ACNP 63rd Annual Meeting: Panels, Mini-Panels and Study Groups

December 8-11, 2024

Phoenix, Arizona

Sponsorship Statement: Publication of this supplement is sponsored by the ACNP. All content was reviewed and selected by the Program Committee, which held full responsibility for the abstract selections. Only disclosures for presenting authors are listed. Individual contributor disclosures may be found within the abstracts.

Mini Panel

1. Amplifying Patient Perspectives in Recovery: Utilizing Patient-Reported Outcomes to Predict Relapse Beyond Conventional Treatment Settings

1.1 Healing and Higher Education: Student Experiences, Identity, and Wellbeing in Collegiate Recovery

Noel Vest

Boston University School of Medicine, Boston, Massachusetts, United States

Background: Collegiate recovery programs (CRPs) provide necessary support for students recovering from substance use disorders (SUDs) and other addictions. These programs create a welcoming environment that fosters academic success and personal growth. While the prevalence of addiction and mental health issues among college students is well-documented, there is limited research on the perspectives and experiences of students in these recovery programs. Understanding these student characteristics is essential to enhance the effectiveness of CRPs and to value the lived experiences of students in recovery.

Methods: This study surveyed 247 college students participating in CRPs across the United States. The survey collected demographic information, recovery pathways, mental health diagnoses, substance use treatment history, and involvement with the criminal justice system. Quantitative and qualitative data were analyzed to characterize student experiences and identify common themes and trends.

Results: A national survey of collegiate recovery programs (CRPs) revealed diverse recovery pathways and significant mental health challenges among students. Specifically, we found that 69.7% were in recovery from substance use disorder and 56.2% in recovery from alcohol use disorder (AUD). The sample was diverse as 13% identified as non-binary, 8.5% were transgender, and 49% were non-heterosexual. Common recovery modalities included 12-step programs (56.2%), therapy/counseling (51.7%), CRP as primary (47.2%), and harm reduction (22.1%). Nearly half (47.8%) had never been to formal SUD or AUD treatment and 22.6% had been prescribed medications for opioid use disorder in their lifetime. Seventy-nine percent had been diagnosed with depression and 76.7% with anxiety. In themes around lived experience driving research common themes were (1) understanding the importance of personal narratives, (2) the program’s impact on identity as a person in recovery, and (3) stigma around harm reduction.

Conclusions: The survey results are the first to report the variety of identities and diversity of recovery pathways among college students. The high prevalence of mental health diagnoses underscores the need for integrated mental health and recovery support. The significant involvement in 12-step programs, lack of formal treatment, and high numbers in therapy reflect the multifaceted approach required for successful recovery.

These outcomes emphasize the critical role CRPs play in supporting students’ recovery and academic success. Valuing the lived experiences of these students in research provides insights that can inform the development of more effective, inclusive, and supportive CRP models. By recognizing and addressing the unique challenges faced by students in recovery, educational institutions can better support their journey toward wellness and academic achievement.

Disclosure: Nothing to disclose.

1.2 Use Social Media Language to Predict Substance Use Disorder Dimensions of Treatment Outlook

Tingting Liu

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

Background: Treatment outcomes for substance use disorder (SUD) are difficult to predict and are influenced by individual psychological factors. This study investigates the link between linguistic patterns and determinants of treatment.

Methods: 504 outpatients (Mage = 33, 33.56% female) completed the Addiction Severity Index (ASI) at the onset of their substance use disorder (SUD) treatment program. Among them, 204 individuals with Facebook posts exceeding 200 words were analyzed. Linguistic features, including categories from the Linguistic Inquiry and Word Count (LIWC) and topics identified using Latent Dirichlet Allocation (LDA) machine learning models, were extracted from Facebook posts spanning two years before treatment. These features were then correlated and compared with key psychological determinants of treatment obtained from the ASI.

Results: Our findings revealed that individuals troubled by cravings or urges to use substances in the past 30 days tended to use more health-related words (d = 0.15) and articles (d = 0.14), and fewer informal expressions (d = −0.12), anger-related terms (d = −0.11), swearing (d = −0.11), and sexual content (d = −0.11), compared to those who did not, after controlling for age and gender. This suggests a shift towards a less informal but more health-focused and detail-driven language during cravings. Additionally, the importance placed on achieving drug abstinence was negatively correlated with numerical expressions (r = −0.26), and with topics about hills and trees (r = −0.414), reading and writing stories and poetry (r = −0.401), and gun shooting (r = −0.343). This suggests that individuals emphasizing abstinence may be less engaged in discussing these specific themes, potentially reflecting a narrower focus on their recovery. Furthermore, compared to those who did not, individuals who experienced medical or psychological problems due to substance use in the past 30 days discussed more topics related to impact and transformation (d = 0.313), such as “mass,” “effect,” “starting,” and “full.” This indicates that individuals facing these issues are more likely to talk about the significant changes and effects of their experiences. All results were controlled for age and gender, and all were significant at p < 0.05 with Benjamini-Hochberg False Discovery Rate correction.

Conclusions: Our study has important implications for understanding the psychological and linguistic profiles of individuals dealing with substance use issues, suggesting that therapeutic interventions could benefit from tailoring communication strategies to promote better recovery outcomes.

Disclosure: Nothing to disclose.

1.3 Sensing and Algorithms for Precision Mental Health With Substance Use Disorders

John Curtin

University of Wisconsin—Madison, Madison, Wisconsin, United States

Background: Patient reported outcomes collected by ecological momentary assessment (EMA) and passively sensed geolocation and smartphone communications can serve as inputs to machine learning models to pursue precision mental health goals. For individuals with substance use disorders, these models can be used personalize support by identifying both when and how best to intervene to prevent return to substance use (i.e., lapses) following initial treatment.

Methods: We followed 151 participants in early remission from moderate to severe alcohol use disorder for up to three months. Participants were committed to alcohol abstinence and reported all alcohol lapses during the study period. We also collected 4x EMAs of their subjective experiences (e.g., craving, affective state, stressors) and passively sensed their geolocation and smartphone communications. We used the EMA and sensed data as inputs to several machine learning models that predict future lapses with increasingly higher temporal resolution (i.e., periods of one week vs. one day vs. one hour). Model performance was evaluated in held-out, test sets using grouped, nested cross-validation. We focus on the Area under the Receiver Operating Characteristic Curve (auROC) as our primary performance metric and use Shapley Additive Explanations (SHAP) to examine global and local feature importance.

Results: Models using EMA as inputs displayed exceptionally high performance when predicting future lapses in the next week (auROC = 0.90), next day (auROC = 0.91), and next hour (auROC = 0.93). The models also perform well (auROCs > 0.85) when lagged to predict lapses up to 2−3 weeks in the future. Global feature importance metrics (mean absolute SHAP values) confirm that all EMA items contributed to predictions and document their relative importance. Local feature importance scores can also be used to identify personal putative risk factors for a lapse for any specific individual and moment in time. Analyses of algorithmic bias indicated poorer model performance for some protected sub-groups due to either low representation in our training data or gaps in domain expertise due systemic under-representation of marginalized groups in past research and theory.

Conclusions: Machine learning prediction models can use EMA and sensed data to identify both when and how to best to support individuals to prevent future alcohol lapses. Our current models already perform at levels sufficient for clinical use but issues related to algorithmic fairness must be addressed prior to implementation. Mappings between locally important features and recommended supports can now be established initially using clinical domain expertise but also eventually derived by reinforcement learning. Features based on sensed geolocation and smartphone communications can contribute additional features that may further improve performance, reduce bias, and/or identify novel risk factors.

Disclosure: Nothing to disclose.

Study Group

2. Data Sharing is Important, but How Should It be Done and Who Will Pay for It?

Robert Innis, Taunton Paine, Gregory Farber, Russell Poldrack, Adam Ferguson, Melanie Ganz Benjaminsen, Melissa Kline Struhl, Dost Ongur

National Institute of Mental Health, Bethesda, Maryland, United States

Study Group Summary: This Study Group will explore the “carrot and stick” motivations to share research data. The “carrot” derives from the fact that shared data have significant benefits for the entire research community and from the belief that sharing increases both transparency and reproducibility. The “stick” comes from the long-anticipated policy that NIH-funded researchers provide “recorded factual material of sufficient quality to validate and replicate research findings” for studies after January 2023.

A Study Group at last year’s ACNP meeting explored the use of checklists to improve methodology, monitoring, and reporting. This Study Group will extend that discussion by explaining the NIH’s current requirements for sharing results and how that can be done in a meaningful way. Simply dumping data into an archive is not sufficient for a researcher to conduct a secondary data analysis. Meaningful sharing requires the data to be in a standard format that is readable by both humans and machines and easily understood by other researchers.

The speakers reflect three relevant perspectives: NIH as a funding agency, researchers who wish to use the data, and archives that can store the data to fulfill NIH and journal requirements.

Russell Poldrack (Stanford) will begin the discussion by reviewing the need for and utility of data sharing to improve transparency and reproducibility. He will outline the FAIR principles (Findable, Accessible, Interoperable, and Reusable) that define best practices for data sharing and show how they are implemented within the OpenNeuro data archive.

Taunton Paine (NIH) will review NIH’s requirements and the various mechanisms that both intramural and extramural researchers use. Greg Farber (NIMH) will review the practices and future directions of the NIMH Data Archive (NDA); all NIMH-funded researchers must now deposit their human subjects data in the NDA.

Adam Ferguson (UCSF) will discuss distributed multi-modal data sharing in two NIH-supported specialist repositories (Open Data Commons for Traumatic Brain Injury and for Spinal Cord Injury). He will present strategies to increase community adoption of data sharing and examples of how data reuse may accelerate both discovery and bench-to-bedside translation of therapeutic strategies.

Melanie Ganz Benjaminsen (U. Copenhagen) will describe OpenNeuroPET, which is both an archive and a data analysis pipeline for PET. The example of PET data sharing is especially relevant, since PET imaging is inherently multi-modal and relies on sharing a combination of PET images, MR images, plasma data, and information on the research radioligands.

Melissa Kline Stuhl (MIT) will describe the Psych-DS standard for organizing and sharing data that include behavioral measurements. She will focus on the tools needed to make it fast and easy for researchers with a variety of backgrounds to begin using the standard, and the immediate “carrots” that standardization can provide to researchers even prior to data sharing. She will also discuss how the Psych-DS standard can support both reproducible analysis and data reuse by enabling merging and extraction of smaller datasets.

Dost Ongur (McLean Hospital and Editor, JAMA Psychiatry) will review the requirements and best practices to share data in publications.

Last, but not least, the speakers and the audience will discuss ways that NIH’s reporting requirement, often viewed as an “unfunded mandate,” can/should be paid for.

Disclosure: Nothing to disclose.

Panel

3. Effects of Adverse Early Life Experiences on Cognitive-Emotional Health: The Role of Gut Microbiome-Neuroimmune Axis

3.1 Abstract Not Included

3.2 Engineered Autoinducer-2 Producing Bacteria Control Susceptibility to Depressive-Like Behaviors

Eleonore Beurel

Miller School of Medicine/University of Miami, Miami, Florida, United States

Background: The gut-brain axis is impacted by stress, and stress affects the microbiome composition to promote stress-related behaviors. Yet the mechanisms whereby microbes mediate the stress response remain poorly understood. We uncovered that a bacterial chemical product called autoinducer-2 promotes susceptibility to depressive-like behaviors in mice by promoting a subset of T helper 17 (Th17) cells.

Methods: Male and female mice were gavaged with engineered bacteria that produce high level of autoinducer-2 or the control strain and 3 days later, assessed for behaviors (e.g., open field, tail suspension, learned helplessness, novel object recognition and Y maze). After behaviors, cells were isolated from the hippocampus for flow cytometry analyses or brains were stained for microglia (e.g., Iba1) or dendritic spine density was evaluated in Thy.1-GFP mice.

Results: Mice receiving engineered autoinducer-2 producing bacteria exhibited increased susceptibility to the learned helplessness paradigm and impairment of learning and memory, while they did not exhibit locomotor deficit or sickness behavior. Associated with these behaviors, there was an increase of Th17 cells and microglia and a decrease of the dendritic spine density in the hippocampus after gavage with the engineered autoinducer-2 producing bacteria. The behaviors, the microglial increase and the decrease of dendritic spine density were normalized in Th17 cell deficient mice and in mice treated with neutralizing antibodies for interleukin-17A, the signature cytokine of Th17 cells, showing that Th17 cells and IL-17A mediate bacterial autoinducer-2 effects in behaviors, microglial increase, and dendritic spine density decrease.

Conclusions: Our findings show that autoinducer-2 promotes susceptibility to depressive-like behaviors by promoting the Th17 cell response. Once Th17 cells are in the hippocampus, they seem to increase microglia and decrease dendritic spine density, contributing to depressive-like behaviors and learning and memory impairment. This study provides a proof of concept that engineered bacteria could be used to modulate behaviors.

Disclosure: Nothing to disclose.

3.3 Abstract Not Included

3.4 Higher Levels of Inflammation are Associated With Accelerated Brain Aging in Adolescents

Tiffany Ho

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

Background: Elevated inflammation has been linked with depressive neurophenotypes (e.g., smaller hippocampal volume, cortical thinning, etc.). Similarly, ‘brain age’ represents a proxy for brain health that may go awry in depression, with subtle evidence of accelerated brain aging in depressed adults. It is unclear, however, whether inflammation may contribute to accelerated brain aging, particularly in adolescents who are less likely to have age-related confounds (e.g., chronic health conditions).

Methods: 100 somatically healthy adolescents (13-18 years) were recruited from the San Francisco Bay Area as part of two independent longitudinal neuroimaging studies on adolescent depression and risk for depression: 57 were diagnosed with current or past history of a depressive disorder, 34 were psychiatrically healthy, and 9 met criteria for an Axis I disorder other than a depressive disorder. All participants completed a high-resolution T1-weighted MR scan at 3T and provided dried blood samples, from which we assayed peripheral markers of inflammation (IL-1β, IL-6, and TNF-α). We used FreeSurfer version 6.0 to estimate gray matter morphometry and applied ENIGMA MDD’s BrainAge algorithm to estimate brain age. We examined linear associations between the difference in the estimated brain age and chronological age (‘brain age gap’) with each peripheral marker of inflammation and tested whether diagnostic group moderated these associations.

Results: Brain ages ranged from 7.73-34.89 years (mean: 23.67 ± 7.11 years); brain age gaps ranged from -7.41 to 24.8 years (mean: 7.42 ± 7.04 years). Higher levels of IL-1β and IL-6 were associated with greater brain age gaps, even after covarying for sex, chronological age, and diagnostic status (all ps < 0.025). History of psychiatric illness did not moderate these associations.

Conclusions: Our results are consistent with prior work in adults showing that brain age gaps are minimal between depressed and non-depressed individuals. Nevertheless, inflammation appears to be an important process that may explain accelerated brain aging, possible through the neurotoxic effects of the kynurenine pathway; future studies are needed to definitively test this formulation.

Disclosure: Nothing to disclose.

Panel

4. Big Data Neuroimaging and the Promise to Predict Illnesses Before Onset

4.1 From Diagnosis to Discovery and Back Again: Dynamic and Multimodal Neuroimaging Signatures of Psychosis and Mood Disorders Along a Continuum Using Flexible Large Scale Data Mining Approaches

Vince Calhoun

Georgia State, Georgia Tech, Emory, Atlanta, Georgia, United States

Background: The field of psychiatric neuroimaging has struggled with capturing heterogeneity while still offering a generalizable modeling framework. Approaches vary between identifying group differences, data-driven clusters, or individual results. Here we propose combining these approaches together by modeling multimodal neuroimaging and clinical data along a continuum. Our goal is to unravel the complex patterns of brain structure and function that correspond to the diverse manifestations of mental illnesses.

Methods: We present a powerful strategy to combine cutting-edge data-driven approaches (e.g., multi-level constrained source separation, dynamic modeling, and deep learning) with big data analysis to derive canonical networks, functionally independent primitives, and dynamic states. We model the data in a way that preserves individual variability, while also allowing us to study the relationship to existing diagnostic categories and enabling derivation of risk or to identify natural clustering in the data. We present new results using structural and functional MRI data collected from a large (N > 6K) transdiagnostic clinical dataset, individuals at risk for psychosis (N > 1K) and in over 10K individuals collected from non-clinical individuals.

Results: Results identify multimodal dynamic neural signatures showing significant differences among psychiatric disorders. Unique signatures include heightened cerebellar connectivity in schizophrenia and sensory domain hyperconnectivity in autism spectrum, with both disorders showing transient reduced cerebellar-subcortical connectivity. Cross-validated prediction shows accuracies above 90%. Results also show smooth transitions between disorders. Notably, certain signatures exhibit disorder-specific characteristics, while others demonstrate commonalities across disorders. These signatures also manifest in a small percentage of non-clinical adolescents and are linked to alterations in behavior.

Conclusions: We present a novel neuroimaging approach combining individual, group, and subgrouping strategies. Our focus on visualization, individual variation, and transdiagnostic relationships highlights the rich information available and reinvigorates the potential of using such approaches to uncover the neural links to mental disorders.

Disclosure: Nothing to disclose.

4.2 Structural and Functional Regional Vulnerability Indices in Youth Psychotic-Like Experiences With and Without Depressive Symptoms

Nicole Karcher

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

Background: Development of early biomarkers of serious mental illness will facilitate early intervention efforts. We have developed a novel biomarker and find that youth persistent distressing psychotic-like experiences (PLEs) show structural neural impairments similar to adult neuropsychiatric disorders, including schizophrenia. According to dimensional perspectives of psychopathology, endorsement of additional depressive symptoms may confer more pronounced impairments. The current work examines whether individuals endorsing persistent distressing PLEs with depressive symptoms, compared to those without, show the greatest neural deviations in the direction of depression and schizophrenia as measured by structural and functional regional vulnerability indices (RVI).

Methods: The first three waves of Adolescent Brain Cognitive Development (ABCD) Study data were utilized to categorize participants into groups based on endorsement of persistent distressing PLEs with or without depressive symptoms. Structural and functional RVI were created by quantifying the similarity of participants’ baseline cortical and subcortical structural and functional (e.g., white matter regional homogeneity) neuroimaging measures to the expected patterns found for in adult neuropsychiatric samples for depression and schizophrenia.

Results: In comparison to the low-level PLEs, the persistent distressing PLEs with depressive symptoms group showed higher schizophrenia structural cortical RVI (95% CI = 0.218-1.546), depression structural cortical RVI (95% CI = 0.078-1.382), and schizophrenia white matter regional homogeneity RVI (95% CI = 0.096-1.509). The persistent distressing PLEs groups with and without depressive symptoms only differed on the white matter regional homogeneity RVI for schizophrenia (95% CI = 0.061-0.253).

Conclusions: Presence of depressive symptoms in healthy pre-pubescent children confers additional impairments above those from persistent distressing PLEs. Persistent distressing PLEs with depressive symptoms may represent a higher risk group for psychopathology that can be targeted in early intervention efforts. Findings indicate that white matter based RVI metrics may be particularly sensitive to impairments for youth with persistent distressing PLEs with depressive symptoms.

Disclosure: Nothing to disclose.

4.3 Incorporating Normative Modeling Techniques into Neuroimaging Summary Scores to Identify Clinical-High Risk Individuals Who Develop Psychosis

Maria Jalbrzikowski

Boston Children’s Hospital/Harvard Medical School, Boston, Massachusetts, United States

Background: We recently developed a neuroimaging summary score (i.e., Psychosis Neuroscore) that integrates multifactorial aspects of brain structure alterations in schizophrenia. In previous work, the Psychosis Neuroscore differentiated between schizophrenia cases and controls (Cohen’s d [d]=0.56) to a greater degree than individual brain regions. Here, we leveraged data from the Enhancing NeuroImaging Genetics Meta-Analysis (ENIGMA) CHR Working Group and examined the ability of Psychosis Neuroscore to differentiate CHR who go on to develop psychosis (CHR-PS + ) from healthy controls (HC) and CHR who do not develop psychosis (CHR-PS-). However, because important structural brain changes occur over development, this “adult score” may not directly translate to accurate identification in young people with CHR. Thus, we used normative brain charts to obtain deviation scores for individual brain regions. We incorporated these deviations into the Psychosis Neuroscore and tested our ability to differentiate between CHR-PS + , CHR-PS-, and HC.

Methods: We pooled cross-sectional T1-weighted magnetic resonance imaging (MRI) data from 26 international sites included in the ENIGMA Clinical High Risk for Psychosis Cohort (N = 2681,9-40 years). CHR status was assessed using the Comprehensive Assessment of At-Risk Mental States or the Structured Interview for Prodromal Syndromes, at the baseline and follow-up visits. We processed MRI scans using harmonized ENIGMA structural protocols.

Results: In comparison to HC, CHR individuals had significantly higher Psychosis Neuroscores (d = 0.24, p = 7.5e−10). CHR individuals who later developed psychosis (CHR-PS + ) did not significantly differ from CHR who did not go on to develop psychosis (CHR-PS-; d = 0.07, p = 0.28). After we integrated measures of deviation from normative brain charts into the Psychosis Neuroscore, CHR-PS+ had significantly elevated Neuroscores in comparison to both HC (d = 0.39, p = 3.2e−08) and CHR-PS- (d = 0.15, p = 0.03).

Conclusions: Predictive indices for psychosis in young must account for both the widespread nature of structural brain alterations observed in psychosis, as well as changes in brain development during adolescence. By combining these methods, we will continue to improve our ability to identify neurobiological markers of psychosis in the ENIGMA CHR cohort.

Disclosure: Nothing to disclose.

4.4 Tracking the Longitudinal Effects of Risk Factors for Schizophrenia on Brain Development and Explaining the Development of Persistent and Disturbing Psychotic-Like Experiences in Adolescents

Yizhou Ma

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

Background: Schizophrenia spectrum disorder (SSD) can have lifelong impact on independence and quality of life. We hypothesized that environmental and genetic risk factors for SSD act insidiously during childhood and adolescence, leading to SSD-specific deficit patterns that bring about the onset of symptoms. We are developing individual Big Data-informed biomarkers to identify individuals who may be at risk for SSD.

Methods: We develop the regional vulnerability index (RVI). RVI is a novel type of biomarker that quantifies the agreement between individual brain patterns with those derived from Big Data adult SSD findings using structural and diffusion-weighted MRI. We tested RVI in healthy adolescents in the longitudinal Adolescent Brain Cognitive Development study (baseline: N ~ 11,800, age 9−10; 2-year follow-up: N ~ 10,400, age 11−12). We examined the relationship between RVI and genetic and environmental risk factors for SSD. We also examined longitudinal changes in RVI and development of persistent and disturbing psychotic like experiences (PLEs), i.e., unusual thought content and perceptual experiences that fall under the psychosis spectrum.

Results: At baseline, elevated RVI was significantly associated with family loading of SSD (p = 3 × 10-3) and environmental risks (pFDR = 1.8 × 10-25-4.9 × 10-2) including delayed verbal development, caregiver unemployment, parental divorce, etc. RVI was affected by interaction between ‘fixed’ (e.g., early motor development) and ‘modifiable’ (e.g., parental warmth) risk factors (pFDR = 1.1 × 10-4-4.9 × 10-2). Longitudinal rise in individual RVI values were significantly associated with increases in environmental risks including sleep problems (pFDR = 7.0 × 10-3), school disengagement (pFDR = 3.1 × 10-2), and stress reactivity (pFDR = 3.1 × 10-2). Finally, baseline RVI showed significant association with PLEs ascertained at the 2-year follow-up (p = 1 × 10-6), and longitudinal changes in RVI approached significance in association with changes in PLEs (p = 0.07).

Conclusions: Big Data-informed indexes of individual brain vulnerability are sensitive to the cumulative effects of genetic and environmental risk factors for SSD and subclinical symptoms of psychosis. Longitudinal changes in individual vulnerability track with changes in environmental risks and worsening of pre-clinical symptoms.

Disclosure: Nothing to disclose.

Panel

5. The Highs and lows of Exploring the Endocannabinoid System as a Novel Target for Stress Related Psychiatric Disorders

5.1 Mesocorticolimbic Cannabinoid Receptor 1 Signaling Regulates Repeated Stress-Cocaine Interactions in Rats

Jayme McReynolds

University of Cincinnati, Cincinnati, Ohio, United States

Background: Clinical evidence implicates stress as an important contributing factor to substance use disorder (SUD) and understanding the neurobiological mechanisms of stress in SUD is critical. Repeated stress daily at the time of cocaine self-administration (SA) produces significant increases in cocaine SA in rats that also leads to increased cocaine-seeking behavior. These effects likely involve signaling systems at the interface of stress and reward, such as the endocannabinoid (eCB) system. We are examining the role of the cannabinoid receptor 1 (CB1R) signaling in mesocorticolimbic brain regions on stress-enhanced cocaine SA and cocaine-seeking behavior.

Methods: Male and female rats underwent cocaine SA in 4 × 30 min SA blocks separated by 4 × 5-min drug-free blocks, during which intermittent footshock stress was administered. Rats underwent extinction training followed by cocaine-primed reinstatement tests. To examine the role of CB1R, rats received intra-ventral tegmental area (VTA), nucleus accumbens (NAc) shell, or prelimbic cortical (PrL) injections of a CB1R antagonist prior to SA or reinstatement testing. We are examining CB1R-regulation of cocaine-evoked dopamine with in vivo fiber photometry measurement of dLight 1.3b in the NAc shell.

Results: Footshock stress significantly increases cocaine SA in male and female rats (p < 0.05). Systemic CB1R antagonist attenuates cocaine SA in male and female rats with a prior history of stress (p < .001). In male rats, we have localized this effect to the NAc shell (p < .05) and VTA (p < 0.05). Using dLight and fiber photometry in the NAc shell, we found that female rats have greater cocaine-evoked dopamine (p < 0.01) than males and CB1R antagonist lower doses attenuate cocaine-evoked dopamine only in females (p < .05). In both males and females, intra-PrL CB1R antagonist attenuates cocaine-primed reinstatement only in rats with a history of stress (p < 05).

Conclusions: Repeated stress likely recruits CB1R signaling in mesocorticolimbic regions to regulate cocaine SA, through regulating cocaine-evoked dopamine, and cocaine-seeking behavior. Females may be more sensitive to CB1R regulation though this requires further exploration. Understanding the unique mechanisms by which stress can drive drug use and seeking is important and these data implicate CB1R signaling as a potential therapeutic target.

Disclosure: Nothing to disclose.

5.2 Abstract Not Included

5.3 Cortical Mechanisms Subserving Endocannabinoid Modulation of Fear Generalization and Extinction

Sachin Patel

Northwestern University, Chicago, Illinois, United States

Background: Excessive fear generalization and impaired extinction after trauma exposure contribute to the pathophysiology of posttraumatic stress disorder. Endogenous cannabinoids (eCBs) have been heavily implicated in the regulation of fear and stress-related bio-behavioral processes and eCB-based therapeutics are being investigated for the treatment of stress and trauma-related disorders. However, the cellular and circuit-level mechanisms by which eCBs regulate fear extinction and generalization are not well understood.

Methods: We administered the 2-AG synthesis inhibitor DO34 to male and female mice to determine the role of 2-AG in the regulation of fear extinction and generalization using an auditory cue and contextual fear conditioning paradigm. In vivo calcium imaging of prelimbic prefrontal cortical neurons was conducted to determine how 2-AG depletion affects neuronal activity and coding across fear conditioning, extinction, and generalization testing.

Results: At the behavioral level, 2-AG synthesis inhibition impaired fear extinction and enhanced generalization in a variety of paradigms cue and context-based experimental designs. Calcium imaging data revealed conditioned stimulus (CS + ) and unconditioned stimulus (US; shock) responsive neurons, and neurons that respond to generalization stimuli (novel tones; NTs). Pharmacological 2-AG depletion increased the proportion of neurons responding to both CS+ and NTs representing a neural correlate of fear generalization. 2-AG depletion also reduced Mahalanobis distance between CS+ and NT in PCA space and reduced within-session representational drift within this sensory-responsive ensemble. Ongoing analyses of PAG-projecting prelimbic cortical neurons and hippocampal cortex neurons during fear extinction will provide additional insights into how 2-AG regulated extinction and generalization at the cellular ensemble and circuit level.

Conclusions: Overall these data suggest that 2-AG signaling is critical for optimal cortical representations of threat-predictive environmental cues and 2-AG deficiency may represent a risk factor for the development of PTSD.

Disclosure: Nothing to disclose.

5.4 The Endocannabinoid as a Novel Therapeutic Target in Trauma-Related Disorders: A Randomized, Controlled Clinical Trial and Cross-Section Evidence From Patients With Trauma Histories

Leah Mayo

University of Calgary, Calgary, Canada

Background: While many individuals are exposed to a traumatic event, only a subset develop a post-traumatic stress disorder (PTSD). The endocannabinoid (eCB) system is a neuromodulatory system that regulates fear and stress, and eCB ligands, anandamide (AEA) and 2-arachidonylylglyercol (2-AG), play distinct roles in gating stress reactivity and fear learning. Thus, the eCB system may be a novel treatment opportunity for individuals following trauma exposure. Here, we explored the eCB system as a therapeutic target in 2 complementary ways, assessing clinical, behavioral, and neural outcomes.

Methods: In study 1, patients with PTSD (N = 100; 85 women) were randomized to FAAH inhibitor JNJ-42165279 or placebo for 12 weeks under double-blind conditions. They underwent fMRI and behavioral tests at week 4, then completed an 8-week course of internet delivered prolonged exposure therapy.

In study 2, adult patients with (N = 51) or without (N = 50; controls) prospectively assessed childhood trauma (N = 101 total). Participants completed fear conditioning and fMRI sessions on separate days. Blood samples were collected for eCB analysis.

Results: In study 1, all participants had improved PTSD symptoms (effect of time: −8.83; 95% CI: −11.94, −5.72, p < 0.001), but in contrast to our hypothesis, there was no effect of drug on PTSD symptoms (-1.06; 95% CI: -5.27, 3.15, p = 0.617) or any secondary outcome (depression, anxiety, sleep; all p < 0.05), even though AEA levels were significantly increased (p < 0.001). More severe PTSD symptoms were associated with greater dlPFC activity during emotion regulation (peak β = -0.023; peak Z-statistic = −3.545) and lower dlPFC-vmPFC connectivity at rest (peak β = −0.014; peak t-statistic = −4.972). Ongoing analyses are exploring whether FAAH inhibition impacted fear learning or extinction.

In study 2, patients with higher 2-AG levels showed better fear discrimination (F1,78 = 4.10, p = 0.046). Better fear discrimination was associated with greater amygdala-dlPFC connectivity (F1,87 = 16.3; p < 0.001) while better fear extinction was associated with greater amygdala-ACC connectivity (F1,87 = 26.1; p < 0.001).

Conclusions: Our findings suggest that elevating AEA via FAAH inhibition is not a suitable novel intervention for individuals with trauma histories. However, other eCB targets, such as 2-AG, may be worth exploring further.

Disclosure: Synendos Therapeutics; Consultant (Self)

Panel

6. From Symptoms and Biomarkers to Clinical Trials of Personalized Treatment Selection for Major Depressive Disorder

6.1 Machine Learning Prediction of Individual Patient Outcomes to Psychotherapy Versus Medication for Major Depression

Devon LoParo

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

Background: First-line treatments for major depressive disorder (MDD) include antidepressant medications and evidence-based psychotherapies, which are approximately equally efficacious in the treatment of non-psychotic MDD outpatients. Individual response to these treatments, however, is variable, implying the presence of factors that differ between patients that could be measured to allow for prospective differential prediction of treatment response (i.e., moderation) and, consequently, personalized treatment recommendations.

Methods: We used a machine learning technique, partial least squares regression (PLSR), to differentially predict MDD treatment response (end of treatment HAM-D) from baseline demographic and clinical items in a randomized controlled trial (PReDICT, n = 344) comparing escitalopram, duloxetine, and cognitive = behavioral therapy (CBT). We created a treatment recommendation tool that used these PLSR-derived variables to create a recommended treatment for each individual and used independent t-tests to compare between individuals who did and did not receive recommended treatments. We also externally validated the PLSR prediction and treatment recommendation approach in a second randomized controlled trial (PET-Predictor, n = 64) comparing escitalopram and CBT.

Results: The model predicted a meaningful proportion of variance in end of treatment depression severity for CBT (39.7%), escitalopram (32.1%), and duloxetine (67.7%), leading to a high level of accuracy in predicting remission (71%). By simulating treatment assignment, patients who were randomized to receive their recommended treatment had significantly lower depression severity after treatment (p = 7.3E−8, Cohen’s d = 0.81) and were much more likely to remit (59% vs 33%, p = 0.002). In the external validation sample, the predictor variables explained 9% of CBT and 24% of escitalopram treatment response variance (balanced accuracy = 64%). Those who received recommended treatment had significantly lower depression severity after treatment (p = 8.7E−5, d = 1.04) and higher remission rates (70% vs 31%, p = 0.004), validating the results from the first sample.

Conclusions: These results represent a highly promising, easily implemented, potential advance for personalized medicine in the treatment of MDD.

Disclosure: Nothing to disclose.

6.2 Identifying a Robust Link Between Spontaneous Neural Oscillations and Depression-Related Symptoms

Faranak Farzan

Simon Fraser University, Surrey, Canada

Background: Patients with depression differ with respect to the symptoms they experience and their severity. This heterogeneity may contribute to the inadequate clinical outcomes that commonly follow antidepressant treatments. Investigating the link between neurophysiology and individual depressive symptoms, as opposed to studying depression as a unitary illness, may further our understanding of this illness and its treatment. We investigated how the power of spontaneous neural oscillations, a feature of popular interest in depression, relates to individual depressive symptoms.

Methods: We combined pre-treatment electroencephalography and clinical data (the self-report Quick Inventory of Depressive Symptomology) from two large and independent datasets (CAN-BIND-1 and EMBARC), providing a final sample size of 393 patients with depression. Canonical correlation analysis was used to assess the multivariate association between power in the five canonical frequency bands and depressive symptoms. This was done in both a descriptive and machine learning framework to assess the statistical significance of any associations and how well they generalize to unseen data.

Results: Across the various analyses conducted, a common association was found between the power of slow (i.e., delta) oscillations and depressive symptoms. Canonical loadings suggested features most strongly linked to this association were delta power in posterior regions and a cluster of symptoms which included hypersomnia, suicidal ideation, and psychomotor slowing.

Conclusions: We report a robust multivariate association between spontaneous power in the delta frequency band and depressive symptoms which was statistically significant and generalizable, thereby revealing a potential neurophysiological signature of specific MDD-related symptoms. This work suggests investigating how neurophysiology relates to individual symptoms could provide novel insight into the pathophysiology and treatment of MDD.

Disclosure: Nothing to disclose.

6.3 Personalized Allocation to Antidepressant Monotherapy or Combination With a Partial Dopamine Agonist, Based on Symptoms and Biomarkers: The OPTIMUM-D Trial

Rudolf Uher

Dalhousie University, Halifax, Canada

Background: Symptom dimensions and brain activity patterns are replicable predictors of response to pharmacotherapy in major depressive disorder (MDD), but it is not known if the combination of symptom- and brain-level predictors can indicate the selection of alternative treatments to improve efficacy.

Methods: We tested depressive symptoms loss of interest, reduced activity, and anhedonia in combination with electroencephalography brain activity features as differential predictors of response to a selective serotonin-reuptake inhibitor (escitalopram) and to augmentation with a partial dopamine agonist (aripiprazole) in 113 adults with major depressive disorder. We implemented the prediction in a nested precision double-randomized trial (OPTIMUM-D, clinicaltrials.gov/study/NCT05017311).

Results: Combination of symptom profile and electroencephalography differentially predicted treatment outcomes with serotonin-reuptake inhibitor monotherapy and partial dopamine agonist augmentation with greater accuracy than either predictor alone. Implementation of the algorithm in the first 40 OPTIMUM-D participants resulted in an equal allocation proportion to monotherapy with a serotonin-reuptake inhibitor and combined treatment with a serotonin-reuptake inhibitor and a partial dopamine agonist (brexpiprazole).

Conclusions: A multimodal algorithm derived form independent data adequately predicts outcomes with alternative treatments for MDD and initial experience suggests feasible implementation in precision trial design that independently evaluates the benefits of personalized allocation from the main effect of combined treatment.

Disclosure: Nothing to disclose.

6.4 Large Language Models as Decision Support Tools for Mood Disorder Pharmacotherapy

Roy Perlis

Harvard Medical School, Boston, Massachusetts, United States

Background: Large language models perform well on a range of academic tasks including medical examinations, and more recently have shown promise in psychopharmacologic decision support for bipolar disorder. The performance of this class of transformer models for treatment selection in pharmacotherapy of major depressive disorder has not been investigated.

Methods: The author and colleagues previously created and validated brief vignettes about adults with major depressive disorder illustrating aspects of psychopharmacologic decisionmaking. Experienced clinicians were asked to rank the appropriateness of a list of antidepressants for each vignette. In the present study, a large language model was presented with each of the 10 vignettes in random order, with random variation in age and inclusion of distractor medications, with results regenerated 5 times to evaluate stability of responses. Results were then compared to those of expert clinicians in terms of (a) whether the optimal expert-curated treatment was identified among the best options, and (b) whether a less good or contraindicated treatment was identified among the best or good options.

Results: With the standard prompt, at least one of the optimal medication choices was included among the best choices in 38/50 (76%) (5/5 for 7 vignettes, 3/5 for 1, and 0/5 for 2). At least one of the poor choice or contraindicated medications was included among the best or good choices in 24/50 (48%) (including 2 vignettes with 5/5, 2 with 4/5, one with 3/5, 3 with 1/5, one with 0/5). Model explanations indicated awareness of FDA labeled indications, need to avoid adverse effects that could exacerbate underlying comorbidities, and drug-drug interactions.

Conclusions: A publicly-available large language model provided generally reasonable medication recommendations for pharmacotherapy of major depression after initial antidepressant nonresponse, but also advised less optimal or contraindicated approaches for all but one vignette. The model appeared to identify and apply a number of common heuristics in clinical practice – avoiding medications that had not shown efficacy in the past, favoring better-tolerated medications, avoiding medications with adverse effects that could worsen medical comorbidities, and avoiding drug-drug interactions.

Disclosures: Psy Therapeutics: Advisory Board (Self). Circular Genomics: Advisory Board (Self). Alkermes: Advisory Board (Self). Swan AI Studios: Advisory Board (Self). Mila Health: Advisory Board (Self). Genomind: Advisory Board (Self). Vault Health: Advisory Board (Self)

Panel

7. Computational Analysis of Spoken Language and Face Expression for Diagnosis Specific Phenotyping and Estimates of Working Alliance

7.1 Scalable Digital Innovations in Real-World Settings: Recent Advances in Digital Biomarkers and Large Language Models

Katharina Schultebraucks

NYU Langone Medical Center, New York, New York, United States

Background: By using deep learning to integrate multi-modal information from smartphones and smartwatches, new horizons emerged such as “digital phenotyping” to define transdiagnostic markers to remotely identify and monitor individuals at risk. While previous studies have examined the relevance of psychological stress response in the emergency department (ED) for predicting long-term PTSD risk, prior prognostic research has been based primarily on time-consuming and costly psychometric assessments, which is difficult if not impossible to operationalize in the challenging clinical milieu of the ED. We present on recent advancements in AI and digital tools for contactless assessment of PTSD and resilience in real-world settings, focusing on digital biomarkers and large language models.

Methods: We enrolled 200 trauma survivors who were admitted to the ED within 24 h of a traumatic event and met DSM-5 PTSD criterion A. We enrolled participants across two Level 1 trauma centers and utilized deep learning and large language models to extract digital biomarker from a video and audio recording directly in the ED to develop a prognostic model for PTSD risk.

Results: Video-and audio-based markers were able to accurately diagnose PTSD (AUC = 0.90; weighted average precision = 0.83; f1-score = 0.83) and predict PTSD symptom severity (R2 = 0.60).

Conclusions: Our results show that digital biomarkers identified in direct clinical observation during free speech can be used to diagnose and risk stratify mental disorders in real world settings. Digital biomarkers can improve the scalability and sensitivity of clinical assessments using low burden, passive patient evaluation. Given the advantages of a flexible approach based on free speech rather than a formal clinical assessment, digital biomarkers emerge as a potentially objective, economical, and ecological valid alternative to assess psychiatric risk. Placing reliable automatable tools for lightweight assessment of PTSD risk in the hands of ED clinicians will improve patient care and will identify patients at risk when they are still in the ED and in contact with the healthcare-system. We will show the great promise of digital biomarkers to be scaled-up for recurrent use in routine practice integrated into everyday life and across daily life contexts.

Disclosure: Nothing to disclose.

7.2 Building Better Behavioral Signatures of Psychopathology: Application to PTSD

Roman Kotov

Stony Brook University, Stony Brook, New York, United States

Background: Artificial Intelligence (AI) models can detect signs of psychopathology in person’s language, speech, and facial expressions. This application of AI promises to produce objective, behavioral signatures of mental disorders. Prior AI studies have made substantial strides in development of such signatures for PTSD and related disorders. However, this work was limited by (1) inadequate sample sizes, (2) lack of rigorous replication, (3) reliance on proxy measures of psychopathology, and (4) absence of clinical validation.

Methods: I will illustrate possible solutions to these challenge in a series of studies to develop AI models of PTSD based on language, speech, and facial expressions. These studies were done in a population of responders to World Trade Center disaster. Responders recorded videos during a routine clinical visit, where they responded to various open-ended topics. AI was trained to predict PTSD checklist score. Models were trained in each modality separately and then across the three modalities together. Models were pre-registered after they were trained on 1,506 responders, and performance of pre-registered models was tested in new 500 responders.

Results: Performance of AI reached r = 0.38 for PTSD checklist and AUC = 0.76 for PTSD diagnosis (p < 0.0001). It was strongest for language, second strongest for speech, and weakest for facial expressions. The multimodal approach improved performance over all individual modalities.

Conclusions: Accuracy of these models was successfully replicated in new sample, a more stringent design than existing studies. Model accuracy is sufficient to be useful for research applications, but is not yet strong enough for clinical use. I will describe methodological features that can enable future studies to substantially improve accuracy and utility of behavioral signatures for mental disorder broadly.

Disclosure: Nothing to disclose.

7.3 Emotional Contextualization of Software-Derived Facial Action Units Using Facial Emotion Estimates Improves Discrimination in Neuropsychiatric Disorders

Stephen Heisig

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

Background: Impairments in facial expressivity and non-verbal communication are cardinal features of schizophrenia (SCZ) and treatment resistant depression (TRD). These may result from deficits in social cognition, impaired emotional reactivity, or have a more neurological basis. Currently the diagnosis and rating of these disorders relies more on clinical impression than objective biomarkers. Software quantification of facial action units is confounded in the sense that an action unit may participate in the production of different expressions. In this work we combine the outputs of models that estimate action unit intensity and emotional expression to produce demonstrably better metrics than either method alone.

Methods: In the SCZ study, open-ended narrative research interviews of roughly 30 min were recorded for N = 74 participants. 28 healthy controls (HC) (20 F, 8 M), 27 clinical high-risk (CHR) (13 F, 14 M), and 19 SZ (10 F 9 M) participants. In the TRD study, N = 15 (8 F, 7 M) patients receiving subcallosal cingulate (SCC) Deep Brain Stimulation (DBS) recorded several minute, spontaneous monologues twice daily. There are currently over 8,000 videos in this dataset. All videos were processed using models from HUMEAI to generate facial action unit and emotional expression estimates.

Results: In the SCZ interviews no action unit features were found to discriminate between cohorts while correlation features between AU7 (lid tightener) and estimates of anxiety, joy, tiredness, and distress survived correction (ANOVA p-values 0.007, 0.008, 0.01, 0.02, 0.02). Regressing to CAT-MH depression scores in the TRD study we saw correlation features outperform action unit features. Typical F-test p-values were: Shame/AU6 and Excitement/AU6 (1e-17 and 1e-16) vs. AU6 mean (1e-14).

Conclusions: We found traditional action unit-based features were out-performed in discrimination and regression tasks in two different disease states by correlation features that add emotional context to the action unit intensities. In some cases, pure emotion expression and gesture estimates out-performed correlation features and actions units.

Disclosure: Nothing to disclose.

7.4 Charting Therapeutic Alliances Turn-By-Turn: A Language Modeling Approach

Baihan Lin

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

Background: The therapeutic working alliance is a critical factor in predicting psychotherapy success. Traditional assessment methods rely on questionnaires, which can be time-consuming and lack scalability for analyzing individual dialog turns. We present COMPASS, a novel framework that directly infers the therapeutic working alliance from the natural language used in psychotherapy sessions.

Methods: We analyzed 950 transcribed psychotherapy sessions covering anxiety, depression, schizophrenia, and suicidality. Our approach utilized deep learning methods and large language models to compare session transcripts with the Working Alliance Inventory. We employed neural topic modeling techniques and generative language prompting to analyze topical characteristics of different psychiatric conditions, and incorporated temporal modeling to capture topic evolution at a turn-level resolution.

Results: COMPASS successfully mapped patient-therapist alignment trajectories at a turn-by-turn level. We observed significant differences in working alliance estimates between patients and therapists (p < 0.001). Working alliance scores differed significantly between certain psychiatric conditions in both therapist and patient versions (p < 0.001). All four working alliance scales could significantly detect suicidality (p < 0.001). The combined feature of inferred working alliance scores and document embeddings improved classification accuracy of psychiatric conditions (46% vs. 44.3%, p < 1e−5). Topic modeling revealed distinct themes in patient-therapist dialogs, such as emotional states, personal experiences, and self-reflection.

