ACNP 64th Annual Meeting: Poster Abstracts P292-583

January 12–15, 2026

Nassau, Bahamas

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

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

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

P292. Using single cell multi-’omics to discover novel factors in the human brain involved in co-occurring depression and opioid addiction

Madeline Fish, Vijay Cherupally, Snehal Sambare, Chen Fu, BaDoi Phan, Jill Glausier, David Lewis, Marianne Seney, Andreas Pfenning, Ryan Logan

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

Background: The opioid epidemic is a critical public health crisis, with high overdose and relapse rates in opioid use disorder (OUD) and co-occurrence with major depressive disorder (MDD) in ~36% of cases. While shared genetic risk loci link these disorders, cell type–specific mechanisms in the human brain remain unclear. Emerging evidence implicates inflammatory signaling and synaptic plasticity as dysregulated processes in both OUD and MDD, yet the molecular pathways altered in defined cell populations are understudied. We therefore examined the cell type specific transcriptional and epigenetic signatures associated with OUD, MDD, and co‑occurring OUD/MDD (COM) across two brain regions critical to affective processing and motivation.

Methods: We generated matched single‑nucleus RNA and chromatin accessibility profiles from postmortem subgenual anterior cingulate cortex (sgACC) and nucleus accumbens (NAc) using the 10x Genomics Multiome assay. The cohort included strict diagnostic groups for OUD, MDD, COM, and an unaffected comparison group, with verification that individuals in OUD or MDD groups did not have a lifetime diagnosis of the alternative disorder. A machine‑learning workflow (doublet/ambient removal, nucleus‑level outlier scoring, cross‑modal filters) and standardized metrics (complexity, mitochondrial content, TSS enrichment, FRiP) ensured high‑quality data across modalities. Cortical and striatal references facilitated clean, reproducible annotation of excitatory, inhibitory, and glial subpopulations. Differential expression was performed using limma-voom on pseudobulked data within each cell type with covariate control, empirical‑Bayes shrinkage, and FDR correction. Chromatin analyses (differential accessibility, peak‑2-gene links) nominated candidate regulators. Signal concordance between diagnoses was quantified with RRHO2, while pathway enrichment used Metascape, and UpSet intersections prioritized shared genes for meaningful visualizations.

Results: In the sgACC, we found evidence that individuals with OUD+/MDD+ exhibit unique pathology in microglia compared to individuals with only OUD or MDD. Specifically, the OUD+/MDD+ signature had 348 differentially expressed genes (DEGs), 90% of these unique compared to OUD or MDD alone. GSEA revealed that downregulation in synapse-related pathways in microglia is a distinguishing feature of having both disorders (i.e., unique pathology in OUD+/MDD+). Other glial populations were less altered in OUD+/MDD+ compared to controls (85 DEGs in astrocytes, 149 in oligodendrocytes, 79 in OPCs). Notably, threshold-free analyses revealed that, within astrocytes, the upregulated signal in OUD+/MDD+ was driven by OUD rather than MDD. Ongoing efforts are incorporating cell type-specific chromatin accessibility with these transcriptional findings. Analysis in NAc is still ongoing.

Conclusions: Across sgACC and NAc, matched single‑cell multi‑omics delineate robust, cell type–specific alterations linking OUD, MDD, and COM. sgACC exhibits a broad OUD‑driven signature with L6‑selective COM prominence and distinct pathway signatures; NAc shows MSN‑subtype‑dependent concordance and a glial shift toward COM/MDD. Signals converge on glial immune/interferon pathways and neuronal transcriptional‑elongation, proteostasis, synaptic‑plasticity, and bioenergetic programs. These data provide mechanistic hypotheses and candidate regulators for validation, and motivate sex‑aware, precision therapeutic strategies for co‑occurring depression and opioid addiction.

Keywords: Postmortem Human Brain Tissue, Opioid addiction, single-cell genomics, Major Depressive Disorder

Disclosure: Nothing to disclose.

P293. Acute stress augments cue-induced nucleus accumbens activation and nicotine-seeking in cigarette smokers: a pharmaco-imaging study

Atul Ganesh, Reem Tamimi, Emilie O’Mara, Mark Greenwald, Eric Woodcock

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

Background: Cigarette smoking cues have been shown to induce craving and motivate nicotine-seeking behavior. Visual smoking cues also often evoke brain activation in the medial prefrontal and orbitofrontal cortices (mPFC, mOFC), amygdala (Amg), anterior insula (aINS), and nucleus accumbens (NAcc). Separately, acute stress-induction reliably increases drug-seeking behavior, including nicotine-seeking, in preclinical and human laboratory studies. Cue reactivity and stress are widely studied in isolation but may interact to enhance nicotine use. Yet, little is known about the potential interaction of stress and cues among cigarette smokers and to our knowledge, no published studies have investigated the interactive effects of stress and cues on nicotine-seeking behavior. To address this scientific and clinically important question, cigarette smokers completed a visual cue reactivity paradigm during functional neuroimaging followed by a laboratory-based nicotine-seeking paradigm under two pharmacological dosing conditions: placebo vs. acute stress.

Methods: Out-of-treatment cigarette smokers (N = 21) completed two identical experimental sessions (11:00–16:00) on non-consecutive days: placebo vs. acute stress (within-subject random crossover design). An acute stress response (biomarker-verified) was evoked by pharmacological pretreatment with yohimbine (54mg PO) plus hydrocortisone (10mg PO) under double-blind conditions. During each dose condition, subjects completed a 60-min MRI scan (13:00–14:00) including a visual smoking cue reactivity paradigm during multi-echo functional neuroimaging (TEs = 15, 36, 56ms; TR = 2.83s; 2.9mm isotropic; 3T Siemens Verio). Self-reported craving and withdrawal (Questionnaire on Smoking Urges and Minnesota Nicotine Withdrawal Scale) were collected before and after cue exposure. Multi-echo fMRI data underwent standard pre-processing (adapted from afni_proc) before denoising with tedana to isolate BOLD signal. Following group-level analyses of the smoking cues > neutral images contrast, eigenvalues were extracted from five a priori regions of interest (ROIs; mOFC, mPFC, NAcc, Amg, and aINS) for each subject and dose condition. After imaging, subjects completed a computerized money vs. drug (cigarette puffs) progressive ratio choice task (30-min). Earned cigarette puffs were smoked upon task completion (video-verified; preferred cigarette brand). Repeated measures analyses of variance (rmANOVA) evaluated within-subject effects of Stress, ROI, and Stress*ROI on cue-induced brain regional activation. rmANOVA evaluated Cue, Stress, and Cue*Stress effects on self-reported craving and withdrawal (separately). Finally, rmANOVA evaluated Stress effects on nicotine-seeking behavior. Nicotine dependence level (FTND) was evaluated as covariate in each model and retained when significant.

Results: Smokers (mean = 28yrs, 21–34yrs; 86% male; 71% African American) were ‘moderately’ dependent (FTND score = 6 ± 2; 17 ± 6 cigarettes/day). Relative to placebo levels, stress increased nicotine-seeking (p = 0.024), controlling for FTND. Relative to pre-cue levels, smoking cues increased cigarette craving and nicotine withdrawal symptoms (ps < 0.001; Cue effect), whereas Stress and Stress*Cue effects were non-significant (ps > 0.30). Cue-induced brain activation exhibited a significant ROI effect (p = 0.003) and a Stress*ROI interaction (p = 0.03) indicating that stress amplified cue-induced activation in the NAcc (p = 0.04) and mOFC (p = 0.055), but not in the mPFC, aINS, or Amg (ps > 0.15). During placebo, partial correlations indicated that greater cue-induced NAcc activation correlated with greater nicotine-seeking behavior, controlling for FTND (r = 0.50, p = 0.05). During stress, greater aINS activation correlated with greater nicotine-seeking, controlling for FTND (r = 0.52, p = 0.04).

Conclusions: Our findings indicate that: a) acute stress potentiated nicotine-seeking among out-of-treatment smokers; b) smoking cues increased craving and withdrawal; c) stress modulated cue-induced brain activation in a region-specific manner; and d) nicotine-seeking behavior correlated with cue-evoked activation in the NAcc during placebo and the aINS during stress. Our findings show that, in the absence of stress, cued-induced NAcc activation was most closely associated with nicotine-seeking, suggesting a brain-behavior link between appetitive motivation and cigarette smoking. Acute stress increased nicotine-seeking and augmented visual smoking cue-induced neural activation in a region-specific profile (NAcc, mOFC) consistent with enhanced motivational properties. During stress, nicotine-seeking was most closely linked with the aINS activation suggesting an interaction between cue salience and interoceptive signals that contributed to amplified smoking behavior. In conclusion, this study validated a novel pharmaco-imaging experimental model for investigating acute stress effects on brain function and drug-seeking behavior, and findings showed a region-specific pattern by which stress altered cue reactivity and was linked to stress-potentiated smoking behavior with implications for targeted network-based interventions, i.e., neuromodulation.

Keywords: cigarette smoking, drug seeking, Acute Stress, Cue Reactivity, BOLD fMRI signal

Disclosure: Nothing to disclose.

P294. Identification of long non-coding RNA expression signatures across different brain regions in alcohol use disorder

Hooriyah Rizavi, Mark Maienschein-Cline, Subhash Pandey

Center for Alcohol Research in Epigenetics, University of Illinois at Chicago and Jesse Brown VA Medical Center, Chicago, Illinois, United States

Background: Alcohol Use Disorder (AUD) is a debilitating psychiatric disorder that impacts millions of individuals worldwide. Long non-coding RNAs (lncRNAs) represent the dark side of the genome that includes a class of RNA transcripts exceeding 200 nucleotides in length without a functional protein-coding capacity. Notably, they exhibit greater tissue specificity than coding genes, are highly expressed in the central nervous system, and play an essential role in gene expression regulation at transcriptional, post-transcriptional, and epigenetic levels. Despite a growing body of research linking lncRNAs to neuropsychiatric disorders, their role in AUD is less explored. Given the diverse functional roles of lncRNAs, their dysregulation may represent a molecular mechanism underlying the vulnerability to and progression of AUD. Here, we investigated the expression profile of lncRNAs in the prefrontal cortex (PFC) using RNA-seq, followed by validations of emerging lncRNAs in the PFC and other brain regions of AUD subjects.

Methods: For RNA-seq, total RNA was extracted from the PFC (Brodmann area 10) of 30 AUD (20 males and 10 females) subjects and demographically matched 30 control subjects (20 males and 10 females) aged 40–78 years. The postmortem brain samples were received from the New South Wales brain tissue resource center, Sydney, Australia. We used the ribosomal RNA (rRNA) depletion protocol to capture the various populations of RNA species, both coding and non-coding. Illumina’s Ribo-Zero Plus was used for rRNA removal, and libraries were constructed using the TruSeq Stranded kit. The libraries were sequenced on NovaSeq X to generate paired-end reads for accurate alignment. Trimmed reads were aligned to the reference genome hg38 in a splice-aware manner using the STAR aligner, and for annotations, we used ENSEMBL gene and transcript annotations, which include non-coding RNAs in addition to mRNAs. The expression level was quantified using FeatureCounts, and differential expression statistics (fold-change and p-value) were computed using edgeR on raw expression counts obtained from quantification. Validation of lncRNAs in the PFC and further expression profiling in the insular cortex and amygdala were performed using qRT-PCR on RNA isolated from the subjects in the same cohort used for RNA-Seq.

Results: A comprehensive bioinformatic analysis of the transcriptome revealed a total of 37,256 gene features, including 18,749 protein-coding, 7,729 lncRNA, and 10,305 pseudogenes. LncRNAs were filtered using the following transcript biotypes: lincRNA, processed_transcript, processed_transcript_retained_intron, sense_intronic, and retained intron. Significant lncRNAs were determined based on an FDR ≤ 0.2, which resulted in 550 lncRNAs, out of which 193 have a positive fold change in expression and 357 have a negative fold change in expression. For validation, sixteen lncRNAs (H19, MALAT1, NEAT1, FAM225A, FAM225B, FAM27C, LINC01503, C9orf106, SNHG11, SNHG17, FTX, OLMALINC, MIRLET7DHG, MIR3681HG, and lnc-SEPT7-1) were selected based on: significant expression changes, possible functional annotation, sequence conservation between species, protein coding potential, interaction with miRNAs or proteins of interest and potential biological significance. Most lncRNAs displayed significant fold changes in the PFC, where H19, MALAT1, FAM225A, FAM225B NEAT1, FAM27C, LINC01503, C9orf106 and SNHG17 were increased and FTX, OLMALINC, MIRLET7DHG, MIR3681HG, and lnc-SEPT7-1 were decreased in AUD subjects. In the insular cortex, only H19, MALAT1, FAM225A, and FAM225B were increased, and in the amygdala, H19, MALAT1, FAM225A, and FAM225B were increased, while FTX, OLMALINC, and MIR3681HG were decreased. SNHG11 remained unchanged in all brain regions. Additionally, Spearman correlation analysis assessed the strength of association between clinical variables like AUDIT-C score, years of alcohol use, age of first alcohol use, and grams of alcohol consumption with fold changes in the expression of the lncRNAs in all brain regions. Correlation analysis showed a significant positive correlation of AUDIT-C score with MALAT1, FAM225A, and FAM225B expression across PFC, insular cortex, and amygdala. Whereas grams of alcohol consumption were positively correlated with FAM225A and FAM225B in the PFC, insular cortex, and amygdala.

Conclusions: Our data show differential expression of lncRNAs in three key brain regions of AUD subjects. Of particular interest, we observed a significant upregulation of lncRNA FAM225A, FAM225B, H19, and MALAT1 in all three regions. Our findings also highlight distinct regional lncRNA differential expression patterns in AUD and their correlation with specific AUD clinical features. Furthermore, our data suggest that the discovery of novel lncRNAs that are impacted in several brain regions may be of interest to serve as potential therapeutic targets for AUD (Supported by NIH-NIAAA P50AA022538, UO1AA-019971, U24AA024605 [NADIA], and by the VA Senior Research Career Scientist award to SCP).

Keywords: long noncoding RNA, alcohol use disorder, prefrontal cortex, Amygdala, insular cortex

Disclosure: Nothing to disclose.

P295. Cocaine use disorder-related prefrontal dysfunction and childhood trauma-related hippocampal dysfunction during social navigation

Mathew Schafer, Shaun Kohli, Jasper Van Oort, Philip Kamilar-Britt, Vyoma Sahani, Yasmin L. Hurd, Daniela Schiller, Keren Bachi

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

Background: Social behavior can influence the risk, trajectory and severity of substance use disorders, yet there remain gaps in knowledge of the related neurobehavioral underpinnings. Successfully navigating social interactions may depend on social mapping—the compact neural representations of social relationships along abstract dimensions. The hippocampus constructs social maps; the prefrontal cortex may use these maps in decision-making. Cocaine use disorder (CUD) and childhood trauma, commonly co-occurring, feature concurrent social dysfunction and neural abnormalities in these regions, with prefrontal cortex alterations common in CUD and hippocampal alterations common in childhood trauma. We use a naturalistic social interaction task to test whether social relationships are tracked differently in the prefrontal cortex of individuals with CUD relative to healthy controls (HC), and differently in the hippocampus as a function of childhood trauma.

Methods: During functional magnetic resonance imaging (fMRI), participants (CUD n = 32 and HC n = 38) completed the social navigation task where they interact with fictional characters to accomplish social goals (e.g., find a job). Participants’ interaction decisions with each character were modeled as a series of locations (i.e., a trajectory) through a two-dimensional abstract space of affiliation by power to represent the “social navigation” of relationships. To probe the neural correlates of this decision-making, we used representational similarity analysis. For each participant, we estimated the voxel-wise correlations between character-level social trajectories and patterns of brain activity, controlling for the temporal distances between characters. This analysis captured the extent to which similar social trajectories across characters were represented by similar neural patterns, irrespective of the temporal appearance of characters in the narrative. We used regions-of-interest analyses to test two predictions: 1) the social trajectory effect for CUD participants would be reduced in prefrontal cortex, specifically in right inferior frontal gyrus, a region involved in processing navigational demands 2) individuals with higher childhood trauma would have a reduced social trajectory representation in left hippocampus, given the region’s sensitivity to stress and its established role in this task. To test these predictions, we modeled the effect of diagnostic group (CUD vs. HC) and childhood trauma (Childhood Trauma Questionnaire, CTQ) on the trajectory analysis estimates. Additional functional connectivity analyses are in progress.

Results: The behavioral estimates of social location tracked post-task subjective placements of the characters, suggesting the behavioral locations tracked perceptions of social locations (permutation p < 0.001; there was no effect of diagnosis or CTQ score). As predicted, in the fMRI analysis there was a smaller pattern similarity effect in CUD than HC in the right inferior frontal gyrus (t = −2.16, right-tailed p < 0.05), suggesting altered engagement of this region during social navigation in CUD. Also as predicted, childhood trauma negatively correlated with the pattern similarity effect in the left hippocampus (t = −2.88, right-tailed p < 0.01), a region previously shown to track abstract social locations. This effect was independent of CUD status.

Conclusions: Our findings suggest that both CUD-related prefrontal dysfunction and childhood trauma-related hippocampal dysfunction extend to the neural tracking of social relationships. Childhood trauma had similar reductions in left hippocampal effects in both CUD and HC participants, suggesting that the effect of trauma on representing social relationships is separable from the effects of cocaine addiction. CUD and childhood trauma may thus impact the representation of social navigation in different but potentially complementary ways. More work is being done to fully characterize these effects – such as the functional interactions between these regions in both CUD and childhood trauma, as well as implications for real-world social behavior and clinical outcomes. Advancing knowledge of social neurobehavioral phenotypes may inform the development of novel social-target interventions in substance use disorders.

Keywords: social cognition, Childhood trauma, cocaine use disorder, Functional MRI (fMRI), Cognitive Neuroscience

Disclosure: Nothing to disclose.

P296. Leveraging fMRI alcohol cue-reactivity to determine data-driven TMS targets for alcohol use disorder: a case for DMPFC and IPL stimulation

Daniel McCalley, Claudia Padula

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

Background: Delivering Transcranial Magnetic Stimulation (TMS) to a specific brain region as identified through individual fMRI data is a promising strategy to enhance TMS treatment response beyond conventional TMS for psychiatric disorders. In Alcohol Use Disorder (AUD), elevated brain fMRI response to alcohol, relative to neutral cues, is a well-established predictor of relapse. A retrospective analysis of a TMS clinical trial for AUD found that when TMS intersected with an individual’s peak brain response to alcohol cues, individuals were 5.6 times more likely to remain abstinent, relative to sham. In a first step toward prospectively evaluating fMRI-guided TMS for AUD, we aim to generate a robust, clinically meaningful pipeline to derive fMRI-based TMS targets for patients with AUD.

Methods: In an archival sample of Veterans with AUD (n = 87, 23F), task-based fMRI was used to define the cortical location wherein brain response to alcohol (vs. neutral) cues was greatest as the individual TMS target. Target search was restricted to regions reachable by conventional TMS coils (<30mm from the scalp). K-means clustering was performed on the group of coordinates to cluster TMS targets in adjacent anatomical regions. Given we expected a variable topological spread of TMS targets, broad-field TMS electrical field models produced by BrainsWay H-coils (SimNibs v.45) were created to identify how many targets are reached by delivering TMS to cluster centroids. In an exploratory analysis, clinical characteristics of individuals with TMS targets in a given cluster were evaluated. Following principal component analysis, composite scores for reward-, relief-, and compulsion-based drinking, as well as depression, anxiety, and AUD severity were entered as predictors in a conditional binary logistic regression to identify clinical characteristics which may differentiate membership in a TMS target cluster.

Results: K-means clustering yielded an optimal 2-cluster solution with centroids in the right DMPFC and right inferior parietal cortex (Silhouette score, k = 0.55). TMS electrical fields, targeted to DMPFC or IPL centroids intersected with 72.0% of frontal targets and 75.0% of parietal targets with a physiologically relevant field strength (>35 Volts/meter). The model predicting TMS target membership in one of two clusters (DMPFC or IPL) was significant (X2 = 8.7, p = 0.03), correctly identifying 69.5% of cases. Individuals with higher composite compulsion or reward scores were 2.5 (p = 0.009) and 2.4 (p = 0.02) times more likely to have their peak brain response to alcohol cues in the DMPFC or IPL, respectively.

Conclusions: Individually derived TMS targets can be clustered generally within the DMPFC and the IPL, and membership within each cluster can be distinguished by composite scores representing compulsive and reward-based drinking scores. These results demonstrate that broad TMS electrical fields, can capture 70–75% of TMS targets. These preliminary results suggest that individuals with higher compulsion-based reasons for drinking may be better suited for DMPFC TMS, while individuals with higher reward-based motivation to drink may benefit from IPL stimulation. Two ongoing clinical trials from our group will evaluate the relative clinical influence of specifically delivering TMS to an fMRI target with a precise figure-of-eight coil or delivering broad-field TMS with H-coils.

Keywords: Alcohol Use Disorder - Treatment, transcranial magnetic stimulation (TMS), Task-based fMRI

Disclosure: Nothing to disclose.

P297. Effects of GHSR blockade through PF-5190457 on alcohol cue reactivity and peripheral biomarkers in people with alcohol use disorder: results from randomized controlled studies

Ryan Tyler, Mehdi Farokhnia, Lorenzo Leggio

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

Background: The stomach-derived “hunger hormone” ghrelin is the endogenous activator of the growth hormone secretagogue receptor (GHSR) while the liver-expressed antimicrobial peptide 2 (LEAP2) serves as an endogenous GHSR blocker, demonstrating tight regulation of GHSR activity. GHSR modulates not only feeding and food-seeking but also alcohol use and seeking, as well as aspects of endocrine- and immune-related function. The GHSR blocker PF-5190457 has been investigated as a potential novel pharmacotherapy for alcohol use disorder (AUD).

In an initial, phase 1b, single-blind, placebo-controlled human laboratory study, PF-5190457 was found to be safe and well-tolerated in currently heavy drinking individuals, even when co-administered with alcohol. Furthermore, PF-5190457 reduced alcohol craving in a “bar-lab” cue reactivity experiment measured by the alcohol urge questionnaire (AUQ).

Methods: A subsequent double-blind, placebo-controlled phase 2a crossover study in abstinent, treatment-seeking AUD participants (N = 29; 8 females) tested PF-5190457 dosing over 5+ days. Blood was collected daily prior to dosing (T1). On days 4+, behavioral testing occurred post-dosing, followed by an additional blood collection approximately 2 hr post dosing (T2). Pharmacokinetic (PK: PF-5190457 and its active metabolite PF-6870961) and pharmacodynamic (PD: comprehensive panel of endocrine and immune markers) parameters were assessed. Exploratory analyses examined relationships between PK, PD, and behavioral measures.

Results: PF-5190457 and PF-6870961 concentrations peaked at T2 and were elevated at T1 under drug vs. placebo. LEAP2 levels were reduced under drug vs. placebo [F (1, 389.11) = 14.82, p < 0.001] but rebounded on the first washout day. PF-5190457 dramatically reduced growth hormone (GH) at T2, followed by an elevation of GH at T1 the following day [Drug: F (1, 400.27) = 22.05, p < .001]; [Drug x Day: F (5, 400.15) = 3.08, p = 0.01]. No other endocrine or immune markers differed significantly between drug and placebo, demonstrating overall negligible effects of the drug on endocrine and immune systems. PF-5190457 blood concentrations positively correlated with ghrelin at T1 and ΔGH levels from T1 to T2. GH at T1 negatively correlated with the number of calories selected during a cafeteria-like virtual reality (VR) buffet experiment under drug, but not placebo conditions. Both PF-5190457 and PF-6870961 concentrations positively correlated with cortisol levels at T2. Cortisol at T1 positively correlated with alcohol craving during the cue reactivity experiment under placebo, but not drug conditions. However, unlike the earlier study, these abstinent participants did not exhibit reduced alcohol cue-reactivity as an effect of PF-5190457.

Notably, the abstinent group reported approximately half the craving levels observed in the currently drinking group from the phase 1b study, suggesting a floor effect. As such, we conducted a secondary analysis combining the alcohol cue reactivity results from the two studies to explore whether PF-5190457 may reduce cue-induced craving in a sub-group of high “cue-reactive” participants [F (1, 114) = 12.78, p < 0.001]. These analyses revealed that in those who were “cue-reactors” (defined by AUQ scores ≥4 during the alcohol cue exposure relative to baseline), PF-5190457 significantly reduced craving compared to placebo regardless of drinking status (currently drinking vs. abstinent).

Conclusions: These findings reiterate the safety profile of PF-5190457 by showing negligible effects on the endocrine and immune systems. Furthermore, PF-5190457 may help reduce craving for alcohol in a sub-group of alcohol “cue-reactive” individuals regardless of drinking status.

Keywords: ghrelin receptor, Alcohol Use Disorder - Treatment, GHSR, Cue Reactivity

Disclosure: Nothing to disclose.

P298. Great expectations and addiction: characterizing and evaluating the impact of daily non-substance reinforcement in tobacco use

Jason Oliver, Taylor A. Love, Ashton R. Baltazar, Sarah Tonkin, Katelyn Romm, Kyle Simmons, F. Joseph McClernon

University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma, United States

Background: Historically, research on substance use and addiction has emphasized the role of drug valuation (i.e., expected drug reinforcement) on use behavior. While this emphasis has yielded critical advances in our understanding of the matter, understanding it in relation to other rewards can provide a more complete picture that better reflects real-world decision-making. Individuals can engage in a vast array of behaviors at any moment—each of which could be compatible or incompatible with substance use, and each with varying degrees of expected and actual reinforcement. Accordingly, recent models have integrated behavioral economic principles by framing substance use as an effort to maximize reinforcement amidst the sea of potential reinforcers that may be pursued at any given moment. This framing has yielded a number of important discoveries that generally converge on the notion that access to a rich array of non-substance reinforcers is associated with reduced substance use. Unfortunately, literature on this topic has relied heavily on retrospective reports with only limited efforts to understand momentary reinforcers as they occur in the daily lives of people who use substances. Our study sought to rectify this gap, while also accounting for potential distinctions between daily reinforcement expectations (e.g., do you expect to enjoy a good meal today?) and their actual impact (e.g., did you enjoy your dinner?). In a sample of adults reporting combustible cigarette use, we sought to: (1) Characterize the amount, intensity, and total reinforcement (i.e., amount x intensity) of common non-smoking rewards and (2) Evaluate the relationship between specific reward types and overall expectations for daily rewards on affect, stress, and cigarette craving.

Methods: People who smoke cigarettes daily (> 5 cigarettes/day; N = 46) and were not actively engaged in a cessation attempt were recruited for a parent study examining the effects of smoking abstinence on the neurobiology of reward processing. Eligible participants completed 14 days of ecological momentary assessment (EMA). Each morning, participants reviewed 12 non-substance reward categories (e.g., time with family/friends, electronic entertainment) and reported how much time they expected to spend engaged in each category and how rewarding they anticipated it being (e.g., 60 minutes watching a moderately enjoyable television show). Each evening, participants reported the actual amount of time engaged and how rewarding the experience was for each category that day. In addition, participants received six randomly occurring prompts each day, in which they reported their affect, craving to smoke, stress, and engagement in any of the 12 activities over the past 60 minutes. Descriptive analyses characterized actual and expected reinforcement (i.e., time x reward) from each activity. Multilevel models examined within-person associations of each activity with affect, stress, and craving, as well as day-level associations between expectancy violations (i.e., days with more or less actual reinforcement from evening prompts than was expected from morning prompts) and craving.

Results: Participants completed a total of 622 morning, 616 evening and 3,257 random prompts. Overall compliance rates were high (86.0%). Participants’ reports of actual daily reinforcement rates at evening prompts were highest for: (1) Spending time with family/friends (M = 5.6); (2) Accomplishing things (M = 4.0); and (3) Socializing with other people in-person or online (M = 3.1). Expected and actual reinforcement exhibited strong but imperfect associations across reward types (B’s 0.33–0.65; all p’s < 0.001). Analysis of random prompts unsurprisingly revealed that engagement with most rewards was associated with increased positive affect. These effects were strongest for attending events (B = 2.4, p < 0.001), physical activity (B = 1.1, p < 0.001), and spending time with family/friends (B = 0.9, p < 0.001). Effects on negative affect and stress were more mixed. At the day level, the discrepancy between actual and expected total reinforcement was a robust predictor of craving (B = −2.3, p = 0.022), such that days where people experienced reported greater reinforcement in the evening than they expected in the morning were associated with lower average craving.

Conclusions: Critically, these results suggest that days which meet or exceed reinforcement expectations are associated with lower average craving levels. They also offer insight into common sources of daily reinforcement for people who smoke cigarettes, including activities with the largest impact on positive affect. With respect to clinical implications, interventions designed to increase reward engagement (e.g., behavioral activation) have had mixed findings for substance use. The current findings enable more targeted approaches through focus on reinforcers likely to yield the greatest benefit. Our findings also raise the possibility that interventions focused on increasing reward may also increase expectations and unintentionally yield cravings on days those expectations are not met. Future directions for this work include replication in a larger sample, over a longer time period, and relating these findings to clinical outcomes, as well as a deeper examination of the contexts when specific activities do or do not take place concurrently with cigarette smoking.

Keywords: Smoking, Reward, Expectancy

Disclosure: Jazz Pharmaceuticals, Employee, Spouse/Partner

P299. OptiMAT smartphone intervention for MOUD: patient acceptability and neuroimaging task validation

Andrew James, Ronald Thompson, Mary Bollinger, Deborah Hasin, Michael Mancino, Clint Kilts

University of Arkansas for Medical Sciences, Little Rock, Arkansas, United States

Background: We developed OptiMAT (“Optimizing MAT”) as an adjunctive smartphone app intervention to reduce opioid relapse among patients receiving medication-assisted treatment (MAT) for opioid use disorder (MOUD). OptiMAT is theorized to improved treatment outcomes through daily tracking of opioid use, cravings, and mood; daily personalized feedback; and just-in-time GPS-triggered interventions when entering areas of personalized risk for opioid relapse. We are conducting a 6-month RCT to evaluate OptiMAT among patients receiving MOUD across 7 clinics in 4 Arkansas cities. The RCT includes an optional longitudinal neuroimaging sub-study to model changes in brain functional organization during recovery from OUD.

Methods: To date, N = 108 participants (of N = 240 planned) have been enrolled into the OptiMAT RCT, with half randomly assigned to the arm evaluating MAT with adjunctive OptiMAT intervention (MAT+OptiMAT) and half assigned to the MAT only. The 6-month RCT will use percent urinalyses positive for opioid misuse as its primary outcome variable as described per our RCT pre-registration (PMC10071730). Participants assigned to the MAT+OptiMAT arm undergo an exit survey at the 6-month endpoint providing feedback about their experience. Additionally, N = 27 participants (of N = 60 planned) have enrolled in the neuroimaging substudy, which includes MRI scans at Intake, Month 2, and Month 4 to model longitudinal change in brain functional organization associated with recovery from OUD. The MRI substudy consists of three fMRI tasks: Counting Stroop (attentional bias for personalized opioid cues), the Methamphetamine and Opioid Cue Database (craving for visual opioid cues), and Monetary Incentive Delay task (for reward processing).

Results: To date, N = 20 participants assigned to the MAT+OptiMAT arm have completed the exit survey. Participant responses are overwhelmingly positive, with 75% Strongly Agreeing (5 on a 5-point Likert Scale) that OptiMAT helped them maintain their sobriety, 94% Strongly Agreeing that they liked the overall format and design of OptiMAT, and 88% Strongly Agreeing that OptiMAT would be beneficial for other patients receiving outpatient MAT. Participants also provided feedback to improve specific OptiMAT features. Finally, fMRI analysis of the Intake MRI session replicated attentional control, craving, and reward network activation patterns previously reported in the literature, but with greater individual variance than expected.

Conclusions: Our interim findings support the acceptability of OptiMAT as an adjunctive treatment for MOUD, as well as the feasibility for using fMRI to model longitudinal changes in functional brain organization as patients recover from opioid use disorder.

Keywords: Opioid treatment, smartphone apps, Human Neuroimaging, Treatment Acceptability, GPS

Disclosure: Nothing to disclose.

P300. Effects of short-term endogenous and exogenous ketosis on brain metabolites and alcohol responses in humans

Xinyi Li, Sunil Khokbar, Zhenhao Shi, Timothy Pond, Nora Foster, Samnang Kim, Mike Dattilo, Vijay Ramchandani, Henry Kranzler, Ravi Prakash Reddy Nanga, Corinde Wiers

University of Pennsylvania, Philadelphia, Pennsylvania, United States

Background: Previous preclinical and clinical studies suggest that both endogenous (i.e., ketogenic diet, KD) and exogenous (i.e., oral ketone supplementation, KS) ketosis have therapeutic potential in reducing alcohol craving, alcohol consumption, and signs and symptoms of alcohol withdrawal. While a 6-week KD intervention led to significantly elevated blood alcohol levels in response to an alcohol challenge in rats, a single dose of 25 g of KS reduced the intoxicating and rewarding effects of alcohol in both rats and humans, demonstrating differential effects of endogenous and exogenous ketosis on alcohol pharmacokinetics. Here, we compared the effects of a 3-day KD, KS, and standard American diet (SAD) on responses to an alcohol challenge in healthy human participants. Based on previous evidence suggesting that brain GABAergic signaling mediates the sedative effects of alcohol, we also examined brain GABA levels following the diet interventions.

Methods: In this open-label, cross-over study, healthy participants (n = 13, 8 female) received three 3-day interventions in randomized order: (1) KD, (2) SAD with thrice daily KS (10 g of ketones from D-beta-hydroxybutyrate acid and R-1,3 butanediol, Kenetik, Vitanav Inc.), and (3) SAD. On day 4 of the dietary interventions, participants underwent magnetic resonance spectroscopy at 7 Tesla for spectral editing measurements of GABA. Subsequently, responses to an oral alcohol challenge, adjusted for body weight and sex were assessed, targeting a breath alcohol concentration (BrAC) of 0.08%. BrAC, the Biphasic Alcohol Effects Scale (BAES; including the sedation and stimulation subscales), and the Drug Effect Questionnaire (DEQ) were repeatedly administered over 180 min post-alcohol consumption. Motor performance was evaluated using the Grooved Pegboard Task both before and 60 min after alcohol administration (at the approximately peak BrAC). Blood β -hydroxybutyrate (BHB) levels were monitored throughout the study day using finger pricks and commerically available monitor.

Results: Both the KD and KS elevated blood BHB levels (peak mean ± standard deviation: KD = 0.8 ± 0.5mM; KS = 1.2 ± 0.6mM; SAD = 0.2 ± 0.1mM; Intervention effect: F2, 169.8 = 39.6, p < 0.001), indicating nutritional ketosis. In response to the alcohol challenge, there was significant effect of diet intervention on BAES-sedation (F2, 167.4 = 3.9, p = 0.02); post hoc tests showed lower sedation with KD than SAD (p = 0.046), but no significant differences between KS and SAD. However, there were no effects of the diet intervention on BAES-stimulation (F2, 167.8 = 1.5, p = 0.20), BrAC (F2, 349 = 1.5, p = 0.20), DEQ effects of feeling alcohol, feeling high, disliking, liking, or wanting more (all p > 0.05), or performance on the Grooved Pegboard Task (F2, 58.0 = 0.5, p = 0.60). Similarly, there were no effects of diet intervention on brain GABA levels (F2, 17.5 = 1.2, p = 0.3).

Conclusions: We found that a short-term, 3-day KD and KS induced mild ketosis, with blood BHB levels exceeding 0.5 mM. However, these short-term dietary interventions did not significantly alter BrAC responses to an oral alcohol challenge or brain GABA levels. Longer dietary interventions may be needed to evaluate the neuropharmacological effects of ketosis on alcohol intoxication and brain metabolism.

Keywords: Alcohol Response, Ketogenic diet, GABA MRS

Disclosure: Nothing to disclose.

P301. Dose-dependent effects of cannabis edibles on simulated driving performance

Bernard Le Foll, Justin Matheson, Paulina Antwi, Madison Wright, Adam Zaweel, Esther Kim, Pamela Kaduri, Stefan Kloiber, Ahmed Hassan, Beth Sproule, Christine Wickens, Patricia Di Ciano, Bruna Brands

University of Toronto/CAMH/Waypoint, Toronto, Canada

Background: Recent meta-analyses have confirmed a robust association between cannabis and risk of collision involvement, though there remain several limitations with epidemiological studies, such as delays between law enforcement dispatch and blood draws. Human laboratory studies are the most rigorous way of evaluating causality in cannabis-impaired driving research and have generally found that smoked cannabis can decrease speed and lane control and increase reaction time. However, to date, very limited evidence exists to examine effects of edible preparations of delta-9-tetrahydrocannabinol (THC, the primary intoxicating component of cannabis) on driving. Thus, our primary objective was to determine the effects of a range of doses of cannabis edibles on key measures of simulated driving performance and behavior. We hypothesized that driving simulator performance would be dose-dependently decreased by edible THC administration.

Methods: The study was a within-subjects, double-blind, placebo-controlled, randomized and counterbalanced study assessing the effects of a range of doses of cannabis edibles on driving simulator performance. We included four dose conditions: placebo, low dose (2 mg), medium dose (10 mg), and high dose (20 mg). Healthy adults aged 19 to 45 years who used cannabis at least one day/week and consumed edibles at least once per month were recruited from Toronto, Canada. Participants meeting eligibility criteria came to the lab for a total of 10 sessions: an eligibility assessment, a practice session, and 8 testing sessions (involving four different randomly assigned and counterbalanced cannabis conditions along with a 24-hour session to test residual effects). Participants drove in a driving simulator at baseline and then 2 and 5 hours after exposure to THC edibles. During each drive, participants drove three scenarios: one scenario was driven on a rural road and included no distractor tasks; one scenario was identical to the first except that we added a secondary task (counting backwards by 3s from a large, randomly generated number); and a final scenario meant to capture reaction time, which was a simulator-based reaction time judgment task that had participants asked to stop when they see a stop sign directly facing them, but not to stop when they see a stop sign at an angle. Results were similar between single-task and dual-task conditions, so we present results from the single-task condition only. Our primary outcome measure was standard deviation of lateral position (SDLP, in cm), which was based on global performance throughout the entire driving scenario. Secondary driving outcome measures included mean speed (km/h), standard deviation of speed (SDS, km/h), maximum speed (km/h), and reaction time (RT, sec; measured in the reaction time task only).

Results: Forty participants completed at least one drug exposure visit and were included in the analysis (50% assigned female at birth; 50% white; mean [SD] 30.4 [6.9] years old). There was a significant main effect of THC condition on our primary outcome, SDLP (p < 0.001). SDLP was higher in the high-dose condition compared to the medium-dose condition (Δmean = 0.014, p = 0.005), the low-dose condition (Δmean = 0.027, p < 0.001), and placebo (Δmean = 0.033, p < 0.001). SDLP was also higher in the medium-dose condition compared to the low-dose condition (Δmean = 0.013, p = 0.007) and the placebo condition (Δmean = 0.020, p < 0.001). There was a significant main effect of THC on SDS (p = 0.005). SDS was significantly elevated in the high-dose condition compared to the medium-dose condition (Δmean = 0.40, p = 0.014), the low-dose condition (Δmean = 0.52, p < 0.001), and placebo (Δmean = 0.47, p = 0.004). The medium-dose and low-dose conditions did not differ from placebo or from one another. There was also a significant main effect of THC on RT (p < 0.001). RT was significantly elevated in the high-dose condition compared to the medium-dose condition (Δmean = 0.025, p = 0.003), the low-dose condition (Δmean = 0.018, p = 0.033), and placebo (Δmean = 0.034, p < 0.001). The medium-dose condition did not differ from the low-dose condition or placebo. However, RT was significantly higher in the low-dose condition compared to placebo (Δmean = 0.017, p = 0.047), which was the only driving outcome that showed any evidence of impairment at the low dose. There were no THC effects on mean speed or maximum speed, or any evidence of differences in performance at 2 vs. 5 hours after THC exposure.

Conclusions: In line with our hypothesis, SDLP was dose-dependently impaired by acute exposure to edible THC, with the greatest impairment at the highest dose (20 mg), and no significant impairment (i.e., no difference from placebo) at the lowest dose (2 mg). While there was no evidence of THC effects on mean speed or maximum speed, we found that standard deviation of speed was significantly impaired by THC at the highest dose only (20 mg). Reaction time was significantly impaired at the high dose and low dose of THC, which was the only significant evidence of driving impairment at the 2 mg dose. Taken together, our results demonstrate that simulated driving performance is significantly impaired by ingestion of THC edibles, especially at a 20 mg dose, and this impairment is primarily related to less ability to maintain a consistent lane position (i.e., increased SDLP) and speed (i.e., increased SDS).

Keywords: cannabis, Simulated Driving, Cannabis edibles

Disclosure: Indiva, Other Financial or Material Support, Self.

P302. Cue the change: baseline cue reactivity as a predictive biomarker of TMS treatment outcomes in cannabis users

Tonisha Kearney-Ramos, Daniel McCalley, Wiebke Struckmann, Seigo Ninomiya, Logan Dowdle, Gregory Sahlem

Duke University Medical Center, Durham, North Carolina, United States

Background: Drug cue-induced craving (CIC) is a major barrier to abstinence across substance use disorders (SUDs). Neuroimaging studies have consistently shown that drug cue exposure evokes elevated activity in frontal-striatal-limbic circuits involved in reward processing, with activation proportionate to craving and drug-seeking behavior. Across SUDs, studies suggest that repetitive transcranial magnetic stimulation (rTMS) to the left dorsolateral prefrontal cortex (L DLPFC) reduces CIC, supported by neuroimaging evidence showing reduced cue-related activation in these circuits. However, rTMS response is variable, with some individuals showing robust benefit and others little change. Our prior work targeting the frontal-striatal-limbic reward circuitry via the ventromedial prefrontal cortex in cocaine users demonstrated that baseline neural reactivity predicted rTMS effects. Using a network-level independent component analysis (ICA) approach, we found that higher baseline striatal cue reactivity predicted greater reductions in striatal activation after active (but not sham) inhibitory rTMS, while the opposite was true for lower baseline reactivity. These findings align with state-dependent models of rTMS. Here, we extend this analysis to cannabis users, a population in which such investigations have not yet been conducted. We examined rTMS-induced changes in brain networks and tested whether baseline striatal cue reactivity predicted treatment outcomes, evaluating the striatum as a biomarker of treatment response.

Methods: Seventy-one treatment-seeking individuals with cannabis use disorder (CUD) were randomized to active or sham rTMS targeting the L DLPFC (Beam-F3; 10 Hz; 2 sessions/visit, 2 visits/week; 20 sessions total) in a double-blind trial. All participants also received motivational enhancement therapy (MET). Cannabis use was monitored throughout, and functional MRI (fMRI) was acquired pre- and post-treatment. The final analytic sample included 40 participants (21 MUSC, 19 Stanford) with complete imaging. Spatial ICA was conducted on all 80 cue reactivity task fMRI datasets using Group ICA of fMRI Toolbox (GIFT) in MATLAB. Striatal network timecourses were extracted and entered into a general linear model (GLM) to estimate activation during cannabis cue vs. neutral cue conditions. The resulting beta weight provided a single metric of cue-related activation for the network as a whole. Beta weights were entered into linear mixed-effects models (LMMs) to evaluate associations between baseline cue reactivity and post-treatment change.

Results: Group ICA identified a striatal network comprised of bilateral caudate, putamen, and nucleus accumbens which was modulated by the cannabis cue reactivity task. In the LLM, no main effect of treatment group was observed, though trending (t(74) = −1.93, p = 0.058). However, when stratified by baseline cue reactivity, a brain state-dependent response emerged, mirroring our prior cocaine findings: individuals with high baseline cue reactivity showed treatment-induced decreases in striatal cue reactivity; whereas, those with low baseline reactivity showed increases. Importantly, this bimodal pattern was observed in both the active and sham groups, suggesting that the neural response to cannabis cues was modulated by baseline state rather than treatment assignment alone. Regression analysis confirmed that baseline reactivity strongly predicted change in cue reactivity (r = −0.64, p < 0.0001, R²adj = 0.37). Although both active and sham groups showed significant associations between baseline and treatment effects, the slope was numerically steeper for active rTMS, reflecting a greater predictive value of baseline state under active rTMS; however, the group × slope interaction did not reach statistical significance (p = 0.09).

Conclusions: To our knowledge, this is the first study to assess the network effects of rTMS in cannabis users, particularly evaluating state-dependent effects. These data demonstrate that the effects of rTMS on the neural circuitry of craving are not uniform and may depend on an individual’s baseline neural reactivity to cues, consistent with models of homeostatic metaplasticity. We found that baseline striatal cue reactivity robustly predicted treatment-related changes in cue processing, but that this relationship was present in both active and sham groups. This contrasts with our prior findings in cocaine users, where only the active group exhibited a predictive relationship. A key distinction is that all participants in the present study received concurrent behavioral treatment (MET), meaning the sham group was not treatment-free. This likely contributed to overlapping effects across groups and blunted the separation between active and sham in neural outcomes. Despite this, the consistent brain state-dependent effect across both datasets reinforces the value of baseline neural reactivity as a biomarker of treatment responsiveness. These findings underscore the need to incorporate individual variability into rTMS trial designs and interpretation, and highlight the striatum as a transdiagnostic, treatment-predictive biomarker in addiction neuroscience.

Keywords: transcranial magnetic stimulation (TMS), cannabis use disorder, Biomarker, fMRI, Network-Analysis

Disclosure: Nothing to disclose.

P303. Decoding etiological heterogeneity in AUD via subjective response profiling

Emma Childs, Joseph A. Lutz, Michael P. Bremmer, Hyerim Yang

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

Background: There is substantial qualitative and quantitative variation in the subjective experiences of alcohol that likely arises from individual differences in alcohol pharmacokinetics and pharmacology. Alcohol subjective effects are a key factor in alcohol involvement and the progression of drinking, making them an important focus for research into the etiological and clinical heterogeneity of AUD. Understanding the biological basis of variation in alcohol subjective effects and how subjective responses influence drinking progression may reveal novel approaches to prevent and treat disordered drinking. This research aimed to characterize subgroups of binge drinkers based on multidimensional profiles of subjective response (SR) to alcohol and to compare longitudinal changes in drinking between subgroups 1-year post-study.

Methods: Datasets comprised subjective responses and breath alcohol curves obtained during laboratory challenge sessions with 0 and 80mg% alcohol (double-blind administration, randomized order) in binge drinkers (N = 116, 55% male, AUD 0–5). Participants also completed behavioral impulsivity tasks (DDQ, Stop Task), and questionnaires to assess personality traits, alcohol involvement and expectancies, drinking motives, and family history of alcohol problems. Participants were contacted 1-year post-study to obtain information on alcohol use. We performed hierarchical cluster analysis on alcohol subjective response domains (derived from factor analysis of subjective scales) and examined group differences in demographic characteristics, personality, alcohol involvement, family history of alcohol problems, trait behavioral impulsivity, acquisition of alcohol conditioned responses (parent study aims), and longitudinal changes in drinking. We predicted two subgroups of drinkers defined by sensitivity to alcohol-induced stimulation vs. sedation, that would differ in frequency of binge drinking, family history of alcohol problems, and drinking progression (stimulant responders > sedative responders for all hypotheses).

Results: Cluster analysis revealed 3 subgroups that differed significantly on dimensions of alcohol SR (Stimulation, Sedation, Positive Mood, Drug Effects, ps < 0.001). Biphasic Responders (BiR, N = 40) reported brief stimulation as BrAC increased, followed by high prolonged sedation as BrAC declined. Stimulant Responders (StimR, N = 40) exhibited strong stimulation throughout the ascending BrAC limb, then sustained low levels of stimulation, minimal sedation, and prolonged anxiolysis. Low Level Responders (LLR, N = 36) exhibited blunted responses across all SR dimensions. Groups did not differ on sex or BMI yet demonstrated significantly different BrAC responses (p < 0.001, BiR > StimR > LLR), suggestive of differences in metabolic tolerance (i.e., pre-existing or acquired). LLR reported significantly higher maximum drinks on a single occasion (p < 0.05), earlier onset of regular drinking (p < 0.01), and a higher prevalence of family history of alcohol problems (p < 0.05). The subgroups did not differ in acquisition of subjective preference for an alcohol-paired context, though StimR failed to acquire a behavioral preference for the alcohol-paired context despite exhibiting the highest rewarding alcohol effects and alcohol-induced alcohol craving (ps < 0.01). There were no significant group differences in drinking progression 1-year post-study, yet male StimR tended toward greater increases in drinks per binge episode (p < 0.08).

Conclusions: The current analysis identified subgroups of binge drinkers who exhibit characteristic patterns of alcohol subjective response consistent with accepted theoretical models of AUD risk (the Low Level of Response and Differentiator Models). Our findings suggest that pharmacokinetic variation (due to metabolic factors) may play a role in distinct SR profiles and that individuals exhibiting an LLR profile of alcohol responses may be most susceptible to developing hyperkatifeia. In future studies, we aim to probe the pharmacological underpinnings of individual variation in alcohol SR, which may reveal novel biomarkers of AUD risk and also individualized approaches for prevention and treatment.

Keywords: Alcohol Subjective Response, heterogeneity, Risk and Resilience

Disclosure: Nothing to disclose.

P304. Altered brain state dynamics during a drug-themed movie in opioid use disorder

Natalie McClain, Greg Kronberg, Kathryn Drury, Nelly Alia-Klein, Rita Goldstein

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

Background: Drug addiction is characterized by heightened salience/value attributed to drug-related cues over alternative reinforcers, accompanied by impaired inhibitory control–especially in a drug-related context. Underlying these impairments is maladaptive recruitment of large-scale brain networks during drug-related processing. However, because traditional task-based studies focus on regional activation or pairwise connectivity during static and tightly controlled laboratory tasks, it remains unclear how coordinated patterns of whole-brain network activity unfold over time in real-world environments. We sought to investigate how time-resolved brain state dynamics are altered during the dynamic processing of a naturalistic stimulus (a drug-themed movie) in individuals with opioid use disorder (iOUD) as compared to healthy controls (HC).

Methods: Sixty five iOUD (12 women; 42 ± 10 years) and 31 HC (12 women; 41 ± 11 years) passively viewed the first 17 minutes of a drug-themed movie, ‘Trainspotting,’ during fMRI acquisition. The BOLD fMRI time series data was extracted for 14 canonical large-scale brain networks and a Hidden Markov Model (HMM) was used to identify discrete brain states (i.e., patterns of network activity and connectivity) from the network-level signals (temporally concatenated across all subjects). The number of HMM states, a free parameter, was selected by minimizing AIC under the constraint that all states were meaningfully occupied. Each state was then characterized by the mean fMRI signal across the 14 networks and the functional connectivity between them. Fractional occupancy, dwell time, appearance rate, and transition probability matrices were computed from the subject-level state paths to compare state dynamics between groups and scene types (drug and non-drug).

Results: Model selection results revealed that HMM inference with 12 states provided the optimal balance between goodness-of-fit and a meaningful engagement of all states during movie-viewing. Across all subjects, from the mean activation profiles, States 5 and 6 showed a relatively uniform fMRI signal across networks. The remaining states (1–4 and 7–12) exhibited distinct activation profiles, with fMRI signals loading preferentially onto networks supporting attentional control (States 1, 3, and 9), self-referential processing (State 4), visual-auditory processing (States 2 and 9) and language (State 3). The functional connectivity profiles revealed some states with weak inter-network coupling (States 2, 3, and 12), whereas others exhibited strong inter-network connectivity (State 6). Transition matrices revealed the dominance of self-transitions (e.g., State 1 to State 1), as well as a high probability of transitioning between a uniform activity state (State 5) and a high sensory network activity state (State 8).

In group comparisons, the iOUD exhibited higher fractional occupancy, dwell time, and appearance rate in states characterized by high salience network activity (State 1) and high default mode sensorimotor network activity (State 4; ps < 0.05); HC exhibited high state metrics in states marked by low salience and high executive control network engagement (State 2), high visuospatial network activity (State 9), and low executive control and visual-auditory network activity (State 12; ps < 0.05). Within iOUD, states defined by high executive control and visual-auditory network engagement (State 2), high default mode network activity (State 4), and low visuospatial and language network activity (State 7) were visited more frequently during the drug scenes as compared to the non-drug scenes (ps < 0.05).

Conclusions: These findings suggest that iOUD preferentially engage specific large-scale brain states during naturalistic drug cue exposure, namely those characterized by heightened salience and default mode/sensorimotor network activity–as compared to the states characterized by heightened executive control and visuospatial network activity that are preferentially occupied by HC. Altered dynamics of these brain states in iOUD may reflect maladaptive coordination of the networks underlying drug-biased salience attribution and impaired inhibitory control function in ecologically valid, drug-relevant contexts. In future analyses, we plan to probe state dynamics as a function of scene-induced craving, sex differences and hormonal effects in women with OUD.

Keywords: Functional MRI (fMRI), Opioid Use Disorder, Naturalistic Paradigms, Brain States

Disclosure: Nothing to disclose.

P305. Augmented neuronal DNA damage and somatic mutations in the human brain associated with opioid addiction

Tara Delorme, Jennifer Ziegenfuss, Esha Sircar, Marianne Seney, Consuelo Walss-Bass, Michael Lodato, Ryan Logan

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

Background: The genome is constantly subjected to internal and environmental mutagens, leading to DNA damage in the brain and other tissues. Uncorrected DNA damage can result in mutations or genomic variations that disrupt normal cellular function. While neuronal DNA damage has been linked to aging and neurodegeneration, our preliminary findings suggest a connection between opioid use disorder (OUD) and increased neuronal mutation rates. In our recent single-cell RNA sequencing study, we identified enrichment of DNA damage related pathways in the dorsal striatum of individuals with OUD and in rhesus macaques chronically treated with morphine. Using an independent cohort of postmortem human striatum, we aimed to validate these findings by examining 8-oxoguanine (8-oxoG), a mutagenic marker of oxidative DNA damage, and its repair enzyme 8-oxoG DNA glycosylase (OGG1). Unrepaired 8-oxoG lesions, or dysregulated OGG1 activity, can lead to permanent DNA mutations, directly linking oxidative stress to the accumulation of somatic mutations in neurons. These mutations, which occur preferentially in highly transcribed genes, can disrupt gene regulatory networks. Evaluating 8-oxoG, OGG1, and somatic mutation burden provides a framework for understanding how opioid exposure may drive neuronal DNA damage and contributes to OUD pathology.

Methods: Postmortem dorsal striatum brain tissue punches (n = 6 subjects per sex per group: neurotypical and OUD) were processed for immunofluorescence microscopy using antibodies targeting 8-oxoG and OGG1. Mean 8-oxoG and OGG1 immunofluorescence signal was quantified for 100 cells per case using Fiji (ImageJ) software. For somatic mutation analysis, nuclei from the dorsal striatum of the same subjects were sorted into cold 96-well plates containing 3µL of Cell Buffer (BioSkryb), lysed, and DNA was amplified using Primary Template-directed Amplification (PTA). Amplified DNA was purified with in-house carboxyl magnetic beads and quantified using the QuantiFluor dsDNA System (Promega). Quality was assessed by multiplex PCR targeting four genomic loci on different chromosomes and only samples positive at all loci proceeded to library preparation. Libraries were constructed following a modified KAPA HyperPlus protocol (ResolveDNA EA), skipping fragmentation, performing end repair + A-tailing on 500 ng input, and ligating adapters (SeqCap Adapter Kit, Roche). Libraries underwent bead cleanup, on-bead PCR amplification, and 300–600 bp size selection. Fragment size was verified using a 5300 Fragment Analyzer (Agilent Technologies), and libraries were sent to Psomagen for sequencing. Single-cell PTA-amplified genome libraries were sequenced on the Illumina NovaSeq 6000 (150 bpx2) at ≥30x coverage.

Results: Our data indicates increased average nuclear 8oxoG and OGG1 intensity in OUD neurons compared to neurons from unaffected matched controls (Mann Whitney, 8oxoG p < 0.001 and OGG1 p < 0.001). Nuclear 8oxoG intensity did not differ by neuron area for OUD, but higher 8oxoG intensity was associated with smaller neurons in controls (linear regression slope, p = 0.025). Higher OGG1 intensity was associated with larger neurons in both OUD (p < 0.001) and controls (p < 0.001). For somatic mutation analysis, nuclei were sorted using Fluorescence-Activated Nucleus Sorting (FANS) based on DAPI and NeuN staining. For each subject, 20 single nuclei and a bulk sample for positive control were collected. Of the nuclei that underwent PTA amplification, 57/60 NT control nuclei (95% pass rate) and 54/80 OUD nuclei (68% pass rate) passed quality control (QC; ≥4 multiplex PCR bands and > 500 ng DNA). The proportion of nuclei passing QC was significantly lower in OUD samples compared to controls (Fisher’s exact test, p < 0.05). Sequencing of the QC-passed nuclei is currently underway.

Conclusions: By characterizing cellular DNA damage in striatal neurons from individuals with OUD, we can gain insights into the molecular alterations driving the disorder. The accumulation of somatic mutations in neurons may disrupt gene regulatory networks essential for reward processing, potentially reinforcing the persistence and severity of addictive behaviors. Chronic opioid exposure may accelerate neuronal aging, with DNA damage and somatic mutations contributing to premature cellular decline. Assessing DNA damage proteins and somatic mutation burden provides insight into how opioids disrupt neuronal function and gene regulation and may identify novel biomarkers and therapeutic targets to improve treatment outcomes for individuals with OUD.

Keywords: Postmortem Brain Tissue, DNA damage response, somatic mutation, Dorsal striatum, DNA double strand breaks

Disclosure: Nothing to disclose.

P306. The role of cannabis use in smoking and alcohol outcomes in a cessation trial among adults who smoke and drink: secondary analyses of a randomized clinical trial

Emma Brett, Andrea King

The University of Chicago, Chicago, Illinois, United States

Background: Cigarette smoking is highly comorbid with other substance use, including alcohol and cannabis. It is well-established that those who drink alcohol at hazardous levels (e.g. heavy drinking = > 14 drinks per week for men and >7 for women, >1 heavy drinking episode monthly) have greater difficulty quitting smoking compared with those who do not drink at high levels. Research also suggests that cannabis use impedes cigarette smoking quit attempts, which is particularly concerning as rates of cannabis and tobacco co-use continue to rise with increased cannabis legalization and popularity of cannabis vaping. The goal of the current investigation was to conduct secondary analyses to determine whether cannabis use impacted smoking and/or drinking outcomes in the context of a combination patch and varenicline smoking cessation randomized trial in adults who smoke and drink heavily.

Methods: A double-blind, placebo-controlled, randomized clinical trial was conducted in N = 122 adults (Mage = 44 + 12.4 SD yrs; 45% female, 44% Black, 46% White) who reported heavy drinking and a desire to quit smoking. At baseline, participants reported smoking an average of 11.8 (SD = 6.6) cigarettes per day for a duration of 25.2 years (SD = 13.2). Regarding drinking, participants reported an average of 9.3 (SD = 7.6) past month binge drinking episodes at baseline and 57% met criteria for alcohol use disorder. Participants were randomized to receive either varenicline plus nicotine patch or placebo plus nicotine patch for 12 weeks. Primary outcomes published elsewhere revealed significantly higher smoking cessation rates in the active versus placebo condition at 12 weeks (end-of-treatment [EOT]). In the current study, we used logistic and linear regression models with sex assigned at birth as a covariate to examine cannabis use and treatment condition as predictors of continuous smoking abstinence, average number of cigarettes smoked per day, and past week heavy drinking days at end-of-treatment. This study was powered to detect effects of the intervention on primary and secondary outcomes and was pre-registered (NCT02859142).

Results: The majority of the sample (81%) reported lifetime cannabis use, with more than half (53%) reporting past year use and over a quarter (26%) use in the past week. On average participants who reported cannabis use had been using cannabis for 11.5 years. Current cannabis use and medication condition, but not their interaction, predicted the odds of smoking abstinence at 12 weeks such that those in the active condition (OR = 0.35, 95% CI [0.14, 0.83], p < 0.05) and those who were not current cannabis users (OR = 0.29, 95% CI [0.09, 0.94], p < 0.05) were more likely to show biochemically verified continuous smoking abstinence (X2[3] = 8.29, p < 0.05). Treatment condition and cannabis use status were not predictive of average number of cigarettes smoked per day at 12 weeks (p = .35). Regarding alcohol use, cannabis use (p < .01) was a significant predictor of greater number of heavy drinking episodes at 12 weeks (F[2,106] = 3.69, R2 = 0.07, p < 0.05). Biological sex was not related to study outcomes.

Conclusions: Cannabis use was very common in this sample of adults who smoked cigarettes and drank alcohol at risky levels. Adults who reported current cannabis use were significantly less likely to report smoking abstinence at end-of-treatment, though use did not appear to affect reductions in cigarette smoking among those who did not quit. Cannabis use was also a significant predictor of heavy drinking at 12 weeks. Results indicate that cannabis use may interfere with successful quitting and/or harm reduction goals within the context of pharmacotherapy treatment for smoking and heavy alcohol use. Providers and researchers should assess for cannabis use within clinical practice and clinical trials and intervene as appropriate. More research is needed to better understand relationships between poly-substance use and whether individuals are willing to modify their cannabis use to aid in quitting smoking.

Keywords: nicotine addiction, alcohol use disorder, Smoking Cessation, Cannabis-Tobacco Co-use, Randomized-Controlled Trial

Disclosure: Nothing to disclose.

P307. Palmitoylethanolamide mitigates stress-induced craving and pain in opioid use disorder: preliminary evidence from a randomized pilot study

Wilson Julia, Rajasekera Therese, Fida Doruntina, Gonsalvez Irene, Suzuki Joji, Graziano Pinna, Primavera Spagnolo

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

Background: Opioid use disorder (OUD) is marked by high relapse rates and persistent symptoms such as heightened stress reactivity, negative affect, and amplified pain sensitivity—factors that powerfully drive craving and relapse through processes of negative reinforcement. Unfortunately, current medications for OUD (MOUD) do not adequately address these opioid-induced aversive states and their neurobiological substrates. Palmitoylethanolamide (PEA), an endocannabinoid-like lipid that acts as an endogenous agonist of peroxisome proliferator-activated receptor-α (PPAR-α), exhibits anxiolytic, analgesic, and mood-enhancing properties. This pilot study evaluated PEA as an adjunctive therapy for individuals with OUD maintained on MOUD.

Methods: In a double-blind, randomized, placebo-controlled trial, participants with OUD receiving methadone or buprenorphine were assigned to PEA (1200 mg/day) or placebo for 21 days. Stress-induced craving and pain were assessed at baseline and post-treatment, using an experimental paradigm combining psychosocial and physical stressors. Mood and affective responses were evaluated as secondary outcomes, along with occurrence of adverse events and compliance were also examined. Data were analyzed using independent-samples t tests and repeated measures ANOVA, with significance set at p < 0.05.

Results: Twelve participants completed all procedures (PEA = 6, placebo = 6). PEA was associated with trend-level reductions in stress-induced craving (p = 0.06) and pain perception (p = 0.07) compared to placebo. No group differences were observed in adverse events; only two mild, non-drug-related events occurred overall. Compliance was high, with treatment adherence exceeding 98% in both groups.

Conclusions: Preliminary evidence suggests that PEA, which exerts PPAR-α–mediated anti-inflammatory and neurosteroidogenic activity, may reduce stress-induced craving and pain in OUD. These findings highlight PEA’s potential as a novel, safe adjunctive therapy to complement MOUD and target largely unexplored mechanisms underlying relapse risk.

Keywords: opioid use disorder, Opioid craving, Acute and Chronic Stress, inflammation, Pain

Disclosure: Nothing to disclose.

P308. Sleep disturbance is associated with greater subjective and neural negative emotionality in people with alcohol use disorder

Erica Grodin, Dylan Kirsch, Wave-Ananda Baskerville, Lara Ray

UCLA, Sherman Oaks, California, United States

Background: Sleep problems in alcohol use disorder (AUD) are ubiquitous. Sleep disturbances are associated with reward dysregulation, increases in negative emotionality, and poor executive functioning. The Addictions Neuroclinical Assessment (ANA) is a trans-diagnostic, neuroscience-based framework that is designed to help explain heterogeneity in AUD. The ANA encompasses three domains: incentive salience, negative emotionality, and executive function, which reflect the processes disrupted in addiction. Given that sleep problems have been previously associated with these processes, the present study investigated the impact of sleep disturbance on neuroclinical phenomenology and functional brain correlates in individuals with AUD.

Methods: In study one, individuals with AUD (n = 115; 62M/53F, agem = 39.39) completed the Pittsburgh Sleep Quality Index (PSQI) and a phenotypic battery. Participants were classified as “good sleepers (PSQI total score ≤ 5; n = 42) or “poor sleepers” (PSQI total score > 5; n = 73). ANA domains of incentive salience, negative emotionality, and executive function were derived and used to assess neuroclinical phenomenology. A subset of participants (n = 55) completed functional neuroimaging paradigms to assess incentive salience (alcohol cue reactivity) and negative emotionality (negative pictures). In study two, individuals with AUD (n = 102; 61M/41F; agem = 44.26) completed the Insomnia Severity Index (ISI) and a phenotypic battery. Participants were classified as having no insomnia (ISI score < 8; n = 47), sub-clinical insomnia (ISI score 8–14; n = 38), or clinical insomnia (ISI score > 14; n = 17). ANA domains of incentive salience and negative emotionality were derived and used to assess neuroclinical phenomenology. In both studies we tested (1) total sleep disturbance as a predictor of ANA domains using multiple linear regression; and (2) sleep disturbance group differences in ANA domains using ANCOVAs. In study one, sleep disturbance group differences in the neural substrates of the incentive salience and negative emotionality ANA domains were also tested through FSL’s FLAME in FEAT (thresholded at Z > 3.1, p < 0.05).

Results: In study one, PSQI total score (B = 0.09, β = 0.29, p < 0.001), number of AUD symptoms (B = 0.14, β = 0.03, p = 0.002), and cannabis use days (B = 0.003, β = 0.27, p = 0.002), predicted higher negative emotionality scores. PSQI total score did not predict incentive salience (p = 0.28) or executive function (p = 0.62) scores. There were group differences in negative emotionality domain scores (F = 6.22, p = 0.01), such that poor sleepers had higher scores compared to good sleepers. There were no group differences in incentive salience (p = 0.98) or executive function (p = 0.90) domain scores. For the neuroimaging subsample, there were group differences in the neural substrates of negative emotionality, such that poor sleepers had greater negative image elicited activation in the bilateral medial prefrontal cortex, anterior cingulate cortex, frontal pole, and posterior cingulate cortex, compared to good sleepers (Zs > 3.55, ps < 0.001). There were no group differences in the neural substrates of incentive salience. In study two, ISI total score (B = 0.08, β = 0.44, p < 0.001) and number of AUD symptoms (B = 0.18, β = 0.35, p < 0.001) predicted negative emotionality scores. ISI total scores did not predict incentive salience domain scores (p = 0.13). There were group differences in negative emotionality scores (F = 8.99, p < 0.001), such that the clinical insomnia group had higher scores than the no insomnia group (p < 0.001) and the sub-clinical insomnia group (p = 0.04), and the sub-clinical insomnia group had higher scores than the no insomnia group (p = 0.01). There were no group differences on incentive salience domain scores (p = 0.21).

Conclusions: These findings suggest that individuals with AUD and sleep disturbance have an AUD neuroclinical presentation marked by greater subjective and neural negative emotionality responses. These findings were replicated in two independent samples of people with AUD using two separate assessments of sleep problems, suggesting that these findings are generalizable across different sleep problems. This work suggests individuals with AUD who endorse sleep problems may benefit from tailored treatments incorporating affective processing.

Keywords: alcohol use disorder, sleep disturbance, negative emotionality, Addictions Neuroclinical Assessment

Disclosure: Nothing to disclose.

P309. Dopamine D2/3 receptor availability during the transition to addiction: repeat PET studies in high-risk volunteers

Maja Nikolic, Maisha Iqbal, Nicole Palacio, Charlotte Caswell, Sylvia Cox, Natalia Jaworska, Jean Seguin, Gassan Massarweh, Frank Vitaro, Mara Brendgen, Sophie Parent, Michel Boivin, Sylvana Côté, Richard E. Tremblay, Natalie Castellanos-Ryan, Marco Leyton

McGill University, Montreal, Canada

Background: People with substance use disorders (SUDs) have low dopamine D2/3 receptor availability. It remains controversial, however, whether this is a pre-existing vulnerability trait or an effect of heavy drug use.

Methods: [¹⁸F]Fallypride positron emission tomography scans were conducted in longitudinally followed youth at ages 18 (Wave-1, n = 58, 36F/22M) and 25 (Wave-2, n = 47, 26F/21M). Binding potential (BPND) values were derived for 12 striatal and extrastriatal regions of interest (ROIs). A linear mixed-effects model examined effects of Group (Emerging SUD vs Control), Time, and ROI, including all two- and three-way interactions. Post hoc comparisons and correlations were controlled for using Bonferroni and Benjamini–Hochberg corrections, respectively.

Results: At Wave-1, BPND values were higher in the Emerging SUD than Control group in the ventral (VS; p = 0.007), associative (AS; p = 0.013), and sensorimotor striatum (SMS; p < 0.001). At Wave-2, BPND values were lower in the Emerging SUD than Control group in the AS (p = 0.004) and SMS (p < 0.001). Both groups exhibited decreases in BPND values during the intervening seven years, but the magnitude of these decreases was greater in those who had developed an SUD, particularly in the AS (p < .001), SMS (p < 0.001), and VS (p = 0.003). Individual differences in Wave-2 AS and SMS BPND values correlated negatively with lifetime alcohol use occasions (AS: r = –0.363, p = 0.017; SMS: r = –0.332, p = 0.030). None of these effects were influenced by sex or driven by stressful life events (all p > 0.221).

Conclusions: This study provides the first longitudinal evidence in humans that striatal D2/3 receptor availability is elevated in high-risk participants prior to the onset of an SUD and then declines in those who develop an SUD, plausibly reflecting compensatory responses to repeated drug-induced dopamine surges.

Keywords: Dopamine (D2, D3) receptors, substance use disorders, longitudinal multimodal imaging

Disclosure: Nothing to disclose.

P310. Brexanolone to target alcohol use disorder and post-traumatic stress disorder co-morbidity

MacKenzie Peltier, Margaret Altemus, Ismene Petrakis, Sherry McKee

Yale Medical School, New Haven, Connecticut, United States

Background: Alcohol Use Disorder (AUD) and Posttraumatic Stress Disorder (PTSD) frequently co-occur. Notably, there is a bi-directional relationship between alcohol use and PTSD symptomatology; PTSD increases alcohol use and craving, while heavy drinking exacerbates trauma symptoms. Despite this relationship, there are limited pharmacological therapies to target AUD and PTSD concurrently. Brexanolone (an exogenous molecule chemically identical to allopregnanolone) exhibits anxiolytic and antidepressant actions which underlie the neurobiological mechanisms observed in AUD and PTSD and thus may be an advantageous treatment target. The objective of this proof-of-concept, Phase 1, open-label study was to establish the feasibility/safety of administering brexanolone to individuals with concurrent AUD/PTSD and determine the drug’s potential to reduce alcohol consumption and PTSD symptoms.

Methods: Following an intake assessment, five participants (n = 4 men, n = 1 women; mean age = 35.0) with AUD/PTSD were enrolled and admitted to an inpatient research unit, where they received brexanolone (target dose: 90mcg/kg/hour) over a 20-hour infusion. Two days later the participants (n = 3; two participants were disqualified due to positive drug screens) returned to the unit to complete a laboratory session, which included an ad-libitum drinking paradigm. The ability to resist drinking, the amount consumed, and the safety of alcohol consumption following the infusion were evaluated. Participants then completed a 30-day follow-up assessment. Alcohol use, including the average number of drinks per day and number of heavy drinking days, alcohol craving, PTSD symptomatology via the Clinician Administered PTSD Scale (CAPS-5), and overall stress ratings via the Positive and Negative Affect Scale (PANAS) were tracked from intake assessment through a 30-day follow-up visit.

Results: Regarding feasibility and safety, no serious adverse events were reported during the study. Common adverse events, such as fatigue and headache, were rated as minimal to mild in intensity. No participant discontinued or required a dose adjustment during the drug infusion. Additionally, no adverse events were reported in conjunction with the alcohol self-administration session following the drug infusion. We observed a decrease in the average number of drinks consumed per day (over the past 30 days) from study intake to the 30-day follow-up visit (t(4) = 2.37, p = 0.08, Cohen’s d = 1.06) and the number of heavy drinking days consumed during this time period (t(4) = 2.15, p = 0.10, Cohen’s d = 0.96). Similarly, there was an approximately 50% reduction in alcohol craving from drug administration (Mean [M] craving rating = 14.60) to discharge from the drug administration session (M craving rating = 7.06; t(4) = 1.34, p = 0.25, Cohen’s d = 0.16). Of those completing the ad-libitum drinking paradigm, the average lapse until first sip was 18 minutes (range 3 minutes to 45 minutes). Regarding PTSD symptomatology, there was a significant decrease in the number of total PTSD symptoms from the intake to the 30-day follow-up (t(4) = 5.29, p = 0.006, Cohen’s d = 2.36) and a similar decrease in symptom severity for this same time period (t(4) = 4.84, p = 0.008, Cohen’s d = 2.17). Overall stress ratings decreased approximately 14 points from admission to discharge on the drug administration day (t(4) = 1.50, p = 0.21).

Conclusions: These data provide preliminary evidence that brexanolone can safely be administered to individuals with AUD and PTSD, as no serious adverse events were reported. Evidence suggests that brexanolone may decrease alcohol use and heavy drinking among individuals with AUD/PTSD; however, while these results did not reach statistical significance, they did exhibit large effect sizes. Significant decreases in PTSD symptomatology and symptom severity were observed. Accordingly, these findings warrant additional research among a larger sample.

Keywords: brexanolone, allopregnanolone, Alcohol Use Disorder - Treatment, PTSD

Disclosure: Nothing to disclose.

P311. Alterations in resting state functional connectivity following acute inflammatory exposure in individuals with alcohol use disorder

Kaitlin McManus, Erica Grodin, Lara Ray

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

Background: Neurobiological alterations are key components of the pathophysiology underscoring alcohol use disorder (AUD). AUD is associated with changes in resting state functional connectivity (rsFC) in the salience network (SN), default mode network (DMN), and central executive network (CEN; i.e., the triple network model). These networks are responsible for integrating information (SN), regulating emotions (DMN), and making decisions (CEN). The mechanistic basis for these changes in rsFC within individuals with AUD is unclear, however. One potential mechanism is inflammation. Dysregulated communication between peripheral inflammatory markers and the brain may drive the pathophysiology of psychiatric conditions, including AUD. Peripheral inflammatory cytokines (interleukin (IL)-6, IL-8, tumor necrosis factor alpha (TNF-α)) are associated with alcohol consumption. Studies have utilized inflammatory challenges to examine the impact of inflammation on rsFC, but not in the context of individuals with AUD. Therefore, the present study utilized an inflammatory endotoxin challenge to elucidate the role of inflammation on rsFC in individuals with AUD.

Methods: Twenty (12 male/8 female) non-treatment seeking individuals with AUD symptoms completed a randomized, double-blind, placebo-controlled trial of low dose endotoxin. Participants were randomly assigned to receive either low-dose endotoxin (0.8 ng/kg of body weight; n = 10) or placebo (same volume of 0.9% saline; n = 10) via intravenous bolus infusion. Blood samples to measure peripheral markers of inflammation (i.e., TNF-α, IL-6, IL-8) were collected at baseline and four hourly time points post-infusion (T0-T4). MRI acquisition occurred three hours post-infusion (T3). A six-minute, eyes-open resting state scan was collected. Resting state analyses were performed in the CONN Toolbox. Analyses of covariance were conducted to identify a main effect of treatment (endotoxin vs. placebo) on seed-based rsFC in the SN, DMN, and CEN. Placebo > endotoxin contrasts were conducted to compare seed-based rsFC between treatment groups while controlling for body mass index and baseline depressive symptoms. Secondary analyses were conducted in the endotoxin group in networks where primary treatment effects were identified to examine how endotoxin-induced changes in levels of cytokines from baseline to expected peak inflammatory response (T2) predicted seed-based rsFC at the time of the fMRI scan (T3). Results were considered significant at a cluster forming threshold of voxel-level uncorrected p < 0.001 and a cluster-level false discovery rate (FDR) correction of p-FDR < 0.05.

Results: There was a main effect of treatment on rsFC between seed regions within the SN with clusters outside of the SN. A main effect of treatment was identified on rsFC between the left anterior insula and clusters with peak locations in the right angular gyrus (F(1,16) = 45.75, pFDR < 0.001), right orbitofrontal cortex (F(1,16) = 55.52, pFDR < 0.001), and right putamen (F(1,16) = 61.39, pFDR < 0.001). A main effect of treatment was additionally identified on rsFC between the right anterior insula and clusters with peak locations in the left angular gyrus (F(1,16) = 43.71, pFDR < 0.001) and right orbitofrontal cortex (F(1,16) = 61.98, pFDR < 0.001). Lastly, a main effect of treatment was identified on rsFC between the left supramarginal gyrus (SMG) and a cluster with a peak location in the right angular gyrus (F(1,16) = 45.33, pFDR < 0.001). Post-hoc analyses to determine between-group differences revealed that individuals who received placebo had significantly greater rsFC between SN seeds and identified clusters compared to those who received endotoxin (ps < 0.001). There was no significant effect of treatment on rsFC in the DMN or CEN. Within individuals who received endotoxin, greater levels of endotoxin-induced cytokines were significantly associated with greater rsFC between SN seeds (i.e., anterior cingulate cortex (ACC), insula, rostral prefrontal cortex, SMG) and clusters with peak locations in the SN (i.e., ACC, insula, SMG) as well as clusters with peak locations in the dorsolateral PFC, caudate, ventral anterior cingulate cortex, Broca’s area, and visual association (ps < 0.001). Within individuals who received endotoxin, greater levels of endotoxin-induced cytokines were significantly associated with less synchronicity between SN seeds and clusters with peak locations in the orbitofrontal cortex, angular gyrus, medial temporal gyrus, and ventral posterior cingulate cortex (ps < 0.001).

Conclusions: The present study sought to advance the translational understanding of the effect of acute inflammation on brain networks involved in AUD. Within individuals with AUD symptoms, findings suggest that endotoxin-induced inflammation led to less rsFC between SN seeds and regions outside of the SN that are involved with higher order reward processing and cognition. Moreover, higher levels of endotoxin-induced cytokines were predictive of 1) greater rsFC between SN seed regions and regions associated with pain perception and sensorimotor function, but 2) less rsFC between SN seed regions and regions associated with reward and cognitive set-shifting. As such, within individuals with AUD, endotoxin-induced increases in inflammatory cytokine levels appear to disrupt rsFC in networks associated with reward processing and cognition. These inflammation-driven disruptions may contribute to the pathophysiology underscoring AUD.

Keywords: Alcohol Use Disorder, Inflammation and cytokines, Resting State Functional Connectivity, endotoxin

Disclosure: Nothing to disclose.

P312. Large language reasoning model predicts later relapse using transcripts from a qualitative interview on the function of substance use in individuals in treatment for alcohol or opioid use disorder

Gregory Niklason, Colleen McGowan, Kailey Lipske, Anna Zilverstand

University of Minnesota, Minneapolis, Minnesota, United States

Background: Relapse rates for substance use disorders are estimated to be up to 85%. Due to the heterogeneity of reasons underlying substance use in different individuals, there is diversity in process and outcome, making the prediction of relapse a challenge. Here, we hypothesize that self-reflection on the function of substance use in one’s life within a qualitative clinical interview may be informative for predicting relapse. We propose to employ artificial intelligence, specifically a reasoning large language model (LLM) to process this data. The emergence of modern LLMs such as ChatGPT, offers new opportunities to efficiently process such qualitative interview data. Reasoning LLMs, such as DeepSeek-R1, are a specific class of LLM that is trained to exhibit the ability to “think” by producing chain of thought (CoT) reasoning traces, which is a step-by-step sequence of “thoughts” (CoTs) that a model followed to arrive at a prediction. Previous research has aimed to predict relapse in substance users using simpler models and data from social media, but to the best of our knowledge the utility of reasoning LLMs to predict relapse using data from qualitative clinical interviews has yet to be studied. Here, we evaluate applying a locally run open-source reasoning LLM to qualitative interview data collected in individuals with alcohol use disorder (AUD) and opioid use disorder (OUD) to predict relapse.

Methods: Thirty-six participants that were in the first three months of treatment for either AUD (n = 22) or OUD (n = 14) were recruited for an ongoing multi-visit study. All participants completed an approximately 1-hour qualitative interview aimed at self-reflection on the function of substance use in one’s life (‘use interview’). At follow-up visits 3 and 9 months later, participants were then asked whether they had relapsed to using substances (yes/no). The audio recordings of the ‘use interviews’ were first transcribed to text with an encoder-decoder transformer, Whisper-large-v3. Then, we applied a Meta Llama 3.1 8B model (‘DeepHermes-3-Llama-3-8B-Preview’) that was distilled from a decoder-only LLM (‘DeepSeek-R1’) across a variety of tasks. This DeepHermes-3-Llama-3-8B-Preview model is a distillation model that was fine-tuned on CoT reasoning traces generated by DeepSeek-R1. This makes it possible to transfer the reasoning behavior of the much larger DeepSeek-R1 to the smaller locally run open-source reasoning LLM, enabling the local model to run on a single NVIDIA RTX 4090 GPU. Using this approach, the local model estimates the probability (0–100%) of relapse for each individual. Additionally, the model returns the CoT reasoning traces used for the prediction, which can be interpreted as an explanation for the estimated probability of relapse. Importantly, this application was unsupervised and performed in a zero-shot setting, meaning that it was not specifically trained for relapse prediction.

Results: We observed an overall relapse rate of 50% in our sample, with 12 out of 22 AUD and 6 out of 14 OUD participants (total: 18 out of 36) having relapsed. The model predicted relapse with a Receiver Operating Characteristic Area Under the Curve (ROC-AUC) score of 0.72. The overall model accuracy was 0.69, with a precision of 0.70, recall of 0.69, and an f1-score 0.69. Based on a data-driven post-hoc analysis (Youdin’s J), the ideal threshold value for predicting relapse in this sample would be a 48% probability. The CoT reasoning traces that were most often selected by the model as being relevant for relapse prediction were largely linked to commonly known risk and protective factors for substance use disorder. Examples for risk factors that the model selected as being relevant included reflections on indicators of addiction severity such as a) age of onset of addiction, and b) number of previous attempts at recovery, as well as compounding factors such as c) periods of homelessness, d) ongoing struggles with stigma and societal rejection, e) having drug cravings, and f) reporting to use substances to cope with strong emotions. Examples for protective factors that the model selected as relevant to predicting relapse included indices of motivation and support such as, g) the expressed desire to maintain sober, h) having a supportive partner or family, i) attending group meetings in the recovery community and j) active engagement in treatment programs. Another set of protective factors that the model used in its prediction included reflections that k) demonstrate insight into one’s behavior and l) recognize the importance of rebuilding one’s identity beyond addiction.

Conclusions: Our results suggest that large language reasoning models can predict substance use relapse from qualitative interviews on the function of substance use in one’s life. The model selected both well known risk and less often studied risk and protective factors for substance use, suggesting its usefulness as a tool to explore such data. Overall, the results support the validity of reasoning LLMs for tasks such as relapse prediction – despite them not being explicitly trained to do so. We therefore conclude that reasoning LLMs hold promise as a clinical tool for predicting relapse risk, both in a treatment setting and at home – particularly, if they would be specifically trained to do so.

Keywords: artificial intelligence, qualitative interview, Alcohol Use Disorder, Opioid Use Disorder, relapse

Disclosure: Nothing to disclose.

P313. Neural and behavioral mechanisms linking the problem cannabis risk network to cannabis use disorder

Sarah Lichenstein, Che Liu, Sandra Field, Francesca Filbey, Godfrey Pearlson, Sarah Yip

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

Background: Cannabis is the most commonly used illicit drug [1], and substantial evidence indicates that its use is associated with clinically-significant harms for a subset of users [2]. In particular, rates of cannabis use disorder (CUD) are on the rise: contrary to earlier epidemiological data that suggested 10% of individuals that use cannabis develop CUD [3], recent data show that 20% of cannabis users—and 30% of frequent users—will develop CUD [4]. Therefore, improved prevention and treatment approaches are urgently needed [5], yet the neurobiological substrates of CUD remain poorly understood. Our prior work applied connectome-based predictive modeling (CPM), a data-driven, whole-brain, machine learning approach [6], to identify a neuromarker of problem-level cannabis use in a non-clinical sample of adolescents/emerging adults that replicated in two independent datasets [7]. Here we seek to replicate and extend this work in a fourth independent sample including chronic cannabis users with and without CUD (MJ +/− CUD) and controls that completed a cannabis cue exposure task during fMRI scanning [8, 9]. To assess potential behavioral and neural mechanisms linking the problem cannabis risk network to CUD, we assessed whether problem cannabis risk network strength (1) distinguishes MJ + CUD from MJ-CUD and controls, (2) relates to cannabis cue-induced craving, as well as (3) examining connectome-wide differences in functional connectivity between MJ+CUD, MJ-CUD and controls, and assessing neuroanatomical overlap with the problem cannabis risk network.

Methods: The current analyses include N = 268 participants drawn from two studies that administered a cannabis cue exposure fMRI task to regular cannabis-using individuals and controls [8, 9]. Strength of the previously identified problem cannabis risk network [7] was extracted from functional connectivity matrices computed from cannabis cue exposure task data. ANOVA was used to assess sex differences and group differences in network strength between MJ + CUD (n = 51), MJ-CUD (n = 60), and controls (n = 138), and Spearman correlation was used to assess the association between network strength and in-scanner cue-induced cannabis craving. The Network-Based Statistic [10] was used to identify connectome-wide differences in functional connectivity between groups and anatomical overlap with the problem cannabis risk network was assessed based on shared edges and edgewise correspondence with canonical neural networks.

Results: Problem cannabis risk network strength differed significantly between men and women (F(1) = 17.9, p < .001) and across groups (F(2) = 3.5, p = .032); network strength was highest among MJ+CUD, followed by MJ-CUD, and controls. Post-hoc Bonferroni pairwise comparisons demonstrated a significant difference in network strength between the MJ+CUD and control groups (p = 0.009). Network strength was also significantly positively associated with cue-induced cannabis craving (n = 264; rho = 0.27, p < 0.001). NBS analysis identified significant differences in functional connectivity between the MJ+CUD and control groups, such that individuals with CUD displayed greater connectivity between the motorsensory network and cerebellar and visual association networks, as well as between the frontoparietal and salience networks, coupled with reduced connectivity between frontoparietal and visual association networks, within motorsensory network connectivity, and between the motorsensory network and medial frontal and subcortical networks during cannabis cue exposure relative to controls. Neuroanatomical overlap with the problem cannabis risk network was observed with regard to increased connectivity between motorsensory and visual association networks and reduced connectivity between the motorsensory, medial frontal, frontoparietal, default mode and cerebellar networks.

Conclusions: The current findings demonstrate that the problem cannabis risk network distinguishes cannabis users with CUD from controls and that network strength exhibits a stepwise increase between controls, MJ-CUD, and MJ + CUD. In addition, we identify increased cannabis cue-induced craving as a potential behavioral mechanism whereby this network increases risk for problem-level cannabis use and CUD. Furthermore, two complimentary connectome-based analysis approaches converge in highlighting common patterns of altered functional connectivity in association with problem-level use and CUD across distinct datasets. Future work is needed to further elucidate the clinical relevance of these neural mechanisms in order to pave the way for the development of more targeted and effective prevention and treatment approaches.

Keywords: fMRI Replication, task-based functional connectivity, cannabis use disorder

Disclosure: Nothing to disclose.

P314. Preliminary findings of an 8-week proof of concept trial evaluating psilocybin-assisted-psychotherapy in adults with cannabis use disorder

Maryam Rahat, James MacKillop, Emily Vandehei, Beth Patterson, Paige Hopkinson, Meghan Romano, Alyssa Swiderski, Vidhi Patel, Hashim Al-Bya, Michael Van Ameringen

McMaster University, Hamilton, Canada

Background: Cannabis use has become increasingly prevalent in North America and worldwide, particularly among males and young adults. A subset of users develop Cannabis Use Disorder (CUD), (12–31% of active users) characterized by significant impairment in social, occupational, and psychological functioning. [1] Unfortunately, current pharmacological and psychosocial treatments have demonstrated limited efficacy and high relapse rates. Recent research examining the use of psychedelics combined with psychotherapy, particularly psilocybin, suggests potential efficacy in treating substance use disorders, including alcohol dependence and smoking cessation [2–4], with substantial benefits reported after a single psychedelic administration. [2] However, there are no published reports examining psychedelics in the treatment of CUD. Given the evidence of psilocybin’s therapeutic potential in the treatment of substance use disorders and the paucity in the CUD literature, we have been conducting a proof-of-concept study to evaluate the safety and tolerability of psilocybin-assisted psychotherapy for CUD, with secondary aims of examining the feasibility and the impact of treatment on CUD symptoms. The results in this abstract are preliminary as enrollment is ongoing. We anticipate that it will be presented in full when the poster is presented.

Methods: Using an open-label, within-subject, prospective design, we propose enrolling 16 participants aged 18–65 who meet DSM-5 criteria for moderate to severe CUD. The 8-week intervention period consists of Motivational Enhancement Therapy (MET) and supportive psychotherapy (including preparatory and integration sessions) combined with two, 8-hour dosing sessions. During each dosing session, 25 mg of oral psilocybin is administered under controlled conditions with physician supervision. The primary outcome of safety will be measured by the incidence of spontaneously reported adverse events from baseline to the 8-week endpoint; tolerability will be assessed by the % recruited participants who discontinue or are lost to follow up before the 8-week endpoint. Feasibility will be determined by the % eligible and % enrolled. Secondary outcome of efficacy for CUD symptoms will be changes in frequency and amount of cannabis used measured by Timeline Follow-Back (TLFB) reports, using metrics of sessions/week and grams/week; response will be determined by % reduction in use; salivary and urinary cannabinoid levels will be used as verifiers. Exploratory outcomes will involve validated subject-rated measures of withdrawal, craving, anxiety, depression, attention deficit hyperactivity disorder (ADHD), OCD, functional impairment, cognitive functioning and well-being and clinician-rated clinical global impression severity and improvement (CGI-S, CGI-I).

Results: In a preliminary analysis of the first 7 participants, 71% were female, 57% Caucasian, 43% university graduates, 43% worked full-time. The mean age was 29.14 ± 6.9 years; At baseline, the mean CGI-S was 6.0 ± 0.0, indicating severe illness. Comorbidity within the sample was high, with 100% having at least one comorbid psychiatric disorder. The most prevalent comorbidities included ADHD (86%) and generalized anxiety disorder (GAD) (71%). With regard to the primary outcomes, the most common adverse events have included feeling tired, difficulty sleeping and loss of appetite; all of which were transient; two of the seven participants (28.6%) declined their second dose, completing all other components of the study protocol (including MET sessions). Both of these participants had achieved cannabis cessation prior to the second dose. One participant (14%) dropped out after the first dose and withdrew from the study. All participants reported a reduction in cannabis use by Week-8; 71% (5/7) reported at least 50% reduction in cannabis use and 29% (2/7) achieved complete cessation. Urinary cannabinoid tests supported reports of decreased cannabis; salivary samples have not yet been analysed. All (100%) participants had an endpoint CGI-I of 1 or 2 (much or very much improved). In terms of feasibility, 77.8 % of those eligible, actually enrolled in the study (7/9). The mean CGI-S at Week 8 was 3.6 ± 1.3 indicating mild-moderate severity.

Conclusions: Preliminary findings suggest that the protocol has sufficient safety, tolerability and feasibility to warrant an adequately powered, randomized controlled trial. The substantial CUD symptom reductions found in these first participants seem promising.

