate the presence of benzodiazepines during an acute treatment with psychedelic.

Keywords: Psilocybin, benzodiazepine, neuropharmacology, Preclinical models and endpoints, Psychedelics.

Disclosure: Nothing to disclose.

P678. Psilocybin augments dopamine responses during probabilistic reversal learning in female rats

Kyna Conn, Gabriel Wong, Catherine Huang, Laura Milton, Erika Greaves, Claire Foldi

Monash University, Melbourne, Australia

Background: Cognitive inflexibility - the inability to adapt behavioral strategies when circumstances change - represents a fundamental pathological mechanism underlying depression, anxiety, and eating disorders. Ventral striatal dopamine (DA) orchestrates this adaptive capacity by encoding prediction errors and signaling when established reward contingencies shift, making it essential for successful reversal learning. Despite the promising potential of psilocybin for alleviating mental ill health, and our prior demonstration that psilocybin enhances reversal learning in rats, the precise dopaminergic mechanisms remain unexplored. Understanding how psilocybin modulates DA signaling during cognitive flexibility tasks could unlock targeted interventions for the debilitating cognitive rigidity that perpetuates multiple psychiatric conditions.

Methods: In order to determine whether the cognitive benefits of psilocybin arise from enhanced ventral striatal dopamine signaling, adult female Sprague-Dawley rats (N = 17) received viral injections of the dopamine biosensor (GRAB-DA2m) and fiber optic implants targeting the ventral striatum for real-time dopamine monitoring via fiber photometry. Following recovery, rats learned a probabilistic reversal task (80:20 reward contingencies) using nose-poke operant devices. Animals received psilocybin (1.5 mg/kg) or vehicle 24 hours before testing onset. We simultaneously recorded reversal learning performance and ventral striatal dopamine dynamics across 7 consecutive testing sessions to determine whether psilocybin alters trial-by-trial DA responses during flexible learning. Data were analysed using custom code to evaluate the slope, peak and area under the curve of changes in DA release in response to reward receipt, as well as using mixed models analysis across time and between treatment groups.

Results: Psilocybin significantly enhanced ventral striatal dopamine transient amplitude and duration following both expected and unexpected rewards (p < .001 and p < .01, respectively) compared to vehicle controls. Enhanced dopamine responses were also observed following expected losses (p < .001), but not unexpected losses. These neurochemical changes occurred alongside improved behavioral performance, with psilocybin-treated rats showing trends toward increased successful trial completion and reversal acquisition (p = .052). Dopamine signal modulation was sustained across all seven testing sessions, demonstrating persistent neurochemical effects 24 hours post-administration. These data establish that psilocybin robustly alters ventral striatal dopamine dynamics in a probabilistic reversal learning paradigm.

Conclusions: We show that psilocybin enhances phasic dopamine signaling in response to changing reward contingencies, which may underlie improvements in cognitive flexibility. The ventral striatum plays a critical role in processing reward outcomes by integrating motivational and value-related signals that guide adaptive decision-making. Therefore, dopaminergic tuning offers a plausible mechanism for psilocybin's therapeutic potential in disorders characterized by rigid or perseverative behavior. Given that cognitive inflexibility perpetuates treatment resistance psychiatric disorders, our data suggest psilocybin may circumvent traditional therapeutic limitations by directly targeting the neurochemical substrates of maladaptive cognitions, potentially enabling breakthrough clinical responses where conventional interventions fail.

Keywords: Psilocybin, Dopamine, Eating disorders, Reinforcement learning, Cognitive/behavioral flexibility.

Disclosure: Nothing to disclose.

P679. Imaging AMPA receptors in PTSD: a pilot positron emission tomography (PET) study

Krisha Shah, Alan Prossin, Takuya Takahashi, Meixiang Yu, Julia Engelhardt, Andreas Weyland, Chadi Abdallah, Nicholas Murphy, Sanjay Mathew

McGovern Medical School at UT Health, Houston, Texas, United States

Background: Post-traumatic stress disorder (PTSD) is a debilitating illness characterized by a clinical syndrome composed of intrusive traumatic reminders, hyperarousal, and changes in cognition and affective stability. Development of novel, interventions that target specific symptoms are sorely needed to improve the lives of people suffering with PTSD. However, despite decades of efforts, available treatments repeatedly fall short of expectations in the clinic. This is due in part to challenges in translating molecular pathophysiology of animal models of PTSD into robust clinical targets. Alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) glutamate receptors play a crucial role in regulating cellular and behavioral learning mechanisms applicable to fear extinction in PTSD. Emerging evidence also highlights the regional complexity of brain changes in people with PTSD. Whereas enhanced glutamatergic activity in the ventral hippocampus can facilitate fear learning and fear-related behavioral responses, reduced glutamatergic activity has been observed in the prefrontal cortex in patients with PTSD. Enhanced understanding of these mechanisms is sorely needed to develop novel, effective personalized treatment strategies in PTSD. AMPAR-mediated mechanisms have high translational significance owing to their mediation of ketamine’s rapid therapeutic effects in stress-related conditions and their role as target in ongoing early-phase drug discovery programs (e.g., AMPAR positive allosteric modulators). Accordingly, the recent discovery of a selective positron emission tomography (PET) radiotracer ([11C]K-2) for AMPA glutamate receptors has broad potential application for PTSD and CNS disorders marked by imbalances in excitatory/inhibitory synapses. This pilot study pioneers the measurement of AMPAR in females with chronic PTSD using a novel PET tracer to test hypotheses regarding the role of this fundamental glutamate receptor. This work is significant for understanding synaptic mechanisms in PTSD and related conditions, laying important groundwork for enhancing the scope of future drug development.

Methods: 20 female participants (10 PTSD, 10 trauma-naive controls), aged 21–45 were recruited for a cross-sectional study of 11C-K2 PET imaging. Patients met DSM-5 criteria for chronic PTSD and had a CAPS-5 score (in the last month) > 35. Substance use, psychosis, traumatic brain injury, and use of glutamate modulating medications were excluded. PET imaging of the brain was acquired in list mode over 90 minutes using a Siemens Biograph Vision PET/CT Scanner. Participants completed 3-tesla T1 weighted structural MRI for purposes of anatomical co-registration. PET radiotracer modeling was completed in PMOD software (ver. 4.2, Bruker, Zurich, CH), the chosen reference region being an area devoid of AMPA receptors (white matter). Each MRI image was normalized to the MNI template. Resulting transformation maps for each subject were applied to their PET images, normalizing their data. Non-displaceable binding potential (BPND) maps were calculated for each participant as previously described.

Voxel-wise, whole brain statistical analyses were completed within Statistical Parametric Mapping software (SPM25: Wellcome Trust, UK) and MATLAB (Mathworks, Natick, MA) with hypotheses based on a priori regions of interest including: ventromedial prefrontal cortex (vmPFC), hippocampus (Hip), orbitofrontal cortex, dorsolateral PFC, and amygdala. Statistical modeling relied on 1-way ANOVA to determine diagnostic differences (PTSD vs. Healthy Controls) AMPA BPND as quantified by 11C-K2 PET. Significance was set to puncorr < 0.01 for a priori hypothesized brain regions and puncorr < 0.001 for other brain regions.

Results: Significant diagnostic differences were determined wherein AMPA receptor availability (e.g. BPND) was greater in Healthy Controls within the Left Hip (XYZ: 32, −6, −30; T36 = 3.04; p < 0.002). Additional effects were present bilaterally in the Inferior Temporal Lobes (R: T36 = 3.09, p < 0.002; L: T36 = 3.20, p < 0.001) however only the effect in the Left Inferior Temporal Lobe was significant at the p < 0.001 level. Results provide initial evidence of enhanced AMPA receptor activity (lower BPND) in the ventral left hippocampus in patients with PTSD (as compared to healthy controls). No diagnostic differences were found within the vmPFC or other apriori hypothesized brain regions.

Conclusions: The enhanced hippocampal AMPA Receptor activity we evidence in our PET study of humans suffering with PTSD helps to advance translation of evidence from animal models wherein enhanced ventral hippocampus AMPA receptor activity associated with fear response learning. Validation of these effects in follow-up studies will provide additional support to for development of innovative treatment strategies relying on AMPA modulation within the hippocampus as a viable target for pharmacological and neuromodulatory interventions in PTSD.

Keywords: PTSD, AMPA receptors, PET Imaging Study.

Disclosure: Evolution Research Group, Consultant, Self.

P680. Transcranial magnetic stimulation enhances neural mechanisms of fear extinction recall: evidence from vmPFC and dACC modulation

Aditya Pillai, Kai Zhang, Lihan Cui, Zhenfu Wen, Mohammed Milad

The University of Texas Health Science Center At Houston, Houston, Texas, United States

Background: Fear extinction is a fundamental paradigm for probing the mechanisms of Posttraumatic Stress Disorder (PTSD) and related anxiety disorders, as it models the capacity to suppress learned threat responses. However, traditional extinction training is often incomplete, unstable across time, and prone to relapse, limiting its translational value for clinical therapies. Neuromodulation approaches such as transcranial magnetic stimulation (TMS) have emerged as promising tools to directly influence prefrontal–limbic circuitry, with prior work suggesting their capacity to modulate regions implicated in safety and threat learning, including the dorsolateral prefrontal cortex (dlPFC) and ventromedial prefrontal cortex (vmPFC). Yet, whether TMS can causally enhance extinction recall and its underlying neural mechanisms remains largely unknown. Addressing this gap is critical for developing circuit-informed interventions that may augment extinction-based therapies.

Methods: Forty-eight healthy adults completed a three-day fear conditioning and extinction paradigm. On Day 1 (conditioning), participants viewed three conditioned stimuli (CSs), two of which were paired with a mild shock (CS+) and one that was never paired (CS–). Resting-state fMRI was collected to guide stimulation target identification in the dorsolateral prefrontal cortex (dlPFC). On Day 2 (extinction learning), one CS+ was paired with repetitive TMS (rTMS) delivered over individualized dlPFC targets, with each train initiated 100 ms after CS+TMS onset, while cues were presented without shock. On Day 3 (extinction recall/renewal), all cues were re-presented to assess extinction memory. Neuroimaging data were analyzed using General Linear Modeling (GLM) in combination with individualized electric field (E-field) modeling. Specifically, task-based fMRI data were analyzed using General Linear Modeling (GLM) to generate activation maps for extinction recall contrasts (CS+TMS > CS+NoTMS). In parallel, individualized cortical TMS electric field (E-field) distributions were computed using electromagnetic forward modeling in realistically shaped head models, with coil positions derived from navigator data. For group-level analyses, cortical surfaces were nonlinearly registered to an average brain (FreeSurfer), and both E-field distributions and GLM-derived activation maps were averaged across subjects to generate group-level correlation activation maps. Psychophysiological responses (SCR) were also recorded, with extinction-related changes assessed in secondary analyses.

Results: Whole-brain contrasts of extinction recall (CS+TMS > CS+NoTMS) revealed significantly greater activation in the ventromedial prefrontal cortex (vmPFC)—a region critically implicated in safety memory retrieval—and reduced activation in the dorsal anterior cingulate cortex (dACC), a key node of the threat expression network. These effects remained significant after small-volume correction (p < 0.05). In contrast, secondary analyses of psychophysiological responses indicated no significant group differences in skin conductance responses (SCR) between CS+TMS and CS+NoTMS conditions.

Conclusions: This study provides the first evidence that transcranial magnetic stimulation, precisely targeted to the dorsolateral prefrontal cortex, can modulate fear extinction processes in humans. Consistent with our hypothesis, TMS during extinction learning selectively strengthened vmPFC safety-related activation, a key neural substrate of extinction recall. These findings demonstrate the potential of neuromodulation to achieve precise circuit-level regulation of extinction memory, underscoring its promise as a novel therapeutic strategy for PTSD and related anxiety disorders.

Keywords: TMS, Fear conditioning and extinction, General Linear Modeling.

Disclosure: Nothing to disclose.

P681. A thalamic mechanism for PTSD symptom persistence

Muhammad Badarnee, Mohammed Milad

University of Texas-Health Sciences Center at Houston, Houston, Texas, United States

Background: Persistent fear is a core feature of post-traumatic stress disorder (PTSD), but the brain circuits that link initial threat learning to enduring clinical symptoms are not well understood. While the mediodorsal thalamus (MD) and anterior pulvinar (PuA) are known to be critical for fear learning, their specific roles in the neurobiology of PTSD remain unclear. Herein, we investigated the functional roles of these nuclei in PTSD and the relationships with clinical measures during Pavlovian fear conditioning and renewal – cornerstone phases mimicking fear acquisition and relapse after extinction.

Methods: We analyzed the fMRI data of large cohorts of 114 individuals with PTSD, 134 trauma-exposed controls, and 158 non-trauma-exposed controls who underwent Pavlovian fear conditioning in the MRI. We examined activation patterns in thalamic subregions, their associations with clinical symptoms, and connectivity with main regions associated with the fear circuit: subcortical regions (amygdala, hippocampus) and cortical regions (ventromedial prefrontal cortex, subgenual anterior cingulate cortex, and dorsal anterior cingulate cortex).

Results: Our findings reveal that even before fear associations were formed, patients with PTSD exhibited significantly greater activation of the mediodorsal thalamus (MD) during the earliest stage of conditioning (PTSD > Controls, p < 0.05). This early hyperactivation was robustly associated with PTSD symptom severity (r = 0.24, p < 0.01, 10⁴ bootstrapped 95% CI excluding zero). Further analysis localized this effect to the medial subregion of the MD (MDm), which was the only subregion to distinguish PTSD patients from controls (p < 0.05) and correlate with symptom severity (r = 0.25, p < 0.01, 10⁴ bootstrapped 95% CI excluding zero).

Using principal component analysis, we next classified the functional connectivity of the MDm with each fear circuit regions. We identified two distinct components: MDm–cortical and MDm–subcortical components. Mediation model demonstrated a direct effect of MDm–cortical component on PTSD symptom severity, whereas MDm–subcortical component predicted symptoms indirectly via its relationship with trait anxiety (all path p-values < 0.05). Importantly, these relationships were not limited to PTSD but also extended to depressive symptoms.

By contrast, we found no group differences in anterior pulvinar activation or connectivity across phases, suggesting that this survival-related pathway remains intact in PTSD, and that dysfunction is specific to the MD.

Conclusions: Our results establish the MDm as a central hub of thalamo-cortical and thalamo-subcortical circuit dysfunction in PTSD. We propose a new mechanistic pathway where early MDm hyperactivation during threat learning sets the stage for persistent fear responses. This brain-trait-symptom relationship provides a critical insight into how initial thalamic dysfunction translates into lasting psychopathology symptoms. Our findings identify the MDm as a promising circuit-based target for developing novel neuromodulation therapies to alleviate the burden of PTSD.

Keywords: Thalamus, Fear conditioning, Post Traumatic Stress Disorder, Fear renewal, mediodorsal nucleus of the thalamus.

Disclosure: Nothing to disclose.

P682. Effects of sternal vibration during mindfulness meditation on executive control networks in trauma-exposed adults

Travis Fulton, Alfonsina Guelfo, Aziz Elbasheir, Timothy McDermott, Timothy Ely, Vishwadeep Ahluwalia, Greg Siegle, Negar Fani

Emory University, Atlanta, Georgia, United States

Background: Mindfulness-based interventions have demonstrated benefits for individuals with post-traumatic stress disorder (PTSD), potentially via altered function in neural networks involved with executive control and emotion regulation. Emerging research demonstrates that pairing mindfulness with neurostimulation may potentiate these effects. One promising neurostimulation approach is sternal vibration. We have found that, when paired with breath-focused meditation, respiration-synced sternal vibration enhances meditative engagement as well as cognition and emotion regulation. In this ongoing clinical trial, we investigated the effects of vibration-augmented mindfulness compared to non-augmented mindfulness meditation on executive control networks in trauma-exposed adults with varying symptoms of PTSD. We also examined associations of executive control network change with change in performance on tasks of executive functioning.

Methods: One hundred and fifty-three adults (n = 153, 75.2% female) aged 18–61 (Mage = 31.6 SD = 12.3 years) were recruited from the community for a clinical trial examining effects of sternal vibration during mindfulness meditation on interoceptive and attentional control networks (NCT04670640). Participants were recruited on the basis of prior trauma exposure (PTSD Checklist for DSM-V score = 42.3 SD = 15.0) and elevated dissociative symptoms (Multi-scale Dissociation Index Depersonalization = 13.3 SD = 4.8). At pre- and post-intervention, participants completed a computerized executive functioning battery that included a Letter N-back (LNB) task; they also completed pre- and post-intervention MRI scanning, which included an affective Stroop (AS) task during which they viewed aversive distractor images. Following the baseline scan, participants were randomly assigned to one of four intervention conditions, a 2 × 2 design that included vibration (respiration-synced or pulsed vibration) or non-vibration mindfulness meditation conditions (breath-focused mindfulness or open awareness mindfulness meditation). Each participant completed eight, 20-minute intervention sessions on separate days over the course of 1–2 months. Executive networks were meta-analytically derived using Neurosynth, and activation examined from pre- to post-intervention; AS task performance and LNB task performance were examined in relation to network changes.

Results: A linear mixed-effects model revealed a significant time by intervention condition interaction on executive network activation during aversive AS task trials, F(1, 157.79) = 4.37, p = 0.038, explication of this interaction indicated decreased executive network activity between pre- and post-scans in the vibration group and increased executive network activity in response to aversive AS trials in the non-vibration group. No significant main effects of time (F(1, 157.79) = 1.61, p = 0.207) or intervention condition (F(1, 157.79) = 0.54, p = 0.464) were observed. Changes in executive network activation from pre-intervention to post-intervention were not associated with change in AS task performance, (error rate and response time); however, executive network activation reduction significantly associated with improved performance on the LNB 1-back condition (r = −0.170, p = 0.036); decreased executive network activation was observed to improved LNB performance over time.

Conclusions: This study provides preliminary evidence that pairing mindfulness with sternal vibration may reduce executive network activation during performance of an affective attentional control task in trauma-exposed individuals. Given findings that link reduced executive network activation to better task performance, this reduced engagement may suggest enhanced cognitive efficiency. In PTSD, heightened or dysregulated activation of the executive network during exposure to trauma-related or aversive cues can reflect impaired resource allocation, often contributing to difficulties with attention regulation, and performance on demanding tasks. Reduced recruitment of the executive network during attention to emotionally salient stimuli may indicate more efficient processing or adaptive disengagement, allowing for greater cognitive resources to be directed toward task demands. Vibration-assisted mindfulness may enhance the functional efficiency of the executive network by dampening hyper-responsivity to emotionally charged stimuli, thereby facilitating improved cognitive control in non-affective contexts. These data suggest a promising mechanistic pathway by which augmented mindfulness interventions may promote cognitive and emotional regulation in individuals with PTSD.

Keywords: Non-invasive Neuromodulation, PTSD, Mindfulness Meditation, executive network, cognitive control.

Disclosure: Nothing to disclose.

P683. Exogenous estradiol facilitates memory for conditioned safety cues among women with high posttraumatic stress symptoms

Katelyn Oliver, Alyssa Roeckner, Cecilia Hinojosa, Kristina Dahlgren, Justin Santos, Linzie Taylor, Amy Murphy, Colin Johnson, Dasani DelRosario, Helena Zeleke, Timothy Ely, Rebecca Hinrichs, Natalie Merrill, Marisa Young, Andrea Braden, Abigail Powers, Sanne van Rooij, Vasiliki Michopoulos, Jennifer Stevens

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

Background: Fluctuations in ovarian hormones, and in particular low estradiol (E2), have been linked to women’s increased risk for posttraumatic stress disorder (PTSD) and various trauma-related symptoms. E2 targets receptors in key emotional memory-related brain regions and has been shown to have positive memory facilitation effects. Yet very little research has causally investigated E2’s influences on threat-related memory mechanisms in women impacted by trauma exposure.

Methods: We administered exogenous E2 to naturally-cycling, Black women (ages 18–35 years) with a range of PTSD symptom severity (PTSS; based on CAPS-5, M = 14.3 [12.4]) using a randomized, placebo-controlled, double-blinded, cross-over design. N = 114 participants received transdermal E2 or placebo during low E2 phases of the menstrual cycle (early follicular or early luteal) and underwent a categorical threat conditioning paradigm during fMRI. A category of unique images (either animals or tools, counterbalanced across participants) was associated with a loud aversive noise, making this category threat-associated (CS+), while the other category was never paired with aversive stimuli and considered safe (CS-). Following the scan, participants were tested for episodic recognition of individual images seen during conditioning and had their blood drawn. The following month, participants repeated experimental procedures during the same cycle phase but received the opposite patch condition. We utilized mixed effects models to test for PTSS (CAPS-5 scores) and E2 patch effects on post-scan recognition as well as region of interest (ROI) responses during encoding of conditioned stimuli (contrasts: CS+ and CS- > implicit baseline). ROIs investigated were: BA25, amygdala, hippocampus, and hippocampal complex structures: entorhinal cortex (ERC), perirhinal cortex, and parahippocampus. We also conducted whole-brain analyses to investigate alternative regional responses associated with patch, PTSS, and recognition differences. Whole-brain results were thresholded at p < 0.005 and corrected for multiple comparisons using false discovery rate correction.

Results: Transdermal E2 patches successfully increased serum E2 concentrations compared to placebo condition (+91.49 pg/mL in early-follicular, +80.88 pg/mL in early-luteal; main effects of E2 patch: t = 6.33, p < 0.0001). Across both cohorts, there was a small post-scan recognition bias for threat-related category images compared to the safety category (MCS+=0.56 [0.21], MCS- = 0.51 [0.21], t = 3.46, p = 0.0006). Mixed effects models revealed a significant three-way interaction of PTSS x patch condition x cycle phase on post-scan recognition (t = −2.649, p = 0.009 corrected). Post-hoc testing revealed significant effects of exogenous E2 on CS- category recognition in women with greater PTSS during the early follicular phase (t = 3.01, p = 0.0041), but not the early luteal phase. This E2 patch improvement in CS- memory during the early follicular phase was supported by increased CS- activations in the hippocampus (t = 2.623, p = 0.036 corrected), ERC (t = 3.17, p = 0.020 corrected), and left superior occipital cortex (t = 4.46, p = 0.002 corrected) in women with greater PTSS. Additionally, whole-brain analyses revealed that women with greater PTSS taking E2 during the early follicular phase showed greater CS+ recognition with higher CS+ activations in the left thalamus, insula, mid-cingulate, and supramarginal gyrus (p’s < 0.007).

Conclusions: In this double blind, placebo-controlled, crossover trial, we showed that E2 had an enhancing effect on episodic memory for safety-related cues, but only in the early follicular phase and for women with greater PTSS. This enhancement in safety-related memory was supported by increased functional activations in several visual and memory associated brain regions. The early luteal phase cohort may experience confounding influences of progesterone. Whereas endogenous E2 and other ovarian hormones are lowest during the early follicular phase, making this an ideal window to observe causal effects of exogenous E2 on these systems. Our results suggest that E2 may help women with high PTSD symptoms – who typically experience fear overgeneralization and memory deficits – form memories for safety cues, which may help them better process and integrate safety related information. This could have important clinical implications for understanding and developing more effective treatments in women who experience high burdens due to trauma.

Keywords: Functional MRI (fMRI), Threat Conditioning, Transdermal Estradiol, Post-traumatic stress symptoms, recognition memory.

Disclosure: Nothing to disclose.

P684. Do Dissociative symptoms predict PTSD symptom reduction following ketamine infusions?

Amanda Tamman, John Roache, Lynnette Averill, Christopher Averill, Brittany O' Brien, Ralitza Gueorguieva, Paulo Shiroma, Kevin Guthmiller, Karl Lautenschlager, Argelio López-Roca, John McCallin III, Stacey Young-McCaughan, Terence Keane, Alan Peterson, John Krystal, Chadi Abdallah

Baylor College of Medicine, Houston, Texas, United States

Background: Intravenous ketamine has been used off-label to treat various psychiatric conditions, with evidence for its effectiveness in cases where other treatments have failed. However, there is vigorous debate as to whether ketamine effects are necessitated by psychoactive experiences experienced as part of acute effects of the infusion. Some studies in this area suggest that ketamine phenomenological effects, including dissociation, are positively related with outcome, whereas others have shown null effects. Previous work by our group in the CAP Ketamine Consortium data found that standard IV ketamine dose (.5mg/kg, over 40 minutes) reduced depression symptoms across time more than placebo. However, there were no dose-by-time effects on PTSD symptoms. In the present study, we examine whether acute dissociative symptoms experienced post-first treatment infusion moderate psychiatric symptom change over time.

Methods: In this secondary analysis (ClinicalTrials.gov identifier: NCT02655692), we examined a cohort of 158 Veterans or Service Members with PTSD who failed previous antidepressant treatment. Participants were randomized to 8 infusions administered twice weekly of intravenous placebo (n = 54), low dose (0.2 mg/kg; n = 52) or standard dose (0.5 mg/kg; n = 52) ketamine. Participants were assessed at baseline, during treatment, and for 4 weeks after their last infusion. We conducted a series of linear mixed models examining PCL-5 total score with main effects of time, dissociation symptoms reported by patient 30 minutes post-first-infusion, treatment group, and all two- and three-way interactions between these variables. We hypothesized that greater dissociation as measured by the CADSS would be associated with greater symptom reduction on the PCL-5 for the ketamine – but not placebo – groups.

Results: We found a statistically significant effect of visit that showed that self-reported PTSD symptoms symptom decreased over time (F(3, 150) = 18.57, p < 0.001). There was also a visit-by-dissociation interaction, whereby greater dissociation was associated with decreased PCL symptoms over time (F(3,150 = 3.52, p = 0.017). This relationship did not depend on treatment group, with a three-way interaction between visit, dissociation, and treatment being non-significant (p > 0.05).

Conclusions: In support of our hypotheses, experiences of dissociation during first ketamine infusion were associated with reduced PTSD symptoms over time. However, contrary to our hypotheses, the effect of dissociation on PTSD symptoms over time did not vary as a function of treatment. Across all treatment groups, dissociation was associated with PTSD symptom change. Subjective experience during the infusion may therefore be an important factor in reducing PTSD symptoms. Findings may also indicate magnification of subjective effects that are associated with greater placebo response. In other words, participant expectations of the dissociation-benefit relationship may have influenced expectation of receiving active drug. As this study was not designed to test and measure the effects of dissociative effects of a fixed dose, we cannot determine the cause of this relationship. Further, with the CADSS originally developed to assess and test dissociation in context of dissociative identity disorder, other measures assessing phenomenological effects are needed to better capture transient state dissociation during ketamine infusion and determine their benefit for treatment outcomes.

Keywords: IV ketamine, PTSD, Dissociation, subjective effects.

Disclosure: Nothing to disclose.

P685. Insula reactivity to infant cries mediates maternal PTSD and mother-infant connection in trauma-exposed women

Rebecca Lipschutz, Kristina Dahlgren, Elizabeth McAfee, Justin Santos, James Rilling, Patricia A. Brennan, Vasiliki Michopoulos, Jennifer Stevens, Abigail Powers

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 (e.g., crying). Emerging evidence has shown links between maternal stress and trauma with neural reactivity to infant cry sounds, however only one study (n = 11) has examined associations between maternal PTSD and brain activity to child-related stimuli. Studies in healthy postpartum women have found that the circuits involved in emotion regulation and threat sensitivity (i.e. amygdala, insula, dorsal ACC) are also activated in response to infant cries and are further associated with maternal-infant bonding and positive maternal behaviors. Parent-child mutuality, characterized by synchronous dyadic interactions and emotional connection, is important for children’s positive socioemotional development. Less is known about how maternal brain reactivity to infant cues relates to dyadic parent-child measures, such as mutuality, particularly in the context of maternal PTSD. Taken together, understanding brain activation patterns related to maternal PTSD symptoms will help elucidate underlying mechanisms that contribute to disrupted parent-child relationships. This study examined associations between maternal PTSD symptoms, brain activation in response to infant distress (infant cry sounds), and parent-child mutuality in postpartum women.

Methods: The current study examined how maternal PTSD symptoms related to maternal neural reactivity and parent-child mutuality in an ongoing study of n = 72 mother-infant dyads. All mothers endorsed at least one Criterion A trauma to be eligible for the study M(SD) = 5.5(2.8), range = 1–15. PTSD symptoms were assessed with the PTSD Checklist for DSM-5 (PCL-5; M(SD) = 15.5(13.3), range = 0 – 55). At 8–10 weeks postpartum mother-infant interactions during a caregiving task were videotaped and dyadic behavior was coded for mutuality. Mutuality measures the degree to which the interaction is characterized by synchrony, comfort, and mutual pleasure. Mothers attended a fMRI session between 8–12 weeks postpartum and completed an infant cry task that included recordings of their own infant’s cries, unknown infant cries, and control stimuli matched for auditory features. Data were preprocessed using fMRIprep 21.0.4 and analyzed in SPM12. Whole-brain analyses corrected for multiple comparisons were used to detect neural activity during the infant cry task (own infant cry > matched control sound). Region-of-interest (ROI) analyses tested correlations between mean activation values in the amygdala, insula, and anterior cingulate cortex (ACC) with maternal PTSD symptoms and mutuality. Post hoc mediation analyses examined indirect effects using bootstrapping (5000 samples) and bias-corrected 95% confidence intervals.

Results: 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 lower neural reactivity to own infant cry (compared to matched control sounds) in the right amygdala (r = −.30, p = .010), right insula (r = −.30, p = .012), and subgenual ACC (r = −.46, p < .001). Among these regions, right insula reactivity was positively associated with mother-infant mutuality behaviors in the caregiving interaction, r = .30, p = .020. Finally, mediation analyses showed a significant indirect association from maternal PTSD to mutuality through right insula reactivity, 95% bcCI [-0.022, −0.001].

Conclusions: Trauma-exposed mothers with higher PTSD symptoms showed lower reactivity to their own infant cries in salience network regions including the amygdala and insula as well as the subgenual ACC. Mediation results suggest a potential pathway by which maternal PTSD influences mother-infant relationships, through decreased insula reactivity to infant distress. These regions are important for mothers’ threat detection, emotion regulation, empathy, and responsive and sensitive caregiving behaviors. Lower reactivity to infant cry sounds in the insula was in turn linked with lower mother-infant mutuality, suggesting that lower engagement of the salience network or orienting response to infant distress cues may negatively impact the mutual connectedness of mother-infant dyads. Notably, lower engagement of salience-related brain regions differs from typical amygdala hyperreactivity patterns observed in PTSD populations. However healthy postpartum women also exhibit higher salience network reactivity to their own infant cues, suggesting that this pattern is disrupted in the context of PTSD, with important implications for the mother-infant relationship. These findings extend prior work in healthy postpartum women and provide novel evidence for neural mechanisms underlying maternal caregiving in postpartum women exposed to trauma.

Keywords: Posttraumatic stress disorder (PTSD), Postpartum mental health, fMRI, Parent - child dyads, Neural mechanism.

Disclosure: Nothing to disclose.

P686. Poster withdrawnP687. Somatomotor network topology moderates relationship between racial discrimination and ptsd re-experiencing

Aziz Elbasheir, Leland Fleming, Nathaniel Harnett, Timothy Ely, Negar Fani

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

Background: Racial discrimination (RD) is a pervasive stressor that has been consistently linked to elevated risk for post-traumatic stressor disorder (PTSD), with evidence showing that RD is not only associated with greater symptom severity but also predicts PTSD onset and maintenance. Research on PTSD has demonstrated that trauma-related symptoms are associated with alterations in the intrinsic organization of large-scale emotion regulatory resting state (rs-) functional networks such as the default mode network and salience network. More recently, evidence has suggested that PTSD is also characterized by disruptions in networks that support mind-body integration such as interoceptive and somatomotor networks. Emerging evidence suggests that RD may similarly impact these same networks, and these effects may be related to trauma symptomology including symptoms of derealization. However, to date, most neuroimaging studies of RD have focused on region-specific or circuit level changes and few have examined how RD influences broader organization at the level of rs-functional networks. For example, metrics such as network modularity reflect the degree to which rs-functional networks maintain cohesive within network communication while limiting extraneous connections with other networks, a property critical for emotion regulation. Other metrics, such as clustering coefficient, measure how brain regions form tightly knit neighborhoods, supporting specialized/local information processing. As such, the current study examined whether alterations in functional topology of large-scale rs-functional networks (modularity and clustering coefficient) influence the relationship between RD and PTSD symptoms with the objective of identifying network-level neural mechanisms that may link RD and PTSD symptomology.

Methods: Ninety adult Black American women (age range: 18–62) were recruited from the Grady Trauma Project, a long-standing trauma study. Participants completed measures of RD and PTSD symptom severity. RD was assessed using the Experiences of Discrimination Scale and PTSD symptoms were measured with the PTSD Symptom Scale, including total severity and symptom clusters of hypervigilance, avoidance and re-experiencing. rs-functional MRI data were collected and parcellated using the Yeo seven-network atlas. Modularity and clustering coefficient values were then extracted for each network to quantify intrinsic network organization. Bivariate correlations were first conducted to examine associations between RD and network segregation metrics with a Bonferroni corrected p value of p = .007. For any significant networks, follow up partial correlation analysis was conducted controlling for age. For networks showing significant associations with RD, modularity and clustering coefficient values were subsequently entered into moderation models to assess whether network topology moderates the relationship between RD and PTSD symptoms.

Results: Greater RD was associated with reduced segregation of the somatomotor network (SMN), reflected in lower modularity (r = −.28, p = .006). Significance remained after controlling for age (r = −.249, p = .02). RD was also significantly associated with lower SMN clustering coefficient (r = −.31, p = .003) and remained significant after controlling for age (r = −.229, p = .03). Our moderation analysis revealed that SMN clustering coefficient positively moderated the association between RD and PTSD re-experiencing, such that the effect of RD on PTSD re-experiencing was stronger at lower (≤.48) [B = .20, CI (.25, 1.05), t = 3.22, p = .002] and mid-range (.54) [B = .15, CI (.03, .65), t = 2.25, p = .03] SMN clustering values.

Conclusions: These findings suggest that RD is associated with altered SMN organization. Moreover, reduced SMN clustering strengthened the impact of RD on PTSD intrusions, further highlighting the role of SMN dysfunction in PTSD re-experiencing. The current findings align well with a growing understanding of the SMN’s role in integrating sensorimotor information with affective states and how changes in SMN function may heighten bodily re-experiencing of trauma memories; thus, increasing risk for intrusive memories. As such, these results demonstrate how changes in network-level properties may be a potential mechanism linking RD and greater vulnerability to PTSD re-experiencing in Black American populations.

Keywords: Posttraumatic stress disorder (PTSD), functional neuroimaging, Racism and discrimination stress.

Disclosure: Nothing to disclose.

P688. Trauma-related differences in cortical glutamate levels during emotional cognitive control in adolescents

John France, Dalal Khatib, Shaurel Valbrun, Sattvik Basarkod, William Davie, Vaibhav Diwadkar, Jeffrey Stanley, Tanja Jovanovic

Wayne State University School of Medicine, Harper Woods, Michigan, United States

Background: Childhood trauma exposure (TE) may heighten bottom-up negative emotion processing, thereby interfering with top-down cognitive control, and ultimately increase risk for pathological fear and worry including anxiety disorders and posttraumatic stress disorder (PTSD). Cognitive control, including inhibitory control, is facilitated in part by glutamatergic excitatory neurotransmission within the dorsal anterior cingulate cortex (dACC). Using functional magnetic resonance spectroscopy (1H fMRS), we investigated dynamic changes in dACC glutamate (Glu) levels to assess the impact of negative emotion processing on glutamatergic neural mechanisms supporting inhibitory control in TE-adolescents. We hypothesized that adolescents with greater TE, relative to adolescents with less TE, would demonstrate significantly lower dACC Glu levels during inhibitory control performed with emotional stimuli, but not with non-emotional stimuli. We further hypothesized that dACC Glu levels during inhibitory control performed with emotional stimuli would be negatively associated with anxiety and PTSD symptoms.

Methods: A median-split based on TE was applied to 50 adolescents generating two TE-Groups – the Higher TE-Group: 52% female, 13.5 ± 1.5yrs; Mtrauma = 6 ± 1events; and Lower TE-Group: 48% female, 13.0 ± 1.5yrs; Mtrauma = 3 ± 1events. 1H fMRS from the dACC was acquired during a 2 × 2 block design, visually guided inhibitory motor control task requiring participant finger tap responses to stimuli under two Response Modes, “Non-Selective” (100% response rate, involving only motor control) and “Selective” (80% response rate while withholding `prepotent` responses on 20% of trials, involving motor and inhibitory control). Both modes were executed with two Stimuli Conditions, “Squares” (non-emotional red and green square stimuli) and “Faces” (negative emotional face stimuli). Using generalized estimating equations (GEE), a two-step statistical approach was employed. First, Glu modulation (percent change in Glu-to-total signal ratio relative to the flashing checkerboard control) was tested across TE-Group, Stimuli Condition, and their interaction, averaging across Response Modes, to examine whether TE-group differences were specific to emotional stimuli. Second, within each Stimuli Condition, Glu modulation was subsequently tested across Response Modes by TE-Group to determine whether observed TE-Group differences were specific to inhibitory control. Task performance was assessed with response time, hit rate (for Non-Selective responding) and D-prime (for Selective responding). Associations between Glu modulation during the Faces Stimuli Condition and anxiety and PTSD symptoms were also assessed. Across all analyses, statistical significance was determined at α = 0.05. Assuming a small-to-moderate effect size (Cohen’s d = 0.25) and α = 0.05, power analysis indicated 80% power to detect a between-group difference.

Results: There was a 2-way interaction of TE-Group x Stimuli Condition (χ 2 = 4.66, p = 0.031), averaged across Response Modes. Post-hoc tests revealed significantly lower Glu modulation during Faces in Higher TE vs Lower TE (z = −2.28, p = .023). Within Faces, the main effect of Response Mode was significant, suggesting the higher TE-group demonstrated significantly less Glu modulation on average across Non-Selective and Selective responding (χ 2 = 4.33, p = 0.037). There was no significant TE-Group x Stimuli Condition interaction when testing response time (χ 2 = 2.90, p = 0.089). No significant effects were observed for hit rate and D-prime (ps > 0.05). No associations were observed between Glu modulation during the Faces Stimuli Condition and anxiety and PTSD symptoms (ps > 0.05).

Conclusions: We demonstrate for the first time distinct task related changes in dACC Glu levels of TE adolescents during an inhibitory motor control task performed with emotional stimuli. Specifically, the higher TE-Group’s Glu levels were significantly lower during the Faces condition compared to the lower TE-Group, independent of motor or inhibitory control processes). This result potentially supports the framework that negative emotion processing interferes with neural mechanisms related to cognitive control performance in TE-adolescents. Further research is needed to clarify if TE-related differences in dACC Glu modulation reflect a general sensitivity to emotional stimuli and to test the specificity for emotional valence of such an effect.

Keywords: Childhood trauma, Inhibitory control, excitatory/inhibitory balance, functional magnetic resonance spectroscopy.

Disclosure: Nothing to disclose.

P689. Dreaming the trauma: neural correlates of the sensory and emotional intensity of trauma-related nightmares

Isabelle Rosso, Kevin Clancy, Meaghan McKenna, Yara Pollmann

McLean Hospital/Harvard Medical School, Belmont, Massachusetts, United States

Background: Distressing dreams about a traumatic experience are prevalent among trauma-exposed individuals and contribute to the persistence of posttraumatic stress symptoms and poorer clinical outcomes. Like other intrusion symptoms, trauma nightmares are marked by sensory imagery and intense emotional content that often contribute to a vivid sense of reliving the trauma. These phenomenological features may share neural substrates with both daytime intrusive memories and idiopathic nightmares. Prior work has implicated anterior and posterior hippocampal networks: resting-state connectivity between the anterior hippocampus (aHPC) and amygdala (Amyg) has been linked to heightened emotionality during sleep, while posterior hippocampus (pHPC)–visual cortex (VC) connectivity has been associated with sensory-rich episodic memory retrieval. Connectivity between the posterior hippocampus and visual cortex has been linked to sensory-rich episodic and trauma memory retrieval. This study examined intrinsic functional connectivity in these circuits in relation to trauma-related dream phenomenology in symptomatic posttraumatic stress adults. We also assessed the clinical significance of dream properties by examining their associations with clinician-assessed nightmare symptoms.

Methods: Participants were 66 trauma-exposed adults (54 female), who met criteria for posttraumatic stress disorder (PTSD; n = 56, 85%) or had subthreshold PTSD (n = 10, 15%) based on the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5). They completed 2 weeks of daily ecological momentary assessment (EMA), reporting on trauma-related dreams, including their emotional intensity, vividness, visual detail, reliving quality, and fragmentation. Participants then underwent resting-state functional magnetic resonance imaging. Pearson correlations tested associations between dream properties and functional connectivity of aHPC-Amyg and pHPC-VC. Bonferroni corrections were applied. Significant findings were followed by whole-brain seed-to-voxel analyses to determine their spatial specificity.

Results: No significant associations were observed between aHPC-Amyg connectivity and dream properties. In contrast, stronger pHPC-VC connectivity was significantly associated with greater visual details (r = 0.35, p = 0.003) and emotional intensity (r = 0.44, p = 0.0002) of trauma-related dreams. Whole-brain analyses confirmed these effects: higher visual detail ratings were associated with greater pHPC connectivity to a cluster spanning the bilateral parieto-occipital sulcus, cuneus, and lingual gyrus (peak = −14, −74, 28; t = 4.82; k = 1816; cluster-level pFWE < 0.001). Emotional intensity was associated with a broader cluster encompassing visual areas and extending to the retrosplenial cortex (peak = −12, −54, −2; t = 5.01; k = 2975; cluster-level pFWE < 0.001). Effects remained robust when controlling for age and sex, and were not moderated by PTSD diagnosis or symptom severity. Finally, clinician-rated nightmare symptom severity (CAPS-5 Item B2) was significantly correlated with visual detail (r = 0.45, p < 0.001) and emotional intensity (r = 0.36, p = 0.003) of nightmares reported during the EMA period. Additional associations were seen with vividness (r = 0.28, p = 0.025) and reliving (r = 0.27, p = 0.028), but these did not survive Bonferroni correction.

Conclusions: These findings identify pHPC–VC connectivity as a key neural correlate of the visual imagery and affective properties of trauma-related nightmares. As this circuit has been previously implicated in the vivid details of episodic memory and the reliving of trauma-related intrusive memories, this adds to accruing evidence that trauma reexperiencing symptoms may share a common neurobiological basis rooted in posterior hippocampal–visual circuitry, with potential implications for mechanistic models and treatment targets in PTSD.

Keywords: nightmares, Post Traumatic Stress Disorder, visual cor, Resting State Functional Connectivity, visual cortex.

Disclosure: Nothing to disclose.

P690. HIV and PTSD status differentially modulate estrogen-induced gene expression and inflammatory responses in women’s immune cells

Amanda Arnold, Susie Turkson, Paul Howell, Gretchen Neigh, Vasiliki Michopoulos

Emory University, Atlanta, Georgia, United States

Background: Women living with HIV (WLWH) disproportionately experience trauma throughout their lifetimes, with rates of posttraumatic stress disorder (PTSD) significantly higher than those observed in the general population. Both HIV infection and PTSD contribute to immune system dysregulation characterized by chronic inflammation and altered cytokine profiles. Estradiol (E2), the primary circulating estrogen, modulates immune function through estrogen receptors expressed throughout the immune system, influencing inflammatory responses and immune cell activation. Generally, physiological concentrations of E2 are associated with anti-inflammatory effects, including suppression of pro-inflammatory cytokines such as TNF-α, IL-1β, and IL-6. However, the effects of E2 on immune function may be significantly modified by chronic stressors such as HIV infection and PTSD. Understanding how HIV and PTSD status influence estrogen-mediated immune modulation is critical for optimizing treatment strategies for WLWH, particularly given that trauma history may determine whether hormone-immune interactions are protective or dysregulated. We hypothesized that WLWH and women with PTSD would show blunted estrogen responsiveness in immune cells, characterized by reduced E2-induced gene expression changes and enhanced inflammatory cytokine production.

Methods: Participants were women recruited from the Women’s Interagency HIV Study Combined Cohort Study (WLWH n = 50, women living without HIV (WLWoH) n = 23) with documented trauma history. PTSD status was assessed using structured clinical interviews using the Mini International Neuropsychiatric Interview (MINI) (PTSD+ n = 37, PTSD- n = 36). Peripheral blood mononuclear cells (PBMCs) were isolated and treated in vitro with E2 or vehicle control for 24 hours. Gene expression analysis was performed using NanoString technology with a custom panel targeting estrogen-responsive genes. Differential expression analysis identified genes with significant changes (p < 0.05) following E2 treatment, stratified by HIV and PTSD status. Culture supernatants were analyzed for cytokine concentrations (TNF-α, IFN-γ, IL-4, IL-6) using MesoScale Discovery multiplex immunoassay. Multiple linear regression models examined interactions between clinical status and gene expression changes in relation to cytokine responses.

Results: WLWH demonstrated a more robust transcriptional response with 16 E2-induced differentially expressed genes (DEGs), while WLWoH had only 4 DEGs. Women without PTSD showed 14 DEGs following E2 treatment, while women with PTSD only had 2 DEGs. Gene ontology enrichment analysis revealed significant enrichment of the TNF signaling pathway within DEGs stratified by HIV status (FDR = 5.30E-05) and DEGs stratified by PTSD status (FDR = 0.002), including genes ATF2, JUNB, PIK3CA, and PTGS2. Multiple linear regression analyses revealed that HIV status significantly modified the relationship between E2-induced gene expression and cytokine production. In WLWoH, greater E2-induced expression changes in PLCB2, LDHA, and PTGS2 were associated with TNF-α responses (HIV×gene interactions: PLCB2 β = 4.708, p = 0.021, R² = 0.346; LDHA β = −4.499, p = 0.015, R² = 0.09; PTGS2 β = −15.883, p = 0.014, R² = 0.2), while these relationships were absent in WLWH. This HIV-dependent disruption of E2-immune signaling extended across multiple inflammatory pathways: PTGS2 interactions were observed for IL-4 (β = −0.295, p = 0.017, R² = 0.156) and IL-6 (β = −948.825, p = 0.015, R² = 0.203), PLCB2 influenced IFN-γ (β = 17.561, p < 0.001, R² = 0.583) and IL-4 (β = 0.099, p = 0.020, R² = 0.156), and LDHA affected IL-8 responses (β = −2389.64, p < 0.001, R² = 0.352), with stronger gene-cytokine associations consistently observed in WLWoH. For PTSD status, the relationship between NAB2 expression changes and cytokine responses was significant only in women with PTSD for both TNF-α (β = −4.116, p = 0.006, R² = 0.439) and IFN-γ (β = −13.078, p < 0.001, R² = 0.519), indicating trauma-specific alterations in E2-immune signaling pathways.

Conclusions: The current study reveals that HIV and PTSD influence how women's immune cells respond to E2 treatment at both transcriptional and functional levels. HIV infection may enhance E2 immune responsiveness while redirecting responses toward inflammatory pathways, whereas PTSD dampens overall immune responsiveness to E2. The significant interactions between clinical status and gene expression changes in relation to cytokine production suggest that HIV and PTSD modify the mechanisms by which E2 modulates inflammatory responses. These findings highlight the importance of considering both HIV and PTSD when developing hormone-based therapeutic interventions for women. The differential E2-immune interactions observed may contribute to health disparities in WLWH with trauma histories and underscore the need for personalized treatment approaches that account for individual trauma and infection status.

Keywords: PTSD, HIV, cytokines, immune function.

Disclosure: Nothing to disclose.

P691. Posttraumatic stress disorder has unique effects on the behavioral and multivariate neural correlates of contextual fear learning

John Leri, Samuel Cooper, Joseph Dunsmoor, Zachary Stowe, Joshua Cisler

The University of Texas at Austin, Austin, Texas, United States

Background: Although posttraumatic stress disorder (PTSD) has been linked with altered fear learning processes, it is unclear if the effects of PTSD are distinct from the impacts of interpersonal violence exposure or anxiety-related pathology. The current study attempted to isolate the effects of PTSD, relative to interpersonal violence exposure and anxiety disorders, on behavioral and neural indices of fear acquisition during a contextual fear conditioning task. To do so, this study used a contextual conditioning task in which three conditioned stimuli (CS) were all paired with shock across multiple contexts. This experimental design extends extant work, which has traditionally explored the effects of PTSD on fear learning using clearly differentiated CS. As such, the current study sought to understand if PTSD has unique effects of contextual fear learning when all CS were linked with threat.

Methods: Clinical interviews were used to phenotype 120 adults (N = 120, M age = 29.43, SD age = 8.17, 66% female) for lifetime exposure to interpersonal violence (National Trauma Assessment), post-traumatic stress disorder (Clinical Administered PTSD Scale for DSM-5), and anxiety-related psychiatric disorders (Diagnostic Interview for Anxiety, Mood, and Related Neuropsychiatric Disorders). Participants completed a contextual fear conditioning task while in an MRI. During conditioning, participants were exposed to three CS (i.e., images of human faces with a neutral facial expression) that were paired with shock at a reinforcement schedule which uniquely identified four visually discernable contexts (i.e., contexts were labeled with colors and names; two contexts were high-threat and two contexts were low-threat). Throughout the conditioning task, participants provided trial-by-trial shock predictions and intermittent fear expectancy ratings. Previously validated support vector machine classification models were used to quantify the prospective neural reinstatement of the unconditioned stimulus (i.e., shock) in response to each CS presentation within six canonical fear networks (amygdala, hippocampus, salience-, medial cingulate cortex-, medial prefrontal cortex-, and inferior frontal-networks). Linear mixed effect models were used to determine whether healthy controls (n = 34), anxiety diagnosed controls (n = 19), or trauma exposed controls (n = 46) differed from participants with PTSD (n = 21) on behavioral and neural indices of fear acquisition. Linear mixed effect models were also used to determine if the degree of neural shock reinstatement in functional networks was associated with behavioral indices of fear acquisition.

Results: PTSD participants were more likely to predict shock on a trial-by-trial basis. Specifically, they only modulated shock predictions based on the CS threat (t = 8.31, p = .001) and not based on contextual threat cues (t = −0.58, p = .559), resulting in higher overall threat predictions. PTSD participants provided higher shock likelihood ratings (i.e., fear expectancies) than control participants (all group comparison p-values < = .064) across contexts. Neural reinstatement of shock in the salience network discriminated between high- and low-threat CS less among PTSD participants compared to healthy controls (t = 4.11, p < .001). The neural reinstatement of shock in the medial prefrontal cortex (mPFC) was more strongly associated with shock likelihood estimates in healthy controls compared to participants with PTSD (t = 2.98, p = .003).

Conclusions: This study provided insight into the behavioral and neural correlates of fear learning among individuals with PTSD. Participants with PTSD exhibited greater behavioral indices of fear compared to all other participants, in part due to an inability to integrate contextual threat cues. In parallel, the fMRI analyses found that PTSD participants show less neural discrimination between threat cues in the salience network and less correspondence between prospective neural representation of threat and shock likelihood estimates in the mPFC. These results indicate that PTSD interferes with the ability to discriminate between high- and low-threat CS by inflating the threat value of low-threat CS. Taken together, these findings highlight unique effects of PTSD on both behavioral and neural indices of fear learning that are not attributable to interpersonal violence exposure or anxiety-related pathology alone.

Keywords: Post Traumatic Stress Disorder, Fear conditioning, Stress and Trauma.

Disclosure: Nothing to disclose.

P692. Functional gains and mood improvement in PTSD: results from IMPACT-1, a randomized controlled phase 2 clinical trial of TSND-201 (Methylone)

Amanda Jones, Jennifer Warner-Schmidt, Hannah Kwak, Carolyn Macleod, Stephanie Galinos, Terence Ching, Martin Stogniew, Blake Mandell, Murray Stein, Benjamin Kelmendi

Transcend Therapeutics, New York, New York, United States

Background: Post-traumatic stress disorder (PTSD) is a chronic and debilitating psychiatric condition that can develop after exposure to a traumatic event and is associated with significant impairments in daily functioning, relationships, and quality of life. Depression is one of the most common and clinically significant comorbidities in individuals with PTSD. It is estimated that 30–50% of individuals with PTSD also meet criteria for major depressive disorder (MDD). The co-occurrence of PTSD and MDD is associated with worse functional outcomes, increased symptom burden, and elevated suicide risk. Moreover, there is overlap in the neural substrates and neurocircuitry that underlie the pathophysiology and treatment of PTSD and MDD. Available treatments also overlap, since the only FDA-approved pharmacotherapies to treat PTSD are SSRI antidepressants (sertraline and paroxetine).

TSND-201 (methylone) is a highly selective, rapid-acting neuroplastogen that enhances neuroplasticity through the release of serotonin, norepinephrine, and dopamine, and without direct agonist/antagonist activity at serotonin 2A (5HT2A) receptors. Preclinical studies have shown rapid, robust, and long-lasting benefit in behavioral tasks associated with PTSD, depression, and anxiety symptoms. In clinical trials, TSND-201 is non-hallucinogenic and has been generally safe and well-tolerated. Most recently, TSND-201’s rapid, robust, and durable efficacy was demonstrated in a randomized, placebo-controlled Phase 2 study in individuals with severe PTSD (IMPACT-1, ClinicalTrials.gov Identifier: NCT05741710).

Methods: In this 10-week study, participants were randomized, received TSND-201 or placebo once per week for four-weeks without psychotherapy, and were followed until the end of study (Day 64). Data were collected 2 days after each dosing session (Days 3, 10, 17, and 24) and during the follow-up period (Days 29, 36, 43, and 64). The mean (SD) baseline Clinician-Administered PTSD Scales for DSM-5 (CAPS-5) total score was 45.8 (7.11) in the TSND-201 group and 46.0 (5.42) in the placebo group, indicating severe PTSD. We have previously reported that TSND-201 demonstrated statistically significant greater improvement from baseline to Day 64 in CAPS-5 total score (LS mean [SE] change, −23.3 [2.84]) compared with placebo (change, −13.6 [2.95]), with a LS mean difference of −9.64 (p = 0.011), meeting the primary endpoint for IMPACT-1. Based on these results, in July 2025, TSND-201 was granted Breakthrough Therapy Designation by the FDA, underscoring its potential superiority over currently available treatments.

Here we report additional endpoints from IMPACT-1 including improvement in depression symptoms among participants with co-morbid MDD assessed using the MADRS (>20 at baseline) and functional improvements across key domains assessed using the Sheehan Disability Scale (SDS).

Results: Depression Outcomes: Baseline scores (SE) on the MADRS were 32.3 (1.34) and 33.6 (5.96) in the TSND-201 and placebo groups, respectively. At the end of study, there was a −13.9 (2.51) point change from baseline in TSND-201 treated participants, compared with −7.7 (2.51) points in the placebo group, resulting in a significant LS mean difference of −6.21 (90% CI, −12.14 to −0.27, p = 0.043). The improvement in depression symptoms was rapid. On Day 10, TSND-201 showed a −10.2 (1.86) point change from baseline compared with −5.8 (1.89) in the placebo treated group, and a significant LS mean difference of −4.46 points (90% CI, −8.89, −0.03, p = 0.049).

Functional Outcomes: Functional improvements assessed by the SDS total score were greater in the TSND-201 group compared with placebo. At baseline, the mean (SE) SDS total score were 21.3 (1.26) and 21.2 (1.19) in the TSND-201 and placebo groups, respectively. By the end of the study, treatment with TSND-201 resulted in −8.3 (1.71) point improvement, compared to −3.6 (1.78) points in placebo (LS mean difference of 4.72 [90% CI: −8.88 to 0.61; p = 0.030]). Treatment with TSND-201 also improved productivity, by reducing both the number of days of lost per week (−1.4 vs. −0.7 days) and days unproductive per week (−2.6 days vs. −0.9 days) at the end of study.

Safety: SND-201 was well-tolerated with no serious adverse events attributed to study drug. Treatment-emergent adverse events were primarily mild-to-moderate and transient, occurring on the day of dosing.

Conclusions: TSND-201 previously demonstrated significant clinical benefit in patients with PTSD evidenced by improvements on the CAPS-5, the gold-standard measure of PTSD symptom severity. Here, results show that it also led to meaningful reductions in depressive symptoms measured by the MADRS, along with improvements in overall functioning. Together this suggests TSND-201 may have broader effects beyond core PTSD symptoms, supporting its potential as a therapeutic option that addresses multiple domains of patient well-being and its future investigation in Phase 3 clinical trials.

Keywords: TSND-201, PTSD, Transcend Therapeutics, neuroplastogen, Phase 2.

Disclosure: Transcend Therapeutics, Employee, Self, Transcend Therapeutics, Stock/Equity - Privately Held Company, Self.

P693. Massed exposure therapy enhanced with MDMA for PTSD: preliminary effectiveness and impact on fear extinction in an open label pilot study

Jessica Maples-Keller, Seth Norrholm, Anna Wise, Boadie Dunlop, Barbara Rothbaum

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

Background: Two recent Phase III trials found support for MDMA-assisted therapy for PTSD (Mitchell et al., 2023). The therapy model used in these trials has not previously been tested as a standalone therapy and is not an existing evidence based PTSD therapy. Prolonged Exposure (PE) is a gold standard PTSD intervention. Translational research in rodent and healthy humans suggests that MDMA enhances fear extinction retention, the foundation of Prolonged Exposure (PE) therapy, a first line PTSD therapy, suggesting this may be an optimal intervention to combine with MDMA. Research also suggests that psychedelics such as MDMA may reopen a social reward learning critical period for a two-week period (Nardou et al., 2023) and induce neuroplasticity, as such MDMA’s impact may be optimized using a massed therapy approach in which treatment is provided daily for a two-week period.

Methods: We are conducting an open-label, one-arm pilot trial of a two-week intensive PE protocol for treatment of PTSD, with the addition of a 100 mg MDMA dose on Day 2, in which two imaginal exposure sessions are conducted during the MDMA session. Primary outcome is PTSD symptom reduction as assessed by independent raters via the Clinician-Administered PTSD Scale for DSM-5-Revised (CAPS-5-R) from pre-treatment compared to one month follow up. Participants complete an experimental fear potentiated startle paradigm of fear extinction and extinction recall, with a pretreatment (i.e., fear acquisition [Day 1]), extinction (Day 3), and extinction recall (Day 5) and post-treatment assessment (i.e., fear acquisition and extinction [Day 9] and extinction recall [Day 10/final day of treatment]).

Results: Currently 12 participants have been enrolled (N = 12; 41.7% male; 41.7% civilian and 58.3% veteran). All participants have completed all study visits and there have been no serious adverse events. In preliminary analyses, CAPS-5-R decreased significantly from pre (M = 75.50, SD = 14.16) to 1 month (M = 16.48, SD = 20.89) (t = 12.33, df = 11, p < .001). Preliminary results of fear potentiated startle suggest that participants successfully acquired fear potentiated startle and discriminated between the CS+ and CS- during acquisition at both pre and post-treatment time points. Next, participants showed within-session extinction of fear to the CS+ during extinction Training at pre and post-treatment. Then, in these preliminary analyses, participants demonstrate a trend towards a return of fear to the CS+ at Extinction Recall at both pre (trend p = 0.07) and post-treatment (p = 0.009), and there is a trend towards less extinction recall at post-treatment compared to pre-treatment.

Conclusions: MDMA enhanced massed exposure therapy for PTSD shows preliminary safety, feasibility, and efficacy, with a large decrease in CAPS-5 symptoms. Preliminary analyses of fear potentiated startle data indicate a trend towards results in the hypothesized direction, such that participants show enhanced extinction recall after completing treatment, suggesting preliminary support for translating MDMA’s impact on extinction within clinical treatment. Enrollment is ongoing and analyses will be completed in the entire sample by December 2025.

Keywords: Psychedelics, MDMA, PTSD, Extinction, exposure therapy.

Disclosure: COMPASS Pathways, Consultant, Self.

P694. FKBP5 risk alleles diminish the protective effect of perceived internal, but not socioeconomic resources, on posttraumatic symptoms after recent trauma

Kate Webb, Taylor Jackson, Anthony Zannas, Ying Zhao, Nathaniel Harnett, Jennifer Stevens, Ronald Kessler, Karestan Koenen, Kerry Ressler, Samuel McLean, Sarah Linnstaedt

Duke University Medical Center, Durham, North Carolina, United States

Background: Individual susceptibility to adverse trauma outcomes is the result of a complex interplay between environmental and genetic risk factors. Prior work indicates that single-nucleotide polymorphisms (SNPs) at the stress-related FKBP5 gene are associated with increased risk of posttraumatic stress disorder (PTSD), depression, and anxiety. Gene-by-environment studies reveal FKBP5 risk allele carriers who have experienced prior adversity may be more likely to develop these disorders following a new trauma compared to their non-carrier counterparts. Protective factors, such as greater perceived ability to cope with stress (i.e., internal resources) and socioeconomic resources, which independently promote psychological recovery after trauma, may also operate differently depending on FKBP5 variant status. Here, we investigated whether interactions between FKBP5 polymorphisms, perceived internal resources, and socioeconomic resources are associated with PTSD, depression, or anxiety symptom development following a recent trauma.

Methods: As part of the AURORA study, participants (N = 1,851; mean age = 37; 49.4% non-Hispanic Black, 37.2% non-Hispanic White; 10.4% Hispanic) completed assessments of PTSD, depression, and anxiety symptoms at 4 timepoints post-trauma (2-weeks, 8-weeks, 3-months, 6-months). Participants provided a blood specimen used for genotyping at the time of recruitment. We examined a previously identified FKBP5 putative risk SNP rs1360780 (T allele). Individuals with 1 or 2 copies of the risk allele were categorized in the risk group (i.e., dominance model). Perceived internal resources were evaluated using the Connor-Davidson Resilience Scale. Annual household income was used as an indicator of individual socioeconomic resources. Linear mixed-effects models evaluated whether interactions between rs1360780 and self-reported internal resources or socioeconomic resources were prospectively associated with PTSD, depression, or anxiety symptoms after adjusting for sex, age, lifetime trauma, childhood maltreatment, prior PTSD symptom severity, and time.

Results: There was no effect of FKBP5 variant status on internal (t(1849) = 0.33, p = .743) or socioeconomic resources (t(1849) = 0.40, p = .691). In linear mixed-effects models, there was no main effect of rs1360780 on PTSD, depression, or anxiety, after adjusting for covariates (ps uncorrected > .05). Greater internal and socioeconomic resources both independently predicted less severe PTSD, depression, and anxiety symptoms (ps corrected < .05). There was a significant interaction between rs1360780 and internal resources, such that among risk carriers, internal resources were not protective for PTSD (interaction term: B = .25, t(1164) = 2.29, p = .022; risk carrier: B = −0.10; p = .14; non-risk carrier: B = −0.35; p < .001) or anxiety (interaction term: B = .07, t(1162) = 2.83, p = .005; risk carrier: B = −0.02; p = .27; non-risk carrier: B = −0.09; p < .001). For depression, the protective effect of internal resources was attenuated in risk carriers (interaction term: B = .18, t(1162) = 2.95, p = .003; risk carrier: B = −0.12; p < .001; non-risk carrier: B = −0.30; p < .001). This interaction pattern was not observed for individual socioeconomic resources (ps uncorrected > .05).

Conclusions: Our findings identified novel gene-by-environment interactions between rs1360780 and two well-established individual protective factors. While socioeconomic resources were beneficial regardless of FKBP5 variant status, the effectiveness of perceived internal resources was diminished among risk allele carriers. The effectiveness of protective factors often depends on situational demands (e.g., responding to trauma); however, these findings suggest that genetic differences may also play a critical role in how these resources operate. Future research on protective factors and resilience after trauma should examine gene-by-environment interactions in order to better inform targeted prevention and intervention strategies for at-risk individuals.

Keywords: Risk and Resilience, genetics, PTSD, Stress and Trauma.

Disclosure: Nothing to disclose.

P695. Virtual confrontation, real-life recovery: a feasibility pilot for a novel virtual reality exposure therapy in military personnel

Remco van Zijderveld, Elbert Geuze, Antoin de Weijer, Eric Vermetten, Bastiaan Bruinsma

Ministry of Defence, the Netherlands, Utrecht, Netherlands

Background: Military personnel with posttraumatic stress disorder (PTSD) respond worse to current trauma-focused treatments than the civilian population. This reduced treatment response could be ascribed to avoidance coping and behavior. A potential technique to mitigate this behavior during exposure therapy is Virtual Reality (VR). Consequently, a new VR trauma-focused exposure therapy, IMMERSE, has been developed. The IMMERSE treatment offers a multisensory exposure with a head mounted display. Several olfactory, tactile, and auditive stimuli are offered while viewing and discussing the patient’s traumatic imagery during exposure. These sensory elements can be adjusted to the right level of immersion and exposure to ensure that patients won’t be overwhelmed by the traumatic memory. Moreover, the patient is in direct contact with their therapist, who monitors eye-tracking and heart rate.

Methods: This VR exposure therapy was offered to Dutch military personnel and veterans with PTSD (n = 16). Simulator Sickness was measured to assess the acceptability of the IMMERSE treatment. Moreover, interviews were conducted with all participants after completing the treatment. Treatment satisfaction and feasibility were evaluated with a questionnaire called the IMMERSE-Q. To examine the preliminary effectiveness of the treatment, both self-report and clinician-rated measures of PTSD symptoms, depressive symptoms, and well-being were assessed.

Results: The IMMERSE treatment was considered safe and feasible, as treatment satisfaction was high and the burden of simulator sickness was absent. Furthermore, the IMMERSE treatment was accompanied by a reduction in PTSD and depressive symptoms and an increase in well-being.

Conclusions: These findings suggest that the IMMERSE treatment is acceptable and feasible in Dutch military personnel with PTSD. Aside from its potential as a stand-alone treatment, the IMMERSE treatment could also be implemented as a tool in current gold-standard therapies to enhance multisensory exposure.

Keywords: Adolescent PTSD, Veterans, Combat PTSD, virtual reality.

Disclosure: Nothing to disclose.

P696. Brexanolone infusion for PTSD: results of an open-label pilot trial

D. Jeffrey Newport, Erin Richardson, Madeleine Sheehan, Ambreen Rana, Fouzia Sheikh, Charles B. Nemeroff

The University of Texas At Austin, Austin, Texas, United States

Background: Posttraumatic stress disorder (PTSD) is associated with significant morbidity, and novel treatments are profoundly needed. While SSRIs have demonstrated efficacy for PTSD, they lack consistency in treating the various domains of PTSD symptoms (re-experiencing, avoidance, hyperarousal) and are less effective than evidence-based psychotherapies. However, psychotherapies for PTSD are hampered by low rates of delivery and frequent treatment discontinuation. Lengthy delays until symptom improvement (4–6 weeks for SSRIs; 8–12 weeks for psychotherapy) may lead patients to feel discouraged. Current treatments, while efficacious, rarely afford patients full remission from PTSD symptoms.

Rodent studies have demonstrated an association between lower central allopregnanolone concentration and deficits in fear extinction. Similarly, clinical studies have observed lower allopregnanolone cerebrospinal fluid (CSF) concentrations. In women with PTSD, allopregnanolone/DHEA ratios are inversely associated with the severity of re-experiencing symptoms and depression. It has been suggested that sex differences in allopregnanolone may play a role in higher rates of PTSD among women.

Consequently, we hypothesized that brexanolone, a proprietary analog of allopregnanolone and a positive allosteric modulator of GABAA receptors, may provide rapid relief of PTSD symptoms in adult females.

Methods: This open label study examines the efficacy and tolerability of brexanolone, infused intravenously over 24 hours up to 90 μg/kg/hr, for PTSD symptoms. Inclusion criteria are premenopausal women ages 18–50, primary DSM-5 diagnosis of PTSD per MINI International Neuropsychiatric Interview (MINI), and baseline PTSD Checklist for DSM-5 (PCL-5) score > 33. Exclusion criteria are lifetime histories of bipolar disorder or psychosis, substance use disorder within the past 6 months, active suicidality, pregnant, lactating, less than 6 months postpartum, unstable medical illness, renal disease, history of seizures, previous treatment and/or allergy to brexanolone, or any other neuroactive steroid GABAA receptor modulator, current treatment with benzodiazepines or CNS depressants, and initiation of any psychotropic agent within 14 days of screening. Follow-up evaluations, including PCL-5, MADRS, CSSRS, and SDS, were conducted during inpatient stay at 2, 4, 8, 12, 18, 24, and 30 hours after initiation of infusion, and at outpatient visits 1, 2, 4, 8, and 12 weeks after infusion.

Baseline assessment includes MINI diagnostic interview, PCL-5 and Montgomery-Asberg Depression Rating Scale (MADRS) to assess PTSD and depression symptom severity respectively, the Childhood Trauma Questionnaire (CTQ) and Life Experiences Survey (LES) to assess trauma history, Columbia Suicide Severity Rating Scale (CSSRS), and Sheehan Disability Scale (SDS)., as well as medical history, physical exam, lab studies (CBC, CMP, pregnancy test, TSH, UDS), and ECG.

Primary outcome is change from baseline in PCL-5 score at each follow-up evaluation. Secondary outcomes include portion no longer fulfilling DSM5 PTSD criteria, changes in PCL-5 subscale, MADRS, and SDS scores at each follow-up. Continuous outcomes were analyzed using analysis of variance with Tukey post hoc pairwise comparisons (baseline vs. each follow-up measure) and Bonferroni correction for multiple comparisons. Last observation was carried forward for those yet to complete the study.

Results: To date, 8 participants have enrolled in this trial and completed the inpatient brexanolone infusion. Five have completed the entire 12-week follow-up period; the other three remain active in the study with 1, 2, and 3 pending follow-up visits respectively. At baseline, mean scores were: PCL-5: 47.0 ± 7.0 (range: 33–51), MADRS: 22.4 ± 6.3 (range: 11–29), and SDS: 15.1 ± 9.8 (range: 0–24). PCL5 scores were significantly lower than baseline at week 2 (mean: 23.9, p < .04), week 8 (23.1, p < .03) and week 12 (mean: 23.3, p < .03). Whereas all 8 participants fulfilled PTSD diagnostic criteria at baseline, 6 of 8 no longer fulfilled PTSD criteria at post-infusion follow-up visits. PCL5 Cluster E symptom scores were significantly lower at weeks 2, 8, and 12 as well. SDS scores were significantly lower at weeks 1, 4, 8, and 12. Neither MADRS nor PCL-5 Cluster B, C, and D subscale scores were lower than baseline at any time. No elevations in CSSRS score were observed. No serious adverse events have been observed.

Conclusions: This pilot study provides preliminary evidence to indicate that brexanolone infusion may provide rapid and sustained relief from symptoms of PTSD, particularly Cluster E (hyperarousal symptoms), among premenopausal women. Future studies should incorporate randomized placebo-controlled designs and investigate utility and tolerability in other groups including men and postmenopausal women.

Keywords: PTSD, Neurosteroid, Women’s Mental Health, brexanolone.

Disclosure: Sage Therapeutics, Advisory Board, Self, Sage Therapeutics, Contracted Research, Self.

P697. The role of posttraumatic stress disorder symptoms in pain severity, interference, and variability after sexual assault: an ecological momentary assessment study

Nicole Short, Mattea Pezza, Lauren Reyes, Jenny Black, Karen Serrano, Samuel McLean

University of Nevada, Las Vegas, Las Vegas, Nevada, United States

Background: Pain is common after sexual assault and may be influenced by posttraumatic stress disorder (PTSD) symptoms. However, little research has tested these associations in the early aftermath of sexual assault to delineate the temporal precedence of PTSD symptoms prior to persistent pain development. The current study used ecological momentary assessment (EMA) collected in the context of a randomized clinical trial to test associations between PTSD symptoms and pain.

Methods: Hypotheses were tested among 64 women sexual assault survivors presenting for emergency care who were followed over six weeks with 4x/day EMA assessments. Dynamic structural equation modeling (DSEM) analyses were used to account for autocorrelations and the effects of time and condition.

Results: Within-persons elevations in PTSD were concurrently associated with highest pain severity and pain interference. Effect sizes for these models were large at the within-level and small at the between-level. Heterogeneous variable models revealed that elevated between-persons PTSD was associated with increased variability in all pain outcomes. Effects were medium to large in size at the between-level. Vector autoregressive models indicated that PTSD prospectively predicted highest pain severity, while highest pain severity and pain interference prospectively predicted PTSD.

Conclusions: Findings add to the body of research suggesting that PTSD symptoms may influence the development of persistent pain after sexual assault but add specificity to these models by indicating that PTSD symptoms may uniquely drive spikes in pain and pain variability. Simultaneously, highest pain severity and pain interference may drive PTSD symptoms in the early aftermath of sexual assault. Future research should test whether early interventions after sexual assault could interrupt the vicious cycle of PTSD symptoms and pain.

Keywords: PTSD, Pain, interpersonal violence, ecological momentary assessment.

Disclosure: Nothing to disclose.

P698. The effects of extended cannabis abstinence in posttraumatic stress disorder (PTSD)

Justyne Rodas, Maryam Sorkhou, Stefan Kloiber, Tony George, Ahmed Hassan

University of Toronto, Toronto, Canada

Background: This preliminary open-label study examined whether 12 weeks of cannabis abstinence was associated with posttraumatic stress disorder (PTSD) symptom improvement in people with comorbid PTSD and cannabis use disorder (CUD) (N = 21).

Methods: Abstinence was defined as a THC-COOH level < 50 ng/mL with no self-reported cannabis use. To support abstinence, participants received progress contingency reinforcement payments for successful abstinence at weeks 4, 8 and 12. PTSD symptoms were evaluated using the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5), capturing both total severity and symptom count, as well as scores across symptom clusters.

Results: Participants who achieved abstinence (n = 11) experienced significantly greater reductions in total PTSD symptom severity and symptom count relative to those who did not (n = 10). Time-by-group interactions revealed more pronounced improvements in avoidance, negative mood and cognition, and hyperarousal among abstainers. Re-experiencing symptoms improved across both groups over time, with no significant difference by abstinence status.

Conclusions: Our findings suggest that sustained cannabis abstinence is associated with significant reductions in PTSD symptom severity and frequency over 12 weeks. These findings challenge the notion that cannabis use alleviates distress or improves symptoms or functioning in PTSD. Thus, cannabis may produce limited clinical benefit and interfere with recovery in key domains of PTSD. This underscores the need to routinely assess cannabis use during PTSD treatment and to educate patients on the potential consequences of continued use. Larger randomized trials are warranted to replicate these findings and to investigate the neurobiological and behavioral mechanisms through which abstinence may promote clinical improvement in PTSD.

Keywords: PTSD, Cannabis Withdrawal, abstinence, Clinical practice.

Disclosure: Nothing to disclose.

P699. Sex differences in fear extinction learning in trauma-exposed individuals: evidence from a multi-hormone analysis

Andrew Toader, Christina Lee, Angel David Arellano Perez, Lucas Taylor, Natalie Wieczorek, Geetanjali Balaguru, Galle Lipshitz, Thomas Neylan, Sabra Inslicht

University of California, San Francisco, San Francisco, California, United States

Background: Women are nearly twice as likely as men to develop posttraumatic stress disorder (PTSD), and impaired extinction learning has been implicated as a mechanism underlying increased risk.

Sex hormones, especially estradiol, play a crucial role by regulating neural circuits involved in fear conditioning and extinction, notably within the amygdala, hippocampus, and prefrontal cortex. While estradiol receives much attention, other hormones—including progesterone and its metabolite allopregnanolone, and adrenal steroids like cortisol and DHEA—also influence fear memory through GABAergic signaling and modulation of stress reactivity. The interactions between these hormones have received relatively little attention in fear conditioning studies, particularly with trauma-exposed individuals. To address this gap, this study examined the combined effects of multiple hormones and their metabolites on fear memory strength in trauma-exposed men and women undergoing a three-phase fear conditioning experiment, aiming to uncover sex-specific extinction mechanisms and inform hormone-based biomarkers and treatments for PTSD.

Methods: A 2 × 2 (sex-by-PTSD status) between-subjects experiment assessed steroid hormone associations with fear extinction/retention using a conditioning paradigm in 83 trauma-exposed adults (final n = 57 after exclusion for sub-threshold acquisition). CS+ and CS– were colored circles; the US was an individually calibrated electric shock. Skin conductance response (SCR) to CS+ vs. CS– during acquisition, extinction, and retention was the primary measure. Blood samples at three time points enabled hormone/metabolite assays. PTSD severity was modeled continuously. Linear mixed-effects models tested hormone concentration effects on stimulus type by PTSD severity for extinction and retention, analyzing estradiol and cortisol individually and jointly. Principal component analysis (PCA) summarized broader hormonal/metabolite panels, modeling their effects on fear outcomes. Finally, relationships between acquisition-day hormones and later extinction retention were evaluated to distinguish encoding- vs. retrieval-phase hormone effects.

Results: Men and women differed in overall fear-learning outcomes, as male sex was associated with lower conditioned responses (β = –0.071, p = .013) and a weaker relationship between PTSD severity and fear responding (β = –0.075, p = .011) during extinction. During acquisition, hormone x PTSD interactions were absent in women; in men, higher DHEA predicted a stronger PTSD-conditioned response link (β = 0.42, p = .026). During extinction learning, women showed significant three-way interactions among stimulus type, PTSD severity, and estradiol (β = –0.43, p = .017) as well as cortisol (β = –0.32, p = .017), indicating that higher levels of these hormones were associated with smaller differential SCRs (i.e., better extinction), particularly in those with severe PTSD. However, in a joint model including both hormones, neither estradiol nor cortisol remained independently significant (E2 p = .07; cortisol p = .29), and no interactions were observed in men during extinction (all p > .05). In women, the first principal component from the multi-hormone PCA, which explained ~26% of variance and loaded positively on estrone, estradiol, progesterone, AMH, and cortisol and negatively on testosterone (with additional contributions from adrenal/progesterone metabolites), also showed a significant three-way interaction with PTSD severity and stimulus type during extinction (β = –0.16, p = .007). Model comparison indicated that this PCA-based factor improved prediction of extinction outcomes relative to estradiol alone (ΔAIC = 4.08). This PCA-based factor improved extinction prediction over estradiol alone (ΔAIC = 4.08). During extinction retention, women showed a significant three-way interaction between progesterone/estradiol ratio on the acquisition day, stimulus type, and PTSD severity (β = 0.250, p = .027), with no similar delayed hormone effects in men.

Conclusions: Our findings indicate sex differences in fear extinction trajectories that are impacted by hormones in a sex-specific manner. In women, higher estradiol, and to a lesser degree, higher cortisol, were associated with diminished conditioned fear responses during extinction, especially among those with more severe PTSD, indicating a protective effect. Men showed no significant modulation of extinction by these hormones. A composite multi-hormone factor (high estrogen, progesterone, AMH, and cortisol, with lower testosterone and elevations in various adrenal/steroid metabolites) predicted extinction learning in women better than any single hormone, and estradiol remained the strongest individual predictor. Hormones present during fear encoding (acquisition) influenced extinction retention in women, suggesting that timing of learning may affect endocrine effects on fear memory. These results suggest that broad endocrine profiles shape extinction processes in trauma-exposed women, whereas men’s trajectories appear less hormone-sensitive, supporting multi-hormone approaches for more personalized PTSD interventions. Further work is needed to elucidate possible synergistic interactions among multiple hormones impacting fear memory.

Keywords: Fear Conditioning and Extinction, Sex Differences, Hormones.

Disclosure: Nothing to disclose.

P700. Maternal PTSD symptoms and sensitivity during caregiving in early postpartum: the moderating role of resting and reactive RSA in a trauma-exposed sample

Abigail Powers, Rebecca Lipschutz, Elizabeth McAfee, Catherine Abrams, David O'Banion, Vasiliki Michopoulos, Patricia Brennan, Jennifer Stevens

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

Background: Impaired maternal sensitivity may be a risk pathway linking maternal posttraumatic stress symptoms (PTSS) to adverse child outcomes. Respiratory sinus arrhythmia (RSA), a psychophysiological marker of emotion dysregulation, may be a key factor in how PTSS influence maternal sensitivity. Yet, these associations remain untested in early infancy. Examining how PTSS influence maternal sensitivity within the context of psychophysiological functioning during a caregiving task that elicits stress in young infants will help to clarify physiological processes contributing to the intergenerational transmission of trauma-related stress. To address major gaps in research, the current study tested the moderating effects of both maternal resting RSA and RSA response to a caregiving task on the relation between maternal PTSS and maternal sensitivity in a sample of trauma-exposed mothers and their 6–10-week-old infants. We hypothesized that maternal PTSS would be related to lower maternal sensitivity in the context of 1) lower resting RSA and 2) greater RSA reactivity (evidenced as RSA withdrawal or a decrease in RSA).

Methods: Participants were recruited for an NIH-funded longitudinal study of trauma-exposed mothers and their infants across the first year of life via social media advertisements and from the obstetrics and gynecological clinic at an urban publicly funded hospital. Mothers were assessed for study eligibility by phone after giving birth and if they were within 8 weeks postpartum at time of consent. Inclusion criteria for mothers were aged 18+, trauma exposed (based on DSM-5 criterion A), able to provide informed consent, and able to undergo a magnetic resonance imaging (MRI) scan; infants had to be less than 8 weeks old, born at full term (>37 weeks), and free of congenital or heart conditions. Exclusion criteria included the presence of active psychotic symptoms. The diverse sample for this study included seventy-seven mother-infant dyads (maternal Mage = 30.06 years, 61.0% Black; infant Mage = 9.53 weeks) who were enrolled and completed the 6-to-10-week behavioral assessment and had coded parenting behavior at the time of analyses. In this sample, 57% of mothers were below the federal poverty line. Mothers reported on PTSS using the PTSD Checklist for DSM-5 and engaged in a caregiving task; maternal sensitivity was coded by independent coders masked to psychological symptoms of mothers. RSA was measured at rest and in response to the task. Generalized linear models for ordinal outcomes analyses examined the moderating effect of resting RSA and RSA reactivity on the association between PTSS and maternal sensitivity. Parity was included as a covariate in analyses.

Results: Overall, RSA decreased from the baseline, M(SD) = 5.27 (1.07), to the caregiving task, M(SD) = 4.62 (0.94), t(72) = 4.91, p < .001. The association between maternal PTSS and sensitivity was significantly moderated by resting RSA, (B(SE) = 0.03(0.01), p = .033; maternal PTSS was negatively associated with maternal sensitivity only among mothers with higher resting RSA (+1SD above mean), B(SE) = −0.05(0.02), p = .030. There were no significant associations between PTSS and maternal sensitivity at average or above average levels of resting RSA scores (ps > .31). All other main effects and predictors were non-significant (ps > .05). The association between maternal PTSS and sensitivity was significantly moderated by RSA reactivity, B(SE) = 0.03(0.01), p = .022; maternal PTSS was negatively associated with maternal sensitivity only among mothers with greater RSA withdrawal (-1SD below mean RSA reactivity scores), B(SE) = −0.06 (0.02), p = 0.021. There were no significant associations between PTSS and maternal sensitivity at average or above average levels of RSA response scores (ps > .28). All other main effects and predictors were non-significant (ps > .05).

Conclusions: Results showed that both maternal resting RSA and RSA reactivity interacted with maternal PTSS to predict maternal sensitivity. Contrary to our first hypothesis, higher levels of PTSS were associated with lower maternal sensitivity among mothers with higher resting RSA but not average or lower resting RSA. It is possible that higher resting RSA among those with higher PTSS reflects a tendency toward overregulation or disengagement; the presence of physiological overregulation could make it difficult to actively engage in caregiving behavior for mothers with PTSS and may signal a form of disengagement from oneself or the environment. Alternatively, in line with our second hypothesis, higher levels of PTSS were associated with lower maternal sensitivity among mothers with higher RSA reactivity (greater RSA withdrawal) in response to a caregiving task, but not for mothers with lower RSA reactivity (RSA augmentation). Significant downregulation of the parasympathetic nervous system (and sympathetic or ‘fight or flight’ activation) in response to caregiving may get in the way of a mother’s ability to appropriately respond to infant needs in the moment when PTSS are present, thus negatively affecting maternal sensitivity. Overall, our findings highlight autonomic regulation as an important effect modifier in the PTSS-impaired maternal sensitivity pathway in trauma-exposed mothers and their infants. A tendency toward autonomic overregulation and heightened RSA reactivity may serve as relevant factors influencing how PTSS leads to maladaptive parenting behavior in early postpartum.

Keywords: PTSD, maternal sensitivity, heart rate variability.

Disclosure: Nothing to disclose.

P701. Epigenetic age acceleration associates with treatment-resistant PTSD

Alicia Smith, Seyma Katrinli, Dasia Simpson, Sarah Pridgen, Mark Logue, Caroline Nievergelt, Monica Uddin, Jessica Maples-Keller, Philip Held, Barbara Rothbaum

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

Background: While the majority of people treated for PTSD using evidence-based treatments such as Prolonged Exposure (PE) and Cognitive Processing Therapy (CPT) report notable improvement in symptoms, ~30% do not improve even after trying multiple interventions. Early identification of factors that distinguish those that are treatment-resistant is critical for personalizing clinical care. Emerging research supports the association between posttraumatic stress disorder (PTSD) and epigenetic age acceleration. In this study, we examined whether age acceleration associates with treatment-resistant PTSD.

Methods: We leveraged data from the PGC-PTSD Epigenetics Workgroup comprising 1029 trauma-exposed controls, 788 participants with current PTSD, and 129 participants that failed to respond to treatment administered in an intensive outpatient program in which they received either daily PE or CPT for two weeks. Treatment resistance was defined as <10 points change in PCL scores from baseline to program completion. DNAm was assayed from blood samples collected pre-treatment with the MethylationEPIC v1 or v2 BeadChip and used to calculate epigenetic age acceleration for each participant as the residual between DNAm age and chronological age. Age acceleration was tested for association with PTSD group (i.e., controls versus current PTSD vs treatment resistant PTSD) using an ANOVA that controlled for sex, DNA methylation-derived smoking scores, and cell composition.

Results: Horvath (F = 15.7; p < .0001), PhenoAge (F = 41.1; p < .0001) and GrimAge (F = 190.0; p < .0001) acceleration all associated with PTSD group, with all measures of age acceleration showing the lowest levels in controls and the highest levels in those with treatment-resistant PTSD. Posthoc pairwise comparisons show that treatment resistant PTSD is significantly higher in those with treatment-resistant PTSD relative to those with current PTSD for Horvath (p = 0.018), PhenoAge (p = .038) and GrimAge (p < .0001) acceleration.

Conclusions: These data suggest that those with treatment-resistant PTSD have higher age acceleration when they initiate treatment. Examination of the PhenoAge clock, which captures DNAm-based estimates of cardiometabolic indicators of aging, and the GrimAge clock, which captures DNAm-based estimates of all-cause mortality, suggest that those that do not respond to PTSD treatment may have higher physical health comorbidities when they initiate treatment. Future studies should consider multiple morbidity as a predictor of PTSD treatment response.

Keywords: GrimAge acceleration, PTSD, Prolonged Exposure Therapy.

Disclosure: Nothing to disclose.

P702. Recent traumatic stress is associated with greater thickness of early visual cortex

Nathaniel Harnett, Kate Webb, Amy Knight, David Knight, Kerry Ressler, Isabelle Rosso

Harvard Medical School, McLean Hospital, Belmont, Massachusetts, United States

Background: Posttraumatic stress disorder (PTSD) is associated with altered neural function and structure, but limited work in humans has investigated potential acute effects of traumatic stress on neurobiology. Emerging evidence suggests traumatic stress induces acute changes in neural responses of early visual cortex that are linked to PTSD. Here, we investigated if recent traumatic exposure is associated with variability in early visual cortex structure using two independent datasets composed of recent trauma survivors (TE) and demographically similar non-trauma exposed (NTE) controls.