Conclusions: COMPASS demonstrates the effectiveness of leveraging language models to quantify the nuances of therapeutic interactions. By enabling a more granular understanding of the working alliance, our approach can facilitate timely feedback for therapists and improve psychotherapy effectiveness. The integration of computational linguistics and deep learning techniques opens new avenues for advancing personalized mental healthcare.

Disclosure: Nothing to disclose.

Mini Panel

8. Mechanisms Supporting Sex Differences in Opioid Use and Pain: From Preclinical to Clinical Studies

8.1 Sex-Specific Effects of Inflammatory Pain on Fentanyl Self-Administration and Modulation by Ovarian Hormones

Jose Moron-Concepcion

Washington University, St Louis, Missouri, United States

Background: Most US adults will be negatively impacted by pain conditions during their lifetime and while opioid analgesics effectively reduce pain symptoms, they pose a high risk for abuse. The imminent threat of the opioid crisis is exemplified by the trajectory of opioid overdose deaths over the past 20 years – a trend that has disproportionately impacted men. Accumulating evidence also indicates that women of menopausal age are more vulnerable to opioid overdose, suggesting a role for ovarian hormones. Therefore, elucidating the neuronal mechanisms underlying the sex-specific impact of pain on opioid reward processing is critical to develop individualized treatment strategies for pain patients and to curb problematic opioid use.

Methods: In the current study, we build upon our past research and multiple lines of evidence indicating sex differences DA signaling underlying altered reward processing. We developed a translational model of instrumental fentanyl exposure under pain conditions that recapitulates clinically supported sex differences in opioid misuse liability. In addition, we bypassed the limitations of traditional fiber photometry techniques by integrating fentanyl self-administration with wireless in vivo fiber photometry to permit chronic assessment of sex- and pain-specific neuroadaptations within the VTA-NAc pathway across multiple weeks of fentanyl use.

Results: Our results show that pain increases fentanyl intake during late stages of self-administration selectively in males, recapitulating human indices of higher opioid misuse liability in men.

We also show that pain facilitates fentanyl-evoked VTADA responses in a time-dependent manner that parallels enhanced fentanyl self-administration in males. Gonadectomized females display behavioral and neural responses to fentanyl that resemble males’, indicating that ovarian hormones are required for females to maintain stable levels of fentanyl use and evoked DA neuron activity despite the lack of influence of the estrous cycle. Finally, systemic estradiol administration suppresses behavioral and neural responses to fentanyl in males via estrogen receptor beta in the VTA. These results highlight a potential mechanism by which high estradiol states suppress opioid-evoked DA neuron activity underlying increased fentanyl use.

Conclusions: We report, that pain increases fentanyl consumption over time selectively in male rats. This behavioral effect requires enhanced fentanyl-evoked mesolimbic dopamine activity. The ability of pain to produce these neuroadaptations and drive fentanyl use in males mediated by a lack of ovarian hormones, which serve as a protective mechanism in females. Specifically, estradiol, can suppress the effects of pain in males via estrogen receptors within the VTA. Overall, synergistic interactions between the effects of pain and long-term opioid use may underlie hormone- and sex-specific effects of pain that drive maladaptive patterns of opioid use.

Disclosure: Nothing to disclose.

8.2 Sex Differences in How Stress Impacts Opioid Self-Administration and Subjective Drug Effects in Experimental and Clinical Settings

Marie Eikemo

University of Oslo, Norway, Oslo, Norway

Background: Opioid use disorder is more prevalent among men than women with estimates ranging between 65% and 75%, yet the mechanisms of addiction risk in men and women remain poorly understood. Psychological stress has emerged as a key risk factor for addiction. Stress-induced drug seeking is well-established in the preclinical domain and stress is a primary cause for relapse in addiction. Nevertheless—whether acute stress can increase opioid abuse liability in non-addicted humans remains unknown. Here we examined how stress impacts opioid effects in men and women and assessed whether sex-differences can explain men’s increased risk for opioid misuse.

Methods: We performed (i) an experiment using stress inductions and i.v. oxycodone and (b) an two studies in an ecologically valid stressful clinical setting: surgery. The experiment had a double-blind placebo-controlled randomized design where 63 healthy participants (31 men) completed stress or control tasks before the injection of oxycodone (moderate dose, 3.1mg/70kg) or saline across four different sessions (2x2 repeated-measures). We collected subjective, autonomic and endocrine responses and the primary outcome was amount of additional oxycodone obtained in an effortful self-administration task. In two cohorts of surgery patients, we recorded how good and anxious patients felt before and after opioid injections on the operating table before anesthesia. In the first study (N = 269, 152 women), open-label remifentanil (effect site 5ng/ml) or oxycodone (5 mg) was used. The second was an RCT (EudraCT 2020-004920-40, N = 210) of pre-operative effects of morphine (10mg), oxycodone (5mg) or fentanyl (0.1mcg).

Results: Across the sample, stress increased oxycodone self-administration by 5 percentage points (95% credible interval [CrI 1,10], Posterior probability of the effect (Pr) > 0.99), but with a robust sex difference (16 pp; 95% CrI [7, 24], Pr > 0.99). Stress induction only increased self-administration in men, although women showed higher stress responses (mean difference: subjective 18 points/100 CrI [11, 26]; Pr > .99, heart rate: 5.3 CrI [0.2, 10.4]; Pr = 0.98). We found no evidence that oxycodone relieved stress, improved well-being or that stress-enhanced drug wanting was related to drug liking in men or women. In the clinical cohort we observed higher anxiety as well as higher odds for opioid-related anxiety relief in women (aOR 95%CI 1.7 (1.04−2.7), X2 = 4.5, p = 0.033).

Conclusions: Stress-induced drug seeking is well-established in the preclinical domain and stress is a primary cause for relapse in addiction. Nevertheless−whether acute stress can increase opioid abuse liability in non-addicted humans has not been tested. Notably, opioid drugs are typically administered to individuals under considerable psychological stress. We believe this level of enquiry is essential for understanding the mechanisms of addiction onset in humans. The study unveils a putative mechanism for men’s higher vulnerability to opioid addiction. Pre-operative anxiety may have a greater influence on the subjective experience of opioids in women.

Disclosure: Nothing to disclose.

8.3 Male but Not Female-Based Self-Regulated Analgesia is Mediated by Endogenous Opioids

Fadel Zeidan

UC San Diego, La Jolla, California, United States

Background: Several preclinical and clinical studies have demonstrated sex-based differences in pain sensitivity and analgesia. Females, versus males, exhibit chronic pain more prevalently and are less responsive to and prescribed opioids at higher rates. Prior studies on sex differences in biological pain systems, however, measured exogenously driven analgesia (e.g., morphine) and/or opioid system activation following pain evocation in healthy volunteers. Thus, it is unclear if endogenous opioids are differentially engaged between sexes during the direct attenuation of evoked pain in humans, and critically, whether these effects translate to chronic pain populations characterized by opioid receptor down regulation. Based on these converging lines of evidence, we tested the hypothesis that endogenous opioidergic blockade would reverse meditation-induced analgesia in males but not females in both pain-free and chronic low back pain (cLBP) populations.

Methods: This secondary analysis combined two operationally analogous clinical trials (n = 98; 51 females; 47 males) in pain-free (n = 39) and cLBP (n = 59) populations to examine sex-differences during opioidergic-blockade and self-regulated analgesia. After baseline pain testing, participants were randomized to a four-session mindfulness- or sham-mindfulness-meditation (a placebo-mindfulness technique performed by slow, deep-breathing) intervention. Employing a double-blinded crossover design, participants received noxious heat (ten, 12-second plateaus of 49°C interspersed with 20-seconds of innocuous 35°C stimuli delivered to the calf) at rest, and during intravenous high-dose naloxone (opioid antagonist; 15mg/kg bolus +.1mg/kg/hour) and placebo-saline infusion, respectively, while meditating. Pain intensity ratings were acquired after each heat series via a 15cm visual analog scale (“0.0” = “no pain sensation”; “10.0” = “most intense pain sensation imaginable”).

Results: A Repeated Measures ANOVA detected a sex X session (saline vs. naloxone) X rest vs. meditation interaction, F(1, 91) = 4.14, p = 0.045, ηp2 = 0.04. After both mental training interventions, males (p < 0.001; CI95 = −1.20; −0.37) and females (p = 0.07, CI95 = −0.76; 0.04) reported pain reduction during meditation and saline infusion versus rest. In contrast, females (p = 0.008, CI95 = −1.04; −0.16) but not males (p = 0.23; CI95 = −0.74; 0.18) experienced meditation-induced pain relief during naloxone infusion versus rest. In males, meditation-induced analgesia was significantly greater during saline compared to naloxone (p = 0.01, CI95 = −1.00; −0.13) infusion, while females exhibited no differences in pain-relief between saline and naloxone infusion sessions (p = 0.78, CI95 = −0.36; 0.47).

Conclusions: Our findings reveal sex-specific differences in the engagement of the endogenous opioid system during meditation-induced analgesia, suggesting females may engage unique, non-opioidergic biological systems in response to evoked pain. These findings highlight the critical need for additional clinical studies on sex-differences in pain modulation toward the development of personalized and non-opioidergic treatment options for chronic pain.

Disclosure: Nothing to disclose.

Study Group

9. Addressing the Increase in Completed Suicides in the United States: Strategies to Decrease Suicides in Diverse Demographics - The Identification of Questions and Problems in the Field

William Bunney, Alan Schatzberg, Carlos Zarate, Maria Oquendo, J. John Mann, Jordan Smoller, Lauren Weinstock, Victoria Arango

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

Study Group Summary: Suicide is a global public health problem. The World Health Organization estimates that there are approximately 800,000 suicides each year. Last year, the CDC reported 49,449 suicides in the USA. Suicide is the 4th leading cause of death between the ages of 18 and 44. Topics of discussion will include: How can multiple types of biomarker data be integrated to identify individuals at highest risk for suicide? There are no FDA-approved medications to prevent suicide. Will strategies such as intravenous ketamine and ultra-low dose oral buprenorphine reduce suicidality and deliver sustained relief? Suicide research tends to collapse all potential risk factors, regardless of time frame. What are the key indicators of acute suicide risk which can guide treatment decisions? Because suicidal individuals are often excluded from clinical trials, how can we detect if a treatment has a potential effect on suicidal thoughts? Neurobiological research has implicated insula response to suicide specific stimuli (hopelessness, sleep difficulties). Each of these risk factors have been examined in response to ketamine. One aim is to develop biomarkers of acute suicidal risk and response to treatment which can be used to detect if other treatments have an impact on suicidal thoughts and behaviors. How can we implement novel interventions for individuals at high-risk leveraging real-world large-scale electronic health data records and artificial intelligence to accelerate the development of tools for suicide risk stratification? Stress responsive suicidal ideation (brain inflammation) is a critical factor for suicidal behavior. How can this information be utilized to identify high risk individuals and to develop novel treatments? Emerging evidence from the laboratory and the field suggests that 1 in 5 persons who make a suicide attempt do not have a psychiatric disorder and that people with no psychiatric disorders have suicidal ideation. This is a heretofore unaddressed population. What are the best ways to identify these individuals and what kind of interventions should be used? Most of the recently developed interventions for suicide prevention are psychosocial in nature. These are more challenging to deliver with fidelity than pharmacologic interventions. What strategies can be used to maximize the fidelity with which evidence-based interventions are delivered? Suicide disproportionately affects minority and disenfranchised populations including Blacks, Hispanics, and sexual gender minorities. However, these groups are often not included in suicide prevention research. How can suicide prevention for interventions for underserved groups be effective so that evidence-based practices can be brought to scale efficiently with high fidelity for optimal effectiveness? There is a significant percentage of adults who died by suicide who have had recent contact with police, courts, and jails. How can these individuals which are tracked by big data systems be connected to suicide prevention services when they are at risk?

This Study Group Session has an outstanding group of participants, three of whom head NIH-funded suicide centers, and five have major funded grants for suicide research.

The ACNP is the ideal audience for discussing these issues and providing a setting in which many sides of this topic can be presented.

Disclosure: Nothing to disclose.

Panel

10. Psychedelic-Assisted Psychotherapy: From Preclinical Mechanisms to Clinical Application

10.1 Preclinical Models for Studying Psilocybin’s Actions

Kellie Tamashiro

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

Background: Psychedelic compounds have attracted substantial attention for the treatment of psychiatric disorders. Psilocybin is one such compound that is currently being used in multiple clinical trials for a wide range of neuropsychiatric disorders. The biological mechanisms responsible for the therapeutic actions of psilocybin are not well understood and will rely on use of translational models.

Methods: Here we used 5 groups of adult male C57BL/6 mice: 1 naive control and 4 groups exposed to chronic variable stress (CVS, n = 5/group) for 14 days. Mice were treated with vehicle or psilocybin (1 mg/kg, i.p.) with or without pretreatment with the 5HT2A/2C-R antagonist ketanserin (2 mg/kg, i.p. 30 min before psilocybin). Behavioral tests assessed the head twitch response (HTR), exploratory behavior (elevated plus maze, EPM) and anhedonia (sucrose preference test, SPT) beginning 2 days following drug treatment. To determine the neural substrates invovled in psilocybin’s action, we conducted fMRI scans of naive anesthetized mice (n = 5/group) after acute treatment with vehicle or psilocybin (1 mg/kg, i.p.).

Results: Treatment with psilocybin resulted in more HTRs, which was blocked by pretreatment with ketanserin. CVS reliably reduced open arm entries in the EPM (P < 0.05) and psilocybin treatment resulted in recovery of this behavior to be similar to the non-stressed control group and ketanserin pretreatment had no effect. In the SPT, CVS mice displayed reduced preference for sucrose (P < 0.05) and psilocybin treatment restored preference to control levels. Quantification of cerebrovascular blood volume showed a reduction in the amygdala and non-significant increase in the mPFC and thalamic brain regions. Psilocybin-treated mice showed increased open arm entries (P < 0.05) in the EPM compared to vehicle-treated mice when tested 3 days following fMRI scan.

Conclusions: Psilocybin treatment following CVS led to increased exploratory behavior and increased hedonic responding. Psilocybin-associated behavioral changes were not blocked by pre-treatment with ketanserin, although ketanserin blocked the HTR. The data suggest that psilocybin’s mechanism of action in ameliorating behavioral consequences of CVS are not completely reliant on agonism of 5HT2A/2C-R. We also identify candidate brain regions that may be involved in acute actions of psilocybin.

Disclosure: Nothing to disclose.

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

Lea Julia Mertens

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

Background: Psilocybin is intensively researched in the treatment of depressive disorders with promising results.

Methods: This double-blind, phase 2b, active placebo-controlled trial (NCT04670081) was conducted from July 2021 until February 2024 at 2 German study sites. Patients were randomly assigned to one of four treatment arms, receiving (1) placebo (100 mg nicotinamide) first and a high-dose psilocybin (25 mg) second; (2) a low-dose psilocybin (5 mg) first and a high-dose second; (3a) a high-dose psilocybin first and a low-dose second; (3b) a high-dose psilocybin at both sessions. The two dosing sessions were separated by 6 weeks and embedded in 7 psychotherapeutic preparatory and integration sessions over a total period of 13 weeks.

The primary endpoint was treatment response, defined as a minimum of 50% reduction in symptoms, measured with the Hamilton Rating Scale for Depression (HAMD-17; range: 0-52), six weeks after the first dose.

Results: 144 patients (59% male; mean age: 42.6 years, SD = 10.83; 98% white) were randomized into the trial, of which 143 received the first dose and 137 received both dosing sessions. The mean HAMD-17 scores ranged from 21.46 to 22.40 in all treatment groups at baseline, indicating a similar and moderate depression level. 140 patients completed the primary endpoint.

Of those patients who received 25 mg psilocybin first 8 patients (16.7 %) classified as responders at six weeks; in the 5 mg psilocybin arm 4 patients (8.33%) and in the placebo arm 5 patients (10.42%) met the response criterion. The mean change from baseline was −1.63 in the placebo, −3.74 in the 5 mg and −5.97 in the 25 mg group after one dose. The inferential statistical analyses are currently ongoing.

Suicidal ideations occurred in all groups, no suicidal behavior occurred until the primary endpoint (further analyses ongoing). Serious Adverse Events occurred in 4 patients in form of colon cancer, worsening of psychopathology and a hypertensive urgency during dosing.

Conclusions: The study treatment of one or two doses of psilocybin in a psychotherapeutic context was well-tolerated by a large patient group of TRD, with a subgroup of patients benefiting a lot. These first results also show a clear placebo response and let questions arise about the treatment mechanisms at hand in this therapy (pharmacological vs. extra-pharmacological).

Disclosure: Nothing to disclose.

10.3 Preliminary Effects of Psilocybin in Patients With Major Depressive Disorder and Co-Occurring Alcohol Use Disorder

Frederick Barrett

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

Background: Major depressive disorder (MDD) and alcohol use disorder (AUD) are two of the most common co-occurring disorders, with a high prevalence of AUD in patients with MDD (> 23%). We and others have demonstrated potential efficacy of psilocybin for the treatment of mood or substance use disorders and specifically alcohol use disorder. Deficits in cognitive control accompany both disorders. The claustrum may support instantiation of cortical brain networks to subserve cognitive control. Psilocybin reduces claustrum activity and connectivity to default mode (DMN) and frontal parietal control networks (FPN) in healthy participants, but claustrum involvement in MDD and AUD is not yet known.

Methods: 25 male and 25 female adults with MDD and co-occurring AUD received placebo or 25 mg psilocybin during experimental drug administration. 3 months later, all were administered open-label 25 mg psilocybin. Blinded clinician ratings of depression severity were collected at screening and 1 month after each drug administration. The Time-Line Follow Back procedure and liver function tests from peripheral blood samples at screening and 3 months after each session were used to assess self-report and biological measures of drinking. A subset of patients completed a task-switching paradigm during fMRI measurement before and 1 week after drug administration.

Results: Depression (p < 1.05e-09, d = 3.9), heavy drinking (p < 0.0005, d = 1.62), days abstinent (p < 0.00005, d = 1.61), and GGT levels (p < 0.05, d = 0.51) were reduced from baseline following the open-label session. Subjective effects during experimental sessions were negatively correlated with response time (r = −0.34) and accuracy (r = −0.51) during task switching. Stronger subjective effects predicted a post-session shift towards greater effort selection (e.g. choosing higher-effort blocks over lower-effort blocks). We will test the hypothesis that greater effort selection after psilocybin is associated with DMN disengagement and FPN engagement during high-effort trials, and that this is mediated by greater claustrum connectivity to DMN and FPN.

Conclusions: Psilocybin reduces depression severity and both self-report and biological measures of drinking in patients with MDD and co-occurring AUD. This response is accompanied by greater effort selection, possibly reflecting increased meta-cognition.

Disclosure: Nothing to disclose.

10.4 The Role of Psychotherapy in MDMA-Assisted Therapy (MDMA-AT) for PTSD

Jennifer Mitchell

UCSF, San Francisco, California, United States

Background: Previous data indicate that MDMA-assisted therapy (MDMA-AT) can significantly attenuate symptoms of PTSD and depression. While it is often accepted that MDMA drug efficacy is bolstered by the addition of psychotherapy, it has not yet be demonstrated that psychotherapy potentiates the effectiveness of MDMA for PTSD. Qualitative reports on self-experience can offer preliminary data to evaluate the role of psychotherapy in the effectiveness of MDMA-AT.

Methods: Qualitative reports of therapeutic experiences collected during phase 3 will be aggregated and shared during this presentation. The data are currently embargoed (the FDA is considering them as part of a submitted NDA.) but should be available in time for this presentation.

Results: The data show that psychotherapy can enable study participants to express more self-compassion and less self-criticism, suggesting that psychotherapy is a critical component in the effectiveness of MDMA for PTSD.

Conclusions: MDMA should be administered in conjunction with psychotherapy to maximize therapeutic benefit.

Disclosure: Nothing to disclose.

Panel

11. Psychotic Symptom Emergence as a Window Into Pathophysiology

11.1 Elucidating Early Risk Markers of Clinically Relevant Psychotic-Like Experiences

Kenneth Osborne

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

Background: Transient, subclinical psychotic-like experiences (PLEs) typically emerge in late childhood and early adolescence and then decrease across adolescent development. However, for a subset of individuals, emerging PLEs become clinically relevant (i.e., persistent/distressing), and sometimes eventuate into florid psychosis in early adulthood. Identifying risk factors that distinguish clinically relevant PLEs from normative experiences across early development is critical for improving our understanding of the etiopathogenesis of psychosis risk.

Methods: Using the Adolescent Brain Cognitive Development Study data (N = 11,868; Age Range: 8-16; 48% female; 52% White), we used machine learning and longitudinal modeling to test associations between theory-driven risk factors and clinically relevant PLE trajectories in early development including cognition, neural measures (e.g., cortical thickness, resting state functional connectivity), internalizing symptoms, developmental delays, adverse childhood events, and substance initiation. Machine learning analyses employed random forest classification models.

Results: Machine learning analyses revealed that a model distinguishing persistent distressing from low-level PLEs showed the highest accuracy (test sample accuracy = 69.33%; 95% CI:61.29−76.59%). Models generally indicated the importance of baseline internalizing symptoms, adverse childhood events, and cognitive metrics in classifying worsening PLEs. Further, longitudinal modeling of substance initiation demonstrated that adolescents that used substances experienced greater distress from PLEs relative to youth that never used substances (e.g., cannabis: B = 1.17; 95% CI = 0.96,1.39), and that distress from PLEs increased leading up to substance initiation (e.g., cannabis: B  =  0.23; 95% CI = 0.21,0.26).

Conclusions: These findings elucidate important environmental and mental health risk factors that distinguish clinically relevant PLEs from normative developmental experiences and provide key insights into the relationship between adolescent PLEs and the psychosis symptom spectrum.

Disclosure: Nothing to disclose.

11.2 Advanced Computational Modeling of EEG Spectral Features and Mismatch Negativity for Identifying Psychosis Risk in Help-Seeking Youth

Andreea Diaconescu

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

Background: Psychosis spectrum symptoms (PSS) may indicate early signs of schizophrenia spectrum disorders. Similarly, reductions in auditory mismatch negativity (aMMN) observed through electroencephalography (EEG) are a robust early warning sign of transition to psychosis, yet prior studies have focused more on young adults at clinical high risk rather than those with broader PSS.

Methods: We analyzed aMMN data from 103 participants in the Toronto Adolescent and Youth Center for Addiction and Mental Health (TAY-CAMH) Cohort study, a longitudinal study of treatment-seeking youth (mean age = 18.4, 63% female). We aimed to identify differences in aMMN and EEG spectral features between those with and without PSS (PSS+ versus PSS-) using the 12-item PRIME Screen self-report.

First, we modeled single-trial EEG data with the Hierarchical Gaussian Filter model, extracting belief trajectories of low-level sensory prediction errors (PEs) and high-level volatility PEs. Second, we examined EEG signals in the frequency domain by extracting parameters with the Fitting Oscillations and One Over F (FOOOF) model, fitting characteristic 1/f activity, and extracting aperiodic and band-specific periodic components of the power spectral density (PSD).

Results: Using partial least squares (PLS) with bootstrapping and permutation testing, we found a significant difference in PSD between PSS+ and PSS- participants (p = 0.045). This difference was primarily in PSD slope (exponent) across multiple channels, with some channels showing differences in periodic components, including theta (4−8 Hz) and alpha (8−13 Hz). Parametric analyses across all participants also indicated that a lower PSD exponent predicted higher PSS scores. Additionally, an increase in both low- and high-level PEs between 129 and 134ms and 137 and 145ms, respectively ppFWE < 0.05 associated with the aMMN response was observed in PSS+ compared to PSS- individuals.

Conclusions: Our results suggest that both low- and high-level PE expression may be enhanced in help-seeking youth with PSS. Changes in aperiodic PSD components, possibly due to reduced synaptic gain and excitatory-inhibitory imbalance, may precede band-specific differences seen in high-risk and schizophrenia patients. Auditory MMN and EEG-PSD features are highlighted as potential biomarkers for psychosis risk in help-seeking youth.

Disclosure: Nothing to disclose.

11.3 Dynamic Patterns of Psychotic Symptom Emergence Support a Secondary Account of Hallucinogenesis

Catalina Mourgues

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

Background: Hallucinations and delusions are often studied in fully-formed illnesses, but understanding the early stages of symptom emergence could help pinpoint the pathophysiological process leading to psychosis. This is particularly important because these symptoms may arise from different abnormalities in learning and inference. We examined the onset order of CHR-level delusions (SOPS P1 and P2 > = 3) and hallucinations (P4 > = 3) in the NAPLS 2 and 3 cohorts, along with their neural correlates, to better understand mechanisms giving rise to symptoms in CHR.

Methods: Participants from NAPLS 2 and NAPLS 3 were grouped based on their first symptom (D = only delusions, H = only hallucinations, and D and H = delusions and hallucinations) and their symptom evolution to explore the following: (a) patterns of emergence of delusions and hallucinations (SOPS scores 3−5, n = 719, n = 699), (b) circuit dynamics using Dynamic Causal Modeling (DMC) on auditory oddball P300 EEG task using (n = 571, NAPLS 3) and (c) connectivity between 15 psychosis-related ROIs using DCM on resting-state fMRI (n = 95, NAPLS 3)

Results: D, as first symptom, were more frequent than H, both retrospectively (ORs: NAPLS 2 = 4.09, NAPLS 3 = 4.14) and at baseline (ORs: NAPLS 2 = 5.6, NAPLS 3 = 11.11), with stable ratios over time. H exhibited greater volatility than Del (Cohen’s d: NAPLS 2 = 0.105, NAPLS 3 = 0.139). Reduced P300 amplitude in CHR-P was explained by increased superficial pyramidal cell self-inhibition (decreased synaptic gain) in the STG (posterior probability P > 0.95). P300 amplitudes were similar between controls and individuals with D, while lower P300 amplitudes were found in those with D and (d = 0.24) and those with D who developed H (d = 0.41). Follow-ups showed a reduction in P300 was observed after H onset (p = 0.02). fMRI showed D was linked to greater bottom-up connectivity, and a switch from bottom-up to top-down connectivity was observed in individuals with D, who later developed H.

Conclusions: Consistent with emerging computational models of psychotic symptom development, these data support the possibility that hallucinations arise after--and perhaps in response to--the factors driving delusion formation.

Disclosure: Nothing to disclose.

11.4 Abstract Not Included

Panel

12. Brain Molecular and Cellular Dissection and Modeling of Stress Related Pathology

12.1 Pseudo-Longitudinal Gene Profiling Identifies Differential Gene Regulatory Programs Between Sexes During Puberty: A Potential Framework for Sex Differences in MDD

Corina Nagy

Douglas Hospital Research Center, Montreal, Canada

Background: Puberty significantly alters brain dynamics, with both shared and distinct effects on males and females. This period involves intense molecular, physiological, and anatomical reorganization, leading to lasting biological differences and potentially contributing to the sex disparity in major depression. Tracking molecular brain changes during puberty provides insight into brain reorganization and underlying sex differences in behavior. This study identifies key time points and transcriptional changes influenced by sex-specific gene regulation, potentially shaping sex-specific behaviors.

Methods: Male (n = 40) and female (n = 40) mice were weaned at PD21 and sacrificed at 20 timepoints throughout puberty, including pre-vaginal opening and pre-preputial separation. Brains were harvested, and the prelimbic cortex was dissected for bulk RNA-seq. DESeq2 analyzed differential expression across time points using raw count data. Timepoint selection (TPS) was performed using the error between the spline estimation and the true data to determine the final average error iDREM plotted gene trajectories and regulatory events for each sex. Significant time points were further analyzed by snRNA-Seq in FCG mice (n = 120) and humans.

Results: Data-driven gene expression trajectories differ between males and females during puberty, with opposing effects driven by key regulators like the estrogen receptor Esr2, which was found to regulate the same genes in opposite directions in males and females. Genes such as Il-2, Drd2, and Pomc increase expression only in males, suggesting sex-specific regulatory effects. This may underlie sex-specific mood-related behaviors emerging post-puberty. TPS analysis identified PD28, PD38, PD42, PD48, and PD72 as key time points. Low-input single-cell RNA-seq in FCG mice highlights hormone versus chromosome effects on cell-specific gene programs, providing greater resolution than bulk analysis.

Conclusions: Our study dynamically explores how genetics and hormones influence gene transcription. We identified specific time points where gene regulation differs by sex and found that transcription factors can affect gene expression patterns oppositely between sexes. These data suggest an important role of upstream regulation from the sex-specific hormonal milieu during puberty.

Disclosure: Nothing to disclose.

12.2 Systems Biology of PTSD Reveals Mechanisms of Risk and Disease Processes at Brain Multi-Omic, Brain Cell Type, and Blood Levels

Nikolaos Daskalakis

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

Background: Studies on postmortem brains of PTSD and MDD patients, compared to neurotypical controls, reveal genetic overlaps, sex disparities, and immune and interneuron signaling involvement, yet lack integrative analyses.

Methods: To address this gap, we established a brain multi-omic, multi-region database comprising individuals with PTSD, MDD, and NCs (77/group, n = 345). FWe analyzed molecular changes across the central nucleus of the amygdala (CeA), medial prefrontal cortex (mPFC), and hippocampal dentate gyrus (DG) at transcriptomic, methylomic, and proteomic levels. Our approach is supplemented by single-nucleus RNA sequencing (snRNA-seq), genetics, and blood proteomics, aiming for a comprehensive systems perspective.

Results: Our findings highlight predominant molecular changes in the mPFC, with differentially expressed genes and exons carrying disease signals. Notably, methylation alterations were concentrated in the DG for PTSD and CeA for MDD. We observed a moderate overlap between disorders, with childhood trauma and suicide driving molecular variations, and sex-specificity was more notable in MDD. Pathway analyses link disease-associated molecular signatures to immune mechanisms, metabolism, mitochondria function, and stress hormone signaling, with low concordance across omics. Top upstream regulators include IL1B, GR, STAT3, and TNF. Multi-omic factor and gene network analyses suggest latent factors and modules related to aging, inflammation, vascular processes, and stress.

Complementing multi-omics, our snRNA-seq in the dorsolateral PFC and DG reveals dysregulated pathways and upstream regulators in neuronal and non-neuronal cell types, including stress-related genes and emphasizing childhood trauma effects in females. Examining brain multi-omics with blood proteins in the large UK Biobank cohort shows significant correlation and overlap, implying potential blood-based biomarkers. Fine-mapping of PTSD and MDD GWAS reveals limited overlap between risk and disease processes at the gene and pathway level.

Conclusions: Our study unveils shared and unique brain multi-omic molecular dysregulations in PTSD and MDD, elucidating distinct cell-type involvement and paving the way for blood-based biomarker development. As these studies expand sex differences and effects of childhood trauma are elucidated.

Disclosure:

Biogen: Stock / Equity - Publicly Traded Company; (Spouse/Partner). Circular Genomics: Advisory Board (Self). Feel Therapeutics: Advisory Board (Self)., BioVie: Advisory Board; (Self)

12.3 Converging Human-Specific Evidence Implicates Astrocytic Glutamate Regulation in Stress-Related Psychiatric Disorders

Natalie Matosin

The University of Sydney, Camperdown, Australia

Background: Severe stress is the leading environmental risk factor for psychiatric disorders. Yet, we have a limited understanding of what underlies this association in humans. Astrocytes, an abundant non-neuronal cell type, are key regulators of synaptic transmission. Rodent studies indicate that stress impairs the capacity to regulate neurotransmission, which can drive the development of psychiatric-like symptoms. However, psychopathology in humans is complex, and these cells are poorly evolutionarily conserved, necessitating human-relevant evidence to translate these findings.

Methods: We used single nucleus RNA-sequencing (n = 87) and spatial transcriptomics (n = 13) from the orbitofrontal cortex (BA11) to examine human astrocyte diversity. We used this basis of diversity to examine how individuals diagnosed with a psychiatric disorder stratified by profound stress exposure prior to diagnosis differ from those without. Immunohistochemistry (IHC) of was used to further dissect the impact of stress on the morphology of >20,000 astrocytes. The functional impacts of stress were examined using astrocytes (iAsts) differentiated from human pluripotent stem cells. iAsts were treated with 100nM dexamethasone (DEX) for 1−7 days to activate the cortisol response. Neurotransmitter clearance and neurotransmitter Ca2+ response were quantified.

Results: We identified that gray matter astrocytes involved in synapse homeostasis were delineable between stress-stratified cases. Exploratory analysis identified that effects were associated with sex (greater in females) and timing of stress (greater in early-life exposure). EAAT2-expressing astrocytes were denser across the layers of the cortex (P = 0.002-0.016) and covered a greater domain, particularly in layers V-VI (Padj = 2.43x10¬-4). iAsts treated with DEX had reduced glutamate clearance after 3 days (P = 3.5x10-6) and increased Ca2+ response to both glutamate (P = 0.004) and NMDA (P = 0.0047) after 7 days.

Conclusions: Our findings indicate that astrocytes are associated with lasting changes in individuals with psychiatric disorders exposed to profound stress, especially in gray matter astrocytes involved in glutamate regulation. These cells may be key responders to stress and, thus, an important focus for intervention or prevention strategies.

Disclosure: Nothing to disclose.

12.4 Stress Hormone Dysregulation During Prenatal Neurogenesis Shapes Tissue Formation and Cell Regulatory Networks With Implications for Neurodevelopmental Disorders

Cristiana Cruceanu

Karolinska Institute, Solna, Sweden

Background: The brain undergoes important growth and plasticity during perinatal development. Altered activation of the glucocorticoid receptor (GR) system is one factor mediating stress effects, likely through transcriptional and epigenetic dysregulation. To investigate these processes from a human-specific perspective we used state-of-the art in vitro induced pluripotent stem cell-derived 3-dimensional brain organoids to model critical periods of cortical neurogenesis.

Methods: To determine cell-type specific GR activation response, we profiled the transcriptomes of thousands of cells using single-cell transcriptomic analyses following GR activation using three paradigms: acutely, chronically, and a two-hit combinatorial model. We used immunofluorescence high-content confocal microscopy to better understand protein, and cell morphology-level effects. We used and single-cell transcriptomic (scRNA-seq) and chromatin accessibility (scATAC-seq) analyses to understand cell-type-specific gene-level responses and to build gene-regulatory networks of lasting effects on specific cell populations.

Results: Chronic GR activation in cerebral organoids induced a robust cell-type-specific differential response from transcription factors involved in neuronal cell fate regulation: SOX2, PAX6, and TBR1. Lineage analyses identified an over-commitment toward inhibitory neurons, through modulation of lineage driver genes that directed the likelihood of individual cells’ commitment. In vitro findings were consistent across diverse genetic backgrounds and supported by ex vivo human fetal brain data. Our investigations across multiple cell-type-specific data modalities identified activation of gene regulatory networks centered around key developmental transcription factors like PBX3.

Conclusions: Cerebral organoids show responsiveness to GR activation, including a cell-type specific transcriptional regulatory response through key driver genes capable of shifting lineage commitment. The likely outcome of aberrant overexposure is a lasting shift in neuron type distribution and consequently brain architecture. This work sheds light on the mechanisms by which environmental stimuli like maternal stress or pharmacology could lead to lasting changes in brain development and ultimately mental illness vulnerability.

Disclosure: Nothing to disclose.

Panel

13. Novel Non-Abstinence Endpoints/Measures for Clinical Trials in Substance Use Disorders

13.1 Reduction in World Health Organization Risk Drinking Levels as Endpoints for Alcohol Clinical Trials

Katie Witkiewitz

University of New Mexico, Albuquerque, New Mexico, United States

Background: The European Medications Agency currently accepts a change in drinking based on reductions in World Health Organization (WHO) risk drinking levels as a valid endpoint for alcohol clinical trials. However, there is considerable uncertainty how WHO risk levels relate to experiences of stress, sleep, craving, alcohol use disorder remission, and functional outcomes, defined by psychosocial functioning, quality of life, mental health, and drinking consequences, among individuals enrolled in alcohol clinical trials.

Methods: Secondary data analysis of self-report alcohol consumption and functional outcome data in four separate randomized clinical trials for alcohol use disorder will be presented together, with data from the COMBINE study (n = 1383), Project MATCH (n = 1726), the Horizant trial (N = 394), and the Varenicline trial (N = 200).

Results: Results indicated reductions in WHO risk levels were associated with significantly greater psychosocial functioning, mental health, sleep quality, and quality of life and significantly fewer drinking consequences, lower stress, and higher likelihood of alcohol use disorder remission, as compared to those with no change or an increase in WHO risk levels during treatment. Importantly, even a one level decrease in WHO risk levels predicted statistically and clinically significant improvements in functioning. Improvements in functioning were similar to those achieved by abstainers. Results further indicated that reductions in WHO risk levels were stable over time, and the associations between reduced drinking and functional improvements did not vary across levels of alcohol use disorder severity.

Conclusions: The results from the current study provide evidence of reductions in WHO risk levels as significantly associated with functional improvements among patients with alcohol use disorder. This study adds to the growing body of literature demonstrating WHO risk level reduction as a viable alternative to abstinence as an endpoint for alcohol clinical trials. The paper will also discuss the application of WHO risk levels in clinical practice and provide clear guidance for clinicians on the targets for alcohol risk reduction that are most likely to be associated with meaningful improvements in functioning among patients with alcohol dependence.

Disclosure: Alcohol Clinical Trials Initiative: Honoraria (Self)

13.2 Impulsivity in Substance Use Disorders—a Potential Mechanism for Precision Treatment in Addiction?

Katrin Preller

Boehringer Ingelheim, Basel, Switzerland

Background: Patients living with substance use disorders (SUD) often suffer from impulsivity and impulsive relapse. Impulsivity in SUD is associated with suicidal ideation, aggression, family and occupational problems, and impaired quality of life. Furthermore, impulsivity is associated with the development as well as maintenance of SUD. It is therefore conceivable that patients living with SUD showing high levels of impulsivity may benefit from interventions that reduce impulsivity. However, a detailed characterization of different facets of impulsivity in SUD compared to other patient groups and healthy controls would help to inform useful clinical avenues to intervene.

Methods: Participants completed three study visits: a screening visit, a baseline visit, and a follow-up visit four weeks after the baseline visit. A total of 133 participants completed the study (N = 20 cocaine use disorder, N = 35 ADHD, N = 10 Borderline Personality Disorder, N = 50 healthy controls). The participants completed a battery of questionnaires and behavioral measures of impulsivity at the baseline and follow-up visits consisting of: the Barrett Impulsiveness Questionnaire 11 (BIS-11), the UPPS-P Impulsive Behavior Scale, the Immediate Memory Task (IMT), a computerized Delay Discounting Task (DDT), the Stop-Signal Task (SST), and the Go/NoGo Task (GNG).

Results: The BIS, SST, GNG, IMG, and DDT produced stable results over the 4-week interval in cocaine users (difference between timepoints: all p > 0.05). Cocaine users reported higher impulsivity scores than healthy controls on the BIS-11. At both timepoints, cocaine users showed higher delay discounting than all other groups (all p < 0.05), while no significant differences were detected between the other groups. Further differences between cocaine users and other groups were either not consistent across timepoints or not significant.

Conclusions: Patients living with cocaine use disorder consistently show high self-reported levels of impulsivity as well as significantly higher delay discounting than any other patient group investigated. Specifically, delay discounting, i.e. the depreciation of the value of a delayed reward, is particularly pronounced in this patient group and reducing impulsivity in response to immediate rewards may contribute to the improvement of clinical and functional problems in SUDs.

Disclosure: Boehringer Ingelheim: Employee (Self). Vifor Pharmaceuticals, Employee; (Spouse/Partner)

13.3 Dissociating Cognitive Phenotypes for Habit and Waiting Impulsivity in Compulsive Behaviors With Novel Online Tasks

Valerie Voon

University of Cambridge, Cambridge, United Kingdom

Background: Impulsivity and sensation seeking are risk factors for the development of drug and alcohol misuse. In rodent studies, premorbid waiting impulsivity is a strong predictor for habitual cocaine seeking behavior but the link has not yet been shown in human studies. Here we developed novel online waiting impulsivity and reward and aversive habit tasks assessing the relationship between these constructs and alcohol misuse in a normative population.

Methods: Novel online tasks were validated and tested in 600 individuals stratified by number of alcohol units consumed per week. The tasks included a 2-choice serial reaction time task primed by alcohol and neutral cues and an overtraining and devaluation habit task with rewarding and aversive sounds. Subjects completed the Alcohol Use Disorders Inventory Test, Obsessive Compulsive Inventory (OCI), Beck Depression Inventory (BDI) and the UPPS-P including a sensation seeking subscale.

Results: High sensation seeking was correlated with impaired learning from aversive (ρ = −0.22, p < 0.001, CIs: −0.34 to −0.11), but not reward outcomes in the training phase, particularly in those with high alcohol misuse (Z = 2.1, p = 0.03). Sensation seeking was also associated with avoidance habit responses (ρ = -0.18, p = 0.01, CIs: −0.31 to −0.05). Alcohol misuse was correlated with greater waiting impulsivity to alcohol cue contexts. Waiting impulsivity correlated with habitual behavioursl to reward (ρ = -0.38, p < 0.001, CIs:−0.4 to −0.07) but not aversive outcomes, most prominently in those with high alcohol misuse (Z = 2.1, p = 0.02). High OCI was correlated with both greater habitual behaviors, both to aversive (ρ = −0.18, p = 0.009, CIs:-0.33--0.05) and reward (ρ = -0.17, p = 0.04, CIs:−0.37 to −0.02) outcomes. Using exploratory mediation, the relationship between alcohol misuse and OCD with aversive habit were mediated by depressive symptoms although aversive habit and depression were not directly associated.

Conclusions: Sensation seeking was associated with impaired aversive goal-directed learning and habit whereas waiting impulsivity was linked to reward habits, both particularly in high alcohol misuse. These findings dissociate from obsessive compulsive symptoms. These findings provide insight into risk factors and cognitive dimensions for compulsive behaviors.

Disclosure: Nothing to disclose.

13.4 Creating an FDA-Qualified Drug Development Tool (DDT) for Assessing Opioid Craving

Cecilia Bergeria

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

Background: Craving is a fundamental symptom of opioid use disorder (OUD) and is included in the Diagnostic and Statistical Manual. Despite this, there is no agreed upon way to assess opioid craving, relegating it to a neglected treatment end point. We have conducted a series of studies to begin developing a validated Clinical Outcome Assessment (COA) for opioid craving, following the qualification framework outlined by the U.S. Food and Drug Administration (FDA).

Methods: First, we conducted focus groups (n = 75) among individuals in treatment and seeking treatment for OUD to gather concept elicitation on opioid craving. Next, we measured opioid cue-reactivity (n = 48) to examine unique physiological (e.g., heart rate) and psychosocial symptoms which co-occur with craving to establish further construct validity. Integrating data from both studies, we created and administered a preliminary questionnaire to 90 individuals in treatment or seeking treatment for OUD to examine the associations between opioid craving assessment items and demographic and clinical features of OUD.

Results: In focus groups, craving was described as “Preoccupation/Obsessive Thoughts”, "Anticipation of Negative Reinforcement”, “Motivation or Desire to Use”, “Anticipation of Positive Reinforcement”, and “Feeling a Lack of Control”, by 56%, 56%, 41%, 29%, and 19% of participants, respectively. In the cue-reactivity study, individuals who reported increases in craving showed significant increases in self-report and physiological indicators of stress (e.g., Symptom Checklist, heart rate) following opioid cue conditions. Integrating data from focus group and laboratory studies, the preliminary COA consists of 6 dimensions: (1) Preoccupation with opioids, (2) Anticipation of negative reinforcement, (3) Anticipation of positive reinforcement, (4) Drive, motivation to use opioids, (5) Loss of control over opioid use, (6) Tension, anxiety related to opioids. In the final analysis, scores did not differ as a function of gender, sex, nor age. Dimensions 3 and 4 were significantly greater among individuals in treatment relative to those seeking treatment.

Conclusions: We have developed a content valid craving COA which is proceeding to additional validity testing including testing the factor structure, measurement invariance, and confirmatory factor analysis.

Disclosure: Canopy Growth Corporation: Contracted Research (Self). Pear Therapeutics Inc.: Contracted Research (Self). Mind Med Inc.: Consultant (Self)

Panel

14. Neural Mechanisms of Psychological Treatment in Pediatric Anxiety Disorders

14.1 Changes in Tripartite Brain Network Connectivity with Cognitive Behavioral Therapy in Pediatric Anxiety Disorders: A Randomized fMRI Clinical Trial

Kate Fitzgerald

Columbia University, New York, New York, United States

Background: Cognitive behavioral therapy (CBT) is an effective intervention for anxiety disorders in youth, but does not work for everyone. CBT requires youth to engage with feared stimuli (exposure)—a challenge that requires capacity for cognitive control. Cognitive control function is served by tripartite brain networks, including the central executive (CEN), salience (SAL), and default mode (DMN) networks. More positive CEN—SAL connectivity and reciprocal interactions of these networks with the DMN support cognitive control. We posited that increased CEN—SAL connectivity, and decreased connectivity of CEN and SAL with DMN may underlie mechanism of CBT effect in anxious youth.