Keywords: cannabis use disorder, Psilocybin therapy, Open Label Trials

Disclosure: Abbvie, Advisory Board, Self, Bausch Health, Advisory Board, Self, Biogen, Advisory Board, Self, Boehringer Ingelheim, Advisory Board, Self, Lundbeck, Advisory Board, Self, Reunion Neuroscience, Advisory Board, Self, Biohaven, Contracted Research, Self, Clairvoyant, Contracted Research, Self, Neumora, Contracted Research, Self, atai Therapeutics, Contracted Research, Self, UptoDate, Honoraria, Self, Canadian Institute for Health Research, Grant, Self, Michael G. DeGroote Centre for Medicinal Cannabis Research, Grant, Self, Abbvie, Speakers Bureau, Self, Lundbeck, Speakers Bureau, Self

P315. Ibudilast attenuates the association between chronic pain intensity and heavy drinking in alcohol use disorder: a secondary analysis of a randomized clinical trial

Steven Nieto, Erica Grodin, Lara Ray

Wayne State University School of Medicine, Los Angeles, California, United States

Background: Chronic pain frequently co-occurs with alcohol use disorder (AUD) and may influence treatment response. In this secondary analysis of a completed randomized clinical trial (NCT03594435), we examined whether baseline chronic pain intensity moderates the effects of ibudilast, a neuroimmune-modulating agent, on drinking outcomes over 12 weeks.

Methods: Participants (N = 102; 60% male; M age = 44.26, SD = 10.81) with moderate-to-severe AUD were randomized to ibudilast (50 mg twice daily) or placebo for 12 weeks. Chronic pain intensity was assessed at baseline using the Graded Chronic Pain Scale. Piecewise linear mixed-effects models tested the moderating effect of chronic pain intensity on percent heavy drinking days (PHDD), percent days abstinent, drinks per day, and drinks per drinking day, adjusting for early (baseline–Week 2) and late (Weeks 4–12) trial phases.

Results: Chronic pain intensity was not significantly correlated with any drinking outcome at baseline (all |r| ≤ 0.08, p > 0.45). Significant three-way interactions (Medication × Pain Intensity × Time) emerged for PHDD in both early and late phases (ps < .05), indicating that higher chronic pain intensity predicted greater PHDD in the placebo arm (b = 0.10, p = 0.04) but not in the ibudilast arm (b = 0.007, p ≥ 0.89). Similar interactions were observed for drinks per day, though simple slopes were nonsignificant. No significant moderation effects were found for percent days abstinent or drinks per drinking day.

Conclusions: Ibudilast attenuated the positive association between baseline pain intensity and heavy drinking, suggesting that chronic pain intensity may identify an AUD subgroup more likely to benefit from neuroimmune-targeted pharmacotherapy. These findings support incorporating pain phenotyping into AUD treatment research and point toward a precision medicine approach to enhance treatment efficacy.

Keywords: Alcohol Use Disorder—Treatment, chronic pain, randomized clinical trial

Disclosure: Nothing to disclose.

P316. A non-drug alternative for opioid use disorder? A double-blind, double-dummy evaluation of an auricular neuromodulator device for opioid withdrawal and craving

Cecilia Bergeria, Alec Zhang, Katja Schmid-Doyle, Dhathri Srungaram, Andrew Huhn, David Wolinsky, Kelly Dunn, Eric Strain

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

Background: Treatment of opioid use disorder (OUD) is characterized by high discontinuation rates, including during treatment initiation. Novel strategies are needed to effectively mitigate withdrawal symptoms that accompany this critical early treatment period and to facilitate transitions from non-medical opioid use (e.g., fentanyl) to approved treatments (e.g., methadone, buprenorphine, extended-release naltrexone). Auricular neuromodulators are designed to electrically stimulate cranial nerve branches (V, VII, IX, X) in the ear with low power and short pulses, with the goal of modulating autonomic and other neural systems implicated in opioid withdrawal. The NSS-2 Bridge Device is FDA cleared (reflecting the device is safe to use) for opioid withdrawal and has demonstrated potential therapeutic benefit for opioid withdrawal in an open label, non-controlled trial. If effective, such a device would provide a non-drug therapeutic that could aid in opioid withdrawal treatment. Critically, its efficacy had not been evaluated in a randomized controlled trial. The goal of this trial (clinicaltrials.gov: NCT04325659) was to examine the efficacy of the Bridge Device for increasing treatment retention and decreasing opioid withdrawal relative to a true placebo condition (Sham Bridge Device) and an active comparator condition (lofexidine (Lucemyra) – the only non-opioid FDA approved medication for opioid withdrawal).

Methods: Medically cleared adults seeking treatment for OUD were enrolled in a double-blind, double-dummy trial in which they were randomly assigned to receive: (1) Active Bridge Device + Placebo, (2) Sham Bridge Device + Placebo, or (3) Sham Bridge Device + Lofexidine (0.72 mg QID, Active Comparator). Participants were admitted into a residential research facility for a 16–20-day protocol. Prior to randomization to a treatment condition, participants were stabilized on a short-acting opioid for 7–11 days (intramuscular morphine 30mg 4X/day or oral hydromorphone up to 20mg 6X/day). After stabilization, on the first day of the treatment phase, blinded research staff applied either an Active or Sham NSS-2 Bridge Device. The Device was affixed behind the participant’s ear with adhesive, and four wire leads were placed to target cranial nerve branches affixed via superficial needles guided by vascularization. Active Devices delivered electrical pulses, and Sham Devices were identical in appearance but delivered no electrical pulses. Participants were evaluated for the 7-day treatment period, followed by a 2-day post treatment observation period. The Device was worn and the full Lofexidine dose was administered for the first 5 treatment days. For the final two days of the treatment phase, the Device was removed and active lofexidine tapered to decrease risk of rebound hypertension. Analyses focused on days 1–5 of the treatment phase. Throughout, participants were assessed for vital signs and opioid withdrawal four times daily using an observer rated scale comprised of 11 items assessing withdrawal symptoms (Clinical Opiate Withdrawal Scale – COWS, range: 0 - 48). Primary outcomes included retention at the end of the treatment period, daily peak COWS scores across treatment days, and COWS score area under the curve (AUC) during the treatment phase. Analyses and statistics include completers - individuals enrolled on day 5 of the treatment phase - unless otherwise noted.

Results: Thirty-six individuals were enrolled and randomized (n = 12 per condition). Participants were majority male and Black, with opioid use on average 29.0 (SEM = 0.6) days in the past 30 and stimulant use on average 7.5 (SEM = 1.9) days in the past 30. Retention did not differ significantly across treatment groups (58–66%). Analyses of peak COWS scores across days, using both intent-to-treat (ITT) and completers only, found no significant treatment effects (ITT p = 0.59, completer p = 0.32). Descriptively, the Active Device + Placebo group displayed a comparable time course of withdrawal to the Sham Device + Placebo group across treatment days: COWS total scores peaked on day 2 (BD + P: 10.3 ± 0.9, SBD + P: 9.1 ± 1.1) before gradually declining on days 3–5. Conversely, lofexidine demonstrated no change in COWS total scores across days 1–2 (Day 2: 7.7 ± 1.0) followed by decreases on days 3–5. Total COWS scores AUC (Mean ± SEM) significantly differed as a function of treatment group; individuals receiving Active Device + Placebo (1.6 ± 3.4) had significantly greater COWS AUC relative to Sham Device + Lofexidine (−8.6 ± 3.9, p < 0.05). Sham Device + Lofexidine was associated with significantly lower diastolic AUC during the treatment period relative to other conditions. Among individuals assigned to the Sham Device + Lofexidine condition, 5/12 participants declined the medication at least once, 3/12 declined in the Active Device + Placebo condition and 1/12 declined in the Sham Device + Placebo condition.

Conclusions: This randomized controlled trial found no evidence that the NSS-2 Bridge Device reduces withdrawal relative to sham or active comparator. Lofexidine attenuated withdrawal but may be associated with poor acceptability evidenced by the frequency of medication refusal by participants. It may be helpful to clarify whether subgroups could benefit from auricular neuromodulation, but current results underscore the need for sham/placebo-controlled evaluations before clinical adoption of novel interventions.

Keywords: opioid use disorder, Non-invasive Neuromodulation, opioid withdrawal, randomized controlled trial

Disclosure: MindMed Inc, Consultant, Self

P317. Clinical trial of mom power psychotherapy on mothers with opioid use disorder: brain circuits and behavior

James Swain, Shaun Ho, Kristin Bernard, David Garry, Cassandra Heiselman, Maria Hensley, Kimberly Herrera, Yanni Liu, Amanda Levinson, Nicole Miller, Brady Nelson, Rosenblum Katherine, Richard Rosenthal, Joseph Schwartz, Jason Amadio, Elle Eggers, Kaitlyn Reimer, Diana Saum, Maria Muzik

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

Background: Opioid use disorder (OUD) continues to be a fast-growing and devastating epidemic in the US. Pregnant woman have been seriously affected by this epidemic, as exemplified by OUD incidence quadrupling in this group from 1999 to 2014 and conferring substantial risk for serious health problems such as postpartum depression, polysubstance use, relapse, and developmental risks to children. Despite receiving “gold standard” buprenorphine medication treatment to reduce withdrawal, OUD mothers commonly continue to experience cravings, depression, impaired interpersonal interactions and maladaptive parenting behaviors, which are regulated by evolutionarily conserved maternal brain neurocircuits (MBN). Related parenting problems lead to child maltreatment and costly utilization of foster care. We know that parenting interventions such as Mom Power improve parent mood and sensitive caregiving behaviors in high-risk non-OUD women via MBN, but we do not know if parenting intervention psychotherapy may be effective for mothers with OUD or by which brain mechanisms. Thus, in order to improve maternal drug use problems and mental health outcomes, we ran a pilot NIDA R61 project to address this compelling knowledge gap. For mothers with OUD, we studied the intervention Mom Power (MP) for feasibility, utility to improve symptoms and related brain mechanisms in evolutionarily conserved maternal brain neurocircuits (MBN), focusing on functional magnetic resonance imaging (fMRI) and evoked response potential (ERP) tasks for motivated maternal behaviors and arterial spin labeling (ASL) of blood brain barrier permeability (pBBB), focusing on the amygdala (AMY) and hypothalamus (HYP).

Methods: Mothers with OUD (n = 15) completed surveys of mood, post-traumatic and parenting stress, and opioid cravings with multiple neuroimaging modalities, repeatedly pre- and post-MP: a) event-related potential (ERP) tasks to measure N170 and late positive potential (LPP) responses to standardized Crying, Laughing and Neutral faces of unknown children and own child’s faces; b) a functional magnetic resonance imaging (fMRI) task that differentiated maternal empathic “join” vs. “observe” facial expressions of own vs. unknown child; c) ASL imaging to measure pBBB.

Results: From pre- to post-MP, mothers with OUD showed: reduced craving (pre: M = 1.27, s.e. = 0.60, post: M = 0.27, s.e. = 0.21, t(14) = 2.14, p = 0.025), depression (pre: M = 9.87, s.e. = 1.41, post: M = 7.73, s.e. = 1.52, t(14) = 2.19, p = 0.023), post-traumatic stress symptoms (pre: M = 28.40, s.e. = 3.928, post: M = 23.07, s.e. = 4.686, t(14) = 1.962, p = 0.035), and parental stress (pre: M = 33.86, s.e. = 2.33, post: M = 30.71, s.e. = 2.65, t(13) = 2.42, p = 0.016); enhanced empathic attunement-related responses in HYP (z = 3.67, p < 0.001) and bilateral AMY (right, z = 2.56, p = 0.005; left, z = 2.46, p = 0.007) in fMRI; greater N170 responses to crying faces (P8-P7 140–180ms, pre: M = −0.92, s.e. = 1.26, post: M = 0.59, s.e. = 0.85, t(12) = −1.92, p = 0.040). Furthermore, reduction in craving correlated with increases in LPP to Laughing vs. Neutral faces at Pz (r = −0.61, p = 0.020) and Oz (r = −0.54, p = 0.045) and LPP to own child vs. unfamiliar child at Oz (r = −0.68, p = 0.0016) in ERP, empathic responses in bilateral AMY’s (right: z = 2.32, p = 0.010; left: z = 2.93, p < 0.001) in fMRI, and pBBB in HYP (z = 3.63, p = 0.002) and left AMY, (z = 2.46, p = 0.007) in ASL.

Conclusions: For mothers with OUD in this pilot study, MP psychotherapy improved mood, anxiety, stress and cravings while concurrently enhancing brain responses to child stimuli in the MBN. These results may justify a randomized control trial to confirm the efficacy of MP to improve mood, anxiety, stress and cravings via specific brain mechanisms in mothers with OUD - suggesting therapeutic mechanisms and approaches for opioid and other substance use disorders.

Keywords: Opioid abuse, maternal brain, Human Clinical trial, EEG-fMRI, Parent - child dyads

Disclosure: Nothing to disclose.

P318. Intranasal insulin as a neurotherapeutic opportunity: safety evaluation in neurodegenerative and addictive disorders

Stephanie Foster, Bhavani Kashyap, Michael Rosenbloom, Leah Hanson, William Frey II, Robert Swift, Carolina Haass-Koffler

Brown University General Psychiatry Residency Program, Providence, Rhode Island, United States

Background: Intranasal (IN) insulin is increasingly investigated as a treatment for behaviors linked to altered brain glucose metabolism, including cognitive impairment and impulsivity. IN insulin delivery enables direct central administration while minimizing systemic exposure. Although generally well-tolerated among healthy individuals, hypoglycemia remains a theoretical safety concern, particularly among those with metabolic vulnerability. We evaluated the safety of IN insulin in conditions associated with central metabolic impairment, specifically neurodegenerative disorders (ND) and alcohol use disorder (AUD). We hypothesized that compared to control, IN insulin would not significantly decrease blood glucose levels among those with ND, irrespective of insulin resistance status, and that hypoglycemia would not occur in pooled ND and AUD samples.

Methods: ND data includes placebo-controlled clinical trial data of IN insulin in Alzheimer’s disease (N = 35). Blood glucose was measured at baseline, 15 min after, and 30 min after administration of IN insulin (40 IU) or IN saline. Data were stratified by treatment (saline or insulin) and insulin resistance (non-diabetic vs pre-diabetic/diabetic as determined by Hgb A1c). Among non-diabetic participants (N = 23), n = 9 received saline and n = 14 insulin; among pre-diabetic/diabetic participants, n = 7 received saline and n = 5 insulin. To examine AUD, data from an ongoing Phase I/IIa, cross over, placebo-controlled trial of IN insulin in AUD was examined (N = 5).

Results: In ND participants without insulin resistance, blood glucose did not differ significantly by treatment group (p > 0.05). Similarly, no significant glucose difference was observed among pre-diabetic/diabetic participants (p > 0.05). Furthermore, we did not observe any significant change in blood glucose in the AUD population (p > 0.05). Across ND and AUD cohorts, no instances of hypoglycemia were observed following IN insulin administration.

Conclusions: IN insulin does not significantly alter blood glucose levels in individuals with ND, regardless of insulin resistance, and is not associated with hypoglycemia in either ND or AUD. These findings support the safety of IN insulin and strengthen its promise as neurotherapeutic strategy for disorders linked to disrupted brain glucose metabolism.

Keywords: alcohol use disorder, brain glucose metabolism, GLP-1 receptor agonist

Disclosure: Nothing to disclose.

P319. Long-acting naltrexone repairs network connectivity in subjects with opioid use disorder and co-morbid cannabis use disorder

Luke Brier, Daniel Langleben, Corinde E. Wiers, Zhenhao Shi

University of Pennsylvania, Philadelphia, Pennsylvania, United States

Background: Co-morbid substance use disorders (SUDs) are common but difficult to study due to complex interacting and overlapping mechanisms by which they affect brain networks. Many datasets collected to investigate a singular SUD inadvertently contain subjects with extraneous co-morbid SUDs, though large sample sizes often allow for control of these covariates. Although it leads to smaller sample sizes, these datasets contain untapped information regarding specific co-morbid SUDs and response to treatments that are approved for one disorder that could have biological implications for another. For patients with opioid use disorder (OUD), approximately 28% have co-morbid cannabis use disorder (CanUD) in the United States, with this number expected to grow with the spreading illicit use of fentanyl and legalization of cannabis. Prior work has hypothesized that these SUDs co-present because opioid and cannabinoid receptors act as heterodimers or allosteric modulators of each other as they are both G-protein coupled receptors that co-localize on pre-synaptic terminals in limbic, mesencephalon, brain stem, and spinal cord regions. While there are multiple treatments established for OUD, none currently exist for CanUD. Naltrexone (NTX), an opioid receptor antagonist, is an FDA-approved treatment for opioid and alcohol use disorder (AUD), treating the latter by decreasing the euphoric effect associated with alcohol use through modulating b-endorphin activity in the ventral tegmental area and nuclear accumbens. While the underlying molecular mechanism is unknown, prior work has shown that like in AUD, NTX can lead to a sudden decrease in cannabis consumption due to subject-reported decreased euphoria with use within two weeks. This has not been a ubiquitous response to NTX as elsewhere NTX has been shown to be woefully inadequate at treating cocaine use disorder (CocUD) by any mechanism.

Methods: Here, we present a retrospective preliminary analysis of resting-state functional magnetic resonance imaging (fMRI) data in subjects with OUD only (N = 25) and either co-morbid CanUD (N = 10), AUD (N = 6), or CocUD (N = 7). Subjects underwent MRI prior to receiving a therapeutic dose of long-acting intramuscular NTX and two weeks after. The Fisher-Z transformed Pearson correlation coefficient between fMRI signals in regions of interest served as a marker of brain functional connectivity which was compared amongst the various co-morbid SUD subjects and to those who only had OUD. A threshold-free network-based statistics (TFNBS) approach was used to identify specific effects each co-morbid SUD had on brain networks and how these evolved after NTX while correcting for comparisons across a total of 13 brain networks.

Results: At baseline, all subjects with co-morbid CanUD, AUD, or CocUD exhibited significant different patterns of network connectivity compared to subjects who only had OUD. These differences were more pronounced in younger brains (linear regression with age, R2 = 0.64, p = 3.4 × 10-3) and most prominent within the default mode network (DMN, pTFNBS < 0.05). Following the NTX injection, connectivity normalized within nearly every network in subjects with co-morbid CanUD (two-sample t-tests, all p < 0.004). In AUD, changes with NTX were significant within parietal (p = 0.0035), subcortical (p = 8.9 × 10-4), and cerebellar (p = 0.0016) networks. There were no significant network changes in CocUD following NTX.

Conclusions: We present recovered network connectivity in co-morbid OUD and AUD following NTX administration and seemingly no effect in co-morbid OUD and CocUD, which is consistent with prior work and FDA approvals. There is currently no approved treatment for CanUD. This data, combined with prior work showing that NTX can decrease weekly use of cannabis by decreasing euphoric effect is promising for NTX in the treatment of CanUD, potentially through a similar mechanism as it treats OUD or AUD.

Keywords: cannabis use disorder, functional neuroimaging, polysubstance use

Disclosure: Nothing to disclose.

P320. Evaluating value-based decision-making mechanisms in opioid use disorder recovery

Justin Strickland, Amber Copeland, Matt Field, Cecilia Bergeria

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

Background: Reinforcer-specific value-based decision-making (VBDM) is an emerging approach in addiction science to quantify neurobehavioral mechanisms underlying the etiology of and recovery from substance use disorder. VBDM approaches use behavioral tasks and drift diffusion modelling (DDM) to measure “drift rate” and “response boundary” parameters, which in the context of substance use disorder represent the expected value and degree of response caution, respectively, for a given option. Theoretical models predict that substance use disorder may be related to increases in drift rate and decreases in response boundary for drug rewards. Effective treatment is posited to reverse or offset these changes with the recovery process marked by greater allocation of behavior to non-drug rewards. We evaluated the impact of opioid use disorder recovery and cue exposure on reinforcer-specific VBDM parameters in a human laboratory design.

Methods: Participants (n = 47 screening data; n = 37 experimental data) were enrolled in a human laboratory study evaluating individual differences in opioid cue reactivity among people in recovery from opioid use disorder. One screening and three experimental sessions took place in which participants completed a two-alternative forced choice task. Screening sessions included tasks for opioid and non-drug rewards while cue reactivity sessions included only non-drug rewards (to avoid contamination of experimental manipulations by behavioral tasks). Experimental conditions involved presentation of neutral cues, visual opioid cues, or tactile-visual opioids cues (e.g., handling “works” and preparing sham drug). Participants also completed the Addiction Severity Index (ASI) including information on treatment status and treatment history. Data were fitted using DDM to evaluate drift rates and response boundaries.

Results: A significant main effect of length of recovery was observed for opioid VBDM drift rates such that expected opioid value (“drift rates”) systematically decreased with greater length of recovery (p = 0.027). During experimental sessions, a significant effect of experimental cue condition was observed (p = 0.014) for non-drug reward drift rates. This effect reflected large effect size and significant increases in expected non-drug value following visual opioid cue (d = 0.69, p < 0.05) and tactile-visual opioid cue exposure (d = 0.77, p < 0.05). The effect of experimental cue condition on response boundary was not statistically significant. Significant effects of or interactions with participant sex were not observed.

Conclusions: These data demonstrate systematic changes in VBDM parameters based on recovery status and opioid cue exposure. Systematic decreases in opioid drift rates by recovery length provides an objective measure of the decrease in expected drug value that occurs with longer recovery duration. Increased non-drug expected value following opioid cue exposure is suggestive of a potential protective recovery mechanism with reallocation of behavior to non-drug alternatives in periods of stress (i.e., opioid craving). Together, these data demonstrate the utility of VBDM approaches for evaluating recovery processes and provide novel objective markers for treatment evaluation.

Keywords: opioid use disorder, Recovery, Value-based Decision-Making

Disclosure: Nothing to disclose.

P321. When a botanical breaks bad: initial characterization of 7-hydroxymitragynine (7-OH) product use among U.S. consumers

Kirsten Smith, Austin Zamarripa, Naftali Zeilinger, Rachel Rattenni, Edward Boyer

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

Background: Among kratom’s constituents, 7-hydroxymitragynine (7-OH) has received increasing scientific attention as both mitragynine’s active metabolite and as a trace kratom alkaloid. Specifically, in the body, 7-OH is produced by hepatic and intestinal cytochrome-mediated metabolism of mitragynine, whereas in the harvested kratom leaf 7-OH forms through oxidization. Kratom alkaloids act across serotonin, dopamine, opioid, and adrenergic systems. As mitragynine’s primary active metabolite, 7-OH is a highly selective mu opioid receptor agonist. However, the oral bioavailability of 7-OH is low, making it unclear what the effects of 7-OH are in humans who consume 7-OH products via an oral route of administration. Mitragynine and 7-OH (and other kratom alkaloids) can be administered in animals at well-tolerated doses, and at no observable effect levels with toxicity dose-dependent. Kratom alkaloids and metabolites may have lower risk profiles because of receptor activity characteristics that differ from classical opioid agonists.

The behavioral pharmacology and abuse liability of both mitragynine and 7-OH studied in animal models is limited and it is unclear if findings will translate to humans who consume kratom, mitragynine, 7-OH, and mitragynine pseudoindoxyl orally in real-world conditions at self-selected doses and for a variety of reasons. A proliferation of 7-OH products since 2023 has been met with a recent recommendation by the U.S. Food and Drug Administration to the Drug Enforcement Administration for emergency scheduling of 7-OH. Scheduling recommendations appear to be based on preclinical literature as there are few epidemiological or clinical data.

Methods: Between March-May 2025 we conducted an online survey among U.S. adults with 7-OH experience. For this descriptive exploratory report, we characterized their 7-OH use patterns and motivations, and their product conceptualizations in order to provide formative information on consumers’ perceived risk/benefit profile of these emerging kratom-derived products.

Results: The sample included 278 respondents with a mean age of 39.9; they were predominantly white (89.9%) and male (75.2%). Most had histories of opioid use (82.7%) including 30.9% reporting lifetime fentanyl use. Lifetime use of kratom leaf and extract was endemic; 59.7% had also used products containing mitragynine pseudoindoxyl, a secondary metabolite of mitragynine. Past-month 7-OH use was 20.4 days on average compared to kratom whole-leaf (12.0 days) and kratom extracts (7.7 days). Popular 7-OH product formulations consumed included tablets (77.7%), sublingual films (24.1%), and liquid shots (40.6%). Reported perceived 7-OH potency was > for kratom whole leaf or extract but lower than heroin or fentanyl. The perceived acute intoxication from 7-OH was low and did not differ between consumers who were more or less experienced with this compound. On average, respondents dosed 7-OH products 3.2 times per day/19.0 times per week; effects peaked approximately 26 minutes post-dosing.

Similar to prior reports on kratom, consumers of 7-OH widely endorsed DSM-5 substance use disorder symptoms of tolerance, using more than intended, and craving but not symptoms related to hazardous use or impairments in psychosocial functioning. Between 17.1–28.1% used 7-OH as a short-term or long-term replacement for kratom, with others (4.3–13.3%) used as an opioid or alcohol replacement. A nearly equal proportion of respondents (70.0%) perceived 7-OH as “habit-forming” but also “beneficial and helpful.” The most widely cited reason for use was pain management (78.0%). A majority believed 7-OH should remain legal and accessible, but with some regulation.

Conclusions: This survey provides the first consumer-level data on 7-OH use in the U.S., highlighting distinct use patterns, formulations, and perceived effects. Our respondents reported relatively quick onset of 7-OH effects, regular use, and use contemporaneous to mitragynine pseudoindoxyl and kratom. Therapeutic motivations for using 7-OH as well as concerns about tolerance were evident. Although participants perceived 7-OH as more potent than kratom but less potent than heroin or fentanyl, most described its acute intoxication as low and endorsed minimal risks with at least equal perceived benefits. These formative human data suggest that 7-OH is being integrated into existing substance use practices among some U.S. adults. Collectively, our findings and the preclinical literature recommend further investigation into 7-OH use in order to guide evidence-based regulatory and clinical decision-making.

Keywords: 7-hydroxymitragynine, kratom, mitragynine, mitragynine pseudoindoxyl, Mu opioid receptor

Disclosure: Nothing to disclose.

P322. Reducing negative affect improves context-dependent decision-making in early abstinence from methamphetamine use disorder

Maëlle Gueguen, Sarah Qundes, Aardron Robinson, Martin Paulus

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

Background: Drugs are supraphysiological reinforcers thought to co-opt value-based decisions by overshadowing other rewards. This is thought to induce biased context-based reward value representations, hindering appropriate non-drug-related choices. In previous work (Gueguen et al., 2024), we showed that context-sensitive valuation is reduced in opioid addiction, especially during early abstinence, and is associated with negative affective states. Here, we extend this work to Methamphetamine Use Disorder (MUD), a growing public health crisis with no approved medication, high relapse rates, and an urgent need for effective behavioral interventions. Using a behavioral intervention to reduce negative affective states, we assess the causal influence of affective and clinical states on contextual valuation in early abstinent individuals with MUD.

Methods: 62 participants in active treatment for MUD performed a validated contextual reinforcement learning task (cRL, Molinaro and Collins, 2022) before and after undergoing a behavioral interventional aiming at improving their affective state through recalling neutral and positive non-drug-related autobiographical memories. PRE- and POST-intervention questionnaires assessed affective and clinical state (positive and negative affect with PANAS-X, PROMIS-banks for depression and anxiety, withdrawal for amphetamine and stimulants, VAS craving). In the cRL task, participants learn to associate cues with rewards from two trinary contexts, each with a low (L), middle (M), and high (H) value cue. The task plays on context during learning to [1] increase the subjective value of M1 and [2] decrease the subjective value of M2. Context sensitivity leads to a choice bias (M1 preferred to M2) during testing, and an explicit valuation bias (M1 valued higher than M2).

Results: Participants learned to choose the best symbols in all trial types, both PRE- and POST-intervention. MUD participants made value-driven choices and showed contextual sensitivity by preferring M1 (choice rate: 0.61 ± 0.38) over M2 (0.38 ± 0.36; t50 = 2.13, p > 0.1), but to a lesser extent than healthy participants (published study, n = 50, 29.94 ± 0.75 y.o.; choice rate: M1 = 0.57 ± 0.02 over M2 = 0.36 ± 0.02; t49 = 6.53, p < 0.001). MUD participants overestimated the SV of M1 over M2, (SV estimates: M1 = 52.84 ± 2.01 vs. M2 = 46.38 ± 2.50; t50 = 2.69, p < 0.001; actual value = 50). Behavioral patterns are consistent with the sensitivity to contextual valuation, which was observed in healthy subjects, (SV estimates for M1 = 55.65 ± 2.04 vs. M2 = 38.9 ± 2.16; t49 = 6.09, p < 0.001), but to a lesser magnitude. After recalling autobiographical memories (POST-intervention), participants reported reduced negative affect (PANAS-X, p < 0.001), craving (p < 0.05) and withdrawal (p < 0.001), while at the same time exhibiting larger choice bias (p < 0.05) and value bias (p < 0.01) than PRE-intervention, bringing them closer to the behavior of healthy participants.

Conclusions: These preliminary findings suggest that MUD individuals exhibit attenuated context sensitivity in reward valuation compared to healthy controls. Given the test-retest reliability of the cRL task, the improved sensitivity to contextual valuation POST vs PRE-intervention seems to arise from the intervention and its effects on reducing negative affective and clinical states. This highlights the possibility to use affect-based behavioral interventions to causally reduce drug biases in value-based decision-making in addiction and accelerate the recovery of healthy-like behavioral patterns.

Keywords: Methamphetamine use disorder, computational models of decision-making, Negative Affect, Value-based Decision-Making

Disclosure: Nothing to disclose.

P323. Semaglutide and tirzepatide effects on weight and heavy alcohol consumption in overweight and obese hazardous drinking patients

Andrea King, Emily Atkinson, Claire Wellendorf, Daniel Fridberg, Silvana Pannain

University of Chicago, Chicago, Illinois, United States

Background: The incretin system is a promising new treatment target for substance use disorders as glucagon-like peptide-1 receptor agonists (GLP-1 RAs) modulate activity in brain regions involved in satiety signaling and substance-related reward, including the nucleus accumbens and ventral tegmental area. Long-acting GLP-1 RAs are approved for the treatment of type 2 diabetes, obesity, obstructive sleep apnea, and secondary cardiovascular prevention, with some early evidence suggesting they may also affect alcohol drinking behavior. As such, there is a critical need for real world studies to evaluate the effects of GLP-1 RA drugs across the continuum of excessive drinking. The present prospective observational study investigated weight and alcohol use outcomes among patients with overweight or obesity receiving GLP-1 RA therapy at the University of Chicago Weight Loss Clinic and reporting hazardous drinking.

Methods: Thirteen patients with overweight or obesity were prescribed semaglutide [n = 8] or tirzepatide [n = 5] along with standard care dietary counseling and physical activity recommendations. Standard dosing schedule and titration were adopted as recommended by the manufacturer and customized to individual tolerability and response. All but one patient met the body mass index (BMI) cutoff for obesity in the U.S. (n = 12, BMI ≥30 kg/m2; n = 1, BMI = 29.50 kg/m2). All patients screened positive for hazardous drinking based on the Alcohol Use Disorders Identification Test (AUDIT; score ≥8). Weight was measured in clinic and the AUDIT was re-administered on average 9.6 (±4.8) months after medication initiation. Subgroup analyses compared outcomes among patients with very high (score ≥15; n = 4 or high (score 8–14; n = 9) baseline AUDIT scores. Other analyses compared outcomes in patients taking semaglutide versus tirzepatide and those taking the maximum recommended maintenance dose of either medication (subcutaneous: semaglutide 2.4 mg, tirzepatide 15.0 mg) versus a lower dose (subcutaneous: semaglutide 1.0–2.0 mg, tirzepatide 7.5 mg).

Results: From baseline to follow-up, weight (kg) and BMI were significantly reduced (weight 112.8 vs. 96.3 kg: t[12] = 8.0, p < 0.001; BMI 39.9 vs. 34.0 kg/m2: t[12] = 7.8, p < 0.001). There was also a significant overall reduction in total AUDIT scores (14.0 vs. 3.6, t[12] = 6.1, p < 0.001) and AUDIT-C scores, i.e., the sum of the first three AUDIT items assessing quantity and frequency of alcohol drinking (8.00 vs. 3.23, t[12] = 7.4, p < 0.001). The magnitude of the AUDIT score reduction was greater for those in the very high AUDIT subgroup versus the high AUDIT subgroup (t[11] = 3.38, p = 0.006). In all but one case, patients in both subgroups showed a shift to lower-risk drinking (AUDIT < 8) at follow-up. Further, both semaglutide and tirzepatide were associated with significant reductions in weight and AUDIT scores with no difference in the magnitude of the reductions between groups (weight: semaglutide: 94.9 kg vs. 79.6 kg; tirzepatide 141.4 kg vs. 123.0 kg; t[11] = 0.74, p = 0.47; AUDIT: semaglutide, 14.3 vs. 3.3, tirzepatide, 13.6 vs. 4.0, t[11] = 0.35, p = 0.73]. Finally, patients on the maximum maintenance dose versus a lower dose of either medication did not differ on weight and AUDIT score reductions during treatment (weight: at maintenance dose, 112.4 kg vs. 96.6 kg; below maintenance dose, 113.3 kg vs. 95.9 kg; t[11] = 0.34, p = 0.73; AUDIT: at maintenance.

Conclusions: These findings support the association between GLP-1 RA treatment and reduced alcohol consumption, particularly for individuals with very hazardous alcohol use. Both semaglutide and tirzepatide were associated with reduction in weight and AUDIT scores in this sample, with no differences observed between effects of a below-maintenance dose relative to the maximum maintenance dose. Effect sizes were large, indicating a high association between GLP-1 RA medications and reductions in weight, BMI, and hazardous drinking scores. These findings underscore the importance of further research to better understand the acute mechanisms underlying GLP-1 RA therapies and aid in the development of pharmacotherapies for hazardous drinking.

Keywords: GLP-1 receptor agonist, Alcohol, Weight loss treatment

Disclosure: Nothing to disclose.

P324. Unravelling polygenic risk and environmental interactions in adolescent polysubstance use: a U.S. population-based observational Study

Jodi Gilman, Dongmei Zhi, Brandon Sanzo, A Eden Evins, Scott Hadland, Joshua Roffman, Richard Liu, Phil Lee

Harvard Medical School, Boston, Massachusetts, United States

Background: Adolescent polysubstance use (PSU), defined as the concurrent or sequential use of multiple psychoactive substances in a one-year period, is a pressing public health concern. In the United States (U.S.), roughly 20–25% of adolescents report using more than one substance in the past year. PSU exposure during adolescence may lead to more severe and long-lasting consequences than single substance use (SSU), including persistent cognitive impairments, psychiatric comorbidities, and elevated risk for substance use disorders (SUDs) in adulthood. Despite its significant public health impact, the underlying mechanisms of adolescent PSU remain poorly understood. This gap is critical, as adolescence represents a sensitive period when genetic predispositions interact with rapidly changing neurobiological and environmental factors.

Methods: This study used data from the ABCD Study (release 6.1), which included 11,868 participants aged 9 to 10 at enrollment. Substance use outcomes were derived from the Substance Use Interview, Substance Use Phone Interview (Mid-Year), and Hair Drug Toxicology data. We examined the associations of polygenic scores (PGSs) for addiction, derived from GenomicSEM analysis and environmental factors (including perinatal/early developmental events (N = 11), prenatal health exposures (N = 5), life events/lifestyle (N = 18), family (N = 31), neighborhood (N = 8), school (N = 3), peer relationship (N = 10)) with the PSU initiation, while controlling for multiple covariates. We also analyzed functional connectivity (FC) within and between 13 Gordon networks and 19 subcortical regions using resting-state fMRI data, and examined their associations with lifetime PSU initiation.

Results: Of 11,868 individuals recruited at baseline, 5,992 independent unrelated participants with PSU outcome data at Year 5 were included in this analysis. By Year 5, 18% of participants reported lifetime SSU, and 15.6% had initiated PSU. In the EUR group, PGSSUD, representing shared genetic liability across SUDs, was significantly associated with PSU, after covarying for age, sex, genetic PCs, and enrollment sites. The adjusted OR for PGSSUD was 1.43 (95% confidence interval (CI) = 1.30–1.57, PFDR = 2.45 × 10⁻¹²). The strongest risk factors for PSU included peer reputation victimization (OR = 1.44; 95% CI = 1.35–1.53; PFDR = 3.65 × 10⁻²⁶), mature-rated video games (OR = 1.41; 95% CI = 1.30–1.52; PFDR = 1.55 × 10⁻¹⁷), and prenatal tobacco use (OR = 1.35; 95% CI = 1.27–1.44; PFDR = 1.28 × 10⁻¹⁸). Conversely, the strongest protective factors included parental monitoring (OR = 0.63; 95% CI = 0.59–0.68; PFDR = 2.10 × 10⁻²⁹) and planned pregnancy (OR = 0.74; 95% CI = 0.69–0.80; PFDR = 1.22 × 10⁻¹³). When comparing those with PSU to those with no substance use, resting state network differences were few at baseline but robust by Year 4, with increased connectivity between subcortical regions and cingulo-parietal regions, and decreased connectivity between subcortical regions and ventral attention regions (FDR corrected > 0.05).

Conclusions: In this study, we examined how genetic susceptibility (“nature”) and environmental exposures (“nurture”) together influence adolescent PSU and its neural manifestations. To our knowledge, this is the first population-based study to explore both polygenic and environmental factors related to PSU in early adolescence. Our findings suggest that increased genetic liability for common addiction risk, as well as several environmental factors, is linked to a higher risk of PSU during this developmental stage. Although still in its nascent phase, understanding the genetic foundations of PSU combined with environmental factors in adolescents could aid in developing targeted prevention programs.

Keywords: polysubstance use, Adolescence, Brain development

Disclosure: Nothing to disclose.

P325. Recovery of cognitive impairments associated with chronic cannabis use following cannabis cessation: evidence from a prospective abstinence paradigm

Gabriel Ferland, Mathilde Argote, Edden Gitelman, Rachel Rabin

McGill University, Montreal, Canada

Background: Rates of cannabis use and cannabis use disorder are increasing, likely driven by changes in legalization policies. These trends are concerning because chronic cannabis use is associated with impaired academic and occupational performance, psychiatric symptoms, and psychosocial difficulties.

Cannabis binds to cannabinoid receptors which are densely expressed in the prefrontal cortex and hippocampus, regions critical for learning and memory. Notably, learning and memory are the cognitive domains most consistently impaired in individuals who use cannabis. However, it remains unclear whether these deficits persist following protracted abstinence, an important question as such impairments can translate into functional difficulties in daily life. While studies in adolescents using controlled abstinence paradigms have shown improvements in learning and memory with ~28 days of cannabis abstinence, evidence in adults is lacking. To address this gap, we examined whether sustained cessation leads to recovery of visual and verbal learning and memory in adults with chronic cannabis use during 28-days of monitored cannabis abstinence.

Methods: Forty-five adults (aged 19–55) who used cannabis nearly daily and/or met criteria for cannabis use disorder (CUD), without any current comorbid DSM-5 disorders, were enrolled. Twenty-eight participants were randomized to a cannabis-quit arm (CAN-), where they underwent 28 days of cannabis abstinence supported by contingency management and weekly behavioral support sessions; abstinence was verified biochemically via twice-weekly urine samples. Seventeen participants were randomized to a cannabis-as-usual arm (CAN+), where they were instructed to maintain their typical cannabis use patterns. To ensure comparable study procedures, the CAN+ group also received weekly behavioral support and provided twice-weekly urine samples.

Cognitive function was assessed at baseline and 28 days later using the Cogstate computerized battery. Visual learning was measured with the Continuous Paired Associative Learning (CPAL) task, while verbal learning and memory were assessed using the International Shopping List (ISL) task. A comparison sample of healthy volunteers (HV; n = 9) was included to evaluate whether baseline cognitive performance was impaired in people with cannabis use relative to HV. Baseline group differences (HV vs. CAN+ and CAN-) were tested using ANCOVAs. For tasks showing significant group effects, changes in cognition between baseline and day 28 (CAN+ vs. CAN-) were examined using repeated-measures ANCOVAs. All analyses controlled for sex and education, with Bonferroni correction applied to post hoc tests.

Results: Among the CAN- group, 22 (79%) participants were verified as abstinent. At baseline, HV outperformed the cannabis-using group on visual and verbal learning (p < 0.05), but not verbal memory (p = 0.16).

A significant time x group effect for visual learning [F(1,32) = 4.99, p = 0.03] and a trend for verbal learning and memory [F(1,32) = 4.07, p = 0.05] emerged. Post hoc tests revealed significant improvements in the CAN- group during abstinence for both visual (p = 0.015) and verbal learning (p < 0.01), with no significant changes in the CAN+ group (p = 0.43 and p = 0.23, respectively).

Conclusions: These findings suggest that learning deficits associated with chronic cannabis use improve with sustained abstinence, indicating that such impairments reflect functional alterations in brain processes rather than irreversible neurotoxic effects. Clinically, cognitive recovery may enhance motivation for behavioral change and support engagement in treatment. Future research should employ prospective designs following individuals prior to cannabis initiation to determine whether learning performance returns to pre-use levels following abstinence.

Keywords: cannabis use, cannabis abstinence, cognition, learning and memory

Disclosure: Nothing to disclose.

P326. Additive effects of alcohol and cannabis on brain activity and neurocognitive function: preliminary results from an electroencephalography study

Hollis Karoly, Eleftherios Hetelekides, William Gavin, Patricia Davies

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

Background: As a wide variety of highly potent legal-market cannabis products become accessible throughout the US, driving under the influence of cannabis has emerged as an urgent public safety concern. Moreover, alcohol and cannabis are frequently used together, and their combination is known to impair driving ability, though the extent and underlying neurophysiological mechanisms of this impairment have not been well characterized. Data are conflicting regarding whether co-intoxication produces additive, synergistic, or less-than-additive effects on impairment compared to single-substance alcohol or cannabis intoxication. The present study uses electroencephalography (EEG) recorded during two driving-relevant cognitive tasks (the memory update task, designed to assess brain activity related to working memory, and the speeded visual flanker task, which assesses selective attention, prepotent response inhibition, and performance monitoring) to shed light on the neurophysiology of co-intoxication. The present analyses focus on differentiating phases of cortically-mediated processing (using event-related potentials [ERPs]) such as decision-making (P3), error detection (ERN) and post-error adaptation (PE). We hypothesize that co-intoxication will produce at least additive impairments in neurocognitive performance, and suppression of ERPs compared to alcohol-only and cannabis only.

Methods: 31 adults who report heavy drinking, regular cannabis use, and regularly combining alcohol and cannabis, have been recruited to participate in a single session in our mobile lab. Participants were randomly assigned to consume either alcohol (n = 11), cannabis (n = 13), or both alcohol and cannabis (n = 7) during the session. Participants self-administered their own cannabis product ad libitum inside their residence while our team remained in our mobile lab parked outside their residence. Participants received a standard dose of alcohol (designed to bring breath alcohol concentration to 0.06 g/dL), prepared and administered by our staff inside the mobile lab. Cognitive performance and brain activity measures were obtained during the memory update and flanker tasks immediately before alcohol and/or cannabis use (T1), 1-hour post-use (T2), 2 hours post-use (T3), and 4 hours post-use (T4). For the memory update task, EEG recordings were processed to obtain amplitude and latency values for stimulus-locked event-related potential components: (N1, P2, N2, P3). For the flanker task, EEG recordings were processed to obtain amplitude and latency values for two brain measures associated with incorrect decisions, error related negativity (ERN; detection), and error positivity (PE; adaptation). Performance (response time, accuracy) was calculated for both tasks.

Results: For the memory update task, there were no significant group differences for errors and response times. Repeated measures ANOVA (RM-ANOVA) of the P3 amplitude demonstrated a significant effect for time (p < 0.001) and stimulus type (p = 0.049). A group by time interaction demonstrated differential changes in P3 amplitudes across time; the alcohol group had a linear decrease in P3 amplitude across time while the cannabis and co-use groups demonstrated quadradic trends with the co-use group showing a maximal decrease in P3 amplitude at T2 and the cannabis group showing a maximal decrease at T3. For N1 amplitude we found a main effect for time (p = 0.002) and stimulus type (p < 0.001). The time by group interaction showed differential changes in N1 amplitudes across time; the alcohol and co-use group had a quadradic trend with greatest increase for alcohol at T3 and co-use at T2 while the cannabis group demonstrated minimal change across time. For the flanker task, there was a significant linear decrease in errors across time (p = 0.03), but no group or interaction effects. We observed a significant main effect of time (p = 0.027) for response time with alcohol and co-use groups showing a quadratic effect (p = 0.015), but the cannabis group showing linear decreases. RM-ANOVA for ERN amplitude revealed a main effect of time (p = 0.005) and a group by time interaction (p = 0.007). For PE, a significant main effect of time (p = 0.001) was found. All groups showed significant quadratic effects of amplitude attenuation across sessions for ERN and PE (p < 0.005), such that amplitude attenuation increased from baseline to peak intoxication and showed evidence of recovery at the final timepoint. At peak intoxication, the co-use group had greater attenuation than the alcohol or cannabis groups (p < 0.05).

Conclusions: Each group showed unique brain responses during both tasks, but minimal differences in performance, suggesting that neural measures of working memory and attention may be more sensitive to intoxication effects than behavioral measures. The reduced ERN and PE amplitudes across intoxication conditions during the flanker task highlights deficits caused by alcohol and cannabis on inhibitory control as well as demonstrating a differential effect of intoxicants on cognitive processing. Amplitude suppression in ERPs were greater in the co-use group than both single-substance groups, showing evidence of additive effects. Data collection is ongoing, and analyses will be repeated with the full sample. Ultimately, results could lead to future studies aimed at developing targeted behavioral roadside tests to detect driving under the influence of different substances.

Keywords: Alcohol, cannabis, Polydrug Use

Disclosure: Nothing to disclose.

P327. Sociocultural factors, personality and cannabis use: a cross-cultural examination of belief-mediated pathways

Erin Prince, Francesca Filbey

The University of Texas at Dallas, Dallas, Texas, United States

Background: The Ecological Systems Theory emphasizes that individual behavior is deeply embedded within broader social and cultural contexts. In the case of cannabis use (CU), drug use behavior is not solely shaped by individual traits but is also shaped by sociocultural factors, including social norms, cultural attitudes, and evolving regulatory landscapes. For example, recent cannabis legalization in the U.S. has led to shifts in cannabis attitudes, particularly among emerging adults (aged 18–29) who reported more permissive views and lower risk perceptions (Carliner et al., 2017; Waddell, 2021). These changes in cannabis beliefs coincide with rising rates of cannabis use disorder (CUD); notably, 16.6% of users (18–25) now meet CUD criteria (Beirut and Fang, 2025). While individual differences in personality traits, such as conscientiousness, agreeableness, and neuroticism, have been linked to CU (Herchenroeder et al., 2021), how these individual factors interact with sociocultural factors have yet to be determined. Taken together, a need to understand how sociocultural environment interacts with individual factors such as personality is needed to inform prevention and intervention strategies that are culturally and contextually sensitive. The aim of this study was to understand the interaction between personality and sociocultural factors impacting cannabis use behaviors.

Methods: We recruited 244 adults who use cannabis (129 males, mean age = 28.8) from two culturally and legally distinct sites: Texas (US, n = 100), where cannabis remains highly regulated, and Amsterdam (Netherlands, n = 144), where cannabis has long been decriminalized. Participants completed the NEO Five-Factor Inventory, Cannabis Culture Questionnaire, General Ethnicity Questionnaire, and the Cannabis Use Disorder Identification Test – Revised. Network analyses using R packages igraph (v2.1.4) and bootnet (v1.6) were used to examine the relationships between personality traits, sociocultural factors, and CU. Post-hoc mediation analyses using Hayes’ PROCESS macro (v4.2) in SPSS (version 29.0.0.0) were conducted to explore potential mechanisms within significant pathways. Additionally, we examined whether network structure differed between sites and sexes using NetworkComparisonTest (v2.2.2) package in R. Significant results were determined using a p-value < .05, corrected using bootstrapped confidence intervals (CIs) based on 1000 resamples.

Results: Network analyses revealed that personality traits clustered closely with CU behaviors, while cannabis culture belief domains across individual, social network, and societal levels formed a separate cluster. Within this network, individual-level positive cannabis beliefs emerged as a central node directly connecting to CU behaviors. This was reflected in its centrality and a strong positive edge weight to CU behaviors (weight = 0.33). Among personality traits, higher neuroticism (weight = 0.16), lower agreeableness (weight = −0.14), and lower conscientiousness (weight = −0.10) showed direct associations with greater CU behaviors. Post-hoc mediation analyses further demonstrated that positive individual cannabis beliefs significantly mediated the relationship between agreeableness and CU behaviors. Specifically, agreeableness negatively predicted positive cannabis beliefs (b = −0.23, p < .001), which in turn positively predicted cannabis use severity (b = 0.83, p < .001). The indirect effect was significant (b = −0.19, 95% CI [−0.35, −0.06]), indicating that lower agreeableness contributes to greater CU behaviors in part through increased permissive cannabis beliefs. Although there were qualitative differences in network structure between sites and sexes (e.g., female network has denser connectivity and higher centrality for cannabis belief nodes, specifically centrality of individual positive beliefs, perceived positive beliefs from family and friends, and perceived negative societal beliefs), the network comparison results did not reach a level of significance.

Conclusions: Cannabis-related belief pathways demonstrate indirect effects on CU behaviors mediated by individual positive cannabis beliefs, particularly agreeableness. Individuals high in neuroticism and social conformity traits (agreeableness, conscientiousness) may be particularly sensitive to cultural messaging. This suggests that permissive cannabis beliefs are a critical psychological mechanism driving cannabis use behaviors, highlighting the importance of targeting these beliefs in prevention and intervention efforts, especially among individuals with certain personality profiles.

Keywords: cannabis use disorder, Big Five Personality Factors, Sociocultural Factors, Psychometrics, SUD risk

Disclosure: Nothing to disclose.

P328. A randomized, double-blinded trial of a ketone monoester to supplement standard of care in medically supervised detoxification: clinical and neurochemical outcomes

Marco Møller, Anastasia Coppoli, Brian Pittman, Jakob Damsgaard, Karen Kettless, Gitte Moos Knudsen, Graeme Mason, Anders Fink-Jensen

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

Background: The liver converts most alcohol to acetate, which the brain oxidizes. Chronic alcohol may partly shift brain energy from glucose to acetate, potentially worsening withdrawal. A ketogenic diet has been shown to reduce withdrawal symptoms, but it takes several days to reach ketosis. Ketone monoesters (KME) raise blood beta-hydroxybutyrate (BHB, a ketone body) to ketosis within 30 minutes and in mice have been reported to reduce withdrawal symptoms. We hypothesized KME could facilitate alcohol detoxification in people, funded by Mental Health Services in the Capital Region of Copenhagen Foundation.

Methods: We recruited 28 subjects seeking treatment for AUD and supplemented standard care with KME (n = 13) or placebo (n = 15) for 3 days, to assess KME effects on withdrawal and 1-month sobriety. The primary outcome was the quantity of benzodiazepines used during detoxification. Secondary outcomes included withdrawal, craving, physiology, and mood. After detoxification, 4 KME and 9 placebo subjects underwent Magnetic Resonance Spectroscopy (MRS) on a Siemens Magnetom 3T Prisma with a 64-channel head coil to measure brain BHB, lactate, GABA, glutamate + glutamine (Glx), N-acetylaspartate (NAA), and other metabolites before and after a single 25g KME dose. The KME/MRS was repeated at 1 month. The KME/MRS challenge was also done once with light/non-drinkers (LD; n = 13), heavy non-dependent drinkers (HD; n = 9), and people in recovery >6 months (LTS; n = 7).

Results: Benzodiazepine use peaked on Day 1 of withdrawal and decreased on Days 2 and 3 with no KME/placebo differences. KME participants reached 1.5–2 mM blood BHB each treatment day. Benzodiazepine use and all other measures, except MRS, were similar between KME and placebo groups. With the KME/MRS challenge, the HD group showed lower brain BHB post-KME compared to LD (p = 0.03) and AUD (p = 0.02) and LTS (P = 0.07), suggesting potentially faster brain utilization, as previously reported with acetate. No other group differences or acute KME effects were observed.

Conclusions: The absence of significant effects on withdrawal symptoms raises a critical question: how much benefit of ketosis is attributable to ketone bodies vs concurrent carbohydrate reduction? Nutritional therapies for AUD could be very desirable, and to understand efficacy and mechanisms, larger studies are needed.

Keywords: Alcohol Use Disorder - Treatment, Ketogenic diet, 1H MRS, ketone monoester

Disclosure: Merck Sharp and Dohme, Consultant, Self, Merck Sharp and Dohme, Consultant, Spouse/Partner, Leal Therapeutics, Consultant, Self

P329. Modulation of addiction and pain neurocircuitry with transcranial ultrasound

Brian Mickey, Brandon Cooper, Sarah Kwon, Katya Brooun, Eric Garland, Rana Jawish, Vincent Koppelmans, Jessica Ranger, Ara Slaugh, Tom Riis, Jan Kubanek, Taylor Webb

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

Background: Use and misuse of opioids have increased in recent decades, with serious social and medical consequences. A common pathway to opioid use disorder is long-term use of prescription opioids for chronic pain. Non-opioid interventions that target the deep neural circuitry underlying pain and addiction are sorely needed for this population. To address this need, our team has developed a new ultrasound-based device to modulate deep brain networks noninvasively, directly, and precisely in human subjects.

Methods: We are using a non-significant-risk, MRI-compatible device called Diadem that can deliver low-intensity ultrasound to deep brain targets in humans noninvasively with millimeter precision. This unique approach uses two ultrasound arrays to measure and compensate for the acoustic distortion caused by the head, enabling the delivery of a known ultrasound intensity to the intended brain target. Furthermore, the ultrasound focus can be steered to a range of targets without moving the subject or device, allowing sequential or near-simultaneous multi-focal stimulation. For this ongoing study, the total enrollment goal is 31 participants; 14 have enrolled so far and data are available from 10 subjects. Eligible subjects have a diagnosis of chronic pain, have been taking opioid medication for at least 3 months, and are misusing opioids as determined by the Current Opioid Misuse Measure (COMM). Subjects receive sham or real low-intensity focused ultrasound in randomized order under double-blind conditions. Ultrasound is delivered sequentially to bilateral nucleus accumbens (NAc) and anterior cingulate cortex (ACC) during an hour-long stimulation session. Changes in pain intensity, opioid craving, and mood are self-reported with numeric rating scales. To quantify target engagement, functional MRI is performed “online” while ultrasound is delivered to the NAc or ACC using a 10-minute block design with 60-second alternating ON and OFF blocks.

Results: The average age of participants was 51 years (range 34–67) and 60% were female.

The mean COMM score was 13.4 (range 10–17), indicating high risk of misuse of prescription opioids. Ninety percent of participants had a comorbid lifetime diagnosis of major depressive disorder. Prior to ultrasound stimulation, the mean ratings for pain intensity, “wanting” opioid medication, and anxiety were 4.7, 2.0, and 3.2, respectively, on 0–10 numeric rating scales. After stimulation, pain decreased more with real versus sham stimulation (Cohen’s d = 0.44), wanting increased more with the sham condition (d = 0.95), and anxiety decreased more with the real condition (d = 0.50). None of these differences reached statistical significance with this limited sample size (p > 0.05). Individual-subject functional MRI analyses demonstrated significant intensity-dependent deactivation of the targeted ACC region in one participant (p < 0.001, corrected). No significant modulation was detected in 5 other subjects (p > 0.05). Group-level analyses are planned. Forty percent of participants exited the study early because they did not tolerate study procedures, which require the subject to lie still for extended periods across multiple study visits. All participants tolerated the stimulation itself, and no serious adverse events occurred.

Conclusions: Patients with chronic pain who are misusing opioids are at high risk of progression to opioid use disorder. We have developed a safe and flexible ultrasound-based approach that could prevent this progression by applying multi-focal stimulation that concurrently addresses pain, drug craving, and mood disturbances. The findings reported here derive from a small sample and reflect a work in progress. Challenges remain to clinical translation for this population. As with other noninvasive brain stimulation interventions, target engagement may be difficult to demonstrate at the individual-subject level. Optimization of ultrasound delivery and novel imaging methods may be needed to confirm that ultrasound delivered to a deep target is having the expected physical or neurophysiological effects. Additionally, this population of patients may have difficulty tolerating noninvasive interventions that require long periods of immobility. Future approaches that allow freer movement, such as wearable devices, may be needed to scale up ultrasound interventions to a broader range of users. We plan to address these challenges in future work.

Keywords: substance use disorders, Neuromodulation, focused ultrasound, chronic pain, Prescription Opioids

Disclosure: Health Rhythms, Contracted Research, Self, LivaNova, Contracted Research, Self, Compass Pathways, Contracted Research, Self, Abbott, Contracted Research, Self, Inside Edge, Consultant, Self, VML, Consultant, Self, Atheneum, Consultant, Self, Guidepoint, Consultant, Self, Kx Advisors, Consultant, Self, S2N Health, Consultant, Self

P330. Unveiling the neurometabolic signature of addiction: meta-analysis of 1H-MRS studies across substance and behavioral dependencies

Shinichiro Nakajima, Fumihiko Ueno, Takahide Etani, Koki Takahashi, Hideaki Yasuda, Masataka Wada, Shiori Honda, Sunjun Huh, Ryoma Kani, Marie Kitano, Shun Mizutori, Sotaro Moriyama, Kanta Niinomi, Yoshito Saito, keita Taniguchi, Hiroyuki Uchida, Ariel Graff-Guerrero

Keio University, Tokyo, Japan

Background: Addiction is a pervasive disorder affecting a wide range of substances and behaviors, with significant societal impact. However, the effectiveness of treatment options remains limited, mainly due to an incomplete understanding of the underlying mechanisms. Previous studies have noted neurometabolic alterations across various forms of addiction employing proton magnetic resonance spectroscopy (1H-MRS). This study aimed to examine cross-diagnostic characteristics of neurometabolite levels in this population by conducting a meta-analysis of these studies.

Methods: A systematic literature search was conducted using PubMed and Embase, targeting original research articles that compared eight neurometabolites (glutamate [Glu], glutamine [Gln], Glu+Gln [Glx], gamma aminobutyric acid [GABA], myo-inositol [mI], N-acetylaspartate [NAA], choline [Cho], creatine [Cre]) between individuals with addictions (alcohol, drug, poly-substance [concomitant use of alcohol and/or drugs], nicotine/tobacco, gambling, and internet/gaming) and controls. Effect sizes for differences in these neurometabolites across 14 regions of interest (ROIs) were calculated using a random-effects model. Meta-regression analyses were also performed based on participant demographics (average age and male ratio) and clinical characteristics (starting age and duration of substance use/behavior).

Results: A total of 111 articles were identified to be relevant from an initial pool of 5,013 records. Higher levels of Glx were observed in the insula and medial frontal cortex (MFC) among the addiction group, with subgroup analyses revealing elevated Glx in the MFC during active addiction. Glutamate levels were elevated in several regions, including the insula, thalamus, basal ganglia, and occipital cortex, in alcohol and opioid/opiate groups, whereas decreases were noted in the basal ganglia and dorsolateral prefrontal cortex in cocaine and nicotine/tobacco groups, respectively. No significant differences were found for glutamine levels between the groups. GABA levels were decreased in the MFC and occipital cortex in addiction, especially during active addiction. mI levels were higher in frontal white matter (WM) and lower in parieto-occipital cortex in addiction, with abstinence associated with increased mI in frontal gray matter and WM. NAA levels were generally reduced across multiple frontal and temporal regions in addiction, notably lower in active and abstinent subgroups. Cho and Cre levels showed no consistent group differences overall. Participant age, gender ratio, and duration of substance use were significant moderators for various neurometabolite levels.

Conclusions: This comprehensive meta-analysis identifies shared and distinct neurometabolic alterations across different forms of addiction, thereby providing a foundation for targeted therapeutic interventions.

Keywords: 1H MRS, Alcohol and substance use disorders, gambling disorder, Addictive behavior

Disclosure: Nothing to disclose.

P331. Age of onset and co-morbidity of binge eating with alcohol use disorder, affective disorder, anxiety disorder, and suicide in American Indian (AI) and Mexican American (MA) young adults: genetic correlation with GLP-1 and NBEA

Cindy Ehlers, David Gilder, Qian Peng

The Scripps Research Institute, La Jolla, California, United States

Background: Binge eating and binge eating disorder emerge during adolescence and early adulthood and are associated with several other mental health conditions. Binge eating and binge drinking may have common neuro-substrates that are both impacted by GLP-1 therapies. Both American Indians and Mexican American young adults have high rates of obesity, compared to EuroAmericans, however, the comorbidity and clinical course of binge eating in these two groups is less understood. This study sought to document the age of onset and comorbidity of binge eating in young adult Mexican and American Indians.

Methods: The AI participants resided on or near 8 geographically contiguous reservations. MA participants were recruited using a commercial mailing list that provided the addresses of individuals with Hispanic surnames in 11 zip codes in San Diego County. All participants had to be of MA or AI ancestry, be between the ages of 18 and 30 years, be residing in the United States legally, and be able to read and write in English. Each participant completed an interview with the Semi- Structured Assessment for the Genetics of Alcoholism (SSAGA) which collected information including demographics, psychiatric disorders, and symptoms of SUDs. A blood sample was collected for genetic analyses. For the purpose of this study, eating binges were defined as: eating a large amount of food in a short period of time (usually less than 2 hours). They were also queried as to: How old were you when you first went on an eating binge? And to whether they went on eating binges as often as twice a week for at least 3 months (DSM-4 binge eating disorder).

Results: Out of 1280 participants 286 reported binge eating. Those with binge eating did not differ from those who did not endorse this item on age, education, race, gender, employment marriage or econstat. However, those with binge eating were significantly more likely to be diagnosed with alcohol use disorder, anxiety disorder, affective disorder and suicidal thoughts and acts (all p < 0.0001, Odds Ratios: 2.7–3.0). Binge eating had a mean age of onset of 16 yrs, however those who began binging under the age of 15 had a significant higher likelihood of developing binge eating disorder (p < 0.009). A genomic scan of the GLP1A and NBEA genes was conducted for the binge eating phenotype and a significant finding was found on or near both GLP1R and NBEA genes in the American Indians (GLP1R: 1.27 kb downstream p < 0.02; NBEA 21.3 kbp downstream p < 0.004) and Mexican Americans (GLP1R: intron p < 0.003; NBEA intron p < 0.05).

Conclusions: These data suggest that binge eating has a genetic component and that especially with early age of onset can lead to eating disorders and is comorbid with substance abuse and affective/anxiety disorders in Mexican and American Indian young adults, highlighting the importance of screening for this phenotype in these racial groups

Keywords: Alcohol, Suicide risk factors, Binge eating

Disclosure: Nothing to disclose.

P332. Neurofunctional domain profiles and clinical allostatic load markers in individuals with alcohol use disorder and comorbid substance use and mood disorders

Tommy Gunawan, Lacey Donahue, Nicholas Weaver, Melanie Schwandt, Bethany Stangl, David Goldman, Nancy Diazgranados, Vijay Ramchandani

National Institute on Alcohol Abuse and Alcoholism, Bethesda, Maryland, United States

Background: Alcohol use disorder (AUD) is a chronic relapsing disorder associated with substantial heterogeneity in etiology, presentation and outcomes. AUD is commonly associated with comorbid substance use disorders (SUD) and mood disorders (MD), which contribute to variability in neurofunctional characteristics including incentive salience (IS), negative emotionality (NE), and executive functioning (EF). In addition, chronic alcohol misuse may exert an “allostatic load”, or cumulative dysregulation of neuroendocrine, metabolic, cardiovascular and immune functions, that can be quantified using a range of common clinical laboratory test measures. The objective of this project was to evaluate heterogeneity in the neurofunctional domain characteristics of individuals with AUD with and without comorbid SUD and MD, and differences in allostatic load markers between groups.

Methods: Data were collected from 1,494 (42.9% female, 54.6% Non-White) participants enrolled in the NIAAA Natural History Protocol. Psychiatric diagnoses were determined via Structured Clinical Interview for DSM-IV/DSM-5 (SCID). Clinical measures included cardiovascular measures (systolic and diastolic blood pressure, heart rate, total cholesterol, HDL-cholesterol), metabolic measures (hemoglobin A1C, albumin, body mass index), and inflammatory measures (c-reactive protein, neutrophils, lymphocytes). For each measure, the proportion of high-risk individuals, i.e., those with levels above established clinical cut-off values, was determined. A set of validated psychological and behavioral measures of craving, mood, anxiety, stress, personality, aggression, impulsivity, urgency and delay discounting were analyzed using factor analysis to derive neurofunctional domain scores (IS, NE, EF). The clinical allostatic load markers and IS, NE, and EF scores were compared across the non-AUD, AUD-only and comorbid groups. Age, sex and race were evaluated as covariates.

Results: Among 902 individuals with AUD, 20.9% had comorbid SUD (AUD + SUD), 9.6% had comorbid MD (AUD + MD), 14.6% had comorbid SUD and MD (AUD + SUD + MD), and 36.0% had neither MD nor SUD (AUD-only). Individuals with AUD + SUD were older and more likely to be non-White, individuals with AUD + MD were more likely to be female and White (p’s < 0.001). Individuals with AUD + SUD + MD had the lowest income and highest rates of anxiety and post-traumatic stress disorders (p’s < 0.001). The AUD + SUD + MD group exhibited the highest IS and NE scores and lowest EF scores (indicating greater executive dysregulation) (p’s < 0.001). AUD + MD exhibited greater NE, and lower EF scores relative to AUD + SUD (p’s < 0.001), and no differences in IS scores. Comorbid groups exhibited greater high-risk levels of heart rate, HDL cholesterol, albumin, and neutrophils relative to those without comorbidities (p’s < 0.03).

Conclusions: Most individuals with AUD had a comorbid SUD or MD diagnoses. Each comorbidity group exhibited distinct sociodemographic and neurofunctional profiles. AUD + MD showed pronounced NE and EF dysregulation, while AUD + SUD + MD showed the greatest impairments across domains. Elevated risk profiles in clinical allostatic load markers were detected across diagnostic groups, providing unique insights into the psychobiological underpinnings of the heterogeneity of AUD and comorbid SUD and MD. Understanding the sources of this heterogeneity may provide critical insights into the etiology and treatment of addiction and associated psychiatric comorbidities.

Keywords: Alcohol and substance use disorders, mood disorders, allostatic load, biomarkers, neurofunctional domains

Disclosure: Nothing to disclose.

P333. An examination of the SLC39A8 rs13107325 polymorphism with alcohol-related behaviors in individuals with and without a history of alcohol use disorder

Bethany Stangl, Theodore Koide, Shyamala Venkatesh, Nancy Diazgranados, David Goldman, Mariella De Biasi, Vijay Ramchandani

NIAAA, Bethesda, Maryland, United States

Background: Multiple genome-wide association studies have identified the SLC39A8 rs13107325 (C > T) polymorphism as a risk locus for alcohol use disorder (AUD) and related phenotypes. The single nucleotide polymorphism (SNP) rs13107325 is a missense mutation that affects divalent ion transport. In humans, the SLC39A8 rs13107325 C allele is over-represented in non-alcohol dependent controls, while the SLC39A8 rs13107325 T allele is a risk allele for heavy drinking and AUD, and thus it may influence alcohol response and related behaviors. Therefore, our aims were (1) to examine the association of the rs13107325 polymorphism on alcohol response phenotypes in participants with a diverse range of alcohol use and (2) to examine the association of the polymorphism in a subsample that completed an intravenous alcohol self-administration (IV-ASA) session and subjective response measures to alcohol in non-AUD drinkers. We hypothesized that participants carrying the rs13107325 SNP (CT/TT genotype) would have differences in alcohol use and alcohol-related problems and less sensitivity to alcohol compared to individuals carrying the CC allele.

Methods: We examined the association of rs13107325 (CT/TT vs CC) with alcohol phenotypes in 1,474 male and female participants from the NIAAA Natural History Protocol. Given the much higher prevalence of the polymorphism in European ancestry participants (76.99%), subsequent analyses focused exclusively on this demographic (n = 660; CT/TT = 87, CC = 573). A sub-sample of participants were recruited for an Intravenous Alcohol Self-administration (IV-ASA) study (n = 139; CT/TT n = 17, CC n = 122). Alcohol phenotypic measures included the Alcohol Use Disorders Identification Test (AUDIT), Timeline Follow Back (TLFB), Lifetime Drinking History (LDH) questionnaire, the Self-Rating of the Effects of Alcohol (SRE), and the Alcohol Effects Questionnaire (AEFQ). Measures from the IV-ASA experimental session included peak and average breath alcohol concentration (BrAC), number of button presses for alcohol, and total ethanol consumed during the session in grams (EtOH). Subjective responses were assessed using the Drug Effects Questionnaire (DEQ), Alcohol Urge Questionnaire (AUQ), and Biphasic Alcohol Effects Scale (BAES) and an additional sensitivity to alcohol questionnaire was conducted called the Alcohol Sensitivity Questionnaire (ASQ). Our analyses incorporated age, sex, smoking status, and AIMs (Ancestry Informative Markers) scores for African, European, and Asian ancestries as covariates.

Results: In our sample of European ancestry participants, those with the CT/TT genotype reported lower levels on multiple drinking-related variables compared to those with the CC genotype. Specifically, significant genotype differences were observed in AUDIT-Dependence scores (p = 0.047), total drinks (p = 0.021), average drinks per day (p = 0.048), drinks per week (p = 0.017), and drinks per 30 days (p = 0.017). Self-reported alcohol sensitivity also differed by genotype, with lower scores among CT/TT individuals on the SRE-Heaviest (p = 0.036) and SRE-Total (p = 0.046) scales. There was a significant effect of sex and smoking status. For the IV-ASA measures in the subsample, analyses revealed no significant differences between genotypes for all IV-ASA measures and most subjective measures. However, participants with the CT/TT genotype reported significantly less feelings of sedation (p = 0.04) and more drinks needed to feel the descending effects of alcohol (p = 0.04).

Conclusions: Overall, there was substantial ancestry-based variation in the frequency of the SLC39A8 rs13107325 T allele, replicating similar reports in the allele frequency distributions in population-level genetic databases. Findings indicate that the rs13107325 polymorphism is associated with reduced alcohol use and greater sensitivity in CT/TT carriers in those of European ancestry, which aligns with some earlier reports suggesting a protective effect of the CT/TT genotype. While this reduced alcohol use was not replicated in the IV-ASA sample, we did see CT/TT carriers feeling less sedated from alcohol and reporting needing more drinks to feel the descending effects of alcohol. Future studies in larger diverse samples are necessary to better understand these differential effects by genotype.

Keywords: Alcohol, ZIP8, SLC39A8 genotype

Disclosure: Nothing to disclose.

P334. A simple prognostic score predicts pain and overall health in a natural history sample of individuals across the spectrum of alcohol use including alcohol use disorder

Melanie Schwandt, Vijay Ramchandani, Nancy Diazgranados, David Goldman

National Institute on Alcohol Abuse and Alcoholism, Bethesda, Maryland, United States

Background: Chronic pain affects as many as 10%–20% of people in the general population. However, approximately half of individuals with alcohol use disorder (AUD) suffer from chronic pain. The interaction of biological, psychological, and social factors (biopsychosocial model) has been associated with both AUD and chronic pain and serves as a basis for better understanding and treatment of these disorders. A recent study utilizing the UK Biobank identified a set of 99 biopsychosocial measures that were predictive of both chronic pain and the spreading of pain across multiple body sites. Furthermore, a “risk of pain spreading” (ROPS) score based on just six binary coded items assessing sleep disturbance, neuroticism (feeling “fed up”), mood (“feeling tired or have little energy”, and “seen a general practitioner or psychiatrist for anxiety or depression”), life stressors (illness/injury/death of a loved one, divorce, and/or financial difficulties), and body mass index (> 30 kg/m2) performed similarly to the full 99 items, and was associated with physical health, depression symptoms, and the inflammatory marker c-reactive protein (CRP). Predictive validity in a sample enriched for problematic alcohol use has yet to be investigated. The aims of the current study were: 1) to calculate an equivalent “risk of pain spreading” score using the NIAAA Natural History sample, which is enriched for AUD, 2) to evaluate whether ROPS score accurately predicts pain in this population, and 3) to test and potentially replicate associations between ROPS score and number of pain sites, physical and mental health, and inflammatory biomarkers.

Methods: The sample included 1101 participants (mean age = 40.1 years; 55.8% male; 46.6% White; 37.9% Black; 50.6% with current AUD) enrolled in the NIAAA Natural History Protocol. Within the full sample, n = 413 (37.5%) were seeking inpatient treatment for AUD. Participants were assessed using the Structured Clinical Interview for DSM Disorders and completed the Whole Body Pain Scale, as well as assessments of sleep, mood, current life stressors, quality of life (including overall health), and recent alcohol consumption. Biological markers examined included CRP and the systemic inflammation response index (SIRI), a peripheral measure of inflammatory response based on number and type of circulating immune cells. The ROPS score was calculated based on 6 items equivalent to those used in the UK Biobank study, all scored as 1/0 with the resulting score ranging from 0 to 6. Analyses investigated the association of ROPS with presence of any pain using logistic regression, and with the number of pain sites, percent of daily awake time in pain, overall physical health, depression symptoms, and biomarker measures using linear regression. Separate analyses were performed in the full sample and in the subsample of treatment-seeking AUD inpatients (AUD+Tx).

Results: The mean ROPS score was 2.6 (stdev 1.3) in the full sample and 3.5 (stdev 1.2) in the AUD+Tx subsample. Analyzed by itself, ROPS score predicted the presence of any pain with moderate accuracy in both the full sample (AUC = 0.71, p < 0.0001) and the AUD+Tx subsample (AUC = 0.77, p = 0.05). Including demographic measures (age, sex, race), average number of drinks per week, and AUD diagnosis (full sample model only) in the logistic regression led to improved accuracy especially in the AUD+Tx group (full sample: AUC = 0.81, p < 0.0001; AUD+Tx: AUC = 0.93, p = 0.03). ROPS was associated with a greater number of pain sites (partial R2 = 0.19), greater percent of time in pain (partial R2 = 0.16), decreased overall physical health (partial R2 = 0.31), increased depression symptoms (partial R2 = 0.37), and higher CRP levels (partial R2 = 0.05) in the full sample (all p < 0.0001). Similar findings were evident in the AUD+Tx sample for number of pain sites (partial R2 = 0.12), percent of time in pain (partial R2 = 0.06), overall physical health (partial R2 = 0.16), depression symptoms (partial R2 = 0.26), and CRP (partial R2 = 0.01) (all p < 0.0001 except CRP, p = 0.0007). There were no significant associations between ROPS and SIRI.

Conclusions: Our results complement those from the large UK Biobank study by validating the use of a simple score based on six biopsychosocial measures to predict pain, pain spreading, and other physical and mental health outcomes. Importantly, our findings replicate these associations in a racially diverse sample enriched for problematic alcohol use and AUD. These same associations were present in the subsample of individuals seeking treatment for AUD, a clinically severe population with a high prevalence of pain. Identifying vulnerability to pain in treatment-seeking AUD patients can inform interdisciplinary approaches aimed at managing both pain and addiction, potentially improving recovery outcomes.

Keywords: Alcohol Use Disorder, Pain, Inflammation, Depression

Disclosure: Nothing to disclose.

P335. Influence of zolmitriptan maintenance on pharmacodynamics of intravenous cocaine in humans

William Stoops, Joshua Lile, Joseph Alcorn III, Kevin Hatton, Lon Hays, Danielle Anderson, Janet Neisewander

University of Kentucky, Lexington, Kentucky, United States

Background: Serotonin1b (5-HT1b) receptors play an important role in preclinical cocaine effects. Rodent data show that zolmitriptan, a commercially available selective 5-HT1b agonist migraine medication, selectively attenuates the reinforcing and other abuse-related effects of cocaine. The goal of this project was to advance these promising preclinical findings into humans, thereby demonstrating that the 5-HT1b system plays a key role in the abuse-related effects of cocaine in people with cocaine use disorder (CUD). We hypothesized that zolmitriptan would attenuate the abuse-related effects of cocaine.

Methods: Twelve non-treatment seeking individuals (4 women) with CUD participated in this within-subject human laboratory study. Participants were maintained on 0, 2.5, 5 and 10 mg oral zolmitripan/day in random order. After at least 3 days of maintenance on each target dose, participants completed experimental sessions in which the reinforcing, subjective, physiological and cognitive-behavioral effects of 0, 10 and 30 mg/70 kg of intravenous cocaine were determined. Data were analyzed using repeated mixed-effects models, allowing for use of the non-completing subjects’ partial data sets.

Results: Cocaine functioned as a reinforcer and produced prototypic dose-related subjective and physiological effects (e.g., statistically significant main effects of Cocaine Dose only in mixed-effects models on ratings of “Stimulated” and heart rate; p values < 0.05). Zolmitriptan potentiated temperature decreasing effects of cocaine (i.e., statistically significant main effects of Cocaine Dose and Zolmitriptan Dose in mixed-effects model on oral temperature; p values < 0.05) and produced differential effects on working memory, with performance impaired by lower doses and enhanced by the highest dose after administration of 0 mg/70 kg cocaine (i.e., statistically significant interaction of Cocaine Dose and Zolmitriptan Dose in mixed-effects model on proportion correct in the 2-back condition of the n-back; p value < 0.05). Zolmitriptan did not alter any other effects of cocaine.

Conclusions: Data indicate that activating the 5-HT1b system through zolmitriptan maintenance produces limited changes in the pharmacodynamic effects of cocaine in humans, contrasting preclinical findings. Failing to translate from preclinical to clinical models could be due to methodological or species differences, suggesting the field needs to better address this translational gap.

Supported by: R01 DA 052203, UL1 TR 001998.

Keywords: cocaine, Serotonin, Self-Administration

Disclosure: Anebulo, Consultant, Self, Myosin, Consultant, Self

P336. Sex differences and impact of oral contraceptive use on effective (directional) connectivity of face processing and emotion regulation networks in individuals with alcohol use disorder

Christopher Hammond, Liangsuo Ma, James M. Bjork, F. Gerard Moeller, Albert J. Arias

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

Background: Women and men with alcohol use disorders (AUD) vary in their subjective responses to alcohol and exhibit different withdrawal presentations, patterns of co-occurring psychiatric symptoms/disorders, and risk and prognostic factors, suggesting sex-specific phenotypes (Erol and Karpyak, 2015). Neurocircuitry alterations underlying these differences are poorly understood. Divergent effects of sex hormones on brain function and structure and alcohol consumption in men and women may play a key role in sex differences in AUD (Pletzer et al., 2019; Maddern et al., 2024). In the current study, we investigated sex differences in effective (directional) connectivity (EC) during emotional face processing in individuals with AUD and healthy controls, and assessed the relative influence of hormonal oral contraceptive (OC) use on ECs in women.

Methods: We utilized functional MRI data from the Human Connectome Project database obtained during performance of an emotional face perception task in 70 participants meeting DSM criteria for alcohol abuse or dependence (combined as AUD group: mean age = 29.0 [3.5] years; 33 women) and 70 matched control (CON) participants (mean age = 28.8 [3.5] years; 38 women). During the task, participants were visually presented with either pictures of fearful, angry, and neutral faces or shapes and asked to match them to a face picture or shape at the bottom of the screen (Hariri et al., 2001). Focusing on an a priori selected network (Ma et al., 2020) including ventromedial prefrontal cortex (VMPFC), bilateral ventrolateral prefrontal cortex (VLPFC), amygdala (AMY), and fusiform gyrus (FG), and right (R) hypothalamus (HTH) nodes, we performed dynamic causal modeling (DCM) analyses to test for sex (male vs. female), diagnostic group (AUD vs. CON) and group-by-sex differences in ECs modulated by facial emotion content (faces – shapes contrast). Supplemental DCM analyses examined EC differences as a function of OC use status (current OC use [n = 32] vs. non-use/naturally cycling [n = 39]) in female participants and tested for diagnostic group-by-OC use interaction effects. Reliable EC differences were evidenced by Bayesian Posterior Probability (PP, ranging from 0 to 1) exceeding 0.95. Larger PP’s reflect stronger confidence in the corresponding EC strength or linear regression coefficient.

Results: For DCM group-level analyses, sex differences and group-by-sex interaction effects were observed in multiple node-node circuits. Highlighting thresholded EC findings (ECs ≥0.2 Hz, Bayesian-PP’s = 1): Compared with men, women showed greater ECs from the R-VLPFC to R-VLPFC, R-AMY to left (L)-VLPFC and b/l-FG, and R-FG to L-VLPFC and HTH and lower ECs from the L-VLPFC to L-VLPFC and HTH, R-VLPFC to L-VLPFC and R-AMY, R-AMY to VMPFC, L-FG to L-VLPFC and HTH, and HTH to b/l-VLPFC and R-FG. Among these ECs, L-VLPFC to L-VLPFC, R-AMY to VMPFC and b/l-FG, R-FG to L-VLPFC and HTH, L-FG to HTH, and HTH to b/l-VLPFC and R-FG showed reliable AUD vs. CON group differences and group-by-sex interaction effects (ECs ≥0.1 Hz, PP’s = 1). In the supplemental analyses conducted in female participants, multiple ECs varied in relation to OC use status, including some which showed AUD vs. CON group and sex differences (e.g., HTH to R-VLPFC and R-FG). Notably, one EC from the R-VLPFC to L-FG was lower in women than men and AUD than CON, negatively correlated with alcohol consumption and depression, and positively modulated by OC use in women. Another EC from R-AMY to L-FG showed a similar pattern of associations but in the opposite direction (i.e., greater in women than men and AUD than CON, positively correlated with alcohol consumption, and negatively modulated by OC use).

Conclusions: These results suggest sex differences in EC during emotional face processing in top-down emotion regulation and bottom-up emotion perception circuits linking VLPFC, AMY, FG, and HTH regions in individuals with AUD and healthy controls. Some ECs with estrogen receptor-rich brain regions (e.g., AMY, HTH, and VLPFC; see Ruehr et al., 2025) as central nodes, also varied in relation to OC use in women, suggesting hormonal modulation of AUD relevant circuits (Gao et al., 2024). Our findings carry clinical implications for interventions targeting emotion processing deficits in women and men with AUD.

Keywords: sex differences, Facial emotion processing, fMRI Effective Connectivity, Alcohol and substance use disorders

Disclosure: Nothing to disclose.

P337. Single-cell atlas reveals transcriptomic and chromatin dynamics in human AUD amygdala

Matthew Girgenti, CheYu Lee, Delaney McRiley, Catharine Duman, Hang Zhou, Joel Gelernter, Alicia Che, John Krystal, Jing Zhang

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

Background: Regulation of gene expression is a highly coordinated process in both the healthy and pathological brain with unique patterns across a multitude of cell types. Here we present a multi-omic single nucleus study of ~175,000 nuclei in alcohol use disorder (AUD), profiling cell type specific gene expression and chromatin accessibility in the human central amygdala. We identify all major CNS cell types and neuronal subtypes and find inhibitory neurons are particularly affected by AUD. We find high numbers of differentially expressed genes (DEGs) including GABRA2, GRM8, and NCAM1 and show significant enrichment for AUD risk genes within these DEGs. We identified 51,431 cell type-specific, disease associated candidate cis-regulatory elements including an interneuron-associated set of chromatin loops at the AUD risk gene CALN1. Transcription factor footprinting identified Kruppel-like factors upstream of AUD GWAS genes and DEGs. Finally, we also perform cell type-specific fine mapping for AUD GWAS to prioritize variants within functional genomic elements.

Methods: We isolated ~500K nuclei from human postmortem central amygdala from cases and controls for single nucleus RNA sequencing across three diagnostic cohorts: AUD, MDD (Psychiatric control), and neurotypical controls to identify neuronal and non-neuronal cell type clusters and cell type-specific gene expression changes. We then performed paired sequencing of the same samples for ATAC-sequencing, to measure differential chromatin accessibility. We identified open chromatin regions harboring risk alleles for PTSD and through integration of snRNA and snATAC we identified disease specific cis-regulatory elements. We used the most current and largest AUD and MDD GWAS to fine map risk variants within specific cell types.

Results: We identified 14 distinct cell type clusters including neuronal and non-neuronal cell types. We identified 1805 FDR significant differentially expressed genes across many cell types and confirmed expression changes of several genes implicated by problematic alcohol use (PAU) GWAS. We found AUD specific cis-regulatory elements for several genes including CALN1. We constructed a gene expression regulatory landscape of AUD by integrating RNA and ATAC modalities to define cis-regulatory elements (CREs) and link them to specific genes. We also assembled AUD-specific TF regulatory networks for KLF7, KLF6, and KLF16 and link these to specific DEGs. This resource enabled us to fine-map all the PAU GWAS genes into specific cell types.

Conclusions: We discovered selective changes in calcium and glutamatergic signaling that were most pronounced in inhibitory neurons. In addition, we identify vulnerability of proenkephalin (PENK) interneurons in AUD and global shifts in the transcriptome reflecting changes in TF binding. Overall, this work enabled characterization of gene pathways and their dynamics in diverse amygdala cell types and prediction of cis-regulatory logic and associated factors underpinning the molecular etiology of AUD.

Keywords: alcohol use disorder, single-cell genomics, fine mapping, transcription factors

Disclosure: Nothing to disclose.

P338. Molecular impact of psilocybin in opioid use disorder: the role of PBX3 in synaptic plasticity

Ming-Fen Ho, Cheng Zhang, Paul Croarkin, Hu Li

Mayo Clinic, Rochester, Minnesota, United States

Background: Opioid use disorder (OUD) affects over 40 million individuals globally and imposes a substantial social and economic burden. Pharmacological treatment remains the primary therapeutic approach, with the U.S. FDA approving three medications: methadone, buprenorphine, and naltrexone. Despite these interventions, relapse rates remain alarmingly high—exceeding 90% within one year. Emerging work suggests that psilocybin may have promise as a pharmacological intervention for OUD. Patient-derived model systems provide key opportunities to bridge critical knowledge gaps in OUD pathophysiology and to better understand the mechanisms of emerging therapeutic agents, such as psilocybin.

Methods: We collected peripheral blood samples from 35 individuals with OUD and 59 demographically matched unaffected controls for bulk RNA sequencing. To investigate the cellular and molecular effects of heroin and psilocybin, we performed single-nucleus RNA sequencing (snRNA-seq) analysis on iPSC-derived forebrain organoids exposed to these drugs at clinically relevant concentrations. Neuronal activity in iPSC-derived forebrain neurons was measured using multielectrode array (MEA) recordings. Additionally, we carried out functional genomics studies to explore underlying molecular mechanisms, including RNA-seq on iPSC-derived forebrain neurons treated with heroin, psilocybin, or vehicle.

Results: Among the 35 OUD patients, two-thirds were male, and 97.1% reported opioids as their primary drug of choice, with heroin being the most commonly used. Most participants had used multiple controlled substances—such as alcohol, cannabis, and hallucinogens—before the age of 18, with first heroin use typically occurring around age 28. Given the growing interest in psychedelics as potential therapeutics for OUD, we examined the transcriptional responses to heroin and psilocybin in iPSC-derived forebrain organoids using snRNA-seq.

We generated iPSC-derived forebrain organoids from three OUD patients and exposed them to heroin or psilocybin at clinically relevant concentrations for one week. Following quality control, we analyzed 20815 nuclei (6300 from vehicle-treated organoids, 7302 from heroin-treated, and 7213 from psilocybin-treated organoids). We identified 16 transcriptionally distinct clusters encompassing all treatment conditions. These clusters were annotated using established markers for major brain cell types.

Differential gene expression and pathway enrichment analyses revealed that both heroin and psilocybin affected pathways involved in synaptic plasticity, neurogenesis, and axon guidance. However, the specific sets of differentially expressed genes were largely distinct between the two drugs. Notably, PBX3 expression was upregulated by psilocybin in multiple clusters, including glutamatergic and GABAergic neurons, but this effect was absent in heroin-treated cells.

To validate these findings, we analyzed iPSC-derived forebrain neurons exposed to the same treatments. Both heroin and psilocybin reduced neuronal activity (firing rate), synchrony (connectivity) and oscillation (burst of action potentials)—as measured by MEA. However, only psilocybin increased PBX3 expression in these neurons, as confirmed by RNA-seq. These findings suggest that while both drugs impair neuronal function and activity, they likely do so through distinct mechanisms. Given these observations, we investigated whether PBX3 expression could also be detected in peripheral blood samples. Bulk RNA-seq of peripheral blood from 35 OUD patients and 59 controls revealed significantly lower PBX3 expression in OUD patients, as compared to the unaffected control group.

Conclusions: In conclusion, we identified PBX3 as a gene potentially involved in the pathophysiology of OUD and modulated by psilocybin. This work highlights the utility of iPSC-based models in uncovering novel molecular mechanisms and supports further investigation into PBX3 as a target for psychedelic-based therapies in OUD.

Keywords: Opioid addiction, molecular mechanisms, iPSC-derived neurons, Psilocybin

Disclosure: Nothing to disclose.

P339. Development of novel negative allosteric modulators of CRFBP-CRF2 for substance use disorders

Kokila Shankar, Dhanya Panickar, Ranajit Das, Nazzareno Cannella, Carolina Haass-Koffler, Nicholas Cosford, Douglas Sheffler

Sanford Burnham Prebys Medical Discovery Institute, san diego, California, United States

Background: Substance use disorders (SUDs), including alcohol use disorder (AUD), remain a prevalent public health burden. The stress system, mediated by corticotropin-releasing factor (CRF), plays a critical role in the development of SUDs/AUD. In the brain, CRF binds to both the CRF1 and CRF2 receptors with high affinity. While CRF1 antagonism in preclinical studies has been shown to decrease anxiogenic and withdrawal-like behaviors, CRF1 antagonists have not been successfully translated to the clinical setting. Conversely, CRF2 agonism has been shown to decrease ethanol consumption and ameliorate withdrawal-like behaviors in rodents, making CRF2 a promising target for therapeutic development. CRF binding protein (CRFBP) is a 37 kD glycoprotein that interacts with CRF2 to modulate drug-taking/seeking behaviors as well as potentiate NMDA-mediated excitatory postsynaptic currents (EPSCs) in dopaminergic ventral tegmental area neurons; thus, CRFBP may play a role in SUD/AUD development. CRFBP undergoes spontaneous proteolytic cleavage into a 27 kD fragment that binds free CRF and a 10 kD fragment that appears to positively regulate CRF2 signaling. For example, we have previously shown that this 10 kD fragment tethered to CRF2 (CRFBP-CRF2) potentiates CRF-mediated Ca2+ mobilization in the cell. As the CRFBP-CRF2 protein-protein interaction has a potential excitatory function in neurons, we hypothesize that it may be a valuable target for pharmacological modulation of SUD/AUD.

Methods: Our previous high-throughput screen of > 350,000 novel small molecule chemical probes yielded two structurally distinct compounds, MLS-0046818 and MLS 0219419, that act as negative allosteric modulators (NAMs) of the CRFBP-CRF2 complex (Haass-Koffler et al., 2022). Neither had activity against CRF1 and had no effect on CRF2 function in the absence of CRFBP (10kD). Both compounds exhibited brain exposure at 24 h post-dose and blocked NDMA-mediated EPSCs in the VTA. Our current lead optimization efforts have focused on elucidating the structure-activity relationship (SAR) of these compounds through synthesis of additional novel analogs. Analogs underwent in vitro potency screening of these compounds in a chimeric cell-based assay in which CRFBP (10 kD) is tethered to CRF2, and agonism of the complex with CRF can be detected using a fluorescence-based calcium sensor. Compounds with improved potency over the original leads were also administered to wild-type C57BL/6 mice (n = 3, mixed gender) to evaluate safety and availability within the plasma and brain.

Results: To date, we have screened 50+ analogs of MLS 0219419 and identified 10 compounds with promising in vitro NAM activity (<1 µM). Potency improved following multiple modifications at different points along the scaffold. 11 of the screened analogs were chosen for snapshot PK analysis, of which approximately half showed promising brain exposures (> 2x IC50).