Methods: Seventy-eight participants were recruited across two samples (Discovery: n(TE) = 24, n(NTE) = 16; Replication: n(TE) = 20, n(NTE) = 18) and completed a T1-weighted anatomical MRI. We used standardized pipelines in FreeSurfer (7.2.0) to reconstruct the cortical surface and estimate cortical thickness within early ventral visual cortex (i.e., V1-V4). General linear models assessed group differences in cortical thickness with Bonferroni correction across region and hemisphere (critical p-value = 0.00625). Cohen’s d was used to estimate effect size. Sensitivity analyses additionally covaried for age, sex, and trauma load. Findings in the discovery sample guided confirmatory one-tailed hypothesis testing in the replication sample.

Results: In the discovery sample, the TE group showed greater thickness compared to the NTE group for left [t(38) = 3.12, p = 0.003, d = 1.01] and right [t(38) = 3.20, p = 0.002, d = 1.03] V1, which was robust to sensitivity analyses including covariates [Left: t(27) = 2.94, p = 0.007, d = 1.26, Right: t(27) = 3.47, p = 0.002, d = 1.48]. An uncorrected effect was also observed in V3 [t(38) = 2.23, p = 0.032, d = 0.72]. In the replication sample, analyses confirmed greater left V1 [t(36) = 2.22, p = 0.0163, d = 0.72] and V3 [t(36) = 2.24, p = 0.0155, d = 0.73] thickness in the TE group. No effect in right V1 was non-significant but directionally consistent [t(36) = 1.30, p = 0.1, d = 0.42].

Conclusions: The present findings suggest recent trauma exposure may remodel early visual cortex in the acute peritraumatic period, and provide novel insight into the unique neurobiological effects of trauma. Future longitudinal research is needed to determine if changes in visual cortex structure, over time, underly long-term development of PTSD symptoms.

Keywords: Trauma exposure, Visual cortex, PTSD, Structural MRI.

Disclosure: Nothing to disclose.

P703. Dissecting fear and emotional pain in PTSD: from symptom networks to neural signatures

Ziv Ben-Zion, Erin Z Basol, Alexander J Simon, Maayan Abargil, Katherine Samonek, Megan Patterson, Tobias R Spiller, Or Duek, Jakcob N Keynan, Roee Admon, Israel Liberzon, Arieh Shalev, Talma Hendler, Ifat Levy, Jutta Joorman, Dustin Scheinost, Ilan Harpaz-Rotem

Yale University School of Medicine, Haifa, Israel

Background: Posttraumatic stress disorder (PTSD) is a heterogeneous condition with diverse symptom presentations and emotional experiences. While fear is traditionally viewed as central, growing evidence highlights the role of non-fear-based emotions - such as sadness, guilt, and shame - collectively termed Emotional Pain. This study aimed to identify Emotional Pain and Fear-based PTSD symptom profiles and their neural correlates across two independent samples.

Methods: In Study 1 (n = 838), trauma-exposed individuals with probable PTSD completed the PTSD Checklist for DSM-5 (PCL-5) and subjective ratings of Fear and Emotional Pain. Item-level network analysis was conducted to identify central symptoms and relationships. In Study (n = 162), recent trauma survivors with high PTSD symptoms underwent resting-state and task-based functional MRI (fMRI) scans 1-month post-trauma, and completed follow-up clinical assessment at 14-months post-trauma. Connectome-based predictive modeling (CPM) was used to predict chronic symptom severity for Fear and Emotional Pain-based profiles, identified in Study 1.

Results: Emotional Pain was rated as more impairing than Fear by most participants (69%). Symptom networks showed distinct patterns: Fear was associated with flashbacks, nightmares, distressing memories, exaggerated startle, and external avoidance; Emotional Pain was linked to anhedonia, negative beliefs, negative emotions, sleep disturbance and emotional reactivity. CPM predicted chronic Fear-based symptom severity (rho = 0.228, p < 0.001), but not Emotional Pain (rho = 0.167, p = 0.055). Predictive features included connections across anterior default-mode, central executive, salience, motor-sensory and subcortical networks.

Conclusions: Emotional Pain and Fear may represent distinct PTSD dimensions. Disentangling their neural signatures may improve diagnostic precision and guide personalized, mechanism-based interventions for trauma-related psychopathology.

Keywords: PTSD, Fear, Emotional Pain, Network-Analysis, Subtyping.

Disclosure: Nothing to disclose.

P704. Identification of an intrusive-hypervigilant phenotype of posttraumatic stress symptoms with unique stress peptide and amygdala functional connectivity profiles

Kevin Clancy, Caitlin Ravichandran, Victor May, Sayamwong Hammack, William Carlezon, Kerry Ressler, Scott Rauch, Isabelle Rosso

McLean Hospital - Harvard Medical School, Belmont, Massachusetts, United States

Background: Posttraumatic stress disorder (PTSD) is a markedly heterogeneous psychiatric disorder. Beyond the categorical diagnosis of PTSD, a significant number of trauma-exposed individuals report clinically significant distress and impairment through the endorsement of PTSD symptoms that do not meet full DSM-5 diagnostic criteria. Combined, this clinical heterogeneity hinders efforts to identify reliable biological substrates that are needed to advance mechanism-based therapeutics of posttraumatic stress.

Recent work has identified a distinct clinical phenotype of posttraumatic stress that is marked by specific elevations in intrusive reexperiencing and hypervigilance symptoms. This clinical phenotype is characterized by impaired fear extinction retention and elevations in circulating levels of pituitary adenylate cyclase-activating polypeptide (PACAP), a neuropeptide that regulates stress response through signaling within the amygdala complex. PACAP is frequently linked to the pathophysiology of PTSD and has been associated with altered functional connectivity between the centromedial amygdala and posterior regions of the default mode network, a network similarly linked to PTSD through its role in autobiographical memory and directed attention. Therefore, elevated levels of PACAP and its subsequent effects on amygdala functional connectivity may provide insights into the mechanisms of the intrusive-hypervigilant (IH) clinical phenotype of posttraumatic stress. This may represent an important step towards a biologically-grounded clinical subtype to target through mechanism-based interventions in a precision psychiatry approach.

In an independent sample of trauma-exposed adults, we sought to replicate findings of an IH clinical phenotype with elevations in PACAP and build upon their biological profile by providing novel evidence for distinct patterns of amygdala functional connectivity. Using resting-state functional magnetic resonance imaging (rs-fMRI), we tested the hypothesis that an IH phenotype would be associated with exaggerated functional connectivity of the amygdala with networks implicated in trauma memory and arousal.

Methods: 158 symptomatic trauma-exposed adults (106 female) were assessed using the Clinician-Administered PTSD Scale and completed a fasting blood draw followed by a 13-minute eyes-open resting-state fMRI scan. Gaussian mixture modeling was used to identify latent symptom profiles from symptom severity ratings using the 20-item CAPS-5 interview. Circulating PACAP levels were quantified using PACAP38-specific immunoassays. Seed-based resting state functional connectivity (FC) analyses were performed with regions of interest placed in the left and right (l/r) centromedial (CMA) and basolateral (BLA) amygdala. An identified IH group was compared to other symptom groups on circulating PACAP levels and strength of amygdala FC.

Results: A three-profile model of CAPS-5 symptoms demonstrated superior fit: a High Group (n = 73) characterized by elevated severity across most PTSD symptoms; a Low Group (n = 61) with generally low symptom severity; and an Intrusive-Hypervigilant (IH) group (n = 38) with pronounced elevations in intrusion and hypervigilance symptoms.

The identified IH group had significantly higher PACAP levels compared to the combined non-IH group (t(155) = 1.97, p = 0.025 one-tailed, d = 0.39). Individual contrasts confirmed elevated PACAP levels in the IH Group compared to both the High (t(154) = 1.65, p = 0.050 one-tailed, d = 0.35) and Low Groups (t(154) = 1.98, p = 0.025 one-tailed, d = 0.44). The High and Low Groups did not differ in PACAP levels (p = 0.607).

Comparisons of whole-brain amygdala-seeded FC maps revealed stronger rCMA connectivity in the IH Group for three midline clusters: a cluster spanning the intracalcarine cortex (k = 302 voxels, cluster FDR q = 0.001, peak MNI = [−4, −72, 16]), precuneus cortex (k = 287, cluster FDR q = 0.001, peak = [−4, −56, 68]), and medial prefrontal cortex (mPFC) including the anterior cingulate cortex (ACC; k = 210, cluster FDR q = 0.005, peak = [−2, 46, 14]). These clusters were situated predominantly within the Default Mode (k = 212) and Visual Networks (k = 208). Meta-analytic decoding of these clusters revealed spatial overlap with studies ascribing functions such as mental imagery, memory retrieval and encoding, fear and threat, and attention.

Conclusions: We replicated the findings of an IH clinical phenotype characterized by more severe intrusive memories, flashbacks, and hypervigilance, as well as elevated levels of circulating PACAP. Extending this prior work, we provide novel evidence that this IH phenotype is associated with stronger FC of the rCMA with midline cortical regions, particularly within the default mode and visual networks. This IH phenotype may reflect a distinct clinical presentation that is characterized by exaggerated interactions between memory and arousal processes, thus driving specific symptoms that are situated at this intersection, like flashbacks. Moreover, the IH phenotype exhibits biomarkers previously implicated in the PTSD literature, which has been stymied by inconsistent replication across different, diverse samples. Therefore, the IH phenotype may be a candidate clinical presentation for a precision psychiatry framework that targets these biomarkers through pharmacotherapy or neurotherapeutics.

Keywords: posttraumatic stress, PACAP, Amygdala, Biomarker, Clinical Subtypes.

Disclosure: Nothing to disclose.

P705. From polygenic risk to circuit pathology: discovering human neuronal cell types underlying PTSD using well-powered human datasets

Shagayeg Navabpour, Naghmeh Shargh, Hazal Senturk, Tayden Li, Madeleine Salem, Hadar Segal, Laramie Duncan

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

Background: Post-traumatic stress disorder (PTSD) affects approximately 8% of the general population. The development of this condition requires exposure to trauma, but research has also firmly established that there are individual differences in vulnerability to PTSD; like other psychiatric disorders, PTSD is moderately heritable. Regarding mechanisms thought to underlie PTSD, extensive prior work in animals and humans suggests that, in those who develop PTSD, the encoding and retrieval of fear memories is profoundly altered, resulting in heightened arousal, intrusive recollections, and emotional dysregulation. Neuroimaging and molecular studies have implicated the amygdala, hippocampus, and prefrontal cortex in PTSD pathophysiology, brain regions critical for fear processing, memory, and emotional regulation. Large-scale genome-wide association studies (GWAS) have demonstrated the polygenic nature of PTSD, identifying approximately 100 genome-wide significant loci to date. Complementing this, single-cell and spatial transcriptomic data have now enabled characterization of the cell-type diversity across the human brain, enabling the identification of nuanced brain cell types.

Despite these advances, the precise cellular and molecular basis of PTSD pathophysiology in humans is almost entirely unknown. Here, we address this gap by integrating, for the first time, an adequately powered PTSD GWAS (N > 1.2M) with the first high-resolution, brain-wide human single-nucleus transcriptomic atlas (>3.3M nuclei across 461 cell types). This unprecedented integration enables a statistically robust identification of the putative cellular architecture of PTSD and details about the molecular and cellular substrate of trauma-related memories.

Methods: In this study, we applied a brain-wide cell type enrichment framework, integrating the most recent PTSD GWAS (N = 1,222,882, both sexes) with high-resolution single-nucleus transcriptomic data (3,369,219 nuclei across 461 cell types, 3 males and 1 female; Siletti et al., 2023). Using MAGMA (as implemented in our previous publication, Duncan et al., 2025), we mapped polygenic signal onto brain cell types to test for enrichment. We then applied pairwise conditional analyses to identify representative cell types among correlated cell types.

Results: Cell type enrichment analysis identified 38 cell types significantly associated with PTSD risk, all of which were neuronal and encompassed both GABAergic and glutamatergic cell types. These cell types were found primarily in the amygdala, cortex, hippocampus, and thalamus. The most significant cellular association (of all 461 neuronal and non-neuronal cell types tested) was an excitatory (i.e., VGLUT1/VGLUT2) neuron found almost exclusively in the cortex and especially in the fusiform area (implicated in face recognition) of the temporal cortex (p = 9.0 × 10-8). This cell type was annotated as intratelencephalic (IT), suggesting long-range projections, and it has prominent expression of serotonin 3B and 2A receptors. There were also numerous amygdala cell types, a finding concordant with the understanding of PTSD as a fear-related condition. The most significant amygdala cell types included those of the inhibitory LAMP5-LHX6 class (p = 4.4 × 10-7), dopamine D1 receptor medium spiny type neurons (MSN-D1 class, p = 1.9 × 10-6), and other excitatory cell types found primarily in the amygdala that surpassed our Bonferroni correction for the 461 cell types tested (i.e., p < .0001). The top hippocampal cellular association (p = 1.9 × 10-5) was a GABAergic neuron, more specifically localized primarily to the dentate gyrus and CA4 regions of the hippocampus.

Conclusions: Our findings link multiple distinct neuronal cell types to PTSD. Reassuringly, the brain regions harboring these cell types mostly match those predicted based on the current understanding of PTSD as a disorder of fear learning. Specifically, this unbiased brain-wide analysis found that PTSD cell types were most prominent in the amygdala, followed by the cortex, hippocampus, and thalamus. Critically, this work provides the most detailed information ever available about precise human brain cell types contributing to PTSD, as prior studies could not have implicated these nuanced human brain cell types, which were first reported in 2023. Subsequent human postmortem tissue studies of PTSD have examined far more granular categories of cell types (e.g., dozens of cell types as contrasted with the 461 cell types tested here). Rather than focusing on broad categories of cell types (e.g., inhibitory neurons) or subclasses of neurons (e.g., somatostatin interneurons), we focused on the more granular cell types. This approach reflects the now well-established finding that there are at least ten or more distinct types of somatostatin interneurons, for example, each with differences in cortical layer localization, connectivity, and firing patterns. Thus, by providing a detailed map of specific human brain cell types linked to PTSD, we have provided a much more granular view of the likely cellular and circuit pathology underlying PTSD, which can be further probed in targeted postmortem tissue studies as well as experimental study designs in humans and model systems. Details about how to label these precise cell types will be provided in our online browser of cell type results.

Keywords: PTSD, cell types, Transcriptomics, Genomics.

Disclosure: Nothing to disclose.

P706. Post-traumatic stress disorder (PTSD) randomized clinical trials: predicting sex differences in enrollment patterns

Amanda Koire, Angelique Ealy, Janet Rich-Edwards, Nichole Goodsmith, Primavera Spagnolo, Hadine Joffe

Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, United States

Background: Equality in clinical trial enrollment by sex is key for detecting sex-specific differences in treatment safety and efficacy. Within the field of psychiatry, we and others have demonstrated that female participants are not equally represented in the populations of drug trials for psychiatric conditions overall or for substance use disorders. The current study analyzed sex differences in enrollment for all randomized controlled clinical trials (RCTs) in the PTSD Trials Standardized Data Repository (PTSD Repository) and identified trial characteristics associated with sex-specific enrollment disparities.

Methods: Data were obtained from the PTSD Repository, which encompasses all national and international RCTs supported by all funding sources that were conducted in adults with PTSD and published between January 1, 1988 and April 1, 2024. Median enrollment by sex for RCTs (550 trials; 48,174 participants) was compared to proportionality (sex-specific PTSD prevalence as per NIMH) expectations (74.2%) and parity (50% each sex) using one-sample Wilcoxon signed rank tests. Multiple regression models were built regressing female enrollment proportion on the following predictors: trial location (U.S. vs. outside U.S.), timing (years since study publication), size (total number of participants randomized), eligibility criteria, and trial intervention type (pharmacotherapy vs. no pharmacotherapy component).

Results: Female participants were significantly underrepresented (p < 0.001) across all PTSD clinical trials compared to proportionality expectations (53.0% [95% CI: 44.1%-57.5%]). Female participation was notably and significantly (p < 0.001) low by both parity and proportionality expectations for trials open to military-connected populations (10.8% [95% CI: 8.8%-13.9%]) or involving pharmacotherapy interventions (38.5% [95% CI: 23.4%-54.1%]). In multiple regression analysis, higher female enrollment was significantly predicted by having at least one U.S. trial site (β = −0.216, p < 0.001) and being more recent (β = 0.09, p = 0.02), while trial size was not a significant predictor (β = −0.030, p = 0.40). Lower female enrollment was significantly predicted by eligibility open to military-connected populations (β = −0.343, p < 0.001), or exclusive to military-connected populations (β = −0.810, p < 0.001), and inclusion of substance use disorder comorbidity (β = −0.117, p = 0.002). After controlling for the above variables, trials for pharmacotherapy interventions were independently associated with lower female enrollment (β = −0.137, p < 0.001).

Conclusions: PTSD RCT enrollment does not reflect the known sex-specific population PTSD prevalence and female participants are significantly underrepresented. Trial eligibility criteria biased toward PTSD subpopulations with male predominance (i.e., military connection, comorbid substance use disorders) may contribute to these disparities. Priority for PTSD RCTs in non-military-connected populations is warranted.

Keywords: PTSD, clinical trials, Health Equity.

Disclosure: Nothing to disclose.

P707. The interplay of stress regulators: corticotropin-releasing factor and norepinephrine interactions in threat

Jhah Cook, Aakash Basu, Jen-Hau Yang, Yulong Li, Alfred Kaye

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

Background: Post-traumatic stress disorder (PTSD) is a neuropsychiatric condition marked by dysregulated stress responsivity and altered activity in the prefrontal cortex (PFC) and locus coeruleus (LC). Norepinephrine (NE) hyperactivity and corticotropin-releasing factor (CRF) signaling via CRFR1 have been implicated in PTSD, but their real-time interplay during threat evaluation remains poorly understood. LC-NE neurons express CRFR1, and CRF enhances tonic firing, suggesting that CRFR1 modulation may shape moment-to-moment threat processing. Pupillometry offers a translational biomarker for LC-NE activity, providing a non-invasive window into stress-induced arousal shifts. This work investigated how CRFR1 antagonism influences PFC NE dynamics, freezing behavior, and pupil-indexed arousal.

Methods: Male and female C57BL/6 mice (n = 8–10 per group; ongoing n ≈ 30 planned per study) underwent classical fear conditioning and stress-sensitization protocols. For neurochemical assays, mice received stereotaxic injections of GRABNE sensors in prelimbic PFC and optic fiber implants for fiber photometry. Following fear conditioning, animals were administered the CRFR1 antagonist antalarmin (i.p.) or vehicle testing. Defensive freezing behavior was quantified using automated behavioral tracking. For physiological readouts, head-fixed mice underwent tone-shock pairing followed by cross-over systemic antalarmin or vehicle injections, with pupillometry performed to track cue-evoked pupil responses. Data were analyzed using repeated-measures ANOVAs with drug treatment as a within-subject factor. All procedures were approved by the Yale University IACUC and conformed to NIH Guide for the Care and Use of Laboratory Animals.

Results: CRFR1 antagonism significantly reduced cue- and shock-evoked PFC NE release compared to vehicle controls, demonstrating that CRF enhances threat-induced NE dynamics in the PFC. Antalarmin also suppressed maladaptive defensive freezing: mice treated with antalarmin froze significantly less than vehicle-treated animals during threat exposure and during re-exposure to the fear-associated context. Ongoing pupillometry experiments are designed to test whether CRFR1 antagonism reduces stress-induced pupil dilation, thereby providing a translational physiological biomarker of LC-NE modulation. Preliminary data and prior work from our group indicate strong correlations between NE dynamics and pupil size.

Conclusions: These findings demonstrate that CRFR1 antagonism attenuates threat-evoked NE release in the PFC and reduces maladaptive defensive freezing, supporting CRFR1 signaling as a critical driver of stress-sensitized arousal. Pupillometry is expected to validate pupil dilation as a translational biomarker of CRF-NE interactions. Together, these results provide mechanistic insight into the role of CRF in moment-to-moment threat processing and highlight CRFR1 antagonists as promising therapeutic interventions for PTSD and related stress disorders.

Keywords: Acute Traumatic Stress, CRF1 Receptor Antagonist, norepinephrine.

Disclosure: Nothing to disclose.

P708. Molecular traces of childhood trauma in hippocampal cell types

Adrien Gigliotta, Clara Snijders, Artemis Iaoutry, Erin Hisey, Sam Frank, Marina Soliva-Estruch, Charles B. Nemeroff, Joel Kleinman, Kerry Ressler, Nikolaos Daskalakis

Harvard Medical School, McLean Hospital, Mailman Research Center, Belmont, Massachusetts, United States

Background: Childhood trauma (CT) is a significant risk factor for posttraumatic stress disorder (PTSD), but the molecular mechanisms distinguishing PTSD with a history of CT from PTSD without CT (noCT) are poorly understood. This study aimed to investigate the distinct cellular and molecular pathologies in the hippocampus that differentiate these two subtypes of PTSD.

Methods: We generated 380,948 single-nucleus RNA sequencing (snRNA-seq) and parallel single-nucleus ATAC-seq (snATAC-seq) profiles from the hippocampus of individuals with PTSD (CT: n = 18; noCT: n = 17) and neurotypical controls (n = 18). We validated these findings in an independent major depressive disorder (MDD) cohort (n = 53) and assessed regional specificity against prefrontal cortex profiles (n = 47). Signatures were further integrated with bulk hippocampal RNA-seq, DNA methylation, and proteomics data from 301 donors to distinguish them from effects of polygenic risk scores (PRS), trauma load, diagnoses, and death by suicide. Plasma proteomics were used to identify brain–blood-concordant biomarkers. Cross-species validation was conducted using an early-life stress (ELS) mouse model and primary neuronal hippocampal cultures.

Results: Our analysis revealed distinct pathologies for each PTSD subtype. In noCT cases, we observed cytoskeleton downregulation in excitatory neurons, upregulation of exocytosis and energy pathways in microglia, and endothelial inflammation. In contrast, CT cases involved disrupted hormonal transport in oligodendrocytes and altered synaptic regulation by microglia. A direct comparison of CT to noCT PTSD revealed downregulation of synaptic structure pathways in astrocytes and upregulation of fatty acid synthesis and antigen presentation pathways in oligodendrocytes. snATAC-seq confirmed related glial chromatin-accessibility shifts. We identified 119 brain–blood-concordant proteins based on CT history, with the protein FURIN showing consistent dysregulation across snRNA-seq, fine-mapping, and blood proteomics. The ELS mouse model and primary cell cultures recapitulated key human CT pathways, including inflammation, metabolic dysregulation, synaptic remodeling, and altered stress and glucocorticoid signaling.

Conclusions: Together, these multi-modal, cross-species data establish a trauma-stratified, single-cell atlas of PTSD. This work reveals enduring and distinct molecular signatures of early-life adversity in the hippocampus, providing potential biomarkers and a foundational resource for future research.

Keywords: MDD, PTSD, Childhood trauma, Hippocampus, snRNASeq.

Disclosure: Nothing to disclose.

P709. Short-term facilitation at prefrontal somatostatin interneurons confers resilience to traumatic stress in mice

Jensine Coudriet, Yuna Park, Wu Min, Xiaoqiang Shi, Jiawei Wang, Matthew Girgenti, Alicia Che

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

Background: Psychological trauma leads to posttraumatic stress disorders (PTSD) in some individuals, while others are resilient to the same stressor. A critical challenge is to understand the neurobiological mechanisms governing the differences in susceptibility. Recent work using human postmortem genomic and transcriptomic analyses revealed selective vulnerability in GABAergic interneurons, particularly somatostatin-expressing (SST) interneurons in PTSD. However, the mechanisms through which SST interneurons regulate traumatic stress response and susceptibility is unresolved. We investigated the role of extracellular leucine-rich repeat and fibronectin type III domain containing 1 (Elfn1) – an extracellular matrix-associated protein expressed on SST interneurons identified to be downregulated in PTSD human postmortem PFC – and its presynaptic partner metabotropic glutamate receptor 7 (mGluR7) in regulating the activation of SST interneurons and susceptibility to traumatic stress.

Methods: We used single prolonged stress (SPS) model in male and female mice, and measured anxiety-like behaviors using elevated plus maze (EPM), open field test (OFT), and an unsupervised machine learning method Motion Sequencing (MoSeq). Mice were categorized as susceptible (SUS), resilient (RES) or indifferent (IND) based on their behavioral performances, and RNA sequencing (10 males and 10 females each for control and SPS) on medial prefrontal cortex (mPFC) tissue was performed to confirm the delineation of these three distinct groups. Fluorescence in situ hybridization (FISH) was performed on SPS and control mice to quantify Elfn1 and Grm7 (gene encoding mGluR7) levels in SUS versus RES mice. Synaptic dysfunction in SST interneurons was characterized by patch clamp slice electrophysiology. Dual-color fiber photometry was performed to assess the relationship between PFC SST interneurons and pyramidal neuron activation during OFT and EPM. We also assessed anxiety-like behaviors and synaptic function of mice with Elfn1 conditional deletion in SST interneurons. Lastly, we evaluated the effects of single-dose AMN082 (an mGluR7 agonist) on rescuing Elfn1 and Grm7 expression as well as anxiety-like behaviors after SPS.

Results: We found SPS reliably resulted in SUS, RES and IND groups in mice. SUS but not RES mice showed decreased Elfn1 expression, and the level of Elfn1 was significantly correlated to anxiety-like behaviors in SPS but not control mice. Patch clamp slice electrophysiology revealed lower inhibitory synaptic transmission and significantly diminished short-term facilitation in SST interneurons, normally characteristic of this cell type. Simultaneous fiber photometry recording of mPFC pyramidal neurons and SST interneurons mice during EPM and OFT showed elevated excitatory-to-inhibitory (E/I) ratio during exploratory behaviors in SPS mice compared to control. Additionally, mice with SST interneuron-specific Elfn1 deletion showed exacerbated vulnerability to SPS, while treatment with the mGluR7 agonist AMN082 (10 mg/kg) mitigated SPS-induced anxiety-like behaviors, restored Elfn1 expression, and reduced mGluR7-mediated synaptic depression at the inhibitory synapse.

Conclusions: Our findings demonstrate that a mouse model of SPS results in distinct behavioral subgroups that are susceptible, indifferent, and resilient to developing anxiety-like behaviors. Elfn1-mGluR7-mediated short-term synaptic plasticity in prefrontal SST interneurons is disrupted in SUS mice, resulting in reduced inhibitory synaptic transmission that leads to elevated E/I ratio and heightened anxiety-like behaviors, which can be rescued by mGluR7 antagonism. We identified this short-term plasticity at the pyramidal-neuron-to-SST-interneuron synapse as a critical mechanism underlying susceptibility to traumatic stress.

Keywords: Anxiety and PTSD, GABAergic interneurons, metabotropic glutamate receptor, Medial Prefrontal Cortex (mPFC), neural circuits.

Disclosure: Nothing to disclose.

P710. Effects of the hallucinogens psilocybin and salvinorin a on fear extinction in male and female mice

Allison Foilb, Brett Schmidt, Brian Kangas, Kerry Ressler, Jack Bergman, William Carlezon

Harvard Medical School McLean Hospital, Belmont, Massachusetts, United States

Background: Psychedelics have powerful effects on the brain that may offer therapeutic potential for psychiatric illness. Evidence suggests that psychedelics such as psilocybin induce neuroadaptive responses that can lead to reductions in key features of stress and trauma disorders. Psilocybin binds to a variety of receptors, but its actions at 5-HT2A receptors have been associated with hallucinations characterized by illusions, auditory distortions, heightened emotion, and mystical experiences. Although limited, clinical studies have indicated that psilocybin improves outcomes for individuals with depression and anxiety in conjunction with behavioral interventions. We previously demonstrated that sleep disruption—which can lead to hallucinations in people—in fear-conditioned mice reduces fear expression and enhances extinction memory, likely through induction of plasticity in the amygdala, a critical fear-processing brain region. The present studies were designed to determine if administration of psilocybin prior to fear extinction could produce similar fear-reducing effects. To explore whether induction of hallucinations is sufficient to produce these effects, we conducted parallel studies with salvinorin A (salvA), a drug that can also produce mystical experiences while binding almost exclusively to kappa opioid receptors (KORs), thus acting via a different mechanism than psilocybin.

Methods: Adult male and female C57BL/6J mice underwent fear conditioning with 5 tone-footshock pairings. On the following day, mice received drug pretreatment prior to the Fear Recall Test/Extinction Training session. In Experiment 1, mice were pretreated with intraperitoneal (IP) 2.0 mg/kg psilocybin (obtained from National Institute on Drug Abuse [NIDA]) or saline 30 min before testing. In Experiment 2, mice were pretreated with 2.0 mg/kg salvA (IP) or vehicle 15 min before testing. Mice were monitored for somatic responses including head twitches, which are thought to indicate hallucinogenic effects. Mice were then placed in a novel conditioning context for the Fear Recall/Extinction session, with 15 non-reinforced tone presentations, an intentionally weak extinction training procedure intended to increase sensitivity to extinction-enhancing effects. On the following day (24 hr later), mice were returned to the extinction context and tested for Extinction Recall during another 15 tone presentations. Mice were returned to the extinction context 4 days after the initial Fear Recall/Extinction training session and presented with 5 tones. All procedures were approved in advance by McLean Hospital Institutional Animal Care and Use Committee and carried out in compliance with the National Institutes of Health’s (NIH) Guide for the Care and Use of Animals.

Results: In Experiment 1, there were no differences in freezing behavior between mice that had received acute psilocybin or vehicle (n = 8/sex/condition) during the Fear Recall/Extinction session. However, 24 hr later during the Extinction Recall session, psilocybin-treated mice displayed significantly less freezing to the conditioned tone than vehicle-treated controls when sexes were combined (P < 0.01). While analyses in males or females separately did not reach statistical significance, there were strong trends for each sex (P’s < 0.07). When the mice were tested in an Extinction Recall session 4 days later, psilocybin-treated mice continued to display less fear than vehicles when sexes were combined (P < 0.05). In Experiment 2, there was no effect of salvA (n = 10/sex) on Fear Recall/Extinction compared to vehicles (n = 10/sex), although salvA-treated mice displayed increased freezing during inter-trial intervals (P < 0.01). During Extinction Recall tests the following day, there was no overall effect of salvA. However, when examining the sexes separately, significant reductions in freezing were seen in salvA-treated females (P < 0.05) but not males. When mice were re-tested 4 days later, no effects were observed.

Conclusions: Administration of psilocybin prior to the Fear Recall/Extinction session significantly reduces fear responses during Extinction Recall tests conducted 24 hr and 4 days later, suggesting sustained enhancement of fear extinction memories. While salvA did not produce consistent effects on Fear Recall/Extinction or Extinction Recall when the sexes were combined, it produced significant reductions in freezing in females during Extinction Recall, suggesting that it produces enhancement of fear extinction memories. Reductions were not seen 4 days later, however, suggesting that the effects of salvA on fear extinction are less persistent than those of psilocybin. This pattern of salvA effects was surprising, as previous work indicates that females are less sensitive to KOR agonists in motivation and reward-related tests. Together these findings confirm the potential therapeutic-like effects of psilocybin while demonstrating that salvA shares some of these same effects despite no known overlap in receptor binding, raising the possibility that hallucinations may play an important role in the ability of psychedelics to reduce key features of fear- and trauma-related disorders. Further work is needed to determine if these treatments produce common neuroadaptative responses, particularly in brain areas implicated in the fear and extinction processes. A better understanding of shared mechanisms may enable transformative improvements in the treatment of psychiatric illness.

Keywords: Psilocybin, Fear conditioning and extinction, salvinorin A, Amygdala.

Disclosure: Nothing to disclose.

P711. Dynamic duo: serotonin transporter and organic cation transporter 3 regulate basolateral amygdala serotonin clearance and fear memory recall

Lynette Daws, Lauren Honan, Sangbin Shin, W Anthony Owens, Rebecca Horton, Glenn Toney

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

Background: Dysregulation of serotonin (5-HT) neurotransmission in basolateral amygdala (BLA), a key integration hub for processing emotional stimuli, is crucial for aberrant fear memory. The high-affinity, low-capacity 5-HT transporter (SERT), the target of selective 5-HT reuptake inhibitors (SSRIs) commonly used to treat post-traumatic stress disorder (PTSD), is richly expressed in BLA. However, SSRIs often provide suboptimal therapeutic benefit. Organic cation transporter 3 (OCT3), a low-affinity, high-capacity 5-HT transporter is also richly expressed in BLA. However, its contributions to BLA 5-HT clearance and fear memory are unknown. Here, we tested the hypothesis that SERT and OCT3 on 5-HT neurons are important both for BLA 5-HT clearance and for fear memory formation and recall.

Methods: In vivo high-speed chronoamperometry was used to measure clearance of 5-HT from extracellular fluid in BLA. Serotonin was pressure-ejected into BLA to achieve signal amplitudes of either 0.5 µM (“low”), where clearance is thought to be driven primarily by the high-affinity, low-capacity SERT, or 2.0 µM (“high”), a concentration known to engage the low-affinity, high-capacity, OCT3. Sample sizes ranged from 14–20/group. The ability of the SSRI, fluvoxamine, to inhibit 5-HT (“low”) clearance was assessed in all groups (n = 7–14/group). Classical fear conditioning was used to assess fear learning (freezing) as well as fear memory in both cued and context scenarios, measured 2 and 4 days after acquisition, respectively. Sample sizes ranged from 9 to 25/group. All studies included male and female mice. Since no sex differences were detected, sample sizes represent both sexes pooled. In vivo high-speed chronoamperometry data were analyzed by two-factor mixed-effects analysis followed by Uncorrected Fisher’s LSD test. Behavioral data were analyzed by two-way ANOVA.

Results: Consistent with SERT being a high-affinity, low-capacity transporter for 5-HT, we found that 5-HT clearance in BLA was slowed at “low” but not “high” extracellular concentrations and that the SSRI fluvoxamine was without effect on 5-HT in mice with SERT knockout compared to controls. In OCT3 knockout mice, 5-HT clearance was slowed both at “low” and “high” extracellular concentrations, suggesting OCT3 contributes more to 5-HT uptake than previously appreciated. Moreover, fluvoxamine more robustly inhibited 5-HT clearance in OCT3 knockout mice than in wildtype mice, consistent with OCT3 being an important player in BLA 5-HT clearance even when concentrations are “low”, especially when high-affinity clearance mechanisms are pharmacologically compromised. Behaviorally, fear learning was unaffected by SERT or OCT3 knockout on 5-HT neurons. In both SERT and OCT3 knockout mice, however, freezing was reduced during cued recall 2 days later. Context recall tested 4 days later was not significantly impacted by SERT or OCT3 knockout.

Conclusions: These data show that OCT3 plays a more prominent role in 5-HT clearance from extracellular fluid in BLA than previously thought. Moreover, SERT and OCT3 each play an active role in fear memory recall, opening the possibility that pharmacotherapeutics targeting OCT3 and SERT concurrently may have greater therapeutic response than SSRIs alone.

Keywords: basolateral amygdala, fear conditioning, organic cation transporter 3, serotonin clearance, serotonin transporter.

Disclosure: Nothing to disclose.

P712. Sex-specific transcription and translation in Adcyap1+ parabrachial nucleus (PBN) cells during fear learning

Zoe Beatty, Joy Otten, Robert Fenster, Katie Mabbott, Torsten Klengel, Kerry Ressler

McLean Hospital - Harvard Medical School, Boston, Massachusetts, United States

Background: The ability to learn associations between innately fear-evoking stimuli and previously neutral stimuli is crucial to survival and, thus, highly evolutionarily conserved across species. This ability is critically disrupted in Post Traumatic Stress Disorder (PTSD), as PTSD patients fear a wider-range of previously neutral stimuli (i.e. increased fear generalization) and have more difficulty learning that these stimuli are no longer associated with danger (i.e. decreased fear extinction) after traumatic stressor exposure than controls. The neuropeptide PACAP has been repeatedly implicated in PTSD specifically in women, with PACAP blood levels predicting PTSD diagnosis and symptom severity. PACAP is also critically involved in fear learning, as evidenced by a range of studies focused on the PACAP system’s role in fear acquisition, expression, and extinction. PACAP-expressing neurons (or Adcyap1+ neurons) in the parabrachial nucleus (PBN) provide PACAP to a host of brain regions involved in fear, such as the central amygdala (CeA) and bed nucleus of the stria terminalis (BNST), and are thus poised to regulate fear processing.

Methods: Adcyap1-Cre and Rosa26-eGFP-L10a mice were crossed to generate a mouse line with GFP tags on the L10a ribosomal subunit in Adcyap1+ cells (Adcyap1 encodes PACAP). Mice were exposed to fear conditioning (FC, 5 tone-shock pairs) or a tone alone (TA) condition, brains were collected two hours later, and the parabrachial nucleus (PBN) was punched. mRNAs from Adcyap1+ cells that were being translated at the time of tissue collection were isolated using Translating Ribosome Affinity Purification (TRAP) and bulk RNA-seq was conducted to measure these mRNAs. Differential gene expression (DGE) analysis was used to compare translated mRNAs between FC and TA mice, as well as to compare mRNAs between Adcyap1+ and Adcyap1- PBN cells. Several results from DGE analysis were validated using RNAscope in situ hybridization. All studies were conducted in adult animals of both sexes.

Results: We identified 20 differentially expressed genes (DEGs) in Adcyap1+ PBN cells of FC vs TA female mice, 18 of which were upregulated and 2 of which were downregulated in FC females (n = 16, p < 0.05 post FDR-correction). We identified a single DEG in Adcyap1+ PBN cells of FC vs TA male mice, which was upregulated in FC males (n = 13, p < 0.05 post FDR-correction). Many upregulated genes in FC females were immediate early genes, such as Fos, FosB, and Arc. Using in situ hybridization, we discovered that Fos mRNA expression is significantly increased in Adcyap1+ PBN cells upon FC in females, while Fos mRNA expression is not significantly changed upon FC in males. Fos mRNA expression in FC females was localized to a subpopulation of Adcyap1+ cells in the dorsal lateral subnucleus of the PBN (lPBd). DGE analysis comparing mRNA expression in Adcyap1+ and Adcyap1- PBN cells revealed a total of 5,197 genes preferentially expressed in Adcyap1+ cells (n = 32, p < 0.05 post FDR-correction). Adcyap1+ PBN cells preferentially express genes encoding a range of fear and stress-sensitive neuropeptides and their receptors, such as Sst, Nts, Tac1, Sstr2, Sstr3, and Tacr1.

Conclusions: Based on the finding that immediate early genes are preferentially translated in Adcyap1+ PBN cells in FC females compared to TA females, we believe that these cells are likely preferentially active in females upon FC. Given Fos mRNA expression is also significantly increased in Adcyap1+ PBN cells exclusively in females after FC, these results may indicate a sex-dependent role of Adcyap1+ PBN cells in fear learning. Current and future studies in the lab will focus on elucidating the role of Adcyap1+ PBN cells in fear learning (as well as the role of PACAP released from these cells) through cell-type specific manipulations including genetic knockout and chemogenetic inhibition/activation.

Keywords: Fear conditioning, PACAP, Aversive Learning, parabrachial nucleus, neuropeptides.

Disclosure: Nothing to disclose.

P713. Differential correlations between default mode network hub resting state connectivity and GABA+ concentrations in major depressive disorder with antepartum onset vs. healthy postpartum females

Kristina Deligiannidis, Anita Barber, Xiang He, Hengyi Cao, Majnu John, Sunu Jacob, Athanasia Giannopoulos, Humberto Monsivais, Gianna Nossa, Brian Bozymski, Ulrike Dydak

Zucker Hillside Hospital, Northwell Health, Glen Oaks, New York, United States

Background: Major depressive disorder (MDD) with peripartum onset (PPD) affects 1 in 10 birthing females. Altered neuroactive steroid-dependent GABAergic functioning/neuroplasticity in the stress circuit is one part of an integrated hypothesis for the pathophysiology of PPD. PPD is characterized by hyperactivity of the default mode network (DMN), with the most evidence for DMN midline hub dysfunction. Abnormal cortical GABA concentrations have been reported in MDD without peripartum onset but not in PPD where GABA has been solely measured as a concentration within a particular region of interest (ROI). Since inter-individual resting-state GABA concentrations are associated with functional connectivity within and between resting-state networks and DMN BOLD responses are mediated by GABA, we examined the relationship between DMN midline hub GABA and DMN midline hub network dysfunction in PPD. We hypothesized that DMN midline hub connectivity would be correlated with GABA concentrations and with depression severity in females with PPD. We hypothesized that this correlation would be specific to DMN hub connectivity and would be absent in a comparison network.

Methods: We conducted an interim analysis (R01MH120313) of a prospective observational study of peripartum females who presented in late gestation either with MDD with antepartum onset or were euthymic without any past psychiatric history. Participants were evaluated with the Structured Clinical Interview for DSM-5 and the Hamilton Depression Rating Scale (HAMD17) at study entry (28 weeks gestation) and postpartum study exit. Exclusions included: history of hypomania, mania, psychosis, substance or tobacco use disorder and use of medications with psychotropic, neurosteroid or hormonal activity. A postpartum imaging session was conducted on a 3T SIEMENS Prisma whole-body MR system with a 64-channel head coil. T1-weighted high-resolution structural scans were collected using an MPRAGE sequence, producing 208 contiguous images with 0.8mm slice thickness. Resting-state fMRI scans were acquired with AP and PA phase encoding directions, each with a 7:17min acquisition time. Multi-band resting state fMRI acquired 594 volumes with TR = 720ms, TE = 33ms, matrix = 104 × 104, FOV = 231 mm × 200 mm, slice thickness = 2mm, and 72 continuous axial oblique slices (voxel size = 2.2mm × 2.2 mm × 2 mm). Data was motion-corrected, skull stripped, slice-timing corrected, intensity-normalized and spatially smoothed. Scans were examined for motion artifact and noise components were regressed out. Subsequently, temporal filtering and nuisance regression was done. For 1H-MRS, two ROIs were placed midsagittally in the pregenual anterior cingulate cortex (pgACC, 15 ml) and in the occipital cortex (OCC, 27 ml). GABA-edited spectra were acquired using MEGA-PRESS (TR = 2000 ms, TE = 68 ms, 256 averages, with editing pulses applied at 1.9 (ON) and 7.5 (OFF) ppm. GABA spectra were processed using Osprey and the GABA+ (GABA plus macromolecules) signal at 3.00 ppm was quantified with LCModel relative to unsuppressed water. Data were excluded from further analyses if SNR < 10, linewidth > 0.1 ppm, or the standard deviation of GABA+ fits was > 30%. Metabolite levels were corrected for cerebrospinal fluid water content within each MRS voxel and for water relaxation times.20 Using a pgACC seed (5-mm sphere) centered within the participant-specific pgACC MRS ROI, the time series for the seed region was extracted from the residual signal after nuisance regression and both were normalized. Full-brain pgACC connectivity maps were submitted to Fisher r-to-Z transformation and entered into SPM group models which tested for associations with pgACC connectivity, HAMD17 or pgACC GABA+. To determine pgACC specificity we examined tested for correlations between OCC connectivity, HAMD17 and OCC GABA+ concentrations.

Results: 49 participants (mean age 33.32 years, SD = 4.89) completed a postpartum imaging session (mean days postpartum at time of MRI = 59.30, SD = 17.31). pgACC GABA+ was positively correlated with pgACC-precuneus/posterior cingulate cortex connectivity in PPD (n = 21). Connectivity between these regions also positively correlated with HAMD17 (p = 0.008). In healthy postpartum females (n = 27), pgACC GABA+ was negatively correlated with pgACC-precuneus and posterior cingulate cortex connectivity (p < 0.01). In PPD (n = 19), OCC GABA+ was negatively correlated with OCC-pgACC connectivity (p < 0.001) but positively correlated with OCC-pgACC connectivity (p < 0.001) in healthy postpartum females (n = 30) and was not correlated with HAMD17 (p > 0.05).

Conclusions: This study provides the first evidence that pgACC GABA+ concentrations are correlated with DMN hub resting-state connectivity in PPD and negatively correlated in healthy postpartum females. The directionality of these correlations differed in the visual network, a comparison region which lacked a relationship to depression severity. This interim analysis suggests the importance of pgACC GABAergic dysfunction within the DMN midline hub in PPD and supports the use of integrated RSFC and MRS approaches to understand the regional and functionally specific role of GABA within networks and their relationship with MDD.

Keywords: postpartum depression, proton magnetic resonance spectroscopy, Resting State Functional Connectivity, Default mode network (DMN), GABA.

Disclosure: Sage Therapeutics, Consultant, Self, Sage Therapeutics, Contracted Research, Self, Biogen, Consultant, Self, GH Research, Consultant, Self, Neurocentria, Consultant, Self, Gerbera/DuKang Therapeutics, Contracted Research, Self, Reunion Neuroscience, Advisory Board, Self, Lipocine, Consultant, Self, Gerbera Thearpeutics, Consultant, Self, Neuroscience Software, Consultant, Self, Brio Biosciences, Consultant, Self.

P714. Menstrual cycle–related sleep disturbance and experimental hormone supplementation in a high-risk psychiatric sample

Anisha Nagpal, Katja Schmalenberger, Fiona Baker, Tory Eisenlohr-Moul

University of Illinois At Chicago, Chicago, Illinois, United States

Background: Sleep disturbances are a well-established risk factor for suicidal thoughts and behaviors (STBs), particularly among individuals with psychiatric comorbidity. Menstrual cycle–related affective changes arise from abnormal sensitivity to normal ovarian steroid fluctuations, with PMDD affecting ~5.5% of AFABs and PME occurring in up to 60% of those with major depressive disorder. Importantly, hormone sensitivity is heterogeneous: some individuals experience luteal-phase symptoms linked to the progesterone metabolite allopregnanolone, while others are more vulnerable to perimenstrual hormone withdrawal. Converging observational and experimental evidence indicates that perimenstrual ovarian steroid withdrawal precipitates worsening suicidality in a subset of sensitive individuals. Sleep disturbances also worsen perimenstrually in both healthy and clinical populations, but most evidence is observational, limiting causal inference.

This gap is striking given estradiol (E2) and progesterone (P4) act on circadian sleep-wake regulatory nuclei, neurotransmitter systems, and thermoregulation, and that P4 metabolites, such as allopregnanolone, act on the GABA-A receptor, promoting sedation. While E2 improves sleep in menopausal samples, no studies have examined how E2/ P4 manipulation impacts sleep in premenopausal individuals with STBs.

The current study addresses this gap by using novel methods to identify unique patterns of menstrual cycle–related sleep disturbance in psychiatric populations and by testing perimenstrual E2/P4 supplementation in a double-blind, placebo-controlled crossover trial.

Methods: (1) Baseline cycles: 139 naturally cycling females with past-month suicidal ideation (SI) completed daily ratings of trouble sleeping (1 = “not at all” to 6 = “extremely”) across 1–3 menstrual cycles. Cycle time (percent luteal/follicular phase; calculated with menstrualcycleR R package) was modeled as a predictor of trouble sleeping using cyclic splines basis function in generalized additive mixed models (GAMMs).

(2) Subgroups: Trajectory subgroups of trouble sleeping were identified using smooth mixture models fit via an expectation–maximization algorithm. Model selection was based on Bayesian Information Criterion (BIC) and generalized cross-validation. Identified subgroups were compared on demographic and clinical characteristics, including the Inventory of Depression and Anxiety Symptoms (IDAS), via Kruskal-Wallis tests.

(3) Experimental trial: A subset of 60 participants completed a luteinizing hormone–timed, placebo-controlled crossover trial, receiving either placebo or perimenstrual supplementation with a.1mg/d E2 patch plus 200 mg/day oral micronized P4, separated by a washout cycle. Trouble sleeping was modeled with GAMMs as a function of days since LH surge and condition (E2/P4 vs placebo), with condition-specific smooths and participant-level random effects. An exploratory analysis repeated this model in subgroups showing cyclical sleep patterns.

Results: In baseline cycles, trouble sleeping followed a significant, nonlinear perimenstrual peak (edf = 3.127, p = 0.001), with cycle time accounting for 8% of within-person variance. Substantial heterogeneity emerged, with three trajectory subgroups: flat with a mild late-follicular rise (N = 72), perimenstrual peak (N = 53), and midluteal peak (N = 14). Subgroups differed in symptom profiles, with the luteal group showing lower appetite loss (\chi^2(2) = 5.23, p = 0.07) but higher mania (\chi^2(2) = 7.22, p = 0.03) (IDAS subscales). In the experimental trial, E2/P4 supplementation buffered perimenstrual increases in trouble sleeping observed under placebo (edf = 2.81, p = 0.005), with significant individual differences in treatment response (edf = 39.9, p < 0.0001). In the exploratory subgroup analysis (N = 23), cyclical subgroups (i.e. perimenstrual and luteal) showed an amplified treatment effect: perimenstrual increases during placebo were replaced by marked perimenstrual declines during E2/P4 condition (edf = 4.04, p = 0.005).

Conclusions: This study applied novel analytic strategies, nonlinear modeling with GAMMs and smooth mixture models, to characterize heterogeneous, hormone-sensitive patterns of self-reported sleep disturbance across the menstrual cycle. Findings demonstrate that perimenstrual E2/P4 supplementation can buffer cycle-related worsening of sleep self-reports, particularly in individuals with pronounced cyclicity. Future analyses will test whether individuals with improvements in sleep coincide with perimenstrual reductions in suicidal ideation, offering insight into sleep self-reports as a pathway linking ovarian hormone sensitivity to suicide risk.

Keywords: ovarian hormones, Sleep disturbances, Mixture modelling.

Disclosure: Nothing to disclose.

P715. PTSD symptom burden increases risk for labor dystocia in black individuals undergoing labor induction: DNA methylation as a potential pathway

Abby Britt, Seyma Katrinli, Alicia Smith, Nicole Carlson, Alexis Amore, Abigail Powers, Houser Madelyn, Catherine Monk, Vasiliki Michopoulos

Columbia University Vagelos College of Physicians and Surgeons, New York, New York, United States

Background: The United States has higher maternal morbidity and mortality rates than any other high-income country, and there are considerable disparities in maternal mortality rates by race, with Black compared to White pregnant individuals 3.5 times more likely to die a pregnancy-related death. In the U.S., labor dystocia, defined as abnormally slow labor progress, is a significant contributor to maternal morbidity and mortality for Black individuals. Post-traumatic stress disorder (PTSD) increases risk for poor intrapartum outcomes, yet the physiological mechanisms are largely unknown. Trauma, as a social determinant of health, may become biologically embedded via epigenetic modifications of gene expression, influencing perinatal outcomes. PTSD has been associated with alterations in DNA methylation (DNAm) in genes responsible for stress response physiology, which also are integral to labor initiation and progress. While physiological dysregulation in these pathways has been identified in relation to preterm birth, low birth weight, impaired infant bonding, and maternal mental health, their potential impact on term labor outcomes is largely uncharacterized, especially in the U.S. Black population. The purpose of the current study was to evaluate associations between PTSD symptoms, labor dysfunction, and DNAm in a cohort of Black pregnant individuals. We hypothesized that higher PTSD symptom burden would increase risk for labor dystocia, and that DNAm at CpG sites associated with PTSD symptom burden also would be associated with labor dystocia.

Methods: The current study is a secondary analysis of data collected from two prospective cohort studies involving Black pregnant individuals recruited from an academic medical center and a midwife-led care birth center in Atlanta, Georgia. Participants had term labor, a singleton fetus in cephalic presentation, and reached active phase labor (N = 238). Prenatal PTSD symptoms were assessed using the Posttraumatic Stress Disorder Checklist for DSM-5 (PCL-5). Labor outcomes were collected via health record abstraction. In a subset of participants (n = 114), DNAm from maternal whole blood was measured by bisulfite sequencing using Illumina EPICv2 BeadChip. Multiple logistic regression models were used to evaluate associations between PTSD symptoms and labor dystocia in the total sample after stratification by mode of labor onset, as it is known to impact labor duration, and adjusted for sociodemographic and pregnancy/labor related factors. Epigenome wide association studies (EWASs) of labor dystocia and PTSD symptom score were completed using R package CpGassoc. Briefly, DNA methylation Beta values (β) were calculated for each CpG site as the ratio of methylated (M) to methylated and unmethylated (U) signal: β = M/M+U. Beta values were regressed on the independent variables (labor dystocia or PTSD symptom score), adjusting for age, blood cell proportions, ancestry principal components, and DNAm-based smoking score. A Fisher’s exact test was used to conduct an enrichment test with CpG sites nominally associated (p < .05) with both PTSD symptom score and labor dystocia in multiple generalized linear models.

Results: No significant associations were found between PTSD symptoms and labor dystocia in the total sample (N = 238) (OR 1.00, 95% CI [.938,1.02], p = .930). In participants with induced labor (n = 64), in univariate and multivariate logistic regression models, each unit increase in total PTSD symptoms increased the odds of labor dystocia, even after adjusting for sociodemographic and labor factors (aOR 1.07, 95% CI [1.02–1.13], p = .006). In the subset of participants with DNAm data (n = 114), CpGs related to versus not related to PTSD were 2.3x more likely to be associated with labor dystocia (OR [95% CI] = 2.31[2.23,2.38], p < 2.2e-16).

Conclusions: In a sample of Black pregnant individuals, 25% of whom underwent labor induction, greater PTSD symptoms were associated with labor dystocia. Additionally, among those in the overall sample with DNAm data, there was significant overlap in DNAm in nominally significant CpG sites associated with PTSD and with labor dystocia. Such differential DNAm may represent epigenetic pathways by which PTSD may impact maternal morbidity- and mortality-related intrapartum outcomes for Black individuals. Further research is needed to understand the potential associations amongst and biological mediation of trauma, PTSD symptoms, and labor dystocia to inform interventions toward birth equity.

Keywords: pregnancy, PTSD, Birth Outcomes, Epigenetics, Health Equity.

Disclosure: Nothing to disclose.

P716. Prenatal stress moderates the effect of social support on third trimester alcohol use: a pregnancy risk assessment monitoring system study

Andrea Maxwell, Anna Zilverstand, Cresta Jones, Jeffrey Wozniak

University of Minnesota, Minneapolis, Minnesota, United States

Background: Prenatal alcohol use (PAU) is associated with increased likelihood of adverse short and long-term pregnancy outcomes, including obstetric complications and Fetal Alcohol Spectrum Disorders. We previously reported that high social support has a causal effect on lower Alcohol Use Disorder symptom severity in non-pregnant women and that this protective effect is moderated by stress. Whether this interaction effect of social relationships and stress on alcohol use extends to women during pregnancy has yet to be explored. In the present study, we used population-wide data collected by the CDC’s Pregnancy Risk Assessment Monitoring System (PRAMS) to test a moderation model between prenatal stress, social support, and third trimester drinking. We hypothesized that social support would buffer the effect of prenatal stress on drinking during the third trimester.

Methods: The Pregnancy Risk Assessment Monitoring System (PRAMS) is an ongoing, population-based surveillance system conducted by the CDC’s Division of Reproductive Health that assesses multidimensional factors including safe sleep habits, substance use, food insecurity, among others, during the perinatal period. To assess prenatal social support, women reported how many of four, if any, social support behaviors they received in the 12 months prior to their child’s birth, which probed both instrumental (e.g., were loaned money), and emotional (e.g., had someone to talk to) support. To measure prenatal stress, women reported how many of 14 stressors, if any, they experienced during that same period, with indices including moving to a new address or someone close to them having a problem with alcohol or other substances, among others. In a subsample of the 2016–2022 phase of PRAMS, (N = 373 with ≥ 1 alcoholic drink in the third trimester, mean age = 31.37 years, 62.7% white), we tested whether the number of stressful events during the prenatal period moderates the effect of social support on third trimester alcohol use, measured in drinks per week. We used PROCESS v4.3 in R to implement a heteroskedasticity-consistent standard error (HC3) and 95% bias-corrected bootstrapping procedure with 5,000 iterations.

Results: The most frequently reported stressor was moving to a new address (30% of women), although 23% of women endorsed someone close to them having a problem with drinking or other substances. Having someone with whom to talk was the most frequently endorsed support mechanism (85%) while having someone to ask for a loan was the least endorsed of these indices (80%). The number of alcoholic drinks ranged from < 1 drink per week to greater than 14 drinks. 77% of women endorsed drinking < 1 weekly drink in their third trimester while 23% of women reported weekly drinking. Women with low (N = 0 stressors), average (1–3 stressors), and high (4+ stressors) stress reported an average of 1.34, 1.32, and 1.47 alcoholic drinks per week, respectively, and reported receiving an average of 3.30, 3.39, and 3.08 social support behaviors, respectively, during their third trimester. Neither weekly drinks nor social support differed statistically by stress level [Pillai’s trace = 0.012, F(2,370) = 1.156, p = 0.329). Importantly, however, the moderation model was significant [F(3,369) = 4.99, p = 0.0021), R2 = 0.045], with a significant Social Support-by-Stress interaction [b = −0.399, t(369) = −2.930, p = 0.0036]. As hypothesized, conditional effects indicated that at low levels of social support, highly stressed women endorsed more weekly third trimester alcoholic drinks compared to women with low stress. At high levels of social support, however, there was no difference in the number of weekly alcoholic drinks by stress level in pregnant women.

Conclusions: These results suggest that the protective effect of social relationships against stress-related drinking extends to pregnant women during the third trimester. Importantly, they highlight the importance of examining higher-level interactive effects on alcohol use, as neither drinking level nor social support differed by stress when examined separately; yet the interaction between social support and stress significantly predicted drinking level. There was also considerable heterogeneity in the patterns of alcohol use, stress, and social support. Although most women endorsed low-level third trimester use, almost a quarter of women reported weekly alcohol use. Additionally, women endorsed receiving both instrumental and emotional support as well as having experienced variety of stressors, including moving to a new address and having someone close to them have problems with alcohol and substance use. Together, these data illustrate the highly complex system through which both stress and social support affect third trimester drinking. Prenatal alcohol use interventions should account for the dynamic interaction and heterogenous presentations of stress and social support in pregnant women.

Keywords: Alcohol consumption, pregnancy, Prenatal alcohol exposure, Women's Mental Health.

Disclosure: Boston Scientific, Employee, Spouse/Partner.

P717. The epigenetic signature of gestational stress across the postpartum in female wistar rats

Fiona Hollis, Erin Gorman-Sandler, Alaina Mullaly, Nazharee Cloude, Gabrielle Wood

University of South Carolina School of Medicine, Columbia, South Carolina, United States

Background: Stress during pregnancy is a well-documented risk factor for postpartum depression (PPD), a disorder affecting 1 in 5 new mothers. PPD is characterized by mood disturbances, anhedonia, impaired maternal care, and, in severe cases, suicidal ideation. Despite its prevalence, the neurobiological mechanisms underlying PPD remain poorly understood. Mitochondria are dynamic organelles that are essential for energy production and inflammation, but also provide the intermediate metabolites required for generating and modifying epigenetic marks. Epigenetic modifications, such as histone acetylation and methylation, play a critical role in stress-related gene transcription and are implicated in depressive disorders. We hypothesized that gestational stress will alter histone modifications within the PFC across the postpartum, contributing to a PPD-relevant phenotype.

Methods: Nulliparous and time-mated adult female Wistar rats (n = 7–10/group) were divided into stress and control groups. On gestational day 10, stress females were exposed to 10 consecutive days of chronic mild unpredictable stress. Control groups were handled. PPD-relevant behaviors were assessed across the postpartum. Histone H3 acetylation and methylation, and plasma cytokine levels were measured in PFC of females at mid- or late postpartum using immunoblots, and ELISA, respectively.

Results: Gestational stress decreased histone acetylation (p = 0.03) and tended to decrease histone methylation in the mid-postpartum (p = 0.08). In the late postpartum, stress tended to increase histone methylation (p = 0.06), while histone acetylation levels were unaffected. Neither parity nor stress affected total histone H3 levels at either time point. Parity increased plasma levels of interleukin 1-beta (IL-1b; p = 0.0006) that was associated with decreased histone acetylation levels in the mid-postpartum (R2 = 0.14, p = 0.03). In the late postpartum, while IL-1b levels were not significantly different across groups, interleukin 2 (IL-2) levels were increased (p = 0.02) and associated with decreased histone acetylation (R2 = 0.26, p = 0.04).

Conclusions: We showed that gestational stress significantly altered histone modifications in the prefrontal cortex in the mid-postpartum in association with pro-inflammatory cytokine levels and altered postpartum behavior.

Keywords: Gestational stress, Histone acetylation, Pro-inflammatory cytokines, maternal behavior.

Disclosure: MitoQ, Contracted Research, Self.

P718. Prefrontal delta dynamics support the translation of reward value into goal-directed behavior in major depressive disorder

Timothy McDermott, Zoe Steelman, Justin Riddle

Florida State University, Tallahassee, Florida, United States

Background: Anhedonia, a diminished capacity to experience pleasure, is a debilitating and treatment-resistant symptom of major depressive disorder (MDD). Current first-line interventions focus on reducing negative affect, with little impact on restoring reward motivation. This therapeutic gap stems from an incomplete understanding of how the brain translates reward signals into goal-directed behavior. The reward-related positivity (RewP), an electrophysiological signal generated in medial prefrontal cortex (PFC) during reward receipt, is blunted in individuals with anhedonia. However, RewP magnitude alone does not predict subsequent behavior, limiting its translational value. In a series of recent studies, we demonstrated that the motivation to exert effort is modulated by the reward or punishment of the previous trial. Reward-responsiveness, defined as increased willingness to exert effort after reward and decreased effort after punishment, was negatively correlated with anhedonia symptom severity. However, the mechanism by which reward value are translated into goal state is unknown.

Methods: Data from three projects were pooled together for this analysis. Each of the studies was approved by the institution review board at Florida State University and written informed consent was acquired. 110 participants completed an adaptive version of the Expenditure of Effort for Reward Task (Adaptive-EEfRT) while high-density electroencephalography (EEG) was acquired. In the Adaptive-EEfRT, participants choose between exerting more physical effort for a chance at a high monetary reward or less effort for a low reward. A button is pressed with the pinky finger many times for high physical effort exertion and less times with the index finger for low effort. Upon successful completion, a reward is provided at 50% chance. If the participant chose high effort on >85% or <15% of trials for a block, then the difficulty is incremented or decremented. Of the 110 participants, 6 were excluded for noisy EEG data and 9 were excluded because behavior never stabilized between 15% and 85%. Time-frequency analysis for the remaining 95 participants was run on the left prefrontal and frontocentral regions to investigate the role of delta oscillations (2–3 Hz) for win versus loss following reward-receipt and post-win versus post-loss during decision-making. In addition, functional connectivity in the delta band was estimated using the left prefrontal cortex as a seed during decision-making using weighted phase lag index to investigate the presence of connectivity with medial prefrontal cortex.

Results: Time-frequency analysis revealed a significant increase in delta oscillatory coherence for wins versus losses (t(94) = 2.29, p = 0.024, d = .24) in frontocentral electrodes, but not left lateral prefrontal cortex (t(94) = 1.59, p = 0.12, d = .16). During decision-making, we found a marginally significant increase in delta oscillatory power post-win versus post-loss in the left lateral prefrontal cortex (t(94) = 1.853, p = 0.067, d = .19), but not frontocentral electrodes (t(94) = 1.03, p = 0.31, d = .11). These results are consistent with a model that dissociates the roles of medial prefrontal cortex to process reward-outcome and lateral prefrontal cortex to translate reward-outcome into goal-directed behavior. For the connectivity analysis, we found a selective increase in delta-frequency functional connectivity between left lateral prefrontal cortex and medial prefrontal cortex for post-wins greater than post-loss. This finding suggests that delta-frequency functional connectivity might serve as the mechanism for the translation of previous reward-outcome into decision-making.

Conclusions: Delta oscillatory activity was increased for receipt of reward and during subsequent decision-making following reward, which suggested that delta rhythms function as a carrier signal within prefrontal circuits to coordinate the translation of reward value into goal-state. Future work should further investigate whether functional connectivity between lateral and medial prefrontal cortex is reduced in people with major depressive disorder and symptoms of anhedonia. By developing more accurate neural models for how prefrontal delta dynamics are altered as a function of symptom expression, the identified electrophysiological mechanisms could be targeted with next-generation non-invasive brain stimulation for the treatment of anhedonia.

Keywords: Reward, motivation and Anhedonia, reward response, decison-making, neural oscillations, Depression.

Disclosure: Nothing to disclose.

P719. Basal ganglia dopamine availability and immune activation interactively relate to anhedonia severity among youth with depression

Iris Ka-Yi Chat, Kalyan Tripathy, Tina Gupta, Zachary Brodnick, Neil Jones, Helmet Karim, Finnegan Calabro, Rebecca Price, Jihui Diaz, Beatriz Luna, James Murrough, Scott Russo, Erika Forbes

University of California, Los Angeles, Los Angeles, California, United States

Background: Anhedonia emerges in adolescence, occurs across serious psychopathologies, has putative substrates in neural reward and dopamine systems, and is a hallmark of poor clinical course in depression. Based on psychoneuroimmunology models of psychopathology, pathways involving reward and immune alterations are potential mechanisms of its development. However, most studies overlook crosstalk between reward and immune function and also fail to take into account the multiple aspects of anhedonia. Pathways are postulated to involve immune activation impairing dopaminergic function that subserves approach motivation. Thus, immune-reward associations may be relevant to anticipatory and not consummatory aspects of anhedonia.

Methods: This cross-sectional study involved 55 youth with depression (15–25 years; Mage = 21.4 years; Female = 75%). Plasma was assayed for relative quantification of 92 immune proteins, followed by dimensionality reduction via principal component (PC) analysis, forming four PC scores. Dopamine availability was defined as basal ganglia tissue iron, quantified using MRI R2’, a measure detectable in youth. Anhedonia was assessed using the Snaith–Hamilton Pleasure Scale (SHAPS), thought to capture general anhedonia and the Positive Valence Systems Scale (PVSS), which includes a general anhedonia score and subscales for anticipatory and consummatory anhedonia. Covariates included biological sex, body mass index, age, and psychotropic medications.

Results: Higher immune PC1 scores were related to greater SHAPS overall anhedonia (p = .017) and anticipatory anhedonia (p = .027). PC1 scores interacted with basal ganglia dopamine availability: at higher PC1 levels, reflecting general immune activation, lower basal ganglia dopamine availability was associated with greater SHAPS overall anhedonia (p = .019) and anticipatory anhedonia (p = .041). Lower immune PC2 scores, which appear to reflect regulatory control or balance within inflammatory and antiviral processes, were associated with higher PVSS overall anhedonia, anticipatory anhedonia, and consummatory anhedonia (p’s < .048).

Conclusions: These findings underscore nuanced neuroimmune pathways underlying anhedonia among youths. They are consistent with inflammation-based models of depression, which emphasize the interplay between dopaminergic and immune systems. Such mechanisms may vary by anhedonia components.

Keywords: inflammation, dopamine, anhedonia, reward, motivation.

Disclosure: Nothing to disclose.

P720. Intravenous ketamine improves social anhedonia in patients with treatment-resistant depression and suicidal ideation

Hiroe Hu, Cristan Farmer, Alaina Tillman, Laura Waldman, Elizabeth Ballard, Carlos Zarate

National Institute of Mental Health, Bethesda, Maryland, United States

Background: Social anhedonia is a state or trait of reduced ability to experience pleasure with social and interpersonal interactions. While social anhedonia historically has been studied in schizophrenia-spectrum disorders, more recently, it has gained attention as a transdiagnostic clinical phenotype also associated with mood disorders and suicide risk. No effective treatments have previously been reported in the literature for specifically addressing social anhedonia in the context of depression, though entactogens may be a class of compounds that is a theoretically plausible candidate to target social anhedonia.

Recent evidence suggests ketamine to have potential entactogenic effects in humans, defined as any eliciting changes in the perception of self/other, reward processing in altruistic behavior and/or prosocial drive. To further and empirically investigate the entactogenic effects of ketamine, we aimed to analyze the changes in social anhedonia in participants with mood disorders and suicidal thoughts, who received repeated doses of open-label ketamine infusions. We hypothesized that 1) ketamine will improve social anhedonia in a dose-dependent manner, 2) that the improvement in social anhedonia would be independent from the general antidepressant effect, and 3) mediated by ketamine’s transient dissociative effects altering self- and other-perception.

Methods: Participants enrolled in an inpatient clinical trial, Neurobiology of Suicide (NCT02543983) between 2015 – 2025 and received up to five open-label infusions of intravenous racemic ketamine administered at 50 mg/kg, administered 3–4 days apart. Snaith-Hamilton Pleasure Scale (SHAPS) measuring anhedonia and the Montgomery-Asberg Depression Rating Sale (MADRS) measuring depression severity were collected at −60min and +230min, relative to each ketamine infusion. A subset of four items from the SHAPS pertaining to social interactions and altruism were calculated as social anhedonia scores, such as “I would enjoy being with my family or close friends” and “I would get pleasure from helping others.” These items were selected a priori based on face validity and theoretical alignment with constructs of social connectedness and prosocial motivation, in order to capture the aspects of anhedonia most relevant to interpersonal pleasure and altruistic behavior.

The Clinician-Administered Dissociative States Scale (CADSS) was administered at +40min immediately following the completion of the first ketamine administration, when dissociative effects are most experienced. A subset of ten items from the CADSS were used to derive a summary score reflecting alterations in self- and other-perception. Similar to the four items selected from the SHAPS, these items from the CADSS were selected to reflect a theoretical construct related to self-other boundary dissolution occurring during dissociative experiences, which may underlie the neural process for facilitating empathic and prosocial behavior.

Linear mixed models were used to test the above hypotheses, with fixed effects for timepoint, infusion number, and their interaction. Age and sex were entered as fixed covariates in all models. A random subject-level intercept was used to account for repeated measures. An uncorrected alpha value of 5% was used for these exploratory analyses.

Results: A total of 282 observations from 28 participants (age M±SD = 40 ± 14 years; n = 15, 54% female) were entered into analysis. Post-ketamine social anhedonia significantly declined by 1.0 [95% CI: −1.8, −0.2] point after the first infusion (t = −2.5, p = 0.01) and this improvement was maintained but did not increase significantly after repeated infusions. However, after adjusting for MADRS Total Score, the effect of ketamine on social anhedonia was no longer different from zero. While increased CADSS Self-Other score was related to lower Social Anhedonia scores overall (B [95% CI] = −0.1 [−0.2, −0.02], p = .02), we found that it did not mediate ketamine’s effect on Social Anhedonia.

Conclusions: Intravenous ketamine improved social anhedonia, though the additional improvements with repeat doses was non-significant. Contrary to our hypotheses, ketamine’s effect on social anhedonia was not independent from its general antidepressant effect, nor was it mediated by ketamine’s dissociative properties pertaining to perception of self and others. This, however, does not rule out that ketamine’s effect on social anhedonia is pharmacologically distinct from the antidepressant mechanisms, and should be investigated further using neuroimaging correlates or peripheral markers tied to social affiliative behavior (such as oxytocin). Future studies should also replicate analyses using double-blinded studies to overcome the limitation of an open-label study design and incorporate multimodal outcomes, such as neuroimaging with behavioral tasks like social affective forecasting and a more extensive panel of self-report questionnaires measuring social anhedonia.

Keywords: social anhedonia, Depression, IV ketamine, self-other processing.

Disclosure: Nothing to disclose.

P721. Predator odor threat exposure disrupts natural reward processing: investigating the role of L1

Amelia Cuarenta, Margie Stricklin, Kevin Mesape, Debra Bangasser

University of Michigan, Ann Arbor, Michigan, 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 = .006, η2 = .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 = .025, η2 = .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 = .633, η2 = .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: Reward self-administration, Adversity, sex difference, L1 retrotransposons.

Disclosure: Nothing to disclose.

P722. Therapeutic potential of psychedelics in a mouse model of opioid use disorder

Megan Francis, James Welsh, Peter Kayastha, Anna Tsyrulnikov, Alexander Smith

Medical University of South Carolina, Charleston, South Carolina, United States

Background: The US is in the midst of an opioid abuse and overdose epidemic, which has been declared a public health emergency. Psychedelics have recently emerged as promising therapeutics for depression, post-traumatic stress, and substance use disorders. Preclinical models of addiction commonly study relapse induced by drug-associated cues, or by acute stressors. However, in clinical populations cues associated with prior life stress also commonly trigger relapse to drug seeking, and this is not commonly studies in preclinical models. Here we employed a novel mouse model in which restraint stress is paired with a conditioning odor (stress-CS), and following heroin self-administration and extinction exposure to stress-CS exacerbated relapse. We then tested the hypothesis that the 5-HT2A agonist DOI would reduce reactivity to stress-CS and reinstatement of heroin seeking. Using tissue clearing and light-sheet microscopy we went on to perform exploratory studies to identify regions and circuits that underlie these behavioral effects of psychedelics.

Methods: Mice underwent three daily restraint stress sessions (2h each) in the presence of a lemongrass odor, and stress-naïve control mice were exposed to the same odor in the home cage. One week later mice were implanted with jugular catheters, and trained to nose-poke to self-administer heroin (.05–0.1mg/kg/infusion). Following 14 days of self-administration, mice began 14 days of extinction training during which nose-pokes had no programmed consequences. Immediately following extinction sessions 8–10, mice were injected with 1mg/kg DOI or vehicle. Four days following the last DOI injection, reinstatement was induced by a combination of drug-associated cues and stress-CS. Following reinstatement, mice were perfused and brains were cleared and immunostained for c-Fos, allowing identification of brain regions affected by reinstatement and/or DOI treatment. To identify brain regions that are responsive to both acute restraint stress and re-exposure to stress-CS, FosTRAP-tdTomato mice underwent the same stresss conditioning paradigm, and were injected with 4-hydroxytamoxifen (4-OHT) immediately following the first restraint stress session. Seven days later, mice were re-exposed to stress-CS, and then were perfused and brains were stained for both tdTomato, tagging cells activated by acute stress, and c-Fos to label cells activated by stress-CS. Python code was written to identify double-labeled cells in the intact brain in a high-throughput manner. Finally, we performed anterograde circuit tracing from the anterior insular cortex (aIC) in FosTRAP mice to label aIC projections that are activated by acute DOI injection. The TrailMap Python package was used to identify eYFP+ axons, and ClearMap was used to register coordinates to the Allen Brain Atlas.

Results: DOI treatment during extinction significantly reduced reinstatement of heroin seeking (F(1,26) = 8.889, p = .0062), and also attenuated reinstatement-induced c-Fos expression in the nucleus accumbens, prelimbic cortex, and aIC. Whole-brain analysis of double-labeled tdTomato/c-Fos+ cells revealed a significant region*stress interaction (F(3,12) = 9.159, p = .002), and post-hoc tests showed a significant increase in double-labeled cells in the aIC as well as the central amygdala (CeA). Analysis of projections from the aIC that are activated by acute DOI exposure is ongoing, but preliminary results indicate significant projections to several thalamic nuclei, the infralimbic cortex, and claustrum.

Conclusions: These data employ a novel mouse model to examine the intersection of posttraumatic stress and opioid use disorder. We show that stimuli associated with prior life stress significantly increases drug seeking, and that this is associated with an increase in c-Fos in several brain regions implicated in both addiction and stress responses. We show that the psychedelic DOI is able to reduce drug seeking in response to stress-CS, and identify the anterior insula as a brain region that may be an efficacious target for therapeutics to reduce relapse to heroin seeking.

Keywords: Psychedelics, opioid use disorder, heroin seeking.

Disclosure: Nothing to disclose.

P723. Dopaminoceptive cells in the ventral hippocampus control cocaine action

Veronika Kondev, Arthur Godino, Brian Kipp, John F. Fullard, Vishwendra Patel, Sarah Naguib, Adam Ripp, Jacob Abroon, Elizabeth Kahn, Earnest P. Chen, Clementine Blaschke, Angelica Minier-Toribio, Molly Estill, Evelyn Hennigan, Alexa LaBanca, Leanne Holt, Tamara Markovic, Li Gan, Panagiotis Roussos, Eric J. Nestler

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

Background: Dopamine signaling facilitates outcome valence, reward prediction error, and motivation; thus, dysfunctional dopamine signaling has been tied to several neuropsychiatric disorders, including substance use disorder (SUD) and depression. While most research has focused on striatal DA signaling, our lab and others have validated that the ventral hippocampus (vHPC) also exhibits topographical organization of mostly non-overlapping dopaminoceptive neuronal populations: those expressing the DA receptor type 1 (D1) or 2 (D2).

Methods: We used fiber photometry, optogenetics, chemogenetics, and single nuclei RNA-sequencing (snRNA-seq) from D1- or D2-Cre mice to explore how dopaminoceptive cells in the vHPC control cocaine action. Male and female mice were exposed to cocaine conditioned place preference (CPP) or intravenous self-administration (IVSA). 2-Way or 3-Way ANOVA was used to analyze behavioral data, with sex as a biological variable.

Results: We reveal that vHPC D1 cell inhibition is necessary and sufficient for reward-context associations driving cocaine CPP and cue-induced reinstatement of drug seeking in self-administration models. On the other hand, vHPC D2 cell activation promotes positive reinforcement and exacerbates voluntary cocaine consumption and drug seeking. snRNA-seq reveals potential molecular mechanisms, including endocannabinoid system elements, that might contribute to these effects.

Conclusions: Together, these studies broaden our understanding of the role of DA within the larger mesocorticolimbic system and reveal new neuronal and circuit mechanisms underlying how drugs of abuse promote strong associative learning processes and facilitate drug seeking and taking characteristic of SUD.

Keywords: cocaine, addiction, motivation, Reward.

Disclosure: Nothing to disclose.

P724. Social and emotional stress influence decision-making toward reward seeking under threat in male mice

Rodolfo Flores Garcia, Sergio Iniguez

The University of Texas at El Paso, El Paso, Texas, United States

Background: Effective decision-making requires balancing reward seeking with threat avoidance. Stress-related psychiatric disorders are characterized by disruptions in this balance, often shifting behavior toward maladaptive reward seeking or excessive avoidance. In this study, we examined how distinct forms of stress affect approach–avoidance conflict resolution in mice.

Methods: Male mice were exposed to 10 days of social defeat stress (SDS) or vicarious defeat stress (VDS) before training in the platform-mediated avoidance (PMA) task. The first phase of PMA consisted of reward training, during which mice learned to obtain Ensure solution by nosepoking into a reward port during the presentation of a 30-s light cue illuminating the port. In the subsequent phase of the PMA task, conflict was introduced by presenting a 30-s tone that co-terminated with a 2-s footshock. Shocks were avoidable by mounting a platform located away from the reward port. The tone was presented with full, partial, or no overlap with the light cue, generating five pseudorandomly presented trial types: light-only (reward-only), tone-only (avoidance-only), co-presentation (simultaneous conflict), and sequential light→tone and tone→light (approach–avoidance conflict). Reward training consisted of five daily sessions, followed by five days of conflict training, both with access to a 50% Ensure solution. Mice were then given a reward re-exposure session with 100% Ensure in the absence of conflict (reward-only trials), followed by five additional days of conflict training with 100% Ensure.

Results: By day 5 of the reward phase, control, SDS-, and VDS-exposed mice showed comparable levels of nosepoking to the light cue. However, during conflict with 50% Ensure, both SDS- and VDS-exposed mice exhibited greater reward approach than controls (p < 0.05). When the reward was switched to 100% Ensure, control mice escalated reward seeking to levels comparable to those of SDS- and VDS-exposed mice.

Conclusions: Altogether, these results suggest that social and emotional stress shape motivational conflict resolution. Specifically, compared to controls, mice exposed to chronic stress displayed increased approach toward lower-value rewards in the face of threats. Thus, chronic stress appears to alter cost–benefit evaluation in favor of risk-taking, providing a foundation for investigating mechanisms by which stress exposure might promote maladaptive reward seeking and risk-prone behavior relevant to stress-related disorders.

Keywords: approach-avoidance conflict, Chronic social stress, Vicarious defeat stress, Preclinical models and endpoints.

Disclosure: Nothing to disclose.

P725. Juvenile immune activation modulates stress effects on cognitive function and microglia in adulthood

Rachel Rahn, Divija Chopra, Gabriela Manzano Nieves, Christopher N. Parkhurst, Vivian Block, Conor Liston

Weill Cornell Medical College, New York, New York, United States

Background: Epidemiological datasets and genome-wide association studies suggest a possible link between childhood infectious illness and adult propensity to mood disorder diagnosis. However, little is known regarding how exactly alterations to the immune system affect long-term stress susceptibility. We therefore aimed to characterize how juvenile immune activation influences stress responses in adulthood, and test for effects on cognitive behavior, microglial function and neural activity in related cortical regions.

Methods: Our two-hit model consisted of an immune activation during pre-adolescent development (intraperitoneal injection of lipopolysaccharide (LPS) at postnatal day 21, 100 μL/kg body weight), followed by 7-day adult mild unpredictable stress. Behavior and flow cytometry cohorts consisted of adult C57BL/6J mice (n = 7–17 per group); widefield calcium imaging cohorts consisted of adult Thy1/GCaMP6s animals to characterize excitatory pyramidal neuronal activity (n = 3–8 per group). All groups were balanced for sex. LPS+stress mice were compared to stress only, LPS only, and saline control mice, except where indicated otherwise. To examine changes to synaptic phagocytic capacity of microglia, we labeled neuronal material that had been consumed by microglia by injecting virus (AAV-PHP.eb-CAG-tdTomato) retro-orbitally in C57BL/6J mice at 8 weeks of age. After 7-day stress, we perfused mice at 10 weeks of age with PBS and performed a flow cytometry experiment using the left (whole) brain hemisphere and gating on live, CD45int+, CX3CR1+, CD11b+, Tmem119+ cells to evaluate microglial phagocytic function by quantifying tdTomato mean fluorescence intensity.

Fixed-ratio operant and variable effort reward behavioral assays were coupled with widefield calcium imaging to assess developmental LPS and adult stress effects on cognitive circuits. The operant conditioning used a 3-day fixed ratio schedule of reinforcement (FR1), while the 8-day variable effort reward task delivered a sucrose water reward of variable size dependent on proportionally increased licking during the tone period compared to the pre-tone intertrial interval (ITI). Operant task-based neural activity changes were assessed by widefield calcium imaging of excitatory pyramidal neurons in the adult mouse cortex. Task-based imaging of FR operant behavior was performed on LPS+stress and control mice and GLM analysis performed on task epoch data to analyze beta weight differences related to reward period, licking, tone and trial predictor variables. Widefield calcium imaging data was collected using sequential LED illumination and sCMOS detection fluorescence and analyzed in MATLAB. Operant task data included all trials completed during a 30-minute period. Student’s t test was performed on task-based imaging data and 2-way ANOVA on all 4-group behavioral data, with Tukey’s post hoc tests performed to evaluate group differences. Statistical analyses were performed in PRISM.

Results: Flow cytometry analysis detected a long-lasting increase in adult microglial phagocytic activity after P21 LPS compared to controls (main effect of LPS p = .001), confirming that LPS at P21 has a long-term effect on adult microglial phagocytosis related to synaptic pruning. As reported in our 2024 ACNP results, LPS+stress decreased the number of trials completed in an operant reward assay with a fixed ratio schedule of reinforcement (p = .0056), amplifying the subthreshold effect of stress only (p = .076). Building upon this finding and the detection of frontoparietal functional connectivity differences in our previous resting-state studies, GLM analysis of LPS+stress vs. control neural activity during operant reward behavior also identified significant region-specific differences in excitatory neuronal activity, including in frontal and parietal cortex. This suggests that immune activation amplifies stress-induced deficits in reward behavior, and LPS+stress effects are observable in neural activity during cognitive behavior.

Because non-specific licking during the inter-trial period of the operant task was also reduced in LPS+stress mice, we examined group differences in the variable effort task involving tone and pre-tone ITI periods. After the initial learning period (Day 1–4) during which the mice learned to increase their licks during the tone period compared to ITI, the ratio of licks categorized as high reward level compared to very low reward level was significantly decreased in LPS+stress, LPS only and stress only groups compared to controls, with mean licks during ITI showing a main effect of stress (stress p = .044, LPS p = .14).

Conclusions: Our findings demonstrate that developmental immune activation and adult stress impact microglia populations and cognitive function, including effortful reward-seeking behavior. LPS and stress both alter microglia phagocytotic capacity, while LPS+stress mice also display a reduction in non-specific (ITI) licking as well as task-related tone licking compared to controls, suggesting they may have reduced motivation to lick due to the effects of LPS and stress. These results indicate that juvenile immune activation and adult stress both contribute to microglia dysfunction and cognitive deficits in reward behavior, and widefield imaging results demonstrate a corresponding change in frontal cortex activity.

Keywords: Reward seeking, neuroimmune activation, in vivo calcium imaging, Chronic unpredictable mild stress.

Disclosure: Nothing to disclose.

P726. Genomic architecture of action impulsivity, attention, and social reinforcement in heterogeneous stock rats

Paul Meyer, Christopher King, DJay Federico, Thiago Sanches, Apurva Chitre, Connor Martin, Oksana Polesskaya, Keita Ishiwari, Leah Solberg Woods, David Dietz, Abraham Palmer

University at Buffalo, Buffalo, New York, United States

Background: Our findings demonstrate that genetic variation in impulsivity, attention, and social motivation contributes to vulnerability for substance use disorders. By identifying conserved loci and candidate genes, this work bridges rodent behavioral genetics with human psychiatric risk and highlights new targets for prevention and treatment.

Methods: We conducted genome-wide association studies (GWAS) in >2,800 outbred heterogeneous stock (HS) rats across two behavioral domains. First, a reaction time task measured attentional performance (reaction speed, omissions, variability) and action impulsivity (premature responses, false alarms). Second, an operant three-port task measured social reinforcement, in which rats earned access to a familiar conspecific. Low-coverage whole genome sequencing was combined with linear mixed models for GWAS, and brain-region–specific expression QTL (eQTL) analyses were integrated to prioritize candidate genes.

Results: Reaction time and impulsivity traits were moderately heritable (h² ≈ .19–.22), with 22 unique quantitative trait loci (QTLs) identified on chromosomes 1, 2, 5, and 14. Candidate genes included Fancl and Vrk2 (implicated in human GWAS for smoking, depression, antidepressant efficacy), as well as Crebbp and Trap1, which showed eQTLs with the mesocorticolimbic system. Social reinforcement was also moderately heritable (h² ≈ .07–.25), with sex-dependent QTLs identified on chromosomes 5 (males) and 10 (females), encompassing candidate genes such as Oprd1, Mecr, and Ipcef1. Males showed stronger relative preference for social stimuli, while females showed higher overall responding, consistent with partially dissociable genetic influences.

Conclusions: These findings identify shared and sex-specific loci underlying impulsivity, attention, and social reinforcement in HS rats. Integration with eQTL data highlights candidate genes with translational relevance to addiction and psychiatric disorders. Ongoing work is testing whether these loci also contribute to drug-related behaviors, providing a genetic bridge between intermediate traits and addiction vulnerability. Funding Support: DA037844, DA060810, DA060669.

Keywords: endophenotypes, sex differences, behavioral genetics, Risk factors, Neuropsychiatric Disorders.

Disclosure: Nothing to disclose.