Methods: Anxiety severity (Pediatric Anxiety Rating Scale, PARS) and resting state functional connectivity data (rsfcMRI) were collected from 69 anxious youth (ages 7−17, 58 F,) before and after randomization to CBT (CBT; n = 48) or a relaxation-based control therapy (RT, n = 21). Group Iterative Multiple Model Estimation (GIMME), a method that retains individual heterogeneity in connectivity, was used to assess tripartite network connectivity at individual- and group-levels. GIMME was first used to derive connections between atlas-defined nodes of CEN, SAL, and DMN for each participant, then linear regressions were conducted to test for interactions of time (pre-, post-treatment) x treatment condition (CBT, RMT) on connectivity patterns.

Results: A significant interaction of time x treatment condition on CEN-SAL connectivity was observed (b = 0.078), false discovery rate [FDR]-corrected p < 0.001, 95% CI [0.055, 0.10], driven by pre- to post- CBT increases of CEN-SAL connectivity (t(47) = −4.88, p < 0.001) relative to RT-related decreases (t(20) = 5.04, p < 0.001). Post hoc changes in CEN-SAL connectivity were not associated with symptom reduction across patients (p = 0.81) or by treatment group (p = 0.50).

Conclusions: Increased CEN-SAL connectivity occurs after CBT relative to an active, relaxation-base comparator. These findings suggest increasing CEN-SAL connectivity as a mechanism of CBT effect that could be targeted by novel strategies (e.g. neuromodulation) to augment CBT response in pediatric patients with anxiety disorders.

Disclosure: Nothing to disclose.

14.2 Change in Brain Activity to Reward is Associated With Treatment Response Among Youth With Anxiety Disorders

Cecile Ladouceur

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Alterations in reward-related brain activity have been found to predict response to psychological treatment in adolescents with anxiety disorders. It remains unclear, however, to what extent effects are driven by reward anticipation or feedback and if brain activity would change as a function of treatment response. We addressed these questions by testing whether pretreatment brain activity during reward anticipation and feedback predicted treatment response and changed from pre-to-post treatment with treatment response in the context of a randomized clinical trial of cognitive-behavioral therapy (CBT) for anxiety disorders in adolescents.

Methods: This study used an Approach-Avoidance fMRI task to examine reward-related brain activity in youth aged 9-14 with anxiety disorders (ANX; N = 133; 57 female youth; 90% White) before and after 16 wks of CBT or child-centered therapy (active control). Age- and sex- matched healthy comparison (HC) youth (N = 38) also completed scans on a similar timeline. A subset of ANX youth completed a 2-yr follow-up symptom assessment. Region-of-interest (amygdala, insula, dorsal ACC, striatum, medial prefrontal cortex) and wholebrain analyses were conducted using generalized linear model with age and sex as covariates. Multiple comparisons were corrected using cluster threshold (z = 3.1 and corrected p = 0.05) and the Holm method. Trial registration: ClinicalTrials.gov NCT00774150.

Results: Within anticipation, ANX youth showed reduced brain activity compared to HC youth in dorsal striatum and thalamus (t = 4.42−4.45), which normalized with treatment. Reduced pretreatment activation in the precuneus (t = 4.46) predicted treatment response and pre-to-post increases in posterior cingulate cortex (PCC) predicted more depressive symptoms at 2 years (beta = 17.98, t(48) = 2.94, p < 0.01). Within feedback, ANX compared to HC youth had elevated pretreatment activity in angular gyrus, PCC and inferior frontal gyrus (t = 3.85−4.53), which normalized with treatment. Pre-to-post reductions in brain activity in left angular gyrus (t = 4.41) corresponded with treatment response.

Conclusions: Treatment response corresponded with activity in default-mode associated regions (e.g., PCC, angular gyrus), suggesting that self-referential processes interfere may with reward feedback in youth with anxiety disorders.

Disclosure: Nothing to disclose.

14.3 Predicting Treatment Response Using a Gamified Attention Bias Modification Training in Anxious Youth: A Randomized Controlled Trial

Katharina Kircanski

National Institute of Mental Health, Bethesda, Maryland, United States

Background: Combining cognitive behavioral therapy (CBT) with attention bias modification training (ABMT) may maximize therapeutic benefit, as CBT and ABMT may target different dysfunctional threat processes. Yet, the corresponding neural mechanisms are not well understood. The current preregistered double-blind, randomized controlled trial (RCT) assigned youth with anxiety disorders to either an active or sham gamified ABMT, delivered alongside manualized CBT. Treatment group differences in associations between amygdala connectivity and treatment response were examined.

Methods: Youth (N = 121; 8−17 years) with a primary anxiety disorder completed either active or sham ABMT alongside CBT. ABMT integrated a modified dot-probe task with a visual search embedded in a game to enhance attentional engagement. Symptom improvement was measured by the clinician-rated Pediatric Anxiety Rating Scale (PARS) and Clinical Global Impression-Improvement (CGI-I) score. Neurobiological correlates of attention bias were assessed using a dot-probe task during fMRI before and after the intervention in patients and healthy controls.

Results: Over time, PARS decreased significantly in both ABMT groups (F(2, 219) = 70.34; p < 0.001). Active ABMT did not enhance this effect (F(2, 219) = 0.35; p = 0.71). There was no difference in the proportion of CGI-I treatment responders vs. non-responders between active and sham (X2 = 0.76, p = 0.38). Baseline amygdala connectivity differences emerged between patients and healthy controls for inferior parietal, middle cingulate, and middle frontal gyrus clusters, and lower connectivity during threat-congruent trials predicted worse posttreatment PARS for similar regions in patients (Fs(2, 124) > 10.31; ps < 0.001). In the superior temporal gyrus, an association emerged between baseline amygdala-temporal connectivity and posttreatment PARS in the CBT+sham ABMT group (F(1, 62) = 29.48; p < 0.001); no such correlation was seen in the active group.

Conclusions: Hypotheses that youth receiving active ABMT + CBT would show greater improvement in anxiety symptoms were not supported. The pattern of amygdala connectivity suggests that patients with lower baseline connectivity may be less likely to respond to treatment. Findings may inform for whom ABMT is most effective in the context of CBT and contribute to baseline CBT stratification.

Disclosure: Nothing to disclose.

14.4 Changes in Frontoamygdala Circuitry Associated With a Parent-Focused Treatment for Pediatric Anxiety Disorders

Dylan Gee

Yale University, New Haven, Connecticut, United States

Background: Although cognitive behavioral therapy (CBT) can be effective for treating pediatric anxiety disorders, up to 50% of clinically anxious youth do not respond sufficiently to CBT. Parents play a central role in guiding children’s learning about threat and safety and in regulating their anxiety, highlighting the possibility of a parent-focused treatment to enhance clinical efficacy. However, the neural mechanisms that support anxiety reduction in parent-focused treatment remain unknown. Comparing neural correlates of symptom change in parent-focused treatment versus CBT may help to guide treatment optimization and targeted treatment recommendations.

Methods: Parents and their children (N = 76; 6−12 years old with an anxiety disorder; male and female children included) were randomly assigned to Supportive Parenting for Anxious Childhood Emotions (SPACE, delivered to parents) or cognitive behavioral therapy (CBT). Children completed a resting-state fMRI scan before and after treatment (55 with usable pre- and post-treatment resting-state fMRI data). Group Iterative Multiple Model Estimation (GIMME) was used to construct person-specific frontoamygdala networks. From the person-specific networks, we calculated network density by counting the number of individual-level connections between prefrontal regions of interest (ROIs) and bilateral amygdala ROIs. Group-level models included covariates of age and sex.

Results: There were no differences in anxiety reduction among children who received CBT compared with children whose parents engaged in SPACE. Controlling for connectivity at baseline, there was a significant main effect of treatment group on post-treatment frontoamygdala connectivity network density (edge count), b = −0.45, p = 0.02. Specifically, youth whose parents received SPACE had more frontoamygdala edges than youth who received CBT.

Conclusions: Taken together, while both treatments were clinically efficacious, these findings suggest that an entirely parent-based intervention may reduce children’s anxiety by modulating frontoamygdala connectivity in a unique way.

Disclosure: Nothing to disclose.

Panel

15. Opioid Signaling in the Etiology and Treatment of Behavior Relevant to Depression

15.1 The Role of the Endogenous Opioids in Regulating Prefrontal Cortical Threat Processing and Circuit Function

Hugo Tejeda

National Institute of Mental Health, Bethesda, Maryland, United States

Background: The medial prefrontal cortex (mPFC) is a critical hub in limbic circuitry that exerts executive control over adaptive behavior that ensures survival, and its dysfunction has been implicated in psychiatric disorders, including mood and anxiety disorders. Dynorphin (Dyn), signaling via the kappa-opioid receptor (KOR) in limbic circuits has been implicated in mediating negative affective states and mal-adaptive behaviors in response to stress. However, there is a gap in our understanding of the role of this system in shaping the way mPFC circuits integrate information to influence adaptive behaviors induced by threats and physiological stressors.

Methods: Male and female WT and PDyn-Cre were used. We employed fiber photometry and single-cell calcium imaging approaches to monitor mPFC cells that express Dyn (mPFC-Dyn) neuron activity, Dyn release using the genetically-encoded KOR sensor kLight, or pan-neuronal vmPFC network dynamics (n = 5−9). The platform-mediated avoidance (n = 11−20) or a modified looming stimulus task (n = 13−23) were used to measure motivational conflict in control mice and mice with vmPFC Dyn knockdown. Slice electrophysiology experiments were performed on acute brain slices from Rhesus Macaques (n = 8) or from resected human temporal or visual cortex (n = 7). Data were analyzed using two-way ANOVA or t-test.

Results: Here, we characterized mPFC-Dyn neurons in the ventral mPFC (vmPFC) that are activated by stress and release Dyn peptides (p = 0.01). vmPFC-Dyn neuron activation and subsequent dynorphin release promotes active defensive behaviors that are conducive to reward attainment in the face of threats in both the PMA and looming stimulus conflict task (p < 0.0001). We also demonstrate the vmPFC Dyn release facilitates encoding of central fear states in vmPFC networks (p = 0.017). Lastly, we describe a circuit-based cellular framework by which Dyn / KOR signaling in specific excitatory inputs to the vmPFC and at various inhibitory microcircuits gates input-output transformations in a pathway-specific manner in rodents, non-human primates, and in humans.

Conclusions: Together, this work provides a model wherein pathway-specific Dyn / KOR regulation of excitation/inhibition balance differentially gates the activity of mPFC projection neurons that control adaptive behaviors in the face of threats.

Disclosure: Nothing to disclose.

15.2 Endogenous Enkephalin Regulates Aversion in Dorsal Raphe Nucleus

Daniel Castro

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

Background: Over the last several decades, significant effort have been made to understand the neural mechanisms underlying opioid analgesia and addiction. However, this approach has stymied research into other meaningful areas, such as the role of opioids in affect, mood, and motivation. The dorsal raphe nucleus (DRN) has long been associated with these psychological processes, and is highly enriched in both opioid receptors and peptides. However, how opioids in DRN directly impact affect and motivation is largely unknown.

Methods: We use CRISPR-Cas9 mediated knockdown to disrupt enkephalin production and measure changes in appetitive and aversive behaviors (expected n = 12 mice/sex). Behaviors include: thermal and mechanical allodynia, odor preference/avoidance, social interaction, food intake, and naloxone-precipitated withdrawal. All behavioral tests are counterbalanced. When possible, mice are tested within subjects. For all studies, male and female mice are used.

Results: CRISPR knockdown of enkephalin peptide in the DRN of Penk-Cre mice produced a hyperalgesic response to low-dose carrageenan injection compared to controls (Two-Way ANOVA, between subjects, Sidak’s multiple comparison; F (1,14) = 9.482, p = 0.0082; t(8,8) = 4.002, p = 0.0008; Cre- males = 4, Cre- females = 4, Cre+ males = 5, Cre+ females = 3). It also enhanced avoidance of an aversive odor (Welch’s t-test, between subjects, t(4,8) = 2.903, p = 0.0247; Cre- males =6, Cre- females = 2, Cre+ females = 4). There were no differences between Penk-Cre+ and littermate Penk-Cre- controls in appetitive behaviors (social interaction, food intake). CRISPR knockdown of enkephalin in DRN of vGlut2-Cre mice replicated the enhanced allodynia following carrageenan, while knockdown in vGat-Cre or vGlut3-Cre mice did not (Two-Way ANOVA, between subject, Sidak’s multiple comparison; F(4,32) = 5.959, p = 0.0011; t(8,7) = 2.803, p = 0.0238; Penk-Cre- males = 4, Penk-Cre- females = 4, vGlut2-Cre+ males = 3, vGlut2-Cre+ females = 4).

Conclusions: These results indicate that DRN enkephalin peptide acts to buffer responses to aversive stimuli. Future work will further characterize the role of DRN enkephalin during other aversive behaviors. We will also use multi-color single cell imaging to assess online encoding of enkephalin neurons during these behavioral states.

Disclosure: Nothing to disclose.

15.3 Investigating Tianeptine as a Novel Therapeutic for Neuropathic Pain: Normalization of Behavior and in Vivo Metabolite Levels

Elizabeth Pekarskaya

Columbia University, New York, New York, United States

Background: Neuropathic pain (NP), characterized by increased pain sensitivity and highly comorbid with major depression, is currently treated with tricyclic antidepressants and anticonvulsants. However, only 40-60% of patients experience relief, indicating a need for novel therapeutics. We find the atypical antidepressant tianeptine (TIA), a mu-opioid receptor (MOR) agonist, reduces allodynia in male and female mice and affective symptoms in male mice in a spared nerve injury (SNI) mouse model of NP long after its acute analgesic effects have worn off. A central question is how and where tianeptine acts to decrease allodynia and affective symptoms.

Methods: 8-week-old male (N = 39; 20 sham, 19 SNI) and female (N = 38; 20 sham, 18 SNI) C57BL/6J mice underwent either sham surgery or tibial-sparing SNI surgery. 10 weeks after surgery, half of each group was administered TIA (30 mg/kg, 2x/day, i.p.) or saline (2x/day, i.p.) for 2 weeks. Behavioral tests were then conducted, including open field, marble burying, social interaction, novelty-suppressed feeding, von Frey, and hot plate. We used magnetic resonance spectroscopy (MRS) to assess SNI-induced changes in relative major neurotransmitter levels in a separate cohort of male mice (sham = 10, SNI = 8−10). Open field, marble burying, and von Frey tests were performed to correlate with behavioral changes. Animals were then administered tianeptine (30 mg/kg, 2x/day, i.p.) for 2 weeks before repeating imaging and behavioral testing.

Results: SNI mice had increased allodynia in males (p < 0.0001) and females (p = 0.004), which was reversed with tianeptine treatment (M – p < 0.0001; F– p = 0.0163). Male SNI mice showed decreased exploration and increased social avoidance and repetitive digging, which was reversed with tianeptine, while female SNI mice were not distinguishable from controls for affective measures. MRS revealed a decrease in relative GABA (p = 0.018) and GLX (glutamate+glutamine) (p = 0.013) in SNI animals, while glutamate levels were similar between groups (p = 0.489).

Conclusions: Our findings demonstrate tianeptine’s promise as a novel treatment for chronic neuropathic pain. We also identified alterations in excitatory and inhibitory metabolites in the habenula, indicating this region as a possible target for future chronic pain treatments.

Disclosure: Nothing to disclose.

15.4 Early-Life Exposure to Fluoxetine Impacts Endogenous Opioid Modulation of Motivated Behavior

Sarah Canetta

Columbia University, New York, New York, United States

Background: Brief exposure to the SSRI, fluoxetine (FLX), during early life leads to a variety of affective behavioral changes in adulthood, including increased avoidance and despair as well as impaired motivation. We were interested in refining the reward-related behavioral changes following early-life FLX exposure as well as identifying therapeutic strategies for reward-related deficits. To that end, we focused on the opioid system given our prior finding that chronic administration of the mu opioid receptor (MOR) agonist, tianeptine (TIA), improved avoidant behavior in these early FLX-exposed animals.

Methods: Mice were exposed to FLX (10 mg/kg) in early-life (P2-11, EL FLX) or vehicle (EL VEH) and in adulthood (> 90) were tested in the progressive ratio, lickometer and pavlovian-to-instrument transfer tasks. We chronically administered the MOR agonist, TIA (30 mg/kg i.p., 2x/day for 14 days), or the MOR antagonist, methocinnamox (MCAM, 10 mg/kg s.c., 1x/day every 3 days for 3 weeks), in EL FLX, EL VEH or no early exposure control mice prior to behavioral testing. We also tested acute MCAM (10 mg/kg s.c.) in these tasks. Sample sizes ranged from 11 to 29 mice and both sexes were included.

Results: Adult EL FLX mice showed impaired motivation (e.g. decreased breakpoint: p = 0.0148) while hedonic response to reward and pavlovian-to-instrumental transfer was intact. TIA had no effect on motivation but surprisingly chronic MCAM improved motivation in EL FLX mice (p = 0.0059) while trending towards decreasing hedonic response to reward. Chronic MCAM did not affect motivation in control mice but significantly decreased hedonic response to reward (p = 0.0006). Acute MCAM increased motivation in EL FLX (p = 0.0058), but not in EL VEH, mice.

Conclusions: Brief, early-life exposure to FLX impairs adult motivation, while leaving hedonic response to reward and cue-reward learning intact. Surprisingly, the MOR antagonist, MCAM, improves motivation in EL FLX mice. MCAM does not impact motivation in control animals and decreases hedonic response to reward in both control and EL FLX mice. Together, this work suggests that briefly elevating serotonin signaling in early life affects the development of the endogenous opioid system such that it suppresses motivation in adulthood. Ongoing experiments aim to identify the mechanism underlying this effect.

Disclosure: Nothing to disclose.

Study Group

16. Generating Greater Inclusive Excellence at the ACNP: Lessons Learned Through the Lens of Hispanic Scientists and Physicians

Hugo Tejeda, Millie Rincón-Cortés, Gustavo Angarita, Luis Colon-Perez, Veronica Alvarez, Angélica Torres-Berrío, Laura O’Dell, Joseph Cheer

National Institute of Mental Health, Bethesda, Maryland, United States

Study Group Summary: Despite significant progress in increasing diversity and inclusion at ACNP, significant challenges that stymie scientific and professional development of members from disadvantaged groups, including Hispanics, remain. Examining challenges and identifying action plans is necessary to foster inclusive excellence at ACNP, consistent with the college’s core values. The purpose of the proposed study group is to foster a community-based discussion to generate important considerations and strategies that take into consideration diversity within Hispanic groups and the equally complex and diverse challenges they encounter. Goals of this proposed panel include fostering inclusivity of Hispanic researchers at the ACNP, providing participants an extended support network, identifying tailored resources and opportunities to facilitate the growth of Hispanic researchers, and providing a synthesis of problems and solutions faced by Hispanics and URMs broadly. Though Hispanics generally face disadvantages in research and medicine, experiences within the Hispanic community vary dramatically and are the result of complex interactions with a multitude of extraneous factors, including place/nation of birth, gender, limited access to academic and research resources, cultural biases or hurdles, career stage when research in the United States was initiated, sexual orientation, and financial in/stability. Recognizing the complexity of the Hispanic researcher and physician will improve the ability of ACNP to incorporate Hispanics more efficiently into the college and identify principles of relevance for understanding challenges URMs face. These varied complex experiences manifests in stark differences in how professional and scientific development is achieved. Funding opportunities, or the lack thereof, starkly differ for Hispanic United States citizens in comparison to foreign nationals, creating further disparities even within Hispanic researchers. Discussion will include highlighting funding opportunities available to URMs from distinct backgrounds. Likewise, approaches for mentoring Hispanics from diverse backgrounds and the importance of training the next generation of mentors and leaders in the field of neuropsychopharmacology to have maximal impact by capitalizing on their unique perspectives. Generating a support network that fosters mentorship and connects Hispanic and URMs across career stages is necessary to enhance inclusive excellence at the ACNP. We will discuss how formal and informal interactions with minority-serving institutions serve as conduits for identifying talented Hispanics and other URMs and enhance on-going work by the ACNP in this domain. Additionally, we will discuss how Hispanic participation as research subjects and researchers is essential for increasing access to quality mental health care and increase representation of Hispanics and other URMs in clinical research. One of the goals of the study group is to generate a synthesis of the discussion to highlight principles and practical actionable items that will aid in increasing recruitment and retention of Hispanics at the ACNP, provide a mentoring and support network for Hispanic researchers in the field of neuropsychopharmacology, to eventually change how mental health clinical research and practice is carried out for and by Hispanics and URMs to address pervasive health disparities.

Disclosure: Nothing to disclose.

Panel

17. New Insight Into the Disordered Hippocampus in Schizophrenia: From Pathogenic Mechanisms to Therapeutics

17.1 Abstract Not Included

17.2 Magnetic Resonance Spectroscopy Applied to Patients and Mice Implicates the Glutamate Metabolic Cycle as a Pathogenic Pathway and Therapeutic Target in Schizophrenia

Scott Small

Columbia University, New York, New York, United States

Background: The ‘glutamate metabolic cycle’ is a key pathway that regulates synaptic glutamate. Here we test the hypothesis that pathway defects, which are known to differentially target the hippocampus, are pathogenically linked to schizophrenia and that GLS1 inhibitors might have therapeutic potential. In patients and mice use 1H-MRS to test for joint changes in GLX and GABA, reflective of glutamate changes.

Methods: Human Study: 1H-MRS data acquired in a 3 Tesla scanner from the hippocampus of 15 healthy controls and a cohort of 35 patients with attenuated psychotic symptoms (APS) who have prodromal schizophrenia or related disorders. Compared to the controls (mean age = 21.1 years of age), the patient group (mean age = 23.4 years of age) was slightly but significantly older (t = 4.7, p = 0.03). No other significant differences were observed.

Mouse Study: 1H-MRS data acquired in a 9.4 Tesla scanner from the hippocampus of 10 4-week-old C7BL/J6 mice, 5 who were administered 100 mg/kg of a brain-penetrant GLS1 inhibitor, and 5 were administered placebo control.

Results: Human Study: A between-group ANOVA was performed, including group (controls vs. patients) as the independent variable and GLX1, GABA, and GLX + GABA as the dependent variables, and age and sex as covariates. While compared to controls the patient group was found to have elevations in GLX (F = 4.1 p = 0.049) and a trending increase in GABA (F = 3.5 p = 0.068), the joint increase in GLX + GABA most reliably distinguished the groups (F = 7.7, p = 0.008). A Pearson’s correlation analysis revealed that compared to each individual measure, only GLX + GABA significantly correlated with both total positive symptoms (beta = 0.42. p = 0.003) and total negative symptoms (Beta = 0.33, p = 0.028).

Mouse study: A between-group ANOVA was performed, including group (drug vs. placebo) as the independent variable and GLX1, GABA, and GLX + GABA as the dependent variables, While compared to placebo the drug group was found to have decreases in GLX (F = 7.5 p = 0.029) and a trending decrease in GABA (F = 4.5 p = 0.072), a joint decrease in GLX + GABA most reliably distinguished the groups (F = 18.4, p = 0.004).

Conclusions: Collectively and taken together with previous studies these findings implicate the glutamate metabolic cycle as a pathogenic pathway and a therapeutic target in schizophrenia and related psychotic disorders.

Disclosure: Nothing to disclose.

17.3 Hippocampal Hyperactivity From Humans to Rodents: Multi-Scale Methods: and Applications to Psychosis

Gemma Modinos

Institute of Psychiatry, Psychology and Neuroscience, King’s College London, London, United Kingdom

Background: Preclinical evidence has shown that hyperactivity of the hippocampus due to neural excitatory (glutamatergic)/inhibitory (GABAergic) imbalance leads to psychosis-relevant phenotypes, which can be prevented by peripubertal exposure to the GABA-enhancing drug diazepam. In humans, our previous functional MRI work found that hippocampal hyperactivity (by regional cerebral blood flow, rCBF) in individuals at clinical high-risk for psychosis (CHR-P) can be normalized by acute diazepam administration. We also identified a link between specific GABAergic cell-type disruption and increased hippocampal rCBF by back-translation in a genetic rodent model. However, the specific cellular and molecular signatures underlying rCBF abnormalities in CHR-P individuals are unknown.

Methods: We linked case-control rCBF alterations to gene and neuroreceptor expression in individuals at CHR-P (n = 129; mean age = 22.7[SD = 4.1], 57% male) and with schizophrenia spectrum disorders (SSD, n = 122; mean age = 39.5[SD = 11.6], 75% male) versus their matched HC (n = 58; mean age = 24.8[SD =4.3], 55% male; and n = 116 mean age = 36.9[SD = 11.1], 72% male, respectively). Case-control t-maps were parcellated into 122 cortical and subcortical regions (Schaefer+Xiao atlases). Transcriptomic data from the Allen Human Brain Atlas and neuroreceptor maps from 19 publicly available PET images were parcellated into the same 122 regions. Correlations between case-control rCBF alterations and gene expression and neurotransmitter systems were assessed with Spearman’s rank correlations and dominance analyses. Significance was determined using FDR-corrected spatial-autocorrelation preserving null models.

Results: The CHR-P rCBF profile tracked the distribution of oligodendrocyte, oligodendrocyte precursor, and endothelial cell gene markers, as well as of D2, DAT, NET and NMDA receptors (all pFDR < 0.05). In contrast, the SSD rCBF profile tracked the distribution of oligodendrocyte, pericyte, oligodendrocyte precursor and microglia gene markers, and of 5-HT2A and GABA-A receptors (all pFDR < 0.05). rCBF-neuroreceptor co-expression was strongest in subcortical regions including the hippocampus (Cook’s distance).

Conclusions: These findings provide novel evidence linking specific cellular and molecular pathways to rCBF alterations in the psychosis spectrum.

Disclosure:

Financial Relationships Details Boehringer-Ingelheim, Consultant; (Self)

17.4 A Longitudinal Study of Hippocampal Subfield Volumes and Hippocampal Glutamate Levels in Antipsychotic-Naïve First Episode Psychosis Patients

Adrienne Lahti

University of Alabama at Birmingham, Mountain Brk, Alabama, United States

Background: Previous studies have implicated hippocampal abnormalities in the neuropathology of psychosis spectrum disorders. Reduced hippocampal volume has been reported across all illness stages, and this atrophy has been hypothesized to be the result of glutamatergic excess. To test this hypothesis, we measured hippocampal subfield volumes and hippocampal glutamate levels in antipsychotic naïve first episode psychosis patients (FEP) and the progression of volume and glutamate (Glu) levels over a 16-week antipsychotic drug (APD) trial. We aimed to determine if subfield volumes at baseline were associated with glutamate levels, and if baseline glutamate levels were predictive of change in subfield volumes over time.

Methods: We enrolled ninety-three medication-naïve FEP participants and 80 matched healthy controls (HC). T1 and T2 weighted images and magnetic resonance spectroscopy (MRS) data from a voxel prescribed in the left hippocampus were collected from participants at baseline and after 6 and 16 weeks of APD treatment. Hippocampal subfield volumes were assessed using FreeSurfer 7.1.1., while glutamate levels were quantified using jMRUI version 6.0.

Results: Results from group x time general linear mixed models for volume showed significant group x time interactions for the whole hippocampus (F(2, 228) = 5.81, p = 0.0035) as well as for CA1 (F(2,229) = 4.58, p = .0013), CA3 (F(2,240) = 5.52, p = 0.0045), CA4 (F(2,235) = 4.84, p = 0.0087), presubiculum (F(2,241) = 9.34, p = 0.0001), and the molecular and granule cell layer of the dentate gyrus (GC/ML/DG) (F(2,236) = 4.38, p = 0.014) subfields. Post hoc comparisons found regional subfield volume deficits in the CA1 (p = 0.017), and presubiculum (p < 0.01) in FEP at baseline and week 16 (CA1: p = 0.0002; presubiculum: p = 0.0003), that further expanded to include GC/ML/DG (p < 0.01) and CA4 (p < 0.01) by week 16. Baseline hippocampal Glu levels in FEP were not significantly different than those of HC, and there was no effect of treatment on Glu levels. Glu levels were not related to initial subfield volumes or volume changes over 16 weeks.

Conclusions: We report a progressive loss of hippocampal subfield volumes over a period of 16 weeks after initiation of treatment, suggestive of a progressing neuropathology. Our results do not suggest a role for glutamate as a driving factor.

Disclosure: Nothing to disclose.

Panel

18. Unraveling the Impact of Early Life Adversity on Adaptation: Translational Insight Into Novel Molecular and Neural Circuit Mechanisms

18.1 Early Life Adversity and Fear Generalization—a Role for Serotonin in the Dentate Gyrus

Christoph Anacker

Columbia University, New York, New York, United States

Background: Threat generalization can be adaptive for survival, but high generalization in the absence of threat can lead to excessive fear, which is common in psychiatric disorders. Early life adversity (ELA) is a risk factor for psychiatric disorders with a fear generalization component, such as depression, generalized anxiety, or PTSD, but the underlying mechanisms are not known. We investigated ELA effects on serotonin (5HT) regulation of the ventral dentate gyrus (vDG) as a potential neurobiological mechanism for enduring effects of ELA.

Methods: We exposed litters to the limited bedding and nesting (LBN) model of ELA from postnatal day (P) 3−10 and tested male and female mice in adulthood (P56). We first determined LBN effects on 5HT neuron electrophysiological properties using ex vivo slice recordings of GFP-labeled 5HT neurons. We then used mice with dis-inhibited 5HT neurons due to postnatal knockdown of 5HT1A autoreceptors (5HT1A KD) to test how increased 5HT neuron activity modulates LBN effects on contextual fear generalization in which mice receive a foot shock in Context A and freezing behavior is assessed in a new, non-shock associated Context B. We used fiber photometry of vDG Ca2+ signals to test LBN effects on granule cell activity during generalization and how increased 5HT neuron activity in 5HT1A KD mice modulates this effect. Lastly, we used hM4Di DREADD receptors to test how direct inhibition of vDG granule cells modulates LBN effects on generalization.

Results: In females, LBN reduced 5HT neuron firing compared (*p = 0.04, n = 11) and increased fear generalization to standard reared controls (i.e, freezing in Context B; **p = 0.002, n = 15−19). Dis-inhibiting 5HT neurons through 5HT1A KD rescued LBN effects on generalization (***p < 0.001, n = 14−18). In the female vDG, LBN increased granule cell Ca2+ activity during fear generalization (*p = 0.04, n = 6−9) and this heightened activity was inhibited in 5HT1A KD mice (**p = 0.007, n = 6). Direct vDG inhibition using hM4Di activation rescued LBN effects on generalization (**p = 0.009; n = 7−12), indicating that 5HT1A KD effects on generalization may be mediated by 5HT effects on vDG inhibition. No LBN or 5HT1A KD effects were observed in males (all p > 0.05).

Conclusions: Our findings reveal that ELA enhances contextual fear generalization in females by impairing serotonergic inhibition of vDG hyperactivity.

Disclosure: Nothing to disclose.

18.2 A Critical Period for Cognitive Flexibility in the Maturation of the Hippocampus to Medial Prefrontal Cortex Pathway

Maithe Arruda-Carvalho

University of Toronto, Toronto, Canada

Background: The hippocampus (HPC) plays a key role in learning and memory, anxiety and stress. The HPC is segregated across its septotemporal axis into dorsal, intermediate and ventral domains showing unique molecular, projection and downstream pathway-dependent behavioral profiles, and yet very little is known about the ontogenesis of these highly specialized subcircuits.

Methods: Here we used a combination of viral tracing and optogenetic-assisted patch clamping to map changes in domain-defined hippocampus CA1 afferent projections from postnatal day (P)10 to P60 in male and female C57BL/6J mice. Lastly, we inhibited the ventral (v)CA1-medial prefrontal cortex (mPFC) pathway with chemogenetics in juvenility (P15-P30) or adulthood (P60-P75) and assessed its effects on extradimensional attentional set shifting during adulthood.

Results: We found a unique pattern of age-dependent changes in vCA1 terminal density, with opposite changes in intermediate (i)CA1 projections in the lateral septum and the prelimbic (PL) subregion of the mPFC. Optically-evoked CA1 responses onto mPFC showed increased vCA1-PL glutamate release probability from P30 only in females (one-way ANOVA F3, 38 = 8.75 p = 0.0002 ; P21 x P30: p = 0.0008; P21 x P60: p = 0.0004), while vCA1-Infralimbic (IL) release probability remained stable from P15. In contrast, iCA1-PL presynaptic function stabilized earlier, by P21, followed by a relatively delayed maturation of iCA1-IL presynaptic output at P60 (one-way ANOVA F4,63 = 2.83 p = 0.032; P15 vs P60: p = 0.037). vCA1-PL and -IL postsynaptic transmission was stable, but iCA1-PL and -IL synapses increased AMPA:NMDA ratios from P30 (iCA1-PL: F2,43 = 4.10, p = 0.023; P15 x P30: p = 0.017; iCA1-IL: F2,49 = 4.52, p = 0.016; P15 x P30: p = 0.035; P15 x P60: p = 0.032). Chronic inhibition of the vCA1-mPFC pathway during juvenility, but not adulthood, led to a deficit in extradimensional set shifting exclusively in females (two-way ANOVA virus F1, 25 = 23.57, p < 0.0001, sex F1, 25 = 50.56, p < 0.0001 and interaction F1, 25 = 11.67, p = 0.0022; mCherry x hM4D females p < 0.0001).

Conclusions: Our data establish a timeline for the postnatal maturation of HPC efferents in a domain- and target-defined manner, identifying a sex- and pathway-specific developmental sensitive period with critical implications for early life influences on adult cognitive function.

Disclosure: Nothing to disclose.

18.3 Early Life Adversity Heightens Threat Sensitivity Through Epigenetic Priming in NAc Ensembles

Catherine Pena

Princeton Neuroscience Institute, Princeton, New Jersey, United States

Background: Early-life adversity (ELA) sensitizes individuals to subsequent stressors such as social threat. While such sensitization may be adaptive under certain circumstances, threat hypersensitivity and constant vigilance can also lead to anxiety and mood disorders. Within the nucleus accumbens (NAc), ELA sensitizes both cellular and transcriptional response to later stress. However, the molecular mechanisms linking initial activation of neurons by ELA with continued threat sensitivity across the lifespan have been poorly understood.

Methods: We employed activity-dependent neuronal tagging, affinity-purification, and ATAC-seq within both experience-activated and non-activated cell populations, both following ELA and in adulthood. We used a variety of computational approaches to determine the degree and endurance of chromatin changes and how such changes predict gene expression response to later social threat. Finally, we postnatally over-expressed Setd7, encoding an enzyme that monomethylates histone-3 lysine-4, associated with our observed chromatin changes and epigenetic priming, and examined behavioral response to social threat. All experiments included both male and female mice.

Results: Our findings reveal epigenetic priming as a novel biological mechanism through which ELA becomes encoded within the brain. We find: (1) chromatin is more open and accessible following ELA, specifically within stress-activated neurons; (2) ELA increases accumulation of H3K4me1, a histone modification associated with enhancer priming; (3) genes regulated by opened chromatin are more likely to be transcriptionally reactive to future stress; and (4) we functionally show that epigenetic priming during postnatal development by deposition of H3K4me1 is sufficient to mimic ELA and increase behavioral sensitivity to later stress. In addition, we find evidence for accelerated chromatin development following ELA, and that the impact of ELA propagates across development from a small initially-activated cell population to cause chromatin remodeling within the broader NAc cell population by adulthood.

Conclusions: Together, our results show that a molecular memory of ELA is encoded at an epigenetic level through changes in chromatin architecture. This constitutes a novel biological mechanism by which ELA programs lifelong threat sensitivity.

Disclosure: Nothing to disclose.

18.4 The Effect of Maternal Stress and Genetic Risk on Cognitive Flexibility and Depression Development – Translating Between Rodents and Humans

Milenna van Dijk

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

Background: Maternal stress exposure disrupts hippocampal structure and function and increases risk for psychopathology in some but not all offspring. We examined cognitive flexibility, genetic risk and gene x environment interactions to elucidate individual differences in adaptability to adversity. We employ novel cross-species functional genomics to translate hippocampal gene expression data from mouse models to hippocampus-specific expression based polygenic scores (PGS) in humans to assess whether differences in predicted gene expression can explain DG structural difference, cognitive flexibility, and depression in children exposed to adversity.

Methods: We examined ASEBA ASR, KSADS, CBCL and family history to assess maternal and child clinical histories and Wisconsin Card Sorting Task for cognitive flexibility using data (n = 5888- 8540) from the Adolescent Brain and Cognitive Development Study. Depression PGS was calculated from Howard (2018) summary statistics. A gene network associated with maternal stress exposure was extracted from mouse dentate gyrus (DG) RNA sequencing data and translated to a human expression based PGS that predicts expression of the gene network in the human hippocampus (DG ePGS) using GTeX. We extracted DG mean diffusivity and volumes with FreeSurfer and MRtrix. Regressions adjusted for family structure, scanner site, sex, age, population stratification and scanner head motion.

Results: Maternal stress exposure in the offspring associated with decreased cognitive flexibility and DG structure. Cognitive flexibility and DG structure was associated with future symptoms. Depression PGS was associated with cognitive flexibility, and depressive symptoms. Higher DG ePGS was associated with higher rates of depression and DG ePGS interacted with early adversity, such that only children with high DG ePGS who were exposed to adversity had disrupted DG microstructure and higher rates of depression development (all ps < 0.05).

Conclusions: Genetic risk is associated with individual differences in cognitive flexibility, hippocampal changes and psychopathology and increases vulnerability to adversity. ePGS provide an avenue of using gene expression data from mouse models to predict adaptability to stress in live human populations.

Disclosure: Nothing to disclose.

Panel

19. Non-Psychedelic Psychedelics: Preclinical and Clinical Evidence of the Therapeutic Actions of an Oxymoron

19.1 Identification of Signaling Pathways that Discriminate Psychedelics From Non-Psychedelics: A Focus on Serotonin GPCR Signaling

John McCorvy

Medical College of Wisconsin, Milwaukee, Wisconsin, United States

Background: Classical psychedelics, such as LSD and psilocybin, have been designated with “breakthrough therapy” status by the FDA thus ushering in a “psychedelics renaissance” in neuropsychopharmacological research. However, it has been proposed that the psychedelic effects may not be entirely necessary for therapeutic effects, including the rapid-acting antidepressant effects. Thus, investigated newer and older compounds devoid of psychedelic effects have been termed as “non-psychedelics” and are of paramount interest for therapeutic effects. Therefore, understanding both the polypharmacology and signaling pathways necessary for discriminating psychedelics versus non-psychedelics are needed.

Methods: Using a panel of psychedelics, established non-psychedelic, and probe compounds, we employ kinetically-sensitive BRET platforms and in vitro assays to detect signaling pathways that are associated with prediction of in vivo psychedelic and non-psychedelic activity.

Results: Here we show that known prototypical psychedelics and established non-psychedelic agents all show polypharmacology across the 5-HT GPCRome with little selectivity for the 5-HT2A receptor. We also show that psychedelics produce robust and efficacious signaling at the 5-HT2A receptor measuring Gq/11 signaling whereas non-psychedelic agents show partial agonism. We support this finding with custom synthesized probe compounds that are 5-HT2A selective and biased for either Gq/11 and β-arrestin2 signaling pathways, making it clear which receptor(s) and signaling pathway(s) are responsible for psychedelic potential. Finally, we show evidence that predicts psychedelic versus non-psychedelic activity and attempt to reconcile key details about their molecular recognition at the 5-HT2A receptor.

Conclusions: These results establish that 5-HT2A-Gq signaling efficacy is important for psychedelic-like effects and that non-psychedelics can be engineered with partial agonism at the 5-HT2A receptor, thus fine-tuning their therapeutic properties, especially as it pertains toward their neuroplastic potential. Importantly, we outline future strategies of non-psychedelics for targeting other serotonin GPCRs.

Disclosure: Nothing to disclose.

19.2 Characterizing Non-Psychedelic Psychedelics: Phenotypic Fingerprinting of Lisuride and LSD

Scott Thompson

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

Background: Psychedelics provide rapid and effective treatment for a range of neuropsychiatric disorders, including depression. However, their implementation as a widespread clinical intervention is hindered by the characteristic alterations in perception and consciousness they produce (i.e. the trip). Psychedelics are potent agonists at all serotonin receptors (5HTRs). 5HT2ARs mediate psychedelic-induced perceptual alterations and might be essential to the therapeutic response. To harness the clinical benefits of psychedelics while lowering economic barriers, biased 5HT2AR agonists that do not impact perception and consciousness are currently being commercially developed. The head twitch response (HTR) is a widely used preclinical assessment of psychedelic responses. However, it is unknown whether the absence of an HTR is sufficient to identify truly non-psychedelic compounds. Here we carried out a phenotypic comparison of the classic psychedelic, lysergic acid diethylamide (LSD) to the archetypal non-psychedelic psychedelic, lisuride, a potent 5HT2AR agonist that does not cause head twitches in mice.

Methods: We used a combination of locomotor (open field, rotorod) and cognitive (the puzzle box) tasks, and EEG recording from the PFC.

Results: Lisuride (0.5mg/kg), but not LSD (0.1mg/kg), produced a significant impairment of motor behavior in the open field and on the rotarod. Rotorod impairement gradually declined over a one-hour period after injection. Lisuride also significantly impaired performance in the puzzle box. Both motor and cognitive effects of lisuride persisted when 5HT2ARs were pharmacologically blocked with MDL100,907 (0.5mg/kg). Lisuride and LSD both produced alterations in electroencephalograms (EEGs) recorded over the prefrontal cortex (PFC) that were different from each other. No statistically significant sex differences were observed. Experiments are in progress to determine if any actions of LSD and lisuride are effective after pharmacological block of 5HT1ARs.

Conclusions: Combined, these results suggest that although lisuride may not be acting as a traditional psychedelic, it does impact both motor and cognitive function. Further, robust preclinical phenotypic characterization beyond the HTR is important for identifying non-psychedelic compounds to move forward for clinical development.

Disclosure: ProNovo Therapeutics: Founder (Self). Terran Biosciences: Advisory Board (Self). Otsuka Therapeutics: Advisory Board (Self). Althea PBC: Advisory Board (Self)

19.3 DLX-001, a Novel Non-Hallucinogenic Neuroplastogen With the Potential for Treating Neuropsychiatric Diseases

Stephanie McTighe

Delix Therapeutics, Bedford, Massachusetts, United States

Background: Psychedelics such as ketamine and psilocybin have been gaining attention for their potential to treat patients with a wide spectrum of psychiatric conditions, including major depressive disorder. However, the dissociative and hallucinogenic properties of these first-generation psychedelic agents will limit their widespread clinical use. The clinical effect of psychedelics is believed to be due to their ability to promote rapid and enduring effects on structural neuroplasticity. By maintaining beneficial effects on neuroplasticity while removing undesirable effects such as hallucinations, it may be possible to generate “neuroplastogens” that promote neuroplastic changes independent of hallucinogenic potential and, as a result, are able to reach broader clinical populations.

Methods: Effects of DLX-001 were tested in in vitro, ex vivo, and in vivo assays relevant to neuroplasticity and major depressive disorder. All animal studies were approved by the appropriate IACUC and were conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals.

Results: DLX-001 showed significant positive effects on structural neuroplasticity in in vitro and in vivo assays. DLX-001 produced significant increases in both neurite length and branching in a neurite outgrowth assay in primary rat neuronal culture. This effect was shown to be reduced by concomitant application of the 5HT2A/2C antagonist ketanserin or the mTOR antagonist rapamycin. In an ex vivo assay of layer V PFC pyramidal neuron structure, Golgi staining revealed a significant effect of DLX-001 on spine density 24 h after a single administration. Similar to ketamine, DLX-001 exhibited antidepressant-like effects in the rat Forced Swim Test 24 h after a single administration, enduring 7+ days after dosing. In the mouse head-twitch assay, DLX-001 did not induce head twitches in male mice. In a human phase 1 trial, DLX-001 has shown no evidence of hallucinogenic, dissociative, or psychotomimetic activity to date.

Conclusions: DLX-001 is an orally active, brain penetrant small molecule that promotes robust structural neuroplasticity in the PFC. Preclinical data also suggests that DLX-001 will have rapid and enduring antidepressant effects in humans. Further, the lack of hallucinogenic effects predicted preclinically has been confirmed in healthy volunteers.

Disclosure: Delix Therapeutics: Employee (Self). Sage Therapeutics: Stock / Equity - Publicly Traded Company (Self).

19.4 Translational Profile of GM-1020, a Novel Orally Bioavailable NMDA Receptor Antagonist That Achieves Robust Target Engagement Without Dissociation or Sedation

Zoe Hughes

Gilgamesh Pharmaceuticals, New York, New York, United States

Background: GM-1020 is a non-competitive NMDA receptor (NMDAR) antagonist designed to have improved oral bioavailability and tolerability compared to the original rapid acting antidepressant, ketamine. The preclinical profile of GM-1020 indicates that it achieves robust and durable antidepressant effects in rats at doses far below those causing motor impairment or disrupting sensory motor processing. Here we present Phase 1 clinical data which include a broad package of translational data which support the findings of preclinical work.

Methods: The effects of oral administration of GM-1020 (20-360 mg) were assessed in an adaptive, randomized, double blind, placebo controlled first-in-human (FIH) study to evaluate the safety, pharmacokinetics (PK), and pharmacodynamics (PD) in healthy adult volunteers.

The effects of GM-1020/placebo on treatment-emergent adverse events (TEAEs), vital signs, and 12-lead ECGs were recorded. PD assessments included resting state electroencephalogram (EEG) and the NeuroCart battery to assess sedative and ataxic effects. The Clinician-Administered Dissociative States Scale (CADSS) was used to assess dissociative effects.