Conclusions: We present the first small molecule chemical probes for use as research tools to evaluate the role of CRFBP in the brain, with the consequent potential to pharmacologically treat AUD. Using these compounds, we have developed additional chemical probes and are currently investigating the role of negative allosteric modulation of CRFBP-CRF2 in vitro, ex vivo, and in vivo in order to optimize orally active compounds suitable for proof-of-concept studies for SUD/AUD treatment.

Keywords: Corticotropin-releasing factor (CRF), drug discovery/development, Alcohol Use Disorder - Treatment

Disclosure: Nothing to disclose.

P340. Glucagon-like peptide-1 receptor agonists (GLP-1RAs) in substance use disorders: a systematic review of Clinicaltrials.gov

Shruti Patil, Nandini Jha, Manish Jha

The University of Texas Southwestern Medical Center, Dallas, Texas, United States

Background: The urgent need for developing novel pharmacotherapies is underscored by the rapid increases in drug overdose deaths and the high prevalence rates of substance use disorders (SUDs). For several SUDs, such as methamphetamine, cocaine, and cannabis use disorders, there are no current FDA approved treatments. With increasing recognition of the role that glucagon-like peptide 1 (GLP-1) receptors play in brain’s reward system, including by modulating the function of ventral tegmental area and nucleus accumbens, agonists of GLP-1 receptors (GLP-1RAs) have gained recent attention as novel treatments for SUDs. Use of GLP-1RAs in clinical practice is increasing rapidly with newer drugs, such as tirzepatide, a dual GLP-1/glucose-dependent insulinotropic peptide receptor agonist, demonstrating clear superiority over older GLP-1RAs in their metabolic effects. Evidence from electronic health records data suggest improved SUD-related outcomes in patients with diabetes treated with GLP-1RAs as compared to other antihyperglycemic drugs. These studies reveal lower rates of opioid overdose, alcohol intoxication and hospitalizations, and reduced risk for developing any substance use disorder. While there have been recent reviews of published literature on the effects of GLP-1RAs in mental health disorders, there is still a need to survey the landscape of ongoing clinical trials evaluating GLP-1RAs for specific SUDs (such as alcohol, nicotine, stimulants, opioids, and cannabis use disorders) to inform the therapeutic pipeline of novel treatments for SUDs.

Methods: This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta Analysis (PRISMA) standards. A comprehensive search for clinical trials involving GLP-1RAs for substance use disorder was conducted July 3rd using the ClinicalTrials.Gov database. This is registered as open registration on OSF as this is an active research area with potential of updating our research results in a future publication.

ClinicalTrials.Gov was searched to identify trials using GLP-1RAs to treat SUDs from inception until July 2nd, 2025 using root search terms: “GLP-1”, “Semaglutide”, “Exenatide”, “Tirzepatide”, “Liraglutide”, “Dulaglutide”, and “Lisexenatide”. Each root search term was combined with substance search terms including: “nicotine”, “smoking”, “alcohol”, “cannabis”, “marijuana”, “methamphetamines”, “amphetamines”, “cocaine”, “crack”, “stimulant”, “opioid”, “heroin”, and “fentanyl.” Two authors conducted these searches independently in July 2025 and the senior author then reviewed all the search terms.

Trials were first screened by title and study overview, then inclusion criteria. Trials were included if they (a) had endpoints measuring substance use severity reduction (i.e., proportion of negative UDS, Timeline Follow-Back, self-reported craving), and (b) utilized a GLP-1 RA for the treatment of SUDs (opioid use disorder, alcohol use disorder, methamphetamine use disorder, cocaine use disorder, nicotine use disorder, and cannabis use disorder) with or without comorbid conditions (i.e., obesity, type 2 diabetes). Trials were excluded if they excluded individuals with SUDs.

Results: A systematic review of ClinicalTrials.Gov identified 192 trials via search term and title criteria, with 33 trials satisfying inclusion criteria. Included studies were most numerous for alcohol use disorder (N = 15) followed by nicotine use disorder (N = 9), cocaine use disorder (N = 5), opioid use disorder (N = 4), and methamphetamine use disorder (N = 1). There were no clinical trials of GLP-1RAs for cannabis use disorder. Semaglutide was the most studied GLP-1RA (N = 13), followed by exenatide (N = 7), tirzepatide (N = 6), liraglutide (N = 2), dulaglutide (N = 1), and pemvidutide (N = 1).

The most common reasons for exclusion from this review included use of GLP-1RAs for a medical condition other than substance use disorder (e.g., osteoarthritis, diabetes, obesity, cardiovascular disease) and absence of a SUD and/or GLP-1RA treatment group. Thresholds of body mass index for study inclusion were highly variable, with some studies excluding underweight or normal weight individuals. Duration of study was also highly variable with durations ranging from 5 days to 32 weeks. Diagnosis of SUD were based on structured validated diagnostic measures. Primary outcomes focused on reduction of substance use, however, studies varied in how they measured this reduction with some studies focusing on clinical measures (such as the Fagerstrom test for nicotine dependence, cue craving visual analog score, Timeline Followback) and others including biologic data (i.e., Phosphatidyl ethanol values, urine drug screen).

Conclusions: With multiple ongoing clinical trials of GLP-1RAs, these drugs may represent a potential paradigm shift in the treatment of SUDs. However, most studies have used older GLP-1RAs, focused mainly on alcohol or nicotine use disorder, and used heterogenous outcome measures relying on both self-reported and objective measures of substance use severity. There is a paucity of trials on methamphetamine use disorder and cannabis use disorder, two SUDs with no current FDA approved pharmacotherapies. Together, these ongoing clinical trials may fill the existing gap in pharmacotherapy options for SUDs, especially for disorders that have no currently FDA approved treatments.

Keywords: Alcohol and substance use disorders, GLP-1 receptor agonist, systematic review

Disclosure: Nothing to disclose.

P341. A retrospective view of electronic health records: the effect of long-term metformin exposure among cocaine and cannabis users

Edith Hernandez, Abigail Brewer, Michael Evans, Claire Felter, Sade Spencer

University of Minnesota Medical School, Minneapolis, Minnesota, United States

Background: Drug use in the United States has steadily increased in recent decades, with a substantial climb in drug activity following the COVID-19 Pandemic of 2020. Drug-related deaths worldwide continue to rise annually, many occurring as a result of relapse in use after a period of abstinence. Although the field of addiction medicine has seen the emergence of pharmacologic therapies for some substance use disorders (SUD), we have not yet found efficacious pharmacological agents to assist individuals interested in curbing use of cocaine or cannabis. As use patterns increase with a significant increase in access to both drugs, the need for treatment escalates. Our lab has previously shown that metformin (MET), a Type II Diabetes medication, can reduce cocaine cue-induced reinstatement (RST) in rats when centrally injected into the nucleus accumbens core. As the first-line treatment for Type II Diabetes, we aimed to use electronic health records to understand the effects of MET when administered long-term on reported cocaine-motivated behavior. Additionally, we investigated effects of MET on curbing cannabis-associated behaviors as a method of understanding how MET may generalize to other substance use disorders.

Methods: This retrospective review was done using patient clinical files from the University of Minnesota’s Clinical and Translational Science Institute (CTSI). CTSI has electronic health records of more than 4.5 million patients who are seen at 8 different hospitals and more than 40 clinics through Fairview Health Services. The source population is individuals who were seen at University of Minnesota Medical Center (UMMC) Fairview Health facility in the state of Minnesota between January 1st, 2011 to December 31st, 2023. We chose data from UMMC Fairview because it is the largest Fairview hospital system with the most diverse patient representation. We collected both quantitative and qualitative data from patient records. Data collected from provider visits, including patient demographics, zip code, patient medical history, diagnoses, medications, obstetric episodes, and provider notes were used. Within the exposed group, there is a subgroup per drug class: those who are prescribed MET and report cocaine or cannabis use). Within the comparison group, there will be two subcategories: those who are prescribed metformin and report no cocaine/cannabis use and those who are not prescribed MET and do report regular drug use.

Results: Eligibility criteria included individuals over the age of 18 who have been seen at a Fairview health facility. Individuals were randomly selected from the patient pool that meet eligibility criteria and divided into control and exposed groups based on inclusion criteria. Preliminary data showed that 36.5% of participants (N = 60) who were receiving metformin reported having an uncomplicated dependence to cocaine in comparison to 52.2% (N = 59) who were not receiving metformin. Additionally, 21.8% of participants (N = 36) reported being in remission from cocaine use while using metformin, while only 13.3% of participants (N = 15) reported entering remission without metformin exposure. Similarly, 11.5% of cannabis users (N = 26) reported entering remission while under metformin exposure, while only 5.7% (N = 8) without metformin reported entering remission.

Conclusions: This study aimed to find evidence showing that long-term metformin exposure may be able to affect ongoing cocaine- or cannabis-motivated behavior. Preliminary data review of retroactive electronic health records from the University of Minneosta health system showed that there may be a suggested impact of metformin in assisting patients entering remission from both cocaine and cannabis use. Further data evaluation will be needed to yield concrete correlations between variables, however preliminary effects show a trend in support of the therapeutic value of metformin as a treatment for substance use disorders, specifically in individuals that use cocaine and cannabis.

Keywords: metformin, cocaine, cannabis, electronic health record (EHR)

Disclosure: Nothing to disclose.

P342. Sex-stratified brain microRNA study of problematic alcohol use

Najeah Okashah, Jiaqi Liu, Selina Vattahil, Thomas Wingo, Aliza Wingo

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

Background: Problematic alcohol use (PAU), routine alcohol use above recommended healthy levels, is a major contributor to global morbidity and mortality, yet the molecular mechanisms linking alcohol exposure to long-term brain changes remain poorly understood. MicroRNAs (miRNAs), small non-coding RNAs regulating post-transcriptional gene expression, are potential mediators of alcohol’s neurobiological effects. Thus far, small sample sizes have limited prior studies of brain miRNAs in alcohol use disorder due to the difficulties of obtaining postmortem brain samples. In this study, we aimed to identify miRNAs whose expression levels in the human brain are altered in PAU and to determine whether these expression patterns are sex-specific.

Methods: MicroRNAs were profiled from the dorsolateral prefrontal cortex (DLPFC) of brain donors recruited by the Religious Orders Study and the Rush Memory and Aging Project (ROSMAP) and annotated using the MirGeneDB database. After quality control, microRNA profiles from 458 donors encompassing 528 miRNAs were included for downstream analysis. PAU was defined as the peak Lifetime Daily Alcohol Intake (LDAI) of ≥1 drink/day for females, ≥2 drinks/day for males. Controls were defined as participants who reported < 1 drink/month. Voom-limma was used for differential expression, adjusting for cognitive diagnosis, postmortem interval (PMI), RNA integrity number (RIN), and batch. Analyses were performed jointly (with sex as a covariate) and in a sex stratified manner. Significant miRNAs (FDR < 0.1) in the joint and sex-stratified analyses were further assessed for sex-by-miRNA interactions (p < 0.05) to determine if there are sex differences in brain miRNAs associated with PAU. Predicted targets will be extracted from TarBase, which contains experimentally supported miRNA targets, and TargetScan 8.0, which predicts biological targets of miRNAs. We will select predicted targets from TargetScan 8.0 with cumulative weighted context scores (CWCS), an estimate of the total predicted action of the miRNA on its target genes, with a value of less than −0.4 for further analysis to minimize potential false positives among the predicted targets.

Results: miR-5683 was significantly downregulated in males with PAU (logFC = –0.80, FDR = 0.06). No significant miRNA changes were observed in females. Testing of the sex-by-miRNA interaction for miR-5683 was significant (interaction term p = 0.0067), indicating sex-dependent expression in PAU. Tarbase contained 62 protein-coding genes, and TargetScan predicted 77 protein-coding genes as downstream targets of miR-5683. The predicted targets were enriched for the Synaptic Gene Ontology (SynGO) term “Integral Component Of Post-synaptic Membrane” (GO:0099055, FDR p = 0.044), suggesting that the miRNA targets play a role in post-synaptic structures. Of the target proteins, 7 had nominally significant (p < 0.05) differential expression: SUGT1, UBE2D1, TTC27, LRRC41, RBM39, FAM169A, and ANKRD17.

Conclusions: This is the largest brain miRNA study of PAU to date and the first to apply sex-stratified analyses in a deeply phenotyped aging cohort. Our findings robustly demonstrate a sex-specific decrease in miR-5683 expression in PAU and suggest potential miRNA-mediated changes involved in the effects of alcohol on the brain. These results highlight the utility of integrating miRNA and behavioral data in human postmortem studies and underscore the importance of considering sex in genetic studies of alcohol use.

Keywords: Alcohol and substance use disorders, MicroRNA, Sex-specific effects, Postmortem human brain study

Disclosure: Nothing to disclose.

P343. Impact of stress-induced CCR5 dysregulation on behavioral responses to opioids

Carole Morel, Hsiao-Yun Lin, Lyonna Parise, Scott Russo, Jun Wang

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

Background: Over the past decades, opioid use disorder (OUD) has emerged as a growing and challenging epidemic, underscoring the urgent need for new treatment strategies and a deeper understanding of its etiology and vulnerability factors. Traumatic life experiences and psychosocial stress are well-established drivers of drug consumption, seeking, and relapse. In addition to their detrimental behavioral and neurophysiological effects, both chronic stress and opioid exposure engage neuroinflammatory processes. Neuroimmune interactions may represent a critical nexus through which stress potentiates opioid action in the brain and drives behavioral responses to opioids. However, the mechanisms by which stress and drug exposure influence neuroimmune interactions to promote opioid-related behaviors remain poorly defined. Here, we aim to bridge this knowledge gap by investigating the role of chemokine receptor 5 (CCR5) in the impact of psychosocial stress on microglia activity, cortical dynamics, and subsequent behavioral response to opioids.

Methods: We use a chronic social defeat stress paradigm, which induces long-lasting and complex behavioral outcomes, to capture motivational, cognitive, and affective deficits in mice. Following social trauma, we conduct a longitudinal assessment of approach–avoidance behaviors in both female and male mice and perform opioid-conditioned place preference (CPP) and self-administration assays to evaluate behavioral responses to opioids (e.g., morphine and oxycodone). Using cell-specific probing techniques, in vitro electrophysiological recordings, behavioral assays, and molecular approaches, we investigate the neuroimmune mechanisms underlying the detrimental interplay between social stress and opioid exposure.

Results: We first observed that, in the absence of stress, mice require a high dose of morphine to develop conditioned place preference (CPP) and do not establish CPP at lower doses. Conversely, following chronic social stress, mice establish CPP even at a low dose of morphine, confirming that stress exposure enhances morphine sensitivity (two-way ANOVA, F(1,46) = 7.819, post hoc test p < 0.05, n = 9–16 mice). When interrogating the contribution of the neuroimmune response, we found that social trauma increases the expression of C-C chemokine receptor 5 (CCR5) within the prefrontal cortex (PFC; p = 0.0018). This upregulation was specific to microglia and associated with their overactivation. Our preliminary data also show that mice exposed to a high dose of morphine exhibit increased CCR5 expression, specifically in PFC microglia, a finding we recapitulated with oxycodone exposure. Moreover, selective blockade of CCR5 signaling in the PFC successfully prevented stress-induced neurophysiological changes and behavioral maladaptation (two-way ANOVA interaction: F(3,800) = 31.79; F(1,50) = 21.89; p < 0.0001; n = 11–15 cells). Lastly, we established that CCR5 downregulation within the PFC prevents opioid-induced CPP (two-way ANOVA, F = 7.561, post hoc t-test p < 0.05, n = 6–9 mice).

Conclusions: Clinical observations and experimental studies indicate that psychosocial stress plays a crucial role in the progression of OUD. Together, these findings suggest that PFC microglia and the CCR5 neuroimmune system act as coupling mechanisms linking stress to opioid responses.

Keywords: Chronic social stress, Opioids, Neuroimmune, Microglia

Disclosure: Nothing to disclose.

P344. A pre-clinical model of e-cigarette use reveals higher nicotine vapor self-administration and seeking in adolescent versus adult rats

Laura ODell, Tiffany Gonzalez-Gutierrez, Ian Mendez

The University of Texas at El Paso School of Pharmacy, El Paso, Texas, United States

Background: The widespread use of electronic cigarettes (i.e., e-cigarettes) among adolescents is a public health concern. Nicotine exposure during adolescence, a critical period of brain development, has detrimental effects on health and behavior. Pre-clinical studies over the last 10 years have begun to identify potential health risks of vaping. Specifically, studies have shown that early exposure to nicotine vapor can disrupt brain function in areas that control reward processing, motivation, and executive functions. Valid animal models are essential for investigating this issue, and emerging work has begun to explore the factors that promote reliable nicotine vapor self-administration (NVSA) in adolescent rodents. Prior studies from our laboratory suggest that age at the time of exposure may impact NVSA in rodent models of e-cigarette use. Thus, the goal of the present study is to build on our previous work by assessing self-administration of vehicle vapor alone or in combination with a range of doses of nicotine (6, 12, and 24 mg/ml) in adolescent and adult rats.

Methods: Adult (n = 21, no earlier than PND 60 at start of experiment) and adolescent (n = 15, PND 30 at start of experiment) rats were passively exposed to ambient air, vehicle vapor (50/50 propylene glycol/vegetable glycerin) plumes, or 24 mg/mL nicotine vapor plumes for 6 consecutive days. Following passive exposure, rats were allowed to self-administer vehicle or nicotine vapor on a fixed ratio 1 schedule with a 2-min timeout period after delivery, in 90-minute daily sessions. Rats self-administered vehicle vapor or nicotine vapor for a total of 24 days, with varying nicotine concentrations delivered in the following order for the nicotine group: 24 mg/mL nicotine for 6 days, 6 mg/mL nicotine for 12 days, 12 mg/mL nicotine for 6 days. All rats then underwent 17 days of extinction training before being tested for the reinstatement of vapor-seeking behavior across 2 days, following the presentation of a single vapor plume (vehicle plume or 24 mg/mL nicotine plume delivered to vehicle groups and nicotine treatment groups, respectively). Total nicotine vapor plume deliveries and active port nosepokes were compared between adolescent and adult rats using mixed-model ANOVA, followed by t-test post-hoc analyses.

Results: Results reveal that while both adult and adolescent rats self-administered vehicle or nicotine vapor plumes across all NVSA days, only adolescent rats delivered significantly more nicotine vapor plumes than vehicle vapor plumes. Significant main effects of age (F(1,53) = 25.94, p < 0.001) and treatment (F(2,53) = 9.44, p < 0.001), as well as an interaction of age and treatment (F(2,53) = 6.82, p < 0.01) were observed across all 24 NVSA days. Specifically, post-hoc analysis reveal that adolescent rats had significantly more deliveries of vehicle (p < 0.05) and nicotine vapor (p < 0.001), when compared to adults. Furthermore, analysis of nicotine seeking behavior across 17 days of extinction testing revealed a main effect of age (F(1,47) = 17.86, p < 0.001), such that relative to adults, adolescents had more nosepokes in the active port on days 2, 4, 8, 9, 10, and 12–15 of extinction testing (ps < 0.05). Lastly, tests of drug-primed reinstatement immediately after extinction showed that adolescents had significantly more nicotine vapor-seeking behavior than adults, such that adolescents displayed more nosepokes in the active port than adults during each of the 2 testing days (ps < 0.01).

Conclusions: The present study describes a preclinical model of nicotine vapor self-administration that results in robust and persistent nicotine vapor-seeking and delivery in adolescents. Findings suggest that the reinforcing effects of nicotine vapor may differ based on age of exposure, with adolescent rats displaying higher levels of NVSA and nicotine seeking during extinction and drug-induced reinstatement. These findings are in line with clinical studies reporting that the highest rates of e-cigarette use are seen in the adolescent and young adult population. Higher reinstatement of nicotine vapor-seeking behavior was also observed in adolescents relative to adults, suggesting that nicotine vapor intake during adolescence may result in higher rates of nicotine craving and seeking. The development of a valid and reliable pre-clinical model of e-cigarette use will be necessary for identifying the neurobiological, behavioral, social, and economic variables driving e-cigarette use early in life and throughout the lifespan.

Keywords: Nicotine, E-Cigarette, Self-Administration, Adolescence, Rats

Disclosure: Nothing to disclose.

P345. Naltrexone reduces choice for heroin in female rats and alters the prefrontal cortex transcriptome

Jamie Peters, Victoria Chang, Nathaniel Robinson, Jeremy Day

The University of Alabama At Birmingham, Birmingham, Alabama, United States

Background: There are several approved therapeutics for the clinical treatment of opioid use disorder (OUD). Naltrexone (NTX), an opioid receptor antagonist, is used to diminish opioid craving and relapse and is often prescribed as an extended-release formulation, in order to maintain continuous blockade of opioid receptors. Compared to other OUD treatments such as buprenorphine or methadone, NTX has some advantages in that it has not been associated with the development of tolerance or dependence and has no known abuse liability. Over time, OUD patients on NTX often report reduced opioid craving and have reduced opioid relapse. Additionally, it has been successfully utilized as a treatment in conjunction with contingency management. The goal of the present study was to test the therapeutic efficacy of NTX in a preclinical model of heroin choice, where rats must choose between a limited number of intravenous heroin infusions versus food pellets. In addition, we examined the prefrontal cortex (PFC) transcriptome as a potential signature of NTX therapy in this model.

Methods: Male and female Wistar rats (aged P55–60 at the start of the experiment) were surgically implanted with intravenous catheters and then subsequently trained to self-administer intravenous heroin on one lever and food pellets on an alternate lever in daily behavioral training sessions (150 min, 10 sessions), progressing from a fixed ratio 1 (FR1) schedule of reinforcement to an FR3. Next, rats proceeded to the choice phase of the experiment, wherein there were a limited number of trials (14 max) in which the levers were extended. After the FR3 requirement was met on either lever, both levers would retract and the house light would turn off for a 10 min time out in between trials. After 3 baseline sessions (150 min/session), rats began 7 consecutive days of the treatment phase. NTX (3mg/kg injection) or saline was administered 10 min prior to the start of each choice session, and again mid-session (80 min later), during the treatment phase. Heroin choice was calculated as the number of trials in which the rat chose heroin over the total number of trials completed for each session. The day after completing the treatment phase, rats underwent a cued reinstatement test (60 min) wherein the discrete cues (tone or light) associated with each reward (heroin vs. food) were available on an FR3 schedule of reinforcement, but neither reward was available. A second cued reinstatement test was conducted 3 days later to assess the persistence of NTX treatment. Finally, the rats were placed back onto heroin/food self-administration for reacquisition, followed by another series of choice sessions to assess any lasting effects of NTX on heroin reacquisition and choice. One day after the last choice session, rats were euthanized and brains were extracted and flash frozen on isopentane for subsequent RNA extraction and bulk RNA sequencing of the ventromedial PFC.

Results: NTX treatment significantly reduced heroin choice over the course of the treatment phase, with a significant difference emerging on days 5 through 7 of treatment compared to saline animals [2-way RM ANOVA main effect of session F(7, 287) = 17.75, p < 0.0001, and session x treatment interaction F(7, 287) = 4.864, p < 0.0001, Fisher’s LSD post-hoc comparisons on days 5 through 7 p < 0.05]. We analyzed the behavioral data disaggregated by sex and discovered that this effect was primarily driven by females [2-way RM ANOVA main effect of session F(7, 140) = 12.68, p < 0.0001, and session x treatment interaction F(7, 140) = 4.510, p = 0.0001, Fisher’s LSD post-hoc comparisons on days 5 through 7 p < 0.05] and not males. NTX treatment during choice also had a lasting therapeutic effect on heroin cued reinstatement, or relapse, after 24 hours of washout [2-way RM ANOVA main effect of lever F(1, 41) = 5.468, p = 0.0243, main effect of treatment F(1,41) = 26.52, p < 0.0001, and lever x treatment interaction F(1, 41) = 28.48, p < 0.0001, Sidak’s post-hoc comparisons between NTX and saline groups p < 0.0001] and after 96 hours of washout [2-way RM ANOVA main effect of lever F(1, 41) = 22.77, p < 0.0001, main effect of treatment F(1,41) = 26.38, p < 0.0001, and lever x treatment interaction F(1, 41) = 16.19, p = 0.0002, Sidak’s post-hoc comparisons between NTX and saline groups p < 0.0001]. These lasting effects of NTX on heroin cued relapse were present in both sexes.

Differentially expressed genes (DEGs) between NTX and saline groups were only detected in the PFC of females and not males. Examples include muscarinic receptor 4 (Chrm4), protein tyrosine phosphatase receptor type B (Ptprb), and adenlyate cyclase 10 (Adcy10). Furthermore, NTX-regulated gene expression correlated with behavioral metrics of heroin choice and relapse in females.

Conclusions: Repeated treatment with NTX effectively reduced choice for heroin over a natural reward, predominantly in females, and reduced relapse rates in both sexes after washout for at least 96 hours. NTX treatment resulted in lasting changes in PFC gene expression in females, but not males, and these gene expression changes correlated with behavioral metrics of OUD in this model. Thus, the PFC appears to be an important brain site whereby NTX elicits its therapeutic action in females with a history of chronic heroin use. Whereas it is an effective anti-relapse therapeutic in males as well, NTX may be less effective in altering heroin choice in males and may exert its therapeutic actions through other brain sites.

Keywords: heroin choice, naltrexone, prefrontal cortex, RNAseq

Disclosure: Delix Therapeutics, Consultant, Self

P346. Prenatal alcohol and cannabinoid co-exposure rewires striatal transcriptomic networks underlying alcohol-seeking and hyperactivity

Siara Rouzer

Texas A and M University College of Medicine, Bryan, Texas, United States

Background: Alcohol and cannabinoids are among the most frequently consumed psychoactive substances during pregnancy, and polysubstance exposures are common in real-world contexts yet remain minimally investigated. The dorsomedial striatum (DMS) is a corticostriatal hub regulating reward-seeking and motor activity, which are both disrupted by developmental drug exposures. While single-substance models have linked prenatal alcohol or cannabinoid exposure to striatal dysfunction, it is unknown whether co-exposures impose distinct alterations to DMS transcriptomic networks, and whether these changes translate to altered alcohol-seeking or hyperactivity phenotypes. Weighted Gene Co-expression Network Analysis (WGCNA) offers a systems-level approach to address these questions, with sex as a critical biological variable.

Methods: Pregnant C57BL/6J dams were assigned to one of four exposure conditions between gestational days 12–15: drug-free control (CON), ethanol vapor (ALC), the cannabinoid agonist CP-55,940 (0.75 mg/kg i.p.; CB), or ethanol vapor + CP-55,940 (ALC + CB). Adult offspring were tested for locomotor activity in the open field at 3 months old, and five months later, assessed for alcohol-seeking behaviors using operant self-administration assays. At ~1 year of age, DMS tissues were microdissected for RNA-sequencing (n = 5–8/sex/exposure). RNA-seq data were normalized (DESeq2 variance-stabilizing transformation) and used for WGCNA network construction. Outlier samples were removed based on Mahalanobis distance, and modules were defined from topological overlap. Differences in module expression were analyzed in a 3-way ANOVA (Sex × ALC × CB), and module eigengenes were correlated with within-subject behaviors. Fisher’s Z-tests assessed between-group differences, and significance was set at p < 0.05.

Results: Independent of sex, co-exposed offspring demonstrated greater locomotor activity and velocity in the open field compared to single-drug exposed offspring. In middle age, co-exposed males lever-pressed for and consumed more alcohol than all other groups under a progressive ratio paradigm, while all female offspring with prenatal drug exposure showed this increase in willingness to work for and consume alcohol.

Transcriptomic analyses revealed that prenatal exposures altered both baseline expression of DMS networks and their associations with prior behaviors. A 3-way ANOVA showed that alcohol exposure broadly upregulated the brown module (synaptic adhesion / neurodevelopment), downregulated the green-yellow module (extracellular matrix / nucleotide metabolism), and reduced the tan module (cognition / synapse assembly), with this reduction blocked by co-exposure.

Correlational analyses revealed that prenatal exposures did not simply weaken existing network–behavior relationships, but reprogrammed their polarity, altering whether transcriptional modules acted as protective or risk-promoting signals. For locomotor activity, co-exposure inverted normative relationships: the brown module, which in controls served as a positive correlate of motor activity, became a negative predictor in ALC + CB offspring, while the tan module, normally linked to reduced activity in controls, was reprogrammed into a positive driver of hyperactivity under ALC + CB conditions. These flips suggest that exposures during a sensitive developmental window may fundamentally change the role of striatal networks in shaping motor output.

Alcohol-seeking outcomes revealed even stronger evidence of developmental reprogramming. In males, prenatal co-exposure flipped the brown module from a negative correlate in controls to a positive driver of drinking, inverted the green-yellow module from a positive predictor in controls into a negative predictor under ALC + CB, and turned the tan module from a pro-drinking signal in controls into an anti-drinking signature under cannabinoid exposures. In females, the polarity was reversed: the brown module shifted from supporting alcohol-seeking in controls to suppressing it following prenatal alcohol exposure, while the green-yellow and tan modules - either weak or negative in controls - were flipped into strong positive drivers of drinking under ALC + CB.

Together, these findings show that prenatal alcohol and cannabinoid exposures not only reshape module expression patterns but also fundamentally reprogram their behavioral functions. Furthermore, the direction of this reprogramming is sex-dependent: males show co-exposure–driven inversions into negative predictors, whereas females display reinforcement into strong positive predictors. This highlights distinct transcriptional mechanisms through which developmental co-exposure may program vulnerability to hyperactivity and alcohol-seeking.

Conclusions: Prenatal alcohol and cannabinoid exposures induce long-lasting, sex-dependent reconfigurations of dorsomedial striatal networks. Rather than producing additive effects, co-exposure qualitatively reverses or amplifies transcriptional–behavioral relationships, with opposite polarity in males and females. These results emphasize the importance of considering both sex and polysubstance exposure history when identifying molecular targets, and suggest that targeted manipulation of module-associated pathways may offer novel strategies to reduce vulnerability in individuals with developmental teratogenic exposures.

Keywords: neurotranscriptomics, Alcohol-seeking behavior, Prenatal Drug Exposure, Sexual Dimorphism, Synthetic Cannabinoids

Disclosure: Nothing to disclose.

P347. Oxytocin modulates nucleus accumbens dopamine dynamics and motivation to seek oxycodone in oxycodone-experienced rats

Marek Schwendt, Harrison Blount, Elizabeth Colley, Lori Knackstedt

University of Florida, Gainesville, Florida, United States

Background: Problematic opioid use and opioid use disorder (OUD) represent a continued public health crisis, with only partially effective therapies available. Thus, translational preclinical models are needed to explore novel pharmacological interventions, based on relevant neurobiological mechanisms that drive OUD. Oxytocin, a neuropeptide recognized for its prosocial and stress-alleviating properties across species, has recently been shown to reduce alcohol- and opioid-seeking in preclinical models. Our previous research showed that oxytocin reduces motivation to seek the commonly abused opioid oxycodone in rats. Critically, this effect was observed in a translational model of economic drug demand when the effort required to obtain the drug was high, and only in male, but not female rats. The present study aimed to investigate the behavioral and neurobiological mechanisms that influence the ability of oxytocin to reduce oxycodone seeking in males but not females. In particular, the ability of oxytocin to modulate nucleus accumbens dopamine signal produced by oxycodone or during high-effort oxycodone seeking was evaluated.

Methods: Adult male and female Sprague-Dawley rats (n = 32) were implanted with intravenous catheters for oxycodone delivery. For fiber photometry, rats were bilaterally injected with AAV9-hSyn-GRAB_DA2m into the nucleus accumbens core, followed by implantation of a fiber optic cannula to the same site. A subgroup of rats received non-nontingent infusions of oxycodone (0.5 mg/kg/infusion) during fiber photometry recordings in the presence of oxytocin (15 or 45 IU/kg IP) or vehicle. Another subgroup of rats was trained to self-administer oxycodone (0.1 mg/kg/infusion) for 3 hours/day using a fixed ratio (FR)-1 and then a FR-3 schedule, each for 7 days. Next, rats were tested in progressive ratio (PR) schedule sessions using a within-subjects design. FR-3 responding was re-established between the PR tests. Rats received systemic oxytocin (15 or 45 IU/kg IP) or vehicle prior to progressive ratio sessions and recording sessions. Dopamine transients were measured using a fiber photometry approach and analyzed using pMAT software.

Results: Non-contingent IV administration of oxycodone increased dopamine signal in the nucleus accumbens. Oxytocin (at both doses tested) augmented the oxycodone-induced dopamine signal in both male and female rats [F (1.787, 33.96) = 8.498, p = 0.0014]. In rats seeking oxycodone, 15 IU/kg oxytocin only reduced the breakpoint for intravenous oxycodone at the trend level (p = 0.06) in male rats. Despite the lack of behavioral effect, oxytocin (15 IU/kg) reduced the dopamine response (AUC) for the two seconds prior to and after the first lever press of the progressive ratio session [t(5) = 2.808, p = 0.037].

Conclusions: These results indicate that while oxytocin increases the nucleus accumbens dopamine response to intravenous oxycodone, this effect likely does not underlie oxytocin’s effects on demand elasticity, as it was observed in both male and female rats. Oxytocin’s ability to reduce oxycodone seeking under high effort conditions may stem from its ability to reduce the nucleus accumbens dopamine activity immediately prior to/after a lever press for oxycodone. The follow-up studies will investigate additional mechanisms through which oxytocin regulates operant oxycodone-seeking.

Keywords: oxycodone, oxytocin, Dopamine, fiber photometry, drug seeking

Disclosure: Nothing to disclose.

P348. Mechanisms behind microbial regulation of psychostimulant use disorders

Samuel Mabry, Xixi Cao, Yanqi Zhu, Caleb Rowe, Shalin Patel, Camila Gonzalez-Arancibia, Tiziana Romanazzi, David Saleeby, Hui Wu, Aurelio Galli, Angela Carter

The University of Alabama At Birmingham, Birmingham, Alabama, United States

Background: The abuse potential and psychomotor stimulant properties of amphetamines (AMPHs) are associated with their ability to increase extracellular dopamine levels. This increase is mediated, at least in part, by reversal of dopamine transporter (DAT) function, causing non-vesicular dopamine (DA) release (efflux).

Dysbiosis, an imbalance in the host microbiome, is associated with drug abuse. Microbial fermentation products, such as the short-chain fatty acid butyrate, are suspected to play a fundamental role in pathogenesis of substance use disorders. Fusobacterium nucleatum (F.n.), a native human microbiome member considered an opportunistic pathogen, is a bacterial species that secretes butyrate and whose abundance is increased by AMPHs in both rodents and humans. Whether F.n. plays an active role in promoting AMPH use disorders and the mechanisms driving this phenomena are unknown.

Methods: Fusobacterium nucleatum (Fn, ATCC 23726) were grown in tryptic soy broth supplemented with 1% Bacto Proteose Peptone No.3 plus 0.05% freshly made cysteine hydrochloride in an anaerobic environment.

All Drosophila melanogaster lines were obtained from Bloomington Stock Center and maintained on standard cornmeal-molasses fly food at 25 °C under a 12:12-hour light-dark schedule.

DA efflux was monitored using amperometry in whole brains of Drosophila. DAT mRNA expression was investigated using qPCR of Drosophila heads and RNAScope of Drosophila brains. DAT protein from Drosophila heads was analyzed using immunoblotting. Behavioral assays performed include locomotion, courting, and drug preference.

Results: Colonization of the intestinal tract of gnotobiotic Drosophila with F. nucleatum significantly enhances DAT expression, AMPH-induced DA efflux and associated behaviors, including locomotion and sexual motivation. This potentiation of AMPH actions, by F. nucleatum, was paralleled by oral administration of butyrate. Further, both pharmacological inhibition and genetic knockdown of histone deacetylase (HDAC) 1 increased DAT expression and AMPH-induced DA efflux as well as AMPH-induced locomotion.

Conclusions: These data support the hypothesis that F. nucleatum enhances AMPH-induced behaviors through secretion of butyrate, inhibition of HDACs, elevation of DAT expression, and increased DA efflux. Together, this suggests microbes, and F. nucleatum in particular, as actionable drug targets, for treatment of AMPH use disorders, that exist outside of the blood-brain barrier.

Keywords: Substance abuse disorders, microbiome, behavioral mechanisms

Disclosure: Nothing to disclose.

P349. Nicotine modulation of dopamine release dynamics underlies sex-specific responses to environmental stimuli

Adora Norman, Taylor McNeal, Grace Bailey, Jennifer Tat, Janet Mariadoss, Erin Calipari, Lillian Brady

The University of Alabama At Birmingham, Birmingham, Alabama, United States

Background: Substance Use Disorders (SUDs) involve disruptions in neural dopamine circuits governing motivation and reward, resulting in maladaptive behavioral patterns. Biological sex is a key factor in SUD vulnerability, with epidemiological data consistently showing sex-dependent differences. Tobacco use disorder, driven by the strong addictive properties of nicotine, remains one of the most prevalent SUDs and continues to be a leading cause of preventable mortality worldwide. Although traditional cigarette smoking had been declining, the rapid growth of e-cigarette use has reignited a nicotine use epidemic. Notably, nicotine alone functions as a relatively weak reinforcer, but when paired with environmental stimuli, it elicits robust responses. The present work aimed to investigate how environmental stimuli interact with nicotine to alter behavioral responses and to elucidate dopamine release mechanisms underlying the effects of this interaction in a sex-specific manner.

Methods: We took a multifaceted approach to determine sex-dependent neurochemical mechanisms that underlie the role of repeated nicotine exposure on operant sensory reinforcement in male (n = 44) and female (n = 44) mice. First, mice were trained for sensory reinforcement, where they emitted a nose poke response for the presentation of a visual and auditory stimulus. Next, mice were treated with saline or nicotine via subcutaneous injection for five days. Additionally, separate cohorts of male (n = 10) and female (n = 10) mice housed in enriched or standard home cage environments were treated with saline or nicotine (0.05, 0.1, and 0.2 mg/kg/mL) daily for four days. Following each injection, mice were placed in an open field environment for 30 minutes to determine how acute nicotine exposure and home cage environmental enrichment altered locomotor activity in a sex-specific manner. Furthermore, we assessed dopamine release dynamics and the effects of nicotine on these dynamics using fast scan cyclic voltammetry in the nucleus accumbens. Finally, we correlated the effects of nicotine on dopamine release with behavioral responses in the same animals that underwent the sensory reinforcement and open field tasks, using a variety of statistical methods (two-way ANOVA, regression analysis, etc.) to assess significance.

Results: Mice showed high rates of responding that were significantly higher than in the unreinforced condition. Interestingly, sensory stimuli were more reinforcing in female mice than in males. After mice were treated with saline or nicotine for five days, we found a significant increase in sensory reinforcement in response to nicotine in male mice, but not in female mice. Notably, correlating the effects of a physiologically relevant concentration of nicotine on dopamine release with the effects of nicotine on reinforced responses to sensory cues revealed a significant negative correlation in males (p = 0.027) but not females (p = 0.343). This was observed when tonic dopamine release decreased as reinforced nose poke responses to sensory cues increased. Additionally, assessing the effects of acute nicotine exposure and environmental enrichment on locomotor activity displayed a more robust reduction in overall distance traveled and average velocity in males compared to females. Interestingly, we observed a sexually dimorphic correlation between the effects of nicotine and environmental enrichment on locomotor activity and dopamine release dynamics, depending on nicotine concentration [30 nM Nicotine (% of 1P dopamine release) vs total distanced traveled (cm); p = 0.034 (males), p = 0.812 (females)] and dopamine release frequency [phasic/tonic ratio (0 nM Nicotine) vs total average velocity (cm/s); p = 0.748 (males), p = 0.003 (females)].

Conclusions: These findings indicate that sex differences in nicotine modulation of dopamine release contribute to sex-specific responses to environmental stimuli. Furthermore, these results suggest that nicotinic receptor activation versus desensitization may differentially influence observed behavioral outcomes.

Relevance: Evidence from both clinical and preclinical studies demonstrates significant sex differences in drug intake, progression to substance use disorder, and sensitivity to environmental cues, supporting a biological basis for sex-specific vulnerability. Clarifying how nicotine shapes sex-dependent interactions between environmental cues and behavior, and identifying the underlying neural circuits driving these differences, will be critical for informing the development of targeted treatment strategies for nicotine dependence and, more broadly, substance use disorders.

Keywords: Dopamine, Nicotine exposure, Sex-specific effects, enriched environment, Reinforcement learning

Disclosure: Nothing to disclose.

P350. Individual differences in oral oxycodone escalation drive lasting vulnerability signatures

Brian Kipp, Trevonn Gyles, Leanne Holt, Veronika Kondev, Elizabeth Kahn, Earnest P. Chen, Eric Nestler

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

Background: Opioid use disorder (OUD) continues to be a complex and burgeoning healthcare issue with an estimated annual cost greater than $700 Billion USD in 2018, and an increasing rate of overdose deaths in the years since. Prescription opioids have contributed to the ongoing opioid epidemic and remain a potent gateway drug in substance use disorders. Among prescription opioids, Oxycodone has a greater abuse potential due to, in part, higher likeability scores and fewer adverse side effects compared to other opioids. Of particular concern, are the lasting effects that drive opioid seeking during periods of abstinence.

Methods: We sought to characterize a homecage prescription opioid abuse model by measuring behavioral and transcriptional changes following oral oxycodone self-administration. Singly housed adult male and female C57BL/6 mice were exposed to increasing oxycodone concentrations in drinking water prior to 14 days of 1mg/mL oxycodone two-bottle choice or water (CON). A clustering-based approach was used to identify subpopulations of oxycodone consuming mice, namely those that escalate in their consumption (High Escalators—HE), versus those who maintain stable oxycodone consumption (Low Escalators—LE) throughout the paradigm. We then performed RNA sequencing on tissue from the prefrontal cortex (PFC) and nucleus accumbens (NAc), followed by Multiscale Embedded Gene Co-expression Network Analysis (MEGENA) 30 days following cessation of oxycodone self-administration.

Results: We reveal that HE mice demonstrate lasting behavioral differences in withdrawal compared to LE mice and control mice, marked by increased oxycodone seeking (p < 0.05) and negative affective behaviors (p < 0.01) that resemble clinical presentation of opioid use disorder. HE mice also show naloxone-mediated conditioned place aversion, even during abstinence (p < 0.05). RNA-sequencing revealed distinct, region-specific transcription networks that segregate with the HE and LE phenotypes, identifying potential molecular mechanisms in hub genes that could underlie these behavioral effects and subsequent susceptibility for forming opioid use disorder.

Conclusions: These data demonstrate that the 2-bottle choice model recapitulates essential features of opioid addiction and offers interesting avenues for understanding the formation of prescription opioid dependence in clinical populations, and the lasting behavioral consequences in withdrawal.

Keywords: Opioid addiction, Transcriptomics, prefrontal cortex, nucleus accumbens (NAA)

Disclosure: Nothing to disclose.

P351. Preclinical models of polysubstance use involving fentanyl and methamphetamine

Robert Seaman, Simon Tangen, Gregory Collins

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

Background: The “fourth wave” of the overdose epidemic is defined by high rates of multi-drug overdoses. Though fentanyl still predominates, overdoses involving fentanyl and stimulants (cocaine or methamphetamine [METH]) account for ~40% of all drug-related deaths. Motivations and patterns of opioid-stimulant co-use are complex, with individuals indicating they commonly co-use fentanyl and METH both as mixtures (“goofballs”) to achieve a greater high, and as independent entities to achieve a more balanced effect (e.g., using METH to counter the sedative effects of fentanyl). Despite the fact that polysubstance use is the norm rather than the exception, most research continues to examine drugs in isolation, underscoring the urgent need to develop procedures capable of defining interactions between commonly co-used drugs to better inform treatment of polysubstance use disorders.

Methods: Male and female Sprague-Dawley rats (n = 12) were surgically prepared with indwelling venous catheters in the left and right femoral veins to allows for concurrent self-administration of two drugs. All rats were initially trained to self-administer 0.1 mg/kg/inf METH and 0.0032 mg/kg/inf fentanyl under a concurrent FR5:TO 5-sec schedule of reinforcement. Once responding for these two doses stabilized, the dose of the less-preferred drug was increased and decreased by ½-log units to establish a dose-response for drug preference. Subsequently, a subset of rats (n = 4) were used to determine preferences between their preferred drug and mixtures of METH+fentanyl. All other rats (n = 8) responded under a concurrent behavioral economic demand procedure to define economic interactions (e.g., compliments, substitutes, independents) between METH and fentanyl.

Results: When rats had access to 0.1 mg/kg/inf METH and 0.0032 mg/kg/inf fentanyl, most responded exclusively for either METH (n = 4) or fentanyl (n = 4), though a subset of rats (n = 4) responded for both drugs, either within the same session or from session to session. Increasing the dose of the less-preferred drug shifted preference towards that drug in all rats, though some rats required larger doses than others. Reducing the dose of the less-preferred drug resulted in exclusive choice of the preferred drug in all rats. When rats had a choice between their initially preferred drug and a mixture of METH+fentanyl, some rats (n = 2) strongly preferred the mixture, whereas other rats (n = 2) preferred to respond for the single drug alternative, regardless of the constituent doses. When the ratio for the preferred drug was increased in the concurrent demand procedure, all rats treated METH and fentanyl as economic substitutes, though a subset of rats (n = 2) increased responding for the less-preferred drug while continuing to respond for the preferred drug, a pattern of co-use we termed imperfect substitution.

Conclusions: Concurrent use of fentanyl and stimulants is common amongst people who use opioids. These studies used a concurrent access model of polysubstance use and found that subsets of rats exhibited behavior indicative of a co-use phenotype. When mixtures of METH+fentanyl were available, rats with shallower preference functions (i.e., histories of responding for both drugs in the same session) chose the mixture over either drug alone. Though increasing the “price” of the preferred drug resulted in all rats reallocating behavior towards the less preferred drug, for some rats this substitution was incomplete and resulted in rats responding for both drugs across a number of conditions. Although the determinants of these co-use phenotypes are currently unknown, understanding these relationships would advance our understanding of vulnerability to polysubstance use while also providing a platform to evaluate candidate medications to more effectively treat individuals suffering from a polysubstance use disorder.

Keywords: methamphetamine, Fentanyl, Self-Administration, polysubstance use

Disclosure: Headspace, Employee, Spouse/Partner, Weedmaps, Employee, Spouse/Partner, Equulus Therapeutics, Consultant, Self

P352. Early life adversity modulates methamphetamine-induced place preference and extinction: implications for dopamine dynamics in the nucleus accumbens

Brenna Williams, Hannah R. Hyland, Kevin G. Bath

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

Background: Experiencing early life adversity (ELA) increases the risk of substance use disorders (SUDs) and stress-induced relapse following abstinence. Among SUDs, methamphetamine (MA) use and abuse has been steadily increasing and has emerged as a major public health concern with profound psychiatric, medical, and psychosocial complications. In 2021 alone, over 64% of MA users met criteria for MA-SUD for which there are currently no approved pharmacological treatments. Psychostimulants, such as MA, primarily exert their effects by increasing extracellular dopamine (DA), and the pursuit of reward and reinstatement (a model of relapse) has been shown to be mediated by DA release.

Conditioned place preference (CPP) is an established model to study reward-related behaviors and the association of drugs of abuse with contextual cues, which has been implicated in context-induced craving and relapse. This association can be extinguished, and either drug injection or stress can elicit reinstatement of CPP. Prior studies have shown that entry into a drug-paired context evokes robust DA release in the nucleus accumbens (NAc) and activation of the D1R-mediated direct pathway. However, the impact of ELA on MA-induced CPP, extinction, and the underlying DA dynamics remains unclear.

Methods: Four days following birth (PD4) until PD11, ELA assigned litters and their dam were placed in a cage with no bedding, a wire mesh cage bottom, and 3/4ths of a cotton nestlet. Control pups were reared in standard conditions. In adulthood, male and female C57Bl/6N mice were subjected to an 8 -day CPP protocol and tested for methamphetamine (MA) preference. Mice underwent one day of pre-conditioning testing (pretest) for 15 min to determine initial side preference; drug pairings were assigned to the non-preferred side (biased design). Animals received an i.p. Injection of MA (1.0 or 2.0 mg/kg) or saline and were restricted to the respective compartment for 20 min. Day 8, animals underwent a 15 min posttest to examine side preference. Extinction began the following day in which the animals were placed into the chamber with free access to both compartments for 15 min per day for 4 days. Animals were then tested for drug-induced reinstatement with a half dose (0.5 or 1.0 mg/kg) i.p. injection of MA. MA preference was calculated as the difference in time spent between the chambers (i.e. MA preference = [time in drug-paired] – [time in saline paired]) and final scores were calculated as the change in preference between test sessions (i.e. CPP score = [MA preference, posttest] – [MA preference, pretest]). Extinction and reinstatement were calculated as both the CPP score for each day as well as the % change for the time spent in the drug paired side compared to posttest. CPP scores were analyzed using a 2-way ANOVA with a post-hoc Fisher’s LSD. Extinction was analyzed using a 2-way repeated measures ANOVA with Šidák correction for multiple comparisons.

Results: We found that both control and ELA animals develop robust preference for the MA-paired chamber after 4 days of conditioning. At 1.0 mg/kg MA (N = 69), CPP was comparable between groups (n.s.), but at 2.0 mg/kg (N = 46), ELA animals showed heightened preference relative to controls (p = .02). During extinction, ELA animals treated with 1.0 mg/kg dramatically decreased their preference for the drug-paired side from posttest levels (Extinction x Rearing: p = .02) reaching a > 20% decrease in time spent on the drug paired side by Day 4 and demonstrating a significant difference compared to controls (% change from Day 8, p = .008). Both groups reinstated to posttest levels of CPP after MA re-exposure (n.s.). Animals treated with 2.0 mg/kg were slower to extinguish, but ELA animals reached > 20% decrease by Day 4 and demonstrated a significant decrease from Day 1 (p = .01). Both groups challenged with 1.0 mg/kg also demonstrated even greater preference for the drug-paired side during reinstatement compared to posttest (n.s.). There was no effect of sex on any of these measures.

Conclusions: These findings suggest ELA alters behavioral sensitivity to MA dose and extinction dynamics, potentially via changes in NAc DA signaling. In my ongoing work, I will employ fiber photometry and a genetically encoded DA sensor to directly assess how ELA modulates DA release during MA-induced CPP and extinction. Together, these results may shed light on mechanisms underlying SUD risk in vulnerable populations and inform new intervention strategies.

Keywords: early-life adversity, Methamphetamine use disorder, conditioned place preference

Disclosure: Nothing to disclose.

P353. Ethanol effects on dorsal striatal cholinergic signaling

Armando Salinas, Charles Levy, Jacob Mitcham, Morgan Schichtel, Logan Slade

Louisiana State University Health Sciences Center Shreveport, Shreveport, Louisiana, United States

Background: Alcohol use disorder (AUD) affects over 29.5 million people in the United States and has adverse health and socioeconomic impacts totaling over $240 billion each year. Despite this prevalence, FDA approved medications for AUD are limited, warranting further study towards understanding of neurobiological mechanisms underlying AUD. Acetylcholine is critical for cognition and decision making. Given the dysfunction in decision making in patients with intoxicating blood ethanol levels as well as in AUD patients, acetylcholine (ACh) signaling becomes an attractive pharmacotherapeutic target. Therefore, the experiments in this study aimed to determine how acute and chronic alcohol affect acetylcholine signaling in the dorsal striatum (DS).

Methods: To examine the acute effects of alcohol on ACh release, we first performed cell-attached electrophysiological recordings of dorsal striatal cholinergic interneurons (CINs) from ChAT-Cre-TdTomato mice. We then assessed ACh release following acute ethanol using a genetically-encoded fluorescent biosensor, iAChSnFR, in in vitro and in vivo experiments using brain slice photometry and in vivo fiber photometry, respectively. To model AUD, we used a mouse model of ethanol dependence, the chronic intermittent ethanol vapor exposure model (CIE), and treated mice for four weeks of CIE. We then used iAChSnFR and exposed mice to CIE treatment and performed brain slice photometry three to five days following the final ethanol exposure period (to ensure that there was no residual ethanol on board at the time of these experiments). To assess chronic ethanol-induced loss of CINs, we treated mice to CIE and harvested brains three days following the final ethanol exposure. We then performed immunofluorescent staining for choline acetyltransferase (ChAT) and counted all ChAT-ir positive cells.

Results: We found that acute ethanol (40mM) treatment depressed CIN firing rate ~24% (n = 5 cells). Accordingly, we also found a decrease in evoked acetylcholine release following acute ethanol treatment in dorsal striatal brain slices (n = 6 slices) and in the dorsal striatum of freely moving mice (2g/kg ethanol; n = 4 mice). Following CIE treatment, we found that evoked acetylcholine release was depressed in the dorsomedial, but not dorsolateral, subdivision of the striatum (n = 6–10 slices). To assess the cause of this striatal subregion specific acetylcholine deficit, we performed a stereological cell count and found that ChAT-ir positive neurons were decreased in the dorsomedial, but not dorsolateral, striatum of CIE treated mice relative to controls (n = 6–10 mice). We also counted ChAT-ir positive neurons in several other cholinergic nuclei but found no significant differences between control and ethanol treated mice.

Conclusions: Altogether, our data show that both acute and chronic ethanol depress dorsal striatal acetylcholine release in subregion dependent manner. Further, we also showed that chronic alcohol has pernicious effects in a dorsal striatal subregion-dependent manner.

Keywords: acetylcholine, alcohol use disorder, Dorsal striatum

Disclosure: Nothing to disclose.

P354. Modulation of opioid reward via galanin signaling within the ventral tegmental area

Brittany Pate, Stephanie Foster, Ewa Galaj, Katharine McCann, Emma Foster, Andy Chen, Sergi Ferre, David Weinshenker

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

Background: Over three million Americans suffer from opioid use disorder (OUD), and opioid overdoses account for more than 75% of all drug-related overdose deaths in the United States. Not only are the effects of OUD devastating for individuals and families affected, but this epidemic also poses a significant economic burden on society. The cost of OUD in the United States was estimated at nearly $1.5 trillion in 2020, and these costs are only expected to increase as a consequence of the pandemic among other factors. Despite this national public health crisis, there are only three FDA-approved medications for the treatment of OUD, and each of these have side effects including respiratory depression, diversion for illicit use, and a necessary detoxification period. Therefore, it is critical that we identify novel pharmacotherapies for OUD. One approach to developing new treatments involves targeting endogenous neurobiological circuits that modulate opioid addiction-related behaviors. Opioid reward is mediated, in part, by suppressing GABA transmission in the ventral tegmental area (VTA) via µ opioid receptor (MOR) activation, thus disinhibiting local dopamine (DA) neurons. Importantly, galanin is an endogenous neuropeptide that regulates opioid responses, most notably opposing opioid reward. Prior studies have demonstrated that mice lacking galanin globally exhibit increased morphine-induced locomotor activity and conditioned place preference (CPP), whereas ICV infusion of galanin has the opposite effect and suppresses morphine CPP. Finally, galanin receptor 1 (GalR1) can form heteromers with MOR, and galanin signaling via MOR-GalR1 heteromers suppresses MOR-induced increases in extracellular DA. These findings suggest that galanin transmission in the VTA may represent a therapeutic target for OUD, but the neurocircuitry and molecular mechanisms underlying the ability of galanin to oppose opioid reward have not been identified. The current experiments were designed to identify 1) the impact of galanin signaling in the VTA on morphine CPP, 2) the co-expression of GalR1 and MOR in critical nodes of the reward pathway such as the VTA, rostromedial tegmental nucleus (RMTg), and the nucleus accumbens (NAc), and 3) areas of the VTA most sensitive to galanin signaling via GalR1.

Methods: To examine the role for VTA galanin in morphine CPP, male and female mice (n = 4 male/ 4 female per group) were implanted with indwelling bilateral cannulae in the VTA and received galanin (400 ng/side) or vehicle 20 min prior to morphine during CPP acquisition. Behavioral data from these experiments were analyzed via 2-way ANOVA. To verify that GalR1 and MOR are co-expressed within the same neurons and thus capable of forming heteromers in brain regions critical to reward seeking, we used RNAscope in situ hybridization to identify mRNA expression of GalR1 and MOR in GABAergic (GAD1+) and non-GABAergic (GAD1-) cells in the VTA, RMTg, and NAc (n = 3–4 per region). To map galanin and GalR1 expression throughout the VTA, adult, immunohistochemistry was used to determine co-localization of galanin fibers and GalR1 throughout the VTA using male mice (n = 3) that express an mCherry-tagged version of GalR1.

Results: In experiment 1, we found that intra-VTA galanin infusion attenuated morphine CPP (p = 0.02). Additionally, we identified galR1 and MOR mRNA co-expression in the VTA, RMTg, and NAc, with greater co-expression in GABAergic vs. non-GABAergic cells. Ongoing analyses will map the co-expression of GalR1 and galanin fibers throughout the VTA.

Conclusions: Together, these findings suggest that galanin signaling in the mesolimbic pathway represents a viable target for modulating opioid reward and shed light on the mechanism by which galanin modulates morphine CPP. These behavioral findings lay the foundation for investigating galanin signaling within the VTA as novel therapeutic target for OUD. To explore the translational relevance of these findings, future experiments will determine the impact of intranasal galanin administration on morphine CPP.

Keywords: galanin, Mu opioid receptor, conditioned place preference

Disclosure: Nothing to disclose.

P355. Precision medicine for AUD: P2X4 receptor modulation reduces alcohol consumption in genetically heterogeneous rats

Paola Campo, Michelle Doyle, Ran Qiao, Thiago Sanches, Benjamin Johnson, Abraham Palmer, Marsida Kallupi, Giordano de Guglielmo

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

Background: Alcohol Use Disorder (AUD) affects ~16% of the ~175M Americans who drink annually, with genetic variation contributing to individual differences in vulnerability and medication response. Heterogeneous stock (HS) rats model human-like genetic and behavioral diversity and enable genetics-guided target discovery.

Methods: We phenotyped 500 male and female HS rats for addiction-like behaviors: fixed-ratio self-administration of 10% ethanol (v/v) to stable intake, ethanol preference, progressive-ratio motivation, and quinine-adulteration (compulsivity). Ethanol dependence was induced by chronic intermittent vapor (14 h/day, 4 weeks; blood ethanol 150–250 mg/dL). During acute withdrawal (6–8 h post-vapor) we quantified escalated intake, motivation, compulsivity, sensitivity, somatic signs, and hyperalgesia to derive an Addiction Index. We then performed a genome-wide association study (GWAS) that implicated P2rx4 (P2X4 receptor). In an independent CIE-exposed cohort (n = 130), we predicted whole-brain P2rx4 genetically regulated expression (GReX) using RatGTEx cis-eQTLs with haplotype-aware allelic fold-change modeling and dichotomized animals into high- vs low-GReX groups. To test functional relevance, dependent rats (n = 32; 16/sex) received ivermectin (1, 2.5, 5, 10 mg/kg) or vehicle 4 h before self-administration. Ex vivo whole-cell recordings from central amygdala (CeA) neurons assessed ivermectin effects on spontaneous inhibitory postsynaptic currents (sIPSCs).

Results: GWAS identified P2rx4, encoding the P2X4 receptor, as a candidate gene; higher P2rx4 GReX predicted greater post-vapor intake and escalation. Ivermectin reduced ethanol intake dose-dependently, significantly at 5–10 mg/kg. Reduction magnitude tracked baseline intake, enabling responder stratification (non-, mild-, high-responders). In CeA, ivermectin enhanced GABAergic inhibition, increasing sIPSC frequency overall, with sustained frequency elevations and selective amplitude reductions in high-responders, transient effects in mild-responders, and slowed kinetics (prolonged rise/decay) across groups.

Conclusions: Large-scale behavioral genetics in HS rats nominates P2rx4 as a liability locus for withdrawal-driven drinking. Integrating GWAS with RatGTEx-based GReX links genetically predicted P2rx4 upregulation to dependence-like escalation and identifies ivermectin, a positive allosteric modulator of P2X4, as a precision-guided intervention that normalizes CeA inhibition and reduces alcohol intake in genetically susceptible subjects. These data support P2X4 modulation as a mechanistically anchored, stratification-ready approach for AUD.

Keywords: Alcohol Use Disorder - Treatment, GWAS, GReX, Precision Medicne neuropsychiatric diseases

Disclosure: Nothing to disclose.

P356. Ventral pallidal regulation of heroin seeking versus refraining behavior

Brittany Kuhn, Jordan L. Hopkins, Eric Dereschewitz, Brice Ayissi Owona, Ava Berckmueller, Stephen Walterhouse, Dongjun Chung, Peter Kalivas

Baylor University, Waco, Texas, United States

Background: Substance use disorder (SUD) is a chronic neuropsychiatric disorder whereby compulsive drug seeking and relapse occurs regardless of attempts to refrain from drug use. The neurobiological mechanisms underlying drug seeking versus refraining (i.e., withholding from seeking) are necessary to understand how to better treat those afflicted with SUD. The ventral pallidum (VP) exhibits opposing regulation of appetitive and aversive motivated behaviors, including cell-specific regulation over cocaine seeking versus refraining behavior. Selective activation of VP GABA or enkephalin (Penk) cells augments cocaine-seeking behavior, whereas VP glutamate (Glu) activation attenuate seeking behavior. These findings prompted investigation into how VP cellular phenotypes moderate opioid-seeking behavior, and furthermore what distinct VP cellular pathways contribute to these effects.

Methods: Male and female transgenic mice (5–10/group) expressing Cre recombinase selectively in either GABA, Penk or Glu cells were used. Mice underwent training for heroin self-administration, extinction and cued reinstatement. During tests for cued reinstatement, cell-specific chemogenetic inhibition or activation occurred. Next, an intersectional chemogenetic approach was used to assess the contribution of cell-specific VP efferent pathways in mediating heroin seeking versus refraining behaviors. We hypothesized that VP-Glu cells and projections would modulate refraining behavior whereas VP-GABA would potentiate seeking behavior. Repeated measures ANOVA was used to assess heroin self-administration and extinction behavior, while 2-way repeated-measures ANOVA and paired t-tests were used to assess reinstatement behavior.

Results: Selective chemogenetic stimulation of VP GABA (p = 0.001) or enkephalin (p < 0.05) cells enhanced cue-induced reinstatement of heroin-seeking behavior, whereas VP glutamate neuron stimulation promoted refraining behavior (p = 0.003). Selective inhibition had opposing effects (all p < 0.05). Activation of VP glutamate projections to the lateral hypothalamus, a critical component of the motive circuitry, attenuated heroin seeking (p = 0.04). However, stimulation of GABAergic projections in this pathway had no effect (p = 0.99). Contrary to expectations, activation of VP GABAergic efferents to the subthalamic nucleus, a region involved in reward seeking, also reduced heroin-seeking behaviors (p = 0.008) and were not the result of a reduction in general locomotor behavior (p = 0.77).

Conclusions: These results emphasize the complex VP cellular and circuitry contributions to heroin seeking and refraining behaviors. Extending upon these findings, using a rat model capturing individual variation in the behavioral and diagnostic heterogeneity akin to human SUD, VP cell-specific neuronal activation in vulnerable versus resilient rats following cued reinstatement is currently under investigation. The objective of this experiment is to elucidate how VP cell types naturally engage in promoting seeking behavior in different addiction phenotypes.

Keywords: Opioid addiction, Ventral Pallidum, GABA, glutamate, subthalamic nucleus

Disclosure: Nothing to disclose.

P357. Downregulation of FGFR1 decreases compulsive alcohol self-administration in rats

Li Xu, Estelle Barbier, Leon Höglund, Dayne Mayfield, Nihal Salem, Robert Messing, Markus Heilig

Linköping University, Linköping, Sweden

Background: A key behavioral hallmark distinguishing individuals with alcohol addiction from casual drinkers is the persistence of alcohol consumption despite adverse consequences, commonly referred to as compulsive alcohol use. This behavior can be modeled in rodents using a footshock-punished alcohol self-administration procedure in which alcohol delivery is paired with an electric footshock. Our previous findings identified GABAergic neuronal ensembles characterized by expression of protein kinase C delta (PKCδ) in the central amygdala (CeA), activated in rats resistant to punishment. However, the molecular mechanisms responsible for the heightened activity of PKCδ+ neurons remain unknown. Here, we evaluated the functional significance of a candidate emerging from our transcriptomic analysis, Fgfr1. We used a Prkcd-Cre transgenic rat line developed in our laboratory (PMID 38977300), in combination with a Cre-dependent viral vector strategy to selectively manipulate Fgfr1 within PKCδ + CeA neurons.

Methods: 32 male Wistar rats were trained to self-administer alcohol (20% (v/v)) for 30 minutes sessions, followed by a punishment phase in which alcohol-self-administration behavior was paired with an aversive footshock (0.1 to 0.25 mA, 0.5 s), The resistance score (RS) was calculated as follows: (punished alcohol deliveries)/(punished alcohol deliveries + mean alcohol deliveries of the last three non-punished sessions). Rats with an RS above 0.45 were selected as the punishment resistant group, and rats with an RS below 0.3 were considered as the punishment sensitive group. Rats were immediately sacrificed after last session, and brains were fresh frozen and sectioned for expression analysis using RNAscope in situ hybridization.64 male Prkcd-cre rats were trained using the same procedures, and both resistant and sensitive groups received cranial viral vector injections bilaterally into CeA. Animals were re-exposed to the footshock punishment procedure 4 weeks after surgery. After all behavior procedures finished animals were transcardiac perfused and brain sections were used for RNAscope.

Results: RNAscope results showed that approximately 65–72% of Fgfr1+ neurons also express PKCδ. We detected a significant upregulation of Fgfr1 mRNA in PKCδ+ neurons in punishment-resistant compared to resistant sensitive rats, confirming a snRNA-seq analysis that identified Fgfr1 as a candidate target (n = 14, p < 0.05). Injection of the Cre-dependent AAV vector encoding an shRNA targeting Fgfr1 bilaterally into the CeA of Prkcd-cre rats resulted in ~40% decrease in Fgfr1 expression in PKCδ+ neurons (n = 20, p < 0.01). This targeted knockdown significantly decreased resistance to punishment as demonstrated by a decreased resistance score in Fgfr1 KD rats compared to scrambled controls (n = 20, p < 0.05).

Conclusions: Fgfr1 signaling in CeA PKCδ+ neurons contributes to compulsive alcohol intake, identifying a druggable target for alcohol addiction therapeutics.

Keywords: Alcohol addiction, Animal models of addiction, drug target, RNAscope fluorescence in situ hybridization

Disclosure: Nothing to disclose.

P358. Sex differences in endocannabinoid regulation of repeated stress-cocaine interactions and cocaine-evoked dopamine in rats

Jayme McReynolds, Andrew Gaulden, Erin Tepe, Sierra Rollins, Nicolas Wiles, Claire Lopez

University of Cincinnati, Cincinnati, Ohio, United States

Background: Clinical evidence has identified stress as an important contributing factor to substance use disorder (SUD). This is particularly problematic as stress is unavoidable in daily life. Therefore, understanding the neurobiological mechanisms that underlie the contribution of stress to SUD is critical. In rats, we can model the impact of stress on cocaine-related behaviors by incorporating repeated daily stress at the time of self-administration (SA). This paradigm produces an increase in cocaine SA in a glucocorticoid-dependent manner. This stress-induced escalation of SA is likely influenced by sex, as there are notable sex differences in both SUD and stress reactivity, and likely involves neurobiological mediators that connect stress-responsive and reward systems in the brain, such as the endocannabinoid system (eCB). We hypothesize that repeated stress at the time of SA induces a persistent increase in eCB signaling, particularly in regions critical for both cocaine taking and seeking, that results in increases in cocaine SA and increased susceptibility to later reinstatement and that these effects are influenced by sex.

Methods: Male and female rats were trained to SA cocaine (0.5 mg/kg/inf) on a FR 4 schedule in 4 X 30 min SA sessions separated by 5-min drug-free periods. Some rats received intermittent electric footshock stress in the SA chamber during the 5 min drug-free period over 14 days. We examined the involvement of endocannabinoid signaling in stress-escalated cocaine intake by administration of a cannabinoid receptor type 1 antagonist systemically. To explore the role of endocannabinoid signaling in regulating cocaine-evoked dopamine, rats received infusions of a dopamine biosensor, dLight1.3, into the nucleus accumbens (NAc shell) with a fiber optic cannula implanted in the same region to measure changes in cocaine-evoked dopamine via in vivo fiber photometry. Rats were pre-treated with either a systemic CB1R antagonist or a MAGL inhibitor, to elevate levels of 2-AG, followed by passive administration of iv cocaine (0, 1, 3 mg/kg). Additional rats were tested for changes in extinction training and increased susceptibility for reinstatement to various stimuli. To test for the role of endocannabinoid signaling in cocaine-seeking behavior, rats received either a CB1R antagonist or DAGL inhibitor, to reduce levels of 2-AG, administered directly into the prelimbic cortex (PrL) prior to cocaine-primed reinstatement (10 mg/kg, i.p.).

Results: Footshock stress significantly increases cocaine SA in male and female rats (n = 16–18 rats/group; p < 0.05) and females stress rats show increased front-loading behavior and increased non-reinforced responding. Systemic CB1R antagonist attenuates cocaine SA in male and female rats with a prior history of stress with females showing increased sensitivity to CB1R antagonism (n = 8–10 rats/group, p < 0.05). Using dLight and fiber photometry in the NAc shell, we found that female rats have greater sensitivity to CB1R and 2-AG regulation of cocaine-evoked dopamine than males (n = 6–9 rats/group; p < 0.05). Rats with a history of stress show augmented cocaine-primed reinstatement in males, but not female, rats (n = 8–10 rats/group). However, in both sexes, intra-PrL CB1R antagonist attenuates cocaine-primed reinstatement only in rats with a history of stress. Additionally, this effect is mediated by the endocannabinoid 2-AG as reduction of 2-AG levels with intra-PL administration of a DAGL inhibitor attenuates cocaine-primed reinstatement in male rats with a history of stress (n = 5–8 rats/group).

Conclusions: Repeated stress induces a glucocorticoid-dependent enhancement of cocaine SA that is the result of persistent neuroadaptation. These data suggest that repeated stress recruits endocannabinoid signaling in mesocorticolimbic brain regions to regulate cocaine SA, likely through regulating cocaine-evoked dopamine, and cocaine-seeking behavior. Females may be more sensitive to endocannabinoid regulation of cocaine-related behaviors through regulation of cocaine-evoked dopamine, though this requires further examination in our stress-enhanced cocaine SA model. Understanding the unique mechanisms by which stress can drive drug use and seeking is important and these data implicate endocannabinoid signaling as a potential therapeutic target.

Keywords: cocaine, Acute and Chronic Stress, endocannabinoid system, Dopamine, Medial Prefrontal Cortex

Disclosure: Nothing to disclose.

P359. Novel 5-HT2R allosteric ligands demonstrate efficacy in preclinical models of cocaine use disorder

Christina Merritt, Robert Fox, Noelle Anastasio, Jia Zhou, Kathryn Cunningham

Center for Addiction Sciences and Therapeutics, University of Texas Medical Branch, Galveston, Texas, United States

Background: Approximately 30% of FDA-approved drugs target G protein-coupled receptors, and GPCRs with solved structures guide targeted medications development. Serotonergic GPCRs, including the 5-HT2A receptor (5-HT2AR) and 5-HT2CR are key modulators of cortical circuitry and promising targets for neurotherapeutic drug discovery. However, the close homology of orthosteric binding domains across 5-HT2Rs complicates the design of receptor-selective therapeutics. Allosteric sites offer new opportunities to optimize 5-HT2AR and 5-HT2CR signaling in disorders marked by cortical dysfunction. To circumvent issues with direct agonists, we developed 5-HT2R allosteric ligands (ALs). We report a first-generation series of 5-HT2AR ALs (e.g., AB0124, Exp. 1) and dual 5-HT2AR/5-HT2CR ALs (e.g., JPC0323, Exp. 2) and evaluate receptor selectivity in the in vitro calcium (Ca²⁺) release assay and “disease state” models of cocaine-taking and reinstatement of cocaine-seeking in rats.

Methods: Newly synthesized AL molecules were screened in 5-HT-induced Ca²⁺ release in CHO cells stably expressing h5-HT2AR, h5-HT2BR, or h5-HT2CR (Exp 1a, 2a). Ligand hits underwent in vitro and in vivo DMPK profiling, CNS panel binding, and behavioral testing in the cocaine intravenous self-administration (IVSA) model. Male and female Sprague-Dawley rats (n = 8/sex) underwent jugular catheterization surgery and were trained to on IVSA cocaine (0.25 mg/kg/infusion) to stability (<10% variation in infusions earned over three days) on an FR5 schedule of reinforcement, after which pharmacological AL testing commenced. Doses of AB0124 (1 mg/kg) and JPC0323 (1, 3, 10 mg/kg) were selected based on prior locomotor activity studies. Exp 1b: In intake studies, the 5-HT2AR AL AB0124 was given in combination with an effective (0.1 mg/kg) and ineffective dose (0.03 mg/kg) of the potent 5-HT2AR agonist (-)-DOI. Exp 2b: In IVSA and cue-reinstatement studies, a full dose-response of the dual 5-HT2AR/5-HT2CR AL JPC0323 was administered prior to test sessions. All statistial analyses were conducted using a within-subjects 1-way ANOVA or mixed effects ANOVA with Šídák’s or Dunnett’s multiple comparisons analysis, where appropriate; statistical significance was defined as α = 0.05, with analyses powered to β = 0.8.

Results: Exp 1a: The 5-HT2AR AL AB0124 enhanced 5-HT-evoked calcium efflux (Emax) at h5-HT2AR (p < 0.05), but not at the 5-HT2BR or 5-HT2CR (n.s.), with no evidence of intrinsic agonist activity at any 5-HT2R (n.s.). Exp 1b: In male and female rats trained to self-administer cocaine, 0.1 mg/kg of (-)-DOI suppressed cocaine intake (p < 0.05), whereas 0.03 mg/kg alone or in combination with AB0124 failed to alter cocaine intake (n.s.). Exp 2a: JPC0323 enhanced 5-HT-evoked calcium efflux (Emax) at h5-HT2AR (p < 0.05) and h5-HT2CR (p < 0.05), but not the 5-HT2BR (n.s.), and showed no intrinsic agonist activity at any 5-HT2R (n.s.). Exp 2b: In male rats trained to self-administer cocaine, JPC0323 displayed efficacy to dose-dependently reduce cocaine intake (p < 0.05) and reinstatement of cocaine seeking (p < 0.05) after extinction training.

Conclusions: Concurrent enhancement of 5-HT2AR and 5-HT2CR signaling may represent an optimal strategy for acute pharmacological intervention to curb cocaine-taking and -seeking in preclinical models. These compounds are currently under patent protection and we are actively seeking academic and industrial partners to support advancement from preclinical development to clinical application. These studies were funded by the NIH National Institute on Drug Abuse (R21MH093844, R01DA038446 and T32 DA007287) and the non-profit organization Cure Addiction Now (Innovation Award).

Keywords: 5-HT2 Receptors, Allosteric Ligands, Cocaine Self-Administration

Disclosure: Nothing to disclose.

P360. Cannabis use changes conditioned stress responses by altering the morphological plasticity of astrocytes through extracellular matrix signaling in the nucleus accumbens core

Ritchy Hodebourg, Eric Dereschewitz, Peter Kalivas

Medical University of South Carolina, Charleston, South Carolina, United States

Background: The growing legal recognition of cannabis and the significant comorbidity observed between cannabis use disorder and post-traumatic stress disorder (PTSD) necessitate a comprehensive understanding of the interactions between stress and cannabis within the brain. Although using cannabis to self-medicate PTSD has been suggested, the literature is mixed on whether cannabis helps or worsens PTSD symptoms. Using acute restraint stress combined with a rat cannabis self-administration paradigm, I recently found that cannabis use promotes two primary PTSD-like symptoms, avoidance coping behaviors and the generalization of stress-coping responses to a neutral stimulus not previously associated with stress exposure. These changes were accompanied by a reduction in spine density in the nucleus accumbens core (NAcore) and a further decrease in spine head diameter after exposure to the stress-conditioned stimulus (stress-CS). Here, we sought to assess whether the interaction between stress and cannabis use influences other components of multipartite synapses, including astroglial plasticity, synapse density, and extracellular matrix catalytic signaling in the NAcore.

Methods: Rats were restraint stressed for 2h and simultaneously exposed to an odor that became the stress-CS. Control rats were exposed to the same odor in the home cage. Three weeks after acute stress, rats self-administered cannabinoids (delta9-tetrahydrocannabinol+cannabidiol; THC + CBD) or vehicle for 10 days. After 10 days of withdrawal, we first evaluated how cannabinoid use affected coping strategies through a defensive burying task (DBT). We then used in vivo zymography, confocal microscopy and digital rendering techniques to quantify astrocyte morphology, colocalization of astrocytes with the presynaptic marker Synapsin-I, and the gelatinolytic activity of matrix metalloproteinases (MMP-2,9) in the NAcore. Finally, to further evaluate the role of MMP-2 and MMP-9 on the multipartite plasticity induced by stress and cannabis, another cohort of rats was treated with either MMP-2 or MMP-9 inhibitors before the DBT.

Results: We demonstrated for the first time that THC + CBD generalized stress responses to a neutral stimulus by causing astrocytes to retract from synapses and decreasing Synapsin-I density in the NAcore. Furthermore, cannabinoid use promoted avoidant coping behaviors in response to a stress-CS by triggering strong activation of MMP-2,9, causing a re-association of astrocytes to synapses along with an increase of Synapsin-I density and astrocyte atrophy. Finally, although both MMP-2 and MMP-9 inhibitors diminished the re-association of astrocytes to synapses induced by stress-CS, only the MMP-2 inhibitor prevented the increase in Synapsin-I density and astrocyte atrophy.

Conclusions: Taken together, these findings emphasize potential therapeutic targets such as MMPs and astrocytes for treating co-occurring CUD/PTSD.

Keywords: cannabis use disorder, Posttraumatic stress disorder (PTSD), astrocytes, extracellular matrix, Matrix Metalloproteinases

Disclosure: Nothing to disclose.

P361. Reactivation of hippocampal cocaine engram cells is required for drug context-induced reinstatement

Shuyi Qi, David A. Soto, Avery Y. Pruitt, Jobe L. Ritchie, Dylan A. Reeves, Rita Fuchs

Washington State University, Pullman, Washington, United States

Background: Cue-induced drug relapse involves the retrieval of cue-drug associative memories. The underlying neurobiological mechanisms of this phenomenon are poorly understood. Here we tested the hypothesis that the functional integrity of cocaine engram cells in the dorsal hippocampus cornu ammonis region 3 (dCA3) is necessary for drug context-induced reinstatement in an instrumental rodent model of drug relapse. To this end, we used a virus-mediated transient activation of active population (TRAP) approach to target halorhodopsin or control protein expression to dCA3 cells that expressed c-Fos at the time of cocaine or extinction memory reconsolidation. We predicted that selective optogenetic inhibition of cocaine engram cells at the time of long-term memory retrieval would reduce cocaine memory strength and cocaine-seeking behavior. We predicted that optogenetic inhibition of extinction engram cells would not alter cocaine-seeking behavior.

Methods: Male and female adult Sprague-Dawley rats (8–10/sex/group) received received bilateral AAV2/8-cFos-ERT2-Cre-ERT2-WPRE-hGH polyA and AAV5-EF1a-DIO-eNpHR3.0-mCherry virus infusions and optic fiber implants into the dCA3 and a jugular catheter implant. After recovery, the rats were trained to lever press for cocaine infusions in a distinct environmental context, and their responding was extinguished in a different context over seven days. Rats received a 4-hydroxy tamoxifen injection (4-OHT; 5 ug/kg, IV) to activate tamoxifen-dependent Cre recombinase-expression and initiate Cre-dependent halorhodopsin expression in cells exhibiting c-Fos expression immediately after the last extinction training session or on the following day, immediately or six hours after re-exposure to the cocaine-predictive context for 15 minutes. All rats then received seven additional extinction training sessions followed by a test session in the extinction context and a drug-free test session in the cocaine-predictive context. During the one-hour test sessions, rats were connected to the optogenetics apparatus with or without intermittent laser light illumination (10 mW; 5 s on, 5 s off). c-Fos expression at test was assessed in the dCA3 using immunohistochemistry and imaged using confocal microscopy. Rats without sufficient virus spread and correct optic fiber placement were excluded from the study. Behavioral data (lever presses, response latency) for each test session were analyzed using analyses of variance (ANOVA) with sex and treatment as between-subjects variables and time as a within-subject factor followed by Tukey’s post-hoc tests, where appropriate. Alpha was set at 0.05 for all analyses.

Results: Exposure to the cocaine-paired context elicited more active lever responses than exposure to the extinction context at test. Optogenetic inhibition of TRAP-ed cocaine engram cells reduced cocaine-seeking behavior independent of sex (p < 0.05) relative to no inhibition. This effect was selective to the cocaine-predictive context and depended on 4-OHT treatment during memory destabilization, halorhodopsin expression, and laser light illumination (p < 0.05). Cell counting is in progress to determine c-Fos immunoreactive cell density at test in the cocaine-paired context as well as the extent of c-Fos expression in previously TRAP-ed cocaine and extinction engram cells in the dCA3. These data will also be analyzed using ANOVA.

Conclusions: Thus far, our results indicate that dCA3 neurons are recruited as engram cells during the reconsolidation of labile cocaine and extinction memories. Signaling in dCA3 cocaine engram cells is necessary for subsequent cocaine seeking in the cocaine-predictive context. Ongoing cellular analyses will determine the extent to which the retrieval of long-term memories during cocaine seeking engages previously recruited cocaine engram cells per se. This research provides insight into the mechanisms of cocaine memory maintenance, and it has the potential to inform anti-relapse treatment development.

Keywords: cocaine, drug seeking, memory engram cell, drug Relapse

Disclosure: Nothing to disclose.

P362. The transcriptional changes in the prefrontal cortex that drive heroin seeking after prolonged abstinence

Zi-Jun Wang, Lu Chen, Shuwen Yue, Fengwei Yang, Tianshi Yu

University of Kansas, Lawrence, Kansas, United States

Background: Opioid use disorder (OUD) stands as one of the most pressing and severe public health crises confronting society. Patients with OUD often become trapped in a persistent cycle of drug use and drug-seeking behavior, driven by intense cravings that can trigger relapses even years after cessation. The medial prefrontal cortex (mPFC), a pivotal node in the brain’s reward circuitry, plays a central role in exerting top-down control over reward processing. Studies have shown that mPFC is involved in opioid-seeking behavior. However, the underlying molecular mechanisms is still unclear. In addition, few studies have characterized the molecular changes that occur during early and late abstinence and contribute to relapse after prolonged abstinence. Therefore, we hypothesize that transcriptional changes in the mPFC occurring during late abstinence play a critical role in driving vulnerability to opioid relapse-like behaviors following prolonged abstinence.

Methods: C57BL/6J mice were purchased from Jackson Laboratory. Both male and female mice were used in this study. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Kansas and were conducted in accordance with the National Institutes of Health guidelines in Association for Assessment and Accreditation of Laboratory Animal Care–accredited facilities.

Animals were subjected to either 1 day (1D) or 14 days (14D) of abstinence following 10 days of saline or heroin self-administration (50 µg/kg/infusion, 3 h/session). Then, mice were sacrificed for mPFC tissue collection and mRNA, which were used for RNA sequencing (RNA-seq) experiment. A different cohort of mice that underwent the same self-administration training were subjected to the validation experiment of top differentially expressed genes (DEGs) using histological staining and PCR analysis. To examine the functional role of top DEGs, mice received viral-mediated manipulation of top DEGs and were assessed in a 1-h heroin-seeking test after forced abstinence.

Results: Our analysis identified 163 significantly upregulated and 137 significantly downregulated DEGs in the mPFC following 1-day abstinence, and 5 significantly upregulated and 36 significantly downregulated DEGs after 14-day abstinence (padj < 0.05, fold change > 1.5). Notably, the transcriptional changes observed during 1-day abstinence included a broader set of genes, whereas 14-day abstinence revealed a more selective set of transcriptional alterations, suggesting dynamic and time-dependent molecular adaptations. Functional enrichment analysis demonstrated that these DEGs are strongly associated with pathways involved in neuroplasticity, emphasizing their relevance to reward-related processes and relapse vulnerability. To further explore the regulatory landscape, we employed an integrated bioinformatics approach to construct gene networks and identify key mRNA axes within the heroin-associated reward circuitry. Fatty acid–binding protein 7 (Fabp7), identified as a hub gene among the DEGs after abstinence, was downregulated after 14 days (n = 4 mice/group, t(6) = 8.893, P < 0.0001, unpaired t-test) but not 1 day (n = 4 mice/group, t(6) = 0.115, P = 0.913, unpaired t-test of abstinence. Overexpression of Fabp7 during abstinence significantly attenuated heroin seeking (n = 10–11 mice/group, t (19) = 2.672, P = 0.0151, unpaired t-test.

Conclusions: In summary, these analyses highlight unique transcriptional signatures that emerge during prolonged abstinence and may drive susceptibility to relapse-like behaviors. Our findings provide critical insights into the time-dependent molecular reprogramming within the mPFC that underpins relapse vulnerability during late abstinence.

Keywords: Medial Prefrontal Cortex (mPFC), gene transcription, Opioid addiction

Disclosure: Nothing to disclose.

P363. Establishment of a naturalistic aversive assay in Sprague-Dawley rats

Arturo Zavala, Christina Nguyen, Adrian Sedillo, Teresa Falcon

California State University - Long Beach, Long Beach, California, United States

Background: Conditioned Place Preference (CPP) studies typically employ an unbiased apparatus, which allows for assessing the rewarding effects of psychostimulant drugs and drug-seeking behavior (i.e., spending more time in an environment previously associated with the drug). In the present study, we aimed to create a naturally aversive environment by illuminating one side of the apparatus with white LED lights to assess the persistence of drug-seeking behavior despite the environment’s aversive nature.

Methods: A two-chamber apparatus was modified to have one compartment with a solid black wall pattern and a perforated acrylic floor featuring horizontal grooves, and the second compartment had a white wall patterned with small black dots and a perforated acrylic floor with round holes. The preference for these two compartments was then assessed for 10 consecutive days during 30 min sessions beginning on postnatal day (PD) 38. The apparatus was then modified so that the white and circle-patterned side was illuminated with white LED lights (650 lux), while the solid black patterned side was illuminated with red LED lights (20 lux). The preference under these two conditions was then examined across four additional days during 45 min sessions.

Results: The time spent in the two compartments across the initial 10 days revealed that male and female rats consistently preferred the solid black compartment to the white one with a circle-patterned design (p < 0.5). During days 11–14, rats showed a strong aversion to the white illuminated compartment, despite an increase in the amount of time they were allowed to explore both compartments (p < 0.5. Importantly, this avoidance of the white illuminated compartment was evident even if animals were initially confined to the white compartment during the first 10 days.

Conclusions: By establishing this new behavioral assay, drug-seeking behavior can be examined using this apparatus to determine whether this behavior persists in an aversive compartment. This approach will enhance our ability to study the neurobiological underpinnings of addiction, given that in humans, substance use is characterized by continued use despite adverse consequences.

Keywords: Animal models of addiction, Conditioning Place Preference, persistent avoidance

Disclosure: Nothing to disclose.

P364. Transcriptomic adaptations during heroin abstinence and motivated drug-seeking highlight actin cycling in the female nucleus accumbens

Treefa Shwani, Newaz Ahmed, Kayla Lilly, Ana Lidia Garcia-Lopez, Lisa Strand, Ian Maze, David Dietz

State University of New York at Buffalo, Buffalo, New York, United States

Background: Opioid Use Disorder (OUD) is a chronic, relapsing disease that has been shown to progress differently between men and women. Despite these differences, the neurobiological mechanisms underlying female-specific vulnerability remain poorly understood. To address this gap, we sought to identify transcriptional and cellular mechanisms associated with heroin motivation in female rats.

Methods: Female Sprague-Dawley rats were trained to self-administer heroin for 10 consecutive days under an escalating fixed-ratio (FR) schedule. Following self-administration (20ug/kg/inf), motivation was assessed on abstinence day 1 (AD1) and abstinence day 14 (AD14) using a progressive-ratio (PR) procedure. Male rats underwent corresponding behavior testing. Subsequently, tissue punches of the nucleus accumbens (NAc) were collected at AD1, AD14, and 24 hours after PR testing at both timepoints for RNAseq and actin assays.

Results: Female rats displayed persistent heroin motivation during abstinence. At AD14, they responded significantly more at their final ratio and were consistently closer than males to advancing to the next ratio, a defined measure of ‘PR remaining.’ Transcriptomic profiling of female NAc revealed robust transcriptional remodeling under PR conditions, with 234 to 243 genes significantly altered across heroin and saline comparisons, in contrast to minimal changes without PR. Gene set enrichment analysis identified pathways converging on actin regulation and immune signaling, including WASP family proteins, Rho GTPases, and interferon-gamma. To further validate these findings, we performed F:G actin assays. Consistent with our transcriptional data, we find a temporally regulated change in actin cycling following exposure to heroin.

Conclusions: Taken together, these findings provide a mechanistic framework for female vulnerability to OUD and highlight actin regulators as potential targets for sex-specific therapeutic strategies.

Keywords: opioid use disorder, female-specific vulnerability, motivation, nucleus accumbens, actin remodeling

Disclosure: Nothing to disclose.

P365. Cocaine disrupts the estrous cycle during conditioned place preference in female rats

Marian Sepulveda-Orengo, Alexia Sanabria, Lenin Godoy

Ponce Health Sciences University, Ponce Research Institute, Ponce, Puerto Rico

Background: Cocaine is a highly addictive psychostimulant that affects men and women differently. Clinical studies show that women may experience a more rapid increase in substance use, stronger cravings, and greater difficulty in quitting, compared to men. It has also been shown that cocaine-dependent women tend to have lower levels of salivary estradiol throughout their menstrual cycles, than women who do not use cocaine. However, the relationship between estradiol, progesterone, and addiction remains unclear. This project aims to investigate how cocaine influences the estrous cycle, including levels of estradiol and progesterone, and the effects of such changes on cocaine-seeking behavior. We hypothesize that female rats exposed to cocaine will experience changes in the stages of their estrous cycle involving alterations in estradiol and progesterone levels, which will subsequently affect their behavior in seeking cocaine.

Methods: To conduct this study, rats underwent a three-day habituation phase, followed by twelve days of conditioned place preference (CPP) training that involved alternating injections of 10 mg/kg cocaine and saline. This was followed by a fifteen-day extinction period and two reinstatement sessions, during which the rats received cocaine doses of 5 mg (Rst1) and 10 mg (Rst2) to assess cocaine-seeking behavior. Throughout the experiment, vaginal smears were collected to monitor the estrous cycles of the rats. Additionally, ELISA assays were performed to measure levels of estradiol and progesterone in both blood and brain tissue after the second reinstatement session (Rst2).

Results: Our results indicated that cocaine induced changes in the stages of the estrous cycle during both the conditioning and extinction phases of conditioned place preference (CPP). In contrast, no such changes were observed in the saline control group. Notably, we found significant daily differences in estrous stages between the cocaine and saline groups. Interestingly, rats exposed to cocaine during the estrus stage in the Rst2 group exhibited decreased brain estradiol concentrations compared to the saline group; however, serum concentrations and estrous stages remained unaffected. While serum estradiol concentrations—unlike brain concentrations—showed a positive correlation with the reinstatement preference ratio score for Rst2, no such correlation was observed for Rst1. Additionally, preliminary data on brain and serum progesterone concentrations in the cocaine group did not reveal any correlation with the reinstatement preference ratio.

Conclusions: This study represents the first reported investigation of the relationship between the physiological effects of cocaine and the estrous cycle, using the Conditioned Place Preference (CPP) model. Our findings indicate that cocaine alters estradiol concentrations differently in the brain and serum, which suggests that higher levels of estradiol in the body may enhance the motivational and reinforcing effects of cocaine. Higher estradiol levels may increase cocaine’s rewarding effects, which could lead to a greater susceptibility to drug-seeking behavior.

Keywords: cocaine, estrous cycle, estradiol, conditioned place preference

Disclosure: Nothing to disclose.