P727. The role nociceptin opioid peptide in dorsal raphe nucleus on appetitive motivation

Laura Masso-Quinones, Kathryn Braden, Daniel Castro

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

Background: In the United States, approximately 59.3 million people experience mental illness across their lifetime. Many neuropsychiatric disorders are characterized by dysregulation of affective or motivated circuits in the brain. The dorsal raphe nucleus (DRN) is considered one of the most important brain regions in modulating affective and motivated behaviors. This brain region is a frequent focus of neuropsychiatric research due to its role as the primary source of serotonergic neurons. However, the DRN is not only a major serotonergic hub, but also a region with high expression of endogenous opioids, including dynorphin, enkephalin, and nociceptin. Despite the established behavioral significance of the endogenous opioid system on motivation and affect, its function within the DRN, particularly that of the nociception opioid peptide, has been largely understudied. Considering that nociceptin opioid peptide receptors are highly expressed on DRN serotonin neurons, it is likely that DRN nociceptin may play an important modulatory role on affective and motivated behaviors.

Methods: To investigate the role of nociceptin in DRN, we used a cre-dependent CRISPR-Cas9 mediated knockdown approach in Pnoc-cre mice (5–12 mice/genotype/sex). All behavioral tests were counterbalanced with several days between each test. When possible, mice were tested within subjects. For sucrose preference testing, mice were habituated to two water bottles in the home cage for 3 days. One water bottle had water, the other was empty, and placement of the bottles was alternated each day. On days 4 and 5, one bottle was filled with water, and the other with 1% sucrose, placement counterbalanced across each day. Total sucrose and water consumed on each day was recorded and percent preference was averaged across the two days. For sucrose self administration, mice were individually housed and exposed to a Feeding Experimental Device (FED) for 24 hours. The FED delivered sucrose pellets ad libitum. Chow and water were also available in the cage ad libitum. On days two through four, mice were trained on FR1, FR3, and FR5 operant responding for sucrose. On day 5, mice were tested on progressive ratio breakpoint. 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 CRISPR knockdown of nociceptin in DRN reduced sucrose preference (Welch’s t-test, between subjects, t(16,9) = 2.565, p = 0.0167, cre- n = 10, cre+ n = 17). In contrast, CRISPR knockdown increased operant responding for sucrose, particularly on FR5 criteria (Two-Way ANOVA, Sidak’s multiple comparison; F (1, 6) = 14.63, p = 0.0087; t(18) = 2.742; p = 0.0134, Cre- n = 4, cre+ n = 4). Although no statistically significant, there was a trend toward a higher breakpoint in CRISPR knockdown mice (t(3,3) = 1.711, p = 0.1694, cre- n = 4, cre+ n = 4).

Conclusions: Overall, we found that loss of nociceptin opioid peptide in dorsal raphe nucleus oppositely impacted affective versus motivated behaviors. Specifically, we found that although mice seemed to find sucrose less preferable, they were willing to work harder to obtain the reward. Such a discrepancy has been observed in reward deficit models, perhaps suggesting that mice are willing to work harder to achieve a certain amount of overall subjective reward. These results have major implications for substance use disorder and mood disorders. Future work will use DREADDs to manipulate these nociceptin neurons, and biosensor imaging via NOPLight to determine when nociceptin peptide is being released.

Keywords: Nociceptin/orphanin FQ, Reward, motivation and Anhedonia, Alcohol and substance use disorders.

Disclosure: Nothing to disclose.

P728. Cocaine intravenous self-administration in near-isogenic spontaneously hypertensive rat substrains and QTL analysis of cocaine locomotor sensitivity and gene expression (RNA, protein) in an F2 cross

Britahny Baskin, Yahia Adla, Emma Sandago, Jacob Porter, Mengyuan Liu, Rylin Lubash, Abraham Kielar, Nina Garbarino, Hong Choi, Bridgette Reilly, Caroline Topping, Grace Whitney, Lara Saba, Hao Chen, Andrew Emili, Kathleen Kantak, Camron Bryant

Northeastern University, Boston, Massachusetts, United States

Background: Psychostimulant use disorders (PUDs) have resurged with increased use and overdose deaths, even without polysubstance use with opioids. Despite exhibiting a ~40–50% heritability, very few PUD-associated loci have been identified. Quantitative trait locus (QTL) mapping in F2 crosses between near-isogenic rodent substrains showing variation in addiction-relevant behavioral and molecular traits facilitates identification of causal genes. We reported enhanced cocaine locomotor sensitivity and self-administration, impulsivity, and compulsivity in SHR/NCrl vs. SHR/NHsd rat substrains purchased from their respective vendors (Charles River Laboratories, Harlan-Innotiv Laboratories). Here, we sought to replicate these observations using substrains bred in-house to further implicate genetic contributions, and to rule out environmental explanations (e.g., differences in vendor conditions). Furthermore, we generated and tested a population comprising a reciprocal F2 reduced complexity cross (RCC) to map QTLs underlying trait variation in novelty response and cocaine locomotor stimulant sensitivity, as well as RNA (eQTLs), and protein expression (pQTLs) from striatum and frontal cortex.

Methods: All experiments were conducted in strict adherence to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committees at Boston University and Northeastern University. Female and male adult rats from both substrains (SHR/NCrl and SHR/NHsd; ns = 5–9 per substrain per sex) were tested for the rate of spontaneous acquisition (no pellet training) of cocaine (0.3mg/kg) self-administration. To test for motivation to acquire cocaine, requirements progressed through FR1-FR5 and a progressive ratio schedule with exponentially increasing required lever presses for one cocaine infusion. Different cocaine doses (1.0, 0.1, and 0.01mg/kg) were then substituted to establish a dose-response curve and finally, rats completed extinction training and were assessed for cue-induced reinstatement. Another cohort of rats bred in-house, including both SHR/NCrl and SHR/NHsd (ns = 13–28) and F2 offspring (n = 292), underwent a two-week cocaine locomotor paradigm during which rats received saline on Monday and Tuesday, and high dose of cocaine (20mg/kg; i.p.) on Wednesday-Friday. For the second week, rats received received saline on Monday and Tuesday and repeated injections of low dose cocaine (5mg/kg; i.p.) on Wednesday-Friday. DNA from tail tissue was sequenced and analyzed with RNA and protein expression from striatum and frontal cortex of F2 rats.QTL analysis was performed in R/qtl and R/MatrixEQTL in order to triangulate on genetic loci and candidate genes underlying RNA, protein, and behavioral variation.

Results: SHR substrains did not differ in days to acquire self-administration (p = 0.87) but SHR/NCrl rats earned more infusions on FR1 (p = 0.05), FR3 (p < 0.01), and FR5 (p < 0.001) and exhibited higher active lever responding on FR3 (p < 0.05) and FR5 (p = 0.01). SHR/NCrl earned higher progressive ratio breakpoints as manifested in a higher last FR completed (p < 0.05), more infusions earned (p < 0.05), and more active lever responses (p < 0.05) vs. SHR/NHsd. Furthermore, although substrains did not differ in active lever responses following extinction (p = 0.66), SHR/NCrl showed greater reinstatement, via more active lever responses on a cue-induced reinstatement session (p < 0.05). QTL analysis of novelty-induced locomotion and cocaine locomotor stimulation identified a genome-wide significant QTL on chr 20 indicating a progressive increase in LOD score across repeated high-dose cocaine injections for cocaine-induced locomotion (LOD = 3.88; p = 0.054; 18.11Mb-53.59Mb). There were also trending QTLs on chr 9 for context-induced conditioning (locomotor behavior in response to a saline injection in a previously cocaine-paired environment; LOD = 3.5; p = 0.10; 76.98Mb-102.06Mb) and on chr20 (LOD = 3.68; p = 0.08; 2.79Mb-9.13Mb) corresponding to novelty responsivity. QTL analysis of RNA expression in striatum and frontal cortex tissue identified 34 transcripts with eQTLs in both areas (ps < 0.01). Furthermore, striatal protein QTL analysis identified 825 proteins with pQTLs (p < 0.01). Triangulating on behavior, mRNA, and/or protein led to the identification of 1 candidate gene (Shc3 at 89.00Mb) that had significant QTLs for striatal protein, striatal RNA, and FC RNA, and was located in a the region on chr9 identified by behavioral QTL analysis which correlated with cocaine context memories.Shc3 is implicated in amphetamine and nicotine SUD traits, and is predicted to be related to learning in memory and could underlie trait variation in novelty and cocaine stimulant sensitivity.

Conclusions: We found significant substrain behavioral differences on assays probing cocaine sensitivity and motivation to self-administer cocaine. SHR/NCrl rats exhibited higher active lever responding than SHR/NHsd and consistently earned more infusions across different schedules of responding replicating the substrain pro-addiction phenotype. Additionally, preliminary QTL analysis of cocaine locomotor sensitivity traits in F2 rats implicates Shc3 as a candidate gene for validation.

Keywords: GWAS, cocaine self-administration, genetics, Rodent models, Cocaine sensitivity.

Disclosure: Nothing to disclose.

P729. The explore-exploit tradeoff and model-based and model-free learning strategies predict subsequent alcohol use behaviors

Summer Thompson, Colin Johnston, Israa Mustafa, Genevieve Thibodeau, Stephanie Groman, Daeyeol Lee, Jane Taylor

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

Background: Aspects of reinforcement learning that contribute to decision making strategies are often disrupted in psychiatric disorders. Two key strategies for solving multi-stage decision-making (MSDM) problems are model-free and model-based learning, which involve choices guided exclusively by reinforcement history versus choices guided by a combination of reinforcement and state transition dynamics, respectively. The balance of model-based and model-free learning decision-making strategies has been argued to be linked to substance use disorders including alcohol use disorder. Another key axis of decision-making strategies that is altered in psychiatric disorders is the explore-exploit tradeoff. However, how these strategies in alcohol-naïve individuals relate to future alcohol use patterns is not well understood, and it is important to disentangle susceptibility from consequence in chronic substance exposure. Here, we used a MSDM task for mice to investigate how decision-making strategies under naïve conditions predict individual differences in subsequent alcohol use patterns.

Methods: Forty-seven adult male and female C57BL/6J mice were first trained on a simple version and then tested on the full MSDM task in operant chambers with 20% sweetened condensed milk as the reinforcer. Several reinforcement learning models were fit to these MSDM choice data and compared to estimate explore-exploit, model-based, and model-free contributions to decision making. Following this evaluation of alcohol-naïve decision-making strategies, mice self-administered either a solution of 10% alcohol + 0.1% saccharin or 0.1% saccharin for 4 weeks on an escalating fixed-ratio schedule (FR-1 × 7 days; FR-3 × 7 days; FR-5 × 14 days). Saccharin was removed from the reinforcer solution for the final self-administration session to test sensitivity of actions to the value of outcomes. Measures of escalation of responses, sensitivity of choices to task structure, and sensitivity to outcome value were calculated for each mouse.

Results: The best fitting reinforcement learning model incorporated several free parameters for trial type-dependent (i.e., the state transition and outcome experienced on the preceding trial) inverse temperature, which represents the degree to which agents explore versus exploit, in addition to weights for model-free and model-based action values. Female mice acquired the training version of the task faster, during which they exhibited a greater likelihood of repeating rewarded choices than males; effects that were driven by a greater likelihood of exploration following unrewarded relative to rewarded trials. During testing on the full MSDM task, no sex differences were observed in the contribution of explore-exploit tradeoff, model-free, or model-based strategies to decision-making. Patterns of self-administration varied by sex and reinforcer. Female mice consumed more and escalated responding for alcohol to a greater degree than male mice. Alcohol induced greater reinforcing efficacy and lower sensitivity to outcome revaluation than saccharin alone. Individual differences in decision-making strategies predicted future patterns of self-administration. Specifically, the propensity to exploit following unrewarded outcomes predicted greater reinforcing efficacy of alcohol, but not saccharin, whereas other explore-exploit parameters were unrelated to this measure. Model-based learning also predicted future reinforcing efficacy of alcohol, but not saccharin, within female mice. By contrast, model-free learning was not associated with reinforcing efficacy but instead predicted escalation of responding and sensitivity to outcome revaluation for alcohol, but not saccharin. Finally, mice that exhibited higher levels of exploration following unrewarded outcomes during MSDM decision-making performance were more likely to exhibit streamlined action sequences within a constant reinforcement schedule during subsequent alcohol self-administration.

Conclusions: Overall, we found that model-based and model-free learning strategies, as well as the explore-exploit tradeoff, in alcohol-naïve mice predicted future sex-specific patterns of self-administration. Our findings suggest that select aspects of decision-making strategies may serve as markers of susceptibility to maladaptive alcohol use.

Keywords: Value-based Decision-Making, Alcohol, Computational Reinforcement Learning Model, sex differences.

Disclosure: Nothing to disclose.

P730. Hugs or heroin? investigating ventral pallidal cell-types in social vs. drug choice

Erica Ramirez, Shreeya Walawalkar, Ryan Rokerya, Cyril Sumarinas, Sofia Lopez-Orraca, Anapurna Germain, Evelyn Arellano, Stephen Mahler

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

Background: Opioid addiction is a major societal problem, and we still have few tools available to prevent or treat it in those who suffer—especially when it comes to tools derived from understanding of the basic neurobiology of addiction. Part of the reason for this failure to translate neuroscience findings into the addiction clinic might be due to a lack of translationally-relevant models of opioid addiction that incorporate complexities of the disorder, such as social factors that can protect or exacerbate addiction, and which are rarely studied in preclinical studies. Indeed, social factors play key roles in the initiation and progression of addiction, and the COVID-19 pandemic laid this bare, with marked changes in drug use, abuse, overdose, and addiction occurring in the early 2020s.

Methods: Here we address this knowledge gap by asking how GABA neurons in the ventral pallidum (VP), a brain region playing a key role in food and drug reward, are involved in social reward, and social modulation of drug reward. In transgenic GAD1-Cre transgenic Long Evans male and female rats, we chemogenetically inhibited or stimulated VP GABA neurons during social behaviors, and during an operant social/drug self-administration task where rats are asked to choose between social and drug rewards. Specifically, we targeted inhibitory and excitatory DREADDs (hM4Di and hM3Dq) to VP GABA neurons, allowing bidirectional control of them during behavior. Wildtype and DREADD-expressing animals were socially isolated from their same-sex, sibling cage-mate for one week before being undergoing assessments of 1) social conditioned place preference, 2) social play and investigation, 3) social communication via ultrasonic vocalizations, 4) operant social contact self-administration, and 5) choice of social vs heroin reward. Additional experiments using c-Fos immunohistology and fiber photometry were conducted to assess VP GABA and non-GABA neuronal activity during social behaviors.

Results: Our findings demonstrate that chemogenetic manipulations of VP GABA neurons bidirectionally modulate social reward seeking, social behaviors, and communication in rats (p < 0.05). Unlike our previous findings in food vs. heroin choice, rats overall preferred heroin over social interactions and VPGABA manipulations did not seemingly shift this choice.

Conclusions: These results highlight the underappreciated role of the VP in social behaviors, and takes steps toward understanding how social factors modulate opioid drug seeking in addiction more generally. It is our hope that by untangling the neural circuit substrates involving VP that impact social and drug rewards, we may develop new neural intervention strategies for leveraging social factors in preventing or treating addiction.

Keywords: Social reward, Animal models of addiction, Ventral Pallidum, chemogenetics, Heroin.

Disclosure: Nothing to disclose.

P731. Dopamine ramps in the medial nucleus accumbens encode discounted future value on a moment-by-moment basis

Johannes de Jong, Yilan Liang, Stephan Lammel

UC Berkeley, Berkeley, California, United States

Background: Dopamine (DA) ramps are gradual increases in DA concentration observed during motivated behavior. They contrast with phasic DA release, which is typically associated with signaling reward prediction errors (RPE). While RPE-related DA transients align with traditional reinforcement learning models, where DA encodes the difference between expected and received rewards, the functional significance of dopamine ramps is less understood. We have previously shown that DA ramps occur both at the level of DA cell bodies in the ventral tegmental area (VTA) and at axon terminals in the nucleus accumbens (NAc), specifically in the medial VTA and medial NAc (NAcMed). In contrast, lateral VTA DA neurons projecting to the lateral NAc (NAcLat) encode the temporal derivative of these ramps. However, the precise information that is encoded in these ramps remains subject to intense debate.

Methods: We present results from two behavioral tasks assessing reward-seeking behavior in mice. The first is a head-fixed virtual reality task in which an abstract cue (a horizontal bar on a computer screen) denotes the distance to a reward. Mice run on a running wheel to move the bar down the screen. They learn to discriminate between two cues, indicated by a solid or dashed bar. CS+ trials result in reward delivery (3 μl of 1% sucrose), whereas CS- trials end without reward delivery, either when the mouse completes the trial distance or after a fixed interval expires. During this task, we use GRABDA-mediated fiber photometry (FIP) to record simultaneous DA release in the NAcMed and NAcLat.

The second task is a freely moving foraging paradigm. Here, mice are placed in a large (150 cm diameter) octagonal arena in a dark room. They learn to collect rewards from eight different ports when availability is signaled by a cue LED above the port. We then use wireless FIP in the NAcMed and NAcLat to monitor DA release as mice approach the reward.

Results: In both behavioral tasks we observe robust DA ramping in the NAcMed as mice approach the reward, which cannot be modeled using the TD-RPE framework. Instead, these DA ramps reflect the temporally discounted expected trial outcome, specifically, the value of the reward discounted by time-to-reward, defined as distance to reward divided by current locomotor speed. This indicates that the animals take their own behavior into account when estimating discounted expected reward. In contrast, DA release in the NAcLat shows transient peaks at cue onset and reward delivery, corresponding to the temporal derivative of discounted value, a continuous assessment of TD-RPE that does not depend on discrete state evaluations.

Conclusions: This parallel encoding of future expected value and its temporal derivative challenges current views of how the mesolimbic DA system supports motivated behavior and reinforcement learning. By reframing TD-RPE as an emergent property of a broader control process, our findings clarify how dopamine integrates past experience with ongoing outcomes to guide future behavior. This perspective refines reinforcement learning models and offers potential insights into disorders of motivation, including addiction and eating disorders.

Keywords: Dopamine, reward prediction error, motivation, Reinforcement learning.

Disclosure: Nothing to disclose.

P732. The effects of intermittent access to oxycodone intravenous self-administration and co-use of alcohol on the motivation to seek oxycodone in rats

Lori Knackstedt, Aaron Claypool, Sydney Dick, Barry Setlow, Linda Cottler, Marek Schwendt

University of Florida, Gainesville, Florida, United States

Background: Opioid use remains a nationwide public health emergency in the US. Co-use of alcohol with opioids worsens treatment outcomes and increases the likelihood of overdose. Preclinical rodent models are essential to uncover the neurobiology of drug-seeking in the polysubstance use condition. We conducted interviews with 141 persons using opioids alone or with alcohol to assess past-30 day daily and hourly patterns of opioid and alcohol use, finding that the majority of individuals using opioids do so “intermittently” in one-hour increments 1–3 times per day, with inter-opioid intervals approximating 1.5 half-lives of the drug of choice. Alcohol use typically occurs after opioid use. Prescription opioids such as oxycodone were the most prevalent opioids reported. Here we back-translated these real-world patterns of opioid-alcohol polysubstance use (PSU) into rats and assessed the impact of such patterns on the motivation to seek intravenous oxycodone.

Methods: Male and female rats (n’s 10–12/condition) were trained to self-administer intravenous oxycodone (0.1 mg/kg/infusion) for 3 hr/day for three days. Rats assigned to the “continuous” group experienced an additional 21 days of intravenous self-administration (IVSA) on the same schedule. Rats assigned to the “intermittent” group also experienced an additional 21 days of oxycodone IVSA, but with three 1 hr daily sessions, each separated by 2 hours. A subset of rats was permitted 5 hours access to alcohol in the home cage at the conclusion of daily IVSA sessions. A dose-response curve for oxycodone IVSA was then determined over the course of 5 days, followed by self-administration on a progressive ratio schedule for two consecutive days. Next, after re-establishing IVSA, rats experienced 14 days of drug abstinence prior to a 1-hr cue-primed relapse test.

Results: The intermittent schedule of oxycodone IVSA resulted in a trend for higher oxycodone consumption relative to the continuous schedule (p = 0.06), and a significant upward shift in the dose-response curve for oxycodone IVSA [F (1.899, 24.69) = 7.437, p < 0.01). There were no group differences in breakpoint for oxycodone on a progressive-ratio schedule and cue-primed relapse was similar across groups.

Conclusions: In rats, the intermittent pattern of oxycodone self-administration that is seen in humans is accompanied by increased reinforcing efficacy of intravenous oxycodone. However, this does not translate into greater motivation to seek oxycodone or oxycodone-associated cues. Future work will examine whether such patterns, including oxycodone-alcohol polysubstance use alter the neurobiology underlying oxycodone-seeking.

Keywords: Addiction comorbidity, Animal models of addiction, Opioid addiction.

Disclosure: Nothing to disclose.

P733. Vta GABA, not dopamine, neuron activity mediates social buffering of stress

Urszula Skupio, Djemila Compaore, Nekhama Riznyk, Alexander Harris

Columbia University, New York State Psychiatric Institute, New York, New York, United States

Background: Stress is a leading risk factor for psychiatric disorders, and its impact is exacerbated by social isolation. Stress causes reward processing deficits, a hallmark of psychiatric illnesses that is often resistant to treatment. Social interaction is a potent buffer of stress, but the neural and behavioral mechanisms of social buffering remain unclear. Objectively assessing the effects of social buffering in mice on reward processing deficits has so far been challenging. To address these gaps, we developed an experimental framework to test social buffering and uncover its underlying neural mechanisms in mice.

Methods: We employed a novel, reward-based behavioral paradigm that combines a Pavlovian reward task, restraint stress, and social interaction with familiar partners, to quantify social buffering in mice. We combined this approach with fiber photometry recordings of ventral tegmental area (VTA) dopamine and GABA neuron activity as well as optogenetic manipulations to investigate how social experiences shape stress-related behavior through the reward circuitry.

Results: We found that brief social interaction with a familiar, unstressed partner after acute restraint restores reward-seeking behavior in both male and female C57BL6/J mice (3-way ANOVA, Fstress = 13.95, Fsi = 9.415, Pstress < 0.001, Psi < 0.01, Psex = 0.31, PstressXsi < 0.01, N = 13–18 mice/group, 50% female). Given prior evidence implicating VTA DA and GABA neurons in stress-induced reward blunting, we examined their activity during social interactions with either familiar partners. We found that VTA GABA, but not DA neurons exhibited distinct activity patterns during interactions with familiar partners under control and stress conditions (N vgat = 7 mice **P < 0.01, N DAT = 6 mice P > 0.05, Wilcoxon rank sum). Furthermore, brief optogenetic inhibition of VTA GABA, but not stimulation of VTA DA neurons after restraint, mimicked the stress-buffering effects of social interaction, suggesting a key role for VTA GABA neurons in mediating these effects.

Conclusions: Together, these findings highlight the critical involvement of VTA GABA neurons in facilitating the recovery from stress through social experiences. These results provide new insights into the neural mechanisms underlying stress-buffering effects and suggest potential targets for therapeutic interventions.

Keywords: Social buffering, Reward, motivation and Anhedonia, neural circuitry.

Disclosure: NeuroKaire, Board Member, Self.

P734. Mounting motivation when the time is right: circadian and internal state influences on timed reward seeking

Eloise Kuijer, Qijun Tang, Hamid Taghipourbibalan, Sofia Shirley, James Edgar McCutcheon, Samer Hattar, Hugo Tejeda, Paule Joseph

NIH/NIAAA, Bethesda, Maryland, United States

Background: Despite advances in understanding acute anticipatory reward-seeking mediated by discrete cues, mechanisms driving internally generated circadian-based reward-seeking are unknown. Food anticipatory activity (FAA) is characterized by heightened activity in anticipation of scheduled feeding and reflects behaviour driven by circadian influences on interoceptive states in the absence of external cues. A current gap in knowledge is whether FAA reflects non-specific circadian-driven locomotion or a state of heightened motivation. Here, we tested the hypothesis that FAA driven by timed operant palatable meals represents a state of heightened motivation that is modulated by internal state (satiety vs hunger).

Methods: C57/BL6j mice (n = 23; 12F/11M) were trained to nosepoke for chocolate rewards on an FR1 schedule using Arduino-based Feeding Experimentation Devices v.3 (FED3) during a one-hour timewindow in the animals inactive cycle (ZT 18–19; lights off). FAA is measured by running wheels to capture activity before the feeding period. Motivation over the circadian cycle was measured progressive schedules during 1 hr semi-exponential progressive ratio (PR) sessions at the following times of the day: during the FAA window (ZT 17–18), active phase (ZT 5–6) or inactive phase (ZT 14–15), before and after taking away chow. A separate cohort of mice (n = 21; 10F/11M) underwent the same PR testing paradigm but either got trained on one-hour reward-window at fixed (ZT 18–19) or random times during the light-on phase (ZT 12–24) (n = 10–11/group). Brains were taken from these animals for whole brain clearing and cFOS analyses. A third cohort (n = 12; 6F/6M) was trained on the same FR1/PR schedule, but initially time food restricted and acutely stated.

All animals underwent all PR times and data was analysed using repeated measures ANOVA or mixed model when due to technical errors where trial data was missing. Wheel running was analysed using paired t-test comparing revolutions during FAA period versus inactive phase.

Results: Fixed, but not random, time-restricted palatable foods promote a high motivation state characterised by increased motivation to run on a running wheel and effort-based reward responding, dependent on internal state. Sated mice exhibit minimal anticipatory wheel running before timed feeding and effort-based motivation (F(1.9, 30.1) = 0.71, p = 0.50). However, fasted mice express anticipatory wheel running ramping up rotations by 40.8-fold during FAA compared to inactive phase (p = 0.0002). They also express heightened effort-based motivation in the PR test during the FAA time window (F(1.7, 29.3) = 25.2, p < 0.0001) with increasing the breakpoint during FAA compared to inactive and active cycle by 75.5% (p < 0.0001) and 38.8% (p = 0.0004), respectively. Interestingly, learned anticipatory behaviour in sated mice that never demonstrated FAA is rapidly unmasked by acute hunger prior to post-hunger timed reward suggesting satiety masks a latent self-generated motivated state. The control of internal state over FAA is bidirectional: a separate cohort of mice that are trained under food restriction and subsequently sated acutely lost FAA. Mice under random timed rewards increase locomotor activity in response to hunger but fail to organize a heightened motivational state to obtain rewards (p = 0.83). Analyses show that increased motivation during anticipatory behavior differs from circadian-dependent shifts in motivation state. Preliminary studies suggest differential brain-wide activity, including in frontalcortical-ventral striatal circuits, assessed using cFOS differs in ventral striatal circuitry between mice over the course of the FAA period, which is an internally-generated motivation state and that genetic ablation of D1 receptors in ventral striatal circuits impairs FAA.

Conclusions: These results suggest that internal state gates anticipation to timed rewards, which is dynamic in nature and when ungated produces an emergent state of enhanced motivation in absence of external cues. Together, these findings identify a mechanism by which circadian processes can autonomously enhance motivational states, which bears implication for disorders characterised by dysfunctional motivation, interoception, and circadian function, such as depression and substance use disorders.

Keywords: motivation, circadian, emergence, food reward, wheel running.

Disclosure: Nothing to disclose.

P735. Accumbal dopamine integrates information about salient stimuli and actions to promote learning in dynamic environments

Stephanie Cajigas Gabriel, Sophia Dunkin, Michael Leonard, Devan Gomez, Hannah Elam, Soren D Emerson, Maxime Chevee, Erin Calipari

Vanderbilt University, Nashville, Tennessee, United States

Background: Cue-outcome associations are formed based on an animal’s — often incomplete — understanding of relationships between cues, actions, and outcomes. Through their actions, animals acquire new information about these relationships, allowing them to update how predictive cues influence behavior. This process is particularly relevant in stress-related disorders, where inaccurate interpretations can lead to learned helplessness (behavior characterized by a failure to avoid aversive stimuli, even when escape is possible).

Methods: The optical dopamine sensor dLight, in conjunction with in vivo fiber photometry, was used to monitor real-time dopamine responses in the nucleus accumbens core (NAc) during aversive learning. A negative reinforcement paradigm consisted of 15, 30-minute training sessions. Each trial began with a 15-second tone and lever presentation, followed by a train of five 0.5-second, 0.5mA shocks, separated by 1 second. Mice could respond at any time from cue onset to the last shock presentation (i.e, during the discriminative cue). After which, a house light went on (also presented with a lever press and cessation of future shocks), termed the trial termination cue or cue-end. There was a variable ITI (30–90sec), then a new trial began. We timestamped each component of the behavior task (cues, shocks, and conditioned responses). This allowed us to obtain precise temporal information about when behavior occurred, along with a custom pipeline for robust and reproducible analysis. To determine the causal nature of trial termination cue-evoked dopamine on learning phenotype, excitatory optogenetics was performed to excite the terminals of dopaminergic axons that project from the ventral tegmental area to the nucleus accumbens core. A 1s, 20Hz (470nM, ~8mW) laser pulse was paired with either the discriminative cue, the trial termination cue following no response, or the trial termination cue following a response. Sample sizes ranged from 21 to 7 mice, depending on the experiment, and included both males and females.

Results: We found that dopamine release to the discriminative cue increased in animals that learned to avoid footshocks; however, this only occurred after performance was maximal, suggesting it did not drive cue-based predictions that developed during learning. Instead, the key predictive signal was dopamine release to the outcome (the trial termination cue), which was strongest early in training in animals that learned to avoid shock and declined over time—but only when it followed an action. This signal did not reflect reward or relief but rather was evoked by salient events that guided the updating of behavior across a range of conditions. Furthermore, optogenetic stimulation of dopamine at this time point enhanced learning, but only when the signal followed an action and not to the same trial termination cue at any other time in the task.

Conclusions: Our work illustrates that behavioral changes dependent on learning ability and changes in dopamine signaling over time may underlie how a subject navigates their environment and perceives the availability of behavioral outcomes (ability to escape shocks versus perceived inescapability). These findings identify a dopamine-driven mechanism for updating cue-outcome associations based on how organisms interpret environmental structure and control, providing insights into how these processes may break down in stress-related psychiatric disorders.

Keywords: negative reinforcement, In vivo fiber photometry, optogenetics.

Disclosure: Nothing to disclose.

P736. Adolescent oral contraceptive drug exposure alters alcohol consumption in mice

Laurel Seemiller, Damini Nair, Camilla Hughes, Nicole Crowley

Penn State University, University Park, Pennsylvania, United States

Background: Most women in the United States have used the oral contraceptive pill in their lifetime, and this medication is particularly common among adolescents and young adults. Oral contraceptive pills typically contain synthetic estrogen and progestin components, and related endogenous hormones have been shown to influence alcohol-seeking in humans and mice.

Methods: Here, we tested how a mouse model of combined oral contraceptive usage in adolescence and young adulthood affected alcohol consumption across two paradigms: drinking in the dark (DID) and two-bottle choice (2BC). From PND 43–74, female C57BL/6J mice voluntarily consumed daily combined contraceptive drugs (ethinyl estradiol and levonorgestrel suspended in a 250 μl droplet of 10% sucrose in H2O) or a hormone-free control solution (10% sucrose in H2O). On PND 50–74, mice received access to alcohol through either intermittent access DID or continuous access 2BC paradigms. Drinking data were analyzed via 2-way (treatment, session) mixed model ANOVA.

Results: We found that oral contraceptive exposure reduced binge-like alcohol consumption across four cycles of DID (n = 6, F1,10 = 5.841, p = 0.036). Similarly, in a separate cohort of mice, 2BC testing revealed dramatically reduced alcohol preference throughout contraceptive treatment (n = 9, F1,16 = 294.1, p < 0.0001). Ongoing experiments are testing the role of ovarian steroidogenesis in mediating these behaviors.

Conclusions: Overall, our data suggest that adolescent oral contraceptive usage alters alcohol-seeking behavior and may have overlooked systemic effects on female physiology.

Keywords: hormonal contraceptive use, Adolescent Alcohol, Neuroendocrine, electrophysiology, Addiction-like behaviors.

Disclosure: Nothing to disclose.

P737. Dopaminergic representations of behavioral flexibility

Natalie Zlebnik, Brandon Oliver

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

Background: Midbrain dopamine (DA) signaling in the striatum is critically involved in reward-based learning and behavioral flexibility, processes essential for adapting to dynamic environments and implicated in psychiatric conditions such as substance use disorder (SUD), obsessive-compulsive disorder (OCD), and schizophrenia. Impaired behavioral flexibility is a prominent feature of various psychiatric disorders, notably substance use disorder (SUD), where persistent maladaptive decision-making contributes significantly to continued drug use and relapse vulnerability. Understanding the neurobiological underpinnings of behavioral flexibility, particularly the dynamics of striatal DA during reversal learning, is crucial because current therapeutic strategies often lack specificity, failing to adequately address the neural mechanisms driving maladaptive behaviors. Identifying precise DA signaling dynamics could enable more targeted, effective treatments for disorders characterized by impaired flexibility. We hypothesize that phasic striatal DA release during the initial stages of reversal learning facilitates updating response-outcome associations, thus improving decision-making and adaptive behavior.

Methods: We employed an operant probabilistic reversal learning (PRL) paradigm in two complementary experiments. In Experiment 1, we recorded phasic DA dynamics using fiber photometry in the dorsal striatum (DS) and nucleus accumbens (NAc) of mice expressing the genetically-encoded dopamine biosensor, GrabDA. Animals were trained to discriminate between levers with distinct reward probabilities, and once trained, the reward contingencies were reversed across the levers. Experiment 2 employed a closed-loop optogenetic approach to causally probe the role of striatal DA during reversal learning. We selectively activated DA neurons projecting from the ventral tegmental area (VTA) and substantia nigra pars compacta (SNc) to the NAc and DS, respectively.

Results: We observed dynamic alterations in DS and NAc DA signaling during the early phases of reversal, indicative of DA-mediated behavioral updating. Optogenetic stimulation during select trials during reversal learning significantly enhanced accuracy and promoted optimal choice strategies.

Conclusions: Collectively, these findings demonstrate the crucial involvement of phasic striatal DA signaling in supporting behavioral flexibility. Such insights into DA’s role in adaptive learning have significant implications for understanding and ameliorating cognitive deficits observed in disorders like SUD.

Keywords: Dopamine, Cognitive / behavioral flexibility, probabilistic reversal learning, circuit optogenetics, Neurotransmitter release.

Disclosure: Nothing to disclose.

P738. Hunger accelerates reward learning through a movement specific dopamine signal

Maxime Chevee, Megan E Altemus, Soren D Emerson, Jennifer Tat, Hye-Jean Yoon, Nevin Crow, Courtney J Kim, Tara F Chalasani, Shemuel Roberts, Stephanie Cajigas Gabriel, Michael Leonard, Ainoa Konomi-Pilkati, Olivia X Murfield, Erin Calipari

Vanderbilt University, Nashville, Tennessee, United States

Background: Modulating behavior depending on the internal state of an organism is critical to survival. For example, hungry animals seek food. In patients suffering from eating disorders, the relationship between food consumption and energy demand can become compromised and lead to obesity or anorexia. While it is intuitive that hunger drives food seeking, clinical studies and our preliminary data in mice show that it also impacts learning in non-food seeking contexts. Thus, hunger must alter general learning mechanisms. The goal of this study is to identify the neural mechanisms that allow hunger to accelerate learning.

Methods: At the center of learning and internal state-dependent modulation in behavior is dopamine release and regulation in the striatum. We first used systematic operand conditioning procedures in both male and female mice to identify the conditions under which hunger effectively accelerated learning. We then used a range of methods, including fiber photometry, voltammetry, single cell RNA sequencing and optogenetics to identify the effects of hunger on striatal dopamine and to understand how this effect accelerates learning. Our approach allowed us to compare the effects of hunger within subjects.

Results: We show that food restriction enhances performance only when learning requires actions and those actions are reinforced by appetitive outcomes - both caloric and non-caloric. This is mediated by the potentiation of a dopamine signal at movement offset that integrates information about actions and the spatial location of rewards. Hunger increases dopamine release far from locations with high reward probability to drive future approach - attracting mice toward the goal and increasing opportunities to engage. Hunger-modulated dopamine transients are specific to the dorsomedial striatum and are not a general amplification of reward-related signaling. Rather, they only emerge in contexts where reward locations are known from prior experience.

Conclusions: These findings reveal a dopamine-based mechanism through which hunger biases spatially guided approach towards rewards, thereby accelerating reinforcement learning. Disruptions in this mechanism may contribute to maladaptive learning and decision-making in psychiatric disorders such as obesity, anorexia, or addiction.

Keywords: motivation, Action-outcome associations, Dorsomedial striatum.

Disclosure: Nothing to disclose.

P739. Sex differences in the dose-dependent effects of cannabis edibles on blood cannabinoid concentrations and subjective drug effects

Justin Matheson, Madison Wright, Paulina Antwi, Adam Zaweel, Esther Kim, Pamela Kaduri, Stefan Kloiber, Ahmed Hassan, Beth Sproule, Christine Wickens, Patricia Di Ciano, Bruna Brands, Bernard Le Foll

Centre for Addiction and Mental Health, Toronto, Canada

Background: Studies in animal models have found sex differences in the endocannabinoid system, in acute responses to exogenous cannabinoid administration, and in pharmacokinetics of cannabinoids, including delta-9-tetrahydrocannabinol (THC). However, human laboratory studies have provided very mixed evidence regarding human sex differences in pharmacological responses to acute THC administration. Many of these previous studies have relied on inhaled exposure to THC, which introduces significant variability in dose of THC administered that might systematically differ by sex. The goal of this study was to examine sex differences in blood THC concentrations and self-reported subjective drug effects after exposure to a range of fixed doses of oral THC, administered as edible gummies.

Methods: The study was a within-subjects, double-blind, placebo-controlled, randomized and counterbalanced human laboratory experiment assessing the acute pharmacological effects of a range of doses of cannabis edibles. Healthy adults aged 19 to 45 years who used cannabis at least one day per week and consumed edibles at least once per month were recruited from Toronto, Canada. After screening and a practice visit, participants attended four drug exposure visits where we administered one of four dose conditions: placebo, low dose (2 mg), medium dose (10 mg), and high dose (20 mg). All participants were characterized as female or male based on self-reported sex assigned at birth. Blood samples were collected at baseline and then 2 and 5 hours after exposure to THC to measure concentrations of THC and two of its metabolites (11-COOH-THC and 11-OH-THC). Visual analogue scale (VAS) ratings of subjective drug effects were measured at baseline, immediately after consuming the edible, and then at 0.5, 1, 1.5, 2, 3, 4, and 5 hours after exposure. VAS items were measured on a scale of 0 (no effect) to 100 (maximum effect) and included “I like this drug effect” (Drug Liking), “This feels like cannabis” (Feels Like), “I feel this effect” (Effect), “I feel this high” (High), “I feel the good effects” (Good Effects), “I feel the bad effects” (Bad Effects), and “I feel the rush” (Rush). Peak blood cannabinoid concentrations and VAS ratings were determined in order to remove the dimension of time. Data were analyzed using linear mixed models (LMMs), which included participant as a random factor, along with the following fixed factors: sex (female, male), condition (placebo, low dose, medium dose, high dose), period, sequence, and the interaction of sex and condition.

Results: Forty participants (n = 20 females, n = 20 males) completed at least one drug exposure visit and were included in the analysis (50% white; mean 30.4 [SD: 6.9] years old; n = 3 non-binary participants, two assigned female at birth and one assigned male at birth). The female and male groups did not statistically differ in age, race, education, age of initiation of cannabis use, current tobacco use (yes/no), or current alcohol use (yes/no). There were no main effects of sex or sex by condition interactions for THC, 11-COOH-THC, or 11-COOH-THC in whole blood (p > 0.05). In the LMMs, there was no significant evidence of sex differences in VAS Drug Liking, Effect, High, Good Effects, or Rush. There was a significant sex by condition interaction for Feels Like (p = 0.039). Post-hoc tests indicated that male participants had a significantly higher mean peak rating of Feels Like in the placebo condition compared to female participants (mean difference = 40.8, t = 4.5, p < 0.001, Cohen’s d = 1.4). Given this large difference in mean peak rating in the placebo condition, we examined post-hoc comparisons for other VAS items and the same trend was observed for VAS Liking, Effect, High, Good Effects, and Rush, where males had significantly higher mean peak VAS ratings in the placebo group compared to females. In reviewing the allocation of study drug order, we observed that males were more likely to be randomized to a drug order that started with placebo in the first session, which could explain the higher self-reported drug effects in males in the placebo condition. Finally, there was a significant sex by condition interaction for VAS Bad Effects (p = 0.007). In the high dose condition only, females had significantly higher mean peak ratings than males (mean difference = 26.4, t = 2.6, p = 0.014, Cohen’s d = 0.8).

Conclusions: Under the present experimental conditions, male participants had greater peak ratings of subjective drug effects than female participants when administered placebo, though this was likely due to more males starting with placebo in their first drug exposure visit. There was limited evidence of sex differences in self-reported drug effects in the active THC conditions, except for greater ratings of bad drug effects in female participants at the high dose (20 mg). We found no significant evidence of sex differences in blood THC or metabolite concentrations. Overall, there was minimal evidence of sex differences in pharmacological effects of oral THC administered as gummies, though the greater ratings of bad drug effects in females at the 20 mg condition may warrant further attention.

Keywords: cannabis, sex differences, subjective effects.

Disclosure: Nothing to disclose.

P740. Sex differences in gene expression networks in the cingulate cortex of mice exposed to chronic neuropathic pain

Ana L. Simal, Giannina Descalzi

University of Guelph, GUELPH, Canada

Background: Chronic pain is a debilitative disease affecting 1 in 4 American adults, and it is a major risk factor for anxiety and depression. Despite significant advancements revealing that chronic pain corresponds with robust neuroplasticity in several brain regions of the emotion-pain brain network, our understanding of the molecular mechanisms involved in these events, and how they promote the transition from acute to chronic pain is incomplete. Consequently, there is a woeful lack of efficacious treatment for people living with chronic pain, with less than half of all people undergoing treatment reporting any control of their pain. Tragically, the rate of suicide is twice as high in the population living with chronic pain, and families of people that have died by opioid poisoning report that unrelieved pain contributed to their mental health condition. Given the current dearth of available treatments for individuals living with chronic pain and comorbid mental illnesses, there is a critical need for research into the molecular mechanisms involved in order to discover novel treatment targets and therapeutic avenues.

Neuroimaging studies of people with chronic pain provide an opportunity to employ reverse translational strategies to test hypotheses of pain chronification. Repeatedly, neuroimaging studies have identified structural and functional changes to multiple regions of the emotion-pain brain network, including the anterior cingulate cortex (ACC), nucleus accumbens (NAc), and amygdala, in people suffering from chronic pain. Accordingly, numerous studies of analogous brain regions in rodent models of chronic neuropathic pain have also revealed changes in neuronal excitability and structure. Within this network, neuroplasticity of the rodent ACC has emerged as a critical step for chronic pain development. Nevertheless, although these experiments provide insight into the plasticity-related events that are present in chronic pain states, they apply a narrow focus and biased approach, that concentrates on a relatively small fraction of molecular events associated with chronic pain. Moreover, the causal mechanisms responsible for the transition from acute to chronic pain remain elusive. RNA sequencing (RNAseq) provides an unbiased approach towards characterizing the presence and quantity of RNA in a biological sample at a given moment, providing direct evidence of the relationship between the physiological state and molecular changes in cells. Measures across different stages of pain chronification allow for the analysis of the continuously changing cellular transcriptome, thereby providing snapshots of molecular network activity across time. In this study, we use transcriptomic profiling to determine changes in gene expression networks across time in female and male mice exposed to nerve injury.

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 pain thresholds at baseline (prior to injury), and at 5, 14, 30, and 60 days post injury, collecting ACC tissue from separate groups of mice at each time point (n = 5/sex/timepoint). RNA was extracted from ACC tissue and sequenced by the Centre for Applied Genomics (SickKids, Toronto) using Illumina next-generation sequencing. Reads were then aligned to the GRCm39 reference genome, and differential gene expression was analyzed using DESeq2 (p < 0.05).

Results: Despite similar behavioral responses to SNI across sexes, results show distinct sex-specific gene expression patterns over time. When differentially expressed genes (DEGs) were compared female versus male, within the same timepoint an interesting pattern was observed. At acute timepoints (5 days), similar DEGs that were upregulated in females, were down regulated in males, showing an inverse correlation. In contrast, direct correlations were observed at more remote timepoints following SNI. These findings indicate that neuropathic pain induces sexually dimorphic transcriptional responses in the murine ACC.

Conclusions: Robust transcriptome-wide changes are observed in the cingulate cortex of female and male mice exposed to nerve injury. Initial gene expression networks differ across sexes at acute time points post nerve injury, with increased overlap in expression patters as female and male mice spend more time experiencing pain. These findings suggest that different treatment targets may be optimal depending on time post-injury, with distinct targets for females and males at early time points. Our results highlight the importance of including both females and males in chronic pain research and contribute to deepening our knowledge of molecular changes underlying pain chronification.

Keywords: chronic pain, Anterior Cingulate Cortex (ACC), sex-differences, RNAseq.

Disclosure: Nothing to disclose.

P741. A longitudinal reference map of sex-biased development of regional brain volume in mice

Linh Pham, Autessa Anoosheh, Simon Beggs, Lisa Bradley, Lindsay Cahill, Mallar Chakravarty, Tiffany Chien, Jacob Ellegood, Myrto Lavda, Gail Lee, Brian Nieman, Mark Palmert, Stephen Scherer, John Sled, Bilal Syed, Lily Qiu, Armin Raznahan, Jason Lerch

National Institute of Health/NIMH, Bethesda, Maryland, United States

Background: Neurodevelopmental conditions have differential risk profiles based on sex, but how the developmental influences of sex interact with the brain to alter these risks remains difficult to determine. Identifying brain regions exhibiting normative sex differences and mapping their sex-specific developmental trajectories is an important first step towards understanding how sex affects neurodevelopmental conditions. Because sex differences in the brain are often subtle and entangled with gender effects - requiring large sample size to be reliably assessed - accomplishing this goal with existing human neuroimaging databases is challenging. Given that mice and humans share some sex biases in the brain, a high-powered, longitudinal mouse neuroimaging study presents a suitable alternative for studying the underlying principles governing sex differences in regional brain size and developmental patterns. Moreover, mouse models enable experimental parsing of mechanisms for these sex differences in a way that is not possible in humans.

Methods: Volumetric measures were obtained for 590 regions from 1,417 longitudinally acquired structural MRI scans in wild-type and control mice from postnatal day 3 to 182 (females: n = 655; males: n = 762). Regions with volumetric sex differences in adulthood and volumetric developmental trajectories were identified using generalized additive models with total brain volume as a covariate. Developmental trajectory sex differences were k-means clustered to identify sets of regions that share similar patterns of sex differences over time.

Results: Volumetric sex differences in developmental trajectories are present in 42% of the mouse brain regions examined (q < 0.05). Many of these regions, such as the somatosensory and somatomotor areas, were not previously identified as having sex differences in their developmental trajectories. Some of the regions exhibiting developmental trajectories sex differences also show volumetric sex-biases in adult mice, such as the olfactory areas. Other regions exhibiting both of these characteristics, such as the bed nuclei of the stria terminalis and the medial amygdala, are also hot spots for volumetric sex biases in humans. Sex differences trajectories can be formed into several distinct groups based on k-means clustering.

Conclusions: In this high-powered, longitudinal mouse imaging study, we identified new regions and patterns of sex differences in the mouse brain by asking which regions show sex differences in developmental trajectories and clustering those findings. We provide a developmental timetable for the emergence of adult sex-differences in regional brain volume – establishing an important empirical foundation for probing potential intersections between sex differences and developmentally operative risk factors for mental illness.

This research was supported [in part] by the Intramural Research Program of the National Institutes of Health (NIH). The contributions of the NIH author(s) were made as part of their official duties as NIH federal employees, are in compliance with agency policy requirements, and are considered Works of the United States Government. However, the findings and conclusions presented here are those of the author(s) and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services.

Keywords: sex-differences, neuroimaging, Brain development, mouse brain.

Disclosure: Nothing to disclose.

P742. Assessing cell type-specific cannabinoid CB1 receptors in psychosis and THC exposure using postmortem human brain samples and rodent models

Shinnyi Chou, Sonia Borker, Alaina Buerger, Tanvi Challa, Jamie Durbin, Shih-Hsuan Ku, Hruthika Lingutla, Jason Newman, Abigail Obeng, Mary Torregrossa, Robert Sweet

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Schizophrenia and cannabis use appear to be closely linked, and alterations in the cannabis receptor target, the cannabinoid CB1 receptor, have been implicated in the pathophysiology of both disorders. CB1 is ubiquitous across cell types. However, its alteration has not been investigated in a cell type-specific manner across the two disorders. Given the unique contribution of CB1 on functional outputs in different cell types, assessing cell type-specific CB1 alterations in psychosis and THC exposure may increase our understanding of CB1’s role in the pathophysiology of each disorder individually and combined, and provide potential therapeutic targets.

Methods: Postmortem brain samples of subject pairs with and without (n = 10) schizophrenia (n = 10 pairs) were investigated for CB1 levels in excitatory and inhibitory neurons through multi-label immunohistochemistry. A separate subject cohort consisting of control, schizophrenia-only, and schizophrenia with cannabis use disorder (n = 16 triads) were also investigated for cell type-specific CB1 levels. Rats subjected to cell type-specific CB1 level alterations and THC self-administration were additionally assessed for working memory to determine functional significance of cell type-specific CB1 alterations. Analysis of variance and covariance (ANOVA and ANCOVA) were used to analyze cell type-specific CB1 levels and behavioral measures.

Results: CB1 level was significantly higher in all inhibitory neurons compared to excitatory neurons, p < 0.001. In schizophrenia samples, CB1 level was 35.5% higher in excitatory neurons compared to unaffected individuals, while CB1 level was 14.9% lower in inhibitory neurons compared to unaffected individuals, p < 0.001 for both. Increased excitatory CB1 in conjunction with THC exposure disrupted working memory performance in rodents.

Conclusions: Examining the cannabinoid CB1 receptor through a cell type-specific lens in psychosis and THC exposure identified unique patterns of receptor alterations, suggesting a more nuanced approach is necessary when considering its possibility as a therapeutic target in the future.

Keywords: Cognitive impairment associated with schizophrenia, Schizophrenia and related disorders (SRD), cannabis, cannabinoid receptor type 1.

Disclosure: Nothing to disclose.

P743. Investigating the association of tandem repeat variation with schizophrenia in whole genome sequenced cohorts

Celine Manigbas, Bharati Jadhav, Paras Garg, Andrew Sharp, Rebecca Birnbaum

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

Background: Rare tandem repeat expansions are known causative factors for over 50 disorders, including Huntington’s Disease and spinocerebellar ataxias, while common tandem repeat variation is increasingly being identified as significantly associated with complex disease and gene regulation. Though previous one-off studies suggest a role of TRs in schizophrenia (SCZ), large-scale systematic and genome-wide analyses have not yet been reported. Overall, the past dearth of investigation of TRs in SCZ is partly due to technical difficulty in high-throughput TR genotyping, as well as limited access to whole genome sequenced cohorts. Therefore, in the current analysis, we leveraged the UK biobank whole genome sequenced (WGS) cohort and a recently developed next generation sequenced based algorithm to investigate TRs for SCZ association.

Methods: Within the UK biobank 1,176 individuals with SCZ were identified as well as 150,660 neurotypical controls, matched for age, sex and European ancestry. Highly polymorphic, genome-wide TRs were genotyped using the ExpansionHunter algorithm, an advantageous approach that can detect both common TR variation as well as rare expansions that may exceed the 100–150bp read length. Of 165,000 genotyped TRs, 3,767 TRs are within 100 kilobases of SCZ GWAS risk loci, overlapping previously reported SCZ risk loci. REGENIE was used to perform association analyses, regressing SCZ status against TR variant copy number, utilizing mean TR copy number for common variant analyses and outlier status, or length of the longest allele, for analyses of rare expansions.

Results: While common variation in TR elements did not surpass genome-wide significance for SCZ association, top-ranked TRs within 100kb of two SCZ GWAS loci included a poly(AC) intergenic repeat between MFAP3 and GALNT10 (p = 1.07 × 10⁻⁴) and a poly(TG) intronic repeat in NALCN (p = 1.14 × 10⁻⁴), a gene encoding a voltage-independent, cation channel that also regulates neuronal resting membrane potential and excitability. Interestingly, for the TR expansion analyses, at a 99.9th percentile outlier threshold, the GCCCCG (6-mer) intronic repeat in C9orf72, known to be pathogenic for ALS/FTD, was identified for significant association with SCZ (p = 2.5 × 10⁻⁵, FDR = 0.04), with 7 individuals harboring a TR expansion ranging from 52–80 repeats.

Conclusions: The systematic study of both common and rare TR variation in SCZ has a very high likelihood of uncovering novel genetic risk factors, with potential diagnostic and therapeutic implications. Future analyses will increase scale and power, by expanding the current TR analysis to additional SCZ cases and neurotypical controls within other multi-ancestry WGS cohorts, including the NIH AllofUs biobank and the Genomic Psychiatry cohort. Specifically, the rare C9orf72 expansion currently identified as associated with SCZ, will be queried in other datasets. In addition, future analyses will leverage fine-mapping approaches to discern and compare the effect of TR elements within SCZ GWAS loci to already reported single nucleotide risk variants.

Keywords: Schizophrenia (SCZ), whole genome sequencing, tandem repeat variation.

Disclosure: Nothing to disclose.

P744. Mechanistic and translational roles of PTPRG: prefrontal knockdown drives behavioral activation phenotypes in mice and reveals sex-specific alterations in schizophrenia and bipolar disorder

Uma Chatterjee, Michael Cahill

University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin, United States

Background: Protein Tyrosine Phosphatase Receptor Type G (PTPRG) is a receptor-type phosphatase implicated in synaptic organization and highly enriched in cortical regions. GWAS have identified PTPRG variants associated with schizophrenia, bipolar disorder, major depressive disorder, autism spectrum disorder, ADHD, and neuroticism, yet causal mechanisms and translational data remain limited. Because orbitofrontal (OFC) and medial prefrontal (mPFC) circuits govern valuation, compulsivity, and executive control, we tested whether prefrontal Ptprg loss alters behavior. We also quantified PTPRG protein levels in prefrontal cortical homogenates in subjects with schizophrenia and bipolar disorders versus matched controls.

Methods: Mice (OFC cohort; males; females forthcoming): C57BL/6J; stereotaxic AAV-Ptprg-shRNA-GFP vs scrambled-shRNA-GFP targeted to OFC (independent non-behavior cohort for knockdown validation). Behavioral battery: Open Field (locomotion, spontaneous grooming), Object In Place, Nestlet shredding, Dig test, Splash/Spray groom, Marble burying, Elevated Zero Maze (EZM), Y-maze, Social Preference, Social Novelty, Forced Swim Test; Knockdown and control groups were compared with the appropriate two-group statistical tests for each measure. mPFC cohort collection/scoring is ongoing with the same design.

Human post-mortem Western blots: DLPFC and VLPFC from schizophrenia, bipolar with psychosis, bipolar without psychosis, and matched controls, in males and females; within-sex ANOVAs with Dunnett's T3 multiple comparisons analyses per dataset.

Results: Target engagement (mouse OFC): Western blots validated Ptprg shRNA knocked down Ptprg protein levels (non-behavior cohort, hemispheric within-subject comparisons; n = 10 vs 10, p < 0.0001).

Behavior (mouse OFC; males, knockdown n = 11 vs control n = 10): Open Field showed a hyperactivity-like profile with higher total distance (p = 0.0005), outer-zone distance (p = 0.0002), average speed (p = 0.0005), total time mobile (p < 0.0001), and outer-zone mobile time (p < 0.0001). EZM also showed a hyperactivity-like profile, indicated by both increased exploration and risk-assessment (more head dips at 5 min, p = 0.0326; more risk assessments at 10/15 min, p = 0.0201/0.0240; more entries at 15 min, p = 0.0383; trends at 5–10 min, p≈0.07–0.09). Grooming assays showed lower latency to groom in the Splash/Spray test (p < 0.0001) and altered bout duration (p = 0.0039) with a trend for total duration (p = 0.0686); spontaneous grooming differences were observed after controlling for locomotion. Nestlet tests showed increased behavioral activation with reduced latency (p = 0.0401) but no differences in nestlet shredding. Social Preference showed hyperactivity phenotypes through group differences in both social investigation time (p = 0.0476) and trending empty chamber investigation time (p = 0.0966). Clean negatives included Dig test (all p ≥ 0.25), Marble burying (p = 0.4816; 0.7583), Forced Swim (latency and immobility windows p = 0.958/0.849/0.353/0.413), and Y-maze alternation (p = 0.3045) with a trend toward more arm entries (p = 0.0824).

Human post-mortem tissue: In males, DLPFC showed a diagnostic effect driven by schizophrenia vs control (p = 0.0434), while bipolar disorder (collapsed across ± psychosis) vs control was null (p = 0.6997; N = 40; 24 controls, 16 bipolar). Conversely, in male VLPFC, bipolar (collapsed) differed from controls (ANOVA p = 0.0208; N = 41; 26 controls, 15 bipolar), whereas schizophrenia vs control was not significant (p = 0.7221; n = 24 vs 26); within bipolar, BP-psychosis vs Ctrl was significant (p = 0.0090; n = 6 vs 26) and BP+psychosis vs Ctrl trended (p = 0.0803; n = 8 vs 26). In females, DLPFC showed no group differences (p = 0.9456; 0.9407; N = 18–26) and VLPFC was also null (p = 0.976; 0.9704; N = 18–27), yielding a sex- and region-specific diagnostic pattern (male DLPFC schizophrenia effect; male VLPFC bipolar effect; females null).

Conclusions: Validated Ptprg knockdown in OFC yields a behavioral activation/hyperactivity signature with increased exploration/risk-assessment and lowered grooming thresholds, while classic anxiety-/perseveration-style assays and FST are negative, thus constraining the phenotypical scope. Human Westerns indicate region- and sex-specific alterations: male DLPFC shows a schizophrenia-related difference (with bipolar vs control null), male VLPFC shows a bipolar-related reduction (significant in BP-psychosis; BP+psychosis trending), and female DLPFC/VLPFC are null. Ongoing mPFC mouse work and addition of female cohorts will further elucidate circuit/sex specificity and whether frontal PTPRG modulation contributes to compulsivity/activation phenotypes and psychosis-spectrum biology.

Keywords: Schizophrenia (SCZ), Bipolar Disorder, compulsivity, psychosis, Hyperactivity.

Disclosure: Nothing to disclose.

P745. The emergence of distressing paranoia and hallucination psychotic experiences throughout early adolescence: contributions from genetic liability and medial prefrontal cortical thinning

Benson Ku, Ying Xu, Yinxian Chen, Anke Huels

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

Background: Hallucinations and delusions are often grouped as positive symptoms of psychosis. However, recent evidence suggests that they may be driven by distinct genetic and neural mechanisms. This study characterizes the trajectories of paranoid-like and hallucination-like experiences and identifies the pathways through which polygenic risk for schizophrenia (PRS-SCZ) and medial prefrontal cortex (mPFC) may contribute to these trajectories.

Methods: This study used data from the Adolescent Brain and Cognitive Development (ABCD) study version 5.1. Participants aged 9 to 10 at baseline were recruited from 21 sites in the US. PRS-CSx-meta was used to calculate multi-ancestral PRS-SCZ and summary statistics of schizophrenia from European, East Asian, and African subpopulations were extracted from the latest schizophrenia Genome-Wide Association Study (GWAS). Cortical thickness was computed for bilateral mPFC regions using FreeSurfer from T1 MRI images. Group-based trajectories were identified for paranoia and hallucination psychotic experiences from ages 9 to 13. The structural equation model examined the associations between PRS-SCZ, mPFC cortical thicknesses, and trajectories of distressing psychotic experiences. Models adjusted for age, sex, principal components for PRS-SCZ, intracranial volume, and study sites.

Results: This study included 11666 participants (52.1% male; mean [SD] age at baseline: 9.9 [0.6] years). The subgroups of early adolescents who had increasing paranoia (n = 329) and hallucination psychotic experiences (n = 323) were modestly correlated (r = 0.211). Higher PRS-SCZ was significantly associated with reduced cortical thickness in the right frontal pole (β = –0.028, 95% CI [–0.041, –0.013]; p = 0.002), which in turn was associated with increasingly distressing paranoia (aOR = 0.95, 95% CI = [0.91, 0.99], p = 0.044). Reduced thickness in the right superior frontal cortex was associated with increasingly distressing hallucination (aOR = 0.93, 95% CI = [0.90, 0.98], p = 0.006).

Conclusions: There may be distinct genetic and neural pathways driving the emergence of paranoia-like and hallucination-like experiences. Future work should examine the relationship between the mechanisms underlying these symptoms and their utility for diagnosis and treatment.

Keywords: psychotic-like experiences, Polygenic scores, Structural MRI.

Disclosure: Nothing to disclose.

P746. EEG signatures of ultra-treatment-resistant schizophrenia

Andor L. Bodnár Bodnár, Yifan Mo, Claire McCaulley, Arlene Cuerdo, Lilian Zhong, Robert Zeigler, Yael Ben Chaim, Edward Golob, Russell Margolis, Sridevi Sarma, Frederick Nucifora

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

Background: Schizophrenia is a complex, heterogeneous neuropsychiatric disorder affecting approximately 1% of the population. Around 70% of patients respond to standard antipsychotics (non-treatment-resistant schizophrenia; non-TRS), while the remaining ~30% do not and are classified as treatment-resistant schizophrenia (TRS), typically requiring clozapine. Notably, up to 30% of TRS patients fail to respond even to clozapine, forming a subgroup termed ultra-treatment-resistant schizophrenia (UTRS). UTRS represents one of the most severe illness trajectories, with persistent symptoms despite clozapine treatment. We hypothesize that UTRS is characterized by a distinct electrophysiological signature that can be differentiated from TRS, non-TRS patients, and healthy controls using conventional EEG.

Methods: o test this hypothesis, we applied a novel auditory spatial Stroop EEG paradigm to examine event-related potential (ERP) responses and spectral features across four groups: non-TRS (N = 10, PANSS = 41.4), TRS (N = 14, PANSS = 48.4), UTRS (N = 13, PANSS = 83.8), and healthy controls (HC, N = 10, PANSS = 30.5). ERP analysis focused on P300 components associated with attentional control. Resting-state EEG was also assessed during eyes-open and eyes-closed conditions.

Results: In the auditory spatial Stroop experiment, UTRS patients showed increased P300 amplitude (330–660 ms) at mid-central sites (p = 0.003) compared to TRS, non-TRS, and HC, suggesting compensatory attentional engagement. Resting-state EEG revealed that UTRS patients exhibited elevated low gamma power (31–40 Hz; p < 0.05), distinguishing them from all other groups across resting-state conditions.

Conclusions: These findings highlight a distinct electrophysiological signature of UTRS, characterized by elevated frontal gamma power that correlates with greater symptom severity and impaired cognition. Conventional EEG provides a practical and scalable means of identifying neurobiological markers of UTRS. By differentiating UTRS from TRS and non-TRS patients, spectral biomarkers may help stratify patients earlier in the course of illness, inform treatment decisions, and guide the development of novel interventions.

Keywords: Treatment resistant schizophrenia, Ultra-treatment resistant schizophrenia, EEG, Clozapine.

Disclosure: Nothing to disclose.

P747. Brain immune alterations in first-episode psychosis: a combined analysis using CSF samples and [18F]DPA-714 PET imaging

Yuya Mizuno, Ines Carreira Figueiredo, Toby Pillinger, Guy Hindley, Luke Baxter, Sita Parmar, Maria Lobo, Jacek Donocik, Oliver Hargrove, Giulia Gambino, Sami Jeljeli, Joel Dunn, Alexander Hammers, Ramla Awais, Kerstin Sander, Erik Årstad, Julia Schubert, Mattia Veronese, Federico Turkheimer, Valeria Mondelli, Melanie Hart, Andrew Church, Tiago Reis Marques, Oliver Howes

Institute of Psychiatry, Psychology and Neuroscience, Kings College London, London, United Kingdom

Background: Immune dysfunction has been hypothesized to underlie inadequate response to antipsychotic treatment in schizophrenia. To test this, we carried out a cross-sectional study to compare cytokine levels in the cerebrospinal fluid (CSF) of patients with first-episode psychosis who were presenting with ongoing symptoms despite antipsychotic treatment, relative to healthy volunteers. Moreover, we examined associations between CSF cytokine levels and a brain imaging marker of 18-kDa translocator protein (TSPO); a protein highly expressed by microglia in neuroinflammatory conditions.

Methods: This case-control study was carried out as part of the Inflammatory Response in Schizophrenia (IRIS) study (NCT03093064). We recruited patients with a diagnosis of DSM-5 schizophrenia or related psychotic disorder, who had less than 5 years of illness duration. All patients were symptomatic despite receiving stable doses of dopamine blockers. Healthy volunteers, who were matched for age and sex, were also recruited. All participants received a lumbar puncture under fasting conditions. We used the MSD V-PLEX Proinflammatory Panal 1 assay to quantify a total of 10 CSF cytokines. Participants also received [18F]DPA-714 PET scans with total grey matter TSPO binding quantified using the distribution volume ratio (DVR).

Results: CSF samples were collected from a total 43 patients with first-episode psychosis (median age 24, 70% male, median duration of illness 20 months, median PANSS total score 58), and 25 healthy controls. The patient group had significantly higher concentrations of pro-inflammatory cytokines in the CSF including TNF-α (z = −4.815, p < 0.001) and IL-8 (z = −2.639, p = 0.008). Interestingly, cytokines that are known to be anti-inflammatory, including IL-10 (z = −3.009, p = 0.003) and IL-13 (z = −3.538, p < 0.001), were also significantly higher in the CSF of patients compared to controls. However, there was no significant correlation between these cytokines and TSPO PET levels in the total grey matter (all p > 0.05).

Conclusions: We found concurrent elevation of pro- and anti-inflammatory cytokines in the CSF of patients with first-episode psychosis, who were presenting with ongoing symptoms despite antipsychotic treatment. This may indicate a complex picture of pro-inflammatory and compensatory anti-inflammatory signaling in early-course illness. However, these alterations were not linked to PET TSPO levels in the grey matter, indicating that cytokine dysregulation in the CNS is unlikely to be linked to microglia alterations in this cohort.

Keywords: first-episode psychosis, cytokines, cerebrospinal fluid, PET imaging, microglia.

Disclosure: Nothing to disclose.

P748. The role of current cannabis use on metabolism, cognition, and brain activity in individuals with psychosis

Tyler Lesh, Joshua Rhilinger, Lia Gugava, Stacey Tevelev, Megan Enos, Emily Childers, Andrea Danila, Cameron Carter

University of California - Davis, Sacramento, California, United States

Background: Cannabis is the most used federally illicit substance by patients with psychosis and is consistently associated with worse outcomes. Ongoing cannabis use is associated with worse symptoms, poorer medication adherence, and increased risk and length of relapse. Furthermore, many studies have shown that regular heavy cannabis use in adolescence is an independent risk factor for developing psychosis and is associated with an earlier age of onset of illness. More frequent use of high potency cannabis is associated with even higher risk for developing a psychotic illness. However, a history of cannabis use has also been linked to higher cognition and brain activity in people with psychosis. Despite its well-known effects on appetite, in healthy subjects and in animal models, cannabis use is also associated with decreased body mass, less insulin resistance and more favorable cardio-metabolic profiles. Recent studies in cannabis-using individuals with psychosis have shown smaller abdominal girth and lower blood pressure, blood glucose and lipids. Consequently, one explanation for higher cognition could be the positive impact of cannabis use on cardio-metabolic profiles or inflammatory pathways.

Methods: Seventy-three individuals with psychosis (schizophrenia, schizoaffective disorder, bipolar disorder with psychotic features) were recruited. While cannabis use disorder was allowed, participants were excluded for other substance use disorders within the past 6 months. Diagnoses were confirmed with the SCID-5 and the majority of participants took part in a MRI scanning session (3T Siemens Trio) where they completed the AX-CPT, a test of cognitive control. Cognitive control performance was measured using the d’-Context score and the primary fMRI contrast focused on high control (CueB) versus low control (CueA) cues. Whole-brain statistical maps were thresholded at p < .001 and cluster-corrected at p < .05. Urine THC metabolite levels and lipid panels were analyzed and individuals were separated into non-use versus current cannabis use. Cognitive and cardio-metabolic data were statistically evaluated with general linear models, including sex and age as covariates (SPSS v30). Functional MRI data were analyzed using SPM-12.

Results: Current cannabis users had non-significantly lower BMI (p = .08) and waist circumference (p = .09). Patients with current cannabis use did not show any significant differences in symptomatology, social or role functioning (p > .05). Total cholesterol was significantly lower in the cannabis using group (p = .022). Interestingly, within the cannabis using group higher levels of THC metabolite detected in urine was associated with lower total cholesterol (p = .04). Finally, current cannabis users showed significantly higher cognitive control performance on the d’-Context score (p < .05) as well as higher activity on CueB versus CueA trials in bilateral premotor/posterior insular cortex.

Conclusions: Results suggest that current cannabis users with psychosis tend to show a more favorable metabolic profile compared to those who do not currently use cannabis. Additionally, subtle differences reflecting higher cognition in users may provide some preliminary evidence for the role of positive cannabis-mediated metabolic effects on cognition. However, alternative explanatory factors should be considered, which could include genetic and environmental influences that may have complementary influences on both cannabis use propensity and cognition/metabolic health. Furthermore, these data must be reconciled with the known consequences of co-occurring substance use that are typically associated with poor clinical and functional outcome.

Keywords: cannabis, early psychosis, fMRI, metabolism.

Disclosure: Nothing to disclose.

P749. Risk factors for death in long-term follow-up of clinical high risk youth

Tara Trujillo, Jean Addington, Carrie Bearden, Tyrone Cannon, Barbara Cornblatt, Matcheri Keshavan, Rachel Loewy, Daniel Mathalon, Diana Perkins, William Stone, Elaine Walker, Scott Woods, Kristin Cadenhead

University of California - San Diego, San Diego, California, United States

Background: Clinical High Risk (CHR) for psychosis research has provided important insights into risk factors for psychotic conversions and other adverse outcomes that occur within 2–3 years. The question of outcome after 5+ years has not been extensively studied. Our group previously reported on the high rate of suicidality in clinical high risk youth in the third iteration of the North American Prodrome Longitudinal Study (NAPLS-3) at ~30% and the relationship to short-term outcomes including conversion. In this project we examine long-term outcomes including death and suicide as well as risk factors for the development of these outcomes in CHR youth.

Methods: In an ongoing project to perform long-term (5–20 year) diagnostic, symptom and functional assessments of up to 2000 individuals who previously met CHR criteria, we recontacted prior participants across 9 NAPLS sites. We assessed all participants for adverse life events and collected broad clinical data. In addition, we obtained information on outcomes including death from family interviews or public records. Additional analyses were done with baseline data from these participants to look at risk factors for death and suicide.

Results: As of June 2025, N = 566 participants have been interviewed and information about an additional N = 100 individuals has been obtained from family members or public records. There were N = 45 total deaths from any cause, which represents 6.8% of the sample. Within this group, there were N = 6 confirmed deaths by suicide; N = 2 overdose deaths; N = 4 accidental deaths; and N = 3 deaths of medical etiology. N = 3 of the participants who died by suicide had converted to a psychotic disorder. Most deaths (N = 30) were by unknown cause. Death by any cause was significantly associated with a variety of baseline measures, including history of trauma (psychological, physical, and sexual abuse) (p’s < 0.04–0.01); poorer functioning (p < 0.01); and total number of undesirable life events experienced (p < 0.01). Death was significantly associated with tobacco use at baseline (p < 0.01) and with age at first cannabis use (p < 0.01) but not with other substance use. There were no significant sex or age differences. There were no significant differences in SOPS attenuated psychotic symptoms total score or symptom domain sub-scores at baseline; however, death was significantly associated with any family history of psychosis (p < 0.01).

Conclusions: Consistent with findings in first episode psychosis, there is a much larger rate of death in CHR youth compared to the general population of adolescents and young adults. Most deaths in this group are by unknown cause; therefore, it is likely that the actual rate of death by suicide is higher. Specific risk factors including trauma, poor functioning, negative life events, and a family history of psychosis are significantly associated with early death, underscoring the importance of early intervention to prevent adverse outcomes in this vulnerable population.

Keywords: Clinical high risk state for psychosis, Suicide risk factors, suicide prevention.

Disclosure: Nothing to disclose.

P750. Impact of early psychosis program participation on service user symptoms and overall life satisfaction

Valerie Tryon, Kathleen Burch, Mark Savill, Sabrina Ereshefsky, Katherine Pierce, Merissa Kado-Walton, Khanh Linh H. Nguyen, Maliha Safdar, Alicia Assang, Jas Ganev, Nikki Motabar, Yi Zhang, Daniel Tancredi, Tara Niendam

University of California, Davis, Sacramento, California, United States

Background: California’s EPI-CAL project (https://epical.ucdavis.edu) joined several university- and community-based early psychosis (EP) treatment programs to create a statewide learning health care network and contributed to the NIMH EPINET database. The programs in the EPI-CAL network practice the Coordinated Specialty Care (CSC) model of EP care for service delivery. CSC is an evidenced-based practice for early psychosis (EP) that has been demonstrated to improve key outcomes for individuals experiencing early psychosis. The effectiveness of CSC outside of clinical trials continues to be evaluated, and our primary objective was to examine whether service users experienced improvement in key outcomes, such as symptoms and life outlook, over time in California’s EP programs.

Methods: Service users across EPI-CAL clinics completed self-report surveys about outcomes related to their life and mental health treatment. Surveys are administered at enrollment in the EP program and every 6 months thereafter via the Beehive application. Descriptive analyses summarize the demographic and outcomes data collected related to service users’ self-reported functioning, including survey responses regarding life outlook and symptoms. Symptom scores were measured on the Modified Colorado Symptom Index (MCSI) and life outlook was measured by single overall life outlook question from the Personal Wellbeing Index (PWI). Paired-samples t-tests were conducted to compare total MCSI symptom score at baseline and follow up (either 6 months or 12 months timepoint) and overall life outlook score at baseline and follow-up timepoints. The family-wise error rate was adjusted to correct for multiple comparisons.

Results: To date, 88 service users across 19 EPI-CAL clinics have completed surveys on life outlook and/or psychiatric symptoms at baseline and at least one follow-up timepoint. Services users received diagnoses associated with a first episode psychosis (FEP; n = 67, 76%), clinical high risk (CHR; n = 18, 20%), or FEP/CHR status not confirmed (n = 3, 3%). Ages ranged from 13 to 39 (M = 21.14 years, SD = 4.82), and 53% of the sample was female sex (n = 47). MCSI total scores were significantly lower at a follow up timepoint (M = 18.60, SD = 1.58) compared to baseline (M = 22.73, SD = 1.83); (t (59) = 4.03, p < .001). We did not observe significant differences in the life outlook score at a follow up timepoint (M = 5.68, SD = 2.47) compared to baseline (M = 5.63, SD = 2.70); (t (84) = 0.16, p = .87).

Conclusions: Findings show improvement in clinical symptoms from baseline to follow-up for individuals served in CSC programs across the state. Factors that influence change in life outlook need further exploration. These results show that EP services provided in various treatment programs across community and university settings can achieve positive outcomes for individuals in the early stages of psychosis and support broader implementation of CSC statewide.

Keywords: early psychosis, coordinated specialty care, Patient Reported Outcomes.

Disclosure: Nothing to disclose.

P751. Novel extracellular autoantibodies in schizophrenia identified by high-throughput screening

Katlyn Nemani, Aaron Ring, Jillian Jaycox, Leon Furchtgott, Adrienne Lahti, Donald Goff

Nathan S. Kline Institute for Psychiatric Research/NYU School of Medicine, New York, New York, United States

Background: While diverse inflammatory abnormalities in schizophrenia have long been described, anti-neuronal cell surface autoantibodies (AAbs) have garnered increasing interest since the discovery of anti-NMDAR encephalitis and other immunotherapy-responsive encephalitides associated with psychosis. These autoantibodies typically produce severe neurologic symptoms and are uncommon in schizophrenia. Few studies have focused on identifying novel AAbs in patients with isolated psychotic symptoms. We used a high-throughput platform called Rapid Extracellular Antigen Profiling (REAP)—an innovative technology that can detect AAbs to over 6,500 extracellular antigens—to (j) quantify AAb burden in schizophrenia spectrum disorders (SSD), (ii) identify SSD-enriched antigens, (iii) localize targets by tissue and cell type, and (iv) explore relationships with antipsychotic response.

Methods: Serum from a metacohort of 352 individuals with SSD (median age 37.5 years [25–54]; 33% female) and 971 controls without severe mental illness (median age 46 years [23–60]; 61% female) was screened using REAP. SSD subgroups included 247 inpatients with chronic psychosis, 43 outpatients with early psychosis and 62 medication-naïve first-episode psychosis (FEP) patients. Primary analyses compared per-subject AAb totals between groups/subgroups. For antigen-level associations, we binarized REAP scores (REAP > 1) and fit logistic models adjusted for age and sex; significance required OR > 5 and q < .05. Tissue/cell-type attribution used Human Protein Atlas gene sets. In the FEP subgroup, we examined the association between baseline AAb burden and treatment response following 16 weeks of risperidone based on change in the Positive and Negative Symptoms Scale (PANSS) with a linear model adjusted for baseline PANSS, age, and sex. We also trained a regularized logistic regression classifier (10-fold CV) to predict SSD vs control based on autoantibody profiles.

Results: SSD had a markedly higher AAb burden than controls (median 34.0 [23.0–47.0] vs 19.0 [11.0–28.0, p < .001). This finding remained significant after correcting for age and sex (p < .001). Relative to controls, AAb totals were highest in chronic psychosis (36.0 [26.0–49.0], p < .001), followed by antipsychotic-naïve early psychosis (29.0 [22.2–43.5], p < .001) and antipsychotic-treated early psychosis (26.0 [15.5–41.0], p = .01). Autoantibodies to 30 extracellular antigens were enriched in SSD after FDR correction (q < .05) (median OR = 7.11; range 5.13–20.09), including AAbs targeting neuronal ion channels, GPCRs, and myelin/ECM proteins; no AAbs were enriched in controls. Brain-enriched proteins explained the largest share of the SSD-associated burden (~20%) with excitatory and inhibitory neurons the top cell types (16% and 14%, respectively). In the FEP risperidone cohort, higher baseline Aab burden predicted a smaller reduction in PANSS scores (p = 0.016), suggesting an inferior response to antipsychotic treatment. Using AAb profiles, a 10-fold cross-validated model distinguished SSD from controls (AUC = 0.884; sensitivity = 0.680 at 95% specificity).

Conclusions: Our data reveal a robustly elevated, neuronally biased extracellular AAb burden in schizophrenia. AAbs to 30 extracellular antigens—including ion channels and GPCRs with plausible functional impact—were enriched in SSD; the presence of this signal in medication-naïve patients suggests argues against medication confounding. AAb burden was highest in treatment-refractory chronic psychosis and predicted smaller PANSS improvement on risperidone, supporting AAb load as a candidate stratification biomarker. These findings are consistent with extracellular autoimmunity in a subset of schizophrenia and support biomarker-guided trials of antibody-targeted interventions.

Keywords: Schizophrenia (SCZ), Antipsychotic-naïve first-episode schizophrenia, Autoantibody, autoimmunity, treatment resistant schizophrenia.

Disclosure: Nothing to disclose.

P752. Interaction between linguistic and theory of mind processing in schizophrenia spectrum disorders during socially relevant naturalistic video viewing

Emily Przysinda, Aaron Nidiffer, Jin Dou, Jillian O’Malley, Abigail Strugger, Rachel Seo, Abhishek Saxena, Bridget Shovestul, Emily Dudek, Stephanie Reda, Xiaoyu Dong, David Dodell-Feder, Edmund Lalor

University of Rochester Medical and Dental School, Rochester, New York, United States

Background: Language perception is impaired in schizophrenia spectrum disorders (SSD) at multiple levels including lower level acoustic features and higher level lexical features of speech. Patients with SSD also have known deficits in social processing, especially theory of mind (ToM), that result in difficulty inferring what others are thinking or feeling. Given that social processing depends on clear verbal and nonverbal communication, it is possible that language perception difficulties in SSD may be contributing to impaired social ToM processing.Traditional paradigms usually only examine one aspect of speech or social processing at a time, often using simple stimuli that are not representative of what we experience in the real world. Naturalistic paradigms, such as video viewing, provide an opportunity to better capture neural data in a more realistic context, and can also allow us to examine multiple aspects of neural processing simultaneously. Neural tracking of acoustic and lexical features during naturalistic video have been established by previous work using electroenceophalography (EEG), which showed significant decreases in neural tracking of patients with SSD (Przysinda et al., 2025). In the functional magnetic resonance imaging (fMRI) domain, previous studies use a similar paradigm of participants watching an episode of “The Office” TV show in order to examine how specific well known ToM regions were tracking the continuous awkwardness ratings collected from control participants (Pantelis et al. 2012 and Przysinda et al. 2023). These awkwardness ratings were used as a proxy for ToM understanding, as the feeling of awkwardness is often related to perceived violations of social expectations that are dependent on understanding of what others might be feeling or thinking. Here we examine how measures of language perception may interact with social ToM processing using multimodal neuroimaging and behavioral data collected from participants watching “The Office” TV show.

Methods: Our participants were confirmed to have an SSD (n = 23) or to be in the control group, Non-SSD (n = 22), using the structured clinical interview for the DSM-5. Both groups were balanced for age and sex. Using linear modeling, we obtained measures of how well the brain (via EEG) is tracking both the acoustic envelope and a measure of context-based language processing, known as lexical surprisal, during episodes of The Office TV show, as per Przysinda et al. 2025. Next we looked at how these speech processing measures correlate with question scores that ascertain participants’ ability to use ToM and mentalize about the characters during a separate episode not included in the EEG analysis. Finally, we examined the interaction between our language EEG and our previously published ToM fMRI processing measures with overlapping subjects (Non-SSD: n = 16, SSD: n = 19).

Results: We found that across subjects, there was a correlation between the mentalizing ToM questions and our measure of neural tracking of lexical surprisal but not acoustic envelope (r(43) = .33, p = .028). While we found no significant differences in the correlations across all subjects between our speech measures and ToM regions of interest tracking with awkwardness, there was a marginally significant correlation between surprisal tracking and how well the dorsal medial prefrontal cortex tracked with awkwardness (r(33) = .32, p = .061). Follow-up tests within groups did not yield any significant correlations.

Conclusions: These results show a potential link between measures of lexical processing of naturalistic video and both behavioral and fMRI measures of ToM processing in a multimodal imaging study. While analysis of the association between the EEG and fMRI findings for the groups individually were limited by low power, future analyses will attempt to parse out associations between these measures within the SSD and Non-SSD group respectively. We hope that these findings could be used in a clinical context to inform language and social interventions to help improve daily functioning of patients with SSD.

Keywords: Language, schizophrenia spectrum disorders, Social Processing, theory of mind.

Disclosure: Nothing to disclose.

P753. A phase 3 double-blind, placebo-controlled study of valbenazine added to standard of care antipsychotic treatment in patients with schizophrenia (ATS3019)

Eduardo Dunayevich, Jean-Pierre Lindenmayer, Eiry Roberts, Pradeep Nathan, Ashok Parameswaran, Hui Zhang, Christoph Correll

Neurocrine Biosciences, Inc., Cambridge, United Kingdom

Background: Valbenazine is a uniquely selective vesicular monoamine transporter 2 (VMAT2) inhibitor approved for the treatment of tardive dyskinesia (TD) and chorea associated with Huntington’s disease. It has been well-established that there is an increase in pre-synaptic striatal dopamine (DA) synthesis in patients with schizophrenia (Howes et al, Arch Gen Psychiatry 2012) which is correlated with psychotic symptoms. VMAT2 inhibition reduces monoamine transport into presynaptic vesicles and efflux in the striatum and medial prefrontal cortex, suggesting potential efficacy in the treatment of psychosis (Huang et al, Pharmacol Biochem Behav 2020). In a rodent model of antipsychotic effects (conditioned avoidance response), synergistic effects were observed when valbenazine was co-administered with risperidone and with olanzapine, suggesting a potential role for valbenazine in schizophrenia (Grigoriadis et al, ACNP 2017). To evaluate the potential of a VMAT2 inhibitor for improving symptoms of schizophrenia, we conducted a multicenter Phase 3 clinical trial (NCT05110157) of once-daily valbenazine in patients who had ongoing symptoms of schizophrenia while receiving stable antipsychotic therapy.

Methods: The study included a 4-week screening period; a 2-week, double-blind (DB), placebo run-in period; an 8-week, randomized, DB treatment period with valbenazine or placebo added to current antipsychotic treatment and a 2-week washout of study treatment. Valbenazine was initiated as adjunctive treatment in participants with schizophrenia who had an inadequate response to standard of care antipsychotic treatment, starting at 40 mg (1 week), increased then to 80 mg and maintained at the maximum tolerated dose for the remainder of the blinded treatment period. The primary endpoint was analyzed in the efficacy analysis set (EAS), defined as participants who had a Positive and Negative Syndrome Scale (PANSS) total score ≥70 at screening and baseline and ≤15% change from baseline to Week 2 (end of DB run-in), as well as stable background antipsychotic treatment. The primary endpoint was defined as the change from baseline at Week 10 (end of DB treatment) in PANSS total score, analyzed using ANCOVA with multiple imputation. Safety assessments including treatment-emergent adverse events (TEAEs) were analyzed using descriptive statistics.

Results: In the full analysis set, which included all 427 randomized participants, 210 (49.2%) were from US study sites, mean age (±SD) was 45.4 (±12.3) years, 278 (65.1%) were male, 319 (74.7%) were White, and the mean antipsychotic dose was 5.0 (±1.4) mg/d in risperidone equivalents. Baseline characteristics were similar in the EAS (N = 326) and well-balanced between the randomized treatment groups (161 valbenazine, 165 placebo). In the EAS, mean (±SD) PANSS scores at baseline were 80.7 (±9.5) and 80.6 (±9.5) for valbenazine and placebo, respectively, and 85.7% of valbenazine-treated participants were taking 80 mg/d from Weeks 4 to 10 (end of DB treatment). Most participants in both groups completed DB treatment (89.4% valbenazine, 90.9% placebo). At Week 10, least-squares (LS) mean changes from baseline [with 95% CIs] in PANSS total score demonstrated greater improvement with valbenazine (−8.0 [−9.6, −6.4]) versus placebo (−6.6 [−8.3, −4.9]), with no statistically significant LS mean difference (LSMD) between treatment groups (−1.4 [−3.6, 0.9], P = 0.23). The change in the PANSS Positive Symptoms subscale indicated a greater LS mean reduction with valbenazine relative to placebo (LSMD −0.9 [−1.7, −0.2]; nominal P = 0.017). In the safety analysis set (N = 425), the incidence of TEAEs was similar between treatment groups (34.4% [73/212] valbenazine, 31.0% [66/213] placebo), with somnolence being the only TEAE reported in ≥5% of valbenazine-treated participants (8.0% vs. 2.8% for placebo). Few participants had a serious TEAE (0% valbenazine, 1.9% placebo) or a TEAE leading to study discontinuation (1.4% valbenazine, 1.9% placebo).