Results: After oral dosing of GM-1020 (20-360 mg) exposure was dose proportional with low variability and consistent with high oral bioavailability. At the doses tested, GM-1020 was generally safe and well-tolerated with TEAEs consistent with NMDAR antagonism seen at higher doses

GM-1020 exerted changes in resting state EEG (rsEEG) at doses ≥60 mg. There was good separation between this quantitative, indirect measure of target engagement and side effects including dissociation, sedation and ataxia.

Conclusions: GM-1020 was orally bioavailable and produced significant quantitative and subjective PD effects without causing dissociation, sedation or ataxia at exposures expected to have antidepressant effects. This distinguishes GM-1020 from existing ketamine-based therapies. A Phase 2 study to explore the potential of GM-1020 as a novel, differentiated treatment for depression in patients with MDD is ongoing.

Disclosure: Gilgamesh Pharmaceuticals: Employee (Self)

Panel

20. The Dorsal Anterior Cingulate Cortex: Breaking the Agony Circuit

20.1 Cingulum Bundle Pathways Provide New Basis to Understand the Clinical Outcome of Anterior Cingulotomy

Marina Celestine

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

Background: Cingulotomy is an effective neurosurgical treatment for refractory obsessive-compulsive disorder (OCD) and major depression disorder (MDD) as well as chronic pain. This procedure involves bilateral lesions in the dorsal anterior cingulate cortex (ACC) and can also impact the cingulum bundle (CB). Despite the efficacy of cingulotomy, it is unclear which corticofugal axons are interrupted by the lesions: either cortical-cingulate connections or fibers passing through the CB. The goal of this study was to delineate the pathways that pass through the cingulotomy target to better understand the neurocircuitry underlying the effectiveness of its treatment. Here, we present the fiber organization of PFC through the CB, which fiber pathways are impacted by cingulotomies and are responsible for clinical changes, and how that interruption compares to fibers impacted at the subgenual white matter target for MDD.

Methods: We analyzed the trajectory of PFC fibers through CB in nonhuman and human primates, based on tracer injections in monkeys and high-resolution diffusion MRI in both species. An unprecedented pre-surgery and follow-up assessment was conducted for 24 patients undergoing cingulotomies for treatment-refractory OCD, MDD and chronic pain. The patients were assessed by using tractography diffusion MRI analysis and neuropsychological evaluation both pre- and post-surgery. PFC pathways through the CB as well as the location of the lesion were compared for each patient. The clinical outcomes will be addressed and linked to the specifics pathways interruptions.

Results: In monkeys, PFC fibers enter ventrally in CB through three main points that are replicable in Human. Dorsal vs. ventral non-cingulate fibers take different routes to leave the CB. Human’s cortical tracts follow similar organization rules. Knowing which circuits are replicable and reliably in Human, we will show which circuits are likely to be interrupted in cingulotomy and involved in clinical changes.

Conclusions: CB is one of the primary route by which non-cingulate fibers can connect to other cortical areas of the brain. Thus, targetting CB lead to neuromodulation of both cortical and cingulate fibers. Clinical changes after cingulotomy will be discussed in relation to neural circuitry and implications for surgical ACC targeting.

Disclosure: Nothing to disclose.

20.2 Structural Connectivity Predictive of Outcomes for Bilateral Anterior Cingulotomy for Chronic Pain Generalises to Obsessive Compulsive Disorder and Major Depressive Disorder

Harith Akram

University College London Queen Square Institute of Neurology, London, United Kingdom

Background: Stereotactic dorsal anterior cingulotomy is a neurosurgical intervention for refractory chronic pain, obsessive compulsive disorder (OCD), and major depressive disorder (MDD). However, the mechanism by which lesioning modulates symptom-specific networks remains unknown

Methods: This multicentre prospective study used lesions from 26 chronic pain patients (52 hemispheres) to generate whole-brain probabilistic tractography with patient-derived diffusion-MRI (dMRI). Connectivity patterns associated with changes in pain scores were identified. The derived computational model was used to predict outcomes in individual chronic pain patients as well as out-of-sample patients with OCD and MDD, respectively (n = 7)

Results: Pain, OCD, and MDD symptoms were significantly reduced post-cingulotomy as measured by NRS (median decrease 50%; IQR 20.0−66.9%; P < 0.0001), Y-BOCS (median decrease 72.6%; IQR 65.6-79.5%), and HDRS (median decrease 21.1%; IQR −3.0 to 33.3%). Comparison of pre- and post-operative Neuropsychological Assessment Battery (NAB) scores identified no significant neuropsychological changes across patients. Structural connectivity seeded from the lesions converged onto a previously identified chronic pain network – involving the medial prefrontal cortex (mPFC), dorsolateral PFC, insula, mediodorsal thalamus, amygdala, putamen, and nucleus accumbens (NAc) – and reliably predicting outcomes in individual chronic pain patients, robust to leave-one-out cross-validation, for morphine equivalent daily dose (MEDD; R = 0.779; P = 0.001) and NRS (R = 0.761; P = 0.002). This same chronic pain-specific connectivity model also robustly predicted Y-BOCS and HDRS outcomes across hold-out OCD and MDD patients (R = 0.395; P = 0.007), respectively. Permutation analyses of the predictive model identified lesion connectivity with regions known to be central to processing aversion and reward as most predictive of outcome (P < 0.05), including the left orbitofrontal cortex, NAc, putamen, and insula

Conclusions: This is the largest stereotactic cingulotomy study and the first study to identify a network predictive of post-neurosurgical chronic pain, OCD, and MDD symptom alleviation. This overlap suggests that similar underlying neural circuits contribute to the effects of cingulotomy across these conditions

Disclosure: Nothing to disclose.

20.3 Deep Brain Stimulation Induces White Matter Remodeling and Functional Changes to Brain-Wide Networks

Peter Rudebeck

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

Background: Deep brain stimulation targeting subcallosal anterior cingulate cortex and adjacent white matter (SCC-DBS) is a promising therapy for treatment resistant depression (TRD). However, little is known about the anatomical and functional mechanisms that underlie this therapy. Thus, our aim was to establish the brain-wide network-level anatomical and functional effects of DBS.

Methods: Modeling the approach used to successfully treat TRD patients, we implanted SCC- DBS electrodes in two rhesus macaques. Specifically, we identified the confluence of the cingulum bundle, forceps minor, and uncinate fasciculus using diffusion tractography imaging (DTI). We then implanted a DBS lead unilaterally in this location, the other hemisphere serving as a control. One month after electrode implantation, stimulation (5mA, 130Hz, 90μsec) began and was maintained for 6 weeks. DTI and whole brain resting-state functional MRIs (rs-fMRIs) were acquired before electrode implantation and following 6 weeks of SCC-DBS stimulation to reveal the anatomical and functional effects of SCC-DBS. Fractional anisotropy (FA) was calculated from DTI data to investigate the macro-level anatomical white matter changes. Additionally, we investigated the micro-level structure changes by histological assessments using immunofluorescence staining of oligodendrocytes and electron microscopy analysis of myelin. Functional data were analyzed using a seed-based comparative-connectome approach where SCC-DBS stimulation induced changes in functional connectivity (FC) were determined.

Results: Compared to before stimulation, we found that six weeks of chronic SCC-DBS enhanced white matter integrity in the midcingulate portion of cingulum bundle selectively. We corroborated this effect at the microscopic level, finding a significant increase in the numbers of oligodendrocytes and thickened myelin in the same region. Additionally, we also found that SCC-DBS significantly changed FC between stimulated SCC and multiple brain networks’ hubs, mainly in the default mode network which is connected through the cingulum bundle.

Conclusions: Our data reveal the specific effects of SCC-DBS showing that it causes myelin remodeling and brain network level functional changes both of which likely contribute to the therapeutic effects of this emerging neuromodulation approach.

Disclosure: Nothing to disclose.

20.4 Abstract Not Included

Panel

21. Development, Mechanism, and Mental Health: Effects of Cannabis on Brain and Behavior

21.1 Preliminary Evidence that Anhedonia is an Important Individual Difference Factor in Neural Reward Response to Δ9-Tetrahydrocannabinol Among Young Adult Cannabis Users

Natania Crane

University of Illinois at Chicago, Chicago, Illinois, United States

Background: Δ9-Tetrahydrocannabinol (Δ9-THC), the main psychoactive component of cannabis, activates brain reward circuitry. This may be especially true among individuals with anhedonia, a focal component of depression specifically linked to deficits in nucleus accumbens (NAcc) reward processing, but no studies have examined this. This presentation will highlight novel data on how Δ9-THC acutely interacts with anhedonia to impact brain reward processing and subjective drug reward.

Methods: In a within-subject, randomized, double-blind, counterbalanced, placebo-controlled design, 26 young adult cannabis users (11 female, 15 male) completed the Monetary Incentive Delay task during fMRI, approximately 120 min after ingestion of placebo or 7.5mg oral Δ9-THC (dronabinol). Participants completed a measure of trait anhedonia (TEPS) at baseline and then completed a subjective drug reward measure (ARCI-MBG) at regular intervals throughout the drug administration visits. Analyses used linear mixed models.

Results: Overall, there were no significant effects of Δ9-THC (vs. placebo) on NAcc reward anticipation or subjective reward response (p-values > 0.05). There were significant interactions between baseline anhedonia and drug on bilateral NAcc reward anticipation activation (p values < 0.05), but not on subjective drug reward (p > 0.05). Follow up analyses found that individuals with greater anhedonia at baseline demonstrated greater bilateral NAcc reward anticipation activation during Δ9-THC (vs. placebo) (right: β = 0.47, p = 0.02; left: β = 0.46, p = 0.02), even after controlling for past month cannabis use and other depression symptoms. Greater bilateral NAcc reward anticipation activation during Δ9-THC was significantly associated with greater subjective euphoric response to Δ9-THC (vs. placebo) (right: β = 0.52, p = 0.007; left: β = 0.58, p = 0.002).

Conclusions: There are significant individual differences in the effects of Δ9-THC on brain reward response and subjective reward response, which are linked, among young adult cannabis users. Anhedonia is an important individual difference factor in Δ9-THC’s effect on brain reward response. It is possible that individuals with anhedonia are more sensitive to Δ9-THC’s effects on reward processing due to pre-existing reward-processing deficits.

Disclosure: Nothing to disclose.

21.2 Recent and Lifetime Effects of Cannabis Use on Domains of Brain Function

Joshua Gowin

University of Colorado Anschutz Medical Campus, Aurora, Colorado, United States

Background: Cannabis use has increased globally but the effects on brain function are not fully known, highlighting the need to better determine the acute and long-term effects of cannabis use on brain activation. We sought to determine if recent and lifetime cannabis use are associated with differences in brain function during cognitive tasks.

Methods: This was a cross-sectional study using the Human Connectome Project dataset. Of the 1206 adults (ages 22−36) assessed, 1005 had data necessary for inclusion in the analysis (i.e., imaging, urine toxicology, and cannabis use). History of heavy cannabis use was assessed using the semi-structured assessment for the genetics of alcoholism (SSAGA), with variables for lifetime history and diagnosis of cannabis dependence. We grouped individuals as heavy lifetime cannabis users if they had > 1000 uses, moderate lifetime users if they had 10-999 uses, and non-users if they had fewer than 10 uses. Participants provided urine samples on the day of the scan to assess recent use. Diagnosis of DSM-IV cannabis dependence was also included. Brain activation was assessed during each of the seven tasks administered during the MRI session (working memory, reward, emotion, language, motor, relational assessment, and theory of mind).

Results: For lifetime history criteria, 88 (8.8%) participants were classified as heavy cannabis users, 179 (17.8%) were classified as moderate users, and 736 (73.4%) were classified as non-users. One-hundred six individuals (10.5%) tested positive for tetrahydrocannabinol, indicating recent use. Both heavy lifetime use (d = −0.28, 95% CI: −0.50, −0.06) and recent use (d = -0.45, 95% CI: −0.65, −0.25) were associated with lower activation on the working memory task. Heavy lifetime use was also associated with lower activation on the theory of mind task.

Conclusions: Heavy lifetime cannabis use was associated with lower brain activation during working memory and theory of mind tasks. Recent cannabis use was associated with lower brain activation during working memory and motor tasks. These findings were not explained by differences in demographic variables, age of first cannabis use, alcohol, or nicotine use. Cannabis use appears to have both short- and long-term effects on brain function, especially during working memory tasks.

Disclosure: Nothing to disclose.

21.3 The Central Role of Cannabis in Polysubstance Use: A Causal Discovery Analysis

Anna Zilverstand

University of Minnesota, Minneapolis, Minnesota, United States

Background: Substance use disorders remain a persistent health and social problem impacting nearly all communities across the country. The nature of the factors driving high level of substance use is highly complex.

Methods: We used data-driven Causal Discovery Analysis (CDA) with data from a community data set of young adults, the Human Connectome Project (N = 1185, age: 22−35, 54% female). The data-driven CDA algorithm (Greedy Fast Causal Inference) generated an integrated causal model relating psychosocial and mental health factors to high substance use levels (alcohol, tobacco, cannabis, hallucinogens, cocaine, opioids, sedatives, stimulants), alcohol and cannabis dependence (DSM-IV). We tested the stability of the model (100 iterations) and evaluated the model fit using structural equation modeling (SEM).

Results: All the edges in the CDA model were significant in the SEM model at p < 0.01, with RMSEA = 0.06 and Tucker-Lewis Index = 0.86 indicating a good fit. Based on the stability analysis (present in > 50% of iterations), we found that externalizing psychiatric symptoms (rule breaking, antisocial behavior, hyperactivity and aggression) were consistently causally linked to high substance use levels, including high recreational cannabis use levels, as well as alcohol and cannabis dependence. In contrast, higher levels of internalizing psychiatric symptoms or negative emotional states were not consistently causally linked to higher substance use levels. Second, use levels of any substance (except tobacco) were causally linked to use levels of other substances. High cannabis use levels played a central role in this polysubstance use, since only high cannabis use levels were causally linked to high use levels for all other illicit substances.

Conclusions: First, these findings support a causal role of mental health factors, specifically of externalizing psychiatric symptoms, in higher levels of substance use, including cannabis use. Second, high levels of cannabis use were causally linked to higher use levels for all other illicit substance use, providing empirical evidence that cannabis may be conceptualized as an “entry drug” into polysubstance use.

Disclosure: Nothing to disclose.

21.4 Abstract Not Included

Mini Panel

22. Harnessing Generative AI for Advancing Mental Health Diagnosis and Care: Industry and Academic Perspectives

22.1 Abstract Not Included

22.2 Capability of Large Language Models to Measure Psychiatric Functioning

Isaac Galatzer-Levy

Google, Brooklyn, New York, United States

Background: Risk prediction based on naturalistic language data holds significant promise to impact the public health burden of common psychiatric illnesses that are overlooked because of limited screening capabilities. Natural language models hold significant promise as psychiatric information is communicated verbally in an unstructured manner. However, natural language models have limited utility because traditional analysis methods require large corpuses of data to train machine learning algorithms that do not generalize to contexts where speech is structured differently than the training data. Large language Models (LLMs) and other forms of generative artificial intelligence(GenAI) are novel machine learning methods that can produce probabilistic predictions without learning from explicit examples. Rather they generate novel specific solutions by learning vast networks of representations on very large data sources and apply attention to these existing representations to infer the most probabilistic output.

Methods: we test the capability of LLMs to assess the most common psychiatric conditions to present in primary care settings (Depressive, Anxiety, Psychotic, trauma and stress, Addictive disorders). Specifically, we test the use of an LLM trained on large corpuses of medical data (MedPalm 2) to assess depression (n = 145) and PTSD (n = 115) from open clinical interviews through zero shot-learning where the model is not provided with any training examples. Next we tested the accuracy to predict the correct psychiatric diagnosis from clinical case descriptions, also using zero short-learning (n = 46).

Results: Results demonstrated high consistency with human raters on standardized assessments (MDD Kappa = 0.55; r = 0.55; p < 0.01; PTSD Kappa = 0.33; r = 0.41; p < 0.01). Further, across clinical case descriptions, the strongest performing LLMs predicted the correct diagnosis significantly > chance (χ² (1,37) = 233.52; p < 0.0001).

Conclusions: Results demonstrate the potential for general LLMs as a generalizable digital measurement tool to flexibly predict psychiatric risk based on free descriptions of functioning from both patients and clinicians.

Disclosure: Google: Employee (Self)

22.3 AI Agents for Robust Empirical Data Analysis and Cross-Referencing With Literature: The LIBR-TU Research Agent Approach

Martin Paulus

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

Background: Identifying robust relationships in empirical data sets is critical for embedding individual findings into existing research and directly relating participant characteristics to those observed in large data and literature. Traditional methods often fail to generalize findings across diverse contexts due to limited screening capabilities and require large corpuses of data. Artificial Intelligence (AI) agents, particularly large language models (LLMs), offer a novel approach to overcoming these limitations by leveraging extensive networks of representations to produce probabilistic predictions without explicit examples.

Methods: We developed an AI agent, referred to as the LIBR-TU Research Agent, to probe the reliability and robustness of empirical findings in individual databases by comparing observed relationships with prior studies. The agent integrates dataset analysis with literature review, utilizing two paths: T500 Agent for analyzing the T500 dataset and PubMed Agent for retrieving supporting literature. The LIBR Research Agent uses tools such as similarity search and regression analysis to find relevant variables and assess relationships. It also generates and summarizes API queries to PubMed to cross-reference findings. The final output combines insights from both paths into a comprehensive report.

Results: The LIBR-TU Research Agent demonstrated high internal consistency (94.6%) across 10 re-runs of the same prompt and substantial external consistency (81.3%) across differently phrased questions. The LLM-as-Judge evaluation, using GPT-4, rated the model’s output with an average score of 3.8 on a scale of 0 to 5, indicating mostly correct answers with minor errors. The AI agent’s ability to retrieve and match data relationships from the T500 dataset with PubMed references showed significant potential for enhancing the accuracy and utility of empirical findings.

Conclusions: The integration of dataset analysis and literature review via AI agents like the LIBR-TU Research Agent demonstrates a promising approach to improving the robustness and reliability of empirical findings. High consistency and positive LLM-as-Judge ratings underscore the potential for these agents to provide accurate, meaningful insights, thereby enhancing the embedding of individual findings into broader research contexts. Future work will focus on refining the system’s sensitivity and adaptability to evolving data formats and research needs.

Disclosure: Nothing to disclose.

Mini Panel

23. Synaptic Density—A Transdiagnostic Marker of Cognitive Impairment Across Different Psychiatric Conditions

23.1 Abstract Not Included

23.2 Synaptic Density and Cognitive Function in Treatment-Seeking Youth With Cannabis Use Disorder: PET Analysis With [18F]SynVesT-1

Isabelle Boileau

Center for Addiction and Mental Health, Toronto, Canada

Background: Studies indicate a correlation between adolescent cannabis use and cognitive decline, potentially linked to its influence on synaptic development. Preclinical findings of diminished synaptic protein markers in THC-exposed rodents support this notion. Furthermore, a recent PET study in adults with cannabis use disorder revealed reduced hippocampal synaptic density, further reinforcing this association. Here, we explore the link between cannabis use disorder, synaptic density and cognitive function in youth seeking mental health treatment using PET radiotracer [18F]SynVesT-1. We hypothesized lower synaptic density in CUD and a link with cognitive deficits.

Methods: Participants of the Toronto Adolescent and Youth (TAY) Cohort Study were recruited to complete a PET scan with the radiotracer [18F]SynVesT-1. MRI scans were conducted to define six regions of interest (ROIs) associated with cannabis use disorder (CUD): prefrontal, anterior cingulate, and temporal cortices, amygdala, hippocampus, and striatum. Time-activity curves from these ROIs were analyzed using a simplified reference tissue model, with the centrum semiovale serving as the reference region (PMOD 4.2). The non-displaceable binding potential (BPnd), indicative of synaptic density, was calculated for each region and compared across groups. Additionally, all participants underwent a comprehensive battery of neurocognitive tests evaluating various cognitive domains.

Results: Forty-two participants (16 males, 26 females, mean age 20.6 years) completed both the PET scan and cognitive testing. The sample had high prevalence rates of mood and anxiety disorders (90%), substance use disorders (50%), and trauma- and stress-related disorders (68%). 33 participants (14 males, 19 females, mean age 20.3 years) had no history of cannabis use disorder (CUD), while nine participants (2 males, 7 females, mean age 21.8 years) met DSM-5 criteria for CUD. Although not significant (p = 0.2), [18F]SynVesT-1 non-displaceable binding potential (BPnd) was substantially lower in CUD group compared to the non-CUD group across all brain regions examined, ranging from -6% (amygdala) to -11% (hippocampus). Correlations with the battery of cognitive tests are currently being analyzed.

Conclusions: This study investigates synaptic density in adolescents with CUD and psychiatric disorders, exploring its correlation with cognitive function. Given the connection between cognitive deficits, mental health outcomes, and cannabis use, this research highlights the importance of further exploration.

Disclosure: Nothing to disclose.

23.3 Preliminary Data Indicate Positron Emission Tomography (PET) Derived [18F]SynVesT-1 Binding as a Potential Marker of Neurocognition in Autistic Youth

Christin Schifani

Center for Addiction and Mental Health, Toronto, Canada

Background: Impairments in functioning and cognition are features of autism but heterogeneity in impairment severity and clinical presentation creates barriers to biomarker discovery. Rising evidence implicates synaptopathology in autism. Initial SV2A PET data indicate lower synaptic density in depression and schizophrenia that relates to cognition. This talk will feature early results on synaptic density in an autistic sample with SV2A PET data. The talk will also present data relating PET SV2A to a widely accessible MRI-derived marker (Neurite Orientation Dispersion and Density Imaging [NODDI]) that may provide a proxy measure of synaptic density in settings/samples where PET is not feasible.

Methods: Autistic youth (16-35 years, IQ ≥ 70) and age-/sex(assigned at birth)-matched controls (CON) were recruited from an existing NIMH-study sample (pool of 117 Autism/43 CON). Cognition in the parent study was assessed with the MATRICS cognitive consensus battery (MCCB); functioning with ABAS-III. All participants complete a 120-min arterial [18F]SynVesT-1 PET scan. Distribution volume (VT) is estimated with the 1-tissue compartment model. Neurite architecture (i.e., neurite density (NDI) and orientation dispersion (ODI)) is derived from MRI NODDI. VT/NDI/ODI is extracted from prefrontal (PFC), cerebellar cortex (CC), fusiform gyrus (FG), striatum, amygdala (strong evidence for gray matter alterations in ASD). Relationships with MCCB were examined in PFC.

Results: The current talk will present data for 17 autistic participants (10M/7F) and 4 CON that completed all procedures successfully (incl. 2h PET). Autistic participants featured a range of cognitive performance scores (MCCB composite: mean ± SD = 48 ± 8). [18F]SynVesT-1 VT was lower across ROIs in ASD vs CON (Cohen’s d = -0.32 (small effect size); PFC:-3.8 ± 0.5%, amygdala:-2.7 ± 0.4%, striatum:−5.3 ± 0.7%, FG:-8.1 ± 1.2%, CC:−3.8 ± 0.4%). PFC VT was positively related to processing speed (r = 0.63, p = 0.028) and reasoning/problem solving (r = 0.74, p = 0.006) in autism. VT and ODI in amygdala were positively related in autism (r = 0.92, p = 0.0013).

Conclusions: Our initial findings indicate that SV2A VT in PFC relates to cognition in autism as in other conditions. Initial associations found between VT and ODI showcase ODI’s potential as a proxy marker of synaptic density.

Disclosure: Nothing to disclose.

Panel

24. Cell Type Specific Epigenetic Regulation of Complex Neurobiological Disease

24.1 Insights Into Cell-Type Specific Susceptibility of Neuronal Populations in Rett Syndrome

Harrison Gabel

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

Background: Neurons in the mammalian brain are enriched for non-CG DNA methylation at CA dinucleotides (mCA), and evidence indicates disruption of gene regulation mediated by mCA underlies a subset of neurodevelopmental disorders (NDD). Excitatory and inhibitory neurons show widely varying global mCA levels, suggesting differential susceptibility to disrupted mCA-mediated regulation in NDD. We therefore investigated the impact of loss of the mCA reader, MeCP2, across major classes of neurons in a Rett syndrome mouse model.

Methods: We assessed gene expression (RNA-seq, n = 4) and enhancer activity (H3K27ac ChIP-seq, n = 4) in excitatory (Layer IV, Layer V) and inhibitory (Pvalb + , SST + ) neurons isolated from cortex of wild-type and MeCP2 knockout (KO) mice. MeCP2-regulated transcripts and enhancers were defined by differential count analysis (RNA: DESeq2, FDR < 0.1; H3K27ac: EdgeR, FDR < 0.1). Functional impact was assessed by gene ontology (FDR < 0.01). Single-cell spatial transcriptomics assessed altered gene expression in distinct brain regions (n = 4). Studies_x0008_ were performed on P42-P56 male and female mice.

Results: We find high mCA cell types display more severe gene dysregulation in MeCP2 KO compared to lower mCA cell types (P < 0.001, 1-way ANOVA). We identify genes recurrently dysregulated across multiple cell types and show that these genes are located in cell-type invariant, megabase-scale genomic regions enriched for mCA. In contrast, cell-type specific demethylation excludes genes from MeCP2 regulation. H3K27ac profiling links altered gene expression to loss of mCA-dependent repression of cell-type specific enhancers. Gene ontology reveals disruption of receptors and ion channels that facilitate the distinct functions of each neuronal type within cortical circuits, and spatial transcriptomic analysis shows this dysregulation occurs in part due to a breakdown of subregion-specific gene programs.

Conclusions: Our results suggest neuron classes with high mCA levels (e.g. SST+ inhibitory neurons) are susceptible to disruption in Rett syndrome. Loss of gene regulation by MeCP2 in these classes leads to altered subtype-specific and spatially-regulated gene programs in multiple brain regions. These findings provide insights into mechanisms of gene regulation by neuronal mCA and suggest key cell types that drive disease.

Disclosure: Nothing to disclose.

24.2 Neural Epigenetic Mechanisms of Early Life Adversity and Exercise Intervention: A Role for KDM7A

Autumn Ivy

Johns Hopkins School of Medicine, Kennedy Krieger Institute, Baltimore, Maryland, United States

Background: Early-life adversity (ELA) can predispose to long-term cognitive impairments, yet effective treatments to buffer against these consequences do not exist. Our previously published data demonstrate that voluntary wheel running specifically during a juvenile developmental stage in mice (Juv EX) results in lasting improvements in hippocampal memory and synaptic plasticity, well after exercise cessation. We have also found that Juv EX intervention can mitigate memory deficits found in middle-aged ELA mice, warranting further investigation into the molecular mechanisms underlying Juv EX intervention in preserving cognitive functions. Employing the Emx1-NuTRAP transgenic mouse line, we performed transcriptomic and epigenomic sequencing in hippocampal neurons and identified the lysine demethylase enzyme KDM7A as a candidate epigenetic regulator of Juv EX impact on improved hippocampal function. Here, we test the hypothesis that Juv Ex mitigates cognitive impairments after ELA by promoting a transcriptionally permissive epigenetic state, and expression of memory-related genes, via the activity of KDM7A.

Methods: Emx1-NuTRAP mice (n = 6−10 mice/sex/group) were used to populate four experimental groups: Control, ELA, Juv EX, and ELA+Juv EX intervention. ELA consisted of limiting bedding and nesting material in the cage during postnatal days (P) 2−9. On P21, Juv EX mice were pair-housed in running wheel-equipped cages and running distances were monitored until P42. In adulthood (10-14 months), mice underwent a battery of hippocampus dependent and independent behavior tasks. Mice were sacrificed at P21, P42, or adulthood, and bilateral dorsal hippocampi were collected for molecular studies. DESeq2 and SEACR (sequencing studies) and Two-way ANOVA with post hoc tests (mass spec, western blot and RT-qPCR data) were used.

Results: KDM7A was significantly increased in hippocampal neurons after Juv EX. There was a higher abundance of H3K27me2 after ELA exposure, whereas Juv EX groups (with or without prior ELA exposure) had greater abundance of H3K27me1. Demethylation of H3K27 can result from Juv EX-induced expression, or change in activity, of KDM7A.

Conclusions: These data implicate an epigenetic mechanism by which Juv EX modulates hippocampal memory functions in adulthood, potentially through KDM7A demethylase activity at H3K27.

Disclosure: Nothing to disclose.

24.3 Cell-Type Specific Epigenetic Regulation of Cartpt Attenuates Cocaine Reward Behavior

Elizabeth Heller

University of Pennsylvania, Philadelphia, Pennsylvania, United States

Background: The nuclear receptor, Nr4a1, and its target gene, Cartpt, are altered in the nucleus accumbens (NAc) in rodent models of cocaine use disorder and in patient postmortem brain. Activation of Nr4a1 represses cocaine reward behavior in mice and activates Cartpt in late abstinence. Cartpt activation in late abstinence is associated with enrichment of Nr4a1 and loss of repressive epigenetic modification, H3K27me3.

Methods: We developed a novel tool, dCas9-FOG1, for cell-type specific epigenetic editing of H3K27me3. Cre-dependent dCas9-FOG1 (or CRISPR, dCas9-VP64) expression plasmid is stereotaxically delivered to NAc of A2a- or Drd1-Cre;LSL-Sun1-GFP mice with either control non-targeting (nt), Nr4a1- or Cartpt-sgRNA. Outcome measures include Nr4a1 and Carpt mRNA and protein expression and H3K27me3 enrichment, cocaine conditioned place preference, extinction, and reexposure. We apply our novel approach, ICuRuS, for MSN-subtype specific mRNA and epigenetic profiling.

Results: 1. A2a+ MSN Nr4a1 CRISPRa activated Nr4a1 expression specifically in Cre+ (A2a + ) MSNs (A2a + ) relative to NT (2-way ANOVA, Effect of sgRNA: Nr4a1 F (1, 10) = 12.05 P = 0.0060).

2. A2a+ MSN Nr4a1 CRISPRa attenuated cocaine conditioned place preference (CPP) (n = 3−6; 2-way ANOVA Effect of Drug: F (1, 7) = 0.005593, P = 0.9425; Effect of sgRNA: F (1, 7) = 21.65, P = 0.0023).

3. Cartpt mRNA expression was elevated at 2 weeks of abstinence following repeated i.p. cocaine exposure (Mixed effects analysis, time: p = 0.0937, treatment: p = 0.5173, time x treatment: p = 0.1647; Sidak’s test, saline = ns, cocaine 4d vs 14d abstinence * p = 0.0261)

4. Cartpt dCas9-FOG1 in NAc repressed Cartpt mRNA expression by 4-days FC = 0.34, Welch’s test, unpaired two-tailed, P < 0.0001

5. Cartpt dCas9-FOG1 in NAc increased Carpt H3K27me3 (Pearson correlation: r(10) = 72, p = 0.008).

Conclusions: Using ICuRuS and CRISPR epigenetic editing, we functionally connect cell-type specific epigenetic regulation of Nr4a1 and Cartpt to cocaine reward and abstinence.

Disclosure: Nothing to disclose.

24.4 Leveraging Single-Cell Epigenomics for the Development of Novel Pain Treatments

William Renthal

Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, United States

Background: Currently available pain treatments such as opioids are often ineffective and associated with unacceptable side effects including respiratory depression and addiction. A major goal for new pain therapeutics is to inhibit the sensory neurons which transmit pain signals (nociceptors) selectively without affecting other neurons involved in innocuous sensation or the central nervous system. However, nociceptor-specific therapeutic approaches remain in their infancy. Viral-based gene therapy offers several attractive advantages in treating refractory pain, as viruses can be engineered to deliver a wide range of molecules, can be administered locally or systemically. While nociceptor-specific viral tools do not presently exist, recent advances in single-cell epigenomics have enabled our group to begin developing these tools.

Methods: Single-nucleus RNA and assay for transposase accessible chromatin sequencing (snATAC-seq) were employed to identify nociceptor-specific molecular features. Specifically, dorsal root ganglion (DRG) neurons from male and female mice were processed using snATAC-seq and analyzed. Differential chromatin accessibility analysis was performed and nociceptor-specific chromatin regions were identified and cloned into an AAVs. AAVs were delivered intracerebroventricularly to mice and DRGs were dissected 3 weeks later. In situ hybridization of GFP and nociceptor markers was used to determine which genomic regions drive reporter expression selectively in nociceptors.

Results: Our single-nucleus ATAC-seq analyses identified > 10,000 nociceptor-specific peaks of chromatin accessibility. Motif analysis of these data identified several transcription factors that are significantly enriched in DRG nociceptors including ISL1, ISL2, RUNX1, and GBX1 (Log2FC > 0.5, FDR < 0.05). After prioritization and screening of putative nociceptor-specific enhancers, we identified three that exhibit preferential expression in DRG nociceptors compared to other neuronal subtypes (our lead AAV drives the expression of a reporter, > 80% are nociceptors).

Conclusions: By leveraging single-cell epigenomics, we have identified gene regulatory regions that drive nociceptor-specific expression in the setting of an AAV in vivo. We are optimistic that these studies will lead to targeted gene therapy for refractory pain.

Disclosure: Teva: Grant (Self)

Study Group

25. The Covid Counterfactual: Measuring Neurodevelopmental Trajectories That Cross Through the Pandemic Era

Joshua Roffman, Deanna Barch, Moriah Thomason, Ian Gotlib, Ran Barzilay, Antonia Kaczkurkin, Henning Tiemeier

Massachusetts General Hospital, Charlestown, Massachusetts, United States

Study Group Summary: Prospective observational studies of children offer rich opportunities to define normative and altered trajectories of neurodevelopment, their underlying psychosocial and biological mediators, and their linkage to clinical outcomes. Ongoing multisite cohort studies that focus on developmental stages from perinatal through late adolescence—such as ECHO, ABCD, and HCP-Development—have promised to help define these trajectories in unprecedented detail, by virtue of large and inclusive samples, harmonized protocols, cutting-edge methods, repeated measures, and deep phenotyping. Equally unprecedented, though, was the degree to which the Covid-19 pandemic – which cut through each of these (and numerous other) studies mid-stream—disrupted study protocols, affected the psychosocial milieu of participants and their families, and potentially altered trajectories of brain development and psychopathology risk on a population level. This Study Group will focus on the challenges and opportunities posed by the pandemic for collecting, analyzing, and interpreting data from longitudinal neurodevelopmental studies in the Covid-19 era.

Discussion will focus on three related themes. First, we will consider practical issues that are critical to incorporate as part of analytic strategies, regardless of whether they center on the pandemic (for example: How did the pandemic influence attrition? How might site-related differences in school- or other state-level policy have affected outcomes?). Second, we will explore whether and how pandemic-related effects on neurodevelopmental trajectories may be isolated from those that the studies were designed to measure (e.g., those related to normative development or to other risk-modifying exposures). Third, we will consider the specific impact of the pandemic on study populations that have been historically underrepresented in developmental cohort studies, and who in many cases were exposed disproportionately to psychosocial adversity during the pandemic. A cross-cutting question for these themes is whether we should see the pandemic as a period effect, a possible effect modifier, or a confounding problem. Or, alternatively, do pandemic-related shifts in environmental factors salient to brain development (e.g., socioeconomic status) provide an “unfortunate experiment of nature” from which we might glean unanticipated insights?

Panel participants are experts in multi-site neurodevelopmental cohort studies ranging from perinatal to late adolescence; whose work cuts across brain imaging, developmental psychopathology, exposome, and -omics measures; and who represent numerous academic institutions and early, mid, and senior career levels. To facilitate discussion, each participant will be asked to describe an example of how the pandemic may have influenced a neurodevelopmental measurement of particular interest to them, and the strategies they considered to isolate or account for pandemic effects. The overarching goal is to bring together emergent themes, practical solutions, and novel scientific ideas that may be useful for investigators whose long-term observational studies crossed through the pandemic era.

Disclosure: Nothing to disclose.

Study Group

26. Translational Research and Collaborative Efforts to Radically Improve Patient Outcomes in Bipolar Disorder

Katherine Burdick, Mark Frye, Hilary Blumberg, Alexandra Vinson, Andrew Smith, Cara Altimus

Brigham and Women’s Hospital/Harvard Medical School, Boston, Massachusetts, United States

Study Group Summary: Historically, bipolar disorder (BD) is under-funded relative to illnesses like schizophrenia and unipolar depression, despite the fact that it consistently ranks among the top ten causes of disability worldwide. Our understanding of the mechanisms that drive disease is limited and clinical care is suboptimal for most patients. Because BD is a dynamic, highly heterogeneous, and exceptionally complex disease—progress toward precision care will require interdisciplinary collaboration at both depth and scale. On average, it takes ~15 years to translate basic scientific discoveries into clinically actionable insights in psychiatry but there are emerging initiatives that are designed to accelerate this translational process.

Among these, Breakthrough Discoveries for thriving with Bipolar Disorder (BD-squared) is a philanthropically-seeded non-profit foundation dedicated to building a translational platform around a common goal – to improve outcomes and positively impact the lives of people with BD. Basic scientists, clinical researchers, clinicians, and people with lived experience have contributed to the build of an ecosystem that we believe will allow us to take innovative research findings into the clinic in near real time. An entirely ‘open science’ model will make this resource widely available to rise to a global-level impact. Although this is not the only collaborative initiative by any means, we will use this as a template for considering several key issues to be discussed in this session.

Disclosure: Breakthrough Discoveries for thriving with Bipolar Disorder: Board Member (Self). Merck: Advisory Board (Self).

Panel

27. Barrier Crossings: An Understudied Target in Neurological and Neuropsychiatric Disorders

27.1 Abstract Not Included

27.2 Regulation of Claudin-5 at the Blood Brain Barrier

Matthew Campbell

Trinity College Dublin, Dublin, Ireland

Background: It is now becoming apparent that disruption to the cerebral microvasculature may be a key driving force of numerous neurological and neuropsychiatric disorders. Additionally, targeted regulation of the blood brain barrier (BBB) might hold immense therapeutic potential. Claudin-5 is the most abundant tight junction (TJ) protein at the BBB. Disrupted levels of claudin-5 have been reported previously in post-mortem and surgically resected tissue from patients with major depressive disorder, schizophrenia and epilepsy, however there is a paucity of studies investigating the therapeutic potential of restoring BBB properties in these disorders.

Methods: Using Cldn5 heterozygous mice (n = 10) compared to WT mice (n = 10), we undertook a series of behavior tests to determine a range of behavioral traits in mice. Additionally, we also developed an adeno-associated virus (AAV) to over-express claudin-5 stabilize the BBB.

Results: Here, we show that endothelial-specific deletion of one copy of Cldn5 leads to cognitive impairment, reduced social interaction and sensorimotor gating deficits. Importantly, these impairments are reversible following restoration of claudin-5 via endothelial-specific gene therapy. Endothelial-specific overexpression of claudin-5 prevented seizures in inducible claudin-5 knockdown mice as well as in a kainic acid model of acute seizures.

Conclusions: We propose a threshold level of claudin-5 expression required to ensure BBB integrity. Once this threshold is breached, neuroanatomical-specific pathology will ensue. Gene therapy-mediated restoration of BBB function represents a new generation of drugs designed to treat multiple neurological disorders.

Disclosure: Nothing to disclose.

27.3 Pericytes Modulate Vascular Integrity in Social Stress and Depression

Caroline Menard

Université Laval, Quebec City, Canada

Background: Dissimilarities in prevalence, symptoms, and treatment response point at major sex differences in major depressive disorder (MDD). We reported that blood-brain barrier (BBB)-related changes underlie stress responses and resilience in mice and depression in human tissue. Alterations are sex-specific with loss of BBB integrity observed in the male nucleus accumbens (NAc) and female prefrontal cortex (PFC), two brain regions involved in emotion regulation. Pericytes ensheath endothelial cells and promote neurovascular functions. Endothelial cell: pericyte crosstalk is critical for BBB integrity and mainly occurs through the PDGFB:PDGFRB pathway. However, the functional role of pericytes in stress responses and mood disorders, as well as underlying sex differences remain unknown.

Methods: Pericyte gene expression and associated morphology were assessed in qPCR, immunofluorescence and electron microscopy following chronic social defeat stress, a mouse model of depression in male and female NAc and PFC. Morphology of pericytes was quantified in post-mortem tissue from individuals with MDD. In parallel, functional manipulations of pericytes were conducted to confirm a causal role of altered pericyte function in the development of anxiety- and depression-like behaviors at baseline and after stress exposure. Finally, we use ELISA to unravel potential vascular markers of MDD, in a sex-specific manner.

Results: After social stress exposure, pericyte-related genes are decreased only in the PFC of stress-susceptible female mice, along with decreased pericyte coverage of capillaries. This was confirmed in brains of depressed women. Partial ablation of PFC pericytes in female mice promoted anxiety- and depressive-like behaviors, despite no exposure to stress. Conversely, promotion of pericyte-endothelial crosstalk through viral-mediated increase in brain endothelial Pdgfb evokes anxiolytic and pro-social behaviors, in the presence or absence of prior stress exposure. Serum levels of PDGF-BB positively correlates with several symptoms of depression in women but not men with mood disorders.

Conclusions: Our results provide the first characterization and functional interrogation, in a sex-, region-, and cell-specific manner, of the role played by pericytes in stress responses, resilience and mental health.

Disclosure: Nothing to disclose.

27.4 Pervasive Neurovascular Dysfunction in the Ventromedial Prefrontal Cortex of Females With a History of Early Life Adversity

Naguib Mechawar

McGill University, Verdun, Canada

Background: Early life adversity (ELA) is a significant global public health concern, with profound pathophysiological implications for affected individuals. Studies suggest that ELA contributes to endothelial dysfunction, bringing into question the functional integrity of the neurovascular unit in brain regions vulnerable to chronic stress. Despite the importance of the neurovasculature in maintaining normal brain physiology, human neurovascular cells remain poorly characterized, particularly with regard to mediating the effects of ELA.

Methods: We recently developed a protocol to enrich and isolate microvessels from archived snap-frozen human cerebral cortex using mechanical homogenization and centrifugation-separation that preserves the structural integrity and multicellular composition of microvessel fragments. Using this approach, microvessels were isolated from postmortem ventromedial prefrontal cortex samples from healthy controls (13M/8F; mean age: 49.1 ± 18.7 years [± s.d.]) and depressed suicides with histories of ELA (17M/7F; mean age: 42.5 ± 14.7 years [± s.d.]) and comprehensively investigated as structural units using RNA sequencing.

Results: Our findings point to substantive differences between males and females. While microvessels in males appeared to be much less affected, females exhibited widespread gene expression changes at the neurovascular unit, including downregulation of the key tight-junction CLDN5 (p = 0.012), as well as vascular nodal regulators KLF2 (p = 8.28 × 10 − 4) and KLF4 (p = 2.99 × 10 − 4), which regulate numerous vascular transcription programs. Samples from females with ELA also displayed a broad downregulation of immune-related pathways at the neurovascular unit that was reflected by a similar downregulation within the brain parenchyma.

Conclusions: These results suggest that the neurovascular unit plays a larger role in the neurobiological consequences of ELA in females, implicating pathways pertaining to basic vascular identity and function as well as immune regulation.

Disclosure: Nothing to disclose.

Panel

28. The Sound of Silence: Emerging Data of Transcranial Focused Ultrasound for Neuromodulation

28.1 Focused Ultrasound Amygdala Neuromodulation: A Brain Imaging-Coupled Open-Label Treatment Trial in Mood and Anxiety Disorders

Gregory Fonzo

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

Background: Mood and anxiety disorders are characterized by exaggerated emotional reactivity, thought to be mediated by hyperactivity of the amygdala to emotional cues. Given the relevance of amygdala-mediated neurobehavioral processes to mood and anxiety disorder development and expression, the amygdala remains a promising therapeutic target. However, existing neuromodulatory approaches are limited in capacity to modulate the amygdala due to attenuation of electromagnetic energy at depth. Low-intensity transcranial focused ultrasound (tFUS) has recently been demonstrated to non-invasively modulate brain function in subcortical areas, including the amygdala. Here, we tested repetitive daily administration of tFUS (rtFUS) to the left amygdala for therapeutic potential in individuals with mood and anxiety disorders.

Methods: Individuals with mood and anxiety disorders (N = 29) underwent structural MRI, MRI-guided targeting of tFUS to the left amygdala, and task-based fMRI with an emotional face matching task (anger, fear, happy, neutral, and a shape-processing control condition). Participants then underwent open-label (OL) once-daily left amygdala rtFUS treatment (using neuronavigation and a 10-min protocol previously found to inhibit the amygdala) 5 days a week for 3 weeks to assess rtFUS therapeutic potential. Participants completed a post-treatment task-based fMRI assessment to measure pre- to post-treatment changes.

Results: There were significant symptom reductions on the primary outcome (Mood and Anxiety Symptom Questionnaire General Distress sub-scale; F = 12.98, p < 0.001, d = 0.77) and most secondary outcomes (PTSD Checklist for DSM-5, Beck Depression Inventory, Beck Anxiety Inventory, and others; all p < 0.05). Imaging analyses revealed a significant attenuation (corrected p < 0.05) of amygdala activation across all task conditions in the left and right amygdala, with an emotion-specific attenuation of activation to angry faces in the left basolateral amygdala. Greater attenuation of right amygdala activation was associated with greater reductions on the primary outcome.

Conclusions: Though tempered by the OL design, these findings provide compelling initial evidence that rtFUS to the left amygdala may be a safe and effective intervention method for mood and anxiety disorders with persisting effects on brain function.