P366. Temporal dynamics of cocaine-elicited dopamine govern how drug cues gain value

Michael Leonard, Thomas Clarity, Ainoa Konomi-Pilkati, Kimberly Thibeault, Erin Calipari

Vanderbilt University, Nashville, Tennessee, United States

Background: The rate of cocaine delivery dramatically influences its acute reinforcing effects, and is a key determinant for the risk of developing cocaine use disorder. In preclinical models, animals preferentially respond to obtain fast drug infusions over slower administration of the same dose. Rapid cocaine infusions also selectively potentiate plasticity markers in the striatum, despite producing comparable peak drug concentrations and elicited dopamine levels. It, therefore, remains unclear how the temporal dynamics of cocaine’s actions within the striatum mediate distinct neurobehavioral outcomes.

Methods: Using microendoscopic and photometric imaging in vivo, we examine how the rate of drug administration (3.0mg/kg; delivered IV over 3, 30 or 100s) influences patterns of neural activity within the NAc of mice expressing either the genetically-encoded calcium sensor GCaMP8 or a fluorescent dopamine sensor (dLight1.2).

Results: We find that cocaine has reproducible effects on cellular activity – increasing the activity of a small population of neurons, while robustly decreasing activity in most remaining cells. While this general pattern was observed across all conditions, the time-course and number of cells that were sensitive to cocaine’s effects were dynamically modulated by injection speed. DA release was also temporally defined by drug delivery rate. Rapid (3s) drug infusions elicited peak DA levels ~6-times faster than when the same dose was infused over 100s, while producing only a modest increase in maximal DA. After repeated pairings with auditory/visual cues, stimuli associated with rapid cocaine infusion (3s-Cue) acquired more potent conditioned stimulus properties than those paired slower drug delivery (100s-Cue). 3s-Cues evoked more robust neural activation, and mice elicited more vigorous instrumental responses to produce the 3s drug-stimulus, even in the absence of cocaine itself (i.e., conditioned reinforcement)—suggesting that cues associated with rapid drug administration exert greater motivational control. Interestingly, delaying the onset of a 3s cocaine infusion was sufficient to diminish the conditioned reinforcing effects of rapid drug delivery.

Conclusions: Collectively, these studies characterize how cocaine pharmacokinetics differentially influence neural dynamics within target striatal circuitry. Rapid drug delivery may potentiate drug-conditioned cue associations by increasing the temporal contiguity between drug-elicited dopamine and the actions/cues that predict it.

Keywords: cocaine, Dopamine, nucleus accumbens core, in vivo calcium imaging, associative learning

Disclosure: Nothing to disclose.

P367. Emerging mechanisms underlying Zhx2 loss-of-function in sex-specific alterations in brain oxymorphone concentration and oxycodone addiction model traits

Sophia Miracle, Morgan Hofmeyer, Ava Glavine, Isabella Conti, Sophia Pavlidis, Bryce Axe, Priyanka Thareja, Aleksandra Gorelik, Kaylie Kaneshiro, Praveen Kulkarni, Christopher Reilly, Craig Ferris, Camron Bryant

Northeastern University, Boston, Massachusetts, United States

Background: Opioid use disorder is highly heritable and remains a public health crisis. Misuse of oxycodone (OXY; active Oxycontin© compound) is a significant source of the opioid crisis. OXY is metabolized to a much more potent, efficacious agonist at the mu opioid receptor, oxymorphone (OMOR) and could enhance OXY addiction liability. We genetically mapped and validated Zhx2 (zinc fingers homeoboxes 2) loss-of-function as a genetic mechanism underlying increased brain OMOR concentration and OXY addiction model behaviors in BALB/cJ females, including increased OXY locomotor sensitivity and changes in OXY reward via conditioned place preference. Here, we further tested the hypothesis that Zhx2 loss-of-function enhances OXY addiction model traits. We also explored neurobiological mechanisms linking Zhx2 KO with increased brain OMOR and OXY behaviors via brain imaging, transcriptomics, and informed, assessment of biomarkers for brain structural dysfunction.

Methods: All experimental protocols in animal studies were approved by the Institutional Animal Care and Use Committee and were conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals, or by the applicable requirements of Northeastern University. To model emotional-affective opioid withdrawal, conditioned place aversion (CPA) to naloxone (NLX)-precipitated withdrawal was tested in Zhx2 knockouts (KOs) vs. wild-type (WT) littermates on a BALB/cByJ background. Mice were first assessed for initial aversion for the NLX-paired side on Day 1. On Days 2 and 4, mice received either saline (SAL; IP) or OXY (40 mg/kg, IP) at 0900h and SAL or NLX (1 mg/kg, IP) at 1300h and were confined to the right side for 30 min. On Days 3 and 5, all mice received SAL in the a.m. and p.m. and were confined to the left side for 30 min. On Day 8, mice were assessed for drug-free NLX-CPA via time spent on the NLX-paired side following a SAL injection in the a.m. and p.m. and placement on the SAL-paired side. On Day 9, mice were assessed for state-dependent NLX-CPA following SAL in the a.m. and NLX or SAL in the p.m. and placement on the SAL-paired side.

To assess brain structural integrity in Zhx2 KOs vs WTs, diffusion tensor imaging (DTI) was employed. Mice were anesthetized and placed in a 7-tesla scanner for 40 min where anatomy and DTI scans were run. We also conducted functional magnetic resonance imaging (fMRI) and functional connectivity (FC) scans to identify the acute OXY response in vivo. Mice were acclimated to the scanner pulse sequence for 5 days for an increasing duration (10–50 min). On Day 6, awake mice were placed in the scanner for fMRI and FC scan (45 min). Five min into the fMRI scan, mice received OXY (1.25 mg/kg) via IP catheter. Blood-oxygen-level-dependent (BOLD) signal change was recorded for 30 min post-OXY. After fMRI scan, an FC scan (10 min) was initiated to assess the effect of acute OXY on region-specific connectivity. At 60 min post-OXY, mice were euthanized, and brain, liver, and tissues were collected for analysis. Transcriptome analysis via paired-end bulk RNA-seq of ½ brain was conducted via NOVA-seq. Deseq2 was used for differential gene expression analysis, in which we focused on enrichment analysis of genes showing Genotype x Treatment x Sex interactions (p < 0.05) to illuminate biological pathways associated with NLX-CPA. Based on imaging results, to assess potential blood brain barrier (BBB) dysfunction, we examined mRNA expression of Claudin5 (Cldn5) via real-time quantitative PCR of half-brain tissue in KO vs. WT mice. Cldn5 codes for a highly enriched tight junction protein and is a reliable marker for BBB integrity.

Results: Zhx2 KOs (n = 22; 11M, 11F) showed enhanced drug-free NLX-CPA vs. WTs [n = 19; 10M, 9F; Genotype x Day: F(2, 42) = 6.035; p = 0.005; Day 8: p = 0.0002]. For brain imaging, Zhx2 KO females (n = 5–8) showed reduced apparent diffusion coefficient (ADC) vs. WT females (n = 5–8), in fornix (p = 0.01), extended amygdala (p = 0.02), and nucleus accumbens core/shell (p = 0.04), and enhanced negative bold signals in zona incerta (p = 0.014), extended amygdala (p = 0.025), and substantia nigra (p = 0.037). In contrast, no differences in structural traits or fMRI BOLD signals were observed in Zhx2 KO males (n = 6–8) vs. Zhx2 WT males (n = 3–4). Enrichment analysis of genes showing a Genotype x Treatment x Sex interaction were highly enriched for granulocyte, neutrophil, and macrophage function (padj < 0.05). Female KOs (p = 0.036), but not male KOs (p = 0.40) showed a reduced Cldn5 expression vs. their WT counterparts.

Conclusions: These findings further support increased brain OMOR concentration underlying enhanced OXY addiction model behaviors in female Zhx2 KOs. Reduced ADC in DTI combined with reduced Cldn5 expression in Zhx2 KO females implicates disrupted brain structural and BBB integrity leading to increased brain OMOR concentration and enhanced negative BOLD response following systemic OXY. RNA-seq results suggest enhanced inflammation and immune cell infiltration resulting from BBB disruption. Future studies include increasing sample sizes for brain imaging, assessment of BBB integrity in vivo and in vitro using cultured endothelial cells, assessment of both operant oral OXY self-administration, shifts in brain reward threshold during spontaneous OXY withdrawal via intracranial self-stimulation, characterization of brain region- and cell type-specific expression of Zhx2 throughout the mouse brain, and snRNA-seq in brain regions with robust Zhx2 expression, including striatum and hippocampus.

Keywords: opiate, substance use disorder, GWAS, pharmacokinetic, fentanyl

Disclosure: Nothing to disclose.

P368. Long-term Δ9-tetrahydrocannabinol (THC) vapor exposure increases anxiety-like behavior, white matter damage, and neuroimmune dysregulation in female rats

Catherine Moore, Riddhi Patel, Hawley Helmbrecht, Eric Chin, Balaji Vijayakumar, Elise Weerts, Lauren Jantzie

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

Background: Vaping of cannabis and cannabis extracts containing Δ-9-tetrahydrocannabinol (THC, the primary psychoactive constituent of cannabis) is on the rise, though we still lack a full understanding of the behavioral and biological consequences of chronic vaporized THC use.

Methods: Female Sprague-Dawley rats (N = 24; 12 per group) had 10 months of exposure to either THC vapor (50 mg/ml) or vehicle vapor (VEH: 100% propylene glycol) on an intermittent (e.g., every other day) basis via passive delivery (1 month; pre-exposure) followed by vapor self-administration. Rats were then tested in the open field test of anxiety-like behavior 24-hrs following their last vapor self-administration session. At the termination of behavioral studies, blood and brains were collected via transcardial perfusion (N = 4–5/group). Diffusion Tensor Imaging (DTI) was performed ex vivo on fixed brains. Brains were then sectioned and processed using immunohistochemistry to quantify myelin basic protein (MBP), a marker of mature myelin, and to examine microglial morphology using Imaris software. Blood was analyzed for inflammatory markers using a multielectrochemiluminescence Immunoassay (MECI: MesoScale Discovery). Data were analyzed with unpaired Student’s T-tests or Mann-Whitney U-tests in the case of non-normal distributions; α=0.05.

Results: Female rats who self-administered THC vapor displayed increased anxiety-like behavior in the open field test (t(18) = 2.99, p < 0.01). Analysis of DTI metrics revealed decreased fractional anisotropy (FA) and increased radial diffusivity (RD) in white matter tracts of rats exposed to THC vapor compared with VEH vapor (FA: t(6) = 5.98, p < 0.01 in corpus callosum, t(6) = 2.5, p < 0.05 in external capsule; RD: t(6) = 2.8, p < 0.05 in corpus callosum, t(6) = 2.70, p < 0.05 in external capsule). MBP was reduced in both the cortex and fimbria of THC vapor-exposed rats (p < 0.0001). THC vapor-exposed rats had higher inflammatory markers in serum (IL-6: t(7) = 3.16, p < 0.05; CXCL1: Mann-Whitney U = 0; n1 = 4, n2 = 5, median1 = 27.15, median2 = 140.1, p < 0.05; TNFα: t(7) = 1.6, p = 0.15). Further, THC-exposed rats had more microglia in the fimbria and cortex compared to controls. Microglia from THC animals were more spheroid, less complex with fewer and shorter processes, consistent with phagocytic activity and molecular activation.

Conclusions: This pilot data demonstrates that chronic exposure to THC vapor produces white matter damage, altered immune function, and increases in anxiety-like behavior. Future work will investigate the link between increased inflammation and effects on white matter and behavior. Preclinical models of THC vaping provide critical translational insights into the effects of cannabis constituents and the consequences of chronic THC vapor exposure.

Keywords: cannabis use, vaping, delta-9-tetrahydrocannabinol, Diffusion Tensor Imaging (DTI), neuroimmune mechanisms

Disclosure: Nothing to disclose.

P369. RMTg-projecting lateral habenula neurons regulate withdrawal-induced hyperalgesia following chronic ethanol exposure

Hyerim Yang, D Nikki Kinarasri, Joseph Pitock, Karl Bosque-Cordero, Abby Hernandez, Elizabeth Glover

University of Illinois Chicago, Chicago, Illinois, United States

Background: Individuals with alcohol use disorder (AUD) often struggle with withdrawal symptoms including anxiety and heighted pain sensitivity that emerge during early abstinence. These symptoms frequently promote relapse as individuals often resume drinking to alleviate their discomfort. Although widely reported, the precise mechanisms mediating these symptoms are still unknown. To bridge this gap, we assessed withdrawal-induced changes in pain sensitivity in a preclinical model of AUD.

Methods: Adult Long-Evans rats were tested for thermal and mechanical pain sensitivity using Hargreaves and Von Frey, respectively. Following baseline testing, rats were rendered dependent using a 14-day chronic intermittent ethanol (CIE) vapor exposure paradigm while control rats were treated with room air (AIR).

Results: Female rats did not show any changes from baseline in either thermal (p > 0.1) or mechanical pain sensitivity (p > 0.05). In contrast, male rats exhibited a significant increase in thermal pain sensitivity that emerged during acute withdrawal and persisted into protracted withdrawal (p < 0.05). Withdrawal was also associated with a significant increase in cFos expression in the rostromedial tegmental nucleus (RMTg; p < 0.0001) and in RMTg-projecting lateral habenula (LHb) neurons (p = 0.029). Selective chemogenetic inactivation of RMTg-projecting LHb neurons ameliorated withdrawal-induced thermal hyperalgesia in a sex-dependent manner (p < 0.05).

Conclusions: Together, these data reveal a critical role for the RMTg and the input it receives from the LHb in governing nociceptive symptoms of withdrawal from chronic ethanol exposure. Supported by NIH grants R01 AA029130, R01 AA031003, R21 AA030867, P50 AA022538, and F31 AA031178.

Keywords: Pain, alcohol withdrawal, Lateral Habenula, RMTg, sex differences

Disclosure: Nothing to disclose.

P370. Changes in temporal firing patterns in the lateral habenula during acute withdrawal from chronic intermittent ethanol

Karl Bosque-Cordero, Shikun Hou, Elizabeth Glover

Center for Alcohol Research in Epigenetics, University of Illinois at Chicago, Chicago, Illinois, United States

Background: Alcohol Use Disorder (AUD) affects millions, with ~50% experiencing withdrawal symptoms that trigger relapse. The lateral habenula (LHb) has emerged as a critical brain region involved in encoding aversive signals, regulating negative affect, and modulating pain sensitivity including during withdrawal. As a predominantly glutamatergic nucleus, the LHb serves as a major output hub, projecting to midbrain structures including the rostromedial tegmental nucleus (RMTg), which in turn inhibits ventral tegmental area (VTA) dopamine neurons. LHb neurons display heterogeneous firing patterns—silent, tonic regular, tonic irregular, and bursting—that are functionally distinct and reflect underlying differences in intrinsic membrane properties, synaptic input strength, and channel conductance profiles. Despite this heterogeneity, very little work has examined changes in LHb temporal dynamics during withdrawal from chronic ethanol exposure.

Methods: Male and female Long-Evans rats underwent chronic intermittent ethanol (CIE) vapor exposure (14 hours/day, 14 days). LHb neuronal excitability and temporal firing patterns were measured 24 hr after the final vapor exposure session using cell-attached and whole-cell patch-clamp configurations to measure spontaneous and evoked firing, respectively. Current clamp recordings measured CIE-induced changes in intrinsic excitability using 500 ms current steps (−100 to +220 pA in 20 pA increments) while neurons were held at −70 mV. Voltage clamp recordings assessed spontaneous synaptic activity, with cells clamped at −55 mV to record spontaneous excitatory postsynaptic currents (sEPSCs) and at +10 mV to record spontaneous inhibitory postsynaptic currents (sIPSCs). Excitatory/inhibitory (E/I) balance was quantified using frequency ratios and synaptic drive calculations.

Results: Intrinsic excitability was significantly reduced in the LHb of withdrawn rats compared to controls (*p < 0.05), while spontaneous excitatory and inhibitory currents remained unchanged. Temporal firing pattern analysis revealed highly dynamic transitions within a single LHb neuron, with increased firing pattern heterogeneity observed in CIE-exposed rats compared to controls. Ongoing analyses are examining the potential for distinct effects of CIE exposure on specific firing patterns.

Conclusions: Acute withdrawal is associated with reduced intrinsic neuronal excitability in the LHb in combination with increased firing pattern heterogeneity. These neuroadaptations may contribute to the regulation of withdrawal symptoms and provide insight into the neural mechanisms underlying vulnerability to relapse in individuals with AUD. Supported by NIH grants R01 AA029130, R01 AA031003, R21 AA030867, P50 AA022538, and T32 AA02657.

Keywords: Lateral Habenula, chronic alcohol, alcohol withdrawal

Disclosure: Nothing to disclose.

P371. Pacap and PAC1 receptors facilitate opioid induced hyperalgesia

Amynah Pradhan, Elizaveta Mangutov, Akhilesh Kotiyal, Yaseen Awad Igbaria

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

Background: Withdrawal and prolonged abstinence from chronic opioid use produces a negative affective state, which includes hypersensitivity to pain. Chronic opioid exposure and withdrawal can result in opioid induced hyperalgesia (OIH), which is observed in individuals experiencing opioid use disorder (OUD), maintained on methadone or buprenorphine, and in chronic pain patients. This hypersensitivity to pain can contribute to opioid escalation, efforts to seek treatment, and relapse. Using an unbiased proteomic screen, our lab has identified pituitary adenylate cyclase activating polypeptide (PACAP) as a promising target for OIH. PACAP binds to the excitatory G protein coupled receptor, PAC1, and previous studies from our lab have demonstrated the effectiveness of a peptide PAC1 antagonist in preclinical models of OIH. An antibody targeting PACAP recently showed clinical benefit in a Phase II clinical trial for the treatment of migraine. In addition, novel small molecule PAC1 antagonists have also been recently developed. One aim of this study was to test PACAP targeting antibody and novel small molecule PAC1 antagonists in models of OIH. Our lab has also identified the ventrolateral periaqueductal grey (vlPAG) as a key site of PACAP dysregulation in OIH. A further aim of this study was to determine if modulation of PAC1 receptor expressing cells in this region is critical for the development of OIH.

Methods: All experimental protocols in animal studies were approved by the WashU Animal Care and Use Committees and were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. C57Bl6/J and PAC1-Cre mice were used in these studies. PAC1 antagonists, PA-915 and BAY2686013, and a PACAP targeting antibody were tested in models of OIH. Muliplex in situ hybridization was used to characterize the distribution of PACAPergic and opioid peptides and receptors in the vlPAG. Stimulating DREADDS were used to selectively activate PAC1 expressing cells, and lentivirus expressing siRNA were used to knockdown PAC1 receptor in the vlPAG and determine effects in OIH.

Results: The PACAP antibody effectively blocked established allodynia in OIH models, and pretreatment also prevented the development of OIH. Similar results were observed with PAC1 antagonist treatment. PAC1 was highly expressed in the vlPAG, with significant co-expression with mu and delta opioid receptors. Chronic chemogenetic activation of PAC1 expressing cells in the vlPAG produced long lasting allodynia, similar to that observed in OIH models. Correspondingly, knockdown of PAC1 receptors in the vlPAG prevented the development of OIH.

Conclusions: Together, our results suggest that chronic opioids produce hypersensitivity through PACAP-PAC1 in the vlPAG, and that PACAP or PAC1 receptor targeting therapies are an effective therapeutic target for OIH.

Keywords: Opioid abuse, Opioid-sparing effects, PACAP, opioid dependence, pharmacotherapy, animal model, withdrawal, Non-opioid, Non-addictive Therapeutics

Disclosure: Lundbeck, Contracted Research, Self

P372. Lynx2 allosteric modulation of nicotinic receptors impacts alcohol drinking behavior in mice

Julia Lagomarsino, James Fowler, Joshuah Arellano, Gaby Isabella Almaraz-Roman, Jibran Khokhar, Christie Fowler

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

Background: Dysregulated alcohol use leads to millions of deaths each year and has been an associated risk factor diseases and injury conditions worldwide. Thus, defining and targeting the mechanisms that mediate alcohol’s effects in driving consummatory behaviors are essential to understanding alcohol use disorder (AUD) in humans. Given the complex nature of alcohol’s actions in the brain and variations in use patterns with different patient characteristics, there is a critical need for novel basic science approaches that will inform the most effective treatments for various clinical populations. Alcohol has been shown to directly act on the cholinergic system in the brain via the nicotinic acetylcholine receptors (nAChRs). Interesting, the lynx2 protoxin is an endogenously expressed protein that has been shown to directly associate with nAChRs. Lynx2 directly and selectively interacts with specific subtypes of nAChRs and serves as a ‘molecular brake pad’ to limit receptor function. We have previously shown that lynx2 is expressed in a distinct brain region- and cell-specific manner, including regions associated with alcohol’s actions, cognition, and social behavior.

Methods: In these studies, male and female mice lacking lynx2 (lynx2 knockout mice) were examined for differences in alcohol drinking behavior, as compared to wildtype littermates, with the two-bottle choice drinking paradigm. The studies included analysis of both single housed and group housed drinking conditions. As a control, lynx2 knockout mice and wildtype littermates were also examined in the elevated plus maze and with sucrose drinking behavior.

Results: Lynx2 knockout mice exhibited a decreased preference to consume alcohol in the two-bottle choice paradigm (p < 0.01). Interestingly, this effect was mitigated due to social housing condition (p < 0.01), in which group housed subjects increased the total amount of daily alcohol consumed. These effects could not be attributed to phenotypic differences in drinking behavior, since no differences between genotypes were found with sucrose consumption. While lynx2 knockout mice exhibit increased anxiety-related behavior at baseline levels (p < 0.05), the genotype groups did not differ in their behavior on the elevated plus maze during daily alcohol access.

Conclusions: Taken together, these preliminary data provide evidence for lynx2’s modulation of nAChRs in regulating alcohol consumption.

Keywords: Alcohol, nicotinic acetylcholine receptors, negative allosteric modulator

Disclosure: GATC Health, Contracted Research, Self, GATC Health, Advisory Board, Self

P373. Repeated fentanyl abstinence intensifies opioid withdrawal and induces a proinflammatory state in striatal microglia

David Bergkamp, Kevin Coffey, Aliyah Dawkins, John Neumaier

University of Washington, Puget Sound VA Health Care System, Seattle, Washington, United States

Background: Opioid withdrawal is a serious obstacle to abstinence for individuals that want to reduce their consumption or stop using altogether. Anecdotal evidence suggests that previous experiences of opioid withdrawal may exacerbate the severity of withdrawal signs upon opioid discontinuation.

Methods: Male and female mice were treated with a single or five repeated cycles of daily, escalating fentanyl injections followed by abstinence; control mice were treated with repeated saline injections on the same schedule. Withdrawal-associated behaviors were assessed with tail flick (hyperalgesia), open field (anxiety-like behavior), and sucrose preference (anhedonia) tests. Microglia morphology was assessed histochemically. The RiboTag method was used to perform deep sequencing of the microglial translatome and was compared to the whole striatal transcriptome in each mouse. RNA was sequenced from 30 mice (n = 4–10/group). Key changes were confirmed by RTqPCR of RiboTag RNA.

Results: Repeated bouts of fentanyl treatment and withdrawal impacted striatal microglia much more than a single cycle of fentanyl followed by withdrawal. Multiple withdrawal cycles reduced ramification of microglial processes, suggesting a more reactive cell state. Five cycles of fentanyl withdrawal induced more severe behavioral withdrawal signs in mice. While there were limited effects of fentanyl withdrawal on the microglial translatome after a single cycle of withdrawal, five cycles of fentanyl and withdrawal led to dramatic changes in RNAs actively undergoing translation. Five cycles of fentanyl withdrawal increased the expression of gene networks associated with microglial homeostasis, disease associated, and an array of cytokine signaling related genes. Indeed, 100% of the genes associated with the “microglia core sensome”, genes associated with sensing extracellular perturbations, were upregulated after five cycles of withdrawal.

Conclusions: Mouse striatal microglia initiate an inflammatory response following five, but not one opioid withdrawal experience and suggest that drug therapies targeting microglial innate immune signaling systems may mitigate the severe withdrawal associated with repeated exposures to opioid tolerance and withdrawal.

Keywords: Fentanyl withdrawal, microglia, RNAseq, striatum, opioid

Disclosure: Nothing to disclose.

P374. Behavioral signatures of hyperkatifeia: mapping opioid withdrawal in genetically diverse mice

Lucy Ward, Adriana Gregory-Flores, Leandro Vendruscolo, Hugo Tejeda, George Koob, Renata Marchette

Neurobiology of Addiction Section, NIDA IRP, Baltimore, Maryland, United States

Background: The ongoing opioid overdose crisis underscores the urgent need to define the neurobiological mechanisms driving opioid misuse and withdrawal. Chronic opioid exposure produces a negative emotional state during withdrawal, termed hyperkatifeia, which encompasses dysphoria, anxiety, irritability, sleep disturbances, and heightened physical and emotional pain. Hyperkatifeia is hypothesized to arise from dysregulated opponent processes within brain reward and stress circuits. Here, we aimed to systematically characterize the behavioral features of hyperkatifeia in both opioid-dependent and opioid-vulnerable mice, using a genetically diverse population to enhance translational relevance.

Methods: Sixty-four male and female Diversity Outbred (J:DO; Jackson Laboratories) mice were assessed for baseline nociception and anxiety. Mice were randomized to saline (n = ~21) or chronic fentanyl administration (0.12–4 mg/kg, subcutaneous, twice daily; n = ~43). Following fentanyl treatment, acute withdrawal was evaluated through nociception and anxiety measures. After three weeks of withdrawal, we re-assessed nociception, anxiety, anhedonia, and irritability. At the conclusion of behavioral testing, serum, spleen, adrenal glands, and brain tissue were collected for proteomic analyses.

Results: Chronic fentanyl exposure induced hyperalgesia during acute withdrawal, evident in cold sensitivity (F1,58 = 3.183; p = 0.079) and significant mechanical sensitivity (F1,54 = 6.916; p = 0.011). Acute withdrawal was also associated with anxiogenic-like behavior (F1,60 = 3.922; p = 0.052). During protracted withdrawal, mice displayed heightened irritability (F1,54 = 4.694; p = 0.035). No significant differences emerged in anhedonia or startle response, and no sex-dependent effects were detected.

Conclusions: These findings provide the first systematic behavioral characterization of hyperkatifeia in a genetically diverse opioid dependence model. The results highlight distinct withdrawal-associated behavioral domains—acute hyperalgesia, anxiety, and protracted irritability—that may serve as quantifiable dimensions of hyperkatifeia. Ongoing work will generate a composite hyperkatifeia score and integrate it with proteomic profiling to identify candidate biomarkers of withdrawal severity. This translational approach may inform the development of targeted therapies to mitigate withdrawal symptoms and reduce relapse risk in opioid use disorder.

Keywords: Addiction-like behaviors, Addiction phenotypes, Fentanyl withdrawal, Outbred mice, hyperalgesia

Disclosure: Nothing to disclose.

P375. Dynamic mitochondrial disruptions within the dorsal striatum in relation to opioid exposure

Kion Winston, Konrad Dabrowski, Alexandra Chisholm, Yasmin Hurd

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

Background: Opioid misuse has continued to be a major public health crisis within the United States, contributing to the growing overdose deaths of ~100,000 people yearly. While current treatments available for opioid use disorder (OUD) have primarily focused on mitigating the maladaptive behavior associated with reward and craving, other complex phenotypes of OUD remain poorly understood. Recent studies have linked long-term heroin exposure to neurodegenerative disease-like changes including the accelerated proliferation of hyperphosphorylated Tau, the abnormal state of the tau protein that is a hallmark feature of neurodegenerative diseases. Potential mechanisms that might facilitate the relationship between opioid use and neurodegenerative disease risk remain unclear. Here, we focused on the dorsal striatum a brain region central to habitual behavior and in which Tau pathology was previously identified in human heroin users.

Methods: We conducted bulk RNA-sequencing on RNA extracted from the putamen of 16 postmortem human heroin users and 14 age-matched controls (n = 30). Functional ontology analysis was performed using the Enrichr database and a False Discovery Rate (FDR) of adjusted-q < 0.05. Ingenuity Pathway Analysis (IPA) was used to identify canonical biological pathways significantly changed in heroin as compared to control subjects. Subsequently, we studied male Long Evans rats (n = 40) that underwent a heroin self-administration paradigm for 15 days. Brains were obtained at 1hr and 24hr following the last drug exposure. Subregions of the dorsal striatum (posterior dorsal medial striatum(pDMS), posterior dorsal lateral striatum (pDLS), anterior dorsal medial striatum (aDLS), and anterior dorsal lateral striatum (aDLS)) were collected and RNA sequencing conducted. An in vitro model was also leveraged to investigate the causative role of heroin on molecular changes identified in the rat heroin self-administration and human conditions. Primary rat striatal cultures and striatal-cortical co-cultures were prepared from Long Evans embryos. Cells were treated with 10 μM and 40 μM concentrations of morphine (from DIV8 to DIV14). qPCR was used to measure the expression of genes of interest in the cell culture model.

Results: Bulk RNA Sequencing of the dorsal striatum of postmortem heroin users identified 2347 significant (p-value < 0.05) differential expressed genes (DEGs) as compared to controls. Functional ontology analysis of the postmortem human heroin users identified significant (p-value < 0.05) alterations in multiple neurodegenerative ontologies which included Alzheimer, Huntington, and Parkinsonism. The top canonical pathways identified through IPA included electron transport (p < 0.0001), oxidative phosphorylation (p < 0.0001), and mitochondrial dysfunction (p < 0.0001). A tertiary analysis, bolstering the MitoCarta 3.0 database, further substantiated significant (FDR < 0.05) alternations of the OXPHOS subunits and mitochondrial function. The rodent self-administration model, confirmed similar changes in nuclear-encoded mitochondrial genes in relation to heroin use. In addition, the animal model demonstrated unique molecular phenotypes with significant upregulation of transcription factor A mitochondria (Tfam) (p-value = 0.033) 24hr, compared to 1hr, after the last heroin self-administration session. Finally, primary cell culture experiments identified a significant (p-value < 0.05) downregulation of the nuclear encoded mitochondrial genes Tfam and Ndufb6 in morphine treated striatal cultures (Cohen’s d: 1.693, 3.529). However, there was a significant upregulation of the same genes in the striatal-cortical co-culture (Cohen’s d: 2.772, 0.951). Further studies are being completed to understand the time course of mitochondrial alterations in the rat and in vitro models.

Conclusions: The transcriptome signature within the dorsal striatum of human heroin users suggests significant mitochondrial perturbation as potentially underlying the neurodegenerative-like phenotype previously established in earlier studies of the human brain. Both the animal and cell culture models causally confirm opioid-related mitochondrial perturbations in striatal neurons. Overall, the results continue to raise concern regarding the link between opioid exposure and subsequent neurodegenerative risk.

Keywords: neurodegneration, opioid dependence, mitochondrial dysfunction

Disclosure: Nothing to disclose.

P376. AI-guided development identifies a novel serotonin receptor modulator for opioid use disorder

Valeria Lallai, Samuel Samuel Kho, Alexa Martin, James Fowler, Madison Roach, Kevin Wang, Mina Palaniappan, Malia Bautista, Allison Mogul, Jinjutha Cheepluesak, Vaishnavi Narayan, Jayson Uffens, Waldemar Lernhardt, Saman Mirzaei, Ian Jenkins, Arturo Zavala, Jonathan Lakey, Robert Tinder, Christie Fowler

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

Background: The opioid epidemic continues to cause devastating morbidity and mortality, with fentanyl driving unprecedented overdose rates. Despite existing treatments, relapse remains high, highlighting the urgent need for novel therapeutics. Here, we applied an innovative artificial intelligence (AI)-based drug discovery platform to identify molecular signatures of dysregulation in human postmortem brain tissue from individuals with opioid use disorder (OUD) and to generate preclinical therapeutic candidates.

Methods: The AI platform predicted serotonin receptor-targeting compounds, of which GATC-021 and GATC-1021 were synthesized and validated for receptor selectivity in vitro. In vivo efficacy was tested in male and female rats (n = 12–13/group) trained to self-administer fentanyl. Compounds were administered, and effects on intake, locomotor behavior, gene expression, and hippocampal dendritic spine morphology were assessed. Statistical significance was determined using repeated measures two-way ANOVA. All animal studies were approved by the Institutional Animal Care and Use Committee (IACUC) and conducted in accordance with NIH guidelines.

Results: GATC-1021 robustly reduced fentanyl self-administration across multiple doses (25, 40, and 70 mg/kg, p < 0.0001). GATC-021 produced more modest effects at higher doses (50 mg/kg, p < 0.01). Neither compound altered locomotor activity, indicating minimal off-target effects. At the molecular level, GATC-1021 normalized fentanyl-induced transcriptional changes in addiction-related brain regions, increasing Bdnf and Camk2a expression in the prefrontal cortex (p < 0.05 to p < 0.01). Structural analyses revealed altered hippocampal dendritic spine profiles characterized by an increase in thin spines and a decrease in stubby spines (p < 0.05 to p < 0.01), consistent with synaptic remodeling.

Conclusions: These preclinical findings highlight the potential of AI-driven drug discovery to rapidly generate novel candidates for OUD. GATC-1021 reduced fentanyl intake without adverse behavioral effects and engaged molecular and structural markers of plasticity associated with opioid exposure. Given the urgency of the opioid crisis and the novelty of this AI-guided approach, these results represent a highly innovative preclinical direction with translational potential, supporting further development and eventual clinical evaluation.

Keywords: opioid use disorder, Fentanyl Self-Administration, Artificial Intelligence, Analysis, NeuroScience, Clinical NeuroScience, Pharmacology, serotonin receptors, Preclinical assessments

Disclosure: Nothing to disclose.

P377. Effects of N-desethyl-fluornitrazene (D-FNZ) in squirrel monkeys

Alex Myers, Maria Zarcone, Juan Gomez, Agnieszka Sulima, Kenner Rice, Michael Michaelides, Hank Jedema

National Institute on Drug Abuse Intramural Research Program/NIH, Baltimore, Maryland, United States

Background: The novel µ-opioid agonist N-desethyl-fluornitrazene (DFNZ) has been reported to have high intrinsic activity and a unique pharmacological profile in rodents. The effect of DFNZ has not been studied in non human primates.

Methods: We examined the effect of subchronic pretreatment of multiple doses of DFNZ (30 and 100µg/kg) on economic choice between natural reward (sweetened condensed milk) and intravenous administration of the ultrashort acting opioid remifentanil in male and female squirrel monkeys performing a touch screen-based task.

Results: Consistent with the effect of other full agonists such as morphine or methadone, choice allocation between drug and milk reward was not affected by pre-treatment with DFNZ. Choice allocation remained consistent throughout the session. The 100µg/kg dose of DFNZ reduced overall progress on the task, primarily by increasing omissions and slight reduction in touch accuracy.

Conclusions: Pre-treatment with DFNZ did not affect choice between a natural reward and intravenous remifentanil. Ongoing experiments will determine whether the squirrel monkeys choose to self-administer DFNZ.

Keywords: opioids, non-human primates, behavioral economics

Disclosure: Nothing to disclose.

P378. Punishment sensitivity and reinstatement propensity represent dissociable phenotypes in mouse oxycodone self-administration

Titilope Akinwe, Lisa Fang, Michelle Lynch, Mason Barrett, Lite Yang, Vijay Samineni, Joseph Dougherty, Meaghan Creed

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

Background: Approximately 9 million Americans misuse prescription opioids annually. While some can consume opioids such as oxycodone (OXY) as prescribed, other individuals transition to opioid use disorder. Similarly, rodents escalate their oral oxycodone self-administration (OXY-SA), but only a subset (20–30%) continue drug taking when punishment is introduced (punishment-resistant; PR), while the majority discontinue OXY-SA (punishment-sensitive; PS). Even once abstinence is achieved, individual variation in relapse vulnerability remains poorly understood, despite its critical importance for treatment outcomes. The “incubation of craving” phenomenon may explain this variable relapse vulnerability. While some individuals successfully overcome physical dependence during acute withdrawal, many continue to experience opioid craving, which persists during protracted abstinence and drives opioid relapse. Despite the importance of these behavioral phenotypes, no studies have simultaneously examined individual variability in continued opioid use despite adverse consequences and in propensity to relapse. We hypothesized that mice exhibiting PR OXY intake would also exhibit increased reinstatement following abstinence compared to PS mice.

Methods: C57bl/j mice (n = [CTRL]: 5F/5M, n = [OXY]: 29F/30M) from Jackson Laboratory were group housed with ad libitum access to chow and drinking water in vivarium housing before the start of OXY-SA. The animals were moved to the testing room and single-housed in static cages equipped with an in-cage lickometer apparatus (i.e., sipper) and chow throughout the habituation and OXY-SA phases. In the acquisition OXY-SA phase, mice had ad libitum access to increasing OXY concentrations (0.1, 0.3, 0.5 mg/ml oxycodone hydrochloride) for ‘OXY’ mice or unadulterated drinking water for ‘CTRL’ mice within a single sipper bottle for five days, after which they could choose between the 1.0 mg/ml OXY bottle or their regular water bottle for seven days in the baseline OXY-SA phase. Drinking water was supplied in both bottles for CTRL mice. Mice then underwent three days of punished OXY-SA, where OXY drinking probabilistically triggered 100-ms, 15-PSI air puffs to the spout on 30% of trials based on photobeam breaks. We (1) classified the PR phenotype based on OXY intake during all punished days (> 50% of baseline oxy dose), and we (2) evaluated whether PR mice exhibit enhanced cue-induced reinstatement using sipper counts during a 60-minute probe trial, following protracted withdrawal. Escalation index was calculated as the body weight-normalized OXY intake on Day 12 (the final day of the two-bottle choice phase) divided by the body weight-normalized OXY intake on Day 6, the first day of the two-bottle choice phase (DoseDay12/DoseDay12). For each measure, outliers with values 2*IQR above the 75th percentile or below the 25th percentile were excluded.

Results: Punished OXY intake reveal significant group and sex effects, with PR mice having a significantly higher OXY intake than PS mice (PR = 7F/8M: 111.3 ± 29.84 PS = 23F/22M: 20.04 ± 12.51, p = 0.0081) and female PR mice having a higher OXY dose (PR female = 170.8 ± 20.98, PR male = 94.17 ± 13.90, p = 0.0062). There was no significant group effect on reinstatement sipper counts, indicating that punishment sensitivity and reinstatement are dissociable (PR = 7F/7M: 7.9 ± 11.3, PS = 15F/18M: 8.9  ± 13.4, p = 0.7784). There was a significant effect of sex on sipper count, with female PS mice having a higher sipper count than CTRL mice (p < 0.05). PR mice had a higher escalation index (PR = 2.230 ± 1.994, PS = 1.072 ± 0.7557, p = 0.0025), with PR females driving most of this difference (PR female = 3.199 ± 2.344, PS female = 0.9384 ± 0.6785, p = 0.0004).

Conclusions: Our results support a behavioral model in which an individual who undergoes voluntary withdrawal due to adverse consequences is still susceptible to relapse as cravings occur. We complemented these findings with recent single-cell RNA-seq analysis in the ventral pallidum of C57bl/j mice, an important node for reward-seeking and cue-induced reinstatement. Isogenic mouse models enable us to focus on molecular substrates beyond mutational differences within the DNA sequence, allowing us to evaluate cellular and epigenetic factors that might regulate gene expression in response to drug exposure. This approach is particularly valuable for identifying factors that distinguish PS and PR mice in their OXY-seeking behaviors following protracted withdrawal and cue-induced reinstatement.

Keywords: Single-cell RNA sequencing, opioid use disorder, Reward seeking, reinstatement, Addiction phenotypes

Disclosure: Nothing to disclose.

P379. Prenatal circadian rhythm disruption induces sex-specific substance use and mood-related phenotypes in mice

Nilanjana Saferin, Isabella Seibert, Taylor Stowe, Colleen McClung, Lauren DePoy

University of Toledo, toledo, Ohio, United States

Background: 20% of Americans are at risk for environmental circadian rhythm disruptions (CRD) due to shift work. Shift workers experience substantial negative health outcomes, but females are especially affected with greater vulnerability for substance use (SU) and adverse outcomes associated with pregnancy. These outcomes not only occur during pregnancy, but offspring are affected at birth and later in life. In mice, prenatal CRD (pCRD) recapitulates these risks, increasing adverse pregnancy outcomes and altering behavior in adult offspring. However, it is unknown whether pCRD affects SU in mature offspring.

Methods: C57Bl/6J dams were sham handled or disrupted, by reversing the light/dark cycle 4 times during gestation. Reward- and mood-related behaviors were measured in adult offspring. Cocaine reward was measured using conditioned place preference (5 mg/kg cocaine). Contingency degradation was used to measure decision making. Mice were trained to respond on two levers for food, then the likelihood that one of those levers will be reinforced was degraded. Another cohort was trained to respond for food before jugular catheterization. After recovery, mice were trained to respond on a different lever for cocaine. Acquisition, the reinforcing and motivational properties of cocaine, extinction and cue-induced reinstatement were measured. A separate cohort of mice were tested in the forced swim test, light/dark box, open field and elevated plus maze. Multiple behavioral tests were also performed with cross-fostered mice, in order to examine whether behavioral effects were due to gestational manipulations or postnatal maternal care. Throughout, 2-way ANOVAs were used for data analysis (disruption group by sex), with additional repeated measures as appropriate (i.e. session, lever, dose). Significant interactions were followed up with Sidak posthoc tests. p < 0.05 was considered significant and p < 0.1 trending. Samples sizes of 6–14, depending on the experiment. In order to investigate possible underlying mechanisms, we investigated how gene expression rhythms (RNAsequencing) change in the nucleus accumbens of adult mice with a history of pCRD. Tissue was collected at 6 times of day (ZT2/6/10/14/18/22)(zeitgeber time-time from lights onset) in male and female mice. Procedures were approved by the University of Pittsburgh IACUC.

Results: Interestingly, females exposed to pCRD developed an anhedonic-like phenotype with decreased food self-administration (interaction, p = 0.007), cocaine intake (disruption, p = 0.008) and reinforcing properties of cocaine (interaction, p = 0.01). On the other hand, pCRD males showed a SU-like phenotype with increased cocaine preference (interaction, p = 0.08), higher order food self-administration (disruption, p = 0.099) and cocaine reinforcement (interaction, p = 0.08). Furthermore, while male pCRD mice maintained goal-directed decision making, responding more on a reinforced lever, female pCRD mice did not, indicating habit formation. Together, these results suggest that male and female mice exposed to pCRD respond differently for rewarding outcomes. In order to determine whether these divergent behavioral outcomes are unique to reward I next measured anxiety-like behavior. As expected, results in anxiety- and mood-related behavioral tasks (interaction, p = 0.005) paralleled reward-related behavior. Anhedonic-like female pCRD mice showed increased anxiety-like behavior and pCRD males showed decreased anxiety/increased risk-taking behavior, as well as decreased immobility in the forced swim test. Behavioral outcomes were consistent after cross-fostering, suggesting gestational disruptions and not maternal care contribute to the behavioral changes observed after pCRD. Gene expression rhythms were changed in the nucleus accumbens after pCRD. We found many rhythmic transcripts, about 2,000 in control mice and 500 in shifted mice indicating a large decrease in rhythmicity after pCRD, particularly in male mice. We found several overlapping rhythmic processes in our samples, although the individual pathways enriched in rhythmic transcripts were typically unique. Circadian rhythm pathways were significantly enriched in all groups. Actin monomer binding and extracellular matrix organization were also commonly identified as significantly rhythmic, particularly in male control and female pCRD mice. The most common significantly rhythmic processes identified among all control mice involved GPCR signaling. We also found unique pathways and processes, including ion channel activity and transport, as well as synaptic transmission in male control samples, methylation, phosphatase activity and cell cycle pathways in male pCRD, e-box binding, hormone binding and stimulus pathways in female control and general kinase activity and cell adhesion pathways in female pCRD samples.

Conclusions: These results suggest that pCRD may predispose individuals to distinct psychiatric disorders based on sex, mood disorders in females and SU disorders in males. By understanding how disrupted rhythms during pregnancy affect behavior in adulthood, we can develop novel therapeutic approaches for SU and mood disorders in adults.

Keywords: circadian, sex differences, substance use

Disclosure: Nothing to disclose.

P380. Alcohol disrupts behavioral flexibility through a discrete claustrocortical circuit

Andreas Wulff, Samuel Sheats, Eliza Douglass, Brian Mathur

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

Background: Cognitive impairment is a major component of Alcohol Use Disorder (AUD). Optimal cognitive performance requires the claustrum, a subcortical nucleus that orchestrates cortical activity. Yet, the impact of chronic alcohol exposure on claustrum function is unknown.

Methods: To test this we employed a chronic intermittent ethanol vapor (vs. air) exposure paradigm in mice of both sexes and assessed physiological and behavioral outcomes using slice electrophysiology and chemogenetic approaches, respectively.

Results: Here, we demonstrate in both sexes that in response to chronic intermittent ethanol vapor burst firing Type 2 claustrum projection neurons exhibited increased responsivity to anterior cingulate cortex (ACC) input, which was mediated by pre- and post-synaptic mechanisms. We found that this increased ACC drive was specific to the output target of these Type 2 neurons: those neurons targeting the parietal association cortex, but not the prefrontal cortex, were affected. Examining the behavioral relevance of this hyperexcitatory drive of the ACC-to-parietal association cortex-projecting neuron pathway, we found that like chronic ethanol exposure itself, chemogenetic activation of this pathway decreased performance on a reversal learning task.

Conclusions: These data open a new node of investigation into chronic alcohol effects on executive function and suggest behavioral inflexibility in alcohol use disorder arises in part from alcohol effects on the claustrum.

Keywords: Cognition, Compulsive behavior, Addiction

Disclosure: Nothing to disclose.

P381. Psychosocial stress-induced cocaine seeking in rats is inversely correlated with activation of medial prefrontal cortical neurons projecting to the rostral periaqueductal gray

Hinds Nicole, Ireneusz Wojtas, Corinne Gallagher, Molly Concannon, Daniel Manvich

Rowan-Virtua School of Osteopathic Medicine, Stratford, New Jersey, United States

Background: Psychosocial stress is well-established to elicit cocaine craving and relapse in humans, yet its use in preclinical relapse models has been limited. Consequently, the neurobiological mechanisms and neural circuits by which psychosocial stressors specifically promote cocaine-seeking behavior are not well understood. Our prior work identified the periaqueductal gray (PAG) as a brain region that may contribute to social stress-induced reinstatement of cocaine seeking. The objective of the present study was to examine which neural region(s) may be driving PAG activity during social stress-induced cocaine seeking.

Methods: Adult male Long-Evans rats (n = 36) received injections of a retrograde viral tracer into the lateral and ventrolateral aspects of the rostral PAG (rPAGl/vl) and were subsequently trained to self-administer cocaine (0.5 mg/kg/infusion, IV) over 20 daily 2-hour sessions. On sessions 11, 14, 17, and 20, a discrete tactile cue was introduced in the operant chamber and was conditioned to signal either impending social defeat stress or footshock stress in separate groups. Following extinction training, rats were re-exposed to their respective stress-predictive cues and tested for cocaine-seeking behavior under extinction conditions. Fos immunohistochemistry was used to examine neural activation in brain regions of interest during the cocaine-seeking episode.

Results: Re-exposure to either a social stress- or footshock-predictive cue elicited reinstatement of cocaine-seeking behavior (p < 0.0001). Social stress-induced cocaine-seeking magnitude was associated with activation of the rPAGl/vl (p < 0.05), and was also negatively correlated with the number of Fos-positive rPAGl/vl-projecting neurons within the prelimbic and anterior cingulate cortices (p < 0.05).

Conclusions: These findings lend further support to a potential role for the rPAGl/vl in stress-induced cocaine seeking and suggest that its contributions to cocaine-seeking behavior may be modulated by top-down frontal cortical input.

Keywords: cocaine self-administration and reinstatement, periaqueductal grey (PAG), psychosocial stress, c-Fos, Medial Prefrontal Cortex (mPFC)

Disclosure: Nothing to disclose.

P382. Transcriptomic signatures of cocaine relapse ensembles: toward drug repurposing for relapse prevention

Katherine Savell, Rajtarun Madangopal, Jennifer Bossert, Ava Holmes, Madeline Sagona, Omodolapo Nurudeen, Owura Berko, Drake Thompson, Olivia Drake, Diana Pham, Megan Brenner, Yavin Shaham, Bruce Hope

IRP/NIDA/NIH, Baltimore, Maryland, United States

Background: Relapse remains a major clinical problem, and there are no approved treatments to reduce the risk of relapse to psychostimulants like cocaine. Environmental cues previously associated with drug-taking can provoke craving and cause relapse long after cessation of drug use. One strategy to combat relapse is to disrupt the persistent, maladaptive drug-cue associative memories by targeting molecular mechanisms that are selectively engaged in neurons active during relapse- relapse ensembles. However, the cell types comprising these ensembles, and the molecular processes by which they maintain these associations, remain understudied.

Methods: To address this gap, I developed a new single nucleus RNA-sequencing method (XPoSE-seq) that enables sample multiplexing and enrichment of rare populations like ensembles underlying specific behaviors. We used Fos-mRFP transgenic rats to genetically tag and isolate neurons activated during relapse. Rats trained to self-administer cocaine were tested for relapse to cocaine seeking after 21 days of abstinence. We collected brains 3 hours after the start of the test, and relapse ensemble neurons (and non-ensemble controls) fron the medial prefrontal cortex (mPFC) were isolated and processed for XPoSE-seq. Cell types were annotated using established mPFC marker genes, and cell-type specific relapse differentially expressed genes (DEGs) were identified. The relapse DEGs were ranked prior to a reverse signature analysis against FDA-approved drug-induced signatures to identify candidates predicted to counteract relapse-associated molecular changes.

Results: Rats exhibited reliable cocaine self-administration during training and robust relapse to cocaine seeking on abstinence day 21. Cocaine relapse ensembles were distributed across multiple excitatory and inhibitory cell types in the mPFC and showed altered expression of genes implicated in memory maintenance, including neuropeptides, receptors, and synaptic plasticity-related molecules. The drug repurposing screen identified 35 FDA-approved candidates predicted to reverse these transcriptional signatures.

Conclusions: This study reveals the molecular diversity of relapse-activated neuronal ensembles and provide a framework for precision-targeted interventions. Ongoing work is testing whether these FDA-approved drugs can reduce relapse behavior in preclinical models. This approach may enable repurposing of existing medications for rapid clinical translation.

Keywords: Drug Relapse, Single-cell transcriptomics, drug repurposing

Disclosure: Nothing to disclose.

P383. Targeting PDGFRβ signaling to mitigate fentanyl reward

Matthew Fanelli, Stephanie Puig

UMass Chan Medical School, Worcester, Massachusetts, United States

Background: Opioids remain the gold standard for the treatment of chronic severe pain. Unfortunately, their high addiction liability leads to OUD and has fueled the current opioid epidemic. Selectively targeting pathways downstream of the mu-opioid receptor (MOR, target of clinically used opioids) to combat side effects without altering analgesia remains an ongoing challenge. Historically, opioid side-effects, including tolerance and reward, were thought to occur via direct modulation of MOR signaling. However, our previous work showed that opioid activation of MOR can transactivate platelet-derived growth factor receptor beta (PDGFRβ), a receptor tyrosine kinase (RTK), via platelet-derived growth factor B (PDGF-B) release. Importantly, co-administration of opioids with imatinib, an FDA approved PDGFRb inhibitor, completely blocked opioid tolerance, without altering analgesia. We examined published single cell RNA-seq mouse brain data and found that, along with MOR, PDGFRβ is enriched in the striatum, particularly within GABAergic neurons – a key brain region involved opioid seeking and craving. Therefore, in the current pilot study, we aimed to determine if PDGFRβ could also be targeted to block opioid reward, a side-effect that promotes OUD.

Methods: Adult female and male C57BL/6J mice (8–10 weeks, Jackson Laboratory) were used. We first conducted a preliminary neuroanatomical study, where mice (N = 2/sex) received saline or fentanyl daily injections (0.5 mg/kg, i.p.). Brains were collected and processed for in situ hybridization (RNAscope, ACDBio) in the striatum, a pivotal structure for reward, to: 1) determine baseline Pdgfb, Pdgfrβ, and Oprm1 mRNAs co-expression in D1R vs. D2R medium spiny neurons (MSNs); 2) examine the impact of fentanyl on expression and distribution theses markers. QuPath software was used for cellular segmentation to analyze images acquired with a wide-field fluorescence microscope (BZ-8000, Keyence).

Next, to test if PDGFRβ signaling could be involved in opioid reward, we used a mice conditioning place preference (CPP) assay. In these pilot experiments, mice were conditioned to receive saline or fentanyl in a conditioned environment (N = 4/sex). Imatinib (PDGFRβ inhibitor) was administered on the last day prior to testing preference. CPP behavioral data was analyzed using one-way ANOVA (GraphPad Prism). P < 0,05 was considered significant.

Results: RNAscope study showed that Pdgfrβ and Pdgfb mRNAs are abundantly expressed in sub-regions of the striatum (dorsal striatum (dStr), Nucleus accumbens core (NAc Core) and shell (NAc Shell). Pdgfb and Pdgfrβ were found to be co-expressed with Oprm1 in both D1R and D2R positive MSNs. Interestingly, we found that within D1R+ MSNs, the subpopulation of neurons co-expressing Pdgfb and Oprm1 was increased after 5 days of fentanyl repeated injection. This was not observed in D2R+ MSNs, suggesting a specific effect on Pdgfb+Oprm1 + D1R+ MSNs.

In the CPP assay, imatinib completely blocked fentanyl place preference in males when administered right prior testing, suggesting that PDGFRβ inhibition can block the expression of fentanyl reward. Importantly, this effect was not observed in females, suggesting a sex-specific effect in males only.

Conclusions: We discovered that imatinib, a PDGFRβ inhibitor, can effectively block the expression of fentanyl reward in male but not female mice. Our preliminary neuroanatomical study suggests that the specific population of D1R neurons that co-express MOR and PDGF-B may be involved in the mechanisms of PDGFRβ signaling-mediated fentanyl reward.

These exciting preliminary findings, suggest that PDGFRβ could be targeted to prevent fentanyl reward in males. Future studies are underway to understand the sex-specific effects of these findings and to identify precise brain circuits underlying PDGFRβ-mediated expression of reward.

Keywords: Fentanyl, Platelet derived growth factor, Conditioning Place Preference

Disclosure: Nothing to disclose.

P384. Metabolic transcriptomic consequences of cocaine self-administration in nucleus accumbens circuits

Cali Calarco, Isaiah Williamson, Symphanie Key, Brian Herb, Gautam Kumar, Christina Barrett, Seth Ament, Mary Kay Lobo

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

Background: Exposure to illicit drugs and subsequent chronic use profoundly impacts behavior, neuronal structure and firing, and gene expression in reward-related brain regions. Some of these changes are mediated by altered cellular energy homeostasis and mitochondrial function. Cocaine IV self-administration (IVSA) significantly reduces mitochondrial size in nucleus accumbens (NAc) neurons, and disruption of this process is sufficient to blunt cocaine seeking. Cocaine also has widespread impacts on transcription across the brain, including influencing mitochondrial-related genes, which may influence downstream changes in mitochondrial morphology and functioning. Understanding the cocaine-induced transcriptional changes across reward-related brain regions will improve our understanding of drug-induced plasticity. In this study we examined both bulk and circuit-specific mRNA sequencing data allowing us to examine both unique circuit markers as well as cocaine-induced transcriptional profiles as well as mitochondrial respiratory capacity across brain regions.

Methods: We examined bulk and circuit-specific transcriptional changes relating to cellular metabolism and mitochondria, concentrating on reward-related brain regions and inputs to NAc after cocaine IVSA C57Bl/6 mice. Using differentially expressed genes (DEG) identified in previously published bulk RNA sequencing data sets from mice that had undergone either cocaine IVSA or received acute cocaine (5–8 mice per group) we performed subsequent gene ontology analysis, with a focus on mitochondrial-related ontology terms. The sequencing data included tissue from the prefrontal cortex, NAc, dorsal striatum, ventral pallidum, amygdala, hippocampus, and ventral tegmental area. In a further circuit-specific analysis, we conducted ribotag-based labeling, isolation, and sequencing of mRNA from neurons in the prefrontal cortex, ventral hippocampus, and the ventral tegmental area that project into NAc as well as the input (bulk tissue) fraction of these samples after cocaine IVSA in male and female mice (n = 4 per group/sex). Similar DEG and gene ontology analyses we performed in addition to circuit-selective gene enrichment analysis. Finally, mitochondrial respiration from multiple brain regions was measured (n = 6/group).

Results: In bulk tissue samples we found significant representation of DEGs (corrected p < 0.05) in metabolism and mitochondrial-related ontology terms, with regional and cocaine exposure-related variability. Further, predictive analysis of transcription factors regulating mitochondrial-related genes identified multiple transcription factors that may control cocaine-related changes in metabolic function. Ongoing analysis is examining expression of metabolism-related genes (DEG FDR < 0.05), predicted transcription factors, and projection-specific characterization in the circuit-specific transcripts. Cocaine IVSA impacted respiration in a brain region-dependent manner (p < 0.05).

Conclusions: Together these data define the landscape of metabolism-related transcriptomic changes across reward regions in response to cocaine and provide the first circuit-selective transcriptomic characterization of NAc inputs after cocaine IVSA. Understanding circuit-specific transcriptional changes will inform how cellular metabolism supports responses to cocaine throughout reward circuits.

Keywords: cocaine self-administration, Transcriptomics, substance use disorders, Cellular metabolism, Mitochondria

Disclosure: Nothing to disclose.

P385. From first press to binge: genetic modulation of oxycodone self-administration trajectories

Ryan Bachtell, Caleb Hodges, Eamonn Duffy, Cove Andrews, Luanne Hale, Jonathan Ward, Kyle Brown, Laura Saba, Marissa Ehringer

The University of Colorado Boulder, Boulder, Colorado, United States

Background: Prescription opioids are considered the gold standard in analgesic pain medication, yet consumption of opioid drugs, such as oxycodone, can lead to problematic drug use and the development of opioid use disorder (OUD). Twin and family studies indicate that up to 50% of OUD vulnerability is due to genetic factors, yet relatively little work has systematically assessed the genetic factors that contribute to different stages of OUD development. Here, we use a rodent genetic model to more precisely measure specific phenotypes associated with the initiation of oxycodone intake and the development of compulsive-like escalation of oxycodone use in a self-administration (SA) model.

Methods: Adult male (n = 274) and female (n = 266) rats from 21 inbred rat strains from the Hybrid Rat Diversity Panel (HRDP) were assessed for differences in oxycodone or saline SA. Animals were surgically implanted with indwelling jugular catheters before beginning SA. To assess the acquisition of SA, rats were initially trained to lever press for oxycodone or saline on a fixed-ratio 1 (FR1) reinforcement schedule in 10 daily 2-hour (short-access; ShA) sessions. Rats then progressed to the escalation phase with 10 daily 12-hour (long-access; LgA) sessions under the same FR1:TO20 schedule of reinforcement. Rat strains were evaluated for several distinct parameters to quantify acquisition, escalation, and dysregulated use. Acquisition was evaluated using the overall number of infusions, the number sessions to reach acquisition criterion, and the number of sessions to achieve active lever discrimination. Escalation of oxycodone intake was assessed using the overall number of infusions, an escalation index, and the rate of change in intake. Compulsive and dysregulated oxycodone intake was measured by determining the number of episodes of burst responding (3+ infusion/90 sec), burst responding magnitude (infusions per burst), inter-infusion intervals (min), and perseverative responding during the timeout period across the LgA sessions.

Results: Rats in both the saline and oxycodone SA groups met the acquisition criterion within the ten ShA sessions, although acquisition was faster in animals responding for oxycodone compared with saline (F1,486 = 15.71, p < 0.001) and varied depending on rat strain (F20,486 = 2.28, p < 0.002). Similarly, both saline and oxycodone SA rats discriminated the active from the inactive lever (>65%) with the oxycodone SA group displaying faster rates of discrimination (F1,486 = 15.64, p < 0.001), and rat strain also impacted lever discrimination rates (F20,486 = 3.84, p < 0.001). As expected, rats in the oxycodone SA group displayed an escalation in intake compared with the saline SA group (F1,431 = 198.6, p < 0.001), although rat strain also significantly impacted escalation (F19,431 = 2.67, p < 0.001). Interestingly, some strains showed robust escalation (e.g., M520/N, HXB2/lpcv, WKY/NCrl) while other strains remained static throughout the escalation phase (e.g., BXH6/Cub, HXB23/lpcv, and HXB31/lpcvMcwi). Patterns of compulsive and dysregulated intake were evaluated in a subset of strains displaying either robust escalation, overall high oxycodone intake, or both. Our results suggest that strains with high, escalating oxycodone intake engage in burst responding that is akin to binge-like use. The M520/N strain displayed the highest overall number of burst episodes (F4,733) = 65.13, p < 0.001), but the degree of burst responding varied between these strains with the HXB2/lpcv strain showing more infusions per burst (F4,638 = 19.26, p < 0.001) and the M520/N strain having the shortest inter-infusion intervals (F4,729 = 32.36, p < 0.001). Perseverative responding during timeout periods was also detected in most high, escalating intake strains with the BXH6 strain showing the least perseverative responding (F4,733 = 8.69, p < 0.001).

Conclusions: These findings suggest that genetic factors contribute to the acquisition of operant responding and escalation of oxycodone intake in rats. We find that rat strains prone to higher levels of oxycodone intake display varying patterns of intake with some displaying uncontrolled, dysregulated intake patterns while others engage in more controlled intake. These findings suggest that the underlying genetic contributions to drug use may uniquely determine multiple features of opioid use disorders.

Keywords: Addiction phenotypes, Prescription Opioids, Opioid abuse, Genetic variability

Disclosure: Nothing to disclose.

P386. Adolescent cannabis vapor exposure elicits cognitive inflexibility and sex-specific alterations in the excitability of parvalbumin interneurons in the rat medial prefrontal cortex

Sara Westbrook, Riana Abeshima, Matteya Proctor, Leisa Uelese, Sara Burres, Zachary Fisher, Travis Brown, Kristen Delevich, Ryan McLaughlin

Washington State University, Pullman, Washington, United States

Background: Cannabis is the most used drug among adolescents, with a lifetime prevalence of nearly double that of all other illicit drugs combined. This is alarming as the long-term neurobehavioral consequences of adolescent cannabis use remain poorly understood. Recently, we reported that vaporized cannabis self-administration in adolescence led to long-lasting impairments in medial prefrontal cortex (mPFC)-dependent cognitive flexibility. Parvalbumin interneurons (PV) mediate cognitive flexibility as their inhibitory function tightly regulates mPFC output neurons, and PV interneuron function is supported by perineuronal nets (PNNs), which preferentially surround this cell type. Thus, we hypothesized that exposure to vaporized cannabis during adolescence induces long-lasting aberrations in PV function in the mPFC, possibly by altering PNNs, thereby leading to cognitive inflexibility.

Methods: Adolescent male and female Sprague-Dawley rats (n = 8–12/sex/group) received daily non-contingent vaporized cannabis extract (63.9% THC; extract diluted to 150 mg/ml) or vehicle (polyethylene glycol-400) exposure from postnatal day (P) 35–55 (3-s ‘puff’ every 2 min for 60 min). On ~P58, rats received bilateral microinfusions of a PV enhancer virus (AAV.PHP.eb-S5E2-dTom-nlsdTom; 300nl/side) into the prelimbic (PL) subregion of the mPFC. After a two-week washout period encompassing recovery from surgery, cognitive flexibility testing began on ~P70 using an operant-based attentional set-shifting task. After behavioral testing, the brains of littermates were either collected for immunohistochemistry (IHC) or whole cell patch clamp slice electrophysiology to record from viral-mediated fluorescently tagged PV cells in the mPFC. A separate group of rats (n = 4/group) received microinections of AAV9-S5E2-GCaMP6f and fiber optic cannula implants into the PL to record calcium transients in PV interneurons during set-shifting and reversal learning.

Results: Our findings indicate that cannabis-exposed rats of both sexes were impaired in the set shifting but not reversal learning component of the task (p = 0.011), requiring significantly more trials to reach criterion, and had reduced fluorescent intensity of PNNs surrounding PV cells in the mPFC compared to vehicle-exposed rats (p = 0.045). This reduction in PNN intensity was significantly correlated with trials to criterion during the shift component of the task (r2 = 0.326). Interestingly, ex vivo electrophysiology studies revealed that PV cells from cannabis-exposed females were more excitable (increased cumulative spikes and max firing rate, lower rheobase, higher input resistance) than PV cells from vehicle vapor-exposed females, with no significant differences observed in males (p’s < 0.01). Notably, preliminary results from fiber photometry recordings indicate that activity of PV cells from cannabis-exposed rats was also higher following a correct (but not incorrect) response compared to vehicle-exposed rats, suggesting aberrant recruitment of PV interneurons in the PL that could contribute to cognitive flexibility deficits.

Conclusions: These findings support the hypothesis that adolescent vaporized cannabis exposure impairs mPFC-dependent cognitive flexibility in adulthood and that this coincides with increased intrinsic excitability of PV interneurons and alterations in PNNs that surround them. Thus, normalizing PNNs and PV cell function may be promising targets to alleviate adolescent cannabis-induced mPFC dysfunction. Altogether, these findings highlight potential circuit-level targets for mitigating the enduring neurobehavioral consequences of adolescent cannabis use.

Keywords: adolescent cannabis, Cognitive/behavioral flexibility, Vapor, parvalbumin interneurons/perineuronal nets, sex differences

Disclosure: Nothing to disclose.

P387. Essential role of the NPAS4-inducing ensemble in the nucleus accumbens for cue-reinstated reward seeking

Jessica Huebschman, Eleanor Kunitz-Levy, Rachel Clarke, Jacqueline Paniccia, James Otis, Christopher Cowan

Medical University of South Carolina, Charleston, South Carolina, United States

Background: Associations between reward and environmental cues are a fundamental component of relapse vulnerability in substance use disorders (SUDs), though underlying neurobiological mechanisms remain elusive. Sparsely distributed populations of neurons, referred to as “ensembles” and often defined by expression of immediate early genes (IEGs), are activated during reward-related experiences and are thought to encode task-relevant information to facilitate future behavior. Neuronal PAS Domain Protein 4 (Npas4) is one such IEG, and it is selectively activated by synaptic activity via L-type voltage-gated calcium channel signaling. Our lab previously discovered that cocaine conditioning induces Npas4 expression in a sparse subset of neurons (<2%; primarily D1 and D2 dopamine receptor-expressing medium spiny neurons) within the nucleus accumbens (NAc). Chemogenetic inhibition of the NAc NPAS4-inducing ensemble abolished expression of cocaine conditioned place preference (CPP). Therefore, we hypothesized that a similar ensemble encodes reward-cue associations formed during contingent cocaine self-administration and that activity within this ensemble is required for cue-reinstated cocaine seeking.

Methods: We injected AAV-DIO-GiDREADD-mCherry or DIO-mCherry control virus into the NAc of NPAS4-TRAP (Target Recombination in Activated Populations) male and female mice. Animals underwent acquisition, extinction, and reinstatement phases of either liquid sucrose or intravenous cocaine self-administration in a head-restrained configuration. The NAc NPAS4-inducing neurons were “TRAPed” using 4-hydroxytamoxifen (4OHT) injections immediately following the last three days of acquisition. Following extinction training, all mice were injected with CNO (clozapine N-oxide) prior to the first cue test, and after two days in the home cage, were re-tested in a vehicle-only cue test. Following the second cue test, brains were collected for immunohistochemical (IHC)-based analysis of ensemble size and NPAS4 re-induction. Statistical analysis (unpaired t-tests, two- and three-way analyses of variance with repeated measures) were performed using either GraphPad Prism or Jamovi.

Results: Compared to mCherry-only condition, Gi-DREADD-mediated inhibition of the NAc NPAS4+ neurons TRAPed during late acquisition significantly reduced cue-reinstated sucrose and cocaine seeking. However, Gi-DREADD mice showed normal reinstatement behavior during the second “vehicle-only” cue test. Preliminary IHC results suggest that ~8–10% of NAc neurons were TRAPed during the final 3-days of cocaine or sucrose SA. Interestingly, the size of the NPAS4-TRAP population was not correlated with lever-pressing during late acquisition, but it positively correlated with later cue-induced seeking. Further, only ~20% of the NPAS4-TRAP neurons re-induced NPAS4 during the final “vehicle-only” cue test in the control mice, suggesting that only a subset of the NAc NPAS4-TRAP acquisition ensemble neurons are strongly re-activated during cue-induced seeking.

Conclusions: Neurons in the NAc that induce NPAS4-expression during stable, reward self-administration are required for future cue-induced reinstatement of reward-seeking behavior, suggesting that this subpopulation of neurons is preferentially engaged by drug-cue exposure and is essential for both drug and natural-reward seeking. However, it remains unclear why and how the NPAS4+ ensemble is required for the formation and/or stability of reward-cue associations, and how it differs with reinforcer type or from NAc neurons not recruited to the ensemble.

Keywords: cue-induced-reinstatement, Neuronal ensembles, Npas4, Self-Administration, cocaine

Disclosure: Nothing to disclose.

P388. A novel role of the mitochondrial ubiquitin ligase, MITOL, in the orbitofrontal cortex in incubation of oxycodone craving

Xuan Li, Chloe Matheson, Xiang Luo, Megan Burke, Phoebe Cousens, Emily Meadows, Ramesh Chandra

University of Maryland - College Park, Columbia, Maryland, United States

Background: High relapse rates are a major challenge in treating opioid addiction, such as oxycodone, a commonly abused prescription opioid. In rats, oxycodone seeking progressively increases during abstinence, a phenomenon known as incubation of oxycodone craving. We previously found that the orbitofrontal cortex (OFC) plays a causal role in the incubation of oxycodone craving, but the molecular mechanisms underlying this incubation are largely unknown. Here, we focused on the mitochondrial ubiquitin ligase, MITOL (also known as March5). MITOL is an E3 ubiquitin ligase that ubiquitinates key proteins regulating mitochondrial dynamics to target them for degradation, such as dynamin-related protein 1 (Drp1), a primary fission mediator.

Methods: In Experiment 1, we trained male and female rats to self-administer saline (n = 6) or oxycodone (n = 6) for 10 days (6 h/d), and collected OFC on abstinence day 15 for subsequent qPCR analysis to measure mRNA expression of MITOL. In Experiment 2, we first injected an adeno-associated virus (AAV) expressing MITOL under a ubiquitous promoter (AAV2-CMV-MITOL-GFP, n = 11) or AAV2-CMV-GFP (control, n = 14) bilaterally into the OFC of male and female rats. Next, we trained rats to self-administer oxycodone (0.1 mg/kg, 6 h/d) for 10 days and then tested rats for oxycodone seeking on abstinence days 1 and 15.

Results: We found that MITOL mRNA expression in OFC of oxycodone rats decreased compared with OFC of saline rats, suggesting MITOL in the OFC may play a negative role in incubation of oxycodone craving. Surprisingly, overexpressing MITOL in the OFC significantly increased both oxycodone intake during self-administration and incubated oxycodone seeking on abstinence day 15, but not non-incubated seeking on abstinence day 1, compared with the control group.

Conclusions: Together, these findings suggest that while MITOL mRNA decreases in OFC during incubation of oxycodone craving, its function may increase during abstinence, which drives incubated Oxy seeking. Following up on these findings, current studies are probing how MITOL protein expression in OFC changes during incubation of oxycodone craving.

Keywords: ubiquitin proteasome system, incubation of oxycodone craving, Mitochondrial dynamics, orbitofrontal cortex

Disclosure: Nothing to disclose.

P389. Examining the effects of cocaine on spatial patterns of striatal dopamine release

Hannah Elam, Devan Gomez, Erin Calipari

Vanderbilt University, Nashville, Tennessee, United States

Background: A key feature of cocaine use disorder (CUD) is deficits in dopamine release in the nucleus accumbens core (NAc). These deficits in dopamine release are concomitant with disruptions in the ability to process non-drug stimuli and the severity of these deficits correlates with negative treatment outcomes in patients. Understanding how drug use leads to impaired stimulus processing is critical for both understanding and ultimately treating individuals suffering from CUD. Dopamine is often discussed as a reward or valence signal; however, recent work has demonstrated that in the NAc both appetitive and aversive stimuli increase dopamine release. However, while both aversive and appetitive stimuli induce dopamine release in the same brain region as measured with techniques like voltammetry, microdialysis, and fiber photometry with optical dopamine sensors, these techniques suffer from poor spatial resolution. Thus, it is possible that different stimuli cause dopamine release from different pools that are in the same region, but spatially distinct. We sought to examine spatial patterns of dopamine transmission in the NAc in response to appetitive and aversive stimuli, before and after repeated exposure to cocaine.

Methods: Here, we utilized in vivo two-photon imaging of the dopamine sensor, dLight 3.8 in the striatum, to visualize spatial dopamine release, in awake and behaving mice. We aimed to test the following hypotheses: 1. patterns of dopamine release are spatially distinct in response to appetitive and aversive stimuli, and 2. repeated exposure to cocaine disrupts patterns of dopamine transmission. First, animals were exposed to random, unsignaled presentations of sucrose or quinine during baseline in vivo imaging sessions. Following baseline recordings, animals received cocaine injections (10 mg/kg, i.p.) for 10 consecutive days or saline, as a control, before they were once again exposed to random presentations of appetitive or aversive stimuli.

Results: We find spatial patterns of dopamine dynamics in response to quinine and sucrose at baseline. Across regions of interest, patterns of dopamine release vary in their intensity and duration in response to sucrose and quinine. As expected, repeated exposure to cocaine disrupts dopamine transmission in response to salient stimuli.

Conclusions: Given the critical role of dopamine to control learning outside of the drug context, it is imperative we examine dopamine dynamics in response to non-drug stimuli. Collectively, these studies aim to uncover differences in dopamine release patterns in response to appetitive and aversive stimuli and how these patterns of release are affected by repeated exposure to cocaine.

Keywords: Dopamine, optical biosensors, Two-photon

Disclosure: Nothing to disclose.

P390. Diurnal rhythms in neuronal activity in the nucleus accumbens: underlying mechanisms and impact of cocaine self-administration

Taylor Stowe, Yanhua Huang, Colleen McClung

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

Background: Biological rhythms, including diurnal (light/dark) rhythms, have been found in psychiatric disorders, such as substance use disorders (SUDs). There is a bidirectional relationship between rhythms and SUDs as those with disrupted rhythms are more vulnerable to drug-taking and drug-exposure can disrupt circadian rhythms. Notably, drug-taking patterns can vary throughout the day, indicating that individuals may be more susceptible to drug use at certain times of day. Overall, it is crucial to determine the mechanisms that mediate rhythms in reward-related behaviors, like drug-taking, to better understand vulnerability to developing SUDs. The nucleus accumbens (NAc) plays a key role in reward-related behaviors and is primarily made up of GABAergic medium spiny neurons (MSNs) but also contains cholinergic interneurons (CINs). Our lab has previously shown diurnal rhythms in NAc MSN activity with higher activity during the dark cycle; however, MSNs are not all the same with some expressing dopamine D1 or D2 receptors with stimulation differentially impacting reward-related behavior. For example, activating D1 receptors increases reward-related behaviors while activating D2 receptors decreases motivated behaviors. Additionally, we do not know if the CINs have diurnal rhythms in activity. CINs are the primary source of acetylcholine (ACh) in the NAc, which modulates dopamine (DA) release. Rhythms exist in the ACh modulation of DA, but it remained unknown if there were rhythms in CIN activity. Thus, we wanted to determine if rhythms in neuronal activity in the NAc were cell-type specific.

Methods: Here we expanded on our previous data by measuring activity via ex vivo electrophysiology in the NAc over the 24 hr cycle in specific types of MSNs and CINs. Using pharmacological methods, we also aimed to determine the potential mechanisms driving rhythms in these mechanisms. From our previous RNA sequencing data, we found that HCN channels may mediate diurnal differences in D1 and D2 MSNs. Thus, we applied an HCN channel antagonist and measure excitability in the light and dark cycle. In the NAc, CIN activity is mediated via D2 receptors and nAChRs. We utilized drugs targeting D2 receptors or nAChRs to determine if these potentially play a role in the diurnal rhythms and sex differences in CIN activity. After determining baseline variations and potential mechanisms, we also utilized cocaine IV self-administration to determine how chronic cocaine exposure affects rhythms in MSNs and CINs. Mice were implanted with a jugular catheter and went through 7 days of cocaine (0.5 mg/kg) self-administration on a fixed-ratio (FR) 1 schedule. After acquisition, mice were euthanized the following day for electrophysiology.

Results: Our data show that D1 and D2 containing MSNs may have opposing diurnal variations in excitability. We found that D1 MSNs were more excitable in the beginning of the light cycle while D2 MSNs were more excitable in the beginning of dark cycle. These rhythms may be driven by HCN channels. Additionally, our data suggest that tonic CIN activity is higher during the dark cycle, specifically in male mice. We found that D2 receptors may mediate rhythms in tonic CIN activity. Lastly, we found that chronic cocaine exposure changed rhythms in both D1/D2 MSNs and CINs. D1/D2 MSNs have an overall decrease in excitability in the beginning of the light cycle while there is an increase in CIN activity in both males and females in the beginning of the light cycle.

Conclusions: Given the essential roles of MSNs and CINs in motivated behaviors and local NAc dynamics, rhythmic activity in these cells may influence drug-taking behaviors and contribute to vulnerability to SUDs. These rhythms may also regulate local dynamics, such as DA release. In addition to baseline rhythms in these cell types, we also found that chronic cocaine exposure may change the neural activity based on cell type and time-of-day. These data further support the potential bidirectional rhythms between rhythms and SUDs. Collectively, these novel findings will advance our understanding of how rhythms contribute to the neural mechanisms driving reward-related behaviors associated with SUDs.

Keywords: Alcohol and substance use disorders, circadian rhythms, Nucleus Accumbens

Disclosure: Nothing to disclose.

P391. Binge drinking decreases somatostatin peptide effect in the prefrontal cortex in a pathway-specific manner

Nicole Crowley

Penn State University, University Park, Pennsylvania, United States

Background: Somatostatin (SST) peptide and SST expressing neurons are implicated in psychiatric disorders and are sensitive to alcohol consumption. Our lab has shown that in naive mice, SST peptide hyperpolarizes pyramidal and non-pyramidal cells in a subregion of the prefrontal cortex, the prelimbic cortex (PLC) and decreases intrinsic excitability of these cells in slice electrophysiology experiments. Subsequent work in our lab demonstrated that binge-like alcohol consumption results in a significant loss of SST inhibitory effects in the PFC. To further understand alcohol induced changes in somatostatin signaling, we examined the extent to which binge-like alcohol consumption altered SST effect onto specific prelimbic outputs such as the nucleus accumbens (NAC)- projecting pyramidal neurons.

Methods: Retrograde viral tracer (pAAV-hSyn-EGFP) was injected into the NAC of adult male and female C57BL/6J mice to label PFC à NAC projecting neurons. Mice then underwent 4 weeks of binge-like alcohol consumption or water control using the drinking in the dark (DID) model. 24-hours after the last binge session, whole cell patch-clamp electrophysiology experiments were used to assess intrinsic excitability and SST peptide-mediated hyperpolarization of labeled NAC-projecting PL neurons.

Results: Bath application of SST resulted in significant hyperpolarization and reduced excitability of NAC-projecting neurons in both water-exposed and alcohol-exposed mice. There was not a significant group difference (p’s > 0.31), suggesting that 4 weeks of DID failed to alter SST effect onto NAC projecting PL neurons.

Conclusions: Overall, this suggests that another projection from the PL (not the NAC projection) is responsible for the observed prelimbic-wide reduction in SST effect following binge-like alcohol consumption. Ongoing work is characterizing the effects of SST peptide onto periaqueductal grey (PAG) projecting neurons.

Keywords: neuropeptides, Alcohol, Medial Prefrontal Cortex

Disclosure: Nothing to disclose.

P392. Alcohol abstinence precipitates increased alcohol seeking and compulsive intake in association with increased BNST activity

Marie Doyle, Hye-Jean Yoon, Megan Altemus, Anika Park, Martha Troutman, Laura Grunenkovaite, Danielle Adank, Caitlyn Edwards, Nia Chetkovich, Louise Lantier, Cody Siciliano, Erin Calipari, Danny Winder

UMass Chan Medical School, Worcester, Massachusetts, United States

Background: Alcohol Use Disorder (AUD) is defined by shared diagnostic criteria, yet individuals show strikingly diverse drinking patterns and relapse vulnerabilities, reflecting neurobiological heterogeneity. The bed nucleus of the stria terminalis (BNST), through its cytoarchitecture and integration of stress and reward circuitry, is well positioned to contribute to this variability. Capturing such individual differences in animal models is essential for identifying the neural mechanisms that underlie AUD susceptibility and progression.

Methods: To understand how BNST function varies across individual mice and contributes to alcohol behavior trajectories, we used the Structured Tracking of Alcohol Reinforcement (STAR) operant task to phenotype male and female C57BL/6J mice as high, low, or compulsive (aversion-resistant) ethanol drinkers. Briefly, mice self-administered 20% ethanol in 1-hour sessions on a fixed ratio 5 (FR5) schedule for 14 days, followed by a 28-day forced abstinence period. Ethanol seeking was assessed under extinction conditions on withdrawal days 1 and 28. Mice then resumed ethanol self-administration under the STAR protocol for phenotyping: three baseline drinking sessions were followed by increasing quinine concentrations (0.25, 0.5, 0.75, and 1 mM) added to ethanol to assess aversion resistance, with phenotypes determined as described in Brown et al., 2023 (n = 9–10/phenotype). For in vivo recordings, a subset of mice received unilateral intra-dorsal BNST infusions of AAV9-hSyn-GCaMP7f-WPRE-SV40 along with fiberoptic implantations at the same coordinates (n = 4–5/phenotype).

Results: During initial ethanol self-administration, BNST cFos+ counts correlated with ethanol intake, directly tying BNST activation to operant drinking. Using in vivo fiber photometry, we found that compulsive drinkers displayed elevated BNST calcium transients during drinking bouts, despite similar ethanol intake across groups, pointing to heightened BNST recruitment in this phenotype.

Forced abstinence uncovered striking phenotype-specific adaptations. Ethanol seeking during protracted, but not early, abstinence predicted future compulsive intake. High and low drinkers reduced seeking behavior across abstinence, whereas compulsive drinkers persisted, showing behavioral rigidity. Notably, BNST calcium transients increased during protracted abstinence seeking only in compulsive drinkers, highlighting phenotype-specific plasticity.

We next compared compulsive ethanol intake between mice exposed to forced abstinence following ethanol exposure and those with only operant training. The abstinence group showed increased compulsive drinking, demonstrating that abstinence experience itself precipitates maladaptive drinking.

Finally, this BNST activity appears ethanol-specific, as saccharin drinking was not correlated with BNST cFos+ counts and reflected a trend toward unique dynamics for high ethanol drinkers.

Conclusions: Together, these findings reveal that ethanol abstinence precipitates increased ethanol seeking and compulsive intake, tightly linked to dynamic, phenotype-specific BNST activity. By uncovering how BNST circuits uniquely adapt in compulsive drinkers, this work identifies a critical neural correlate of relapse vulnerability and offers new insight into the biological diversity of AUD.

Keywords: alcohol use disorder, Mouse models, individual differences, bed nucleus of the stria terminalis

Disclosure: Nothing to disclose.

P393. Sex-dependent molecular and circuit mechanisms of the interpeduncular nucleus in alcohol use disorder

Junshi Wang, Stephanie Caligiuri, Amanda Roberts, Mary Heyer, Masago Ishikawa, Lauren Wills, Adam Catto, Paul Kenny

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

Background: Alcohol use disorder (AUD) affects 29.5 million people in the U.S., with women showing an increased vulnerability to developing AUD over the past decade. The interpeduncular nucleus (IPN) is an understudied midbrain region relaying information between the basal forebrain and hindbrain. Emerging evidence suggests the IPN plays an important yet not fully understood role in regulating the motivational properties of substances of abuse. Also, the IPN is involved in female reproductive behaviors and is sensitive to ovarian hormones. IPN neurons are highly heterogeneous and can be subdivided into functionally distinct populations based on transcriptomic and signaling profiles. These neurons utilize a diverse set of signaling mechanisms, including neurotransmitters such as GABA, glutamate, serotonin, and many neuropeptides. Intriguingly, the IPN is densely vascularized. Morphological studies in humans showed that many IPN neurons assembled near the blood vessels and form foot-like expansions onto them, suggesting that these neurons may detect circulating signals and play neurosecretory functions. Taken together, these lines of evidence highlight the need for comprehensive single-cell transcriptomic and connectomic characterization of the IPN, as well as functional studies to define the roles of distinct neuronal subsets in AUD-related behaviors.

Methods: Behavioral experiments were conducted in adult C57BL/6J mice. Ethanol intake was assessed using the drinking-in-the-dark paradigm and operant self-administration procedures. Single-cell RNA sequencing (scRNA-seq) was performed on the 10x Genomics Chromium platform, and data were analyzed with ScanPy. Visium spatial transcriptomics (10x Genomics) data were processed and analyzed using Seurat. Whole-cell patch-clamp recordings were employed to evaluate neuronal activity and synaptic transmission. For electrophysiological studies, Four Core Genotype and gonadectomized mice on a C57BL/6J background (Jackson Laboratory) were used.

Results: To investigate the molecular organization of IPN, we first employed spatial transcriptomics platforms with complementary levels of resolution (Visium, CurioSeeker) to compare the transcriptional profiles of midbrain cells between female and male mice. Among all midbrain areas profiled, the IPN exhibited the highest density of sex-dependent differentially expressed genes (DEGs). Single-cell RNA sequencing of IPN from male and female mice further revealed that chronic alcohol exposure induced transcriptomic plasticity in the IPN in a sex-dependent manner. Gene ontology analysis of IPN DEGs predicted sex differences in axon guidance, synaptic transmission, and circuit connectivity. To assess connectomic difference between males and females, we employed MAP-Seq to analyze IPN efferent projections in males and females at single-cell resolution. We identified sex-dependent differences in the connectivity strength of multiple IPN efferent circuits. Notably, a major subset of IPN neurons projected exclusively to the medial raphe nucleus (MR). Pathway-specific electrophysiological recordings revealed robust sex differences in synaptic strength within this circuit. Using gonadectomized male and female mice and the Four Core Genotypes mouse model, we discovered that gonadal hormones, but not sex chromosome assignment, shaped sex differences in synaptic strength in this circuit. Finally, lesioning the IPN-to-MR circuit increased alcohol drinking in both sexes and abolished sex-related differences in drinking behavior.

Conclusions: In summary, our high-throughput single-cell transcriptomic and connectomic analyses reveal pronounced sex-dependent differences in IPN neurons and their adaptations to alcohol exposure. These findings highlight previously underappreciated non-mesolimbic and sex-specific mechanisms that may contribute to alcohol use disorder (AUD).

Keywords: Alcohol consumption, Neural circuit and animal behavior, Gonadal Hormones, brain connectome, Transcriptome

Disclosure: Nothing to disclose.

P394. Divergent responses in prefrontal cortical and hippocampal blood-brain barrier protein expression in animal models of nondependent drinking and alcohol dependence

Angelica Vandekerkhoff, Sehwa Jung, Hannah Nonoguchi, Dhwani Kharidia, Chitra Mandyam

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

Background: The prefrontal cortex (PFC) and hippocampus are important for the development of alcohol addiction. Neuronal activity in these regions is regulated by immune responses, and neuroimmune responses could be assisted by disruption of the blood-brain barrier (BBB). The detrimental effects of chronic alcohol consumption, abstinence and relapse to alcohol drinking on BBB markers in the PFC and hippocampus have been minimally explored. Moreover, how and if these effects are different or divergent in alcohol dependence and nondependent conditions is not known. Here we seek to answer these questions.

Methods: Adult female and male rats were made ethanol dependent by chronic intermittent ethanol vapor (CIE) and ethanol drinking (ED) procedure or only experienced ED as nondependent drinking. CIE rats were euthanized during acute withdrawal (6–8h after CIE), protracted abstinence (2 weeks) or after relapse session and nondependent rats were euthanized at matched time points. Brain tissue homogenate of the PFC and hippocampus were analyzed for BBB disruption via Western blotting. This allowed us to associate the levels of drinking in male and female nondependent and dependent rats and expression of proteins important for tight junctions (Cld5) and adherens junctions (Cdh5). NF-kB, a marker of inflammation was also evaluated in each of these conditions.

Results: CIE increased ED, with rats having higher ethanol consumption during CIE and relapse to ethanol drinking sessions compared with pre-CIE sessions. ED was not altered in nondependent rats. Under both conditions, females consumed more ethanol than males. Western blotting indicated enhanced expression of Cdh5 in CIE rats, an effect that was significant during acute withdrawal in males and females in the PFC and hippocampus. In the PFC, Cdh5 levels decreased below control levels during abstinence and were stimulated after the relapse session. The reduction in Cdh5 was associated with increases in activated NF-kB during abstinence. This effect was specific to the PFC and was not observed in the hippocampus. Cld5 was reduced in both brain regions in both sexes under CIE conditions and the reduction was more pronounced during acute withdrawal. Nondependent drinking did not alter Cdh5 or Cld5 in either brain regions in either sex.

Conclusions: These results reveal significant effects of ethanol dependence in adherens junction protein Cdh5, and its association with neuroimmune factor NF-kB in the PFC, and emphasize their possible role in relapse to ethanol seeking in ethanol dependent rats.

Keywords: Blood-Brain-Barrier, prefontal cortex, Alcohol addiction, Hippocampus

Disclosure: Nothing to disclose.

P395. Cannabidiol Attenuates Heroin Seeking in Male Rats Associated With Normalization of Discrete Neurobiological Signatures Within the Nucleus Accumbens With Subregional Specificity

Alexandra Chisholm, Jacqueline-Marie Ferland, Randall J. Ellis, Yasmin L. Hurd

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

Background: Opioid use disorder involves cycles of compulsive use, abstinence, and relapse. Cannabidiol (CBD), a non-intoxicating cannabinoid, is under investigation as an anti-relapse treatment. CBD was shown to attenuate cue-induced heroin-seeking in a rodent model of relapse and reduce cue-induced craving and anxiety in abstinent individuals with heroin use disorder. The neurobiological mechanisms by which CBD may exert its anti-relapse effects are still unknown. The objective of the current study was to evaluate the effects of CBD administration on heroin-seeking in conjunction with transcriptomic profiling in the nucleus accumbens core (NAcC) and shell (NAcS), subregions of the NAc highly implicated in different aspects of addiction.

Methods: Male Long Evans rats (n = 32) self-administered heroin for 15 days, followed by 14 days of abstinence. Rats were injected with vehicle or CBD (5 or 10 mg/kg, i.p) 24 hours before a drug-seeking session. Tissues were extracted 1.5 hours following the drug-seeking session. NAcC and NAcS tissue were dissected, and bulk RNA sequencing was performed. Analyses of the drug-seeking session were assessed using two separate one-way ANOVAs with lever (active or inactive) serving as the independent variable. The threshold for statistical significance was set at p < 0.05. Standard RNAseq pipelines were used for alignment, quality control, and gene counting. Differential gene expression was examined in relation to heroin-seeking behavior. Differentially expressed genes (DEGs) were defined as a nominal p < 0.05. We assessed the effect of heroin (saline-vehicle vs. heroin-vehicle (H)), the effect of heroin in combination with CBD (saline-vehicle vs. heroin-CBD10 (H + CBD), and the effect of CBD relative to the effects of heroin (heroin-vehicle vs. heroin-CBD10 (H vs. H + CBD)) as separate analyses using the DESeq2 package in R. Functional ontology analysis was completed using the Enrichr database using a False Discovery Rate (FDR) of adjusted-q- < 0.05. Overlap analyses were completed by selecting group DEGs with a nominal p < 0.05. Predicted canonical pathway and upstream regulator analysis were performed using Ingenuity Pathway Analysis. Canonical pathways and upstream regulators considered for analyses met a Benjamini-Hochberg-corrected p-value of p < 0.05. A Pearson correlational analysis was conducted to examine the relationship between gene expression levels and active lever responses during the drug-seeking session for all vehicle animals. Significant correlations were determined using a nominal p-value < 0.05.