Conclusions: In this placebo-controlled clinical trial of adjunctive valbenazine for patients with ongoing symptoms of schizophrenia, the primary efficacy endpoint was not met. An efficacy signal in favor of valbenazine was observed in the positive symptoms subscale of the PANSS. Most participants receiving valbenazine achieved the 80 mg dose and continued it throughout the double-blind treatment period. Safety and tolerability findings were consistent with the established safety profile of valbenazine.

Keywords: valbenazine, VMAT2, Schizophrenia (SCZ).

Disclosure: Neurocrine Biosciences, Inc., Employee, Self.

P754. Altered brain bioenergetics in schizophrenia: potential for treatment targets

Laura Rowland, Frank Gaston, Christopher Marano

Maryland Psychiatric Research Center, University of Maryland School of Medicine, Baltimore, Maryland, United States

Background: Current antipsychotic treatments for schizophrenia (SZ) are often insufficient, particularly for cognitive and negative symptoms, underscoring the need for novel treatment targets. We previously reported higher brain lactate levels that correlated with lower cognitive performance, and greater brain glucose levels that correlated with neuronal insulin resistance and poorer memory performance in SZ. We present evidence from an ongoing study showing impairments in peripheral and brain glucose metabolism in SZ.

Methods: This ongoing study (34 SZ, 35 HC) investigated 1H-MRS measures of lactate and glucose across three brain regions (anterior cingulate (AC), occipital cortex (OC), and thalamus(Th)), brain global venous oxygenation and oxygen extraction fraction with TRUST, cognitive function assessed with the MATRICS Consensus Cognitive Battery (MCCB), and HOMO-IR. Males and females were included.

Results: Results showed 1H-MRS glucose levels across all three regions were higher in SZ vs HC (p = 0.023). Higher glucose levels were related to higher HOMO-IR levels in all three regions (AC: r = 0.42, p < 0.001;OC: r = 0.362, p = 0.001; Th: r = 0.31, p = 0.005). Lactate, brain global venous oxygenation, and oxygen extraction fraction were higher in SZ vs HC but not statistically significant.

Conclusions: These findings provide further support for abnormal brain bioenergetics in SZ. Higher brain glucose levels coupled with higher HOMO-IR in SZ likely reflects lower brain glucose utilization related to insulin resistance. Higher levels of brain lactate, although not statistically significant, may reflect altered mitochondrial function. Brain metabolic pathways may offer new avenues for intervention.

Keywords: Schizophrenia (SZ), 1H-MRS, Energy metabolism.

Disclosure: Nothing to disclose.

P755. Multiple sources of causal data converge on a common circuit for psychosis

Andrew Pines, Garance Meyer, Frederick Schaper, Calvin Howard, Nicole Chiulli, Clemens Neudorfer, Konstantin Butenko, JiJun Wang, Yingying Tang, Alik Widge, Darin Dougherty, Joan Camprodon, David Silbersweig, Andreas Horn, Michael Fox, Miklos Argyelan, Shan Siddiqi

Brigham and Women's Hospital, Boston, Massachusetts, United States

Background: Rare cases of brain lesions and deep brain stimulation have been shown to cause schizophrenia-like psychosis, offering unique causal insights into the neuroanatomy of the disorder. Clinical trials of TMS and DBS have extended this approach, testing circuit-based interventions for treatment. Electroconvulsive therapy (ECT) is a well-established treatment for schizophrenia and is known to increase brain volume, yet the anatomical correlates of symptom change remain unclear. Here, we find that neuroanatomically localized evidence from lesions, TMS, DBS, and ECT to identify a common brain circuit that modulates psychotic symptoms.

Methods: ECT: 31 patients with schizophrenia underwent MRI 1 week before and 8 weeks after ECT. Symptoms were tracked with BPRS. Grey matter volume change before and after ECT was quanitifed with CAT12 and FreeSurfer. A linear mixed-effects models tested associations with symptom improvement, controlling for baseline psychotic score and change in total brain volume.

TMS: 55 patients at clinical high risk (CHR) for psychosis were randomized to active vs sham TMS to the temporoparietal junction. Conversion to psychosis was assessed at 1 year. Each individual’s stimulation site was mapped to a large connectome (n = 1000) and correlated with the lesion-derived psychosis network. Linear models tested the interaction between treatment group and circuit connectivity.

DBS: A systematic review and multicenter collaboration identified 36 cases in which DBS either caused or improved psychosis. Electrodes were reconstructed with Lead-DBS, stimulation volumes modeled (OSS-DBS), and overlap with anatomical pathways quantified using atlases and connectomes. Common connections were mapped to candidate psychosis circuits.

Lesions: 153 published cases of lesions causing secondary psychosis were mapped to a normative connectome and tested against 1156 control lesions to determine functional connections specific to psychotic symptoms. Results were validated in an independent lesion cohort.

Results: ECT: Across all FreeSurfer DKT parcellations, the top predictor of total psychotic symptom change was grey matter volume change in the lesion-derived psychosis circuit (LME p = 0.0014; linear regression p = 0.0059, r = –0.525).

TMS: Connectivity between the individual stimulation site and the lesion-derived psychosis circuit predicted conversion in CHR patients (p = 0.0017).

DBS: Stimulation sites that caused psychosis had significantly higher network damage scores to the lesion-derived psychosis circuit than sites not associated with psychosis. Across all 36 cases, electrodes converged on two candidate circuits involving the hippocampus–VTA–NAc loop and the amygdala–MD thalamus pathway.

Lesions: 153 cases of lesion-induced psychosis mapped to a common circuit centered on the posterior hippocampus.

Conclusions: Convergent evidence from lesions, TMS, DBS, and ECT demonstrates that psychotic symptoms map to a common brain circuit centered on the hippocampus and its subcortical connections. This circuit provides a causal, anatomically grounded target for future neuromodulation in schizophrenia.

Keywords: Schizophrenia (SCZ), BOLD fMRI signal, TMS, DBS, electroconvulsive therapy.

Disclosure: Nothing to disclose.

P756. Simultaneous attenuation of recurrent cortical excitatory and inhibitory connections can account for visual contrast detection deficits in psychotic disorders

Baktash Babadi, Daphne Holt, Roger Tootell

Massachusetts General Hospital, Boston, Massachusetts, United States

Background: Individuals with psychotic disorders (PD) have repeatedly been shown to exhibit impairments in several domains of visual perception, including visual contrast detection, relative to healthy controls (HC). While these perceptual deficits are well-replicated across studies, their underlying neural mechanisms remain insufficiently understood. Identifying circuit-level alterations that account for these impairments could provide valuable insights into the broader pathophysiology of psychosis.

Methods: We tested visual contrast detection in 52 individuals with PD and 58 demographically matched HC participants. Stimuli consisted of 2D filtered white noise across a range of spatial frequencies and contrast levels. Individual psychometric functions were estimated to quantify the performance of each subject in the task. To probe the neural basis of these deficits, we constructed a biologically inspired computational model of primary visual cortex (V1) comprising excitatory (E) and inhibitory (I) neuronal populations, recurrent and feed-forward connections, and a downstream readout neuron. The model received the same filtered white noise stimuli as the participants. For each participant, synaptic parameters of the model were iteratively adjusted until the output of the readout neuron reproduced the psychometric function observed behaviorally. Group-level comparisons of synaptic strengths were then conducted across the fitted networks.

Results: Relative to HC, the PD group demonstrated consistently higher contrast detection thresholds and flatter psychometric slopes (all p < 0.0008) across the full range of tested spatial frequencies (0.1–12 cycles/deg). These behavioral differences indicate both reduced sensitivity to low-contrast stimuli and diminished ability to discriminate fine gradations of contrast. The fitted V1 models successfully captured psychometric functions in both groups (R² > 0.7, p < 0.02). Critically, the models tuned to PD participants revealed significantly weakened recurrent connectivity within both excitatory (E-to-E: −21%, p = 0.027) and inhibitory (I-to-I: −18%, p = 0.012) populations compared with HC models. In contrast, cross-population connection strengths (E-to-I and I-to-E) did not differ significantly between the two groups.

Conclusions: Our findings suggest that concurrent attenuation of recurrent excitatory and inhibitory circuitry in V1 may underlie contrast detection deficits in psychotic disorders. This pattern of network alteration provides a mechanistic explanation that aligns well with the existing biological models of PD, which implicates the dysfunction of both excitatory and inhibitory components. By linking behavioral psychophysics to circuit-level modeling, this study highlights a potential pathway through which cortical microcircuit alterations contribute to perceptual abnormalities in psychosis. Further studies are required to investigate whether similar recurrent connectivity deficits extend to other sensory domains and cognitive functions in PD.

Keywords: Psychotic Disorders, visual contrast sensitivity, computational modeling.

Disclosure: Nothing to disclose.

P757. Efficacy of xanomeline and trospium chloride across symptom domains in adults with schizophrenia: results from the 52-week, open-label EMERGENT-5 clinical trial

Tracy Hicks, John Kane, James Appio, Monica Elias, Pierre Nicolas, Amy Claxton, Lara Shirikjian

Bristol-Myers Squibb, Fair Lawn, New Jersey, United States

Background: Xanomeline and trospium chloride (X/T) combines the dual M1/M4-preferring muscarinic receptor agonist xanomeline with the peripherally restricted muscarinic receptor antagonist trospium chloride and is approved for the treatment of schizophrenia in adults by the U.S. Food and Drug Administration. X/T demonstrated significant improvement in Positive and Negative Syndrome Scale (PANSS) total score compared with placebo in the pivotal phase 3, 5-week, double-blind, placebo-controlled, randomized EMERGENT-2 (NCT04659161) and EMERGENT-3 (NCT04738123) clinical trials. Additionally, continued long-term improvement in PANSS total score was observed with X/T treatment in the 52-week, open-label extension EMERGENT-4 trial (NCT04659174), which enrolled participants who completed the treatment period of EMERGENT-2 and EMERGENT-3. In this post hoc analysis, long-term efficacy of X/T is assessed across schizophrenia symptoms in the 52-week, open-label EMERGENT-5 (NCT04820309) trial using PANSS Marder 5-factor analysis.

Methods: EMERGENT-5 was a phase 3, multicenter, 52-week, open-label outpatient trial enrolling adults with a confirmed diagnosis of schizophrenia who had stable symptoms with antipsychotic use and no prior exposure to X/T. Participants were required to be aged 18–65 years with a primary DSM-5 diagnosis of schizophrenia, a PANSS total score ≤80, and a Clinical Global Impression-Severity (CGI-S) score ≤4. Participants had received oral antipsychotic therapy within 30 days of screening and discontinued these medications prior to their baseline visits. After initial assessments, participants were initiated with twice-daily (BID) oral doses of xanomeline 50 mg/trospium 20 mg and titrated to a maximum dose of xanomeline 125 mg/trospium 30 mg BID for the remainder of the 52-week treatment period. Change from baseline in PANSS total score was examined in the modified intention-to-treat population (mITT), which included all participants who received ≥1 dose of trial medication and had one baseline and ≥1 postbaseline PANSS assessment. Post hoc analysis evaluated change from baseline in PANSS Marder 5-factor scores. Adverse events were assessed in the safety population, defined as all participants who received ≥1 dose of trial medication.

Results: The safety and mITT populations consisted of 566 and 558 participants treated with X/T, respectively. The mean±standard error baseline PANSS score in the mITT population was 66.0 ± 0.44, reflecting a mildly ill population. Long-term treatment with X/T resulted in improvements in PANSS total score (change from study baseline at week 52 (CFB): −5.5 ± 0.66). Additionally, score reductions from baseline to week 52 were observed in each of the PANSS Marder factors. The largest improvement was in the PANSS Marder positive factor, with a −2.1 ± 0.23-point change from baseline resulting in a mean score of 18.2 ± 0.28 at 52 weeks. Similarly, score reductions from baseline to week 52 in PANSS Marder negative (15.6 ± 0.29; CFB: −1.1 ± 0.27), uncontrolled hostility (6.3 ± 0.15; CFB: −0.3 ± 0.17), disorganized thought (14.0 ± 0.22; CFB: −0.8 ± 0.20), and depression/anxiety (6.1 ± 0.16; CFB: −1.2 ± 0.18) factors indicate broad and sustained efficacy with X/T over the 52-week trial period. No new safety signals were observed; in addition, X/T was not associated with motor symptoms, somnolence, weight gain, or other adverse events often reported with use of D2 receptor-acting antipsychotic treatments.

Conclusions: In the long-term, open-label, outpatient EMERGENT-5 trial, X/T demonstrated consistent efficacy across symptoms among adults with schizophrenia who transitioned from other antipsychotic treatments. X/T was generally well-tolerated, and no new safety signals were observed.

Keywords: Schizophrenia (SCZ), efficacy, xanomeline.

Disclosure: Bristol Myers Squibb, Employee, Self.

P758. Microstructural and diffusion tensor imaging of clozapine for treatment-resistant schizophrenia

Samira Raminfard, Tate Overbey, Annie Blazer, Leigh Pearcy, Michael Pupi, Charles Kahn, K.N. Roy Chengappa, Brian Coffman, Amelia Versace, Deepak Sarpal

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Clozapine (CLZ) demonstrates superior efficacy for treatment-resistant schizophrenia (TRS), but mechanisms underlying its effects remain unknown. Pathophysiologic links between cortical regions and the basal ganglia (BG) characterize schizophrenia and are implicated in the mechanism of CLZ’s action. Here, we examined CLZ’s efficacy with diffusion weighted imaging (DWI) to examine microstructural and white matter-related measures within and between cortical and BG regions.

Methods: Twenty-six participants with TRS and moderate-to-severe psychosis underwent DWI scanning while starting CLZ, and nineteen were re-scanned after twelve weeks of treatment. Symptoms across treatment were measured via the Brief Psychiatric Rating Scale. DWI scans were processed to derive an array of microstructural measures (neurite density index, orientation dispersion index, mean kurtosis, and radial kurtosis) in frontal and BG gray matter parcels and white matter-related measures in corticostriatal tracts (fractional anisotropy, axial diffusivity, and radial diffusivity). Exploratory analyses linked results with published functional connectivity findings. Findings were considered significant if they met false discovery rate correction (p < 0.05).

Results: We found significant links between CLZ response and increased mean kurtosis in the left dorsolateral putamen, left ventral caudate, and left globus pallidus. Additionally, significant relationships between striatal segments of corticostriatal white matter tracts and CLZ response were observed, including increased mean kurtosis, decreased axial diffusivity, and decreased fractional anisotropy. Exploratory findings linked changes in BG microstructure with published corticostriatal connectivity findings.

Conclusions: Our results implicate microstructural changes within gray matter regions of BG and white matter changes within the striatum as mechanisms underlying CLZ treatment. These findings further elucidate CLZ’s mechanism of action and warrant further analyses that explore pharmacologic effects at the microstructural level.

Keywords: clozapine response, Diffusion Weighted Imaging, microstructural imaging, treatment-resistant schizophrenia.

Disclosure: Nothing to disclose.

P759. Source-based morphometry of cerebellar gray matter volume reveals altered structural networks in schizophrenia compared with healthy controls and bipolar disorder

Mahmoud Rashidi, Jayden Morales, Vince Calhoun, Bryon Mueller, Caroline Demro, Scott Sponheim, Jessica Turner

Ohio State University, College of Medicine, Columbus, Ohio, United States

Background: Research on schizophrenia has largely focused on cortical and subcortical structures. In contrast, the cerebellum remains comparatively understudied, despite its well-established role in cognition, emotion, and motor coordination. Source-based morphometry (SBM) involves independent component analysis (ICA) of gray matter volume data, allowing the identification of spatially distinct structural patterns that covary across individuals. This multivariate approach captures network-level anatomical organization beyond what is revealed by traditional univariate methods such as voxel-base morphometry (VBM).

Methods: This work was conducted as part of the Psychosis Human Connectome Project (P-HCP) using the data collected at the University of Minnesota. Our analysis included 168 participants: 85 with schizophrenia, 36 with bipolar disorder with psychotic features, and 47 healthy controls. We used the ENIGMA Cerebellum Volumetrics Pipeline for preprocessing the T1-weighted images. Automatic Cerebellum Anatomical Parcellation using U-Net with Locally Constrained Optimization (ACAPULCO) was used to isolate the cerebellum from the rest of the brain. The Spatially Unbiased Infratentorial Template (SUIT) toolbox was used to normalize the cerebellum into a standardized space and segment the cerebellum into gray and white matter. Images were smoothed using a 3-mm full-width at half-maximum (FWHM) Gaussian kernel. We used the SBM module of the GIFT Toolbox (http://trendscenter.org/software/gift) to perform ICA decompositions. Independent vector analysis (IVA) was chosen for its ability to jointly analyze multiple datasets or multiple subjects while preserving dependencies between corresponding components. General linear models controlling for age, gender, and frame-wise displacement were used to compare the diagnostic groups on anatomical metrics.

Results: Two independent components (ICs) of cerebellar gray matter showed significant diagnostic group effects (p < 0.05). The first IC showed significant difference between schizophrenia patients and healthy controls (p = 0.018). It comprised the network predominantly in bilateral X, right VIIb, and bilateral crus I. The second IC demonstrated significant differences between schizophrenia patients and healthy controls (p = 0.01) as well as bipolar patients (p = 0.039). The network predominantly comprised vermis VIIIa, X, and left VIIIb. A partial correlation analysis adjusted for age and gender revealed a significant association between the second IC and cognitive measures including the Wide Range Achievement Test (WRAT; r = −0.327, p < 0.001) and Wechsler Adult Intelligence Scale (WAIS; r = −0.222, p = 0.007).

Conclusions: This study demonstrates that cerebellar gray matter structural networks are altered in schizophrenia compared with healthy controls and, in some networks, bipolar disorder with psychotic features. The use of source-based morphometry revealed spatially distinct networks that would likely be overlooked by traditional univariate approaches. Alterations were observed in regions including bilateral lobule X, VIIb, crus I, and vermis VIIIa, X, and left VIIIb, highlighting the cerebellum’s involvement beyond motor functions. Notably, one of the identified networks showed significant associations with cognitive performance, suggesting a link between cerebellar structural organization and cognitive deficits in schizophrenia. These findings support the growing recognition of the cerebellum as a key node in the neural circuits underlying psychosis. Future studies integrating functional and structural data could further elucidate cerebellar contributions to cognitive and clinical features in schizophrenia. Overall, the results emphasize the importance of including cerebellar analyses in neuroimaging studies of psychotic disorders.

Keywords: cerebellum, Independent Component Analysis, psychosis.

Disclosure: Nothing to disclose.

P760. Prognostic value of individual-level heterogeneity in the psychosis spectrum: predicting transition and remission in individuals at ultra-high risk for psychosis

Tina Kristensen, Louise Glenthoej, Karen Marie Sandø Ambrosen, Cecilie Lemvigh, Kirsten Borup Bojesen, Mette Nielsen, Merete Nordentoft, Birte Glenthoj, Warda Syeda, Bjørn Ebdrup

Center for Neuropsychiatric Schizophrenia Research (CNSR), Glostrup, Denmark

Background: Predicting illness trajectories within the psychosis continuum remains challenging due to pro-nounced heterogeneity and diagnostic overlap. Traditional group-level comparisons of biopsycho-social variables often miss critical individual differences. This study introduces Affinity Scores (AS), a statistical framework designed to profile within-illness heterogeneity and quantify individual affini-ty to diagnostic groups across the psychosis continuum. AS integrate multimodal data, including cognitive, clinical, and biological markers. We assess whether heterogeneity features derived from AS can predict illness progression, specifically remission or transition to psychosis, in individuals at ultra-high risk for psychosis.

Methods: We included 627 participants aged 18–60: 240 healthy controls [HC], 201 individuals at ultra-high risk [UHR], 129 patients with first-episode psychosis [FEP], and 57 patients with schizophrenia [SCZ]. Clinical, cognitive, and demographic data were standardized using population-based z-scores. Glob-al functioning was assessed using the SOFAS. UHR symptoms severity was measured in UHR and SCZ using the CAARMS interview. Level of psychopathology were assessed in FEP and SCZ using the PANSS. Cognitive performance was assessed using WAIS-III (estimated Verbal IQ, Performance IQ, and Total IQ. BACS were used to assess verbal learning and memory (List learning); verbal working memory (Digit sequencing); verbal fluency (Fluency); processing speed (Symbol coding). CANTAB was used to assess spatial working memory (SWM); planning (SOC); set shifting / flexibility (IED); reaction time (RTI); and sustained attention (RVP-A’).

Individualized AS were calculated for each variable using participant-specific neighborhood graphs, assigning weights to neighbors within a defined proximity, capturing individual-level variability. Pairwise distances between participants were computed in z-score space, with AS reflecting the frequency of neighbors within the hop size for each participant. Multivariate AS were calculated using the mean affinity score across all variables as input into k-means clustering algorithm for classification.

Individual-level heterogeneity indices were derived from normalized AS, aiming to quantify the ex-tent to which an individual's profile aligned with one specific diagnostic group versus being spread across the four diagnostic subgroups of HC, UHR, FEP, and SCZ. Features included entropy, spread, standard deviation, Gini index, dominance gap, and distances to diagnostic groups (Euclidean and cosine). An additional UHR group rank feature was also computed capturing an individual’s affinity to UHR symptoms. After z-scoring, heterogeneity features were entered into elastic net logistic regression models predicting transition to psychosis and remission of UHR individuals. Model opti-mization was performed using stratified 3-fold cross-validation with repeated stability checks and bootstrap confidence intervals. Out-of-fold probabilities were calibrated using Platt scaling to en-hance reliability. Model performance was evaluated with ROC AUC, PR-AUC, Brier scores, and cali-bration plots.

Results: Affinity Scores demonstrated robust predictive performance, achieving an overall classification ac-curacy of 0.84 to diagnostic group. The individualized profiles revealed unique patterns of hetero-geneity across the psychosis continuum, while multivariate AS provided deeper insights into cross-domain interactions, highlighting subgroup-specific variations relevant to illness progression. Two heterogeneity features emerged as most predictive of both transition and remission: UHR-rank, a rank-weighted score reflecting the individuals positioning relative to the UHR-group (rank 1 mean-ing highest affinity to UHR and rank 4 lowest affinity to UHR). The second heterogeneity feature selected was dominance gap, reflecting the alignment to a specific diagnostic subgroup (calculated as the distance between the two highest ranked AS, where a large dominance gap reflects more distance between the AS / low heterogeneity, while a small dominance gap reflects lesser distance / a more mixed /heterogeneous profile). For UHR with transition (18/627 cases), the calibrated model achieved ROC AUC = 0.865 (95% CI: 0.828–0.919), PR-AUC = 0.106 (95% CI: 0.045–0.174), and Brier = 0.027. For remission (32/627 cases), ROC AUC = 0.813 (95% CI: 0.786–0.865), PR-AUC = 0.144 (95% CI: 0.127–0.206), and Brier = 0.045 (See ROC-curves in Figure 1.) In both outcomes, the hetero-geneity index captured distributional imbalance and rank-based positioning of individuals within the affinity space, suggesting that reduced diversity (low affinity to UHR, high dominance gap) and UHR positioning carry prognostic value.

Conclusions: AS offer a transparent and interpretable framework for addressing heterogeneity in psychosis by enabling individual inferences from multivariable data. By integrating scores into classification and prediction models based on heterogeneity features which carry meaningful predictive information, AS may offer a viable tool for advancing precision psychiatry, paving the way for more personalized diagnostics.

Keywords: Psychosis spectrum symptoms, individual variability, multivariate analysis, clinical outcome prediction.

Disclosure: Nothing to disclose.

P761. Value representation in youth psychopathology: evidence of a transdiagnostic risk mechanism for psychosis

Zachary Millman, James Gold, Jason Schiffman, Lauren Ellman, Elaine Walker, Albert Powers, Scott Woods, Steven Silverstein, Vijay Mittal, Philip Corlett, Gregory P. Strauss, James Waltz

McLean Hospital - Harvard Medical School, Belmont, Massachusetts, United States

Background: Negative symptoms of schizophrenia (SZ) are associated with deficits in representing the value of actions, observed in reinforcement learning (RL) tasks as impaired learning from gains but intact learning from losses. This RL profile contrasts with depression, where enhanced loss sensitivity is common. Whether a schizophrenia-like RL pattern characterizes negative symptoms in youth at clinical high-risk (CHR) for psychosis—who show modest psychosis transition rates but high rates of co-occurring depression—remains unclear. Distinguishing shared from risk-specific impairment profiles in this population is essential for understanding the developmental mechanisms of psychopathology, improving prognostic prediction, and informing treatment selection. Thus, we tested whether CHR youth show a schizophrenia-like RL profile and whether RL parameters relate differentially to negative vs. depressive symptoms across clinical groups by quantifying RL in CHR, clinical controls (CC) with other psychopathologies, and healthy controls. To strengthen inferences of specificity, we also tested whether RL performance relates to neuropsychological and prognostic indicators of psychosis vulnerability.

Methods: Data were drawn from the five-site Computerized Assessment of Psychosis Risk study. The sample included youth at CHR (n = 292), CC (n = 241), and healthy controls (n = 175). Risk status was measured with the Structured Interview for Psychosis-risk Syndromes; negative symptoms with the Negative Symptom Inventory – Psychosis Risk; depressive symptoms with the Center for Epidemiologic Studies Depression scale; and premorbid intelligence with the Wide-Range Achievement Test – Reading. Additional measurements were taken to obtain scores for the validated North American Prodrome Longitudinal Study’s psychosis risk calculator, which uses clinical, neurocognitive, demographic, and historical data (e.g., trauma exposure) to produce individualized estimates of psychosis transition probability.

Participants completed a probabilistic RL task shown to be sensitive to negative symptoms in SZ. The task requires learning from gains vs. potential losses over blocks of trials. On each trial, participants were presented with one of four pairs of stimuli. Two pairs led to a potential gain or neutral outcome and two led to a loss or avoidance of loss. For each valence condition, choice of the optimal stimulus was reinforced with 80% probability in one stimulus pair and 90% in the other. We calculated win-stay (percent of trials in which the participant stayed with the same choice after an optimal outcome) and lose-shift (percent of trials in which the participant shifted after a suboptimal outcome) rates separately for gain and loss-avoidance pairs, resulting in four primary outcome variables. Lower win-stay rates on gain trials suggests reduced value representation ability, particularly when paired with evidence of intact performance from the other RL parameters.

The three participant groups were compared on overall RL performance (the mean percentage of choices of the optimal stimulus across all trials of a given type). Correlational analyses determined the interrelations between RL parameters, negative and depressive symptoms, neurocognitive performance, and risk calculator scores. We then used multivariate regression models to estimate the unique effects of each RL parameter on symptom severity while controlling for other clinical factors. Group × RL interaction terms estimated CHR-CC differences in the magnitude of the relation between RL and symptom severity.

Results: Although overall RL performance was similar across groups, in both CHR and CC youth, higher negative symptoms were selectively associated with reduced win-stay behavior on gain trials (b = −0.01[0.04], t = −2.32, p = .021), suggesting impaired use of reward outcomes to guide subsequent actions. The group × RL term was not significant, suggesting the strength of the effect was similar across CHR and CC (b = 0.09[0.06], t = 1.58 = 1.58, p = .115). No relations were observed between depressive symptoms and any RL parameter, with very small correlation coefficients (−.002–.058). Lower win-stay rates were also linked to lower premorbid intelligence (CHR: rho = .287, p < .001; CC, rho = .202, p = .003) and higher psychosis risk calculator scores (CHR: rho = −0.137, p = .034; CC: rho = −.250, p < .001).

Conclusions: In this study, a schizophrenia-like RL profile was unrelated to depression but associated with multiple indicators of psychosis risk and negative symptoms across clinical groups, suggesting a transdiagnostic psychosis risk mechanism that is distinct from affective disturbance.

Keywords: Clinical high-risk for psychosis, Negative symptoms, Depression, Diagnostic specificity, Reinforcement learning.

Disclosure: Nothing to disclose.

P762. Cross-sectional pathways from dysfunctional beliefs to negative symptoms and functioning in schizophrenia spectrum disorders

Sarah Saperia, Alex Prosserman, Emilia Flores Anaya, Mahavir Agarwal, Margaret Hahn, Michael Best, Sean Kidd, Konstantine Zakzanis, George Foussias

Centre for Addiction and Mental Health, Toronto, Canada

Background: The negative symptoms of schizophrenia spectrum disorders (SSDs), marked by impairments in motivation and emotional expression, are key predictors of poor functional outcomes for affected individuals. The cognitive model of negative symptoms delineates a set of dysfunctional beliefs (i.e., defeatist performance beliefs, asocial beliefs, low expectancies for success, low expectancies for pleasure, internalized stigma, and perception of limited resources) as key factors linked to the development and maintenance of negative symptoms in SSDs, which may serve as potential mechanistic targets to treat negative symptoms. This study sought to comprehensively evaluate the cross-sectional pathways between these dysfunctional beliefs, negative symptoms, and community functioning in individuals with SSDs.

Methods: The study sample consisted of 132 participants with SSDs who were administered assessments of each dysfunctional belief system, along with evaluation of negative symptom severity and community functioning. Statistical analyses consisted of path analyses to evaluate cross-sectional pathways from dysfunctional beliefs to negative symptoms to impaired community functioning. Additionally, principal components analysis was utilized to evaluate underlying components of dysfunctional beliefs in an omnibus path model from dysfunctional beliefs to community functioning through negative symptoms.

Results: Path analyses revealed significant direct effects for asocial beliefs, expectancies for success and pleasure, and perception of limited resources, and to some degree defeatist performance beliefs, with negative symptoms, with better model fit particularly with the diminished motivation subdomain of negative symptoms (|β| = 0.22 to 0.31, p’s < 0.03). In addition, these dysfunctional beliefs also demonstrated significant indirect effects on community functioning specifically through their effects on diminished motivation (|β| = 0.14 to 0.21, p’s < 0.02). Principal components analysis revealed that all dysfunctional beliefs loaded on a single factor. This dysfunctional belief factor demonstrated a significant direct effect on negative symptoms, and in particular diminished motivation (β = 0.32, Z = 3.7, p < 0.001), and a significant indirect path to community functioning (β = −0.20, Z = −3.4, p = 0.001).

Conclusions: This investigation of the role of dysfunctional beliefs posited within the cognitive model of negative symptoms in SSDs revealed that some dysfunctional belief systems, specifically asocial beliefs, expectancies of success and pleasure, the perception of limited resources, and to a lesser degree defeatist performance beliefs, are linked to negative symptoms and through this to impairments in community functioning for individuals with SSDs. The findings from our principal components analysis of a single dysfunctional belief factor also raises the question of whether there is a central underlying dysfunctional belief domain driving these effects, with future work needed to disentangle the relationships among these dysfunctional belief systems. Overall these findings provide further support for the cognitive model of negative symptoms, and suggest potential modifiable targets for therapeutic innovation for these debilitating symptoms and their functional sequelae for individuals with SSDs.

Keywords: Negative Symptoms, functional outcomes, Dysfunctional beliefs.

Disclosure: Nothing to disclose.

P763. In vivo evidence of reduced muscarinic m1 receptor availability in schizophrenia

Tommaso Volpi, Mika Naganawa, Nabeel Nabulsi, Henry Huang, Richard Carson, Rajiv Radhakrishnan, Deepak D'Souza

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

Background: Converging lines of evidence from postmortem studies provide strong evidence that brain muscarinic M1 receptor deficits are present in a subset of schizophrenia (SZ) patients (Paul et al. 2024, https://doi.org/10.1016/j.biopsych.2024.03.014). M1 receptors are thus considered an important target for the treatment of SZ; recently the FDA approved COBENFY, which contains the muscarinic M1/M4 agonist xanomeline, as the first non-dopaminergic medication for the treatment of schizophrenia (Kaul et al. 2024, https://doi.org/10.1016/s0140-6736(23)02190-6).

M1 receptors can be imaged in vivo with the positron emission tomography (PET) radiotracer 11C-LSN3172176, which has excellent kinetic properties (good M1 selectivity, high uptake and high specific binding), as previously shown in healthy controls (HC) (Naganawa et al. 2021, https://doi.org/10.2967/jnumed.120.246967).

In this work, we examined in vivo brain muscarinic M1 receptor availability in SZ for the first time.

Methods: 36 adult participants (SZ patients, n = 16, 39 ± 15 years, 2F; age-matched HCs, n = 20, 44 ± 17 years, 9 F) underwent a 2-h PET scan with 11C-LSN3172176 on a HRRT scanner. Arterial blood sampling was performed in a subset of participants (HC, n = 17; SZ, n = 11). Region-of-interest (ROI) time-activity curves (TAC) were generated with the AAL atlas for the main cortical (frontal (FRO), temporal (TMP), parietal (PAR), occipital (OCC)) and subcortical (caudate (CAU), putamen (PUT), pallidum (PAL), ventral striatum (VST), thalamus (THA), cerebellum (CER)) ROIs. The centrum semiovale (CS) white matter was used as pseudo-reference region (2mL ROI, generated as in Rossano et al. 2020, https://doi.org/10.1177/0271678x19879230), after confirming CS binding was not significantly different between HC and SZ (t-test, p = 0.78). Kinetic analysis was performed using the simplified reference tissue model with parameter coupling (SRTMC), and the one-tissue model (1T) for the subset with arterial sampling. The main outcome kinetic parameter was the distribution volume ratio relative to CS (DVR(CS-2mL)). We also evaluated the 1T total distribution volume (VT).

Group differences (HC vs. SZ) were evaluated with linear mixed-effects models (LMM) globally and for each individual ROI (F-test, p < 0.05, %difference (%diff.) from HC mean value). Since age was correlated with M1 binding (Pearson’s r, p < 0.05), it was included as a covariate in LMMs. Individual patient differences (%diff(i)) in cortical M1 binding (average DVR(CS-2mL) and VT of FRO, TMP, PAR, OCC ROIs) were calculated against the mean of the corresponding HC values.

Results: SZ patients had significantly lower 11C-LSN3172176 binding than HCs (DVR(CS-2mL): main effect of diagnosis [F(1,428) = 15.3, p < 0.001], main effect of ROI [F(1,428) = 27.5, p < 0.001]; VT: main effect of diagnosis [F(1,360) = 5.5, p = 0.02], main effect of ROI [F(1,360) = 41.4, p < 0.001]). When evaluating individual ROIs, significantly lower DVR(CS-2mL) was found in all ROIs except PAL, with %differences of up to −20% (AMY, p < 0.001, %diff. = −20%; CAU, p = 0.002, %diff. = −19%; HIP, p < 0.001, %diff. = −18%; FRO, p < 0.001, %diff. = −18%; TMP, p < 0.001, %diff. = −18%). Significantly lower VT values were also detected in multiple ROIs (AMY, p = 0.02, %diff. = −17%; OCC, p = 0.01, %diff. = −16%; HIP, p = 0.02, %diff. = −15%; TMP, p = 0.02, %diff. = −15%).

6 out of 16 SZ patients had DVR(CS-2mL) %diff(i) ≤ −20%, as well as 3 out of the 11 patients with available VT (37% and 27% of the SZ sample, respectively).

Conclusions: This is the first in vivo PET imaging study of muscarinic M1 receptor availability in schizophrenia using 11C-LSN3172176. A somewhat global reduction in M1 receptor availability, as assessed by 11C-LSN3172176 DVR(CS-2mL) and VT, was detected in SZ. The results were consistent irrespective of the kinetic analysis method. Additionally, approx. 30% of patients exhibited > 20% reduction in M1 cortical availability (consistent with a “muscarinic receptor deficit intermediate phenotype”, Dean and Scarr 2024, https://doi.org/10.1016/j.psychres.2020.112989). Our findings of lower muscarinic M1 receptor availability across several cortical and subcortical regions overlap with postmortem data and support M1 as a target for SZ treatment. The relationship between M1 receptor availability and SZ symptoms will now be a focus of investigation.

Keywords: Schizophrenia (SZ), Positron Emission Tomography (PET) Imaging, M1 Muscarinic Receptors, Schizophrenia- Novel Treatment.

Disclosure: Nothing to disclose.

P764. Changes in self-reported distress across experimental sessions, in people with psychosis, are predicted by changes in activations of salience and reward network nodes

James Waltz, Maahee Patel, Julia Sheffield, Jacob Nudelman, Olivia Hutchinson, Eric Neutzling, James Gold

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

Background: Rather than being designed to attenuate psychosis symptoms directly, psychosocial interventions in schizophrenia spectrum disorders (SSDs) are often designed to reduce the distress associated with symptoms. Despite our growing understanding that people with SSDs are distinguished from other people with psychotic-like experiences by the distress associated with their symptoms, and that treatments like Cognitive Behavioral Therapy for Psychosis (CBTp) are often successful in reducing distress, the neural correlates of these reductions in distress have not been established. Using images from the International Affective Pictures Set (IAPS; Lang et al., 2005), our goal was to examine how both immediate and delayed responses to pleasant, aversive, and neutral stimuli related to self-reports of symptoms related to distress. In order to investigate how changes in responses to pleasant, aversive, and neutral stimuli related to changes in self-reports of symptoms related to distress, we had participants perform the Emotional Experience Task (EET; Ursu et al., 2011) twice, on separate study days (once 10 minutes after presentation of an acute stressor, once 10 minutes after a control task). We hypothesized that changes in self-reported distress across experimental sessions, in people with psychosis, would be predicted by changes in activations of the anterior insula (AI) and other salience network nodes.

Methods: In our study, 52 patients with schizophrenia spectrum disorders (SZ) and 32 healthy volunteers (HVs) performed the EET twice, at least a week apart. The Cold Pressor Task (CPT; Schwabe et al., 2008) was used as the acute stressor. In the stress condition, participants submerged their non-dominant arm in a bucket of ice water for up to three continuous minutes. The control condition used warm water. During performance of the EET, subjects viewed 72 IAPS picture stimuli (24 pleasant, 24 unpleasant, 24 neutral) in four runs of 18 trials each. On each trial, subjects were asked to rate both the “positivity” and “negativity” of the emotion experienced during the presentation of a picture, on a 5-point scale, by means of a four-button response unit. In both sessions, between the second and third runs of the task, participants were prompted to report their subjective feelings on a 7-point scale along dimensions such as “stress”, “pressure”, “sadness”, and “paranoia”. Changes in subjectively-rated feelings between sessions ranged from −6 to +6.

Results: We observed that changes in self-reported distress across experimental sessions, in people with psychosis, correlated significantly with changes in evoked activations to unpleasant pictures in left AI (r = 0.314; p = 0.034) and right AI (r = 0.353; p = 0.016). Changes in self-reported distress across experimental sessions, in people with psychosis, correlated significantly with changes in delayed activations to both pleasant (r = 0.300; p = 0.043) and unpleasant pictures (r = −0.320; p = 0.030) in dorsomedial prefrontal cortex (dmPFC). Changes in self-reported feelings of being “under pressure”, across experimental sessions, correlated significantly with changes in delayed activations to pleasant pictures (r = 0.355; p = 0.016) in left AI. Finally, changes in self-reported feelings of “suspiciousness”, across experimental sessions, correlated significantly with changes in evoked activations to unpleasant pictures in ventromedial prefrontal cortex (vmPFC; r = 0.342; p = 0.020).

Conclusions: We observed that changes in self-reported distress across experimental sessions, in people with psychosis, correlated significantly with changes in activations of the anterior insula, bilaterally, dmPFC, and vmPFC. These relationships between changes in brain activations and changes in reports of subjective experience may represent a neural correlate of distress that is impacted by psychosocial interventions such as CBTp. The ability to improve the lives of people with psychotic illness would benefit from an improved understanding of how both pharmacological and psychosocial interventions attenuate symptoms and the distress associated with symptoms.

Keywords: Acute Stress, Reward, Functional MRI (fMRI), IAPS.

Disclosure: Nothing to disclose.

P765. Neuroinflammation is associated with resting-state intrinsic hypoconnectivity in early psychosis

Annalisa Lella, Tyler Lesh, Jason Smucny, Kathleen Carlos, Laiya Thorpe, Kamalakannan SO M Vijayakumar, Jaylen Lee, Joni Ricks-Oddie, Vince Calhoun, Cameron Carter, Theo van Erp

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

Background: Immune dysregulation-related neuroinflammation is thought to play a role in the pathophysiology of early psychosis, but its functional consequences remain to be fully determined. Extracellular free water (FW), which can be estimated using free water mapping, has been proposed as a diffusion magnetic resonance imaging-based measure of neuroinflammation. Whole brain, white matter (WM), and gray matter (GM) FW have been found to be higher in individuals with psychosis compared to healthy controls. Moreover, recently, high gray matter FW was shown to be associated with better treatment outcome in early psychosis. Psychosis is also considered to be a disorder of brain dysconnectivity, and many resting-state functional imaging studies have reported aberrant brain connectivity in individuals with psychosis compared to controls, including studies on intrinsic connectivity as measured by fractional amplitude of low frequency fluctuations (fALFF). To advance our understanding of the functional consequences of neuroinflammation in psychosis, this study compared GM and WM FW between individuals with affective or non-affective early psychosis, and examined their relationships with positive and negative symptoms, and fALFF in the Human Connectome Early Psychosis sample. We hypothesized higher GM and WM FW in individuals with psychosis and a negative association between FW and fALFF.

Methods: This study analyzed MRI and clinical data from 118 individuals with psychosis [28 affective (primary mood disorder with psychotic features), 90 non-affective (primary psychotic disorder based on SCID-5-RV; 45 women, mean age (SD) = 22.7 (3.6)] and 53 healthy controls [18 women, mean age (SD) = 22.7 (3.6)], collected by the HCP-EP10 (release 1.1, August 19, 2021). Diffusion MRI data was preprocessed using “preprocessing and analysis of q-space images” (QSIprep), which also generates GM and WM masks. FW was estimated using a Diffusion Imaging In Python (DIPY) pipeline and fALFF was estimated using ENIGMA (Enhancing Neuro Imaging Genetics through Meta-Analysis) HALFpipe; both FW and fALFF measures were averaged within the GM and WM masks for each subject. Positive and negative symptom severity were based on the Positive and Negative Syndrome Scale (PANSS14 sum of items 1 to 7, and 8 to 14, respectively). ANCOVAs predicted FW (GM and WM) with diagnosis (early psychosis, control), with age, sex, and site (scanner) included as covariates. Kendall partial correlations examined relationships between FW (GM and WM) and PANSS positive symptoms, negative symptoms, and fALFF (GM and WM). Threshold-free cluster enhancement (TCFE) analyses examined relationships between FW (GM and WM) and voxelwise fALFF.

Results: Both GM FW (t = 2.17, p = 0.032) and WM FW (t = 2.46, p = 0.015) were higher in individuals with Early Psychosis compared to Controls. WM FW [r = 0.14, p = 0.031] showed a significant positive correlation with positive but not negative psychotic symptoms. Both GM FW (t = −7.27, p < 0.001) and WM FW (t = −6.86, p < 0.0001) showed significant negative associations with fALFF (p < 0.0001). TCFE analysis results examining the relationships between FW and voxelwise fALFF showed key regions involved in psychosis including the middle frontal gyrus / dorsolateral prefrontal cortex, superior frontal gyrus, anterior cingulate cortex, and superior parietal lobe.

Conclusions: This study replicates prior findings of higher GM and WM FW in early psychosis based on high-quality HCP-EP diffusion MRI data and found a positive relationship between WM FW and positive symptoms that corroborates an earlier finding in chronic psychosis. The significant associations between GM and WM FW and brain connectivity -here measured with fALFF- suggests possible new avenues for the treatment of the functional consequences of neuroinflammation in psychosis using neurostimulation.

Keywords: neuroinflammation, early psychosis, connectivity.

Disclosure: Nothing to disclose.

P766. Impact of CYP2D6 metabolizer status on efficacy and tolerability of xanomeline and trospium chloride in adults with schizophrenia: a post hoc analysis

Yi Luo, Aditi Bal, John Schwarz, Scott Vuocolo, Ken Kramer, Sarah Harris

Bristol-Myers Squibb, Morristown, New Jersey, United States

Background: CYP2D6 is a highly polymorphic member of the cytochrome P450 family of enzymes involved in drug metabolism. The combination of CYP2D6 alleles predict several metabolizer phenotypes: poor metabolizer without enzymatic activity, intermediate metabolizer with decreased enzymatic activity, normal metabolizer, and ultrarapid metabolizer with increased enzymatic activity. In EMERGENT-2 (NCT04659161) and EMERGENT-3 (NCT04738123), xanomeline and trospium chloride (X/T) demonstrated superior efficacy compared with placebo in adults with schizophrenia. Population pharmacokinetic analysis showed that xanomeline exposure was higher in intermediate metabolizers compared with normal metabolizers. The current post hoc analysis of data pooled from EMERGENT-2 and EMERGENT-3 evaluated the impact of CYP2D6 metabolizer status on the efficacy and tolerability of X/T.

Methods: DNA was isolated from whole blood samples collected from participants who consented for genetic testing in the EMERGENT-2 and EMERGENT-3 clinical trials. Whole genome sequencing was performed using Illumina NovaSeq S4 flow cell platform. CYP2D6 variants were evaluated using consensus sequences from 3 separate analysis tools: StellarPGx, Aldy, and Cyrius. Metabolizer status for CYP2D6 was then determined following guidelines from the Clinical Pharmacogenetics Implementation Consortium and Dutch Pharmacogenetics Working Group. Changes from baseline to weeks 2, 3, and 5 of the Positive and Negative Syndrome Scale (PANSS) total and PANSS subscale (positive, negative, and general psychopathology) scores, and frequency of adverse events were compared among participants with different CYP2D6 metabolizer status using the Mann-Whitney U test.

Results: CYP2D6 metabolizer status was evaluated for 324 participants (X/T: n = 169; placebo: n = 155). Most participants were normal (n = 168) or intermediate metabolizers (n = 116); there were few poor (n = 7) or ultrarapid metabolizers (n = 10). The changes from baseline for PANSS total and all subscale (positive, negative, or general psychopathology) scores at all visits were similar across the metabolizer groups. Adverse event frequencies were not significantly different across the metabolizer groups.

Conclusions: CYP2D6 metabolizer status does not significantly affect the efficacy and tolerability of X/T in adults with schizophrenia, indicating a consistent treatment effect across genetic variations.

Keywords: xanomeline, CYP2D6, whole genome sequencing.

Disclosure: Bristol-Myers Squibb, Employee, Self.

P767. Results from a phase 4, open-label, multicenter, two-cohort, two-period, slow-titration and food effect trial to assess the safety and efficacy of xanomeline and trospium chloride in schizophrenia

Jenna Hoogerheyde, David Walling, Kimball Johnson, Rachel Dyme, Allison Gaudy, Rasika Suryawanshi, Andrew Miller, Ranjan Tiwari, Ken Kramer

Bristol-Myers Squibb, Madison, New Jersey, United States

Background: In pharmacokinetic studies of xanomeline and trospium chloride (X/T), dosing with food reduced trospium bioavailability, potentially impacting tolerability. This study was designed to demonstrate taking X/T with food after a slow up-titration is safe and tolerable.

Methods: An inpatient, open-label, 2-cohort trial (NCT06572449) enrolled adults with schizophrenia and PANSS total score ≤80. In period 1, participants began X/T BID treatment on an empty stomach at 50mg/20mg and uptitrated to a maximum 125mg/30mg (cohort 1) or 100mg/20mg (cohort 2). In period 2, treatment was administered within 30 minutes of food intake. Safety, efficacy, and X/T pharmacokinetics pooled from both cohorts in both periods were assessed.

Results: 64.2% and 39.0% of participants reported ≥1 TEAE during periods 1 and 2, respectively. The most common TEAEs (≥5%) in period 1 were nausea (22.5%), dyspepsia (15.6%), headache (15.0%), constipation (12.7%), and vomiting (11.6%); the incidence of all new onset TEAEs was less in period 2 when taken with food. All TEAEs were mild or moderate in intensity; none were serious. PANSS total score decreased over periods 1 and 2; CGI-S score decreased over period 1 with no further change over period 2. Dose-normalized AUC and Cmax of trospium pooled from cohorts 1 and 2 decreased by 36% and 43%, respectively, when taken with food vs fasting. No clinically significant differences were observed in xanomeline exposure.

Conclusions: After slow uptitration on an empty stomach, there was no increase in the incidence or severity of TEAEs when taking X/T with food; efficacy was maintained.

Keywords: xanomeline, food effect, Pharmacokinetics, panss.

Disclosure: Bristol-Myers Squibb, Employee, Self.

P768. Interventions to mitigate psychotropic associated weight gain using an algorithmic measurement-based integrated clinical pathway: a retrospective chart review study

Nicolette Stogios, Femin Prasad, Riddhita De, Kateryna Maksyutynska, Akash PrasannaKumar, Vittal Korann, Marcos Sanches, Leah Burton, Mohammed Alarabi, Tanner Isinger, Valerie Powell, Michelle Desanti, Rebecca Oliviera, Daphne Korczak, Ariel Graff-Guerrero, Gary Remington, Margaret Hahn, Mahavir Agarwal

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

Background: Individuals with severe mental illness (SMI) have a 20% reduced life expectancy compared to the general population, largely owing to an increased risk for cardiometabolic comorbidities. Psychotropic drugs undoubtedly contribute to this increased cardiometabolic risk with their propensity to cause metabolic adverse effects. Several non-pharmacological and pharmacological interventions are available to mitigate these metabolic adverse effects. However, data on the utility of these interventions in real-world settings are limited. This study presents real-world data from a Mental Health and Metabolism Clinic that specializes in providing psychiatrist-led care for metabolic dysfunction in those with mental illness.

Methods: We conducted a retrospective chart review of all patients attending the Mental Health and Metabolism Clinic at the Centre for Addiction and Mental Health (CAMH) in Toronto, Canada, between 2016 and 2023 (N = 914). This study received approval by the CAMH Research Ethics Board (REB #: 072/2021). Clinical records were reviewed for demographic, anthropometric, and metabolic data. These variables were extracted for each patient, if available, at baseline (date of clinic consultation), 3-, 6-, 9-, and 12-months post-baseline (± 1.5 months for each time point). Our primary outcome was the change in body weight over 12 months of treatment, as well as the percentage of patients that achieved clinically significant weight loss (≥5%). Descriptive statistics were conducted to describe the baseline profile of the sample. Means, standard deviations (SD), and frequencies were calculated for clinical and demographic variables, as appropriate. A mixed model analysis with subjects as random effects was used for our primary and secondary outcome measures at 3, 6, 9 and 12 months after the initial visit. Time, intervention and the interaction between intervention and time were included as predictor variables while controlling for age, sex, and the baseline value of each outcome. The percentage of patients that lost ≥5% of their baseline weight at any timepoint were analyzed using chi-square tests and odds ratio (OR). The significance level for all tests was established at p < 0.05.

Results: The final analysis consisted of 378 patients (males: 46.3%, mean age: 34.6 +/− 11.81 years) who were followed in the clinic for at least 6 months. The mean baseline body weight of the sample was 97.3 +/− 23.1 kg. Overall Clinic Effect: A significant effect of the clinic (including all interventions) was observed on body weight over time (F = 7.72, p < 0.001). The mean change in weight was −2.45 +/− 0.52 kg (p < 0.001) at 12 months. Moreover, 132 (37%) patients lost ≥5% of their baseline body weight during the 12-month study period. Effect by Intervention: In this sample, 285 were receiving add-on therapy with a pharmacological intervention (Metformin: N = 227, Topiramate: N = 15, Semaglutide: N = 9, Other treatment combinations: N = 34), while the remaining 93 received lifestyle counselling only. No significant change in weight was observed at any of the timepoints for patients on non-pharmacological (lifestyle only) interventions. Among the pharmacological interventions, significant weight loss was observed at 12 months with metformin (−2.12 +/− 0.58 kg, p < 0.001), topiramate (−4.16 +/− 1.46 kg, p = 0.01), and semaglutide (−9.61 +/− 2.55 kg, p = 0.001). The presumptive number needed to achieve ≥5% weight loss was 3 for lifestyle interventions, 3 for metformin, 3 for topiramate, and 2 for semaglutide. Among the remaining patients on combination treatment (N = 35), there were very small samples for each of the individual combinations, precluding meaningful analysis.

Conclusions: This is the largest study to provide real-world clinical data on the value of a specialized psychiatrist-led metabolic clinic for patients with SMI experiencing psychotropic-associated weight gain. The results derived from this chart review demonstrate the effectiveness of an array of interventions to facilitate clinically significant weight loss, thereby indicating that clinical trial evidence translates well to clinical practice. The findings suggests that treating psychiatrists may be best positioned to monitor patients with SMI and intervene at the earliest signs of abnormality, and even prevent the onset of metabolic dysfunction, with psychotropic medications. This approach may streamline the implementation of monitoring strategies and interventions in a systematic manner. Nonetheless, further investigation in large controlled clinical trials is warranted, with additional timepoints collected beyond the 12-month mark.

Keywords: antipsychotic-induced metabolic dysfunction, Semaglutide, Cardiometabolic Risk, Metabolic Psychiatry, Schizophrenia Spectrum Illness.

Disclosure: Nothing to disclose.

P769. Symptom stability during treatment transition to xanomeline and trospium chloride: post hoc analysis of an inpatient trial in schizophrenia

Rachel Dyme, Jenna Hoogerheyde, Ayesha Pavithran, Ranjan Tiwari, James Appio, Scott Vuocolo

Bristol-Myers Squibb, Lawrenceville, New Jersey, United States

Background: Individuals with schizophrenia often switch antipsychotics. A trial of xanomeline and trospium chloride (X/T) requiring washout of prior antipsychotics (pAPD) provided an opportunity to examine safety and disease symptom stability during this transition.

Methods: An open-label, 2-cohort inpatient trial (NCT06572449) enrolled symptomatically stable adults with schizophrenia with PANSS total scores ≤80. Following a washout period, participants began treatment with X/T BID on an empty stomach for 4 weeks starting at 50mg/20mg and were then uptitrated to a maximum 125mg/30mg (Cohort 1) or 100mg/20mg (Cohort 2). Treatment continued for an additional 4 weeks at the maximum tolerated dose with food. PANSS total scores from the 2 cohorts were pooled and analyzed by pAPD and washout duration (WOD) during the transition from pAPD to X/T.

Results: A total of 168 participants had calculable washout periods at baseline. Median (min, max) WOD was 3 (1, 6) days. Mean±SD PANSS total scores were stable from screening (63.4 ± 10.9) to baseline (64.1 ± 10.7) and independent of WOD (r = 0.05). Mean±SD changes from screening to baseline in PANSS total score for individuals previously on aripiprazole, olanzapine, quetiapine, or risperidone were 1.5 ± 4.8 (n = 11), −0.4 ± 3.2 (n = 36), 0.9 ± 6.2 (n = 48), and 0.8 ± 3.5 (n = 35), respectively. PANSS total score improved from baseline to day 56 with X/T regardless of pAPD and WOD.

Conclusions: Overall, increases in PANSS total scores from screening to baseline were not seen when switching individuals from pAPD to X/T. These data suggest that stable individuals can be safely transitioned from pAPD to X/T without destabilization of schizophrenia symptoms.

Keywords: xanomeline, Treatment transition, Symptom stability.

Disclosure: Bristol-Myers Squibb, Employee, Self.

P770. Non-rapid eye movement sleep oscillations and cognition in aging people with schizophrenia

Bengi Baran, Amanda McCleery, Isabella Martinez, Hazal Arpaci, Stephanie Orellana, Brennan Kelly, Morgan Lott, Tess Filip, Marie E. Gaine, Krystal Parker, Aislinn Williams, Sarah Akhras, Sophia Koesterer

University of Iowa, Iowa City, Iowa, United States

Background: Cognitive deficits are a core feature of schizophrenia (SZ), with emerging evidence for a steeper cognitive decline than in typical aging. Non-rapid eye movement (NREM) sleep oscillations, particularly sleep spindles and their temporal coupling with slow oscillations (SO), are critical for memory and cognitive functioning. Prior work in young and middle-aged adults with SZ revealed marked reductions in spindle density and impaired SO-spindle coupling, both linked to cognitive deficits. Obstructive sleep apnea (OSA), highly prevalent and undertreated SZ, is independently associated with age-related cognitive decline. The goal of this study is to investigate the effects of OSA and SZ on NREM sleep oscillations and cognition older adults.

Methods: Adults aged 30–65 years with SZ and demographically matched non-psychiatric comparison (NC) subjects (n = 34; recruitment ongoing; matched for age, sex, BMI, and OSA presence/severity) completed one night of high-density polysomnography (PSG) with OSA monitoring and a sleep-dependent memory consolidation task. Clinical interviews and MATRICS Consensus Cognitive Battery (MCCB) were administered on subsequent days. We tested group differences and associations of age and apnea hypopnea index (AHI) with sleep, cognition, and NREM oscillations.

Results: Preliminary analyses reveal similar PSG-derived sleep architecture but reduced spindle density and spindle-SO coupling in people with SZ compared to NC. Across groups, older age was associated with reduced spindle density (r = −.53, p = .01) and spindle-SO coupling (r = −.63, p = .04); and higher AHI correlated with reduced spindle density (r = −.62, p = .002). People with SZ had reduced sleep-dependent memory consolidation (p = .03) and impaired performance across all MCCB domains (p’s < .019) except reasoning/problem solving (p = .2). Age showed a marginal negative correlation with global cognition in SZ (r = –.52, p = .06) but not NC. While people with SZ reported worse sleep quality and increased insomnia symptoms, actigraphy showed no group differences in objective sleep quality.

Conclusions: Preliminary findings suggest stronger age-related cognitive decline in SZ, potentially reflecting unique neurodegenerative trajectories. NREM oscillations demonstrate expected aging and OSA associations, though group differences are subtle in this early dataset. Ongoing analyses will formally test whether SZ and OSA jointly exacerbate age-related cognitive deficits via disrupted NREM sleep oscillations.

Keywords: Cognitive impairment associated with schizophrenia, NREM sleep, sleep-dependent learning, obstructive sleep apnea, aging with schizophrenia.

Disclosure: Nothing to disclose.

P771. Preliminary evaluation of synaptic density in first episode psychosis individuals with minimal exposure to antipsychotic medication: a [18F]SynVesT-1 positron emission tomography imaging study

Christin Schifani, Stephanie Ameis, Isabelle Boileau, Nicholas Neufeld, Wei Wang, Kim Desmond, Neil Vasdev, Aristotle Voineskos, George Foussias, Muhammad Husain

Centre for Addiction and Mental Health, Toronto, Canada

Background: Psychotic disorders, including schizophrenia, are associated with significant distress, poor functioning and premature mortality. Functional outcomes in psychotic disorders are influenced by duration of untreated psychosis, and interventions delivered at the earliest stages of illness can significantly impact functional trajectories in affected individuals. Schizophrenia has long been considered a neurodevelopmental disorder resulting from excessive synaptic pruning during late adolescence and early adulthood. Compelling evidence for reduced synaptic density in schizophrenia has come from recent in vivo positron emission tomography (PET) imaging studies using tracers targeting the synaptic vesicle glycoproteins 2A (SV2A). Reduced SV2A levels (i.e., synaptic density) have been reported in chronic schizophrenia and recently in first episode psychosis (FEP). Reduced synaptic density was associated with greater symptom severity and poorer cognition. Herein, we present early data replicating and extending initial findings of reduced synaptic density in FEP individuals with minimal antipsychotic exposure (<3 months). We also report on associations between synaptic density, functioning and disability.

Methods: FEP participants were recruited from the Early Intervention for Psychosis inpatient and outpatient services at the Centre for Addiction and Mental Health, Toronto, Canada. Eligibility: FEP participants had a DSM-5 diagnosis of psychotic spectrum disorder, were within 5 years of initial diagnosis, and had a total lifetime exposure of antipsychotic medication < 3 months. Healthy control (HC) participants had no DSM-5 psychiatric disorder, life-time exposure to psychotropic medication, and first-degree relative with a current/past DSM-5 diagnosis of schizophrenia spectrum or other psychotic disorders, autism, or affective disorder with psychotic features. Participants completed a 120-minute arterial [18F]SynVesT-1 PET scan (187 ± 10MBq) and T1-weighted MRI scan (for ROI delineation). After reconstruction, correction of attenuation and motion, PMOD was used to delineate grey matter regions of interest (ROIs) on each participant's T1-weighted MRI using the Hammers atlas and subject-specific tissue segmentation. Each participant’s MRI was co-registered to their summed [18F]SynVesT-1 PET images, the ROI delineation resliced with the resulting transformation and the time activity curves (TACs) generated. Distribution volume (VT) was estimated using the 1-tissue compartment model with arterial input function in 3 pre-specified ROIs: frontal cortex (FC), anterior cingulate cortex (ACC), and hippocampus (HPC). Disability was evaluated with the WHO-Disability Assessment Schedule (WHO-DAS) 2.0 and the Columbia Impairment Scale (CIS), and function with Birchwood Social Functioning Scale (SFS) and Global Functioning: Social and Role Scales (GF: Social and Role). Due to the initial smaller sample size, group differences in VT are presented as % difference and effect size (Cohen’s d). At time of presentation, we will show results of a mixed effect model with VT as dependent variable, group as predictor, and ROIs as control variable.

Results: 16 FEP (7 inpatients/9 outpatients, 11M/5F, Age = 23.8, illness duration < 6m) and 21 HCs (11M/10F, Age = 24.4) successfully completed the arterial [18F]SynVesT-1 PET scan and MRI. [18F]SynVesT-1 binding across ROIs was lower in FEP (10.4 ± 1.4%; Cohen’s d = 1.12, large effect size) compared to HCs. Differences in [18F]SynVesT-1 binding ranged from ~9–12% lower in FEP (FC: −8.8 ± 1.2%, ACC: −12.3 ± 1.8%, HPC: - 10.0 ± 1.1%) vs HCs. Across both groups, lower synaptic density in ACC was related to greater disability (WHO-DAS: R = −0.46; CIS: R = −0.54), and poorer social (GF: Social: R = 0.47; SFS: R = 0.38) and role functioning (GF: Role: R = 0.47). Although the results are preliminary (n = 16 FEP), with ongoing recruitment we anticipate we will have additional data on ~8 participants at the time of presentation based on our current recruitment rate.

Conclusions: Our findings support evidence of lower synaptic density across brain regions in FEP and the association of reduced synaptic density with greater disability and poorer functioning. More definitive evaluation of synaptic density in FEP will be possible once we have completed data collection and analysis of our desired sample (30 FEP vs 30 HCs). At that stage, we will be able to more accurately investigate relationships between synaptic density, cognition and functioning in individuals with FEP.

Keywords: synaptic density, first episode psychosis, everyday functioning, Synaptic Vesicle Glycoprotein 2A (SV2A), Positron Emission Tomography (PET) Imaging.

Disclosure: Nothing to disclose.

P772. Toward a comprehensive characterization of anxiety in schizophrenia spectrum disorders: integrating categorical and dimensional perspectives

Marren Jenkins, Victoria Fox, Stephan Heckers, Neil Woodward, Jennifer Blackford, Brandee Feola

Vanderbilt University Medical Center, Nashville, Tennessee, United States

Background: Anxiety is highly prevalent in people with schizophrenia spectrum disorders and is associated with numerous adverse outcomes, including greater symptom severity, elevated suicide risk, increased likelihood of comorbid substance and alcohol use disorders, and reduced quality of life. Emerging evidence suggests that anxiety is a significant factor in the development and maintenance of schizophrenia spectrum disorders. Despite its clinical significance, anxiety—both as a categorical diagnosis and a dimensional construct—has been relatively understudied in schizophrenia spectrum disorders. Therefore, the current study aims to provide a comprehensive characterization of anxiety in schizophrenia spectrum disorders using both categorical and dimensional frameworks and examine its associations with psychosis symptoms.

Methods: As part of an ongoing neuroimaging study, detailed interview-based and self-report measures of anxiety and psychosis symptoms were collected among three groups: individuals with schizophrenia spectrum disorders and co-occurring anxiety disorders (SZ+ANX, n = 47); individuals with schizophrenia spectrum disorders without anxiety disorders (SZ-ANX, n = 49); and individuals with no history of psychiatric disorders (HC, n = 55). Interview-based measures included the Anxiety Disorders Interview Schedule (ADIS) to diagnose anxiety disorders and the Positive and Negative Syndrome Scale (PANSS) to assess psychosis symptoms. Self-report questionnaires measured anxiety symptoms (state anxiety, social anxiety, worry, perceived stress), negative affect traits (trait anxiety, intolerance of uncertainty, inhibited temperament), and paranoia. ANOVAs were used to examine group differences (SZ+ANX, SZ-ANX, HC) in anxiety symptoms, negative affect traits, and psychosis symptoms with age, sex, and race as covariates. Correlations were conducted to determine associations between anxiety and psychosis symptoms.

Results: Within the SZ+ANX group, social anxiety was the most common anxiety disorder, with 80.49% of the individuals in the SZ+ANX group having a current social anxiety disorder. When the three groups were compared, SZ+ANX disorders had the highest levels of anxiety symptoms, negative affect traits, and paranoia. Yet, SZ-ANX also had significantly elevated anxiety symptoms, negative affect traits, and paranoia relative to HC. More specifically, the SZ+ANX > SZ-ANX > HC gradient pattern was shown for anxiety symptoms [state anxiety (p < .001; η²_p = .26), social anxiety (p < .001; η²_p = .43), worry (p < .001; η²_p = .33), stress (p < .001; η²_p = .34)]; negative affect traits [trait anxiety (p < .001; η²_p = .17), intolerance of uncertainty (p < .001; η²_p = .30), inhibited temperament (p < .001; η²_p = .19)]; and paranoia (p < .001; η²_p = .20). In addition, the SZ+ANX group had higher levels of PANSS distress symptoms than the SZ-ANX group (p < .001; η²_p = .26). Correlation analyses revealed positive correlations between anxiety measures with self-reported paranoia (ps < .001; r = .51-.79), PANSS positive symptoms (ps < .01; r = .26-.48), and PANSS distress symptoms (ps < .001; r = .40-.66).

Conclusions: The current study highlights prominent anxiety symptoms and negative affect traits in people with schizophrenia spectrum disorders, including people with and without co-occurring anxiety disorders, that are associated with greater severity of positive psychosis symptoms. Further characterizing anxiety in schizophrenia spectrum disorders may enhance our understanding of the underlying mechanisms of schizophrenia, inform current treatment strategies, and support the development of more targeted and effective interventions.

Keywords: Anxiety, Psychosis, Negative Affect, Vulnerability Traits.

Disclosure: Nothing to disclose.

P773. A single-channel wearable EEG device accurately measures sleep oscillations and their deficits in schizophrenia

Ahmad Mayeli, Chloe Huston, Omeed Chaichian, Cynthia Cheng, Caitlin Moore, Sabine Janssen, Fabio Ferrarelli

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Sleep spindles and slow waves are the two main oscillatory activities characterizing deep non-rapid eye movement (NREM) sleep. Recent studies using high-density EEG and polysomnography (PSG) have demonstrated marked reductions in both spindles and slow waves in individuals with SCZ. However, most evidence comes from single-session laboratory studies using EEG systems that are not well suited for home recordings and repeated, longitudinal assessments, which are critical for understanding the onset and progression of sleep oscillatory deficits in SCZ in real-world settings. Individuals with early course SCZ (EC-SCZ) represent a critical population where sleep biomarkers may enable timely interventions. Single-channel wireless EEG offers transformative potential for longitudinal sleep monitoring, providing accessibility for repeated assessments while maintaining measurement precision. Here, we assessed the suitability of a single-channel, wearable EEG device for measuring sleep oscillations against gold-standard PSG in individuals with EC-SCZ versus healthy controls (HC).

Methods: We collected simultaneous single-channel wireless EEG (Fpz channel) and PSG recordings (F3, F4 channels) during adaptation (Night 1) and baseline (Night 2) sleep recordings in N = 30 HC (12F, age 24.4 ± 4.8 years) and N = 31 individuals with EC-SCZ (11F, age 24.3 ± 4.7 years). Participants were 18–40 years of age, with individuals with EC-SCZ having ≤5 years from illness onset and currently meeting Diagnostic and Statistical Manual of Mental Disorders-5 (DSM-5) diagnosis of SCZ, schizophreniform, or schizoaffective disorder. Sleep spindle density, spindle duration, and slow wave density were quantified from baseline night recordings using automated detection algorithms. For PSG recordings, we also computed F3-F4 averages representing bilateral frontal activity. Statistical analyses included unpaired t-tests comparing groups, Cohen’s d calculation, Pearson correlations examining wireless EEG with PSG measures separately for each group, and partial correlations controlling for group effects to assess overall device validation.

Results: Individuals with EC-SCZ demonstrated significant reductions across all sleep oscillation parameters compared to HC. For slow wave density: wireless EEG (HC: 37.5 ± 1.9 vs SCZ: 34.9 ± 4.1, p = 0.002, d = 0.83), F3 (HC: 37.3 ± 2.4 vs SCZ: 36.0 ± 2.2, p = 0.028, d = 0.58), F4 (HC: 37.2 ± 1.4 vs SCZ: 35.4 ± 3.2, p = 0.009, d = 0.69), and F3-F4 average (HC: 37.2 ± 1.8 vs SCZ: 35.7 ± 2.6, p = 0.009, d = 0.69). For spindle density: wireless EEG (HC: 2.5 ± 0.6 vs SCZ: 1.9 ± 0.7, p = 0.001, d = 0.88), F3 (HC: 3.3 ± 0.9 vs SCZ: 2.6 ± 0.9, p = 0.002, d = 0.81), F4 (HC: 3.1 ± 1.1 vs SCZ: 2.2 ± 1.0, p = 0.001, d = 0.86), and F3-F4 average (HC: 3.2 ± 1.0 vs SCZ: 2.4 ± 0.9, p = 0.001, d = 0.86). For spindle duration: wireless EEG (HC: 1.0 ± 0.2 vs SCZ: 0.8 ± 0.1, p = 0.005, d = 0.75), F3 (HC: 1.1 ± 0.2 vs SCZ: 0.9 ± 0.2, p = 0.010, d = 0.69), F4 (HC: 1.0 ± 0.2 vs SCZ: 0.9 ± 0.2, p = 0.035, d = 0.56), and F3-F4 average (HC: 1.1 ± 0.2 vs SCZ: 0.9 ± 0.2, p = 0.017, d = 0.63). Furthermore, wireless EEG measurements showed highly significant partial correlations with PSG F3-F4 averages, after controlling for diagnostic group: slow wave density (partial r = 0.78, p < 0.001), spindle density (partial r = 0.82, p < 0.001), and spindle duration (partial r = 0.80, p < 0.001). Group-specific Pearson correlations between wireless EEG and PSG F3-F4 averages were also highly significant in both HC (r = 0.68–0.87, all p < 0.001) and individuals with EC-SCZ (r = 0.73–0.82, all p < 0.001), demonstrating consistent device accuracy across diagnostic groups.

Conclusions: These findings provide the first validation that sleep spindle and slow wave deficits can be accurately detected using a wearable, single-channel wireless EEG system in individuals with SCZ. The highly significant correlations between wireless EEG and gold-standard PSG, after controlling for diagnostic group as well as within each group separately, validate the clinical utility of portable devices for assessing sleep oscillatory deficits in SCZ. As such, this approach offers transformative potential for at-home, longitudinal monitoring of sleep biomarkers in SCZ, thus enabling early detection of sleep oscillatory deficits, treatment response evaluation, and sleep-targeted interventions in real-world clinical settings.

Keywords: Schizophrenia (SCZ), Sleep oscillations, Wearable EEG.

Disclosure: Nothing to disclose.

P774. Neuromelanin MRI contrast in patients with treatment-resistant schizophrenia: a cross-sectional study

Fumihiko Ueno, Shiori Honda, Shinichiro Nakajima, Clifford Cassidy, Yusuke Iwata, Guillermo Horga, Edgardo Torres-Carmona, Jianmeng Song, Takahide Etani, Saki Honma, Shunya Sekihara, Sakiko Tsugawa, Vincenzo de Luca, Sri Mahavir Agarwal, Gary Remington, Philip Gerretsen, Ariel Graff-Guerrero

University of Ottawa, Ottawa, Canada

Background: Neuromelanin-sensitive MRI (NM-MRI) enables in vivo assessment of monoaminergic activity in brainstem regions including the substantia nigra/ventral tegmental area complex (SN-VTA) and locus coeruleus (LC), which reflect dopaminergic and noradrenergic function, respectively. We investigated whether NM-MRI can distinguish treatment-responsive and treatment-resistant schizophrenia (TRS) subtypes, including clozapine-unresponsive cases.

Methods: Seventy-eight participants were enrolled across four groups: clozapine-unresponsive TRS (ultra-resistant schizophrenia [URS], n = 16), clozapine-responsive TRS (non-URS, n = 16), first-line antipsychotic responsive schizophrenia (FLR, n = 20), and healthy controls (HCs, n = 26). NM-MRI contrast ratios (CR) in the SN-VTA and LC were compared across groups using ANCOVA controlling for age and sex, with correction for multiple comparisons. Voxel-wise analyses were performed in the SN-VTA for subgroups with significant CR differences.

Results: Data from 76 participants were analyzed. SN-VTA CR was significantly higher in URS (Cohen’s d = 1.01) and FLR (Cohen’s d = 0.94) compared to HCs, but not in non-URS. Voxel-wise analyses confirmed increased signal localized to the dorsal SN-VTA in both URS (549/1807) and FLR (556/1807). LC CR was significantly higher in URS relative to FLR (Cohen’s d = 0.99), with no other group differences. No significant associations were found between NM signal and clinical characteristics.

Conclusions: Elevated NM signals in the SN-VTA were found in both URS and FLR, suggesting dopaminergic hyperactivity underlies both treatment response and resistance. Only URS showed elevated LC NM signal, suggesting additional noradrenergic dysregulation. NM-MRI may offer a non-invasive biomarker to differentiate treatment-responsive subtypes in schizophrenia.

Keywords: Neuromelanin-sensitive MRI, treatment-resistant schizophrenia, substantia nigra, Locus coeruleus.

Disclosure: Nothing to disclose.

P775. Auditory and visual hallucinations arise from modality-specific learning dynamics

Alexandria Bond, Peter Waade, Eren Kafadar, Victoria Fisher, Christoph Mathys, Albert Powers

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

Background: The fundamental question of whether hallucinations in different sensory modalities share common mechanisms or arise from distinct computational processes remains unresolved. Understanding this distinction has critical implications for psychosis research: modality-specific mechanisms would implicate sensory system disruption as an upstream deficit driving hallucination development, while modality-general mechanisms would point to higher-level cognitive processing abnormalities. Previous work has established that hallucinations may arise from excessive weighting of prior beliefs relative to sensory evidence within a Bayesian framework of perception, but whether this pattern operates uniformly across sensory modalities or exhibits modality-specific characteristics has not been systematically investigated in large samples.

Methods: We administered auditory and visual Conditioned Hallucination (CH) tasks to 727 sex-matched participants with varying hallucination symptomatology. Participants were categorized into four groups: those experiencing both auditory and visual hallucinations (A+V+, n = 367), auditory-only (A+V-, n = 111), visual-only (A-V+, n = 60), and neither (A-V-, n = 189). The conditioned hallucination tasks employed Pavlovian associative learning paradigms where participants learned associations between difficult-to-detect targets in one modality and salient cues in another, then reported target presence/absence across trials with varying stimulus intensities and target-absent conditions. We used hierarchical Bayesian computational modeling (the Hierarchical Gaussian Filter) to quantify perceptual belief dynamics, capturing both belief formation during conditioning and persistence during absent sensory input. This approach revealed modality-specific computational mechanisms including belief updating rates (ω), prior precision (ν), and decision noise parameters. We also conducted sensory noise simulations to test whether differential noise across modalities could generate the observed computational signatures.

Results: Participants exhibited robust modality-specific biases aligned with their real-world hallucination profiles. Those with auditory hallucinations showed enhanced auditory sensitivity and significantly higher auditory conditioned hallucination rates with greater confidence (∆μ = 0.068, p < 0.0001), while those with visual hallucinations demonstrated the opposite pattern (∆μ = −0.028, p < 0.0001). Computational modeling revealed that these biases arose from two key modality-specific mechanisms: rapid belief updating in response to confirmatory sensory evidence and sustained, hyper-precise priors that maintained beliefs despite contradictory evidence. Both groups demonstrated hyper-precise priors in their symptom-congruent modality. In participants with multimodal hallucinations, these computational dynamics tracked with symptom dominance–prior precision and belief updating were stronger in the sensory domain matching their predominant hallucination type. The resulting asymmetric perceptual system showed enhanced belief formation during early learning followed by resistance to belief updating when stimuli were absent, creating a self-reinforcing cycle. Sensory noise simulations confirmed that modality-specific noise imbalances could generate the specific computational signatures observed in empirical data, supporting theoretical accounts linking cortical hyperexcitability to reduced signal-to-noise ratios and a compensatory overreliance on priors.

Conclusions: These findings suggest that hallucinations may arise from modality-specific imbalances in the temporal dynamics of belief formation and maintenance, rather than from domain-general cognitive abnormalities. The results reveal a computational motif present across psychotic illness: rapid formation of new beliefs followed by resistance to updating despite contradictory evidence. This asymmetric learning system creates self-reinforcing loops between expectation and perception that may maintain hallucination-prone states. The modality-specific nature of these computational biases suggests that sensory system disruption may be a critical upstream factor in hallucinogenesis, with differential sensory noise across modalities determining the balance of computational biases and consequently the frequency of hallucinatory percepts in each channel. However, the frequent coexistence of cross-modal hallucinations indicates broader network involvement, likely through interactions between modality-specific computational biases and large-scale network dysfunction affecting salience processing, thalamic gating, and shared attentional resources. These insights provide new targets for understanding the computational origins of hallucinatory experience and suggest that effective treatments may need to address both modality-specific sensory processing abnormalities and broader network-level dysfunction.

Keywords: psychosis, hallucinations, computational psychiatry.

Disclosure: Nothing to disclose.

P776. Contribution of visual dorsal attention network dysconnectivity to cognitive deficits in first-episode psychosis

Alfredo Sklar, Dylan Seebold, Jenay Kocsis, Lauren Fowler, Hayley Rhorer, Jack Kavanagh, Brian Coffman, Dean Salisbury

University of Pittsburgh Western Psychiatric Institute and Clinic, Pittsburgh, Pennsylvania, United States

Background: Cognitive impairment represents a source of significant disease debility and long-term functional impairments in schizophrenia. Despite its onset prior to formal diagnosis and evidence for a superior response to cognitive training during early illness stages, few efficacious treatments exist targeting these deficits in first-episode psychosis (FEP). Impaired selective attention, the prioritization of stimuli in our environment for enhanced processing based on internal goals, is a core feature of the disorder and one of its most consistently identified cognitive deficits. Neural connectivity is critical to the execution of complex cognitive functions with communication across frontal and parietal nodes of the dorsal attention network (DAN) implicated in the deployment of visual attention to items in space. The current study sought to examine the integrity of functional connectivity between the DAN and visual cortex in FEP and its contribution to visual perceptual modulation and disease burden proximal to illness onset.

Methods: Nineteen FEP and 25 healthy controls (HC) were included in the study. Magnetoencephalography (MEG) was recorded during a covert visual spatial attention task. Gabor patches, rotated slightly off vertical, were presented at 4 contrast levels on the left or right side of the screen. Trials included either neutral or valid location cues prior to gabor presentation. Eye position was monitored throughout the task and structural MRIs were obtained for each participant to localize MEG sensor activity to the cortical surface. Responses to gabor targets were isolated within primary visual cortex (V1) and the weighted phase-lag index was used to examine functional connectivity between V1 and the DAN during cue presentation. Clinical assessments included MATRICS consensus cognitive battery (MCCB) and Global Functioning: Role and Social (GFR/GFS) scales for all participants and the Positive and Negative Syndrome Scale (PANSS) for patients.

Results: Despite equivocal maximal V1 responses between groups across trial type (p = .18), FEP exhibited an impaired enhancement of these responses during valid cue trials (t18 = 0.37, p = .71) relative to HC (t24 = 3.27, p = .003). FEP also exhibited impaired high alpha-band (10 – 12Hz) functional connectivity between both the frontal (t41 = 2.49, p = .02) and parietal (t41 = 2.16, p = .04) nodes of the DAN and V1 relative to HC. In HC, DAN—V1 connectivity was associated with V1 response enhancement (frontal: rho = .53, p = .008; parietal: rho = .56, p = .004). While this relationship was not present in FEP (p’s > .4), there was a relationship between impaired DAN—V1 connectivity and lower MCCB scores (frontal: rho = .46, p = .047; parietal: rho = .58, p = .01) as well as a trend-level relationship with higher PANSS disorganization factor scores (frontal: rho = −.40, p = .1).

Conclusions: Results from this study indicate an inability of FEP to engage endogenous attention networks to modulate visual processing in the service task goals. Importantly, disruptions in the communication between the DAN and V1 was associated with cognitive impairments as measured by both standardized neurocognitive measures and symptom assessments. These findings elucidate sources of cognitive impairments in FEP and, potentially, therapeutic targets for future interventions to ameliorate these debilitating symptoms which are currently lacking. This study also highlights the benefits of using neurophysiological techniques with high spatio-temporal resolution to identify frequency-specific interactions between sensory and executive brain networks occurring on a millisecond timescale. Future work will examine integrity of the ventral attention network subserving exogenous attentional reorienting and its coordination with the DAN in providing flexible control of attention in FEP based on task and environmental demands.

Keywords: first-episode psychosis, magnetoencephalography, Cognitive impairment associated with schizophrenia, task-based functional connectivity, Dorsal Attention Network.

Disclosure: Nothing to disclose.

P777. Associations between event-related potentials and theta oscillations assessed from a mismatch negativity paradigm in NAPLS2 individuals at clinical high-risk for psychosis and healthy controls

Jessica Hua, Brian Roach, Holly Hamilton, Peter Bachman, Aysenil Belger, Tyrone Cannon, Ricardo Carrion, Erica Duncan, Jason Johannesen, Gregory Light, Margaret Niznikiewicz, Jean Addington, Carrie Bearden, Kristin Cadenhead, Diana Perkins, Elaine Walker, Scott Woods, 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. 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 earlier psychosis onset.

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 event-locked EEG epochs. These neuro-oscillatory approaches provide an improved method for translating response patterns across species. Notably, TFA studies suggest that theta oscillatory abnormalities contribute to MMN schizophrenia deficits.

Using this same sample of participants from the 8-site North American Prodrome Longitudinal Study 2 (NAPLS2), we examined group differences in MMN amplitude as well as theta oscillations in future CHR-P converters (CHR-C), CHR-P who were followed for 24 months and did not convert (CHR-NC), and healthy controls (HC). In a previously published paper, we found MMN amplitude reduction in future CHR-P converters (CHR-C) relative to non-converters (CHR-NC) and healthy controls (HC). Moreover, in a paper under review, we found CHR-C had reduced theta ITC for standards and deviants relative to HC and CHR-NC as well as reduced theta total power relative to HC.

However, it is unclear to what degree MMN and theta deficits account for each other in terms of sensitivity to underlying pathophysiological processes in CHR-P. Results from our prior study in an early schizophrenia sample (ESZ; within five years of psychosis onset) showed that covarying for one (MMN amplitude or theta) did not fully account for deficits in the other, raising the possibility of their sensitivity to dissociable pathophysiological processes In the current study, we examined associations between theta oscillations and MMN amplitude.

Methods: CHR-P (n = 580) and HC (n = 241) were recruited as part of the NAPLS2 study. 77 CHR-P converted to psychosis (CHR-C), and 238 were followed clinically for ≥24 months without converting (CHR-NC). Participants completed a MMN paradigm consisting of frequent standard sounds (85%) and three different types of infrequent (5%) deviant sounds.

To dissociate the effects of MMN and theta oscillations, we used analyses of covariance (ANCOVAs) to assess whether theta deficits persisted when covarying for corresponding MMN. As MMN is measured from ERP deviant-standard difference waves, ANCOVAs examined whether the theta deviant-standard difference score, or the average score across the standards and mean of 3 deviants (thereby giving equal weight to standards and deviants), accounted for the MMN deficit. MMN amplitude, averaged across deviant types, was compared between CHR-C and CHR-NC in four separate ANCOVA models, covarying for theta ITC or total power averaged over stimulus types or as difference scores between deviants and standards (i.e., the mean of the deviants minus standards).

Results: The previously reported significant deficit in overall MMN amplitude, averaged across deviant types, in CHR-C, relative to CHR-NC, was no longer significant (Group β = .02, p = .695) after accounting for theta ITC averaged across all stimuli (ITC β = −.53, p < 001). However, this group difference remained significant (Group β = .12, p = .022) when covarying for the theta ITC deviant-standard difference score (ITC β = −.32, p < .001). Similarly, the CHR-C MMN amplitude deficit, relative to CHR-NC, was no longer significant (Group β = .07, p = .187) when covarying for theta total power averaged across all stimuli (power β = −.38, p < .001) but remained significant (Group β = .11, p = .039) when covarying for the theta total power deviant-standard difference score (power β = −.28, p < .001). Interpretation of these ANCOVA results is facilitated by examination of the correlations of MMN amplitude z-scores with theta ERO deviant-standard difference and cross-stimulus average z-scores, which show that MMN amplitudes tended to be more strongly related to theta ITC cross-stimulus scores than theta deviant-standard difference scores, a pattern that was similarly evident across groups. A similar pattern was evident when raw amplitudes and ITC were correlated. For theta total power, MMN amplitudes are similarly or more strongly related to theta power cross-stimulus scores for CHR-C and CHR-NC than theta deviant-standard difference scores.

Conclusions: While controlling for theta deviant-standard difference scores did not eliminate the significant reduction in MMN in CHR-C, controlling for theta ITC or total power averaged across stimuli fully accounted for CHR-C MMN deficits. MMN was also more highly correlated with theta oscillatory measures averaged across stimuli than with deviant-standard difference scores, a pattern that was similarly evident across all groups. Thus, these data suggest that abnormal theta oscillations in general are a more significant driver of MMN deficits in CHR-C than the differential theta response to deviants relative to standards.

Keywords: Clinical high-risk for psychosis, Auditory Mismatch Negativity, time-frequency analysis, event-related oscillations, theta band oscillatory measures.

Disclosure: Nothing to disclose.

P778. Electrophysiological biomarker for working memory and cognitive function in schizophrenia

Xiaoming Du, Yizhou Ma, Ann Summerfelt, Alina Siatka, Ankeeta LNU, Bhim Adhikari, Peter Kochunov, Elliot Hong

The University of Texas Health Science Center At Houston, Houston, Texas, United States

Background: Cognitive deficits are core features of schizophrenia alongside positive and negative symptoms. Over 75% of patients with schizophrenia experience cognitive deficits with impaired working memory being chief among them. Identifying electrophysiological biomarkers for cognitive dysfunction in schizophrenia is critical for advancing our understanding of the neural underpinnings of these deficits and for guiding precision treatments. The present study used electroencephalographic (EEG) recordings to examine alterations in working memory in schizophrenia.

Methods: We recruited 190 patients with schizophrenia spectrum disorders (SSD) and 184 healthy controls (HC) to complete a 4-minute auditory continuous performance (ACP) task. 64-channel EEG data were recorded during the auditory 1-back task. During the task, participants listened to a series of tones and were asked to compare the current tone with the previous one. They responded by pressing one of two buttons to indicate whether the two tones were “same” or “different.” Response accuracy, reaction times, and d-prime were calculated as indices of ACP task performance.