Disclosure: SynapseBio AI: Consultant (Self)., Alto Neuroscience: Consultant (Self). Alto Neuroscience: Stock / Equity - Publicly Traded Company (Self)

28.2 Transcranial Low-Intensity Focused Ultrasound of the Ventral Capsule/Ventral Striatum for Treatment-Resistant Obsessive-Compulsive Disorder

Darin Dougherty

Massachusetts General Hospital/Harvard University, Charlestown, Massachusetts, United States

Background: Obsessive-compulsive disorder (OCD) is characterized by unpleasant intrusive thoughts, images, and/or (mental) rituals. Deep Brain Stimulation (DBS) of the ventral capsule/ventral striatum (VC/VS) can reduce OCD symptom severity and modulates cortico-striato-thalamo-cortical (CSTC) network activity. However, DBS has neurosurgical risks; transcranial focused ultrasound (tFUS) is a newer form of noninvasive stimulation with the potential to focally modulate deep brain regions previously inaccessible by current noninvasive methods.

Methods: Across two study sites, we first explored the safety, feasibility, and tFUS parameter space across 48 sessions of tFUS in n = 12 healthy individuals (6 women and 6 men). In an ongoing open-label study, we have conducted 30 tFUS sessions in n = 4 individuals with treatment-resistant OCD (3 women and 1 man). Participants underwent blood oxygen level-dependent (BOLD) resting-state functional magnetic resonance imaging (fMRI) before and after 6 tFUS sessions (pulse repetition frequency = 125 Hz, pulse-width = 0.2 ms, duty cycle = 2.5%) targeting the left VC/VS.

Results: There were no serious adverse events in the healthy individuals; one individual with OCD could not tolerate buzzing sensations associated with tFUS. Three out of the four individuals with treatment-resistant OCD were classified as responders (>35% reduction on the Yale-Brown Obsessive Compulsive Scale), although the average decrease in Y-BOCS scores was at a trend-level of statistical significance (p = 0.107, Cohen’s d = 0.91). After VC/VS tFUS, individuals with OCD showed increased resting-state functional connectivity between the VC/VS and caudate (Cohen’s d = 1.72, p = 0.043), as well as decreased VC/VS resting-state functional connectivity with the globus pallidus (Cohen’s d = 1.11, p = 0.023) and thalamus (Cohen’s d = 3.09, p = 0.033). There were also trend-level decreases in VC/VS functional connectivity with the putamen (Cohen’s d = 1.402, p = 0.086) and anterior cingulate cortex (Cohen’s d = 1.066, p = 0.097).

Conclusions: VC/VS tFUS can safely modulate the target VC/VS and regions within the larger CSTC network in healthy individuals and individuals with treatment-resistant OCD. Future studies should further explore whether VC/VS tFUS can significantly decrease OCD symptoms in this treatment-resistant population.

Disclosure: Boehringer-Ingelheim: Advisory Board (Self). Celanase: Advisory Board (Self). Livanova: Advisory Board (Self). Omniscent: Advisory Board (Self). Inner Cosmos: Stock / Equity - Privately Held Company (Self). Neurable: Stock / Equity - Privately Held Company (Self). Intrinsic Powers: Stock / Equity - Privately Held Company (Self)

28.3 Ultrasonic Neuromodulation of Subcallosal Cingulate Elicits Rapid Antidepressant Response

Brian Mickey

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

Background: Severe depression has been linked to excessive subcallosal cingulate (SCC) activity. Disruption of SCC activity with surgically implanted electrodes can relieve depressive symptoms, but current noninvasive techniques cannot directly and selectively modulate deep targets. We have developed a new noninvasive approach that delivers focused ultrasound to deep circuits. The objectives of this study were to demonstrate that this approach can engage the SCC and to characterize the immediate short-term effects of SCC neuromodulation on mood and depressive symptoms.

Methods: Twenty-two subjects (64% female) with treatment-resistant depression participated in a randomized, double-blind, sham-controlled study. Low-intensity focused ultrasound was delivered transcranially using two transducer arrays that measured and compensated for aberration of ultrasound by the head. To quantify target engagement, ultrasonic stimulation was delivered to bilateral SCC during concurrent “online” blood oxygenation level dependent (BOLD) imaging (n = 16). Mood state was measured with the Sadness subscale of the expanded Positive and Negative Affect Schedule immediately before and after 40 min of real or sham SCC stimulation (n = 18). Change in depressive symptoms was measured with the 6-item Hamilton Depression Rating Scale (HDRS-6) at 24 h and 7 days (n = 19).

Results: Ultrasonic stimulation caused an SCC-specific decrease in BOLD activity (d = −0.61, p = 0.028). SCC neuromodulation was detectable at the individual-subject level with a single 10-minute scan in 8 of 16 participants. Sadness improved more with real stimulation than with sham stimulation (d = −1.15, p = 0.027). Real stimulation was superior to sham for HDRS-6 at 24 h (d = −1.08, p = 0.031) but not statistically significant at 7 days (d = −0.59, p = 0.22). Stimulation was well tolerated. No serious adverse events occurred.

Conclusions: Noninvasive ultrasonic stimulation can modulate SCC activity, immediately improve mood, and rapidly reduce depressive symptoms. Clinically meaningful antidepressant effects may be possible with repeated treatment and optimization of targeting. The capability to noninvasively and selectively target deep brain areas creates new possibilities for future development of circuit-based diagnostics and therapeutics.

Disclosure: LivaNova: Contracted Research (Self). Health Rhythms: Grant (Self)

28.4 Low-Intensity Focused Ultrasound of Human White Matter Tracts: An Opportunity to Advance Precision Neuromodulation Targeting in Major Depression

Salvador Guinjoan

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

Background: Treatment resistance in depression is ubiquitous and there is an utter need to develop precision neuromodulation methods that safely, reversibly, and noninvasively modify the neural activity of anatomically discrete circuits presumably involved in the formation of symptoms in this disorder. Burgeoning in vitro evidence shows that low-intensity focused ultrasound (LIFU) operates mechanosensitive K+ channels in nodes of Ranvier, thus impairing saltatory conduction in myelinated axons. We translated these findings to the neuromodulation of white matter tracts in humans.

Methods: In a preregistered double-blind, randomized, sham-controlled trial of LIFU of tracts in the right anterior limb of the internal capsule (NCT05697172; FDA non-significant risk determination), we targeted with LIFU the coordinate with maximum number of tractography streamlines connecting thalamus and prefrontal limbic cortices (80-s stimulus, 10% duty cycle, 500 kHz, estimated ISPPA 2.26 W/cm2) with and without sound absorbing material applied to the stimulation transducer, and measured changes in resting state fMRI connectivity, state affect (PANAS-X) and rumination (BSRI), and heart rate variability in patients with Major Depressive Disorder (n = 21; 17 female; mean ± SD age: 34 ± 11 years).

Results: Seed-to-voxel whole-brain thalamic connectivity was significantly decreased after active LIFU in the ventromedial prefrontal and posterior cingulate cortices, compared with sham LIFU. Rumination changes were indistinguishable between post-active and post-sham LIFU. PANAS-X Joviality scale was greater after active LIFU (d = 0.268) vs. sham LIFU (d = -0.137). In a preliminary analysis of the initial set of patients (n = 10) we observed an increase in sympathovagal balance onto the heart after sham stimulation, which was not present after active LIFU (F[1,27] = 4.25, p < 0.05, n = 10).

Conclusions: LIFU administered to deep white matter in humans (in a region overlapping with a historical surgical target for depression) produces functional disconnection of gray matter hubs linked by the sonicated axonal tracts. We also obtained direct evidence of acute emotional changes, including a PANAS-X item possibly analogous to mirth, and top-down autonomic effects of LIFU, probably representing a blunted autonomic response to the stress of the neuromodulation session.

Disclosure: Nothing to disclose.

Panel

29. 1 + 1 = 3: Understanding the Neurobiology of Polysubstance Use

29.1 The Effects of Prenatal Co-Exposure on the Placental-Fetal Brain Axis and its Cognitive Implications

Valeria Lallai

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

Background: The increasing prevalence of concurrent nicotine and THC use among pregnant women is influenced by the widespread availability of e-cigarettes marketed as a ‘safer’ alternative and the growing accessibility of THC-infused edibles. Nicotine and cannabinoids affect dopaminergic signaling through distinct receptors within overlapping cellular groups, suggesting potential diverse consequences from their combined consumption. This complex interplay during prenatal development raises questions about their cumulative impact on embryonic growth, neurodevelopment, and overall well-being.

Methods: To mirror human usage patterns, female Wistar rats underwent pre-exposure before mating and were then administered daily doses of nicotine vape and oral THC throughout pregnancy. Maternal blood samples were collected to quantify nicotine and cannabinoid metabolites, ensuring accurate assessment of drug exposure. Offspring from the first cohort underwent cognitive behavioral tests during adolescence and drug intake assessments during adulthood. For the second cohort, placental tissue and fetal brains were collected on gestational day (GD) 18 for RT-q-PCR analysis of mRNA expression levels of nicotinic subunits, cannabinoid receptors (nAChRs and CBR), dopamine receptors and transporters (D3 and DAT), and sex-specific markers.

Results: Our findings demonstrate differential effects of prenatal exposure to e-cigarette nicotine vape and/or edible THC on cognitive function, with varying impacts within male and female groups. Our umpublished data confirmed the presence of cotinine, nicotine’s principal metabolite, and THC in maternal blood, placenta, and fetal brain. Additionally, our results reveal a specific distribution of nAChRs across various placental regions, with fetal sex-dependent influences on these expression patterns following drug exposure. RNA scope analysis further revealed a distinct distribution of CBR, primarily localized within the junctional zone and decidua, with notable absence in the labyrinth zone.

Conclusions: This research unveils the intricate relationship between maternal e-cigarette usage, drug exposure, and developmental outcomes in rats, offering potential insights to enhance public health strategies and interventions for the benefit of future generations.

Disclosure: Nothing to disclose.

29.2 Nicotinic Cholinergic Mechanisms in Models of Alcohol Use Disorder: Potential Insights Into Co-Use

Anna Lee

University of Minnesota, Minneapolis, Minnesota, United States

Background: Nicotine and alcohol co-use is highly prevalent and nicotine and alcohol use disorders share common molecular mechanisms. One common molecular mechanism is the involvement of nicotinic acetylcholine receptors (nAChRs), which are the target of nicotine and modulate the effects of alcohol. Here we present new data on the involvement of cholinergic neurons, expression and upregulation of nAChRs after alcohol administration and withdrawal in pre-clinical mouse models.

Methods: We used two passive alcohol administration procedures in adult male and female C57BL/6J mice, saline, 2, or 4 g/kg i.p. alcohol for 1, 5, or 15 days (3-5 mice per group per sex), and 2.5 g/kg i.p. or saline, both paired with alcohol dehydrogenase inhibitor 4-methylpyrazole (9 mg/kg ip) for 9 days (9-12 mice per group per sex). We measured neuronal activation with cFos immunohistochemistry or Fos in situ hybridization in the ventral tegmental area (VTA) and the mesopontine tegmentum (MPT). Behavioral assessments of alcohol withdrawal included somatic withdrawal signs, elevated zero maze, marble burying test, sucrose consumption test and social interaction test.

Results: We observed a sex difference in MPT cholinergic neuron activation (FsexXtreatment(2, 18.9) = 1.275, p = 0.30), with an increase in activated cholinergic neurons in the MPT in males but not in female mice. In males, there was an increase in alpha6, alpha4 and beta2 nAChR subunit transcript expression in cholinergic neurons of the brainstem after chronic alcohol administration measured by fluorescent in situ hybridization. For the 9-day procedure, male mice showed increased withdrawal signs at 7h, 24h and 7days, with different signs occurring at different times (FtimeXtreatment(2, 35) = 6.521, p = 0.004). Male mice had increased anxiety-like behavior at 24h but not 7d. Female alcohol-treated mice showed no differences in withdrawal signs or anxiety-like behavior at 24h compared. Instead, there was a significant 65% deficit in social interaction behavior in alcohol-treated female mice at 24h withdrawal (t = 3.053, p = 0.01).

Conclusions: Our data indicate a sex difference in the behavioral manifestation and in the cholinergic mechanisms of alcohol withdrawal. The development of cholinergic pharmacotherapies for alcohol withdrawal and co-use of alcohol and nicotine may need to be different between sex.

Disclosure: Nothing to disclose.

29.3 Exploring the Interplay of Nicotine and Opioids: Insights From Mouse Behavioral Assays

Brandon Henderson

Marshall University School of Medicine, Huntington, West Virginia, United States

Background: Tobacco dependence is correlated in individuals that are opioid use dependent (OUD) as 95% of OUD individuals are heavy smokers. Despite the high rate of co-use among opioids and tobacco, little is known about either drug altering the reinforcement-related properties of the other. Therefore, we investigated how opioid exposure altered nicotine reinforcement-related behaviors in a mouse model. Our primary hypothesis was that nicotine and opioid co-exposure produced unique changes in nicotinic receptors and this would contribute to subsequent changes in nicotine reinforcement.

Methods: We trained male and female adult mice in a e-Vape self-administration assay. Mice responded on a fixed-ratio 1 (FR1), FR2, and FR3 schedule. Mice were also given saline or 10 mg/kg morphine (ip) on an alternating daily schedule. Self-administration occurred in two different paradigms: (1) morphine/saline administration continued for the entire 15 session paradigm or (2) morphine/saline injections were given for 10 sessions and then self-administration continued for 15 sessions during morphine abstinence. Next, brains were extracted for the use of confocal microscopy or patch-clamp electrophysiology.

Results: In paradigm 1, morphine-injected mice exhibited a significant decrease in nicotine self-administration. In paradigm 2, morphine-assigned mice exhibited decreased nicotine self-administration. However, during abstinence from morphine, morphine-injected mice exhibited a significant increase in nicotine self-administration. Confocal imaging revealed that morphine-injected mice exhibited an attenuation of nicotine-induced upregulation of nicotinic receptors. Electrophysiology and supplemental microscopy assays revealed that morphine exposure stabilizes low-sensitivity nicotinic receptors.

Conclusions: Morphine-induced stabilization of low sensitivity nicotinic receptors reduces reward associated with nicotine. However, during abstinence the increase in these low sensitivity receptors may drive greater nicotine intake due to the higher concentration that is needed to fully activate low sensitivity nicotinic receptors. These data suggest that high smoking rats in OUD populations may be due to changes in receptor assembly that requires higher nicotine loads to reach satisfaction.

Disclosure: Nothing to disclose.

29.4 Assessing PSU in Humans for Backtranslational Models in Rodents

Linda Cottler

University of Florida, Gainesville, Florida, United States

Background: Polysubstance use remains a substantial public health concern. One reason could be that studies do not include concurrent, simultaneous or sequential use. Studies on common patterns of substance use among humans are needed to backtranslate to rodent models to help us understand neuro and cognitive effects.

Methods: in a previous R21, 148 community members who used cocaine in the past 30 days were interviewed with the Polysubstance Use Temporal Patterns Survey to assess use of cocaine, alcohol and marijuana on a typical day, hour by hour, along with dosage and route of administration. The Substance Abuse Module was used to assess DSM criteria associated with each of 13 substance classes. For a new R61, 300 individuals from the community will be assessed for the patterns and consequences of opioid and alcohol use. These patterns will be backtranslated to rodents to evaluate effects on cognition and neurodevelopment.

Results: In the R21, over 50% met DSM-5 criteria for cocaine or alcohol use disorder; 47% met criteria for cannabis use disorder. Hierarchical clustering of common sequences found a five-cluster solution: limited cocaine/cocaine+alcoho; extensive marijuana then all three; limited alcohol+cannabis then cocaine+alcohol; extensive cocaine+ marijuana use then all three; extensive cocaine then cocaine+alcohol. There were no differences by PSU pattern for reported number of days of cocaine use, amount of cocaine per day, or cocaine use disorder. There was an effect on number of rocks of cocaine used.

Based on this, we are conducting an R61 to assess the effects of opioids and alcohol. Focus groups have been completed to date on 20 participants. We are now beginning interviews with 300 people who use opioids to evaluate the differential effect of opioids and alcohol on cognition.

Conclusions: In the R21, we identified five distinct patterns of cocaine PSU, though they were not accompanied by differential timing, order or duration of use. Most people who used cocaine reported use for periods shorter than 6 h, and for 1−2 days/week on average. These models can inform animal models of substance use that reflect real-world patterns of human behavior. Data collection on human patterns of opioid and alcohol use have begun. We will focus on methods to assess PSU patterns in humans and how these patterns can inform rodent modeling.

Disclosure: Nothing to disclose.

Mini Panel

30. Emerging Novel Targets for Schizophrenia

30.1 Abstract Not Included

30.2 Synaptic Density in Early Psychosis and Clinical High Risk and its Association With Negative Symptoms

Romina Mizrahi

McGill University, Montreal, Quebec, Canada, Montreal, Canada

Background: Synaptic dysfunction is involved in schizophrenia pathophysiology. However, whether in vivo synaptic density is reduced in early stages of psychosis, including its high-risk states remains unclear.

Objectives:

First, to investigate whether synaptic density (SV2A BPND) is reduced in first-episode psychosis (FEP) and in clinical high-risk (CHR). Second, to investigate the effect of cannabis use on synaptic density and examine its relationship with psychotic symptoms and gray matter microstructure across groups.

Methods: This is a cross-sectional study which was performed in a tertiary care psychiatric hospital from July 2021 to October 2023. Antipsychotic-free/minimally exposed FEP, CHR and healthy controls with clean urine drug screen (except cannabis).

Main Outcomes and Measures:

Synaptic density (SV2A BPND) was quantified with dynamic 90 min [18F]SynVesT-1 PET scans across prioritized brain regions of interest (ROIs) delineated in individual Magnetic Resonance images (MRI). Cannabis use was confirmed with urine drug screens. Gray matter microstructure was assessed using diffusion-weighted MRI to estimate neurite density.

Results: A total of 49 participants, including 16 FEP patients (mean[SD] age, 26.1[4.6] years; 7 females, 17 CHR (mean[SD] age, 21.2[3.5] years; 9 females)), and 16 healthy controls (mean[SD] age, 23.4[3.6] years; 9 females) were studied. Synaptic density was significantly different between groups (F(2,273) = 4.02, p = 0.02, Cohen’s F = 0.17; ROI: F(5,273) = 360.18, p < 0.01, Cohen’s F = 2.55) with a group by ROI interaction (F(10,273) = 2.67, p < 0.01 Cohen’s F = 0.32). Synaptic density was lower in cannabis users (F(1,272) = 5.31, p = 0.02 Cohen’s F = 0.14). Lower synaptic density (SV2A BPND) across groups was associated with more negative symptoms (PANSS Negative Scores: F(1,81) = 4.31, p = 0.04, Cohen’s F = 0.23; SOPS negative scores: F(1,90) = 4.12, p = 0.04, Cohen’s F = 0.21). SV2A BPND was significantly associated to neurite density index (NDI) (F(1,138) = 6.76, p = 0.01, Cohen’s F = 0.22).

Conclusions: Synaptic density reductions are present during early stages of psychosis and its risk states and associated with negative symptoms. These findings position SV2A as a potential molecular target for the treatment of negative symptoms. Future studies should investigate the impact of cannabis use on synaptic density in CHR longitudinally.

Disclosure: Nothing to disclose.

30.3 Marker of Glial Cell Activation and Modulation by a Mononuclear Antibody: In Vivo Evidence in Schizophrenia

Oliver Howes

King’s College London, Institute of Psychiatry, London, United Kingdom

Background: Multiple lines of evidence implicate microglia and related immune cells in the pathophysiology of schizophrenia. Activated microglia express increased levels of the translocator protein (TSPO), which can be measured in vivo using PET imaging. However, studies to date in first episode schizophrenia (FES) have been relatively small and findings inconsistent. In view of this, we aimed to address this in a large first episode sample. Natalizumab, a monoclonal antibody, has been shown to suppress activation of microglia and to reduce PET TSPO binding in people with multiple sclerosis but has never been tested in FES. Thus, we used natalizumab to test if natalizumab can suppress microglial activation in FES and if this is related to symptom improvements.

Methods: [18F]DPA-714 PET and MRI imaging scans were acquired in103 volunteers (62 patients with first episode schizophrenia (FES) and 41 controls) for case-control comparisons. Patients entered a longitudinal trial of natalizumab over 3 months administered by IV infusion. All patients received clinical measures at baseline and follow-up and 31 patients received baseline and follow-up PET-MR scans following natalizumab treatment.

Results: [18F]DPA-714 PET relative uptake was higher in the total gray matter (GM p < 0.05, η2 = 0.043) and temporal lobe GM (p = 0.05, η2 = 0.057) of patients with FES relative to controls.

Natalizumab treatment was associated with a significant decrease in the primary clinical outcome of PANSS total score from baseline to follow-up (p < 0.05, Cohen’s d = 0.40). An exploratory analysis of subscales, showed a significant decrease in positive (p = 0.002, d = 0.53) and negative (p = 0.010, d = 0.43) symptoms, which survived multiple corrections correction, but not general symtpoms. Analysis of biomarker data pre and post 3 months of natalizumab showed there was no significant change in mean PET measures in the total GM (Mean ± standard deviation (SD) change: +0.001 ± 0.018, p = 0.825) or temporal lobe GM ( + 0.002 ± 0.017, p = 0.497). However, there was a positive correlation between percent change in PANSS total scores and change in DVR in the total GM (rho = 0.682, p < 0.05).

Conclusions: These data indicate TSPO levels are higher in gray matter in people with schizophrenia relative to reference tissue in comparison to healthy controls, and, in absolute terms, this is particularly marked in the temporal cortex. Natalizumab treatment was associated with improvement in symptoms and improvement in symptoms was associated with a reduction in the marker of relative TSPO levels in gray matter. Together these findings are consistent with the hypothesis that glial activation plays a role in the pathophysiology of schizophrenia and identify them as a potential novel treatment target. Placebo controlled data (currently being analysed) will help determine the specificity of effects to natalizumab. We will consider the implications of these findings for the synaptic hypothesis of schizophrenia and novel drug treatments for schizophrenia.

Disclosure: Angellini, Autifony, Biogen, Boehringer-Ingelheim, Delix, Eli Lilly, Elysium, Heptares, Global Medical Education, Invicro, Jansenn, Karuna, Lundbeck, Merck, Neumora, Neurocrine, Ontrack/ Pangea, Otsuka, Sunovion, Teva, Recordati, Roche, Rovi and Viatris/ Mylan: Advisory Board (Self)

Mini Panel

31. Neural Mechanisms Underlying Mammalian Transmission of Affective and Motivated Behaviors Across Generations

31.1 Abstract Not Included

31.2 Prolonged Abstinence From Morphine Restores Sperm Small Non-Coding RNA Content in Sires and Protects Offspring From Multigenerational Transmission of Addiction-Related Phenotypes

Mathieu Wimmer

Temple University, Philadelphia, Pennsylvania, United States

Background: The vulnerability or resilience to develop neuropsychiatric diseases is modulated by complex interactions between genetic and environmental factors. Substance use disorders (SUDs) are heritable and genetics account for 50−70% of SUD risk. It is inherently difficult to tease apart the various contributions to SUD risk in human and clinical studies and establishing causality is not possible in this context. Rodent studies have begun to untangle whether parental consumption of cocaine, cannabis, or opioids prior to conception is sufficient to produce long-lasting changes in future generations. In our multi-generational model, male rats have daily access to morphine for 60 days (the duration of rat spermatogenesis) prior to mating with drug-naïve females. Our studies demonstrate that paternal morphine exposure leads to an increase in the reinforcing efficacy of morphine in adult male progeny. Moreover, adolescent males produced by morphine-treated sires show reduced rough and tumble play, a critical developmental milestone that influences later risk for SUD-related endophenotypes. Interestingly, if sires are allowed to recover for 90 days following chronic morphine taking prior to mating with drug naïve females, changes in play and morphine self-administration dissipate in male progeny. We leveraged these observations to better understand the mechanisms underlying transmission of opioid exposure from sires to future generations.

Methods: Male Sprague Dawley rats had daily access to morphine self-administration (0.75mg/kg/infusion; 3 hours/day) for 60 days; controls received saline. Sperm was collected from the cauda epididymis 24 h after the last morphine self-administration session. A separate cohort of animals underwent the same morphine regimen followed by 90 days of forced abstinence in their home cage prior to sperm collection (morphine n = 7; saline n = 7, saline + 90 days abstinence n = 5; morphine+90 days abstinence n = 6). RNA was extracted from all samples and small RNA libraries were prepared and sequenced on the Illumina HiSeq 4000 platform. Following FAST QC, pre-processing and annotation to the miRbase and Ensembl ncRNA rat databases were performed using sRNAbench. Planned pairwise comparisons between experimental groups were performed using DESeq2 to identify differentially expressed transcripts (Benjamini-Hochberg adjusted p value < 0.1 and log2 expression fold change > |0.58 | ).

Results: Two small RNAs, rno-miR-150-5p and a snoRNA annotated to Snora42, were differentially expressed exclusively in sperm of morphine-exposed males. Morphine exposure was associated with an increase in rno-miR-150 and a decrease in Snora42 in sperm. Importantly, expression of these transcripts was not altered by morphine exposure followed by abstinence or by abstinence from saline treatment.

Conclusions: Changes in small RNA expression in sperm elicited by chronic morphine self-administration are transient and recover with extended forced abstinence. Similarly to germline signatures, the behavioral consequences of paternal morphine exposure dissipate when the sires undergo extended forced abstinence prior to conception. The current studies lay the foundation for functional delineation of mechanisms underlying inheritance of paternal morphine exposure.

Disclosure: Nothing to disclose.

31.3 Abstract Not Included

Study Group

32. Will Recent Therapeutic Advances Change the Clinical Course of Alzheimer’s Disease

Davangere Devanand, Gary Small, Lauren Gerlach, Doris Molina-Henry, Luis D Medina, Helen Kales, Jennifer Gatchel, Lon Schneider

New York State Psychiatric Institute, New York, New York, United States

Study Group Summary: Recent developments with lifestyle interventions, novel biomarkers, and approved anti-amyloid drugs have raised hopes for improved diagnosis and intervention, but controversies have arisen because of the varied results and risk of adverse effects with some of these new strategies. We will review the evidence for the use of cognitive training, physical exercise, and other lifestyle interventions for both the prevention and treatment of cognitive decline and dementia in older adults.

Key Questions: Do these lifestyle interventions really work and what are the underlying mechanisms? Does combining them lead to improved cognition and function?

The new biomarkers are based on both brain imaging and peripheral blood markers of amyloid and tau brain pathology, which are the hallmarks of Alzheimer’s disease. Lecanemab was recently approved by the FDA and it markedly reduces brain amyloid and is associated with less cognitive decline than placebo, but carries the risk of MRI amyloid-related imaging abnormalities (ARIA).

Key Questions: Is it worth prescribing lecanemab and similar anti-amyloid agents (if approved in the future) when the benefit to risk ratio is small and cost is a major issue? Will these putative disease-modifying medications actually alter the clinical course of Alzheimer’s disease? What about for patient populations with long-standing psychiatric histories and comorbid cognitive decline?

Psychiatric symptoms are common in dementia and diagnosis and management of psychosis and agitation remain major clinical challenges. The recent approval of brexpiprazole to treat agitation in Alzheimer’s disease is a step forward that needs to be considered in the context of other available behavioral and medication strategies.

Key Questions: Is brexpiprazole unique in reducing agitation or is its approval an artifact of timing with respect to changes in regulatory priorities? Do antipsychotics have a role in the treatment of psychosis and agitation in dementia even though side effects can be prominent, including an increased risk of mortality?

The study group will briefly review these latest developments in the field of Alzheimer’s disease and then address these important questions with active audience participation in the clinical management of patients with Alzheimer’s disease and early cognitive decline. The study group leaders and members are well-known researchers in this field and have contributed to the growing evidence base while linking this information to practice across a variety of clinical disciplines.

Disclosure: Acadia: Advisory Board (Self). BioXcel: Advisory Board (Self). Eisai: Advisory Board (Self). GSK: Advisory Board (Self)

Panel

33. Motivated by Fear or Hope: The Role of Reward Circuitry in Anxiety Treatment Response

33.1 Abstract Not Included

33.2 Striatal Activity During Loss of Reward Predicts Anxiety Reductions During Behavioral Activation and Exposure-Based Therapies for Adults With Generalized Anxiety Disorder

Hannah Berg

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

Background: An emerging body of research has revealed anxiety-related alterations in reward circuitry, suggesting that reward systems may also play a role in anxiety treatment. In the present study, we examined reward-related neural predictors of behavioral treatment outcomes for generalized anxiety disorder (GAD).

Methods: In this randomized clinical trial, adults with GAD completed the Monetary Incentive Delay task during functional magnetic resonance imaging (fMRI) pre-treatment, then were randomized to complete 10 weekly sessions of behavioral activation (BA) or exposure-based therapy (EXP). Of the 94 participants allocated to treatment, 70 completed treatment and 49 (26 BA, 23 EXP, 90% female) met data quality standards for inclusion in fMRI analysis. Neural activations during anticipation and receipt of reward gain and loss were assessed in bilateral caudate and nucleus accumbens as a priori regions of interest (ROIs), and across the whole brain. GAD symptoms were assessed before, during, and after treatment. Linear mixed-effects models were conducted to examine interactions of time, treatment-arm, and neural activity on GAD symptom trajectory.

Results: Lower left caudate activity during loss receipt predicted poorer symptom improvement in EXP but not BA (F(1, 428) = 5.24, p = 0.023, d = −0.22), such that individuals with lower levels of activation in this region were more likely to benefit from BA than EXP. Whole brain analyses echoed this finding and identified similar patterns in functional nodes of the salience (dorsomedial prefrontal cortex, right anterior insula) and frontoparietal networks (dorsolateral prefrontal cortex, right intraparietal lobule).

Conclusions: BA is shown to be more effective than EXP at reducing GAD symptoms for individuals with blunted striatal and cortical reactivity during loss. One interpretation of these findings is that BA, unlike EXP, may serve to compensate for blunted loss reactivity, perhaps due to its focus on monitoring affective reactions to salient events. Activity in the caudate and in salience network and frontoparietal regions may reflect loss-related aversive arousal, avoidance motivation, and/or emotion regulation. Further research is warranted to investigate whether psychological treatments can be optimized by addressing imbalanced reward and loss reactivity.

Disclosure: Nothing to disclose.

33.3 MRI and Ecological Momentary Assessment Proxies for Dopamine Availability: Predictors of Adolescents’ Anhedonia and Anxiety Improvement With Reward-Targeted TMS

Erika Forbes

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Standard transcranial magnetic stimulation (TMS) for depression targets negative affect systems, does not incorporate multimodal approaches, and fails to address disruption in reward systems. TMS to reward systems could particularly benefit adolescents, given ongoing development of these systems. Frontostriatal hyperconnectivity provides a viable target, as it is postulated to underlie anhedonia, which involves low reward motivation, anticipation, or enjoyment and predicts poor treatment response. Dopamine (DA) availability could be leveraged by such treatment. Anxiety, which often accompanies adolescent depression and involves altered reward systems, could also improve with reward-focused TMS. We examined whether neural and behavioral characteristics reflecting DA availability predict response to continuous theta burst stimulation (cTBS; brief TMS thought to reduce cortical response) to dorsomedial prefrontal cortex (dmPFC) with augmentation by positive affect (PA)-enhancing behavioral training.

Methods: Youth with depressive disorder (N = 45, 15−25 years, 79% female) completed 10-visit, 2x/visit cTBS with dmPFC target, plus PA exercises (Craske et al., 2016) between cTBS administrations. Participants completed SHAPS for anhedonia and GAD7 for anxiety, 3T magnetic resonance imaging of DA availability via a brainstem neuromelanin scan and an R2’ scan for striatal tissue iron, and 1 week of 7x/day ecological momentary assessment (EMA) of mood and symptoms. LASSO regressions predicted anhedonia or anxiety improvement from DA variables and EMA variables postulated to reflect DA availability (energetic or happy mood, 0-100 scale; anticipatory or motivational anhedonia, yes/no).

Results: Both anhedonia and anxiety decreased with TMS. Anhedonia improved for those with pre-TMS higher happy and energetic mood, lower anticipatory anhedonia, higher baseline tissue iron in caudate, and lower left substantia nigra (SN) neuromelanin. Anxiety improved for those with pre-TMS lower happy mood, lower anticipatory and motivational anhedonia, and higher caudate and accumbens tissue iron and higher SN neuromelanin.

Conclusions: cTBS to dmPFC could be valuable for treating adolescent depression. Findings

Disclosure: Nothing to disclose.

33.4 Abstract Not Included

Panel

34. Spirited Aging: What’s New With Alcohol and Brain Fluids

34.1 Choroid Plexus Perfusion and CSF Secretion is Blunted in Middle-Aged Rats Chronically Exposed to Alcohol

Helene Benveniste

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

Background: Enlarged cerebral ventricles is associated with alcohol use disorder (AUD). While ventricular enlargement in AUD likely reflects brain tissue loss, it might also indicate abnormal fluid homeostasis which in itself can contribute to damage and cognitive impairment. Dysregulation of fluid homeostasis by cEtoh might involve cerebrospinal fluid (CSF) secretion/reabsorption processes. The major portion of CSF is produced in the choroid plexuses (ChP). Here we highlight novel data on age-related changes in ChP blood perfusion and CSF secretion rates in a rat AUD model.

Methods: Two groups of Fisher (CDF) 344 female rats were studied. One (‘cEtoh’, N = 14) was treated with Etoh vapor 8h/day, 5days/week from 12-15 month of age; and the other was administered air (N = 14) in vapor chambers. Etoh vapor levels were ~14 mg/L to yield breath Etoh of ~100 mg/dL. The vapor exposure approach was designed to mimic “risky” alcohol consumption (3−4 drinks over 1−2 h). To quantify ChP perfusion and blood-to-CSF-barrier (BCSFB) mediated water exchange (surrogate for CSF secretion), we implemented arterial spin labeling (ASL) MRI. All rats underwent two scan sessions: One at baseline and the second at the end of the 3-month Air/ Etoh exposure. Normalized ChP perfusion and BCSFB weighted signals were fitted to derive ChP blood perfusion, and CSF secretion rates.

Results: The levels of Etoh measured in the cEtoh rats by breath averaged ~100 mg/dL. In the air-exposed rats, ChP blood perfusion and CSF secretion rates increased in an age-dependent manner by ~4% (p value = 0.041) and 24% (p value = 0.030), respectively, from 12-15 months. In contrast, no changes in neither ChP perfusion (p value = 0.332) nor CSF secretion rates (0.807) were observed in the cEtoh rats after the 3-month Etoh exposure. There was no significant group*age interaction effects for ChP perfusion (p value = 0.456) or for CSF secretion rates (p value = 0.176).

Conclusions: Understanding CNS fluid dynamics in AUD is pivotal, due to its many functions. We uncovered that ChP perfusion and CSF secretion rates increased in normal rats from 12 to 15 month of age implying that cerebral fluid homeostasis adjusts even in middle-age. However, in middle-aged cEtoh rats, the expected age-related increases were obliterated, implying that Etoh levels that mimic ‘risky’ drinking in humans alters cerebral fluid homeostasis.

Disclosure: Nothing to disclose.

34.2 Brain Morphometric Changes Including Local CSF Expansion Observed in a Rat Model of Alcohol Use Disorder

Sunil Koundal

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

Background: Clinically, chronic alcohol consumption is associated with enlarged cerebral ventricles and tissue loss. Intriguingly, in subjects with alcohol use disorder (AUD), CSF volume changes are dynamic and can reverse with abstinence. Here we investigated whether long-term exposure to ‘risky’ high levels of alcohol exposure in middle-aged rats would result in similar morphometric changes observed clinically.

Methods: We studied 12-month-old female Fisher 344 rats exposed to air (N = 14) or ethanol vapor (N = 14) for three months (5daysX8hr/week). Ethanol vapor was maintained at approximately 14 mg/L to achieve a blood/breath ethanol concentration of ~100 mg/dL, exceeding the U.S. legal limit for drunk driving. To assess morphometry, 3D proton density-weighted (PDW) brain scans were acquired at a spatial resolution of 300 µm3 at baseline (12 months) and post-exposure (15 months) using a Bruker 9.4T/30 MRI system. All PDW images were segmented into gray matter (GM), white matter (WM), and CSF tissue compartments. The Jacobian determinant for the deformation fields were calculated, followed by creation of % volume change from baseline maps for both Air- and Etoh-exposed groups. The % volume change maps were smoothed and compared between groups (Air vs Etoh) using t-test within the statistical framework of a general linear model.

Results: The analysis uncovered ventricular dilation and ‘redistribution’ of the CSF volume in Etoh compared to Air rats. Specifically, in the Etoh rats the CSF compartment expanded in the 3rd ventricle, aqueduct and olfactory subarachnoid space (SAS) and shrank in SAS at the level of the circle of Willis, pineal recess and ambient cistern. Furthermore, in the Etoh group shrinkage was evident in the hippocampus, thalamus, hypothalamus and several cortical regions.

Conclusions: Our study provides detailed insights into the deleterious effects of AUD in middle-aged female rats revealing expansion of the CSF compartment locally in the 3rd ventricle and aqueduct and shrinkage in regions functionally involved in reward and memory functions.

Disclosure: Nothing to disclose.

34.3 Compromised Brain Lymphatic Drainage Identifies a Presynaptic Hot Spot of Metabolic Redox Vulnerability

Laura Santambrogio

Weill Cornell Medical College, New York, New York, United States

Background: The brain anatomical and functional complex architecture comes at a high metabolic cost, thus producing a substantial amount of metabolic by-products and interstitial waste products which are removed by a glia-dependent system of peri-venous conduits which will drain into the meningeal lymphatic vessels with subsequent transport to the cervical lymph nodes

Methods: To comprehensively analyze the biological implications of impaired CNS lymphatic drainage we used three different models: (i) mice where the afferent lymphatics draining to the cervical node were ligated (ii) mice treated with the photodynamic drug Visudyne (which render the brain lymphatics functionally impaired) and (III) rats where the afferent lymphatics entering the deep cervical nodes were cauterized. Both male and female mice were used, 6−8 mice were used in each experimental group. Metabolomic and proteomic analysis were performed on the CSF and hyppocampus is all groups. Values, from biologically independent replicates, were statistically analyzed using a one-way ANOVA (alpha = 0.05), followed by Tukey’s multiple comparisons test. Significance levels are reported as (*p  <  0.05, **p  <  0.01, ***p  <  0.001, ****p  <  0.0001).

Results: Disfunction in lymphatic drainage resulted in a CSF proteomic and metabolomic signature of redox imbalance. The cellular metabolic stress was associated with increased CSF release of damage-associated molecular patterns and a mild inflammatory signature. This was mirrored in a signature of reduced brain proteostasis as highlighted by the presence of protein microaggregates in the 1x10^9 kDa range, modified by oxidative post-translational modifications (PTMs), including carbonylation, sulfation, citrullination and nitrosylation. Proteins from pre-synaptic mitochondria as well as a synaptic vesicular trafficking proteome were the most represented in the PTMs-modified microaggregates, indicating that presynaptic regions are a hot spot for metabolic stress following decreased brain waste disposal.

Conclusions: Overall, our findings quantitatively map at the metabolic and proteome level the relevance of lymph-driven waste disposal on brain homeostasis and proteostasis.

Disclosure: Nothing to disclose.

34.4 Effects of Aquaporin Genotype on Sleep Quality and Cerebrospinal Fluid in Alcohol Use Disorder

Corinde Wiers

University of Pennsylvania, Philadelphia, Pennsylvania, United States

Background: Alcohol use disorder (AUD) is associated with sleep disturbances and marked changes in brain structure that have been ascribed in part to alcohol’s acceleration of aging, including increases in cerebrospinal fluid (CSF). CSF flow through the brain parenchyma is facilitated by the astrocytic water channel aquaporin 4 (AQP4). Genetic variance in the human AQP4 gene have recently been associated with sleep quality in healthy volunteers, and with cognitive decline in patients with Alzheimer’s disease. Here, we aimed to study the effects AQP4 on sleep quality, depressive mood, and CSF volumes in patients with AUD.

Methods: We analyzed the effects of AQP4 SNP rs335929 on sleep quality (Pittsburgh Sleep Quality Index) and depressive mood (Montgomery-Asberg Depression Rating Scale) measured as part of a screening and natural history study of 630 participants with AUD and 507 non-dependent healthy control (HC) participants. In a subset of participants (n = 142 AUD, n = 180 HC), we evaluated the effect of rs335929 on brain CSF levels, as quantified using structural MRI and analyzed with a standard FreeSurfer segmentation pipeline.

Results: Patients with AUD reported poorer sleep quality (F = 302.1, p < 0.0001) higher depression scores (F = 452.1, p < 0.0001), and larger CSF volumes than HC (F = 2.09, p = 0.055); corrected for age and sex. There was a significant AUD x rs335929 interaction effect on sleep quality (F = 6.64, p = 0.010), such that minor C-allele carriers demonstrated poorer sleep quality compared to major allele AA homozygotes in the AUD (F = 5.43, p = 0.020), but not the HC group. A similar interaction effects of AUD x rs335929 was present for depression scores (F = 6.35, p = 0.013), with higher depression scores in the AUD minor compared to major allele group. In a subset of patients who underwent MRI, the AUD x rs335929 interaction effect on CSF volumes was significant (F = 5.94, p = 0.015), with smaller CSF volumes in the AUD minor allele group. Throughout the cohort, CSF volumes correlated with both poorer sleep quality (r = 0.142, p = 0.011) and depression scores (r = 0.181, p = 0.001).

Conclusions: We found first evidence for a link between genetic variation in AQP4, sleep quality, and depressive mood in AUD. The associations between AQP4 and CSF further suggest a potential mechanism of AQP4 in affecting CSF volumes thereby contributing to poorer sleep and depression.

Disclosure: Nothing to disclose.

Panel

Parent-Child Stress Transmission: Translational Models for the Study of Psychopathology Risk in Youth

35.1 The Adverse Early Environment of Paternal Deprivation, in the Biparental California Mouse, Leads to Sex-Specific Social Behavior Deficits and Neuroimmune Suppression

Erica Glasper

The Ohio State University, Columbus, Ohio, United States

Background: Nearly one-quarter of US children are reared in a home lacking appropriate parental care. This can increase the likelihood of social behavior problems and serious health outcomes. In the absence of paternal care (i.e., paternal deprivation) in the biparental California mouse (Peromyscus californicus), vigilance-avoidance behavior (i.e., social anxiety-like) was observed in males and females. This behavior was associated with reduced pro-inflammatory cytokines in brain regions involved in socio-cognitive and stress-related behaviors in males, but not females. The extent to which neuroinflammatory processes underlie paternal deprivation-related alterations in sociability is unknown. We hypothesized that paternal deprivation induces lasting sex-specific impairments in neuroimmune function and social anxiety-like behaviors.

Methods: Using a three-chamber apparatus, control-reared and paternally-deprived adult male and female California mice (n = 12−18/group) performed a social preference test to assess preference for a novel mouse versus a novel object. Inflammatory gene expression was determined in brains (n = 3/group), via NanoString nCounter, to identify pathways and processes involved in neuroinflammation. Differentially expressed genes (DEGs) were compared to the Simons Foundation Autism Research Initiative gene database to identify DEGs related to social and cognitive impairments. Ingenuity pathway analysis of significant master regulators were compared within each sex.

Results: Male mice exhibited a shorter latency to enter the novel mouse chamber, compared to the novel object chamber(p ≤ 0.001). Paternally-deprived females exhibited a shorter latency to enter the novel mouse chamber, compared to the novel object chamber (p ≤ 0.0001), and paternally-deprived females exhibited a shorter latency to enter the novel mouse chamber, compared to the control-reared females (p ≤ 0.001). Neuroimmune-related canonical pathways were significantly inhibited in paternally-deprived females, compared to control-reared females (p ≤ 0.05). Paternal deprivation did not inhibit any neuroimmune-related canonical pathways in males (p ≥ 0.05).

Conclusions: Paternal deprivation may induce greater social impairments in females that are accompanied by neuroimmune suppression.

Disclosure: Nothing to disclose.

35.2 Altered Neural Encoding of Threat and Modification of Learned Threat Through Parental Observation in Normative and Trauma-Exposed Adolescents

Sara Heyn

University of Wisconsin—Madison, Madison, Wisconsin, United States

Background: The use of social cues to attenuate prior learned fear associations is a critical skill for normative development. Impairments in extinction learning underlie emerging models of posttraumatic stress disorder (PTSD) in adults. However, few have investigated how trauma may impact the unique neural mechanisms of observed-parent vicarious safety learning during adolescence.

Methods: Typically developing (TD, N = 43) and trauma-exposed (TE, N = 40) parent-child dyads underwent a three-day neuroimaging paradigm characterizing the transmission of parental social cues that may modify established threat associations. Youth and parents complete acquisition (Day 1), direct (CS+Direct) and vicarious extinction by watching a video of parent direct (CS+Vicarious) on Day 2, and extinction recall and reinstatement on Day 3. Whole-brain voxelwise activation analyses were used to characterize normative and trauma-related neural correlates of acquisition and observational extinction.