Results: During the drug-seeking session, heroin-experienced animals showed robust active lever responding (M = 122.40, SD = 54.62), [F(3,28) = 13.73, p < 0.0001] compared to saline-vehicle counterparts [(M = 20.13, SD = 6.96), t(28) = 6.21, p < 0.0001, d = 2.63]. Administration of CBD5 [(M = 68.00, SD = 27.77), t(28) = 3.31, p = 0.0156, d = 1.26] and CBD10 [(M = 47.89, SD = 24.23), t(28) = 4.65, p = 0.0004, d = 1.76] significantly reduced active lever responding compared to heroin rats given vehicle [(M = 122.40, SD = 54.62)]. RNA-sequencing of the NAcC and NAcS revealed shared transcriptomic alterations the NAc subregions in response to heroin, with a more robust impact evident in the NAcS. Though CBD had minimal impact on the heroin-induced perturbations in the NAcC, it normalized components of the transcriptomic signature altered by heroin in both NAc subregions including transcripts that correlated with heroin-seeking behavior. The most substantial effects were seen in the NAcS where CBD normalized a particular subset of genes with a strong correlation to heroin-seeking behavior. Those genes were specifically linked to the extracellular matrix, astrocyte function, and their upstream regulators related to immune function.

Conclusions: These findings underscore the NAc subregional signatures of heroin-induced neurobiological perturbations and provide novel biological targets relevant for CBD’s apparent anti-relapse effects.

Keywords: Cannabidiol, Nucleus Accumbens, drug seeking

Disclosure: Nothing to disclose.

P396. The role of orbital frontal and intralaminar thalamic glutamate inputs to dorsal striatum on cognitive flexibility and punished cocaine taking

David Bortz, Mary Torregrossa

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Cognitive inflexibility is a trait marker that predicts substance use disorder (SUD) development across species, but the underlying neural mechanisms relating the two are unknown. Glutamate (Glut) and dopamine (DA) interactions in the dorsomedial (DMS) and dorsolateral (DLS) striatum are involved in the transition of behavior from flexible to outcome-insensitive during cognitive and drug seeking tests. We hypothesized that a common mechanism within this striatal interaction could predict both cognitive inflexibility and risk for developing SUD phenotypes.

Methods: Rats learned a simple discrimination, then performed a strategy switching test. We used pathway-specific DREADDs and multi-channel fiber photometry to simultaneously measure Glut and DA release in the DMS and DLS during learning and strategy switching when either orbital frontal or intralaminar thalamic inputs were silenced. We then determined the effects silencing each projection on cocaine taking despite foot shock punishment.

Results: Elevations in Glut release in DLS at trial initiation (r = −0.46, P = 0.039) and in DMS prior to a correct choice (r = −0.38, P = 0.085), with reductions in DA release, predicted both better performance on the cognitive flexibility task and cocaine taking that was sensitive to punishment ((DLS: r = −0.68, P = 0.008); (DMS: r = −0.75, P = 0.001)). Inhibition of orbital frontal cortex inputs reduced the Glut measured in both the DMS and DLS prior to a correct choice (DMS: t(4) = 3.06, P = 0.019 and DLS: t(4) = 1.97, P = 0.072), whereas inhibition of intralaminar thalamus inputs reduced Glut measured in both regions at trial initiation (DMS: t(4) = 1.66, P = 0.087 and DLS t(4) = 2.80, P = 0.025). The inhibition of both projections impaired strategy switching (F (2, 30) = 18.4, P < 0.0001) and increased punishment resistant cocaine use (F (1, 9) = 8.93, P = 0.015).

Conclusions: We propose that projection-specific elevations in Glut release at specific test epochs represent a “signature” of flexibility that predicts resilience to SUD phenotypes, whereas elevations in DA release at those same test epochs predicts inflexibility and vulnerability to SUD phenotypes.

Keywords: cognitive flexibility, cocaine self-administration, dorsal striatum, orbital frontal cortex, intralaminar thalamus

Disclosure: Nothing to disclose.

P397. Molecular adaptations in the parabrachial nucleus following opioid exposure and withdrawal

Isaiah Williamson, Cassie Nelson, Bridget Asare-Owusu, Krystal Flores-Felix, Barbara Juarez

University of Maryland, Baltimore, Washington, United States

Background: Opioid dependence and withdrawal induces a spectrum of physiological, behavioral and psychological symptoms and neural circuit adaptations that are a contributor to opioid relapse. Understanding the molecular basis of these adaptations is critical to furthering treatments. We have discovered that calcitonin gene-related peptide neurons of the parabrachial nucleus (PBNCGRP), a critical in relaying information about aversive experiences to forebrain regions, are inhibited by opioid exposure and activated by opioid withdrawal in mice. Inhibiting activity of this population during opioid withdrawal dampens distinct opioid withdrawal symptoms. Here, we sought to understand the potential molecular mechanisms that may underlie these neurophysiological adaptations we observed.

Methods: Male and female adult C57BL/6J mice were exposed to a twice daily injection regimen of escalating doses of morphine to induce opioid dependence. Control mice (the nondependent groups) received twice daily injections of saline. Opioid dependent mice were either kept as opioid dependent, exposed to naloxone to precipitate withdrawal or underwent spontaneous withdrawal for 24 hours. Non-dependent mice experienced one of the two conditions. PBN tissue was collected at time points to compare molecular adaptations across five groups in total: Nondependent-saline, Nondependent-naloxone, Dependent-saline, Dependent-naloxone (having experience naloxone withdrawal), and Dependent-SpontaneousWithdrawal (having experienced 24 hours of spontaneous withdrawal). We performed RNA sequencing using the Illumina Novaseq6000 platform. Genes that met a false discovery rate (FDR) of < 0.05 were deemed significantly differentially expressed and were then used to perform a gene ontology analysis along with transcriptomic profiling of differentially expressed genes.

Results: We found the opioid exposure and withdrawal induces gene expression changes in the PBN. We discovered 481 differentially expressed genes in total across group comparisons, with 255 upregulated genes and 226 downregulated genes. Inducing opioid dependence in mice specifically resulted in 31 altered genes in the PBN when compared to non dependent mice. Naloxone precipitated withdrawal induced 181 altered genes in the PBN when compared to non dependent mice. Spontaneous withdrawal induced 43 altered genes in the PBN when compared to non dependent mice. This suggests transcriptional differences between modes of opioid withdrawal. When comparing the two different withdrawal groups, 100 genes were found to be differentially expressed. Gene ontology analysis revealed transcriptional regulation in genes involved in GPCR signaling, monoaminergic transport regulation, and calcium ion transport.

Conclusions: Our findings indicate that the PBN is transcriptionally and molecularly active during the induction of opioid dependence and opioid withdrawal. Through gene ontology analysis, we discovered differences in classes of genes involved in activity regulation. This could be the mechanism that regulates the neurophysiological adaptations we have previously observed. Interestingly, we discovered that genes were differentially expressed between opioid withdrawal groups. This highlights the need for further investigation into the mechanisms underlying these two aversive states in order to tailor better clinical therapeutics.

Keywords: RNA Sequencing, parabrachial nucleus, opioid dependence, opioid withdrawal

Disclosure: Nothing to disclose.

P398. A robust and convenient framework for studying the neurobiological substrates underlying individual, strain, and sex differences in fentanyl use disorder in mice

Lukas MacMillen, Todd Appleby, Sierra Schleufer, Erica Sanchez, John Neumaier, Sam Golden, Kevin Coffey

University of Washington, Seattle, Washington, United States

Background: Opioid use disorder represents a significant public health crisis in the United States, with fentanyl in particular contributing to more than half of the 80,391 overdose-related fatalities in 2024. Fentanyl is now the drug of choice for most opioid users, particularly for younger individuals and those suffering psychological distress. Fentanyl poses a unique risk compared to other opioids due to its extreme potency, low cost, and ease of administration. Current approved treatments for fentanyl use disorder are similar to those used for other opioids, primarily opioid replacement therapies or opioid receptor blockade, but these strategies have unfortunately proven to be insufficient to stem the tide of fentanyl related suffering and death.

Methods: Opioid use disorder (OUD) is complex and difficult to model preclinically, and two major facets of OUD that are often understudied are 1) the substantial individual differences in severity of opioid use and relapse along with the predictors of those differences, and 2) individuals’ opioid use and relapse risk profiles may involve diverse brain circuits and behavioral patterns. To address these issues, we developed a robust and convenient fentanyl self-administration (SA) model that 1) is useful for studying individual differences in risk and resilience toward volitional fentanyl intake and reinstatement in male and female mice, and 2) is easily compatible with cutting-edge whole brain genetics and activity mapping techniques. We are also providing these methods as an open community resource that allows labs to implement this model quickly and easily, using relatively low cost and widely available equipment and have made software tools available to analyze complex fentanyl-associated behaviors under a common framework that includes both traditional operant metrics and modern video-based behavioral classification.

Results: On a group level, mice in our model show classical evidence of volitional drug intake, including escalating intake, strong association of conditioned cues, and reinstatement to drug seeking. They also displayed evidence of drug-level preference and titration during both standard SA and variable dose “between-sessions thresholding”. Utilizing both inbred C57BL/6 and outbred CD1 (male and female) mice, we found profound individual, strain, and sex differences in fentanyl self-administration. Some notable differences are that while C57BL/6 mice escalate intake, CD1 mice maintain relatively high intake throughout training. And while C57BL/6 mice show increased lever pressing during cued reinstatement, CD1 mice show reduced head entry latency. Using light sheet microscopy and tissue clearing, we also found unique whole brain activity patterns among animals exposed to fentanyl and those undergoing cued reinstatement.

Conclusions: Standardized methodologies for researching the effects of fentanyl preclinically are necessary to map whole brain neural activity during different stages of fentanyl use. The oral self-administration model presented here shows how different strains of mice may be utilized to study unique components of fentanyl use disorder and will be a pivotal resource in the development of new treatments.

Keywords: Fentanyl Self-Administration, machine learning, individual differences, Whole-Brain Rodent Imaging, sex differences

Disclosure: Nothing to disclose.

P399. Continued studies of cocaine polysubstance use in socially housed female and male monkeys: effects of nicotine on cocaine- and cocaethylene-induced reinstatement

Michael Nader, Mia Rough, Brianna Roberts

Wake Forest University School of Medicine, Winston-Salem, North Carolina, United States

Background: The goal of this research is to enhance our understanding of the behavioral and neuropharmacological effects of cocaine in socially housed nonhuman primate models of cocaine use disorders (CUD). Estimates indicate that ~80% of people who use cocaine co-use nicotine and nearly 90% of people with CUD co-use alcohol. Despite this, very few preclinical studies incorporate either drug in animal models of CUD. Cocaethylene is an active metabolite formed when alcohol and cocaine are co-used in close succession. Recently, we showed that cocaethylene had equal reinforcing strength and potency to cocaine when studied under progressive-ratio and concurrent drug vs. food choice schedules of reinforcement (Allen et al., Drug Alcohol Dep., 2023). The goal of the present study is to extend this comparison of cocaine and cocaethylene to another animal model – reinstatement of drug seeking.

Methods: Typically, under a reinstatement model, drug self-administration is extinguished and pretreatments are tested for their ability to increase saline self-administration (i.e., reinstate drug seeking). For these studies, socially housed female (N = 11) and male (N = 12) monkeys in same-sex social groups of 4 monkeys per pen, self-administered cocaine or cocaethylene under a concurrent FR 30 schedule of reinforcement with food as the reinforcer; note that not all monkeys in a social group were studied. After completing self-administration dose-response curves for both cocaine and cocaethylene (0.01–0.3 mg/kg/injection), saline was substituted for drug; typically, this occurred at the lowest cocaine dose that resulted in > 80% drug choice over food. When stable choice occurred (3 days in which the mean did not vary by ± 15% and choice was at least 80% food), various doses of cocaine or cocaethylene (both at ranges of 0.01–0.3 mg/kg, iv) were administered 1-min before a food vs. saline choice session; each dose was double determined (Experiment 1). Next, nicotine (0.01–0.056 mg/kg, iv) was co-administered with the highest dose of cocaine and cocaethylene that did not result in at least 75% saline choice; test sessions were typically 2 times per week (Experiment 2). There was a return to food vs. cocaine choice after 2–3 test sessions before returning to food vs. saline choice. Reinstatement was operationally defined as drug choice > 75%.

Results: As noted previously, both cocaine and cocaethylene increased drug choice in a dose-dependent manner; there were no differences in reinforcing potency between the two drugs. When saline was substituted for drug, responding was reallocated to almost exclusively the food option within 2–3 sessions. For the reinstatement pretreatment studies, we hypothesized that both cocaine and cocaethylene would reinstate drug seeking, with greater potency in subordinate monkeys and that nicotine would potentiate these effects. In Experiment 1, cocaine reinstated responding in 5 of 7 dominant monkeys, but only 1 of 5 subordinate monkeys. In females, cocaine reinstated responding in 2 of 5 dominant and 3 of 6 subordinates. Cocaethylene was less effective, only reinstating drug choice in 4 of 23 monkeys. In preliminary findings, nicotine enhanced cocaine-induced reinstatement in 2 of 6 dominant males and no subordinate males, and in 3 of 4 females tested so far. Nicotine increased cocaethylene-induced reinstatement in only 3 of 15 monkeys of both sexes.

Conclusions: There are some noteworthy outcomes from these studies. In Experiment 1, there appeared to be sex and social rank differences in the ability of cocaine to induce reinstatement, with dominant male monkeys appearing the most sensitive of all phenotypes; cocaethylene, in contrast to having equal reinforcing effects to cocaine, was ineffective in reinstating drug-seeking behavior. In preliminary findings from Experiment 2, also in contrast to self-administration studies, nicotine was ineffective in potentiating the ability of either cocaine or cocaethylene to reinstate responding; this was the case irrespective of sex and social rank of the monkeys. If reinstatement studies are modeling subjective-like effects, these findings suggest that reinforcing effects are more sensitive indicators of polysubstance use than other effects, including subjective effects. It is clear that preclinical studies designed to understand the neuropharmacology and, ultimately, development of novel treatments for CUD need to include the co-use of alcohol and nicotine. While we have shown significant interactions with nicotine on the effects of a behavioral intervention designed to decrease cocaine and cocaethylene self-administration, future studies are required to examine the effects of behavioral and pharmacological interventions on nicotine interactions with cocaine- and cocaethylene-induced reinstatement.

Keywords: cocaine reinstatement and taking, polysubstance use, cocaine, nicotine, sex differences, Animal models of addiction, Nonhuman Primate Models

Disclosure: Nothing to disclose.

P400. Antiviral and anti-inflammatory effects of cannabidiol in HIV/SIV infection

Alysha Ellison, Jose Javier Rosado-Franco, Labib Mamun, Stephen Knerler, Ryan Vandrey, Elise Weerts, Michael Corley, Lishomwa Ndhlovu, Dionna Williams

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

Background: Chronic, immune activation is one of the hallmarks of HIV in the modern era, despite the effectiveness of antiretroviral therapy (ART) in suppressing viral replication. Importantly, the chronic inflammation that occurs contributes to the development of comorbid disorders that hinders the quality of life of infected people and increases mortality. Unexpectedly, ART is not sufficient to restore markers of immune activation to their pre-infection levels. Thus, there is a critical need to identify novel adjunctive agents to limit inflammation and its consequent effect on comorbid disease. We aim to characterize the impact of cannabidiol (CBD), a component of the cannabis plant, on HIV-associated chronic immune activation. We hypothesize that, through activation of the endocannabinoid system, CBD will decrease inflammation and decrease viremia.

Methods: Four juvenile rhesus macaques (n = 2 male, n = 2 female, aged 4 years old, power of 73.5%) were infected with 20AID50 of SIVmac251 for seven days, after which time oral CBD was administered. As oral CBD administration had not been evaluated previously in rhesus macaques, escalating doses of 10, 20, and 40 mg/kg CBD were administered once weekly for a total of three weeks. Macaques were necropsied in the fourth week. Blood was collected and plasma and immune cells isolated. Cytokines and the SIV p27 protein were measured in plasma, cerebrospinal fluid (CSF), and tissues. The frequency of peripheral blood mononuclear cells and markers of their activation were evaluated. The CBD metabolite, 7-OH-CBD, was quantified in urine. Eight tissues were collected and fourteen endocannabinoid receptors evaluated. Animals and procedures in this study were approved by and in compliance with the Johns Hopkins and Emory University Animal Care and Use Committees. Animal handling and euthanasia were conducted as stated under the NIH Guide for the Care and Use of Laboratory Animals and USDA Animal Welfare Regulation.

Results: SIV infection promoted viremia, as determined by detection of the p27 protein in plasma, which peaked to 2128+/−1073 pg/mL at 2-weeks post-inoculation. However, CBD significantly decreased viremia to 713+/−440 pg/mL at the terminal timepoint. Decreased viremia coincided with a significant decrease in 15 plasma, cerebrospinal fluid, and tissue (spleen, kidney, lymph node, adipose tissue and heart) cytokines involved in the antiviral response, including interferons, IL-1b, and TNF-a between 10–60 fold, often to levels below the limit of detection. CBD also slowed CD4+ T cell decline and polarization as well as decreased CD14+ CD16+ monocyte expansion known to occur during acute HIV/SIV infection. Urinalysis demonstrated that the 10 mg/kg CBD dose best reflected the pharmacokinetic profile of 7-OH-CBD that occurs in humans, notably the earlier peak that occurs in females (2 hours earlier) and the overall higher concentration that occurs in males relative to females (1.5 fold). Unexpectedly, CBD did not alter plasma or CSF endocannabinoids (anandamide or 2-arachidonoylglycerol) or 31 endocannabinoid-related/oxylipin lipids in the eight tissues that were evaluated, suggesting its effects were mediated by a change in signaling rather than expression profiles. To evaluate this, we evaluated 15 endocannabinoid receptors in five brain regions and six organs by qRT-PCR in uninfected, SIV-infected, and SIV-infected animals that received CBD. Surprisingly, SIV infection did not change any of the 15 endocannabinoid receptors in the evaluated organs and brain regions, relative to uninfected macaques. However, CBD administration significantly modulated eight endocannabinoid receptors in peripheral organs and brain, in an organ- and region-specific manner compared to SIV-infected animals that did not receive CBD. Notably, CBD modulated CB1, PPAR-a, PPAR-g, and TRPV-2 in most evaluated brain regions and organs. We performed single cell multiomics (scRNA-Seq + scATAC-seq) to more comprehensively evaluate CBD’s effects on the endocannabinoid system by comparing hippocampi from uninfected animals with SIV-infected macaques that received CBD, focusing on endocannabinoid receptors, modulators, transporters. Consistent with our qPCR findings, CB1 was enriched in SIV-infected animals that received CBD which was primarily restricted to neuronal populations. Further, single cell multiomics demonstrated CBD administration impacted endocannabinoid modulators responsible for anandamide synthesis, 2-arachidonoylglycerol degradation and cytosolic phospholipase A2, which regulates endocannabinoid precursors. We also found selective enrichment of fatty acid binding protein 3 that facilitates endocannabinoid intracellular transport. Importantly, the antiviral and anti-inflammatory effects of CBD were confirmed with in vitro experiments involving HIV infection of primary human macrophages, T cells, and microglia suggesting translational potential in people.

Conclusions: Our data indicate oral CBD administration as a promising adjunctive therapeutic strategy during HIV infection that can decrease viremia, plasma markers of chronic inflammation, and immune cell antiviral responses, even in the absence of ART. These findings suggest a clinical benefit of CBD that may improve quality of life in people living with HIV and decrease risk of comorbid disease.

Keywords: Cannabidiol, endocannabinoid system, rhesus macaques, HIV and Inflammation

Disclosure: Nothing to disclose.

P401. Microglial acid-sensing drives panic/PTSD-relevant outcomes during withdrawal from chronic ethanol in mice

Emily Cicon, Shawn Lam, Myles Allred, Haritha Karthikeyan, Pragnya Munisetti, Katherine McMurray

University of Illinois At Chicago, Chicago, Illinois, United States

Background: Alcohol use disorder (AUD), panic disorder and PTSD are highly comorbid psychiatric disorders with limited treatments. Emerging evidence suggests increased homeostatic stress sensitivity may be a shared mechanism in AUD, panic and PTSD. Regulating physiological homeostasis (e.g. pH) is critical for survival and threats to homeostasis elicit behavioral, emotional and physiological responses directed toward this goal. Homeostatic stressors that produce acid-base imbalance (e.g. non-hypoxic CO2 inhalation) elicit greater fear/anxiety, hyperventilation, and cardiovascular effects in panic disorder and PTSD patients. Alcohol use evokes acidosis which positively correlates with withdrawal severity. CO2 sensitivity is increased in people with a history of alcohol dependence and during alcohol withdrawal, while acute alcohol use reduces CO2 sensitivity. This also suggests dysregulated acid-sensing mechanisms contribute to AUD, panic and PTSD pathophysiology and may facilitate the increased risk to develop panic or PTSD in previously dependent individuals. Investigating the effect of alcohol use on CO2 sensitivity may yield unique and impactful findings, increase our understanding of comorbidity and lead to identification of new treatment targets. To do this, we recently developed a model to test the hypothesis that chronic alcohol use and withdrawal increases CO2 sensitivity in mice. Next, we sought to identify molecular mechanisms and neurocircuitry driving this effect. Previous work found microglial acid-sensing receptor T cell death association gene 8 (TDAG8) regulates CO2-evoked defensive responding. We recently found that acute ethanol increases TDAG8 in the brain. Thus, we sought to test the hypothesis that withdrawal from chronic ethanol increases CO2 sensitivity in mice in a TDAG8-dependent manner. We also sought to identify regional neuronal activation and inter-regional co-activation patterns that may drive these effects.

Methods: Adult male and female C57BL/6J mice received control (4-methylpyrazol (4-MP) 9mg/kg) or ethanol (2g/kg) + 4-MP injections for 9 days then underwent the CO2 inhalation paradigm. On the last day of treatment (day 9), mice were habituated to the inhalation chamber for 7m. They received their last injection immediately after removal from the chamber and were returned to their homecages. The next day at 24h withdrawal mice were exposed to breathing air or 5% CO2 inhalation for 10m. After 90m, brain tissue was collected and processed for quantification of cFos as a marker of neuronal activation using either immunohistochemistry and brightfield microscopy in slice (n = 10–12/group) or light sheet fluorescence microscopy in cleared whole brains (n = 8–16/group). We quantified fear-relevant defensive behaviors (freezing and rearing) in response to CO2 inhalation. In an additional study, mice received control or ethanol treatments as above for 14 days then received 0, 10 or 20 mg/kg of a TDAG8 inhibitor 30m prior to CO2 inhalation at 24h withdrawal. Behavioral studies were analyzed using 2 or 3-way ANOVA as appropriate, while cFos was quantified (regional and interregional co-activation) using ANOVA, Spearman or Pearson’s correlations as appropriate.

Results: We found significant overall effects of ethanol treatment (p < 0.01), air/co2 inhalation (p < 0.01), and sex (p < 0.05) on freezing in the CO2 inhalation paradigm. Withdrawal from chronic ethanol increased CO2-evoked defensive behaviors (increased freezing, reduced rearing) in both male and female C57BL/6J mice (p < 0.05). Post hoc tests revealed that within each sex, there were significant differences between the ethanol treated, CO2-exposed mice and all other treatment groups (vehicle treated/air, vehicle/CO2, and ethanol/air; p < 0.05). There were no other significant differences between groups. We also found a significant overall effect of TDAG8 inhibitor treatment (p < 0.01) and interaction of withdrawal and treatment on CO2-evoked freezing (p < 0.05). Post hoc tests revealed the 20mg/kg TDAG8 inhibitor treatment significantly reduced CO2-evoked freezing during withdrawal from chronic ethanol (p < 0.05).

Conclusions: These data suggest chronic ethanol use and withdrawal increases panic and PTSD-relevant behaviors in male and female mice. They also suggest a role for microglial acid-sensing receptor TDAG8 in mediating the effect of ethanol-withdrawal on CO2 sensitivity. Ongoing studies are investigating the molecular mechanisms and circuitry underlying this effect. Overall, these data support the use of this highly translational chronic alcohol use/withdrawal and CO2 inhalation model to improve our understanding of shared mechanisms underlying comorbid AUD, panic and PTSD.

Keywords: alcohol use disorder, Comorbidity, panic disorder, Posttraumatic stress disorder (PTSD), neuroinflammation

Disclosure: Chemular, Consultant, Spouse/Partner

P402. Pet Imaging of camp signaling: fentanyl-induced modulation of brain PDE4B activity and its reversal by naloxone in rats

Nathaniel J Burkard, Ailen Costamagna-Soto, Woochan Kim, Cheryl Morse, Jeih-San Liow, Susovan Jana, Shyam Samanta, Sung Won Kim, Leandro F Vendruscolo, Victor W Pike, George F Koob, Nora D Volkow

National Institute on Alcohol Abuse and Alcoholism, Bethesda, Maryland, United States

Background: Fentanyl is a potent opioid used for managing severe pain and anesthesia. However, fentanyl has high addiction potential when misused and has become the main contributor to the ongoing overdose crisis. Fentanyl’s activation of mu opioid receptors (MOR) decreases cyclic adenosine monophosphate (cAMP) levels which in turn, are regulated by phosphodiesterases (PDEs). Recently, [18F]PF-06445974 was developed as a PDE4B radioligand to image cAMP signaling in vivo. In this study, we employed [18F]PF-06445974 positron emission tomography (PET) in rats to investigate the effect of fentanyl on the cAMP signaling pathway.

Methods: Male Wistar rats received either fentanyl (0.05 mg/kg, SC; n = 4, FEN group) or saline (n = 4, CTL group) 2.5 ± 0.3 hours before undergoing 120 minutes of dynamic PET imaging with [18F]PF-06445974. In a second experiment, rats received an identical fentanyl pretreatment, followed by administration of the opioid receptor antagonist naloxone (2.8 mg/kg, IV; n = 3, FEN/NLX group) at 5 minutes before imaging. A transmission CT scan was performed before each PET scan and was utilized for attenuation correction and co-registration for image analysis. After image reconstruction, radioactivity was normalized with injection dose and body weight, generating time-activity curves for each brain region of interest (ROI). The area under the curves (AUCs) were used to evaluate statistical differences in brain uptake between groups (Student’s t-test).

Results: Fentanyl administration reduced [18F]PF-06445974 brain uptake (FEN group; AUC, 9729.8 ± 456.83 g/mL·min) by 14.1% compared to saline administration (CTL group; AUC,11324 ± 331.57 g/mL·min). In the second experiment, whole brain uptake of [18F]PF-06445974 was higher in the FEN/NLX group than in the FEN group (p = 0.013). Comparison of the FEN/NLX group to the CTL group showed no significant difference, indicating that naloxone restored brain uptake to baseline levels (p = 0.37).

Conclusions: The reduced brain uptake of [18F]PF-06445974 following fentanyl administration shows that fentanyl acutely modulates intracellular cAMP signaling in vivo, likely via cAMP-dependent regulation of PDE4B activity. Notably, the immediate and complete reversal of this effect by naloxone provides direct evidence that these changes are mediated through actions on MORs. Our study demonstrates that this in vivo approach could provide insight into the downstream neuroadaptations to acute and chronic opioid exposure, including the temporal dynamics and reversibility of cAMP signaling alterations.

Keywords: Positron emission tomography, cyclic AMP, Phosphodiesterase-4B (PDE4B), Rodents, Fentanyl

Disclosure: Nothing to disclose.

P403. Focal transcranial magnetic stimulation (TMS) of rat medial prefrontal cortex prevents incubation of cocaine craving

Hanbing Lu, Samantha Hoffman, Ying Duan, Zilu Ma, Hieu Nguyen, Aidan Carney, Taylor Scott, Olivea Varlas, Md Mohaiminul Haque, Elliot Stein, Zheng-Xiong Xi, Yavin Shaham, Yihong Yang

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

Background: Relapse to drug use is a major challenge in addiction therapy, highlighting the need for innovative therapeutic approaches. In humans, cocaine craving and relapse are frequently triggered by re-exposure to drug-associated cues and environments. This relapse vulnerability is modeled in rodents through the incubation of cocaine craving model—a progressive increase of cocaine-seeking behavior during abstinence.

Transcranial magnetic stimulation (TMS) has emerged as a promising therapeutic strategy, however randomized controlled studies in humans have proven difficult. We have developed high-density theta-burst stimulation (hdTBS) technology and a novel TMS platform, enabling repeated, focal stimulation of the rat medial prefrontal cortex (mPFC in awake rats. Here we leveraged these technological advances to assess the effects of TMS in a rat model of cocaine dependence.

Methods: Behavioral training: Rats (n = 24) were trained to self-administer intravenous cocaine (0.5 mg/kg/infusion) over 15 days (3 hours/day). This was followed by 3 weeks of homecage abstinence. A cocaine-seeking test was conducted on abstinence day 1, during which presses on the active lever triggered a tone and houselight previously paired with cocaine infusion, but no cocaine was delivered.

TMS treatment: Following day 1 testing, the rats were divided into two groups: one group received active high-density theta burst stimulation (hdTBS) once daily from abstinence days 14 to 20 (n = 13), while the other group received sham hdTBS during the same period (n = 11). A second cocaine-seeking test was conducted on abstinence day 21. Additionally, resting-state fMRI scans were performed on abstinence days 2 and 22—one day after each behavioral test.

TMS administration: A headpost was implanted on the rats’ skull to guide the hotspot of the TMS coil to mPFC (prelimbic and cingulate cortex). For active TMS, the rat’s head was position near the coil surface, and stimulation power was set to 120% motor threshold. For sham TMS, the rat’s head was placed 4 cm below coil surface, and the TMS power was 70% motor threshold. The resulting electric field inside rat head was essentially zero while acoustic noise levels of the TMS pulses were similar under these two conditions. Rats were habituated to handling and the TMS setup for 5 days, followed by daily TMS administration in awake state for 7 days.

fMRI scan: Resting-state fMRI data were acquired on a Bruker 9.4 Tesla scanner using an echo-planar imaging sequence. Scan parameters were the following: matrix size = 80 × 80, field of view = 30 × 30 mm2, slice thickness =0.6 mm, repetition time =1500 ms, echo time = 5 ms, 19 slices, slice gap = 0.1 mm. During imaging, the rats were lightly anesthetized with low isoflurane dose (0.5%) and dexmedetomidine (0.015 mg/kg i.p.), an anesthetic regimen shown to maintain large-scale brain networks.

Results: All rats tolerated TMS well with no signs of changes in eating, drinking, and grooming. No TMS-induced seizure occurred during or after TMS administration. Linear mixed effects modeling showed significant interaction between treatment conditions (active/sham) and time (abstinence day 1 pre-TMS, day 21 post-TMS) (F[1, 22] = 7.85, p = 0.01]. In the sham TMS group, there was a significant increase in cocaine seeking abstinence day 21 compared to day 1 (two-tailed paired t-test, p = 0.04), reflecting an “incubation” effect. In contrast, rats that received active TMS treatment showed no significant difference in cocaine seeking between abstinence day 21 and day 1 (p = 0.12), indicating TMS-induced prevention of the incubation effect. Furthermore, the incubation score, defined as the difference in cocaine seeking between abstinence day 21 and d1, was significantly higher in the sham vs the active TMS group (two-tailed t-test, p = 0.004).

We conducted voxel-wise functional connectivity analysis using the anterior prelimbic cortex as the seed region (A-P: +4.8 mm), followed by linear mixed-effects (LME) modeling with two fixed effects: hdTBS type (active vs. sham) and time point (withdrawal day 2 vs. day 22). Three brain regions—right prelimbic cortex, right bed nucleus of the stria terminalis (BNST), and left thalamic nuclei—showed a significant interaction effect (p < 0.05, corrected for multiple comparisons).

We next compared functional connectivity between early and late abstinence phases. In rats receiving sham TMS, a significant reduction in functional connectivity was observed across multiple brain regions, including orbital frontal cortex, cingulate cortex, prelimbic cortex, and nucleus accumbens. In contrast, in rats receiving active TMS, no significant reduction in functional connectivity was observed, suggesting that the TMS treatment normalized functional connectivity.

Conclusions: Our results showed that daily hdTBS sessions for 7 days during abstinence prevented the incubation of cocaine seeking and restored incubation-related decreases in functional connectivity. To the best of our knowledge, this is the first preclinical study showing the therapeutic potential of TMS for addiction with coil focality and stimulation parameters closely mirroring those used in clinical applications. As TMS leading form of neuromodulation treatment for depression and other neuropsychiatric disorders with an established safety profile, our results open novel possibilities for brain circuit-based cocaine addiction therapy.

Keywords: repetitive transcranial magnetic stimulation (rTMS), cocaine, Incubation of drug craving, Functional MRI (fMRI), TMS

Disclosure: Nothing to disclose.

P404. Sex differences in cue-induced fentanyl seeking and R-ketamine efficacy across withdrawal

Eleanor Towers, Ivy L. Williams, Raiya Nathan, Reena R. Sayani, Abhinaya Kodali, Richard Paget-Brown, Morgan A. A. Gatti, Wendy J. Lynch

University of Virginia, Charlottesville, Virginia, United States

Background: Women develop substance use disorders and other drug-related health consequences after fewer years of drug use compared to men, a pattern known as the telescoping effect. This phenomenon has been consistently observed across epidemiological studies of stimulants, alcohol, and opioids and replicated in preclinical models with cocaine. We recently reported similar findings in a rat model of opioid use disorder, with females developing an enhanced motivation for fentanyl and adverse health consequences more readily than males. The current study aimed to determine whether cue-induced fentanyl seeking, another key feature of addiction, also emerges earlier in females than in males. Additionally, we evaluated the efficacy of R-ketamine, an NMDA receptor antagonist hypothesized to attenuate drug seeking during periods of heightened craving, when glutamate signaling is presumed to be elevated.

Methods: Sexually mature male (N = 17) and female (N = 21) Sprague-Dawley rats were trained to self-administer fentanyl (0.25 µg/kg/infusion; fixed-ratio 1). Following acquisition, rats completed 10 days of extended, intermittent-access fentanyl self-administration (24 hr/day; 2 trials/hr, 5 min each). Immediately after the final session, rats received an acute injection of R-ketamine (20 mg/kg, i.p.) or vehicle (water). Cue-induced fentanyl seeking was assessed in a 2-hr session during early withdrawal (day 2), to capture potential sex differences in the emergence of drug seeking, or protracted (day 14) withdrawal, to evaluate both the persistence of R-ketamine and possible incubation of craving.

Results: Among controls, females exhibited greater cue-induced fentanyl seeking during early withdrawal compared to males (sex × withdrawal timepoint × treatment group: F₁,₂₇ = 5.3, P < 0.05). By protracted withdrawal, responding was similar between control females and males (P > 0.05), driven by increased seeking from early to protracted withdrawal in males. Preliminary findings indicated a tendency for R-ketamine to decrease responding in females tested during early withdrawal (P = 0.08), but this effect was transient and absent by protracted withdrawal. In males, R-ketamine was ineffective at both withdrawal time points.

Conclusions: These findings provide additional evidence for a biological basis of the telescoping effect across drug classes, with females showing heightened cue-induced fentanyl seeking earlier in withdrawal and males showing delayed increases in fentanyl seeking. R-ketamine reduced fentanyl seeking in females when craving, and likely glutamate signaling, was high during early withdrawal, but was ineffective in males, who exhibited lower craving at that time. These differential behavioral and pharmacological effects highlight sex-specific differences in the temporal trajectory of disease progression and the underlying neurobiological mechanisms.

Keywords: Incubation of craving, Fentanyl self-administration, Opioid use disorder, Sex differences, R-Ketamine

Disclosure: Nothing to disclose.

P405. Endocannabinoid mediated ethanol effects on striatopallidal synapses

Yasmeen Lowe, Emilce Artur de la Villarmois, Anthony James, Shana Augustin

Northwestern University, Chicago, Illinois, United States

Background: The endocannabinoid system, which comprises endocannabinoids and CB1 receptors, plays a crucial role in regulating synaptic transmission. CB1 receptors have been implicated in the rewarding effects of misused substances like alcohol. CB1 receptors are expressed throughout the brain within basal ganglia structures like the striatum and external segment of the globus pallidus (GPe), where they can modulate GABA release at specific synapses. The GPe is comprised of numerous cell types, with parvalbumin neurons being one of the main cell types in the region. However, it is unclear whether ethanol interacts with the endocannabinoid system at striatopallidal synapses to contribute to ethanol reward.

Methods: In this study we used a fluorescent endocannabinoid biosensor, pharmacology, and slice photometry to evaluate ethanol induced changes to endocannabinoid mobilization at striatopallidal terminals. To determine which cell types in the GPe express endocannabinoid synthesis machinery, we used fluorescent in situ hybridization (FISH). Additionally, whole-cell patch-clamp recordings were performed to determine how the endocannabinoid system modulates the effects of ethanol on synaptic transmission in labeled GPe parvalbumin neurons at striatopallidal synapses using optogenetics.

Results: We found that ethanol decreases endocannabinoid mobilization at striatopallidal terminals, where endocannabinoid mobilization is predominantly 2-AG mediated. In the GPe, our FISH data show that parvalbumin neurons express the necessary molecular machinery to synthesize endocannabinoids. Using whole-cell patch-clamp recordings, we observed that CB1 receptor activation modulates both spontaneous synaptic activity and evoked inhibitory postsynaptic currents (IPSCs) in parvalbumin neurons. Interestingly, ethanol does not alter evoked IPSCs at striatopallidal synapses formed onto parvalbumin neurons, suggesting that another cell type in the GPe may contribute to the ethanol-induced changes in endocannabinoid mobilization.

Conclusions: Our findings suggest a mechanism through which activation of CB1 receptors at striatopallidal synapses regulate ethanol effects. Ongoing experiments aim to identify the specific GPe cell type involved in CB1 receptor–mediated ethanol transmission. This potential mechanism can provide insight into synapse specific contributions to ethanol related behaviors.

Keywords: endocannabinoid system, Alcohol, Globus Pallidus

Disclosure: Nothing to disclose.

P406. Sex-and withdrawal-dependent remodeling of the synaptic proteome in the nucleus accumbens and prefrontal cortex of mice after cocaine exposure

Yun Young Yim, Alexa LaBanca, Tamara Markovic, Rita Futamura, Arthur Godino, Tukit Lam, Matthew Rivera, Leanne Holt, Corrine Azizian, Brian Kipp, Veronica Kondev, Elizabeth Kahn, Angelica Minier-Toribio, Trevonn Gyles, Eric Nestler

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

Background: Substance use disorders (SUDs) remain a pressing public health challenge, with relapse driven by complex and poorly understood neurobiological processes. The nucleus accumbens (NAc) and medial prefrontal cortex (PFC) are key brain regions driving drug seeking behavior and relapse. While transcriptional adaptations have been widely studied, the synaptic proteome, which more directly reflects functional and structural plasticity, remains poorly characterized. Gaining insight into these proteomic changes may reveal novel therapeutic targets for SUD treatment. Synaptosomal preparations, which preserve both pre- and postsynaptic elements, provide a critical platform to study localized molecular adaptations at the synapse.

Methods: In this study, we used synaptic neuroproteomics to investigate cocaine-induced proteomic remodeling across sex, brain region, and withdrawal duration. Adult male and female C57BL/6J mice received daily intraperitoneal injections of cocaine (20 mg/kg) or saline for 7 days, followed by either 24 hours (acute withdrawal, 1WD) or 30 days (extended withdrawal, 30WD) of forced home-cage abstinence (n = 8 per condition). Synaptosomes were isolated from the NAc and PFC and analyzed via LC-MS/MS with label-free quantification at the Yale/NIDA Neuroproteomics Core. Differentially expressed proteins were identified via two-way ANOVA with Type III sums of squares in R. For ephexin1 validation and functional studies, adult male and female C57BL/6J mice were used (N = 4–20 per sex per experiment). Ephexin1 was overexpressed using AAV or knocked down using lentivirus in vivo, and Western blot, qPCR, and various rodent behavior assays, such as elevated plus maze, open field, social interaction, conditioned place preference (CPP), saccharin self-administration (SA), and cocaine SA assays, were performed to examine the role of ephexin1 in baseline behavior and cocaine-seeking. Statistical analyses were conducted using unpaired T test or two-way ANOVA, as appropriate.

Results: We observed robust, sex- and region-specific synaptic proteomic changes. At 1WD, male mice showed ~2.5-fold more differentially expressed proteins (minimum 20% fold change and p < 0.05) in the PFC, primarily downregulated. At 30WD, female mice exhibited ~2.5-fold more changes in the NAc compared to males. While most cocaine regulated proteins were neuron-derived, a significant subset was enriched in astrocytic and microglial compartments, indicating important non-neuronal contributions to synaptic plasticity. Among the differentially expressed proteins, we identified ephexin1, a guanine nucleotide exchange factor (GEF), as a key candidate. Ephexin1 was upregulated in both sexes at 1WD but at 30 WD, it remained upregulated in males and was downregulated in females in the NAc (Q = 0.05). To test its functional relevance, we manipulated ephexin1 and tested in baseline behavior and preclinical models of cocaine-seeking. Overexpression of ephexin1 in the NAc of male mice induced anxiety-like phenotypes and enhanced conditioned place preference (p < 0.05), while reducing cocaine seeking and intake in cocaine self-administration without affecting seeking for the natural rewards (saccharine self-administration). Conversely, knockdown of ephexin1 in the NAc of female mice reduced conditioned place preference (p < 0.01), whereas no significant effect observed in males. Additional behavioral studies on ephexin1 knockdown are ongoing to further clarify its role in relapse vulnerability.

Conclusions: Altogether, these findings identify ephexin1 as a key modulator of cocaine-induced synaptic plasticity and a potential target for relapse prevention. This work provides a high-resolution, systems-level framework for understanding how cocaine reshapes synaptic networks across sex, brain region, and withdrawal duration. Building on this foundation, future studies will leverage neuroproteomics, behavioral pharmacology, and genetic tools to uncover cell type– and circuit-specific proteome change driving relapse vulnerability and to guide the development of targeted therapeutic strategies for substance use disorders (SUDs). Furthermore, this experimental pipeline offers a broadly applicable approach for identifying novel molecular targets across psychiatric and neurological conditions with unmet therapeutic needs.

Keywords: brain proteomics, cocaine addiction, Nucleus Accumbens

Disclosure: Nothing to disclose.

P407. Probing the differential abuse liability and therapeutic potential of ketamine enantiomers

Marjorie Levinstein, Reece Budinich, Sedona Ewbank, Ingrid Schoenborn, Oscar Solís, Tommaso Di Ianni, Jordi Bonaventura, Carlos Zarate, Yavin Shaham, Raag Airan, Mike Michaelides

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

Background: (R,S)-ketamine, a racemic mixture made of (R) and (S) enantiomers, has been used for decades as an analgesic and anesthetic. More recently, it has been used as an antidepressant. Further, (S)-ketamine is an FDA-approved pharmaceutical for treatment-resistant depression and depression with suicidal ideation. (R)-ketamine has shown promising results in preclinical studies on treating depression-like behavior, showing greater and longer lasting antidepressant efficacy, and less psychotomimetic effects than (S)-ketamine. (R,S)-ketamine and its enantiomers are commonly regarded as noncompetitive N-methyl-D-aspartate receptor (NMDAR) antagonists, but they also interact with other receptors. (S)-ketamine has distinct pharmacokinetics and pharmacodynamics, and it has about fourfold higher affinities at NMDARs and μ opioid receptors (MORs) than (R)-ketamine.

Recently, our lab found that while rats will learn to self-administer (S)-ketamine at 0.5 mg/kg/infusion dose, they will not learn to self-administer (R)-ketamine at this dose. We also showed that this dose of (S)-ketamine was sufficient to increase brain activity in the nucleus accumbens (NAc). Further, we observed that repeated exposure of 20 mg/kg/day of (S)-ketamine had effects on MOR density and function. Here we delve into (R)-ketamine’s effects using equimolar and equipotent doses to compare to our previous (S)-ketamine studies. We also recently showed that this interaction with MOR drives (S)-ketamine’s abuse liability and that repeated administration of (S)-ketamine decreases subsequent heroin reinforcement. We now investigate (S)- and (R)-ketamine’s abilities to curb heroin seeking in rats.

Methods: Rats (n = 12) underwent (R)-ketamine intravenous self-administration (IVSA) training (2 mg/kg/infusion) and their IVSA dose-response curve was determined with and without the presence of the opioid receptor antagonist naltrexone (3 mg/kg, SC). Rats were then exposed to additional IVSA sessions prior to extinction training.

Functional ultrasound imaging (fUSI), a neuroimaging modality, was used to measure brain-wide changes in cerebral blood volume (CBV). Rats (n  =  9–10) were pretreated with saline or naltrexone (3 mg/kg, SC) followed by a single IV injection of saline, 0.5 mg/kg, or 2 mg/kg (R)-ketamine 15 min later while imaging the mPFC and NAc.

Rats were catheterized and underwent IVSA of heroin. They were then tested for heroin seeking and then divided into groups to receive 8 days of saline (n = 15), (S)-ketamine (20 mg/kg/day, IV, n = 10), or (R)-ketamine (20 mg/kg/day, IV, n = 6). 24 hours and one week after the last infusion, the rats again were tested for heroin seeking.

Fiber photometry was used to measure dopamine responses in the nucleus accumbens. Rats (n = 4) were implanted with optic fibers coated in a mixture of the dopamine sensor, AAV9-hsyn-GRABDA. While being recorded, rats were pretreated with naltrexone (3 mg/kg SC) or saline and infused with saline or (S)-ketamine (0.5 mg/kg, IV). Rats then received 8 days of (S)-ketamine (20 mg/kg/day, IV). 24 hours after the last infusion, rats were again recorded while receiving an infusion of (S)-ketamine (0.5 mg/kg, IV).

Results: Rats acquired (R)-ketamine IVSA at a unit dose 4 times that of S-ketamine (2 mg/kg/infusion) and showed an inverted U-shaped dose-response function with highest responding to a 1 mg/kg/infusion unit dose. Naltrexone (3 mg/kg, SC) significantly blunted this dose response at the unit doses of 1 mg/kg (P = 0.0375) and 0.5 mg/kg (P = 0.0012). MK-801 (0.1 mg/kg, SC) significantly decreased (R)-ketamine infusions at the peak dose of 1 mg/kg/infusion (P < 0.001). Additionally, rats extinguished on the first day dropping from 57 active lever presses on the previous session to 8.

Unlike with (S)-ketamine, we found no change in peak change of CBV in either the shell or the core of the NAc after a single IV infusion of either 0.5 (equimolar dose to (S)-ketamine) or 2 mg/kg (equipotent dose to (S)-ketamine) (R)-ketamine. Additionally, we found that in the cingulate area and the secondary motor cortex, 0.5 mg/kg (R)-ketamine did not cause a change in activity. However, the 2 mg/kg dose increased CBV significantly above baseline in Cg1 (P = 0.165) and above both baseline (P < 0.001) and the 0.5 mg/kg (P = 0.035) dose in M2. Interestingly, naltrexone pretreatment did not affect the peak of the activation in Cg1 or M2 at the 2 mg/kg dose; however, it elongated the time that the activation was above baseline and significantly increased the area under the curve in M2 (P = 0.019).

Rats acquired heroin IVSA at a unit dose of 100 μg/kg and increased their infusions once the unit dose was decreased to 50 μg/kg. Rats that received (S)-ketamine infusions significantly decreased their heroin seeking compared to rats infused with saline (P = 0.003); however, this difference was only present at the 24-hour test.

Preliminary findings of our fiber photometry experiments indicate an acute infusion of (S)-ketamine, equivalent to the IVSA training dose, may increase dopamine release in the NAc, as shown by an increase in the area under the curve. Additionally, this change in dopamine may be blunted when animals are pretreated with naltrexone.

Conclusions: Our findings indicate that (R)-ketamine has reinforcing effects at sufficiently high doses and these effects may be due to its effect on the opioid system with different regional brain activity patterns than (S)-ketamine. (S)-ketamine decreases heroin seeking, suggesting potential for treating opioid use disorder.

Keywords: R-ketamine, Esketamine, opioid use disorder, Intravenous self-administration

Disclosure: Nothing to disclose.

P408. Aberrant effects of metabolic state on cognitive control and effort valuation in bulimia nervosa

Laura Berner, Jiulin Dai, Blair Shevlin, Saren Seeley, Andrew Westbrook, Thalia Viranda, Maia Chester

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

Background: Prior research has assumed that bulimia nervosa (BN) is maintained by stable cognitive control deficits. However, the episodic nature of excessive intake alternating with restricted intake in BN suggests intermittent and aberrant control oscillations. In addition, emerging work suggests that failures to deploy control depend on decisions about whether exerting control is worth the costly cognitive effort. Here, we investigated how cognitive control deployment and control-related effort valuation may abnormally fluctuate across fasted and fed states in BN.

Methods: In Study 1, adult females with BN (n = 28) and group-matched healthy controls (HC; n = 28) completed a Stop Signal Task, a classic assessment of response inhibition, in two states: after a 16-hour fast and after a standardized meal. In Study 2, a partially overlapping sample (N = 35) completed the Cognitive Effort Discounting task in both states. The task assesses the discounting of cash offers to complete more demanding levels of an N-back. Steeper discounting indicates higher estimated effort costs. Sequential sampling models examined which aspects of the effort-related decision-making process may have been influenced by fasting and eating.

Results: Study 1: After eating, HC showed increased correct go responses (p = 0.004) and unchanged response inhibition, while the BN group showed worse performance: increased errors on both go (p < 0.001) and stop (p = 0.003) trials (Group x State x Trial type b = −0.27, p = 0.004). Study 2: The effects of food consumption in HC resembled those previously observed after dopamine-releasing pharmacological agents like methylphenidate: increased subjective values of cognitive effort. However, the BN group maintained low subjective values of cognitive effort after eating (Group x State b = 0.12, p = 0.006). This steeper discounting of cognitive effort in the fed state was linked to more frequent binge eating (z = −0.12, p = 0.034). Drift diffusion modeling of decisions on the final task phase revealed that while eating increased the healthy controls’ initial bias toward choosing options that required more cognitive effort, it increased the BN group’s initial bias towards low cognitive-effort options (Group x State b = −0.05, p = 0.037). In addition, regardless of state, the HC group considered the benefits of cognitive effort before its costs, while the BN group considered cognitive-effort costs first (p = 0.025).

Conclusions: Our results suggest that eating has aberrant effects on both cognitive control deployment and underlying effort-related computations in BN, perhaps perpetuating out-of-control binge/purge episodes that alternate with periods of dietary restriction. Metabolic state-specific interventions—including inhibitory control training in the fed state and interventions that increase the subjective value of cognitive effort, specifically after eating–may prove fruitful for BN.

Keywords: Computational Cognitive Neuroscience, Inhibitory control, Cognitive effort, Metabolic Psychiatry, bulimia nervosa

Disclosure: Juniver, Ltd., Consultant, Self

P409. Resting state functional connectivity correlates of perseverative negative thinking and anhedonia in adolescents with anorexia nervosa

Sasha Gorrell, Tracy Luks, Anne Collins, Daniel Le Grange, Tony Yang

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

Background: Only 40% of adolescents with anorexia nervosa (AN) achieve remission following evidence-based treatment. The refractory nature of AN has been attributed to dysregulation in reward processing, and cognitive features such as perseverative negative thinking (PNT) that might contribute to reward-related affective deficits, e.g., anhedonia. PNT and anhedonia are hallmark features of depression and other psychiatric disorders, but to date, neurobiological features underpinning these symptoms in AN remain unspecified. This study examined resting state brain activity with functional magnetic resonance imaging (rs-fMRI) among adolescents with AN and controls (HC). Compared to HC, those with AN were hypothesized to show increased connectivity in orbitofrontal cortex (OFC) and caudate, regions implicated in PNT and affective processing.

Methods: Thirty-one subjects (16% male; 16 AN; 15 HC; Mage[SD] = 15.5[1.3]) underwent 3T rs-fMRI and self-reported anhedonia and PNT with the Snaith-Hamilton Pleasure Scale, and Perseverative Thinking Questionnaire, respectively. Images were analyzed using the CONN toolbox. Regions-of-interest (ROIs) for bilateral OFC and caudate (AAL atlas) were used for seed-based connectivity analyses, controlling for age and BMI (FDR-cluster correction, p < 0.05).

Results: Group-based interaction effects indicated that among those with AN, greater R and L caudate connectivity was associated with lower anhedonia scores whereas for HC, greater connectivity was associated with higher anhedonia scores. Further, L OFC connectivity with L and R superior parietal lobe, precuneus, L supramarginal gyrus, and bilateral occipital cortex, showed an interaction between group and PNT. Specifically, for HC, greater connectivity was associated with higher PNT scores, while for those with AN, this connectivity was not associated with PNT scores.

Conclusions: Adolescents with AN differed from HCs in caudate and OFC connectivity, potentially relating to the role that these regions serve in integrating cognition and emotion. These preliminary data suggest a role for the OFC and cognitive-affective integration in neurobiological substrates of PNT and anhedonia among young people with AN. Results implicate a novel neurobiological model of eating pathology, with neural signatures similar to obsessive-compulsive disorder.

Keywords: anorexia nervosa, orbitofrontal cortex, Caudate, Adolescent, Eating disorders

Disclosure: Nothing to disclose.

P410. Amygdala and anterior cingulate cortex engagement during emotion conflict adaptation scales with binge-eating severity

Kelsey Hagan, Nandini Datta, Victoria Franco, Noam Weinbach, James Lock, Cara Bohon

Virginia Commonwealth University, Medical College of Virginia, Richmond, Virginia, United States

Background: Binge eating is characterized by consuming an unambiguously large quantity of food over a short period of time and experiencing a subjective loss of control over one’s eating. Recurrent binge eating is a hallmark diagnostic criterion for binge eating disorder and bulimia nervosa, which are prevalent and impairing psychiatric disorders. Classic theoretical conceptualizations posit that binge eating functions to regulate and/or escape aversive emotions and is negatively reinforced (and maintained) by providing temporary relief. Moreover, research has documented that individuals with binge-type eating disorders experience difficulties in adaptive emotion regulation. Emotion regulation can occur explicitly (through deliberate, conscious effort) or implicitly (without awareness); however, most research to date has focused on explicit emotion regulation, and little is known about implicit emotion regulation and its neural correlates among individuals with binge-type eating disorders.

Methods: Participants were 58 women (M = 27.13, SD = 4.94) with binge eating (n = 41) and without binge eating (n = 17). To assess binge-eating frequency and determine eating disorder diagnosis, participants completed the Eating Disorder Examination clinical interview and height and weight measurements with trained research staff. Body mass index (BMI) < 18.5 kg/m2 was exclusionary. All women with binge eating met criteria for a DSM-5 eating disorder (n = 18 bulimia nervosa; n = 15 binge eating disorder; n = 8 low-frequency bulimia nervosa or binge-eating disorder). Women with and without binge eating did not differ on BMI and age (ps > .05). Women with binge eating reported an average of 2.63 (SD = 1.67) weekly binge-eating episodes. To interrogate implicit emotion regulation and its underlying neural correlates, participants completed an emotional conflict task (ECT) while undergoing a functional magnetic resonance imaging (fMRI) scan. The ECT is a modified version of the classic color-word Stroop task wherein photographs of human facial expressions (happy or fearful) are overlaid with the word “happy” or “fear” in red text. The overlaid emotion word is congruent with the facial expression (e.g., the word “happy” over a happy face) or incongruent (e.g., the word “fear” over a happy face). Participants were instructed to ignore the word and, via button press, indicate whether the facial expression was happy or fearful. fMRI data were preprocessed with FMRIPrep. In FSL, first-level fMRI analyses were modeled on a trial-by-trial basis, with each trial coded according to the congruency of the current trial and the immediately preceding trial, yielding four conditions: congruent preceded by congruent (cC), congruent preceded by incongruent (iC), incongruent preceded by congruent (cl), and incongruent preceded by incongruent (iI). The primary contrasts of interest were iI > cI (indexing conflict adaptation) and iC > cC (capturing persistence of conflict-related processes from the previous trial). Binge-eating and non-binge-eating groups were compared using voxel-wise whole-brain analysis (Z > 3.1, FWE cluster significance threshold of p = .05) in FSL. Additionally, region of interest (ROI) analysis was conducted in FSL using Harvard-Oxford Atlas anatomical masks of the left and right amygdala, rostral anterior cingulate cortex (rACC), and caudal anterior cingulate cortex (cACC). Mean percent signal change was extracted from these anatomical ROIs during il > cl and iC > cC contrasts. Mean percent signal change was compared between binge-eating and non-binge-eating groups and correlated with weekly binge-eating frequency.

Results: Voxel-wise whole-brain analyses yielded no group differences in neural activation for either contrast during the ECT. Further, there no group differences in percent signal change in the amygdala, rACC, nor cACC were observed for the contrasts. Although there were no categorical group differences in percent signal change in these ROIs, binge-eating frequency scaled with percent signal change for the iC > cC (or persistence of conflict-related processes). Specifically, percent signal change in the left amygdala (rs = 0.401, p = 0.009), left rACC (rs = 0.477, p = 0.002), right rACC (rs = 0.432, p = 0.005), left cACC (rs = 0.317, p = 0.044), and right cACC (rs = 0.320, p = 0.041) for the iC > cC contrast was correlated with the average number of weekly binge-eating episodes. There were no significant correlations with percent signal change in these regions and weekly binge-eating episodes for the il > cI (conflict adaptation) contrast.

Conclusions: Although women with and without binge eating did not differ in neural response during conflict adaptation, dimensional analyses revealed that binge-eating frequency was positively associated with greater neural activation in the amygdala, rACC, and cACC for the iC > cC contrast, but not for the il > cl contrast. These findings suggest that binge-eating severity corresponds to difficulty flexibly disengaging affective and control-related circuitry following conflict, but not with the adaptive recruitment of these circuits during conflict. Taken together, these results elucidate the neural underpinnings of implicit emotion regulation in binge eating and dovetail prior findings that suggest rigidity in control-related processes and affect regulation in binge eating.

Keywords: Binge-eating disorder, Bulimia nervosa, Emotion regulation, Emotion Conflict Task, Task-based fMRI

Disclosure: Nothing to disclose.

P411. Cannabidiol in regulating eating disorder symptoms in anorexia nervosa: safety, pharmacokinetics, and symptom change

Neha Sahota, David Grelotti, Walter Kaye, Tyler Nguyen, Skylar Swindle, Megan Shott, Guido Frank

Keck School of Medicine University of Southern California, Los Angeles, California, United States

Background: Anorexia nervosa (AN) is a severe psychiatric disorder marked by an intense fear of gaining weight and persistent body dissatisfaction, both during periods of low body weight and following weight restoration. A possible target of pharmacological therapy to improve treatment outcomes in AN is the endocannabinoid system (ECS). Prior research into the ECS demonstrates potential therapeutic benefits, particularly with respect to anxiety, cognitive rigidity, and fear memory reconsolidation. This randomized controlled trial of cannabidiol (CBD) investigated safety, tolerability, pharmacokinetics, and symptomatic improvement in AN.

Methods: Women with AN or Atypical AN were recruited from the UCSD Eating Disorder center and the community. They were randomized to receive an FDA approved oral CBD extract known as Epidiolex (n = 16) or placebo solution (n = 16) over the course of 21 days. The dose was up-titrated weekly, starting from 1.25mg/kg twice daily in the first week, 3.75 mg/kg twice daily in the second, and 6.25mg/kg twice daily in the third week. The end of each week was marked by a study visit at which CBD and metabolite levels, liver function, severity of eating disorder, and depression anxiety symptoms were assessed.

Results: Age at baseline was similar between the CBD and placebo group (22.9 ± 2.8 years, vs 22.5 ± 3.5 years), as was body mass index (BMI, kg/m2, 20.1 ± 2.5 vs 19.4 ± 1.8). There were no significant differences between the experimental and placebo group with respect to baseline years of education, EDI-3 Drive for Thinness, EDI-3 Body Dysmorphia, STAI-State and Trait Anxiety levels, EDE-Q Global, Restraint, EDE-Q Eating Concern, comorbidity or medication use measures.

Concentrations of CBD and metabolites (OH-CBD, COOH-CBD, and AEA) were measured at 7, 14, and 21 days in the CBD and placebo group via venous blood samples. Plasma levels of CBD and its metabolites increased each week, reflecting the increased dose. As expected, there was no detectable level of CBD, OH-CBD, and COOH-CBD in the placebo group. Levels of the naturally occurring AEA were similar in both the CBD and placebo group (p = 0.14) throughout the course of the study. There were no significant correlations between weekly CBD, CBD metabolite or AEA levels and behavior outcome measures. The comparison of individuals with AN versus Atypical AN within study groups did provide significant results, but the study was not powered to provide those data.

Both groups showed modest decreases on the EDE-Q and FCQ scales, BDI and STAI over time. Repeated measures ANCOVA indicated a small but significant group by time interaction for BMI increase in favor of CBD (F = 3.039, p = 0.046, partial η² = 0.252). In the CBD group, mean BMI was 20.1 at baseline (SD = 0.7), 20.4 (0.7) at assessment visit 1, 20.4 (0.7) at visit 2, and 20.5 (0.7) at visit 3. By contrast, in the placebo group the mean BMI was 19.3 (0.7) at baseline, 19.3 (0.7) at visit 1, 19.4 (0.7) at visit 2, and 19.4 and (0.7) at visit 3. The difference in the BMI of the CBD group at assessment visits 1, 2, and 3 when compared to baseline BMI was considered significant (p = 0.002, p = 0.004, and p = 0.001, respectively). Wilk’s Lamda for the CBD group and BMI over time was 0.247, F = 7.167, p = 0.015, partial η² = 0.754, power 0.842. In contrast, for the placebo group, Wilk’s Lambda for BMI over time was 0.637, F = 1.328, p = 0.340, partial η² = 0.363, power 0.223. Effect sizes were large but nonsignificant for improvements favoring CBD with respect to FCQ-S Lack of Control (p = 0.063, partial η² = 0.233), EDE-Q shape concern (p = 0.132, partial η² = 0.185), and STAI (p = 0.146, partial η² = 0.178).

Generally, CBD was well tolerated. Acceptability for the treatment assessed through the Treatment Acceptability and Adherence Scale (TAAS) revealed that over the three visits, the mean TAAS score for the CBD group was 60.90 (SD = 5.20), and the placebo group had a mean TAAS score of 62.47 (SD = 4.49), which was not significantly different from the CBD group.

Conclusions: The results of this randomized, double-blind, placebo-controlled, pilot study indicate that CBD is well tolerated in the AN and Atypical AN population. The CBD group showed a dose-dependent increase in both CBD and its metabolites, supporting normal bioavailability. Additionally, CBD was associated with a small but significant increase in BMI over time compared to placebo and there was some indication of improving eating disorder psychopathology. The findings in this study offer a critical starting point for identifying the limitations and potential use of CBD for treating AN. As the first randomized trial of CBD in patients with AN, this study demonstrates the need for further research into the treatment before definitive clinical recommendations can be made. Future studies may need to focus on a larger sample size and longer course to make definitive conclusions about the mechanism through which CBD impacts AN pathophysiology as well as the maximal dose at which effects are still seen. Additionally, combining CBD with another treatment strategy (such as a specific dietary alteration) may be necessary to appreciate its full therapeutic capacity.

Keywords: anorexia nervosa, Cannabidiol, Eating disorders, CBD, Randomized-Controlled Trial

Disclosure: Nothing to disclose.

P412. Symptom recovery during ketogenic therapy in weight normalized anorexia nervosa

Guido Frank, Barbara Scolnick, Megan Shott, Jong Rho

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

Background: Anorexia nervosa (AN) is a severe psychiatric disorder marked by extreme food restriction, significant weight loss, and a distorted perception of being overweight. It is a chronic condition with a high risk of relapse, substantial disease burden, and considerable treatment costs. Even after weight restoration, individuals with AN frequently continue to experience intense fears of eating and weight gain and body dissatisfaction. These symptoms can be as severe — or even more pronounced — than during the underweight phase, increasing the risk of relapse. Recent theories suggest that the underlying mechanisms contributing to AN may involve metabolic abnormalities, leading some to propose that AN is a “metabolic disorder of psychological origin.” Several studies have identified metabolic irregularities in AN, including signs of increased oxidative stress. Recent genetic research indicates that metabolic traits may contribute to the risk of developing AN. A neurobiological model centered on energy homeostasis proposed that heightened anxiety and stress play central roles in the development and maintenance of the disorder, and that ketogenic therapy (KT) could improve symptoms. In this trial, we aimed to evaluate the effectiveness, tolerability and safety of KT in a larger sample of individuals with wrAN. Specifically, we hypothesized that the intervention would lead to fast improvement in eating disorder symptoms, and that symptom improvement could be related to biological markers of nutritional ketosis.

Methods: We recruited individuals aged 18 to 45 years with a history of AN (any subtype) according to DSM-5 criteria, but who were weight-restored at the time of study initiation (body mass index [BMI] > 17.5 kg/m2). To ensure significant illness burden, study participants were required to have an elevated Eating Disorder Examination Questionnaire (EDE-Q) Global score > 2.09 for an excellent discrimination validity for AN. Subjects were administered the structured clinical interview for DSM-5 diagnoses (SCID-5), and completed at the begin and end of study the Eating Disorder Examination Questionnaire (EDE-Q), Eating Disorder Inventory-3 (EDI-3), Beck Depression Inventory 2 (BDI), Spielberger State-Trait Anxiety Inventory (STAI) and Temperament and Character Inventory (TCI). For weekly symptom monitoring, we adapted the EDE-Q scoring for past week. The study dietician (SK) introduced participants to the KT, emphasizing three meals and two snacks per day with macronutrient targets of 70% fat, 20% protein, and 10% carbohydrates, promoting hunger- and satiety-based eating. Participants submitted post-meal photographs for dietitian review to estimate consumption. Nutritional ketosis was monitored via daily blood beta-hydroxybutyrate (BHB) measurements during the initial 14-day induction period, then weekly thereafter (Keto-Mojo, target range: 0.5–3.0 mmol/L). Weekly self-assessments—to monitor AN symptoms, mood, anxiety and suicidality, ketone levels, and weight—were conducted in the evening of the same weekday, prior to meeting with the psychiatrist and dietitian. Descriptive statistics and independent t-tests evaluated demographic variables between completer and non-completer individuals. Chi2 tests evaluated frequency of comorbid conditions and medication use. Repeated measures MANCOVA (intent-to-treat analysis, carrying forward data from individuals who dropped out of the study) tested weekly change over time, including comorbidity and medication use in the model. Pearson’s correlations analyzed relationships between demographic and behavioral and blood BHB data. Significant correlations were false discovery rate multiple comparisons corrected.

Results: Twenty-two individuals (100% female sex) started the KT, 18 of whom completed all 14 weeks (82%); two individuals dropped out after three, and two after five weeks in the study. Non-completers were younger and had a shorter duration of illness. Free and total carnitine was lower in completers. The repeated measures MANCOVA (Wilk’s lambda = 0.165, F = 2.383, p < 0.001), showed significant time effects for EDE-Q Global (F = 12.892, p < 0.001), Restraint (F = 5.081, p < 0.001), Eating Concern (F = 11.316, p < 0.001), Shape Concern (F = 11.443, p < 0.001) and Weight Concern (F = 8.425, p < 0.001), as well as BDI (F = 10.216, p < 0.001). Comorbid conditions and medication did not have significant effects.

A pre-post KT paired samples t-tests in completers indicated significant improvements in EDE-Q, EDI-3, CIA, BDI, STAI, and TCI Harm Avoidance. Baseline total carnitine correlated positively with baseline ketone levels (r = 0.436, p = 0.043). Seventy-two percent (13) of the completers (59% of the full sample) had an EDE-Q Global score in the recovered range at study end. A one-Way ANOVA explored whether state or trait behaviors (TCI, EDI-3, BDI and STAI) could differentiate groups. Participants who did not have normalized EDE-Q Global scores at end of study were characterized by low self-esteem (EDI-3).

Conclusions: This study indicates that KT in wnAN is well tolerated and associated with fast decrease in symptoms for both eating disorder and depression symptoms. Depression scores were improved in all, and EDE-Q and EDI-3 eating disorder symptom scores within normal range at study end in most study completers. Elevated low self-esteem scores distinguished those with poor AN symptom improvement.

Keywords: Anorexia nervosa, Ketogenic diet, Acute treatment, Energy metabolism., Recovery

Disclosure: Nothing to disclose.

P413. Shared genetic architecture between anorexia nervosa and metabolic markers

Carolina Makowski, Alexey Shadrin, Anders M. Dale, Ole A. Andreassen, Dennis van der Meer

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

Background: Anorexia Nervosa (AN) is a severe psychiatric disorder with high mortality rates and a chronic illness course in many individuals, yet existing treatments remain limited in their efficacy. Insights from genomics have shown that AN is highly heritable and has a prominent metabolic component, independent of its association with other psychiatric traits. Within this metabolic lens and given the diagnostic criterion of low body weight, Body Mass Index (BMI) in particular has garnered attention in studies aiming to understand the genetically-informed etiology of AN. However, it is still unclear how the biological factors shaping BMI may increase risk for AN and/or confound genetic signals with AN.

Methods: We leveraged two sets of Genome-Wide Association Studies including i) 207,836 individuals from the UK Biobank across 249 circulating plasma metabolites; and ii) 72,517 individuals (16,992 cases with AN) from the Psychiatric Genomics Consortium; both males and females were included. We computed genetic correlations using linkage disequilibrium score regression to assess the genetic overlap between metabolites and AN to commonly comorbid psychiatric conditions (e.g., anxiety, depression) and cardiometabolic traits (e.g., BMI, type II diabetes, coronary artery disease). To pinpoint shared genetic loci, we leveraged the comprehensive metabolite panel and applied conditional false discovery rate (cFDR) to discover shared genetic variants between AN and various metabolites, and in turn, identify shared genes and biological pathways through a gene ontology analysis. Finally, we used bidirectional Mendelian randomization (MR), to investigate the causal relationships between AN, BMI, and metabolic markers.

Results: Strong genetic correlations were found between AN with 249 circulating metabolites, which were opposite in direction to metabolic traits such as BMI and T2D (rs~−0.94). Anxiety, a commonly comorbid trait with AN, had weaker correlations with the metabolites, emphasizing the relatively stronger metabolic component of AN. CFDR revealed shared genetic architecture between AN and metabolic markers, particularly lipid-based metabolites, with 57.1% of shared variants having opposite effects on AN versus the metabolites. The highest number of overlapping genetic variants was found between AN and cholesteryl esters, and cholesterol and triglyceride relative lipoprotein lipid concentrations (FDR-corrected p-value < 0.05). AN-metabolite overlapping variants were mapped to genes involved in lipid-related cell signals, developmental growth, inflammation (e.g., cytokine production) and canonical cellular processes (e.g., tyrosine kinase signaling) (q < 0.05). MR analyses found no direct significant relationship between AN and metabolic markers. However, there were significant bidirectional causal relationships between 6 metabolites, including lipoproteins and fatty acids, and BMI (p < 0.05) and between AN and BMI (BMI to AN: β = −0.44, p = 1.06e-19; AN to BMI: β = −0.06, p = 4.26e-4), suggesting that metabolites may exert their influence on AN through BMI, and vice versa.

Conclusions: Our findings point to strong associations between AN and metabolic markers, similar in strength but opposite in direction to other cardiometabolic traits. Our results suggest that developmental and lipid-based biological processes, as well as a mediating role of BMI, may be underlying the relationship between AN and metabolic markers. The shared genetic architecture and biological pathways between AN and metabolites offer a novel perspective on the complex etiology of AN and provide new research directions for treatment targets and weight restoration approaches.

Keywords: Eating disorders, Genomics, Metabolic Psychiatry, genome wide association study, anorexia nervosa

Disclosure: Nothing to disclose.

P414. Establishing the impact of a high-fat/high-sugar diet on drug preference in a rodent model of binge eating

Kimberlei Richardson, Asia Smith, Haley Warren, Leikwaivion Davis, Alexa Ryan, Alyssa Roach

Howard University, Washington, District of Columbia, United States

Background: There is incongruency in the literature concerning which drugs of abuse may be preferentially consumed by individuals with eating disorders. By gaining a better understanding for how increased consumption of high-fat, high-sugar diets may affect the reward pathway and influence the vulnerability to the use of drugs of abuse, we seek to investigate how a chronic high-fat, high-sugar diet impacts drug reward. The purpose of this study is to determine whether rats with varying feeding phenotypes for high fat, high sugar pellets have a distinct drug preference (morphine versus cocaine) and whether the nutritional state of rats increases drug preference scores.

Methods: Female Sprague Dawley rats (n = 38) were given standard chow ad libitum and 30g of palatable food (PF) for 4 hours. After nine intermittent tests, the feeding phenotypes were established through evaluation of the PF consumed across multiple tests. Rats in the upper tertile were labeled as high preference (HP) and the rats in the lower tertile were labeled as low preference (LP), the remaining rats were considered neutral and used further in the study. For three days, rats (n = 7–8/group) were alternatively injected with morphine (10mg/kg, i.p.) and saline or injections of cocaine (10mg/kg, i.p.) and saline until the conditioning period concluded. In a separate group, rats (n = 7–8/group) received alternative injections of cocaine (10mg/kg, i.p.) and saline for seven days, and then underwent the feeding tests to establish phenotypes. The preference tests were conducted after drug conditioning.

Results: HP rats consistently consumed significantly more PF than LP rats during the nine feeding tests (p < 0.0001). There was a significant difference in drug preference for cocaine versus morphine in rats with prior PF intake (p = 0.0123). Rats that received cocaine CPP training before PF exposure displayed lower CPP scores than rats that received no CPP conditioning before PF exposure.

HP rats displayed a higher cocaine preference score than LP and Chow rats (p = 0.0559).

Conclusions: Rats exposed to cocaine prior to PF had a lower preference for the drug, compared to rats that received PF before CPP. This finding suggests that the order of reward exposures (food vs drug) may affect preference for the drug. It was concluded that prior exposure to a high fat, high sugar diet increased preference for cocaine versus morphine in a rodent model of binge eating.

Keywords: conditioned place preference, Binge Eating, morphine, cocaine

Disclosure: Nothing to disclose.

P415. Ambulatory parasympathetic and sympathetic psychophysiological associations with self-reported overwhelming sensory experiences, masking, and stimming behavior in autistic adults over 14 Days

Lauren Bylsma, Xin Hu, Caitlin Conner, Carla Mazefsky, Lori Scott

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

Background: Sensitivity to sensory stimulation, masking, or stimming behavior are common among autistic adults who often report that overstimulation leads to dysregulation. Autistic adults also often report that stimming (i.e., repetitive or rhythmic movements involving one or more senses) is self-regulating, while masking (i.e., conscious or unconscious efforts to hide autistic traits, like stimming) is experienced as depleting. High frequency heart rate variability (HF-HRV), a measure of parasympathetic activity associated with emotion regulation capacity, may reflect depletion of self-regulation capacity associated with sensory overstimulation and masking, as well as increases in regulation capacity associated with stimming. Mean skin conductance level (SCL) may be useful as an objective marker of sympathetic arousal associated with dysregulation during periods of overstimulation. However, the psychophysiological correlates of these experiences and behaviors in the daily life of autistic adults is unknown. We examined psychophysiological associations with daily life sensory experiences, stimming, and masking behavior among autistic adults using ecological momentary assessment (EMA) with ambulatory psychophysiology. We predicted that (1) lower HF-HRV and higher mean SCL would be associated with report of recent overwhelming sensory experiences, (2) stimming behavior would be more likely to occur during periods of dysregulation (e.g., lower HF-HRV and/or higher mean SCL) and would be associated with subsequent increases in HF-HRV and reductions in mean SCL, and (3) masking behavior would be associated with reductions in HF-HRV and increases in mean SCL.

Methods: Participants were recruited from the Pittsburgh community as part of an ongoing Autism Center of Excellence grant (MH130957). Participants in the current analyses include 102 autistic adults (age = 31.32 ± 9.07 years, 65.6% biological female, 32.3% biological male) who completed a 14-day ecological momentary assessment (EMA) using study-provided Android smartphones with concurrent ambulatory psychophysiological monitoring using Movisens ECGMove4 and EDAMove4 sensors. Participants received 8 prompts during their waking hours using the MovisensXS app where they self-reported recent overwhelming sensory experiences as well as masking and stimming behaviors over the last 60 minutes. During the same period, participants also wore the electrocardiogram (ECG) sensor using a chest strap or chest electrodes to assess HF-HRV, as well as the electrodermal (EDA) sensor on the instep of their foot to assess skin conductance level (SCL). Mixed-effects models examined within-individual associations between self-reported sensory experiences, masking, and stimming with psychophysiological indices derived from the 60 minute periods preceding and subsequent to each prompt. Age and biological sex were considered as co-variates.

Results: As expected, recent overwhelming sensory experiences in the past 60 minutes predicted a significant reduction in HF-HRV from the 60 minutes pre to post prompt (p = 0.029, t = −2.21). Recent overwhelming sensory experiences also predicted higher mean SCL both pre (p < 0.001, t = 2.98) and post prompt (p = 0.002, t = 3.2), but these experiences did not predict change in SCL from pre to post. Lower HF-HRV and greater mean SCL in the pre 60 minute period both independently predicted greater report of stimming in the same time period, controlling for recent sensory experiences (p = 0.013, t = −2.53; p = 0.006, t = 2.78). Stimming behavior predicted greater reductions in mean SCL from the 30 minutes pre to 60 minutes post period (p = 0.041, t = −2.07), but stimming did not predict changes in HF-HRV over the same interval. Masking predicted greater reductions in HF-HRV from pre to post 60 minutes (p = 0.015, t = −2.5), but there was no effect of masking on changes in mean SCL. Results were consistent with and without co-variates.

Conclusions: This is the first study to examine ambulatory psychophysiology with self-reported sensory experiences, stimming, and masking in the daily life of autistic adults. As expected, recent report of overwhelming sensory experiences was associated with reductions in parasympathetic activity and greater sympathetic arousal, indicating that these experiences are linked to physiological dysregulation for autistic adults. Lower parasympathetic and higher sympathetic activity in the past hour, indicative of greater dysregulation, was associated with greater stimming in the same time period, which supports the idea that autistic adults engage in more stimming when experiencing higher levels of dysregulation. Partially consistent with predictions, stimming also predicted greater reductions in sympathetic arousal over time (but not changes in parasympathetic activity), indicating support for the purported regulatory effect of stimming behavior. Furthermore, masking behavior did predict reductions in HF-HRV, which is consistent with autistic individuals typically reporting masking to be depleting. These findings have important conceptual and methodological implications for understanding physiological processes underlying sensory experiences and regulatory behaviors in autistic adults. Results are preliminary as data are still being collected, with a target sample size of 200 autistic adults. Future analyses will also explore effects in other time windows using longer lags and compare results with a non-autistic sample also being collected.

Keywords: autism, psychophysiology, heart rate variability, emotion regulation, EMA

Disclosure: Nothing to disclose.

P416. Beta-endorphin release, subjective pain, and affective change during nonsuicidal self-injury: a real-world test of endogenous opioid models

Johanna Hepp, Lisa M Störkel, Ryan W Carpenter, Christian Schmahl, Inga Niedtfeld

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

Background: Nonsuicidal self-injury (NSSI) is a highly prevalent and impairing behavior, with lifetime prevalence of up to 17% in adolescents and approximately 5% in adults. In clinical populations, particularly those with mood, personality, or trauma-related disorders, rates are substantially higher, reaching 40–80% depending on diagnosis and setting. While traditionally conceptualized as a behavioral response to affective dysregulation, emerging biological models posit that NSSI may engage the endogenous opioid system (EOS), especially β-endorphin release, as a mechanism of short-term relief. Laboratory studies have shown altered pain processing and attenuated baseline β-endorphin in individuals with NSSI, yet real-world evidence linking opioid activity, subjective pain appraisal, and affect regulation is lacking. Initial findings suggest a hypo-opioidergic state prior to NSSI and a post-injury rebound in β-endorphin. However, little is known about how momentary pain intensity and its cognitive-affective appraisal contribute to the emotional relief that reinforces NSSI. This study investigates the interplay of β-endorphin release, subjective pain appraisal, and affective benefit following NSSI in daily life.

Methods: Fifty-one adult cis-women with DSM-5 NSSI disorder participated in a 15-day ambulatory assessment protocol. Participants completed five semi-random prompts per day and self-initiated reports following NSSI acts. Saliva samples were collected during high-urge states, directly after NSSI and at a 10, 20, and 30 minute follow-up and, if possible, preceding NSSI engagement. This yielded a total of 1054 valid β-endorphin assays and 136 tissue-damaging NSSI episodes. Each report included ratings of pain intensity, pain appraisal (pleasant vs. unpleasant), injury severity, and momentary affect as well as state dissociation. Multilevel linear models with random intercepts and slopes were used to analyze within-person associations between β-endorphin, pain, and affect regulation trajectories following NSSI.

Results: Consistent with a hypo-opioidergic model, β-endorphin levels were significantly lower before NSSI acts than on non-NSSI days (p = 0.004), and following NSSI, β-endorphin levels increased significantly (Est. = 0.62, SE = 0.20, p = 0.032; Cohen’s d = 0.82), with higher injury severity predicting greater increases (Est. = 0.39, SE = 0.15, p = 0.009, Störkel et al., 2021). The momentary association between β-endorphin levels and pain intensity was not significant (p = .199), suggesting a possible dissociation between peripheral opioid release and nociceptive experience. Subjective pain levels during NSSI were low (M = 2.63, SD = 2.08) and 17.7% of NSSI events entailed complete subjective analgesia. In contrast, NSSI pain was generally appraised as pleasant (M = 5.31, SD = 3.04), with 81.6% of pain episodes falling in the pleasant range. Multilevel models showed that greater pain unpleasantness predicted steeper declines in negative affect in the 30 minutes following NSSI (Est. = −0.12, SE = 0.05, p = .014), with pain intensity showing a trend-level effect (Est. = −0.09, SE = 0.05, p = 0.068), suggesting pain appraisal as a possible affective moderator. Dissociative symptoms were frequently endorsed but of low intensity and were not significantly associated with β-endorphin release (p = .074) or affective recovery when controlling for injury severity.

Conclusions: This study extends previous work on the biological underpinnings of NSSI by linking peripheral opioid activity and subjective pain experience to affective change in real time. While β-endorphin levels increased following NSSI, particularly after more severe injuries, subjective pain was often minimal and frequently appraised as pleasant. Exploratory models suggest that how pain is cognitively appraised, rather than its intensity alone, may shape affective relief following NSSI. These findings point to a potential role for endogenous opioid signaling in reinforcing self-injury and highlight pain appraisal as a possible target for future mechanistic and intervention studies.

Keywords: Non-Suicidal Self-injury (NSSI), affective components of pain, Ambulatory Assessment, pain analgesia

Disclosure: Merck Healthcare KGaA, Employee, Spouse/Partner

P417. Outdoor air pollution exposure and fear processing in adolescents: evidence for heightened reactivity and altered prefrontal engagement

Clara Zundel, Samantha Ely, Reem Tamimi, Hilary Marusak

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

Background: Outdoor air pollution, including fine particulate matter (PM2.5), has been linked to adverse neurodevelopmental and mental health outcomes, including heightened anxiety symptoms in youth. Adolescence represents a sensitive window for both pollutant exposure and psychiatric risk. Impairments in fear extinction—a process central to cognitive behavioral therapy—may be a key mechanism linking air pollution to anxiety. We examined whether past-month PM2.5 exposure was associated with impairments in fear extinction learning and its later recall in youth.

Methods: One hundred and two youth (10–17 yrs, 55% female) completed a two-day fear extinction task; a subset completed the task during functional magnetic resonance imaging (N = 68). On Day 1, participants underwent fear conditioning with two conditioned stimuli paired with an adverse white noise burst (CS + E, CS + U) and one unpaired stimulus (CS-), followed by extinction of the CS + E. On Day 2, 24 hours later, extinction recall was tested with all cues presented without reinforcement. Past-month ambient PM2.5 concentrations (mean = 9.1 ug/m3) were estimated for each participant’s residential address using a random forest machine learning model and dichotomized using a median split (low vs. high exposure). Fear indicators included skin conductance response (SCR), avoidance behavior, and fear ratings. Linear mixed-effects models tested PM2.5 group differences in fear indicators over time and by conditioned stimulus type (CS), adjusting for age, sex, and neighborhood deprivation. For fMRI analyses, linear regression models adjusted for age and sex examined associations between PM group and brain activation to the CS + E during extinction learning and recall. A whole-brain threshold was applied (p < 0.005, k > =10).

Results: Adolescents with higher exposure exhibited greater SCRs across stimuli during fear conditioning (F(1, 49) = 5.23, p = 0.026) and greater avoidance behavior during extinction learning, regardless of CS type (F(1, 99) = 5.15, p = 0.025). No group differences were observed between late extinction learning and early extinction recall. However, across extinction recall, adolescents with higher exposure maintained high fear ratings, whereas those with lower exposure showed a significant decrease over time (F(1, 493.08) = 4.63, p = 0.032). Neuroimaging analyses revealed greater right dorsolateral prefrontal cortex (PFC) activation in the high exposure group during fear conditioning (xyz = 49, 23, 27; z = 4.62, k = 20) and during extinction learning for CS + E trials (xyz = 9, 43,44; z = 4.15, k = 16). During extinction recall, youth with higher exposure also showed greater activation in the right anterior PFC for CS + E trials (xyz = 46, 46, 21; z = 3.62, k = 38).

Conclusions: Higher PM2.5 exposure was associated with heightened physiological and behavioral reactivity during conditioning and persistent fear during extinction recall. Altered recruitment of prefrontal regulatory circuity across conditioning, extinction, and recall in high-exposure youth may reflect compensatory engagement that is insufficient to support successful fear reduction. This pattern parallels compensatory PFC activation observed in adolescents exposed to other chemical toxicants (e.g., heavy metals) and trauma. Together, this pattern suggests disrupted fear regulation is a neurobiological pathway through which outdoor air pollution may increase risk for anxiety in adolescence.

Keywords: Air Pollution, Adolescence, Fear conditioning and extinction, prefrontal cortex, Adolescent Anxiety

Disclosure: Nothing to disclose.

P418. Childhood maltreatment and cortisol levels influence changes in emotion dysregulation over pregnancy

Vasiliki Michopoulos, Rebecca Lipschutz, Mariana Rocha, Alicia Smith, Abigail Powers

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

Background: Black pregnant females experience disproportionately higher rates of childhood maltreatment and resulting adverse mental health sequalae, including depression and posttraumatic stress disorder (PTSD). One mechanism by which childhood maltreatment contributes to trauma-related psychopathology is through heightened emotion dysregulation, which is linked to activity in the hypothalamic pituitary adrenal (HPA) axis. Neuroendocrine changes induced by pregnancy may also play a key role in modulating emotion dysregulation in women exposed to childhood maltreatment. More specifically, levels of cortisol, which typically increase during pregnancy, may influence how emotion dysregulation changes over gestation. The current study assessed how emotion dysregulation changes over pregnancy and how childhood maltreatment and cortisol independently impact the trajectory of emotion dysregulation in a sample of under-resourced pregnant Black females. We hypothesized that childhood trauma would influence the change in emotion dysregulation over pregnancy and that changes in emotion dysregulation over pregnancy would be related to cortisol levels during gestation.

Methods: Pregnant female participants (N = 260) were recruited and enrolled between 2018 and 2022 from the obstetrics and gynecological clinic at Grady Memorial Hospital, a publicly funded hospital primarily serving minoritized individuals with few economic resources in downtown Atlanta, Georgia. All study procedures were approved by the Emory Institutional Review Board and the Grady Hospital Research Oversight Committee. Participants completed up to three interviews assessing PTSD symptoms over pregnancy; the enrollment interview assessed exposure to childhood maltreatment via the Childhood Trauma Questionnaire (CTQ). Cortisol concentrations were measured in a subset of participants (N = 90) who provided venous blood samples that were collected in EDTA tubes by trained research staff using standard techniques. Plasma cortisol concentrations were quantified with liquid chromatography-triple quadrupole tandem mass spectrometry (LC-MS/MS). To determine the effect of childhood maltreatment on changes in emotion dysregulation over the course of pregnancy, we conducted conditional growth curve mixed model where time was defined by weeks’ gestation at the respective study visit (calculated using pregnancy due date in each participant’s medical record) with the model intercept set to conception. Conditional models also examined the effect of cortisol on emotion dysregulation over time, measured in weeks gestation. Post hoc simple slopes analyses examined conditional associations at the mean and +- 1SD. Maternal age, recent stressors, and parity were included as covariates. Growth curve models were conducted using linear mixed models in Jamovi(2.5) and alpha levels for analyses were set to p < 0.05.

Results: Childhood maltreatment exposure was associated with higher initial emotion dysregulation (B = 0.369, p < 0.001). Childhood maltreatment significantly moderated the change in emotion dysregulation over gestation (B = −0.009, p = 0.008), such that greater levels of childhood maltreatment were associated with a more negative slope in emotion dysregulation over pregnancy. Post-hoc simple slopes analyses showed that individuals with higher levels of childhood maltreatment showed steeper declines in emotion dysregulation over the course of pregnancy (p’s < 0.001). Individuals with no/low levels of childhood maltreatment did not show changes in emotion dysregulation over pregnancy (p = 0.241). Cortisol significantly moderated the trajectory of emotion dysregulation (B = 0.042, p = 0.040), such that lower levels of cortisol were associated with a more negative slope in emotion dysregulation. Post-hoc simple slopes analyses showed that individuals with the lowest levels of cortisol experienced steeper declines in emotion dysregulation over the course of pregnancy (p = 0.002). Individuals with the highest levels of cortisol did not show changes in emotion dysregulation over pregnancy (p = 0.907).

Conclusions: The current findings indicate that individuals exposed to high levels of childhood maltreatment demonstrate higher levels of emotion dysregulation symptoms early in pregnancy and suggest that emotion dysregulation decreases from early to late gestation. Our findings also suggest that lower levels of cortisol during gestation may contribute to decreases in emotion dysregulation over pregnancy. Overall, the findings provide evidence that cortisol childhood trauma influences the change in emotion dysregulation over pregnancy and that changes in emotion dysregulation over pregnancy are related to cortisol levels during gestation.

Keywords: pregnancy, Emotional dysregulation, Childhood Adversity

Disclosure: Nothing to disclose.

P419. Oral contraceptive use, emotion regulation and depressive symptoms: the importance of going beyond the one-size-fits all and considering parameters of use

Juliette Fortier, Alexandra Brouillard, Lisa-Marie Davignon, Marie-France Marin

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

Background: The use of oral contraceptives has been associated with emotional regulation difficulties and higher levels of psychological distress, notably through higher risk for depression. However, the effects remain highly heterogeneous, suggesting that different factors may modulate these associations. Notably, our team and others have shown that age of initiation and duration of use may influence some of these effects. More recently, researchers have suggested that the hormonal composition of oral contraceptives, such as the hormonal components’ dose, is also an important factor future studies should consider. Considering that estradiol plays an important modulatory role on emotional processes, it is likely that that ethinyl estradiol (synthetic estradiol), and more specifically its dose, is a key variable in understanding how oral contraceptives affect emotional regulation and psychological distress. We therefore tested whether age of initiation, duration of use, and dose of ethinyl estradiol contributed to emotional regulation and depressive symptoms.

Methods: Using data from a larger research project, we narrowed our analyses on women currently using oral contraceptives (n = 71) aged between 23 and 35 years old ($\bar{{{\rm{x}}}}$ = 26.24, s = 2.87). None of the women had a past or current diagnosis of psychopathology and none were taking other medication. Emotional regulation was assessed using the Emotion Regulation Questionnaire, which yields a score on cognitive reappraisal and emotional suppression, the former generally considered as the most evolved emotion regulation strategy. Depressive symptoms were assessed with the Beck Depression Inventory. Three linear regressions were performed (one per dependent variable: cognitive reappraisal, emotional suppression, depressive symptoms), with age and education as covariates, age of initiation, duration of use, and dose of ethinyl estradiol as predictors, as well as 2-way interactions between these predictors.