Event related potential (ERP) components were extracted from the grand averages of correct trials. The N1 and P2 components were identified at frontal sites (e.g., Fz), while the P3b component was most prominent at parietal sites (e.g., Pz, CPz). Behavioral performances and ERP components were compared between SSD and HC groups. Correlation analyses, controlling for age and gender, were conducted to examine associations between ERP components, behavioral performances, and symptom severity. False discovery rate correction was applied for multiple comparisons (q = 0.05).

Results: The SSD and HC groups matched in age (SSD: 36.3 (mean) ± 3.3(standard deviation) vs HC: 35.9 ± 13.2 years) while there were more males in SSD group (77% male) than in HC group (51% male) (χ² = 29.3, p < .001). The HC group showed better performance: d prime, t = 7.0, p < .001; response accuracy, t = 6.1, p < .001; and reaction times, t = −2.0, p < .05. Significantly reduced N1 (t = −4.3, p < .001), P2 (t = 3.0, p = .003 at Fz), and P3b (t = 3.4, p = .001 at CPz) components were observed in SSD patients compared to HC. Larger P3b was associated with better non-auditory working memory (SSD: partial r = .28, p < .001; HC: partial r = .17, p = .04) in both groups and faster processing speed (partial r = .35, p < .001) in SSD, and also better auditory working memory ACP performance (d prime: partial r = .30, p < .001; accuracy: partial r = .36, p < .001) in patients with SSD.

Conclusions: Performing the auditory continuous performance task requires sustained attentional control and working memory. Patients with schizophrenia spectrum disorders exhibited poorer performance in auditory working memory performance and reduced P3b component. Moreover, P3b component, an ERP marker associated with attentional resource allocation and working memory updating, robustly predicted working memory as well as process speed deficits in patients. These findings underscore P3b as a promising electrophysiological biomarker for characterizing cognitive dysfunction in schizophrenia.

Keywords: schizophrenia, EEG, biomarker, working memory, P3b.

Disclosure: Nothing to disclose.

P779. Event-related theta responses as candidate markers of cognitive impairment associated with schizophrenia: independent replication of large case-control differences and cognitive associations in a large patient sample

Guhan Sundar, Chao Wang, Xiaoming Du, Ann Summerfelt, Yizhou Ma, Hemalatha Sampath, Ankeeta LNU, Adam Savitz, Elliot Hong, Amit Etkin, Patricio O'Donnell

Alto Neuroscience, Cambridge, Massachusetts, United States

Background: Cognitive impairment associated with schizophrenia (CIAS) remains a core, disabling feature with limited treatment options. Rapid, mechanism-sensitive biomarkers could accelerate early drug development, where clinical outcomes often emerge slowly and variably. In a recent analysis of 1,266 participants from the Bipolar-Schizophrenia Network for Intermediate Phenotypes (BSNIP) studies, theta-band inter-trial coherence (ITC) and event-related spectral perturbation (ERSP) emerged as the most blunted electroencephalography (EEG) markers in patients with schizophrenia, outperforming traditional event-related potentials (ERPs) and resting-state EEG measures in their association with cognitive deficits (the strongest being with processing speed). Here, we sought to replicate and extend these findings in an independent cohort.

Methods: Data from 155 individuals with schizophrenia and 272 healthy controls from the University of Texas Health Science Center, Houston / University of Maryland Baltimore were analyzed. Theta ITC and ERSP were derived from an auditory single-stimulus paradigm. Resting-state theta power was estimated during an eyes-closed condition, and mismatch negativity (MMN) amplitude was obtained from a passive auditory oddball task. Cognitive performance was assessed using the digit symbol substitution test (processing speed). Case–control effect sizes were calculated, and partial correlations between EEG markers and cognitive measures were computed separately within each group, adjusting for age, sex, race, and education.

Results: Theta ITC and ERSP showed the largest case–control differences (Cohen’s d = 0.64 and 0.78, respectively; p < 0.001), exceeding MMN (d = 0.41, p < 0.001) and resting theta power (d = 0.40, p < 0.01). Both theta ITC and ERSP correlated with processing speed in schizophrenia patients (ITC: r = 0.17, p < 0.05; ERSP: r = 0.23, p < 0.01). No significant associations were observed between processing speed and either MMN or resting theta power (p > 0.05).

Conclusions: Theta ITC and ERSP strongly differentiate patients with schizophrenia from healthy controls and show consistent associations with processing speed, replicating previous findings from BSNIP in a large independent dataset. These results strengthen the argument for theta-band responses as candidate biomarkers of disease-related processes and potential targets for CIAS drug development, providing further support for their outperformance of traditional ERP and resting EEG biomarkers.

Keywords: schizophrenia, EEG, ERP, cognition, biomarkers.

Disclosure: Alto Neuroscience, Employee, Self, Alto Neuroscience, Stock/Equity - Publicly Traded Company, Self.

P780. Study medication adherence and cognition outcomes in cognitive impairment associated with schizophrenia trials: insights from the erudite study with luvadaxistat

Ni Khin, Reuben Fan, Tintting Ge, Hans S. Klein, Mahnaz Asgharnejad, Jacob S. Ballon, Satjit Brar, Philip D. Harvey, Joshua T. Kantrowitz, Richard S. E. Keefe, Venketesha Murthy, Eiry Roberts, Jaskaran B. Singh

Neurocrine Biosciences, Inc., San Diego, California, United States

Background: High medication adherence is critical for interpreting efficacy data accurately. Digital tools, such as a smartphone app(‘Platform’), can be used to assist with monitoring adherence. Previously, we presented findings from the INTERACT study (NCT03382639), evaluating luvadaxistat as adjunctive therapy in negative symptoms of schizophrenia, in which daily doses of luvadaxistat 50 mg showed a potential efficacy signal in cognitive performance. In the INTERACT study, high medication adherence was demonstrated using multiple monitoring methods: a video analysis-based Platform, traditional pill counts, and detectable investigational drug (luvadaxistat) plasma levels with 90% concordance. As the cognitive outcome signal seen with the 50 mg dose of luvadaxistat in INTERACT was not replicated in the phase 2 ERUDITE study (NCT05182476), we conducted exploratory analyses using data from ERUDITE on adherence parameters and their correlation to cognitive performance.

Methods: ERUDITE was a randomized, double-blind, placebo-controlled study that assessed the effects of luvadaxistat on cognitive impairment associated with schizophrenia (CIAS). Following a 2-week placebo run-in, participants (aged 18–50 years) who were already receiving stable antipsychotic medication were randomized 2:1:1 to either placebo, luvadaxistat 20 mg, or luvadaxistat 50 mg for a 12-week treatment period. The primary efficacy endpoint was change from baseline (Day 14) to the study’s endpoint (Day 98) in the Brief Assessment of Cognition in Schizophrenia (BACS) composite score.

The efficacy analysis set (EAS) included all randomized participants who demonstrated treatment compliance (≥75% adherence between Day 1 to 14, measured via the Platform app, pill count, and interviews) and stable symptoms (Positive and negative syndrome scale [PANSS] scores within 85–115% of the mean across Day 1 to 14).

Platform adherence was defined as the ratio of the number of pills logged on the Platform vs. the expected number of pills. Pill count compliance was also determined by taking the ratio of the actual vs expected counts based on the number of pills dispensed minus the number of pills returned at each study onsite visit. To measure concordance between Platform use and pill count, adherence thresholds were set at 75% and 90%, respectively. Concordance with detectable luvadaxistat plasma levels was based on samples collected at the final study visit (Day 98) and adherence data from the final 2 weeks of the double-blind treatment period (Day 84 to 98).

Results: Of 214 enrolled participants, 203 were randomized and 173 met the EAS criteria; 21 (12.1%) of these participants discontinued treatment prematurely. Baseline demographics (mean age 37.5 yrs, mostly Whites, majority were males: 58% in placebo, ~70% in luvadaxistat arms) and clinical characteristics (mean PANSS total score ~61 and CGI-S 3.5) were similar across treatment arms.

Of 173 participants in the EAS, 160 (94.1%) had Platform data available for analysis. Within the first 4 weeks of the treatment period, 139/160 participants (86.9%) were able to achieve > 75% of the Platform adherence (high-adherent group) they attained during the placebo-lead in period, while 21 (13.1%) showed < 75% Platform use (low-adherent group). Of these, 13 (9.3%) from the high-adherent group and 5 (23.8%) from the low-adherent group discontinued early.

Of 152 participants in the EAS with both Platform data and pill count data available, 123 (81%) had both > 75% adherence to Platform and ≥90% pill count compliance: of these, 60 (49%) received placebo and 63 (51%) received luvadaxistat.

Of 65 participants in the EAS with Platform adherence data and plasma samples available for analysis, 60 (92%) in the luvadaxistat arms had detectable luvadaxistat levels. In total, 55 of 65 participants (~85%) met all 3 thresholds across all 3 measures (≥75% Platform use, ≥90% pill-count, and detectable plasma luvadaxistat).

Overall, the efficacy results showed no significant benefit of 20 mg or 50 mg doses of luvadaxistat versus placebo on mean BACS composite scores (20 mg: 1.9; 50 mg: 2.1; placebo: 2.6). When considering subsets of participants with higher Platform adherence thresholds (e.g., >80%, >85%, >90%) or both Platform (>75%) and pill count (90%) high adherence, the mean change from baseline to study endpoint in BACS composite scores in both luvadaxistat dose groups did not appear to show significant improvement compared with placebo in this study. At baseline, the luvadaxistat 50 mg arm had a numerically higher mean (SD) BACS composite score of 36.0 (12.2) versus 32.8 (13.0) for the placebo in the EAS. The high Platform and pill count adherent group in the luvadaxistat 50 mg arm also had a numerically higher baseline mean BACS of 36.4 (11.4) versus 32.2 (13.0) for the placebo. During the placebo run-in period, ~28% of participants showed highly variable changes in BACS composite score (>0.5 SD beyond practice effects).

Conclusions: Luvadaxistat 20 mg or 50 mg did not show a treatment effect on cognitive outcome. Adherent behaviors observed early in the ERUDITE study seemed to predict continued adherence throughout the treatment period; this observation was supported by the integration of the Platform, pill count and detectable luvadaxistat plasma levels. High confidence in study medication adherence was helpful in interpretation of efficacy outcomes. Further exploratory analyses of cognitive variability may enhance CIAS trial design.

Keywords: Cognitive impairment associated with schizophrenia, Medication adherence, Clinical trial, clinical efficacy.

Disclosure: Neurocrine Biosciences Inc, Employee, Self.

P781. Acute diazepam modulation of hippocampal activity and functional connectivity interactions in psychosis risk: preliminary evidence

Nicholas Livingston, Eva Smith, Amanda Kiemes, Owen O'Daly, Samuel Knight, Paulina Lukow, Luke Jelen, Thomas Reilly, Aikaterini Dima, Maria Nettis, Cecilia Casetta, Gabriel Devenyi, Thomas Spencer, Andrea De Micheli, Paolo Fusar-Poli, Anthony Grace, Steve Williams, Philip McGuire, Mallar Chakravarty, Alice Egerton, Gemma Modinos

Institute of Psychiatry, Psychology and Neuroscience, King's College London, London, United Kingdom

Background: Hippocampal hyperactivity, particularly from the CA1 subfield, has been proposed to drive downstream dysfunction within a cortico-limbic-striatal circuit involved in the emergence of psychosis. Preclinical evidence has demonstrated that the GABA-enhancing drug diazepam can prevent the emergence of a psychosis-relevant behavioural and physiological phenotype, through downregulating hippocampal hyperactivity and normalising neuronal activity in this circuit. Our recent work forward-translated this mechanistic evidence in humans, demonstrating that acute diazepam normalised hippocampal hyperactivity and CA1 to cortico-limbic-striatal dysconnectivity in individuals at clinical high-risk for psychosis (CHR-P). However, it remains unclear whether 1) the relationship between hippocampal activity and downstream functional connectivity (FC) is different in CHR-P individuals compared to healthy controls (HC), and whether 2) diazepam can normalise this relationship.

Methods: In this randomised, double-blind, placebo-controlled study, 24 CHR-P individuals underwent multi-modal magnetic resonance imaging twice, once following a 5mg dose of diazepam and once following a placebo. They were compared to 22 HC who did not receive diazepam/placebo. Following quality control, 18 CHR-P and 20 HC were included in the analysis. From the hippocampal CA1 subfield, regional cerebral blood flow (CBF) was sampled and FC with key cortico-limbic-striatal regions (ventromedial prefrontal cortex (vmPFC), amygdala, nucleus accumbens (NAc)) was calculated. Multiple regression models investigated the effect of group (CHR-P placebo/diazepam vs. HC) and condition (CHR-P diazepam vs. placebo) on the association between CA1 CBF and downstream FC (CBFxFC). Exploratory models also investigated positive and negative CBFxFC associations within each group/condition. Significance was set at pFWE < 0.05 using peak-level coordinates, and FDR correction was applied to the statistical outputs for the primary multiple regression models.

Results: Relationships between CA1 CBF and cortico-limbic-striatal FC did not significantly differ between CHR-P placebo and HC. Within groups, HC showed a positive CA1 CBFxFC interaction with the vmPFC (T = 3.2, pFWE = 0.049), while CHR-P placebo showed a negative CA1 CBFxFC association with NAc (T = 2.7, pFWE = 0.047). In CHR-P individuals, diazepam significantly enhanced CA1 CBFxFC coupling with the amygdala (T = 4.0, pFWE and FDR = 0.024) and NAc (T = 3.1, pFWE and FDR = 0.030) compared to placebo. As a result, CHR-P diazepam showed positive CA1 CBFxFC interactions with both the amygdala (T = 4.1, pFWE = 0.020) and NAc (T = 2.4, pFWE = 0.090), although the latter did not reach significance, and compared to HC demonstrated stronger CA1 CBFxFC coupling with the NAc, although this did not survive FDR correction (T = 2.7, pFWE = 0.049).

Conclusions: In this study, the relationship between hippocampal activity and cortico-limbic-striatal connectivity was not significantly different in CHR-P individuals compared to HC (although there were differences in direction of within-group associations), potentially highlighting limited power for this case-control comparison. Our finding that acute diazepam modulates the relationship between CA1 activity and downstream connectivity with the ventral striatum and amygdala provides proof-of-concept evidence for a mechanism through which GABA-enhancing compounds could modulate psychosis symptoms. These preliminary findings support further studies in larger samples to investigate whether GABAergic modulation of this circuit is associated with clinical symptom changes in order to clarify the therapeutic potential of enhancing GABA signalling in early psychosis.

Keywords: Clinical high-risk for psychosis, benzodiazepines, Multimodal Neuroimaging, Resting State Functional Connectivity, Cerebral Blood Flow.

Disclosure: Nothing to disclose.

P782. Efficacy, engagement, and safety results from CONVOKE: a pivotal phase 3 RCT evaluating a first-in-class digital therapeutic for experiential negative symptoms of schizophrenia in adults

Abhishek Pratap, Cassandra Snipes, Wenbo Tang, Harold Murck, René S. Kahn, Jean-Pierre Lindenmayer, Faith Dickerson, Jamie Winderbaum Fernandez, Greg W. Mattingly, John Torous, Gregory P. Strauss, Shaheen E. Lakhan, Christoph von der Goltz, Cornelia Dorner-Ciossek

Boehringer Ingelheim Pharmaceuticals Inc., Ridgefield, CT, United States, Ridgefield, Connecticut, United States

Background: There is an urgent need for patient-centered treatments to improve outcomes for people living with schizophrenia, especially for those with negative symptoms (NS), for which there are no FDA-approved pharmacotherapies specifically indicated. Previous research shows that psychosocial interventions can improve NS by targeting defeatist beliefs and enhancing motivation and goal-oriented behavior. However, effective access and uptake of psychosocial interventions are limited in this population.

Digital therapeutics (DTx) offer the potential to deliver digital adaptations of evidence-based psychosocial interventions in an accessible and scalable manner. CT-155/BI 3972080 (CT-155) is an investigational DTx being developed to treat NS of schizophrenia adjunctive to antipsychotics. CT-155 is designed to remediate deficits in reward pathways linked to impaired goal-directed behavior and motivation. CT-155 targets defeatist beliefs through cognitive restructuring and delivers therapeutic interventions aimed at helping patients successfully achieve their goals (social skills training, positive affect training, distress tolerance skills). It also uses a tailored adaptive goal setting approach aligned with a patient’s current functioning level, promoting engagement in real-world activities (behavioral activation). The design and development of CT-155 was patient-centered, informed by feedback from people with schizophrenia. Following encouraging results from exploratory single-arm studies using abbreviated versions of CT-155, a confirmatory Phase III trial, CONVOKE (NCT05838625), was initiated to assess efficacy and safety of CT-155. CONVOKE is the largest randomized controlled study to date of a DTx for NS of schizophrenia.

Methods: A Phase III, multicenter, randomized, double-blind, parallel-group, controlled, 16-week study was conducted between March 2023 and July 2025 across 82 clinical sites in the United States. The study had 2 arms; CT-155 (active treatment) and a digital control (which included disease education and daily check-ins). To maintain blinding, study arms were labeled digital therapeutic A and B. Participants were told they would receive one of two digital interventions, that both may be effective, and that the study aimed to compare their effectiveness. Key inclusion criteria were stable positive symptoms on standard-of-care antipsychotic medication, age ≥18 years, experiential NS of schizophrenia, and smartphone ownership. Participants were randomized 1:1 to CT-155 or digital control.

The study was powered to detect an effect size of Cohen’s d = 0.35 at 90% power with an assumption of 20% early termination, resulting in 432 participants (~216 per arm). The primary endpoint was change from baseline to Week 16 in experiential NS using Clinical Assessment Interview for Negative Symptoms, Motivation and Pleasure subscale (CAINS-MAP) total score. CAINS was administered by a centralized rater team blinded to treatment allocation to avoid bias. CAINS is a second-generation NS scale recommended by experts, designed to comprehensively assess NS deficits with robust psychometric properties sensitive to changes in symptom severity over time. Secondary endpoints included change from baseline to Week 16 in Positive and Negative Syndrome Scale (PANSS) positive subscale score. Engagement and safety were assessed throughout the study. All analyses were performed using the intention-to-treat set.

Results: A total of 457 participants were randomized to CT-155 (n = 227) or digital control (n = 230). Among those, 277 (60.6%) were male, 244 (53.4%) were Black or African American, 181 (39.6%) were White, 127 (27.8%) were Hispanic or Latino, mean (SD) age was 45.8 (12.46) years, and 290 (63.5%) had an annual income of < $25,000. The median (range) number of days participants used CT-155 was 76.0 (0–112) and the digital control was 92.0 (1–112).

The primary endpoint was met, with a statistically significant reduction in CAINS-MAP total score for CT-155 vs digital control (least squares [LS] mean change from baseline to Week 16 [standard error]: −6.8 [0.51] vs −4.2 [0.50]; Cohen’s d = −0.36; p = 0.0003). Sensitivity analyses using a multiple imputation method based on missing-at-random assumption and by the tipping point method demonstrated the robustness of the primary endpoint analysis. Positive symptom severity (assessed using PANSS positive subscale to monitor stability of positive symptoms) was not significantly different between arms (LS mean change from baseline to Week 16 [standard error]: (−0.8 [0.24] vs −0.9 [0.24]; p > 0.5). Adverse events (AEs) were reported in 19 (8.3%) participants in the CT-155 arm, and 31 (13.4%) in the control arm. Serious AEs (SAEs) were reported for 3 (1.3%) participants in the CT-155 arm and 5 (2.2%) in the control arm and were not linked to treatment. No participants in the CT-155 arm and 2 (0.9%) in the control arm had AEs/SAEs that led to study discontinuation.

Conclusions: CONVOKE is the first trial to demonstrate statistically significant improvement in experiential NS of schizophrenia using a DTx adjunctive to antipsychotics. High participant engagement in both study arms shows internal validity of the digital control, which is critical for rigorous evaluation of DTx in confirmatory trials. In adults with NS of schizophrenia, CT-155 was effective, well tolerated, and represents a novel first-in-class therapy option in this area of critical need.

Funding: Boehringer Ingelheim (CT-155-R-001).

Keywords: digital therapeutics, negative symptoms, randomized clinical trial, schizophrenia, patient-centricity.

Disclosure: Boehringer Ingelheim, Employee, Self.

P783. Antipsychotic potential of low-dose methotrexate: longitudinal evidence from a real-world data network and potential neuroimmune targets

Fabiana Corsi Zuelli, Maxime Taquet, Bill Deakin, Rachel Upthegrove

Department of Psychiatry, University of Oxford, Ribeirão Preto, Brazil

Background: Low-dose methotrexate (LD-MTX) - the cornerstone treatment for rheumatoid arthritis (RhA) - is known to restore the function of regulatory T cells (Tregs), specialised CD4+ T cells that promote immune tolerance, curtain inappropriate inflammation, and infiltrate the healthy human brain from early development to orchestrate glial function. In a previous trial, weekly LD-MTX added to standard antipsychotic treatment significantly reduced psychotic symptoms and improved global functioning. Motivated by these findings, we investigated whether LD-MTX use in RhA is associated with reduced long-term risk of psychosis (primary outcome) and other psychiatric disorders (secondary outcomes) compared with alternative anti-rheumatic therapies.

Methods: Using TriNetX US Collaborative Network, we identified all individuals with a diagnosis of RhA (ICD-10-CM: M05–M06) who initiated pharmacological treatment with LD-MTX (2–20mg) or comparator drugs commonly prescribed NSAIDs (naproxen, diclofenac, celecoxib), biologic DMARDs (anti-TNF: infliximab, adalimumab, etanercept; anti-IL-6: tocilizumab), and the conventional synthetic DMARD hydroxychloroquine. Analyses were restricted to adults aged ≤45 years at treatment initiation. Outcomes assessed at five-year follow-up were psychotic disorders, bipolar disorder (BD), major depressive disorder (MDD), and anxiety disorders. Cohorts were matched 1:1 using greedy nearest-neighbour propensity score matching across 35 baseline covariates, including sociodemographics, comorbidities (cardiovascular and metabolic disease), prior or concurrent corticosteroid use, and family history of psychiatric disorders. Outcome-specific exclusions were applied to minimise misclassification (e.g., excluding prior MDD/BD for psychosis analyses). Survival analyses employed the Kaplan–Meier estimator, and results were expressed as restricted mean time lost ratios (rRMTL). The rRMTL estimates the average time lost to an outcome during follow-up in the treatment group relative to comparator. Sensitivity analyses examined negative control outcomes (dog bite, ingrowing nail, viral warts) and compared all-cause mortality between groups. Statistical significance was set at two-sided p < 0.05. Where results were significant, we performed exploratory analyses stratified by sex (male, female).

Results: Following matching, cohort sizes ranged from 958 to 18,033 individuals, with mean ages between 33.3 ± 9.0 and 36.1 ± 7.5 years. Psychosis: Contrary to expectation, LD-MTX did not reduce risk relative to NSAIDs overall. A significant protective effect was observed only in the LD-MTX vs naproxen comparison (rRMTL = 0.56, 95% CI: 0.41–0.77, p < 0.001), with consistent findings in both sexes. Analyses were limited by low event numbers. BD: LD-MTX reduced risk compared with all NSAIDs: naproxen (rRMTL = 0.52, 95% CI: 0.41–0.66), diclofenac (rRMTL = 0.58, 95% CI: 0.47–0.71), and celecoxib (rRMTL = 0.72, 95% CI: 0.54–0.96), all p < 0.05. Protective effects were replicated in sex- and age-stratified models. Among DMARDs, LD-MTX reduced BD risk relative to hydroxychloroquine (rRMTL = 0.79, 95% CI: 0.66–0.96, p = 0.017). MDD: LD-MTX lowered risk compared with NSAIDs naproxen (rRMTL = 0.78, 95% CI: 0.72–0.85), diclofenac (rRMTL = 0.78, 95% CI: 0.71–0.85), celecoxib (rRMTL = 0.84, 95% CI: 0.76–0.94), all p < 0.001. A protective effect was also seen relative to infliximab (rRMTL = 0.57, 95% CI: 0.44–0.74, p < 0.001), replicated in both sexes. Anxiety disorders: LD-MTX conferred protection compared with NSAIDs naproxen (rRMTL = 0.76, 95% CI: 0.70–0.82), diclofenac (rRMTL = 0.76, 95% CI: 0.71–0.82), celecoxib (rRMTL = 0.76, 95% CI: 0.69–0.84), all p < 0.001, consistent across sexes and age strata. Among DMARDs, LD-MTX reduced risk relative to infliximab (rRMTL = 0.69, 95% CI: 0.56–0.86, p < 0.001; effect confined to females) and adalimumab (rRMTL = 0.85, 95% CI: 0.77–0.93, p < 0.001; replicated in both sexes). Sensitivity analyses: No associations with negative control outcomes suggested limited residual confounding (rRMTL rage between 0.67, 95% CI = 0.43 – 1.05 to 1.12, 95% CI = 0.75 – 1.68, p < 0.05). All-cause mortality was lower in the LD-MTX group (range rRMTL = 0.32–0.79).

Conclusions: In this large real-world cohort, LD-MTX was not consistently protective against incident psychosis at five years, although analyses were constrained by event scarcity. Nonetheless, robust and replicable protective associations emerged for more prevalent outcomes - MDD, anxiety disorders, and BD -when LD-MTX was compared with commonly prescribed NSAIDs, and in some cases with targeted biologic DMARDs. These findings suggest that broad, non-specific anti-inflammatory strategies may be insufficient to mitigate psychiatric risk, whereas Treg-modulating therapies such as LD-MTX may exert transdiagnostic benefits. The results underscore the need for targeted intervention trials in high-risk populations and in established psychiatric illness.

Keywords: Low-dose methotrexate, Regulatory T cells, Psychosis, Immunopsychiatry, Anti-inflammaory drugs.

Disclosure: Nothing to disclose.

P784. Translational evidence for dopaminergic rewiring of the basal ganglia in persons with schizophrenia

John Williams, Philip Tubiolo, Roberto Gil, Clifford Cassidy, Natalka Haubold, Yash Patel, Sameera Abeykoon, Zu Jie Zheng, Dathy Pham, Najate Ojeil, Kelly Bobchin, Eilon Silver-Frankel, Greg Perlman, Jodi Weinstein, Christoph Kellendonk, Guillermo Horga, Mark Slifstein, Anissa Abi-Dargham, Jared Van Snellenberg

Renaissance School of Medicine at Stony Brook University, Stony Brook, New York, United States

Background: In prior work, a transgenic mouse model of the striatal dopamine dysfunction observed in persons with schizophrenia (PSZ), the D2-receptor-overexpressing mouse (D2-OE), exhibited dopamine-related neuroplasticity in the basal ganglia. These mice show a highly plastic increase in the density of axonal collaterals from direct pathway D1R-expressing medium spiny neurons of the associative striatum to the globus pallidus externus (GPe; indirect pathway). However, this phenotype of increased “bridging collaterals” has not yet been demonstrated in PSZ.

Consequently, we aimed to test the translational hypothesis that unmedicated PSZ show analogous alterations in connectivity between dorsal caudate (DCa) and GPe using functional magnetic resonance imaging (fMRI) acquired both in the resting state (rs-fMRI) and in a working memory (WM) task state (ts-fMRI), the latter collected during a self-ordered WM task. As striatal neurons are hyperpolarized at rest, we predicted that connectivity alterations would be most pronounced during task engagement, while the neurons are actively engaged in information processing. Further, we examined whether these alterations are associated with deficits in both WM task performance and dopamine function, assessed by neuromelanin-sensitive MRI (NM-MRI) and positron emission tomography (PET) using [11C]-(+)-PHNO (PHNO) PET.

Methods: We conducted a case-control multimodal study of unmedicated PSZ (antipsychotic-naïve or -free for at least 3 weeks) and matched healthy controls (HC; matched on age, parental socioeconomic status, race, ethnicity). Resting-state fMRI data were acquired in 37 unmedicated PSZ and 30 demographically matched HCs; a subset of 29 PSZ and 29 HC also completed ts-fMRI. Four runs each of rs-fMRI and ts-fMRI data (approximately 30 minutes for each paradigm) were collected with a repetition time of 850 ms, 2 mm isotropic voxels, and multiband factor of 6. Additionally, 22 PSZ and 20 HC completed NM-MRI, and 7 PSZ and 4 HC completed PHNO PET with amphetamine challenge. Primary outcome measures included: 1) group differences in ts-fMRI and rs-fMRI functional connectivity between DCa and GPe, 2) NM-MRI contrast ratio in substantia nigra-ventral tegmental area (SN/VTA) voxels associated with psychotic symptom severity, and 3) baseline and amphetamine-induced change in PHNO binding potential (ΔBPND) in DCa.

Preprocessing was conducted with the Human Connectome Project Minimal Preprocessing Pipeline version 4.2.0. WM ts-fMRI data were post-processed by residualizing to task-evoked responses, and only “task on” volumes (defined as beginning 2 s after task block onset to 4 s post-block) were analyzed. Resting-state functional connectivity (rs-FC) and task-state functional connectivity (ts-FC) between DCa and GPe were calculated as partial correlations within a basal ganglia-thalamo-cortical network, controlling for signal in non-pair nodes (dorsolateral prefrontal cortex, globus pallidus internus, substantia nigra, mediodorsal thalamic nucleus). Dunn–Šidák correction was performed for the two primary comparisons (rs-FC and ts-FC). We performed robust multivariable regressions to relate DCa-GPe ts-FC to WM capacity, NM-MRI contrast in psychosis-associated subtantia nigra (SN) and ventral tegmental area (VTA) voxels, baseline PHNO ΔBPND, and amphetamine-induced PHNO ΔBPND in DCa; models included terms for age, biological sex, and diagnosis.

Results: During WM task performance, PSZ exhibited higher DCa-GPe ts-FC than HC (PSZ mean[SD] = 0.11[0.10], HC = 0.05[0.09]; P = 0.0252). DCa-GPe rs-FC did not differ between groups (PSZ median[IQR] = 0.08[0.12], HC = 0.05[0.16]; P = 0.233). Across all participants, greater DCa-GPe ts-FC was associated with lower WM capacity (β*[SE] = −0.31[0.13]; P = 0.020) and greater NM-MRI contrast in psychosis-associated SN/VTA voxels (β*[SE] = 0.40[0.17]; P = 0.023). In the PET subset, greater DCa-GPe ts-FC was associated with DCa baseline D2R availability, measured as PHNO BPND (β*[SE] = −0.45[0.17]; P = 0.039), as well as increased estimated DCa dopamine release capacity (more negative ΔBPND; β*[SE] = −0.82[0.27]; P = 0.021).

Conclusions: Unmedicated PSZ demonstrate a task-dependent DCa-GPe hyperconnectivity that covaries with WM impairment and multiple markers of striatal dopamine dysfunction (NM-MRI and PHNO PET). These converging results provide in vivo translational evidence for a dopamine-associated abnormality in associative striato‑pallidal circuitry in schizophrenia, consistent with compensatory “bridging collateral” plasticity observed in D2-OE mouse. Further, this phenotype suggests a potential neurodevelopmental mechanism of working memory deficits in schizophrenia, representing a critical step towards developing treatments for cognitive deficits.

Keywords: Schizophrenia and related disorders (SRD), fMRI Functional Connectivity, Neuromelanin-sensitive MRI, PET Imaging, working memory fMRI.

Disclosure: Nothing to disclose.

P785. Heightened sleep disturbance vulnerability to cumulative overnight stress hormone in schizophrenia

Alia Warner, Yizhou Ma, Xiaoming Du, Bhim Adhikari, A. Ankeeta, Antonio Pagan, Keiko Kunitoki, Oluwabunmi Akindona, Alina Siatka, Francisco Pallares Solano, Morgana Cowley, Peter Kochunov, Elliot Hong

The University of Texas Health Science Center At Houston, Houston, Texas, United States

Background: Stress dysregulation is implicated in the pathophysiology of schizophrenia spectrum disorders (SSD). Cortisol represents a biomarker of stress and HPA axis function and can influence sleep architecture, another prominent area of dysfunction among schizophrenia patients. However, measuring the effect of daytime stress on sleep is indirect, at best. Here, we developed a new method by assessing cumulative cortisol levels from overnight urine collection to estimate the stress level/regulation from daytime and during sleep, while using in-home wearable sleep EEG recording during the same night, to test whether cumulative cortisol levels overnight impact sleep differently in SSD compared to healthy controls. Examination of relationships between cortisol and objective sleep metrics may elucidate the complex role neurophysiological stress response plays in sleep abnormalities, psychopathology, and physical wellness.

Methods: Seventy-nine adults frequency-matched on age (p > 0.05) but not sex (p < 0.05) (SSD n = 48, M Age = 36.37, 60.4% male; healthy controls (HC) n = 31, M Age = 41.32, 32.3% male) participated in the study. Cumulative cortisol level was measured by overnight urine, and sleep metrics were recorded during the second night wearing the Sleep ProfilerTM EEG sleep monitor at home. Participants also completed questionnaires measuring symptoms of depression (Patient Health Questionnaire-9), anxiety (Generalized Anxiety Disorder-7), insomnia, and daytime sleepiness, and provided medication history. Between-group comparisons, partial correlational analyses (controlling for age and/or sex as appropriate), and multiple (multivariable) linear regressions were conducted. Linear regression analyses were conducted to examine whether cumulative cortisol, diagnosis, and cortisol × diagnosis interaction predicted sleep variables, covaried for age and sex.

Results: The SSD group slept longer, endorsed higher severity of depression and anxiety, and reported more frequent sleep-related prescription medication use compared to HC (all p < 0.05).

Sleep efficiency, as defined by the percentage of time in bed spent asleep, was explained by a significant cumulative stress x diagnosis interaction (p = .035): higher cortisol was negatively associated with sleep efficiency (r = −.32, p = .015) in SSD and positively correlated for HC (r = .33, p = .077). The correlation coefficients were significantly different between groups (z = −2.81, p = .005).

Across sleep stages, hours of stage 2 sleep showed significant main effects for diagnosis. In regression models limited to SSD or HC, cortisol had significant main effects on stage 1% (p = .005, β = .396), REM% (p = .006, β = .382) and stage 3% (p = .037, β = −.290) in participants with SSD. Partial correlation coefficients were significantly different between groups for stage 1% (z = −1.74, p = .041); for SSD, higher cortisol was associated with more time asleep spent in stage 1. Antipsychotic dose (daily olanzapine equivalents) for the SSD group was unrelated to sleep metrics or cortisol.

Conclusions: Cumulative cortisol, measured via overnight urine, has a differential impact on sleep between patients and controls. In HC, higher cumulative cortisol appears linked to more consolidated sleep, perhaps reflecting a more normative sleep response to a more strenuous or stressful day time. In contrast, for SSD, higher cumulative cortisol was associated with poorer sleep efficiency, despite the overall similar overnight urine cortisol levels, possibly reflecting increased vulnerability to sleep disturbances in SSD in response to cumulative stress.

Keywords: Schizophrenia (SCZ), Sleep, Cortisol response to stress.

Disclosure: Nothing to disclose.

P786. A Prospective 12-week randomized controlled trial of remotely- delivered customized adherence enhancement for poorly adherent individuals with schizophrenia vs enhanced treatment as usual

Martha Sajatovic, Jennifer Levin, Salayna Abdallah, Emma Church, Gracie Howard-Griggs, Jessica Surdam, Jacqueline Krehel-Montgomery, Sakshi Priya, Nicole Fiorelli, Farren Briggs

Case Western Reserve University, Cleveland, Ohio, United States

Background: Behavioral interventions that address common barriers to medication-taking have the potential to modify long-term medication attitudes, behaviors, and outcomes among patients with schizophrenia. This prospective 12-week pilot randomized controlled trial tested an adherence promotion approach called customized adherence enhancement for individuals with schizophrenia (CAE-S) vs. enhanced treatment as usual (eTAU) in 36 poorly adherent individuals.

Methods: Patients with schizophrenia were randomly assigned (1:1 ratio) to receive either CAE-S or eTAU after baseline assessment and were reassessed at 12-weeks follow-up. CAE-S consists of 6 remotely-delivered sessions addressing psychoeducation, communication with providers; medication routines, and modified motivational interviewing to address substance use. eTAU was a time-and-attentional control comprising 6 remotely-delivered sessions of wellness content. The primary feasibility outcomes were attendance (session attendance out of possible 6) and patient satisfaction using Likert scales (Aim 1). The primary outcome was change in schizophrenia symptoms from baseline to 12 weeks, measured by the Positive and Negative Symptom Scale (PANSS) (Aim 2). An exploratory evaluation (Aim 3) examined mechanistic underpinnings by assessing change in psychotropic medication adherence from screening to 12 weeks, measured by past 7-day Tablets Routine Questionnaire (TRQ). Secondary measures included the Clinical Global Impression (CGI), the Global Assessment of Functioning (GAF) and Strauss-Carpenter Level of Functioning Scale (SCLFS). Pre/post analysis examined change from baseline to 12-week follow-up on study completers using non-parametric Wilcoxon signed-rank test. A two-sided alpha of 0.05 was considered statistically significant. The study was registered on clinicaltrials.gov under NCT06061952 and approved by the local institutional review board (IRB).

Results: Participants had a mean age of 44.9 (Standard deviation/SD 12) years, and included 25 women (69.4 %). The mean age for schizophrenia diagnosis was 25.1 (SD 9.1) years. One-third (N = 12) lived alone, and the mean number of previous schizophrenia hospitalizations was 5.6 (SD 5.5). Overall, 12-week attrition was 19.4% with 3 individuals withdrawing from study and 4 lost to follow up by study endpoint. There were no study-related adverse events.

For Aim 1 outcomes, mean session attendance was 4.42 (SD 2.2). At screening, mean past 7-day TRQ (proportion of missed drug) was 29.7% (SD 23.8) for individuals enrolled in the CAE arm and 41.7% (SD 26.5) in the eTAU arm. By baseline, mean TRQs substantially improved, decreasing to 11.3 (SD 15.8) in CAE-S, and to 19.3 (SD 25.7) eTAU. Of16 individuals in the CAE arm who provided end-of-study data, satisfaction was high: 93.8% (N = 15) reported the intervention was useful, 87.5 % (N = 14) felt benefit exceeded attendance burden, and all stated they would recommend the program to others.

For Aim 2 outcomes, mean PANSS scores improved significantly from baseline to 12 weeks in both CAE-S [60.9 (SD10.5) to 53.8 (SD9.0)] and eTAU [66.0 (SD9.7) to 54.2 (SD11.2)], with no significant difference between arms. Mixed models that adjusted for covariates found similar results, indicating no significant difference across arms.

Aim 3 outcomes did not demonstrate a significant change in TRQ for either CAE-S or eTAU. Mean CGI scores improved significantly from baseline to 12 weeks in both arms, decreasing from 3.6 (SD 0.8) to 3.1 (SD 0.6) in CAE-S and from 3.5 (SD 0.7) to 2.9 (SD 0.6) in eTAU, with no significant difference between groups. Mean GAF scores also improved significantly from baseline to 12 weeks in both arms, increasing from 60.1 (SD 8.1) to 65.3 (SD 4.8) in CAE-S and from 61.1 (SD 5.5) to 64.5 (SD 5.0) in eTAU, with no significant difference between groups. Mean SCLF scores improved from baseline to 12 weeks in both arms, increasing from 22.6 (SD 4.1) to 25.7 (SD 4.0) for CAE-S and from 23.7 (SD 3.6) to 25.5 (SD 3.9) for eTAU, with results favoring CAE (p < .001).

Conclusions: CAE-S is a brief, remotely delivered adherence enhancement promotion approach that is feasible and highly acceptable among poorly adherent patients with schizophrenia. Medication adherence improved rapidly with monitoring and attention in this brief-duration study, which could potentially explain the observed improvement in schizophrenia symptoms and largely similar outcomes across intervention arms. To further examine the utility of CAE-S, future studies will aim to recruit more robust samples while ensuring representation across diverse sociodemographic groups.

This project was supported in part by a research grant from Investigator-Initiated Studies Program of Merck Sharp and Dohme LLC, a subsidiary of Merck and Co., Inc., Rahway, NJ, USA. The opinions expressed in this paper are those of the authors and do not necessarily represent those of Merck Sharp and Dohme LLC, a subsidiary of Merck and Co., Inc., Rahway, NJ, USA.

Keywords: Medication adherence, Schizophrenia (SCZ), Behavioral approach.

Disclosure: Neurelis, Intra-Cellular, Merck, Otsuka, Alkermes, Teva, Contracted Research, Self, Otsuka, Lundbeck, Janssen, Teva, Medscape, Consultant, Self, Springer Press, Johns Hopkins University Press, Oxford Press, UpToDate, Royalties, Self.

P787. Immunometabolic contributions to deficits in motivation and pleasure in patients with schizophrenia

David Goldsmith, Courtney Ning, Evanthia Wommack, Elaine Walker, Michael Treadway, Jennifer Felger, Andrew Miller

Emory University, Atlanta, Georgia, United States

Background: Negative symptoms of schizophrenia, particularly deficits in motivation and pleasure, are significant barriers to functional recovery. Moreover, there are no current treatments available for these symptoms. We have previously demonstrated associations between inflammation and negative symptoms specific to motivation and pleasure. We have also previously shown that inflammation and metabolic dysfunction have been association with anhedonia in depressed patients. Thus, both metabolic dysfunction and inflammation have been implicated in psychiatric disorders, nevertheless their additive and interactive impact on negative symptoms in schizophrenia remains largely unexplored.

Methods: Negative symptoms were measured in 56 medically healthy adults with schizophrenia using the Brief Negative Symptom Scale (BNSS), which assesses the two negative symptom dimensions: deficits in Motivation and Pleasure (MAP) and deficits in Expressivity (EXP), as well as the subdomains of each dimension (avolition, anhedonia, and asociality for MAP). Fasting blood samples were collected to measure C-Reactive Protein (CRP) and markers of metabolic dysfunction (glucose, insulin, resistin, adiponectin, leptin). A metabolic composite score (MCS) was calculated by adding the z-transformed variables. To test the contribution of both metabolic dysfunction and inflammation, an immunometabolic composite score (IMCS) was calculated by adding z-scores of the metabolic variables and CRP. We also tested the interaction of metabolic dysfunction and inflammation by calculating an interaction term (IMINT). Four linear regression models were tested with MAP as the dependent variable and (1) CRP; (2) MCS; (3) IMCS, and (4) IMINT. Age, sex, race, BMI, smoking, depression, and antipsychotic-dose were included as covariates in all models.

Results: There were significant relationships between BNSS MAP and CRP (beta = 0.342, p = 0.011) as well as BNSS MAP and MCS (beta = 0.335, p = 0.012), though the strongest relationship was found between BNSS MAP and IMCS (beta = 0.399, p = 0.003). There was no significant relationship between BNSS MAP and IMINT (beta = −0.119, p = 0.387). No significant findings were found for any models with BNSS EXP. Similar findings were seen for Avolition (CRP: beta = 0.317,p = 0.018; MCS: beta = 0.307, p = 0.021; IMCS: beta = 0.371, p = 0.005) and Anhedonia (CRP: beta = 0.259, p = 0.056; MCS: beta = 0.292, p = 0.029; IMCS: beta = 0.337, p = 0.012). Of the individual items in the MCS, resistin showed the strongest association with BNSS MAP (beta = 0.358, p = 0.013).

Conclusions: These findings support a role for both inflammation and metabolic dysfunction in negative symptoms, with the IMCS combination demonstrating the strongest relationship. There was no association with IMINT, which suggests that the effects of inflammation and metabolic dysfunction on negative symptoms are additive. These associations were specific to MAP, in particular avolition and anhedonia, consistent with previous literature demonstrating an effect of inflammation and metabolic dysfunction on motivational anhedonia. These findings are also consistent with previous literature suggesting a role for both inflammation and metabolic dysfunction in worse negative symptom trajectories in individuals with first episode psychosis. Resistin, which showed the strongest effect on BNSS MAP, is produced in adipose tissue and may not only contribute to insulin resistance, but may also induce production of inflammatory cytokines, potentially serving as a key intermediate link between disturbances in metabolic dysfunction and inflammatory processes that jointly impact the brain and behavior. As such, immunometabolic approaches to MAP negative symptoms may be important for considering potential novel treatments.

Keywords: Immunometabolism, Negative Symptoms, Schizophrenia (SCZ), CRP, Reward, motivation, and anhedonia.

Disclosure: Nothing to disclose.

P788. Functional outcomes in schizophrenia subtypes based on clinical symptoms or endophenotypes

Robert Chen, Yash Joshi, Tim Bigdeli, Tiffany Greenwood, David Braff, Laura Lazzeroni, Neal Swerdlow, Monica Calkins, Robert Freedman, Ruben Gur, Raquel Gur, Keith Nuechterlein, Allen D. Radant, Jeremy M. Silverman, William S. Stone, Catherine Sugar, Ming Tsuang, Bruce Turetsky, Gregory Light, Debby Tsuang

University of Washington, Seattle, Washington, United States

Background: Schizophrenia is diagnosed based on clinical symptoms, yet patients show significant heterogeneity, suggesting distinct subtypes may exist. Endophenotypes—quantifiable, heritable traits like neurocognitive and electrophysiologic markers—offer a more objective alternative, relying on measures less prone to subjectivity. This study explored whether unsupervised machine learning of clinical symptoms or endophenotypes could identify meaningful subgroups in schizophrenia.

Methods: A total of 1,415 outpatient individuals with schizophrenia or schizoaffective disorder were recruited from five US academic hospitals via the Consortium on the Genetics of Schizophrenia 2 (COGS-2). Clinical symptoms were assessed using the Scale for the Assessment of Positive/Negative Symptoms (SAPS/SANS) while endophenotypes were derived from neurocognitive testing (e.g., attention, memory) and electrophysiologic features measured by electroencephalography (EEG). Unsupervised graph clustering, which captures complex variable relationships, identified subgroups based on clinical symptoms (‘clinotypes’) or endophenotypes (‘biotypes’).

Results: We identified three statistically distinct clinotypes and three biotypes which were largely non-overlapping. Clinotypes were linked to global functional impairment whereas biotypes were associated with impairments in specific functional domains (e.g., finances, planning) and clinical outcomes (e.g., hospitalizations, clozapine usage, anticholinergic burden). Partial correlation network analysis revealed SANS subscales ‘anhedonia’ and ‘attention’ as key mediators linking clinical symptoms and endophenotypes to function.

Conclusions: These findings suggest integrating clinical and endophenotypic data could provide complementary information that would enhance outcome prediction and pave the way for precision therapy in schizophrenia.

Keywords: Biomarker, Cognitive impairment associated with schizophrenia, Schizophrenia (SCZ), Machine learning clustering.

Disclosure: Nothing to disclose.

P789. Symptom trajectories in schizophrenia spectrum disorders during long-acting injectable antipsychotic treatment: an individual participant data meta-analysis

Koki Takahashi, Shinichiro Nakajima, Mehrdad Sadri, Hiroyoshi Takeuchi, Raiki Yoshimura, Taiyo Oura, Shingo Iwami, Hiroyuki Uchida, John Kane, Jose Rubio

Keio University School of Medicine, Shinju-ku, Japan

Background: Schizophrenia spectrum disorders (SSD) are chronic conditions in which symptoms can fluctuate over time, and long-term antipsychotic maintenance is often required. Even during maintenance antipsychotic treatment, symptom trajectories are heterogeneous, and a subset may fail to sustain stabilization—potentially reflecting diminishing treatment response (e.g., dopamine supersensitivity). Because medication non-adherence is common in SSD and a major driver and confounder of unfavorable changes, we sought to examine symptom trajectories under maintenance with long-acting injectable (LAI) antipsychotics, which more reliably maintain continuous exposure. We hypothesized that an identifiable subset would show delayed, progressive worsening despite reliably maintained exposure, indicating a pathway toward secondary treatment resistance.

Methods: This study utilized clinical trial data from the Yale University Open Data Access (YODA) project and was approved by the YODA administration (approval No. 2024-0556). We selected clinical trials which targeted SSD (i.e., schizophrenia or schizoaffective disorder) and conducted LAI treatment for relapse prevention for 6 months or longer. Participants who met study-defined symptom stability, had baseline positive and negative syndromes scale (PANSS) score below 75 (corresponding to moderate severity), and treated with LAIs were included. On these cohorts with guaranteed continuous stable antipsychotic treatment we conducted a data-driven stratification of symptom trajectories by applying latent class mixed models (LCMM) to the longitudinal PANSS scores in each participant. We fit 2 to 5 natural cubic spline curves, set the criteria for minimum class size (5%) and model convergence, and selected the model with the least bayesian information criterion (BIC). The population-level (i.e., fixed-effect) trajectory was estimated using a generalized additive mixed model (GAMM) including random intercepts and slopes. Additionally, we searched for baseline clinicodemographic predictors of the symptom trajectories by logistic regression. We checked the effects of the differential length of observation in each study and treatment discontinuation on observed trajectories. We performed sensitivity analysis for the trajectories of the subscales of PANSS.

Results: We included 4,289 patients with SSD (39.3 ± 12.2 years old, 1,730 female patients [40%], 268 patients with schizoaffective disorder [8% of available record]) treated with LAI risperidone or paliperidone from 17 studies (duration of treatment = 332 ± 174 days, final observation at 945 days). We identified four distinct symptom trajectories with high confidence in classification (average posterior class probability = 0.91): (1) marked worsening from the outset and further worsening by 24 points in total (n = 350 [8%]), (2) stable course followed by slight worsening by 3.8 points (n = 2,494 [58%]), (3) milder improvement by 14 points (n = 1,150 [27%]), and (4) profound improvement by 28 points (n = 295 [7%]). The delayed worsening in the first two trajectories both occurred after 14 months on LAI treatment. In GAMM analysis, the PANSS total score trajectory followed a slightly U-shaped course, declining from 56.4 at baseline to a minimum of 52.3 around day 365, then increasing to 57.1 by day 945. The lower bound of the 95% confidence interval of the estimated trajectory exceeded the estimated minimum score from baseline until day 262 and day 500 until day 945. This trend was not affected by censoring of each study and treatment discontinuation. The two trajectories with delayed worsening were associated with higher severity on tardive dyskinesia (TD) scales (OR = 1.10, p = 0.026), higher age (OR = 1.02, p < 0.001), male sex (OR = 0.81, p = 0.002), and lower baseline PANSS total score (OR = 0.96, p < 0.001) than the two other trajectories with symptom improvement. In a sensitivity analysis, the trajectories for the PANSS positive subscale, both in GAMM and LCMM analyses, showed the same results.

Conclusions: While most patients with SSD remain symptomatically stable while on guaranteed antipsychotic treatment, at least over a timescale of months to a year, we demonstrated heterogeneity in such course of illness. Of note, 8% of previously stabilized patients experienced symptom worsening over months despite guaranteed treatment, and in the most common trajectory a stable trajectory was followed by a discrete worsening over time. Tardive dyskinesia – a phenomenon linked to dopamine supersensitivity – was associated with worsening trajectory, suggesting it as a possible mechanism of loss of efficacy of antipsychotic drugs.

Keywords: Schizophrenia spectrum disorders, Long-acting injectable antipsychotics, Clinical heterogeneity.

Disclosure: Nothing to disclose.

P790. Comparison of change from screening to baseline for an abbreviated 10-item vs. the full 30-item panss in an acutely exacerbated adult schizophrenia clinical trial population

David G. Daniel, Alan Kott, Xingmei Wang, Robert Findling, Eric Youngstrom, Joshua Langfus, Joan Busner

Signant Health, Virginia Commonwealth University, Avondale, Pennsylvania, United States

Background: The Positive and Negative Syndrome Scale (PANSS) is the gold standard for assessing symptom severity and treatment response in schizophrenia clinical trials. The length and complexity of the 30-item PANSS are sometimes perceived as burdensome in both research and clinical settings. We have previously reported and replicated promising psychometric qualities for an abbreviated 10-item version of the PANSS extracted post-hoc from the full PANSS in both pediatric (Findling et al, 2023; Youngstrom et al, 2024; Langfus et al, 2024; Busner et al, 2025) and adult clinical trial populations (Daniel et al, 2023; Daniel et al, 2025).

Instability of scores during the clinical trial screening period may be associated with post-baseline changes (Targum et al, 2020; Kott et al, 2025). As a putative placebo response mitigation procedure it is not uncommon for acute schizophrenia clinical trials to exclude patients who improve by 20% or more between screening and baseline. In the current retrospective analysis, we compare change from screening to baseline for an abbreviated 10-item vs. the 30-item PANSS in an acutely exacerbated adult schizophrenia clinical trial population.

Methods: Data were obtained from 13 acute schizophrenia clinical trials, with screening and baseline PANSS scores extracted for all participants. For each subject, percent change from screening to baseline was calculated for both the full PANSS and the abbreviated PANSS-10 extracted post-hoc from the full PANSS.

To compare the two scales, we conducted separate paired t-tests by each protocol, a linear regression analysis of the difference between percent change scores, including clinical trial protocol as a categorical predictor and calculated the pairwise correlation coefficient.

Symptom change was further classified into seven categories of worsening and six categories of improvement (up to 10%, 20%, 30%, 40%, 50%, and above 50% improvement) for both PANSS versions and the agreement of the classification between the two scales was assessed using polychoric correlations for the ordinal categories and Spearman rank correlations.

Finally, we compared the proportion of subjects with at least 20% improvement between the PANSS total and the abbreviated PANSS-10 total using a test on the equality of proportions. All analyses were performed in Stata19.5.

Results: Data from 4,619 subjects across 13 clinical trials were analyzed. The mean percent change from screening at baseline in total score was 1.1% (SD = 9.9) for the PANSS and 0.5% (SD = 14.1) for the PANSS-10. Significant differences in mean change scores were observed in 2/13 studies between the full and the abbreviated PANSS change scores from screening at baseline.

The mean difference in percent change between the full PANSS and PANSS-10 was −0.5% (95% CI: −0.8 to −0.2), with differences observed across protocols. The Pearson's pairwise correlation coefficient was 0.74 (p < 0.001).

Agreement between categorical change classifications for the full PANSS and PANSS-10 was moderate to strong, with a polychoric correlation coefficient of 0.76 and a Spearman’s rho of 0.60 (both p < 0.001). Significant differences were identified in the ability to identify subjects with a 20% symptom improvement between the 2 scales. While the full PANSS identified only 1.7% of subjects, the PANSS-10 identified 5.7% of subjects, the difference in proportions being significant (p < 0.001).

Conclusions: In our analyses the total scores of the abbreviated 10 item PANSS and the full 30 item PANSS performed comparably in assessing symptom change from screening to baseline among adult patients with acute schizophrenia. The observed mean difference in percent change between the two scales was minimal, and the correlations between categorical change classifications were exceptionally strong. These findings are consistent with clinical utility for the PANSS-10 in adult acutely exacerbated clinical trial populations with schizophrenia. Moreover, the brevity of the PANSS-10 makes it likely less fatiguing to both raters and patients compared to the full PANSS.

Given the different contribution of a single point change to the full PANSS vs. the PANSS-10 it is not surprising that the PANSS10 change would be more sensitive. This study has several strengths, including a large and diverse sample size (N = 4,619) drawn from 13 clinical trials, the use of both continuous and categorical analytical approaches, and adjustment for protocol differences to account for trial heterogeneity. There are many limitations to interpretation of these results. The PANSS-10 was derived post-hoc from the full PANSS. It is possible that an interview directed at all 30 PANSS items would better inform scoring of the PANSS-10 items compared to an interview limited to the PANSS-10 items. The analyses were not pre-specified, the categorization of change thresholds was somewhat arbitrary, and the findings may not generalize to non-acute or negative symptom predominant schizophrenia populations. Moreover, the data were blinded and do not provide insight as to the relative ability post-baseline to distinguish drug from placebo.

In summary, the PANSS-10 appears to offer practical advantages, such as reduced rater burden and faster administration, while maintaining measurement fidelity. Future research should aim to determine if these findings extend to the PANSS-10 administered as a standalone scale and to real-world clinical settings and more diverse patient populations.

Keywords: Schizophrenia (SCZ), clinical outcome assessments, clinical trials methodology.

Disclosure: Signant Health, Employee, Self.

P791. Multiscale auditory network inefficiencies underlying input failure in auditory hallucinations in schizophrenia

Andrew Murphy, Michael Avissar, Yadi Chen, Antigona Martinez, Gaurav Patel, Daniel Javitt

Columbia University and New York State Psychiatric Institute, New York, New York, United States

Background: Most resting-state fMRI studies of auditory verbal hallucinations (AVH) in schizophrenia have focused on connectivity between cognitive systems, providing insight primarily at the large-scale network level. Here, we extend this approach by applying multiscale tools to assess auditory system organization at the network, mesoscale (intermediate), and regional levels. In addition, we use non-invasive brain stimulation (TMS) to evaluate whether perturbation of the language system leads to causal changes in auditory–language coupling and auditory network dynamics. Clarifying how auditory network inefficiencies contribute to AVH, and whether they can be modulated by TMS, may help refine mechanistic models and inform future therapeutic strategies in schizophrenia.

Methods: We studied 56 stable outpatient participants with schizophrenia on or off medications who underwent structural and resting-state fMRI and clinical assessment with the PANSS.

Network level: Functional connectivity (FC) between auditory and language systems was computed using the Glasser atlas parcels with Cole-Anticevic functional labeling.

Mesoscale: We derived auditory subnetwork architecture was derived via a modularity-maximization approach, and integration–segregation ratios were calculated as the mean within-module FC divided by mean between-module FC.

Regional level: Within-auditory signal propagation efficiency was quantified using mean-first-passage time (MFPT), the average time for a signal to travel between randomly selected auditory regions, with longer MFPT indicating less efficient communication.

Statistical analysis: Associations with AVH severity (PANSS P3) were tested using regression models, with significance determined via a non-parametric permutation-based null model. Specificity was assessed relative to parallel measures in the language system.

TMS cohort: In a second cohort of 4 patients, we examined fMRI network measures pre- and post-TMS to the left perisylvian language area, with resting-state fMRI acquired at both timepoints.

Results: Network-level: Lower auditory–language FC predicted greater hallucination severity (β = −1.26, p = 0.04).

Mesoscale: Weaker auditory network integration was associated with both reduced auditory–language FC (r = −0.20, p = 0.04) and greater hallucination severity (β = 1.1, p = 0.04). No effects were observed in the language network.

Region level: Longer auditory system MFPT was linked to reduced auditory–language FC (r = −0.34, p = 0.01) and greater hallucination severity (β = 0.24, p = 0.01). Language MFPT showed no associations.

TMS cohort: Following stimulation, hallucination score significantly decreased (t = 2.7, p = 0.03), auditory–language functional coupling was strengthened, mesoscale analyses showed increased auditory network integration and reduced segregation, and region level MFPT decreased relative to pre-TMS imaging (all exploratory, not powered for significance).

Conclusions: These findings highlight the utility of investigating functional brain organization across multiple scales—network, mesoscale, and regional—in relation to AVH. The addition of TMS provides evidence for causal influences within distributed functional systems, complementing correlation-based rsfMRI analyses. The present data are in-line with the input failure interpretation of AVH (wherein impoverished auditory input to the language network may heighten sensitivity to internally generated reverberations): hallucination severity was consistently predicted by weaker auditory–language coupling, weaker auditory network integration, and inefficient within-auditory signal propagation. Future studies with larger samples and multimodal approaches may help refine mechanistic models and guide development of targeted interventions for persistent AVH in schizophrenia.

Keywords: Schizophrenia (SCZ), Auditory hallucinations, network neuroscience.

Disclosure: Nothing to disclose.

P792. TMS motor cortical inhibition: preliminary analysis mapping the cognitive correlates in people with schizophrenia

Stephanie Hare

Maryland Psychiatric Research Center, University of Maryland School of Medicine, Catonsville, Maryland, United States

Background: Preclinical, clinical and post-mortem research findings suggest that the balance of brain excitation/inhibition (E/I) is altered in people with schizophrenia. Paired-pulse transcranial magnetic stimulation (TMS) can be used to probe brain E/I balance by obtaining measures, such as the short-interval intracortical inhibition (SICI). In those with schizophrenia, SICI scores are consistently elevated, reflecting reduced inhibition. Yet, few investigations have gone beyond reporting this potentially meaningful patient-control difference and explored the relationship between SICI scores and cognitive and/or clinical outcomes. In this preliminary analysis, our goal is to explore the relationship between SICI scores and cognitive performance in patients with schizophrenia. We hypothesize that the SICI score – a physiological measure of motor inhibition – will be correlated with performance on the Stop-Signal motor inhibition task. We also test the alternative hypothesis that SICI scores may be associated with additional cognitive performance measures from the Matrics Cognitive Consensus Battery (MCCB).

Methods: This preliminary analysis draws on data from 15 patients recruited as a part of the ongoing NIMH 1K01MH133116-01A1 study (target recruitment N = 70 over 5 years). To be eligible for participation in this study, patients have a primary diagnosis of schizophrenia or schizoaffective disorder confirmed by the DSM-V.

To calculate SICI score and resting motor threshold (RMT), we use a Magstim Bistim2 system. We use neuronavigation (Rogue Brainsight) to locate the participant’s hand motor hotspot from their T1 structural brain image, and then calculate RMT as the minimum stimulator intensity to elicit a peak-to-peak motor evoked potential of 50 microvolts or greater in 5 out of 10 trials (measured from index finger FDI muscle). We deliver 24 SICI trials, where a conditioning pulse at 0.8RMT precedes a 1.2RMT test pulse separated by 2 ms. We also deliver 24 single-pulse test trials at 1.2RMT and calculate SICI as the ratio of the average of paired-pulse SICI trials divided by the average of the single-pulse test trials.

In our study, participants perform a Stop-Signal motor inhibition task with 288 trials (216 Go, 72 Stop) presented in 5 test blocks (32/64/64/64/64), with a variable 1000–1400 ms fixation period; stimulus onset asynchrony starts at 200 ms and steps upwards and downwards by 50 ms upon correct and incorrect responses. We selected the Stop-Signal task, and the Stop Signal Reaction Time (SSRT) as the primary dependent measure in this study, based on the use of this task in a prior investigation of SICI and schizophrenia (Lindberg et al. 2016).

With respect to statistical design, we perform bivariate Pearson correlation analyses to explore the relationship between SICI and cognitive measures (SSRT and 10 tests from the MCCB assessing processing speed, attention, working memory, verbal and visual learning, problem solving and social cognition). Linear regression analyses will also be performed to examine whether SICI predicts cognitive performance with age and sex as covariates.

Results: To date, there are no results to report at this early stage of the K01 study. Dr. Hare plans to perform this interim exploratory analysis on the first 15 patients in October 2025 to present the results at the 2026 ACNP meeting.

Conclusions: There is a critical need to use innovative technologies like single-pulse and paired-pulse TMS to understand how brain E/I balance is altered in people with schizophrenia spectrum disorders. The hope is that we can then use repetitive (or theta-burst) TMS to restore disrupted E/I balance and help treat symptoms. This project takes a critical first step in asking whether SICI is a clinically significant biomarker, by mapping out which symptoms and cognitive functions it relates to. Future directions of this research will be discussed including the launch of a pilot mechanistic trial planned for October 2025 in which we will apply 5 sessions of repetitive TMS with the goal of enhancing inhibition (indexed by SICI) in people with schizophrenia.

Keywords: Schizophrenia (SZ), TMS, SICI, Biomarker, Cognition.

Disclosure: Nothing to disclose.

P793. Predicting psychoticlike experience severity in youth with longitudinal features using long-short term memory neural networks: an adolescent brain and cognitive development study

Jason Smucny, Nicole Karcher, Cameron Carter

University of California - Davis, Sacramento, California, United States

Background: Previous work suggests that a combination of baseline demographic, clinical, and neuroimaging features can predict psychotic-like experience (PLE) severity based on the last 3 assessment waves in children in the Adolescent Brain and Cognitive Development (ABCD) study. Here we tested the hypothesis that incorporating both baseline and two-year follow-up data using the same features would predict persistent distressing PLEs with greater success than the previous two-year study. As it is specialized for analyzing time series (longitudinal) data, we used long-short term memory (LSTM)-based recurrent neural network (RNN) architecture to train the model. Notably, this differs from the vast majority of previous work using machine learning to predict clinical outcomes in that it uses longitudinal (as opposed to cross-sectional) features.

Methods: Consistent with prior work, children aged 9–10 years at baseline from the ABCD study (release 5.1, which includes half of the four-year follow-up data) were classified as having either persistently elevated distressing PLEs (Z score ≥ 1.96 based on the total distress score for the Prodromal Questionnaire-Brief Child Version for two out of the last three assessments) or low-level PLEs (Z < 0.50 at the last three assessments). Consistent with the previous work, predictors included speech and motor delays (developmental milestones; only available at baseline), family history, internalizing symptoms, adverse childhood experiences, fluid and crystallized cognition, average prefrontal cortical thickness, and average within-network functional connectivity of the cingulo-opercular and default-mode networks. A secondary model including age, sex, and race/ethnicity measured at baseline as predictors. Features for models included both baseline and two-year follow-up data (e.g., internalizaing symptoms at baseline and two-year follow up) except where noted. Family ID was also included in the model. Data were standardized and an LSTM model was tuned to minimize focal loss due to class imbalance. The model was trained and tested using leave-one-site-out cross-validation. Feature importances were calculated using a decision tree model using the LSTM predictions as output.

Results: After excluding children with missing longitudinal data, 160 had severely distressing PLEs at four-year follow-up and 3993 did not. The LSTM model showed 96% accuracy (95% CI: 95%-97%) with 98% negative predictive value (NPV), 45% positive predictive value (PPV), and a receiver operating characteristic area under the curve (AUC) of 82%. The secondary model with demographics showed 96% accuracy (95% CI: 95%-97%), 97% NPV, 53% PPV and an AUC of 82%. In order of importance, for the model without demographics the five most salient features were internalizing symptoms (two-year follow-up), internalizing symptoms (baseline), adverse experiences (follow-up), crystallized intelligence (baseline), and adverse experiences (baseline). For the model with demographics, the five most salient features were internalizing symptoms (two-year follow-up), internalizing symptoms (baseline), race/ethnicity, adverse experiences (follow-up), and sex.

Conclusions: These results suggest that LSTM models based on data collected during childhood can effectively predict persistent distressing PLEs in adolescence. These results build upon prior results that used a random forest model to predict persistent distressing PLEs using the past three assessment waves in ABCD with ~70% accuracy and ~75% AUC. The performance enhancement observed here (96% accuracy and 82% AUC) also illustrates the utility of incorporating features from more than one timepoint using an LSTM architecture. The importance of internalizing symptoms and adverse experiences to prediction is also consistent with prior results, suggesting their predictive value is stable and significant over time. Future ABCD analyses will continue to use LSTM approaches to predict persistent distressing PLEs at later timepoints and using expanded feature sets.

Keywords: machine learning, psychosis, ABCD study, Psychosis-Risk.

Disclosure: Nothing to disclose.

P794. Differences in prefrontal glutamate levels according to illness phase in never-medicated patients with psychosis

Luis F. Rivera-Chávez, Daniela González-Sangabriel, Luz González-Manríquez, Tomás Moncada-Habib, Pablo León-Ortiz, Melanie Malacara, Francisco Reyes-Madrigal, Camilo de la Fuente-Sandoval

Instituto Nacional de Neurologia y Neurocirugia, Mexico City, Mexico

Background: Schizophrenia is a chronic psychotic disorder that is one of the leading causes of disability worldwide, with up to one-third of patients exhibiting resistance to treatment with dopamine postsynaptic D2 receptor blockers (namely antipsychotic drugs). These medications reduce symptoms, particularly positive symptoms. On the other hand, schizophrenia has been linked to hypofunction of N-methyl-D-aspartate (NMDA) glutamate receptors, since NMDA antagonists can produce the full range of positive, negative, and cognitive features of schizophrenia in healthy individuals. Although proton magnetic resonance spectroscopy (1H-MRS) studies in unmedicated patients with early psychosis have shown increased glutamatergic metabolite levels and that these initially increased levels decrease with antipsychotic treatment, some studies suggest that glutamate levels tend to decrease with age at a faster rate than in healthy individuals. On the other hand, some studies have reported that patients with treatment-resistant schizophrenia, as well as those who do not respond to a clinical trial with an antipsychotic, exhibit excessive glutamatergic system activity; however, pharmacological trials with metabotropic glutamate receptor 2/3 (mGlu2/3) agonists (pomaglumetad) have failed.

With the aim of examining the temporal nature of glutamatergic alterations, we studied individuals with psychosis on the schizophrenia spectrum, who were treatment-naïve, at different stages of the illness.

Methods: The Ethics and Scientific Committees of the Instituto Nacional de Neurologia y Neurocirugia approved this study. All participants provided written informed consent, and minors provided assent with written informed consent provided by both parents. Two groups of patients were recruited between November 2017 and July 2025: (1) 51 individuals experiencing their first episode of non-affective psychosis (FEP, defined as a duration of untreated psychosis of less than 74 weeks), and (2) 32 individuals with chronic schizophrenia (defined as a duration of untreated psychosis of 74 weeks or more). Neither group had ever received antipsychotic medication. Patients with current substance use (except for caffeine and nicotine) or any comorbid disorders were excluded from this study. A group of 60 healthy controls matched for age and sex was also recruited. Clinical assessments included the Positive and Negative Syndrome Scale for Schizophrenia and the MATRICS Consensus Cognitive Battery.

Glutamate levels were obtained by 1H-MRS at 3T in a 2.5 × 2.5 × 2.5 cm voxel centered on the bilateral medial prefrontal cortex. Water-suppressed spectra were quantified with Osprey using LCModel for fitting and corrected to account for the tissue composition of the voxel.

Type III Analysis of Covariance was used, with glutamate as the dependent variable and age, sex, and education as covariates, and Bonferroni correction for post hoc pairwise comparisons. Statistical significance was set as P < .05 (2-tailed). Exploratory Spearman rank correlations were used to examine associations between glutamate levels, symptom severity, and cognitive variables.

Results: Glutamate levels differed among the three groups (F = 7.5, P = .0008), without any significant effect of age, sex, or education in the model (P = .54, P = .64, and P = .88, respectively). Post hoc pairwise comparisons revealed higher glutamate levels in the FEP group compared to both chronic schizophrenia patients (P = .003; Cohen’s d effect size, 0.69) and the control group (P = .008; Cohen’s d effect size, 0.83). There were no significant differences in glutamate levels between the schizophrenia group and controls (P = 1).

Higher glutamate levels were associated with lower verbal (ρ = −0.29, P = .04) and visual learning scores (ρ = −0.29, P = .04) in the FEP group. Furthermore, negative correlations were found between glutamate levels and attention-vigilance (ρ = −0.22, P = .01) and visual learning scores (ρ = −0.17, P = .04) across all study participants; however, these correlations did not survive correction for multiple comparisons.

Conclusions: To our knowledge, this is the first in vivo study to investigate cortical glutamate levels in never-medicated patients with first-episode psychosis and chronic schizophrenia. The differences in glutamate levels suggest that these alterations occur primarily in the early stages of the disease; this finding has implications for resuming clinical trials with targeted mGlu2/3 agonists, taking into account both the disease phase and the glutamatergic alterations measured by magnetic resonance imaging.

Keywords: Antipsychotic-naïve first-episode psychosis, Schizophrenia (SCZ), Duration of Untreated Psychosis, glutamate.

Disclosure: Nothing to disclose.

P795. Increased TNF superfamily members in neuroinflammatory schizophrenia and bipolar disorder midbrains

Gerardo Mendez Victoriano, Yunting Zhu, Layla Neuhaus, Frank Middleton, Mitsuyuki Matsumoto, Bart J. L. Eggen, Adam Walker, Maree Webster, Iris Sommer, Cynthia S. Weickert

Neuroscience Research Australia, Sydney, Australia

Background: Neuroinflammation is a key neuropathological finding in schizophrenia and bipolar disorder, as increased cytokines are found in the midbrain of these individuals. Yet, the most upregulated inflammatory cytokine and most activated signaling pathway(s) remain unidentified.

Methods: We aimed to identify the most robust transcriptional change in the midbrain of individuals with schizophrenia by bulk RNA-seq and to confirm the cellular source and magnitude of change by single-nuclei RNA-seq, RT-PCR, and immunohistochemistry in 61 healthy controls, 63 individuals with schizophrenia, and 33 individuals with bipolar disorder, all of both sexes, stratified into low and high inflammation. Our study was approved by the Human Research Ethics Committee at the University of NSW (#HREC: HC230253).

Results: By RNA-seq, mRNAs of the TNF superfamily (TNFSF) pathway were among the most changed in high-inflammation schizophrenia (all p ≤ 0.01), with TNFSF receptors (TNFR1, TNFR2 and FAS) being most highly expressed in astrocytes and microglia. By RT-PCR, we confirmed that five TNFSF receptor mRNAs (TNFR1/2, DR4, FAS, and TWEAKR) were increased in high-inflammation schizophrenia/bipolar disorder compared to low-inflammation controls (all p ≤ 0.01). Furthermore, significantly increased mRNAs encoding cell death-related (P53, CASP1, 7, 8, all p ≤ 0.05) and cell survival-related (BCL2 and MCL1, all p ≤ 0.01) proteins acting downstream of TNFRs were found mainly in high-inflammation schizophrenia. All five TNFSF receptor mRNAs positively correlated with effector mRNAs (all p ≤ 0.05) and with the astrocyte-related marker GFAP mRNA (all p ≤ 0.001). Indeed, FAS+ cells had astrocyte-like morphologies, and GFAP mRNA was significantly increased in high-inflammation schizophrenia/bipolar disorder compared to low-inflammation controls (all p ≤ 0.05).

Conclusions: Our results suggest TNFSF transcripts represent the main activated inflammatory pathway in the midbrains of people with schizophrenia, which overlaps with bipolar disorder. This supports that anti-inflammatory interventions targeting TNF/TNFSF receptors may be of therapeutic benefit in psychiatric patients displaying elevated inflammation.

Keywords: Schizophrenia (SZ), Bipolar Disorder, tumor necrosis factor super family, midbrain, mRNA.

Disclosure: Nothing to disclose.

P796. Age-normative cerebellar functional connectivity across the lifespan and its associations with psychosis

Hengyi Cao, Halil Aziz Velioglu, John Cholewa, Majnu John

Feinstein Institute for Medical Research, Queens, New York, United States

Background: Cerebellar dysconnectivity is strongly implicated in psychosis development and predicts psychosis onset. To better understand its mechanisms, it is essential to build lifespan charts of cerebellar connectivity. Such charts would serve as a valuable reference for evaluating age-standardized cerebellar “normality” at the individual level and its clinical relevance.

Methods: We analyzed resting-state fMRI data from 2330 healthy individuals (aged 6–100 years) across three large-scale Human Connectome Project (HCP) cohorts – the HCP Development (601 subjects), the HCP Young adult (1011 subjects), and the HCP Aging (718 subjects). Cerebellar connectivity was computed for ten cerebellar functional systems defined by the CAB-NP atlas. We used the Generalized Additive Model for Location, Scale and Shape (GAMLSS) to estimate connectivity changes for each cerebellar system across the lifespan, separately by sex. The resulting age-normative models were further applied in an independent clinical dataset of early psychosis (HCP-EP, 115 patients and 49 healthy controls, aged 16–36 years). Z scores quantifying connectivity deviations from the age-normative models were calculated for each individual.

Results: All ten cerebellar systems showed significant age-related effects (p < 0.001). In females, progressively declined cerebellar connectivity was observed across the lifespan in the sensory systems, while males showed relatively stable connectivity in these systems after midlife. For key cognitive systems such as the language, frontoparietal, and default-mode systems, relatively stable connectivity was present in females and slightly increased connectivity was present in males after midlife. In patients, significantly lower z scores (more negative deviations) were observed in the primary visual (p = 0.03), secondary visual (p = 0.005), sensorimotor (p < 0.001), default-mode (p = 0.002), cingulo-opercular (p = 0.02), frontoparietal (p = 0.02), attention (p = 0.02), and multimodal integration (p = 0.01) systems. The patients also showed significantly larger proportion of extremely negative z scores (<5th percentile) compared with controls (p = 0.02).

Conclusions: This is the first study to build age-normative models of cerebellar connectivity across the lifespan. Greater negative deviations were observed in psychosis, aligning with the “synaptic over-pruning” and “accelerated aging” hypotheses of schizophrenia.

Keywords: cerebellum, Resting State Functional Connectivity, Normative Modeling, psychosis.

Disclosure: Nothing to disclose.

P797. Mega-analysis of neurometabolites in antipsychotic-responsive versus non-responsive psychosis

Alice Egerton, Bridget King, Matthew Kempton, 1H-MRS Treatment Response in Schizophrenia Collaborators

King's College London, Institute of Psychiatry, London, United Kingdom

Background: Understanding the biological mechanisms which contribute to antipsychotic response or non-response in psychosis may aid in predicting clinical outcomes and refining targets for novel interventions. Several individual 1H-MRS studies have investigated whether regional neurometabolite levels, and particularly elevated glutamate in the medial frontal cortex (MFC), may be associated with a poor antipsychotic response. Here we present a collaborative mega-analytic approach, combining individual-level data to determine the association between 1H-MRS neurometabolites and antipsychotic response in a large sample.

Methods: A systematic literature search was conducted using the Web of Science database to identify ¹H-MRS studies reporting glutamate, Glx (glutamate plus glutamine), glutamine, N-acetyl aspartate, choline, myo-inositol, GABA, or glutathione in people with a diagnosis of psychosis, schizophrenia or related disorder, classified according to antipsychotic response (responders and non-responders) and healthy controls. Authors of eligible studies were invited to provide individual participant-level neurometabolite data. For studies with multiple timepoints, baseline data were used in the analyses. Neurometabolite levels were compared between groups using linear mixed models, with metabolites as dependent variables, group as a fixed effect, and study as a random effect, while controlling for age and sex. Regions of interest included the medial frontal cortex (MFC, comprising the anterior cingulate cortex [ACC] and midcingulate cortex [MCC]), dorsolateral prefrontal cortex (DLPFC), thalamus, and basal ganglia. Subgroup analyses examined the influence of patient classification (treatment-resistant vs. non-responsive but below treatment-resistant criteria vs. responders), study design (cross-sectional vs. prospective), and MFC voxel placement (ACC vs. MCC).

Results: Data were available in a total of 1,189 participants, including 427 treatment responders, 476 treatment non-responders, and 286 controls. In the MFC, levels of Glu (Estimate (E) = 0.340, P = 0.016), Glx (E = 0.606, P = 0.002), myo-inositol (mI; E = 0. 355, P = 0.001), and choline (E = 0. 075, P = 0.0278) were elevated in non-responders compared to responders. MFC Glx remained elevated in non-responders compared to responders when analysis was restricted to studies employing prospective designs (E = 0.824, P = 0.002). Additionally, MFC mI levels were highest in treatment-resistant samples compared to non-responders (below the TRS threshold) (E = 0.654, P = 0.001), responders (E = 0.788, P < .001), and healthy controls (E = 0.985, P < .001).

Conclusions: This mega-analysis provides the most definitive evidence to date that elevations in MFC glutamate metabolites related to and potentially predict antipsychotic non-response in psychosis. Additionally, elevations in MFC myo-inositol and choline in non-responders may indicate that increased glial activation, membrane turnover and neuroinflammation could also contribute to a poor antipsychotic response. Our findings support the continued investigation of glutamate-acting and inflammatory pathway-associated interventions for psychosis and schizophrenia.

Keywords: glutamate, myo-inositol, magnetic resonance spectroscopy (MRS), Antipsychotic response, Schizophrenia (SZ).

Disclosure: Leal Therapeutics, Consultant, Self.

P798. Novel linguistic markers of loneliness in people living with schizophrenia and non-psychiatric comparison participants

Andrea Coppola, Varsha Badal, Ellen Lee

University of California San Diego, San Diego, California, United States

Background: Loneliness – distress arising from a discrepancy between desired and actual relationships – is prevalent in people living with schizophrenia and is associated with significant functional consequences including poor physical health and quality of life. This is often challenging to assess accurately in real time due to symptom severity, time-consuming clinical interviews, impression management/social desirability biases, and difficulty obtaining repeated measures. Linguistic markers may offer the potential for a more ecologically valid metric of loneliness that could address these assessment challenges. Natural language processing (NLP) is an AI approach to identify linguistic features in unstructured speech data. NLP allows for analysis of pronoun use, lexical complexity, parts of speech, and general topic themes. Our prior work has demonstrated the utility of NLP features to predict social and cognitive functioning in healthy aging populations, though this has been underexplored in populations aging with schizophrenia. We hypothesized that first-person plural pronouns and lexical diversity would be associated with better social and cognitive functioning, respectively, in both diagnostic groups. We explored other linguistic features related to both functioning outcomes.

Methods: We recruited individuals (age 40–70) with schizophrenia/schizoaffective disorder (PwS) and without lifetime history of major psychiatric illness (NCs). Participants completed the UCLA loneliness scale and a semi-structured interview conducted by trained study staff. The semi-structured interviews included questions about relationships and sleep patterns. Interviews were recorded and transcribed. We used Linguistic Inquiry and Word Count (LIWC) software to extract standard (e.g., first person pronoun use, third-person plural pronoun use) and lexical diversity features (e.g., measure of textual lexical diversity, type-token ratio, corrected TTR, Herdan’s c, Uber’s U). We conducted independent sample t-tests and chi-square tests to compare the groups and Spearman’s correlations and general linear models to examine relationships within each group.

Results: The sample included 21 PwS and 20 NCs (M age = 53.7 years, SD = 7.7 years). The groups were similar in age, sex, and race. PwS were less likely to be married and reported higher levels of loneliness and social isolation compared to NCs. In terms of linguistic markers, loneliness was associated with greater first-person singular pronoun usage regardless of diagnostic group (r = .32, p = .04). Among PwS, greater loneliness was associated with greater usage of third-person plural pronouns (r = .54, p = .01) and social isolation was associated with greater use of third-person plural pronouns (r = −.56, p = .01). General linear models revealed an interaction effect between third-person plural pronoun use and diagnostic group on loneliness such that greater third-person plural pronoun use was associated with higher levels of loneliness for PwS (B = 7.72, SE = 2.14, p = .001), but not for NCs. Loneliness was also associated with lower lexical diversity (r = −.51, p = .02) and lower usage of leadership/status words (r = −.56, p = .01), prosocial behavior words (r = −.75, p < .001), and family words (r = −.45, p = .048) in PwS.

Conclusions: These preliminary results provide evidence for potential linguistic markers of loneliness among people with schizophrenia.

Keywords: Schizophrenia (SZ), natural language processing (NLP), loneliness.

Disclosure: Nothing to disclose.

P799. False alarms in memory undermine reward pursuit in people with schizophrenia and may be reduced via transcranial magnetic stimulation

Samantha Abram, Andrew Moses Lee, Susanna Fryer, Judith Ford, Daniel Mathalon

University of California, San Francisco, San Francisco, California, United States

Background: People with schizophrenia have motivation deficits that may relate to underlying issues with episodic memory, particularly recall of contextual details that help create rich memories that might motivate pursuing the experiences again. In an active study, we are using a cross-species translational reward task (Web-Surf) to test whether individual differences in reward pursuit relate to performance on an episodic memory task in participants with and without schizophrenia; we consider both the contributions of signal (i.e., accurate recall) and noise (i.e., false alarms) to memory performance given elevated false alarm rates alarms in schizophrenia. We are also testing whether repetitive transcranial magnetic stimulation (rTMS) to the parietal cortex, a key node of the episodic memory network, can rescue poor episodic memory performance in people with schizophrenia, and if rTMS-induced memory changes have downstream effects on reward pursuit.

Methods: Participants with schizophrenia (SCZ; n = 14) and without a history of psychiatric illness (controls; n = 13) completed a word-based episodic memory task and Web-Surf. SCZ completed two additional memory/reward testing sessions after receiving rTMS to the parietal cortex or a control site (vertex), on separate days (counterbalanced order). During the encoding phase of the episodic memory task, participants viewed a series of words while asked whether the word was alive or pleasant. During the retrieval phase they were then asked whether they recognized the word from earlier (yes/no) and if yes, from which context (animacy/pleasant). We then calculated: 1) yes/no recognition accuracy, 2) context memory accuracy, and 3) false alarms. During Web-Surf, participants cycled between four video galleries (e.g., kittens, landscapes, dance, sports) that each displayed a short video clip after a variable delay (3–30 seconds). The study session was fixed at ~25 minutes. We calculated the number of earns from more preferred relative to less preferred video categories as a proxy for successful reward pursuit. We measured anhedonia from clinical rating scales (e.g., Scale for Assessment of Negative Symptoms).

Results: At baseline (before rTMS), SCZ exhibited reduced context memory accuracy (t[25] = 2.62, p = .01, d = 1.01) and higher false alarms (t[25] = −2.83, p = .008, d = −1.09) relative to controls. The groups did not differ in yes/no recognition. This points to a specific deficit in recalling contextual information. rTMS to parietal cortex reduced the false alarm rate in SCZ (Chisq[2] = 10.03, p < .001) to levels akin to controls (p = .52), but did not impact yes/no recognition or context memory (both ps > .05). Follow-up tests indicated that inhibitory parietal stimulation reduced false alarms when compared with both baseline performance (p = .002) and inhibitory stimulation to the control site (vertex; p = .02). SCZ who made more false alarms, at baseline, were more likely to earn videos from less preferred categories (between-group interaction: t[23] = −2.41, p = .02; effect in SCZ only: r[12] = −.58) and reported greater anhedonia (r[12] = .45). Finally, SCZ who showed greater reductions in false alarms following inhibitory parietal stimulation, relative to baseline, also earned more preferred rewards (r[6] = −.69).

Conclusions: Preliminary results suggest fewer false alarms in memory supports more successful reward pursuit and less anhedonia. Successful reward pursuit during Web-Surf depends, in part, on the ability to effectively delineate between more and less preferred reward options. False judgements may make it more difficult for SCZ to effectively reject less preferred offers. Contrary to our expectation, contextual memory did not directly correlate with reward pursuit. Rather, increased noise (versus reduced signal) may be a driver of poor motivation in SCZ. Inhibition of the parietal cortex may normalize the false alarm rate in SCZ to what is seen in controls; it is possible that overactivity in this region in schizophrenia may lead to a false confidence that new stimuli were previously encountered. Future research can help clarify how the underlying function in relevant brain networks (via MRI) inform rTMS-induced changes. Together our initial findings demonstrate a causal role for the parietal cortex in episodic recognition functions that have consequential impacts on the pursuit of enjoyable and meaningful experiences in schizophrenia.

Keywords: episodic memory, Reward-based decision-making, repetitive transcranial magnetic stimulation (rTMS).

Disclosure: Nothing to disclose.

P800. Real-world effectiveness of clozapine augmentation with mood stabilizers: evidence from two nationwide cohorts

Anssi Solismaa, Heidi Taipale, Dan Siskind, Antti Tanskanen, Ellenor Mittendorfer-Rutz, Christoph Correll, Jari Tiihonen

Karolinska Institutet, Stockholm, Sweden

Background: Approximately one third of people with schizophrenia are treatment-resistant. Clozapine is the recommended antipsychotic medication for this group, but only 40% respond adequately. Mood stabilizers have been used to augment treatment in non-responders, but evidence on the long-term outcomes of this approach remains limited. We aimed to study the effectiveness and safety outcomes associated with mood stabilizer augmentations (valproate, lamotrigine, lithium) to clozapine in two nationwide cohorts.