Results: Activation in dorsolateral prefrontal (K = 366, F = 6.10, PCORR < 0.05) and posterior insula (K = 343, F = 7.65, PCORR < 0.05) cortices underlie CS + /CS- differentiation during early fear acquisition, where activation magnitude uniquely encodes CS-type. TE youth are less successful at discriminating threat and safety during acquisition than TD youth (F = 4.31, P = 0.014) and exhibit generalized threat associations of all CS-types through extinction recall (F = 5.0, P = 0.025). Atypical stimulus differentiation during acquisition is related to differential activation in visual association areas (V3, K = 122, F = 10.81, PCORR < 0.05; precuneus, K = 129, F = 11.28, PCORR < 0.05) in TE youth, which is enhanced by increasing depression (F = 7.91, PCORR = 0.02) and trauma load (F = 9.81, PCORR = 0.003). Finally, aberrant threat generalization in TE youth is supported by a complete reversal of reactivity patterns to threat and safety during direct (fusiform gyrus, K = 81, F = 13.21, PCORR < 0.05) and vicarious extinction (vmPFC, K = 89, F = 16.51, PCORR < 0.05).

Conclusions: Youth can directly modify learned threat associations and vicariously through parent observation via unique patterns of reactivity. Trauma distinctly impacted the early encoding of visual stimuli, reactivity patterns of threat and safety during extinction, and was marked by persistent threat generalization.

Disclosure: Nothing to disclose.

35.3 Observational Fear Learning in Children of Trauma-Exposed Mothers: Preliminary Data on the Moderating Role of Age and Sex

Marie-France Marin

Université du Québec à Montréal, Research Center of the Montreal Mental Health University Institute, Montréal, Canada

Background: Children of parents with PTSD are at increased risk for fear-related disorders. The prevalence of such disorders increases at puberty, when girls are twice as likely as boys to develop them. This study aimed to examine children’s observational fear learning abilities as well as fear regulation patterns as a function of the mother’s trauma history, as well as the child’s sex and age.

Methods: To date, 176 healthy children between 8 and 16 years of age (93 girls, 83 boys) have been recruited with their mother, who either had a history of trauma (n = 97, Trauma + ) or not (n = 79, Trauma-). On Day 1, children watch two videos: one of their mother undergoing a fear conditioning protocol in which a cue (e.g., blue lamp; CS+Parent) is paired with a shock and a cue (e.g., yellow lamp) is not reinforced (CS-), and a video of a stranger undergoing a similar procedure in which a different cue (e.g., red lamp; CS+Stranger) is reinforced. After this observation period, the children are exposed to the 3 cues to test the acquisition of fear. No shock is administered. After several presentations of the unreinforced cues, extinction learning takes place. The next day, all cues are presented again to test extinction recall. At each phase, a Stimulus X Sex X Age X Trauma ANOVA was conducted on skin conductance responses.

Results: On Day 1, no main effect of Trauma or interaction with this variable was found. On Day 2, during extinction recall, a Stimulus X Age X Trauma was found, F(1.81, 309.66) = 5.83, p = 0.004, ηp2 = 0.033. Young children (8−11 years old) of Trauma+ mothers tended to show a greater return of fear regardless of stimulus compared to young children of Trauma- mothers, F(1, 99) = 3.03, p = 0.085, ηp2 = 0.030. Among older children (12-16 years), there was a greater return of fear to the CS- for those whose mothers were Trauma+ compared to children of Trauma-, p = 0.045, ηp2 = 0.057.

Conclusions: These preliminary data suggest that maternal trauma history may affect the child’s ability to regulate fear. In younger children, there appears to be an overall deficit in the retrieval of information formed during extinction learning. In older children, the deficit appears to be specific to the stimulus signaling safety. This highlights the importance of considering developmental effects when investigating the consequences of intergenerational transmission of fear in trauma-affected families.

Disclosure: Nothing to disclose.

35.4 Maternal Childhood Trauma and Fear Circuit Connectivity in Her Child: Intergenerational Effects of Trauma

Tanja Jovanovic

Wayne State University School of Medicine, DETROIT, Michigan, United States

Background: Childhood experiences of trauma in mothers can have intergenerational effects on her children. Maternal early trauma may be particularly critical, as it impacts mental health, parenting, and brain structure and function in their children. Several studies have investigated brain development associated with trauma-related psychopathology in children. This research shows evidence of intergenerational effects of trauma on offspring brain development, however, the mechanisms of this transmission are not yet clear. The current study examined associations between maternal child abuse history and their children’s resting-state functional connectivity (rsFC) between the amygdala (AMY) and anterior cingulate cortex (ACC), two structures involved in regulating fear responses and psychopathology. We hypothesized that maternal childhood trauma exposure would be associated with increased connectivity within this circuitry.

Methods: Black mother-child dyads, N = 52, completed trauma interviews and magnetic resonance imaging (MRI) scans. The children (N = 52, 27 girls; 8- to 14-years-old, MAge = 10.74, SD = 1.61) underwent resting-state functional MRI in a Siemens 3T scanner with eyes open for 7min.

Results: Mean rsFC between left and right AMY to ACC was positive, MLeft = .173, SDLeft = 0.213, MRight = 0.140, SDRight = .185. For left AMY-ACC rsFC, all models were significant for the Block 1 and 2 predictors, all ps < .003. Maternal physical abuse explained additional variance, ΔR2 = .101, ΔF(1,39) = 7.642, p = .009, and was associated with increased connectivity, B = 0.343, t = 2.764. Emotional abuse explained marginally more variance, ΔR2 = .056, ΔF(1,39) = 3.884, p = .056, and was linked to increased rsFC, B = 0.251, t = 1.971. Sexual abuse was not significantly associated with left AMY-ACC rsFC, p > .350. In contrast to these associations in the left hemisphere, right AMY-ACC rsFC was not significantly associated with any type of maternal childhood abuse, all ps > .09.

Conclusions: These results provide novel evidence that maternal childhood abuse history is associated with increased left frontoamygdala connectivity in school-aged children. These results suggest that trauma may have intergenerational effects and highlight the need for further study of how maternal trauma exposure confers risk via biological pathways that modulate neurocircuitry.

Disclosure: Nothing to disclose.

Panel

36. Manipulation of Aversive Events and Memories by Psychedelics

36.1 Region-Specific Effects of Psilocin on Stimulus Reactivity and Threat Responding Behavior in Male and Female Rats

Melissa Herman

University of North Carolina School of Medicine, Chapel Hill, North Carolina, United States

Background: Current research suggests that psychedelic compounds may have therapeutic efficacy in the treatment of psychiatric conditions including anxiety, treatment resistant depression, and substance use disorders, however the underlying neurobiology and brain region-specific effects remain poorly understood. Many psychiatric disorders involve stress or dysregulated responsivity of the stress circuitry. The paraventricular nucleus of the hypothalamus (PVN) is a midline hypothalamic nucleus that plays an integral role in stress reactivity, autonomic functioning, social behavior, and many other affective processes.

Methods: We used immunohistochemistry and fiber photometry recordings of awake behaving rats to investigate the effects of psilocin (2 mg/kg, sc) on PVN activity and threat response behavior. Adult male and female Sprague Dawley rats (200-400 g) were injected with pGP-AAV-syn-jGCaMP7f-WPRE and fibers were bilaterally implanted into the PVN. During recording, rats were exposed to a brief air-puff stimulus and time-locked changes in fluorescence (ΔF/F) and behavior were measured and compared between male (psilocin n = 9; vehicle n = 9) and female rats (psilocin n = 11, vehicle n = 9). Recordings were performed during acute drug treatment and at 2 and 7days following injection. At 10 days post injection rats were subjected to a 20 min restraint prior to a final recording session.

Results: Psilocin increased c-Fos in the PVN of both males (t(8) = 2.603, p = 0.03) and females (t(8) = 2.324, p = 0.04). Psilocin-injected males had greater PVN stimulus reactivity [F(1,16) = 5.373, p = 0.03] and decreased threat response [velocity: F(1,21) = 6.246, p = 0.02; distance traveled: F(1,21) = 4.896, p = 0.03] as compared to vehicle controls. There were no differences in PVN reactivity in females or in either sex at 2- and 7-days post-injection or following an acute restraint stress. Experimental groups were sub-classified by baseline threat responding and psilocin-induced increases in male PVN reactivity were found to be driven by active threat responders.

Conclusions: Collectively, these data indicate that psilocin acutely engages the PVN in a sex-specific manner and time-dependent manner but reactivity following acute stress is preserved. Further, these data suggest that treatment responsiveness may be related to baseline threat responding behavior.

Disclosure: Nothing to disclose.

36.2 Behavioral and Neural Correlates of Psilocybin’s Effect on Fear and Punishment Learning

David Jacobs

Oregon Health and Sciences University, Portland, Oregon, United States

Background: Aversive events can exacerbate symptoms of mental illness. Moreover, the interplay between reward seeking and risk of aversive outcomes is critical for optimal planning of goal-directed actions. We posit that a better understanding of how psychedelics affect processing and formation of aversive experiences is critical for understanding their therapeutic and side-effects.

Methods: Adult male and female rats were used in two novel operant tasks to assess the effects of psilocybin on various aspects of punishment learning. The first one utilized a probabilistic punishment task where animals learned the association of risky versus safe reward-motivated actions (N = 31). We examined the effect of psilocybin on both learning and post-learning performance. The second task was a Kamin blocking paradigm where animals learned that a single or a compound conditioned stimulus predicted footshock (N = 36). Psilocybin was given in the first session of compound training and animals were later tested for prediction error signaling. We also used chemogenetics to investigate the role of thalamocortical projections from mediodorsal thalamus to prelimbic cortex, a pathway implicated in Kamin blocking, on behavioral performance.

Results: Psilocybin (1 mg/kg) produced divergent effects during different phases of the probabilistic punishment task. Over the learning phase, it reduced responding in a sex-specific manner (female p =0.046; male p > 0.50). After learning, psilocybin increased responding in both sexes (p = 0.01). In Kamin blocking, psilocybin enhanced fear learning during the compound training phase (p = 0.03) but not at test (p = 0.30). A 3 mg/kg dose of psilocybin, however, produced a disruption at test for prediction error as indicated by enhanced fear expression in block and control groups (p = 0.03). Chemogenetic inhibition of thalamocortical projections produced comparable changes in prediction errors during Kamin blocking.

Conclusions: We find that the impact of psilocybin on punishment processing is context, experience, and sex specific. The Kamin blocking results further indicate that psilocybin can engender aberrant learning. These findings highlight the context specific effects of psilocybin in aversive processing and may provide mechanistic insight into psilocybin’s effects on learning and aberrant prediction error signaling.

Disclosure: Nothing to disclose.

36.3 Psilocybin Facilitates Fear Extinction: Importance of Dose, Context, and Serotonin Receptors

Alex Kwan

Cornell University, Ithaca, New York, United States

Background: A variety of classic psychedelics and MDMA have been shown to enhance fear extinction in rodent models. This has translational significance because a standard treatment for posttraumatic stress disorder is prolonged exposure therapy. However, few studies have investigated psilocybin’s potential effect in fear learning paradigms. More specifically, the extents to which dose, timing of administration, and serotonin receptors may influence psilocybin’s effect on fear extinction are not understood.

Methods: In this study, we used an auditory delay fear conditioning paradigm to determine the effects of psilocybin on fear extinction, extinction retention, and fear renewal in male and female mice. We tested various doses, timing of administration, and co-treatment with serotonergic receptor antagonists, in total using a sample size of 112 male and female mice.

Results: Psilocybin robustly enhances fear extinction when given acutely prior to testing for all doses tested. Psilocybin exerts long-term effects to elevate extinction retention and suppress fear renewal in a novel context, though these changes were sensitive to dose. Administration of psilocybin prior to fear learning or immediately after extinction yielded no change in behavior, indicating that concurrent extinction experience is necessary for the drug’s effects. Co-treatment with a 5-HT2A receptor antagonist blocked psilocybin’s effects for extinction, extinction retention and fear renewal, whereas 5-HT1A receptor antagonism attenuated only the effect on fear renewal.

Conclusions: Collectively, these results highlight dose, context, and serotonin receptors as crucial factors in psilocybin’s ability to facilitate fear extinction. The study provides preclinical evidence to support investigating psilocybin as a pharmacological adjunct for extinction-based therapy for PTSD.

Disclosure: Empyrean Neuroscience: Advisory Board (Self). Freedom Biosciences: Advisory Board (Self). Psylo: Advisory Board (Self). Intra-Cellular Therapies: Contracted Research (Self)

36.4 The Neural Circuit Mechanisms Underlying the Effects of LSD on Pain Perception

Tanya Sippy

New York University, New York, New York, United States

Background: Conventional therapeutic agents for chronic pain including opioid and non-opioid analgesics have been associated with adverse side effects and issues with addiction. Novel agents are urgently needed to treat pain. LSD, or lysergic acid diethylamide, has been previously investigated for pain control in humans and animals but a strong rodent model of this has not been established.

Methods: We used conditioned place aversion (CPA), in which rats learn to a avoid a chamber in which they have been exposed to pin prick in their forelimb. We then dosed rats with LSD (85ug/kg, n = 14) or saline (n = 6) IP and directly into the anterior cingulate cortex (ACC) (n = 7, saline; n = 7), and tested them in the CPA assay 30 min and 7 days later. We directly administered LSD to S1 (n = 5), and infused a serotonin 5HT2A antagonist, MDL 100907 (n = 8) into the ACC. We performed longitudinal Ca2+ imaging of ACC neurons during pain perception with IP injection (n = 556 neurons from 13 rats) or LSD (n = 546 neurons, 13 rats), and two weeks after LSD (n = 745 neurons, n = 13 rats).

Results: While rats dosed IP with saline displayed pain avoidance after the initial injection (p = 0.0003), rats that received an IP injection of LSD demonstrated a decreased avoidance (p = 0.6096). Rats that received an intracranial injection of saline in the ACC demonstrated avoidance (p = 0.0007), while rats that received intracranial injection of LSD did not (p = 0.8535). This effect remained after 7 days (saline, p = 0.0050; LSD, p = 0.4100). Rats infused with saline in S1 displayed avoidance (PP, p = 0.0003), as did rats that were infused with the 5HT2A antagonist, MDL 100907 in the ACC before IP injection of LSD. Ca2+ imaging showed that ACC neurons had a significant decrease in peak activity in response to noxious stimuli that returned to baseline after 2 weeks. (Baseline vs. LSD: p = 0.0386; Baseline vs. 2 Weeks after LSD: p = 0.3323).

Conclusions: Our results show that LSD injected IP can reduce pain aversion in rats and that infusion of LSD directly into the ACC is sufficient for this effect. This effect is specific to ACC, given that LSD infused in S1 did not produce such aversion. The 5HT2A receptor in the ACC is necessary for the abolishment of aversion to painful stimuli. An underlying mechanism for this behavioral change is a reduction of the activity of ACC neurons to noxious stimuli during acute LSD administration.

Disclosure: Nothing to disclose.

Panel

37. Tuning in, not Shutting Down: Transdiagnostic Mechanisms and Treatments for Physiological Blunting and Withdrawal Responses

37.1 The Role of Amygdala Reactivity and EEG Theta/Beta Ratio in the Association Between Acute Dissociation and Affective-Cognitive Processing in Borderline Personality Disorder

Annegret Krause-Utz

Leiden University, Leiden, Netherlands

Background: Problems in stress regulation and reduced cognitive (attentional) control have been linked to acute dissociation in borderline personality disorder (BPD). Previous neuroimaging research suggests that an imbalance in fronto-limbic regions (e.g., prefrontal regions, amygdala) may underlie these BPD symptoms. However, the impact of acute dissociative states on affective-cognitive functioning in BPD is not yet entirely understood. Moreover, no study so far investigated the role of EEG Theta/beta Ratio (TBR) in the association between acute dissociation and affective-cognitive processing in BPD. TBR, the ratio between slow and fast brainwaves at rest, may be an important biological marker of affective-cognitive control, as it has previously been linked to automatic attentional bias towards emotional (threat) stimuli (see Angelidis et al., 2016).

Methods: This study aimed to investigate if TBR moderates the effect of acute dissociation on cognitive control of emotional stimuli in BPD. Participants with BPD (N = 40) underwent a script-driven imagery procedure (exposure to either a neutral script or the recall of an autobiographical situation that involved dissociation). Participants without BPD and dissociation (N = 40) were the control group. Resting-state EEG was recorded with frontal theta and beta activity being obtained from the Fz, F3, and F4 10/20 positions using Ag/AgCl electrodes of the ActiveTwo BioSemi system. Afterwards, patients performed a modified Sternberg recognition task with emotional material presented as distractors during functional magnetic resonance imaging.

Results: A preliminary analysis of the neuroimaging data (FWE corrected) revealed that patients with dissociation showed impaired cognitive performance (F = 4.43, p = 0.018, η2 = 0.17) and lower reactivity in the amygdala (F = 5.36, p =.008, ƒ2 = 0.20) during the task. Reactivity in the amygdala to aversive pictures was fully mediated by severity of dissociation [B(SE) = −0.010, SE = 0.004, 95%CI = [−0.0175,−0.001]]. Findings on the effect of TBR are currently processed and will be presented.

Conclusions: Acute dissociation had a dampening effect on threat processing, while impairing general cognitive functioning. During (exposure) treatment, dissociation may hinder therapeutic processes aimed at processing stressful memories and experiences.

Disclosure: Nothing to disclose.

37.2 Abstract Not Included

37.3 Restoring Large-Scale Functional Connectivity Patterns in Trauma-Related Dissociation Through a Mind-Body Intervention

Ruth Lanius

University of Western Ontario, London, Canada

Background: The dissociative subtype of post-traumatic stress disorder (PTSD) is characterized by trauma-related dissociation and distinct functional connectivity patterns. However, varying findings across numerous studies emphasize the need for a unified understanding of the neurobiology of dissociation. In addition, it remains critical to examine these connectivity patterns pre- and post-treatment.

Methods: In a retrospective study, we advanced towards this goal by performing one of the most extensive region of interest (ROI)-to-ROI analyses in PTSD to date, assessing resting-state functional MRI data from 192 participants. In follow-up, we examined connectivity patterns before and after a mind-body intervention (n = 52), deep brain reorienting, that has recently been shown to be effective for the treatment of PTSD through a randomized controlled trial (NCT04317820).

Results: Small differences in functional connectivity (maximum effect size 0.27) were observed between participants with PTSD and controls, particularly in the temporal regions and the right frontoparietal network. However, participants with the dissociative subtype exhibited a significantly different pattern of widespread functional hyperconnectivity compared to controls (maximum effect size 0.46), affecting subcortical regions, sensorimotor areas, and other intrinsic connectivity networks. Additionally, an analysis of latent dimensions, considering both ROI-to-ROI brain results and various behavioral and clinical measures, identified three clinically relevant dimensions: two associated with dissociation and one with PTSD symptoms. Critically, subcortical (including brainstem and cerebellum) to cortical connectivity patterns appeared to be restored after a mind-body intervention in individuals with the dissociative subtype of PTSD (pFDR < 0.001), which was associated with a decrease in symptoms of trauma-related dissociation.

Conclusions: These findings enhance our understanding of the neurobiology of dissociation, depicting its brain connectivity as a deviation from the typical small-world organization. Moreover, these results add to emerging data showing the importance of mind body treatments in restoring brain connectivity patterns related to trauma-related dissociation.

Disclosure: Nothing to disclose.

37.4 Mechanisms of “Shutting Down” in a Marginalized Population and Response to a Physiologically-Augmented Intervention

Negar Fani

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

Background: Dissociation is physiological “shutting down” response to chronic trauma, including racial discrimination (RD), which disrupts attention and interoceptive processes. However, these processes can be improved with exteroceptive signaling. In this talk, we discuss associations of RD, dissociation and attentional control in a Black, trauma-exposed community sample, as well as findings from an intervention study for dissociation that employed vibroacoustic breath-synced feedback delivered on the sternum during mindful breath focus (MBF).

Methods: Clinical (N = 487 Black women) and MRI (N = 88 men and women) data were collected in trauma-exposed adults (18−65 years old); n = 88 had clinically significant depersonalization and completed the intervention study, which involved 8 sessions of meditative practice; n = 24 were randomly assigned to receive MBF with breath-synced vibroacoustic feedback (VBMM) delivered via a wearable subwoofer. They also performed an affective attentional control (affective stroop, AS) task during fMRI at pre- and post-intervention to assay change in executive control network activation to trauma-relevant cues.

Results: In N = 487, RD frequency associated with dissociative severity, particularly depersonalization. Depersonalization moderated associations between RD and attentional control, revealing more negative associations among those who reported more RD. In n = 88, a linear mixed effects model revealed that that intervention group moderated the relationship between change in executive network activation during AS task and change in PTSD symptoms. In VBMM, increased activation corresponded with PTSD symptom decrease (r = −0.56, p = .004) and interoceptive awareness (r = −0.56, p = .043); no significant associations were observed in other intervention groups (ps > 0.05).

Conclusions: RD associated with depersonalization and attentional disruptions, and vibration augmentation of MBF may enhance executive control, which, in turn may positively influence attention and emotion regulation and lead to PTSD symptom reduction. Vibration augmentation of MBF may address attention and interoceptive disruptions in dissociative trauma-exposed people (including marginalized groups) and may be used either as a stand-alone treatment or to better prepare individuals for intensive trauma-focused treatments.

Disclosure: Nothing to disclose.

Mini Panel

38. In Sickness and in Health: Using Non-Disease Phenotypes to Elucidate and Decompose Psychiatric Disorders

38.1 Genome-Wide Association Study of Delay Discounting in 134,935 Individuals Reveals Common and Specific Biological Associations Spanning Psychiatric and Somatic Traits and Disorders

Abraham Palmer

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

Background: Delay discounting (DD), which is the depreciation of a reward’s value with its delayed receipt, is positively correlated with several psychiatric conditions including substance use disorders, attention-deficit hyperactivity disorder (ADHD), and schizophrenia, and is negatively correlated with anorexia nervosa and obsessive compulsive disorder.

Methods: We performed a genome-wide association study (GWAS) of DD in 134,945 male and female individuals clustered as European based on genetic similarity. DD was assessed via the 27-item Monetary Choice Questionnaire and quantified by a temporal discounting value (k) wherein higher values indicate greater DD. Our GWAS assumes an additive model with age (inverse-normal transformed), sex, the top 5 principal genotype components, and genotype platform indicator variables as covariates. Genetic variants identified via GWAS were mapped to genes via Multimarker Annotation of Genomic Annotation (MAGMA). DD heritability and genetic correlations between DD and 110 other traits were determined using Linkage Disequilibrium Score Regression (LDSC) and Local Analysis of [co]Variant Annotation (LAVA). Molecular processes shared across DD and several genetically correlated traits were explored by coupling network propagation of MAGMA-identified genes with Gene Ontology. GWAS-by-subtraction and genomic Structural Equation Modeling (genomic SEM) were used to parse DD genetic architecture from correlated cognitive-related traits (i.e., executive function, educational attainment, intelligence). Phenome-wide association analysis (PheWAS) was conducted to test associations between 1,380 medical diagnoses and DD polygenic scores calculated using an electronic medical records database (N < 69,447).

Results: DD heritability was 9.85%. GWAS identified 14 significant DD loci, many of which were pleiotropic with psychiatric disorders and cognitive phenotypes. Gene-based analyses identified 93 DD candidate genes. Global genetic correlations revealed positive genetic correlations with 45 traits, including externalizing behavior, impulsivity, substance use and disorder, ADHD, antisocial behavior, and chronic pain, and 28 negative genetic correlations with cognitive and socioeconomic variables, anxiety, schizophrenia, bipolar, anorexia nervosa, and obsessive compulsive disorder, among others. The NCAM1-TTC12-ANKK1-DRD2 gene cluster, which is implicated in numerous psychiatric, neurological, and biobehavioral conditions, showed the most extensive pleiotropy between DD with other traits. Network propagation identified shared biological networks with several traits including educational attainment, addiction risk factor and body mass index. Using gSEM to account for DD genetic variance from cognitive-related variables, we found that positive genetic correlations with substance use, impulsivity, externalizing, and other psychiatric disorders persisted. PheWAS revealed 245 medical associations including substance use, mood, anxiety, and bipolar disorders, as well as with many somatic traits.

Conclusions: Our study indicates that DD is influenced by numerous genetic variants shared with other psychiatric and health traits.

Disclosure: Nothing to disclose.

38.2 Cross Species Studies of Heritability of 24-Hour Motor Activity as a Window Into Mood Disorders

Kathleen Merikangas

National Institute of Mental Health, Bethesda, Maryland, United States

Background: There is increasing evidence that dysregulation of motor activity and sleep are key components of mental disorders, particularly bipolar disorder. The use of accelerometers as an objective index of physical activity, sleep and circadian rhythms has provided valuable data to track these domains in real time. However, there is limited information on the genetic and environmental factors that influence these systems and their interactions.

Here we examined the patterns of heritability of accelerometry derived features in a large family study of the affective spectrum, a population sample of adolescent twins and a study of non-human primates in naturalistic settings. The combination of three approaches enabled us to estimate both the heritability and environmental influences in 24-hour patterns of activity as well as their associations with mood disorder subtypes in the human studies.

Methods: The samples including: 403 individuals from the NIMH Family Study of Affective Spectrum Disorder; 624 individuals from the Queensland Twin Study of Adolescent Development; and 989 monkeys in 30 pedigrees including 1310 parent-offspring pairs and 127 siblings and 7826 half siblings from the Oregon primate study. Physical activity was measured with the GeneActiv device worn on the nondominant wrist for 2 weeks in the family and twin studies and the neck-worn triaxial Actical accelerometer was worn for 3 days in the monkey study. The actigraphy data were processed by GGIR 2.2.0 (Migueles et al., 2019) and mMARCH.AC 2.4.0.2 pipeline (Guo et al., 2022) and measures of physical activity, sleep and circadian rhythms were extracted. Functional principal components and Joint and Individual Variance Explained (JIVE) models were employed to characterize the overall structure of the data. Polygenic heritability estimates defined as the proportion of the total phenotypic variance explained by additive genetic (or familial) effects were calculated, adjusted for age and gender using the variance components approach implemented in the SOLAR package. We further examine the genetic correlations with clinical and biologic phenotypes in the family study and twin samples.

Results: The heritability of total activity was 0.39 (SE = 0.21, P = 0.03), relative amplitude (RA) was 0.6 (SE = 0.18, P = 0) an objective measure of rest-activity rhythmicity, 0.36 (SE = 0.17, P = 0.01) and 0.36 (SE = 0.17, P = 0.01) for acrophase. The heritability of sleep midpoint was 0.72 (SE = 0.15, P = 3.4e−05) and sleep efficiency 0.42 (SE = 0.19, P = 0.02). In the twin sample, heritability of total activity was 0.83(0.03,P = 2.1e−35); factors was 0.35 (se = 0.19); sleep midpoint was 0.88(0.02,P = 3.5e−36); acrophase was 0.81(0.03,P = 1.6e−30) and RA was 0.81(0.03,P = 1.6e−30). Comparable heritability estimates were found in the monkey sample: 0.40 (se = 0.24) for total activity count; 0.62 (se = 0.28) for acrophase; and 0.42 (se = 0.26) for RA.

Conclusions: These findings demonstrate the strong familial and likely genetic influence on sleep and motor activity that may comprise risk factors or core features of bipolar disorder. This work highlights the importance of employing objective measures of the core systems underlying clinical phenotypes for psychiatric disorders and the potential of genetic epidemiologic study designs, particularly across species, to inform genetic studies of psychiatric disorders.

Disclosure: Nothing to disclose.

38.3 Dissecting the Phenotypic and Genomic Structure of Normal and Psychopathological Personality in the Genetics Links to Anxiety and Depression or GLAD Study

Gerome Breen

Institute of Psychiatry, Psychology and Neuroscience; King’s College London, London, United Kingdom

Background: Normal and extreme personality traits represent quantitative traits that could inform both treatment selection and development for psychiatric disorders. We investigated normal and psychopathological personality measurements and their genomic associations in a large sample of severe depression and anxiety cases and controls, diagnostic psychopathological traits in a subset of the Genetic Links to Anxiety and Depression and its associated control samples (GLAD + ) within the UK NIHR BioResource. The GLAD+ sample has sample sizes between 14,190 and 20,144 with genome-wide association study (GWAS) data, and both normal and extreme personality measures.

Methods: Personality was measured via two scales, the six factor model of personality HEXACO normal personality scale and the Personality Inventory for DSM-5 (PID-5) (actually PID-5 BF) designed to measure extreme personality relevant to psychiatric disorders. Trait heritabilities and genetic correlations were obtained using GCTA-GREML (genomic-relatedness-based restricted maximum-likelihood) and LD (linkage disequilibrium) score (LDSC) regression. To avoid potential confounding by diagnostic status or known differences such as between sexes, genome-wide association studies were performed in sexes, cases and controls separately and then meta-analysed. We assess genetic correlations between normal and extreme personality, and their genetic correlations with a wide range of psychiatric disorders.

Results: The six HEXACO dimensions obtained GCTA GREML heritabilities as Honesty-Humility: 15% (0.037), Emotionality: 10% (0.038), Extraversion: 16% (0.038), Agreeableness: 9% (0.038), Conscientiousness: 15% (0.038), and Openness to Experience: 14% (0.038). Heritabilities of the five factors in PID-5 were estimated as Negative Affect: 9% (0.027), Antagonism: 11% (0.027), Detachment: 9% (0.027), Disinhibition: 11% (0.026), and Psychoticism: 8% (0.027). LDSC genetic correlations (rg) revealed strong relationships between PID-5 Negative Affect and normal range Neuroticism (rg = 0.77, s.e.=0.33) while HEXACO Honesty-Humility showed a broad range of significant, large negative genetic correlations with extreme personality including Antagonism (rg = −1.00 [capped], s.e.=0.35), Psychoticism (rg = −0.83, s.e.=0.31), Negative Affect (rg = −0.79, s.e.=0.34). Within psychiatric disorders, there were significant, large genetic correlations between PTSD and multiple aspects of personality including Negative Affect (rg = −0.67, s.e.=0.17), Disinhibition (rg = −0.61, s.e.=0.17), and Antagonism (rg = −0.5, s.e.=0.19).

Conclusions: Our study revealed significant heritabilities and genetic correlations between normal and extreme personality traits, and their associations with psychiatric disorders, especially PTSD. These results suggest that personality traits, both normal and extreme, could play a crucial role in understanding the genetic basis of psychiatric disorders, which could be vital for enhancing treatment selection and developing new therapeutic strategies.

Disclosure: Nothing to disclose.

Mini Panel

39. Capturing and Understanding Affective Disruption in Adolescent Affective Disorders: A Step Toward Personalized, Developmentally Informed Clinical Approaches

39.1 Two-Year Trajectories of Anhedonia in Adolescents at Transdiagnostic Risk for Severe Mental Illness: Association With Clinical Symptoms and Brain-Symptom Links

Tina Gupta

University of Pittsburgh, Glen Ellyn, Illinois, United States

Background: Anhedonia emerges during adolescence and is characteristic of severe mental illness (SMI). To understand how anhedonia first emerges, changes with time, and relates with clinical symptoms such as depression and suicidal ideation (SI), there is a need to understand patterns of this symptom’s course reflecting change or stability—and associations with symptoms and neural reward circuitry in adolescents at risk for SMI.

Methods: 113 adolescents (57% girls) at low/high familial risk for developing SMI (schizophrenia/schizoaffective disorders, bipolar disorder, major depression) completed clinical measures at up to five timepoints across two-years and fMRI scanning during a guessing reward task. Functional neuroimaging data were collected using a 3.0 Tesla Siemens Prisma MRI scanner. Preprocessing and fMRI analyses were performed using SPM. ROI (ventral striatum [VS], a key region of reward) to whole brain functional connectivity (FC) analyses were applied using gPPI Interaction analysis (win outcome > loss outcome). Clusters were extracted using the principle eigenvariate method, thresholded at p < 0.001, 20 contiguous voxels. Latent Class Growth Analysis was used to identify anhedonia trajectories using Mplus. Trajectory differences in symptoms, FC of the VS, and brain-behavior associations were examined using multivariate logistical regression and moderation analyses. We predicted three trajectories would emerge--worsening, stable, and decreasing--and they would differ in clinical features, VS FC to regions of both reward valuation and motor and decision-making, and brain-behavior associations.

Results: Three anhedonia trajectories emerged: High and Worsening anhedonia, Moderate and Stable anhedonia, and Low and Decreasing (slope ps = <0.001−0.51). Lower depression severity predicted trajectory membership for the Low and Decreasing anhedonia trajectory (n = 79, Est = −0.30, SE = 0.14, p < 0.05). For the Moderate and Stable anhedonia trajectory (n = 28), higher SI was related to negative VS FC with the paracentral lobule, p = 0.02, 95% CI [0.85, 8.91], coeff = 4.88.

Conclusions: Results highlight anhedonia’s course which may impact symptoms and neural reward circuitry particularly for certain patterns of anhedonia development. Findings of VS FC with the paracentral lobule (implicated in sensorimotor processing) may reflect preoccupation with suicidal thoughts which may lead to disinterest and disengagement with one’s environment.

Disclosure: Nothing to disclose.

39.2 Different Reward Processing Markers Predict Unipolar or Bipolar Anhedonia in Adolescence

Roselinde Kaiser

University of Colorado Boulder, Boulder, Colorado, United States

Background: Reward processing abnormalities are associated with, and predict, symptoms of anhedonic depression in adolescence. However, it remains unclear which (overlapping or distinctive) reward abnormalities predict anhedonia in the context of unipolar versus bipolar clinical presentations. In this longitudinal study, we aimed to distinguish multiple dimensions of reward processing and predict anhedonic depression in the context of unipolar or bipolar clinical presentations.

Methods: Participants (n = 112, ages 13−19, risk-enriched for family history of mood disorders) completed behavioral testing and functional magnetic resonance imaging at baseline, followed by daily diary evaluation of anhedonic depression (in nine 21-day bursts, 1x/3 months) and clinical diagnoses (annually) over a two-year follow-up. Anhedonia was measured with the Mood and Anxiety Symptom Questionnaire. Behavioral tasks included the Probabilistic Reward Task, Two-Armed Bandit, and Instrumental Learning Task. A modified Card Guessing Task was administered during neuroimaging. Reward processing was estimated using reinforcement learning drift diffusion models for behavioral dimensions, and task-modulated general linear models for neural dimensions. Dimensions of reward processing included: decision processing (factor estimated from drift rate/tau across tasks), learning rate (factor estimated from alpha across tasks), response bias (from the Probabilistic Reward Task), and corticostriatal response to reward (functional connectivity in corticostriatal networks in response to reward, Card Guessing Task). Mixed effects analyses tested associations between reward dimensions and future anhedonic depression, and compared such associations between adolescents who developed unipolar versus bipolar diagnoses. Analyses were FDR corrected.

Results: Lower reward decision processing predicted higher anhedonia for adolescents who developed mood disorders [B = −0.09, F = 9.88, q < 0.01], and specifically bipolar disorders [B = -0.16, F = 4.88, q = 0.03]. Higher corticostriatal reward sensitivity also predicted higher bipolar anhedonia, especially in response to life stress [B = 0.68, F = 9.38, q < 0.01]. In contrast, lower reward response bias predicted higher anhedonia specifically for adolescents who developed unipolar disorders [B = −0.19, F = 7.33, q = 0.02]. Lower reward learning rate was a common predictor of bipolar or unipolar anhedonia [B = −0.16, F = 6.25, q = 0.02], although in post hoc analyses, this effect depended on task context.

Conclusions: Adolescents with lower reward decision processing and amplified corticostriatal reward sensitivity reported higher levels of bipolar anhedonia over follow-up, whereas adolescents with blunted reward response bias reported higher unipolar anhedonia over follow-up.

Disclosure: Cingo Health: Founder (Self). Bruin Health: Advisory Board (Self)

39.3 Harnessing Smartphone Sensors to Predict States of Elevated Negative Affect in Adolescents

Christian Webb

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

Background: Disturbance in negative affect (e.g., heightened sadness, anxiety, or irritability) is a core feature of affective disorders. Over 90% of Americans (and 85% of adolescents) own a smartphone, equipped with an array of sensors which can continuously and unobtrusively measure behaviors (e.g., activity levels, location, and phone usage patterns) which may predict increases in negative affect in real-time in individuals’ daily lives.

Methods: Forty-nine individuals with a primary mood or psychotic disorder completed daily emotion surveys for over a year, on average (mean 490 days; total surveys collected = 10,278). At the same time, semi-continuous collection of smartphone accelerometer, GPS location, and screen usage data, along with movement tracking from an actigraphy watch, was conducted for the duration of the study. A range of statistical approaches, including a novel personalized ensemble machine learning algorithm, were compared in their ability to predict states of heightened negative affect.

Results: A personalized ensemble machine learning algorithm outperformed other statistical approaches (including a linear mixed effects model and fully idiographic models), achieving an area under the receiver operating characteristic curve (AUC) of 0.72 (for irritability) −0.79 (for loneliness) in predicting different negative emotions. Smartphone location (GPS) variables were the most predictive features overall. Critically, there was substantial heterogeneity between individuals in the association between different smartphone features and negative emotional states, which highlights the need for a personalized modeling approach.

Conclusions: Findings support the use of smartphones coupled with machine learning to detect states of heightened negative emotions. The ability to predict these states in real-time could inform the development and timely delivery of emotionally beneficial smartphone-delivered interventions which could be automatically triggered via a predictive algorithm.

Disclosure: Nothing to disclose.

Study Group

40. Mental Health Consequences of the Evolving Cannabis Policy Landscape in the US; Findings From the 2024 National Academies of Sciences Engineering and Medicine (NASEM) Report

Ziva Cooper, Yasmin Hurd, Beau Kilmer, Rosalie Pacula, Magdalena Cerda, Dustin Duncan, Kelly Young-Wolff

UCLA Center for Cannabis and Cannabinoids, Jane and Terry Semel Institute for Neuroscience and Human Behavior, Los Angeles, California, United States

Study Group Summary: Cannabis regulation is undergoing rapid changes in the United States and globally. Since 1996, 40 states have legalized cannabis for medical purposes; 24 of these states have also legalized cannabis for non-medical use. Despite widespread state legalization, cannabis remains federally illegal as a Schedule I drug, resulting in a patchwork of policies across the states that regulate the manufacturing, sales, and use of cannabis. With an evolving cannabis marketplace and growing access, cannabis initiation and frequency of use is increasing. Given the known associations between cannabis use and psychopathologies, there is an urgent need to understand how public policies related to cannabis regulation may directly or indirectly impact mental health endpoints relevant to the ACNP audience. In September 2023, the National Academies of Sciences Engineering and Medicine (NASEM) convened an ad hoc committee to review and evaluate the public health implications of differing policies regulating cannabis, identify gaps in knowledge, and develop recommendations to mitigate societal harms and inform policy research. This Study Group brings together 7 of the committee members with diverse areas of expertise; study group Chair and Co-chairs, Dr. Yasmin Hurd and Dr. Ziva Cooper, are full and associate ACNP members, respectively. They will introduce the NASEM report with a brief overview of the neuropsychopharmacology of cannabis and cannabinoids; they will also moderate the panel, facilitating audience engagement and discussion. Dr. Beau Kilmer, drug policy researcher, will describe the impact of legalization on the regulated and illicit cannabis market, prices, diversification of cannabis products (including rising strengths), and consumer behavior. Dr. Rosalie Liccardo Pacula, an economist, will then probe the relevance of state regulations for influencing public health outcomes and the role of industry in countering public health regulations, which ultimately contribute to mental health outcomes. The relationship between cannabis legalization policy and mental health will be highlighted by Dr. Magda Cerdá, expert in social and psychiatric epidemiology. Dr. Dustin Duncan, epidemiologist focused on health equity-based research, will discuss the need for mental health equity within the framework of diverse cannabis policies. Dr. Kelly Young-Wolff, clinical psychologist, will describe the impact of cannabis legalization policies in pregnant individuals, a vulnerable population with significant implications regarding mental health risk. The NASEM report will be published in the fall of 2024, thus very timely for this panel to lead a vibrant, audience engaged discussion related to the report’s conclusions and recommendations with attention given to the mental health impact of changing cannabis policies.

Disclosure: Canopy Growth Corporation: Other Financial or Material Support (Self). True Terpenes: Other Financial or Material Support (Self). Stroz and Bickel: Other Financial or Material Support (Self)

Panel

41. Neurobehavioral Alterations Underlying Avoidance: From Animal Models to Psychiatric Disorders

41.1 Subcortical Mechanisms for Persistent Safety-Seeking Behavior

Justin Moscarello

Texas A and M University, College Station, Texas, United States

Background: Anxiety triggers persistent safety-seeking responses that can easily become maladaptive. Signaled active avoidance (SAA) is a rodent behavioral paradigm that models the connection between anxiety-like threat processing and sustained safety-seeking behavior. In SAA, subjects learn to shuttle across a divided chamber during a warning signal (WS) to prevent an aversive outcome. An interesting feature of AA is that shuttling persists after the offset of the WS, suggesting that a sustained form of avoidance is governed by uncertain/ambiguous (anxiety-like) threats associated with the AA context. We hypothesized that the ventral hippocampus (VH), a limbic region known to integrate aversive and contextual information, regulates the expression of these inter-trial responses (ITRs). Because VH sends an excitatory, nonreciprocal projection to the bed nucleus of the stria terminalis (BNST), an extended-amygdala structure that coordinates reactions to uncertain threat, we further hypothesized that BNST also contributes to ITR expression.

Methods: Male rats (n = 25) received 4 days SAA training prior to an SAA test preceded by muscimol or vehicle infused into VH via chronic cannula. We followed this with an experiment of a similar behavioral design in which, prior to test, CNO was administered to male (n = 10) and female (n = 14) rats expressing GFP or the excitatory hM3Dq receptor in VH CamKII+ principal neurons. Finally, we performed an identical behavioral experiment in which CNO was administered to male (=17) and female (n = 16) rats expressing the inhibitory hM4Di receptor or GFP in BNST neurons. Results were analyzed with ANOVA followed by appropriate post hoc tests.

Results: Muscimol inactivation of VH significantly attenuated ITRs (p < 0.05), whereas CNO excitation of VH principal neurons significantly facilitated ITR expression (p < 0.01). CNO inhibition of BNST also significantly attenuated ITRs (p < 0.01).

Conclusions: These data suggest a model in which a subcortical circuit mechanism including VH and BNST connect uncertain/ambiguous contextual threat to the expression of inter-trial responses, a persistent safety-seeking behavior.

Disclosure: Nothing to disclose.

41.2 The Midline Thalamus Modulates Motivational Aspects of Avoidance Behavior via the Ventral Striatum

Mario Penzo

National Institute of Mental Health, Bethesda, Maryland, United States

Background: Changes to motivational drive are pervasive amongst neuropsychiatric disorders such as depression (decreases motivation or anhedonia) and addiction (heightened motivation for substances of abuse). Understanding the neural mechanisms that underlie motivational processes is thus critical for the development of more effective therapies for these conditions. While changes in local dopamine (DA) concentration within the ventral striatum (VS) has been recently identified as a features of motivational state, the neural circuit mechanisms that account for this process is unknown. The paraventricular thalamus (PVT) is a putative candidate region for driving local VS DA release during motivated behaviors. The PVT makes functional connections onto DA-driving interneurons, and has been implicated in motivated behaviors including reward seeking and threat avoidance.

Methods: To investigate the potential role of PVT-VS DA in avoidance motivation in mice, we utilized an active avoidance task (AA), where subjects learn to shuttle across two compartments during an auditory warning cue to prevent a foot shock. Next, we used fiber photometry and optogenetics to respectively monitor and manipulate the activity of PVT input to the VS while mice performed in our AA task. Lastly, we used statistical methods such as functional linear mixed modeling and reinforcement learning models to probe the relationship between behavior-related variables and neural circuit dynamics.

Results: We found that the development of active avoidance responses was associated with safety outcome-related responses in PVT−VS axon terminals (F11,1062 = 51.80, P < 0.0001, one-way ANOVA, n = 6 mice). Importantly, optogenetic inhibition of these terminals during the safety period following avoidance (but not failure) trials significantly impaired task performance (F1,144 = 24.28, P < 0.0001, Two-way, repeated measures ANOVA, n = 14 mice/group). Ongoing research suggests that PVT projections to the VS shape the motivation to engage in avoidance behavior by controlling safety-related fluctuations in DA.

Conclusions: Our findings highlight the importance of PVT-VS circuitry in regulating motivational processes and suggest a potential target for therapeutic interventions aimed at modulating motivational drive in conditions such as depression and addiction.

Disclosure: Nothing to disclose.

41.3 Abstract Not Included

41.4 Sensitivity to Endogenous Threat and Avoidance One Year Later in a Mixed, Internalizing Sample: An Event-Related Potential Study

Annmarie MacNamara

Texas A and M University, College Station, Texas, United States

Background: Errors can be costly and potentially even life-threatening. Therefore, errors can be thought of as endogenous threats. Avoidance of endogenous threats may play a key role in internalizing psychopathology such as anxiety. Indeed, an individual’s sensitivity to endogenous threats, especially when accompanied by future avoidance of these threats, might prompt a broader repertoire of avoidant behaviors (i.e., in everyday life) aimed at reducing unpleasant, internal sensations. EEG event-related potentials, the error positivity (Pe) and error related negativity (ERN), can be used to measure the degree to which individuals perceive errors as threatening. Here, we investigated whether the Pe and ERN would predict increases in avoidance across one year, in a mixed internalizing sample.

Methods: Participants (N = 36; all female, M = 23 years; SD = 6.52) with a range of internalizing psychopathology self-reported on avoidance of internal experiences and completed an EEG during a computerized (flanker) task designed to elicit errors. These measures were collected at baseline (Time 1) and approximately one year later (Time 2). A factor measuring avoidance at each of Time 1 and Time 2 was derived from items on the Panic Disorder Severity Scale and the Contrast Avoidance Questionnaire-Worry. The Time 1 Pe and Time 1 ERN were assessed as predictors of Time 2 avoidance controlling for Time 1 avoidance. Moderation analyses were also performed.