Results: For cognitive reappraisal, the regression model was not significant, F(8, 58) = 1.02, p = 0.432. For emotional suppression, the model was significant, F(8, 58) = 2.11, p = 0.049; adjusted R² = 0.119. More precisely, education was negatively associated with emotional suppression, β = −0.278, t = −1.98, p = 0.053, while ethinyl estradiol dose was positively associated with emotional suppression, β = 0.371, t = 2.52, p = 0.015. Finally, for depressive symptoms, the model was significant, F(8, 52) = 2.25, p = 0.038; adjusted R² = 0.143). Dose was positively associated with symptoms (β = 0.261, t = 1.99, p = 0.052), and an interaction between age of initiation and ethinyl estradiol dose was also obtained (β = 0.342, t = 1.95, p = 0.057). Simple slopes showed that when oral contraceptive use was initiated early (−1 SD; 14.6 years old), the ethinyl estradiol dose had a positive and significant effect (β = 0.599, t = 2.53, p = .015). However, when oral contraceptive use was initiated later (+1 SD; 19.0 years old), the relationship between the dose of ethinyl estradiol and depressive symptoms was no longer significant (p = 0.695).

Conclusions: Overall, our results suggest that higher dose of ethinyl estradiol may contribute to less adaptive emotional regulation strategies and higher depressive symptoms, even in a non-clinical population. Moreover, initiating oral contraceptive use early may accentuate the deleterious effect of dose. The results highlight the importance of considering factors intrinsic to the use of the oral contraceptives to better nuance their effects and more accurately guide prescribing recommendations.

Keywords: hormonal contraceptive use, age of initiation, ethinyl estradiol, emotion regulation, depressive symptoms

Disclosure: Nothing to disclose.

P420. Exploring neural correlates for the culture of dissemblance, an American black feminist concept examining sociopolitical effects on emotional processing in black women

Linzie Taylor, Trinidi Prochaska, Amanda Arnold, Amy Murphy, Megan E. Huibregtse, Angela Li, Abigail Powers, Negar Fani, Vasiliki Michopoulos, Jennifer S. Stevens

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

Background: The culture of dissemblance is a widely recognized Black feminist concept that describes how sociopolitical positioning places Black women at a greater risk of being exposed to psychosocial stress while dismissing their emotionality. This culturally learned masking of emotional expression may contribute to mood and stress disorders in Black women. Suppression is an emotional regulatory strategy that masks one’s behavioral emotional response. Emotional regulatory literature suggests that habitual suppression reduces behavioral expression but increases physiological response, potentially impacting health over time. Additionally, African Americans report habitual use of suppression strategies more than White counterparts, prompting a need to explore if there is overlap between neural regions engaged in emotional suppression and neural correlates of the culture of dissemblance. The current study hypothesized that social visibility during emotional contexts influences neural engagement in regions involved in emotional suppression (vmPFC/amygdala).

Methods: Forty Black women (ages 40–55) completed an adapted International Affective Picture System task with “watched” and “unwatched” conditions to simulate demands for dissemblance. In watched conditions, participants were told they were being watched with a video camera previously shown to them and there was a red recording symbol present to indicate the camera was turned on. In the unwatched conditions, participants were verbally notified that the camera was off and the red recording symbol was no longer present. Participants later reported how watched they felt during the task as a manipulation check. Participants viewed negative, positive, and neutral images while undergoing fMRI. Self-report measures assessed habitual emotion regulation strategies, discrimination experiences, Black identity associations, and extroversion levels. BOLD responses in the vmPFC and amygdala were analyzed using contrast analyses for watched vs. unwatched conditions across emotional valences.

Results: Significant effects of the watched condition on BOLD response emerged in the right (B = 0.40, p < 0.005) and left amygdala (B = 0.37, p < 0.005) for negative > neutral scenes only. The unwatched condition showed significant effects for negative > neutral in bilateral amygdala (right: B = 0.32, p < 0.005; left: B = 0.24, p = 0.001) and vmPFC (B = −0.28, p = 0.03). Women reported comparable use of both suppression (range: 17.8 – 89.2, mean: 48, sd: 9.3) and reappraisal scores (range: 22.8 – 60, mean: 49.6, sd: 17.1). Women reporting greater suppression strategies compared to reappraisal showed lesser amygdala activation during watched > unwatched negative scenes compared to neutral (right: B = 0.18, p = 0.04; left: B = 0.21, p = 0.04).

Conclusions: Our findings suggest that the culture of dissemblance may engage similar neural mechanisms as emotional suppression during highly visible and negative emotionally salient contexts. Women who report greater habitual use of suppression compared to reappraisal demonstrated greater down regulation of amygdala reactivity during watched > unwatched for negative > neutral scenes. Suggesting that suppression may be an effective strategy for Black women experiencing negative emotions when external perception is high. This is a novel exploration of how a black feminist interpretations of emotion engagement intersect with existing neuroscientific literature on emotion regulation strategies. While emotion regulation literature demonstrates habitual suppression as maladaptive, the present findings demonstrate there may be culturally informed contexts in which use of suppression is useful. Future research should explore how this culture surrounding emotional engagement in Black women may help to contextualize health disparities in mood and stress disorders in this population.

Keywords: emotional processing, Human Neuroimaging, Interdisciplinary

Disclosure: Nothing to disclose.

P421. Endogenous opioid receptor-mediated regulation of prefrontal cortex microcircuitry and valence processing

Hector Yarur, Chloe Noh, Sofia Shirley, Emilya Ventriglia, Miguel Arenivar, Huikun Wang, Brenda Shield, Karl Deisseroth, Grégory Scherrer, Andre Berndt, Michael R. Tadross, Hugo Tejeda

NIH/NIMH, Bethesda, Maryland, United States

Background: The medial prefrontal cortex (mPFC) plays a critical role in coordinating goal-directed behavior and executive function, processes often disrupted in neuropsychiatric conditions. These functions include learning context-dependent rules and integrating past experiences and outcome valence into schema representations that guide future behavior. Both opioid transmission and dysregulated mPFC activity have been linked to inhibitory-control deficits, which are characteristic of addiction and binge-type eating disorders. While endogenous opioids and their receptors are abundantly expressed in the mPFC, it remains unclear whether endogenous opioid transmission within this region directly modulates inhibitory control. In this study, we employed cutting-edge techniques to investigate the opioid system in mPFC circuitry.

Methods: Optogenetic stimulation and in vivo calcium imaging were performed in male and female mice to assess neuronal activity (Penk-Cre, n = 10) and Enk release dynamics (WT, n = 10) during associative learning tasks. Whole-cell patch-clamp recordings in acute brain slices from mice (n = 3–4), non-human primates (n = 3–6) and humans (n = 9–12) were conducted to measure the effects of Enk, MOR, and DOR activation on interneuron excitability. The pharmacological blockade of opioid receptors in the PFC was used to determine the impact of disrupted Enk signaling on reward-related behavior for WT (n = 10), SST-Cre (n = 7) and PV-Cre (n = 8) mice. Statistical analyses were performed using ANOVA or t-tests.

Results: Enkephalin (Enk), mu-opioid receptors (MOR), and delta-opioid receptors (DOR) were expressed in glutamatergic and GABAergic neurons in the mPFC (Unpaired t-test, p < 0.0001). Enk was enriched in VIP- and SST-interneurons (one-way ANOVA, p < 0.0001), MOR in SST interneurons, and DOR in both SST- and PV-interneurons (one-way ANOVA, p < 0.0001). Electrophysiology and optogenetics showed that Enk signaling inhibited SST outputs by MOR and DOR activation (Unpaired t-test, p < 0.0001), whereas PV outputs were selectively reduced by DOR activation (Unpaired t-test, p < 0.0001). Enk was also found to suppress inhibitory synaptic transmission onto layer V pyramidal neurons through a presynaptic mechanism in both the non-human primate (NHP) PFC and the human cortex (one-way ANOVA, p < 0.0001). Monitoring the activity of Enk-expressing neurons and the dynamics of Enk release during a head-fixed associative learning task revealed that these cells are likely engaged and release peptides in response to task-related rewards and punishments. Moreover, pharmacological blockade of opioidergic transmission within the PFC led to an increased lick frequency in response to rewards (paired t-test, p = 0.003). Blocking opioidergic signaling in SST interneurons increased reward-related licking, while blockade in PV interneurons had no effect (Unpaired t-test, p = 0.003). Additionally, MOR activation in the mPFC reduced food intake but increased locomotor activity in food-restricted mice (one-way ANOVA, p < 0.0001).

Conclusions: These findings indicate that Enk release selectively inhibits MOR- and DOR-expressing interneurons, engaging a disinhibitory mechanism that shapes mPFC microcircuit function and provides insight into how opioids regulate the exploitation–exploration dilemma, influencing reward consumption, aversive responses, and motivated behavior.

Keywords: enkephalin, endogenous opioids, Medial Prefrontal Cortex, Inhibitory interneurons

Disclosure: Nothing to disclose.

P422. Neuropeptidergic transmission shapes emergent properties of prefrontal cortical circuits underlying learning

Miguel Arenivar, Nathanyal Ross, Rodolfo Flores Garcia, Sultan Zakariya, Zachary Wang, Hector Yarur, Rufina Kore, Wilma Richiez-Mateo, Aaron Limoges, Huikun Wang, Hector Bravo-Rivera, Diana Agbor-Enoh, Juan Enriquez-Traba, Bruno Averbeck, Yulong Li, Gu Xinglong, Hugo Tejeda

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

Background: Prefrontal cortex (PFC) inhibitory interneurons are essential for top-down control of affect and are dysfunctional in psychiatric disorders. Interneurons expressing somatostatin (SST), a neuropeptide used as a cellular marker, have been implicated in cognition, affect, and disease. However, a critical knowledge gap remains in understanding somatostatin’s function as a neuropeptide transmitter specifically in regulating emergent properties of PFC circuit dynamics in-vivo.

Methods: We employed a viral/genetic strategy to selectively delete SST expression in the medial prefrontal cortex (mPFC) of male and female SST-loxP mice. The effects on cued threat discrimination were assessed (n = 15–17), while SST interneuron activity and neuropeptidergic dynamics during this task were monitored using single-cell Ca²⁺ imaging (n = 9) and fiber photometry with a genetically encoded SST sensor (GRABSST; n = 8), respectively. In addition, we evaluated the role of SST transmission in shaping mPFC pan-neuronal single-cell Ca²⁺ dynamics during cued threat discrimination (n = 4–6) in control and SST-loxP mice. Statistical comparisons were conducted using ANOVA, Fisher’s Exact test, or t-tests.

Results: Control mice showed robust freezing to the footshock-paired cue (CS+) but not to the neutral cue (CS−). In contrast, mPFC-SST knockouts froze to both CS+ and CS− cues (Two-way ANOVA, Genotype × Trial Interaction, p = 0.0006). GRABSST recordings revealed responses during threat acquisition to both footshock and the predictive CS+, but not to the neutral CS− (paired t-test, p = 0.0049). Pharmacological blockade of SST receptors within the mPFC impaired the acquisition, but not the expression, of cued threat discrimination against aforementioned cues (Two-way ANOVA, Manipulation × Stimulus Interaction, p = 0.0046). Furthermore, mPFC-SST neuropeptidergic signaling was required for neurons to integrate footshock outcomes with additional task variables to construct configural representations (Fisher’s Exact Test, p = 0.0005) and to support population-level encoding linked to discriminative learning (unpaired t-test, p = 0.0425).

Conclusions: Together, these findings demonstrate that endogenous mPFC-SST signaling shapes learning and drives emergent properties of prefrontal circuits, underscoring the importance of a neuropeptidergic system that has been traditionally regarded as an interneuron marker. This is particularly relevant given that reduced SST expression and immunoreactivity in the PFC have been implicated in multiple psychiatric disorders, including depression, bipolar disorder, and schizophrenia.

Keywords: prefrontal cortex, Somatostatin, interneurons

Disclosure: Nothing to disclose.

P423. The corticostriatal circuit mediates the behavioral response of frustrative nonreward (FNR)

Aijaz Naik, Xiaoyu Ma, Ellen Leibenluft, Zheng Li

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

Background: Frustrative nonreward (FNR), the emotional and behavioral response to omission, reduction, or delay of an expected reward, is a cross-species phenomenon present in animals and humans. Aberrant responses to FNR are central to the pathophysiology of irritability, a transdiagnostic and impairing symptom that negatively impacts mental health, particularly in youth. Little is known about the neural mechanism of FNR, hampering the clinical studies and treatment development for irritability. We previously developed the Alternate Poking Reward Omission (APRO) task, an instrumental paradigm that can introduce frustrative nonreward in juvenile mice within a relatively short time frame. In this study, we combined APRO with whole-brain cFos staining, in vivo electrophysiology, and chemogenetics to investigate the circuit mechanism of FNR.

Methods: APRO was applied to juvenile mice in a closed running track with a water dispensing port and a LED light installed at each end. The mouse runs on the track freely for a 15-minute session every day. The task has three phases: (1) Non-frustration phase (three sessions) during which mice receive rewards in all trials that they poke the two ports alternately (100% reward); (2) Non-frustration phase (one session) during which mice receive nonreward in 50% of correct alternations with a fixed pattern (50% reward); nonreward is expected, thus nonfrustrative; (3) Frustration phase (one session) during which 80% of correct alternations result in nonreward with a random pattern (20% reward); nonreward is unexpected, thus frustrative. Control mice receive 100% reward in all phases. To identify regions activated by FNR, brains were removed 90 minutes after the frustration phase for whole-brain clearing and cFos staining. For in vivo recording of neural activity, mice were implanted with multi-electrode arrays in the anterior cingulate cortex (ACC) and nucleus accumbens (NAC) for recording of local field potentials during APRO. For chemogenetics, ACC was injected with AAV expressing hM4Di; CNO was injected intraperitoneally 20 minutes before the frustration phase. All animal procedures followed the U.S. National Institutes of Health guidelines for using animals in intramural research and were approved by the National Institute of Mental Health Animal Care and Use Committee.

Results: FNR increased the number of cFos positive cells in ACC and NAC (two-tailed Student’s t-test; p values corrected by FDR; p = 0.013 for ACC and 0.043 for NAC; n = 6 control mice and 7 FNR mice). Neural oscillations are increased more by nonreward trials than by reward trials in ACC (13 mice; Two-tailed paired Student’s t-test; p < 0.05) and NAC (6 mice; Two-tailed paired Student’s t-test; p < 0.05). Chemogenetic inhibition of ACC abolished increases in locomotion (one-way ANOVA, F(2,26) = 4.8176, p = 0.0166; n = 8 control mice injected with saline, 12 FNR mice injected with saline, 9 FNR mice injected with CNO) and aggression in FNR mice (one-way ANOVA, F(2,15) = 3.63, p = 0.05; n = 8 control mice injected with saline, 12 FNR mice injected with saline, 9 FNR mice injected with CNO). Statistics: 1. cFos staining. Two-tailed Student’s t-test. P values were corrected by FDR. P = 0.013 for ACC and 0.043 for NAC; n = 6 control mice and 7 FNR mice; 2. Electrophysiology. Two-tailed paired Student’s t-test. P < 0.05 for ACC (n = 13 mice) and NAC (n = 6 mice). 3. Chemogenetics: One-way ANOVA. Locomotion: F(2,26) = 4.8176, p = 0.0166; n = 8 control mice injected with saline, 12 FNR mice injected with saline, 9 FNR mice injected with CNO. Aggression: F(2,15) = 3.63, p = 0.05; n = 8 control mice injected with saline, 12 FNR mice injected with saline, 9 FNR mice injected with CNO.

Conclusions: These findings indicate that FNR activates the ACC-NAC circuit to increase motor activity and aggression.

Keywords: irritability, Nucleus accumbens, Anterior cingulate cortex, Frustrative nonreward, APRO

Disclosure: Nothing to disclose.

P424. Circulating endocannabinoids and fear generalization in adolescents

Samantha Ely, Carmen Carpenter, Reem Tamimi, Clara Zundel, Hilary Marusak

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

Background: Anxiety-related disorders are often hallmarked by fear generalization, or difficulty discriminating between threatening and safe stimuli. Recent evidence has implicated the role of the endocannabinoid (eCB) system in fear generalization. In adults, higher concentrations of circulating 2-arachidonoylglycerol (2-AG) are related to less fear generalization. However, whether this relationship is present in youth remains unknown. This represents a critical gap given that anxiety disorders commonly emerge in adolescence—a sensitive developmental window marked by dynamic changes in fear-related learning and eCB signaling. This study examined associations between circulating 2-AG and fear generalization during an extinction learning task, using multiple indices of conditioned fear responding (behavioral, subjective, physiological) as well as brain activation in adolescents.

Methods: Ninety-seven adolescents (M = 13.35 ± 2.20 years; 54.6% female) provided plasma samples to assess circulating 2-AG. Then, adolescents performed a virtual reality fear extinction paradigm, with a subset (n = 67) undergoing concurrent functional neuroimaging. During fear conditioning, a stimulus (CS+) was paired with an aversive white noise burst (“threat cue”), while another (CS-) was unpaired (“safety cue”). After a rest period, both cues were presented without reinforcement in a novel context (extinction). Indices of conditioned fear included behavioral avoidance (distance from the CS), subjective fear ratings, skin conductance responses (SCRs), and brain activation to the CS+ and CS-. Linear regressions controlling for age, sex, and time of day tested associations between 2-AG and each index.

Results: Higher circulating 2-AG concentrations were related to greater avoidance of both the threat cue (R2 = 0.120, β = 0.237, t(90) = 2.270, p = 0.026) and safety cue (R2 = 0.139, β = 0.210, t(90) = 2.030, p = 0.045), as well as higher subjective fear ratings toward each (threat cue: R2 = 0.100, β = 0.259, t(90) = 2.451, p = 0.016; safety cue: R2 = 0.110, β = 0.222, t(90) = 2.113, p = 0.037). No significant associations emerged between 2-AG and SCRs. Whole-brain analyses (p < 0.005, k > 10) revealed that higher 2-AG was associated with lower activation in the cerebellum to both the threat (xyz = −12, −76, −20, k = 37) and safety cues (xyz = −12, −76, −20, k = 24), and lower activation in the right superior frontal gyrus to the safety cue (xyz = 15, 5, 64, k = 11).

Conclusions: In contrast to adult findings, higher circulating 2-AG in adolescents was associated with greater avoidance, higher subjective fear, and reduced brain activation to both threat and safety cues. These results suggest that eCB-related mechanisms of fear learning may differ across development—potentially protective in adulthood but conferring greater vulnerability to fear generalization in youth. Replication and further longitudinal work are needed to clarify developmental shifts in the role of the eCB system in anxiety risk.

Keywords: Endocannabinoids, Fear extinction, Child development, Anxiety and PTSD, Human Neuroimaging

Disclosure: Nothing to disclose.

P425. The relationship between endocannabinoids and brain function following acute alcohol intake: a placebo-controlled, within-subjects study

Ryann Tansey, Gavin Petrie, Raegan Mazurka, Elisabeth Paul, Markus Heilig, Leah Mayo

University of Calgary, Calgary, Canada

Background: Alcohol is a widely used psychoactive substance with the potential for harmful or maladaptive use. The subjective effects produced by alcohol can contribute to an individual’s personal behaviours regarding alcohol intake. Like many drugs, the pleasurable subjective effects of alcohol are linked to an individual’s motivation to drink. Individuals who consume more alcohol in their daily lives report liking its effects more than those who engage in lower levels of drinking, and also report higher desire to drink more after consuming alcohol. Alcohol also appears to have an effect on activation and glucose metabolism in the striatum, a key area for reward in the brain, but studies have been limited.

The endocannabinoid (eCB) signalling system is involved in a number of fundamental processes in the brain. There are two main ligands for the eCB system: anandamide (AEA) and 2-arachidonoylglycerol (2-AG), which are agonists for the cannabinoid 1 receptor (CB1R). eCBs modulate activity at mesolimbic reward synapses, such as those in the ventral tegmental area and the nucleus accumbens. CB1Rs are also expressed in the prefrontal and cingulate cortices, which are also involved in reward activity. Further, in humans, preliminary functional magnetic resonance imaging (fMRI) evidence suggests individuals with higher concentrations of AEA also show greater activation to reward in the putamen.

Alcohol intake behaviours may also be influenced by eCB signalling in the brain. Most work investigating this link has been conducted in preclinical models, which suggest that CB1R agonism is related to increases in alcohol self-administration and motivation to obtain alcohol, while CB1R antagonism results in decreased preference for alcohol and self-administration.

Given the important role that eCBs appear to play in both reward and alcohol intake, it is possible that the rewarding effects of alcohol (which in turn influences how much an individual drinks) are mediated at least partially through eCB signalling. Here, we investigated whether the eCB ligands AEA and 2-AG are associated with any changes to reward processing in the brain following alcohol intake. This study elucidates the fundamental relationships between reward, alcohol intake, and eCBs in the human brain, which can serve as an important foundation for clinical research targeting problematic alcohol use behaviours.

Methods: To investigate the intersection of eCBs, neural indices of reward function, and acute alcohol intake, we conducted a preliminary single-blind, within-subjects study in healthy individuals. Participants (n = 32; F = 16, M = 16) completed two experimental sessions lasting 4–5 hours each. In order to account for drug expectancy effects, participants were informed at the onset of each session that they could receive alcohol, a sedative, caffeine, or placebo, though they only ever received alcohol or placebo. The order of the sessions was randomized between subjects and counterbalanced. In order to account for differences in body water content between sexes, the concentration of the alcoholic drinks were set at 0.6 g/kg of alcohol for males and 0.4 g/kg for females. The placebo was cranberry juice with a 1% alcohol float to mask for smell and taste. At the very beginning of the session, participants were fitted with an IV catheter to facilitate serial blood sampling throughout the experiment. These samples were used to determine the concentrations of eCB ligands (AEA and 2-AG) in plasma throughout the course of the session.

At approximately 110 minutes following ingestion, participants completed an MRI scan. Task and resting-state functional MRI (fMRI) data were obtained. One of the tasks run in the scanner was the monetary incentive delay (MID) task, which assesses reward anticipation. Reward anticipation is measured by presenting participants with a cue that indicates whether the upcoming trial is one where they could gain money (potential reward), lose money (potential loss), or neither gain nor lose money (neutral), depending on their performance. A blood sample was also obtained at this time point, so that plasma eCB levels while in the scanner could be determined.

fMRI data was analyzed using AFNI. Linear mixed effects models were applied using the program 3dLMEr in order to account for repeated measures within-subjects. Models included covariates of interest accounting for session (alcohol vs. placebo), stimulus (reward, punishment, or neutral) and eCBs (AEA or 2-AG, in separate models). Head motion and sex were included as control covariates. Clusters were defined using a voxelwise threshold of p = 0.002 and a cluster forming threshold of α = 0.05, corresponding to a cluster size of 23 voxels.

Results: There was a significant interaction between stimulus (reward, punishment, neutral) and plasma AEA concentrations in two clusters: in the right temporo-parieto-occipital junction (cluster size: 51 voxels; peak χ2 = 22.63; peak MNI coordinates = [52, −50, 34]), and the right dorsolateral prefrontal cortex (cluster size: 25 voxels; peak χ2 = 20.15; peak MNI coordinates = [37, 46, 16]), such that individuals with higher AEA showed greater activation to the reward and neutral conditions but lower activation to the loss condition. There were no significant interaction between session and AEA or session, stimulus, and AEA. There were no significant effects of 2-AG.

Conclusions: AEA may have an influence on reward processing in higher-order cortical areas, regardless of acute alcohol intake.

Keywords: Endocannabinoids, Anandamide, Functional MRI (fMRI), Alcohol intake

Disclosure: Nothing to disclose.

P426. Preliminary efficacy of a novel non-addictive cannabidiol (CBD) with delta-9-tetrahydrocannabinol (THC) formulation for chronic pain and related symptoms

Rajita Sinha, Elcin Sakmar, Keisha Smith, Nia Fogelman, Alexa Torrens, Gretchen Hermes, Daniele Piomelli, Joseph Guarnaccia

Yale University, New Haven, Connecticut, United States

Background: There is significant interest in cannabis-based compounds as therapeutics for various pathologies, including chronic pain-related illnesses, post-traumatic stress disorder (PTSD) as well as epilepsy, multiple sclerosis, and cancer. However, uncertainty remains regarding a number of factors including effective formulations for specific symptoms, optimal dosing of specific cannabinoids and the risk of misuse. This project evaluating specific CBD and CBD + THC therapeutic dose ranges to target acute and chronic pain, as well as associated symptoms of anxiety and sleep disturbance. Here we assessed specific therapeutic doses for acute and chronic pain and related symptoms of anxiety and sleep problems. Study 1 assessed acute pain and anxiety responses to single oral tablets dose of two doses of cannabidiol (CBD) (40 mg, 100 mg), and three CBD/delta-9-tetrahydrocannabinol (THC) combination doses (CBD40mg/THC10mg, CBD40mg/THC20mg and CBD100mg/THC30 mg) and matching placebo (PBO) in a randomized, double-blind, cross-over design. Study 2 enrolled patients with chronic pain and evaluated 7-day chronic dosing of placebo versus the Study 1 doses that produced reductions in acute pain and anxiety and also determined sustained therapeutic benefit for pain and related symptoms, and assessed plasma levels of the endocannabinoids 2-AG and anandamide.

Methods: Study 1 enrolled 8 non-medical users of cannabis (5 male, 3 female; mean age 28.2 years) who were randomized to receive a single dose of 40 mg CBD alone (CBD40), 100 mg CBD alone (CBD100), or combination 40 mg CBD/10 mg THC (CBD40/THC10), 40 mg CBD/20mg THC (CBD40/THC20), and 100 CBD/ 30 mg THC (CBD100/THC30), or placebo on separate sessions one week apart over six weeks. Study 2 included 24 patients with chronic pain from musculo-skeletal disorder, multiple sclerosis or opioid use disorder who were randomly assigned to 7-day b.i.d. dosing of CBD40/THC10 or CBD100/THC30 or placebo in a parallel group design. Study 1 utilized the Yale Pain Stress Test (YPST) adapted from Cold Pressor Test (CPT) with three repeated pain-stress and no pain-stress trials per condition to assess acute pain and anxiety in each of the 6 sessions. Study 2 implemented the YPST on day 1 and day 7 to assess acute pain and anxiety and 2-AG and anandamide levels (by LCMS-MS) and assessed real world pain, anxiety and sleep daily during the 7-day dosing (study days) and over a 2-week follow-up period. Linear Mixed Effect (LME) models were utilized to assess meddose X Day X repeated time effects for YPST subjective pain and anxiety, and 2-AG and anandamide and daily pain and sleep ratings in study 2.

Results: Study 1 results showed significant decreases in average YPST provoked pain for the CBD40/THC10 and the CBD100/THC30 doses relative to placebo (p’s < 0.05), but not for CBD alone (40 mg or 100 mg) and not for the CBD40/THC20 doses. Relative to placebo, significant reductions in provoked anxiety during YPST were only significantly reduced in the CBD100/THC30 dose (p < 0.01) and not for any other dose. Feelings of high were only significant for the CBD40/THC20 (p < 0.008) and the CBD100/30THC (p < 0.05) doses relative to placebo. Study 2 comparing chronic 7-day b.i.d. dosing of placebo versus CBD40/THC10 and CBD100/THC30 doses showed Meddose X YPST day effect (F[2,342] = 6.07, p < 0.003) for acute pain with increases for placebo and CBD100/THC30 doses but not the CBD40/TH10 dose. Similarly, YPST acute anxiety showed Meddose X Day effect (F[6,155] = 3.09, p < 0.04) with significant increases in anxiety on both day 1 and 7 for placebo and CBD100/THC30 but not the CBD40/THC10 dose. Study 2 daily ratings of pain showed significant and sustain reductions in pain for the CBD40/THC10 dose from baseline to study period and during follow-up (p’s < 0.001) but increases in pain (p’s < 0.01) after initial drop in pain from baseline to study period (p’s < 0.001) for the placebo and the CBD100/THC30 doses. Notably, sleep ratings showed significant improvements from baseline to study period (p < 0.05) and follow-up period (p < 0.001) only for the CBD40/THC10 and not for placebo not the CBD100/THC30 doses. Meddose X Day interactions were observed for 2-AG (Χ²[2] = 7.12, p < 0.03) and for anandamide (Χ²[2] = 7.66, p < 0.02). Elevations in 2-AG levels by day 7 for CBD40/THC10 dose compared to placebo (p < .055) and CBD100/THC30 (p < 0.009) were observed, but lower anandamide levels were observed with placebo (p < .05) and nonsignificant increases in anandamide for the CBD40/THC10 dose. All doses were well-tolerated, and no serious adverse events were reported.

Conclusions: Across both studies, the combined CBD40/THC10 dose produced significant reductions in acute YPST pain and anxiety in non-medical users of cannabis and in patients with chronic pain. This dose also showed sustained improvements in pain, anxiety and sleep during the study and follow-up periods. Together, we present preliminary evidence for a specific, non-addictive moderate CBD and low THC combined formulation that warrants further evaluation as a novel therapeutic for chronic pain, anxiety, and sleep disturbances.

Keywords: delta9-tetrahydrocannabinol, cannabidiol, chronic pain treatment, Anxiety, Sleep disturbances

Disclosure: Aelis Farma, Contracted Research, Self, CT Research Pharma, Grant, Self, Imbrium Therapeutics, Consultant, Self, Tenacia Biotechnology, Other Financial or Material Support, Self, Menda Health, Advisory Board, Self

P427. Endocannabinoids and related lipid biomarkers for interventions: results from pilot pharmacological and behavioral studies in youth

Hilary Marusak, Jeanne Barcelona, Sophia Hopkins, Kevin Crombie, Cecilia Hillard, Aimee Luat

Wayne State University, Detroit, Michigan, United States

Background: The endocannabinoid (eCB) system plays a key role in neurodevelopment, psychiatric risk, and mental health (e.g., anxiety, mood). Growing evidence supports promise of circulating eCBs and related lipids as biomarkers for predicting—and potentially modulating—treatment response and mental health improvements, not just in adults but increasingly in youth. Here, we present findings from two pilot studies illustrating this potential.

Methods: Study 1 was a 4–6 week open-label trial of chronic dosing with FDA-approved cannabidiol (Epidiolex®, up to 10–25 mg/kg/day) in 12 youth with pediatric epilepsy (6 female; M age = 12.17 ± 5.17 years). Study 2 was a randomized controlled trial examining the acute effects of 30 minutes of moderate-intensity treadmill walking, light-intensity stretching, or seated meditation in a community sample of 68 youth (31 female; M age = 13.21 ± 2.34 years). Plasma samples were analyzed using liquid chromatography with tandem mass spectrometry to quantify circulating eCBs (2-arachidonoylglycerol [2-AG]; N-arachidonoylethanolamine [AEA]) and related lipids (e.g., palmitoylethanolamide [PEA]; linoleoylethanolamide [LEA]). Across both studies, we assessed pre-to-post changes in mood, anxiety, seizure frequency (Study 1 only), and lipid concentrations using one-sample t-tests. In Study 2, group differences were evaluated using one-way ANOVA (condition: treadmill, stretching, meditation).

Results: In Study 1, participants showed reductions in generalized anxiety and seizures (p’s < 0.001), and increases in 2-AG (p = 0.047). The latter two findings were only significant when excluding one participant who demonstrated an atypical elevation in seizure frequency. In Study 2, mood and anxiety improved across all conditions (p’s < 0.004), with elevations in PEA and LEA scaling with physical activity intensity (p = 0.032). Interestingly, youth with greater increases in AEA showed greater elevations in anxiety (p = 0.026), though this effect was attenuated when controlling for baseline anxiety symptoms (p = 0.124).

Conclusions: These findings support the notion that concentrations of circulating eCBs and related lipids are promising biomarkers for predicting and tracking response to both pharmacologic and behavioral interventions in youth.

Keywords: Endocannabinoids, interventions, pediatric, Anxiety, epilepsy

Disclosure: Nothing to disclose.

P428. Protective effects of cannabidiol on cognitive outcomes following acute use of cannabis for chronic pain

Renée Martin-Willett, Samantha Melendez, Kent Hutchison, Angela Bryan, L. Cinnamon Bidwell

University of Colorado, Boulder, Boulder, Colorado, United States

Background: An estimated two-thirds of medical cannabis users in the U.S. are motivated by chronic pain, but clinical guidance remains guarded as to whether the potential benefit of cannabinoids for this indication outweighs potential risks, such as effects on cognition. One recent review suggests the impact of tetrahydrocannabinol (THC) may be low to moderate. Other newer pre-clinical studies suggest cannabidiol (CBD) may be protective against cognitive impairment in the context of neuropathic pain. However, more research is sorely needed. Thus, the primary aim of this study was to evaluate the effects of differing cannabinoids on cognition following acute use for chronic, non-specific low back pain.

Methods: These analyses were among the primary aims of a large, pre-registered study of cannabinoid use and chronic pain (R01DA04413; PI: Bidwell). Participants were in one of three self-selected groups using edible cannabis: a) CBD-dominant, b) THC-dominant, or c) a largely equal ratio of CBD to THC. The study design included assessments at pre-use, 1 hour post-use use, and 2 hours post-use of participants’ legal market cannabis products. Outcomes of interest included scores on tests of executive function (Flanker Inhibitory Control and Attention Test), episodic memory (Picture Sequence Memory Test), working memory (List Sorting Working Memory test), verbal learning (International Shopping List Task (ISLT)), and processing speed (Pattern Comparison Processing Speed Test). Analyses utilized an intent to treat approach. Sample size was selected to permit analysis of the primary research questions at two-tailed alpha of 0.05 and power level of 0.95. Results suggested a total n = 162 would allow for detection of an interaction effect as small as f = 0.14. First, linear mixed effects models were constructed such that time (pre-use, 1 hour post-use, and 2-hour post use), group (CBD, THC, or CBD + THC) and their interactions predicted each outcome while controlling for age and education level, as cognitive performance is known to be sensitive to these individual characteristics. Analyses were conducted in R Studio using the lmerTest and emmeans packages.

Results: 184 participants who used cannabis for chronic pain were included in the study (71 in the CBD-dominant group, 24 in the THC-dominant group, and 89 in the group using both CBD and THC; 56.5% female; Mage = 45.4, 44.3% bachelor’s degree). Results demonstrated participants in the CBD and the CBD + THC groups did not have significant changes in scores between the pre-use and post-use timepoints on any cognitive measures (ps > 0.05), while participants in the THC group had significant, negative changes in scores following use for Flanker, incorrect ISLT responses, Pattern Comparison, and Working Memory (ps < 0.05). The CBD group performed better compared to the average of the whole sample on Pattern Comparison (b = 2.12, 95% CI[0.17, 4.07], p = .034) and Working Memory (b = 2.33, 95% CI[0.54, 4.13], p = .011). Participants in the CBD group also significantly outperformed participants using THC on Working Memory (d = 0.62, 95% CI[0.15, 1.10], p = .032) and had significantly more correct responses (d = 0.39, 95% CI[–0.025, 0.806], p = .0029) and fewer incorrect responses (d = –1.34, p = .0124) on the ISLT 1 hour following use. Finally, participants in the CBD group scored significantly higher on Working Memory compared to the group using CBD + THC 2 hours post-use (d = 1.21, 95% CI [0.27, 2.16], p = .028).

Conclusions: In general, participants in the group using CBD products experienced few negative effects of use on performance, as well as faster, and often superior functioning after use, both in pairwise comparisons, and in comparison to the sample overall. These findings largely replicate previous work supporting negative cognitive effects of THC use and provide novel evidence supporting a role for CBD in potentially mitigating those effects for participants in the CBD + THC group.

Keywords: Cannabidiol, delta9-tetrahydrocannabinol, Cognition, chronic pain

Disclosure: Nothing to disclose.

P429. “Green chemistry” approach for facile generation of CBD/THC combination products for medical use

Michael Palfreyman

Palfreyman BioPharm Advisors, Inc., St. Petersburg, Florida, United States

Background: The therapeutic value of the major phytocannabinoids, tetrahydrocannabinol (THC) and cannabidiol (CBD), has been demonstrated well beyond their FDA-approved and narrow range of indications. However, barriers such as drug scheduling, regulatory constraints, and limited availability of pharmaceutical-grade materials hinder rigorous clinical trials and the establishment of medical guidelines.

The very definition of “medicinal cannabis” remains elusive, typically resting on the judgment of individual physicians. They must choose from a wide range of products with inconsistent provenance, variable potency, and diverse delivery systems (oral, intraoral, inhaled, or smoked). Many of these products are excessively potent and poorly controlled.

Recent trends have sought to mitigate the risks of high-THC formulations, often associated with side effects and prolonged intoxication, while exploring the potential synergy between THC and CBD. Notably, Sativex, a 1:1 THC/CBD oromucosal spray, is approved in multiple jurisdictions (though not in the US) for multiple sclerosis-related spasticity and neuropathic pain. Other studies have evaluated CBD-dominant formulations, with THC ratios as low as 1:50, and incorporated titration protocols to optimize dosing for elderly or cannabis-naïve patients. Neuropsychiatric treatments require varying doses of CBD and THC, particularly those with lower THC to CBD ratios, addressing the limitations of many plant-based materials with high THC content, contaminants, and uncertain analysis.

Methods: CBD’s intrinsic propensity to cyclize into THC under acidic conditions has been recognized for decades. This conversion is typically performed in organic solvents using Brønsted or Lewis acids under heating, followed by extraction and neutralization. Reaction outcomes depend on the catalyst and conditions, with THC-delta 9 favored as the major product and THC-delta 8 and other isomers minimized.

However, this traditional approach has critical limitations: the use of organic solvents, catalyst contamination, and the challenge of meeting GMP-compliant standards for food- or pharmaceutical-grade products.

To overcome these barriers, we developed a “green chemistry” method utilizing solid-state catalysis with food-grade aluminosilicates and controlled heating. These catalysts are inexpensive, commercially available, and free of leachable contaminants.

Process: Highly pure crystalline CBD (> 99%) is mixed with the catalyst and heated at 100–110 °C for 3–10 minutes.

Formulation options:

• The resulting CBD/THC mixture can be blended with edible oils (e.g., olive oil) for sprays, lozenges, capsules, or topical creams.

• Alternatively, CBD can be combined with edible oil and the catalyst before heating (5–15 minutes), followed by filtration to remove the catalyst.

• A device-based approach allows CBD pre-loaded with the catalyst in cartridges to be heated under controlled conditions, yielding low, medium, or high THC conversions.

Analysis: Reaction products are analyzed by HPLC and GC/MS, enabling the generation of certificates of analysis.

Results: Controlled CBD-to-THC ratios were reproducibly achieved using CBD: catalyst ratios of 20:1 to 5:1 at 100–110 °C. By varying the catalyst load and heating time, THC yields up to 25% were obtained.

Example 1: 25 mg CBD + 5 mg catalyst, heated for 3 min, yielded ~10% THC.

Example 2: 50 mg CBD + 10 mg catalyst in 1 mL olive oil, heated 10 min at 100 °C, yielded a mixture of 88% CBD, 8% THC-delta 9, 3% iso-THC, and trace amounts of THC-delta 8 and iso-THC isomers. Analysis with HPLC/MS and GC/MS consistently yielded results with >95% purity of the active ingredients CBD and THC-delta 9, and no other unidentified side-products. Prolonged heating or higher temperatures reduced purity and increased side-product formation.

Products of the reaction were filtered to remove the catalyst and then can be prepared as oral sprays, lozenges, capsules, and topical ointments. Stability studies on oil-based solutions have shown suitable characteristics for therapeutic development, pre-clinical pharmacology, GLP-safety, and toxicology evaluation, and eventual clinical use. Thus, the process is optimal for producing low-THC medicinal products, less prone to diversion for recreational use.

Conclusions: Although fewer than 4 million individuals in the US currently receive prescribed medical cannabis, many more use recreational products for therapeutic purposes. Demand for high-THC products used recreationally has overshadowed the need for low-dose, high-quality CBD/THC medicines, particularly for vulnerable populations. Our catalytic process enables consistent production of CBD/THC formulations with limited byproducts, mainly iso-THC. The defined CBD/THC mixtures can be delivered in the final product as orobuccal sprays, lozenges, oral capsules, or used topically as creams or ointments. A device for delivering controlled ratios of low, medium, and high ratios of THC to CBD is in development. Further pharmaceutical development and refinement of this method could expand access to safe, standardized low-THC therapeutics for neuropsychiatric conditions such as PTSD, anxiety, autism spectrum disorder, sleep disorders, chronic pain, other neurological indications, and nausea and cachexia associated with cancer.

Acknowledgement: The chemistry, analysis, and patents on these methods were developed by Alex Nivorozhkin, PhD, and licensed to C-Click Life Sciences, Inc.

Keywords: cannabidiol (CBD), tetrahydrocannabinol (THC), neuropsychiatric indications, catalysis, click-chemistry

Disclosure: Palfreyman BioPharm Advisors, LLC, Founder, Self, C-Click Life Sciences, Inc., Board Member, Self

P430. Psychiatric safety profile of CFTR modulators: a multi-source real-life assessment

Diane Merino, Samuel Cohen, Julie Bulsei, Nouha Ben Othman, Marie-Noëlle Osmont, Tessa Pietri, Sophie Gautier, Fanny Rocher, Milou-Daniel Drici, Alexandre Gérard, Laurent Chouchana

Children’s Hospitals of Nice, CHU-Lenval, Nice, Franceversity Hospital CHU de Nice, Nice, France c Université Côte d’Azur, CoBTeK, Nice, France

Background: Cystic fibrosis (CF) is a severe, multisystem genetic disorder caused by pathogenic variants in the Cystic Fibrosis Transmembrane conductance Regulator (CFTR) gene. The resulting defective chloride and bicarbonate transport across epithelial cells leads to thickened secretions, impaired mucociliary clearance, and a cascade of respiratory, gastrointestinal, and metabolic complications. Although symptomatic management has historically prolonged survival, median life expectancy remained limited until the emergence of CFTR modulators (CFTRms). While clinical trials mainly addressed hepatic and respiratory adverse drug reactions (ADRs), clinician observations and post-marketing monitoring have increasingly suggested the occurrence of psychiatric ADRs following CFTRm initiation, especially suicidal ideation. To help identify high-risk subgroups and to guide clinical practice, we conducted an analysis of CFTRms-related psychiatric ADRs, leveraging three complementary real-world datasets.

Methods: (1) We queried VigiBase®, the WHO pharmacovigilance database, for psychiatric ADRs with ivacaftor, lumacaftor/ivacaftor, tezacaftor/ivacaftor, and elexacaftor/tezacaftor/ivacaftor (ETI) from 2012–2025. Disproportionality analyses (reporting odds ratio [ROR], information component [IC]) assessed global safety signals, so that a drug-ADR association was considered as a significant signal only if both IC025 > 0 and the ROR 95% CI lower bound >1. We conducted subgroup analyses according to age and a sensitivity analysis relying on cases reported by healthcare professionals. We also described the distribution of the most reported psychiatric ADRs according to the age. VigiBase® ensures preservation of the anonymity of patients and notifiers.

(2) We reviewed the French Pharmacovigilance Database (FPD, 2012–2024) for narratives providing clinical context, psychiatric history, co-occurring ADRs, and therapeutic management.

(3) We analyzed the French National Health Data System (FNHDS, 2020–2023) to identify incident psychiatric disorders after CFTRm initiation in a population-level cohort, excluding patients with a history of psychiatric condition(s) in the year prior to CFTRm initiation, or in the month following initiation.

Statistical analyses were performed with R (version 4.5.1).

Results: VigiBase®: Among 3,407 psychiatric ADR reports, median age was 19 years; 50% involved males and 799 involved youths. Median time to onset was 1–2 months, and 28% of cases were serious, including 29 deaths. Dechallenge resolved 67.6% of psychiatric ADRs, while rechallenge led to 25.5% recurrence. There was a signal for an association between CFTRms and psychiatric ADRs, with a disproportionality signal (ROR 1.73 [1.67–1.80]), even strengthened when restricting to healthcare professional reports (ROR 2.38 [2.27–2.49]). This signal was higher in youths (ROR 2.01 [1.87–2.17]) than in adults (ROR 1.48 [1.38–1.58]) with a relative ROR of 1.36 [1.23–1.51]. Sleep disorders predominated in young children (42.5%) and older adults (40.7%). Mood disorders affected about one-third of youths (8–17 years) and declined with age. Anxiety rose in early adolescence (25.9%) and stabilized in adulthood (~27%). Suicidal behaviors peaked in adolescents, especially those aged 15–17 (14%).

FPD: We analyzed 116 high-quality cases, among which 58% were pediatric, with a median age of 15 years. Narratives highlighted sleep disturbances (61%), mood disorders (37%), anxiety (27%), and suicidality (12%). Psychiatric history was present in 23/25 evaluable cases, often mood or neurodevelopmental disorders. Co-occurring somatic ADRs (abdominal pain, rash, hepatic cytolysis) were frequent (40.5%). Supratherapeutic plasma levels were documented in half of monitored cases. Changing drug timing improved symptoms in 62% of attempts.

FNHDS: In a cohort of 4,248 CF patients (median age 19 years, 54% male) initiating CFTRms between 2020–2023, 12% developed new-onset psychiatric disorders. Their mean time to onset was 1.4 years, but they could occur as late as 3.9 years.

Conclusions: This study represents the most comprehensive, multi-source, real-life assessment to date of psychiatric ADRs associated with CFTRms, integrating global pharmacovigilance signals, national narrative-rich case reports, and pharmacoepidemiological data. Our findings bring to light varying age-dependent profiles. Younger patients seem to be more prone to develop psychiatric disorders with CFTRms. While sleep disorders were leading in terms of frequency, mood disorders and suicidal behaviors were peaking in late adolescence, carrying major clinical concern. Qualitative evidence from FPD supported recurrence on rechallenge, improvement after dose modification, and potential relation to supratherapeutic levels. Population-based FNHDS data confirmed a substantial incidence of new-onset psychiatric disorders in real-life practice. Together, these findings emphasize the need for baseline psychiatric assessment, structured monitoring throughout treatment, and collaboration between CF and mental health specialists. Age-specific vulnerability, especially in adolescents, highlights the importance of patient and caregiver education. Future prospective research should provide more definitive answers, especially in youths, to optimize treatment benefits while safeguarding psychological well-being across the lifespan.

Keywords: Pharmacovigilance, Pharmacoepidemiology, CFTR modulators, Adverse drug reaction, Safety profile

Disclosure: Nothing to disclose.

P431. State of Ohio Adversity and Resilience (SOAR) Wellness Discovery Survey: Bio-Psycho-Socio-Environmental Predictors of Persistent Distress in a N = 15,000 Abs Representative Population Sample of Ohio

Anthony King, Steven Gavazzi, Brett Klamer, Hyoshin King, Soledad Fernandez, Canada Keck, Randall Olsen, Stephanie Gorka, Marybel Gonzalez, Grace Maynard-Wentzel, Nina Kraguljac, Tim Crawford, Cory Cronin, Jean Forney, Paul Hershberger, Larrilyn Grant, Susan Brown, Kai Nomaguchi, K. Luan Phan, Scott Langenecker

The Ohio State University, Columbus, Ohio, United States

Background: Persistent psychological distress is major public health problem, impacting personal wellbeing and psychological health as well as the broader community, contributing to disability and family and occupational problems. Psychological distress is associated with childhood adversity and lifetime trauma and adversity and independently predicts deaths due to suicide and drug overdose. The State of Ohio Adversity and Resilience (SOAR) study is a prospective, longitudinal, multimodal, integrated familial study supported by the State of Ohio Mental Health and Addiction Service, designed to identify biological, psychological, and social risk and resilience factors and processes leading to disorders of the brain, including overdose, suicide and psychological distress.

Methods: The SOAR Wellness Discovery Survey is an address-based random population epidemiological sample of 15,312 respondents (unique households) representative of the state of Ohio collected between 2024 and 2025. Construction of the ABS sampling frame (random sampled from Marketing Systems Group database of all household addresses receiving mail in Ohio) and all fieldwork, data collection, data management, cleaning, and census-based weighting were performed by CHRR (Ohio State University). The sampling frame subdivided Ohio proportionately into 15 sampling regions by urban/rural, geographic region, race and ethnicity. We sent > 320,000 postcards to random sampled addresses, with QR codes to access the survey for online completion (the survey could also be completed via a phone). A call center also contacted households directly based on location and demographics. The survey took ~50 minutes to complete. The survey assessed psychiatric symptoms (DSM-5 cross-cutting measures), history of suicidal ideation and substance use, pain, sleep quality, personal and family psychiatric and health history, demographics, social functioning, childhood adversity and protective experiences, address at age 10 (for geo-coding of childhood environment), lifetime experience of trauma and discrimination, economic outlook, resilience and coping skills, and other psychological factors related to risk and resilience. Psychological distress was measured using the 6-item Kessler 6 scale, previously used in national surveys to assess nonspecific psychological distress and serious mental illness in the general population. Identification of predictors of psychological distress utilized Lasso regression (R packages glmnet, ordinalNet) and Random Forest (R ranger, ordinalForest: trtf).

Results: Analyses of the SOAR Wellness Discovery survey are ongoing. Preliminary analyses of the statewide survey found 66% endorsed low, 18% moderate, and 13% high levels of current psychological distress. Predictors of psychological distress risk included being a gender and sexual minority, social (loneliness, discrimination, childhood adversity), psychological (cognitive control inflexibility), biological (pain, poor sleep), and environmental (economic pessimism) factors. Protective factors included social (social network size, religious attendance, protective childhood experiences, psychological (cognitive flexibility, ability to decenter from negative emotions, meaning in life, comfort in religion, hope), and environmental (household income, educational attainment).

Conclusions: While analyses are ongoing, preliminary results indicate both risk and protective factors associated with psychological distress across biological, psychological, social, and environmental domains. Further planned and ongoing work will test multi-level bio-psycho-social-environmental models combining self-report and geo-coded contextual/“neighborhood” level factors (e.g. indices such as Area Deprivation Index, Ohio Opportunity Index and specific census/exposure/crime data) in directional asynchronous graphs (DAGs) and other models, modeling family history/biological loading, early life adversity (maltreatment, threat, deprivation) and protective factors, contextual factors at age 10 (address based), life-time trauma and discrimination, and current factors.

Keywords: psychological distress, Social epidemiology, Risk and Resilience

Disclosure: Nothing to disclose.

P432. Teen and young adult mental health trends over time in the Population Assessment of Tobacco Health (PATH) Study, 2013/2014–2024

Wilson Compton, Michael Parks, Carlos Blanco, Kevin Conway, MeLisa Creamer, John Kingsbury, Daniela Marshall, Colm Everard, Kaylin Greene, Heather Kimmel

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

Background: Increasing mental health problems among youth ages 12–18 have been widely documented over the past decade, but data from nationally representative samples have been limited and few studies include both youth and young adults (i.e. ages 19–25). Moreover, most research on mental health trends among youth have been based on samples drawn from schools, and school-based samples have limitations. Both documenting overall trends and uncovering potential mechanisms that explain the increases are important for developing effective strategies and policies to enhance youth mental health. This study uses nationally representative data from the Population Assessment of Tobacco and Health (PATH) Study to document recent trends in the mental health of young people from 2013/2014 (Wave 1) to 2024 (Wave 8), using both mean scores on mental health measures as well as clinically meaningful score thresholds.

Methods: The PATH Study is an ongoing, nationally representative, longitudinal cohort study in the U.S. Analyses included data collected from youth (ages 14–18) and young adults (ages 19–25) from Waves 1 (2013/2014) through Wave 8 (2024). Mental health problems were assessed via the Global Appraisal of Individual Needs-Short Screener (GAIN-SS), and specifically, the Internalizing and Externalizing Disorder Screener subscales. Taking advantage of the clustering of internalizing and externalizing disorders, the GAIN-SS subscales are designed to quickly assess the presence of clinically meaningful internalizing disorders (e.g. somatic, depressive, anxiety and trauma-related disorders) and externalizing disorders (e.g. conduct disorders, attention deficit/hyperactivity disorders). GAIN-SS subscales are included the NIH PhenX Toolkit Behavioral Health Screening measures and have documented good reliability and validity. Mean GAIN-SS scores (range: 0 to 4 internalizing symptoms, 0 to 7 externalizing symptoms) and dichotmous measures, which are based on clinically meaningful “high symptoms” dichotomous measures (3+ symptoms for both measures), were assessed. Both measures capture past-year symptoms. Full-sample and replicate weights were used to account for the complex sample design and nonresponse. Design-based significance tests were used to test differences in mental health status within waves and change scores (CSs) across waves. Differences across sex and age groups were assessed. Regression analyses and joinpoint analyses were also used to assess trends overall and across sex and age groups.

Results: Among youth and young adults ages 14–25, GAIN Internalizing subscale mean scores increased during the study period (2013/2014 to 2024) from 1.72 to 1.86, and prevalence of high GAIN Internalizing subscale symptoms increased from 35.3% to 41.4% (respectively, CS = 0.14, p < 0.001; CS = 6.2%, p < 0.001). Similarly, GAIN Externalizing subscale mean scores increased from 2.16 to 2.26, and prevalence of high symptoms increased from 40.6% to 44.8% (respectively, CS = 0.10, p < 0.001; CS = 4.2%, p < 0.001). There was a recent decline (2022/W7 to 2024/W8) in GAIN Internalizing subscale but not GAIN Externalizing subscale scores. Mean GAIN Internalizing subscale scores declined from 2022/W7 to 2023/W7.5 (CS = −0.10, p < 0.001) and from 2023/W7.5 to 2024/W8 (CS = −0.08, p < 0.001), while percent with high GAIN Internalizing symptoms declined by 2.9 percentage points (p < 0.001) from 2022/W7 to 2023/W7.5 and 1.4 percentage points (p = 0.02) from 2023/W7.5 to 2024/W8. For both youth and young adults, joinpoint analyses generally corroborated wave-to-wave results, with a few exceptions in relation to statistical significance. Across age groups and sex, GAIN Internalizing and Externalizing subscale scores were higher among youth compared to young adults across the study period, but increases in both subscales were steeper for young adults. While wave-to-wave analyses demonstrated that this difference in slopes was present for mean scores and percent with high symptoms for both GAIN Internalizing and Externalizing subscales, joinpoint analyses demonstrated that slopes for increases in percent with high symptoms were parallel for both youth and young adults. Comparing sexes, females had higher mean levels and prevalence of high symptoms for both subscales. Increases over time were also steeper for females. These results for differences in slopes for both GAIN Internalizing and Externalizing subscales (for mean scores and percent with high symptoms) across males and females were found in wave-to-wave and joinpoint analyses.

Conclusions: Internalizing and externalizing symptoms increased among U.S. young persons ages 14–25 from 2013/2014 to 2024, especially among females and young adults ages 19–25. By 2024, high internalizing symptoms (i.e. clinically meaningful levels) were seen in 41.4% and high externalizing symptoms in 44.8% of U.S. young persons. These findings emphasize the important need for targeted treatment and intervention services to improve the mental health of young people. There was also a recent decline in internalizing problems, suggesting that the increasing trend may be slowing or reversing. Future research should identify mediating factors that explain population-level changes in mental health, both within individuals and between individuals over time.

Keywords: Mental Health, Teen and Young Adult, Population Assessment

Disclosure: Pfizer Inc., Stock / Equity - Publicly Traded Company, Self, 3M Companies, Stock / Equity - Publicly Traded Company, Self

P433. Safety of high and low frequency cannabis use: a national prospective study

Carlos Blanco, Pablo Fernandez-Navarro, Celia Talavan-Gonzalez, Mark Olfson

National Institute on Drug Abuse, Washington, District of Columbia, United States

Background: Although cannabis use is increasing and widely perceived as harmless by the general population, especially if used infrequently, little is known about the impact of low frequency cannabis use on psychosocial functioning.

Methods: We drew on data from the participating in Waves I (2001–2002) and II (2004–2005) of the National Epidemiologic Survey for Alcohol and Related Conditions, a nationally representative sample of non-institutionalized adults assessed prospectively. Frequency of cannabis use was assessed by self-report categorized as no use, low frequency use (3 or fewer times per month) and high frequency use (4 or more times per month) and psychosocial function was assessed with the Short Form 12, version 2 which yields 4 subscales: a physical component summary (PCS), mental component summary (MCS), social functioning, and role emotional all of which have a range of 0–100 with lower scores indicating worse psychosocial functioning.

Results: The cohort included 20,030 participants of whom 53.7% (n = 17,320) were female and 46.3% (n = 11,710) were male. At baseline, all frequencies of cannabis use were associated with lower (worse) scores on all SF-12 subscales after adjusting for age, sex, race and ethnicity, and educational attainment. At Wave 2, frequency of cannabis use was not associated to PCS scores but was associated with lower scores on the MCS, social functioning scale and role emotional scores, which were all significant. Furthermore, for all 3 scales, higher cannabis use frequency was associated with greater decrements in subscale scores.

Conclusions: At the population level, even low frequency cannabis use is associated cross-sectionally and longitudinally with lower psychosocial function.

Keywords: cannabis, Alcohol and substance use disorders, addiction

Disclosure: Nothing to disclose.

P434. Internal state modulates ventral tegmental area astrocyte activity during reward-seeking

Ashley Holloway, Amritha George, Jenny Klein, Hao Li, Reesha Patel

Northwestern University, Chicago, Illinois, United States

Background: Internal states, like hunger, are critical for guiding reward-seeking. Elucidating the mechanisms by which internal states shape reward-seeking is essential for advancing our understanding of how the brain integrates changes in systemic physiology with changes in neural activity to drive adaptive behaviors. Astrocytes are uniquely situated within the brain to sense peripheral changes in circulating hormones and nutrients accompanied by changes in internal state via their interaction with the blood-brain barrier. Astrocytes are also embedded in local neural circuits where they can sense variations in neurotransmitters and neuromodulators. Thus, astrocytes are perfectly poised to act as effectors of internal state changes, integrating signals from within the brain and outside of it to alter neural plasticity and behavior. The ventral tegmental area (VTA) is a key node in the neural circuitry controlling reward-related behaviors. Recent studies have reported that optogenetic activation of astrocytes within the VTA can induce both negative and positive valence states, as indicated by place avoidance and preference. While these results are exciting for understanding what optogenetic VTA astrocyte activation can do to an animal’s internal state to shape behavior, it’s unclear if endogenous VTA astrocyte activity is altered by internal state to affect behavior. Clarifying the endogenous activity dynamics of VTA astrocytes in response to reward under different internal states will add to our understanding of how these cells contribute to reward-seeking behavior, and advance our understanding of the mechanisms by which internal states alter behavior. To this end, we used fiber photometry in conjunction with a Pavlovian reward conditioning paradigm to elucidate the endogenous activity dynamics of VTA astrocytes in response to reward under two different internal states: hunger and satiety. We compared VTA astrocyte responses to reward under each internal state, and sought to understand how individual differences in astrocyte reward-related activity related to reward-seeking in each state.

Methods: We used dual-color fiber photometry in combination with viruses encoding an astrocyte-specific green fluorescent calcium indicator (AAV5-gfaABC1D-cyto-GCaMP6f) and a neuron-specific red fluorescent calcium indicator (AAV9-hSyn-jRGECO1a) to track calcium activity of ventral tegmental area (VTA) astrocytes and neurons simultaneously in mice of both sexes (n = 11; 6 males and 5 females). We recorded cellular activity as mice underwent Pavlovian reward conditioning while food restricted and while sated. For Pavlovian reward conditioning, mice were food restricted and subjected to 25 tone-sucrose pairings each day for six days. Mice were then ad libitum fed for 3–6 days to establish a sated internal state, and then subjected to one more day of Pavlovian reward conditioning.

Results: We found that astrocytic and neuronal activity reflected port entry outcomes (i.e. rewarded and unrewarded) during training under both hungry and sated states (Two-way ANOVA with Port Entry Type and Internal State as factors: Neurons, Main effect of Port Entry Type, p < 0.0001; Astrocytes, Main effect of Port Entry Type, p = 0.002). Intriguingly, astrocytes exhibited state-specific activity in response to rewarded port entries (Two-way ANOVA, Main effect of Internal State, p = 0.03), such that the magnitude of astrocyte activity in response to reward when mice were sated was significantly lower than that in response to reward when mice were food restricted (Fisher’s LSD Multiple Comparison’s, p = 0.04). There was no effect of internal state on astrocytic or neuronal encoding of unrewarded port entries. These results indicate that VTA astrocyte activity increases in response to reward in a manner that is modulated by internal state.

We then investigated how astrocyte activity during rewarded port entries related to reward-seeking behaviors. When mice were food restricted, we found that the magnitude of astrocyte activity in response to reward was negatively correlated with the latency of mice to enter the reward port after the conditioning tone started (Linear regression between latency to enter port after cue and peak astrocyte activity during rewarded port entries, r2 = 0.52, p = 0.01). We also found a positive correlation between the magnitude of astrocyte activity in response to reward and the rate of cue port entries made by mice under food restriction (Linear regression between cue port entries per minute and peak astrocyte activity during rewarded port entries, r2 = 0.36, p = 0.05). Interestingly, these relationships were abolished when mice were sated. Moreover, we did not find any significant relationship between neuronal activity in response to reward and reward-seeking behaviors. These results suggest that VTA astrocyte activity reflects motivational vigor for rewards when mice are hungry, but not sated.

Conclusions: Overall, our studies indicate that while both VTA astrocytes and neurons differentially encode port entry outcomes during reward conditioning, astrocyte responses to reward are selectively modulated by internal state. Additionally, our findings suggest that VTA astrocyte activity reflects motivation to acquire rewards during an appetitive task when mice are food restricted, but not sated. Altogether, our results suggest that VTA astrocytes represent a key cell type for translating internal state changes into alterations in reward-seeking.

Keywords: Reward seeking, Astrocyte, Ventral Tegmental Area (VTA)

Disclosure: Nothing to disclose.

P435. Adolescent social isolation stress induces microglial phagocytosis, astrocyte atrophy, and social behavior dysfunction in adult rats

Jonathan VanRyzin, Kathryn Reissner

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

Background: Early life adversity during adolescence increases the risk of psychiatric disorders later in life. Although the underlying mechanisms remain unclear, stress can dysregulate microglia and astrocytes—glial cells essential for neural circuit development and function—which may contribute to long-term behavioral consequences. Here, we used adolescent social isolation as a rodent model of early life adversity and hypothesized that social isolation stress affects later-life social behavior by impairing glial function in the nucleus accumbens (NAc), a brain region responsible for integrating social, reward, and stress information.

Methods: Experiment 1: To determine the effects of adolescent social isolation on later-life behavior, male and female Long Evans rats were isolated (vs group-housed controls, n = 6 per group) from weaning until adulthood (P21-P56). From P57-P58, rats underwent social behavior testing with a novel age- and sex-matched conspecific in a social interaction test and 3-chamber preference test. Brains were then collected for immunohistological analysis of astrocytes and microglia using GFAP and Iba1, respectively. Experiment 2: To determine the involvement of stress hormones during adolescent social isolation on glial dysfunction and behavior, male and female Long Evans rats were isolated from weaning until adulthood (P21-P56). Rats were treated with the glucocorticoid receptor antagonist RU486 (or vehicle control, n = 8 per group) once daily during the isolation period (20 mg/kg; s.c.). From P57–58, rats underwent social behavior testing, and brains were collected for analysis as described above.

Results: Socially-isolated rats spent more time interacting with a novel conspecific compared to group-housed controls (t(10) = 4.484; p = 0.007) in the social interaction test, while there were no differences in social preference index between groups (t(10) = 0.3563; p = 0.729). We found GFAP+ astrocytes within the NAc of socially-isolated rats had decreased structural complexity (F(1, 30) = 30.0; p < 0.001) as assessed by Sholl analysis of GFAP branching patterns, and a corresponding increase in the percentage of phagocytic Iba1 + /CD68+ microglia (t(249.4) = 6.522; p < 0.001). Moreover, phagocytic microglia had significantly more inclusions that co-labeled with the astrocyte marker GLT-1 (t(215.6) = 4.680; p < 0.001). RU486 treatment both prevented the isolation-induced increase in social interaction time (t(14) = 4.016; p = 0.0013), and rescued GFAP+ astrocyte structural complexity (F(1, 14) = 28.15; p = 0.001. Ongoing studies will determine the effects of RU486 on evidence for microglial phagocytosis.

Conclusions: Our results demonstrate that adolescent social isolation stress induces aberrant phagocytosis of astrocytes and reduced astrocyte complexity within the NAc, as well as increased social interaction, by a mechanism reliant on glucocorticoid receptor function. We hypothesize that microglia pruning of astrocytes provides a putative mechanism by which isolation stress drives astrocyte atrophy and behavioral dysregulation in adulthood. Our findings highlight the sensitivity of the neuroimmune system to developmental stress and propose a novel model of microglia/astrocyte pruning by which neuropsychiatric disorders may arise.

Keywords: microglia, Astrocyte, Social Behavior, Adolescent social isolation, early-life adversity

Disclosure: Nothing to disclose.

P436. Microglia depletion reduces motivation for fentanyl

Samara Vilca, Benjamin Williams, Ryan Logan

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

Background: The pronounced prevalence of opioid use disorder (OUD) continues to be a major public health concern, with high rates of relapse. The synthetic opioid fentanyl, which is far more potent than other opioids, contributes to the vast majority of the number of deaths due to drug overdose. Despite significant research, the underlying mechanisms of fentanyl use and relapse are not fully understood. Several studies have shown that microglia, the resident immune cell of the brain, impacts opioid reward, tolerance, and withdrawal. We previously found that chronic opioid use disrupts neuroinflammation which may drive opioid use, suggesting that microglia may be involved in OUD. However, the precise role of microglia in regulating opioid-related behaviors has not been examined. Here, we use fentanyl intravenous self-administration (IVSA) in mice to examine whether ablating microglia impacts fentanyl seeking, craving, and relapse behaviors.

Methods: Adult male (n = 12) and female (n = 12) mice were fed either CSF1R inhibitor PLX5622 (1200 ppm in chow) or control diet (AIN-76A) 10 days prior to self-administration. Then, mice were implanted with an indwelling jugular catheter. Following recovery, mice were trained to self-administer fentanyl for 15 days in daily 2-hr sessions (5 d/week) under a fixed ratio 1 (FR1) schedule. Successful completion of criteria (active nose pokes) resulted in delivery of fentanyl (3 μ g/kg/inf) and activation of a cue-light and 2.5 kHz tone, as well as a 20-sec timeout period in which active nose pokes were recorded but carried no scheduled consequences. Following this, a progressive ratio (PR) session was conducted to test motivation for fentanyl, wherein the number of active nose pokes required to obtain a single infusion increases according to the formula [5e(0.20inj number)]-5. Immediately following last IVSA session, brains were perfused or flash frozen for downstream immunohistochemistry (IHC) and RT-qPCR. All experimental procedures were approved by the University of Massachusetts Chan Medical School’s IACUC and were conducted in accordance with the NIH’s Health Guide for the Care and Use of Laboratory Animals.

Results: Here, we show PLX5622-treated mice display reduced fentanyl self-administration, deescalating their intake (Session x Treatment Interaction: F (17, 271) = 2.741, P = 0.0003) and nose poking (Session x Treatment Interaction: F (17,271) = 3.333, P < 0.0001) compared to controls. Mixed-effects models revealed that mice exhibit robust discrimination between the fentanyl active and inactive nose poke ports (F (2.67, 20.99) = 3.46, P = 0.04). When components of fentanyl self-administration were operationalized, many behaviors were decreased in PLX5622-treated mice. Specifically, seeking (P = 0.0788), intake (P = 0.0701), and motivation (P = 0.0622) were modestly reduced, while escalation was significantly blunted (P = 0.0005) in PLX5622-treated mice. Additionally, Spearman’s rank-order correlation revealed that while PLX5562-treated mice showed reduced fentanyl-related behaviors, their behaviors more strongly correlated with each other compared to control mice. Z-scoring the behaviors for each mice within treatment group and calculating the sum of those Z-scores produced an FUD “Severity” score, which revealed that PLX5622-treatment significantly mitigated fentanyl use severity (P  =  0.0071). Furthermore, principal component analysis (PCA) of all fentanyl-related behaviors revealed that PLX5622-treated mice cluster together and oppose many of the behavioral trajectories associated with the control mice cluster. From IHC analysis, we found PLX622 treatment robustly depleted microglia across the brain (P = 0.0002). Finally, we did not find any significant sex differences, neither behaviorally nor molecularly.

Conclusions: Our findings indicate that microglia depletion with PLX5622 significantly reduces fentanyl-related behaviors, especially escalation of drug-taking. Further analysis of depleted tissue may reveal microglia-dependent molecular changes. Moreover, previous studies have indicated opioid-induced gene expression changes in microglia are robust and long-lasting, which can now be examined using our mouse model of OUD. Finally, these studies further our understanding of the role of microglia in fentanyl use in the hopes of developing novel neuroimmune therapies for this debilitating disease.

Keywords: Fentanyl Self-Administration, microglia, opioid use disorder

Disclosure: Nothing to disclose.

P437. Acute cannabis exposure alters microglia and oligodendrocyte precursor cell density, morphology and ultrastructure in a sex-specific manner

Colin Murray, Hayley Thorpe, Hakan Kayir, Haley Vecchiarelli, Sophia Loewen, Emiko Osborne, Sean Foster, Marie-Eve Tremblay, Jibran Khokhar

Western University, London, Canada

Background: Recreational cannabis use, cannabis legalization and cannabis potency have been increasing worldwide over the past decade. With the advent of more etiological preclinical administration models, we are now able to understand better how vaporized cannabis exposure (the most common route of administration) alters cells in the central nervous system. Importantly, it is becoming increasingly understood that glial cells, including microglia and oligodendrocyte lineage cells (including oligodendrocyte precursor cells), express cannabinoid receptor(s) and can respond to cannabis administration. These cells can potentially contribute to the intoxicating and acute effects of cannabis. As well, these cells may be a mechanism by which cannabis exerts potential beneficial or deleterious effects.

Methods: Whole cannabis plant was administered to young adult, male and female, C57BL/6J mice for 15 min (one 15-second puff every 5 min; 3 puffs total; 0.15 g flower/puff). Four groups were utilized, mice that received control air vapor, and mice that were exposed to either: high CBD/low THC [CBD], high THC/low CBD [THC], or balanced THC/CBD [Balanced] cannabis chemovars. Brains were isolated 30 min post-cannabis administration onset, when THC levels peak in the brain. We stained the tissue with antibodies to identify microglia (IBA1-microglia and macrophages; TMEM119-more specific for microglia) or oligodendrocyte lineage cells (Olig2-oligodendrocyte lineage cells; CC1-mature oligodendrocytes; PDGFRα-oligodendrocyte precursor cells). Correlative light and scanning electron microscopy was utilized to assess microglia or oligodendrocyte precursor cell density and distribution, morphology and ultrastructural features in the prelimbic area and forceps minor (anterior projections from the genu of corpus callosum), respectively.

Results: For microglia in the prelimbic area, all types of cannabis altered microglial density and distribution in males, but this was not due to a change in infiltration of peripheral immune cells. In terms of microglial morphology: exposure to CBD cannabis leads to reduced arbor perimeter and reduced soma size versus Balanced; exposure to THC cannabis leads to reduced soma size versus Balanced; and exposure to Balanced cannabis leads to increased soma size, reduced complexity and increased symmetry. For microglial ultrastructure, there were no changes in the percentage of microglia containing lysosomes, microglia at satellite positions to neurons, or microglia at blood vessels. However, THC cannabis exposure did increase the proportion of microglia containing lipid droplets. However, these changes were only observed in male mice. Neither the density of mature oligodendrocytes nor oligodendrocyte precursor cells was altered in the forceps minor with any type of cannabis exposure. Similarly to the effects on microglia, in males exposed to Balanced cannabis, there was an increase in oligodendrocyte precursor cell soma size and sphericity and nucleus size, and in males exposed to CBD cannabis, there was a reduction in oligodendrocyte precursor cell branching.

Conclusions: Our preliminary data indicates that acute cannabis exposure modifies microglial and oligodendrocyte precursor cell distribution, morphology and ultrastructural features in the prelimbic cortex and forceps minor, respectively, particularly in males. We are currently using scanning electron microscopy to investigate potential changes in organelles (e.g., mitochondria) and interactions with parenchymal elements. This work will lay the foundation for understanding how vaporized cannabis exposure alters glial form and function. Future work will focus on the duration of these changes as well as the outcomes following chronic or repeated exposure.

Keywords: cannabis, microglia, oligodendrocytes

Disclosure: Nothing to disclose.

P438. Targeting glymphatic pathways to ameliorate cognitive impairments in an aged mouse model of perioperative neurocognitive disorder

Yuto Hasegawa, Robyn Wiseman, Gianluca Ursini, Yannan Li, Feiyi Xiong, Sohan Gummadi, Xiaolei Zhu, Tomoyo Sawada, Barbara Slusher, Atsushi Kamiya

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

Background: Perioperative neurocognitive disorder (PND) is a frequent complication in older adults following major surgery and is associated with impairments in learning and memory. Despite its clinical significance, the molecular and cellular mechanisms underlying its pathophysiology remain poorly understood, limiting advances in prevention and treatment. Increasing evidence suggests that surgery-induced peripheral inflammation can trigger neuroimmune responses that compromise brain homeostasis. Among the central mediators of these effects are glial cells, particularly astrocytes, which regulate glutamate levels and mediate cerebrospinal fluid (CSF) dependent glymphatic clearance via aquaporin-4 (AQP4) at perivascular end-feet. Although aging and inflammation are known to disrupt these processes, it remains unclear how surgical intervention affects astrocyte function and glymphatic flow, leading to cognitive impairments. In our current study using an aged mouse model of PND, we investigate whether abdominal surgery-induced systemic inflammation increases glutamate levels in the central nervous system, which may alter AQP4 signaling in perivascular astrocytes and lead to aberrant glymphatic waste clearance and cognitive impairments. We also examine whether pharmacological inhibition of specific enzymes involved in glutamate regulation can normalize astrocyte phenotypes and improve cognitive performance, and whether these processes contribute to sex-specific vulnerability, as epidemiological evidence suggests that older males are at greater risk for postoperative delirium and related cognitive decline.

Methods: 18–20 months old C57BL/6 mice of both sexes were used in this study, and data from male and female mice were analyzed separately to assess potential sex differences. Animals were divided into the following three groups; The first control group was naïve mice without anesthesia and surgical intervention. The second control group (sham) received 2–2.5 % isoflurane in 0.5–1 L/min O2 flow anesthesia for 15 minutes. The surgery group received the same anesthesia setting as the sham group during abdominal surgery involving one minute of intestine manipulation. After two days of recovery from surgery, histochemical, biochemical, and behavioral assays were performed to identify surgery-induced changes in glutamate levels, astrocyte morphology, and AQP4 signaling in perivascular astrocytes, as well as cognitive disturbances, particularly in spatial learning and memory, in aged mice.

Results: Inhalation anesthesia followed by abdominal surgery produced male-specific impairments in recognition and spatial memory in aged mice, as shown by deficits in novel object, novel location, Barnes maze tests. These impairments were absent in females. In males, surgery induced excess hippocampal glutamate levels (p = 0.0343, t = 2.728, df = 6), morphological alterations of perivascular astrocytes (p = 0.0002, t = 5.570, df = 10), loss of AQP4 polarization (p = 0.0002, t = 5.570, df = 10), and reduced glymphatic tracer influx (p = 0.0343, t = 2.728, df = 6). These changes may be related to alterations in astrocyte expressed enzyme involved in glutamate production, which was more abundant in aged mice compared to young adult mice, consistent with postmortem human brain data. Pharmacological inhibition of this enzyme normalized astrocyte structure and ameliorated cognitive performance in males. These findings implicate astrocyte-mediated glutamate accumulation and glymphatic impairment in PND pathophysiology, with heightened vulnerability in aged males.

Conclusions: We identified that inhalation anesthesia followed by abdominal surgery in aged mice resulted in male-specific cognitive impairments that were linked to glymphatic impairment and astrocytic dysfunction. Pharmacological targeting of glutamate production ameliorated astrocyte phenotype and cognitive impairments in males, suggesting a potential therapeutic approach for PND. These findings support the idea that surgery-induced peripheral inflammation may drive sex-dependent alterations in astrocyte and glymphatic function. We are currently investigating whether astrocyte-specific genetic inhibition of glutamate production can ameliorate the phenotypes and conducting astrocyte-specific single cell RNA sequencing to identify sex-dependent transcriptional changes, with the goal of guiding novel therapeutic strategies for PND in the elderly.

Keywords: glymphatic system, Post-operative neurocognitive disorders, Astrocyte

Disclosure: Nothing to disclose.

P439. Chemotherapy-induced estradiol reduction is associated with less relief of anxiety in breast cancer patients

Angela del Aguila, Lindsey Strehle, Seth Adarkwah Yiadom, Robert Wesolowski, Stephanie Gorka, Rebecca Andridge, Leah Pyter

Ohio State University, Columbus, Ohio, United States

Background: Chemotherapy-induced apoptosis of mature ovarian follicles in breast cancer patients likely results in decreased circulating estradiol (E2), although this is rarely measured. Outside of cancer, sudden E2 reductions (e.g., post-childbirth) are associated with increased anxiety and depressive symptoms. About half of breast cancer patients undergoing chemotherapy report anxiety and depressive symptoms, which is a major reason for discontinuing or changing chemotherapy treatment. This study investigated whether reductions in circulating E2 during chemotherapy were associated with increased anxiety and depressive symptoms. Intolerance of uncertainty was assessed as a secondary outcome.

Methods: Seventy-seven breast cancer patients provided blood samples and completed validated questionnaires (PROMIS Anxiety and Depression Short Forms 8a, and the Intolerance of Uncertainty Scale-12) before and during chemotherapy. Plasma E2 concentrations were measured via ELISA. Patients were stratified based on whether E2 levels decreased (n = 51) or did not decrease (n = 17) during chemotherapy. Longitudinal trajectories of the outcomes were analyzed using linear mixed models with random subject-level intercepts, controlling for age and menopausal status, and applying the Kenward-Roger adjustment.

Results: Chemotherapy-induced reductions in circulating E2 were associated with limited relief of pre-chemotherapy anxiety symptoms (p = 0.07) and intolerance of uncertainty (p = 0.43) during chemotherapy, regardless of age and menopausal status. Conversely, participants whose E2 concentrations did not decrease experienced relief from pre-chemotherapy anxiety symptoms (p = 0.0001) and intolerance of uncertainty (p = 0.01) during chemotherapy. Similar trends were observed for depressive symptoms (interaction over time p = 0.06).

Conclusions: This study indicates that chemotherapy-induced reductions in circulating E2 may be associated with less relief from anxiety symptoms in female breast cancer patients undergoing chemotherapy. Also, these findings suggest that monitoring circulating E2 levels may provide valuable mechanistical insights into patients’ psychological responses during chemotherapy and could inform more personalized approaches to supportive care.

Keywords: adult clincial anxiety, estradiol, breast cancer chemotherapy, uncertainty, Depression

Disclosure: Nothing to disclose.

P440. Poster Withdrawn

P441. Psy-Map.com: a genetics browser for neuroscientists

Naghmeh Shargh, Hazal Senturk, Laramie Duncan

Stanford University, Stanford, California, United States

Background: In the past decade, tremendous progress has occurred in the field of psychiatric genetics, owing primarily to the coordinated efforts of the Psychiatric Genomics Consortium (PGC). This international group of thousands of researchers achieved the pooled analysis of data from unprecedentedly large genome-wide association studies (GWAS), with over a million participants. Thanks to the statistical power afforded by these studies, hundreds of risk loci have been identified across psychiatric disorders including schizophrenia, bipolar disorder, PTSD, and all other major psychiatric disorders. These studies have also compellingly demonstrated that psychiatric disorders are highly polygenic, and the fact that (at the level of the population) individual genetic variants have exclusively small effect sizes (owing to evolutionary forces). Despite this undeniable progress, there have been two major barriers to greater translation of psychiatric GWAS results in neurobiological investigations: 1) accessibility of results and 2) lack of a mechanistic framework for systematically studying the effects of polygenic risk for psychiatric disorders. We offer here a solution to these longstanding challenges.

Methods: To address the fact that GWAS results in papers are oftentimes not detailed enough, and in contrast, that full results files can be complex to navigate (e.g. 40 million rows in individual variant files), we have computed gene-level associations for all major psychiatric phenotypes (using MAGMA) and created a browser, Psy-map.com, to display these results. We also briefly describe the appropriate interpretation and limitations of these results. To provide a framework for studying polygenic risk in neurobiological investigations, we have mapped polygenic risk to specific human brain cell types (specifically, each of the 461 human brain cell types reported in Siletti et al. 2023) using our validated GWAS + snRNAseq analysis method (Duncan et al. 2025).

Results: Searchable online in a user-friendly format, one can now search psychiatric GWAS results via three routes of inquiry: 1) by phenotype, 2) by gene, or 3) by cell type. For example, one may initiate a search for schizophrenia by gene, and this will generate a rank-ordered table of all protein coding genes, quantifying their associations with schizophrenia from the best powered GWAS of schizophrenia available to date (Trubetskoy et al. 2022). One may also view, in graphical or tabular format, the full cell-type association results for a disorder such as schizophrenia, enabling prioritizing of specific cell types for follow up studies (e.g., top cell type #239, cortical somatostatin interneuron, p = 4.3e-17). Alternatively, given interest in a particular gene, one may search for the gene and automatically generate a table of gene-level associations for any major psychiatric disorder of interest (e.g. schizophrenia, bipolar disorder, and/or depression). Genes can also be viewed from the perspective of expression across the 461 human brain cell types (from the most comprehensive atlas of human brain cell types available to date, Siletti et al. 2023). For example, NeuN (RBFOX3 gene) is used to label mature neurons. Our browser shows that while this gene is certainly preferentially expressed by neurons as contrasted with non-neurons, expression is highly variable in different neuron types, with a few neuronal cell types with little to no expression of this gene. This enables researchers to confirm that selected marker genes will likely label the desired targets, and it also reveals specific possibilities for off-target labeling.

Conclusions: This freely available browser, Psy-map.com, is offered as a community resource for neuroscience, psychiatry, and related disciplines. By making GWAS and human brain cell type results more accessible, we hope to accelerate progress in translating psychiatric genetic results into greater neurobiological understanding of these conditions and, ultimately, better treatments. A prototype of the browser is available now, and we anticipate preprint release on or before ACNP 2026.

Keywords: GWAS, CELL TYPE SPECIFIC, Schizophrenia (SCZ), snRNA-seq, psychiatric disorders

Disclosure: Nothing to disclose.

P442. Uncovering latent brain imaging phenotypes leveraging genetic pleiotropy

Todd Lencz, Upasana Bhattacharyya, Jibin John, Max Lam

Zucker Hillside Hospital, Glen Oaks, New York, United States

Background: Genome-wide association studies (GWAS) of imaging derived phenotypes (IDPs) have recently obtained sufficient power to yield significant results. However, the large number and high collinearity of IDPs has hampered interpretation of GWAS results, and has limited the ability of researchers to dissect pleiotropic relationships with neuropsychiatric phenotypes. Here, we present a novel analytic approach to identifying the latent genetic architecture of high-dimensionality pleiotropic phenotypes: DIMPLE-GWAS (Dimensionality reduction and Integrated Multi-Phenotype Landscape Explorer for GWAS), developed to enhance gene prioritization, downstream functional annotation, and association with neuropsychiatric phenotypes.

Methods: We obtained GWAS summary statistics from 3935 neuroimaging IDPs spanning multiple modalities (T1 structural, DTI, fMRI) in N = 33,224 European-ancestry participants from the UK Biobank. After removing phenotypes with heritability not significantly different from zero, we extracted 86,885 LD-pruned SNPs from 2135 IDPs. The DIMPLE-GWAS framework was then applied to the resulting 2135*86885 data matrix: First, using UMAP, we constructed a robust low-dimensional representation of IDP co-heritability profiles. Next, we applied HDBSCAN-based clustering to this UMAP embedding, identifying genetically homogeneous IDP clusters. Finally, summary statistics for all IDPs within each cluster were meta-analyzed via genetic PCA. For each set of cluster summary statistics, independent loci were identified using FUMA. We performed further biological annotation and characterization using eQTL/pQTL analysis, and associations with psychiatric, neurological, and cognitive traits were tested using genetic correlation (LDSC) and Mendelian randomization.

Results: DIMPLE-GWAS identified 25 genetically homogeneous IDP clusters with 422 genome-wide significant loci; of these, 189 loci were unique to a single cluster, and the rest were shared amongst two or more clusters. While derived entirely based on genetic data, the clusters were marked by either regional or modality characteristics. For example, four clusters represented frontal, temporal, parietal, and occipital volumes/areas, while two additional clusters exclusively contained dorsal stream or ventral stream structural volumes. Other clusters represented distinct white matter tracts and resting-state functional networks. Heritability for each of these clusters [median = 0.267] was substantially larger than for the constituent IDPs [median = 0.185].

Dozens of novel associations were identified in downstream analyses, yielding biological insights relevant to neuropsychiatric phenotypes. As one example, we (and others) have previously reported association of higher circulating levels of ITIH1 protein with greater risk for schizophrenia, bipolar disorder, and cognitive deficits; this same protein was associated with lower ventral stream volumes. Mendelian randomization indicated that dorsal stream volumes were associated with cognitive task performance, while subcortical volumes were correlated with reaction time.

Conclusions: The DIMPLE-GWAS framework provides a scalable approach for mapping complex pleiotropic relationships in imaging genetics. This pipeline facilitates robust gene and pathway prioritization by leveraging dimensionality reduction, genetic clustering, and probabilistic fine-mapping. Our results highlight novel links between brain structure/function and mental health, suggesting new avenues for biological discovery in the era of large-scale multi-phenotype genomics.

Keywords: Imaging genomics, Pleiotropy analysis, GWAS, Mendelian Randomization

Disclosure: Nothing to disclose.

P443. Genetic Architecture of Puberty-Related and Sex Hormone Phenotypes, and Genetic Associations With Psychiatric Traits

Megan Patterson, Naomi Friedman, Matthew Keller, Andrew Grotzinger

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

Background: Genome-Wide Association Studies (GWAS) of psychiatric traits, puberty-related traits, and sex hormones have identified associated genetic loci. However, little is known about their genetic relatedness despite well-established phenotypic associations between pubertal markers and mental health. Genomic Structural Equation Modeling (Genomic SEM) is a statistical package the leverages existing GWAS summary statistics to identify genomic and sampling covariance matrices to which structural equation models can be fit, enabling nested model comparisons to identify the optimal architecture of genomic associations between traits. The current study uses Genomic SEM to identify the genetic architecture of puberty-related traits including sex hormones, and subsequently uses QTrait analyses to investigate whether genetic associations of psychiatric traits with puberty-related and hormone phenotypes operate through the identified latent factors.

Methods: Puberty-related GWAS summary statistics included age of menarche (AAM, N = 252,514); relative age of facial hair (FH, N = 186,626) and voice break (VB, N = 178,532); and sex-stratified pubertal growth (female age 10 height [H10F], N = 6974; male age 12 height [H12M], N = 6,986; height difference age 8-adult [HDTotalF, HDTotalM], female N = 5756, male N = 5043; height difference age 14-adult [HDAdolF, HDAdolM], female N = 4946, male N = 4282); as well as sex-stratified sex hormones, including, sex hormone-binding globulin (female SHBGF, N = 189,473; male SHBGM, N = 180,094), estradiol (female E2F, Neff = 115,613; male E2M, Neff = 48,629), and testosterone (female TF, N = 230,454; male TM, N = 194,453). Recent Psychiatric Genomics Consortium GWAS summary statistics were used for psychiatric traits, including anxiety (Neff = 51,867), depression (Neff = 1,105,086), PTSD (Neff = 447,559), alcohol dependence (Neff = 19,112), schizophrenia (Neff = 114,792), bipolar (Neff = 101,963), and anorexia nervosa (AN, Neff = 46,322). Mean effective sample size (Neff) is reported for case-control studies. Genomic SEM model comparisons were used to identify the best fit model for to the puberty-related and sex hormone phenotypes. Following this, QTrait indices (chi-square differences) compared models in which each psychiatric trait was associated with the latent factor(s) (i.e., homogenous associations operating via factor) vs. models in which each psychiatric trait was associated with individual indicators (i.e., heterogenous associations operating on specific indicators). Associations with psychiatric traits were considered significant at a Bonferonni corrected threshold of p < .007 for seven psychiatric traits.

Results: A correlated factors model was identified as the best fit model to the puberty-related and sex hormone phenotypes, identified by three factors: female puberty (AAM, l = 0.87; H10F, l = −0.39; HDTotalF, l = 0.65; HDAdolF, l = 0.76; ps < 0.001), male puberty (FH, l = 0.74; VB, l = 0.90; HDAdolM, l = 0.91; ps < 0.001; VB-FH residual correlation r = 0.69, p = 0.01), and sex hormones (SHBGF, l = 0.96; SHBGM, l = 0.90; E2F, l = 0.19; TM, l = 0.74; E2F p = 0.01, all other ps < 0.001), all of which were significantly correlated (female-male, r = 0.90; female-sex hormones, r = 0.14; male-sex hormones, r = 0.17; ps < 0.001). QTrait indices indicated largely heterogenous associations (QTrait ps < 0.05), with the exception of significant homogenous (i.e., factor-level) associations of schizophrenia with the sex hormone factor (B = .07), and depression and PTSD with the male puberty factor (depression, B = −0.09; PTSD, B = −0.10). Heterogeneous significant associations were noted between anxiety and AAM (B = −0.12); depression and female puberty (AAM, B = −0.08, H10F, B = −0.12) and sex hormone indicators (SHBGF, B = −0.09; SHBGM, B = −0.04; TM, B = −0.07); and PTSD and anorexia nervosa and sex hormone indicators (PTSD: SHBGF, B = −0.13; SHBGM, B = −0.06; TM, B = −0.07; anorexia nervosa: SHBGF, B = −0.28; SHBGM, B = −0.16; TM, B = −0.18).

Conclusions: The genetic architecture of puberty-related phenotypes was best captured by separate latent factors for female and male puberty, which were highly positively correlated with one another, indicating a large deal of genetic overlap across sexes. These factors were also positively correlated a sex hormones latent factor, which was identified by SHBG in both sexes, and the dominant sex hormone in each sex (estradiol in females and testosterone in males). Associations with psychiatric traits were primarily found for internalizing disorders, namely depression, anxiety, PTSD, and anorexia nervosa, such earlier pubertal timing and lower levels of sex hormones were associated with greater genetic liability for psychiatric diagnosis, largely consistent with phenotypic research. The male puberty latent factor, and the specific indicators of age of menarche, SHBG in both sexes, and male testosterone were each linked to more than one psychiatric trait, indicating some similar genetic risk conferred across disorders. The exception to this pattern was schizophrenia, which was positively genetically associated with the sex hormones latent factor, such that higher levels of sex hormones were associated with genetic liability for schizophrenia. This suggests that schizophrenia may have a different pathway of genetic risk linked to sex hormones than internalizing-type disorders.

Keywords: psychiatric genomics, Puberty, Sex Hormones

Disclosure: Nothing to disclose.

P444. Multi-Ancestry Genome-Wide Association Study of Short-Term Weight Change Following Atypical Antipsychotic Exposure in the All of Us Research Program

Allison Lake, Lea Davis, Maria Niarchou, Jeffrey Neul

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

Background: Atypical antipsychotics (ATAP) are the mainstay of treatment for schizophrenia and are commonly used as adjunctive therapy in other psychiatric conditions including bipolar disorder and depression. Several ATAP medications are associated with adverse metabolic effects including weight gain, with rapid weight gain typically occurring in the first several weeks of treatment. Common genetic variants have been associated with ATAP-induced weight gain (AIWG), but the precise genetic mechanisms of this process remain unclear. Furthermore, pharmacogenomic testing holds promise for individualizing antipsychotic treatment to reduce risk of weight gain, yet previous genome-wide association studies (GWAS) are limited by small sample size and limited generalizability across psychiatric diagnoses and populations.