Methods: We identified clozapine-treated patients with schizophrenia or schizoaffective disorder from nationwide registers in Finland (years 1996–2017) and Sweden (years 2006–2023). These registers include data on inpatient and specialized outpatient care, disability pensions and sickness absence. Clozapine and augmentation prescriptions were identified from the Prescribed Drug register and modelled to drug use periods. Patients were followed from their first clozapine prescription until death or the end of data linkage. The primary outcome was psychosis relapse, measured by hospitalization for psychosis, and the secondary outcome was hospitalization for physical conditions. We compared periods of augmentation to clozapine-only treatment by calculating adjusted hazard ratios (aHRs) using a within-individual design with stratified Cox models. Each national cohort was analyzed separately, and the results were combined thereafter using random-effects meta-analysis. Since mood stabilizers are titrated slowly, we analysed the results also by omitting the first 30 days in exposure and non-exposure periods.

Results: The analysis included 23,206 clozapine users: 14,053 in Finland (mean age 39.1 ± 13.3 years, 58.2% male) and 9,153 in Sweden (mean age 45.6 ± 12.6 years, 60.0% male). Concerning mood stabilizers, a significantly decreased risk of relapse was observed for augmentation of clozapine with valproate (aHR 0.86, 95% CI 0.82–0.90). In the 30-day omission analysis, the risk of relapse was decreased for valproate (aHR 0.85, 95% CI 0.81–0.89), and also for lamotrigine (0.92, 0.87–0.97). Conversely, valproate and lamotrigine were also associated with a slightly increased risk of somatic hospitalization (aHR 1.19, 95% CI 1.00–1.42 for valproate, and 1.15, 1.01–1.31 for lamotrigine).

Conclusions: This meta-analysis from two nationwide cohorts indicates that among patients using clozapine, valproate augmentation therapy is associated with a substantially decreased risk of severe relapse but also possible deleterious somatic outcomes.

Keywords: clozapine, Schizophrenia (SCZ), Pharmacologic augmentation.

Disclosure: Janssen, Grant, Self, Janssen, Consultant, Self, Lundbeck, Consultant, Self, Orion Pharma, Consultant, Self, Teva, Consultant, Self, Lundbeck, Honoraria, Self, Janssen, Honoraria, Self, Orion Pharma, Honoraria, Self, Otsuka, Honoraria, Self, Teva, Honoraria, Self.

P801. Age-specific cognitive trajectories for individuals with schizophrenia across the lifespan: a community-based longitudinal study of people with schizophrenia and non-psychiatric comparison participants

Ellen Lee, Lucy Shao, Tsungchin Wu, Xin Tu, Lisa Eyler

University of California San Diego, LA JOLLA, California, United States

Background: Middle-aged and older people with schizophrenia (PwS) have up to 20-fold higher prevalence of dementia compared to the general population, though the evidence is mixed for accelerated cognitive decline among PwS. This study evaluates cognitive functioning among a cohort of 210 community-dwelling PwS and 213 non-psychiatric comparison participants (NCs), with mean follow-up over 3.78 years (range 0–12.3 years). We examined age-specific cognitive trajectories across multiple cognitive domains and evaluated associations with age, sex, race, educational attainment, and diagnostic group over time. We hypothesized that 1) PwS would have worse cognitive functioning over time compared to NCs, and 2) older groups of PwS would have steeper decline of cognitive functioning compared to NCs.

Methods: We recruited people (age 26–70 at baseline) with schizophrenia/schizoaffective disorder (PwS) and without lifetime history of major psychiatric illness (NCs). Participants were categorized by age at their first assessment (26–40, 41–50, 51–60, and 61–70 years) to capture cohort and time effects, and completed the MATRICS Cognitive Consensus Battery (attention/vigilance, verbal learning, working memory, visual) and the Delis-Kaplan Executive Functioning System (DKEFS; executive functioning, speed of processing). We conducted linear mixed effects models for each cognitive outcome, covarying for age group, sex, race, education, diagnostic group, time, and record count (for practice effects). We included the following interactions: age group by time, age group by diagnostic group, and diagnostic group by time. Models were trimmed using the backwards elimination for most reliable inference.

Results: The sample (mean age 49.6 years, 50.1% female) had on average 4.2 study visits. The diagnostic groups were comparable by age and sex. The PwS group had more years of education and more non-white participants. Among PwS, the 60–70 age group had more education years, older age of onset, and lower daily antipsychotic doses compared to the younger groups – though depressive, positive, and negative symptoms were comparable across age groups.

For all cognitive domains, there were significant main effects of diagnostic group and race, such that PwS and non-white individuals had worse cognitive performance at all timepoints.

For executive functioning, there were significant age group by time and diagnostic group by time interactions such that the 41+ age groups had steeper declines compared to the 26–40 group (p’s < .009) and PwS had steeper declines compared to NCs (p = .046).

For speed of processing, there was a significant diagnostic group by time interaction such that NCs had a steeper decline compared to PwS (p = .043).

For attention/vigilance, there were significant main effects of older age (51+), and education such that older individuals and those with less education had worse performance at all timepoints.

For verbal learning, there was a significant age group by diagnostic group interaction such that PwS aged 51–60 had less decline over time compared to NCs aged 51–60 (p = .01).

For working memory, there was a significant age group by time interaction such that the 41–50 age group had declining performance compared to the 26–40 age group (p = .008).

Conclusions: Our findings support steeper cognitive decline among PwS and older adults for executive functioning and among older adults for working memory, but not for other cognitive domains. The study is ongoing, and we will present updated findings on clinical correlates of the cognitive trajectories.

Keywords: Psychosis spectrum symptoms, Cognitive Functioning, Aging; Cognition; Stress; Acetylcholinesterase; Splice Variants.

Disclosure: Bristol Myers Squibb, Advisory Board, Self.

P802. LB-102 for acute schizophrenia in adults: results from the phase 2 NOVA1 clinical trial, with a focus on PANSS marder factor scores

Christoph Correll, Anna Eramo, David P. Walling, Rishi Kakar, Niccolo Bassani, Leslie Callahan, Baker P. Lee, Zack Prensky, Andrew R. Vaino, John Kane

Donald and Barbara Zucker School of Medicine at Hofstra/Northwell, Glen Oaks, New York, United States

Background: LB-102 is a novel benzamide antipsychotic under development for schizophrenia and other neuropsychiatric diseases. An antagonist of D2/D3/5-HT7 receptors with low off-target activity, preclinical assays highlight a receptor-binding, pharmacokinetic, and behavioral modification profile for LB-102 that is similar to amisulpride. A human dopamine receptor occupancy study demonstrated a similar profile for LB-102 50 mg to amisulpride 400 mg. LB-102 was generally safe and well-tolerated in a phase 1 trial in healthy volunteers. The phase 2 NOVA1 clinical trial in acute schizophrenia demonstrated a significant effect across LB-102 doses versus placebo on Positive and Negative Syndrome Scale (PANSS) and Clinical Global Impressions–Severity of illness (CGI-S) scores after 4 weeks of treatment. This analysis explored the impact of LB-102 through the PANSS Marder factor model, which provides a more nuanced framework for understanding the complexities of schizophrenia symptoms and evaluates treatment outcomes with greater precision.

Methods: NOVA1 was a double-blind, placebo-controlled trial (NCT06179108) in adults (18–55y) with acute schizophrenia who required hospitalization or continued hospitalization for a current exacerbation of psychotic symptoms and had PANSS total score of 80–120, PANSS Positive Symptoms subscale score of ≥4 on ≥2 key items, and CGI-S score of ≥4. Participants were randomized 3:3:3:1 to oral once-daily placebo, LB-102 50 mg, 75 mg, or 100 mg (exploratory). The primary efficacy endpoint was change from baseline to week 4 in PANSS total score. Secondary endpoints included changes from baseline to week 4 in PANSS Marder factor scores. Safety was assessed through treatment-emergent adverse events (TEAEs; coded using MedDRA 26.1).

Results: 359 participants were randomized (placebo, n = 108; LB-102 50 mg, n = 107; 75 mg, n = 108; 100 mg, n = 36), with 293 (82%) completing the 4-week treatment phase. LB-102 met the primary endpoint, with least squares mean (LSM) change from baseline to week 4 in PANSS total score of −9.3 (placebo), −14.3 (50 mg, p = 0.0009 vs placebo), −14.0 (75 mg, p = 0.0022 vs placebo), and −16.1 (100 mg, nominal p = 0.0017 vs placebo).

A significant treatment-related effect was observed across all LB-102 doses versus placebo on the PANSS Marder Positive Symptoms factor score, with LSM changes at week 4 of −2.5 (placebo), −4.4 (50 mg, p = 0.0020), −4.4 (75 mg, p = 0.0020), and −5.4 (100 mg, p = 0.0020). A non-significant treatment effect was observed for each LB-102 dose versus placebo on the PANSS Marder Negative Symptoms factor score at week 4 (LSM change from baseline: placebo, −1.6; 50 mg, −2.5, p = 0.0774; 75 mg, −1.9, p = 0.4958; 100 mg, −3.0, p = 0.0669) and PANSS Marder Anxiety/Depression factor score (LSM change: placebo, −2.4; 50 mg, −2.7, p = 0.5437; 75 mg, −3.2, p = 0.0926; 100 mg, −3.5, p = 0.1296). All LB-102 doses were associated with significantly greater improvements from baseline to week 4 on the PANSS Marder Disorganized Thought factor score (LSM change: placebo, −2.0; 50 mg, −3.6, p = 0.0002; 75 mg, −2.9, p = 0.0301; 100 mg, −3.3, p = 0.0447) and PANSS Marder Uncontrolled Hostility/Excitement factor score (LSM change: placebo, −0.7; 50 mg, −1.5, p = 0.0203; 75 mg, −1.5, p = 0.0248; 100 mg, −1.9, p = 0.0290).

TEAEs were reported in 56% (placebo), 69% (LB-102 50 mg), 57% (75 mg), and 75% (100 mg) of participants. Ten participants reported TEAEs that led to withdrawal (placebo, n = 2; 50 mg, n = 2; 75 mg, n = 3; 100 mg, n = 3), and 5 had serious TEAEs (placebo, n = 2 [death, n = 1]; each LB-102 arm, n = 1). Elevated prolactin levels at Day 28 compared to baseline were reported across all treatment arms (placebo, 5.2%, +1.3ng/ml; 50 mg, 92.2%, +59.1ng/ml; 75 mg, 94.2%, +50.3ng/ml; 100 mg, 87.5%, +51.3ng/ml). Clinical adverse events related to prolactin increase were reported in 5 participants, including galactorrhea (50 mg, n = 2; 75 mg, n = 1), breast enlargement (100 mg, n = 1), and erectile dysfunction (100 mg, n = 1). The incidence of extrapyramidal symptom (EPS) incidence was low (placebo, n = 2; 75 mg, n = 1; 100 mg, n = 1), as was dystonia (placebo, n = 1; 75 mg, n = 3; 100 mg, n = 1). SAS, BARS, and AIMS were stable over time, with little to no emergence of EPS, akathisia, tardive dyskinesia, or other involuntary motor side effects.

Conclusions: This phase 2 clinical trial provided robust evidence demonstrating the efficacy and safety of LB-102 for acute schizophrenia. LB-102 demonstrated greater improvements than placebo across all five PANSS Marder factor scores, which was statistically significant for three of the factor scores (Positive Symptoms, Disorganized Thought, and Uncontrolled Hostility/Excitement). There was a lack of statistical significance for Negative Symptoms and Anxiety/Depression which may reflect a non-enriched patient population. These data highlight a potentially broad, multifactorial antipsychotic effect for LB-102, further informing the phase 3 clinical program.

Keywords: Schizophrenia (SCZ), Schizophrenia (SZ), Phase II clinical trial, placebo-controlled trial, clinical efficacy.

Disclosure: AbbVie, Alkermes, Boehringer-Ingelheim, Cerevel, Click Therapeutics, BMS, Lundbeck, Merck, Newron, Novartis, Otsuka, Sumitomo, Teva, Terran, Advisory Board, Self, Lundbeck, Janssen, Otsuka, Sunovion, Grant, Self, HealthRhythms (private/stock options), LB Pharmaceuticals, Inc. (private/stock options), North Shore Therapeutics (private/stock), Vanguard Research Group (private); NW Pharmatech (private/stock options), Saladax (private/stock options), Reviva (stock options); Terran (private/stock options), Stock/Equity - Privately Held Company, Self, MedinCell, Stock/Equity - Publicly Traded Company, Self, UpToDate, Royalties, Self, AbbVie, Alkermes, Allergan, Boehringer-Ingelheim, Bristol Meyer-Squibb, Cerevel, Dainippon Sumitomo, H. Lundbeck, HealthRhythms, HLS Therapeutics, Indivior, Intracellular Therapies, Janssen Pharmaceutical, Johnson and Johnson, Merck, Minerva, Neurocrine, Novartis, Otsuka, Roche, Saladax, Sunovion, Teva, Honoraria, Self, HLS Therapeutics, LB Pharmaceuticals, Mapi, Maplight, Newron, NW PharmaTech, Teva, Consultant, Self.

P803. Plasma antipsychotic levels in remitted patients with schizophrenia treated with long-acting injectable antipsychotics: preliminary results

Robert Zipursky, Ofer Agid, Gary Remington, Olivia Spandier, Nicole Sung, Cristiana Stefan, Sheng Chen, Wei Wang

University of Toronto, Toronto, Canada

Background: Problems with medication adherence are the most common reason why patients with schizophrenia relapse after having achieved remission of their psychotic symptoms. Offering treatment in the form of long-acting injectable antipsychotic medication (LAIs) has the potential to minimize relapses by enhancing adherence. Measuring plasma antipsychotic levels in patients receiving LAIs may optimize relapse prevention if it can be determined that the LAI dose being administered is providing sufficient levels of medication. LAIs may also be more acceptable to patients if minimum effective plasma levels can be defined that would lower the risk of medication side effects. This study was carried out to estimate the 48-week risk of breakthrough psychotic symptoms when patients with schizophrenia are treated with second generation LAIs after experiencing a remission of psychotic symptoms. In this analysis, we investigated whether antipsychotic plasma levels were associated with the risk of relapse, and with measures of symptom severity, functioning and medication side effects.

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 for the past three months. 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 Scale (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. Trough antipsychotic plasma levels, measures of clinical symptoms, functioning and side effects were assessed at baseline, 24 weeks and 48 weeks. Linear mixed models analysis was used to assess severity of clinical symptoms and side effects over time.

Results: 60 participants consented to participate, with 51 (19F, 32M) completing the baseline assessment including 27 receiving paliperidone (Invega Sustenna or Invega Trinza), 22 receiving aripiprazole (Abilify Maintena) and 2 receiving risperidone (Risperdal Consta). Mean age of the sample is 38.3 years (s.d. = 11.8 years). Nine subjects dropped out prior to the baseline visit and 21 subjects dropped out subsequently. Follow-up assessments have been completed for 30 participants at 24 weeks and 24 participants at 48 weeks.

To date, 4/44 participants who have had follow-up visits post-baseline have experienced breakthrough symptoms, resulting in an estimated 48-week probability of relapsing of 9.1%. Mean baseline plasma levels for participants on paliperidone was 33.9 (s.d. = 21.9) ng/ml and for aripiprazole 200.7 (s.d. = 90.2) ng/ml. Of the four subjects who relapsed, two were treated with paliperidone and two with aripiprazole. Participants who relapsed while on paliperidone had been on their current doses for 12 weeks and 16 weeks, and had plasma levels of 19.6 ng/ml and 26.9 ng/ml, respectively; those on aripiprazole had been on their current doses for 26 and 88 weeks, and had plasma levels of 217 and 232 ng/ml, respectively. Participants who relapsed on aripiprazole but not paliperidone met criteria for SCID-5 criteria for current cannabis use disorder and tested positive for cannabis on their urine drug screen completed at baseline. Levels of paliperidone and aripiprazole were not significantly associated with ratings on the BPRS total score, CGI-SCH Severity, or Personal and Social Performance Scale (PSP). Using linear mixed model analysis, plasma paliperidone levels were associated with scores on the Arizona Sexual Experiences Scales (p < 0.05) while aripiprazole levels were associated with scores on the Barnes Akathisia Scale (p < .01).

Conclusions: In this study, we have been assessing the 48-week 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. Relapses on paliperidone palmitate were limited to those who had been treated for less than 6 months, while relapses on aripiprazole occurred within 24 months of starting the LAI and were associated with documented cannabis use during the follow-up period.

These findings suggest that the risk of relapse on second generation LAIs is low once effectiveness is established in the first two years of treatment. Our findings are also consistent with recent work suggesting that cannabis use may increase the risk of relapse even in those who are adherent with LAI treatment. The significant associations reported between plasma levels and adverse effects suggests that monitoring plasma levels in patients treated with LAI paliperidone and aripiprazole may be of value in minimizing side effects and optimizing the acceptability of these medications for long term treatment.

Keywords: Schizophrenia (SCZ), Long-acting injectable antipsychotics, relapse prevention, therapeutic drug monitoring, drug side effects.

Disclosure: Janssen Inc, Honoraria, Self, Boehringer Ingelheim Pharmaceuticals, Inc., Consultant, Self.

P804. Impact of gluten-free diet on psychiatric symptom severity and neuroimaging findings in a randomized double-blind inpatient study of a persons with schizophrenia and having elevated antigliadin antibodies (AGA IgG)

Deanna Kelly, Christopher Lee, Laura Rowland, Andrea Wijtenburg, Peter Kochunov, Bhim Adhikari, Gopal R. Vyas, Robert Buchanan, Daniel Roche, Valerie Harrington, Monica Vladut Talor, Heather Adams, Matthew Glassman, Ann Marie Kearns, Korrapati Sathyasaikumar, Robert Schwarcz, Fang Liu, Sarah M. Clark, William Eaton

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

Background: Schizophrenia and related disorders (SRD) are characterized by positive and negative symptoms, such as anhedonia and avolition as well as deficits in cognition. There are no current FDA approved treatments for negative symptoms, which is a critical gap in our treatment of people with SRDs, since they are a major determinant of functional impairment. An emerging literature suggests that SRDs have a relationship with immune function and inflammation. Recently an SRD subgroup with high inflammation and elevated levels of anti-gliadin antibodies (AGA) of immunoglobulin G type (IgG), T cell dysregulation and high inflammatory markers have been characterized. Negative symptoms and cognitive improvement have been previously observed with gluten removal in this subgroup in two small clinical trials.

Methods: We conducted a 5-week confirmatory double-blind placebo-controlled trial of a gluten free diet (GFD) versus gluten-containing diet (GCD) for negative symptoms in people with SRD who have elevated AGA IgG (NCT03183609). Participants were between the ages of 18–64 years, had a diagnosis of schizophrenia or schizoaffective disorder. Those included had AGA IgG > 20 U, no serologic evidence of celiac disease, and stable antipsychotic treatment and dose. All participants were inpatients, received a GFD and were randomized to 30 grams of gluten or rice flour daily delivered in protein shakes. The Clinical Assessment of Negative Symptoms Motivation and Pleasure scale (CAINS MAP) was used as the primary negative symptom outcome. Selected kynurenine metabolites (HPLC) and cytokines (multiplex Luminex) were quantified in plasma, and in n = 20 participants cerebral blood flow (CBF) using Arterial Spin Labeling, and neurochemistry using magnetic resonance spectroscopy.

Results: We included 39 participants (N = 21 GFD and N = 18 GCD). There was a significant improvement over time in the CAINS MAP (treatment X time F = 2.78, df = 30.1, p = 0.045) in the GFD compared to GCD. There was a trend for BPRS positive symptoms (p = 0.08) improvement in the GFD relative to GCD. In this subgroup with neuroimaging (N = 20) we were not powered to find changes, however we observed increased CBF changes in the insula region approaching significance (F = 4.12, df = 1,16, p = 0.059, Cohen’s d = 0.751). Figure 1 shows Cohen’s d differences between groups favoring increased blood flow in frontal regions. Also, we found a significant time X treatment interaction on the mixed model ANCOVA showing an increase in medial frontal creatine in the GFD relative to the GCD group (F = 10.89, df = 1,15, p = 0.0049) over the 5 weeks.

Additionally, relative to a GCD, a GFD significantly decreased kynurenic acid (KYNA; df = 32, F = 9.51, p = 0.0042), kynurenine (df = 32, F = 11.45, p = 0.0019), and interleukin (IL)-2 (df = 27.9, F = 3.5, p = 0.019). KYNA and CAINS MAP change scores were correlated (r = 0.350, p = 0.039).

Conclusions: This is the first large scale double-blind randomized clinical trial in SRD with elevated AGA IgG. The results are consistent with smaller studies and indicate that negative symptoms, particularly anhedonia and avolition, may be improved with GFD. We also observed significant increases in frontal CBF and creatine, and significant decreases in kynurenine and KYNA in the GFD. Our previous finding of cognitive improvements was not replicated in this study. The study was conducted during COVID-19 quarantines and lockdowns which may have hindered the impact of negative and cognitive symptom improvement in this inpatient study. Given no current treatments for negative symptoms, better identifying, understanding and treating this subgroup is of growing importance.

This project was funded by NIMH R01 R01MH113617 (DL Kelly and WW Eaten MPI).

Keywords: Gluten-Free Diet, Schizophrenia and related disorders (SRD), Antigliadin antibodies (AGA IgG), Human Neuroimaging, Inflamation.

Disclosure: Boehringer Ingelheim, Advisory Board, Self, Karuna, Advisory Board, Self.

P805. Single-subject network covariance perturbations of cerebral blood flow and GABA-A α5 receptor availability in clinical high-risk and first-episode psychosis

Samuel Knight, Julia Schubert, Mario Severino, Paulina Lukow, Amanda Kiemes, Nicholas Livingston, Fernando Zelaya, Andrea De Micheli, Zerrin Atakan, James Davies, Thomas Spencer, Paolo Fusar-Poli, Jacek Donocik, Natasha Vorontsova, Eugenii Rabiner, Anthony Grace, Federico Turkheimer, Steven Williams, Mattia Veronese, Owen O'Daly, Philip McGuire, Gemma Modinos

Institute of Psychiatry, Psychology and Neuroscience, King's College London, London, United Kingdom

Background: Preclinical evidence suggests that hyperactivity of the hippocampus due to neural excitation (glutamatergic)/inhibition (GABAergic) imbalance leads to dysfunction within a corticolimbic network involved in psychosis-relevant phenotypes. We recently identified that hippocampal hyperactivity (by regional cerebral blood flow, rCBF) in individuals at clinical high-risk for psychosis (CHR-P) can be normalized by acutely enhancing GABAergic signalling with diazepam. Using [11C]Ro15-4513 PET, however, we found no significant regional differences in hippocampal GABAA α5 receptor (α5-GABAAR) availability between individuals at CHR-P, people with first-episode psychosis (FEP), and healthy controls (HC). Traditionally, analysis of rCBF and PET data relies on univariate statistics examining group differences in individual regions or voxels of interest independently, limiting the ability to capture subtle and complex network-level interactions. Here, we applied network covariance statistics to investigate corticolimbic rCBF and α5-GABAAR network perturbations in early psychosis for the first time, compared to a reference connectome based on normative modelling.

Methods: We recruited 24 CHR-P (mean ± SD = 25.2 ± 4.6 years, 15 female), 10 FEP (mean ± SD = 29.2 ± 5.2 years, 3 female), and 24 HC participants (mean ± SD = 25.2 ± 4.4, 15 female). All participants underwent a single-session scan using a GE Signa 3T PET/MR scanner at Invicro (Imperial College London). rCBF data was acquired via 3D pseudo-continuous arterial spin labelling (ASL) to obtain CBF maps in standard physiological units (ml blood/100gm tissue/min). α5-GABAAR availability (volume of distribution, VT), was quantified through [11C]Ro15-4513 PET with arterial blood sampling. Both PET and ASL scans were registered to a T1-weighted IR-FSPGR image. PET data were analysed with MIAKAT 4.3.24 in MATLAB R2017a. After QC, the final ASL sample was 23 HC, 24 CHR-P, and 10 FEP, and the final PET sample was 22 HC, 22 CHR-P, and 10 FEP. The corticolimbic network of interest included six anatomically pre-defined bilateral ROIs: amygdala, anterior cingulate, associative striatum, hippocampus, midbrain, and orbitofrontal cortex.

For rCBF and α5-GABAAR separately, individual z-score matrices were constructed to quantify deviations in ROI-ROI covariance relative to the reference group (HC). For each modality, the HC reference matrix was computed using partial Pearson’s correlations across the six selected ROIs adjusting for age and sex (and global CBF for ASL). An individual covariance perturbation matrix was generated for each CHR-P and FEP participant, by iteratively adding them to the HC reference matrix. The difference between this augmented connectivity matrix and the original reference matrix was calculated and standardised into a z-score matrix. This matrix reflects the magnitude and direction of covariance differences for each ROI pair (edge) relative to the normative control network, per subject. To compare covariance perturbations at each ROI edge, we calculated the absolute z-scores of deviations, averaged across participants. Group differences (HC, CHR-P, FEP) in network perturbations were first assessed by comparing the mean magnitude of deviations across all network edges, and from hippocampal edges only, using non-parametric permutation tests (10,000 permutations), in which observed F-statistics were compared to a null distribution generated by shuffling group labels. Where significant effects were found, pairwise group comparisons were conducted using two-tailed permutation tests on mean differences. Next, mean positive- and negative-only deviation z-scores were compared across the whole network and in hippocampal edges only to assess the direction of the perturbations. Significance was set at p < .05, FDR-corrected.

Results: rCBF covariance was overall significantly perturbed in both CHR-P and FEP individuals for hippocampal edges (F = 5.763, p = .006), but not within the wider corticolimbic network (F = 2.331, p = .102). Splitting the analysis by the directionality of the deviations revealed a significant group effect for negative but not for positive deviation z-scores, reflecting lower network covariance, both within the whole network (F = 4.006, p = .032, CHR and FEP) and for hippocampal edges (F = 4.471, p = .015, CHR only).

The magnitude of overall α5-GABAAR covariance perturbations did not significantly differ across the network or in hippocampal edges only (F = 1.996, p = .145; F = 1.937, p = .245, respectively). However, splitting the analysis by the directionality of the perturbations revealed again a significant effect for negative but not positive z-scores, both within the whole network (F = 6.265, p = .026, CHR and FEP) and for hippocampal edges (F = 3.14, p = .047, CHR only).

Conclusions: People at high-risk and with first-episode psychosis show alterations in corticolimbic network organisation in both neural activity and α5-GABAAR availability, reflecting lower covariance between regions and particularly for the hippocampus. A multivariate individualised covariance perturbation approach provides a promising way to investigate network level dysfunction psychosis.

Keywords: first-episode psychosis, Cerebral Blood Flow, GABA receptors, clinical high-risk for psychosis, multi-modal neuroimaging.

Disclosure: Johnson and Johnson, Honoraria, Self.

P806. Efficacy of low-dose adjunctive methylphenidate extended-release on cognition and functioning in schizophrenia: a randomized open-label trial

Naista Zhand, Ridha Joober, Alain Labelle, David Attwood, Fatima Alquraish, Fatima Iftikhar, Carrie Robertson, Elizabeth Kozyra, Esther Carefoot, Phillip Harvey

University of Ottawa, Ottawa, Canada

Background: Cognitive impairment, affecting ~75% of individuals with schizophrenia, severely disrupts functioning and recovery. No approved pharmacological treatments currently target these deficits. Methylphenidate extended-release (ER), a psychostimulant enhancing prefrontal dopamine, shows promise for cognition in attention-deficit/hyperactivity disorder but has limited, inconsistent evidence in schizophrenia. This study investigates low-dose methylphenidate ER’s effects on cognitive and functional outcomes in schizophrenia, addressing a critical therapeutic gap.

Methods: In an 8-week, open-label, randomized crossover trial, 24 stable adults (18–55 years) with DSM-5 disgnosis of schizophrenia spectrum disorder received 4 weeks of methylphenidate ER (18 mg/week 1, 36 mg/weeks 2–4) or treatment-as-usual (TAU), with crossover at week 4, and follow-up at week 12. The primary outcome was improvement in functional capacity, measured by the Virtual Reality Functional Capacity Assessment Tool (VRFCAT) total adjusted time, while secondary outcomes included cognitive performance, assessed by the Brief Assessment of Cognition in Schizophrenia (BACS), and symptom severity, evaluated with the Positive and Negative Syndrome Scale (PANSS-6). Generalized Linear Models assessed time and sequence effects, controlling for baseline scores (power = 0.8, effect size d = 0.5–0.8, p < 0.05), with sex considered as a biological variable.

Results: The study enrolled 24 participants (mean age 35.66 ± 10.56 years, 70.8% male; 79.2% schizophrenia, 20.8% schizoaffective disorder; mean illness duration 14.20 ± 9.30 years; 45.8% on clozapine), with no significant baseline differences in gender, age, or PANSS-6 scores across arms (p > 0.05). VRFCAT scores improved significantly over time (F(2,20) = 10.00, p < 0.001, ηp² = 0.32); in the first period (baseline to week 4), the medication-first arm showed improvement (163.29s, t(10) = 2.08, p = 0.05, d = 0.60) vs. TAU-first (134.08s, t(11) = 1.66, p = 0.11, d = 0.48), with overall gains from baseline to week 8 of 303.47s (28.8%, p = 0.002) and 159.91s (17.1%, p = 0.078), respectively, sustained post-medication. BACS showed significant improvements in the TAU-first arm during the medication phase (weeks 4–8) for Symbol Coding (4.75 points, t(11) = 2.59, p = 0.017, d = 0.75) and Tower of London (t(11) = 3.03, p = 0.006, d = 0.87), with no gender- or diagnosis-specific effects (p > 0.05). PANSS-6 improved significantly while on study medication (p < 0.001, d = 1.49 medication-first arm in the first period; d = 1.30 TAU-first arm in the second period, weeks 4–8), notably in delusions and social withdrawal, without psychosis exacerbation. Covariate analysis (age, sex, illness duration, clozapine use) confirmed findings. At 2-month follow-up, 75% (18/24) resumed methylphenidate ER (10 on 36 mg, 8 on 18 mg).

Conclusions: While results should be interpreted cautiously due to the open-label design and small sample size, this trial suggests low-dose methylphenidate ER may enhance functional capacity (VRFCAT), specific cognitive domains (processing speed, executive function), and symptoms (PANSS-6) in schizophrenia without exacerbating psychosis. The amplified learning effect on VRFCAT indicates potential neuroplasticity-based benefits, though larger, double-blind, placebo-controlled trials are needed to validate efficacy and optimize dosing for broader applicability.

Keywords: Cognitive impairment associated with schizophrenia, psychostimulants, schizophrenia and cognition, Schizophrenia; Functional Capacity; Antipsychotic Treatment.

Disclosure: Nothing to disclose.

P807. Alteration of 3D-genome architecture and gene expression by ultra rare structural variants associated with schizophrenia

Paola Giusti-Rodriguez, Luz Porras, Martilias Farrell, Isabelle Rodriguez- Lausell, Gabriel Iglesias-Maldonado, Emily Val F. Tuliao, Chelsey Leveque, NaEshia Ancalade, Jin P. Szatkiewicz

University of Florida, College of Medicine, Gainesville, Florida, United States

Background: Alterations in topologically associated domains (TAD) boundaries may lead to disarrangement of enhancer-promoter contacts and dysregulation of gene expression and have been associated with multiple disorders. We previously found that ultra-rare structural variants identified via whole genome sequencing of schizophrenia patients were enriched at TAD boundaries. In this study, we sought to determine of impact of these TAD boundary disrupting SVs on the three-dimensional organization of chromatin (i.e. 3D-genome) and gene expression.

Methods: Using CRISPR-Cas9 we engineered an ultrarare ~500 kb deletion on chromosome 13 (DEL-chr13), initially identified in the schizophrenia cohort, in isogenic human induced pluripotent stem cells (iPSCs), and differentiated these into neural precursor cells (NPCs), excitatory glutamatergic neurons (EGNs), and medium spiny neurons (MSNs). We performed high-resolution, in situ Hi-C and RNA sequencing to assess changes in 3D-genome organization and gene expression across these three cell types. High-resolution Hi-C datasets were processed with Juicer, topologically associated domains (TAD)/insulation analysis was done using FAN-C, loop calling via HICCUPS, and differential loop detection with HICCUPSDiff. Clustering analysis showed that transcriptomic and chromatin profiles grouped primarily by cell type, and replicates (DEL vs. control) showed high similarity, confirming robust differentiation and reproducibility.

Results: Locally, we found that the TAD boundary overlapping the deletion region persisted across all cell types, as confirmed by insulation scores using Arrowhead, and direct visualization in Juicebox. High-resolution loop calling revealed consistent formation of a novel loop at the deletion site in all deletion lines, and was associated with altered expression of nearby genes, suggesting regulatory rewiring.

Globally, we observed widespread alterations in chromatin compartments, TADs, frequently interacting regions (FIREs), and loops, with 3.9–5.2% of loops affected depending on the cell type. These structural changes correlated with differential gene expression and were enriched in schizophrenia-relevant pathways, including synaptic signaling, neurogenesis, and cytoskeletal organization, particularly in MSNs.

Among the dysregulated genes, those consistently connected to loop anchors altered by the deletion, showed significant expression changes in corresponding RNA-seq data. Key candidates included SLITRK4, schizophrenia-associated protocadherins (PCDHA13, PCDHA4, PCDHA5), and neurodevelopmental regulators (NKX6-2, ARX, ABI3BP). These genes emerged from integrated loop-gene and RNA-seq analyses, particularly in MSNs.

Conclusions: Our findings demonstrate that ultrarare structural variants can reshape 3D genome architecture and transcriptional networks in a cell-type-specific manner, with medium spiny neurons (MSNs) showing the greatest impact. This highlights chromatin remodeling as a key mechanism by which structural variants contribute to schizophrenia pathogenesis.

Keywords: functional genomics, Hi-C chromatin conformation, structural variants.

Disclosure: Nothing to disclose.

P808. The role of C4A gene expression in synaptic formation

Agnieszka Kalinowski, Marcus Ho, Reenal Pattni, Alexander Urban

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

Background: Schizophrenia is an impairing chronic mental illness for which we have no disease-modifying treatments and a lack of understanding of which pathophysiological mechanism could be interrupted for therapeutic benefit. Multiple forms of evidence point to a glutamatergic synaptic deficit as a primary pathophysiologic mechanism in SZ. In order to try to reverse the glutamatergic synaptic deficit, we need to understand the mechanisms that contribute to it. One influential hypothesis explaining the synaptic deficit is the excessive synaptic pruning hypothesis, which posits that SZ arises due to excessive elimination of synapses by microglia during adolescence: more copies of the A-type complement 4 (C4A) gene in individuals with SZ results in more C4 protein that tag synapses for elimination by microglia. Indeed, C4A gene expression (GE) is increased in SZ dorsolateral prefrontal cortex postmortem samples and a mouse model demonstrated that a high number of C4A gene copies results in low synaptic density. However, more recent studies demonstrate that high C4A GE may contribute to the synaptic deficit in ways other than through synaptic pruning, such as by interfering with synaptic receptor trafficking and changing transcriptomic patterns.

An unexplored area of investigation is whether C4A GE may impact synaptic formation. Thus, we hypothesize that environmental perturbations in early development or adolescence - expressed through the elevation of systemic IFN-Ɣ and/or IL-6 - would impair synaptic formation through synaptic gene downregulation due to an amplified expression of the C4A gene because individuals with SZ have a high number of C4A gene copies. IFN-Ɣ and IL-6, cytokines that increase in the context of stress and infection, are both known to increase C4A GE and are elevated in serum and cerebrospinal fluid in individuals with SZ.

Methods: IPSCs were cultured and differentiated into induced neurons according to established protocols. Briefly, iPSC cells were cultured on matrigel-coated plates, then virally transfected using lentiviruses containing the rtTA, GFP and Ngn2 plasmids. C4A expression was induced using chemical induction (IFNg) or transfected with a C4A-overexpressing plasmid (Origene, Rockville, MD). Cells were collected for RNA analysis on day 7. C4A gene copy number was determined using established protocols with digital-droplet PCR. iN cells were collected in TRIzol reagent (Invitrogen). RNA was isolated using Direct-zol (Zymogen) and sent to Novogene for bulk RNA-sequencing (Sacramento, CA). Differential gene expression and primary component analyses (PCA) were performed using the nf-core/rnaseq pipeline (https://nf-co.re/rnaseq/3.12.0). Lists of differentially expressed genes were further analyzed using Syngo (https://www.syngoportal.org/). Synaptic density was determined by co-localization of vGlut1 and SAP97 in immunofluorescent samples.

Results: Differential gene expression (DEG) analyses of RNA-seq of the untreated iNs compared to those treated with IFN-γ or transfected with C4A overexpressing plasmid overlap with GWAS-loci in Syngo (annotated database of synaptic genes). More specifically, C4A protein overexpression downregulates a fraction of the DEGs compared to IFN-γ, suggesting that the C4A protein (or mRNA) may be exerting an effect on synaptic gene downregulation.

Conclusions: The importance of C4A overexpression in the pathophysiology of SZ is illustrated by convergent findings in several different data types. The implication is that C4A expression causes excessive pruning. However, blocking pruning by blocking the action of microglia by using minocycline has not been effective as a clinical treatment and alternative consequences of increased C4A gene expression are being identified. This work highlights an additional mechanism by which increased C4A gene expression in SZ could be leading to pathological effects.

Keywords: Schizophrenia (SZ), Complement 4, Induced neurons.

Disclosure: Nothing to disclose.

P809. SnRNA-seq investigation of schizophrenia reveals neurodevelopmental and synaptic dysfunction in upper and deep layer excitatory neurons of the dlPFC

Lauren Schramm, Anton Schulmann, Yash Patel, Ningping Feng, Qing Xu, Ajeet Mandal, Dowon Kim, Bhaskar Kolachana, Clifton Dalgard, Michael Kelly, Pavan, K. Auluck, Stefano Marenco

National Institute of Mental Health, Bethesda, Maryland, United States

Background: Schizophrenia (SCZ) is a psychiatric disorder characterized by hallucinations, delusions, disorganized thought, social withdrawal and dulled affect. It causes long-term disability and early death due to high rates of suicide. Treatments have targeted symptomatology, as the underlying mechanisms of disease are poorly understood. Prior studies have implicated cell-type specific gene expression changes in excitatory neurons of the prefrontal cortex in the pathogenesis of SCZ, and functional enrichment analyses have implicated neurodevelopmental and synaptic processes.

Methods: To further elucidate the cell type-specific genomic mechanisms underlying SCZ, we performed single nucleus RNA sequencing (snRNA-seq) using 10X Genomics Chromium Single Cell 3’ Platform on the dorsolateral prefrontal cortex (dlPFC) of 23 cases with SCZ (7F, mean age = 55.1+12.3) and 21 controls (7F, mean age = 50.8+14.5). We initially captured 637,000 nuclei, and following quality control measures for doublets, mitochondrial percentage, transcript count and gene count per cell, our final dataset consisted of 393,000 nuclei across 16 excitatory, 19 inhibitory, and 11 non-neuronal cell classes. We conducted differential gene expression analysis with Dreamlet, controlling for sex, race, post-mortem interval, RNA integrity number, mitochondrial RNA percentage, ribosomal RNA percentage, and cellular detection rate. Functional enrichment analysis was conducted using gprofiler2, and ontology terms were semantically clustered using rvvgo.

Results: Our analysis revealed transcriptional dysregulation in ~3,000 genes (padj < 0.05; |logFC| > 0.231) across all 46 cell types, primarily in superficial (n = 1,020, 3 cell types) and deep layer (n = 273, 1 cell type) excitatory neurons. Inhibitory neuron subclasses showed minimal differential expression, apart from one class of parvalbumin expressing interneurons (n = 285). Consistent with recent findings (Ruzicka et al., 2023, Ling et al., 2024), gene ontology analysis identified downregulation of genes involved in synapse and axon assembly, neurogenesis, and CNS development in superficial and deep layer excitatory neurons.

Conclusions: These results suggest involvement of these cell types in dysregulation of cell differentiation and maturation in SCZ. Our ongoing analysis is focused on testing the hypothesis that excitatory neurons undergo divergent or stunted developmental trajectories in SCZ. We conclude that developmental pathways appear dysregulated in superficial and deep layer excitatory neurons in individuals with SCZ.

Keywords: Schizophrenia (SCZ), single nucleus RNA-seq, Excitatory Synapses, Brain development.

Disclosure: Nothing to disclose.

P810. CeLl Type-specific Effects Of Psychosis On The Medial Thalamus

Anton Schulmann, Ningping Feng, Pavan K. Auluck, Nirmala Akula, Gabi Dugan, Anthony D. Ramnauth, Madeleine Peng, Maura Boldrini Dupont, Stefano Marenco, Francis J. McMahon

Columbia University and NYSPI, New York, New Hampshire, United States

Background: Converging evidence implicates the thalamus, particularly the mediodorsal (MD) nucleus and its projection to the prefrontal cortex, in the pathophysiology of psychosis. Neuroimaging studies demonstrate reduced thalamic volume and altered thalamocortical connectivity as early features of psychosis, with decreased connectivity to prefrontal cortex and increased connectivity to somatomotor cortices. These changes are observed in schizophrenia (SCZ), SCZ prodromal syndromes, and bipolar disorder (BD) with psychotic features. However, the cellular and molecular underpinnings of these thalamic abnormalities remain poorly understood. We previously generated the first cell type atlas of the adult human MD and adjacent thalamic nuclei. To gain insight into the molecular underpinnings of psychosis, we now examined transcriptional alterations in the medial thalamus across SCZ and BD with psychotic features.

Methods: We analyzed postmortem medial thalamus tissue from 12 individuals with SCZ, 12 with BD, and 12 neurotypical controls. Following nuclei isolation with myelin removal, we performed single-nucleus multiomic profiling using the 10x Genomics Chromium platform. After quality control and outlier filtering, the final dataset comprised ~70,000 nuclei per diagnostic group. Data were clustered and annotated in Seurat, with excitatory subpopulations further resolved by integration with spatial transcriptomics and published thalamic datasets. Differential gene expression was assessed using dreamlet with batch as a random effect, and functional enrichment was performed using Zenith and GSEA.

Results: We identified 13 excitatory neuron clusters that mapped onto the magno- and parvocellular divisions of MD, paraventricular (PVT), intralaminar, laterodorsal (LD), ventrobasal (VB) thalamus as well as the lateral habenula, including an atypical mixed GABAergic population. A large population of SOX14+ inhibitory thalamic interneurons and major glial populations were also recovered. Transcriptional alterations were cell type- and diagnosis-specific. In SCZ, neuronal effects were strongest in the magnocellular division of MD (MDmc) and adjacent primary-type nuclei (LD, VB), whereas BD showed stronger effects in the parvocellular division of MD (MDpc) and PVT. Glial perturbations in SCZ were more pronounced in oligodendrocytes and microglia compared to BD. Functional enrichment analyses revealed upregulation of mitochondrial energy metabolism pathways in the MDpc, particularly in BD, along with broader alterations across both SCZ and BD in genes related to synaptic function, ion channels, axon guidance, cilia, and immune response.

Conclusions: Our findings demonstrate that psychosis spectrum disorders exert distinct cell type- and subnuclei-specific effects in the medial thalamus. SCZ and BD with psychosis show a high degree of functional convergence, but with distinct cell type-specific vulnerabilities, with BD showing greater involvement of limbic-projecting nuclei, such as the PVT, and SCZ showing more prominent effects in primary-type nuclei. Limitations include modest sample size and the potential influence of treatment and secondary disease effects. These results provide initial insights into the cellular and transcriptional landscape underlying thalamic abnormalities in psychosis. Future work will analyze accompanying ATAC-seq data to uncover regulatory mechanisms and integrate transcriptional changes with common genetic risk for SCZ and BD.

Keywords: Thalamus, Postmortem Human Brain Tissue, Single-cell transcriptomics, Schizophrenia (SCZ), bipolar disorder (BD).

Disclosure: Nothing to disclose.

P811. M1 receptor signaling in SST interneurons of the prefrontal cortex: implications for treatment of schizophrenia

Zixiu Xiang, Jonathan W. Dickerson, Margaret Wilson, Kendall Jensen, Insung Kim, Ashley Kim, Shalini Dogra, Arisa Timoll, P. Jeffrey Conn, Jerri Rook

Vanderbilt University, Nashville, Tennessee, United States

Background: Muscarinic acetylcholine receptor subtype 1 (M1) activators have robust efficacy in enhancing cognition and reverse deficits in multiple domains of cognitive function in NMDA receptor hypofunction rodent models involving cortical disruption similar to that observed in schizophrenia patients. However, the precise cellular and circuit mechanisms by which M1 activation improve different domains of cognitive function remain unclear. Early stages of schizophrenia are associated with hyperactivity of the prefrontal cortex (PFC) and this is thought to be important for some of the cognitive and negative symptoms associated with the disease. Interestingly, recent studies suggest that the profound increase in disordered activity in the PFC in schizophrenia patients and animal models may be mediated by two mechanisms. First, multiple studies suggest that schizophrenia is associated with a loss of long-term depression (LTD) at excitatory inputs to the PFC from the hippocampus. Loss of normal LTD could contribute to an increase in excitatory drive to the PFC that has been observed at hippocampal inputs to the PFC. Second, studies suggest that loss of GABAergic inhibitory transmission in the PFC may play a critical role in the pathophysiological changes in the PFC and underlying cognitive deficits in schizophrenia patients and animal models. Interestingly, we have found that M1 activation may reduce excessive excitation of the PFC by combined actions on both excitatory and inhibitory transmission. First, M1 activation induces long-term depression (M1- LTD) of transmission at excitatory projections to the PFC from the hippocampus. In addition, activation of M1 induces a robust increase in spontaneous inhibitory postsynaptic currents (sIPSCs) in PFC pyramidal cells. Based on previous and new preliminary studies, we postulate that M1-LTD and M1-induced increases in synaptic inhibition in the PFC are mechanistically distinct and depend on activation of M1 in different neuronal populations. Developing an understanding of the precise cellular mechanisms of M1 regulation of PFC function and associated behaviors could be of critical importance in understanding the actions and optimal profiles of M1 activators for inducing specific behavioral responses. We have generated novel mouse lines that allow us to selectively delete M1 from specific neuronal populations. Using these novel tools, along with optogenetic and chemogenetic approaches, will allow us to rigorously evaluate the roles of M1 expressed in each of these major neuronal populations in regulating inhibitory and excitatory transmission in the PFC and in specific behavioral responses to M1 activators.

Methods: Both male and female mice (n = 3–16) ≥ 7 weeks of age were used in the studies. A subset of mice underwent stereotaxic surgery for AAV5 viral injection. Electrophysiology field excitatory postsynaptic potential (fEPSP) and whole cell recordings were conducted in coronal brain slices containing the mFPC. For optically evoked fEPSPs, a blue light was delivered on brain slices expressing channelrhodopsins, ChR2. Optogenetic silencing of neurons expressing halorhodopsin, eNpHR3.0, was achieved by applying combination of yellow (565 nm) and blue (470 nm) LED light.

Med Associates chambers were used for locomotor activity monitoring over 90 minutes. Novel object recognition was conducted using a 10 minute exposure to two identical ‘familiar’ objects, followed by a 1 hour intertrial interval and another 10 minute exposure period with one familiar object being replaced with a novel object. Progressive ratio testing was performed in Med Associate operant chambers using mice food restricted to 85% of free fed body weight. Mice were trained to nose poke to obtain food reward on a FR1, FR3 and FR6 schedule. Animals were then tested in a progressive ratio paradigm for three consecutive days.

Statistical analysis was performed using Prism (GraphPad Software), and were presented as mean ± SEM. Two-tailed Wilcoxon matched-pairs signed rank test, two-tailed Mann-Whitney test, or one-way ANOVA with Bonferroni or Tukey post-test was used as appropriate.

Results: Electrophysiology in mouse mPFC slices revealed that muscarinic agonist oxotremorine-M (oxo-M) induced long-term increase in frequency of spontaneous inhibitory postsynaptic currents (sIPSCs) and reduced E/I ratios in PNs—effects blocked by M1 antagonist VU0255035, suggesting that these effects were mediated by M1 receptors. Using optogenetic silencing approach, we identified SST-INs as primary mediators of M1-driven inhibition in PFC layer 5 PNs. Additionally, oxo-M induced sustained-increase in excitability of SST-INs, but not PV-INs. Furthermore, conditional M1 knockdown (KD) in SST-INs diminished oxo-M-induced sIPSC increases and impaired novel object recognition (NOR) without altering locomotion. Female mice with M1 KD in SST-INs exhibited progressive ratio task deficits, revealing sex-specific motivational impairments.

Conclusions: Here we demonstrate that M1 activation on somatostatin interneurons (SST-INs) induces long-term enhancement of GABAergic inhibition in PFC layer 5 pyramidal neurons (PNs), which might rectify E/I imbalance in pathophysiologic conditions. These results establish SST-INs as critical cellular substrates for M1-mediated cortical inhibition, suggesting that dysfunction of M1 in SST-INs might be one of the mechanisms underlying cognitive deficits and negative symptoms associated with schizophrenia.

Keywords: Cognitive impairment associated with schizophrenia, muscarinic M1, schizophrenia negative symptoms, Medial Prefrontal Cortex (mPFC).

Disclosure: Nothing to disclose.

P812. Peripheral dopamine receptor signaling in pancreas: novel mechanisms of antipsychotic drug actions on metabolism

Caroline Ward, Feng Li, Jenesis Kozel, Ashley Nilson, Eva Martin-Solana, R. Benjamin Free, David Sibley, Michael Jurczak, Zachary Freyberg

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

Background: Dopamine (DA) signaling is increasingly recognized as a metabolic modulator. Most studies focused on DA D2-like receptors in metabolically-relevant brain regions like striatum and hypothalamus. Yet, discovery of D2-like receptors outside the brain has expanded the scope of DA’s roles as a metabolic modulator. In pancreas, D2-like receptors are expressed in insulin-secreting beta-cells. To date, it has been difficult to disentangle metabolic contributions of brain vs peripheral DA signaling. Thus, we created new genetic tools to selectively target peripheral D2-like receptors including beta-cell-specific D2R conditional knockout (cKO) mice. Moreover, we explore D2R expression and signaling in additional pancreatic islet cell types including alpha- and delta-cells.

Methods: Pancreatic beta-cell D2R cKO: Tamoxifen ip for 5 days induced D2R cKO in Ins1-Cre/ERT;Drd2-flox and Gcg-Cre/ERT;Drd2-flox mice.

Glucose tolerance: Plasma glucose and insulin were measured after a 6h fast followed by ip glucose bolus.

Cyclic AMP (cAMP) assays: Intracellular cAMP was measured via the LANCE ultra cAMP kit.

Results: Using newly-developed beta cell-specific D2R cKO mice, we discovered sex differences in beta-cell D2R-mediated metabolic regulation. Female beta-cell D2R cKO mice became dysglycemic while males did not, particularly in response to high-fat diet challenge. These results mirror females' predisposition for antipsychotic drug (APD)-induced dysglycemia, providing a new mechanism for how antipsychotic drugs cause diabetes. We also identified expression of D2R in glucagon-secreting alpha-cells and somatostatin-secreting delta cells. Unlike beta-cells, alpha-cell D2R cKO did not modify glycemic control in females. Finally, APD treatment of a delta-cell line modified cyclic AMP (cAMP) signaling.

Conclusions: We find sex-specific mechanisms of pancreatic beta-cell D2R-mediated regulation of glycemic control, explaining why females are more vulnerable to APDs' metabolic effects. In contrast, D2R cKO in alpha-cells did not similarly cause dysglycemia in females, indicating that beta-cell D2R may drive APD-induced dysglycemia. Our work in delta-cells also indicates that APDs concurrently act on multiple islet cell types to modify glycemic control.

Keywords: Metabolic Psychiatry, insulin, Antipsychotic drugs, D2 dopamine receptor, APD-induced metabolic disease.

Disclosure: UPMC Enterprises, Contracted Research, Self.

P813. Neural and behavioral evidence for therapeutic utility of targeting KCC2 in schizophrenia

Arghya Mukherjee, Navdeep Bajwa, Jonathan Scott, Noell Cho, Huiwen Zhu, Toshiya Nishi, Zhong Zhong, Stephen Moss, Michael Halassa

Children's Hospital of Philadelphia and University of Pennsylvania, Philadelphia, Pennsylvania, United States

Background: Prefrontal cortex (PFC) abnormalities and cognitive impairments are widely documented in schizophrenia. Existing evidence links cognitive deficits to PFC excitatory-inhibitory (E-I) imbalance, particularly GABAergic dysfunction. However, a conserved molecular explanation of PFC E-I imbalance remains elusive, limiting therapeutic advances. Potassium/chloride co-transporter (KCC2) are crucial regulators of chloride homeostasis and postsynaptic inhibitory efficacy at GABAergic synapses. Here, across multiple models of schizophrenia, we probe whether PFC KCC2 function is impaired, leading to E-I imbalance as well as cognitive deficits and are rescued with a small molecule activator of KCC2.

Methods: Using immunohistochemistry, in vivo slice physiology and photometric measurements of chloride gradients we quantified PFC KCC2 function across 22Q11DS and chronic methamphetamine mouse models of schizophrenia. Subsequently, using optogenetics and multi-electrode recordings we measured changes in E-I balance and network activity in mutants. These mutants when tested on a two-alternative forced choice task (2AFC), revealed cognitive deficits that could be causally linked to KCC2 function via selective prefrontal knock down. Finally, we use the same analytical pipeline to precisely quantify the physiological as well as cognitive outcomes of treatment with a novel agonist of KCC2 transporter.

Results: All mutants show a reduction in PFC KCC2 expression in excitatory neurons and a consequent depolarization of GABA reversal potential. We further identified in vivo consequences of KCC2 dysfunction as delayed chloride dynamics and diminished fidelity of network activity patterns. We further found that mutants exhibit deficits in working memory, and task switching that were mirrored by the KCC2KD cohort, confirming KCC2 dysfunction as causal to observed impairments. Finally a direct agonist of KCC2 transporter activity rescued physiological and cognitive deficits across both the models. n = 3 to 6 animals genotype, p < 0.05 across all experiments.

Conclusions: Our findings link KCC2 dysfunction to physiological and cognitive deficits across multiple schizophrenia models, identifying KCC2 as a promising pharmacological target for cognitive enhancement.

Keywords: Cognitive impairment associated with schizophrenia, excitatory / inhibitory balance, Prefrontal cortex, KCC2, Pharmacology.

Disclosure: Nothing to disclose.

P814. Characterization of hippocampal E/I balance in novel models of schizophrenia and pharmacological intervention

Vivien Zell, Coralie Brifault, Lawrence Fourgeaud, Bartosz Balana, Pascal Bonaventure

Johnson and Johnson Innovative Medicine, San Diego, California, United States

Background: Schizophrenia is a debilitating neuropsychiatric condition with high prevalence that we still don’t fully understand at the cellular level. Nevertheless, the hippocampal excitation/inhibition balance (E/I) is now a well-accepted hypothesis leading to downstream effects on positive, cognitive and negative symptom domains of the disease. With the emergence of novel mouse models of schizophrenia in the past few years, an effort has been made to characterize these at a cellular level to refine our understanding of the role of E/I in disease well as its preclinical intervention.

The genetic Grin2A mouse model (Farsi et al., 2023) has been shown to display phenotypes relevant to the study of schizophrenia (increased locomotor activity, elevated gamma band power, cognitive deficits…), but the characterization of hippocampal E/I remains elusive. The pharmacological subchronic LPS mouse model (Jung et al., 2023) has been shown to display a selective loss of the vesicular GABA transporter in the hippocampus together with an E/I imbalance believed to lead to the cognitive deficits in these mice. This present study offers a new characterization of the hippocampal E/I in both models as well as an attempt to rebalance the shifted E/I in the Grin2A model with a positive allosteric modulator (PAM) of the GluN2D subunit of the NMDA receptor, a target previously shown to be specifically expressed in inhibitory interneurons (Yi et al., 2020). We believe that this body of work will offer a comprehensive characterization of hippocampal E/I in these models as well as showing the feasibility of a pharmacological intervention aimed at rebalancing altered E/I towards a healthy output.

Methods: To characterize E/I in each model, we recorded spontaneous excitatory and inhibitory postsynaptic currents (sI/EPSCs) from hippocampal pyramidal neurons using ex vivo slice patch clamp electrophysiology.

We used voltage clamp recordings of pyramidal neurons in Grin2A+/+ (wild type, WT, N = 4 mice, n = 20 neurons), Grin2A+/− (heterozygotes, HET, N = 4, n = 17) and Grin2A-/- (knockouts, KO, N = 4, n = 25) at −65 mV of holding to record sEPSCs and at 0 mV to monitor sIPSCs. We also measured the impact of a bath application of PTC-174 (GluN2C/D PAM, 1 µM) on the hippocampal sIPSC frequency of Grin2A KO.

In the LPS model, mice were injected on 2 consecutive days with LPS or PBS and hippocampal pyramidal neurons were recorded at 0 mV (N = 5, n = 21 and N = 4, n = 13, respectively). Additionally, we also measured the electrically triggered IPSCs (eIPSCs) from the same hippocampal pyramidal population in both LPS- and PBS-treated animals.

All experimental protocols in animal studies were approved by the 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,

Results: In the Grin2A genetic mouse model, we measured a significantly higher sEPSCs frequency in the KO compared to HET and WT animal (F2,59 = 1.25, p = 0.035) while no difference has been observed in sIPSCs (F2,60 = 0.94, p = 0.714). Preliminary recordings in Grin2A KO mice showed that PTC-174 (1 µM) largely increased sIPSC frequency (n = 2).

In the subchronic LPS model, the frequency of sIPSCs was significantly lower in LPS-treated animals compared to control PBS (p = 0.0096), while we measured no significant difference in eIPSCs amplitude between treatments.

Conclusions: In this study we used ex vivo electrophysiology to characterize the hippocampal E/I in novel models of schizophrenia. In the Grin2A KO mice, the higher sEPSC frequency is representative of a disease state E/I imbalance (hyper excitatory output). In the LPS model, the decrease in sIPSCs frequency is reflecting the decrease in hippocampal VGAT expression previously reported (Jung et al., 2023). Both models thus displayed an E/I imbalance at the hippocampal level, in line with the general hypothesis. The bath application of PTC-174 increase sIPSC frequency displaying the feasibility of rebalancing E/I in the hippocampus with selected pharmacological agents. Follow up studies should focus on the impact of rebalancing E/I in translational assays (Electroencephalogram, mismatch negativity…) to link the cellular mechanism to the behavioral output characteristic of schizophrenic diseases. Taken together this study shows the importance of hippocampal E/I in these mouse models and shows the feasibility of rebalancing the network towards healthier output by targeting the hippocampal inhibitory tone.

Keywords: Schizophrenia (SCZ), Mouse model, E/I blanace, Hippocampus.

Disclosure: Johnson and Johnson Innovative Medicine, Employee, Self.

P815. NREM sleep signatures of bipolar disorder

Naihua Gong, Nataliia Kozhemiako, Shaun Purcell, Jen Pan

Brigham and Women's Hospital, Boston, Massachusetts, United States

Background: Sleep disruptions are ubiquitous across psychiatric diseases, and abnormalities in specific aspects of sleep architecture can provide insight into neurocircuit dysfunction. Prior work demonstrated specific disruptions in NREM sleep in schizophrenia that could reflect circuit-level disruptions. Bipolar disorder (BP) is a mood disorder characterized by episodes of mania and depression. Here, we expand our prior work to identify NREM biomarkers of BP, and determine whether they are shared with SCZ or disease specific.

Methods: Overnight polysomnography with 64-channel high-density EEG was performed on 59 bipolar disorder patients, 59 schizophrenia patients, and 85 healthy controls (both sexes included, age- and sex-matched) at Wuxi Mental Health Center, China. Additionally, 24-hour EEG/EMG recordings were obtained from female Akap11 wild-type, heterozygous, and knockout mice (n varied by genotype) using surgically implanted electrodes. Sleep staging and microarchitecture analysis including spectral power, spindle detection, and connectivity measures were performed using Luna open-source software for both human N2/N3 and mouse NREM sleep. Statistical comparisons between groups were conducted in R, adjusting for age and sex, with additional medication-specific analyses controlling for individual drug classes in the bipolar disorder cohort.

Results: BP patients exhibited decreased N1 duration and distinct spectral power changes including decreased delta (0–4 Hz) and fast sigma (13–15 Hz) power, with increased slow sigma (10–12 Hz) power that differentiated BP from both controls and SCZ patients. During N3 sleep, BP showed increased 11 Hz spindle density (most significant effect at channel CP5, standardized beta = 0.86, empirically-adjusted p-value < 0.0001), amplitude (most significant effect at channel FT7, standardized beta = 0.60, empirically-adjusted p-value < 0.0001), and duration (most significant effect at channel Fp1, standardized beta = 0.74, empirically-adjusted p-value < 0.0001) compared to controls and SCZ. Systematic lowering of detection thresholds revealed increased density of low-amplitude 11 Hz spindles in BP, which was replicated in Akap11 knockout mice (a genetic model of BP susceptibility) but not in Cacna1g knockout mice (associated with SCZ risk). These BP-specific spindle abnormalities remained significant after controlling for most medications, with the exception of benzodiazepines affecting spindle density measures.

Conclusions: We find distinct changes in BP NREM sleep that may reflect sleep-specific signatures of disease and neurocircuit changes. Specifically, changes in slow sigma power and slow spindle metrics were largely unaffected by medication and differentiated BP from SCZ. These findings demonstrate the utility of studying sleep biomarkers of disease. The consistency of these findings with the Akap11 mouse model, which also showed an increase in low-amplitude spindles, provides support for its use as a promising genetic model for BP, a tool that is currently lacking in the field. This model will enable future studies to dissect the specific neurocircuitry underlying these BP-specific sleep spindle changes.

Keywords: EEG biomarkers, sleep spindles, schizophrenia and bipolar disorders, High Density Sleep EEG, genetic mouse model.

Disclosure: Nothing to disclose.

P816. Sleep disordered breathing and ADHD—is there a connection?

Iliyan Ivanov, Muhammad Parvaz, Dana Prodanova, Jeffrey Newcorn, David Kouassi, Isabelle Moger, Ronan Cunningham, Oghene Tejiri Smith, Johanna Jarcho

The Icahn School of Medicine at Mount Sinai, New York, New York, United States

Background: Estimates show that 25 to 50% of children with ADHD experience sleep issues. Sleep disordered breathing (SDB) is one of the most prevalent sleep disorders in this population and may play a particularly important role in ADHD for two reasons: 1) accumulating evidence suggests that SDB exacerbates pre-existing ADHD, and 2) SDB independently produces symptoms that mimic ADHD. SDB encompasses a range of breathing difficulties during sleep, most commonly associated with mouth breathing rather than nasal breathing. One of the anatomical features that contributes to dominant mouth breathing in SDB is misalignment between the maxilla (upper jaw bone) and mandibular (lower jaw bone), which can be ameliorated with orthodontic intervention. Thus, orthodontic intervention may be a novel, non-pharmacological treatment to decrease risk and expression of ADHD. Importantly, ADHD in adolescents is frequently comorbid with disruptive behavior disorders, particularly oppositional defiant disorder and conduct disorder, highlighting the broader externalizing burden associated with the condition. Moreover, aggression—especially impulsive and reactive aggression—is a common and clinically significant comorbidity in youths with ADHD, further compounding functional impairments and treatment challenges. A critical first step towards advancing these treatment innovations is demonstrating that SDB is more prevalent among children with more severe ADHD symptomology and associated externalizing behaviors.

Methods: A community sample of 10–15 year olds (N = 85; F = 55%; age: 12.39 ± 1.60 years) with no lifetime stimulant use underwent a structural MRI scan. Parents rated the severity of their child’s mental health problems on the Child Behavioral Checklist (CBCL). We used MRICron software to extract structural images and quantified the distance between the tip of the upper lip and the tip of the chin, a proxy measure of the misalignment between the maxilla and mandibular. A linear regression was performed to test the relation between misalignment and ADHD. Secondary analyses tested the same relation with other externalizing symptoms (conduct problems and aggression) that are commonly comorbid with ADHD. All regressions controlled for age and sex at birth.

Results: The greater maxilla-mandibular misalignment significantly predicted higher severity of parent reported ADHD (β = 0.54, p = 0.01). Secondary analyses demonstrated similar relationship between the greater maxilla-mandibular misalignment and more severe conduct problems (β = 0.32, p < 0.05), and aggression (β = 0.44, p < 0.05).

Conclusions: This is the first study to demonstrate that measures of maxilla to mandibular misalignment, a key anatomical feature that contributes to SDB, is positively correlated with severity of ADHD and other externalizing symptoms. These preliminary results open the door to a new line of research that could result in the much needed development of non-pharmacological treatment for, and possible prevention of, ADHD and comorbid externalizing symptoms in youth.

Keywords: ADHD, sleep disordered breathing, mulocclusion.

Disclosure: Nothing to disclose.

P817. Molecular rhythm alterations in the human nucleus accumbens in subjects with major depressive disorder

Megan Perez, RuoFei Yin, Micah Shelton, Nicole Horan, Madeline Scott, Jill Glausier, David Lewis, George Tseng, Colleen McClung, Marianne Seney, Kyle Ketchesin

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Major depressive disorder (MDD) is a highly prevalent and debilitating psychiatric disorder and remains a leading global cause of disability. In addition to core affective symptoms, MDD is consistently associated with major disruptions in sleep and circadian rhythms. While behavioral and physiological rhythm disruptions in MDD are well-documented, much less is known about how diurnal patterns of gene expression are altered in the human brain in this disorder. The nucleus accumbens (NAc), a key striatal region involved in motivation and reward processing, has been strongly implicated in the pathophysiology of depression. Here, we investigated how diurnal rhythms in gene expression are altered in the NAc in subjects with MDD and further explored DNA methylation as a potential epigenetic mechanism contributing to these alterations.

Methods: RNA-sequencing and DNA methylation profiling were performed on postmortem NAc tissue from subjects with recurrent MDD (n = 43) and unaffected comparison subjects (n = 43). Differential gene expression analyses were initially performed without considering time of death (TOD), including sex, age, race, and RIN as covariates. Given prior evidence for circadian disruption in MDD, we then stratified subjects by TOD (day vs. night) and performed additional differential expression analyses. Genes were considered differentially expressed at p < 0.05 and a |log2 fold change| > 0.26 (corresponding to > 20% change in expression). Diurnal rhythmicity in gene expression and DNA methylation was assessed using a nonlinear regression model based on individual TOD (p < 0.05), with the coefficient of determination (R-squared) used to quantify goodness-of-fit. Gain or loss of rhythmicity was determined by comparing differences in R-squared values between groups. Pathway enrichment was performed using Metascape.

Results: We identified many differentially expressed genes in the NAc of MDD subjects, with more observed in those who died at night (356 genes) compared to the day (216 genes). Genes related to myelination were significantly upregulated in MDD subjects who died during the day. In contrast, MDD subjects who died at night showed significant upregulation of cilia-associated genes and downregulation of vascular and angiogenesis-related genes. Across both groups, we identified large sets of rhythmic genes (MDD: 772 genes; comparison: 636 genes) with minimal overlap (77 genes) between MDD and comparison subjects. Genes that lost rhythmicity in MDD were enriched for translation and protein folding, whereas genes that gained rhythmicity in MDD were enriched for vascular and immune-related functions. Preliminary DNA methylation analyses similarly revealed distinct sets of rhythmic CpG sites in MDD and comparison subjects, with ongoing work aimed at determining whether these changes are associated with gene expression rhythm alterations.

Conclusions: Our findings show significant diurnal alterations in gene expression and DNA methylation in the NAc of individuals with MDD, suggesting that molecular rhythmicity is disrupted in this brain region. These alterations may contribute to striatal dysfunction relevant to the pathophysiology of depression. Ongoing analyses are focused on integrating the transcriptomics and epigenetic data to determine whether rhythmic DNA methylation contributes to dysregulated gene expression rhythms in MDD.

Keywords: Circadian rhythms, Postmortem, Depression, Nucleus Accumbens, Transcriptomics.

Disclosure: Nothing to disclose.

P818. Short-term effects of lemborexant on subjective sleep parameters in older adults with chronic insomnia

Margaret Moline, Dinesh Kumar, Jocelyn Y. Cheng, Taro Kishi, Michinori Koebisu

Eisai Inc., Nutley, New Jersey, United States, Nutley, New Jersey, United States

Background: Insomnia disorder is a chronic condition characterized by difficulty falling asleep, staying asleep, or both, at least 3 nights per week for at least 3 months that is associated with an impact on the patient’s functioning. Despite this chronicity, many clinical development programs have included phase 3 studies of short duration such as those for zolpidem extended-release (approved 2007; 3-week duration, but data analysis at 2 weeks) and zaleplon (approved 1999). By contrast, phase 3 studies of the 3 commercialized dual orexin-receptor antagonists (DORAs), such as lemborexant (LEM), were longer in duration, with 1- to 6-month double-blind, placebo-controlled efficacy components. Data has been previously reported based on analyses for the first 7 nights and last 7 nights of treatment for one month Study E2006-G000-304 (Study 304; NCT02783729). However, 2 drugs under development for the treatment of patients with insomnia (dimdazenil [benzodiazepine derivative and a partial positive allosteric modulator of the GABAA receptor] and vornorexant [DORA]) included 2-week trials with patient-reported (subjective) endpoints in their registration programs. Therefore, it was of interest to evaluate LEM data from Study 304 across the second 7 nights of treatment based on sleep diaries completed by older participants with insomnia disorder.

Methods: Study 304 was a global, multicenter, randomized, double-blind, parallel-group study comparing 2 dose levels of LEM (5 mg, 10 mg) with placebo and zolpidem (not reported here). Study 304 enrolled males aged ≥65 years and females aged ≥55 years who met the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria for insomnia disorder, specifically including sleep maintenance difficulties. Participants also could report difficulties with sleep onset based on DSM-5 criteria for insomnia disorder. Exclusion criteria included, among others, a current diagnosis of other sleep disorders, unstable medical or psychiatric conditions, or concomitant use of medications that would pose a drug interaction risk. After the first screening visit and during the entire study, sleep was assessed with sleep diaries. Confirmation of study eligibility was obtained from sleep diary data and data from 2 consecutive nights of baseline polysomnography (PSG) during the run-in period. After a minimum of 2 nights following the baseline PSGs, the run-in period ended, and the baseline period took place. Participants who continued to meet eligibility criteria entered the treatment period and were randomized (5:5:5:4 ratio) to receive LEM 5 mg (LEM5), LEM 10 mg (LEM10), zolpidem, or placebo (PBO). Data from the sleep diaries were analyzed for subjective sleep onset latency (sSOL), sleep efficiency (sSE), wake after sleep onset (sWASO), and total sleep time (sTST) by week, with 7 days from the run-in period serving as baseline, against which the first and second weeks of treatment were compared.

Results: Study 304 randomized 208 subjects to PBO, 266 to LEM5, and 269 to LEM10. Baseline values for sSOL (mean [SD]) were 55.90 (37.39), 65.79 (43.53), and 60.88 (42.51) for PBO, LEM5 and LEM10, respectively. Across the first week of treatment, the decreases from baseline were larger with LEM than with PBO: −6.83 (23.04), −22.54 (32.81), −21.88 (29.27). These differences persisted during the second week of treatment: −9.42 (26.82), −21.84 (32.30), −23.26 (30.37). Baseline values for sSE (mean [SD]) were 56.08 (17.34), 56.05 (17.09), 54.31 (18.32). The results across the first week of treatment showed increases from baseline that were larger with LEM than with PBO: least squares mean (LSM) (SE): 7.05 (0.93), 10.86 (0.83), 13.91 (0.84) and persisted during the second week of treatment: 7.59 (1.02), 11.01 (0.92), 13.99 (0.92). Baseline values for sWASO (mean [SD]) were 170.89 (80.68), 166.76 (82.05), and 175.35 (83.45). Across the first week of treatment, the decreases from baseline (LSM [SE]) were larger with LEM than with PBO: −28.21 (3.94), −40.85 (3.52), −54.60 (3.55). These differences persisted during the second week of treatment: −30.56 (4.39), −37.85 (3.89), −49.32 (3.92). Baseline values for sTST (mean [SD]) were 276.23 (87.65), 275.74 (83.65), 266.10 (92.16). The results across the first week of treatment showed increases from baseline (LSM [SE]) that were larger with LEM than with PBO: 32.52 (4.63), 51.83 (4.18), 67.22 (4.19). These differences persisted during the second week of treatment: 35.76 (5.09), 54.28 (4.56), 71.14 (4.57).

Conclusions: These data support the view that LEM works quickly to address difficulties with falling asleep and staying asleep, with persistent early-term efficacy, in this population of older adults with chronic insomnia.

Keywords: lemborexant, insomnia, insomnia disorder, pharmacotherapy, sleep diary.

Disclosure: Eisai Inc., Employee, Self.

P819. Seasonal patterns of gene expression in psychiatric disease

Katherine Lyman, Madeline Scott, Kyle Ketchesin, Kaitlyn Petersen, RuoFei Yin, Jaehyoung Choi, Xiangning Xue, Jill Glausier, David Lewis, Ana Andreazza, Marianne Seney, George Tseng, Colleen McClung

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

Background: Psychiatric disorders such as bipolar disorder (BD) and major depressive disorder (MDD) often have seasonal patterns of episodes with manic episodes in BD being more prevalent in the spring/summer and depressive episodes in BD and MDD being more prevalent in fall/winter. However, the mechanisms that lead to these patterns remains unknown. Our lab has previously shown that individuals with psychiatric disorders, when compared to subjects without psychiatric disease, have altered circadian rhythms in gene expression in postmortem brain, demonstrating that there are disruptions to normal rhythmic processes in the brain. Given these findings, we sought to determine whether there are seasonal expression patterns of gene expression and if so, do they differ in subjects with psychiatric disease.

Methods: Postmortem brain tissue samples were taken from 33 BD subjects (21 male, 12 female), 50 schizophrenia (SCZ) subjects (47 male, 13 female), 83 MDD subjects (58 male, 26 female), and 83 non-psychiatric comparison (NPC) subjects (58 male, 25 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) and anterior cingulate cortex (ACC) 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). Lists of significantly rhythmic genes were analyzed for the presence of particular gene pathways and peak expression times.

Results: Seasonal gene expression peak timing was found to vary across diagnostic groups and region. Interestingly, we found few genes that significantly varied with season in NPC subjects in either brain region (82 in DLPFC, 624 in ACC). In contrast, we found a large number of genes with seasonal rhythms in the DLPFC of BD subjects (6229 genes) which primarily peak in the late summer, with many fewer seasonally rhythmic genes in the ACC (132 genes). In MDD we also found a large summertime peak in gene expression in the ACC, (2395 rhythmic genes) with a smaller winter/spring peak in the DLPFC (456 genes). We also found a smaller winter/spring peak in the BD group in the ACC (132 genes). SCZ subjects had overall fewer seasonally rhythmic genes (388 in DLPFC; 401 in ACC) compared to the BD and MDD groups but still had some weaker peaks in the fall/winter in the DLPFC and spring/summer in the ACC. When examining specific genes which displayed a seasonal rhythm, we found that genes involved in mitochondrial function play a prominent role. This was particularly true in the DLPFC, where the genes with the highest amplitude rhythmicity belonged to the mitochondrial genome (MT-RNR1 and MT-RNR2) in both the BD and SCZ groups; interestingly, these peaked in January-February. Mitochondrial pathways within the rhythmic genes showing a summer peak in the DLPFC were further analyzed, and we found that mtRNA metabolism and mtDNA maintenance were both significantly over-represented amongst seasonally rhythmic genes in BD, but not in MDD, SCZ, or NPC.

Conclusions: The varied timing of peak gene expression amongst psychiatric diagnoses suggests that seasonal transcriptional differences may be related to seasonal symptom patterns observed in psychiatric disease. For example, subjects with BD display a peak in gene expression in the ACC during the winter months, while their peak expression in the DLPFC occurs during summer months, in concordance with seasonal symptom patterns. As mitochondria are involved in seasonal change responses, i.e. via thermogenesis and energetic adaptations to changing photoperiod, the unique timing of mitochondrial gene expression peaks throughout the year may also indicate a particular mitochondrial role in this process.

Keywords: Seasonality, biological rhythms, Postmortem Brain Tissue Gene Expression, mood disorders, Bipolar Disorder.

Disclosure: Nothing to disclose.

P820. Manipulating kynurenic acid in the lateral hypothalamus alters orexinergic activity and sleep-wake behavior in rats

Maria V. Piroli, Snezana Milosavljevic, Mindal L. Reese, Hayley A. Nicholson, Charlie A. Grant, Katherine M. Rentschler, Sujit Pujhari, Jim R. Fadel, Ana Pocivavsek

University of South Carolina School of Medicine, Columbia, South Carolina, United States

Background: Sleep disturbances are prevalent in neurocognitive disorders such as schizophrenia and age-related dementias. Kynurenic acid (KYNA), an astrocyte-derived metabolite of the kynurenine pathway, is increasingly recognized as a contributor to the pathophysiology of these disorders. Modest increases in KYNA, which antagonizes N-methyl-D-aspartate (NMDA) and α7 nicotinic acetylcholine (α7nACh) receptors, result in cognitive impairments and altered sleep architecture, specifically negatively impacting rapid eye-movement (REM) sleep and increased wakefulness (Pocivavsek et al. Sleep 2017). We hypothesized that elevations in brain KYNA impair sleep via activation of orexin neurons in the lateral hypothalamus (LH). To test this, we employed pharmacological and viral strategies to manipulate KYNA synthesis by targeting kynurenine aminotransferase II (KAT II), the primary KYNA-synthesizing enzyme in the brain.

Methods: Experiment 1: Immunohistochemistry was used to evaluate cFos expression in the LH orexinergic neurons. Wistar male and female rats were injected with vehicle, kynurenine (100 mg/kg), KAT II inhibitor PF-04859989 (30 mg/kg) alone, or inhibitor prior to kynurenine challenge and brain tissues collected at 2 hrs (N = 5/group).

Experiment 2: KYNA was infused intracerebroventricularly (icv) in a dose response manner (0µM, 1µM, 3µM, 10µM) at Zeitgeber time (ZT) 0, corresponding to lights-on, in Wistar rats (N = 12 female, 9 male) and electroencephalogram (EEG) and electromyogram (EMG) polysomnography was obtained with telemetry devices to evaluate sleep-wake behavior. Vigilance states—wake, REM sleep, and non-REM (NREM) sleep—were classified using an artificial intelligence neural network (Smith et al. NPP-DPN 2025) and validated by expert scoring.

Experiment 3: Virus-mediated gene transfer was employed to locally enhance expression of Aadat, the gene that encodes KAT II. An astrocyte targeting (GFAP promoter) adeno-associated virus serotype 5 (AAV5) vector encoding rat Aadat and mCherry reporter was bilaterally injected into the LH of rats. Microdialysis was performed in the LH to evaluate basal and stimulated (25 mg/kg kynurenine treatment) levels of extracellular KYNA via high-performance liquid chromatography (N = 6 males/group). Separate animals were used to evaluate EEG/EMG polysomnography and sleep-wake behavior (N = 5 males/group).

Results: Acute kynurenine administration significantly increased cFos expression in LH orexin neurons (P < 0.001). This effect was attenuated by pre-treatment with the KAT II inhibitor PF-04859989, implicating KYNA synthesis in orexinergic activation. Vigilance state durations were evaluated from ZT 0–4 and we found a dose-dependent impact of KYNA on sleep duration. KYNA infusion (10µM) at ZT0 increased wakefulness by 30% (P < 0.05), while reducing REM and NREM sleep by 48% and 26%, respectively (P < 0.05). Effects were dose-dependent and transient, with sleep architecture normalizing by ZT 2–4. AAV5-mediated overexpression of KAT II in LH astrocytes increased basal and kynurenine-stimulated KYNA levels (P < 0.05). This manipulation reduced REM sleep and altered vigilance state distribution during the light phase (ZT6-12).

Conclusions: These findings support a mechanistic role for KYNA in promoting wakefulness and disrupting REM/NREM sleep via orexinergic activation in the LH. We established a unique astrocyte-targeted AAV approach for manipulating KYNA synthesis in vivo, enabling new insights into sleep regulation. Our study highlights KYNA and the kynurenine pathway as pharmacologically targetable mechanisms for treating sleep disturbances in neuropsychiatric conditions and further draws attention to the orexin system as a target for treatment of sleep dysfunction in illnesses like schizophrenia.

Keywords: orexin/hypocretin, Lateral hypothalamus, REM sleep, Adeno-associated Virus, Astrocyte.

Disclosure: Nothing to disclose.

P821. Body mapping inner-body sensations associated with low to high mother-child face-to-face emotional synchrony

Andreane Lavallee, Quinn Manning, Lilian Bryan, Hope Hendry, Ginger Atwood, Austen Curtin, Jennifer Warmingham, Nicole Shearman, Vitoria Chavez, Jill Owen, Katrina Fuller, Yasmin Leon-Mateo, Elizabeth Matei, Paul Curtin, Dani Dumitriu

Columbia University Medical Center, New York, New York, United States

Background: Biobehavioral synchrony, i.e., coordination of physiological, neural and behavioral interpersonal signals during reciprocal mother-child interactions, is critical for building socioemotional skills. Mounting evidence suggests that mother-infant dyads reach biobehavioral synchrony during reciprocal face-to-face interactions, but uncertainty remains regarding the conditions under which biobehavioral synchrony emerges. Evidence from the adult literature offers insight into the potential of shared emotions, or emotional synchrony, as the facilitator of cardiac (ECG) and neural (EEG) synchrony. Hypothesizing that engaging in emotional synchrony may be an environmental input that promotes mother-child biobehavioral synchrony, we use an emotional connection screener to reliably quantify the strength of emotional synchrony using blind observers’ own subjective feelings. Nonetheless, emotional states are notoriously difficult to access experimentally. Here, the goal was to define emotional states as perceived by mothers as they interact with their infant in relation to observational rating of emotional synchrony.

Methods: Using methods proposed by Nummenmaa et al. in a series of PNAS papers showing that emotion-associated inner-body sensations can be topographically mapped to body areas, we tested the hypothesis that the strength of parent-child emotional synchrony during face-to-face interactions is associated with distinct subjective maternal feelings. As part of the COMBO Initiative at Columbia University, 211 mothers and their 2-year-old child completed a video visit where they were asked to interact face-to-face for 3min with child sitting in mother’s lap without the use of toys. Right after the interaction, mothers were asked to color on a body map “where in their body they felt their emotions during the interaction”. The videotaped face-to-face interactions were independently coded for emotional synchrony using the Welch Emotional Connection Screen (WECS) by trained blinded coders. Body maps were overlayed using imageJ to create four normalized density maps grouped based on emotional synchrony score quartiles.

Results: Upon visual inspection, mothers in the lowest quartile of emotional synchrony most distinctively perceived inner-body sensations in the brain area. Mothers in the higher quartile of emotional synchrony was most distinctively perceived inner-body sensations in the chest area. The hands and gut area were uniquely colored by mothers in the higher and lower emotional synchrony quartiles. Preliminary investigation of the statistical significance of these differences suggests that there are localized differences in entropy structures across these matrices (D = 0.005, p < 0.001).

Conclusions: These preliminary findings provide insight into the investigation of maternal subjective emotional experiences during mother-child interactions. Here, we find that inner-body sensations activate in distinct areas that may be closely aligned with low to high emotional synchrony as rated by blind observers. Specifically, mothers who exhibited higher levels of emotional synchrony with their child reported localized sensations in the chest, whereas mothers in lower synchrony quartiles more distinctly localized sensations to the brain. These results suggest that emotional synchrony may not only be observable in behavior but also embodied in mothers’ own subjective experience, potentially serving as an environmental input that facilitates broader biobehavioral synchrony. Further work investigating the statistical robustness of these findings is needed. Nonetheless, body mapping of subjective feelings is a novel approach for representing emotion-dependent sensations contributing to an in-depth understanding of what emotional synchrony emerging from face-to-face interactions entails, and more specifically how it is subjectively embodied. Ultimately, this work may lead to the identification of intervention targets for promoting biobehavioral synchrony and child socioemotional development.

Keywords: emotion, face expression synchrony, Perinatal development.

Disclosure: Nothing to disclose.

P822. Computational signatures of craving social connection across symptom dimensions

Shawn Rhoads, Kaustubh Kulkarni, Xiaosi Gu

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

Background: Globally, loneliness and social isolation have become an epidemic and a significant challenge for public health. Social craving, or the desire to connect with others, is a core feature of loneliness that motivates social activity seeking to relieve the distress associated with a perceived lack of social connection. Understanding mechanisms related to social craving is thus crucial for clarifying its connection to mental health symptoms such as depression and anxiety. Identifying how social motivation computations are disrupted across symptom dimensions may inform early interventions and targeted approaches for psychiatric disorders.

Methods: Here, we developed a novel paradigm and computational model to characterize the computational basis of momentary social craving. In a pilot study, participants (N = 93) completed a social cue-induced craving task in which they learned to make choices that either placed them in social cue-rich environments or not and rated their momentary social craving. They also completed a battery of self-report measures assessing various psychiatric symptoms, including depression and anxiety. We tested how symptom dimensions varied as a function of social craving computations.

Results: Using a dynamic computational model of decision-making and subjective craving with detailed model comparison, we found that social craving arises algorithmically as a weighted combination of bottom-up signals (i.e., cue-reactivity) and top-down signals (i.e., prior expectations), which in turn impacts future updating about social expectations. Furthermore, we found that depression and anxiety symptoms had opposing effects on social craving computations. The impact of current social expectations on future social craving was reduced in depression (b = −.26, t = −2.00, p = .048; controlling for social anxiety), but elevated in people with higher social anxiety (b = .30, t = 2.31, p = .026; controlling for depression).

Conclusions: These preliminary results provide a computational account of momentary social craving, demonstrating how craving emerges from expectations about social environments. Building on this framework, symptom associations reveal differential computational phenotypes that may explain why unresolved loneliness is a major risk factor for the onset and persistence of psychiatric symptoms. Together, these findings position social craving as a transdiagnostic marker bridging mental health dimensions and underscore the potential of computational psychiatry to inform targeted, personalized interventions.

Keywords: loneliness, Social Behavior, computational psychiatry, computational modeling, Transdiagnostic.

Disclosure: Nothing to disclose.

P823. Neural and hormonal mechanisms of altruism across digital and in-person social interactions

Vanessa Jeske, Gari Walkowitz, Matthias Uhl, Nevena Toporova, Ekaterina Schneider, Johannes Schultz, Marla Dressel, Abigail Marsh, Rene Hurlemann, Nina Marsh

School of Medicine and Health Sciences, University of Oldenburg, Oldenburg, Germany

Background: Social connectedness is a key protective factor for mental health, whereas deficits in empathy and altruism are central to many psychiatric disorders. At the same time, digitalization is transforming social life, with face-to-face encounters increasingly replaced by digital communication. Although empathy and altruism are critical mechanisms through which social interactions support mental well-being, it remains unclear how digital versus in-person encounters influence prosocial behavior. Oxytocin (OT), a neuropeptide central to social bonding, has been implicated in these processes, but its effects are highly context-dependent. We therefore examined how different interaction modes affect altruistic donations and how these effects are moderated by empathy and OT.

Methods: We conducted two experiments with a total of 223 healthy male participants using an established donation paradigm after one of three priming conditions: (i) in-person face-to-face conversation (PI), (ii) text-based digital conversation (DI), or (iii) no interaction (NI). In Experiment 1 (n = 103), participants were randomly assigned to one of the three priming conditions, and endogenous OT was measured before and after priming to test whether interaction type modulates peripheral OT and subsequent donation behavior. In Experiment 2 (n = 120), participants were randomly assigned in a double-blind design to receive intranasal OT (24 IU) or placebo and then performed the donation task during fMRI. The study was approved by the Institutional Review Board of the Medical Faculty of the University of Oldenburg and conducted in accordance with the Declaration of Helsinki.