Results: Individuals with increased elaborative processing of errors at Time 1 (Pe) showed greater increases in avoidance from Time 1 to Time 2 (β = 0.412, p = 0.009). Moderation analyses (β = 0.371, p = 0.016) revealed that this association was evident only for individuals with attenuations in early error processing (ERN) from Time 1 to Time 2 (t = 3.70, p < 0.001). Attenuated Time 2 ERNs were also associated cross-sectionally with greater Time 2 avoidance (r = 0.379, p = 0.023).

Conclusions: Heightened sensitivity to endogenous threat (Pe) may lead to the emergence of more avoidant thoughts and behaviors, particularly for individuals who show a shift towards more blunted early response to endogenous threat over the same time period. Results highlight a novel psychophysiological risk factor and potential mechanism for worsening avoidance in the internalizing disorders over a one-year period.

Disclosure: Nothing to disclose.

Panel

42. Molecular, Cellular and Behavioral Diversity of Lateral Septum Function

42.1 Septo-Hypothalamic Regulation of Binge-Like Alcohol Consumption by the Nociceptin System

Thomas Kash

University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States

Background: High intensity alcohol drinking during binge episodes overwhelmingly contributes to the socioeconomic burden created by Alcohol Use Disorders (AUD). Novel interventions are needed to add to the current therapeutic toolkit and nociceptin receptor (NOP) antagonists have shown promise in reducing heavy drinking days in patients with an AUD. However, an endogenous locus of nociceptin peptide and discrete sites of NOP action underlying this effect remains understudied.

Methods: We utilized a combination of systems neuroscience approaches to explore the role of the nociceptin system in both male and female mice. Viral approaches were used to manipulate and monitor neuronal functional in mice during a range of behaviors. Pnoc-cre mice (N = 153) were used for fiber photometry, electrophysiology, chemogenetic, optogenetic experiments, and circuit mapping studies. In situ hybridization and genetic deletion studies used N = 118 NOPfl/fl mice.

Results: Pnoc-expressing neurons of the LS (LSPnoc) are tuned to stimuli associated with negative valence and display increased excitability during withdrawal from binge-like alcohol drinking. LSPnoc activation was found to have aversive qualities and also potentiates binge-like drinking behavior, suggesting a convergence of circuitry that promotes aversion and drives alcohol consumption. Viral mediated tracing and functional assessment of LSPnoc projection fields revealed GABAergic synapses locally within the LS, and downstream within the lateral hypothalamus (LH) and supramammillary nucleus (SuM). Genetic deletion of NOP from the LS attenuated binge-like alcohol intake in male mice while NOP deletion from the LH and SuM decrease alcohol intake in females.

Conclusions: Together, these finding demonstrate an endogenous population of nociceptin-expressing neurons that contributes to alcohol consumption and identifies sex-dependent modulation of alcohol drinking

by NOP.

Disclosure: Nothing to disclose.

42.2 The Lateral Septum Regulates State-Dependent Inhibition of Fear Memory Across the Estrous Cycle

Elizabeth Lucas

University of Alabama at Birmingham, Birmingham, Alabama, United States

Background: Clinical research implicates cycling reproductive hormones as a primary mechanism driving enhanced female susceptibility to post-traumatic stress disorder (PTSD). Hormone levels at the time of trauma are correlated with subsequent development of PTSD, and women report less severe symptoms when estradiol peaks midcycle. We tested the hypothesis that cyclic hormone fluctuations contribute to state-dependent memory, using cued fear conditioning in adult male and female C57Bl/6J mice as a translational preclinical paradigm.

Methods: Female behavior was targeted to high (proestrus, P) or low (diestrus, D) hormone states with fear conditioning and memory recall occurring in the same or opposite estrous stage, resulting in 5 experimental groups (n = 6−14/group): male, P > P, P > D, D > D, D > P. We quantified regional c-fos expression following fear conditioning and recall (n = 4−16/group) and used graph theory methods to construct neural networks. Targeting our focus to the lateral septum (LS), we used chemogenetics to activate LS neurons during behavior (n = 5/group) and performed single nucleus RNA sequencing (n = 4/group) followed by multiplex fluorescent in situ hybridization (mFISH; n = 4/group) to identify the LS fear memory ensemble.

Results: Fear memory recall was indistinguishable between males and P > P females, whereas all other female groups exhibited enhanced cued-induced freezing (p < 0.001). Graph theory-derived analyses revealed the rostral LS as a hub region in P females. LS c-fos expression was increased in P females after conditioning (p = 0.03) and recall (p < 0.001). To assess the causal link between LS activity and state-dependent memory, we used the excitatory chemogenetic construct hM3D to activate LS neurons during recall in P > D females, which robustly decreased cued-induced freezing compared to vehicle controls (p = 0.007). LS sequencing revealed 52 transcriptionally distinct cellular clusters. Two neuronal clusters exhibited immediate early gene expression in conditioned P females: somatostatin-neurotensin (SST-NTS) and corticotropin-releasing hormone receptor 2 (Crhr2). mFISH revealed state-dependent engagement of LS SST-NTS (p = 0.02), but not Crhr2, neurons.

Conclusions: The LS contains a unique neuronal ensemble solely engaged in high hormone states to protect against the overexpression of fear memory in females.

Disclosure: Nothing to disclose.

42.3 How Developmental Origin and Local Cell Environment Shape Neural Diversity in the Septum

Corey Harwell

UCSF, San Francisco, California, United States

Background: The lateral septum (LS) is a nucleus in the ventral forebrain that modulates complex social and affective behaviors. Several distinct neuronal types have been described in the LS; however, the full extent of this cellular and molecular diversity remains unclear. We address this gap by profiling the transcriptional identity of mature LS neurons originating from two progenitor lineages defined by their anatomical location and expression of the transcription factor Nkx2.1

Methods: A total of six mice (3 male and 3 female) across three replicates were used for this study. Nkx2.1-Cre mice crossed with the conditional Sun1GFP allele were 35 days old at the time of sacrifice. Nuclei double positive for GFP and Ruby Dye were separated from nuclei only positive Ruby dye using a SONY SH800S FAC sorter. After collecting a minimum of 50,000 events, the nuclei were loaded onto a 10X Chromium Single Cell 3’ chip (v3 chemistry) at the specified concentration. This was followed by adherence to the standard 10X Chromium Single Cell 3’ protocols (v3 chemistry). Finally, the resultant libraries were sequenced using a NovaSeq 6000 S4 flow cell, achieving a read depth of approximately 50,000 reads per nucleus.

Results: Subsequently we performed single nucleus RNA-seq (snRNA-seq) on each sorted population and obtained a total of 25,173 cells that expressed an average of 3,014 unique genes. Using principal component analysis (PCA) we constructed a neighborhood graph which allowed us to visualize the data in two dimensions. Our sequencing approach enabled us to distinguish between GFP-positive cells, indicating a history of Nkx2.1 expression (Nkx2.1 + ) and GFP-negative cells that had never expressed Nkx2.1 (Nkx2.1-). We describe 12 molecularly distinct subtypes of LS neurons that fall into two main groups: those with a history of Nkx2.1 expression and those without. We discovered that LS neurons from the Nkx2.1 lineage share an enrichment of select cell adhesion and communication molecules.

Conclusions: In this study, we resolve the multimodal features and organizational principles of lateral septal neurons (LSNs) by detailing their transcriptome, developmental origin, synaptic partners, and physiological properties. We identified 12 molecularly distinct LSNs types that are shaped by both their developmental lineage and anatomical location.

Disclosure: Nothing to disclose.

42.4 Conserved and Divergent Molecular Signatures Associated With Neuropsychiatric Disorders Across Species in Spatially Defined Cellular Populations of the Lateral Septum

Keri Martinowich

Lieber Institute for Brain Development, Baltimore, Maryland, United States

Background: The lateral septum (LS) is a predominantly GABAergic region in the basal forebrain that controls behaviors related to motivation, affect, social cognition and reward, which are often dysregulated in psychiatric disorders. Single nucleus RNA-sequencing (snRNA-seq) studies in mice have begun to identify how molecularly and anatomically heterogeneous LS neuronal populations confer these diverse behavioral functions. However, understanding the molecular composition of these populations in human LS is essential for understanding their function in psychiatric disorders.

Methods: We dissected fresh frozen postmortem human brain tissue (n = 5 samples, both male and female neurotypical donors) to construct spatio-molecular maps of human LS. We confirmed LS inclusion on the tissue blocks using RNAscope for marker gene TRPC4. Tissue blocks were scored to isolate the LS, and cryosections were collected for downstream snRNA-seq or placement on Visium or Xenium slides. For snRNA-seq nuclei were labeled with PI and NeuN to enrich samples for neuronal nuclei, and molecular profiles were generated with the 10X Genomics Chromium platform. The 10X Genomics Visium and Xenium platforms were used to construct spatial transcriptomics maps. We integrated the human data with publicly available mouse LS snRNA-seq datasets to assess cell type conservation, identify regional enrichment across species, and quantify enrichment of risk for psychiatric disorders across cell types and spatial domains.

Results: Four transcriptionally distinct neuronal populations in human LS were identified based on TRPC4 and DGKG expression. Differential expression analysis revealed a population marked by OPRM1, and we identified a novel, human-specific marker gene, FREM2. We further identified several transcriptionally distinct neuronal subtypes that mapped to discrete positions along the anterior-posterior and dorsal-ventral axes of the human LS, and which varied according to their genetic risk enrichment for both psychiatric disorders and conservation with subtypes linked to behavior in functional mouse snRNA-seq datasets.

Conclusions: LS cell types are transcriptionally distinct and spatially organized in humans with subpopulations that are enriched for psychiatric disorder risk, and associated with behaviors related to mood, addiction and social behaviors.

Disclosure: Nothing to disclose.

Panel

43. Expectations, Contextual, and Placebo Effects: Brain Mechanisms and Tangible Applications

43.1 Abstract Not Included

43.2 Neural Circuit Mechanisms of Conditioned Placebo-Like Analgesia in Mice

Matthew Banghart

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

Background: Placebo analgesia is a prominent form of cognitive pain modulation. Functional imaging studies in humans have led to models of the underlying neural circuit mechanisms, but these have not been rigorously tested with loss-of-function manipulations. This presentation describes a mouse model of placebo analgesia and its use with targeted experimental manipulations and recordings of neural activity to reveal circuit mechanisms of placebo-like analgesia.

Methods: Male and female mice were contextually conditioned with either morphine or saline prior to assessment of pain thresholds on a hot plate. The placebo condition involved administration of morphine in the saline context. Both nociception and pain affect were assessed. Genetically and anatomically-targeted chemogenetic loss-of-function manipulations were used to probe the role of brain regions and projection pathways implicated in human studies. Fiber photometry using genetically encoded sensors for Ca2+ and opioid neuropeptides were used to measure the temporal dynamics of neural and neurochemical signaling. Photopharmacology with a light-activated opioid antagonist (PhNX) was used to probe the necessity of endogenous opioid neuropeptide signaling in target brain areas.

Results: Only male mice demonstrated placebo-like response (p = 0.002, n = 10M, 8F, M-W test). The placebo effect ranged from 40-70% of the morphine effect. In addition to antinociception, conditioning attenuated non-conditioned metrics of pain affect (p < 0.0001, F(5,54) = 12.81, 1way ANOVA, n = 10 placebo, n = 10 ctl; p < 0.0001, Sidak’s). Chemogenetic silencing identified multiple neural pathways that support placebo-like analgesia (p = 0.006, < 0.001, 0.0002, 0.003, n = 8-12 HM4Di and n = 8-12 mCh ctl each, t-test or M-W test). Fiber photometry revealed endogenous opioid release in the vlPAG on placebo trials (p = 0.02, n = 8, sal vs. placebo, Wilcoxon). Local photoantagonism of opioid receptors revealed a causal role for endogenous opioid signaling in the vlPAG and cortex (p = 0.02, 0.04, 0.001, n = 8−9 per site, sal vs PhNX, Wilcoxon).

Conclusions: Our mouse model of placebo analgesia shares similarities to human subjects, including the sex difference. The mechanistic insights gleamed support a model in which conditioning increases noxious stimulus-evoked endogenous opioid release in the vlPAG to drive analgesia.

Disclosure: Nothing to disclose.

43.3 Abstract Not Included

43.4 Naloxone Blocks Observationally Induced Placebo Effects

Luana Colloca

University of Maryland Baltimore, Baltimore, Maryland, United States

Background: Observing others, which begins at birth, is a powerful way to gain information. It influences a variety of behaviors including pain and social threat. Using the opioid antagonist naloxone and in vivo receptor binding of μ-opioid receptors, a series of neuropharmacological studies have corroborated and confirmed the notion that the opioid system is involved in the formation of placebo (analgesic) effects. Based on previous psychoneurobiological studies and the strong neural results supporting our hypotheses, we determined the role of endogenous opioids in observationally induced hypoalgesia by using the opioid antagonist naloxone.

Methods: This project was conducted in 68 healthy study participants from Dec 2021 to August 2023 at University of Maryland School of Medicine, Baltimore, USA. The study design was a 2 (group: naloxone vs. saline) X 2 (condition: placebo vs. control) double-blind mixed experimental design with group as the between-subjects factor and condition as the within-subjects factor (NCT03897998). Each participant was randomly assigned to either the naloxone or the saline group by the Investigational Drug Service (IDS). Investigators, staff, and participants were blinded to the intranasal naloxone vs saline treatment options. The study paradigm started with the observation phase, consisting of two runs of 12 trials each (treatment and control run). The runs were randomized across participants to account for sequence effects. After the observation phase, participants were asked to rate their outcomes.

Results: We found a main effect of treatment (F1,804.04 = 6.3, p = 0.012): those in the saline group had greater placebo effects (6.4 ± 1.6) than those who received naloxone who showed no placebo effects (0.3 ± 1.6), independently of sex (F1,803.27 = 0.43, p = 0.51) and race (F1,803.27 = 0.59, p = 0.079). Placebo effects were significantly correlated with expectations (r = 0.83, p < 0.001). However, naloxone did not block expectations (F1,788.99 = 2.3, p = 0.13), suggesting that observationally induced effects are also independent of expectations.

Conclusions: These findings indicate that observationally induced placebo effects are opioid-mediated. Given that humans are highly social organisms and that our behaviors are affected by social learning, this finding is helpful to enhance endogenous analgesic responses.

Disclosure: Nothing to disclose.

Panel

44. Cross-Species Perspectives on the Neurodevelopment of Reward Processing in Adolescence

44.1 Reduction of Grin2a in Adolescent Dopamine Neurons Disrupts Prediction Error Signaling in Adult Rats

Michelle Kielhold

Oregon Health and Science University, Portland, Oregon, United States

Background: Genetic variants resulting in a loss of function of the GluN2A subunit of the NMDA receptor have been implicated in neurodevelopmental disorders linked to disruptions in dopaminergic signaling during prediction error such as autism spectrum disorder (ASD) and schizophrenia. To better understand how genetically driven reductions in GluN2A could contribute to disrupted striatal dopamine (DA) release, we assessed the behavioral and physiological consequences of a reduction of Grin2a in dopamine neurons using a CRISPR-generated virus.

Methods: DA neuron-specific loss of function of Grin2a was modeled in TH-Cre+ Long Evans rats using Cre-inducible AAVs. Juvenile animals were infused with either a control or Grin2a conditional knockout (Grin2a cKO) virus and GRABDA into the ventral tegmental area (VTA). We used fiber photometry to explore causal mechanisms linking Grin2a cKO with aberrant DA release during prediction error by measuring accumbal DA in adult rats during a flexible contingency learning task. In this task, animals associated conditioned stimuli (CS) with positive (reward) and negative (shock) outcomes under shifting contingencies. Behavior was assessed by calculating the likelihood of food trough entry during either CS and GRABDA fluorescent response was determined by averaging z-scores across two, 2s periods: CS and outcome onset.

Results: During initial association learning, controls exhibited the expected phasic DA response to the reward- and punishment-predicting CS, however, the Grin2a cKOs had a reduced response to the reward-predicting CS (p < 0.0001) and increased response to the punishment-predicting CS (p = 0.0015). Phasic responses to the actual reward or shock were similar between groups. After the contingencies switched, Grin2a cKOs were less likely than controls to approach the food trough during the reward-predicting CS (p = 0.0104) and had lower DA response to CS onset (p < 0.0001) and reward (p = 0.0164).

Conclusions: We find that Grin2a cKO in DA neurons during adolescence impairs DA prediction error signaling in adults. The Grin2a cKO animals had DA disruptions during initial association learning and after the contingency switch, consistent with deficits in prediction error encoding. These disruptions may have ultimately contributed to behavioral differences observed in the Grin2a cKOs.

Disclosure: Nothing to disclose.

44.2 Adolescence as a Critical Period in the Maturation and Stabilization of Decision-Making Circuits

Stephanie Groman

University of Chicago, Chicago, Illinois, United States

Background: The neurodevelopmental changes that occur during adolescence may underlie the improvements in decision-making functions that co-occur during this developmental period. The biology that mediates these age-related improvements in decision-making are not known, but evidence indicates a key role of the orbitofrontal cortex (OFC), amygdala, and nucleus accumbens. The current study sought to identify new protein targets that regulate the maturation and stabilization of decision-making circuits.

Methods: Decision-making functions were assessed in male and female rats (N = 37) in a three-choice, probabilistic reversal learning (PRL) task at specific adolescent ages (P30 = 10; P50 = 10; P70 = 7; P90 = 10). Rats were then euthanized and tissue collected from the OFC, amygdala, and nucleus accumbens for proteomic analyses. We used an FDR-corrected, cross-correlation analysis to identify protein targets whose abundance was correlated (1) with age, (2) with performance in the PRL task, and (3) with the specific reinforcement-learning process that improves during adolescence. This analysis identified perineuronal nets (PNNs) as a potential target and so in a second experiment, we used immunohistochemistry to quantify the density of PNNs in the OFC under typical and atypical adolescent developmental conditions.

Results: Decision-making improved as a function of age and computational analyses revealed that this was due to an improvement in the ability of rats to effectively use positive outcomes. Mass spectrometry identified ~3000 proteins in each of the brain regions, but only 15 proteins met out criterion. Notably, five proteins are involved in perineuronal nets (PNNs) which are an extracellular matrix structure involved in the stabilization of circuits. The density of PNNs in the OFC increased throughout adolescence and intermittent access to alcohol led to the precocial emergence of OFC PNNs and premature stabilization of decision-making functions.

Conclusions: These data indicate that adolescence is an important developmental period in the maturation of reinforcement-learning functions that guide decision-making. Our findings suggest that PNNs play a critical role in the stabilization of decision-making circuits and ongoing work is investigating how PNNs may regulate circuit-specific encoding of reinforcement-learning mechanisms.

Disclosure: Nothing to disclose.

44.3 Abstract Not Included

44.4 Distinct Prefrontal-Striatal-Amygdala and Cinguloopercular Brain Networks Mediate Exploration to Maximize Rewards Versus Avoid Losses in Human Adolescents

Jeremy Hogeveen

University of New Mexico, Albuquerque, New Mexico, United States

Background: Adolescence is crucial for exploring new rewards, learning about the environment, and shaping our adult preferences. However, maladaptive exploratory behaviors can increase risk for behavioral problems and consequences. Limited research exists on the neurocomputational mechanisms balancing reward exploitation and exploration in adolescents. We integrate behavioral assays, computational modeling, and functional MRI to study exploratory decision making in reward versus loss contexts, using a translational approach linking human and nonhuman primate studies.

Methods: Adolescents and young adults completed a 3-armed bandit task during fMRI, which was used to assess reward exploitation and novelty-directed exploration. The task manipulated reinforcement valence to separately model decision-making for gains versus avoiding losses. A computational model generated estimates of exploration and exploitation values in win/loss contexts. We also collected a task-based exploration assay and a battery of other clinical and behavioral measures.

Results: Decision-making shifted in win versus loss contexts. Participants weighted the immediate value of exploiting familiar options more in win than in loss contexts. Conversely, they weighted the future value of exploring novel options more in loss than in win contexts. Exploration value in win related to age, with older participants engaging more effectively in directed exploration to earn rewards. Exploration in win contexts was encoded in the prefrontal cortex, amygdala, ventral and dorsal striatum, with prefrontal responses increasing with age. Exploration in loss contexts was encoded in insular and anterior cingulate regions and was not age-associated.

Conclusions: Prefrontal, striatal, and amygdala regions uniquely encoded the value of exploring something new when it was associated with the possibility of gaining rewards. The current findings suggest this mechanism matures from adolescence to young adulthood. In contrast, cinguloopercular and salience networks encoded the relative future value of novel choice options in loss contexts, which increased the motivation to explore across the current age range. These data are the first to parcellate the neurocomputational bases of explore-exploit decision-making in win versus loss environments in human adolescent participants.

Disclosure: Nothing to disclose.

Panel

45. Using Pre-Clinical Animal Models to Understand Mechanisms Underlying Neuro-Plastic and Behavioral Effects of Brain Stimulation

45.1 A Prefrontal-Insular Circuit Mediates the Effects of Intermittent Theta Burst Stimulation on Motivated Avoidance Behavior

Conor Liston

Weill Cornell Medical College, New York, New York, United States

Background: Converging evidence suggests that transcranial magnetic stimulation (TMS) may act by enhancing plasticity locally within the prefrontal cortical (PFC) target or through network-level effects on downstream areas, but establishing causal mechanisms in clinical studies has been challenging.

Methods: We validated an optogenetic model of accelerated intermittent theta burst stimulation (aiTBS) in C57/BL6j mice expressing an ultrafast opsin (hSyn/Chronos or eGFP ctrl) in prelimbic (PL) neurons, tested on the forced swim test (FST) after 30 aiTBS sessions. SHIELD tissue clearing, c-Fos immunolabeling, and light sheet imaging were used to map whole-brain aiTBS responses. Inhibitory hM4Di DREADDs were used to inactivate downstream insular responses during aiTBS and test the necessity of this circuit in driving behavioral effects.

Results: 30 sessions of aiTBS targeting the PL-PFC increased motivated avoidance behavior (t = 3.34, P = 0.004, N = 9−10/grp vs. eGFP ctrl) and increased PFC postsynaptic spine density (P < 0.0001 by LME). Whole-brain c-Fos mapping revealed iTBS-evoked network activity the insula, orbitofrontal, anterior cingulate, amygdala, and hypothalamus (t > 4.0, FDR Q < 0.05, N = 8−10/grp) with the most pronounced effects in insular cortex, which was confirmed by photometry. DREADD inhibition of insula during iTBS blocked effects on FST behavior (F(2,57) = 6.1, P = 0.003, N = 20/group).

Conclusions: Together, these results establish show that aiTBS acts through both local effects on synapse density in PL projection neurons, as well as downstream network effects. aiTBS-induced activation of a PFC-insular circuit is required for supporting antidepressant-like behavioral effects.

Disclosure: Delix Therapeutics: Advisory Board (Self). Brainify.AI: Advisory Board (Self). Janssen Neuroscience: Advisory Board (Self)

45.2 A Cell-Type Specific Mechanism Underlying the Antidepressant Effect of Transcranial Magnetic Stimulation

Scott Wilke

Semel Institute for Neuroscience and Human Behavior at UCLA, Los Angeles, California, United States

Background: Repetitive transcranial magnetic stimulation (rTMS) is a non-invasive method for clinical brain stimulation. When employed to treat depression and other neuropsychiatric disorders, rTMS pulse sequences are thought to drive lasting changes in dysfunctional brain networks. However, we still lack a detailed mechanistic understanding for how rTMS generates therapeutic effects. This is in large part due to a lack of preclinical models with strong face validity for how rTMS is deployed in the clinic.

Methods: We developed a novel preclinical model of rTMS, enabling us to deliver highly focal clinical treatment protocols in awake, head-fixed mice. We applied an accelerated intermittent theta burst stimulation (aiTBS) protocol (1800 pulses x 10 sessions; 1 day) in a mouse model of chronic stress (chronic corticosterone in drinking water x14 days). All experiments used statistical methods to compare the effect of aiTBS versus a simultaneously performed sham stimulation control. To study the effect of aiTBS on dorsomedial prefrontal cortex (dmPFC) circuits, we used viral tools to target intratelencephalic (IT) and pyramidal tract (PT) projection neurons. Using this approach, we applied cell-type specific photometry, dendritic spine imaging, and chemogenetics to discover how aiTBS drives rapid antidepressant-like effects using mice of both sexes.

Results: Applying aiTBS to dmPFC rapidly reversed behavioral deficits associated with chronic stress (n = 10−11 mice; p < 0.01). Both IT and PT projection classes were activated during aiTBS, but only IT neurons exhibited persistently increased activity during stimulation and subsequent depression-related behavior (n = 5−8 mice; p < 0.05). Moreover, aiTBS reversed stress-related loss of dendritic spines on IT, but not PT neurons (n = 24−36 dendrites; p < 0.05). Finally, chemogenetic inhibition of dmPFC IT neurons during rTMS blocked the antidepressant-like behavioral effect of aiTBS (n = 5−7 mice; p < 0.05).

Conclusions: Using our novel preclinical model of rTMS, we discovered that aiTBS can selectively drive cell type specific structural and functional plasticity in dmPFC IT neurons. rTMS-driven activation of IT neurons is required for the antidepressant-like behavioral effect of aiTBS. Similar cell type and circuit specific changes may underlie the therapeutic effects of clinical rTMS.

Disclosure: Nothing to disclose.

45.3 Understanding How Stimulation Parameters of Theta Burst Stimulation Affects Neural Plasticity of Single Neurons in Rat Medial Prefrontal Cortex

Dhakshin Ramanathan

San Diego VA Medical Center, UCSD, San Diego, California, United States

Background: Theta Burst Stimulation (TBS) is a form of patterned neural stimulation that was first discovered as a robust method of inducing long-term-plasticity (LTP) in hippocampal slices. TBS targeting prefrontal cortex is currently FDA-cleared for depression treatments in humans. However, the degree to which this form of patterned stimulation results in plasticity of neuronal responses in prefrontal cortex in vivo is not well understood.

Methods: In this study, we combined electrical stimulation of prefrontal cortex with optical imaging of single neurons (using an implanted 1-photon miniscope). In one group of animals (n = 4 female, implanted with miniscope, ~800 neurons at present), we characterized the effects of theta-burst-stimulation on activity of neurons before, during and after stimulation. We also compared/tested three different types of TBS: continuous theta burst stimulation (cTBS: 40 seconds of TBS; iTBS, which consists of 2 seconds of TBS followed by an 8 second pause, repeated for 20 cycles; experimental TBS, consisting of 2 seconds of TBS followed by a 40 s puase, repeated for 20 cycles).

Results: 1) Findings

1. Using the standard FDA-cleared iTBS paradigm, about 38% of neurons were activated and 29% neurons were suppressed during stimulation. Neurons that were activated during stimualtion on average showed a long-term increase in activity measured up to 40 min after stimulation (mean Z-score of 0.45 + /- 0.02 above baseline measured at 40 min post stimulation). Neurons that were suppressed during stimulation showed a significant reduction in activity (mean z-score of -0.5 + /- 0.02 at 40 min post stimulation).

2) Finding 2. Continuous theta burst stimulation (cTBS, with no pause between theta bursts) resulted in a higher proportion units being suppressed during stimulation compared to iTBS and an overall significant reduction in plasticity observed for both activated and suppressed neurons (rmANOVA, stim type x time, p < 0.01). The experimental TBS resulted in a higher proportion of cells being activated, and overall greater levels of plasticity in both activated and suppressed cells compared to the standard iTBS (repeated measures stim type x time, ANOVA, p < 0.01).

Conclusions: We show how TBS parameters drives plasticity of single neurons in PFC that may give insight into novel TBS patterns with greater clinical efficacy.

Disclosure: Nothing to disclose.

45.4 Effects of Chronic rTMS on Neurotransmitter Profiles in Vivo

Jennifer Rodger

University of Western Australia, Crawley, Australia

Background: Repetitive transcranial magnetic stimulation (rTMS) has shown therapeutic effects for a range of psychiatric conditions however, brain mechanisms remain poorly understood. We aimed to determine whether chronic rTMS affects excitatory and inhibitory neurons within the cortex, providing a possible mechanism for alterations to subcortical neuromodulatory systems.

Methods: Three cohorts of adult male and female C57Bl6j mice underwent 14 consecutive days of 10Hz low intensity (LI)-rTMS or sham stimulation using custom miniaturised coils. After the final stimulation session, brain samples were collected: (1) after 15 min for liquid chromatography-tandem mass spectrometry to quantify DA, DOPAC, HVA, 5-HT, 5-HIAA, glutamate and GABA; (2) after 15 min for western blot to assess the expression key metabolic enzymes, tyrosine hydroxylase (TH), tryptophan hydroxylase 2 (TPH2) and monoamine oxidase A (MAO-A); (3) after 100 min to investigate the co-expression of c-Fos with excitatory (CaMKII) and inhibitory (GAD67) neuronal markers in the cortex beneath the coil.

Results: Chronic LI-rTMS reduced the degradation of DA and 5-HT in the striatum (p = 0.008) and hippocampus (p = 0.008) respectively. In the hippocampus, TPH2 expression was increased by chronic LI-rTMS (p = 0.015) but there was no change in TH. LI-rTMS increased glutamate (p = 0.033) but not GABA within the cortex. Moreover, the density of cells co-expressing c-Fos and GAD67 was decreased in cortical regions adjacent to or directly below the maximally induced electric field of the LI-rTMS coil (various regions p < 0.05).

Conclusions: Changes to 5-HT degradation may be related to altered expression of enzymes involved in its synthesis, but the mechanism of change to DA turnover remains unclear. The changes in subcortical neurotransmitter biochemistry were associated with changes in cortical neuron activation: we found an increase in glutamate within the cortex following chronic LI-rTMS, and a consistent decrease in the activity of inhibitory neurons within layer 1 of the cortical regions proximal to the maximally induced electric field. Our results suggest that reduced inhibition within the cortex following chronic LI-rTMS may indirectly alter the neurochemistry of hippocampal serotonergic and striatal dopaminergic circuits, potentially contributing to therapeutic effects.

Disclosure: Nothing to disclose.

Mini Panel

46. The Missing Piece in the Endocannabinoid Puzzle: Sex Regulation of Endocannabinoids Signaling in PTSD, Alcohol and Cannabis Use Disorders

46.1 Abstract Not Included

46.2 Sex Differences in Endocannabinoid Response to Stress in Individuals With Cannabis Use Disorder

Erin Martin

Medical University of South Carolina, Charleston, South Carolina, United States

Background: Stress contributes to relapse in cannabis use disorder (CUD) differentially by sex, and sex differences in subjective and hypothalamic−pituitary−adrenal (HPA) responses to stress in individuals with CUD have been shown. The endocannabinoid (eCB) system modulates HPA response to stress, is dysregulated in CUD, and is sexually dimorphic. The present study evaluated sex differences in eCB response to stress in individuals with CUD to inform pharmacotherapeutics development for stress-induced relapse to cannabis.

Methods: Non-treatment-seeking individuals with CUD (n = 17; 10 male, 7 female) were enrolled in a six-day laboratory study. Participants completed the Trier Social Stress Test (TSST) on study day six after three days of abstinence from cannabis. Subjective stress and desire to use cannabis, as well as plasma samples for cortisol and eCB analyses, were collected prior to and at four time points (0, 10, 30, 60 min) following the TSST.

Results: Both sexes reported increased subjective stress following the TSST relative to baseline (Males: 0-min: p < 0.01; 10-min: p < 0.01; Females: 0-min: p < 0.01; 10-min: p < 0.01). Females reported increased craving (60-min: p = 0.03) and difficulty resisting cannabis if offered (-20-min: p < 0.01; 0-min: p = 0.03; 30-min: p < 0.05; 60-min: p = 0.01) relative to males. Males, but not females, showed increased plasma cortisol following the TSST relative to baseline (0-min: p < 0.01; 10-min: p < 0.01; 30-min: p < 0.01). Similarly, males, but not females, showed elevations in the eCBs AEA, 2-AG, LEA, and DEA at all time points following the TSST relative to baseline (all p’s < 0.05); similar elevations were evident for males at the 0-, 10-, and 30-min time points for PEA and the 10- and 30-min time points for OEA (all p’s < 0.05).

Conclusions: While individuals with CUD report increased stress following the TSST regardless of sex, females report increased desire to use cannabis and males show increased plasma cortisol and eCB tone. Blunted physiological responses implicated in stress recovery in females with CUD may underlie increased risk of stress-induced relapse and present a potential target for pharmacotherapeutic agents.

Disclosure: Nothing to disclose.

46.3 The Impact of Sex and Sex Hormones on Endocannabinoid Signaling in Response to Acute Stress and in Post-Traumatic Stress Disorder

Primavera Spagnolo

Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, United States

Background: The endocannabinoid (eCB) system is an integral regulator of stress and fear responses and dysregulations in this system have been implicated in various stress-related psychiatric disorders, including depression and post-traumatic stress disorder (PTSD). However, increasing evidence of sex-differences in eCB signaling, partly due to estradiol regulation of this pathway in females, raises the question of whether this system modulates stress and fear responses in a sex-specific manner. Here, we conducted two studies aimed at investigating the impact of acute stress and estradiol withdrawal on peripherical eCB levels (study 1), and at examining sex-differences in levels of eCBs and pro-inflammatory markers in men and women with PTSD and non-psychiatric controls (study 2).

Methods: In study 1, we obtained serum samples and carried out a secondary analysis in 26 healthy premenopausal women (age 21-45 years) [NCT03047330], who completed a stress-evoking sleep fragmentation (SF) procedure lasting 3 nights, preceded by 2 nights of unfragmented sleep (US). Participants underwent this paradigm during the mid-follicular phase (estrogenized) and after pharmacological estradiol suppression (hypoestrogenized). Serum samples collected after the second night of UF and the first night of SF during each estrogen condition were analyzed to assess eCBs levels (AEA, 2-AG, 2-OG, OEA, PEA). In study 2, we obtained serum samples from individuals, aged 18- 45, with a diagnosis of PTSD based on ICD-10 criteria (n = 88; 42 F; 67% White), and sex- and age-matched non-psychiatric controls (n = 88; 40F; 65% White). Serum samples were processed to assess levels of eCBs [(AEA, 2AG, OEA, AA]) and pro-inflammatory molecules (IL-1b, IL-6, IL-8, IL-18, and TNF-α).

Results: In study 1, repeated measure ANOVAs revealed no significant main effect of estrogen condition or sleep fragmentation, or their interaction, on eCB levels (p’s > .05). Sleep fragmentation compared to US significantly increased total mood disturbance, as measured by the Profile of Mood States questionnaire, in the hypoestrogenized condition (F = 4.54; p = 0.03). In study 2, we found a significant decrease in AEA, OEA and AA levels in male PTSD patients compared to male controls, which no significant differences were observed between female individuals with PTSD and female controls (p’s  <  0.001; FDR corrected). Conversely, significant differences in IL-6 and IL-8 were found in women with PTSD, who exhibited significantly higher levels than female controls (p’s  <  0.01; FDR corrected), while no differences were observed between the male subgroups. We next grouped the eCBs and pro-inflammatory markers in two clusters and examined changes in their levels across subgroups (male and females with PTSD and controls). Our results confirmed that eCBs were significantly reduced in males but not females with PTSD (p = 0.0001; FDR corrected), while both males and females with PTSD showed elevated pro-inflammatory markers levels, although the difference with the control subgroup was significant only for females with PTSD (p = 0.02; FDR corrected).

Conclusions: Our findings show for the first time that exposure to acute and chronic stress may recruit the eCB system in a sex-specific manner. Furthermore, they suggest that the neurobiological mechanisms underlying PTSD may vary by sex, which may have important therapeutic implications.

Disclosure: Nothing to disclose.

Panel

47. Current Landscape in Digital Precision Psychiatry: From Objective Digital Biomarkers for Patient Selection to Evidence-Based and Patient-Centered Digital Therapeutics

47.1 Abstract Not Included

47.2 Cognitive Biomarkers for Fast Acting Antidepressants

Catherine Harmer

University of Oxford, Oxford, United Kingdom

Background: Ketamine has rapid antidepressant effects in depression and bipolar disorder. While potential mechanisms of action have been well characterised in animal models, there is much less evidence in humans. In particular, animal models support a rapid effect on burst firing of the lateral habenula (the ‘anti-reward’ center) as well as changes in affective memory.

Methods: We assessed the effect of randomised double blind administration of ketamine (0.5mg/kg) vs placebo in a sample of healthy volunteers (n = 66 complete data sets). Effects on habeula responses during monetary reward and punishment (shock, or loss of money) were assessed using a Pavlovian learning task 24 h after administration using high field 7T fMRI. A battery of computerised tasks measuring emotional processing and memory were also administered in the same session. This study was preregistered at ISRCTN13629652 10.1186/ISRCTN13629652.

Results: Ketamine reduced BOLD responses to the expected value of shock associated stimuli (relative to reward) in the habenula (F = 5.7, p = .02). There was also a reduction in the aversiveness rating of stimuli associated with shock (P < .05). Ketamine also reduced negative bias of memories encoded prior to drug administration, indicating a possible effect on affective memory consolidation (P < .05).

Conclusions: These results translate key findings from preclinical animal models to humans. They support a rapid effect of ketamine on the habenula, 24h post infusion, during aversive conditioning. They also highlight effects on emotional memory consolidation, which may contribute to ketamine’s antidepressant action.

Disclosure: P1vital: Consultant (Self). J and J: Consultant (Self). UCB: Consultant (Self). Lundbeck: Consultant (Self). GH Pharma: Employee (Spouse/Partner). J and J: Grant (Self). Zogenix: Grant (Self).

47.3 Abstract Not Included

47.4 Cash for Neurobehavioral Code: A Cruise Through the Funding Frenzy of Computational Psychiatry

Michele Ferrante

DTR/NIMH/NIH, Rockville, Maryland, United States

Background: Dr. Ferrante serves as the Program Officer for Computational Psychiatry at the National Institute of Mental Health (NIMH). A new focus of the computational psychiatry program is translational digital mental health. Particularly, the early-stage development of Software as Medical Devices (SaMD) such as Prescription Digital Therapeutics (PDT) for treating, diagnosing, monitoring, or preventing various mental health conditions. PDTs are evidence-based neurobehavioral treatments that require FDA clearance and a prescription from a clinician. In his talk, Dr. Ferrante will first outline the most common regulatory pathways to market for SaMDs and PDTs. He will then delve into various digital therapeutic mechanisms of action and underscore the importance of a stakeholders-centered approach. The discussion will also encompass potential strategies to enhance the safety and effectiveness of SaMDs and PDTs. Additionally, Dr. Ferrante will highlight available funding opportunities to support computational methods translating basic science knowledge to discover phenotypic characterization, etiology, pathophysiology, trajectory, and interventions for mental conditions.

Methods: Natural Language Processing was applied to NIH funded grants to identify emergent topics relevant to digital therapeutics a network graph analysis of NIH PIs funded in this area will also be presented.

Results: NIH-led funding efforts have fostered a vibrant, highly clustered network of NIH-funded PDT researchers. The network’s short path length ensures quick communication and efficient information exchange between any two investigators. This small-world structure promotes innovation and effective knowledge sharing across the various NIH-funded health topics that will be depicted.

Conclusions: NA. This presentation, brought to you by a key funder, the National Institute of Mental Health (NIMH), shifts the focus from scientific outcomes to the financial underpinnings of computational psychiatry and digital mental health. It aims to navigate the landscape of these emerging fields, highlighting their challenges and opportunities, identifying scientific gaps, and exploring potential solutions. The goal is not just to fund, but to foster a deeper understanding and facilitate progress in these crucial areas of mental health.

Disclosure: Nothing to disclose.

Panel

48. Towards a Computational Framework for Cognitive Models of Treatment Response in Depression Across Multiple Treatment Modalities

48.1 Effect of Ketamine on Model-Based Planning in the Two-Step Task

Shabnam Hossein

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Cognitive and behavioral inflexibility has been suggested to have an important connection with mood disorders. In depression, for example, one of the cardinal presentations of inflexibility include inflexibility in decision-making and lack of mental response shifts and adaptation to environmental demands. Ketamine, a glutamatergic agent, has shown rapid antidepressant effects. While ketamine has exhibited some evidence to improve cognitive flexibility, its effect on goal-directed decision-making, indicatives of flexible cognitive responding, is unknown. In this study, we used the two-step task, a computer task in which participants learn a model of the dynamic structure of the environment and use this model to plan and shift their choices, to study the effect of ketamine on habitual (model-free) and goal-directed (model-based) behavior.

Methods: We used data from three studies. In study 1, a total of 104 adult patients with treatment resistant depression were randomly assigned to either ketamine (n = 70) or saline infusion (n = 34). Studies 2 (n = 15) and 3 (n = 38) were open-label ketamine studies where depressed patients received a ketamine infusion. At pre-infusion and 24 h post-infusion participants in all studies played the two-step task, and their depressive symptoms were assessed. Hierarchical logistic regression (HLR) models were conducted to examine model-based and model-free behaviors. Moreover, to study individual differences model-based index (MBI) and model-free index (MFI) were extracted from the regression models at each time point pre- and post-infusion.

Results: In all studies ketamine (in study 1 relative to saline) significantly improved depressive symptoms post-infusion (ps < 0.05). We observed evidence for model-free and model-based behaviors at pre and post infusion for both saline and ketamine arms in study 1, as well as in studies 2 and 3 at pre and post infusion. In study 3 and not study 2 MBI significantly increased post infusion (p = 0.047).

Conclusions: These results suggest that while in this study and previous literature model-based and model-free indices are not associated with the severity of depressive symptoms; ketamine seems to have an effect on model-based decision making through a potentially different mechanism than that of ketamine on depressive symptoms.

Disclosure: Nothing to disclose.

48.2 Machine Learning Prediction of Real World Treatment Response to Internet-Delivered Cognitive Behavioral Therapy vs. Patients Receiving Antidepressant Medication

Claire Gillan

Trinity College Dublin, Dublin, Ireland

Background: Depression is a complex disorder with no universal treatment approach. Predictive models using machine learning (ML) may aid treatment decisions by matching patients with treatments to which they are most suited. Prior studies have suggested model predictions might be specific to different antidepressant medications, but none has assessed if ML models can differentiate treatments of different modalities. We aimed to investigate, using real world data, whether ML predictions of response in psychotherapy are treatment-specific or generalisable to antidepressant medication.

Methods: The Precision in Psychiatry (PIP) study assessed patients initiating internet-delivered cognitive behavioral therapy (iCBT) and antidepressants for 4 weeks. They provided diverse self-report (socio-demographic, clinical, psychosocial, treatment, health and lifestyle) and cognitive test data at baseline, which were used to train a host of candidate ML models (elastic net regression, random forest, and extreme gradient boosting machines) to predict depression improvement in iCBT (N = 543). We tested the best performing model on hold-out iCBT data (N = 233) and antidepressant data (N = 110).

Results: The best performing model was an elastic net regression with 27 predictors that explained 14.6% variance in depression improvement, surpassing a benchmark model with baseline severity, age, and sex by 5.6% (p < 0.001). Key predictors included baseline depression, treatment expectation, as well as transdiagnostic and cognitive measures. In real-world data, where concurrent treatments are possible, the model generalised to hold-out iCBT data (R2 = 18.5%) and antidepressant data (R2 = 17.7%). However, when the model was retrained to eliminate cross-over, the model predictions were specific to the iCBT test set (True R2 = 19.3% vs Benchmark R2 = 9.5%) compared to the antidepressant-only test set (True 10.8% vs Benchmark 12.8%).

Conclusions: Machine learning models that predict response to internet-based cognitive behavioral therapy in real world settings also predict response to antidepressant medication, due to spillover from concurrent treatments. Training on more precise groups yield treatment-specific results. Our findings identify a host of baseline self-report and cognitive characteristics predictive of treatment response to iCBT.

Disclosure: Nothing to disclose.

48.3 A Stratified Precision Medicine Trial Targeting Alpha 2A Agonism as a Treatment for the Cognitive Biotype of Depression: The Biomarker Guided (BIG) Study for Depression

Laura Hack

Stanford University School of Medicine, Sierra Pacific Mental Illness Research, Education, and Clinical Center (MIRECC), VA Palo Alto Health Care System, Stanford, California, United States

Background: Although cognitive impairments contribute to serious and long-lasting psychosocial dysfunction in major depressive disorder (MDD), we lack a mechanistically selective treatment targeted to these impairments. We evaluated guanfacine immediate release (GIR) as a novel treatment for selectively engaging the mechanism of alpha 2A (α2A) receptor agonism to improve dorsal prefrontal brain circuit function and cognitive control impairment in a cognitive control biotype (hereafter referred to as a cognitive biotype for brevity) of depression.

Methods: A total of 17 patients with MDD (47% female, mean age 31.4 years, age range 18.7−57 years) met both mechanistic circuit (C) and cognitive behavioral (B) criteria for the cognitive(CB) biotype of depression and completed treatment with GIR. Functional magnetic resonance imaging of the cognitive control circuit, defined by patient-level quantification of the dorsolateral prefrontal cortex (dLPFC) and/or the dorsal anterior cingulate cortex (dACC) was used as the mechanistic criteria. Multimodal measures were repeated after 8 weeks of treatment with GIR. The primary mechanistic outcome was change in circuit function and secondary outcomes were defined using measures of depression, cognitive behavioral performance, self-reported cognition, function, and global life satisfaction.