Methods: All of Us Research Program participants with available whole-genome sequencing data, longitudinal anthropometric data, and evidence of ATAP treatment were identified. Pregnant individuals, those with cancer diagnoses or treatment codes, those undergoing bariatric surgery or amputation, and those with evidence of treatment in a critical care facility at the time of ATAP initiation were excluded. The final analysis cohort consisted of 2,228 adults categorized by genetic similarity to global reference populations into African (AFR, N = 653) and European (EUR, N = 1575) ancestry groups. BMI change within the first 12 weeks of ATAP initiation was calculated relative to baseline pre-ATAP BMI for each participant. Ancestry-stratified GWAS of percent change in BMI per month of follow-up was performed in REGENIE.

Results: Our GWAS cohorts were primarily female (AFR, 57.4%; EUR, 59.8%), with a median age of approximately 48 years. The majority of participants carried at least one mental health diagnosis code (AFR, 96.6%, EUR, 94.4%), with a preponderance of depression (AFR, 77.5%; EUR, 80.1%), bipolar disorder (AFR, 53.0%; EUR, 47.6%), and schizophrenia spectrum disorder (AFR, 46.2%; EUR, 19.5%) diagnoses documented in the 30 days prior to ATAP initiation (diagnosis codes were not mutually exclusive). On average, patients underwent a 0.67% - 0.84% increase in BMI per month of follow-up, with 21.2%–23.0% of participants undergoing clinically significant weight gain (≥7% BMI change per month of follow-up). Preliminary GWAS revealed two genome-wide significant loci (p < 5e-8) associated with percent change in BMI per month: rs601315 (AFR cohort, nearest gene RNU1-45P) and rs12444803 (EUR cohort, nearest gene TLDC1).

Conclusions: Our study contributes to a small but rapidly growing body of AIWG genetics literature by providing the largest GWAS of AIWG in African-ancestry individuals to date, addressing a crucial gap in the literature. Furthermore, the inclusion of a diverse range of psychiatric conditions improves cross-disorder generalizability over prior work which has primarily been limited to schizophrenia. Our preliminary results identify two novel loci, with ongoing external replication, meta-analysis, and functional analyses in progress.

Keywords: pharmacogenomics, Antipsychotic induced weight gain, GWAS

Disclosure: Nothing to disclose.

P445. Rare coding variation implicates 36 large-effect risk genes in obsessive-compulsive disorder and tic disorders

Belinda Wang, Matthew Tran, Sheng Wang, Yuting Liu, George Wang, Nawei Sun, Emily Olfson, TIC Genetics, Foundation for OCD Research, Kerry Ressler, Carol Mathews, Jay Tischfield, Gary Heiman, Nenad Sestan, A Moses Lee, Thomas Fernandez, A Jeremy Willsey, Matthew State

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

Background: Obsessive-compulsive disorder (OCD) and chronic tic disorders (TD), including Tourette syndrome, are early-onset neuropsychiatric conditions that affect 1–3% of the population. Both are highly heritable (OCD ~50%, TD ~77%), but genetic discoveries have not yet clarified mechanisms or advanced clinical care. OCD and TD frequently co-occur and share phenomenological features, supporting their conceptualization along a shared spectrum. Common variants account for most population-level risk, yet there remain significant challenges in translating non-coding alleles of small effect to actionable biology. Rare pathogenic coding variants that implicate large-effect risk genes provide a complementary and more direct route to mechanistic insight. To date, only four genes (two per disorder) have reached accepted significance thresholds (FDR < 0.1) in whole-exome sequencing (WES) studies.

Methods: We conducted a WES meta-analysis, roughly doubling prior sample sizes, focusing on rare and de novo protein-disrupting mutations. The case cohort included 2,418 parent–proband trios (1,550 TD, 1,493 OCD, 2,418 combined ‘OCDTD’) and 1,546 OCD singletons, of which 872 trios (36%) and 1,080 singletons (70%) were newly sequenced. Controls comprised 1,734 neurotypical trios. De novo and rare variants were called using a unified pipeline and classified by predicted functional impact, with analyses focused on likely gene-disrupting (LGD: frameshift, stop-gain, essential splice site) and damaging missense (Mis3) mutations. Risk genes were identified using the Bayesian Transmission and De Novo Association (TADA) framework, which integrates evidence across variant classes while accounting for effect sizes and gene-specific mutation rates. Identified risk genes were further assessed for overlap with OCD common variant risk from the largest genome-wide association study (GWAS), tissue and cell type-specific expression in transcriptomic atlases, and functional connectivity in PCNet, a multimodal gene interaction database.

Results: We identified 36 high-confidence (hc) genes (FDR < 0.1) associated with OCD and/or TD (12 hcOCD, 10 hcTD, 34 hcOCDTD), four of which overlapped genes from a recent OCD GWAS. Comparisons with hc genes from autism spectrum disorder, developmental delay/intellectual disability, and schizophrenia revealed both shared and distinct gene sets. Transcriptomic analyses showed significant risk gene overrepresentation (Benjamini-Hochberg adjusted p < 0.05) in genes specifically expressed in the brain and telencephalic projecting glutamatergic neurons, and nominal overrepresentation (p < 0.05) in striatal medium spiny neurons, consistent with cortico-striatal involvement. Risk genes formed significantly connected networks in PCNet (hcOCD: p = 0.0184; hcTD: p = 0.03; hcOCDTD: p = 0.0028), and their interactors were enriched for both rare OCD/TD and common OCD variants.

Conclusions: This work increases the number of high-confidence risk genes from 4 to 36, demonstrates overlap with common variant risk, and shows shared genetic risk with other neurodevelopmental disorders. Risk genes are highly expressed in cortical excitatory projection neurons and striatal medium spiny neurons, implicating cortical-striatal circuits. Their interaction networks are highly connected and enriched for common variant risk, suggesting shared biological mechanisms.

Keywords: Human Genetics, Obsessive Compulsive Disorder, Tic disorders, Tourette syndrome, Rare genetic variation

Disclosure: Nothing to disclose.

P446. Pilot study of the genetics of intense imagery movements

Sarah Abdallah, Emily Olfson, Tamsin Owen, Osman Malik, Raj Seraya Bhatoa, Tammy Hedderly, Thomas Fernandez

Yale University, Child Study Center, New Haven, Connecticut, United States

Background: Complex motor stereotypies (CMS) are childhood-onset motor disorders described as repetitive, rhythmic movements, usually without premonitory urges. These movements can lead to functional impairment or self-injury, and there are no clear pharmacologic treatment options. Although their pathophysiology is not well-understood, aggregation of stereotypies within families suggests a genetic basis. Our group recently conducted whole-exome DNA sequencing of primary CMS in 129 affected children and their unaffected parents, demonstrating an increased rate of de novo predicted-damaging genetic variants compared to controls and identifying the first high-confidence risk gene for CMS.

The phenotypic heterogeneity of CMS poses a challenge in understanding their biology. Investigation of sub-phenotypes may parse out stronger genetic signals, as has been theorized for other heterogeneous disorders like autism spectrum disorder (ASD). Clinicians in the United Kingdom have identified a unique subset of children with motor stereotypies who engage in concurrent acts of imagery or imagination, termed Intense Imagery Movements (IIM). These episodes can cause distress or functional impairment. Although IIM aggregates in families, there are not yet any reported genetic studies of IIM. Importantly, given their remarkable and consistent clinical presentation, IIM are considered a more homogenous subset of stereotypies.

This study is a preliminary examination of genetic variants in 23 children with IIM and their parents through whole-exome sequencing of DNA samples. This is the first study to our knowledge that will examine genetic variation in this specific phenotype. Earlier studies of ASD and obsessive-compulsive disorder in 20 affected individuals successfully detected increased mutation rates, potentially causative mutations, and relevant biological pathways in these disorders.

Methods: DNA samples were collected from 22 families comprising children with IIM and their unaffected parents (one family contained two affected children). DNA sequencing and genotyping data from these samples was compared against 115 sex-matched control families comprising unaffected children and their parents from the Simons Simplex Collection. Exome capture and DNA sequencing of IIM samples was performed using the Illumina NovaSeq 6000 platform. We employed our well-validated analysis pipeline to detect rare de novo and inherited single-nucleotide variants and indels from whole-exome DNA sequencing data using the latest Genome Analysis Toolkit Best Practices guidelines. We then annotated variants to determine the predicted effect of each variant, gene location, and variant frequency in large publicly available databases. Frameshift, nonsense, and canonical splice site mutations were defined as likely gene disrupting (LGD). We filtered out variants that did not pass quality control, those present in more than two samples within our set, and those with frequency > 0.00005 in the public database gnomAD. We removed samples with an excess of de novo variants (more than 10). We calculated rates of inherited and de novo single-nucleotide and insertion-deletion variants per callable base pair using the R package rateratio.test. We will use this tool to compare the rates of each type of genetic variant in children with IIM compared to unaffected children. Following annotation of rare variants, downstream pathway and expression analyses were conducted using the ConsensusPathDB and CSEA tools, with genes overlapping rare variants in children with IIM as the input.

Results: This analysis identified 245 rare LGD variants (8 de novo) among 23 IIM probands. The rate of all rare LGD variants per callable base pair in children with IIM (1.2 × 10^-7) was not significantly increased compared to controls (1.7 × 10^-7), rate ratio = 0.69, p > 0.05. Similarly, the rate of de novo LGD variants in IIM (4.2 × 10-9) was not significantly increased compared to controls (6.9 × 10-9), rate ratio = 0.61, p > 0.5. The overall rate of de novo variants in IIM (4.2 × 10^-8) was not significantly increased compared to controls (4.7 × 10^-8), rate ratio = 0.89, p > 0.5. Among all rare damaging variants (LGD and damaging missense) in children with IIM we find overlap with four genes previously implicated in ASD (CHD8, SHANK3, DYNC1H1, and TAF4). Preliminary downstream analysis of genes containing rare LGD variants in IIM revealed top significant pathways were Codeine and Morphine Metabolism (2.14 x10^-5), Synaptic vesicle cycle (0.000193), Mycophenolic Acid Metabolism Pathway (0.000622), and Na + /Cl- dependent neurotransmitter transporters (0.00105). Specific expression analysis of all rare LGD variant genes in IIM revealed top nominal significance in cerebellum middle-late childhood (p = 0.007), cerebellum early childhood (p = 0.017), and thalamus adolescence (p = 0.02).

Conclusions: This study is the first genetic analysis of IIM and identifies several likely damaging de novo variants in children with IIM. A limitation of this analysis is the underrepresentation of female children with IIM in the pilot sample, which prohibits sex-specific analyses. Future sample ascertainment would be aimed at collecting a more representative IIM sample. Still, this pilot study serves an impetus for further investigation in a larger sample of IIM families with the hopes of using the identified genetic variants to implicate risk genes and biological pathways that contribute to this understudied neurodevelopmental phenotype.

Keywords: motor stereotypies, Rare genetic variants, Whole exome sequencing

Disclosure: Nothing to disclose.

P447. Mapping acute effects of cannabis on multiple memory domains: a double-blind, placebo-controlled study

Carrie Cuttler, Ryan McLaughlin

Washington State University, Pullman, Washington, United States

Background: One of the most robust detrimental effects of acute cannabis intoxication is impairments in memory. However, memory is not a unitary construct. Rather, memory is a complex system comprised of multiple, distinct, subsystems that are interrelated, have unique characteristics and developmental courses, employ different neural mechanisms, and can be selectively impaired. Nevertheless, the overwhelming bulk of previous research examining acute effects of cannabis on memory have focused on verbal memory and working memory. A paucity of research has considered impacts of acute cannabis use on more ecologically valid domains of memory that better reflect the events and tasks people need to remember in everyday life, such as temporal order memory, prospective memory, source memory, false memory, and episodic memory. Moreover, to our knowledge no prior research has examined all these aspects of memory in a single study. This is problematic as considerable variability in methodologies exist across studies, including location (lab vs. field), method of administration (oral, smoked, vaped), potency of cannabis, and lack of inclusion of a placebo control group, which complicates cross-study comparisons within different memory domains.

Methods: A total of 121 healthy cannabis-using participants were randomly assigned to vaporize placebo (n = 40), 20 mg delta-9-tetrahydrocannabinol (THC) (n = 40), or 40 mg THC (n = 41) using a Volcano Vaporizer. However, one participant experienced an adverse event (vomiting) and was excluded. The final sample of 120 participants (56 men, 52 women, 12 transgender/non-binary) subsequently complete a battery of memory tests designed to measure the following memory domains: event-cued prospective memory (the ability to remember to execute tasks upon the occurrence of a specific event), time-cued prospective memory (the ability to remember to execute tasks at a specific time), verbal free recall (the ability to remember lists of words), verbal short-term memory (the ability to immediately recall verbal information), verbal working memory (the ability to store and manipulate verbal information), visuospatial memory (the ability to recall visual information), spatial short-term memory (the ability to immediately recall spatial information), spatial working memory (the ability to store and manipulate spatial information), source memory (the ability to remember the source of previously learned information), false memory (the recollection of items that were not previously presented), episodic memory (the ability to recall past events), and temporal order memory (the ability to recall the order in which past events occurred). The study was pre-registered on clinicaltrials.gov [#NCT05488509] and was approved by Washington State University’s Institutional Review Board (IRB), the Food and Drug Administration (FDA), and Drug Enforcement Agency (DEA).

Results: Results of a series of one-way ANOVAs revealed significant detrimental effects of THC on event-cued prospective memory (p = .032), source memory for pictures (p < .001) and words (p = .027), false memory (p < .001), temporal order memory (p = .004), immediate (p = .004) and delayed (p = .023) visuospatial memory, immediate (p = .003) and delayed (p = .023) verbal free recall, and verbal working memory (p = .003). There were no significant effects of THC detected on verbal short-term memory (p = .179), spatial short-term memory (p = .05), spatial working memory (p = .135), time-cued prospective memory (p = 897), or episodic memory (p = .462). There were no significant difference in the memory test performance of participants administered 20 mg vs. 40 mg THC, which likely reflects the fact that both groups reported comparable levels of intoxication and drug effects.

Conclusions: This is the first study to detect detrimental effects of acute cannabis intoxication on event-cued prospective memory and temporal order memory, which is critical as these tests better reflect everyday life memory and better predict daily functioning. Overall, results suggest that acute cannabis intoxication has fairly broad impacts on multiple domains of memory rather than selectively impairing only a limited number of memory domains.

Keywords: Cannabis, Memory, Randomized clinical trial

Disclosure: Nothing to disclose.

P448. Altered stimulus-response learning and updating in posttraumatic stress disorder

Krystian Loetscher, DT Nguyen, John Krystal, Stephanie Groman, Elizabeth Goldfarb

Yale University, New Haven, Connecticut, United States

Background: Posttraumatic stress disorder (PTSD) is a debilitating illness characterized by paradoxical memories. To date, memory research in PTSD has focused on memories for details of and emotional reactions to cues. However, nonhuman animal and acute stress findings indicate that another type of memory – “habits”, or gradually acquired, rigid, instrumental associations between stimuli and responses – can be enhanced by stress. Such memories have been theorized to drive maladaptive cue-induced responses like avoidance, yet stimulus-response learning and updating have not been investigated in PTSD. Here we tested whether this novel memory system is generally enhanced with trauma or specifically engaged in PTSD.

Methods: We recruited individuals meeting diagnostic criteria for PTSD (N = 35; mean = 31.9 years old, 71.4% female/28.6% male) and an age- and sex-matched control group of individuals who had experienced trauma but did not meet current or lifetime criteria for PTSD (TC, N = 35; mean = 29.6 years old, 65.7% female/34.3% male). Sample sizes were determined based on a power analysis of successful probabilistic learning (80% power, alpha = 0.05). After an intake appointment to determine eligibility, participants returned to complete a novel multiphase probabilistic learning game designed to measure acquisition and updating of stimulus-response associations while undergoing fMRI. In the learning phase, participants learned to place objects in the correct rooms (left or right door), with optimal responses yielding an object-unique outcome (an image of a fruit, indicating an addition to their monetary bonus). Each participant learned 6 probabilistic object/door associations (80% reinforcement, 300 trials). Following learning, participants entered an unsignaled reversal phase, in which 2 of the 6 objects switched their room association, requiring flexible response updating (300 trials). Finally, participants completed an explicit devaluation procedure in which they were instructed that 2 of the 6 fruit outcomes were “rotten” and would detract from their bonus, requiring value updating (30 trials, no feedback presented).

Results: Participants successfully completed this novel multiphase learning task: they learned probabilistic associations (main effect cue presentation: p < 0.001), switched responses for reversed cues (cue type [reversed or not]: p < 0.001), and decreased responses to cues for which the outcome was devalued (cue type [devalued or not]: p < 0.001). Surprisingly, we found that individuals with PTSD were slower to form probabilistic S-R associations (group x cue presentation: p = 0.037). However, once they did form these associations, they were highly resistant to updating them with a new response (group x cue type [reversed or not]: p < 0.001). There were no differences between groups in responses following outcome devaluation, suggesting comparable value updating (p = 0.11). Examining learning more closely, we found that individuals with PTSD were more likely to misapply feedback to change their behavior in irrelevant situations. Ongoing computational modeling work aims to disentangle how these learning biases may relate to over-reactivity to negative feedback or difficulties with credit assignment.

Conclusions: These results demonstrate novel alterations in learning and updating that characterize individuals with PTSD. Explicating ways in which habit-like learning differs with trauma and PTSD provides an important bridge from nonhuman animal models showing that stress and trauma can enhance neural and behavioral correlates of this memory system to highlight their clinical relevance. Together, this work underscores the need to expand the mechanisms by which memory may contribute to the development and maintenance of PTSD.

Keywords: Probabilistic learning, PTSD, probabilistic reversal learning, habit, trauma

Disclosure: Nothing to disclose.

P449. Causal evidence for visual cortical disinhibition in human fear memory

Joshua Brown, Yijia Ma, Yifan Wang, Mingzhou Ding, Raymond Dolan, Chenyu You, Wen Li

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

Background: Fear learning and memory are central to anxiety and stress disorders, yet the neural mechanisms in humans remain incompletely understood. Animal models highlight sensory cortical disinhibition—a transient release of primary neurons from inhibitory control—as a key mechanism for fear memory formation, but direct human evidence is lacking.

Methods: We combined high-definition α-frequency (10 Hz) transcranial alternating current stimulation (α-tACS) over the visual cortex with multimodal neuroimaging (simultaneous EEG-fMRI and high-density EEG) and AI-based transformer decoding. Participants (N = 81, 53 females) underwent fear conditioning while receiving α-tACS or sham stimulation (randomly assigned). Memory traces were probed using EEG and fMRI at 15 minutes, 24 hours, and 7 days post-conditioning to assess early formation, consolidation, and retention of fear memory.

Results: α-tACS reduced visual cortical excitability (Z = 3.98, FDR p = 0.040) and attenuated CS+-potentiated alpha-band (8–12 Hz) event-related desynchronization (F = 8.56, p = 0.004). Transformer-based decoding of multimodal neural data revealed disrupted fear memory traces in the α-tACS group, including impaired decoding in the amygdala, anterior insula, and secondary visual cortex (V2), as well as diminished EEG-based decoding across all assessed time points (all p’s < 0.05). Importantly, the magnitude of CS + -potentiated αERD at 15 min post-conditioning predicted decoding accuracy at 24 hr and 7 days, linking early visual cortical disinhibition to long-term fear memory consolidation and retention.

Conclusions: This study provides the first causal evidence in humans that sensory cortical disinhibition drives fear memory formation. By showing that α-tACS disrupts this mechanism, the findings establish a novel target for mechanism-based interventions in anxiety and stress disorders.

Keywords: Fear learning, fear memory, EEG-fMRI, tACS, Neural decoding

Disclosure: Nothing to disclose.

P450. ACSS2 inhibition blocks reinstatement of cue-induced drug seeking in rodent self-administration models

Natalie Belle, Thomas Kim, Katherine Nicholson, Miriam Shin, Ethan Anderson, Christopher Cowan, Keith Shelton, Paul Kenny, Philipp Mews

Chobanian and Avedisian School of Medicine at Boston University, Boston, Massachusetts, United States

Background: Substance use disorder (SUD) can be understood as a disorder of learning and memory in which drugs of abuse hijack associative mechanisms to form persistent cue–drug memories that drive craving and relapse long after abstinence. These maladaptive associations depend on epigenetically regulated gene programs, including histone acetylation, in mesolimbic circuits. Acetyl-CoA synthetase 2 (ACSS2) localizes to neuronal chromatin and generates acetyl-CoA in situ, supporting activity-dependent histone acetylation and immediate early gene expression critical for drug-cue memory consolidation. Genetic or pharmacological disruption of ACSS2 modulates memory formation without producing broad neurological deficits, positioning it as a selective neuroepigenetic control point. To directly test whether targeting ACSS2 could weaken maladaptive drug–cue memories, we employed rat self-administration models of nicotine, cocaine, and oxycodone, in which drug intake is paired with discrete cues that later provoke relapse-like behavior. We hypothesized that brief, brain-penetrant ACSS2 inhibition administered during drug-taking sessions—when drug-paired cues are repeatedly encoded and re-engaged—would disrupt reconsolidation of these associations and attenuate cue-induced reinstatement across drug classes. This approach targets the underlying cue memory process rather than a specific receptor system, offering a novel transdiagnostic therapeutic strategy for relapse prevention.

Methods: Adult male rats were implanted with intravenous catheters and trained to self‑administer nicotine, cocaine, or oxycodone in standard operant chambers. Presses on the active lever delivered an infusion paired with a discrete light (±tone) cue; an inactive lever had no scheduled consequence. After acquisition to stable intake, animals received a short‑acting ACSS2 inhibitor (ACSS2i) during the drug‑taking phase: intraperitoneal dosing occurred 20 min before and again immediately after each daily self‑administration (SA) session for the final 5 consecutive days (for oxycodone, the post‑session dose followed ~30 min). No ACSS2i was administered during abstinence, extinction, or the reinstatement tests to avoid acute confounds. Following the dosing during SA, rats underwent extinction (no drug, cues withheld) to prespecified criteria, then a drug‑free cue‑induced reinstatement test in which active‑lever presses produced the previously drug‑paired cues without infusion; in the oxycodone cohort, a 6‑s non‑contingent cue was also presented at test onset. Across experiments, sample sizes per arm (vehicle vs. ACSS2i) were nicotine n = 12, cocaine n = 8, and oxycodone n = 12, males only. Primary outcomes were active‑lever responses during reinstatement; inactive‑lever responses were analyzed as controls for nonspecific activity. Statistical analyses used mixed/repeated‑measures ANOVA for behavior across the dosing‑during‑SA and extinction phases, and planned between‑group t‑tests for reinstatement (α = 0.05).

Results: Transient ACSS2 inhibition (ACSS2i) delivered during daily self‑administration (SA) sessions did not alter ongoing drug taking or extinction prior to relapse testing, with no acute treatment effect and inactive‑lever presses unchanged compared to controls. In subsequent drug‑free cue‑induced reinstatement tests, prior ACSS2i treatment consistently reduced relapse‑like seeking across drug classes: nicotine (vehicle vs ACSS2i ≈60% reduction; p < 0.0001), cocaine (vehicle vs ACSS2i ≈55% reduction; p = 0.0028), and oxycodone (vehicle vs ACSS2i ≈50% reduction; p = 0.0099).

Conclusions: Brief, well‑timed ACSS2i exposure during the drug‑taking phase produces a robust, cross‑drug attenuation (~50–60%) of cue‑induced reinstatement while sparing acquisition/maintenance of drug taking and extinction dynamics. Together, these findings position ACSS2 as a transdiagnostic neuroepigenetic target and support testing ACSS2i as a temporally precise adjunct to cue‑exposure–based therapies for relapse prevention.

Keywords: Epigenetics, cue-induced-reinstatement, Self-Administration

Disclosure: EpiVario, Inc., Founder, Self

P451. Associative mechanisms of midbrain acetylcholine release

Serena Miller, Savatdy Selena, Ye Liv, Russell Gabrielle, Kurt Fraser

University of Minnesota, Minneapolis, Minnesota, United States

Background: Dopamine release arising from midbrain neurons in the ventral tegmental area is thought to underlie the attribution of motivational value to cues paired with rewarding and aversive outcomes. The regulation of dopamine neuron activity, thus, is paramount to associating and regulating behavioral responses to reward- and fear-paired stimuli. The activity of VTA dopamine neurons is tightly regulated with limited sources of direct excitation onto these neurons. However, acetylcholine release arising from the brainstem cholinergic neurons in groups C5 and C6 provide a direct excitatory input to VTA dopamine neurons. It has been established that these inputs generally control rudimentary features of reward yet we fail to understand the psychological consequences of midbrain acetylcholine release

Methods: To better understand the function of midbrain acetylcholine release we implanted rats with cannula targeting the VTA and assessed acetylcholine signaling in Pavlovian fear and reward conditioning. In fear conditioning, Long-evans rats (n = 32, 16 male 16 female; 8 per group) were infused with 300 nL of saline, mecamylamine (a nicotinic receptor antagonist), scopolamine (a muscarinic receptor antagonist), or physostigmine (an inhibitor of acetylcholine clearance, which locally elevates acetylcholine levels) into the VTA prior to fear conditioning and then tested later for fear discrimination in tests of conditioned suppression. For reward conditioning (n = 24 rats, 12 male, 12 male) we assessed 1) consumption of different rewarding concentrations of sucrose 2) Pavlovian conditioned responsing, and 3) contextual renewal following intra-VTA saline, mecamylamine, scopolamine, and physostigmine.

Results: Antagonism of VTA muscarinic signaling selectively abolished contextual memory for an aversive context but preserved cue memory. In tests of reward, antagonism or enhancement of VTA acetylcholine altered sucrose preference. VTA muscarinic antagonism disorganized Pavlovian conditioned responding such that rats responded equally to a CS+ and CS- and further resulted in a loss of contextual renewal.

Conclusions: These findings lay the initial groundwork for an associative and circuit-level account for the functions of midbrain acetylcholine release. Collectively we aim to provide a comprehensive account of the function of midbrain acetylcholine release in the acquisition, expression, and modification of cue-generated expectations.

Keywords: Dopamine, Acetylcholine, Ventral Tegmental Area (VTA), Aversive Motivation, incentive motivation, Context

Disclosure: Nothing to disclose.

P452. Somatostatin interneurons drive extinction memory encoding in the infralimbic cortex

Rodrigo Campos-Cardoso, Hunter Franks, Noelle Potter, Faith Brown, Zephyr Desa, Brianna Fitzgerald, Kirstie Cummings

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

Background: The ventromedial prefrontal cortex is critical for the extinction, or suppression, of fear memories across species. In rodents, glutamatergic signaling in the infralimbic (IL) subregion of the ventromedial prefrontal cortex has long been known to be required for extinction behaviors. We previously found that the neighboring prelimbic cortex requires the activity and plasticity of somatostatin-interneurons (SST-INs) for proper fear memory encoding and expression. Whether interneuron activity and plasticity is required for extinction memory processes in IL is largely understudied.

Methods: Adult (P > 60) male and female mice were used for this study. We performed cFos immunohistochemistry (n = 5–6 mice/group), activity-dependent neural tagging (n = 5–6 mice/group), brain slice electrophysiology (n = 14–23 cells/group; 3–5 mice/group), fiber photometry (n = 9 mice/group), and in vivo optogenetics (n = 8–12 mice/group). Groups include naive mice, fear conditioned mice, and fear extinguished mice. Parametric statistics were used if data passed Levine’s and Shapiro-Wilk tests. If either failed, non-parametric equivalents were used.

Results: cFos immunohistochemistry (p < 0.0001, one-way ANOVA) and activity-dependent neural tagging (p < 0.006, Kruskal-Wallis) revealed that IL somatostatin-interneurons (SST-INs) are strongly recruited to the prefrontal fear extinction neural ensemble. Brain slice electrophysiology revealed that the frequency of miniature excitatory postsynaptic currents onto IL SST-INs is enhanced after extinction compared to naive or fear conditioned groups (p < 0.0001, Kruskal-Wallis). Fiber photometry calcium imaging revealed that IL SST-INs exhibit enhanced cue-related activity across extinction learning and retrieval (p < 0.01, paired t-test). In vivo optogenetic manipulations revealed that activating or silencing IL SST-INs during extinction memory encoding enhances or suppresses fear extinction encoding and expression, respectively (p < 0.001, Kruskal-Wallis). We additionally measured IL microcircuit alterations and observed strengthening of inhibition from SST-INs onto PV-INs after extinction training, driving disinhibition of IL PNs (p < 0.0001, Kruskal-Wallis). Finally, we measured experience-dependent alterations of basolateral amygdala inputs onto IL SST-INs and found that fear memory extinction drives enhanced synaptic strength of glutamatergic inputs (p < 0.0001, Kruskal-Wallis). Silencing these inputs during extinction memory encoding blocks extinction behavior and the associated learning-related plasticity onto IL SST-INs (p < 0.0001, Mann-Whitney U-test).

Conclusions: Overall, infralimbic cortex SST-INs exhibit activity, connectivity, and plasticity alterations that are required for successful fear memory extinction.

Keywords: Extinction, Somatostatin-expressing interneurons, infralimbic cortex, parvalbumin interneurons

Disclosure: Nothing to disclose.

P453. Neuromodulatory control of entorhinal–hippocampal circuits in sensory processing for memory

Heechul Jun

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

Background: The entorhinal cortex serves as a critical interface between the hippocampus and neocortex, integrating sensory information into memory representations. The lateral entorhinal cortex, in particular, transforms external cues into learning-relevant signals. Neuromodulatory inputs such as acetylcholine (ACh) and serotonin (5-HT) are increasingly recognized as key regulators of associative learning, yet their coordinated dynamics across the entorhinal–hippocampal circuit remain poorly understood. Understanding these dynamics is essential for elucidating how new sensory information is incorporated into existing memory frameworks. Here, we examined the spatiotemporal patterns of ACh and 5-HT signaling in entorhinal and hippocampal networks during memory-guided behavior and sleep.

Methods: Male and female C57BL/6J mice received viral delivery of genetically encoded GRAB sensors for ACh and 5-HT along with chronic implantation of silicon probes. Fiber photometry was used to monitor neuromodulatory signals while simultaneously recording hippocampal and entorhinal local field potentials and unit activity at subsecond to submillisecond resolution. Animals performed a delayed odor match-to-sample task, with recordings collected during pre-task sleep, task engagement, and post-task sleep to compare neuromodulatory dynamics across brain states.

Results: As reported previously, hippocampal sharp wave ripples were preceded by rapid decreases in ACh levels. Notably, ACh and 5-HT fluctuations were strongly correlated between entorhinal and hippocampal regions during NREM sleep, whereas during wakefulness and task performance, correlations diminished and region-specific dynamics emerged. These findings reveal state-dependent divergence in neuromodulatory coordination, with synchronized activity during sleep and differentiated signaling during behavior.

Conclusions: This study provides the first simultaneous recordings of ACh and 5-HT dynamics in hippocampal and entorhinal circuits. The strong inter-regional correlation during NREM sleep suggests a coordinated neuromodulatory state that supports memory consolidation. In contrast, decorrelated signaling during wake and task performance may reflect selective gating of entorhinal inputs to the hippocampus, aligning external sensory representations with ongoing memory processing. Together, these findings highlight how neuromodulatory systems orchestrate entorhinal–hippocampal interactions across brain states to regulate learning and memory.

Keywords: Serotonin, Acetylcholine, Memory and Learning, Entorhinal-hippocampal circuitry

Disclosure: Nothing to disclose.

P454. Egocentric coding of geometric features in the mouse anterior cingulate cortex

Indrajith Nair, Kritika Ramesh, Sachie Ogawa, Takashi Kitamura

The University of Texas Southwestern Medical Center, Dallas, Texas, United States

Background: Animals perform action as a motor output in the self-perspective. For this to happen, the allocentric spatial map is transformed into an egocentric spatial map, and is then used by the animals to perform motor action by the secondary motor cortex (M2). Retrosplenial Cortex (RSC) is implicated in the transformation of allocentric to egocentric framework. However, it remains unclear how the information in the egocentric map is transformed for action. Anatomical studies have shown that Anterior Cingulate Cortex (ACC) receives input from RSC and is projected to M2 and is responsive to objects and social cues. These results suggest that ACC could be the site for map to action transformation. Therefore, we hypothesize that ACC could encode a wide variety of geometric features in egocentric fashion.

Methods: To study the representational schema of the ACC, we expressed GCaMP6f in the ACC neurons using AAV infection and implanted a GRIN lens to monitor calcium activity in the ACC during spatial navigation.

Results: We found that a subset of ACC neurons encoded borders, convex and concave corners, objects, and social cues (P < 0.05 compared with shuffled data, N = 6 mice). Beyond these features, ACC neurons also categorized diverse environmental elements. Moreover, we identified ACC neurons that represented not only environmental boundaries but also spatial connectivity, indicating that the ACC supports topological spatial representation. Notably, these geometry-encoding cells exhibited egocentric responses, suggesting that the ACC represents the abstract concept of space for goal-directed behavior. Finally, the emergence of these representations required repeated exposure to the environment (N = 6 mice).

Conclusions: Our data suggests that the ACC is potentially acting as a gateway to successful motor output by representing geometric features on the environment, social and objects in an egocentric fashion, much like a contour map which provides a ‘birds eye view’ of the space to the animal.

Keywords: Anterior Cingulate Cortex (ACC), Spatial Navigation, egocentic representation, Hippocampus

Disclosure: Nothing to disclose.

P455. The contributions of the circulating endocannabinoids and beta-endorphins to placebo analgesia

Rossi Tomin, Kevin Murray, Georgia Hadjis, Omar Khalil, Christine Sexton, James Khan, David Finn, Lauren Atlas, Massieh Moayedi

University of Toronto, Toronto, Canada

Background: Placebo analgesia reduces pain in response to inert treatment through psychosocial manipulation. This analgesia results from top-down activation of endogenous pain modulatory systems. There are individual differences in placebo responsivity and the extent of analgesia in responders. Ample evidence indicates that placebo analgesia is mediated by endogenous opioids. However, the opioid antagonist naloxone does not completely abolish placebo analgesia, suggesting the involvement of non-opioidergic systems. The endogenous cannabinoid (eCB) system has emerged as a potentially significant contributor to placebo analgesia.

The eCB system comprises cannabinoid receptors, ligands, including arachidonoylethanolamide (also called anandamide; AEA), 2-arachidonoylglycerol (2-AG), and metabolizing enzymes (e.g., fatty acid amide hydrolase—FAAH). N-acylethanolamines (NAEs), such as N-palmitoylethanolamide (PEA) and N-oleoylethanolamide (OEA), can also be considered part of the eCB system as they can elevate AEA via substrate competition at FAAH. Preclinical evidence has shown that eCBs/NAEs are present at every level of the nociceptive system, as well as key nodes of the descending modulatory pathway.

The overall aim of this study was to determine whether circulating eCBs/NAEs contribute to placebo analgesia. Specifically, we aim to determine whether baseline levels of eCB/NAE ligands predict individual differences in placebo analgesia. Given the known involvement of the endogenous opioid system in placebo analgesia, we also aim to determine whether baseline circulating beta-endorphin (BE) levels are associated with placebo analgesia. We further tested whether the presence of a FAAH C385A allele, sex, and cannabis use affect placebo analgesia. Next, we aimed to determine whether dynamic changes in circulating eCBs/NAEs, induced by a placebo paradigm, are associated with the extent of analgesia produced. We further investigated the relationship between placebo-induced changes in BE and the extent of analgesia produced. Finally, we explored whether there is an interaction between circulating BE and eCB/NAE levels in placebo analgesia.

Methods: Forty-eight healthy adult participants (23 female, 25 male) recruited from the University of Toronto and surrounding environment underwent procedures approved by the University of Toronto’s Human Research Ethics Board (Protocol # 45552).

Participants fasted for 12h (overnight) before attending the lab session. Participants provided a baseline blood sample. Next, they underwent placebo conditioning, where they were told an inert cream applied to the volar forearm was a potent analgesic, and a second identical cream applied to an adjacent skin region was a control cream. In the manipulation phase, the placebo cream was paired with low intensity heat stimuli (4/10) while the control cream was paired with a high intensity heat stimulus (8/10), to generate the belief that the placebo cream was analgesic. In the test phase, identical heat stimuli (6/10) were applied to a new set of placebo and control sites on the volar forearm. The order of placebo and control conditions was counterbalanced across participants. We collected blood samples before and after the placebo condition, and before and after the control condition. Circulating levels of eCBs/NAEs in serum were determined with liquid chromatography-tandem mass spectrometry for each timepoint, and we calculated changes in each condition (placebo, control). Serum BE was quantified by ELISA. As the FAAH substrates (AEA, OEA and PEA) were highly correlated, we performed a data reduction step with principal components analysis (PCA).

Statistical Analysis: We first tested if we induced a placebo effect (paired t-test). We then performed 5 sets of linear mixed models (LMMs) to test whether individual differences in placebo analgesia are associated with (1) baseline FAAH substrates and 2-AG, (2) baseline BE, (3) changes in FAAH substrates and 2-AG, (4) changes in BE, and (6) changes in FAAH substrates and BE, to determine whether they interact. LMMs included sex, cannabis-use, FAAH C385A genotype, order of conditions (placebo, control), and testing time to account for circadian effects. Significance was set to p < 0.05.

Results: We successfully induced placebo analgesia (p < 0.0001). The PCA yielded a single component of FAAH substrates. LMMs found that baseline eCB (p > 0.05) and BE levels (p > 0.05) collected prior to the task did not predict analgesia. However, changes in FAAH substrates (post-task – pre-task) were associated with placebo analgesia (B = –0.64, SE = 0.306, t(57.6) = –2.09, p = 0.041). Surprisingly, changes in BE levels were not associated with placebo analgesia (p = 0.417). We did, however, find that the change in BE levels moderated the relationship between FAAH-substrates and placebo analgesia (B = 0.75, SE = 0.356, t(26.2) = 2.12, p = 0.044), such that those with high BE reported placebo analgesia, whereas those with lower BE showed placebo analgesic effects associated to changes in FAAH substrates levels.

Conclusions: During a placebo task, changes in circulating FAAH substrates, but not BE, are associated with individual differences in placebo analgesia. Furthermore, the relationship between FAAH substrates and placebo analgesia is moderated by BE. These findings suggest a synergistic effect between FAAH substrates and BE in placebo analgesia, and provide novel evidence of interactions between the eCB and opioidergic systems in pain modulation.

Keywords: pain analgesia, Placebo Effect, Endocannabinoids, endogenous opioids, FAAH

Disclosure: Nothing to disclose.

P456. Brain metabolite changes following MDMA-assisted therapy in veterans with comorbid PTSD and alcohol use disorder comorbidity

Lewis Nunez Severino, Brian Gully, Erica Eaton, Christy Capone, Mollie Monnig, Michael Warden, Robert Swift, Carolina Haass-Koffler

Brown University, Providence, Rhode Island, United States

Background: The co-occurrence of posttraumatic stress disorder (PTSD) and alcohol use disorder (AUD) is highly prevalent among U.S. veterans and contributes to a reinforcing cycle of hyperarousal, emotional avoidance, and maladaptive alcohol use. Existing psychotherapies and pharmacological treatments for PTSD-AUD remain limited, with high rates of persistent symptoms and relapse. A promising intervention is 3,4-methylenedioxymethamphetamine (MDMA) assisted therapy (MDMA-AT), which may enhance emotional processing, reduce fear responses, and increase openness in therapy. Neuroimaging studies using magnetic resonance spectroscopy (MRS) have reported elevated myo-inositol (mI), a glial marker of neuroinflammation, in the amygdala and frontal cortex of individuals with PTSD and AUD. By targeting these processes, MDMA-AT may reduce alcohol misuse while disrupting the reinforcing cycle sustaining PTSD-AUD.

Methods: This is an open-label pilot trial in U.S. veterans with comorbid PTSD-AUD (n = 12) to evaluate the therapeutic effects of MDMA-AT and examine associated brain changes. Participants completed an 11-week treatment cycle consisting of three preparatory sessions, two MDMA-AT experimental sessions, and integration therapy. PTSD and alcohol use outcomes were assessed with the Clinician-Administered PTSD Scale (CAPS-5) and the Timeline Follow-Back (TLFB), a structured interview that quantifies daily alcohol use. A subset of participants (n = 9) also underwent single-voxel 1H-MRS at baseline and post-treatment to assess metabolite concentrations in fronto-limbic regions (amygdala, frontal cortex) and a control occipital voxel. MRS data were processed using the Osprey pipeline with LCModel fitting. Voxels were co-registered to each participant’s anatomical image and segmented into gray matter, white matter, and CSF fractions. Metabolite estimates were then tissue-corrected and expressed as ratios relative to total creatine (tCr), with myo-inositol to creatine (mI/tCr) examined as a marker of glial and neuroinflammatory processes. Clinical and neural measures were compared before and after MDMA-AT.

Results: Compared to baseline, we found a reduction in PTSD symptom severity (CAPS-5) and alcohol use (TLFB) following MDMA-AT treatment. Tissue-corrected MRS analyses revealed a decrease in the mI/tCr ratio in the amygdala and an increase in the frontal voxel, consistent with region-specific modulation of glial activation. Within-subject changes in mI/tCr appeared to align with improvements in clinical outcomes, suggesting that metabolite shifts linked to neuroinflammatory processes may serve as candidate biomarkers of treatment response.

Conclusions: These preliminary findings suggest that MDMA-AT not only reduces PTSD and AUD symptomology but also modulates neurometabolites implicated in neuroinflammation. Ongoing work will incorporate peripheral cytokine assays to further characterize the inflammatory profile of treatment response. Together, these results provide early insight into the neural effects of MDMA-AT and support the potential of MRS-based biomarkers to guide the development of targeted interventions for PTSD-AUD comorbidity.

Clinicaltrials.gov: NCT05943665

FDA: IND 165005

Funding: Brown Office of the Vice president for Research OVPR) and P20 GM130414

Keywords: Combat PTSD, Alcohol Use Disorder - Treatment, MDMA, 1H MRS, neuroinflamation

Disclosure: Nothing to disclose.

P457. Radiogenetic neuromodulation of cell activity

Aurelio Galli

The University of Alabama At Birmingham, Birmingham, Alabama, United States

Background: Biological techniques (e.g., optogenetics, chemogenetics) that enable the regulation of genetically defined neurons offer powerful opportunities to investigate how neural circuits control specific behaviors. Optogenetics allows for precise control of neuronal activity and neurotransmitter release in living tissue and freely moving animals. However, it has several limitations: it requires permanent intracranial implants, its effects are confined to the illuminated region of the brain, and the intensity of stimulation is difficult to regulate. To address these limitations, we developed a novel technology platform using X-ray (Xray) irradiation.

Methods: Xray irradiation of physiological (aqueous) solutions rapidly generates H₂O₂.

Results: Our findings show that the increases in intracellular Ca²⁺, membrane depolarization/action potential firing, and neurotransmitter release induced by Xray in dTRPA1(A)10b-expressing preparations are mediated through Xray-induced H₂O₂ production and subsequent activation of dTRPA1 channels. Xray stimulation does not activate mouse TRPA1 (mTRPA1) or induce neurotransmitter release in murine brain preparations. Therefore, targeted viral expression of dTRPA1(A)10b in specific neurons in murine models may allow Xray to control their activity and associated behaviors. Finally, we began translating this technology to mammals by virally expressing dTRPA1(A)10b in vivo in defined mouse brain regions, enabling Xray to selectively activate these neuronal populations and drive neurotransmitter release.

Conclusions: We provide a mechanistic description of how Xray activation of dTRPA1(A)10b leads to increases in intracellular Ca²⁺, membrane depolarization, neurotransmitter release, and behavioral changes, using cells, neurons, and Drosophila/mouse as model systems. We define this approach as radiogenetic neuromodulation.

Keywords: Non-invasive Neuromodulation, Dopamine, Circuit-function

Disclosure: Nothing to disclose.

P458. Hippocampal input-prefrontal interneuron interactions supporting spatial working memory

Shana Silverstein, Thomas Clarity, Meena Deshpande, Erik Vaughan, Shoshana Novik, Hector Yarur, Shiliang Zhang, Valerie Tsai, Gabriel Loewinger, Madeline Hsiang, Francisco Pereira, Marisela Morales, Vikaas Sohal, Hugo Tejeda, Joshua Gordon, David Kupferschmidt

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

Background: Dynamic communication between the ventral hippocampus (vHPC) and medial prefrontal cortex (mPFC) is essential for various learning and memory processes, including spatial working memory (SWM). Disrupted vHPC-mPFC connectivity has been linked to psychiatric disease-related SWM deficits. Discrete classes of inhibitory mPFC interneurons support SWM and vHPC-mPFC coupling as well as receive direct excitatory inputs from vHPC. While interactions between vHPC projections and mPFC interneurons, and their plasticity, are well-positioned to influence SWM, the nature of these interactions remains unclear.

Methods: We combined in vivo optogenetic stimulation of vHPC inputs to mPFC with fiber photometry calcium recordings of mPFC vasoactive intestinal polypeptide (VIP)- or somatostatin (SST)-expressing interneurons in transgenic mice (VIP-Cre, SST-Cre) to record stimulation-evoked responses. Male and female mice underwent six stimulation-recording sessions across 50 days with a separate group receiving an additional 12 sessions of high frequency stimulation (HFS) (VIP n = 11,12; SST n = 12,13). Ex vivo whole-cell electrophysiology was performed on mPFC tissue from mice with or without prior HFS to characterize synaptic adaptations in VIP and SST interneurons (VIP n = 16,12 cells; SST 8,12 cells). Circuit-level computational modeling assessed the influence of altered synaptic weights due to vHPC input stimulation. Behavioral and calcium activity dynamics were measured during a delayed non-match-to-sample T-maze task in mice with or without prior HFS (n = 27,24). Task-dependent calcium activity was compared between wild-type and 22q11.2 deletion syndrome mice (n = 8,9). Finally, electron microscopy assessed the post-synaptic targeting of monosynaptic vHPC inputs on mPFC interneurons in wild-type and 22q11.2 deletion mice (n = 6,6).

Results: Repeated vHPC stimulation persistently depressed mPFC VIP activity [2-way ANOVA, Day: F(3.349, 70.33) = 12.81, p < 0.05] and potentiated SST interneuron activity [2-way ANOVA, Day: F(1.934, 44.49) = 29.48, p < 0.0001]. Ex vivo recordings revealed that prior stimulation reduced vHPC monosynaptic connectivity to VIP interneurons biophysically [AMPAR oEPSCs: p < 0.005, U = 33; NMDAR oEPSCs: p < 0.01, U = 40] and pharmacologically [AMPAR oEPSCs: p < 0.05, U = 15; NMDAR oEPSCs: p < 0.01, U = 11], but not SST. Computational modeling of weakened vHPC input onto VIP interneurons could plausibly enhance SST interneuron responses, matching in vivo results. During T-maze training, prior vHPC input stimulation potentiated VIP activity in early learning during the delay epoch when spatial information must be maintained [Stimulation x Stage: F(1, 19) = 7.135, p < 0.05]. This activity profile was recapitulated in 22q11.2 deletion mice [Genotype x Stage: F(1, 15) = 9.657, p < 0.01]. Electron microscopy confirmed reduced vHPC targeting of VIP interneurons in 22q11.2 deletion mice [t(10) = 2.71, p < 0.05]. In both HFS and 22q11.2 deletion mice, aberrant VIP activity during early training correlated with poorer SWM performance [HFS: F(1, 19) = 7.453, p < 0.05, R2 = 0.2817; 22q11.2: F(1, 15) = 14.73, p < 0.005; R2 = 0.4955].

Conclusions: These findings reveal cell-type-specific plasticity in vHPC-mPFC circuit and demonstrate how reweighting excitatory inputs to VIP interneurons can reshape local inhibitory dynamics and disrupt SWM. Parallel alterations in the schizophrenia-predisposing 22q11.2 deletion syndrome mice suggest impaired inhibitory microcircuit targeting may represent a shared mechanism contributing to both cognitive and psychiatric dysfunction.

Keywords: GABAergic interneurons, Hippocampal-prefrontal, Spatial working memory, 22q11 deletion syndrome

Disclosure: Nothing to disclose.

P459. The paraventricular thalamus gates coding of salient experiences in the hippocampus

Mark Gergues, Gregory Telian, Kelsey Logas, Mona Li, Arsine Kolanjian, Jeremy Biane, Shazreh Hassan, Jasmine Tai, Victoria Turner, Mazen Kheirbek

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

Background: Recent work has revealed the hippocampus (HPC), which has classically been implicated in learning, for its role in encoding emotionally charged environments and generating appropriate behavioral responses. Yet how and which inputs are involved in the generation of emotion-related representations in the ventral HPC remain unknown. Here, we used high-throughput single-neuron tracing, opto- and chemogenetic manipulations, and one and two-photon calcium imaging to characterize a direct projection from the anterior paraventricular thalamus (aPVT) that modulates emotional stimuli processing in the ventral hippocampus (vHPC).

Methods: We used a Cre-dependent strategy to label vHPC-projecting cells of the paraventricular thalamus (PVT) with the inhibitory DREADD receptor hM4Di. We injected retroAAV-Cre in vHPC and a Cre-dependent chemogenetic inhibitor, hM4Di, into PVT, both bilaterally. Four weeks later we injected animals with CNO (3mg/kg) 15 minutes before running the elevated plus maze and open field test. To assess the collateral projections of vHPC-projecting cells in PVT, we used anatomical tracers (a Cre-dependent expression of Synaptophysin-mRuby in PVT and retroAAV-Cre in vHPC). Four weeks later, animals were sacrificed, brains extracted, followed by whole-brain histology to identify axon fibers to other brain regions. For MAPseq studies, we injected mice with Sindbis Virus expressing the viral barcode library into PVT and extracted brain regions of interest to be sent for RNA sequencing. 1-photon freely moving behavior was done with Inscopix nVoke with 0.6mm diameter GRIN lenses implanted above vHPC with DREADD inhibition of PVT, then animals were run through the EPM. 2-photon experiments consisted of a 3-odor associative learning, where each odor is paired with either a sucrose reward, a tail shock, or a null.

Results: In this study, we identified how an extrahippocampal input, specifically, the anterior paraventricular thalamic nucleus (aPVT), shapes ventral hippocampal representations of salient stimuli and associated behavioral responses. In addition, we describe the anatomical characterization of this projection, which has not previously been done. In both our synaptophysin and MAPseq experiments, we identified axon collaterals to the nucleus accumbens, central amygdala, bed nucleus of the stria terminalis, and zona incerta.

Inhibition and excitation of the PVT-vHPC pathway resulted in reduced time in the open and center of the elevated plus maze (p < 0.05) and reduced time in the center of the open field test (p < 0.05). Our freely moving calcium imaging reveals that the inhibition of PVT reduces selectivity for anxiogenic contexts in vHPC (p < 0.05). Associative learning experiments under 2-photon imaging show that coding for salient cues predicting rewarding and aversive outcomes is impaired at the population level (p < 0.05).

In sum, our optogenetic, chemogenetic, and in vivo calcium imaging approaches, we demonstrate that dynamic communication between the aPVT and vHPC is essential for appropriately balancing responses to appetitive and aversive stimuli, and for regulating exploratory versus avoidance behaviors. Inhibition of this pathway enhanced vHPC representations of innately rewarding and safe stimuli (e.g., sucrose, closed arms), while dampening responses to innately aversive stimuli (e.g., tail shocks, open arms). These findings suggest that aPVT input biases vHPC activity toward processing threat-related information, potentially at the expense of reward-related encoding.

Conclusions: Here, we provide direct evidence demonstrating that the anterior PVT-to-vHPC projection constitutes a critical circuit for hippocampal encoding of motivationally salient stimuli and modulating anxiety-related behavior. Our findings suggest that this direct thalamic-hippocampal pathway updates the classic Papez model, offering new mechanistic insights into the neural substrates of emotional memory encoding and anxiety. By identifying and characterizing the function of this pathway, we uncover how midline thalamic input shapes hippocampal processing of salience at the circuit level, providing a foundation for targeted neuromodulation strategies to correct dysregulated avoidance and anxiety-related behaviors.

Keywords: paraventricular nucleus of the thalamus, Hippocampus, Anxiety

Disclosure: Nothing to disclose.

P460. Prefrontal cortex bidirectionally modulates periaqueductal gray activity via glutamatergic and dynorphinergic mechanisms

Zahra Farahbakhsh, Lindsay Cain, Zev Jarrett, Jose Zepeda, Brad Grueter, Cody Siciliano

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

Background: Prior work has demonstrated that aberrant activity patterns in medial prefrontal cortex (mPFC) neurons projecting to the dorsal periaqueductal gray (dPAG) are a prodromal marker of alcohol use disorder and putative causative agent mediating punishment insensitivity. mPFC-PAG cells bidirectionally encode the valence of unconditioned stimuli, displaying inhibition to appetitive stimuli such as sucrose and excitation to aversive stimuli such as footshock. Functional studies using channelrhodopsin-assisted circuit mapping have demonstrated that photostimulation of mPFC terminals in the dPAG elicits fast excitatory post-synaptic potentials (EPSPs); however, following stimulation there is a long-lasting inhibition in the rate of spontaneous EPSPs such that the net effect of stimulation is inhibitory. Intriguingly, human neuroimaging suggests that two regulatory pathways between the PFC and PAG, one excitatory and one inhibitory, differentially contribute to valence coding and alcohol use disorder pathophysiology. Here, we sought to determine how valence-specific information routed through the mPFC→dPAG circuit is transferred to post-synaptic targets in the dPAG.

Methods: GCaMP6m was expressed under a general neuronal promoter in the dPAG and the excitatory red-shifted opsin ChrimsonR was expressed in the mPFC of 6 male C57BL/6J mice. Chronic optical access was gained via a gradient-index (GRIN) lens implanted above the dPAG, allowing for single-cell calcium imaging of dPAG neurons and simultaneous local photostimulation of mPFC terminals. Within a single session to maintain the field of view, single-cell recordings were obtained while freely moving mice had 3 15-minute blocks: open access to sucrose, unsignaled footshock, and optogenetic activation of mPFC terminals (5 minute baseline, 5 minutes of stimulation [blocks of 5 seconds of 20 Hz, 25 seconds off], 5 minutes post-stimulation). To determine the receptor system most likely implicated in the functional effect of mPFC terminal activation, we analyzed single cell RNA-sequencing of projection-identified mPFC neurons whereby multiple anatomically-defined populations had been labeled and isolated (n = 2 – 5 per experiment) prior to sequencing (accession number: GSE161936). For whole-cell patch-clamp electrophysiology experiments, recordings were conducted in the dPAG in acute ex vivo slices. Spontaneous excitatory post synaptic currents (sEPSCs) were recorded at baseline and after bath application of 1 µM U50,488, a selective kappa opioid receptor (KOR) agonist. To interrogate mPFC inputs, chanelrhodopsin-2 was expressed in mPFC neurons and presynaptic terminals were photostimulated in slices before and after bath application of a NorBNI (10 nM), a selective KOR antagonist. For all pairwise comparisons between two conditions or groups, we utilized t-tests and comparisons across three or more variables were made using one- or two-way ANOVAs. All tests were two-sided and alpha was set to 0.05.

Results: Through simultaneous photostimulation of mPFC inputs and calcium imaging of dPAG neurons, we were able to dissect circuitry through an ‘anterograde phototagging approach’ and classify cells based on upstream functional connectivity. Aligning to photostimulation onset revealed excitatory, inhibitory, and unresponsive profiles which were used to categorize the cells into ‘Network+’, ‘Network-’, and ‘Non-Network’, respectively. Network+ cells were bidirectionally modulated by valence such that the largest proportion of Network+ cells were activated during sucrose consumption and inhibited by unsignaled shock (X2 = 18.81, p < 0.0001). The cell types were also differently responsive to positive valence (One-way ANOVA, F(1, 178) = 5.736, p = 0.0039) with Network+ cells increasing activity during sucrose licks more than Network- (Šídák’s test, t178 = 2.73, p = 0.02) and more than Non-Network cells (Šídák’s test, t178 = 3.237, p = 0.004). In addition to acute modulation, there was a long-lasting reduction in event rate in Network+ and Network- cells, and a reduction in the amplitude of response in Network+ cells after repeated stimulations. Single cell sequencing analysis revealed that, of mPFC-dPAG cells, 33% expressed detectable levels of PDYN mRNA, a greater proportion than those expressing any other inhibitory neuromodulator, including the other endogenous opioid ligands, PENK (X2 = 6.383, p = 0.0115) and POMC (X2 = 7.37, p = 0.007). Neurons were also enriched in PDYN mRNA compared to the other projection populations tested (Kruskal-Wallis ANOVA, statistic = 109, p < 0.0001). Whole-cell recordings in dPAG revealed that repeated photostimulation of mPFC inputs decreased the frequency of sEPSCs (one-sample t-test, H0 = 0, t11 = 5.6, p = 0.0002) mirroring the effects of photostimulation of mPFC terminals in vivo. KOR activity was sufficient to recapitulate decreased sEPSCs frequency, as measured by U50,488 application (paired t-test, t4 = 3.39, p = 0.028) and was necessary for the inhibitory effect of photostimulation of mPFC terminals as demonstrated by application of NorBNI (one-sample t-test, H0 = 0, t12 = 1.9, p = 0.08).

Conclusions: Together, the results suggest that mPFC inputs to the dPAG bidirectionally modulate valence-responsive cells through both glutamatergic and dynorphinergic mechanisms. This work elucidates a novel neuromodulatory circuit that dynamically controls neurons critical to adaptive and maladaptive motivated behaviors.

Keywords: Dynorphin, Alcohol and substance use disorders, periaqueductal grey (PAG), Medial Prefrontal Cortex (mPFC)

Disclosure: Nothing to disclose.

P461. Distinct encoding of reward and aversion by D1 cell subtypes in the nucleus accumbens

Matthew Pomrenze, Nicholas Denomme, Jinhee Baek, May Wang, Gavin Touponse, Jason Tucciarone, Neir Eshel, Robert Malenka

Stanford University, Stanford, California, United States

Background: The nucleus accumbens (NAc) is a key integration site for the control of motivated behaviors. Not only is diverse information routed to the NAc by glutamate and neuromodulator (i.e. dopamine (DA)) inputs, a heterogenous assortment of cell-types locally integrate and process information related to reward, aversion, and motivation. Canonical models of the NAc describe two major medium spiny neuron subtypes, expressing DA D1 or D2 receptors, which promote reward and aversion, respectively. However, recent advances in single-cell RNA sequencing have revealed numerous subpopulations of neurons within these major families. We recently identified an eccentric cell-type in the D1 family, expressing Tshz1, that operates distinctly from canonical D1 neurons, who express Pdyn. How these D1 cell subtypes modulate local NAc activity to encode reward and aversion remains poorly understood.

Methods: We used a combination of strategies in female and male transgenic recombinase driver mice. Cell activity was monitored in vivo with genetically-encoded DA, acetylcholine (ACh), and calcium sensors (n = 6–8 mice), while activity was manipulated with optogenetic tools (n = 8–14 mice). Analysis of single-cell RNA seq data collected from NAc tissue from naïve mice and post-mortem humans was performed. Electrophysiology was performed in acute slices harvested from the NAc (n = 10–25 cells/4–8 mice). Appetitive behavior was tested with social interaction and pavlovian and operant sucrose conditioning, while aversive behavior was tested with shock conditioning (n = 8–16 mice). All studies were analyzed with unpaired, two tailed-tests, two-way ANOVAs, repeated measures when appropriate, and mixed effect linear models when appropriate.

Results: Single-cell RNA seq identified distinct populations of D1 neurons, including those expressing Tshz1 or Pdyn, in mouse and human NAc. Optical stimulation of Tshz1 neurons in the NAc medial shell produced sustained inhibition of DA release and was aversive, while stimulation of NAc Pdyn neurons increased DA release and was reinforcing. Stimulation of Tshz1 neurons also generated sustained inhibition of neighboring Pdyn and ChAT neurons, but less so in D2 or PV- and SST-expressing interneurons. Surprisingly, light offset triggered an excitatory response in SST interneurons. Stimulation of Pdyn neurons also inhibited ChAT neurons, but this inhibition decayed within 1–2 seconds. Both Tshz1- and Pdyn-induced inhibition of ChAT cell activity produced an excitatory rebound after light was turned off, and the magnitude of this rebound was correlated with the magnitude of inhibition. Photometry recordings revealed increased activity in both Tshz1 and Pdyn neurons during social interaction (Tshz1: t(8) = 6.1, p = 0.0003, Pdyn: t(7) = 8.8, p = 0.0001). During palatable food intake in the home cage, Tshz1 neurons displayed inhibition (t(8) = 4.9, p = 0.0026) while Pdyn neurons were excited (t(7) = 3.7, p = 0.0065). ChAT neurons presented with complex waveforms during both tests. In pavlovian and operant sucrose tasks, Tshz1 neurons consistently showed reduced activity when sucrose was consumed (t(8) = 4.6, p = 0.0012) and Pdyn neurons showed enhanced activity (t(7) = 5.7, p = 0.0002). ChAT neurons again displayed multi-phasic responses during sucrose cue presentation. In response to foot shocks, both Tshz1 and Pdyn neurons were excited (Tshz1: t(8) = 3.2, p = 0.01, Pdyn: t(7) = 10.1, p = 0.0001) while ChAT neurons showed strong inhibition followed by a large rebound. RNA seq data revealed enrichment of Oprm1 in Tshz1 neurons, suggesting that endogenous opioid release may contribute to the feeding-induced inhibition.

Conclusions: We have identified divergent responses to appetitive and aversive stimuli in two distinct D1 neuronal populations in the NAc. These data uncover how distinct cell subtypes in the NAc may be encoding and controlling appetitive and aversive behavioral responses. Future studies that further dissect the local control of neuronal activity will clarify the dynamic roles the NAc plays in affective behavior.

Keywords: Nucelus accumbens, Cell-type diversity, microcircuit processing, reward and aversion

Disclosure: Nothing to disclose.

P462. Chronic alcohol effects on nociception and neural properties of descending midbrain projections during withdrawal

I. Pamela Alonso-Vazquez, Carter Frankum, Jake Shapiro, Jason Middleton, Nicholas Gilpin

Louisiana State University Health Sciences Center, New Orleans, Louisiana, United States

Background: Pain and alcohol use disorder (AUD) are highly prevalent and comorbid conditions. When they are present together, pain worsens clinical prognoses and reduces abstinence rates in individuals with AUD. Although acute alcohol is analgesic, chronic alcohol exposure leads to hyperalgesia during withdrawal and increases sensitivity to subsequent pain challenges. The emergence or exacerbation of pain during withdrawal can drive relapse.

The periaqueductal gray (PAG) is a brain region involved in descending pain modulation, and it mediates the antinociceptive effects of endogenous opioids and cannabinoids. Within the PAG, neurons in the ventrolateral subdivision (vlPAG) that project to the rostral ventromedial medulla (RVM) have a central role in descending pain regulation. Our lab has previously reported that rats chronically exposed to alcohol in adulthood or adolescence exhibit lasting hyperalgesia that is mediated by lower activity of forebrain GABAergic inputs to the vlPAG. While inhibition of glutamatergic neurons in the vlPAG increased nociception, GABAergic activation has similar pronociceptive effects. However, the role of specific RVM-projecting vlPAG circuits in alcohol withdrawal-induced hyperalgesia remains poorly understood. Here, we examined how chronic alcohol affects the intrinsic and synaptic properties of specific subsets of RVM-projecting vlPAG neurons during withdrawal in adult male and female mice.

Methods: Adult C57BL/6J, Vgat-ires-cre, and Vglut2-ires-cre mice were exposed to chronic intermittent ethanol (CIE) vapor or air (controls) for 2 weeks (8 sessions). Mechanical and thermal nociception was assessed using von Frey and Hargreaves tests at baseline and 24, 48, and 72 hours after the last vapor exposure. Whole-cell patch-clamp recordings were performed in vlPAG slices 4–7 days into withdrawal. Viral genetic (Vgat-ires-cre, Vglut2-ires-cre) and retrograde tracing approaches (AAVrg-FLEX-tdTomato) were used to identify RVM-projecting vlPAG neurons.

Results: We found that (1) male and female mice developed mechanical and thermal hypersensitivity after 2 weeks of CIE. (2) Unidentified cells in the vlPAG showed a significant decrease in the amplitude of spontaneous excitatory postsynaptic currents on withdrawal day 4 to 7. (3) In GABAergic vlPAG neurons projecting to the RVM, a history of CIE increases the frequency of spontaneous inhibitory postsynaptic currents. Ongoing experiments are assessing effects on glutamatergic neurons.

Conclusions: CIE model produced hyperalgesia during withdrawal and cell type-specific alterations in vlPAG. Given that the PAG is a major site of analgesic drug action, these findings may have important implications for the development of new treatments for individuals with AUD and comorbid pain.

Keywords: Pain sensitivity, Alcohol Abstinence, midbrain

Disclosure: Nothing to disclose.

P463. Mesolimbic endocannabinoids actuate dopamine-based action selection

Miguel Angel Lujan, Reana Young-Morrison, Joseph Cheer

University of Illinois College of Medicine, Chicago, Illinois, United States

Background: Quickly adapting behavior to constantly evolving environmental demands is key for survival. In the brain, dopamine signals influence behavior by reinforcing actions that procure rewards and avoid adverse outcomes. However, the neuronal and molecular mechanisms within the circuits that orchestrate these processes remain largely unexplored.

Methods: Here, we exploit a novel viral-genetic approach (Lujan, Covey et al, Nature Comms 2023) to selectively delete the enzyme diacylglycerol lipase alpha (DGLa) in VTA TH-positive neurons, therefore impairing the synthesis of 2-AG in dopamine neurons. We also use fiber photometry to measure phasic dopamine release in the NAc and eCB release in the VTA during active avoidance and conditioned food training. To identify the presynaptic target of this 2-AG retrograde signal, we delete cannabinoid receptors type 1 (CB1R) from ventral pallidum (VP) terminals projecting to the VTA. We use causal inference tools such as functional linear mixed models (FLMM, Loewinger et al., eLife 2024) and random forest classifiers to determine the functional relevance of the photometric signals recorded during behavior.

Results: We show that midbrain endocannabinoid mobilization is the conditio sine qua non of dopamine-mediated behavioral invigoration evoked by external cues (t-test; p = 2.1 × 10⁻⁵). Reward- and punishment-predictive signals elicit retrograde release of the principal brain endocannabinoid, 2-arachidonoylglycerol (2-AG), from ventral tegmental area (VTA) dopamine neurons (Satterthwaite; p = 7.2 × 10⁻¹²), a process we show is necessary for striatal dopamine release to engage rapid conditioned responses [two-sided permutation test; FDR-corrected (q = 0.01) p = 0.022]. This temporally-constrained signal targets cannabinoid type-1 receptors (CB1Rs) located on inhibitory terminals arising from the ventral pallidum. Exploiting a novel statistical-computational framework to interpret continuous brain signals, we describe that 2-AG/CB1R communication does not influence dopamine-based error signaling (Satterthwaite; p = 0.10) but rather second-to-second adjustments in action selection required to learn from detected errors (Satterthwaite; p = 0.0105).

Conclusions: These findings reveal an endocannabinoid-mediated temporal filter exploited by dopamine neurons to overcome incoming inhibitory synaptic drive, orchestrate rapid striatal dopamine release events, and enable adaptive action selection informed by exteroceptive cues.

Keywords: Endocannabinoids, Dopamine, negative reinforcement

Disclosure: Nothing to disclose.

P464. Dendritic fields and axon collaterals of modality-specific locus coeruleus projection neurons

Barry Waterhouse, Haven Predale, Daniel Chandler, Nicholas Plummer, Patricia Jensen

Virtua Health College of Medicine and Life Sciences at Rowan University, Berlin, New Jersey, United States

Background: The brainstem nucleus locus coeruleus (LC) projects broadly throughout the forebrain, brainstem, cerebellum, and spinal cord and is a major source of norepinephrine (NE) release in these regions. A long-standing question concerning the LC is the specificity of its afferent and efferent connectivity with respect to its terminal field projections. In addition, while we know much about LC organization with respect to sensory and cognitive circuitries and the impact of LC output on sensory guided behaviors and executive function, less is known about LC-NE influences on motor network operations and movement control. To begin closing this gap in understanding, we used a viral-genetic method (TrAC – Plummer, Chandler et al, 2020) to characterize the dendritic fields and axon collaterals of LC neurons that project to supraspinal motor centers including cerebellum, primary motor cortex, red nucleus, and ventrolateral (VL) thalamus.

Methods: We used male and female En1Dre; DbhFlpo; RC::RFLTG triple heterozygous mice that were engineered to express tdTomato in NE-containing LC neurons (Plummer, Chandler et al., 2020). A retrograde tracer, Canine Adenovirus Type 2 Cytomegalovirus encoding Cre recombinase (CAV2-CMV-Cre), was injected into target areas of the cerebellum. The tracer was taken up by the axons of the NE-containing LC cells, and transported retrogradely to the cells of origin, prompting them to switch from expression of tdTomato to green fluorescent protein (GFP). Eight weeks post viral infusion, the mice were transcardially perfused, and brains were extracted and sectioned in the coronal plane (40 µm from prefrontal cortex thru cerebellum) on a freezing microtome. Every third section of brain tissue from each animal was stained to accentuate tdTomato and GFP fluorescence. Injection sites were verified by staining tissue sections with Anti-Cre Recombinase antibody. Imaging was performed with fluorescent imaging and confocal microscopy.

Results: The mouse LC consists of a core of tightly clustered NE-containing soma surrounded by a dense dendritic field extending 60–800 µm into the peri-coerulear space. Dorsally projecting dendrites extend to the border of the 4th ventricle, suggesting direct contact with the cerebrospinal fluid (CSF). The rostral and caudal ends of the LC are tubular in shape whereas the primary cluster of LC-NE cells is lens shaped along the dorsal-ventral axis. To visualize LC neurons projecting to the cerebellum we injected CAV2-CMV-Cre into cerebellar terminal fields of our mouse model, causing LC projection neurons to flip their expression from tdtomato to GFP. This approach reveals a dense, bilateral distribution of LC cells that send axons to deep cerebellar nuclei (avg = 85–180 cells; 68% ipsi, 32% contra) and anterior or posterior cerebellar cortex (avg = 53–80 cells; 60% ipsi, 40% contra). Cerebellar-projecting LC neurons are concentrated rostro-caudally within the intermediate to ventral 2/3 of the nucleus, with only scattered cells identified in the dorsal LC. These results contrast with the more scattered, sparse, and predominantly ipsilateral (95%) distribution of the NE-containing cells that project from LC to 1° motor cortex. The dendrites of LC-cerebellum projection neurons extend in all directions from the LC core, but with a preference for medial and lateral sub-regions of the peri-coerulear space. Axon collaterals of LC-cerebellum projection neurons are observed in not only motor circuits (1o motor cortex, VL thalamus) but also in cognitive (prefrontal cortex, hippocampus) and sensory (dorsal lateral geniculate) regions of the forebrain.

Conclusions: These results indicate that: 1) the cerebellum receives bilateral projections from LC, 2) this output is driven by inputs to lateral and medial sub-regions of the peri-coerulear space as well as perhaps information transmitted via the CSF, 3) axon collaterals of LC-cerebellum projecting neurons are observed in motor, sensory and cognitive forebrain circuits. Thus, activation of LC-cerebellum projecting cells can promote release of NE in multi-modal forebrain regions via broadly distributed axon collaterals.

Keywords: Locus coeruleus, Norepinephrine, Cerebellum, Dendritic fields, Axon collaterals

Disclosure: Nothing to disclose.

P465. Sex-specific role of corticotropin-releasing factor signaling in the medial prefrontal cortex in fear generalization

Dahee Jung, Anastasia Anastasia, Xianru Yu, Morgan Chaljub, Luis Rosas-Vidal, Hao Li, Reesha Patel

Feinberg School of Medicine, Northwestern University, Chicago, Chicago, Illinois, United States

Background: Fear generalization, the inappropriate extension of fear to non-threatening cues, is a hallmark of post-traumatic stress disorder (PTSD) and contributes to maladaptive avoidance and impaired daily functioning. Corticotropin-releasing factor (CRF) signaling is dysregulated in PTSD, with increased CRF release from inhibitory neurons and signaling through its type-1 receptor (CRFR1) in excitatory neurons in the prefrontal cortex. Notably, mice show conserved cell-type expression patterns in the prelimbic cortex (PL), positioning the PL CRF-CRFR1 system as a potential driver of generalized fear. However, its functional contribution to fear behaviors, especially sex-specific roles, remains poorly understood.

Methods: We examined mPFC CRF-CRFR1 signaling in adult male and female C57BL6J, CRF-Cre, and CRFR1-Cre mice. We 1) measured mPFC CRF release using fiber photometry with a genetically encoded CRF biosensor (N = 9 males, 9 females), 2) tested the causal role of mPFC CRF-expressing (CRF+) neurons using optogenetic inhibition (N = 7–9 males/group, 6 females/group), and 3) monitored mPFC CRFR1-expressing (CRFR1+) neuronal activity using cellular resolution calcium imaging (n = 169 cells from N = 6 males, n = 157 cells from 7 females). Mice underwent fear conditioning (9 tone-shock pairing; CS + : 4 Hz tone, 0.5 mA footshock) and generalization testing (1–3 Hz non-threatening, novel tones) 24 hours later. Freezing behavior was scored.

Results: During auditory fear conditioning, PL CRF release increased in both sexes and tracked freezing behavior (males: r = 0.47, p < 0.05; females: r = 0.52, p < 0.05). However, during generalization, PL CRF release in response to the 3 Hz novel tone correlated with freezing only in males (r = 0.45, p < 0.01), suggesting a male-specific CRF signal associated with overgeneralization. Optogenetic inhibition of PL CRF+ neurons selectively reduced freezing to 2 and 3 Hz novel tones in males (two-way ANOVA, F (1,56) = 11.69, p < 0.01), with no effect in females, and did not impact acquisition or recall, indicating a targeted role for PL CRF+ neurons in generalization. At the downstream CRF receptor-expressing cells, PL CRFR1+ activity ramped to the CS+ during acquisition in both sexes. Yet only in males did CS+-evoked PL CRFR1+ activity correlate with freezing (r = 0.68, p < 0.01), identifying a sex-specific link between the CRFR1+ population responses and fear expression. During generalization, CS-excited PL CRFR1+ neurons tracked freezing to 2 and 3 Hz novel tones in males (2 Hz: r = 0.82, p < 0.01; 3 Hz: r = 0.95, p < 0.01), consistent with optogenetic findings, while no correlations were seen in females. Anatomical tracing showed that PL CRFR1+ neurons send projections to several downstream fear-processing brain regions including the basolateral amygdala (BLA). However, these BLA-projecting PL CRFR1+ neurons were not preferentially recruited during generalization, suggesting that distinct PL CRFR1+ output pathways mediate this behavior.

Conclusions: Together these findings uncover a male-specific prefrontal neuromodulatory CRF-CRFR1 mechanism selectively amplifying responses to ambiguous cues underlying fear generalization, a core feature of PTSD.

Keywords: fear generalization, Sex-specific effects, Corticotropin-releasing factor (CRF), Medial Prefrontal Cortex

Disclosure: Nothing to disclose.

P466. Paradoxical network-mediated excitation during optogenetic stimulation of inhibitory interneurons in the mouse dentate gyrus

Thomas Hainmueller, Adetta Elisabeth Heyhold, Nicholas Paleologos, György Buzsáki

Stanford, Palo Alto, California, United States

Background: Probing neuronal networks through genetically targeted optogenetic manipulations has become a staple of contemporary circuit neuroscience. These tools enable rapid, cell-type specific manipulations of neuronal networks and can thereby reveal insights into the roles of distinct neuronal populations on network activity and their potential implications in cognitive operations. Here, we used these tools to study the role of different subtypes of GABAergic interneurons in the mouse dentate gyrus (DG), a subsection of the hippocampus, and discovered unexpected dynamics within the DG network.

Methods: To manipulate the spiking activity of the two main DG interneuron classes, we injected parvalbumin (PV)-Cre (Pvalbtm1(cre)Arbr/J) and somatostatin (SST)-Cre (B6J.Cg-Ssttm2.1(cre)Zjh/MwarJ) mice with adenoviruses expressing the light-excitable cation channel channelrhodopsin-2 (ChR; n = 8 PV-Cre, n = 7 SST-Cre) or the chloride pump enhanced- halorhodopsin 3.0 (eNpHR; n = 2 PV-Cre, n = 2 SST-Cre) to optogenetically excite or inhibit PV or SST expressing interneurons, respectively. Mice were chronically implanted with silicon probes and optical fibers to manipulate interneuron activity while recording their activity and that of other surrounding DG neurons. Mice were recorded in their home-cages and during various spatial behavior tasks. At the end of each recording day, brief light pulses (10,000 pulses of 20 ms for eNpHR, 1,600 – 3,200 pulses of 10 ms for ChR; Methods) were applied to identify light-responsive neurons. Furthermore, 1-second light pulses were applied to study the impact of these light-responsive cells on the surrounding network activity.

Results: We used genetically targeted optogenetics to investigate the roles of PV expressing, soma-targeting, and SST expressing, dendrite targeting, interneurons in the DG of freely behaving mice. To study synaptic and network interactions of these interneurons amongst each other and with other DG cell types, we first developed a machine-learning based classifier to distinguish the different cell types based on electrophysiological criteria and validated it with optogenetic labeling data. Our classification distinguished GCs, MCs, PV, and SST interneuron classes with 90.04% accuracy, using 10-fold cross-validation, and can therefore be used to identify all four cell types simultaneously in the same recordings while studying the impact of different interneuron manipulations.

We first examined the impact of interneuron manipulations on the transmission of synchronized excitation from the entorhinal cortex, which manifest prominently as ‘dentate spikes’ (DS) in the DG local field potential. Optogenetic inhibition of SST interneurons with 1-second-long light pulses modestly but significantly increased the rate of DS, while PV interneuron suppression had no impact on DS rates. Conversely, we tested whether optogenetic activation of DG interneurons could reduce the rate of DS. Activation of SST cells dramatically reduced DS to 23.94 ± 15.74% of their baseline rates. In contrast, PV interneuron stimulation reduced DS rates to only 79.07 ± 27.53%, significantly weaker than the effect seen in SST-Cre mice (p = 1.24 × 10⁻¹⁰, Wilcoxon rank-sum test). We therefore conclude that SST interneurons exert powerful control over the transmission of entorhinal excitation during DS, and that low SST interneuron activity immediately preceding DS may be necessary to permit their passage.

We next examined local changes in the DG resulting from increasing or decreasing inhibition from the respective interneuron types with 1-second light pulses. Opsin-expressing interneurons were strongly modulated throughout light-pulses. However, firing of a wide variety of presumably non-opsin-expressing cells, which had not responded to brief light stimulation, was also modulated during prolonged perturbation of interneuron activity. Particularly, activation of PV interneurons via ChR stimulation inhibited most surrounding cells but, paradoxically, a subset of presumably non-opsin-expressing PV and SST interneurons was excited instead of inhibited. Notably, such excitatory responses during PV interneuron stimulation were also observed in a small fraction of excitatory GCs and MCs. Stimulation of SST interneurons suppressed firing in most surrounding GCs, MCs, and PV interneurons. However, MCs and PV cells also included a small group that was excited during SST interneuron stimulation. Further experimental analyses and computational modeling support the idea that such paradoxical excitatory effects are no artifact of the experimental setup and emerge through network-inherent dynamics.

Conclusions: Our results demonstrate that increasing inhibition in hippocampal circuits can result in complex dynamics that go far beyond reducing overall network excitation. In the DG, increasing inhibition may boost winner-take-all dynamics to support the selection of unique GC assemblies, which can dominate the network. Increasing firing rate inequality through excessive PV-mediated inhibition paradoxically enhanced the net spiking output of the DG network by promoting the unleashed activation of a small, “privileged” subset of GCs and MCs. Such non-linear network effects of interneuron output may be critical for their role in cognitive functions and need to be considered when interpreting studies using optogenetics as a tool to investigate the impact of interneurons on downstream brain areas or animal behavior.

Keywords: dentate gyrus, Hippocampus, Inhibitory interneuron signaling, Parvalbumin fast-spiking GABAergic interneurons, Somatostatin-expressing interneurons

Disclosure: Nothing to disclose.

P467. A history of high-fat diet alters reward processing and dopamine signaling in the nucleus accumbens

Claire Corbett, Leighelle Adrian, Samantha Bozarth, Brianna Linneman, Mark Niedringhaus, Elizabeth West

Rowan University, Stratford, New Jersey, United States

Background: With the increased prevalence of consumption diets high in fat in modern societies, it is critical to consider the potential adverse effects on brain function and deficits in motivated behavior. A chronic history of a “junk food” diet interferes with reward processing, and cue-triggered motivation is stronger in obesity-susceptible rats. Here, we examined how a history of a high-fat diet alters reward processing, associative learning, and NAc dopamine dynamics in behaving rats.

Methods: Male and female Long-Evans rats were divided into three groups (n = 15–16/group): a control group fed a standard chow diet, a group fed a high-fat diet (n = 16), and a group fed a low-fat/high-sugar diet (n = 15) for 12 weeks. At week 10, we measured rats’ preference for high-concentration sucrose (12 %) compared to low-concentration sucrose (4 %) for 7 daily sessions. Following the preference tests, rats underwent stereotaxic surgery to inject the genetically encoded GPCR-activation-based-DA (GRABDA) and optical ferrules into the NAc and switched to a standard chow diet. Rats underwent 15 days of Pavlovian conditioning in which one cue predicted a sucrose pellet (CS+) and a different cue did not (CS-) to evaluate associative learning. We measured NAc dopamine dynamics throughout conditioning using fiber photometry on Day 1 and Day 15 of Pavlovian conditioning.

Results: We found that the high-fat diet led to a significant decrease in preference for the high concentration of sucrose in a 2-bottle choice test (4% vs. 12%), while both the control diet animals p < 0.0001) and the high-sugar/low-fat rats (p = 0.0042) preferred the 12% over the 4% sucrose concentration. On Day 15 of Pavlovian conditioning, all groups of rats discriminated between CS+ vs CS- as shown by increased time spent [(main effects of diet (p < 0.0001) and cue type (p < 0.0001)], and increased entries into the food cup during the CS+ vs CS- [main effects of diet (p = 0.02) and cue type (p < 0.0001)]. The group with a history of high-fat diet spent less time in the food cup compared during both cues compared to the rats fed the standard chow diet, suggesting diminished motivated behavior, but intact learning after the diet history. On Day 15 of Pavlovian conditioning, rats show abolished NAc dopamine dynamics to a reward-predictive cue compared to rats with access to standard chow or a low-fat/high-sugar diet. In addition, the area under the curve (during the cue) negatively correlates with the percent (%) change in body weight from week 1 to week 8 with the percent (%) change in body weight from week 1 to week 8 on the diets (R2 = 0.30, p = 0.007), i.e., the more weight gained the greater the loss of dopamine dynamics during the cue presentation.

Conclusions: We found that all groups of rats showed the ability to discriminate the cues (measured as the percent of time spent and number of entries in the food cup to CS+ vs CS-), indicating preservation of learning mechanisms. However, in rats with a history of a high-fat diet, behavioral responses to the cues were diminished (i.e., less time spent in the food cup during the presentation of both CS+ and CS-) compared to control rats. Finally, rats exposed to the high-fat diet exhibited an abolition of NAc dopamine dynamics during reward-predictive cues (CS+) relative to control rats, but increased dopamine dynamics in response to the reward cue. Taken together, a high-fat diet alters reward processing, motivated behavior, and NAc dopamine signaling in rats with high-fat diet exposure.

Keywords: Nucleus Accumbens, high fat diet, Pavlovian conditioning

Disclosure: Nothing to disclose.

P468. Targeted ablation of striatonigral D1-MSNs shifts reward-seeking behavior in mice

Lee Peyton, Jeongoen Lee, Allison Sereno, Maria Traver, David Lovinger

National Institute on Alcohol Abuse and Alcoholism, Rockville, Maryland, United States

Background: The dorsal striatum, home to medium spiny projection neurons (MSNs), is a critically important input structure of the basal ganglia uniquely involved not only in motor control but also associative learning and decision making. The dorsal striatum is compartmentalized into the dorso-medial striatum (DMS), and the dorso-lateral-striatum (DLS). While both regions receive cortical input, the DLS receives major inputs from somatosensory and motor-cortical regions. Direct pathway projecting MSNs express the dopamine D1 receptor (D1-MSNs) and are known to project to the substantia nigra pars reticulata (SNr), the entopeduncular nucleus (EPN) and to the globus pallidus external segment (GPe). Classical experimentation has shown that the DMS controls with goal-directed actions while the DLS controls habit formation and compulsive actions. Using viral tracing strategies, it has been noticed that DMS D1-MSNs projections target the medial SNr neurons while the DLS D1-MSNs target the lateral portion of the SNr. The exact molecular nature of the neurons in the SNr receiving these topographic projections is not fully defined, however approximately 85% of lateral SNr neurons express the marker parvalbumin (PV). Misused substances including ethanol are known to alter synaptic plasticity in the DLS; however, it is still unknown whether the DLS D1-MSN dominated striatonigral circuit plays a role in habitual ethanol seeking. Additionally, it is currently unclear whether PV+ neurons of the SNr are differentially innervated along a medial-lateral striatal axis, and how this differential innervation may play a role in habitual or compulsive behaviors. Preliminary data indicate that DLS-lateral SNr synapses are strengthened by chronic ethanol exposure. We hypothesize that discrete subcircuit (DMS vs DLS) D1 MSNs project topographically to the medial vs lateral SNr where they differentially engage PV+ GABAergic neurons and this engagement facilitates habitual behavior.

Methods: We expressed the adeno-associated virus AAV2/5-Ef1a-DIO-Synaptophysin-GCaMP7b in the DMS and DLS of EY262 mice to measure presynaptic calcium activity of D1-MSN projections to the medial and lateral SNr using acute slice photometry. To test the causal role of these projections, we performed retrograde, projection-specific caspase ablation to selectively delete D1-MSNs innervating either the medial or lateral SNr and then trained mice on a random interval operant schedule. Finally, we used in vivo photometry to monitor topographically organized calcium activity in the SNr during operant stages of random interval training.

Results: In EY262 mice, Cre-dependent expression of the presynaptic calcium sensor synaptophysin-GCaMP7b in D1-MSN terminals produced comparable stimulation-evoked calcium transients in the medial and lateral SNr as measured with acute slice photometry. In contrast, expression of GCaMP8f in a separate cohort of EY262 mice revealed distinct stimulation-evoked calcium dynamics across medial and lateral SNr terminals. We next performed caspase-mediated ablation of D1-MSN projections from the dorsal striatum to the medial and lateral SNr to assess their role in operant learning and random interval (RI) performance. Ablation of DMS D1-MSN projections to mSNr reduced lever pressing and the number of reinforcements earned under both fixed ratio 1 (FR1) and RI schedules compared to controls, without impacting the sensitivity to outcome devaluation or habit index. In contrast, ablation of DLS D1-MSN projections to the lSNr did not alter lever pressing or reinforcements earned under FR1 or RI schedules but resulted in an increased habit index despite intact performance under valued and devalued conditions compared to control animals. In vivo fiber photometry of DLS D1-MSN presynaptic terminals in the lSNr revealed time-locked calcium transients to CS+ cues during Pavlovian conditioning.

Conclusions: Our findings demonstrate that selective ablation of D1-MSN projections to the SNr is sufficient to alter performance and reward valuation. These results provide a refined view of striatonigral circuitry and its roles in modulation reward-seeking behavior.

Keywords: Basal Ganglia, Direct pathway, Operant behavior, Photometry

Disclosure: Nothing to disclose.

P469. Social isolation engages opposing lateral habenula mechanisms to drive depression-like behavior in males and females

Meghan Flanigan, Jordan Hopkins, Sanghoon Choi, Grace Dodis

Medical University of South Carolina, Charleston, South Carolina, United States

Background: Experiences of social isolation promote depression, which can subsequently lead to voluntary social avoidance that worsens depression symptoms. A host of prior studies in rodents indicate that enhanced lateral habenula (LHb) burst firing drives depression-like behavior, and that ketamine reduces LHb bursting to exert its rapid antidepressant actions. Moreover, serotonin dampens LHb bursting to promote behavioral resilience to acute stress, indicating that SSRIs could exert their effects via this mechanism in the LHb as well. However, these studies were performed exclusively in males, leaving a critical gap in our understanding of sex differences in depression pathophysiology.

Methods: First, we assessed the effects of 30d of chronic social isolation (CSI) on depression-like behaviors, anxiety-like behaviors, and social behaviors in male and female mice. Following behavior, we collected LHb tissue and performed qPCR for serotonin receptors, Fos, and the RNA-editing enzymes ADAR1 and ADAR2 (n = 8/sex/group). In a second experiment, we assessed the effects of chronic social isolation on LHb tonic firing, evoked firing, burst firing, and responses to serotonin using ex-vivo patch-clamp electrophysiology (n = 4–8/sex/group). Finally, we determined the effects of LHb 5HT1a deletion on depression-like, anxiety-like, and social behaviors (n = 8–12/group/sex).

Results: We found that CSI reduced sucrose consumption, consistent with anhedonia. This was associated with increased 5HT1a (p = 0.08) and decreased Fos (p < 0.05) and ADAR2 (p < 0.05) expression in the LHb of males, and increased ADAR2 (p < 0.05) and decreased 5HT1a (p < 0.05), 5HT2a (p < 0.01), 5HT2c (p < 0.05), 5HT4 (p = 0.057), and Fos (p < 0.05) in the LHb of females immediately following social reunion with a novel conspecific. We further found that CSI induced opposing adaptations in LHb physiology in males and females. In males, CSI increased burst firing, consistent with previous reports demonstrating a causal role for LHb bursting in depression-like behavior in males. However, this was accompanied by a stark decrease in tonic spontaneous firing. CSI also shifted the response of serotonin on LHb neurons from depolarization to no effect in males, indicating increased engagement of inhibitory 5HT1a receptors. In females, CSI did the opposite–shifting the effects of serotonin from hyperpolarization to no effect, consistent with decreased engagement of inhibitory 5HT1a receptors. Furthermore, serotonin increased evoked excitability of LHb neurons in CSI, but not group housed, females. Remarkably, CSI did not alter bursting in females, but rather robustly increased tonic spontaneous firing. Finally, LHb 5HT1a deletion induced a depressive-like state in females, reducing sucrose preference as well as social preference. On the other hand, this manipulation had no effects on depression-like or social behaviors in males, but reduced anxiety-like behavior.

Conclusions: These findings indicate that the LHb undergoes sex-specific alterations in serotonin receptor expression and neuronal physiology following CSI to promote depression-like behavior. While CSI-induced increases in LHb 5HT1a expression may underlie depression-like behavior in males, CSI-induced decreases in LHb 5HT1a expression may underlie depression-like behavior in females. These data are highly informative for the development of sex-specific treatments for depression targeting the serotonin system.

Keywords: Depression, Lateral Habenula, social isolation stress

Disclosure: Nothing to disclose.

P470. Social context-dependent dopamine dynamics in the medial prefrontal cortex and nucleus accumbens core during social interactions in mice

Albert Li, Nayla Jimenez, Charles Ye, Meghan Cum, Sequioa Smith, Elizabeth Illescas-Huerta, Sophie Dehaut, Marisa Ponder, Nancy Padilla-Coreano

University of Florida, Gainesville, Florida, United States

Background: Dopamine (DA), a modulatory neurotransmitter, plays a critical role in motivated behaviors and reward responses. Given that certain social interactions are rewarding, DA is implicated in both social behavior and decision-making. However, little is known about how DA impacts social interactions in a circuit-dependent manner. Recent work demonstrates that DA dynamics in the nucleus accumbens core (NAcc) encode social novelty, motivation and valence during mating, parenting and aggressive encounters (Dai et al., 2022). Yet how NAcc DA correlates with social interactions in long-term partners with established social hierarchies remains unexplored. Furthermore, no studies have examined how DA dynamics in the medial prefrontal cortex (mPFC) contribute to social interactions. The mPFC is critical for social decision-making and processes relevant social information, including social rank. Little is known about the general role of mPFC DA, but emerging evidence indicates that it responds to both positive and negative valence stimuli and signals novelty (Ellwood et al., 2017; Melugin et al., 2024). To address these knowledge gaps, we measured real-time DA signaling in the mPFC versus NAcc during social interactions, while systematically controlling for familiarity, hierarchy, and competitive context.