Results: In Experiment 1, donation behavior was predicted by an interaction between priming condition and salivary OT concentration (F(2,68) = 6.70, p = 0.002, η² = 0.02), which further interacted with empathy (F(2,68) = 7.54, p = 0.001, η² = 0.15). This three-way interaction was replicated in Experiment 2 (treatment x priming x empathy: F(2,88) = 4.25, p = 0.017, η² = 0.09). In the NI condition, highly empathic participants donated more under OT than placebo (F(1,25) = 3.72, p = 0.065, η² = 0.13), whereas in the PI condition, OT was associated with reduced donations in empathic participants compared to placebo (F(1,32) = 4.33, p = 0.046, η² = 0.12). Whole-brain fMRI analyses (donations > no donations) revealed that DI elicited activity in the right parainsular/superior temporal cortex [46 −52 52] (T = 5.96, Z = 5.45) and dorsolateral prefrontal cortex [42 12 50] (T = 5.73, Z = 5.27; [42 28 40] (T = 4.84, Z = 4.55), whereas PI recruited the right superior parietal lobule [46 −50 52] (T = 5.47, Z = 5.06) and precentral gyrus [42 12 50] (T = 5.40, Z = 5.00) (all p’s < 0.05, FWE-corrected).

Conclusions: OT’s effects on altruistic behavior are not uniform but depend on empathy and the type of social interaction. The finding that digital versus in-person encounters engage distinct neural systems suggests that increasing digitalization may reshape the biological underpinnings of empathy and altruism. These insights are particularly relevant for psychiatry, highlighting the importance of tailoring interventions for social dysfunction to both social context and individual disposition. More broadly, the results underscore the need to integrate digital communication into psychiatric research and treatment within a precision psychiatry framework.

Keywords: oxytocin, altruism, social Interactions, digital psychiatry, fMRI.

Disclosure: Nothing to disclose.

P824. The association between social media use and mental health: a longitudinal adolescent brain cognitive development (ABCD) study

Hannah Reiter, Kelly Hay Yee Chan, Neslihan Yildiz Ozhan, Nora Penzel, Marek Kubicki, Yogesh Rathi, Suheyla Cetin-Karayumak, Benson Ku, Johanna Seitz-Holland

Mass General Brigham, Harvard Medical School, Sommerville, Massachusetts, United States

Background: Social media use has become a pervasive aspect of daily life for today’s children and adolescents. Although preliminary evidence suggests potential adverse mental health outcomes associated with social media use, longitudinal investigations using large, representative datasets remain limited. In this study, we will leverage the Adolescent Brain Cognitive Development (ABCD) dataset and apply epidemiological methods to test our working hypothesis that greater social media use among children and adolescents is associated with increased indicators of risk for psychopathology. Our primary specific aim is to characterize the relationship between baseline social media use and the progression of psychopathology risk indicators over a five-year follow-up period. In addition, we will test the association with various sociodemographic variables on the relationship between social media use and mental health.

Methods: We will examine the relationship between social media use and mental health using data from the Adolescent Brain Cognitive Development (ABCD) Study, a prospective, population-based cohort that began enrolling children aged 9–11 across the United States between 2016 and 2018. The ABCD Study includes repeated assessments of behavioral, cognitive, and environmental factors, with complete data currently available through the five-year follow-up. Please note that we currently have analyzed baseline to year 3 data. However, given the most recent data release, we aim to include baseline to year 5 data for the poster presentation.

The primary exposure of interest is social media use time, assessed via the self-reported 14-item Screen Time Questionnaire (STQ) administered at all ABCD follow-up visits. To quantify social media use, we computed a composite score that includes time spent texting, using social media platforms, and engaging in video chats. For our primary aim, we operationalized social media use as a dichotomous variable, using a cut-off of 30 minutes per day, consistent with current recommendations (8,868 individuals reported less than 30 minutes social media use/day and 455 reported more than 30 minutes social media use/day).

To assess mental health outcomes, we selected measures that reflect early indicators of increased risk for developing psychopathology. The primary outcomes are the internalizing and externalizing scores on the Child Behavior Checklist (CBCL), the total score on the Sleep Disturbance Scale for Children (SDS), and the presence of distressing psychotic like experiences as measured by the Prodromal Questionnaire-Brief Child Version.

Odds ratios assessed the influence of socio-demographic variables on social media use and age, sex, puberty status, race, ethnicity, household income, and socioeconomic status was included in all analyses as covariates to account for potential confounding.

We used linear mixed-effects models to characterize the association between baseline social media use and mental health outcomes across the follow-up time period. Our primary focus was time-by-group interaction effects to assess whether trajectories of mental health indicators differ between social media users and non-users.

Results: The results indicated that female sex, black race, lower household income, the child opportunity index, and neighborhood disadvantage score were associated with higher social media consumption (p < 0.05, adjusted Odds Ratio > 1.5). More social media consumption at baseline was associated with higher externalizing scores on the CBCL, and more distressing psychotic like experiences (p < 0.05 after FDR correction). We will repeat analyses including additional longitudinal data points and correcting for baseline mental health status.

Conclusions: In summary, our findings underscore the importance of demographic and socioeconomic variables in social media consumption. Additionally, preliminary results suggest a correlation between social media consumption and mental health outcomes. Future analyses will focus on disentangling the longitudinal relationships between social media use and mental health in children and adolescents.

Keywords: Social Media Use, Mental health, Adolescence.

Disclosure: Nothing to disclose.

P825. Exposure to early life adversity impairs social cooperative learning in rodents

Allie Cauchon, Henry Kietzman, Noor Nouaili, Alya Bagdas, Genevieve Thibodeau, Robin Bonomi, Boram Cho, Songjun Li, Jane Taylor

Yale University, New Haven, Connecticut, United States

Background: Early-life adversity (ELA) is a major risk factor for neuropsychiatric disorders, yet many individuals show resilience, particularly in social domains. Cooperative behavior, a higher-order social skill requiring coordination and mutual reward valuation, has not previously been studied in ELA-exposed animals. Additionally, entactogens such as MDMA (3,4-methylenedioxymethamphetamine) have emerged as a potential psychopharmacological agent that enhance affiliation and prosociality. In this study, we investigated the impact of ELA on cooperative learning and tested whether MDMA could ameliorate deficits in social cooperation. To probe molecular substrates, we used [¹⁸F]FPEB PET imaging to measure availability of metabotropic glutamate receptor 5 (mGluR5), which regulates synaptic plasticity and has been implicated in multiple psychiatric and neurodevelopmental disorders.

Methods: Male and female Long Evans rats (n = 62) were exposed to either limited bedding and nesting (LBN, postnatal days 2–9), chronic social isolation (CSI, P28–42), both, or neither (control). Animals were subsequently trained in adulthood on Pavlovian and instrumental learning tasks, followed by a novel cooperative task using a semi-naturalistic, automated social decision-making paradigm where animals obtain mutual reward only after coordinating an instrumental action (lever press). Animals also underwent additional testing in a Reciprocal Social Interaction (RSI) assay and yoked trials with a computer-simulated partner. We then performed PET imaging (n = 46) with the [¹⁸F]FPEB tracer to assess mGluR5 receptor availability in the insula, cingulate, orbitofrontal cortex (OFC), striatum, and amygdala. Pharmacological (MDMA, 5 mg/kg IP, either 30min or 24h prior to testing with a conspecific) and behavioral interventions (“tutor” sessions, where animals were paired with conspecifics that had learned to cooperate) were then tested to assess their ability to rescue cooperative behavior.

Results: ELA exposure did not affect Pavlovian (ANOVA, p = 0.4289) or instrumental training (ANOVA, p = 0.8865), but significantly impaired cooperative learning (47.4% ELA vs. 77% control; Fisher’s exact test, p = 0.0367). Among cooperative learners, acquisition speed was comparable to controls (ANOVA, p = 0.6543), and no sex differences were observed (mixed-effects model, p = 0.3543). Deficits were not explained by reduced social interaction, while ELA animals performed well in yoked trials with a computer-simulated partner (ANOVA, p < 0.01; Tukey’s p < 0.0001), indicating intact motivation and reward-seeking in non-social contexts. Interestingly, tutor sessions failed to improve performance in non-learners (n = 16, p > 0.3 across multiple sessions). Next, PET imaging showed a trend for reductions in mGluR5 availability (binding potential, BPnd) across the insula, cingulate, OFC, striatum, and amygdala for animals exposed to ELA compared to controls (ANOVA, p > 0.2). Multivariable regression, controlling for performance in a non-social decision making task, revealed a significant stress × cooperation interaction in the insula (adj R² = −0.5240; p = 0.0243), but no significant effects in other regions. Lastly, MDMA administration (n = 30), either 24h before testing or acutely 30min prior, did not rescue cooperative behavior. Both delayed (RM-ANOVA, p = 0.3298) and acute MDMA exposure (t-test, p < 0.0001) failed to enhance task engagement or lever pressing.

Conclusions: In this pilot study, we found that exposure to ELA significantly impaired acquisition of cooperative behavior in adult rats, despite intact decision making in non-social contexts. These findings highlight a specific vulnerability in higher order social learning rather than general cognitive or motivational deficits. Although our preliminary results failed to show evidence that behavioral “tutor” sessions or pharmacological intervention with acute or prior MDMA had benefit in non-learners, our ongoing studies are focused on why social cooperative learning deficits are resistant to these interventions. As such, PET imaging revealed that insular mGluR5 availability specifically interacted with stress and cooperation outcomes, implicating glutamatergic signaling in vulnerability. Given mGluR5’s role in synaptic plasticity and potentially social-affective processing, allosteric modulators may represent a promising pharmacological strategy to restore cooperative behavior following stress in future translational studies.

Keywords: Early life stress (ELS), social learning, stress resilience and susceptibility, metabotropic glutamate receptor 5 (mGluR5), insula cortex.

Disclosure: Nothing to disclose.

P826. Paraventricular thalamic astrocytes regulate social novelty through glutamate transporter-mediated adaptation in synaptic function and neural circuits

Seungwoo Kang

Augusta University Medical College of Georgia, Augusta, Georgia, United States

Background: Social behaviors play an essential role in maintaining healthy living patterns. Most mammalian species exhibit a natural tendency toward social novelty, demonstrating an inherent drive to pursue and form new social connections with unknown members of their communities. Consequently, impaired social activities for unfamiliar individuals are often characteristic of neurological and psychiatric disorders including autism spectrum disorder. The paraventricular nucleus of the thalamus (PVT) functions as a critical neuroanatomical hub mediating context-dependent salience attribution and aversive stimulus processing, including social behavioral regulation. Indeed, recent studies suggest that the PVT presents a variety of neural signals according to social condition and the activities are significantly correlated with social behaviors.

Accumulating evidence demonstrates that astrocytes not only provide microstructural support within neural circuits but also physically constitute tripartite synapses with adjacent neurons and directly modulate neuronal activity through gliotransmitter release and neurotransmitter reuptake. However, whether coordinated astrocyte-neuron activity in the PVT encodes social behaviors remains unknown.

Methods: To elucidate astrocyte activity-dependent modulation of adjacent neuronal function, its downstream multi-regional circuits, and associated behavioral phenotypes, we employed chemogenetic activation of PVT astrocytes utilizing GFAP promoter-mediated expression of hM3Dq excitatory designer receptors exclusively activated by designer drugs (DREADDs). Initially, we quantified chemogenetically-induced cellular activation in PVT astrocytes expressing GfaABC1D promoter-driven GCaMP8s fluorescence-based calcium indicators. Subsequently, we recorded electrophysiological alterations in PVT neuronal spiking, inter-regional coherence within PVT circuitry, and synaptic transmission properties of PVT neurons.

To assess behavioral outcomes, we subjected mice to social approach assays. Mice were sequentially exposed to three distinct social conditions: (1) absence of social stimuli for habituation and baseline place preference assessment, (2) presence of an unfamiliar conspecific, and (3) simultaneous presentation of both unfamiliar and familiar conspecifics. Social preference was quantified by comparing the duration of interaction between the test animal and either a stranger versus a novel object (sociability) or a stranger versus a familiar conspecific (social novelty).

Results: Using fiber-photometry, we confirmed that chemogenetic activation of the PVT astrocytes potentiated the intracellular calcium transients in the astrocytes of the PVT. In vivo electrophysiological recordings showed that astrocyte activation significantly increased neuronal firings in the PVT. Simultaneous local field potential recordings from multiple PVT downstream targets, including the mPFC, NAc, and amygdala, revealed that PVT local astrocyte activation decreased theta-frequency coherence between these regions and the PVT. Ex vivo electrophysiological recordings showed that PVT astrocyte activation enhanced evoked excitatory postsynaptic currents (eEPSCs) in neighboring PVT neurons. The enhancement of eEPSCs was abolished by pretreatment with DHK, a selective antagonist of the astrocyte-predominant glutamate transporter GLT1 (EAAT2, slc1a2), indicating that GLT1 is, at least, partially required for astrocyte-mediated modulation of PVT neuronal activity.

In Social Behavioral evaluation, the PVT astrocyte activation significantly impaired social novelty preference in the social approach paradigm. Notably, conditional deletion of GLT1 specifically in PVT astrocytes recapitulated the social behavioral deficits induced by astrocyte activation.

Conclusions: Collectively, our observation indicates that PVT astrocyte activity regulates local synaptic function and PVT circuitry, through active modulation of glutamatergic signaling, with GLT1 serving as a partial modulator of these effects on social behavioral output. These results highlight the critical role of astrocytes as an additional regulatory layer in social behavior modulation and suggest that astrocyte-neuron interactions mediated by glutamatergic signaling may represent a novel therapeutic intervention for associated neuropsychiatric disorders.

Keywords: astrocyte-neuron interaction, glutamate transporter (EAAT2), Social Behavior.

Disclosure: Nothing to disclose.

P827. Oxytocin signaling in the BNST dampens anxiety-like internal state and mediates social buffering effects on stress

Hao Li, Ziqing Huai, Talia Fernandez, Reesha Patel, Cameron Good, Callista Polasek, Anagha Kangovi, Yuxiao Wu, Addison Kim, Maya Ramos, Alida Tonga Sani

Feinberg School of Medicine, Northwestern University, Chicago, Chicago, Illinois, United States

Background: Social support provides powerful stress-buffering effects, yet the neural mechanisms underlying social modulation of anxiety- and stress-related behaviors remain poorly understood. The bed nucleus of the stria terminalis (BNST) integrates stress and social information, while oxytocin signaling is crucial for social behavior and stress regulation. We hypothesized that oxytocin receptor (OXTR) neurons in the BNST mediate social buffering of stress-induced anxiety states.

Methods: Male C57BL/6J male mice underwent footshock stress (10x 0.5–0.7 mA footshocks) followed by reunion with unstressed cagemates or alone in their homecage. Social behaviors were quantified before (30 mins) and after (60 mins) footshock stress for the reunion group. Anxiety-like behavior was assessed using marble burying and light-dark box tasks. BNST OXTR neurons were manipulated using DREADDs (Designer Receptors Exclusively Activated by Designer Drugs) with CNO administration. Neural activation was measured using cFos immunohistochemistry and fiber photometry. OXT release in the BNST was measured using an OXT fluorescent sensor. Physiological stress responses were monitored using implantable heart rate sensors.

Results: We found that mice received increased social support after footshocks, as indicated by allogrooming (p = 0.0313) and total social interaction (p = 0.0270), compared to the pre-shock baseline period. This increase in social support behaviors is independent of prior experience with footshock, social hierarchy, or sex. Following reunion with cagemates, we found the stressed mice reduced anxiety-like behavior in the marble burying test compared to the alone group (reunion: 4.2 ± 1.1 vs alone: 7.8 ± 1.5 marbles buried, p = 0.0181). Increasingly, we found that four-week pair-housing enhanced social support behavior compared to one-week housing, as observed in both allogrooming and sniffing behaviors. Using an OXT sensor, we found that BNST OXT sensor activity in response to social support behaviors is enhanced by footshock stress, suggesting increased OXT release in the BNST during social buffering. Furthermore, cFos immunostaining results show that contextual fear predominantly activates non-OXTR-expressing neurons in the BNST, consistent with existing findings that OXTR activation in the BNST is inhibitory. Consistently, we found that DREADDs excitation of BNST OXTR neurons suppressed anxiety-like behaviors in the light-dark box test (p = 0.0016), while the inhibition enhanced anxiety-like behaviors (p = 0.0436). In addition, inhibition of BNST OXTR neurons delayed the recovery in heart rate when exposed to an anxiolytic environment (p = 0.042073).

Conclusions: These findings demonstrate that BNST oxytocin signaling mediates social buffering of stress-induced anxiety. Social reunion reduces anxiety-like behavior and activates BNST OXTR neurons during stress recall. Chemogenetic manipulation confirms that BNST OXTR neurons causally regulate anxiety states and physiological stress responses. This circuit represents a novel target for understanding social modulation of anxiety and developing treatments for stress-related psychiatric disorders.

Keywords: oxytocin, Social buffering, Anxiety and stress, BNST, neuropeptides.

Disclosure: Nothing to disclose.

P828. Sp4 hypomorphic mice exhibit abnormal social behavior and reduced sensitivity to the pro-social effects of MDMA

Tyler Dexter, Shylee Roohian, Richard Sharp, Landon Cunningham, Elizabeth Hoang, Melissa Flesher, Susan Powell, Jared Young

University of California - San Diego, San Diego, California, United States

Background: Deficits in social drive and decision-making are debilitating symptoms present in various psychiatric disorders including schizophrenia, necessitating the development of targeted therapeutic interventions. Importantly, impaired social functioning can manifest in various forms, including reduced social reward valuation, reduced social motivation, and/or impaired ability to use social information to guide decision-making. Thus, the use of animal models to investigate social behaviors relevant to psychiatric disorders should implement behavioral testing paradigms that enable the measurement of multiple social behaviors with translational validity. Truncation at the Sp4 gene has been identified by multiple genome wide association studies as a significant risk factor for the development of schizophrenia. Sp4 hypmorphic mice, which exhibit reduced Sp4 expression, recapitulate multiple behavioral abnormalities associated with schizophrenia, though the presence of social impairments has yet to be tested. Here, we characterized the social profile of the Sp4 hypomorphic mouse model using the classic 3-chamber social interaction test and a recently developed operant paradigm assaying social motivation and decision-making. Additionally, we evaluated whether MDMA, which increases sociability in humans and has garnered interest as a potential therapeutic compound, could modulate social preference in Sp4 and wildtype (WT) mice.

Methods: In addition to social characterizations, Sp4 (n = 10 M, 10 F) and WT controls (n = 12 M, 17 F) were tested in prepulse inhibition (PPI) and a progressive ratio breakpoint task for food (fPRBT), measuring impaired sensorimotor gating and effortful motivation, respectively, which are deficient in schizophrenia. Mice were individually housed for all social experiments and trained to lever press for access to a social partner, before testing in the social PRBT (sPRBT) and social vs. food decision-making task. For MDMA experiments, mice received i.p. injections of saline, 3 or 7 mg/kg MDMA 15 min prior to testing with novel stimulus mice used for each session. Lastly, we tested mice on the 3-chamber social interaction task, which assessed preference for interactions with a novel social partner vs. a novel object.

Results: We first replicated schizophrenia-relevant behavioral deficits of Sp4 mice, which exhibited impaired PPI (F(1, 95) = 6.049, p = 0.02) and fPRBT (F(1, 35) = 7.183, p = 0.01). Interestingly, Sp4 mice exhibited intact social motivation as measured by the sPRBT (F(1, 36) = 1.874, p = 0.18) and social decision-making (F(1, 26) = 0.2451, p = 0.625). During the social decision-making task, MDMA treatment significantly increased social preference and interacted with genotype (F(2, 52) = 5.213, p = 0.009). Post hoc analyses revealed significantly higher social preference in WT mice treated with the highest MDMA dose compared to Sp4 mice (p < 0.01). Lastly, we observed reduced social engagement in Sp4 mice on the 3-chamber social interaction task (F(1, 45) = 12.30, p = 0.001), which did not develop a significant social preference unlike WT mice (p = 0.263 and p = 0.001, respectively). Interestingly, we also observed significant positive correlations between indices of social preference on the operant social task and 3 chamber social task, but only in female mice (R2 = 5.59, p = 0.01).

Conclusions: These data provide the first characterization of social abnormalities in the Sp4 hypomorphic mouse model of schizophrenia. Sp4 hypomorphic mice exhibited reduced social preference in the traditional 3-chamber social preference test, replicable schizophrenia-like deficits in PPI and fPRBT, but not in the operant social paradigms (i.e., sPRBT and social vs. food decision-making). Interestingly, we found a positive correlation between social preference and social reward choices in female but not male mice. Importantly, we observed that MDMA increased social preference in the social vs. food decision-making paradigm but that this pro-social effect was blunted in Sp4 mice compared to WT mice. Thus, alterations associated with reduced Sp4 expression may blunt the pro-social effectiveness of MDMA. Together, these data provide: 1) An important characterization of social abnormalities in a mouse model with high genetic relevance to schizophrenia; 2) Further support for the pro-social effects of MDMA; and 3) Insight into the therapeutic potential of MDMA for social impairments in psychiatric disorders.

Keywords: social cognition, Schizophrenia (SZ), MDMA, Translational Cognitive Testing.

Disclosure: Nothing to disclose.

P829. Neurotrophin signaling in the medial orbitofrontal cortex is essential for socially-guided choice

Trevor Towner, Esther Seo, Alexa Leahy, Elle McCall, Kaisi Xing, Tanvi Anand, Shannon Gourley

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

Background: The neurotrophin, Brain-derived neurotrophic factor (BDNF), within the medial orbitofrontal cortex (MO), is a key regulator of social and decision-making behaviors. We recently developed a task, social incentivization of future choice (SIFC), that allows us to evaluate the process by which prior social experiences promote subsequent reward-seeking behavior, thus connecting two separable functions of the MO. Here we evaluated the role of BDNF within the MO, and its location of receptor binding given its antero- and retrograde transport, in socially-incentivized and non-social decision making.

Methods: Male and female mice were trained to nose poke for two different reinforcers. One reinforcer type was then paired with a social experience whereas the alternative reinforcer type was paired with a novel object. In a subsequent choice test, naïve mice preferentially respond at the nose poke predictive of the socially-paired reinforcer. As a non-social decision-making task, mice were trained to nose for a single reinforcer prior to an instrumental set shift, during which lever pressing became the reinforced operant behavior. Using viral-mediated genetic knockout and knock in, as well as chemogenetic approaches, we asked whether the loss BDNF within the MO and/or the activity of BDNF at its cognate receptor, tropomyosin receptor kinase B (trkB), are uniquely necessary for social but not non-social decision making.

Results: We find that both neuronal and glial reductions in Bdnf within the MO impair the ability of mice to choose reinforcers based on prior social experience, even while they are dispensable for a non-social decision-making task requiring the MO. BDNF binding to its high-affinity receptor trkB within the MO is necessary for SIFC, given that overexpression of a dominant-negative receptor in the MO and on neurons projecting from the MO to the basolateral amygdala (BLA) occlude SIFC performance. In addition, local Bdnf loss results in fewer mature dendritic spines on excitatory neurons within the MO but not on neurons within the BLA. Chemogenetic manipulations indicate that the MO contributes to SIFC by enabling mice to form associations between social experience and external reward. Ongoing investigations are evaluating whether BDNF signaling promotes SIFC through its regulation of social engrams held in the BLA.

Conclusions: BDNF and trkB activation in the MO and more specifically, on MO to BLA projection neurons, regulates the ability to select actions predictive of socially-associated reinforcers. Current investigations are further elucidating whether BDNF/trkB in the MO are influencing social engrams in the BLA, resulting in impaired social decision making.

Keywords: Orbitofrontal, Decision Making, Social Behavior, Neurotrophins.

Disclosure: Nothing to disclose.

P830. An endocannabinoid-regulated social-active ensemble encodes negative affect in the nucleus accumbens of a shank3 autism mouse model

Oakleigh Folkes, Meaghan Donahue, Paola Negrón-Moreno, Yong-Hui Jiang

Yale, New Haven, Connecticut, United States

Background: Social behavior deficits are common symptoms of neuropsychiatric disorders, including autism spectrum disorder, but there are limited pharmacological treatments for these symptoms. Therefore, understanding how neurons encode social information will give insight into identifying novel pharmacological targets to improve sociability.

SHANK3 encodes a postsynaptic scaffold protein and is a risk gene for autism spectrum disorder. Our lab previously developed the Shank3∆e4-22 mouse model, which recapitulates patients’ social deficits and shows abnormal nucleus accumbens (NAc) circuitry, a critical region for social behaviors. Therefore, we tested the hypothesis that Shank3 deletion alters NAc neural response to social investigation.

Methods: Using miniscopes we recorded single-cell calcium activity in the Nucleus Accumbens of Male and female, WT and Shank3e4-22, mice (n = 5–6/ group) during a 3-chamber social interaction task and a social dyadic task. Social investigation was evaluated by calculating the time animals spent in the chamber with a mouse or the time spent in body contact with an unfamiliar, sex-matched mouse. We then injected animals with Monoacylglycerol lipase inhibitor (JZL-184) to increase endocannabinoid tone while recording neural activity during a social dyadic task.

Next, we identified neuronal ensembles activated by social interaction using ArcCreER:Ai14 (ArcTRAP) mice crossed with WT and Shank3e4-22 lines (n = 6–7/group). We captured active neurons during a social dyadic task with an unfamiliar, sex-matched juvenile mouse. We then used cre-dependent in vivo optogenetic tools (DIO-NpHR) to selectively inhibit social ensembles during a real-time place preference assay, and during the 3-chamber social interaction task.

Results: NAc single-cell activity is significantly increased during the first social interaction bout of the 3-chamber task in Shank3e4-22 mice (*p < 0.05) compared to WT controls. Shank3e4-22 mice also have increased neural activity in a subpopulation of up-modulated cells (*p < 0.05), and lower neural activity in down-modulated cells (*p < 0.01). These data indicate that NAc neurons in Shank3e4-22 mice are hypermodulated in response to the first social bout. When we administered JZL184, we restored social preference (**p < 0.01) and decreased neural response to the first social bout (**p < 0.01) in Shank3e4-22 mice.

Next, we investigated the function of social ensembles in ArcTRAP x Shank3e4-22 mice. In the real-time place preference assay, we inhibited social ensembles by turning ON inhibitory opsins in an ON chamber, or OFF in the opposite chamber. Shank3e4-22 mice spend significantly more time in the ON chamber than the OFF (***p < 0.01), whereas WT mice have no preference for either side. Next, we inhibited social ensembles during a 3-chamber task and restored social preference in Shank3e4-22 mice (**p < 0.01) and disrupted social preference in WT mice.

Conclusions: Our work revealed that social ensembles in the Nucleus Accumbens of Shank3e4-22 mice encode a negative affective state. Social deficits were previously attributed to a lack of neural response, but our findings reveal that, in an autism mouse model, social ensembles encode qualitatively different information compared to controls.

Keywords: Shank3-deficient mice, Social Behavior, Neuronal ensembles, Endocannabinoids, Nucleus Accumbens.

Disclosure: Nothing to disclose.

P831. Adolescent social isolation in female mice impairs postpartum caregiving and produces intergenerational deficits in offspring social behavior

Jose Francis de Oliveira, Rinako Tanaka, Shin-ichi Kano, Minae Niwa

University of Alabama at Birmingham, Birmingham, Alabama, United States

Background: Mammalian social bonds, including maternal attachment and pair bonding, significantly influence maternal mental health and offspring socioemotional development. Adolescence is a sensitive period for social brain maturation, during which adverse experiences can lead to lasting vulnerability to dysfunction later in life. We have shown that adolescent social isolation in female mice establishes a subclinical vulnerability that manifests after delivery as impaired postpartum behaviors linked to mood and social cognition, mediated by glucocorticoid receptor-dependent hypofunction of the anterior insula-prelimbic cortex pathway (Kin et al, Nat Commun 2023; Niwa et al, Nat Ment Health 2024). This model of early-life adversity (ELA) allows mechanistic dissection of how adolescent adversity sensitizes maternal circuits and influences intergenerational outcomes. In the present study, we tested whether maternal ELA disrupts postpartum caregiving and alters social behavior in adult offspring.

Methods: Virgin C57BL/6J female mice underwent mild social isolation during late adolescence (postnatal weeks 5–8; visual/auditory/olfactory access maintained). At 8 weeks, females were mated with unstressed males and gave birth to pups. Maternal behaviors were assessed in separate cohorts at postpartum day (PPD) 0 and 7 using the basal maternal behavioral test (BMBT) and pup retrieval test (PRT) (N = 11–12 for BMBT, N = 10–11 for PRT).

Adult offspring (PND 56) underwent the three-chamber social interaction test (SIT; sociability and social novelty trials) and olfactory recognition test (ORT; non-social vs. social odors from familiar or novel mice). (N = 19–23 for SIT, N = 11–14 for ORT). Open field, elevated plus maze, light/dark box, and novel object recognition tests were conducted to assess general locomotor and anxiety-related behaviors (N = 24–33).

Data were analyzed using t-tests, ANOVA, and nonparametric equivalents. A significance level of p < 0.05 was considered. No sex differences were found in offspring, so data were combined unless noted.

Results: The shape score of the nest and duration of caring behaviors (licking and nursing), but not duration of self-caring behaviors (climbing, digging, and self-grooming), in dams exposed to adolescent social isolation (stressed dams) at PPD 7, but not PPD 0, were significantly reduced compared to those in unstressed dams (p < 0.05). We observed that stressed dams spent substantially more time retrieving pups compared to unstressed dams at PPD 0 (p < 0.05). Behavioral deficits in PRT in stressed dams worsened and persisted for at least one week after delivery (p < 0.05). These behavioral deficits in stressed dams might represent prolonged abnormalities in maternal behavior.

Offspring reared by stressed dams (stressed offspring) showed deficits in sociability in SIT at PND 56 (p < 0.05). In ORT, both unstressed and stressed offspring exhibited normal behavior in response to non-social cues (water and coconut/almond odors). In the context of social cues, both unstressed and stressed offspring displayed a normal response to social cues from the test mouse’s home cage. In contrast, stressed offspring showed significant deficits in response to a novel social cue from the stranger’s cage (p < 0.05). Furthermore, we observed that stressed offspring did not show behavioral changes in the open field, elevated plus maze, light/dark box, and novel object recognition tests at PND 56 compared to unstressed offspring. Our data suggested that adolescent social isolation-induced behavioral deficits in stressed dams might be a causal factor for social behavioral deficits in adult stressed offspring.

Conclusions: Adolescent social isolation in future mothers leads to selective and lasting impairments in maternal caregiving, which are associated with specific impairments in offspring social cognition without affecting anxiety-like or non-social behaviors. These findings highlight a developmentally specific maternal adversity pathway with intergenerational consequences and suggest social cue processing as a potential translational target for prevention strategies.

Keywords: Adolescent social isolation, Postpartum caregiving, Intergenerational Transmission, Social Behavior, Early-life stress/adversity.

Disclosure: Nothing to disclose.

P832. Biological and psychological factors as predictors of the course of depression in youth following parental cancer diagnosis

Salahudeen Mirza, Meilin Jia-Richards, Eli Goodfriend, Isabel Novacich, Dara Sakolsky, Kehui Chen, David Brent, Anna Marsland, Nadine Melhem

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

Background: Depression in youth is a significant source of distress and longer-term morbidity, with heterogeneous trajectories. Early intervention relies on a strong understanding of antecedent signs of risk. In this study, we examined the longitudinal trajectories of depression symptoms in children of parents recently diagnosed with cancer and examined their prediction by psychosocial factors and biological markers of stress and inflammation.

Methods: The sample consisted of 332 youth (mean age = 16.2 years, 59% female), distributed as 117 (35%) youth with a recent parental cancer diagnosis (Stress group) and 215 (65%) youth without a parental cancer diagnosis (Control group). The sample was followed up at 6 and 18 months following baseline. At each timepoint, youth completed psychosocial assessments; provided hair for the measurement of hair cortisol concentrations (HCC) in the 3 cm segment closest to the scalp representing the past three months; and provided blood samples to measure cytokines (IL-1RA, TNF-α, IL-6, IL-10) and C-reactive protein (CRP). We used growth mixture modeling (GMM) to identify depression trajectories. First, chi-square test and classic or Welch’s one-way analysis of variance (ANOVA) with post-hoc tests were conducted to compare depression trajectories on group and baseline demographic (including sex as a biological variable), psychosocial, and biological measures. Next, baseline measures which were significantly different between depression trajectories at P < .05 were standardized if continuous and jointly tested in hierarchical multinomial logit regression, with Low trajectory as reference. Group (Stress vs. Control) was included as a key covariate in the final model, and interactions of each measure with group as well as mediation using the Karlson-Holm-Breen method were tested. Linear mixed-effects models, also adjusted for group, were used to identify psychosocial and biological measures whose change over the course of the study differed by depression trajectories, with Low trajectory as reference. Biological variables were skewed and as such, we applied a natural logarithmic transformation.

Results: GMM revealed three depression trajectories: High (n = 42; 13%), Medium (n = 87; 26%), and Low (n = 203; 61%). Youth in the High depression trajectory were more likely to be in the Stress group compared to Low depression trajectory (50% vs. 31%, χ2 = 5.9, df = 2, P = .053). Youth in the High depression trajectory showed higher HCC (F(2, 63.9) = 6.0, P = .004) and lower CRP (F(2, 34.6) = 5.6, P = .007) than Low depression trajectory. Multinomial logit regression with only group showed that Stress group (RRR = 2.22, 95% CI = 1.11–4.42, P = .023) significantly predicted High depression trajectory. However, group was no longer significant in the final multinomial logit model controlling for psychosocial characteristics. Prior mood disorder (RRR = 9.76, 95% CI = 2.77–34.41, P < .001), higher perceived stress (RRR = 7.86, 95% CI = 2.93–21,07, P < .001), higher internalizing problems (RRR = 5.94, 95% CI = 2.68–13.17, P < .001), and lower perceived parental care (RRR = .47, 95% CI = .24-.89, P = .022) significantly predicted High depression trajectory. Only higher perceived stress (RRR = 1.79, 95% CI = 1.14–2.79, P = .010) and higher internalizing problems (RRR = 3.51, 95% CI = 2.30–5.34, P < .001) significantly predicted Medium depression trajectory. When HCC and CRP were each added to this model, HCC (RRR = 3.93, 95% CI = 1.30–11.83, P = .015) significantly predicted High depression trajectory. There were no significant two-way interactions between psychosocial or biological characteristics and group. Significant mediation of the group effect for High depression trajectory was identified for perceived stress (91.44%, P = .001) and internalizing problems (70.38%, P = .004). Mixed-effects models showed that substance use problems significantly increased over time in the High depression trajectory (b = .04, 95% CI = .02-.06, P < .001).

Conclusions: A subset of youth showed a concerning pattern of elevated depression over eighteen months. Pre-existing mood disorder, stress appraisal, emotional problems, parent-child relationship, and HCC predicted this high depression trajectory. Some of these mechanisms mediated the effects of chronic stress from parental cancer diagnosis. Further, youth with elevated depression showed an increase in substance use problems over time. These results have important implications for screening of youth at risk for depression and better understandings of the embedding of chronic stress into psychopathology.

Keywords: Early-life stress/adversity, Adolescent Depression, hypothalamic-pituitary-adrenal axis, Immune Biomarkers, Symptom Trajectories.

Disclosure: Nothing to disclose.

P833. Sex differences in neural correlates of subjective stress and prediction of future real-world stress

Clara Freeman, Cheryl Lacadie, Rajita Sinha

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

Background: Increased sensitivity to stress exposure, including greater stress reactivity and difficulty regulating stress responses, is a transdiagnostic feature of psychopathology. Women may have heightened subjective stress responses compared to men, a difference thought to contribute to greater prevalence of internalizing disorders such as depression and anxiety. However, studies of sex differences in neural markers of acute stress reactivity have yielded mixed results, perhaps due to small sample sizes. Therefore, the mechanisms underlying the experience of higher subjective stress reactivity and its link to sustained distress in women remain unclear. To address this question, the present study assessed neural correlates of subjective stress, both average stress and change in stress over the course of acute stress exposure, in a large sample of men and women. We then tested whether these neural correlates predicted real-world prospective daily stress ratings.

Methods: Participants (N = 118, 52.5% women) completed a well-validated stress provocation task during fMRI scanning in which they viewed two blocks of images (stressful/aversive, neutral control) in a randomized order. Each block contained six one-minute runs of images. Following each run per block, participants rated their stress level on a 9-point Likert scale. Random-intercept linear mixed-effects models tested sex-by-condition (stress, neutral) interactions in stress ratings and in change in stress ratings (last two runs—first two runs). Next, we ran whole-brain 3dLME models in AFNI (p < .001, whole brain cluster corrected at α < .05) to assess main and interaction effects for condition and sex on stress ratings during fMRI. We ran a similar model but with stress rating change scores to understand which regions were associated with increasing stress levels across the stress block. Beta-weights from significant clusters were extracted and used to examine whether the neural correlates of in-scanner stress predicted future real-world daily stress ratings, which were collected via smartphone app for 58 days following the scan.

Results: In-scanner stress ratings were significantly higher in the stress condition (b = 4.57, p < .001) compared to neutral and increased across stress runs only (b = 0.43, p = .031). Women reported greater subjective stress than men during the stress condition (b = 1.11, p = .003) and had larger increases in stress ratings across both blocks (b = −0.24, p = .003). Across men and women, hyperactivation of the right anterior prefrontal cortex (PFC) but decreased activation of the left hippocampus, hypothalamus, putamen, thalamus, insula, and mid cingulate cortex during the stress minus neutral condition were associated with greater average stress ratings. This circuitry was also associated with stress change scores, but largely in the opposite direction: greater activation to stress minus neutral images in these limbic-striatal regions but lower ventromedial PFC activation was associated with increasing stress ratings during the stress block. When examining sex differences in neural correlates of overall stress ratings, we found that in women, stress-related activation in the orbitofrontal cortex, amygdala, and rostral anterior cingulate cortex (rACC) was positively associated with stress ratings while right hippocampus was negatively associated with stress ratings. Stress-related activation of the caudate, dorsolateral PFC and rACC was negatively associated with stress ratings in men. Sex differences in the neural correlates of stress change scores were localized to a large dorsal prefrontal cluster including the ACC where blunted activation was associated with increasing stress in women but decreasing stress in men. Several of the neural correlates of in-scanner stress ratings significantly predicted real-world stress consistent with in-scanner results. Individuals with more blunted left putamen activation reported less real-world stress (b = −10.75, p = .012). Right hippocampal (b = 15.78, p = .045) and ACC (b = 28.08, p = .025) activation interacted with sex such that they predicted greater real-world stress for men but less for women.

Conclusions: Overall, blunted cortico-striatal-limbic circuitry to stress exposure was linked with greater average subjective in-scanner stress, especially for men. However, greater cortico-striatal-limbic activation and lower ventromedial PFC activation was associated with increasing stress ratings across the stress exposure. These findings suggest that increasing striatal-limbic activation with lower ventromedial PFC stress responses increases subjective distress and represents reduced stress regulation. Women were more likely to report increasing stress, which was linked to hypoactive ACC activation to stress exposure, compared to men. Taken together, these results indicate both shared and sex-specific neural correlates of subjective stress and dynamic changes in neural stress response, suggestive of stress regulatory responses. The hippocampus and ACC predicted real-world daily stress in a sex-specific manner, suggesting that this network has important and externally valid relevance for sex differences in stress reactivity. Future work is needed to evaluate whether these biomarkers may help predict the development of stress-related psychopathology, especially those with significant sex differences in prevalence.

Keywords: Perceived stress, sex differences, functional magnetic resonance imaging.

Disclosure: Nothing to disclose.

P834. Pfc valence encoding properties and facial expression features predicts anhedonia prior to stress

Karen Safaryan, Polo Morales, Daniel Leal, Madeline Donahue, Millie Auslender, Nicole Clarin, Austin Coley

University of California - Los Angeles, Los Angeles, California, United States

Background: A critical issue within the mental health field is the lack of granularity in diagnostic practices. Many distinct pathologies are currently being diagnosed as the same disease, and prescribed the same treatments – and that the key to developing effective antidepressants that work for everyone, we need to first identify a strategy to differentiate between heterogeneous conditions. Anhedonia, a condition identified in multiple neuropsychiatric disorders, is described as the inability to experience pleasure and is linked to anomalous medial prefrontal cortex (mPFC) activity. The mPFC is responsible for higher order functions, such as valence processing; however, it remains unknown how mPFC valence-specific neuronal population activity is affected during anhedonic conditions in response to acute severe stress.

Methods: To test this, we implemented the Learned Helplessness (LH) protocol in rodents and examined hedonic behaviors following stress and ketamine treatment. We used unsupervised clustering (k-means clustering, k = 4) as a spectrum-based approach to identify helplessness behavior and delineate individual variability in hedonic values in response to stress. We then performed in vivo 2-photon calcium imaging to longitudinally track mPFC valence-specific neuronal population activity and dynamics during Pavlovian conditioning tasks. Also, we utilized behavioral pose-estimation tracking systems to detect facial features during Pavlovian conditioning tasks during anhedonia.

Results: Our findings showed a significant reduction in reward consumption and sociability in LH mice (Pearson Correlation, r = −0.69, p = 0.03). In 2-Photon experiments, we observed reductions in mPFC valence neurons within anhedonic mice, in contrast to hedonic groups (n = 5 mice, N = 366 neurons, Wilcoxon Rank Sum test, p < 0.05). Using dimensionality reduction analysis, we identified a critical time point that shows an imbalance in mPFC dynamics that predicts anhedonia phenotypes, and rebounds following ketamine treatment. We applied a linear classifier to decode susceptibility to stress based on mPFC valence-encoding properties (p < 0.0001).

Conclusions: These results indicate that mPFC valence encoding properties are predictive of anhedonic states. Altogether, these experiments point to the need for increased granularity in the measurement of both behavior and neural activity, as these factors can decode the induction conditions of stress-induced anhedonia.

Keywords: Acute Stress, Medial Prefrontal Cortex (mPFC), valence.

Disclosure: Nothing to disclose.

P835. Prenatal stress reprograms the fetal brain in a sex-specific manner through placental dysfunction

Valentina Zonca, Veronica Begni, Samantha Saleri, Marco Andrea Riva, Annamaria Cattaneo

University of Milano, Italy, Italy

Background: Growing evidence indicates that gestational stress can induce the onset of a pathological phenotype in the mothers as well as disruptions of the normal foetal neurodevelopment. However, the biological mechanisms underlying the close relationship between stress during pregnancy and fetal brain development are still poorly understood. Prenatal stress (PNS) paradigm is a well-established preclinical model used to study the emergence of depressive-like phenotypes in animals. Therefore, the aims of this study was to uncover the biological mechanisms through which stress during pregnancy can disrupt placenta functioning and possibly affecting fetal brain neurodevelopment, with a sex specific effect.

Methods: Pregnant C57BL/6 female mice (12 animals per group) were exposed to PNS protocol consisting of restraint under bright light for 45 minutes, three times daily, from gestational day 12 (GD12) to GD17. Behavioral assessments were conducted at GD17 to evaluate self-care, locomotor activity, and anxiety-like behaviors. On GD18, dams were euthanized, and placenta samples and fetal brains from males and females offspring were collected for comprehensive molecular analyses, including RNA sequencing. Transcript-level raw counts have been quantified, comparing differential expression between groups (DESeq2). Genes differentially expressed at a log2 fold change cutoff and FDR < 0.01 have been entered in IPA Software for pathway analyses.

Results: Exposure to PNS induced a mild pathological phenotype in stressed dams which associated with altered placenta functioning with a sex specific effect, observed in term of RNA-sequencing: 1660 DETs were differentially modulated in placentas of dams carrying male fetuses and 461 DETs in placentas of dams carrying female fetuses, as compared to controls (FC > |1.2|, p. value < 0.05).

Moreover, RNAseq analyses conduced in the foetal brains revealed 494 DETs in male PNS foetal brains with an enrichment in 49 biological pathways, including an overall upregulation of collagen-related pathways, such as Extracellular matrix organization (z-score 3.6), Collagen chain trimerization (z-score 2.6), Collagen degradation (z-score 2.6), and Collagen biosynthesis and modifying enzymes (s-score 2.6). In female PNS foetal brains 255 DETs were observed (FC > |1.2|, p. value < 0.05), which resulted in 22 pathways including a downregulation of the anti-inflammatory IL-4 signaling pathway (z-score −2.4) and VEGF signaling (z-score −2).

Conclusions: Our results shown that male fetuses present a more robust response than females, with transcriptomic shifts pointing to a upregulation of collagen-related pathways, which are crucial for neurodevelopment and structural maturation of the brain. Their dysregulation may indicate premature or maladaptive remodeling in response to prenatal stress. On the other hand, female foetal brains exhibited fewer transcriptional changes, mainly associated with inflammatory-related pathways. The differential responses observed in males and females may rely on distinct coping strategies during foetal development, which could have lasting implications for vulnerability or resilience to neurodevelopmental and psychiatric disorders later in life.

Keywords: prenatal stress, placenta, foetus brain, trascriptomic profile.

Disclosure: Nothing to disclose.

P836. Shared and sex-specific VTA dopaminergic adaptations across stress

Kailyn Price, Diana Black, Hannah Vaidyan, Abigail Polter

George Washington University, Washington, District of Columbia, United States

Background: Dopaminergic neurons of the ventral tegmental area (VTA) are well characterized for their involvement in motivation and reward processing. However, dopaminergic activity is also altered by stress and aversive stimuli. Our recent work (Bouarab et al, European Journal of Neuroscience, 2025) shows that during six days of subchronic stress, male and female mice undergo shared and distinct adaptations to synaptic inputs onto VTA dopaminergic neurons, ultimately leading to a suppression of dopaminergic activity in females but not males. Here, we investigate whether sex differences in behavioral responses or in activation of dopaminergic neurons during stress may contribute to divergent physiological responses in the VTA.

Methods: To measure behavioral changes during stress, we exposed adult male and female mice to a 6-day subchronic variable stress (SCVS) paradigm with one hour-long session of foot shock, tail suspension, or restraint on each day and recorded stress coping during the tail suspension phases of stress (TSS). To measure dopaminergic responses during stress, we injected an AAV encoding the optical dopaminergic biosensor dLight into the lateral nucleus accumbens shell and implanted fiber optic cannulae above the injection site. After allowing for expression, we again exposed mice to SCVS. Dopaminergic signals were recorded via fiber photometry during pre- and post-stress homecage exploration, the TSS stage of stress, and during a post-stress conflict task of palatable food exploration in an open field.

Results: We found that during the TSS, female mice spent less time immobile than male mice during tail suspension sessions (2-way ANOVA, main effect of sex, F (1, 42) = 4.882, p = 0.03, n = 21). Furthermore, in females, but not males, more immobility correlated with less avoidance in post-stress elevated plus maze testing (females: R2 = 0.43, p = 0.002, males: R2 = 0.14, p = 0.103). Fiber photometry during SCVS TSS sessions showed dopaminergic dynamics over stress, with increases in dopamine levels at the onset of immobility and decreases at the onset of movement. Our preliminary data suggests that these signals are qualitatively similar but may be of larger amplitude in females.

Conclusions: These studies show that male and female mice adopt distinctive coping behavior during stress and dopaminergic dynamics during stress evolve over time and may differ between males and females. Future work will investigate changes in dopaminergic responses in a conflict test following SCVS and pre- and post-stress dopaminergic function.

Keywords: Ventral Tegmental Area (VTA), Dopamine, nucleus accumbens.

Disclosure: Nothing to disclose.

P837. Acute exercise buffers stress-induced metaplasticity in CRH-PVN neurons and mitigates stress-induced defensiveness

Mijail Rojas-Carvajal, Tamas Füzesi, Dinara Baimoukhametova, Nuria Daviu, Sarah Cook, Matthew Hill, Jaideep Bains

Hotchkiss Brain Institute, University of Calgary, Calgary, Canada

Background: Stress imprints biochemical, molecular, and synaptic changes in the brain to promote adaptation. However, these changes can become maladaptive and foster neuropsychiatric diseases. Surprisingly, there is limited understanding on how these imprints can be reversed. In humans, exercise is used to cope with stress despite inducing physiological stress itself. Here we examined the effects of exercise on stress-induced short-term potentiation (STP) of glutamate synapses on corticotropin release hormone cells in the paraventricular nucleus of the hypothalamus (CRH-PVN).

Methods: We leveraged synaptic and behavioral readouts of stress sensitization to understand how exercise modulates the lasting consequences of acute stress. Particularly, we used a combination of ex vivo electrophysiology, genetic, optogenetic, and behavioral approaches.

Results: Here we show that exercise immediately after stress engages synergistic local signaling on CRH-PVN neurons to reverse the behavioral and synaptic consequences of stress. The combination of glucocorticoid increase resulting from stress and exercise-induced brain-derived neurotrophic factor (BDNF) release decreases stress-induced behavioral threat sensitization and synaptic metaplasticity. In the absence of exercise, tropomyosin-related kinase B (TrkB) receptor activation following stress is sufficient to reverse stress-induced threat sensitization and synaptic metaplasticity. We also identified a functional critical period immediately after stress that may provide a therapeutic window for the effects of exercise.

Conclusions: These findings provide the first mechanistic insights into how exercise alleviates the lasting effects of stress and reconcile previously paradoxical observations regarding the timing and efficacy of exercise as a stress intervention.

Keywords: Acute Stress, exercise, paraventricular nucleus of the hypothalamus, sensitization, BDNF.

Disclosure: Nothing to disclose.

P838. Hypothalamic recurrent inhibition regulates functional states of stress effector neurons

Wataru Inoue, Aoi Ichiyama, Samuel Mestern, Lyle Muller

University of Western Ontario, London, Canada

Background: Stress triggers rapid and reversible shifts in vital physiological functions from homeostatic operation to emergency response. However, the neural mechanisms regulating such functional stress states remain poorly understood. Here we identify a novel recurrent inhibitory circuit governing functional states of key stress regulatory neurons: corticotropin-releasing hormone (CRH) neurons in the hypothalamic paraventricular nucleus (PVN).

Methods: We recently reported in vivo spiking activities of opogenetically-identified CRHPVN neurons in mice, and that computational modeling of spiking dynamics revealed that CRH-GABA recurrent inhibition is required for stress-induced functional state transitions. Guided by model predictions, here we experimentally interrogated the recurrent inhibitory circuits using in vivo and ex vivo electrophysiology combined with AAV-mediated circuit mapping in male and female mice.

Results: We demonstrate that PVN-projecting GABAergic neurons constrain CRHPVN neurons to low-activity state at baselines via a prolonged feedback inhibition. Unexpectedly, slow CRHergic excitation was dispensable for driving this prolonged feedback, whereas glutamatergic transmission predominated at CRH→GABA excitatory synapses. Incorporating these findings, we refined our computational model to include fast excitation and slow inhibition, yielding new predictions for circuit operation.

Conclusions: Together, our results establish recurrent inhibition as a fundamental circuit motif controlling CRHPVN neurons functional states and highlight the value of iterative experiment–model integration in advancing understanding of neural circuits functions.

Keywords: paraventricular nucleus of the hypothalamus, latent states, Corticosterone response to stress, Corticotropin-Releasing Hormone.

Disclosure: Nothing to disclose.

P839. An amygdala to anterior hypothalamic circuit gates negative valence and stress vulnerability

Zachary Pennington, Alexa LaBanca, Shereen Abdel-Raheim, Madeline Bacon, Afra Mahmud, Patlapa Sompolpong, Austin Baggetta, Yosif Zaki, Bumjin Ko, Yu Feng, Zhe Dong, 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 the brain’s stress circuits, how these circuits are modified by adverse experience to increase stress vulnerability remains unknown. Here, using an unbiased whole-brain screen, we identify a novel hypothalamic locus whose activity and functional connectivity is shaped by prior adversity to promote stress vulnerability.

Methods: Immediate early gene imaging of cleared, intact, mouse brains was first conducted to identify how prior adversity (10 footshocks) influences brain-wide activity levels in response to a subsequent stressful experience (an auditory startle stimulus 10 days later). After identifying the anterior hypothalamic nucleus (AHN) as a possible regulator of stress vulnerability, in vivo calcium imaging with Miniscopes was performed to assess how AHN neurons dynamically respond to somatosensory as well as auditory stressors. To determine the AHN’s causal role in stress vulnerability, we optogenetically silenced AHN neurons during multiple stress and anxiety-related behavior tests. Chemogenetic activation of AHN neurons was also performed to test if increasing AHN activity is sufficient to augment defensive behavioral responding to stress. Lastly, projection-specific chemogenetic and optogenetic inhibition of inputs to the AHN from the amygdala, lateral septum, and ventral hippocampus were 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 models were used to assess all behavior.

Results: Harnessing unbiased brain-wide activity-mapping in mice, we found that the AHN – a region that to date has received little attention – displayed potentiated responses to stressors in previously stressed mice (FDR controlled). Moreover, the AHN showed increased correlational strength with a limbic stress network in previously stressed mice (p < 0.001). Using in vivo calcium imaging, we then found that AHN activity linearly tracked stress severity, and the proportion of stress-reactive AHN neurons was increased by prior stress. Addressing the causal role of the AHN in stress vulnerability, optogenetic inhibition of AHN neurons blunted the ability of prior stress to increase stress sensitivity. Conversely, chemogenetic excitation of AHN neurons was sufficient to facilitate defensive behavioral responses to stress. Lastly, projection-specific chemogenetic and optogenetic inhibition of inputs to the AHN from the amygdala reduced responding to stressful events in previously stressed mice.

Conclusions: Collectively, this work highlights the AHN as a novel regulator of negative valence and demonstrates that its modification by prior adversity promotes stress vulnerability. Given the AHN’s rich cellular diversity, it represents a compelling potential target for pharmacological intervention.

Keywords: stress vulnerability, hypothalamus, valence, amygdala, stress.

Disclosure: Nothing to disclose.

P840. Social experience in adolescence shapes hippocampo-orbitofrontal interactions in adulthood

Michelle Sequeira, Karinne Cobb, Kathryn Stachowicz, Trevor Towner, Dan Li, Shannon Gourley

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

Background: Adversities like social isolation can disrupt the ability of organisms to integrate new learning into established knowledge later in life, causing a deferral to habit-like behavior. The orbitofrontal cortex (OFC) supports action flexibility in changing environments, while the ventral hippocampus (vHC) is involved in habit-like behavior. Both structures are impacted by stressors, including during sensitive developmental periods. Here, we investigate whether social isolation in adolescence impacts vHC-OFC interactions to induce habit-like response biases.

Methods: Female mice were socially isolated in adolescence and then reintegrated into social groups as young adults. We then tested mice in a task in which they must flexibly deviate from familiar action strategies to obtain rewards. We used chemogenetics, high-resolution cellular imaging, and manipulation of activity-defined neuron populations (identified by the immediate-early gene cFos) to determine whether and how social isolation alters decision-making strategies.

Results: A history of social isolation impeded flexible responding and increased task-related cFos in the vHC. Chemogenetically silencing cFos-expressing cells restored action flexibility, as did chemogenetically suppressing vHC-OFC connections in isolated mice. Meanwhile, stimulating vHC-OFC projections induced habit-like response biases, mimicking the effect of social isolation. Suppressing neural activity in the vHC facilitated learning-related dendritic spine turnover on task-relevant neurons in the OFC. Moderating neurotrophin signaling – which otherwise stabilizes dendritic spines – selectively in these neurons reinstated action flexibility following isolation.

Conclusions: Our findings indicate that vHC inputs to the OFC control action flexibility, with their activity overriding flexible choice in favor of habit-like behavior. Projection activity impacts learning-related dendritic spine plasticity on task-relevant neurons in the OFC, and strategies aimed at disinhibiting this plasticity could combat the long-term consequences of social isolation on flexible choice behavior.

Keywords: Adolescent social isolation, BDNF, Dendritic Spine, orbital frontal cortex, Hippocampus.

Disclosure: Nothing to disclose.

P841. Female California mice (peromyscus californicus) exhibit sustained reduced synaptosome mitochondria respiration in response to social isolation

Amy Wegener, Hannah Fulenwider, Hannah Stadtler, Paul Howell, Nour Abosamak, Susie Turkson, Charlotte Ream, Niyomi Shah, Erica Glasper, Gretchen Neigh

Virginia Commonwealth University, Richmond, Virginia, United States

Background: Social isolation, loneliness, and loss profoundly impact health and quality of life, but the complex nature of these constructs can prohibit mechanistic examination in rodent models. Assessment of the peripheral and neural consequences of social manipulations in a social rodent species, such as California mice (Peromyscus californicus), can facilitate inquiry into the biological repercussions of experiencing different types of loss. Social stressors can compromise neural function in the hippocampus, which can be partially attributed to shifts in mitochondrial function at the synapse. Synaptic mitochondria are essential for facilitating neuronal communication and compromise of mitochondrial respiration capacity can impair neurotransmission and setting the framework for behavioral dysfunction.

Methods: Genetically monogamous male and female California mice were used to evaluate synaptic mitochondrial respiration following removal of a same-sex cage mate for 10- or 30-days (i.e., social isolation). Synaptosome mitochondria were isolated from the hippocampus. Agilent’s Cell Mitochondrial Stress Test was utilized to assess electron transport chain function with oxygen consumption rate (OCR) as the endpoint. OCR outputs were normalized to the protein quantity per well. All data sets were analyzed separately using GraphPad Prism v10.2.3 and an α = 0.05. Data sets were tested for normality using the Shapiro-Wilks test and analyzed using a non-parametric equivalent statistical test if normality test was not passed.

Results: Following social isolation, males, but not females, exhibited physical changes consistent with a prolonged stress response, including reduction in body mass and hypertrophy of the adrenal glands following isolation. There was no effect of social isolation on hippocampal synaptic mitochondria within male California mice (p > 0.05); however, isolated female California mice, compared to paired female California mice, exhibited reduced hippocampal synaptic mitochondrial function in terms of lower basal respiration (F(2,22) = 6.63, p = 0.0056), lower maximal respiration (F(2,22) = 6.92, p = 0.0047), lower spare capacity (F(2,22) = 6.62, p = 0.0056), and lower proton leak (F(2,22) = 7.15, p = 0.0040). Tukey’s multiple comparisons tests indicated that these reductions in mitochondrial endpoints were present both 10 and 30 days after separation. In a follow-up cohort to explore 60-day separation outcomes in females we observed that mitochondrial respiration remained lower compared to paired females for basal respiration (p = 0.0069), proton leak (p = 0.0037), and spare capacity (p = 0.0181).

Conclusions: Collectively, these data suggest that males and females differentially manifest the impact of social stress. While males exhibit peripheral effects of isolation, females demonstrate alterations in synaptic mitochondria respiration in the absence of observable peripheral effects of isolation. These data provide mechanistic insight to our understanding of functional changes within the hippocampus, an important mediator of cognition, neuroinflammation, and stress. Future work should investigate behavioral assays to identify cognitive deficits reflective of altered hippocampal function, as well as behavioral measurements of anhedonia following social isolation.

Keywords: Mitochondria, isolation, California mice, Hippocampus, social loss.

Disclosure: Nothing to disclose.

P842. Stress-enhanced fear learning is influenced by nature of stressor

Robert Aukema, Lauren Seabrook, Paloma Martinez, Olga Ponomareva, Esha Kelkar, Bettina Ventura, Claudia Klengel, Zoe Beatty, Jwo Vy Tseng, Jakob Hartmann, William Carlezon, Kerry Ressler

McLean Hospital, Harvard Medical School, Belmont, Massachusetts, United States

Background: A major risk for development of post-traumatic stress disorder (PTSD) is exposure to multiple traumas experienced over time; likewise, a major symptom of PTSD is an enhanced response to seemingly mild stressors. This is modeled in rodents, where exposure to intense stress leads to inappropriately strong fear conditioning to relatively mild stressors and increased avoidance behaviour even weeks later. The basolateral amygdala (BLA) is strongly implicated in both stress reactivity and fear learning, however mechanisms within the BLA linking prior stress with fear sensitization are unknown. Further, it is unknown whether specific type of acute stressor differentially impacts reactivity to future stressors. We therefore sought to address 3 specific questions: (1) Does stressor type influence sensitization of new fear learning? (2) Do stress-activated BLA neurons get re-activated during recall of a new fear memory? (3) Does inhibiting stress-activated BLA neurons alter reactivity to new stressors?

Methods: EXPERIMENT 1. We first determined whether history of acute stress impacts new fear learning in adult male mice, and if this is influenced by type of stressor. Mice were exposed to one of four conditions (n = 14–16 / group): familiar environment (no stress control), ten unsignalled shocks, one hour restraint stress, or subthreshold social defeat. 2 weeks later, mice underwent single shock-paired contextual fear conditioning in a new context and were assessed fear recall 24 hours later. In a separate cohort of mice, plasma corticosterone was measured at the termination of each stressor (n = 6–8 / group).

EXPERIMENT 2. We next assessed whether the same BLA neurons activated by an acute stressor are re-activated during recall of a new, fundamentally different fear memory. Adult male and female mice underwent acute shock or restraint stress, or no stress controls (n = 6–16 / group), and activity-dependent genetic tools (TRAP2 reporter mice) were used to label BLA neurons with tdTomato that were active selectively within this time window. 2 weeks later, mice underwent single shock-paired contextual fear conditioning in a contextually distinct environment and were assessed for fear recall 24 hours later. Tissue was collected 90 min following recall of new fear and immunohistochemistry for FOS was used to label cells active during recall of this new fear memory. Colocalization of tdTomato (initial stress) with FOS (recall of new fear) would indicate activation of this cell by both events.

EXPERIMENT 3. To assess the functional role of shock-activated BLA neurons, we used TRAP2 adult male and female mice in combination with chemogenetic tools to selectively express the inhibitory receptor hM4Di or control mCherry selectively in shock-activated BLA neurons (n = 12–17 / group). Four weeks later, animals were administered DCZ (to inhibit hM4Di+ cells without effect on mCherry controls) prior to testing and assessed for exploration in the elevated plus maze (EPM), recall of a new fear memory following fear conditioning in a new context, and acoustic startle response.

Results: IN EXPERIMENT 1, the type of acute stressor differentially enhanced new fear learning, as animals with a history of either shock (p < 0.001) or social defeat stress (p < 0.05) – but not restraint stress – froze significantly more during recall of new fear than no stress controls. Paradoxically, this was dissociable from the initial neuroendocrine response to stress, as in a separate group of animals, restraint and social defeat stress drove a significantly greater neuroendocrine response to stress than to either shock or the no stress condition (p < 0.05).

IN EXPERIMENT 2, BLA neurons activated by recall of a new, mild fear memory were significantly more likely to have been previously activated by prior shock (p < 0.05) or restraint stress (p < 0.05) than by a neutral, non-stressful context, as indicated by greater co-expression of FOS and tdTomato.

IN EXPERIMENT 3, inhibition of shock-activated BLA neurons modulated reactivity in various behavioral tests. Compared to mCherry controls, selective inhibition of shock-activated BLA neurons prior to each test significantly increased distance travelled in the EPM (p < 0.01), reduced recall of new fear (p < 0.01), and reduced startle amplitude in the acoustic startle test (p < 0.05).

Conclusions: Our findings suggest that type of stressor differentially sensitizes new fear learning, an effect that is dissociable from the initial corticosterone response to stress. Additionally, a subset of stress-activated BLA neurons are reactivated during recall of a fundamentally new stressor, suggesting that a subpopulation of BLA neurons are broadly reactive to diverse stressors. Functionally, inhibition of shock-activated BLA neurons reduced reactivity in several behavioral tests. Ongoing work is determining the molecular identity of trauma-activated BLA neurons, as this may provide insight into pharmacologically targetable mechanisms for reducing stress reactivity.

Keywords: basolateral amygdala, Fear learning, Acute Stress, TRAP2 mice.

Disclosure: Nothing to disclose.

P843. Stress-induced synaptic modifications in the NAc-VTA circuit underlie vulnerability to stress

Jaehan Kwon, Urszula Skupio, Ava Hargrave, Alexander Harris

Columbia University, New York State Psychiatric Institute, New York, New York, United States

Background: Chronic stress is a well-established risk factor for reward-processing deficits, core symptoms of psychiatric disorders such as depression. The ventral tegmental area (VTA) and nucleus accumbens (NAc) form a key circuit regulating reward processing. However, it remains unclear how stress reshapes this circuit and leads to persistent reward dysfunction. We previously found that stress induces 4 Hz NAc oscillations that entrain medium spiny neurons (MSNs), which inhibit VTA GABAergic neurons and disinhibit dopaminergic neurons, raising the possibility that synchronized presynaptic activity drives NAc–VTA synaptic plasticity. Our study investigated how stress alters NAc-VTA circuit activity, focusing on neuronal firing patterns and stress-induced synaptic plasticity, and how these changes determine individual differences in susceptibility or resilience to chronic stress.

Methods: C57BL/6 mice (10 weeks old, 9 males, 6 females) were trained in a Pavlovian cued-reward task and were exposed to two weeks of chronic restraint stress (CRS). Post-consumption behavior was quantified by sustained licking duration. Based on behavioral outcomes, mice were classified into susceptible and resilient groups using k-means clustering. To assess stress effect on neuronal activity, in vivo single-unit recording was performed in the VTA during three phases: pre-restraint, restraint, and post-restraint. Distinct firing patterns were identified via long short-term memory machine learning (LSTM). To examine stress-induced alterations in NAc MSN-VTA GABAergic neuron synapses, ChR2 was injected into the NAc of VGAT-Cre mice and electrodes and optical fibers were implanted in the VTA. Light-evoked multi-unit activity (MUA) was recorded responding to 4 Hz light stimulation mimicking stress-induced presynaptic synchronization.

Results: K-means clustering revealed two behavioral outcomes to CRS: a susceptible group with attenuated post-consumption responses (5 males, 5 females; p < 0.001), and a resilient group with stable or increased post-consumption responses (4 males, 1 female; p < 0.05). Single-unit recordings revealed 6 distinct VTA firing patterns during stress (n = 168). Ongoing analysis will test whether these firing states predict behavioral outcomes. 4-Hz optogenetic stimulation of MSN terminals induced a reduction of light-evoked MUA in the VTA (n = 6; p < 0.01).

Conclusions: These findings indicate that stress-induced NAc–VTA synaptic modifications reshape VTA firing states, which may underlie individual differences in stress vulnerability.

Keywords: Acute and Chronic Stress, in vivo electrophysiology, limbic system.

Disclosure: Nothing to disclose.

P844. Adolescent stress-induced MMP-9 dysregulation drives parvalbumin interneuron deficits and behavioral abnormalities

Emilly Andres, Francisco Guimaraes, Felipe Gomes

University of Sao Paulo, Ribeirao Preto, Brazil

Background: Adolescence is a critical period of neurodevelopment during which exposure to stress is a major risk factor for the emergence of psychiatric disorders, such as schizophrenia. Using rodent models, we demonstrate that stress during adolescence—but not in adulthood—leads to long-lasting behavioral, molecular, and electrophysiological abnormalities associated with schizophrenia. These include increased anxiety-like behavior, impaired sociability and cognition, heightened responsiveness to psychostimulants, and elevated activity of dopamine neurons in the ventral tegmental area (VTA). These changes are strongly associated with deficits in GABAergic interneurons expressing parvalbumin (PVIs) and their associated perineuronal nets (PNNs) in the ventral hippocampus (vHipp). Multiple mechanisms have been proposed to contribute to PVI/PNN deficits, including oxidative stress and increased activity of matrix metalloproteinases (MMPs), particularly MMP-9. Here, we present findings on the impact of adolescent stress on oxidative stress markers—especially within PVIs—and on MMP-9 activity across development, highlighting the increased MMP-9 activity induced by adolescent stress. Furthermore, we show that pharmacological interventions targeting oxidative stress and MMP-9, as well as viral-mediated knockdown of MMP-9 in the vHipp, successfully prevent stress-induced PVI/PNN deficits and associated behavioral abnormalities.

Methods: MMP-9 activity in the vHipp of male Sprague Dawley rats was assessed by gel zymography at postnatal days (PND) 31, 41, 51, 75, 85, and 110 (n = 6/group). To examine the effects of stress on MMP-9 activity, animals were subjected to a stress protocol starting on PND 31, consisting of daily sessions of 25 foot shocks (1 mA, 1 s) for 10 days, combined with 1-hour restraint sessions on PND 31, 32, and 40. Stress-induced changes in MMP-9 activity were assessed on PND 41, 51, and 75 (n = 6/group). To evaluate the effects of the pan-MMP inhibitor doxycycline on behavioral and molecular outcomes, both treatments were initiated concurrently with the stress protocol (PND 31–40). Doxycycline (2 mg/mL) were administered via drinking water. We also investigated whether viral-mediated knockdown of MMP-9 in the vHipp could prevent the deleterious effects of adolescent stress. In adulthood, animals underwent behavioral testing, including the light-dark box (LDB), elevated plus maze (EPM), social interaction (SI), and novel object recognition (NOR) tests (n = 7–8/group).

Results: MMP-9 activity increased significantly from PND 31 to later time points (PND 41: p = 0.03; PND 51: p < 0.0001; PND 75: p = 0.0004; PND 85: p < 0.0001; PND 110: p = 0.002). Adolescent stress increased MMP-9 activity at PND 45 and 51 (p = 0.04 and p = 0.03, respectively) and decreased the expression of its endogenous inhibitor, tissue inhibitor of metalloproteinases-1 (TIMP-1; p = 0.04). Behaviorally, stressed animals showed reduced social interaction and a lower discrimination index in the NOR test, indicating deficits in sociability and cognition. There was a significant reduction in the number of PV+ cells (F(1,20) = 18.82, p = 0.0003) and PV+/PNN+ cells (F(1,20) = 13.82, p = 0.001) in the vHipp of adult rats subjected to adolescent stress. These behavioral and molecular alterations were prevented by treatment with doxycycline during the stress protocol, as well as by viral-mediated MMP-9 knockdown in the vHipp prior to stress exposure. Additionally, doxycycline attenuated the stress-induced increase in MMP-9 activity.

Conclusions: Our results reveal dynamic changes in MMP-9 activity in the vHipp during development. Adolescent stress increases MMP-9 activity and reduces the number of PV+ and PV+/PNN+ cells in the vHipp, while also producing lasting behavioral impairments. Importantly, interventions with doxycycline or local MMP-9 knockdown conferred protection against these effects. Although further studies are needed to elucidate the precise mechanisms linking MMP-9 dysregulation to the effects of adolescent stress, our findings suggest that it represent a promising target for the prevention of stress-related neurodevelopmental abnormalities. Financial support: São Paulo Research Foundation (FAPESP).

Keywords: adolescent stress, parvalbumin, Matrix Metalloproteinase-9 (MMP-9).

Disclosure: Nothing to disclose.

P845. Decoding stress and the gut microbiome: evidence from mouse models of chronic stress

Larissa Dillmann, Shanlin Ke, Fiona Seibert, Yuan Li, Claudia Klengel, Kerry Ressler, Karestan Koenen, Xuesong He, Yang-Yu Liu, Jakob Hartmann

McLean Hospital, Harvard Medical School, Belmont, Massachusetts, United States

Background: The gut microbiome plays a central role in maintaining physiological balance and contributes to various pathophysiological processes. Studies suggest it is integral to stress responsiveness and affective disorders, including anxiety and major depressive disorder (MDD). Despite growing evidence of gut microbial associations with mental health, the mechanisms connecting the microbiome to stress-related psychiatric disorders remain unclear. Research further indicates that diets targeting specific microbiota or bacterial strains may positively influence mood, anxiety, cognition, stress, and hypothalamic-pituitary-adrenal (HPA) axis function. Consequently, microbiome-based therapeutics hold significant promise for treating stress-related psychiatric disorders. To better understand how stress shapes the gut-brain axis, we investigated the impact of chronic stress on the gut microbiome and brain using mouse models of chronic social defeat stress (CSDS).

Methods: Male and female mice were exposed to CSDS for 21 days. A home cage social interaction test was conducted on the final day of the paradigm, 24 hours after the last defeat to assess stress-related behaviors using machine-learning based analysis (DeepOF). Stool samples were collected on days 1 (pre-defeat), 15, 21, and 22 and underwent whole metagenome sequencing to assess microbial changes.

Results: Stressed male and female mice exhibited increased avoidant and defensive behaviors alongside reduced exploratory activity compared to their respective control group (unpaired t-test; males: B6_stim_following T(36) = 4.869, p < 0.0001, B6_sim_nose2tail T(36) = 4.12, p < 0.001; females: B6_sniffing T(30) = 5.711, p < 0.0001, B6_stim_nose2body T(30) = 2.616, p < 0.05). While CSDS did not affect within-group species richness (alpha diversity), it significantly altered between-group gut microbial composition (beta diversity; PERMANOVA, p = 0.0001), with these changes intensifying over time. Differential abundance analysis revealed microbial species linked to gut barrier integrity, a key factor in maintaining immune and metabolic homeostasis, were altered in the CSDS group compared to controls.

Conclusions: These findings highlight stress-induced microbiome remodeling as a potential driver of psychiatric vulnerability. Ongoing work will test whether targeted microbiome interventions can enhance resilience, laying the groundwork for microbiome-based therapeutics in stress-related disorders.

Keywords: Gut-Brain Axis, microbiome, chronic social defeat stress, stress resilience.

Disclosure: Nothing to disclose.

P846. Stress-induced alterations in synaptic metabolism and their behavioral relevance

Hirotaka Nagai, Kohei Ota, Chisato Numa, Midori Nagai, Wenran Qiu, Masayuki Taniguchi, Tomomi Yamashita, Kotaro Mizuta, Yusuke Kawashima, Nobuhiko Ohno, Yosky Kataoka, Shuichi Shimma, Yuko Mimori-Kiyosue, Taro Kato, Yasunori Hayashi, Tomoyoshi Soga, Tomoyuki Furuyashiki

Institute of Science Tokyo, Tokyo, Japan

Background: Depression and other stress-related mental disorders are highly prevalent worldwide. Chronic social stress is a major environmental risk factor for their onset and progression. Previous studies indicate that chronic stress can induce dendritic atrophy and synaptic alterations in higher-order brain regions such as the prefrontal cortex, with mitochondrial structural and functional abnormalities implicated in their pathophysiology. However, the mechanistic links between stress-induced mitochondrial dysfunction, synaptic structural remodeling, and underlying metabolic disruption remain poorly understood. Here we investigated how chronic social stress alters metabolic processes at prefrontal cortical synapses, and how these changes relate to neuronal structure, function, and depression-related behaviors.

Methods: Male C57BL/6 mice were exposed to acute or chronic social defeat stress (SDS). Following chronic SDS, mice were classified as susceptible or resilient based on social avoidance. Dendritic morphology was assessed by high-resolution confocal imaging with three-dimensional reconstruction. Synaptic ultrastructure and mitochondrial morphology were examined using three-dimensional electron microscopy and expansion microscopy. Synaptic mitochondrial function was evaluated via biochemical assays and respirometry. Proteomic and metabolomic profiling were performed using mass spectrometry, and spatial metabolic changes were visualized by mass spectrometry imaging. Neuronal activity was recorded using in vivo two-photon Ca²⁺ imaging. Molecular interventions targeting central metabolism in prefrontal neurons were applied to assess their effects on synaptic integrity, neuronal activity, and behavior.

Results: Chronic social stress induced dendritic atrophy, altered mitochondrial morphology, and caused shrinkage specifically in mitochondria-containing synapses. Synaptic mitochondrial function was impaired, accompanied by significant alterations in proteins and metabolites linked to central metabolism and mitochondrial function. Mass spectrometry imaging revealed brain region-specific metabolic changes, notably in the medial prefrontal cortex. Targeted modulation of central metabolism ameliorated stress-induced synaptic, neuronal, and behavioral deficits.

Conclusions: Our findings indicate that chronic social stress disrupts central metabolic processes at prefrontal cortical synapses, driving structural and functional neuronal changes that contribute to depression-related behaviors. Modulating these metabolic pathways may offer novel therapeutic strategies for stress-related mental disorders.

Keywords: Chronic social stress, Depression, Synapse, Glycolysis, Mitochondria.

Disclosure: Nothing to disclose.

P847. Region- and cell-type-specific alterations in gabaa receptor α-subunit expression following chronic stress in mice

Thomas Prevot, Cassandra Marceau-Linhares, James Chiella, Michael Marcotte, Etienne Sibille, Carla Mezo-Gonzalez

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

Background: Chronic stress is a risk factor for anxiety disorders and depression. These conditions are marked by dysregulated GABAergic (primary inhibitory neurotransmitter) signaling, particularly in the prefrontal cortex (PFC), hippocampus (HPC), and hypothalamus (HPT). Therapeutic interventions target the GABAergic receptors but seem to lose efficacy in depressed patients, potentially due to changes in GABAA receptor expression in key regions affected by depression. Here, we examined the region- and cell-specific expression of α1, 2, 3, 4, and 5-containing GABAA receptors in a mouse model of chronic stress. We predicted changes in α subunit expression between chronic stress and control groups, potentially marked by decreased α2 and α5 expression and increased α1 levels across brain regions.

Methods: C57BL/6J mice (n = 64, 2 months old, 50% female) were divided into unpredictable chronic mild stress (UCMS) and control cohorts. UCMS was applied daily for 6 weeks, and anxiety-like phenotype and cognitive performances were assessed. The Phenotyper test was used to measure anxiety phenotype weekly, by measuring avoidance of a light stressor in an automatized, overnight test. Working memory was assessed in the Y maze alternation task. Then blood and brains were collected. Blood was processed to measure corticosterone levels, and brain tissue was processed through RNAscope - an RNA fluorescent in-situ hybridization assay - on PFC, HPT, and HPC sections collected from n = 4 per group, using probes for SST, CRH, and GABRA1, 2, 3, 4, and 5 (mRNAs coding for the α1-5 subunits of the GABAA receptor). Slides were imaged and GABRA1, 2, 3, 4, and 5 puncta found inside CRH+, SST+ or SCL17a7 (glutamatergic) cells were quantified.

Results: UCMS induced gradual anxiety-like phenotype and working memory deficits, associated with elevated corticosterone levels. In the CRH+ cell of the paraventricular nucleus of the HPT, there was a significant decrease in GABRA2 (p = 0.0027) and GABRA4 (p = 0.0044) puncta and a significant increase in GABRA3 (p = 0.0084) puncta per cell. In the SST+ cell of the PFC, there was a significant increase in GABRA2 puncta (p < 0.0001). In the SST+ cell of the HPC, there was a significant decrease in GABRA2 puncta (p < 0.0001). In the SCL17a7+ neurons of the PFC and HPC, there was a significant decrease in GABRA5 puncta (p < 0.001).

Conclusions: GABRA1, 2, 3, 4, and 5 mRNA expression profiles are affected by chronic stress suggesting region- and cell-specific changes to GABAergic signaling caused by stress, potentially explaining pathology and response to GABA-targeting treatments.

Keywords: GABA, Chronic variable stress, Mouse models, prefrontal cortex, RNAscope fluorescence in situ hybridization.

Disclosure: DAMONA Pharma, Consultant, Self.

P848. A preliminary latent class analysis to examine within-individual variation in disinhibition among native american adults: associations with suicide ideation

Andrea Wiglesworth, Wesley Vaught, Rayus Kuplicki, Xi Ren, Martin Paulus, Robin Aupperle, Evan White

Laureate Institute for Brain Research, Tulsa, Oklahoma, United States

Background: Suicide is a leading cause of death among individuals in the United States that disproportionately impacts Native American communities. While these risk patterns are well-established, we have insufficient research that examines heterogeneity within samples of Native American individuals to identify particular patterns of suicide risk and resilience. One way to highlight within-group heterogeneity is to use “person-centered” analytical approaches such as profile or cluster analysis when characterizing risk or protective factors. Person-centered approaches aim to identify how characteristics vary within individuals and relate these patterns (i.e., profiles, clusters) to outcomes of interest. Such approaches are particularly helpful when meaningful within-individual patterns of a construct are obscuring group-level effects.

Latent profile analysis (LPA) has been used in previous work to characterize indicators of disinhibition and associations with lifetime suicide ideation among Native American youth. Wiglesworth and colleagues (2024) used LPA to identify five profiles based on seven indicators of disinhibition measured across levels of analysis (i.e., self-report, behavior, brain). In this study, youth with an elevated self-reported negative and positive urgency profile were at higher odds of reporting suicide ideation than “normative” youth. While this study was the first to examine a person-centered approach to modeling disinhibition and examining associations with suicide ideation among Native American youth, additional research has provided evidence that disinhibition may be of interest when studying suicide ideation among Native American adults.

The current study extends upon past research by addressing two primary research questions: 1) What is the number and nature of profiles that best fit the data when using a person-centered analytical approach to examine within-individual variability in multi-level indicators of inhibition? 2) Do resultant profiles differ in rates of suicide ideation and related clinical correlates?

Methods: Data for this study come from the “T-1000” study, which was designed to examine biological and objective behavioral measures to improve assessment and treatment of psychiatric patients in Tulsa, Oklahoma. Our sample includes 91 individuals (ages 18–55) from the first 500 participants in the T-1000 study who self-identified as Native American. Participants underwent clinical interviews and self-report surveys as well as behavioral tasks that were completed during simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) scanning. Suicide ideation was assessed via interview with research assistants. Indicators of inhibition for the current study were informed by past research among Native American youth and adults. These indicators included self-report on three subscales of the UPPS-P Impulsive Behavior Scale, task performance (reaction time and percent correct inhibition responses) on the Stop Signal Task (SST), a behavioral paradigm of inhibition, select event related potentials (ERPs) during contrasts of interest of the SST, and activation in the right inferior frontal gyrus (rIFG) during contrasts of interest of the SST.

In total, we included 8 indicators (three self-report, two behavior, two ERPs, one fMRI) in a LPA conducted in MPLUS, estimated using full-information maximum likelihood (FIML). Indicators were winsorized and z-scored before running the model. After conducting the LPA, we selected the solution that best fit our data based on the Log Likelihood, AIC, BIC, entropy, and VLMR T11 values. We then used chi-square and Kruskal-Wallis tests to examine differences between profiles in reports of suicide ideation and related clinical characteristics.

Results: Fit statistics and theoretical guidelines pointed toward a three-class solution as providing the best fit for our data. Classes 1–3 represented 19.8%, 47.3%, and 33.0% of our sample, respectively. The largest class, Class 2, showed distinct patterns of poor behavioral performance and atypical neurobiological responding to the SST across multi-modal disinhibition indicators as well as the highest reports of suicide ideation across all ideation subtypes (X2 range = [4.7, 5.4], p’s < 0.10) and the highest severity of reported depressive symptoms (K-W X2 = 6.7, p = 0.04). Class 3, which showed distinct patterns of high self-reported impulsivity along with the highest scores of childhood trauma (K-W X2 = 5.2, p = 0.07), reported having a suicide plan with intent at the highest rates across classes, though this difference was not statistically significant.

Conclusions: By using a person-centered profiling approach, we identified patterns of within-group heterogeneity in disinhibition that may differentially impact risk for suicidal thoughts and behaviors among Native American adults. While Classes 2 and 3 both demonstrated increased risk for suicide ideation, their distinguishing characteristics differed, perhaps indicating that the risk pathways, and thus best intervention or treatment approaches, for these groups may also differ. Leveraging multi-level and multi-modal research designs to identify risk patterns is compatible with movement toward precision medicine approaches in the field and is a strength of this study. However, this project is limited by the sample size, highlighting the exploratory nature of this work.

Keywords: Suicide ideation, disinhibition, Native American populations, latent profile analysis, Suicide risk.

Disclosure: Nothing to disclose.

P849. Low dose buprenorphine enhances the antisuicidal effects of ketamine in major depressive disorder: a randomized controlled trial

Igor D. Bandeira, Jason M. Tucciarone, Christine Blasey, Ian H. Kratter, Jarrod Ehrie, Jennifer Keller, Heather Pankow, Maureen Chang, Jessica Hawkins, Audrey G. Evers, Rebecca Bernert, Charles DeBattista, Henry Truong, Carolyn I. Rodriguez, Boris D. Heifets, Alan F. Schatzberg

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

Background: Currently, no pharmacological strategy is established to sustain ketamine’s antisuicidal effects. Evidence from preclinical and clinical studies suggests that mu opioid receptor (MOR) activity contributes to ketamine’s antidepressant and antisuicidal properties. We evaluated whether low-dose sublingual buprenorphine, a partial MOR agonist, could safely extend the benefits of intravenous (IV) ketamine.

Methods: In a single-site, randomized, double-blind, placebo-controlled trial in the United States, adults with major depressive disorder (MDD) and a Scale for Suicidal Ideation (SSI) score ≥6 were randomly assigned (1:1, computer-generated) to receive either buprenorphine (0.2–0.8 mg/day) or matched placebo for four weeks, starting 48 hours after one open-label IV ketamine infusion (0.5 mg/kg over 40 minutes). Participants, investigators, and raters were blinded to allocation. The primary outcome was weekly change in SSI score from day 1 to day 31. Analyses used a modified intention-to-treat population, including all randomized participants who received at least one study dose and completed the week 1 visit. Trial registration: ClinicalTrials.gov (NCT04116528).

Results: Between November 2020 and March 2025, 50 patients received ketamine (68% female). Five discontinued prior to week 1 of follow-on treatment. Both groups showed significant SSI reductions, but improvement was greater in the buprenorphine group (mean change −11.6 [SD 5.8], n = 23) than placebo (−6.3 [7.0], n = 22; Glass’s d = 0.76, 95% CI −1.39 to −0.11). Mixed-effects models confirmed a significant treatment-by-time interaction (p < 0.001). Depression outcomes did not differ significantly between groups. No treatment-related serious adverse events occurred.

Conclusions: Buprenorphine significantly prolonged the antisuicidal effects of ketamine across four weeks in patients with MDD. To our knowledge, this represents the first pharmacological intervention shown to maintain ketamine’s antisuicidal benefit.

Keywords: suicide prevention, Ketamine, Buprenorphine, Randomized-Controlled Trial, Major depression.

Disclosure: Nothing to disclose.

P850. Examining the role of translocator protein 18 kDa (TSPO) in stress-related mood and suicidal ideation with [11C]ER176 positron emission tomography (PET)

Elizabeth Bartlett, Sarah Herzog, Shiv Patil, Mike Schmidt, Hanga Galfalvy, M. Elizabeth Sublette, Francesca Zanderigo, Jeffrey M. Miller, J. John Mann

Columbia University / NYSPI, New York, New York, United States

Background: Positron emission tomography (PET) binding to the translocator protein 18 kDa (TSPO) has been proposed as a marker of immune cell density and inflammatory activation given that TSPO is highly expressed on the mitochondrial membrane of immune cells resident centrally (e.g., microglia, oligodendrocytes, astrocytes), peripherally (e.g., monocytes), and both centrally and peripherally (e.g., macrophages). Elevated brain TSPO has been shown in major depressive disorder (MDD) relative to healthy volunteers (e.g., Setiawan et al., 2015), in individuals with MDD with suicidal ideation vs. without ideation (Herzog et al., 2025), and with more severe depression and life stress (Bartlett et al., 2024). Here, we follow-up on such brain TSPO PET analyses to examine TSPO outside of the brain to understand the central vs. peripheral localization of TSPO effects in individuals who report depression and suicidal ideation in the context of daily life stressors.