Results: There was a significant increase in activity (dACC, t = −2.33, p = 0.033, d = 0.57) and connectivity (dACC - left dLPFC, t = −2.75, p = 0.01, d = 0.67), within the cognitive control circuit over 8 weeks of treatment with GIR. Using the 17-item Hamilton Depression Rating Scale, the response rate was 76.5% and the remission rate, 64.7%, exceeding the rates for conventional antidepressants. GIR also significantly improved cognitive behavioral performance, cognitive symptoms, functional capacity, and global life satisfaction.

Conclusions: This is the first study to provide preliminary evidence for the efficacy and target engagement of GIR as a mechanistically selective treatment for the cognitive(CB) biotype of depression. We also demonstrate that the cognitive(CB) biotype can be identified prospectively using a combined cognitive control circuit and behavioral performance biomarker suited to stratification of individual patients and enrichment of study samples in future target engagement trials.

Disclosure: Roche: Advisory Board (Self).

48.4 Acute Expectancy-Mood Neural Dynamics Predict Long-Term Belief Associated Mood Improvement in Depression

Marta Pecina

University of Pittsburgh Medical Center, Western Psychiatric Institute and Clinic, Pittsburgh, Pennsylvania, United States

Background: Acute experimental models suggest that antidepressant placebo effects depend on reinforcement learning (RL) mechanisms encoded within the salience network (SN) and further reinforced by sensory experience and mood fluctuations. However, the extent to which these dynamics extend to longer timescales remains elusive. To address this gap, we investigated how, during the processing of antidepressant expectancies, connectivity between the SN and the default mode network (DMN) facilitates the switch from salience attribution to contextual cues in the SN to belief-induced mood responses in the DMN, contributing to the persistence of placebo effects.

Methods: Sixty psychotropic-free patients with major depressive disorder (MDD) completed a session of the antidepressant placebo fMRI experiment, which manipulates placebo-associated expectancies and their reinforcement while assessing trial-by-trial mood improvement, before entering an 8-week double-blind, randomized, placebo-controlled trial (RCT) of a selective serotonin reuptake inhibitor (SSRI) or placebo.

Results: RL-informed SN-seed connectivity during the processing of learned antidepressant expectancies revealed increased between-network connectivity with the DMN. Acutely, greater SN-DMN functional connectivity (FC) was associated with significantly higher effects of the expectancy and reinforcement conditions on expectancy and mood ratings. Over time, higher SN-DMN connectivity predicted mood improvement, especially when participants believed they were receiving an SSRI, but were instead receiving a placebo.

Conclusions: These findings underscore the pivotal role of SN-DMN coupling in shaping expectancy-mood interactions, serving as a moderator for the impact of drug-assignment beliefs on mood improvement, both acutely and over time.

Disclosure: Nothing to disclose.

Panel

49. Progress Towards RNA Therapeutics for Brain Disorders

49.1 Treatment with AAV9-shRNA to Knockdown the 5-HT2A Receptor Improves Memory in Vivo and Decreases Excitability in Primary Cortical Neurons

Troy Rohn

Cognigenics, Eagle, Florida, United States

Background: Short-hairpin RNAs (shRNA), targeting knockdown of specific genes, hold enormous promise for precision-based therapeutics to treat numerous neurodegenerative disorders. We designed a AAV9-shRNA targeting the downregulation of the 5-HT2A receptor and recently demonstrated that intranasal delivery of this shRNA, referred to as COG-201, reduced anxiety and enhanced memory in mice and rats.

Methods: Methods Summary: In Vitro MEA and In Vivo Memory Test

In Vitro Multi-Electrode Array (MEA) Analysis:

Cell Preparation: Primary cortical neurons isolated from fresh mouse brain embryos.

Treatment: On day 6 (DIV 6), neurons were treated with an AAV9 vector (COG-201 or scrambled shRNA). MEA Analysis: Performed on day 16 to assess spontaneous electrical activity.

In Vivo

Subjects: Wistar male rats (12 animals per group). Groups were randomly assigned to the vehicle (PBS) or COG-201-treated groups.

Procedure: Test based on rats’ natural tendency to explore novel objects, requiring memory of familiar objects.

Memory assessment was conducted 3 weeks post-treatment.

Memory Discrimination Index Calculation:

(Time exploring novel object - Time exploring familiar object) / (Time exploring novel object + Time exploring familiar object) x 100%.

Data collection and analysis were blinded until the end of the experiments.

Statistical analysis

For PCR, MEA analysis, and Western blot quantification, t-tests for two independent means were calculated using a significance level set at 0.05 and one-tailed hypotheses.

Results: Here we provide additional in vivo data supporting a role of COG-201 in enhancing memory by demonstrating an increase in the discrimination index in a novel object recognition test (p value = 0.00025). In vitro, we also demonstrate knockdown of the 5-HT2A receptor in primary mouse cortical neurons led to a decrease in mRNA expression, protein expression, and spontaneous electrical activity as measured by multi-electrode array (MEA).

Conclusions: The decrease in mRNA and protein expression, along with reduced spontaneous electrical activity in primary mouse cortical neurons, corroborate our in vivo findings and underscore the efficacy of COG-201 in decreasing HTR2A gene expression. This convergence of in vitro and in vivo evidence further suggests the potential of COG-201 as an effective and precisely targeted therapeutic strategy.

Disclosure: Nothing to disclose.

49.2 Brain Delivery of Neuron-Derived Extracellular Vesicles Carrying miR-151a-5p Elicits Antidepressant Response

Gustavo Turecki

McGill University, Montreal, Canada

Background: Cells, including neurons, release extracellular vesicles with specific molecular cargo that can cross the blood-brain barrier and be taken up by target cells producing a molecular response. Here we describe a study using complementary approaches, including human, postmortem brain and animal studies whereby we first identified a neuron-derived extracellular vesicle (NEV) carrying a microRNA that is differentially expressed peripherally according to antidepressant response, then we characterized this NEV and its cargo and study its effect when delivered to the prefrontal cortex of mice.

Methods: We analyzed 430 human plasma samples from a clinical trial (clinicaltrials.gov, NCT01655706). The samples were collected before and after 8-week escitalopram treatment. NEVs were isolated from plasma using size exclusion chromatography following by immunoprecipitation against SNAP25. Small RNA-seq profiling followed by RT-qPCR validation were used to determine miRNA cargo in NEVs. We optimized a method of producing engineered NEVs enriched with specific miRNA cargo. Such engineered NEVs had biological properties similar to naturally occurred NEVs. Antidepressant properties of engineered NEVs were tested in vivo using chronic social defeat, a well-established mouse model that produces behavioral phenotypes similar to depression and anxiety.

Results: Before treatment, depressed patients had lower miR-151a-5p levels in NEVs. However, these levels significantly increased over time to levels similar to healthy controls only in the group of depressed patients who responded to antidepressant treatment. miR-151a-5p levels in NEVs demonstrated good predictive value in discriminating between responders and non-responders. In vitro and in silico experiments revealed that miR-151a-5p negatively regulates a gene set enriched in the prefrontal cortex, responsible for regulating glutamatergic signaling. Engineered NEVs enriched with miR-151a-5p and delivered to mouse prefrontal cortex displayed antidepressant properties in vivo and effectively ameliorated the behavioral deficits induced by chronic social defeat stress.

Conclusions: Increased levels of miR-151a-5p cargo in NEVs mediate antidepressant treatment response, and the delivery of this miRNA to the prefrontal cortex of mice elicits effects similar to antidepressants.

Disclosure: Nothing to disclose.

49.3 Novel Insight on Depression From the Dark Side of the Genome

Orna Issler

NYU Langone Medical Center, New York, New York, United States

Background: Depression is a devastating psychiatric disorder striking women at twice the rate as men, yet our understanding of the molecular mechanism leading to this sex difference is limited. We discovered that long non-coding RNAs (lncRNAs), a prominent class of epigenetic agents, are robustly regulated in post-mortem brains of depression cases in a sex-specific manner. We identified LINC00473 as downregulated in the prefrontal cortex (PFC) of depressed women only. Using animal models and human cells in culture, we concluded that LINC00473 is a CREB effector acting solely in females to drive stress resilience in an unknown mechanism.

Methods: We performed Chromatin Isolation by RNA Purification (ChIRP) followed by Mass Spectrometry (MS) for LINC00473 in human female neuroblastoma cells (n = 4 per group). We generated Herpes Simplex Viruses (HSV) to overexpress one of the ChIRP-MS hits, Aars2 (Aars2-OE), or GFP as a control (GFP-OE) in the PFC of female mice. In several independent cohorts, we tested behavior (n = 12 per group) and mitochondrial gene expression using Magnetic Cell Sorting (MACS) followed by RNA sequencing (n = 4 per group) or MS (n = 5 per group).

Results: The ChIRP-MS results revealed that CREB-mediated induction of LINC00473 recruits it to the mitochondria. There, LINC00473 binds proteins associated with metabolic functions, including alanyl-tRNA synthetase 2 (Aars2), the enzyme that mediates the loading of the mitochondrial alanine tRNA. Like LINC00473, Aars2 is downregulated in the PFC of post-mortem depression cases only in females. To test the causal role of Aars2 in the PFC in female mice stress resilience, we compared Aars2-OE to GFP-OE to find an effect on the behavioral strategy used in lever pressing for liquid reward in operant boxes. Finally, we profiled Aars2-OE on mitochondrial gene expression compared to nuclear gene expression at RNA and protein levels. This allowed testing the effects of Aars2-OE on mitochondrial gene translational efficacy and points to molecular mediators of its effects on female mice behavior.

Conclusions: Our molecular and behavioral findings suggest that Aars2 mediates LINC00473’s pro-resilience role in females. This study sheds light on the female-specific mechanisms associated with individual differences in the response to stress and promotes our understanding of sex differences in depression.

Disclosure: Nothing to disclose.

49.4 Circular RNA-Derived Micropeptides at the Synapse: Potential New RNA Therapeutics for Neurological Disorders

Timothy Bredy

The University of Queensland, Brisbane, Australia

Background: CircRNAs are abundant in the brain and enriched within synapses, with recent studies suggesting a relationship between circRNA activity and cognitive processes. Emerging evidence also indicates that circRNAs can encode micropeptides; small proteins involved in a range of processes including cellular proliferation, differentiation, and migration. Whether there are any circRNA-derived micropeptides involved in learning and memory, remains to be determined.

Methods: We performed ribosome sequencing and mass spectrometry on synapses isolated from behavorally-trained mice, which provided strong evidence for dynamic local circRNA translation in response to learning. By combining these data with learning-induced circRNA expression profiles in the synaptic compartment, we discovered 325 synapse-enriched circRNAs bound to the ribosome, with mass spectrometry revealing direct evidence of translation from 171 of these circRNAs.

Results: Amongst this list was a micropeptide derived from a single exon circRNA that we call P1, the full-length host of which is a methyltransferase involved in protein repair. Importantly, we found that this micropeptide, which is only a third of the size of the full-length protein, is enzymatically active and locally expressed, and that P1 knockdown in the prefrontal cortex impairs memory.

Conclusions: These data imply the existence of an experience-dependent circRNA-derived micropeptide that is critically important for memory and, potentially, a subcellular compartment-specific apparatus for localised plasticity in the adult brain.

Disclosure: Nothing to disclose.

Panel

50. The new Faces of Fear: Cortical and Subcortical Orchestration of Freezing, Flight, Avoidance, and Escape

50.1 Defensive Response Regulation by Central Amygdala Projections to the Retrorubral Field

Jonathan Fadok

Tulane University, New Orleans, Louisiana, United States

Background: Animals exhibit a diverse array of adaptive behaviors in response to threats. When dysregulated, however, these behaviors can become maladaptive, as is observed in patients suffering from posttraumatic stress and panic disorders. The central amygdala (CEA) is vital for defensive responses such as freezing and flight, yet the downstream targets of the CEA mediating high intensity reactions to threat are unresolved. This presentation will highlight the role of a novel projection from the CEA to the midbrain retrorubral field (RRF) in control of defensive behavior.

Methods: Both male and female mice were used in all experiments. Anterograde and retrograde neuroanatomical tracing methods were used to define the details of the pathway. Intersectional caspase-mediated ablation of the pathway was performed, and mice were conditioned in a paradigm that elicits freezing and flight behavior to different components of an auditory serial compound stimulus. Chemogenetic inhibition of GABAergic neurons was conducted in the RRF also in mice conditioned in the flight paradigm.

Results: Robust projections from the CEA to the RRF were observed. A trans-synaptic viral vector approach revealed that CEA neurons innervate both dopaminergic and non-dopaminergic neurons in the RRF. FOS staining revealed that the CEA to RRF pathway is activated during fear conditioning. Caspase-mediated ablation of the pathway significantly lowered both freezing and flight behavior. Chemogenetic inhibition of GABAergic neurons in the RRF led to significant increases in flight behavior.

Conclusions: These data demonstrate that the understudied CEA to RRF pathway is important for defensive behavior. Ongoing experiments are monitoring neuronal activity of CEA projectors and several types of RRF neurons using genetically encoded calcium indicators. These studies advance our understanding of the neural circuits controlling threat responding.

Disclosure: Nothing to disclose.

50.2 Prefrontal Circuits Correlated With Observational Active Avoidance in Male and Female Rats

Maria Diehl

Kansas State University, Manhattan, Kansas, United States

Background: An individual witnessing others respond to danger can gain information via social cues without directly experiencing harm. This “observational fear” and its underlying mechanisms have been demonstrated in rodents, but whether animals can also learn to avoid a threat through observation is unknown. We therefore modified the platform-mediated avoidance task (PMA) to identify the neural basis of observational active avoidance in male and female rats.

Methods: Demonstrators underwent 9 tone-shock trials per day for 10 days. Observers were assigned to either naïve (n = 18−20) or shock-exposed (n = 10−12) subgroups, and then witnessed a Demonstrator undergo PMA during early (Day 1) or late (Day 10) training, followed by a Direct PMA Session 24hr later. 90 min after Tone 1 of Direct PMA, we used cFos immunohistochemistry to quantify neural activation in anterior cingulate cortex (ACC), a region critical for observational learning.

Results: Regardless of prior shock exposure, Day 1 Observers exhibited greater freezing compared to Day 10 Observers (p < 0.05). Additionally, Day 1 Observers with prior shock exposure spent more time on the platform (p < 0.05) and avoided more shocks compared to naive (p < 0.05). Day 10 Observers, regardless of shock, showed similar time on platform and number of shocks avoided. Time on platform in Day 1 Observers with prior shock was positively correlated with the Demonstrator, whereas time on platform of Day 1 Observers with no prior shock was negatively correlated with the Demonstrator. We found the opposite in Day 10 Observers: time on platform in rats with prior shock was positively correlated with the Demonstrator, whereas time on platform in rats with no prior shock was negatively correlated with the Demonstrator. cFos density in ACC was greater in D10 compared to D1 Observers (p < 0.01).

Conclusions: Observers witnessing early PMA use fear strategies that depend on shock experience, whereas Observers witnessing late PMA use avoidance strategies that may not require shock experience. Current results suggest ACC is recruited during late Observational PMA. Ongoing analyses will assess neural activation in established PMA circuits, (prelimbic cortex, basolateral amygdala, ventral striatum). These findings will advance our understanding of prefrontal circuits for observational active avoidance.

Disclosure: Nothing to disclose.

50.3 Hypothalamic Control of Acquisition of Learned Flight Induced by Threat Imminence

Avishek Adhikari

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

Background: Conditioned freezing ahs been extensively studied. Similarly to conditioned freezing, flexible experience-dependent learned escape also has high adaptive value. However, the neural basis of learned escape is not well-understood. Often escape in rodents is elicited by a chasing predator, but this is difficult to simulate in a lab setting. We developed a new assay that induces rapid learning of flight induced by a simulated predatory chase. We also show that hypothalamic cholecystokinin (cck)-expressing dorsal premammillary nucleus (PMd) cells are necessary for the acquisition of learned flight.

Methods: A mouse is placed in a long corridor with a moveable shocking grid. During habituation, the grid is moved alternately from left to right and right to left, and this motion does not induce defensive behaviors. During fear acquisition day, 24-h later, the grid delivers foot shocks during grid movements. Mice learn within < 10 trials to jump by traversing the moving grid to avoid shocks. The mouse learns flight upon grid movement (FUGA). During retrieval, the mouse displays FUGA during the grid movement.

Separate cohorts of mice were trained to learn standard versions of either cued auditory or contextual fear conditioning.

To test if if PMd-cck cells controlled acquisition of these 3 assays, we chemogenetically inhibited these cells during fear learning for contextual fear, auditory fear and FUGA.

We performed fiber photometry recordings in GCaMP6s-expressing PMd-cck cells.

Both sexes were used throughout.

Results: Inhibition of PMd-cck cells during FUGA acquisition impaired behavior during fear retrieval 24-h later. Relative to controls, during fear retrieval, hM4Di mice show decreased number of successful FUGA (49% < controls, p = 0.011) and less freezing (31% < controls, p = 0.008). In contrast, PMd-cck inhibition during acquisition of cued auditory or contextual fear did not impair any behaviors during fear retrieval.

Additionally, PMd-cck cells encode FUGA behaviors, but did not represent neither auditory fear conditioned tone nor freezing in this assay.

Conclusions: These data show a new behavioral assay eliciting conditioned flight. We also identify hypothalamic PMd-cck cells as a key node in the acquisition of these learned escape actions, adding to discoveries showing that the hypothalamus contributes to learning.

Disclosure: Nothing to disclose.

50.4 A Mu Opioid Receptor Mechanism Underlying Sex Differences in Conditioned Fear Behavior

Rebecca Shansky

Northeastern University, Boston, Massachusetts, United States

Background: The behavioral responses elicited by a traumatic event can have implications for long-term clinical outcomes, and so identifying the neurobiological basis of individual differences in threat responding is critical for translational neuroscience. Males and females may be best served by different responses, yet paradigms like Pavlovian fear conditioning usually assess only freezing. Our lab has found sex differences in additional behaviors, including darting, ultrasonic vocalizations (USVs), and shock responsivity, but the mechanisms mediating these differences are unknown. Here we investigated the role of mu opioid receptor (MOR) signaling in driving these effects.

Methods: Experiment 1: Adult male and female Sprague Dawley rats (n = 12/group) were given i.p. injections of MOR antatonist naloxone (5mg/kg) prior to undergoing cued fear conditioning. Locomotor behavior and USVs were recorded.

Experiment 2: Adult male and female Sprague Dawley rats (n = 12/group) underwent stereotaxic surgery to label neurons in the ventrolateral periaqueductal gray (vlPAG) that receive direct input from the anterior cingulate (ACC) using a trans-synaptic viral approach. After fear conditioning, brains were extracted and the vlPAG was processed for cellular phenotyping (MOR and GAD) in virally labeled cells using RNAscope.

Results: Naloxone produced only minor changes to freezing behavior in both sexes. However, naloxone-treated females exhibited a complete elimination of darting, a reduction in shock response (p = 0.0005 vs veh), and an increase in USV emission (p = 0.003 vs veh) that was not seen in males. We also found that female Darters had fewer vlPAG cells receiving ACC input than non-darting females or males, and overall, females had more GAD-expressing cells than males (p = 0.02).

Conclusions: Together, these data point to a novel, MOR-mediated mechanism for driving sex differences in diverse conditioned fear behaviors. The naloxone findings suggest that darting as a conditioned response is MOR-dependent and that measurement of additional behaviors like USVs and shock responding can be highly informative in assessing the emotional state of animals during classic conditioning paradigms. Our RNAscope data demonstrate that vlPAG circuit organization may be sex-dependent and offers a potential site of action by which naloxone produced its effects.

Disclosure: Nothing to disclose.

Panel

51. Isolating Mechanisms of Neuropeptide Neural Circuits Which Shape Motivated Behavior

51.1 A Peri-Ceorulear Neuropeptidergic Pathway for Modulating OFC-Mediated Opioid-Seeking Behavior

Kasey Girven

University of Washington, Seattle, Washington, United States

Background: A relatively unknown neuropeptide, neuropeptide S (NPS), drives reward-seeking through activation of its cognate Gq-coupled protein receptor NPSR1. However these studies lack the tools necessary to understand the system’s role in reward-seeking. We generated both NPS-Cre and NPSR1-Cre driver mouse lines and isolated a population of NPS-containing cells that reside partly within and medial to the locus coeruleus (LC), known as the periLC. We found that the periLC sends excitatory projections to the orbitofrontal cortex (OFC) that connect directly with NPSR1-expressing neurons. Therefore, I hypothesized that periLC-NPS enhances reward-seeking through increased activation of OFC-NPSR1 post reward consumption.

Methods: NPSR1-Cre (n = 12) and NPS-Cre (n = 12) mice received unilateral infusions of AAVDJ-DIO-GCaMP6s to either the OFC or periLC, respectively. In addition, all mice received unilateral fiber photometry implants to either the ipsilateral OFC for OFC-NPSR1 cell body recordings, or the ipsilateral periLC for periLC-NPS cell body recordings. These mice underwent Pavlovian and operant conditioning for sucrose reward, and an oral fentanyl self-administration task where mice received 120min sessions where they could nosepoke for sipper delivery of fentanyl solution (10 ug/ml). 2-Photon Imaging: NPSR1-Cre mice (n = 6) expressing AAVDJ-DIO-GCaMP6s in the OFC with a 1mm diameter GRIN lens positioned above received 10 Pavlovian conditioning sessions for 10% sucrose. Following this, animals were given access to a wheel and trained to make clockwise (active) or counter-clockwise (inactive) rotations for sucrose access. All studies used male and females.

Results: We found that OFC-NPSR1 neuron activity increases to cue delivery (One Way ANOVA, p < 0.05), but reduces during consumption of reward (One Way ANOVA, p < 0.001) showing a bidirectional response. Then post consumption, these neurons have a rebound increase in activity (One Way ANOVA, p < 0.01). We then determined that during this post consumption event, periLC-NPS neuron activity is aligned and activated (One Way ANOVA, p < 0.001).

Conclusions: Examining the mechanisms of less understood neuropeptides, such as NPS in driving reward-seeking will aide in development of more effective therapeutic strategies. These results establish a critical role for the NPS system in reward-seeking.

Disclosure: Nothing to disclose.

51.2 Uncovering the Role of Prefrontal Cortical Somatostatin Signaling in Binge Drinking

Nicole Crowley

Penn State University, University Park, Pennsylvania, United States

Background: Our previous published work found that prefrontal cortical (PFC) somatostatin (SST) neurons release the peptide SST and are engaged during exploratory behavior, and that PFC administration of an SST receptor targeting drug can promote exploratory behavior (Brockway, 2023 Cell Reports). We also found that these SST neurons are affected by an animal model of voluntary binge alcohol consumption (drinking in the dark; ‘DID’) and that alteration of SST neurons modulates binge drinking behavior (Dao, 2021 Neuropsychopharmacology). Given this, we sought to understand the effect of binge drinking via the DID model on SST dynamics in the PFC.

Methods: Adult male and female C57BL/6J, SST-IRES-Cre, and Ai9 reporter mice were used. Mice were maintained on a reverse light cycle and had ad libitum access to food and water. Binge drinking: DID was conducted as previously published (Dao., 2021; Rhodes, 2005; Thiele, 2014). Mice received 20% (v/v) ethanol in tap water, 3 h into the dark cycle for 2 h on three sequential days. On the fourth day, they received EtOH for 4 h. Following the binge day, mice had three days of abstinence before repeating the cycle 3 more times. ELISA for photostimulated release of SST: Acute brain slices for photostimulation were prepared as previously described (Al-Hasani et al, 2015; Dao, 2019). Slice electrophysiology: Pyramidal neurons were identified by location from midline, morphology, and membrane characteristics, consistent with previously published electrophysiology (Cummings and Clem, 2020; Dao, 2021; Lowery-Gionta., 2018).

Results: Following 4 weeks of DID, SST was unable to hyperpolarize PFC pyramidal neurons in male and female mice, while control mice had normal SST-induced silencing. We did not see changes in SST cell number, peptide expression, or optogenetic-induced SST release. This data suggests that alcohol consumption leads to a downregulation in SST receptor expression, which renders the neuropeptide unable to modulate the dynamics of PFC activity. Ongoing experiments are exploring SST receptor expression on pyramidal neurons via RNAseq, cannula administration of SST receptor targeting drugs on binge drinking, and the role of estrogen receptor alpha on SST neurons.

Conclusions: Alcohol consumption leads to a dysregulation of SST peptide signaling, normally essentially for dynamic regulation of the PFC.

Disclosure: Nothing to disclose.

51.3 Substance P Neuromodulation of Striatal Circuit Function and Stress Related Behavior

Tanner Francis

University of South Carolina, Columbia, South Carolina, United States

Background: Despite our understanding of how stress affects our nervous system, causal mechanisms driving stress susceptibility are unknown. We recently discovered that the neuropeptide substance P (SP) scales with the salience of stressors and causes cell-type selective plasticity on Nucleus Accumbens (NAc) medium spiny projection neurons. This provides a mechanism in which SP could enhance stress susceptibility by enhancing stressor salience.

Methods: Male and female mice underwent acute or repeated foot shock or chronic social defeat stress. SP release was assessed with in vivo fiber photometry and GRAB-SP1.0. To assess how SP affects stress-dependent outcomes, mice were injected with the SP receptor antagonist L-733,060 or saline (N = 6−9 mice per group). Glutamatergic plasticity was assessed on dopamine 2 receptor (D2) expressing MSNs, the site of NAc SP plasticity, after 1-day or multiple days of foot shock stress (N = 7−19 cells per group) and cells were filled with neurobiotin. Dendritic spines were reconstructed to assess morphological changes associated with plasticity (N = 3−6 animals per group).

Results: Both foot shock stress and social stress caused scaled release of NAc SP. SP receptor antagonism blocked social stress sucrose preference deficits (ANOVA Interaction F(1,7) = 5.777, p < 0.05), suggesting SP is required for susceptibility to stress. Both 1-day (t(26) = 2.173, p < 0.05) and 7-days (t(13) = 1.826, p < 0.05) of foot shock stress significantly enhanced the amplitude of spontaneous excitatory post-synaptic currents (sEPSCs) to NAc D2-MSNs, in a SP-dependent manner. In addition, 7-day stress also increased the frequency of sEPSCs (t(18) = 2.377, p < 0.05). While no effect was observed on total spine density, stubby spine density increased by day 3 (t(8) = 1.964, p < 0.05) and remained elevated to day 7 (t(6) = 4.314, p < 0.01). We are currently assessing electrophysiological and morphological changes due to SP in chronic and relapsing social defeat stress.

Conclusions: Our results indicate SP is critically involved in stress-dependent outcomes including plasticity changes that lead to motivational deficits. This is likely occurring by enhancing the salience of stressors and perceived intensity of stress. Our work provides a target for suppressing susceptibility to stress with broader implications in depression development and relapse.

Disclosure: Nothing to disclose.

51.4 Diet-Induced Changes in Neurotensin Signaling Shape Hedonic Feeding and Obesity Progression

Neta Gazit Shimoni

UC Berkeley, Albany, California, United States

Background: We conducted a comparative study to examine the neurobiology underlying hedonic feeding behavior among mice on a regular diet (REG) vs chronic high-fat diet (HFD). We focused on a major disinhibitory feedback loop of the brain’s reward system that connects the ventral striatum to the midbrain dopamine system, specifically the lateral nucleus accumbens and its projections to the ventral tegmental area (NAcLat→VTA).

Methods: We used in vivo electrophysiology to record neural activity of NAcLat→VTA cells. Cells were recorded in a hedonic feeding assay in mice subjected to REG or HFD. Optogenetics were used to manipulate this pathway during feeding behavior in REG or HFD mice. Cellular and molecular adaptations triggered by HFD were examined using single-cell patch-sequencing in REG and HFD mice. We developed a novel GPCR fluorescent sensor, ntsLight1.1, which increases fluorescence when the neuropeptide neurotensin (NTS) is bound. Genetic strategies were developed to selectively overexpress NTS in NAcLat→VTA cells of HFD mice. In these mice, NTS release, weight gain and feeding behaviors were assessed. All experiments were conducted in C57Bl6 mice (male and female, 8-12 weeks old) with n = 5−15 mice per diet group.

Results: NAcLat→VTA cell activity increased during hedonic feeding in REG mice (p < 0.001) but not in HFD mice (p > 0.99). Optogenetic activation of the NAcLat→VTA pathway promoted hedonic feeding in REG mice (p < 0.001) but not in HFD mice (p = 0.145). However, hedonic feeding behavior could be restored in HFD mice switched back to a regular diet (p = 0.004). We also explored synaptic and molecular mechanisms and found that NTS mRNA is highly expressed in NAcLat→VTA cells of REG mice but significantly reduced in chronic HFD mice (p = 0.041). Additionally, NTS release in the VTA was reduced in HFD mice when compared to REG mice (p = 0.006). Elevating NTS expression and release in the NAcLat→VTA pathway of HFD mice normalized various aspects of diet-induced obesity, including weight gain (p = 0.018) and feeding patterns (p = 0.005).

Conclusions: Our findings elucidate a neural mechanism through which the neuropeptide NTS normally regulates hedonic feeding behavior and shed light on how disruptions in NTS signaling may play a role in the disordered feeding patterns and weight gain frequently observed in response to a chronic high-fat diet.

Disclosure: Nothing to disclose.

Study Group

52. The Challenges and Opportunities Large Scale Population Based Neuropsychiatric Datasets

Martin Paulus, Chun-Chieh Fan, Terry Jernigan, Damien Fair, Dedra Buchwald, Wesley Thompson

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

Study Group Summary: As the field of neuropsychiatric research contends with historical biases and a lack of representativeness in participant samples, both neuroimaging and genomic studies are now emphasizing inclusivity. This shift results in significantly larger and more socio-demographically diverse datasets. This transformation offers a unique opportunity to rectify past research disparities and promote equality by allowing researchers to examine how structural issues, such as socioeconomic and racial inequalities, impact the human brain. However, increased socio-demographic diversity introduces greater heterogeneity, which, while beneficial, poses significant challenges in analysis and causal inference. These challenges are compounded by recent controversies regarding the conflation of genetic ancestry with race/ethnicity.

The tension between the benefits of diverse samples and the analytical challenges of heterogeneous data is evident. Without clear guidelines, researchers, and clinicians face difficulties in analyzing and interpreting data from diverse samples, which hinders progress in identifying causal factors for neuropsychiatric disorders. The recent National Academies of Sciences, Engineering, and Medicine (NASEM) report recommends disentangling race and ancestry, advising against using discretized race and ancestry as population descriptors. However, many observational studies use race as a metric for diversity recruitment, and polygenic predictions often stratify by continental ancestry. Additionally, race and ancestry are highly correlated with environmental exposures due to historical disparities, leading to confounds and biases in diverse samples, with limited covariates available to control for these confounds.

Through this session, we aim to examine the following key points, including defining “diversity,” contextualizing “representativeness,” and avoiding inferential fallacies despite best intentions. Specifically, we will work towards (1) Establishing best practice guidelines for analyzing diverse neuropsychiatric datasets. (2) Identifying robust methodological approaches to control for confounding factors related to socio-demographic diversity. (3) Proposing effective strategies to disentangle race and ancestry in neuropsychiatric research. (4) Creating a framework to address the structural impacts on mental health, thereby promoting more equitable and accurate scientific outcomes. By fostering a comprehensive dialog, we seek to advance the understanding and utilization of diverse neuropsychiatric datasets in research and clinical practice.

Disclosure: Nothing to disclose.

Mini Panel

53. The Transdiagnostic Food-Mood Connection: From Novel Insights to Targeted Treatment

53.1 Food Reward-Related Hypothalamic-Striatal Connectivity in Response to Psychosocial Stress Differentiates State-Related Appetite Changes in Major Depressive Disorder

Laura Holsen

Harvard Medical School, Boston, Massachusetts, United States

Background: Emotional eating, a tendency to eat in response to negative emotions, including psychosocial stress, has been linked to bulimia nervosa and binge eating disorder, and more recently, major depressive disorder. Compared to those with hypophagic depression (HypoMDD; decreased appetite/weight loss), individuals with hyperphagic depression (HyperMDD; increased appetite/weight gain) exhibit greater medical comorbidities, with recent data suggesting that hyperphagic behaviors stem in part from disruption in mesolimbic regions governing reward and signaling in ghrelinergic pathways. The current study examined the hypothesis that connectivity between the hypothalamus, striatum, and amygdala during hedonic food cue processing following psychosocial stress would be stronger in HyperMDD, compared to HC and HypoMDD, and would be related to ghrelin.

Methods: We studied 90 adults [40 healthy control (HC; 20F/20M), 29 unmedicated Hypophagic MDD (HypoMDD; 15F/14M); 21 unmedicated Hyperphagic MDD (HyperMDD; 9F/12M)], each of whom completed 2 study visits (~1 week apart) involving serial blood sampling (for ghrelin), completion of the Maastricht Acute Stress Task (MAST; an acute psychosocial stressor; Stress Visit) or a non-stressful version of the MAST (Control Visit), and MRI scanning (food incentive delay task) on a Siemens 3T Skyra. Changes in ghrelin, brain connectivity, and connectivity-hormone relationships were examined in the context of response to psychosocial stress.

Results: During Stress vs. Control, for anticipation of food reward (vs. neutral), the HyperMDD group, relative to HC, showed hyperconnectivity between the hypothalamus and putamen [t = 2.81, p(uncorr) = 0.006; p(FDR) = 0.096], while for receipt of food reward (vs. neutral), the HyperMDD group showed hyperconnectivity between amygdala and hypothalamus and between left and right NAcc. Comparing the Stress and Control visits during food reward anticipation, HyperMDD, relative to HypoMDD, exhibited hyperconnectivity between the caudate and putamen [t = 3.14, p(uncorr) = 0.003; p(FDR) = 0.038] and between hypothalamus and putamen, caudate and amygdala, right and left putamen, and putamen and amygdala. Left-right putamen connectivity during stress-enhanced food reward anticipation was negatively associated with ghrelin area under the curve (AUC) in HyperMDD (r = -0.48, p < 0.05), but not in HC (r = 0.30, p = 0.10, Fisher Z: p = 0.004) or HypoMDD (r = 0.14, p = 0.47; Fisher Z: p = 0.018).

Conclusions: As hypothesized, homeostatic-reward region synchrony was enhanced in HyperMDD specifically in response to psychosocial stress, and inversely related to stress-induced ghrelin. As ghrelin potentiates both homeostatic appetitive signaling and dopamine receptor binding through dimerization with D1 and D2 receptors, this could reflect stress-related dysfunction in GHS-R1a-to-D1/D2 binding affiliation, a potential mechanism for stress-induced overeating in this group.

Disclosure: Nothing to disclose.

53.2 Task-Based and Naturalistic Emotion Prediction Errors Influence Urges and Behavior Among Binge Eaters

Laura Berner

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

Background: Outcomes that deviate from expectation (i.e., prediction errors) guide our decision-making and learning. Recent research suggests that emotion prediction errors (EPEs), and in particular feeling worse than expected, may promote maladaptive decision-making and reactive choices in healthy populations. While considerable research has focused on the experience of negative emotion before maladaptive behaviors like binge eating, it is unknown whether the difference between this experienced emotion and emotional expectations prompts these symptoms. Here, across two studies and using task-based and ecological momentary assessments, we tested the hypothesis that momentary fluctuations in EPEs would predict increases in binge-eating urges and behaviors.

Methods: In Study 1, we adapted an online Ultimatum Game to measure fluctuations in expected emotion, experienced emotion, and eating-related urges among demographically-matched adults with and without binge eating (N = 116). In the game, participants chose to punish or forgive others for offers that varied in fairness. Before seeing a partner’s offer, participants predicted the offer amount and rated how they thought they would feel after receiving an offer. They then rated their experienced emotion and eating-related urges after actually getting the offer, and finally chose to accept or reject the offer. EPEs on each trial were calculated as the difference between expected and experienced emotions. In Study 2, adults diagnosed with binge-eating disorder (N = 24) completed 10 days of signal-contingent and binge-eating event-related ecological momentary assessment (EMA). Six times per day, participants rated their momentary predictions of how eating would affect their emotions. At all assessment points, participants rated their current emotion. EPEs were calculated as the difference between the actual post-binge-eating change in emotion and the pre-behavior expected emotional change from eating.

Results: In Study 1, task data indicated that large negative EPEs (i.e., feeling worse than expected) increased trial-wise urges to eat more strongly in the binge-eating versus control group (t = -1.05, p < 0.001). Within the binge-eating group, negative emotion prediction errors specifically increased urges to binge (t = -1.97, p = 0.049), and this link was particularly strong among individuals who reported the strongest sense of loss-of-control over their eating in the past month (t = -2.94, p = 0.003). Experienced negative emotion alone was not associated with urges to binge eat during the task. Study 2 EMA results indicated that individuals with binge-eating disorder who had larger negative EPEs in the natural environment had more frequent binge eating over the 10-day assessment period (z = -3.23, p = 0.001). Experienced emotional change after eating alone was not predictive of binge-eating frequency.

Conclusions: These data challenge traditional models of binge eating that are focused only on experienced emotions, (e.g., emotion-regulation models) or expected emotions (e.g., expectancy models). Instead, the results highlight the value of integrating these features into an emotion-focused expectancy violation model. Our findings could inform new interventions that focus on helping patients more adaptively respond to and learn from surprising emotional states, rather than on regulating strong negative emotional experiences alone.

Disclosure: Juniver, Ltd.: Advisory Board (Self).

53.3 Real-Time fMRI Neurofeedback Targeting Limbic Disturbances Enhances Emotion Regulation in Anorexia Nervosa

Ann Haynos

Virginia Commonwealth University, Richmond, Virginia, United States

Background: Anorexia nervosa (AN) is a serious disorder with elevated morbidity and premature mortality rates. Studies suggest that individuals with AN experience emotion regulation (ER) difficulties that maintain eating disorder symptoms and that limbic system disturbances may contribute to these difficulties. However, interventions targeting ER and associated limbic disruptions in AN are limited. Real-time fMRI neurofeedback (rt-fMRI NF), a type of biofeedback that allows individuals to view and purposefully alter their neural activity, is a novel tool that may bolster ER by directly altering limbic reactivity. We conducted a randomized study comparing the impact of rt-fMRI NF targeting amygdala downregulation compared to sham feedback on limbic activity and affect in AN.

Methods: Participants with AN (N = 23) were randomized to receive rt-fMRI NF (n = 11) or sham feedback (n = 12) across 2 fMRI sessions (Visits 1 and 2). During fMRI, all participants were instructed to either: (1) regulate emotional responses to negative images (Regulate); or (2) passively view negative images (Negative View). In the Regulate condition, NF participants received real-time NF on their amygdala activation; sham received non-contingent feedback. Regions of interest (ROIs) included right and left amygdala (reflecting limbic activation) and dorsolateral prefrontal cortex (DLPFC; reflecting cognitive control). After viewing images in-scanner, they self-reported emotional responses (higher scores = more positive mood). Outside of fMRI, participants reported negative affect at baseline, Visit 1, Visit 2, and follow-up.

Results: Across visits, groups showed no difference in amygdala activation (ps > 0.68), but NF had lower right DLPFC activation (F = 4.31, p = 0.035) for the Regulate > View contrast relative to sham. This group difference was not evident at Visit 1 (p = 0.44), but emerged at Visit 2 (F = 4.31, p = 0.044). For in-scanner affect, a significant group x condition interaction was noted at Visit 1 (Chi-square = 6.48, p = 0.039), in which NF had lower mood than sham in the Regulate condition (B = −0.37, p = 0.110). However, at Visit 2, the NF group reported more positive in-scanner mood than sham across the session (Chi-square = 4.25, p = 0.039) and particularly in the Negative View condition (B = 0.81, p < .001). Similarly, negative affect linearly decreased for NF across assessment points (B = -3.57, p = 0.015); whereas sham showed a quadratic pattern (B = 0.41, p = 0.010) wherein negative affect decreased at Visit 1, then rebounded thereafter.

Conclusions: rt-fMRI NF allows for simultaneously probing neural mechanisms and evaluating a potential intervention for disorders of ER. We found that repeated rt-fMRI targeting amygdala downregulation improved ER in AN. Further, results suggest that successful ER in AN may be facilitated by reducing cognitive control circuit engagement. Future research is needed to investigate rt-fMRI NF as a clinical approach to enhance ER in AN and related disorders.

Disclosure: Nothing to disclose.

Mini Panel

54. Sleep Disruption in Neuropsychiatric Illness: Liabilities and Opportunities

54.1 Abstract Not Included

54.2 Sleep and Circadian Disruptions During Adolescence: Impact on Neurobiology and Vulnerability for Substance Use

Colleen McClung

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

Background: Adolescence is a vulnerable time for the initiation of substance use. There are a number of factors that contribute to this, but one that is often overlooked is the influence of sleep and circadian rhythm disruptions which are all too common during this developmental time period. In our modern society, teens are faced with a number of environmental factors that coupled with biological changes, leave them in a state of circadian misalignment termed “social jet lag” and chronic sleep deprivation. In this talk I will discuss some of the neurobiological changes that result from these disruptions during adolescence and how they lead to increased risk for substance use.

Methods: These studies are part of the translational NIDA P50 Center for Adolescent Reward, Rhythms and Sleep (CARRS). In this talk I will focus on the rodent projects in the center which perform circadian and sleep disruptions in adolescent rats followed by molecular studies (RNA sequencing), behavioral studies (5-choice task and nicotine or THC self-administration) and ex vivo electrophysiology (patch clamp recordings). These studies have focused on the nucleus accumbens (NAc), an important area of the brain which controls reward and motivation.

Results: We have found that adolescent rats exposed to chronic circadian rhythm disruption have an increase in nicotine self-administration which ultimately results in a loss in the normal diurnal rhythm in drug taking (n = 10−15; p < 0.05). We also find significant correlations between sleep duration and motivation for drug taking in progressive ratio tasks, as well as associations between circadian and sleep measures and impulsivity and cognitive performance. When we record from the NAc we find an increase in excitability in medium spiny neurons with rhythm and sleep disruption, as well as a loss in normal rhythmicity in neuronal activity (n = 10−12; p < 0.05). Ongoing RNA sequencing studies indicate a range of changes in NAc in response to sleep deprivation including an increase in orexin receptor expression (Hctr2) and preliminary studies find that orexin 2 antagonists decrease drug self-administration. Results from additional behavioral and electrophysiology studies across time of day as well as ongoing single nucleus sequencing will also be presented.

Conclusions: In conclusion, I will present novel, unpublished data from the CARRS Center which sheds light on the impact of sleep and circadian rhythms disruptions during adolescence on cellular function, neuronal activity and addiction-related behavior. These results are important in understanding how these factors influence the developing brain and the importance of interventions that target these factors.

Disclosure: Nothing to disclose.

54.3 Next-Day Sleep Disruption Reduces Conditioned Fear Recall in Male and Female Mice

Allison Foilb

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

Background: Sleep problems are characteristic of many psychiatric conditions, including Generalized Anxiety Disorder (GAD) and Post-Traumatic Stress Disorder (PTSD). The relationship between trauma and sleep is reciprocal—with sleep loss exacerbating symptoms of these disorders. Existing preclinical research has demonstrated an effect of sleep deprivation on fear memory, but has largely focused on disruption of memory consolidation immediately after exposure to trauma, which has limited real-life therapeutic benefits. Here we examined the effects of sleep disruption on neuroplasticity and prior to fear recall in male and female mice. These findings present an interesting route for future clinical interventions with sleep disruption.

Methods: Adult male and female mice were used to study the effects of 6-h sleep disruption. The gentle stimulation method of sleep disruption was used, which encourages voluntary locomotor activity. Mice were allowed to sleep normally after fear conditioning (for 18 h), and sleep disruption began the next day at the start of lights-on. In a separate cohort of mice, we also utilized qPCR to measure neuroplasticity after 6 h sleep disruption with mRNA encoding brain derived neurotropic factor (BDNF), and enzyme-linked immunoassay (ELISA) to measure corticosterone (CORT) as an assessment of stress induced by this sleep disruption paradigm.

Results: We found that sleep disruption with the gentle stimulation method immediately prior to fear recall testing leads to significantly reduced fear expression compared to mice allowed to sleep normally (n = 14/condition, 7 male, 7 female; P’s < 0.01). Reduced fear expression in sleep disrupted mice compared to controls persists the next day as well (P’s < 0.05). A test of reinstatement of fear after extinction indicates that the effect of sleep disruption at this time point is not a disruption of fear memory consolidation. Analysis of inter-trial intervals during the fear recall test found no differences between sleep disrupted and control mice, indicating that the effect is not purely locomotive. One possible mechanism for this fear reduction is increased neuroplasticity. While we do not observe elevated CORT levels after sleep disruption, we do see increased BDNF in the basolateral amygdala (BLA) in sleep disrupted males and females compared to controls (n = 15−16/condition, 7−8/sex; P < 0.01).

Conclusions: This investigation provides new and exciting evidence that sleep disruption may be used to reduce fear at a timepoint after fear memory consolidation. Our data indicate that this may occur through a mechanism involving enhanced neuroplasticity in fear-related brain regions, such as the BLA. This inhibition of fear and enhancement of neuroplasticity by sleep disruption could lead to novel clinical treatments for fear- and threat-related disorders. An improved understanding of the neural and molecular mechanisms by which sleep disruption methods can reduce fear responding is necessary to reach this broad therapeutic goals.

Disclosure: Day Zero Diagnostics: Employee (Spouse/Partner).