Methods: Dynamic PET scans with the third generation tracer, [11C]ER176, allowing for imaging of all three rs6971 genotypes affecting radioligand binding (low, mixed, and high affinity binders), with concurrent arterial blood sampling, were obtained in currently-depressed participants with MDD (N = 21). Following our standard pipeline, brain tracer total volume of distribution (VT) was quantified in a region representing whole brain grey matter (described in Herzog et al., 2025). We then employed a new approach of additionally quantifying nonbrain TSPO VT (Bartlett et al., 2023). First, k-means clustering was applied to partition the PET field of view into 30 clusters. The algorithm obtains distinct clusters containing voxels with similar time activity curves (TACs) during the course of the PET scan (i.e., similar [11C]ER176 kinetics). The cluster with the greatest average TAC for [11C]ER176 uptake was localized to areas of the neck. A dual-input function (DIF) kinetic modeling approach was applied as proposed (Fujita et al. 1999) and initially validated for [11C]ER176 (Bartlett et al., 2024). The DIF approach accounts for uptake of [11C]ER176 metabolites into the nonbrain cluster (since there is no blood-brain barrier), to allow for quantification of nonbrain VT for parent [11C]ER176 that can bind to TSPO. Psychological assessments included the clinician-assessed Beck Scale for Suicidal Ideation (SSI) and ecological momentary assessment (EMA), where participants reported on stressors, negative affect (NA), and suicidal ideation (SI) approximately 6x/day for 7 days. Longitudinal linear mixed models were fit with either time-varying EMA NA or SI as the model outcome, with main effects of [11C]ER176 VT, whether or not the NA/SI occurred during a time-period where a stressor was also endorsed, and their interaction. A regression was also performed with the clinician-assessed SSI instead of EMA SI. Age, sex, body mass index, and rs6971 genotype were included as covariates.

Results: Participants were 27.2 ± 6.4 years-old, 76.2% female, and moderately depressed (Hamilton Depression Rating Scale 17-item: 19.2 ± 6.1). The automated k-means clustering approach revealed clusters that included areas consistent with parotid glands and brown adipose tissue. Nonbrain TSPO VT was significantly higher than brain VT and the two measures were moderately correlated (r = 0.58, p-values < 0.01). Brain TSPO VT trended towards being positively correlated with clinician-assessed SI severity. Using the longitudinal EMA data, the positive association between brain TSPO VT and both suicidal ideation and negative effect was significantly stronger during periods surrounding daily life stressors than during periods without stress (Herzog et al., 2025). Further results will be presented testing these effects with the nonbrain TSPO cluster newly delineated and validated here.

Conclusions: We validate automated segmentation and modeling of a nonbrain cluster of strong TSPO [11C]ER176 uptake in the neck from TSPO PET scans used for brain quantification. Given that nonbrain and brain TSPO binding were correlated, there may be common effects that regulate TSPO on the mitochondria of immune cells both centrally and peripherally. Further, brain TSPO binding was greater with increasing levels of SI and NA, particularly during periods of daily life stress, indicating that upregulated TSPO expression and/or the density of TSPO-expressing immune cells may be a mechanism by which stress confers risk for negative mood states and SI. Further investigation into the exact localization of nonbrain clusters to brown adipose tissue (BAT) and subsequent testing of the relationship with stress, mood, and SI may suggest new hypotheses for the biological bases for the deleterious effects of stress. For instance, TSPO PET has been validated as a tool to measure brown adipose tissue (BAT) activation/metabolism and BAT dysregulation has been connected to obesity, circadian rhythm dysfunction, and stress responses, lending credence to hypotheses of BAT-CNS cross-talk impacting depression and suicide risk.

Keywords: translocator protein (TSPO), suicidal ideation, PET Imaging, daily stressors, Major Depressive Disorder (MDD).

Disclosure: Nothing to disclose.

P851. fMRI neural decoder of emotion regulation predicts daily life stress exposure: evidence from a DBT vs SSRI trial

Sarah Herzog, Hanga Galfalvy, Noam Schneck, Mike Schmidt, Christina Michel, Barbara H. Stanley, J. John Mann, Beth Brodsky

Columbia University Medical Center, New York, New York, United States

Background: Suicide is a significant public health issue with limited effective treatment options. SSRIs are widely prescribed in populations at risk (e.g., mood disorders, borderline personality disorder), but evidence for reducing suicidal ideation (SI) is mixed (Stone, et al., 2009). In contrast, dialectical behavioral therapy (DBT), a skills-based intervention that emphasizes distress tolerance and emotion regulation, has demonstrated reductions in suicidal ideation and behavior (Linehan, 2015). DBT’s effectiveness may in part reflect its focus on coping with life stress, a potent trigger of increased suicide risk. Few randomized trials have directly compared DBT vs. SSRIs in treating SI, and none have used ecological momentary assessment (EMA) to capture naturalistic experiences of stress and SI in daily life as outcome measures. Neural markers of suicide risk are also lacking in clinical trials. Such measures could enhance clinical trials by providing objective indices of vulnerability and mechanisms of change. To our knowledge, this study is the first to apply EMA data to compare the effect of DBT vs. SSRIs on stress-related SI in a randomized trial, and to test an fMRI-based neural decoder of emotion regulation as a predictor of vulnerability to stress and SI.

Methods: Fifty-seven participants with borderline personality disorder (BPD) and current SI were drawn from a larger randomized clinical study (N = 84) in which participants were assigned to 6 months of DBT (weekly individual sessions) or SSRI (primarily fluoxetine) with weekly clinical management. This subgroup also completed an fMRI task wherein they recalled negative autobiographical memories while engaging a distancing regulation strategy, a form of cognitive reappraisal. A validated fMRI-based neural decoder of emotion regulation (Schneck, Herzog, 2022, Herzog et al., 2024) was applied to fMRI data to quantify engagement of emotion regulation. Decoder output was averaged to yield a participant-level index of regulation engagement. Ecological momentary assessment (EMA) was conducted at baseline and 6 months, with participants reporting on stressors and SI 6 times daily over a 12-hour waking period for 7 days. EMA data were analyzed using longitudinal linear mixed-effects models with random intercepts for participant. SI severity was modeled as both level and change (ΔSI) between consecutive prompts within the same day. Models of stress-related ΔSI included covariates for SI in the prompt prior, and analyses focused on treatment group (DBT vs. SSRI), stressor exposure (stress vs. no stress), and timepoint (baseline vs. post-treatment) as predictors of SI severity and stress-related ΔSI.

Results: The subgroup (n = 57) was all female (mean age 28.4, SD = 9.0, range 18–59); 59% were White, 17.5% Multiracial, 12% Black, and 15% of Hispanic ethnicity. Most had comorbid major depressive disorder (73.7%) and prior suicide attempt (77.2%). Both DBT and SSRI groups showed reductions in overall SI from baseline to post-treatment (β = −2.55, p = .006). The treatment group × time interaction was not significant in the mixed effects model with AR(1) correction. SI severity significantly increased during EMA periods of stress as compared with periods without stress (β = 0.92, p < .001). DBT produced a significantly greater reduction in stress-related SI compared with SSRI (DBT–SSRI Δ slope = −1.33 ΔSI units, 95% CI [−2.53, −0.12], p = .031), even after adjusting for frequency of stress exposure. The fMRI-based neural decoder was not associated with SI severity or stress-related SI change. The decoder predicted ~31% lower odds of stress exposure per 1 SD increase in decoder output (OR = 0.69, 95% CI [0.52, 0.93]; β = −0.368, p = .014), with no interaction with treatment group or timepoint.

Conclusions: Life stressors acutely increase SI and are among the most proximal triggers of suicidal behavior. Targeting stress-linked ideation may be critical for suicide prevention. Here, DBT outperformed SSRIs in reducing stress-related SI in daily life among participants with BPD (majority with comorbid MDD). The neural decoder for emotion regulation—a proxy for cognitive reappraisal—predicted lower likelihood of stress exposure. Potentially, cognitive reappraisal may reduce negative reactions to events, reducing likelihood of perceiving these events as stressors. Results are limited by a moderate all-female sample. Findings highlight the value of integrating EMA and neural decoding in clinical trials. Future studies might apply the decoder at post-treatment as a potential biomarker for treatment-related attenuation in vulnerability to stress-related increases in suicide risk.

Keywords: Neural decoding, ecological momentary assessment, randomized clinical trial, suicidal ideation, life stress.

Disclosure: Nothing to disclose.

P852. Preliminary evidence for a relationship between motor cortical excitability and negative urgency in youth with suicidal thoughts and behaviors

Charles Lewis, Gökçe Uncu, Maya Day, Alana Lieske, Mia Kellman, Paul Croarkin, Kathryn Cullen

University of Minnesota Medical School, Minneapolis, Minnesota, United States

Background: Inhibition of motoric impulses is a key component of regulating behavior. Inhibitory deficits occurring in the context of negative affective states, known as negative urgency, have been associated with various forms of psychopathology and risk behaviors, including suicidal thoughts and behaviors (STBs). Prior work has identified a developmental trajectory of negative urgency, which is thought to peak during adolescence and contribute to emerging suicide risk in youth. However, there has been a lack of research into neurobiological mechanisms underlying negative urgency, particularly in adolescents, and little examination of how motor impulsivity contributes to negative urgency. This preliminary analysis from an ongoing mechanistic study examined the relationship between motor cortex excitability and self-report and behavioral measures of impulse control in youth with STBs.

Methods: Participants were 17 adolescents (12 female, 5 male; 17.53 ± 2.76 years, range 13–21 years) with current unipolar depressive symptoms and histories of suicidal ideation or suicidal behavior, who enrolled in a longitudinal, multimodal study of mechanisms of negative urgency in STBs. Participants underwent clinical assessment, including a structured diagnostic interview as well as assessments of prior and current STBs (Columbia Suicide Severity Rating Scale, C-SSRS), depression severity (Quick Inventory of Depressive Symptoms–Adolescent, QIDS-A), and anxiety symptom severity (Generalized Anxiety Disorder-7, GAD-7). Impulsivity constructs were measured on the self-report Urgency, Premeditation (lack of), Perseverance (lack of), Sensation Seeking–Positive Urgency (UPPS-P) Impulsive Behavior Scale. Participants also completed computerized behavioral testing that included a cued Go/No-Go task to assess motor impulse control. Motor cortex excitability was assessed using single-pulse transcranial magnetic stimulation (TMS) of the left primary motor cortex, recording motor evoked potential (MEP) amplitudes on electromyography of the contralateral abductor pollicis brevis muscle at rest. Participants were instructed to relax and clear their minds to maintain a neutral affective and cognitive state during testing. The resting motor threshold (RMT) was measured according to previous published methods, using serial stimulation of the motor cortex hand knob to determine the minimum magnetic pulse intensity necessary to evoke potentials of 50 μV in 50% of trials. Cross-sectional relationships between RMT and behavioral outcome measures (UPPS-P-measured negative urgency, total and cued commission errors on the Go/No-Go task) were examined with generalized linear models, covarying for participant age and sex. Sensitivity analyses examined potential effects of depressive symptom severity, anxiety severity, and suicidal ideation on the RMT–negative urgency and RMT–commission error relationships.

Results: RMT had a significant inverse relationship with self-reported negative urgency (β = −0.025, SE = 0.0043, 95% CI −0.034 to −0.017, pFDR < .001). Age did not correlate with RMT, but was a significant covariate in the model predicting negative urgency. The RMT–negative urgency relationship remained significant when depressive symptom severity (QIDS-A score), anxiety severity (GAD-7 score), and lifetime and past-month severity, frequency, and controllability of suicidal ideation (measured on the C-SSRS) were included as covariates. There were no significant relationships between RMT and Go/No-Go overall commission errors or between RMT and cued commission errors.

Conclusions: In these preliminary data from youth with depression and STBs, motor cortical excitability correlated with self-reported negative urgency, but not with performance-based measures of prepotent response inhibition. This suggests an involvement of motor circuitry in emotion-relevant impulse control. However, experimental models of response inhibition, such as the basic cued Go/No-Go task, may measure impulsivity constructs that are distinct from negative urgency. Assessment of cortical excitability in other, later-developing brain regions involved in regulation of emotion and context processing, such as medial prefrontal and dorsolateral prefrontal cortices, will yield a more comprehensive understanding of how excitability of affective, executive, and motor networks contribute to negative urgency. Research using other behavioral tasks that simulate emotional context, as well as the development of novel experimental paradigms that induce emotional states immediately before or during excitability testing, are needed to delineate the neurophysiologic basis of negative urgency. Furthermore, longitudinal analyses across the adolescent developmental span will yield a more comprehensive understanding of how cortical network development parallels the maturation of impulse control. Future work will identify brain-based targets for modifying negative urgency and bolstering impulse control in order to reduce risk for STBs in youth.

Keywords: Cortical excitability, Negative urgency, Adolescent, Suicidal behavior, Suicidal ideation.

Disclosure: Nothing to disclose.

P853. OFC resting state connectivity related to aggression in female veterans with a history of suicide behavior: a pilot study

Erin McGlade, Keenan Roberts, James Pewitt Yancey, Jadwiga Rogowska, Deborah Yurgelun-Todd

University of Utah School of Medicine, Huntsman Mental Health Institute, Rocky Mountain MIRECC, Salt Lake City, Utah, United States

Background: Despite significant efforts toward suicide prevention within the US Department of Veterans Affairs (VA) and Office of Suicide Prevention, suicide deaths in Veterans continue and remain a top priority within VA. Prior research has identified numerous risk factors for suicide attempts and death, including mental health and substance use disorders, prior suicide attempts, and aggression toward others. Neurobiological differences also have been associated with SB, including orbitofrontal cortex (OFC) connectivity and volume. However, most research has focused on male Veterans or not examined male and female Veteran groups independently. The current study sought to combine psychosocial and neurobiological findings on SB to examine sex differences in OFC resting state connectivity and aggression in Veterans who reported a range of SB.

Methods: Participants completed clinical assessments, including the Columbia Suicide Severity Rating Scales (CSSRS), Buss-Perry Aggression Questionnaire (BPAQ), Hamilton Rating Scale for Anxiety (HAM-A), and Hamilton Depression Rating Scale (HAM-D). Participants also completed structural and functional magnetic resonance imaging (MRI) on a 3T Siemens Prisma scanner. High-resolution, high contrast structural MRI images were acquired using a three-dimensional (3D), T1-weighted magnetization prepared rapid gradient echo (MPRAGE) sequence with the following parameters: Repetition time (TR) = 2500ms; Echo time (TE) = 2.80ms; Inversion time (TI) = 1060ms; Flip angle = 8°; Number of slices = 176; Matrix size = 256 × 256; Field of view (FOV) = 256 × 256; and Slice thickness = 1.0mm.

Factor analysis was used to examine the main effects of SB, sex, and the SB by sex interaction on brain connectivity. Post-hoc analyses were conducted between participant groups with a significance level set at p = 0.005. Regression analyses examined the relationship between BOLD signal data and aggression scores.

Results: Participants included 50 male and female US Veterans ages 18 to 55. Participants were divided into six groups based on sex and SB. Veterans who denied a history of suicidal ideation and attempts were placed in the ‘no SB’ group; those with a history of suicidal ideation but no attempts were included in the SI group, and those with a history of suicide attempt(s) were included in the SA group. SI and SA groups were combined in the aggression analyses to create the SB group. This resulted in the following groupings: 5 females without SB, 8 males without SB, 13 females with SI, 12 males with SI, 5 females with SA, and 7 males with SA. The majority of Veterans in the sample identified as White and married with some college/specialized training or having graduated from college/university.

Factor analyses examining resting state connectivity in Veterans with SB and those without showed that Veterans without SB showed increased resting state connectivity between the left OFC, cerebellum, supra marginal gyrus, angular gyrus, parietal, and frontal regions compared to Veterans with SI. No between group differences were observed in the right OFC.

Analyses measuring the association between resting state connectivity with aggression by sex and SB showed increased connectivity between the left OFC and left caudate and right thalamus correlated with higher Buss Perry Anger scores in females with SB; increased connectivity between the left OFC and left angular gyrus and left inferior parietal gyrus correlated with increased Buss Perry Total Scores in females with SB; increased connectivity between the right OFC and right frontal mid gyrus, left angular gyrus, and left parietal inferior gyrus correlated with higher Buss Perry Total Scores in females with SB; and decreased connectivity between the right OFC and right fusiform gyrus, right cuneus, and inferior right occipital gyrus correlated with higher Buss Perry Hostility scores in females with SB. Males with SB showed none of these associations between OFC resting state connectivity and aggression.

Additional analyses showed that males who endorsed a lifetime history of SB also endorsed significantly more anger (p = .02). Chi-square tests showed no significant between-group differences on handedness, education, ethnicity, relationship status, or employment status. Males without SB, males with SB, females without SB, and females with SB did not differ on anxiety or depressive symptoms; physical, verbal aggression, or total; or hostility.

Conclusions: As hypothesized, our results showed sex differences in OFC resting state connectivity and aggression by SB in Veterans. Moreover, sex-dependent associations between SB groups for the left and right OFC connectivity were related to brain regions involved in information processing and decision making. These differences in connectivity associations were related to aggression scores in females with SB but not males. These results suggest that OFC connectivity and aggression are related to SB differently in female Veterans and to males. Future studies should continue to explore brain-based changes related to aggression by sex with the goal of identifying more sensitive and accurate risk factors for SB. These findings further suggest potential neurobiological underpinnings of aggression and SB in Veterans, which could lead to the development of novel interventions.

Keywords: Suicide, Resting State Connectivity, Attention, Sex Differences, Veterans.

Disclosure: Nothing to disclose.

P854. Evidence of real‑world effectiveness of ketamine pre-treatment followed by inpatient psychiatric care for rapid reduction of severe suicide risk

Alexandre Paim Diaz, Manivel Rengasamy, Sharvari Shivanekar, Anthony Pizon, Crystal Spotts, Priya Gopalan, Rebecca Price

University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, United States

Background: Suicide is one of the leading causes of mortality globally. Suicide rates are considerably higher after discharge from psychiatric facilities, with approximately 25% occurring within the first month post-hospitalization. Thus, scalable, rapid-acting, effective interventions to reduce suicide risk during this high-risk period are urgently needed.

Methods: This pragmatic, randomized, parallel-arm trial investigated the effectiveness of a single infusion of ketamine in decreasing depressive symptoms and suicidal ideation at 24 hours and across the first month following a serious suicide attempt. Patients hospitalized in a general medical hospital after a serious suicide attempt, regardless of their psychiatric diagnoses, were invited to participate in this randomized study. Participants were randomized to receive either standard care supplemented by one single infusion of 0.5mg/kg IV ketamine pre-treatment (prior to psychiatric inpatient stay) or standard care alone. Standard care during the study period involved inpatient medical stabilization (with or without the study ketamine infusion) followed by transfer to inpatient psychiatric care and discharge to psychiatric outpatient providers. We hypothesized that pre-treatment with a single ketamine infusion could synergistically enhance the clinical outcomes achieved through standard care. The primary outcomes were blinded outcome raters’ evaluations of depressive symptoms measured by the Montgomery–Åsberg Depression Rating Scale (MADRS) and suicidal ideation measured by the Beck Scale for Suicidal Ideation (BSSI). Intent-to-treat analyses using linear mixed-effects models compared trajectories on continuous outcomes between groups. To additionally dichotomize outcomes, response was defined as ≥50% improvement from pre-randomization baseline in each measure.

Results: A total of 101 participants received ketamine plus standard of care, and 99 received standard of care alone. The sample was on average 37 years old (SD 14), predominantly female (60%), White (77.5%), and Non-Hispanic/Non-Latino (91.7%). At baseline, participants had a mean of 3.6 suicide attempts (SD 3.1), a MADRS score of 39 (SD 9) indicating severe depression, and a BSSI score of 25 (SD 6). Borderline personality disorder and substance/alcohol use or dependence, diagnoses commonly underrepresented in ketamine trials, were present in 18% and 67.5% of the sample, respectively. Most participants were using a psychotropic medication at baseline (69.5%), mainly antidepressants (52.5%), mood stabilizers (30%), and antipsychotics (26.5%). The mean number of psychotropics used at baseline was 1.7 (SD 1.6). Ketamine was associated with a significantly larger rapid decrease in depressive symptoms at 24 hours (–4.07, 95% confidence interval [CI] –7.19 to –0.95, p = 0.01) as well as persistently lower depression scores at one month post-randomization (–3.20, CI –5.93 to –0.48, p = 0.02). By one month (but not at 24h), ketamine was also associated with a larger decrease in BSSI scores (–0.09, 95% CI –0.17 to –0.01, p = 0.03). At both 24h and one month post-infusion, a significantly higher proportion of participants in the ketamine group were classified as responders in depressive symptoms (24h: 57.6% vs. 37.6%, p = 0.009; Month 1: 72.9% vs. 52.7%, p = 0.02. At one month, ketamine was also associated with a higher response rate in suicidal ideation (89.6% vs. 73.6%, p = 0.03).

Conclusions: A single pre-treatment dose of IV ketamine was associated with larger decreases in depressive symptoms and suicidal ideation observed during a one month, high-risk period in a transdiagnostic sample of individuals hospitalized for a serious suicide attempt in a real-world setting. A single ketamine infusion, delivered in the course of inpatient medical stabilization and just prior to specialized psychiatric care, may open a window of opportunity that promotes the uptake of therapeutic benefits during standard care.

Trial registration number NCT04578938.

Keywords: Suicidality, IV ketamine, Randomized-Controlled Trial, depressive symptoms, effectiveness.

Disclosure: Nothing to disclose.

P855. Frontolimbic activity related to irritability and suicidality in adults with major depressive disorder

Aaima Cheema, Jessica Sah, Sakina Rizvi, Sidney Kennedy, Katharine Dunlop

University of Toronto, Toronto, Canada

Background: Suicide is the ninth leading cause of death in North America. Irritability, which can be defined as dysregulated emotions and a low threshold for anger and reactive aggression, consistently predicts suicidal ideation severity and risk of attempt. Both irritability and suicidality are associated with interpersonal difficulties. Thus, inducing irritability through negative social experiences may provide valuable neurobiological insights into the relationship between irritability and suicide risk in major depressive disorder (MDD). Therefore, our study aims to characterize the behavioral and neurobiological relationship between irritability and suicidality in adults with MDD using tasks that elicit frustrative nonreward in a social context.

Methods: 90 male and female adults with MDD (30 controls, 30 MDD with a past attempt [MDD+SA], and 30 MDD without a past attempt) completed mood, suicidality and irritability assessments. Participants completed two fMRI paradigms: the Cyberball Task, which simulates social exclusion, and the Dictator Task, which examines retaliatory aggression against excluders. We tested for group differences in self-reported irritability, and task-related behaviors and brain activity controlling for age and sex.

Results: The MDD+SA group reported significantly higher irritability (p < .05) relative to the other groups. Self-reported irritability on a visual analogue scale increased following social exclusion in the full sample (p < .05) with no significant group differences. The MDD+SA group had elevated retaliation towards excluders relative to the other two groups (p < 0.01). Relative to the other groups, irritability in the MDD+SA group was correlated with ventro/dorsomedial prefrontal and dorsolateral prefrontal cortex activity during the Cyberball and Dictator Tasks.

Conclusions: Elevated baseline irritability and greater retaliation of excluders was present in individuals with a higher suicide risk. Irritability could be a valuable behavioral target to assess and mitigate suicide risk in adults with MDD. Medial and dorsolateral prefrontal cortex activity was correlated with irritability in the MDD+SA group, which may set a rationale for future explorations of non-invasive brain stimulation manipulating circuitry as new treatments for suicidality presenting with irritability.

Keywords: Irritability/Aggression, Suicidality, fMRI biomarkers.

Disclosure: Nothing to disclose.

P856. Altered gut microbial metabolism in patients with depression and recent suicide attempts

Faith Dickerson, Emese Prandovszky, Robert Yolken

Sheppard Pratt Health System, Baltimore, Maryland, United States

Background: The human gastrointestinal tract and other mucosal surfaces are populated by many non-pathogenic microorganisms characterized as the gut microbiome. Components of the microbiome can communicate with the brain through the gut-brain axis, a bidirectional communication network between the enteric nervous system (gut) and the central nervous system (brain). The gut-brain axis can influence mood, cognition, and behavior. Recent studies have suggested that alterations in the gut microbiome are associated with metabolic disturbances and inflammatory processes in psychiatric disorders. However, the composition of the gut microbiome in persons with recent suicide behaviors has not been extensively studied.

Methods: The sample in this study included 50 patients with Major Depressive Disorder (MDD) hospitalized for a major depressive episode, 35 with a suicide attempt in the previous 30 days and 15 without any history of suicide attempts. Fecal swab samples were collected from the individuals during the hospital stay. DNA was extracted from the samples and bacterial microbiome were assessed by amplification of the conserved 16S rRNA region as well as whole genome sequencing. Diversity measures were determined using Qiime2, differential abundant bacterial taxa and predicted metabolic pathways were determined using the Human 3 algorithm.

Results: We found that persons who had made a recent suicide attempt exhibited significantly greater microbial richness, distinct beta-diversity patterns, and differences in the relative abundance of individual taxa in the gut microbiome compared to the hospitalized patients with MDD and no suicide attempt history. In terms of alpha diversity, individuals with a recent suicide attempt exhibited significantly richer microflora, as measured by the Faith phylogenetic diversity index (p = 0.045). No significant differences were found in the other three alpha diversity measures. For beta diversity, the greatest difference between the groups was observed in the unweighted UniFrac metric (p = .020). The greatest differences in taxa abundance between groups, meeting the criteria of p < .05 and an absolute value of log fold change (LFC) ≥.20, were Fenollaria timonensis (p < .003, LFC = .56), which was increased, and Corynebacterium aurimucosum (p = .034, LFC = .71), which was decreased in individuals with a recent suicide attempt. Additionally, 25 metabolic pathways differed between groups, with several linked to energy metabolism and amino acid processing, processes previously associated with MDD and suicidal behavior.

Conclusions: These findings suggest that microbiome composition may influence suicide risk through gut-brain axis-mediated pathways. The microbiome can be modified by a number of interventions relating to diet, probiotics, and medications. Future research should explore microbial-targeted interventions as a potential strategy for suicide prevention in individuals with MDD and other high-risk groups.

Keywords: Gut Microbiome, Suicide attempt, biomarkers.

Disclosure: Nothing to disclose.

P857. Applications of dynamical systems theory to identify suicide attempts 24 hours in advance

Melanie Bozzay, Jonathan Butner, Jessica Turner, Rebecca Neufeld, Craig Bryan, Nicole Nugent, Michael Armey

The Ohio State University, Columbus, Ohio, United States

Background: Suicide is a significant public health problem. Currently, there are no clinical strategies available to identify who will engage in suicidal behavior (SB) and when, hindering efforts to deliver targeted, timely interventions to prevent suicides. In recent decades, computationally intensive models have been developed to forecast which patients are at risk of SB, months to years in advance. Despite these advances, existing risk monitoring systems do not reduce SBs or identify clinically actionable time periods of SB risk.

Dynamical systems approaches may provide a critical means of anticipating SB over shorter time periods. The causes of suicide are multifactorial, involving complex dynamical interactions among cognitive processes, emotional states, and environmental stressors. Through the lens of dynamical systems theory, suicide risk can be thought of as a system with low (non-SB) and high (SB) risk states, each demarcated by different patterns of dynamic interrelationships between key risk and protective processes. In other words, modeling competing processes (i.e., dynamics that promote versus resist change), while allowing for differing degrees of stability or instability, could provide a method for describing the complex behavior of suicide risk under various conditions, facilitating anticipation of ‘shifts’ towards SB. The present study examined whether changes in dynamics demarcated which patients made an SB, 24-hours prior to the SB.

Methods: Data were analyzed from 70 participants recruited as part of a larger, ongoing study investigating the role of deficient sleep in short-term suicide risk. Participants completed Ecological Momentary Assessment (EMA) measures up to 5 times per day in the 4 weeks following discharge from a psychiatric inpatient stay for suicide risk. EMA items measured positive and negative affect, suicidal ideation, suicidal intent, and recent SB. To test study aims, the first derivatives of each individual’s EMA measures were estimated (capturing how that individual was changing over time). Multilevel models with a Bayesian estimator predicting these changes from individual values on EMA measures were computed, with a dummy code for presence or absence of SB in each time window. Results were used to construct Jacobian matrices summarizing the dynamics. These dynamics were then visualized as individual networks, and the eigenvalues of the Jacobian matrices were derived from a principal components analysis and used to investigate stability of the SB and non-SB states. These eigenvalues were used to examine whether a stability cut-off value differentiated SB versus non-SB windows.

Results: The dynamics between EMA variables differentiated prior to SB and non-SB windows within individuals. There was also heterogeneity in these relationships, with patients showing individualized dynamic patterns in network models. We also found a change towards more interactions in the 24 hours before SB. When we used principal components analysis as a summary of the dynamics across subjects, we identified a potential decision criterion to anticipate SB, even with all the individual variability. Specifically, a cut-off of −.025 on the first eigenvalue yielded a positive predictive value (PPV) of 34% (95% CI 32.0–36.3) for identifying individuals 24 hours prior to an SB, with a sensitivity of 74% and specificity of 99%, representing a potentially clinically actionable criterion.

Conclusions: This study provides the first evidence that dynamical systems modeling approaches may have utility for anticipating SB risk, 24 hours in advance. Although a pilot model, it outperforms SI screening and newer machine learning models that use data from very large clinical datasets (meta-analytic pooled PPVs = 4–10%), despite using a decision criterion derived using dynamics from only 4 variables, and data from a limited sample. Additional research is needed with a large, ideally nationally representative, sample to confirm these models and examine whether this approach to identify SB risk generalizes beyond high-risk clinical groups. Nevertheless, our findings preliminarily suggest this approach could be used to facilitate strategies for identifying an individual patient’s current probability for transitioning to a higher or lower suicide risk state within clinically relevant timeframes.

Keywords: Suicide prediction, Dynamical Systems, ecological momentary assessment, Suicide Assessment.

Disclosure: Nothing to disclose.

P858. The association between suicidal ideation and negative affect: a 6-month ecological momentary assessment study

Alejandro Porras-Segovia, Lucía Albarracín-García, Enrique Baca-García

Health Research Institute Fundación Jiménez Díaz, Madrid, Spain

Background: Suicide is a significant global public health concern, accounting for over 700,000 deaths worldwide each year. Traditionally, suicide risk has been assessed through clinical interviews and self-report questionnaires. However, these methods have limitations such as recall bias, disregard of context, and emphasis on mid and long-term risk factors at the expense of proximal risk factors or triggers.

Ecological Momentary Assessment (EMA) has emerged as a promising tool to address these limitations. EMA involves repeatedly assessing participants in real time, typically through daily prompts delivered in their natural environments. EMA enables tracking of dynamic changes over time and captures subtle variations that traditional methods may overlook.

However, the “EMA burden”—participant fatigue due to frequent prompts—often reduces adherence and limits the duration of data collection to just a few weeks. This problem may be mitigated by implementing a rotating system of questions to avoid repetition. Using this method, longer follow-up periods may be obtained, which in turn may help us to better capture the proximal risk factors of suicidal behaviour.

The current study aims to investigate the association between suicidal ideation and several variables measured in real time using EMA.

Methods: This is an observational prospective study, embedded in the SmartCrisis 2.0 clinical trial (ClinicalTrials.gov ID NCT04775160). Our sample consists of adult psychiatric outpatients who had a recent (past month) episode of suicidal behaviour. Depression was measured using the Inventory of Depressive Symptomatology–Clinician-Rated (IDS-C); suicidality through the Columbia Suicide Severity Rating Scale (CSSRS), and psychological and physical pain by Visual Analog Scales (VAS).

EMA was administered through an app installed on participants’ smartphones. Each morning, a prompt presented 2 to 4 randomly selected questions from a pool of 34 items measuring suicidal ideation (frequency over-represented), non-suicidal self-injury, negative affect, interpersonal experiences, sleep, and eating habits. Participants were followed for six months.

Intra-individual mean and variability of responses were calculated; variability was assessed using the root mean squared of successive differences (RMSSD). Mixed-effects models were employed to evaluate within-person variability, temporal trends, and the effect of EMA categories on suicidal ideation over time. Multiple linear regression examined whether baseline levels of depression and psychological distress predicted the mean and variability of suicidal ideation.

Results: Characteristics of the sample:

N = 99 participants were included in the final analysis. Mean age was 39.5 ±0.05 years, and there was a 62.6% of females. A total of 13,900 EMA answers were obtained, of which 5,638 belonged to the suicidal ideation category.

Association between EMA variables:

We found a high within-person variance for all EMA variables, with an ICC of 0.17 for suicidal ideation, ICC = 0.24 for negative affect, ICC = 0.26 for interpersonal difficulties, ICC = 0.29 for sleep, and ICC = 0.45 for eating habits. Negative affect (B = 0.193; p < 0.001) and interpersonal difficulties (B = 0.168; p = 0.001) were directly associated with suicidal ideation, and greater intra-individual means of negative affect (R2 = 0.39, B = 0.502; p < 0.001) and interpersonal difficulties (R2 = 0.182, B = 0.359, p < 0.001) were associated with greater intra-individual mean of suicidal ideation.

Intra-individual variability of SI was directly associated with intra-individual variability of negative affect (R2 = 0.15, B = 0.477, p = 0.001), interpersonal difficulties (R2 = 0.15, B = 0.323, p = 0.001), sleep problems (R2 = 0.12, B = 0.259, p = 0.004) and eating habits (R2 = 0.23, B = 0.318, p < 0.001).

Association between traditional baseline measures and EMA variables:

There was an inverse association between baseline maximum psychological pain and variability of negative affect (R2 = 0.12, B = −1.40; p = 0.040). The interaction psychological pain*days of follow up significantly predicted intra-individual mean of EMA-measured negative affect (R2 = 0.24, B = 0.019, p = 0.017). The interaction psychological pain*days of follow up predicted the intra-individual mean of suicidal ideation (R2 = 0.34, B = 0.024, p < 0.001). Higher baseline psychological pain predicted persistently elevated suicidal ideation and negative affect throughout the follow-up.

Conclusions: Negative affect and interpersonal difficulties appeared as significant real-time correlates of suicidal ideation in high-risk outpatients. Our results are consistent with prior research that identified emotions such as anxiety, shame and self-hatred as predictors of subsequent SI using EMA. Our results also echo the core tenets of the Interpersonal Theory of Suicide, which posits that a lack of belonging and the perception of being a burden are fundamental drivers of SI.

The dynamic interplay between negative affect, interpersonal difficulties and suicidal ideation, as observed in our study, supports the need for personalised, real-time monitoring strategies. This reinforces recent critiques of traditional clinical assessments, which often fail to capture moment-to-moment fluctuations in suicide risk. Our results emphasise the value of EMA in capturing these rapid shifts, which standard tools may overlook.

Keywords: Suicide Assessment, ecological momentary assessment, Psychological Pain, suicidal ideation, smartphone apps.

Disclosure: Nothing to disclose.

P859. Levels of matrix metalloproteinase 9 (MMP-9) are elevated in persons with major depressive disorder who have had a recent suicide attempt

Faith Dickerson, Robert Yolken

Sheppard Pratt Health System, Baltimore, Maryland, United States

Background: Suicide attempts are the strongest risk factor for death by suicide and are widely prevalent, especially in persons with major depressive disorder (MDD). While numerous clinical and demographic factors are significantly associated with suicide attempts in this population, it is not currently possible to accurately predict suicide attempts on an individual level. Recent studies have indicated that many individuals with suicide behaviors have evidence of systemic immune activation. The identification of blood-based immune markers thus holds promise for predicting and monitoring suicide behavior. Matrix metalloproteinases (MMPs) are a diverse set of enzymes associated with tissue remodeling, inflammation, and alteration in the blood-brain barrier. Matrix metalloproteinase 9 (MMP-9) is of particular interest since it has been associated with psychiatric disorders including MDD.

Methods: The study population consisted of 265 individuals. These included 86 persons hospitalized with MDD who had a suicide attempt within the previous 30 days, and 94 persons hospitalized with MDD who did not have a recent attempt and 79 persons without a psychiatric disorder. We measured the levels of MMP-9 in blood samples from all of the study individuals and determined their association with the diagnostic groups employing mixed effects models adjusted for demographic and clinical variables. Wald tests were used to calculate groupwise comparisons.

Results: The overall model showed significant differences in levels of MMP-9 among the groups (Wald Test, Χ2 = 11.53, p = .005). Pairwise comparisons indicate higher levels of MMP-9 in the MDD participants with a recent suicide attempt compared with individuals with MDD who did not have a recent suicide attempt (Wald Test, Χ2 = 7.68, p = .006) and with the non-psychiatric comparison group (Wald Test, Χ2 = 9.39. p = .002). There were not significantly higher levels of MMP-9 in the MDD group without a recent suicide attempt vs. the non-psychiatric comparison group (p = 0.4).

Conclusions: Therapeutic approaches to modulate MMP-9 activity are currently being developed and evaluated. Measurement of MMP-9 may inform clinical trials of these modalities for the prevention of suicide behaviors in high-risk individuals. Our findings show the promise of biological markers to identify persons who are at risk for suicide attempts and ultimately develop new treatments. Prospective studies, larger sample sizes, standardized methods, and participation by persons who are at highest suicide risk are all needed to advance this area of research.

Keywords: Suicide attempt, Immune Biomarkers, Major depression.

Disclosure: Nothing to disclose.

P860. Polygenic risk for suicide behavior is associated with amygdala volume in youth with bipolar disorder

Mikaela Dimick, Clement Zai, Bradley MacIntosh, Benjamin Goldstein

Centre for Addiction and Mental Health, Toronto, Canada

Background: Youth with bipolar disorder (BD) are at greatly increased risk for suicide, yet the genetic and neurobiological mechanisms contributing to this risk remain poorly understood. Polygenic risk scores (PRS) provide a quantitative measure of molecular genetic liability to suicide behavior. There is a paucity of knowledge regarding neuroimaging phenotypes that may be implicated in mediating this genetic liability, particularly in youth and in relation to BD. Neuroimaging research has consistently implicated frontolimbic circuitry—including the orbitofrontal cortex (OFC), ventrolateral prefrontal cortex (vlPFC), and amygdala—in the regulation of emotions and decision-making processes that are implicated in suicide risk. Structural alterations in these regions have been associated with suicidal thoughts and behaviors. The present study seeks to address a gap in the literature by examining PRS for suicide behavior in relation to brain structure in youth with and without BD, focusing on the amygdala, OFC, and vlPFC as regions of interest.

Methods: Participants were 114 youth of European ancestry aged 13–20 with BD (n = 67) and control youth without BD (n = 47) who completed structural 3T MRI. Diagnoses and history of self-harm were ascertained via the K-SADS-PL interview with youth and parent. Cortical surface area, volume and thickness (vlPFC, OFC) and amygdala volume were derived with FreeSurfer. PRS for suicide behavior were computed using PRS-CS-auto based on summary statistics from the latest GWAS from the Suicide Attempt Working Group of the Psychiatric Genomics Consortium. Linear regressions tested (1) associations within the BD group, (2) differences between BD youth with and without self-harm, and (3) diagnostic group-by-PRS interactions, adjusting for age, sex, genetic principal components, intracranial volume (for surface area and thickness) and diagnosis. Multiple comparisons were controlled using false discovery rate (FDR) correction at q < 0.05 within each ROI family.

Results: Within the BD group, higher suicide behavior PRS was associated with larger amygdala volume (ϐ = 0.31, 95% CI = 21.20–186.11, p = 0.02). Diagnosis-by-PRS interactions were also significant (ϐ = 0.19, 95% CI = 18.88–111.96, p = 0.006). Post hoc analyses confirmed that in BD, higher suicide behavior PRS was associated with larger amygdala volume (ϐ = 0.31, 95% CI = 21.47–188.54, p = 0.02), whereas in controls, higher suicide behavior PRS showed a non-significant trend toward smaller amygdala volume (ϐ = –0.28, 95% CI = −188.40–0.14, p = 0.05). Self-harm status within BD did not significantly moderate any associations. After FDR correction, both BD main effects and diagnosis-by-PRS interactions remained significant. No significant associations were observed between suicide behavior PRS and OFC or vlPFC.

Conclusions: Findings indicate that polygenic risk for suicide behavior is reflected in amygdala structure among youth with BD, with opposite associations observed in control youth. This suggests that genetic liability to suicide behavior may interact with BD and developmental processes to shape amygdala morphology. Polygenic liability for suicide behavior may act through serotonergic and stress-response pathways that shape frontolimbic circuitry; the amygdala, as a hub for emotion regulation and threat detection, may be particularly sensitive to these effects. Findings represent a preliminary step toward advancing understanding of suicide risk in this high-risk population by focusing on the intersection of genetic risk and intermediate phenotypes putatively implicated in suicide behavior. Identifying neurobiological correlates of suicide behavior PRS in youth with BD may advance mechanistic understanding of inherited risk and support the development of early prevention strategies tailored to individuals at greatest risk for suicide.

Keywords: suicide, brain structure, Polygenetic Risk Score.

Disclosure: Nothing to disclose.

P861. Electrophysiological correlates of suicide risk in selective attention to mortality-related stimuli

Yoojin Lee, Jessica Gilbert, Steven Lamontagne, Laura Waldman, Megan Kenna, Nancy Adleman, David Jobes, Carlos Zarate, Elizabeth Ballard

National Institute of Mental Health, Bethesda, Maryland, United States

Background: Suicide is a leading cause of death. Previous suicide research has been limited by heterogeneity in individual suicide risk profiles as well as our incomplete understanding of the neurobiological mechanisms for suicide risk. Establishing a neurobiological marker of suicide risk is crucial because it can contribute to overcoming the limitations of traditional research methods in suicide research that heavily rely on retrospective self-reports. Recent research suggests that selective attention to emotional cues could be a putative biomarker of depression and suicide. However, the mechanistic understanding of how selective attention is processed in the brain remains limited.

Methods: This study investigates the neural correlates of suicide risk using magnetoencephalographic (MEG) power. Specifically, we examine whether suicide risk moderates the selective attention to death-related (i.e., mortality) cues. We recruited three groups of participants (N = 53): individuals with suicide risk (suicide risk [SR]; n = 17) who either had a recent suicidal crisis, a previous suicide attempt history, or on-going suicidal ideation; individuals with depression but no suicide risk (clinical controls [CC]; n = 17); and healthy controls (HC; n = 19). The task consisted of a 2 × 2 factorial design of conditions: death/life cues for mortality and congruent/incongruent cues for attention allocation to emotional cues. Participants were instructed to focus on mortality-related (death) or non-mortality-related (life) cues, as well as to compare these cues to neutral stimuli. Trials were categorized as congruent when participants attended to death- or life-related cues relative to neutral cues and as incongruent when they attended to neutral cues relative to death- or life-related cues. Participants completed the suicide dot probe task, which measured their selective attention to mortality cues by using both behavioral (reaction times) and neural activity (MEG). Statistical group differences in brain activation were analyzed using linear mixed-effects models (3dLMEr) implemented in the AFNI software, accounting for within-subject variability and covariates such as age and biological sex. Effective connectivity was estimated using the Dynamic Causal Modeling (DCM) in the SPM12 pipeline.

Results: No significant differences in reaction time were observed among the groups in any of the conditions (ps > .05). This suggests that selective attention to mortality cues may not be captured at the behavioral level. Compared to CC group, SR group exhibited decreased activity in the middle temporal gyrus (MTG) and entorhinal cortex (EC) in the beta band (15–29 Hz), especially in the death and incongruent conditions. Compared to CC group, SR group showed decreased beta and gamma (30–58 Hz) band activity in the incongruent condition, while CC group showed increased alpha (9–14 Hz) and beta band activity in the MTG within the congruent condition. Moreover, SR group revealed decreased activity in the superior parietal lobule (SPL) compared to the HC group in the alpha band (ps < .05) within the death and incongruent conditions, whereas CC group revealed decreased activity in the SPL compared to HC group in the alpha band (p < .05) within the congruent condition. These findings indicate that suicide risk can moderate the MEG power in brain regions involved in sensory information processing. Using Dynamic Causal Modeling analysis, we estimated the effective connectivity in each condition, comparing the suicide risk group, clinical controls, and healthy controls. In the death condition, reduced feedforward connectivity was observed from the early visual cortex to EC, then to SPL, and finally to MTG in the SR group, compared to the others (posterior probabilities(posteriorps) > .90), suggesting potential deficits in updating sensory inputs and regulating selective attention. Conversely, in the life condition, reduced connectivity was observed from the MTG to SPL in the SR group, compared to the others (posteriorps > .90), hinting subtle impairments in downregulating attention orientation during life-related cognitive processes. In the congruent condition, MTG exerted a constraining influence on the EC through top-down signal processing in the SR group (posteriorps > .90). On the other hand, in the incongruent condition, those in the SR group showed less constraining in the reduced top-down signaling between the EC and the SPL (posteriorps > .90). These findings may indicate that MTG plays a critical role in signal processing and integration, particularly in the context of mortality and emotional attention orientation.

Conclusions: Our findings suggest that abnormalities in selective attention, due to inefficient signal updating as context updating continues in response to mortality cues, may serve as potential biomarkers for suicide risk. These deficits may not be readily apparent at the behavioral level, highlighting the need for further research employing neuroimaging and other techniques to fully characterize these impairments. These findings have the potential to enhance our understanding of the neurobiological underpinnings of suicide risk and inform the development of more effective prevention and intervention strategies.

Keywords: magnetoencephalography, Suicide risk factors, dynamic causal modeling, selective attention.

Disclosure: Nothing to disclose.

P862. Comparative effectiveness of ECT vs. subanesthetic intravenous ketamine for rapid reversal of acute suicidal depression: a patient centered multi-site study

Amit Anand, Bo Hu

Harvard Medical School, Boston, Massachusetts, United States

Background: In the United States there is a public health crisis: a person dies by their own hand every 10 minutes. It is estimated that 10.7 million adults seriously consider suicide each year, 3.5 million make a plan, and 1.5 million make an actual attempt. Patients suffering from Acute Suicidal Depression (ASD) are usually admitted as inpatients for safety and started on oral antidepressants (which can take 6 – 12 weeks to have an effect) and given nursing-care and psychotherapy. They are then discharged, usually within a few days, as soon as immediate safety concerns are ameliorated or insurance payment runs out. Multiple studies have shown that the immediate post-discharge period is of the highest risk for repeat suicide attempts and completed suicides. An important reason for the inadequate treatment of ASD is the lack of data from large scale comparative studies of available efficacious treatments such as ECT and KET. In the absence of data to guide rational treatment choice, neither treatment is being used adequately. To address this critical knowledge gap, we are conducting a pragmatic randomized non-inferiority single-blind comparative effectiveness trial of ECT vs. KET treatment for ASD. This study is funded by Patient Centered Outcomes Research Institute (PCORI) and sponsored by Mass General Brigham Hospitals.

Methods: Patients (N = 1500) who are referred by their clinicians for rapid reversal of ASD are included in the study and randomized to ECT (3 times a week) or KET (2 times a week) for 4 weeks. The primary outcome measure is remission of suicidal symptoms as assessed by the clinician-rated Scale for Suicidal Ideation (SSI) score. Secondary outcome measures are ratings on depression, quality of life, patient engagement and cognition. A modified intent to treat (mITT) analysis will be conducted and all patients who receive at least one treatment and who complete the end of treatment (EOT) assessment will be included in the analysis. At treatment end, remitters (SSI score < 4) will be started on weekly Collaborative Assessment and Management of Suicide (CAMS) therapy sessions and receive up to 6- 8 sessions with follow-up assessments at 1, 3 and 6 months. Aim 1 is to investigate the differential effects of ECT vs. KET on the SSI and other secondary measures. ECT and KET have been reported to have a differential effect on the young and older populations, therefore Aim 2 is to investigate the effect of age on ECT vs. KET effects. Finally, Aim 3 is to investigate the effect of ECT and KET on effectiveness of CAMS in remitted patients in the post-treatment phase. In the feasibility phase of the study, for the duration of 1.5 years, ascertained recruitment and acceptance of treatments and procedures. Using the experience and data from the feasibility phase, the full phase study protocol has been finalized and implemented. Ten sites, with leading investigators in the field across the country, are already participating in the study and another 5–8 sites will be included. A separate and independent Data Coordinating Center (P.I. Dr. Bo Hu PhD) is in charge of data quality, management and analysis. A Data Safety Monitoring of 5 experts is overseeing the project. Patients, advocacy organizations and third-party payer stakeholders are included in study design and monitoring. Results will be disseminated with the participating patients, general community and the scientific community. This study will be the largest ever prospective clinical trial of acute suicidality associated with depression.

Results: At present, we have completed the feasiblity phase of the study. A total of N = 842 subjects with suicidal depression have been prescreened and N = 491 were excluded due to not meeting inclusion criteria. N = 294 subjects declined to participate predominantly due to not wanting to be randomized or wanting one treatment over the other. N = 145 subjects were consented and N = 125 randomized. Demogrpahics of randomized subjects are as follows - Age 40+/− 14 yrs (Ages 18–35yrs (44%), 35–60yrs (46%), and 60–90 yrs (10%)); 60% females, White: 78%; Black: 4%, Asian 12%, Native American 2%; Hispanic 4%. Patients recruited as inpatients - 66%. 8% subjects withdrew before they received the first treatment, predominantly subjects randomized to ECT. Treatment has been started for 82% of subjects and 47% subjects have completed treatment. An overall withdrawal rate of 8% has been seen. CAMS was started on 25% of subjects initially randomized.

Conclusions: This will be the largest study of acute somatic treatment of sucidal depression ever conducted. Results from this study will provide unique data regarding whether ECT or ketamine should be used as the first line treatment for the life-threatening condition of ASD. The study will provide addtional data regarding pragmatic availability and acceptance of ECT vs. ketamine treatment in day to day clinical service. The most frequent reason for patient declining taking part in the study is that patient prefer to get one treatment or the other. Other challenges in the study have been the burden of frequent assessments with each treatment and that IV ketamine is not reimbursed by insurers as it is still desginated as an investigtional drug by the Food and Drug Administration. From all the lessons learned in the feasibility phase, the study is now moving the 5-year full phase where an additional N = 1375 subjects will be recruited. At least 5–8 addititonal sites will be added to study to accomplish the goal. We continue to invite interested to take part in the study.

Keywords: suicide prevention, Depression, electroconvulsive therapy, IV ketamine, Comparative effectiveness.

Disclosure: Nothing to disclose.

P863. Brain gaba and glutamate are associated with self-reported eating behavior in midlife women

Jessica Busler, Monica Foneska, Aleta Wiley, Vicky Liao, Katherine Breedlove, Laura Holsen, Alexander Lin, Hadine Joffe, Pamela Mahon

Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, United States

Background: The prevalence of severe obesity is highest among midlife women and an increased propensity for obesity in menopause can be related to menopausal factors including hormonal changes, as well as psychological symptoms such as depression, mood swings, and irritability; factors that also have negative impacts on metabolism and energy balance. In addition, obesity-linked fluctuations in mood and stress symptoms promote maladaptive eating behaviors in postmenopausal women thereby potentially exacerbating obesity and related negative outcomes such as cardiovascular disease in this population. Neurobiologically, obesity is linked to aberrant functioning in brain regions involved in impulse control, emotion regulation and mood, such as the anterior cingulate cortex (ACC). Dysregulation in GABAergic (primary inhibitory) and glutamatergic (primary excitatory) systems within this neurocircuitry has been implicated in mood and other disorders, which have increased prevalence in aging women; thus, these systems may play a role in influencing eating behaviors in aging women. However, the relationship of psychological and neurobiological drivers of eating behavior in midlife women is not well understood. Therefore, we aimed to examine brain markers of GABAergic and glutamatergic function in the ACC and their relationship to eating behaviors in midlife women.

Methods: Fifty-one women were recruited at Brigham and Women’s Hospital, with ages 45–65 (M = 55.45, SD = 4.28). Self-report eating behaviors were assessed using the Three Factor Eating Question–Revised 18-item (TFEQ-R18). TFEQ scores for the three subscales representing the domains of Cognitive Restraint (CR), Uncontrolled Eating (UE), and Emotional Eating (EE) were used in analyses to determine relationships with GABA and glutamate. Cognitive Restraint refers to a tendency to consciously restrict one's food intake. Uncontrolled Eating refers to a tendency to feel out of control when overeating. Emotional Eating means the tendency to eat in response to negative emotions. MR spectroscopy, using STEAM at 7 Tesla (TE = 20ms, TM = 10ms, TR = 3000ms), was used to measure markers for GABA and glutamate in the ACC and LCModel was used for neurometabolite quantification. Glu/tCr and GABA/tCr ratios were use as primary MRS measures of interest and TFEQ Cognitive Restraint, Uncontrolled Eating, and Emotional Eating subscale scores as primary eating behavior measures of interest. We conducted separate correlation analyses of each neurometabolic marker with scores from the TFEQ.

Results: We observed a significant inverse correlation between GABA/tCr levels in the ACC and the scores on the Cognitive Restraint subscale of the TFEQ (r = −0.348, p = 0.012) indicating that lower GABA/tCr in the ACC relates to higher cognitive restraint or control of eating. We also observed a trend in the positive association of Glu/tCr levels in the ACC with the Emotional Eating subscale of the TFEQ (r = 0.274, p = 0.051), suggesting that increased eating emotional eating relates to higher Glu/tCr levels in the ACC.

Conclusions: Results suggest that neurometabolic markers in the ACC are linked to eating behaviors in midlife women. We observed that lower GABA levels are associated with higher cognitive restraint indicating the lower GABA in the ACC may be involved in the ability to consciously restrict food intake. We also observed that higher glutamate levels trend toward greater emotional eating implicating the involvement of glutamate in the tendency to eat in response to negative emotions. Interestingly, we did not observe a relationship of GABA or glutamate with uncontrolled eating, suggesting that other mechanisms may better characterize uncontrolled eating. These results provide evidence that GABAergic and glutamatergic function in emotion-regulatory circuitry may contribute to maladaptive eating behaviors that may increase obesity risk in aging women. Further, this work highlights the potential of targeting either or both inhibitory and excitatory neurotransmitter systems in the ACC as novel interventions for obesity and emotional eating in midlife women, a group at heightened risk during the menopausal transition and more susceptible to mood disturbances. Integrating neurobiological markers with behavioral assessments of the control of eating and emotional eating may help identify aging women most vulnerable to weight gain and inform the development of tailored, mechanism-based strategies that address psychological and neurobiological drivers of eating behavior.

Keywords: women's health, proton magnetic resonance spectroscopy, eating behavior, anterior cingulate cortex, cortical GABA and glutamate.

Disclosure: Nothing to disclose.

P864. Neural mechanisms of mother-child dance: a novel approach to maternal depression

Rachel DeLauder, Noor Tasnim, Mackenzie Aychman, Georgia Hodes, Julia Basso

Virginia Tech, Blacksburg, Virginia, United States

Background: One in five mothers experiences a mental health disorder, with profound consequences for both maternal well-being and child development. Pharmacological treatments are ineffective in up to 70% of affected mothers and often produce adverse side effects, underscoring the urgent need for alternative approaches. To address this gap, we developed a mother–child dance intervention designed to reduce depressive symptoms and strengthen the mother–child relationship.

Methods: Mothers with moderate to severe self-reported depression, anxiety, or stress who had a child between the ages of 3 and 5 (n = 26 dyads) were randomly assigned to either a mother–child dance intervention or a mother–child play intervention. Each intervention consisted of ten 1-hour sessions delivered over two weeks (5 sessions per week). Pre- and post-intervention, mothers completed validated self-report measures of mental health. In addition, mother–child dyads underwent hyperscanning electroencephalography (EEG) during interactive tasks, including conversation, cuddling, dancing, and play, to assess changes in neural activity and interbrain synchrony.

Results: We found that compared to the control group, the mother-child dance intervention decreased levels of depression (Beck Depression Inventory, t(14) = −1.791, p = 0.047) and parental stress (Parental Stress Scale, t(15) = −1.974, p = 0.034) and increased positive affect (Positive and Negative Affect Scale, t(15) = 2.107, p = 0.026) and closeness of the parent-child relationship (Child-Parent Relationship Scale, t(5.482) = 2.346, p = 0.031). At the neural level, alpha (8–12 Hz) interbrain synchrony between the mother and child increased from before to after the intervention (t(2) = −8.542, p = 0.013). Finally, those individuals who demonstrated the largest increases in alpha interbrain synchrony had the greatest decreases in depression (r = −0.935, p = 0.032).

Conclusions: These findings suggest that mother–child dance is a promising, non-pharmacological intervention for maternal depression, improving affect, reducing stress, and strengthening the parent–child relationship. Importantly, increases in interbrain synchrony were associated with reductions in depressive symptoms, highlighting a potential neural mechanism through which embodied social interaction promotes mental health.

Keywords: maternal brain, Non-Pharmacological Therapy, mother-child dyads, hyperscanning, interbrain synchrony.

Disclosure: Nothing to disclose.

P865. Temporal subtypes of depressive and anhedonic symptom cyclicity across the menstrual cycle in patients with suicidal ideation

Elizabeth Mulligan, Anisha Nagpal, Tory Eisenlohr-Moul

University of Illinois At Chicago, Chicago, Illinois, United States

Background: Affective symptoms show pronounced menstrual cycle sensitivity in a subset of individuals, with risk for perimenstrual exacerbation (PME) of depression and suicidality documented across multiple psychiatric populations. These exacerbations appear not to result from abnormal hormone levels, but from atypical neural sensitivity to normative ovarian steroid changes, such as midluteal hormone surges and perimenstrual estradiol withdrawal. These sensitivities map onto distinct temporal patterns of affective dysregulation described in the emerging DASH-MC framework, which posits multiple hormone sensitivities with unique neurobiological drivers (e.g., GABAergic and dopaminergic alterations; Peters et al., 2024). Identifying temporal subtypes of symptom cyclicity may clarify heterogeneity in reward and mood circuit vulnerability and help pinpoint individuals at greatest risk for mood worsening and suicidal behavior across the menstrual cycle. The present study, therefore, aimed to (1) determine whether subgroups of temporal symptom cyclicity exist for symptoms of depression and anhedonia, (2) examine the degree of overlap between subgroups across symptom domains, and (3) examine associations between subgroup membership and demographic and psychiatric traits.

Methods: Participants were naturally cycling individuals (n = 155) recruited on the basis of past-month suicidal ideation. Daily ratings of depressed mood (Daily Record of Severity of Problems [DRSP] item 1) and anhedonia/less interest (DRSP item 9) were collected prospectively across ≥1 cycle. Symptoms were cycle-aligned (menses-centered) using the menstrualcycleR R package and modeled using smoothed mixture models to identify latent subgroups of temporal symptom trajectories. This approach allowed us to test whether subtypes correspond to distinct hormone sensitivity patterns described in the DASH-MC framework (e.g., luteal-onset vs. perimenstrual-onset sensitivity). Subgroup solutions were estimated and evaluated separately for depression and anhedonia, and chi-square tests evaluated overlap in membership across symptom domains. Chi-square tests and one-way ANOVAs were used to examine associations between group membership and demographic/psychiatric variables.

Results: While 3- and 4- group solutions exhibited comparable fit, a 3-group solution paralleled the hypothesized hormone sensitivity dimensions according to the DASH-MC framework for both depressed mood and anhedonia (Peters et al., 2024). Regarding depressed mood, we observed a group with perimenstrual exacerbation (n = 47), a group with mid-luteal exacerbation (n = 48), and a group with low and stable symptoms (n = 60). Regarding anhedonia, we observed a group exhibiting perimenstrual exacerbation (n = 41), another group exhibiting early-to-mid-luteal exacerbation with a slight perimenstrual increase (n = 41), and a third group with low and stable symptoms (n = 73). Although both domains exhibited perimenstrual, luteal, and low-stable subtypes, temporal patterns slightly differed across symptoms, suggesting partially distinct mechanistic drivers (e.g., reward vs. mood circuitry sensitivity to ovarian steroids). Subgroup membership was not independent (χ²(4) = 23.49, p < .001). For instance, individuals in the luteal depression group were disproportionately likely to also belong to the luteal anhedonia group, and individuals in the perimenstrual depression group were disproportionately likely to also belong to the perimenstrual anhedonia group. On the other hand, overlap among low/stable depression and low/stable anhedonia groups was a bit weaker. Notably, especially in the case of the anhedonia symptom, the luteal subgroups appeared to partially subsume some perimenstrual exacerbation patterns. Finally, group membership was not associated with demographic variables (e.g., age, SES, education, parity). However, SCID-5 diagnoses of current major depressive disorder (χ²(2) = 5.61, p = .061) and persistent depressive disorder (χ²(2) = 5.73, p = .057) trended toward significance, with greater counts in the mid-luteal exacerbation group for the depressed mood symptom domain.

Conclusions: The current findings suggest that both depressed mood and anhedonia showed distinct temporal subtypes of cyclicity in individuals with suicidal ideation, with overlap strongest between luteal depression and luteal anhedonia subgroups. Temporal subtypes of both symptoms aligned with the DASH-MC framework, potentially reflecting perimenstrual estradiol withdrawal, midluteal allopregnanolone sensitivity, and low-stable symptom courses. Demographics were unrelated to group membership, though depressive disorder diagnoses trended higher in midluteal exacerbation groups. Taken together, these findings highlight heterogeneity in hormone-linked risk processes. In future work, identifying reproducible subtypes of cyclicity may provide a foundation for precision psychiatry approaches, such as guiding biomarker discovery and informing novel preventive strategies (e.g., hormone stabilization, neuromodulation, or targeted behavioral interventions) tailored to distinct hormonally sensitive subtypes and aligned with periods of peak vulnerability.

Keywords: Menstrual Cycle, Depression, Reward, motivation, and anhedonia, suicide.

Disclosure: Nothing to disclose.

P866. Racing hearts and menstrual rhythms: an investigation of cortical processing of cardiac signals during high vs. low progesterone states across the menstrual cycle

Jellina Prinsen, Emily May, Arno Villringer, Julia Sacher

Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany

Background: The resting heart rate in women varies significantly across the menstrual cycle, increasing by on average 2.33 beats per minute (bpm) during the luteal phase. Similarly, cardiac vagal activity, i.e. parasympathetically mediated heart rate variability (HRV), significantly decreases from the follicular to luteal phase. Progesterone (P4), which peaks in the mid-luteal phase, is thought to be a key driver of these autonomic changes. Beyond these peripheral effects, P4 has been shown to modulate activity in cortical areas involved in the processing of cardiac signals. However, it remains unclear whether menstrual cycle-related changes in P4 and cardiac autonomic function are parallelled by altered cortical processing of the heartbeat itself, as indexed by the heartbeat evoked potential (HEP).

Methods: We aim to include 45 naturally cycling, premenopausal female participants. Based on a meta-analytic effect size regarding the 2.33 bpm follicular-to-luteal increase in resting heart rate, this sample size provides 95% power to detect subtle within-person changes (partial η² > .01) in heart rate. At the time of abstract submission, data collection is ongoing; here, we present preliminary data from 37 participants (aged 20–35 years). Each participant underwent two 10-minute simultaneous ECG-EEG resting-state recordings in precisely defined menstrual cycle phases: once in the mid-follicular phase (low P4, expected lower resting heart rate) and once in the mid-luteal phase (peak P4, expected higher resting heart rate). Assessment order was randomized to avoid sequence effects. Saliva samples were also collected to characterize endogenous P4 and estradiol fluctuations.

Assessed cardiac measures of interest include heart rate, vagally mediated HRV, indexed by the root mean square of successive differences in the inter-beat interval (RMSSD), and systolic and diastolic blood pressure. The HEP was obtained by averaging EEG epochs time-locked to the R-peak of the ECG and preprocessed in accordance with recent methodological guidelines. The BrainBeats toolbox was adopted for heart rate analysis and automatic R-peak detection. Within-subject differences in cardiac metrics were assessed using linear mixed models with a random intercept per subject (lmer in R). Potential menstrual phase-related differences in HEPs were evaluated across all 64 EEG electrodes within the 200–600 ms window post R-peak, employing cluster-based permutation testing in FieldTrip.

Results: Within-subject comparisons across menstrual cycle phases confirmed a significant increase in resting heart rate from the mid-follicular to the mid-luteal phase. Specifically, resting heart rate was 3.65 bpm (± 8.87 bpm) per minute higher in the luteal phase (F(1,39.091) = 7.66, p = .008). In contrast, the mostly parasympathetically-mediated HRV, indexed as RMSSD, was significantly lower in the mid-luteal relative to the mid-follicular phase (F(1,41.629) = 4.54, p = .039). These findings emerged in the absence of significant differences in systolic and diastolic blood pressure between cycle phases (all, p > .05). In the current preliminary dataset, cluster-based permutation testing revealed no significant positive or negative clusters in HEP waveform characteristics within the 200–600 ms time window post R-peak, suggesting no detectable differences in cortical processing of cardiac signals between different cycle phases.

Conclusions: These findings further emphasize that the luteal phase of the menstrual cycle is characterized by increased resting heart rate and reduced cardiac vagal activity, suggesting a cycle-related shift in autonomic balance during the P4-dominated luteal phase. Such alterations in peripheral cardiac regulation may contribute to the heightened vulnerability for mental health symptoms observed during this cycle window. In contrast, cortical processing of cardiac signals, as indexed by the HEP, did not show significant follicular-to-luteal modulation. Taken together, these results suggest that menstrual cycle–related changes in heartbeat and cardiac autonomic balance may not directly extend to changes in cortical heartbeat processing, underscoring the need for further work to clarify the conditions under which peripheral–central interactions emerge.

Keywords: Progesterone, Menstrual Cycle, Heartbeat Evoked Potential, Cardiac Autonomic Tone, Women’s Health.

Disclosure: Nothing to disclose.

P867. Hormone-dependent dynamics of hippocampal connectivity across the menstrual cycle

Livia Rühr, Karsten Mueller, Jellina Prinsen, Rachel Zsido, Veronica Witte, Arno Villringer, Julia Sacher

Cognitive Neurology, University of Leipzig, Leipzig, Germany, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany

Background: Ovarian hormone fluctuations parallel neuroplastic changes in brain organization, suggesting a link between hormonal transitions and brain dynamics. The hippocampus, rich in estrogen and progesterone (P4) receptors, exhibits neuroplasticity throughout the lifespan [1,2]. Rodent estrous cycle research indicates neuroendocrine modulation via dendritic spine density, synaptogenesis, axon growth, and dentate neurogenesis [3]. In humans, dense-sampled single-subject studies reveal marked menstrual cycle-linked changes in functional connectivity (FC) [4,5]. Hippocampal FC varies with progesterone (P4) in dorsolateral prefrontal cortex (PFC) and sensorimotor regions [5], and with estradiol (E2) in the superior parietal lobule [6] and default mode network [7]. However, it remains unclear whether these effects replicate in larger samples balancing condition density, menstrual phase, and hormone levels.

Methods: Our sample includes 27 naturally cycling females (aged 19–32), measured at six menstrual cycle phases: menstrual (M), preovulatory (PREO), ovulatory (O), postovulatory (POSTO), midluteal (MIDL) and premenstrual (PREM). Continuous cycle tracking included repeated intravaginal ultrasounds and ovulation tests. At each time point, participants underwent 3T Magnetic Resonance Imaging (MRI) using a Siemens Verio scanner, including anatomical MPRAGE and resting-state functional MRI sequences, behavioral assessments, and serum hormone assessments (E2, P4, lutropin (LH) and follicle-stimulating hormone (FSH)) via liquid-chromatography mass-spectrometry (LC-MS/MS). Imaging data were preprocessed, denoised and band-pass filtered (0.01–0.1 Hz). First- and second-level hippocampus-to-ROI (region of interest) analyses were performed using aal atlas ROIs. First-level FC values were extracted per participant and time point. Linear-mixed effects models (LMEs) tested associations with E2, P4 and time point, including main and interaction effects as fixed terms, and subject as random intercept (lme4 in R). Fixed effects were reduced stepwise. Separate LME sets targeted hippocampal FC with dorsolateral PFC and sensorimotor cortex, and with superior parietal lobule and default mode network. Whole-brain seed-to-voxel analyses assessed hippocampal FC changes across cycle phase pairs, followed by exploratory post-hoc LMEs of significant clusters. The study was preregistered (https://osf.io/kwzxg).

Results: No significant effects were found for P4-only models for hippocampal–lateral PFC FC. However, three-way interaction models (E2*P4*time point) were significant for left hippocampus-right lateral PFC (PREO: ß = −1.182, SE = 0.539, p = .031; PREO*E2: ß = 0.183, SE = 0.091, p = .048; PREO*P4: ß = −0.915, SE = 0.343, p = .009; PREO*E2*P4: ß = 0.158, SE = 0.062, p = .013), and right hippocampus-left lateral PFC (O: ß = 0.926, SE = 0.392, p = .020; O*E2: ß = −0.161, SE = 0.661, p = .017; PREM*P4: ß = 0.453, SE = 0.218, p = .040). No significant models emerged for hippocampal-sensorimotor FC. In the second ROI model set, significant time point*E2 interactions predicted FC between right hippocampus and precuneus (O: ß = −0.820, SE = 0.324, p = 0.013; O*E2: ß = 0.115, SE = 0.051, p = .027). Time point*P4 significantly predicted FC between bilateral hippocampus and ACC (e.g., MIDL: ß = −0.355, SE = 0.111, p = .002; O*P4: ß = −0.355, SE = 0.045, p = .037; MIDL*P4: ß = 0.088, SE = 0.041, p = .035; PREM*P4: ß = 0.165; SE = 0.048, p < .001). Right hippocampus-ACC also showed main P4 effects (e.g., P4: P4: ß = −0.100, SE = 0.038, p = .010). Three-way interaction LMEs were significant for right nucleus accumbens with bilateral hippocampus (e.g., PREO*E2*P4: ß = −0.092, SE = 0.044, p = .040; PREO*P4: ß = 0.742, SE = 0.269, p = .007; PREO*E2*P4: ß = −0.138, SE = 0.049, p = .006; PREM*E2*P4: ß = −0.081, SE = 0.035, p = .024). FC between left hippocampus and PCC was predicted by time point*E2 (O: ß = −0.753, SE = 0.365, p = .041; PREM*E2: ß = −0.084, SE = 0.041, p = .046), and right hippocampus-amygdala by time point*P4 (MIDL: ß = −0.294, SE = 0.110, p = .008). Exploratory seed-to-voxel analyses revealed hippocampal FC changes in middle (MFG) and superior frontal gyri (SFG). Post-hoc three-way interaction LMEs showed significant effects for left hippocampus-left MFG (PREO: ß = 0.113, SE = 0.502, p = .026) and right MFG (PREO: ß = −0.116, SE = 0.534, p = .032; PREO*P4: ß = −0.680, SE = 0.340, p = .048), as well as right hippocampus-left MFG (O*E2: ß = −0.138, SE = 0.067, p = .041; O*P4: ß = −0.668, SE = 0.329, p = .045; PREM*P4: ß = 0.444, SE = 0.329, p = .046). Left hippocampus-right SFG FC was associated with time point*P4 (POSTO*P4: ß = 0.098, SE = 0.041, p = .019; PREM*P4: ß = 0.098, SE = 0.043, p = .045).

Conclusions: Our findings partially replicate prior single-subject evidence suggesting P4 modulation of hippocampus-dorsolateral PFC FC [5]. Whole-brain analyses revealed cycle-related changes in MFG and SFG, regions encompassing dorsolateral PFC, though effects were not independent of E2. We did not replicate prior associations of P4 with hippocampus-sensorimotor FC [5], or E2 with hippocampus-superior parietal lobule FC [6]. Instead, we found E2- and P4-dependent modulation in cingulate and limbic regions. These results support dynamic, hormone-sensitive changes in large-scale brain networks across the menstrual cycle and highlight the complex interplay of ovarian hormone fluctuations in shaping functional brain organization.

Keywords: Ovarian hormones, Human hippocampus, Menstrual Cycle, Human Neuroimaging, Neural Plasticity.

Disclosure: Nothing to disclose.

P868. The retina as a window to sex differences in vascular-brain relationships in youth

Megan Mio, Kody Kennedy, Bradley MacIntosh, Benjamin Goldstein

Centre for Addiction and Mental Health, Toronto, Canada

Background: Imaging studies provide evidence of microvascular dysfunction across organ systems in youth with bipolar disorder (BD). For example, there is evidence that neuroimaging measures of cerebral blood flow (CBF) are sensitive to mood state, and that cerebral metabolic inefficiency contributes to anomalous CBF in BD. Furthermore, epidemiologic studies confirm that the extent to which serious mental illness confers increased cardiovascular risk is more pronounced in women. Retinal imaging comprises an appealing approach to non-invasively studying vascular-brain relationships and related sex differences. Retinal arteriovenous ratio (AVR), a ratio of arteriolar to venular caliber, has been associated with psychopathology, as well as heart-related (blood pressure, body mass index [BMI]) and brain-related measures (brain structure, function) in adults and youth. Population-based studies also confirm that AVR can predict heart attack and stroke risk, above and beyond traditional risk factors.

Methods: 68 youth (N = 38 BD, 30 controls) ages 13–20 years completed 3T arterial spin labeling MRI and retinal fundus imaging on a Topcon 50DX camera at a 50° angle. AVR was computed using standardized procedures. We examined sex differences in AVR and sex differences in the relationship of AVR with global gray matter and region-of-interest CBF (anterior cingulate cortex [ACC], middle frontal gyrus [MFG]). Analyses controlled for age and body mass index. Analyses are underway using Optical Coherence Tomography-Angiography (OCT-A), which yields additional vascular readouts (e.g., retinal vessel density, perfusion density, foveal avascular zone).

Results: Despite a lack of significant sex differences in AVR, greater AVR significantly predicted greater regional CBF in females, but not males, in the overall sample [ACC: Female (β = 0.36 p = 0.03, f2 = 0.36); Male (β = 0.01, p = 0.97, f2 < 0.01), MFG: Female (β = 0.45, p = 0.01, f2 = 0.54); Male (β = −0.14, p = 0.57, f2 = 0.04)]. There was a nominal association between AVR and global CBF in females (β = 0.32, p = 0.06, f2 = 0.25) but not males (β = 0.11, p = 0.57, f2 = 0.09). Diagnosis subgroup analyses were not significant.

Conclusions: In this proof-of-concept analysis, a well-established measure of retinal microvascular health, with relevance to both heart and brain, was associated with cerebrovascular imaging phenotypes specifically in female youth. Lack of diagnosis-related findings should be interpreted cautiously due to small cell sizes. Present findings support the premise that integration of heart-brain perspectives may bolster contemporary treatment approaches among youth, particularly among females. From a methodologic vantage point, retinal imaging circumvents several challenges and barriers related to MRI: it is scalable, inexpensive, quick, and well-tolerated by research participants. Forthcoming OCT-A results will augment the breadth and depth of our vascular phenotyping.

Keywords: Retinal imaging, Cerebral blood flow, Bipolar Disorder, Youth, Sex differences.

Disclosure: Nothing to disclose.

P869. Preeclampsia and early postpartum cognition: a cantab-based analysis of attention, memory, and executive function

M. Mercedes Perez-Rodriguez, Chelsea DeBolt, Valerie Riis

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

Background: Preeclampsia is a major contributor to maternal and neonatal morbidity and mortality. Beyond pregnancy, preeclampsia confers long-term cardiovascular, metabolic, and cerebrovascular health risks, and is associated with dementia and later-life cognitive dysfunction. Despite the impact that cognitive challenges can present postpartum, such as inability to adhere to postpartum health needs and potential to disrupt the mother-infant bond, evidence regarding cognitive function in this vulnerable, early postdelivery period is limited. We examined whether preeclampsia is associated with differences in 3-month postpartum cognitive performance across multiple domains, using the Cambridge Neuropsychological Test Automated Battery (CANTAB).

Methods: We analyzed data from an observational, prospective cohort study of pregnant women, who delivered at an urban, academic institution and completed CANTAB assessment at 3 months postpartum. We compared participants with preeclampsia (n = 62) to those without hypertensive disorders of pregnancy (n = 264). Cases of preeclampsia were defined by standard criteria according to the American College of Obstetricians and Gynecologists: new onset maternal hypertension (≥140/90 mmHg) > 20 weeks’ gestation with or without proteinuria, and/or symptomatic or laboratory evidence of end organ dysfunction. Maternal demographics, medical and obstetric history, and pregnancy outcomes were prospectively collected. Cognitive function was assessed at 3 months postpartum using CANTAB, which provides age-, sex-, and education-adjusted z-scores. We assessed three cognitive domains, including attention (Rapid Visual Information Processing, RVP), memory (Paired Associates Learning [PAL], Spatial Working Memory [SWM], Pattern Recognition Memory [PRM], and Delayed Matching to Sample [DMS]), and executive function (One Touch Stockings of Cambridge [OTS]). Demographic characteristics between those with and without preeclampsia were compared using t-tests or Mann-Whitney U tests (continuous variables) and chi-square or Fisher’s exact tests (categorical variables). Linear regression estimated mean differences in z-scores between groups.

Results: Groups were similar in regard to maternal characteristics, medical and obstetric history, and sociodemographic variables. At 3 months postpartum, mean CANTAB z-scores were lower in the preeclampsia group across the aspects of the memory domain, though differences were not statistically significant. The largest numerical differences were observed for spatial working memory (β = –0.33, 95% CI –0.90 to 0.23, p = 0.26) and delayed match to sample (β = –0.25, 95% CI –0.89 to 0.39, p = 0.44). Rates of mild cognitive impairment across domains (defined as < 1 standard deviation below the mean) were similar between the two groups.

Conclusions: While no statistically significant group differences were detected, women with preeclampsia demonstrated lower mean postpartum cognitive scores in memory, with the greatest differences in spatial working memory and delayed match to sample. These findings suggest subtle memory-related vulnerabilities that merit confirmation in larger cohorts and longitudinal follow-up to evaluate persistence and clinical impact.

Keywords: Cognition, Postpartum Period, pregnancy, Preeclampsia.

Disclosure: Neurocrine, Consultant, Self, Neurocrine, Grant, Self, Mitsubishi Tanabe Pharma, Consultant, Self.

P870. Women using hormonal contraceptives demonstrate greater improvements in emotion regulation and coping skills during residential psychiatric treatment

Seyma Katrinli, Alex Rothbaum, Raneeka DeMoss, Ben Hunter, Abigail Powers, Vasiliki Michopoulos, Alicia Smith

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

Background: Emotional dysregulation (ED), characterized by difficulties modulating emotional responses or using adaptive coping strategies, is a transdiagnostic risk factor for various mood and anxiety disorders. Women in residential treatment exhibit higher ED and greater reliance on maladaptive coping strategies such as rumination, which may be partly exacerbated by ovarian hormone fluctuations. The current study examined the influence of hormonal contraception (HC) on ED and coping strategies in women receiving intensive transdiagnostic evidence-based psychotherapy for ~90 days. We hypothesize that women on HC, who would have relatively stable ovarian hormone levels, would exhibit greater improvements in emotion regulation and coping skills compared to those who are not on HC.

Methods: The study sample included 162 non-menopausal women (ages 18–45), with 75 (46.3%) on HC. Among HC users, 37 used combined methods (e.g., combined oral contraceptive pills, patch), and 38 used progestin-only methods (e.g., progestin-only pills, hormonal IUD, implant). ED was assessed by the Difficulties in Emotion Regulation Scale (DERS), which yields a total score and six subscales: nonacceptance of emotions, difficulties with goal-directed behavior, impulse control difficulties, lack of emotional awareness, limited access to regulation strategies, and lack of emotional clarity. Coping strategies were measured by the Coping Orientation to Problems Experienced Inventory (Brief-COPE), which assesses three overarching coping styles: problem-focused, emotion-focused, and avoidant coping. Associations between HC use, ED, and coping strategies were tested cross-sectionally at admission and discharge using regression models. Longitudinal effects were tested with linear mixed models including an HC × visit interaction term and a random intercept for subjects. Pairwise comparisons evaluated differences across combined HC users, progestin-only users, and non-users.

Results: At admission, HC users reported increased difficulties in engaging in goal-directed behavior when experiencing negative emotions compared to non-users (B[SE] = 0.87[0.43], p = 0.04). HC use was not associated with ED (across total and all subscales) and coping skills at discharge. Women on HC showed greater improvements in overall ED (Bint[SE] = −6.21[2.12], p = 0.004), as well as multiple ED subscales, including nonacceptance of emotional responses (Bint[SE] = −1.27[0.54], p = 0.02), engaging in goal-directed behavior (Bint[SE] = −1.11[0.53], p = 0.04), impulse control (Bint[SE] = −1.00[0.48], p = 0.04), emotional awareness (Bint[SE] = −0.93[0.44], p = 0.04), and emotional clarity (Bint[SE] = −1.05[0.52], p = 0.05). HC use was also associated with greater improvements in problem-focused (Bint[SE] = 2.78[0.83], p = 0.001) and emotion-focused coping strategies (Bint[SE] = 1.82[0.78], p = 0.02). Pairwise comparisons demonstrated greater improvements in nonacceptance of emotional responses (p = 0.02), emotional awareness (p = 0.03), and emotional clarity (p = 0.03) among combined HC users, greater improvements in problem-focused (p = 3e-4) and emotion-focused (p = 0.02) coping strategies among progestin-only users, and greater improvements in overall ED in both combined (p = 0.007) and progestin-only users (p = 0.05) compared to non-users. No differences emerged between combined and progestin-based HC users.

Conclusions: The current findings suggest that while contraceptive-induced low estradiol and progesterone states may initially pose challenges for engaging in goal-directed behavior when experiencing negative emotions, the hormonal stabilization provided by HCs may reduce emotional volatility and facilitate consistent application of therapy skills. Considering ovarian hormone dynamics alongside psychological interventions may help refine treatment approaches for women with severe psychopathology.

Keywords: Emotional dysregulation, hormonal contraceptive use, coping.

Disclosure: Nothing to disclose.

P871. Dynamism of the inactive X chromosome and its role in adult female brain plasticity

Luisa Demarchi, Ana Milosevic, Jordan Rowley, Marija Kundakovic

Fordham University, Bronx, NYC, New York City, New York, United States

Background: Women experience a higher prevalence of brain disorders such as anxiety, depression, and Alzheimer’s disease compared to men, yet the molecular basis of this sex bias remains poorly understood. Two features of female biology—cyclical ovarian hormone fluctuations and the presence of two X chromosomes—are well recognized, but whether and how they converge to shape brain function is unknown. While ovarian hormones are regulators of mood, cognition, and neuronal plasticity, X chromosome biology has been largely underexplored. In particular, the inactive X chromosome (Xi) has long been viewed as repressed and static in the adult female brain, with the exception of a subset of so-called X-escapee genes known to escape X chromosome inactivation. In our recent study, we described putative “estrous cycle dependent X escapees” associated with substantial dynamism in 3D genome structure of the X chromosome in ventral hippocampal neurons across the mouse estrous cycle. This raised a provocative question: could the inactive X chromosome, usually considered permanently repressed, participate in ovarian hormone-dependent regulation of the female brain? If so, hormone-dependent Xi activity might represent an entirely unrecognized layer of sex-specific brain plasticity—reshaping our understanding of female resilience and vulnerability to brain disorders.

Methods: We employed the ΔXist/CAST mouse model that allows us to perform allele-specific genomics analyses and distinguish between active (Xa) and inactive (Xi) X chromosomes and confirm hormone-induced Xi plasticity. Ventral hippocampal (vHIP) mature neurons from 11-week-old females and males were sorted using FACS and profiled using unbiased genomic assays: RNA-seq (gene expression), ATAC-seq (chromatin accessibility), and Hi-C (3D genome organization). Female cohorts included proestrus (high estrogen, low progesterone), diestrus (low estrogen, high progesterone), and ovariectomized (OVX) mice (N = 6/group), with estrous cycle stages determined by vaginal cytology across three consecutive cycles. Gene expression was quantified using RNA-seq with TPM normalization and chromatin accessibility was measured with ATAC-seq using BPM normalization. Chromatin organization features were identified, including CTCF loops (SIP), compartments (POSSUMM), and enhancer–promoter contacts (FitHiC). To assess the functional impact of Xi, we also analyzed 39, X0 females (lacking the inactive Xi) including behavioral assays for anxiety- and depressive-like phenotypes (N = 12/group).

Results: The ΔXist/CAST model enabled robust distinction between Xa (maternal) and Xi (paternal) X chromosomes. We identified 24 high-confidence Xi escapee genes (~3% of X-linked genes), including both traditional and previously unrecognized escapees. Expression of a subset of these genes implicated in intellectual disability, anxiety and depression, including Ddx3x, 5530601H04Rik, and Grpr, varied dynamically across the estrous cycle. Overall, Xi-derived escapees showed higher expression in diestrus than proestrus. ATAC-seq further confirmed allele-specific chromatin remodeling, with escapee promoters displaying increased accessibility during diestrus. Importantly, these cycle-dependent molecular dynamics suggest that Xi activity contributes directly to estrous stage–specific modulation of hippocampal function. Consistent with this, 39,X0 females (lacking Xi) exhibited behavioral changes, underscoring Xi’s critical role in female brain physiology.

Conclusions: We demonstrate an unprecedented dynamism of Xi in the adult female brain, establishing Xi as a previously unrecognized, hormone-responsive layer of female brain plasticity. By uncovering an unexpected dimension of X chromosome biology, this work provides fundamental insights into sex-specific mechanisms of brain function and resilience, with broad implications for female vulnerability and adaptation in health and disease.

Keywords: Sex hormones, X chromosome, Brain plasticity, 3D genome, Epigenetics.

Disclosure: Nothing to disclose.

P872. Antihostility Effects of Xanomeline and Trospium Chloride in Subjects With Schizophrenia: Post Hoc Pooled Analyses of EMERGENT-1, EMERGENT-2, and EMERGENT-3 Trials

Leslie Citrome, Scott Vuocolo, Ronald Marcus, James Appio, Pierre Nicolas, Amy Claxton

Bristol-Myers Squibb, Darnestown, Maryland, United States

Background: In addition to hallucinations and delusions, people with schizophrenia may exhibit hostility, which is associated with aggressive and violent behavior during acute exacerbations. Prior research has demonstrated the specific antihostility effects of several atypical antipsychotics. The objective of this post hoc analysis of 3 acute trials was to assess the efficacy of xanomeline and trospium chloride (X/T) in reducing hostility in subjects with schizophrenia.

Methods: Pooled data from the 5-week, randomized, double-blind, placebo-controlled, inpatient EMERGENT-1 (NCT03697252), EMERGENT-2 (NCT04659161), and EMERGENT-3 (NCT04738123) trials were assessed. Subjects aged 18-65 years experiencing acute exacerbation of psychotic symptoms received placebo or X/T titrated over 7 days to 125mg/30mg twice daily. Mean change in the Positive and Negative Syndrome Scale (PANSS) hostility item (P7) score was assessed, and pseudospecificity was addressed by adjusting for positive symptom change and somnolence.

Results: Subject demographics and disease characteristics were similar across groups. In individuals with baseline PANSS hostility scores ≥2, ≥3, or ≥4, improvement in hostility from baseline to week 5 was greater among those receiving X/T than in those receiving placebo. This improvement was maintained when adjusted for PANSS positive symptom change and somnolence.

Conclusions: This post hoc analysis of data pooled from the short-term EMERGENT trials demonstrates that X/T has a specific antihostility effect compared with placebo in subjects experiencing an acute exacerbation of schizophrenia.

Keywords: xanomeline, Schizophrenia (SCZ), hostility, panss

Disclosure: Bristol-Myers Squibb, Employee, Self.