Methods: Male C57 mice were injected with AAV-dLight1.3b (to visualize DA activity) and implanted with optical fibers targeting the NAcc (n = 10) or mPFC (n = 15). DA was recorded in vivo using fiber photometry. We examined social dominance using a trial-based social competition task in which a 10-second cue signaled an upcoming liquid reward. Recordings were done while subjects were performing the task alone and during competition with a cagemate. In addition, recordings were done during social interactions with different social agents: novel (conspecific stranger), short-term familiar (10-minute prior exposure), long-term (cagemate), and aggressor (male retired breeder CD-1). Behavior was analyzed by hand scoring along with pose tracking for automated distance-based interaction detection.

Results: Both mPFC and NAcc DA signaled social novelty when encountering strangers, with DA levels decreasing with repeated presentations of the novel mouse agent (NAcc: p = 0.0012; mPFC: p = 0.0141) and increasing upon presentation with a second novel agent (NAcc: p = 0.0012; mPFC: p = 0.0003). However, when comparing DA responses across social agents with distinct familiarity levels, we found no difference between responses to cagemates vs novel mice (NAcc: p = 0.8410; mPFC: p = 0.5736). This suggests that mPFC and NAcc DA does not only signal social novelty.

When interacting with cagemates during social competition for rewards, NAcc DA differed with competitive success. Comparing trials in which the subject received the reward (win) and did not receive the reward (loss), we observed greater NAcc DA response to the signal tone, prior to the reward delivery, during win trials compared to loss trials (AUC win vs. loss: p = 0.011). Next, we investigated if overall competitive levels impacted DA dynamics. We found that NAcc DA and mPFC DA, increases in trials when both mice compete for the reward relative to when only the winner attempts to win (NAcc, AUC: p = 0.043; and mPFC AUC: p = 0.035). Finally, during social defeat bouts with a larger CD-1 aggressor, NAcc and mPFC DA show diminished DA dynamics at the onset of defeat encounters.

Conclusions: Our findings suggest that DA dynamics in the NAcc and mPFC play distinct roles in regulating social behaviors in mice. NAcc DA responds to social valence, while mPFC DA is involved in multiple functions including novelty and reward modulation in social context. These findings enhance our understanding of the neural mechanisms underlying social interactions and inform future research on social behavior disorders.

Keywords: Dopamine, Medial Prefrontal Cortex (mPFC), Social Behavior

Disclosure: Nothing to disclose.

P471. POSTER WITHDRAWN

P472. Direct interoceptive input to the insular cortex mediates feeding behavior

Sarah Stern, Zhe Zhao, Binbin Xu, Skylar Anthony, Suganya Subramanian, Bryan Granger, Carolyn Von-Walter, Elisa Mizrachi, Matthew Kidd, Abhishikta Srigiriraju, Isaac McKie, Zhiying Li, McLean Bolton, Stefano Berto

Max Planck Institute, Jupiter, New York, United States

Background: The insular cortex integrates internal state (interoception) and external information, but the mechanism by which the insula receives interoceptive information is not known. We discovered a population of leptin receptor expressing neurons (LepR) in the insula that are not present in surrounding cortical areas. Based on leptin and LepR’s role in detecting body fat levels and hunger state in the hypothalamus, we hypothesized that LepR plays a similar role in the insula. Here we used optogenetics, miniscope calcium imaging and single-cell sequencing to test this hypothesis and investigate the role of leptin receptors in the insular cortex.

Methods: Mice that were infused with leptin into the insula underwent body weight and food intake assessment, as well as single-nuclei sequencing of the insula. LepR+ cells were optogenetically activated during operant nose-poke feeding and drinking, calcium activity was imaged in LepR+ cells during feeding and drinking under both food and water restriction. LepR cell identity was assessed via FISH and ribosome profiling. Ouputs from LepR+ neurons were assessed using viral terminal tracing and output function tested with optogenetic activation at terminals.

T-tests or ANOVA’s were used. Animal numbers: Behavior, n = 6–20; calcium imaging, 4–6; RNA sequencing, 4–6; Criterion for statistical significance: p < 0.05

Results: Infusion of leptin into the insula significantly decreased body weight and food intake, and altered gene expression in layer 2/3 neurons. Optogenetic activation of LepR neurons decreased chow pellet retrieval but had no effect on water intake. Calcium imaging revealed that LepR neurons encode food intake bouts, and hierarchical clustering revealed that activity is dependent on hunger state. Similarly in a food/water choice task, LepR neurons were responsive to food over water, particularly in the food deprived state. We found that LepR was expressed on both glutamatergic neurons and in vascular cells, suggesting active transport through the blood brain barrier. Finally, we found that LepR+ cells make mostly local connections within the insula as well as to the basolateral amygdala, and activation of those terminals also decreased operant feeding.

Conclusions: Overall, our results point to a mechanism for direct interoceptive input to the insular cortex in order to detect hunger states and control complex feeding behavior.

Keywords: insular cortex, food intake, Leptin

Disclosure: Nothing to disclose.

P473. Serotonin modulation of social behavior across striatal subregions

Allen Chen, Mona Li, Michaela Guo, Matthew Pomrenze, Jun Song, Liqun Luo, Robert Malenka

Stanford University, Stanford, California, United States

Background: Serotonin has been heavily implicated in neuropsychiatric disorders, thought to play major roles in the regulation of emotional processing, perception, and social behaviors. However, the neural circuit mechanisms through which serotonin orchestrates these behaviors remains obscure. Serotonin is thought to facilitate prosocial behaviors through brain regions such as the nucleus accumbens and frontal cortical regions. However, other pharmacological and clinical evidence supports a role for serotonin in anxiety and social aversion. It is therefore unclear how serotonin may facilitate complex social dynamics, and what its role is across striatal subregions. Here, we studied the role of serotonin across mouse striatal subregions including the nucleus accumbens, dorsal striatum, and tail striatum.

Methods: We used viral tracing, genetic depletion of serotonin, fiber photometry, pharmacology, and analysis of in situ hybridization data to study the role of serotonin across striatal subregions. Wildtype, SERT-Flp + /−, and Tph2flox/flox mice bred on a C57BL6 background, of both sexes were used. Cholera toxin studies were used to assess neuronal populations projecting to the nucleus accumbens, dorsolateral striatum, and tail striatum (n = 5–8 mice). Genetic disruption of serotonin synthesis via conditional knockout of tryptophan hydroxylase 2 (Tph2) was conducted by injecting a retrograde AAV Cre recombinase construct in respective striatal subregions in Tph2flox/flox mice (n = 18 mice). Fiber photometry was used to monitor genetically-encoded serotonin and calcium sensors (n = 20 mice). Microinfusion of serotonin agonists and antagonists were applied to ascertain receptor function (n = 14 mice). We employed these techniques in mice performing a variety of behavioral tasks, that implicate serotonergic modulation, such as the 3-chamber social preference testing, novel juvenile interaction, novel object recognition, and 3-chamber food pellet testing. Data were analyzed using a combination of parametric and non-parametric statistical tests: paired and unpaired two-tailed t-tests, and Wilcoxon rank-sum tests, used when appropriate.

Results: We employed cholera toxin tracing of three different wavelengths in three separate striatal regions to identify striatal-projecting serotonergic (Tph2-positive) neurons. The nucleus accumbens, dorsolateral striatum, and tail striatum appear to receive mostly non-overlapping inputs from both the dorsal raphe and medial raphe nuclei. Genetic depletion of serotonin neurotransmission at the nucleus accumbens downregulated mouse social interaction whereas depletion of serotonin neurotransmission at the tail striatum had the opposite impact of upregulating mouse social interaction. Rescuing these serotonin depletion deficits, microinfusion of serotonin 1B and serotonin 2A agonists reversed these sociability perturbations, in the nucleus accumbens and tail striatum, respectively. Using fiber photometry, we find that serotonin dynamics in the nucleus accumbens differed in responses to various stimuli, as well as differing in baseline dynamics. Serotonin dynamics in the nucleus accumbens tend to fluctuate positively with social interactions whereas dynamics in the tail striatum did not have a similar effect but correlated with novelty interactions.

Conclusions: Serotonin has been shown to have conflicting roles for learning and social behavior. Our work provides evidence for opposing roles for striatal serotonin in the regulation of social behavior. We characterized the serotoninergic inputs across three striatal regions, demonstrating distinct circuits across subregions. Serotonin in the nucleus accumbens plays a prosocial role whereas in the tail striatum it had the opposite effect. This is suggestive of neural circuit mechanisms through which serotonin may cause imbalances in social dynamics.

Keywords: Serotonin, Tail of the Striatum, Nucleus Accumbens, Social Behavior, neuromodulators

Disclosure: Nothing to disclose.

P474. Serotonin modulates the dorsomedial striatum in a behavioral state-specific manner

Ka Ng, Arati Sharma, Kate Nautiyal

Dartmouth College, Hanover, New Hampshire, United States

Background: The dorsomedial striatum (DMS) is critical for motivating and inhibiting behavioral responses, both of which are key facets of behavior that are dysregulated in a number of psychiatric disorders. The DMS circuitry is complex as it integrates multiple inputs from the cortex, thalamus, and other subcortical structures including midbrain dopamine neurons. Though less studied, serotonin neurons from the dorsal raphe nucleus also richly innervate the DMS, which expresses the majority of the 14 receptor subtypes. In particular, slice electrophysiology shows that the serotonin 1B receptor (5-HT1BR) impacts DMS physiology and plasticity. Also, behavioral experiments show that 5-HT1BR expression modulates impulsivity and other DMS-dependent reward-related behaviors. In these studies, our goal was to investigate the effects of 5-HT1BR on the DMS activity in vivo during reward-related behaviors in mice.

Methods: We examined the calcium activity of individual medium spiny neurons (MSNs) using miniature microendoscopes to image the GCaMP6f calcium indicator expressed virally in the DMS under the CaMKII promotor in male and female 5-HT1BR KO mice (N = 9) and littermate controls (N = 6). Mice were trained in two behavioral tasks using a homecage Arduino operant system (DIY-NAMIC) in which all trials were self-initiated. The first was a simple cue-guided fixed ratio task, in which mice responded with a nosepoke in one of two ports indicated by a response cue. The second task was a 2-choice serial reaction time task used to measure impulsive behavior, in which a delay period was introduced between trial initiation and the response cue. Calcium activity was recorded in 537 and 842 cells in control and 5-HT1BRKO mice, respectively. Activity for each cell was averaged over all rewarded trials within a session and aligned to the operant nose poke response.

Results: We found that knockout of 5-HT1BRs resulted in a significant reduction of neurons showing excitatory calcium responses (χ2(1, N = 930) > 6.07, p < 0.013) following receipt of a reward. Additionally there was also an increase in the proportion of cells with inhibitory calcium responses following reward receipt (χ2(1, N = 930) > 5.62, p < 0.018). This suggests that serotonin may recruit MSN activity in response to reward via actions at 5-HT1BRs. On the other hand, in a behavioral paradigm designed to test impulsivity, we found that serotonin inhibits DMS calcium activity through 5-HT1BRs. Specifically, mice lacking 5-HT1BR expression had a larger proportion of cells showing increased calcium responses during the waiting period of the trial, compared to controls (t13 = 4.26, p = 0.0009). Furthermore, of the cells that showed excitation during the delay period, those from 5-HT1BR KO mice showed higher calcium event rates compared to controls (F1,126 = 11.58, p = 0.0009).

Conclusions: In summary, the data demonstrate that serotonin can modulate the DMS in a behavioral state-specific manner, potentially having opposite effects on neural signaling dependent on the ongoing behavior. This may provide a mechanism for how serotonin effects on behavior are context-dependent. Overall, these results point to the importance of in vivo studies to understand the functional role of DMS serotonin in reward-related behavior.

Keywords: Serotonin 1b receptor, impulsivity, Dorsomedial striatum, Reward, goal-directed behaviors

Disclosure: Nothing to disclose.

P475. Interneuron-driven oscillatory states engage valence-specific basolateral amygdala ensembles and downstream circuits

Kenneth Amaya, Yingchu He, Grant Weiss, Pantelis Antonoudiou, Jamie Maguire

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

Background: The binary assignment of positivity or negativity to cues and outcomes, known as valence processing, is crucial for behavioral approach or avoidance, and thus survival. A critical brain region for valence processing is the basolateral amygdala (BLA), as both appetitive and aversive learning are dependent on BLA function. Recent advances have highlighted the role of interneurons in mediating BLA oscillatory states to alter behaviors like fear expression and reward seeking. Despite their documented ability to govern BLA states and behaviors, how interneurons may be involved in recruiting valence-relevant ensembles to promote information routing remains unknown.

Methods: To address this, we conducted a set of 6 experiments to interrogate how BLA interneurons facilitate valence processing. We first quantified overlap between cFos expression (post fear conditioning or after extinction) and distinct BLA projection populations (BLA to NAc or BLA to BNST, total N = 33) using retroAAV-mCherry in male and female C57BL6/J mice (Exp1). Next, using the same wild-type mice (N = 9), we similarly labeled BLA projection neurons while introducing a DLX-promoted channelrhodopsin to optogenetically stimulate interneurons at differing frequencies (4 Hz or 8 Hz) to assess how rhythmic interneuron entrainment may selectively engage projection ensembles (Exp2). Then, in FosTRAP x Ai14 mice (N = 10), we targeted BLA interneurons for optogenetic stimulation (AAV-DLX-ChR2) at differing frequencies (4 Hz or 8 Hz) to assess how rhythmic interneuron stimulation may activate fear or extinction-tagged ensembles (Exp3). We then used the same transgenic mouse line (N = 12) to tag fear- or extinction-recruited BLA neurons (n = 25) for ex-vivo electrophysiological characterization (Exp4) before imaging ensemble activity using in-vivo single-photon calcium imaging to record ensemble activity (n = 275) during PV-interneuron stimulation (AAV-FLEX-ChRimson) in PV-Cre mice (N = 6) (Exp 5). Finally, we used whole brain clearing and light-sheet microscopy to quantify cFos expression brain-wide following interneuron stimulations at either 4 of 8 Hz, to assess how distinct BLA interneuron stimulations may be activating unique networks (Exp6).

Results: First, we observed differential engagement of BLA projection populations with respect to their activation during fear or extinction learning: BLA to BNST projections were more active during fear learning (W = 72, p < 0.001) while the BLA to NAc projections were more active during extinction (W = 0, p < 0.001). Next, we observed that interneuron stimulations recruit projection ensembles in a frequency-congruent manner. That is, BLA to NAc labeled neurons were activated by 8 Hz interneuron stimulation and BLA to BNST labeled neurons were activated by 4 Hz interneuron stimulation. Here, these stimulation and label conditions were Congruent, and displayed greater cFos overlap than their Non-congruent counterparts (NAc + 4 Hz and BNST + 8 Hz) (W = 102.5, p = 0.002). Interneuron stimulation at 4 or 8 Hz preferentially activated fear or extinction-tagged ensembles, respectively. Congruence was defined as Fear + 4 Hz and Extinction + 8 Hz overlap, while Non-congruence was Fear + 8 Hz and Extinction + 4 Hz overlap. When comparing cFos overlap in the Congruent condition to the Non-congruent condition, there were promising trends in our preliminary cohort in the BLA (W = 3, p = 0.067) and NAc (W = 4, p = 0.11), and a significant difference observed in the BNST (W = 2, p = 0.038). In our ex-vivo data, we sorted recorded neurons into either Fear or Extinction-tagged groups. There, we saw no differences between the groups in intrinsic properties (resting membrane potential, rheobase, peak to trough) and the waveforms of the neurons appeared quite similar. One treatment that we applied involved current clamping patched cells and applying an input stimulus that increased in oscillatory frequency (from 1 Hz to 60 Hz) over 60 seconds of recording. We saw profound differences between the tagged populations when quantifying spike probability, as Fear-tagged cells preferentially fired in the low-theta band while Extinctiopostn-tagged cells fired in high-theta. Similarly, our in-vivo imaging revealed distinct ensembles that are tuned to differing PV interneuron stimulation frequencies. Finally, our whole-brain cFos data show that networks are engaged by these two stimulation frequencies, especially in the accumbens and bed stria of the nucleus terminalis.

Conclusions: In sum, BLA ensembles are naturally recruited by experience in a pathway specific manner, but we can leverage interneuron activity to artificially engage these valence-tuned ensembles. Further, our physiological data suggest that BLA neurons display preferential responsiveness to inputs in a frequency-specific manner, even prior to fear conditioning. Thus, ensemble recruitment appears to depend on the projection target of a principal neuron but also some resonant frequency sensitivity. We argue that valence processing is at least aided by interneuron regulation of oscillatory states as our manipulations demonstrate sufficiency in this domain. Together, our results provide a framework to understand BLA function as a product of microcircuit governance over ensemble engagement, oscillatory state setting, and network activation, with interneurons sitting in a privileged position as conductors of valence processing.

Keywords: Amygdala, Aversive Learning, Neuronal ensembles

Disclosure: Nothing to disclose.

P476. The nucleus accumbens encodes model-based decision-making variables more strongly than the dorsomedial striatum

Matthew Geramita, Chamberlain Brittany, Mark Nicholas, Eric Yttri, Susanne Ahmari

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

Background: Humans and rodents both use a decision-making system known as model-based control (MBC) to flexibly adapt behavior in the face of new information. MBC relies on using a cognitive map of how actions, sensory cues, and outcomes relate to one another to plan the best decision. Impairments in MBC are a hallmark of compulsive disorders such as obsessive-compulsive disorder, addiction, and eating disorders. Additionally, the same striatal networks that display disrupted activity during compulsive behaviors support MBC. Therefore, determining how striatal circuits contribute to MBC will facilitate the discovery of novel treatment targets for compulsive disorders.

In this project, we delineate the neural mechanisms of MBC by monitoring striatal activity in mice using an established assay of human MBC called the two-step task. We are among the first groups to translate this task into mice. In it, mice choose between two actions that each lead to a separate, probabilistically rewarded sensory stimulus (i.e., left choice ♢ tone A ♢ reward; right choice ♢ tone B ♢ no reward). The key element of this task is that, occasionally, an action will lead to the sensory stimulus not typically associated with it through a rare transition (i.e., left choice ♢ tone B). On the next trial, mice are forced to use MBC to select the choice most likely to lead to the more highly rewarded sensory state. Here, we will elucidate how MBC is represented in two areas of the striatum implicated in both MBC and compulsive behaviors – the dorsomedial striatum (DMS) and the nucleus accumbens (NAc).

Methods: Here we present a novel head-fixed mouse version of the two-step task which facilitates dissection of the neural circuits underlying MBC. In our version of the two-step task, head-fixed mice are positioned in front of an LED screen while their front paws rest on a steering wheel. During each trial, the mouse has 10s to signal a left or right choice by moving the wheel. Each choice then leads to a 2s compound sensory stimulus occurs. In 80% of trials, left choices lead to stimulus A (12 kHz tone and vertical bars) and right choices lead to stimulus B (3 kHz and horizontal bars). During rare transitions, which occur in 20% of trials, left choices lead to stimulus B and right choices lead to stimulus A. Each sensory stimulus is subsequently probabilistically rewarded. Reward contingencies vary in a block-wise manner between 10% and 90% Block transitions are triggered 8–12 trials after an animal crosses a threshold of 70% correct choices. Behavioral data are presented from 22 mice (10 female). In a subset of animals (n = 6, 13 sessions), we recorded DMS and NAc activity using Neuropixels probes. Single units were identified using Kilosort and Phy software.

Results: Mice tracked which choice was currently best, completing 7.41 ± 2.98 blocks per session. To assess the extent to which mice used MBC in their decision-making, we explored the factors that predicted whether the animal would repeat their choice on the next trial. Decisions to repeat prior choices most strongly depended on the interaction between the prior trial’s transition and reward (mixed-effects logistic regression – transition x outcome: β = 0.78  ± 0.07, p < 0.001), suggesting that mice used MBC to guide their decisions. Additionally, we fit choice data to several reinforcement learning models that included model-based and model-free algorithms and found that behavior in all animals was best explained by model-based algorithms.

In a subset of mice, we recorded activity from 830 DMS neurons and 1254 NAc neurons while animals performed the two-step task to determine whether the DMS and NAc play distinct roles in MBC. NAc neurons represented both the current trial’s transition (p < 0.001; repeated measures ANOVA) and outcome (p < 0.001; repeated measures ANOVA) more strongly than DMS neurons. However, neither transition or reward-related DMS or NAc activity persisted into the next trial, suggesting that choice behavior may be driven by latent variables such as choice value that incorporate both a trial’s transition and outcome. Therefore, we next used reinforcement learning models of MBC to extract trial-to-trial estimates of choice value. Interestingly, NAc represents MBC-based choice value more strongly than DMS, and these value signals persisted into the next trial suggesting that NAc representations of value may play an important role in guiding future choice.

Conclusions: Compulsive behaviors are associated with disrupted activity in the same striatal circuits that are hypothesized to be important for MBC, yet the specific roles of striatal circuitry in MBC are unknown due to a paucity of in vivo studies of neural activity. Here we present a novel version of the two-step task and show that mice engage in MBC. Additionally, we show that the NAc represents elements of MBC more strongly than the DMS, suggesting that it plays a critical role in MBC. Future work will use optogenetics to determine the necessity of the NAc and DMS in MBC.

Keywords: Reinforcement learning, Striatum, In vivo electrophysiology

Disclosure: Nothing to disclose.

P477. Decoding ventral tegmental area neural activity from EEG using machine learning

Laura Benoit, Jaehan Kwon, Alexander Harris

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

Background: Current circuit neuroscience research heavily relies on invasive techniques in animal models, such as in vivo electrophysiological recordings of local field potentials or fiber photometry, to investigate neural dynamics across brain regions. While these methods have provided invaluable insights into the mechanisms underlying in many behaviors and neuropsychiatric disorders, their invasiveness limits translational applicability to human studies, which require non-invasive approaches.

Functional magnetic resonance imaging (fMRI) has been used in many studies to explore specific brain regions involved in various processes but lacks the resolution to differential neural activity by cell type or activity pattern. Similarly, electroencephalography (EEG) is non-invasive and widely available for human studies. While it has been used to indirectly explore the impact of behaviors or pharmacological interventions, circuit-specific questions have to date been out of reach.

This study aims to develop a machine learning-based classifier capable of decoding EEG signals to distinguish between distinct cell types and activity patterns within the ventral tegmental area (VTA), a key region for motivation and reward procession that has been implicated in numerous psychiatric and neurological disorders.

Methods: In these experiments, we recorded from male DAT-IRES-Cre and VGAT-Cre mice (n = 8 and n = 6, respectively). All animals were injected with a calcium-sensitive fluorescent reporter (GCaMP8f) targeted to the ventral tegmental area (VTA). Mice underwent optogenetic stimulation of either dopamine (DA) neurons at 4 Hz, 20 Hz, or with an aperiodic stimulation pattern or GABAergic neurons at 4 Hz. Concurrent recordings included local field potentials (LFPs) from the VTA and nucleus accumbens (NAc), and electroencephalography (EEG) from three skull screw electrodes placed over the olfactory bulb, cerebellum, and midline. Power spectral density analyses were performed for all animals across stimulation conditions: (1) no stimulation, (2) 4 Hz periodic, (3) 20 Hz periodic, and (4) aperiodic stimulation. We then trained a machine learning classifier (model: bagged trees) on a subset of the EEG spectral data to assess its ability to decode stimulation type and neuronal cell class from EEG recordings alone and tested another subset on the trained model. Both training and testing were conducted with cross-validation.

Results: The classifier successfully distinguished between GABAergic and dopaminergic neuronal stimulation at 4 Hz and 20 Hz using EEG recordings at a value greater than 75%. It also accurately classified distinct dopaminergic stimulation patterns (no stimulation, 4 Hz, 20 Hz, aperiodic). Classifier performance reached 65.15% +/- 9.87%, which was significantly above chance levels (25%, p = 0.0039) and exceeded performance on shuffled control datasets (26.07% +/- 2.03%, p = 0.0002).

Conclusions: These findings demonstrate that EEG signals, when paired with appropriate machine learning approaches, can reliably distinguish both cell type-specific and firing pattern-specific activity within the VTA. This represents a significant step toward bridging the translational gap between invasive rodent recordings and non-invasive human neuroimaging. The approach offers the potential for circuit- and cell type-specific insights using EEG, with wide-reaching implications for understanding the neurophysiology of motivation, reward, and psychiatric disorders.

Keywords: Neurocircuits, Ventral Tegmental Area, EEG, Machine Learning, Neural decoding

Disclosure: Nothing to disclose.

P478. A novel, translational virtual height threat task for rodents reveals the locus coeruleus—prefrontal cortex circuit is critical for threat-based decision-making

Stephanie Staszko, Emi Krishnamurthy, Rohan Lokanadham, Aakash Basu, Alfred Kaye

Yale University, New Haven, Connecticut, United States

Background: The ability to detect and respond to threats is critical for survival, but dysfunction of threat circuits may result in anxiety or trauma-related disorders. Our ability to understand and treat these disorders has traditionally been limited by a lack of highly translatable pre-clinical models due to a lack of threat stimuli which are transferrable between rodents and humans. The visual cliff task is a task which gives the illusion of a drop off, or cliff, in the testing arena. It has been used to study the innate behavioral preference of human infants to avoid high places as well as to study visual depth perception in animals. Recent work evaluating the response of humans to visual threats in virtual reality found that height threat was the only visual stimulus to reliably induce peripheral markers of arousal. Despite this history and recent advances in human neurophysiology of heights, threat from physical height has not been adapted to test threat circuits using systems neuroscience approaches. Thus, we developed a virtual reality height threat task, based off a pole descent task, in which height threat can be manipulated to investigate neural circuits underlying threat-based decision-making. The medial prefrontal cortex (mPFC) is known to be activated by spatial proximity to threat and hypothesized to coordinate escape behaviors, while the locus coeruleus (LC) regulates arousal states. Thus, we hypothesized a LC—mPFC circuit may regulate decision-making in response to visual height threats.

Methods: All experiments were conducted in both male and female mice (C57BL6/J, DBH-Cre, or DBH-ChR2 mice, mean age 4 months at time of behavior). For optogenetic experiments, animals were injected with opsin bilaterally (no virus for DBH-ChR2 mice) and simultaneously implanted with 200 uM ferrules for optogenetic stimulation (LC inhibition AAV1.hSyn1.SIO.stGtACR2.FusionRed, DBH-Cre mice; PFC inhibition AAV1.CKIIa.stGtACR2.FusionRed, C57 mice). Optogenetic inhibition was performed using 5mW of constant stimulation at 450 nM; excitation was performed using 10 mW, 10 Hz, 10 ms pulses. For calcium imaging experiments, mice were injected with 500 nL of the calcium indicator AAV1.Syn.GCaMP8f in mPFC. Two weeks later, a GRIN lens (4.0 × 0.5 mm) was implanted above the mPFC. Animals were fitted with a baseplate to allow for miniaturized microscope attachment.

Prior to experiments, animals received two, 10-minute handling sessions and 10 minutes of acclimation to attachment of the miniscope or fiber optic patch cord in a clean cage. For behavioral experiments, animals were placed at the top of a 20” pole in an enclosed metal box. Visual stimuli were presented on a monitor beneath the mice and included checkerboard visual stimuli where one quadrant was visually proximal to the mouse, and the other three quadrants were projected as close (5cm) or far (20cm) from the proximal quadrant. The pole ends in a cone at which point mice must decide which quadrant to exit onto a Plexiglas surface. Animals receive 48 randomized trials where the target quadrant is rotated as well as the stimulus depth. “Safe” responses are those in which the mouse exits to the proximal, target quadrant.

Results: When the difference between the target, “safe” quadrant and remaining quadrants is small, mice choose the safe quadrant at chance levels, while mice choose the safe quadrant more often when the visual discrepancy is large (p < 0.001, paired t-test). To test the necessity for mPFC in threat-based decision making, we optogenetically inhibited mPFC in half of the trials and found that compared to no-light trials, inhibition of mPFC reduces preference for the safe quadrant in the far, more threatening condition (p = 0.004, RM One-Way ANOVA; p = 0.03 Tukey’s multiple comparisons for far condition light vs no light). Similarly, LC inhibition decreases the safe choice in the far, more threatening condition (p = 0.01 for stimulation, p = 0.03 for stimulation x height, Two-Way ANOVA). Finally, to test the role of the LC—mPFC circuit, we stimulated noradrenergic LC projections in the PFC and found this increased preference for the safe quadrant across all conditions, eliminating the difference between close and far conditions (Three-Way ANOVA, height p = 0.02, genotype x stimulation p = 0.01), suggesting this projection regulates threat-based decision making in this task. We next recorded calcium activity of single neurons while mice performed this task using Inscopix miniscopes. Current analyses demonstrate about 15% of the mPFC population respond to various aspects of the task (stimulus type, outcome), with different neurons responding to different task characteristics. Analysis of this data is focused on identifying neurons based on response subtypes and decoding behavioral choice from neural data.

Conclusions: We developed a novel, virtual height threat task in which neural circuits underlying threat-based decision making can be investigated. We demonstrate the necessity of LC and mPFC individually, as well as a LC-mPFC circuit, for optimally performing threat-avoidance behavior in this task. Current analyses are focused on determining the coding properties of individual neurons in mPFC neurons and which ensembles might underly decision making. Importantly, this task is ethologically relevant and has high potential for translation to humans in virtual reality, providing the opportunity to more carefully characterize the circuits underlying threat processing in health and disease.

Keywords: norepinephrine, systems neuroscience, Visual threat detection, threat processing

Disclosure: Nothing to disclose.

P479. Dorsolateral striatal acetylcholine signals threat and primes ensemble activity for upcoming aversive events

Munir Kutlu, Oyku Dinckol, Noah Wenger, Charlie Maddox, Taylor Good, Jake Caselli, Zehra Bozdag

Temple University School of Medicine, Philadelphia, Pennsylvania, United States

Background: Acetylcholine (ACh) in the central nervous system is critical for attention, learning, and adaptive responses to salient stimuli. While previous research implicates ACh in negative valence processing, its role in encoding threat—both social and non-social—and in shaping local striatal network activity remains unclear. The dorsolateral striatum (DLS) receives dense ACh input from local interneurons and brainstem projections and is central to action selection. We hypothesized that DLS ACh release encodes valenced information and proactively organizes single-cell ensemble dynamics for impending aversive events.

Methods: We used an integrative approach combining fiber photometry with a genetically encoded ACh sensor (ACh.Sn.Fr) and single-cell calcium imaging to simultaneously measure DLS ACh release and neuronal ensemble activity in behaving C57BL/6J mice. Threat contexts included both social (e.g., tube dominance, competitive warm spot, novel same-sex conspecific interactions) and non-social (fear conditioning, extinction, appetitive conditioning, and punishment) paradigms. To establish causality, we employed optogenetic stimulation of DLS cholinergic interneurons during behavioral tasks and concurrent single-cell imaging, enabling direct assessment of how elevated ACh release modulates ensemble reorganization and threat-related behaviors.

Results: Using in vivo fiber photometry in male and female C57BL/6J mice, we found that DLS ACh release increased in response to aversive stimuli and the conditioned cue during fear conditioning, with this response progressively diminishing as conditioned fear extinguished. In contrast, during appetitive conditioning, DLS ACh responses to a reward-predictive cue decreased over time and remained stable with learning, indicating a valence-dependent effect of ACh signaling. Strikingly, this reward-related suppression was reversed when reward-seeking was punished with footshock. ACh dynamics were absent or markedly reduced in non-threatening contexts, even when locomotor activity and sensory cues were matched. In social settings, DLS ACh also acted as a perceived threat and competition signal, increasing during encounters with novel same-sex conspecifics and during competitive interactions. Importantly, we showed that the increasing ACh release via optogenetic stimulation of DLS cholinergic interneurons (CINs) enhanced fear conditioning and impaired fear extinction. Furthermore, simultaneous optogenetic stimulation of CINs and single-cell calcium imaging revealed that elevated anticipatory ACh release preceded the reorganization of DLS ensemble activity, biasing population states toward threat-related configurations and attenuating responses to the predicted aversive outcomes.

Conclusions: Our findings reveal that DLS ACh functions as a general threat signal across both social and non-social contexts, encoding perceived threat and competition while priming local neuronal ensembles for impending aversive events. This anticipatory cholinergic modulation likely serves as a key mechanism linking environmental threat detection to rapid and adaptive action selection. Disruption of this process may contribute to maladaptive threat processing seen in disorders such as PTSD, anxiety disorders, and social phobias.

Keywords: acetylcholine, Threat neurocircuitry, dorsolateral striatum

Disclosure: Nothing to disclose.

P480. Active coping elicits distinct corticostriatal response profiles in male and female rats

Connor McNulty, Gianni Bonnici, Lia Farrell, Steven Maier, David Root, Robert Rozeske, Michael Baratta

The University of Colorado Boulder, Boulder, Colorado, United States

Background: Coping strategies are associated with human resilience and are a frequent target of therapeutic intervention. Learning how one’s actions (R) influence outcomes (O) is central to this process; however the neural basis supporting this form of learning has largely been studied in appetitive contexts. Here we examine corticostriatal responses in male and female rats during an aversively motivated instrumental learning task in which subjects can control (R) the termination of each stressor exposure (O). Though males and females perform the instrumental controlling response with the same efficacy, recent work suggests that the stress-buffering effects of control are only present in males, not females.

Methods: To manipulate controllability, male and female Sprague-Dawley rats were run in a triadic design. One subject of each triad received escapable shock (ES) in which performing an instrumental response (R) terminates shock (O) once the fixed-ratio requirement was met. A second subject received yoked-inescapable shock (IS), and a third subject received no shock (HC). Thus, ES and IS subjects were exposed to physically identical stressors. EXP1: Trial-by-trial prelimbic (PL) cortex pan-neuronal activity was assessed with synapsin-driven expression of GCaMP8m and a recording fiber implanted into the PL. Fluorescent traces were averaged across the first 10 (acquisition) and last 10 (end-of-session) trials (males/females, n = 5). EXP2: We subsequently examined responses in PL projections to the dorsomedial striatum (DMS) by injecting retrograde GCaMP8m into the DMS and placing a recording fiber in the PL (males/females, n = 4). EXP3: Similarly, we monitored dopamine (DA) binding to the corticostriatal pathway by delivering a retrograde GrabDA1h biosensor in the DMS and fiber in the PL (males/females, n = 5–6). EXP4: The contribution of mesocortical DA to the behavioral outcome of controllability was assessed using an intersectional chemogenetic approach. Retrograde cre recombinase was injected into the PL and a cre-dependent hM4Di or mCherry control virus was injected into the ventral tegmental area (VTA; females, n = 8–11). CNO was administered prior to the controllability procedure. Animals were tested for juvenile social exploration levels the following day when CNO was no longer active.

Results: EXP1: Males and females displayed similar positive PL GCaMP responses to shock onset during both acquisition and maintenance of the controlling response. EXP2: Males produced significantly greater AUC (p = 0.0059) and peak dF (p = 0.0306) in corticostriatal calcium fluorescence than females during ES acquisition trials (two-way ANOVA). The differential GCaMP response was only present during epochs in which instrumental responses terminated shock (R-O), but not during epochs in which performance of R was independent of O. No differences were observed between male and female IS groups. EXP3: In contrast, females exhibited a robust dopamine response in the corticostriatal pathway that was absent in males during acquisition trials (AUC, p = 0.0105; peak dF, p = 0.0027; two-way ANOVA). EXP4: Females expressing mCherry in mesocortical neurons showed social avoidance across stress groups (ES, p = 0.0040; IS, p = 0.0100). Chemogenetic silencing of the VTA→PL pathway prevented this social avoidance phenotype in the ES condition but not in IS (p = 0.0200, two-way ANOVA).

Conclusions: Using an animal model of coping, our findings suggest sex differences in the circuitry engaged during aversively motivated instrumental learning. Notably, these differences emerged (1) during R-O associative events and (2) only when corticostriatal, but not pan-neuronal PL activity, was examined. Our data also suggest that mesocortical activity may bias females away from engaging the goal-directed corticostriatal system, thereby preventing the typical stress resilience conferred by behavioral control. Thus, the circuitry recruited during the controlling experience determines its impact, rather than the act of control itself.

Keywords: Medial Prefrontal Cortex (mPFC), Dopamine, resilience, coping, Corticostriatal circuit

Disclosure: Nothing to disclose.

P481. Proteomic dysregulation associated with psychosis in Alzheimer’s disease

Leslie Nucifora, Bobana Samardžija, Laura Gomez-Isaza, Russell Margolis, Christopher Ross, Robert Sweet, Julia Kofler, Gwenn Smith, Benjamin Orsburn, Frederick Nucifora

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

Background: Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by protein aggregation and cognitive decline. Many patients with AD experience neuropsychiatric symptoms, including hallucinations and delusions, also prominently observed in schizophrenia (SZC), that are associated with poor clinical outcomes, in part due to the lack of safe and effective treatments. The neurobiological basis of psychotic symptoms in AD is not well understood. Understanding the mechanisms of psychotic symptoms in SCZ, may provide an improved understanding of psychosis in AD (AD + P).

An observation that is a logical focus of study for AD + P comes from our recent work in SCZ, which showed increased protein aggregation in a subset of SCZ patients using postmortem brain. A preliminary comparison of the insoluble proteins in this SCZ subtype identified in our study demonstrated overlap with proteins identified in a similar AD proteomic insolubility study. Furthermore, we have recently published evidence that directly correlates protein aggregation in SCZ to cognitive deficits. These results suggest a common proteomic signature in AD and SCZ that may be related to cognitive deficits and psychosis. Therefore, we hypothesize that the overlapping proteomic changes associated with protein aggregation in schizophrenia and Alzheimer’s disease will implicate genes and proteins, and thus specific cellular processes, associated with neuropsychiatric symptoms and psychosis observed in Alzheimer’s disease. This work may provide evidence that the mechanisms responsible for aggregation in schizophrenia contribute to the neuropsychiatric symptoms that occur in Alzheimer’s disease.

Methods: The superior temporal gyrus (STG) region of postmortem brains from patients with SCZ, AD + P and AD - P were obtained from the University of Pittsburgh Medical Campus Alzheimer’s Disease Research Center (ADRC) and matched control brains were obtained by the JHUSOM Alzheimer’s Disease Research Center (ADRC). A brain fractionation protocol was used to obtain a purified insoluble protein pellet from a 10% homogenate brain sample. The insoluble protein fraction has been analyzed by LC-MS/MS on an Orbitrap XL™ Hybrid Ion Trap-Orbitrap Mass Spectrometer (Thermo Scientific) using a label free proteomic approach. Proteome Discoverer 2.2. Beta versions has been used to perform all data processing and finalized in 2.20.388. SequestHT was used for the identification of peptides and Percolator has been used to assess the false discovery rate (FDR) at the peptide spectral match (PSM) level and filtered to a 1% estimate cutoff. MaxQuant, MetaMOrpheus, MSFragger and Proteome Discoverer was used to analyze the raw proteomic data to maximize the number of proteins identified and determine significant protein enrichment. Gene Ontology Enrichment Analysis, Ingenuity Pathway Analysis (Ingenuity Systems; Redwood City, CA, USA), and weighted correlation network analysis has been used to analyze the enriched protein data to determine most relevant proteins and biological processes, and networks associated with protein enrichment.

Results: We have obtained promising preliminary data through a comparative proteomic investigation of the insoluble proteins in a SCZ subtype, characterized by protein aggregation, with proteins identified in a previous similar proteomic insolubility study in AD, which suggests overlapping proteins and pathways between the two neuropsychiatric disorders. These results suggest a common proteomic signature in AD and SCZ.

Conclusions: By elucidating cellular mechanisms of protein aggregation observed in SCZ, a disease characterized by psychosis and cognitive decline, and the wealth of proteomic and genomic data related to AD, the cellular abnormalities leading to psychosis associated with AD can be better understood. Understanding of etiology and pathogenesis may ultimately result in the identification of new therapeutic targets for the treatment of AD + P.

Keywords: Alzheimer’s disease, Proteomics, Translational research, Psychosis, Schizophrenia

Disclosure: Nothing to disclose.

P482. Resting-state and oddball EEG biomarkers of SPG302 network modulation in AlS with randomized trial results and translational expansion to Alzheimer’s and schizophrenia

Ernest Pedapati, Peter Vanderklish, Sharron Gargosky, Craig Erickson, Stella Sarraf

Cincinnati Children’s Hospital/University of Cincinnati, Cincinnati, Ohio, United States

Background: SPG302 is a first-in-class investigational agent hypothesized to normalize synaptic function implicated in neurodegenerative disorders. Resting-state and event-related potentials (ERPs) Electroencephalography (EEG) including resting-state and event-related potentials (ERPs) can provide mechanistic insight and translational endpoints. We present randomized controlled trial (RCT) and serial exposure findings in ALS up to 24 months, with Alzheimer’s disease (AD) and schizophrenia cohorts underway.

Methods: In ALS (N = 23; NCT05882695), participants (males and females) were randomized 3:1 to SPG302 vs. placebo for 4 weeks, followed by open-label extension. Source localized resting-state EEG (eyes-open/eyes-closed) quantified spectral power (delta–gamma) and amplitude envelope correlation (AEC) connectivity. Oddball paradigms measured N100, N200, and P300 components, harmonized to a drug-exposure timeline. RCT and longitudinal outcomes were modeled using mixed-effects frameworks. Clinical endpoints included ALS Functional Rating Scale-Revised (ALSFRS-R) and Edinburgh Cognitive and Behavioral ALS Screen (ECAS).

Results: SPG302 produced regionally specific oscillatory modulation, with significant increased delta and alpha power. Resting connectivity analyses demonstrated robust reductions in theta/gamma hyperconnectivity, particularly during eyes-open rest. N100 amplitudes were stable but SPG302 participants demonstrated a relative improvement (shortening) of N100 latency compared to placebo. P300 amplitude was maintained under SPG302, contrasting with placebo-related decline. Importantly, increases in delta power at week 4 and 8 correlated with slower ALSFRS-R decline measured at 24 weeks. Placebo-shifted drug exposure data demonstrated similar EEG power trends over 24-weeks. AD and schizophrenia cohorts are ongoing, with preliminary findings to be reported.

Conclusions: SPG302 engages neurophysiological biomarkers of dysfunction in ALS, with direct correlations to clinical function. Preservation of P300 amplitude, N100 latency acceleration, and resting-state spectral and connectivity normalization highlight convergent electrophysiological signatures. These findings support SPG302 as a cortical network modulator with cross-disease translational potential.

Keywords: ALS, EEG, CNS drugs, Drug Discovery—new approaches

Disclosure: Spinogenics, Consultant, Self

P483. KCNA4 expression is reduced in dual glutamatergic-serotonin neurons of the dorsal raphe nucleus, which are especially vulnerable to tau pathology in Alzheimer’s disease

Louis Kolling, Nagalakshmi Balasubramanian, Shafa Ismail, Alexander Feller, Jessica Alberhasky, Ruixiang Wang, Marco Hefti, Catherine Marcinkiewcz

University of Florida, Gainesville, Florida, United States

Background: Previously, we found that transgenic mice expressing human tau (htau mice) recapitulate many of the neuropsychiatric features of early Alzheimer’s disease (AD), though we did not report any changes in the action-potential properties of htau 5-HT neurons. We then used the Visium spatial gene expression platform to profile the htau dorsal raphe nucleus (DRN) and characterized computationally-distinct subregions. Within the centromedial (CM) DRN, we identified differential expression of ion-channel genes, which regulate the electrical activity and excitability of neurons. Pathological tau (Ptau) is known to disrupt neuronal excitability, which may concomitantly exacerbate its trans-synaptic spread.

Methods: The gene expression dataset from a previous Visium reaction was analyzed to identify transcriptomic changes in the CM DRN of htau mice. We used RNAscope to validate, and ex vivo electrophysiology to assess, findings of interest in dual glutamatergic-serotonin (5HT/glut) neurons. We next used in vivo chemogenetics to determine the role of neuronal excitability in Ptau accumulation of htau mice. Lastly, we used RNAscope and IF to assess these findings in the DRN of Braak 2 post-mortem human brain tissue, AD vs cognitively-normal controls (CN).

Results: The CM DRN is characterized by 5HT/glut neurons that differentially express ion-channel genes in the htau mouse, particularly Kcna4 (****p < 0.0001), Slc24a5 (*p = 0.0120) and Scn4b (*p = 0.0351). This is associated with an increase in evoked excitability (**p = 0.0055), higher normalized AP frequency (*p = 0.0492) and altered AP df/dt (****p < 0.0001). Chronic hyper-excitability enhances Ptau accumulation within the DRN of htau mice (**p = 0.0044) and confers enhanced novel-object recognition (*p = 0.0396). KCNA4 is reduced in 5HTglut neurons of AD vs CN controls (*p = 0.0437). Across all human tissue, 5HTglut neurons are especially vulnerable to tau pathology compared to 5HT-nonglut neurons (*p = 0.0003). Correlational assessment for KCNA4 and SLC24A5 with Ptau is ongoing.

Conclusions: Tau-mediated dysfunction of the DRN may be driven by changes in ion channel activity that concomitantly enable the spread of Ptau in Braak progression. Modulation of Kv1.4 activity and gene expression in 5HT/glut neurons may represent a critical therapeutic target for the treatment of tau-based AD.

Keywords: tau, spatial transcriptomics, chemogenetics, Preclinical Alzheimer’s Disease, translational Alzheimer’s Disease

Disclosure: Nothing to disclose.

P484. Sodium-glucose cotransporter-2 inhibitors and neuropsychiatric outcomes in individuals with psychiatric conditions

David Liebers, Tianshe He, Betensky Rebecca, Kaitlin Swinnerton, Chunlei Zheng, Paola Gilsanz, Mary Brophy, Nhan V. Do, Ricardo S. Osorio, Nunzio Pomara, Antonio Convit, Donald C. Goff, Dan V. Iosifescu, Nathanael R. Fillmore, Jaime Ramos-Cejudo

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

Background: Individuals with psychiatric disorders are at increased risk of dementia. Incidence of dementia is two- to three-fold higher among those with mood disorders and may be even higher among those with psychotic disorders. Impaired brain bioenergetic metabolism is increasingly understood to be a transdiagnostic feature of neuropsychiatric disease that forms part of the etiological basis major depressive disorder, bipolar disorder, schizophrenia and Alzheimer’s dementia. Sodium-glucose cotransporter-2 (SGLT2) inhibitors, antidiabetic medications that improve mitochondrial function via increased ketogenesis, have associated with reduced risk of depression and dementia in population-based studies, and have emerged as therapeutic candidates. In this study, we evaluated whether SGLT2 inhibitors associate with incidence of dementia and proxies of psychiatric acuity, including psychiatric emergency evaluations and hospitalizations, among individuals with psychiatric diagnoses.

Methods: We used a target trial emulation approach in a retrospective dataset from the national Veterans Health Administration. The study window was from January 2016 to June 2024, and included 122,725 patients over the age of 65 with a diagnosis of major depressive disorder, bipolar disorder or a schizophrenia spectrum disorder, without prior dementia diagnosis or history of sodium-glucose cotransporter-2 inhibitor prescription. The intervention was initiation and sustained use of an SGLT2 inhibitor for ≥3 months. Analyses were conducted on a per-protocol and intention-to-treat basis. Our primary outcome was incident all-cause dementia defined by ICD-coded diagnoses. Our exploratory outcomes included emergency psychiatric evaluations and psychiatric hospitalizations – proxies of psychiatric acuity. Covariates included demographics, comorbidities, psychiatric diagnoses, and medication use. Analyses used marginal structural models weighted by inverse probability of treatment and censoring.

Results: SGLT2 use associated with reduced odds of dementia diagnosis (OR, 0.60; 95% CI, 0.48–0.76) and psychiatric ED visits (OR, 0.71; 95% CI, 0.55–0.93). ITT analysis also showed lower odds of dementia (OR, 0.63; 95% CI, 0.54–0.76) and psychiatric ED visits (OR, 0.75; 95% CI, 0.62–0.91). No significant association was observed for psychiatric-related hospitalizations.

Conclusions: Among older individuals with psychiatric conditions, SGLT2 inhibitor use associated with reduced incident dementia and risk of emergency psychiatric evaluation. There was no significant association with risk of psychiatric hospitalization.

Keywords: metabolism, Dementia, Ketogenic diet, emulated target trial

Disclosure: Nothing to disclose.

P485. Alcohol metabolism: a key driver of brain aging and Alzheimer’s disease risk?

Nagalakshmi Balasubramanian, Shafa Ismail, Yu Xu, Benjamin Hartman, Catherine Marcinkiewcz

University of Florida, Gainesville, Florida, United States

Background: Alzheimer’s disease (AD) is projected to affect 152 million people worldwide by 2050, with chronic alcohol use recognized as a modifiable risk factor. Excessive drinking in individuals with alcohol use disorder (AUD) accelerates cognitive decline, hippocampal atrophy, and metabolic aging, yet the mechanisms linking alcohol metabolism to AD remain poorly understood. Importantly, AUD and AD also share genetic risk factors, most notably the ALDH2*2 (rs671; E504K) variant, carried by ~40% of East Asians and ~8% of the global population. This mutation impairs alcohol metabolism, causing acetaldehyde accumulation even after minimal alcohol exposure. While the flushing response can be protective against AUD, elevated acetaldehyde promotes cellular damage and neurodegeneration. Reports of early-onset AD in ALDH2*2 carriers further implicate impaired alcohol metabolism as a direct driver of AD risk and raise concerns that pharmacological ALDH2 inhibition—proposed for AUD treatment—may inadvertently exacerbate neurodegeneration. These findings suggest that alcohol and its toxic metabolites synergistically accelerate metabolic aging and vulnerability to AD. To test this mechanism, we employed ALDH2*1 knock-in mice (heterozygous) as a translational model of impaired alcohol metabolism.

Methods: Adult male and female C57BL/6J wild-type (WT) and ALDH2*1 mice (2 months; n = 8/group) were exposed to a modified chronic intermittent ethanol (CIE) vapor paradigm (10 h/day, 4 days/week, 4 weeks) without ethanol loading or pyrazole pretreatment. Blood ethanol and acetaldehyde concentrations were quantified weekly by GC–MS. A subset of WT and ALDH2*1 mice underwent indirect calorimetry (Promethion chambers) with an acute ethanol challenge (1.5 g/kg, i.p.). In a separate cohort, WT and ALDH2*1 mice received hippocampal (CA3) injections of AAV-P301L tau, followed by CIE exposure and histopathological analysis of tau pathology via immunofluorescence (p-tau [AH36, Ser202/Thr205], NeuN, MAPT, astrocytic markers). All data are presented as mean ± SEM. Data were analyzed using two-way ANOVA with factors of genotype (WT vs. ALDH2*1) and sex, or CIE exposure (air vs. ethanol) where appropriate. Post hoc comparisons were conducted using Tukey’s multiple comparisons test to determine pairwise differences between groups. Main effects and interactions were considered significant at p < 0.05. For metabolic data following acute ethanol challenge, repeated measures were visualized, and two-way ANOVA was applied to specific timepoints (e.g., 6 h post-injection).

Results: WT females exhibited higher blood ethanol concentrations (BECs) than males [F(1,24) = 10.41, p = 0.036], with no sex difference in ALDH2*1 mice (sex × genotype interaction: p = 0.0248). For acetaldehyde, two-way ANOVA revealed a sex × genotype interaction [F(1,24) = 4.375, p = 0.0472]. Post hoc tests showed (a) elevated acetaldehyde in ALDH2*1 vs. WT mice (males: p < 0.001; females: p = 0.0427), and (b) higher acetaldehyde in ALDH2*1 males vs. females (p = 0.0065).

Promethion analysis revealed persistent metabolic impairments (reduced RER up to 6 h) after ethanol challenge. Two-way ANOVA of RER at 6 h showed main effects of ethanol [Males: F(1,16) = 25.16, p = 0.0001; Females: F(1,17) = 19.79, p = 0.004] and genotype in females [F(1,17) = 5.552, p = 0.0307]. Post hoc analysis confirmed significant impairments in ethanol-treated WT [Males: p = 0.0373; Females: p = 0.0330] and ALDH2*1 [Males: p = 0.0002; Females: p = 0.0009] vs. saline, with stronger effects in ethanol-exposed ALDH2*1 mice [Males: p = 0.0350; Females: p = 0.028].

Immunofluorescence showed increased neuronal p-tau in the hippocampus of CIE-exposed WT and ALDH2*1 males. Two-way ANOVA revealed a main effect of CIE [F(1,18) = 25.16, p < 0.0001]. Post hoc analysis showed elevated neuronal p-tau in ventral dentate gyrus (DG) of WT (p = 0.0255) and ALDH2*1 (p = 0.0002) CIE mice vs. air controls, with a greater effect in CIE-ALDH2*1 vs. CIE-WT (p = 0.0301).

While in males, CIE consistently elevated neuronal p-tau, in females, two patterns emerged: neuronal p-tau (n = 4) or astrocytic p-tau (n = 4). For neuronal tau, two-way ANOVA revealed main effects of CIE [F(1,18) = 36.07, p < 0.0001] and genotype [F(1,18) = 13.66, p = 0.0017]. Post hoc analysis confirmed increased neuronal p-tau in ventral DG of WT (p = 0.0019) and ALDH2*1 (p = 0.0001) CIE mice, with stronger effects in CIE-ALDH2*1 vs. CIE-WT (p = 0.0104). For astrocytic tau, two-way ANOVA showed main effects of CIE [F(1,18) = 14.67, p = 0.0012] and genotype [F(1,18) = 21.92, p = 0.0002]. Post hoc analysis revealed higher astrocytic p-tau in CIE-ALDH2*1 vs. controls (p = 0.0006) and CIE-WT (p = 0.0006). Notably, control ALDH2*1 females exhibited baseline astrocytic p-tau, which was exacerbated by CIE.

Conclusions: These findings demonstrate that excessive alcohol use and impaired alcohol metabolism accelerate metabolic dysfunction and tau pathology, suggesting a strong mechanistic link to early AD risk. ALDH2*1 mice represent a translational model for dissecting alcohol metabolism–driven vulnerability to AD. The data also reveal sex-specific differences, with females showing unique susceptibility to astrocytic tau pathology, underscoring the need to investigate astrocytes as contributors to alcohol-related AD risk.

Keywords: Alcohol Use Disorder - Treatment, tau, Alzheimer’s dementia, metabolism, Alcohol consumption

Disclosure: Nothing to disclose.

P486. Targeting neuropsychiatric symptoms in aging and Alzheimer’s: from benzodiazepines to psychedelic therapies

Kameron Kaplan, Elise Webber, Lainey Toennies, Daniel Steinbrenner, Smriti Gupta, Grace Stutzmann, Holly Hunsberger

The Chicago Medical School At Rosalind Franklin University, North Chicago, Illinois, United States

Background: The presence of neuropsychiatric symptoms (NPSs) in cognitively normal individuals is one of the earliest signs of Alzheimer’s disease (AD). We’ve shown earlier anxiety-like behavior and cognitive decline in female APP/PS1 (AD) mice, which correlated with an unbalanced brain-wide network compared to males and controls. We’ve also validated these preclinical findings with the Alzheimer’s disease neuroimaging (ADNI) human dataset. To treat NPSs in the aged population, physicians often prescribe benzodiazepines (BZDs). However, these drugs can cause memory loss, and continued use may lead to neurodegenerative diseases. We’ve shown disrupted memory and altered neuronal ensemble activation and memory engrams in male and female mice after a single injection of Alprazolam. Here, we will explore the mechanisms behind BZD-induced amnesia and discuss psychedelics as an alternative.

Methods: Benzodiazepine study:

We hypothesize that during CFC encoding, BZDs potentiate inhibitory signals coming from upstream brain regions projecting to somatostatin (SST) interneurons. This consequently leads to their silencing and disinhibition of dCA1 principal cells, resulting in the observed enhanced encoding ensemble activation in the dCA1 of BZD-treated adult female mice. We injected 2–3 month-old male and female mice (n = 5/group) with saline or alprazolam (1mg/kg) 30 minutes before 3-shock contextual fear conditioning (CFC). Mice were sacrificed 1 hour after CFC training to determine activity dependent activation of SST interneurons (overlapping with c-fos). A separate cohort of SST-Cre mice (n = 5/group) were then bilaterally injected with cre-dependent excitatory DREADD virus to stimulate SST interneurons during alprazolam administration to rescue memory impairment.

Psilocybin study:

6–8 month-old control and APP/PS1 mice (male = 8–10, female = 5–10) were injected i.p. with psilocybin (3mg/kg) 1 week before either behavior testing or electrophysiology. To test avoidance and exploratory behaviors, mice were administered an open field and zero maze. Fear memory was measured using 3-shock CFC. Freezing behavior was used as a proxy for memory. For electrophysiology, acute brain slices (400 um) were prepared in oxygenated artificial cerebrospinal fluid. A stimulating electrode was placed along Schaffer collaterals at CA3, recording electrode placed along Schaffer collaterals at CA1. The Input/Output curve assessed basal synaptic excitability, Paired pulse ratio assessed vesicle release properties upon short-interval paired stimulation, and Long-term potentiation assessed increased excitatory post-synaptic potential increase post-high frequency tetanic stimulation. Whole-cell patch clamp was then performed using Ryanodine receptor evoked calcium release (Caffeine 10mM) and measuring spontaneous miniature excitatory postsynaptic currents (sEPSCs) and inhibitory postsynaptic currents (sIPSCs). IPSCs from human AD patients, differentiated into glutamatergic neurons were then bathed in different doses of psilocybin to determine caffeine evoked calcium release.

A three or two-way ANOVA (drug x Tg x sex) (drug x sex) was run for all behaviors and cell counts with RMANOVA run across time in the fear conditioning test.

Results: BZD-treated males and females show decreased SST interneuron activation after CFC encoding (F(1,16) = 33.25, p = 0.0001). We hypothesize that DREADD stimulation of SST+ interneurons will compensate for the BZD-induced silencing of these interneurons and allow for proper memory formation in the dCA1.

We’ve found increased exploratory behavior in female control and AD mice treated with psilocybin compared to saline in the zero maze. Psilocybin treatment increased distance traveled and mobility (F(1, 12) = 20.76, p = 0.0007). Psilocybin-treated male control mice exhibited increased time spent in the open areas of the zero maze compared to saline-treated mice (F (1, 17) = 5.930, p = 0.0262), indicating less anxiety-like behavior. While female mice, regardless of drug, exhibited similar freezing levels, male control and AD mice treated with psilocybin showed enhanced freezing compared to saline mice (F (1, 17) = 5.798, p = 0.0277). Electrophysiology revealed decreased synaptic hyperexcitability in the hippocampus as shown by I/O curves and Paired Pulse Ratio (two-way ANOVA, (F(1,32) = 7.216, p = 0.011). Patch clamp also revealed decreased evoked release of calcium (t-Test, p < 0.0001) and increased sIPSCs. Caffeine-evoked calcium release in IPSC-derived glutamatergic neurons corroborated with previous results and also revealed decreased activation (T-test from baseline, p < 0.0001).

Conclusions: Long-term use of BZDs could continue to dysregulate E/I, leading to cognitive decline. Hyperexcitability (E/I imbalance) has been observed in human AD patients and rodent models of AD before cognitive decline. Therefore, decreasing this hyperexcitability could be a beneficial treatment option in the early stages of AD. Psilocybin or drugs that mimic psilocybin could represent novel drugs for treating the neuropsychiatric symptoms associated with Alzheimer’s disease and, in turn, cognitive decline. This would give alternative options to those suffering, as BZDs have adverse side effects and do not work for all patients. However, we did not observe any rescue of memory in female AD mice treated with psilocybin, which future studies will need to address at different doses.

Keywords: Alzheimer’s Disease, Neuropsychiatric symptoms (NPS), Sex differences, Psilocybin, Benzodiazepines

Disclosure: Nothing to disclose.

P487. Disrupted aperiodic neural activity signatures in Rett syndrome across human EEG and mouse electrophysiology

Christian Cazares, Krissy Lyon, Setareh Metanat, Devorah Kranz, April Levin, Michela Fagiolini, Charles Nelson, Bradley Voytek

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

Background: Rett syndrome, caused by mutations in the X-linked Mecp2 gene, is characterized by typical early development followed by developmental regression around 18 months of age, leading to severe cognitive and motor dysfunction. While abnormalities of oscillatory neural activity have been reported in Rett syndrome patient EEG, aperiodic neural activity patterns remain poorly understood, despite their emerging role as biomarkers of cortical dysfunction in neurodevelopmental conditions. One such marker is the aperiodic exponent, which reflects the slope of the 1/f power spectrum decay and is thought to serve as an index of cortical excitation-inhibition balance. In typical development, the aperiodic exponent diminishes with age, leading to spectral “flattening”, which has been attributed to developmental changes in inhibitory drive. However, little is known about whether aperiodic measures differ in Rett syndrome compared to age-matched neurotypical children, if they can serve as markers of developmental divergence, or if these alterations can be recapitulated in widely used mouse models of the condition.

Methods: Here, we analyzed task-free EEG data from 99 participants (61 girls with Rett syndrome in active regression and post-regression stages and 38 neurotypical controls, aged 2–14 years) alongside task-free, intracranial recordings from primary visual cortex in female Mecp2-/+ mice and littermate controls (2 per group, aged 14 weeks).

Results: Using spectral parameterization methods, we found that girls with Rett syndrome exhibited elevated aperiodic exponents compared to neurotypical controls, indicating “steeper” spectral slopes that contrast with the neurotypical “flattening” developmental trajectories reported in the literature. Complementary findings from Mecp2-/+ mice also revealed elevated aperiodic exponents relative to wild-type littermates, suggesting conserved alterations in aperiodic neural activity across species with consistent directional changes.

Conclusions: These cross-species patterns suggest that Mecp2 dysfunction fundamentally disrupts electrophysiological measures of cortical excitation-inhibition balance, potentially reflecting altered inhibitory function that contrasts with typical age-related decline in inhibitory drive. Our findings support the inclusion of task-free aperiodic neural activity measures as candidate biomarkers of neurodivergence in Rett syndrome that could provide objective assessments of therapeutic efficacy in clinical settings and accelerate treatment discovery in associated mouse models.

Keywords: EEG, Rett Syndrome, MeCP2, Local field potentials, Genetic mouse model

Disclosure: Nothing to disclose.

P488. Structural similarity analysis identifies 7q11.23 gene dosage effects on neural circuits underlying social and visuospatial functions

Madeline Garvey, J. Shane Kippenhan, Michael Gregory, Philip Kohn, Tiffany Nash, Carolyn Mervis, Daniel Eisenberg, Anna Kelemen, Darby Krugel, Petra Vertes, Edward Bullmore, Karen Berman

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

Background: Identifying neurogenetic mechanisms in debilitating neuropsychiatric conditions like autism and schizophrenia is complicated by substantial clinical and genetic heterogeneity. Studying copy number variation (CNV) disorders in which the affected genes are known and the associated cognitive and behavioral phenotypes are well-defined enables reduction of the genetic search space by several orders of magnitude, offers clear focus on neural circuits, has potential to uncover mechanistic threads that translate across scales from genes to cells to circuits, and that may inform targeted treatments. Rare reciprocal CNVs at chromosomal location 7q11.23 confer clear contrasts in two important cognitive and behavioral domains, visuospatial processing and social motivation. Williams syndrome (WS) is caused by a hemideletion (leaving only one copy of ~27 affected genes) and is characterized by increased social drive (Klein-Tasmin et al., 2003) coupled with severe weakness in visuospatial abilities (Mervis et al., 2000), whereas the 7q11.23 duplication syndrome (Dup7) is caused by duplication of these same genes (resulting in three gene copies) and is characterized by severe social anxiety and shyness (Mervis et al. 2015) along with relatively preserved spatial processing. Multimodal neuroimaging can serve as an intermediate in bridging the gap between these known genotypes and behavioral phenotypes. Here, we used a novel structural brain analytic approach called Morphometric INverse Divergence (MIND), which synthesizes several MRI-measured brain parameters to quantify inter-regional “similarities” across the cortical brain surface. Prior work has shown that these “structural similarity” measurements can be related to underlying genetic and cytoarchitectural profiles and can reveal coordinated patterns across brain regions (Seidlitz et al., 2018; Sebenius et al., 2023). Here, MIND was applied to structural MRI scans of children and adolescents with WS and Dup7, as well as typically developing individuals (TDs), with a goal of identifying brain networks that are affected by 7q11.23 copy number.

Methods: 34 participants with WS (mean age = 12.0 ± 4.0 years, 23 females), 21 individuals with Dup7 (mean age = 14.1 ± 2.9 years, 9 females), and 112 age- and sex-matched TDs (mean age = 12.8 ± 3.3 years, 66 females) completed T1-weighted multi-echo MPRAGE structural MRIs. Three scans for each individual were visually inspected for quality, averaged into one image per participant, and processed with Freesurfer software to obtain five different measurements (volume, cortical thickness, surface area, mean curvature, and sulcal depth) at each vertex across the cortical surface. Sets of these vertex-level measurements were then parcellated based on the Desikan-Killiany anatomical atlas into 68 distinct cortical regions. Each region’s set of measurements was subsequently compared to multivariate distributions of every other region’s set of measurements using Kullback-Leibler divergence, resulting in a single measurement of “similarity,” or conversely of “specialization” for each region. General linear modeling was subsequently used to test for 7q11.23 gene dosage effects (1 gene copy [WS] vs. two copies [TD] vs. three copies [Dup7]) while controlling for age and sex. Results were then Bonferroni-corrected for the 68 regions (p < 0.000735).

Results: We found significant 7q11.23 gene dosage effects in structural similarity for twelve of the 68 cortical regions: bilateral fusiform gyri, left isthmus of the cingulate, left posterior cingulate, left paracentral lobule, bilateral postcentral gyri, left medial orbitofrontal region, right cuneus, right parahippocampal region, right pars opercularis, and right superior temporal gyrus. Regions with negative t statistics - reflecting areas with relatively less morphological specialization in people with WS and greater specialization in those with Dup7 - included left paracentral lobule, bilateral postcentral gyri, right cuneus, and right parahippocampal gyrus. In contrast, regions with positive t statistics - reflecting relatively greater specialization in people with WS and less specialization in those with Dup7 - included bilateral fusiform gyri, left isthmus of the cingulate, left posterior cingulate, right pars opercularis, left medial orbitofrontal region, and right superior temporal gyrus.

Conclusions: MIND identified directionally opposite effects of 7q11.23 gene dosage on cortical regions that overlap with circuits supporting social motivation (fusiform/cingulate/superior temporal regions) on the one hand compared to circuits supporting visuospatial processing (postcentral/parahippocampal/paracentral/cuneus regions) on the other hand. These results not only assort with 7q11.23 copy number but also align with known behavioral phenotypes of people with WS and Dup7. The identified patterning of results in circuit-level brain signatures thus represents an important intermediate between genes and behavior, helps to bridge these investigative scales, and may lead to discovery of cross-scale mechanisms. Future directions will (1) relate these findings to cortical spatial patterns of gene expression in an effort to identify specific genes in the 7q11.23 locus that may be driving these effects and (2) investigate how these patterns of cortical similarity change longitudinally across development.

Keywords: 7q11.23 copy number variation, Neurodevelopmental Disorders, Structural neuroimaging

Disclosure: Nothing to disclose.

P489. Cannabidivarin (CBDV) improves repetitive behaviors compared to placebo in Prader-Willi Syndrome

Eric Hollander, Casara Ferretti, Blerona Shaipi, Sam Rashkovich, Yin Zhao

Albert Einstein College of Medicine, Bronx, New York, United States

Background: Prader-Willi Syndrome (PWS), a rare neurodevelopmental disorder caused by the loss of paternally expressed genes on chromosome 15q11–q13, is characterized by hyperphagia, severe compulsivity and rigidity, and temper tantrums.

CBDV is a GPR55 agonist, targeting LPI-induced excitability and helping to restore excitatory/inhibitory (E/I) balance. CBDV binds to GPR55 with six times greater potency than CBD.

Methods: Fourteen youth with PWS (ages 7–27) enrolled in a 12-week randomized, double-blind, placebo-controlled study of CBDV.

Subjects received a weight-based dosing of 10mg/kg/day of CBDV or placebo (up to 800 mg/day) administered BID with food.

The Montefiore Einstein Rigidity Scale-Revised-PWS (MERS-R-PWS) assessed three domains of rigid behavior in individuals with PWS: Behavioral Rigidity, Cognitive Rigidity, and Protest Domain.

The Repetitive Behavior Scale – Revised (RBS-R) is 44-item self-report questionnaire used to measure repetitive behavior, consists of 6 subscales: Stereotyped Behavior, Self-injurious Behavior, Compulsive Behavior, Routine Behavior, Sameness Behavior, and Restricted Behavior.

The Children’s Yale-Brown Obsessive Compulsive Scale (CY-BOCS) is 10-item clinician measure designed to assess the severity of obsessive-compulsive symptoms in children and adolescents.

Results: Reductions in repetitive behaviors and rigidity were observed on the Montefiore-Einstein Rigidity Scale-Revised-PWS (MERS-R-PWS), Repetitive Behavior Scale-Revised (RBS-R), and the Children’s Yale-Brown Obsessive Compulsive Scale (CY-BOCS) in the CBDV group relative to placebo. Overall, the data reflect a very large treatment effect (Cohen’s d range from 1.4 to 2.4) favoring CBDV, supporting its potential to target core behavioral symptoms in Prader-Willi Syndrome.

Conclusions: This is the first study to examine the effects of cannabidivarin (CBDV) on repetitive behaviors and rigidity in individuals with Prader-Willi Syndrome (PWS). Results demonstrated a large reduction in participants receiving CBDV vs placebo in behavioral rigidity and protest-related responses. Improvements were evident across multiple caregiver- and clinician-reported outcome measures.

These findings align with preclinical evidence supporting CBDV’s modulatory role in excitatory/inhibitory balance and inflammation, mechanisms believed to underlie rigidity and compulsivity in neurodevelopmental conditions.

While the sample size was small, this feasibility trial supports the potential utility of CBDV in treating key behavioral symptoms in PWS and provides a strong rationale for future, larger-scale studies to further evaluate efficacy, optimal dosing, and long-term safety.

Keywords: neurodevelopmental disorders, cannabidiol, CBDV, translational models, clinical trials, Prader Willi Syndrome, Compulsivity

Disclosure: Jazz Pharmacueticals, Contracted Research, Self, Department of Defense, Contracted Research, Self

P490. Prenatal bisphenol exposure and neural correlates of attention in infancy

Amy Margolis, Katrina Simon, Huiyu Yang, Fifer William, Fox Nathan, Rauh Virginia, Champagne Frances, Beebe Beatrice, Herbstman Julie

The Ohio State University, Columbus, Ohio, United States

Background: Bisphenols (BP) are endocrine-disrupting chemicals commonly used as plastic additives in food packaging materials. Data from the National Health and Nutrition Examination Survey (NHANES) study show that 93% of individuals in the United States have detectable BP in their urine. Animal models consistently show causal effects of prenatal exposure to BP-A and its replacements (BPF, BPS) on attention and on gene expression in brain regions critical to attention, including the prefrontal cortex, with effect modification by sex. Human studies link prenatal BP exposure with altered attention, and one study reported associations between prenatal BP exposure and reduced volume in the inferior frontal and superior occipital gyri in 2–7-year-olds. No study to date has examined if prenatal exposure to BP is associated with infant neural function underlying attention.

Herein we tested associations of prenatal BP exposure and 9-month infant neural function during an auditory oddball task using electroencephalography (EEG) and event-related potentials (ERPs). The mismatch response (MMR) compares brain activity elicited by a frequently repeated “standard” sound versus an infrequent “deviant” sound. The P3, a positive going waveform, measures neural response to novel stimuli. The MMR and P3 have been linked with altered attention and risk for Attention-Deficit/Hyperactivity Disorder. We tested if prenatal exposure to BPA, BPF or BPS was associated with the MMR or P3 in 9-month infants. We hypothesized that higher prenatal BP exposure would be associated with reduced MMR and P3 amplitudes, indicative of reduced attentional processing. Given the endocrine-disrupting properties of BP and prior reports of sex differences in effects of exposure, we explored sex differences in stratified analyses.

Methods: Pregnant women and their infants were enrolled in a prospective birth cohort study. Women provided 3 urine samples over the course of 48 hours in the 3rd trimester of pregnancy from which BPA, BPF and BPS were extracted. Infants completed an auditory oddball paradigm at ~ 9-months old. The task presented a series of counterbalanced tones at ~75 dB peak SPL (stimulus length: 200 ms; interstimulus interval: 300 ms). Tones were presented in two blocks, each consisting of 400 tones with three different sound types (standard, deviant, or novel). Novel sounds consisted of complex sounds (e.g., a car horn or cow mooing). Within each block, standard, deviant, and novel tones appear 80%, 10%, and 10% of the time, respectively. N = 78 infants (M = 10.68 months; 59% male) had useable BP and EEG/ERP data. Ninety-one percent of mothers identified as Hispanic/Latino. Most mothers reported having at least a high school degree or equivalent (85.9%), with 24.4% having completed a 4-year college degree or higher.

Statistical analyses were run in R Statistical Software version 2024.12.1. Log-transformed bisphenol measures were entered into a principal components analysis (PCA) to identify patterns of exposure. One-sample t-tests compared the mean amplitude of the difference wave of the MMR or P3 at each region of interest (ROI) to 0. Separate linear regression models examined associations between prenatal BP and either the MMR or P3, adjusting for infant age at visit, sex, maternal education, and birthweight-for-gestational age z-score. Primary analyses were adjusted for multiple comparisons using the Benjamini-Hochberg false discovery rate (FDR) procedure. Post-hoc power analysis showed 76 percent power to detect moderate effects at this sample size.

Results: PCA revealed three distinct patterns of BP exposure. The first principal component (PC1) constituted general overall exposure to all three BPs, with moderate positive loadings across BPA, BPF, and BPS. A clear MMR was detected between 200–400 ms post-stimulus at the regions of interest (T7/T8 F7/F8, FCz), reflecting a larger response to deviant compared to standard sounds at the ROIs. The mean amplitude of the MMR at T7/T8 and F7/F8, but not FCz, was significantly different from 0, (T7/T8: t(77) = 4.50, p < 0.001; F7/F8: t(77) = 5.14, p < 0.001; FCz: t(77) = 1.38; p = 0.17). As such, we examined the MMR at T7/T8 and F7/F8. The mean amplitude of the P3 differed significantly from 0 at all clusters (ps < 0.05).

Total BP exposure (PC1) was significantly associated with MMR mean amplitude at F7/F8, spanning the frontolateral left and right cortex, with higher exposure associated with a reduced MMR mean amplitude (β = −0.32; p = 0.016, FDR-corrected). No other pattern of BP exposure (PC2 or 3) was associated with MMR mean amplitude (all corrected ps > 0.05). In stratified analyses, higher prenatal overall BP exposure was associated with reduced MMR amplitude at F7/F8 in females (β = −0.58; p = 0.004) but not in males (β = 0.13, p = 0.40). Prenatal BP exposure was not associated with mean amplitude of the P3 in primary or stratified analyses.

Conclusions: The study contributes two novel findings to the literature on prenatal BP exposure and attention: 1) potential effects of BP on attention are observable as early as infancy, and 2) the MMR may serve as a neural marker of prenatal BP exposure. Sex-stratified analyses indicated this association may be specific to females. Results suggest that prenatal BP exposure may perturb neural processes underlying attention, with effects observed by 9–10 months of age. Documenting these associations will inform public health interventions aimed at reducing exposure and aid in developing screening tools to identify infants at risk for attention problems.

Keywords: Environmental exposures, Attention, EEG/ERP electrophysiology, Neurodevelopmental and psychiatric disorders, ADHD

Disclosure: Nothing to disclose.

P491. Distinct alterations in frontal oscillatory networks uniquely distinguish autism subgroups with intellectual disability: a deep-learning analysis of resting state EEG data

Amy Rapp, Jennifer Foss-Feig, Alexander Kolevzon, Tess Levy, Paige Siper, Joseph Buxbaum

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

Background: Meta-analytic findings show several differences in oscillatory measures derived from resting state EEG data among individuals with autism spectrum disorder (ASD), although these studies have largely excluded individuals with co-occurring intellectual disability (ID; i.e., IQ < 70–75). Further, most of these studies used region-averaged, univariate measures that reduce the richness of high-dimensional neural dynamics. Recent machine learning approaches have sought to advance beyond standard spectral analyses by integrating measures of oscillatory power with connectivity measures to provide insights into synchronization of neural regions.

Thus, the present study applied a deep-learning framework to resting state EEG data to reveal neurophysiological markers that uniquely characterized individuals with ASD and rare genetic disorders (RGD) associated with ASD who also meet criteria for ID, populations that are understudied and consequently, underserved. Specifically, an approach that leveraged multivariate power topographies and machine learning classifiers was used to identify distributed spatial-frequency patterns that best discriminated ASD subgroups and typically developing (TD) controls.

Methods: Participants (N = 69) included children and adolescents belonging to three groups: (1) ASD with ID without a known genetic cause (n = 26), (2) rare genetic disorders associated with ASD with ID (i.e., Phelan-McDermid syndrome, ADNP syndrome; n = 24), and (3) TD controls (n = 19).

Continuous eyes-open resting state data were recorded using a 128-channel Hydrocel Geodesic Sensor Net. Standardized pre-processing scripts were applied. Data were epoched into 10 second windows, and segments were transformed using complex Morlet wavelets. Per-subject feature sets (i.e., regularized covariance matrices and phase-locking value [PLV] connectivity within canonical frequency bands) were derived from these time-frequency representations.

A multinomial logistic regression model with elastic net regularization was trained on the extracted power features to classify the three groups. Hyperparamaters were tuned using repeated stratified 3-fold cross-validation, optimized for balanced accuracy. Signed model coefficients were extracted to estimate feature importance. A positive coefficient for a certain group indicated that higher values of the feature increased the likelihood of classification into that group, whereas a negative coefficient indicated that the feature was suppressed in that group.

Network analyses were then conducted on group-averaged PLV matrices to identify group-discriminative connections. Pairwise group contrasts were computed edgewise correcting for multiple comparisons.

Results: The classifier performed with 78% overall accuracy, similar to the harmonic mean of precision and recall in each group (F-1 score: ASD with ID: = 0.78; RGD with ID = 0.77; TD = 0.80), suggesting balanced performance across groups. The top features most strongly weighting the classifier included alpha, beta, delta, and theta power across a small set of electrodes in the frontal/prefrontal regions.

The ASD with ID and RGD with ID groups showed distinct power and connectivity signatures that differentiated them from both TD youth and one another. Alterations in theta power and synchrony uniquely characterized the RGD with ID group. Specifically, this group demonstrated a pattern of strong theta at channel 20 (~AFz), β = 1.54, which was highly uncharacteristic of the TD group, β = −2.16, and virtually absent in the ASD with ID group, β = 0.06. A similar pattern was observed at channel 8 (~Fp2), such that strong theta was seen for the RGD with ID group, β = 0.93, which was highly uncharacteristic of the ASD with ID group, β = −0.96, and virtually absent in the TD group, β = 0.03. Consistent with this, edgewise PLV contrasts revealed that the RGD with ID group showed significantly weaker synchrony in the theta band between AFz and Fp2 and other prefrontal/dorsolateral prefrontal sites (e.g., AF8, F2) compared to the TD and ASD with ID groups, all pFDR < .02.

Alterations in alpha power and synchrony most strongly characterized the ASD with ID group. Specifically, this group showed a distinct pattern of strong alpha at channel 16 (~AFz), β = 1.29, that was highly uncharacteristic of the TD group, β = −1.69, and virtually absent for the RGD with ID group, β = 0.30. Similarly, the ASD with ID group showed strong alpha at channel 12 (~AF4), β = 1.26, which was highly uncharacteristic of the RGD with ID group, β = −1.34, and virtually absent in the TD group, β = 0.08. Synchrony between AFz and AF4 with other frontal sites (e.g., FCz, Fp2) was significantly weaker compared to the RGD with ID and TD groups, all pFDR < .03.

Conclusions: Resting state EEG power and connectivity features differentiated autism subgroups from one another and TD youth with high accuracy, revealing distinctive neural signatures for each group. ASD subgroups were characterized by a local-dominant, network-fragmented pattern of exaggerated oscillatory power in frontal/prefrontal regions coupled with weak connectivity that was observed in the theta band for the RGD with ID group and in the alpha band for the ASD with ID group. Collectively, results highlight distinct frontal oscillatory network alterations as candidate biomarkers for distinguishing autism subgroups and point to potential phenotypic differences in executive and cognitive functioning. Future directions include examining syndrome-specific profiles.

Keywords: Autism Spectrum Disorder (ASD), Intellectual disability, EEG, Deep learning

Disclosure: Nothing to disclose.

P492. Comorbid autism spectrum disorder in youth presenting with early psychosis: clinical significance and underlying neurobiology

Cameron Carter, Kathleen Carlos, Jason Smucny, Joshua Rhillinger, Theo van Erp, Bethany Lazo, Kayla Tepes, Marjorie Solomon, Tyler Lesh

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

Background: Despite advances in understanding of the mechanisms and causes of psychotic disorders, they have increased in prevalence and remain a major cause of suffering and disability worldwide. Early intervention improves outcomes. However, treatment resistance is common and up to 40% of EP individuals have a poor response following one year of treatment. A barrier to effective treatment is the early identification of individuals who do not respond well to standard treatments. A potentially important source of heterogeneity in EP is the presence of comorbid autism spectrum disorders (ASD). ASD comorbid with EP (ASDEP) is widely recognized by clinicians familiar with individuals entering EP care. However, the clinical impact and underlying neurobiological mechanisms of this source of heterogeneity has received relatively little attention to date. An initial, study by our group, as well as review of the published literature, shows that more than 20% of EP referrals can be identified as having ASD. A small number of published studies suggest that ASDEP patients have poorer clinical outcomes than noncomorbid EP individuals. The present study is a systematic analysis of baseline symptoms, and in a subset of participants following 12 months of treatment. Importantly, we also use neuroimaging to investigate underlying neurobiological differences between EP subjects with and without ASD, using two measures we have found previously associated with treatment outcomes in EP (the free water (FW) fraction acquired using Diffusion Weighted Imaging, which is a putative biomarker of neuroinflammation, and the neuromelanin (NM) contrast, a proxy measure of dopamine activity in the midbrain dopaminergic nuclei).

Methods: Consecutive male and female subjects aged 12–35 were recruited from individuals with EP enrolling in the EDAPT clinic, a specialty care program at UC Davis. Clinical Diagnosis was established using the SCID and the presence of ASD comorbidity was established based upon clinical interviews and confirmatory documentation of ASD in school and regional center records. Clinical assessments included the BPRS, SAN, SAPS and GAF and were acquired at baseline (EP 189, ASDEP 35) and in a subset (EP 72, ASDEP 22) at 12 months. Additionally baseline FW (EP, 72 ASDP, 27) and NM (52 and 26) were acquired on a 3T Siemens scanner in a subset of subjects.

The multi-shell diffusion sequence was acquired with the following settings: TR = 11,400-msec, echo time = 92.4-msec, field of view = 240mm, 1.7mm isotropic voxels, 69 directions acquired P-A b-shells: 12 x b = 0, 15 x b = 500, 23 x b = 900, 31 x b = 1400. For NM a 2D GRE sequence with a MTC pulse at 300-degree flip angle: TE/TR = 3.94/313ms, 13 contiguous slices, 512 imaging matrix, FOV = 162 × 200 mm2, voxel size = 0.4 × 0.4 × 3.0 mm3, 11 measurements, FA = 40 degrees, MTC pulses (FA = 300 degrees, 1.2 kHz off-resonance, 10 ms duration), and 410 Hz/pixel receiver bandwidth.

Preprocessing of DWI and NM data utilized previously published approaches. FW was measured in gray (GM) and white matter (WM) masks. Neuromelanin (NM) contrast was measured in the Substantial Nigra pars compacta (SNpc) and reticulata (SNr) and the Ventral Tegmental area (VTA).

Clinical and functional differences between EP and ASDEP were tested with the Mann–Whitney U test. Relationships between neuroimaging (midbrain neuromelanin and free water) and diagnostic group were examined using linear regression adjusted for age, sex, and scanner type. Post-estimation analyses were run to assess stability of regression models.

Results: Compared to EP, ASDEP had higher BPRS scores at baseline (z = −2.03, p < 0.05) and follow up (z = −2.14, p < 0.05), higher negative symptoms (z = −2.10, p < 0.05), and lower GAF scores (z = 2.07, p < 0.05) at follow-up. There were no differences in positive symptoms at either time point. ASDEP had lower NM in the SNc (β = −0.97, p < 0.001), and SNr (β = −1.68, p < 0.001) than EP and interestingly higher NM in the VTA (β = 1.24, p < 0.001). ASDEP had lower WM FW in bilateral cerebral hemispheres (β = −0.017, p < 0.001). There were no FW differences in GM.

Conclusions: These initial findings add important new insights into the heterogeneity of early psychosis (EP), highlighting the common occurrence of this comorbidity in youth seeking EP care. ASDEP individuals had more severe overall symptoms at baseline and at follow up than noncomorbid individuals, and worse negative symptoms and functioning at follow up. Interestingly, our neuroimaging data suggest that EPASD show two features previously associated with poor treatment outcomes in EP. ASDEP show significantly lower levels of WM FW than noncomorbid subjects. We have previously shown that HIGH FW is associated with better outcomes in EP. Second, we show that ASDEP show significantly lower levels of NM in the SNpc and SNr that noncomorbid subjects. Previous work has shown that higher NM in the SN is associated with better treatment response in EP. Interestingly we also found that ASDEP had higher NM than EP in the VTA, a novel finding that needs to be confirmed and that may have treatment implications for this cohort. Identifying clinical and neurobiological differences in ASDEP (lower levels of neuroinflammation, a different pattern of midbrain dopamine dysregulation) sheds new light on a common but neglected aspect of the heterogeneity of EP. Additional prospective studies using gold standard EP and ASD related clinical assessments and treatment relevant neuroimaging biomarkers are needed to confirm and extend these findings.

Keywords: early psychosis, Autism Spectrum Disorder (ASD), MRI, Free water imaging, Neuromelanin-sensitive MRI

Disclosure: Nothing to disclose.

P493. Evaluation of regional vulnerability index for autism spectrum disorder in normally developing children demonstrates links to known risks and cognitive deficits

Antonio Pagán, Elliot Hong, Paul Thompson, Shuo Chen, Yizhou Ma, Si Gao, Xiaoming Du, Keiko Kunitoki, Bhim Adhikari, Ankeeta LNU, Alia Warner, Neda Jahanshad, Jair Soares, Peter Kochunov

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

Background: Large-scale neuroimaging meta-analyses report stable neurobiological signatures in Autism Spectrum Disorder (ASD) known for substance clinical and behavioral heterogeneity. Case-control studies in ASD also confirm consistent and replicable cognitive deficits, especially in the domain of fluid intelligence. We developed a Regional Vulnerability Index (RVI) to quantify an individual’s similarity to the ASD neuroanatomical signature. Using the Adolescent Brain Cognitive Development (ABCD) Study, we tested if the RVI-ASD in typically developing youth was associated with known risk factors and cognitive profiles, and evaluated its specificity against RVIs for Schizophrenia (SSD) and Major Depressive Disorder (MDD).

Methods: We analyzed neuroimaging, cognitive, and behavioral data for N = 7,442 ABCD participants at baseline (age ~10) and follow-up (age ~12). We calculated multi-modal RVIs for ASD, SSD, and MDD based on large-scale ENIGMA consortium meta-analyses. We tested associations between the RVI-ASD and risk factors at baseline and longitudinally, and evaluated the link between all three RVIs and NIH Toolbox cognitive scores.

Results: Parent-reported ASD diagnosis (N = 20; 16 M/4 F) was not associated with RVI-ASD (p = 0.218). However, whole group analysis demonstrated that elevated RVI-ASD tracked with the exposure of known risks and outcomes. At baseline, the whole-brain RVI-ASD was significantly associated with lower socioeconomic status (β = −0.006, p < .001) and pre-and-perinatal complications (β = 0.017, p < .001). While paternal age showed no overall effect, a strong positive association was found in fathers who were 35 years and older (r = 0.20, p < .001) and contrasting trend for mothers (age = 30 and above: r = −0.05, p = .003). Longitudinally, the RVI-ASD was stable (ICC = 0.91), and its change in two years was not predicted by these baseline risks. Critically, RVI-ASD was robustly associated with lower total cognitive performance (r = −0.11, p < .001), an effect not seen for the RVI-SSD (r = −0.04, n.s.) or RVI-MDD (r = −0.01, n.s.).

Conclusions: The ASD neuroanatomical signature is a specific endophenotype linked to both foundational risks and cognitive outcomes in the general pediatric population. The dissociation between baseline risk predictors and longitudinal change suggests the signature is largely stable by early adolescence. The specific link between the RVI-ASD and cognition, not present for SSD or MDD signatures, validates its use for studying the cognitive dimensions of autism liability. This population neuroscience approach is a powerful tool for disentangling the distinct biological pathways of neurodevelopmental conditions.

Keywords: Autism Spectrum Disorder (ASD), Human Neuroimaging, Cognition, Population neuroscience, Adolescent Brain Cognitive Development study

Disclosure: Nothing to disclose.

P494. Maturation of dorsal association tracts during preadolescence links to concurrent and future cognitive performance and transdiagnostic psychopathology

Danni Wang, Christopher Hammond, Betty Jo Salmeron, Xiang Xiao, Laura Murray, Hong Gu, Tianye Zhai, Annika Quam, Justine Hill, Hieu Nguyen, Hanbing Lu, Amy Janes, Thomas Ross, Yihong Yang

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

Background: Adolescence is a critical window for the emergence of psychiatric disorders and coincides with rapid white matter (WM) maturation, supported by energy-intensive processes such as myelin synthesis and axon-oligodendrocyte interactions. However, it remains unclear whether deviations from normative WM development contribute to cognitive deficits or psychopathology. In particular, how mitochondrial functioning contributes to these brain-behavior associations is poorly understood.

Methods: In this study, we developed 30 whole-brain or tract-specific brain age models using tract-wise diffusion MRI features from 10,227 individuals (4,901 females; age 5–21 years) across three large-scale developmental cohorts, including longitudinal Adolescent Brain Cognitive Development (ABCD), Lifespan Human Connectome Project and Healthy Brain Network pediatric mental health studies. For each participant, the difference between the predicted age and chronological age is defined as tract- brain age gap (BAG), which reflects whether the WM development appears less or more mature than expected relative to his/her chronological age. In ABCD cohort, we conducted sparse canonical correlation analysis to identify latent dimensional associations of the tract-specific developments with multidomain cognitive and psychopathological assessments. We further included the independent assessments of follow-up cognitive measures and psychiatric diagnoses assessed by parent-reported Kiddie Schedule for Affective Disorders and Schizophrenia for DSM-5 (KSADS-5). Finally, high-resolution postmortem mitochondrial maps were incorporated to test whether brain–behavior associations aligned with regional mitochondrial content and respiratory capacity.

Results: All 30 tract-based models robustly predicted chronological age in both cross-validation (e.g. whole-brain model: R² = 0.70, MAE = 1.805 years) and independent testing. We found that tract-specific deviations in WM development of association and limbic/subcortical WM systems were linked to concurrent cognition and psychopathology, controlling for age and sex. The pattern of the association system showed significantly spatial consistency with the distributions of mitochondrial content and respiratory capacity within WM. Of note, more advanced maturation of the tracts within the association system significantly predicted better cognitive performance and less risks of KSADS-5 disorders (F = 10.52, pFDR < 0.001) two or three years later.

Conclusions: This study identifies tract-specific maturation as a robust biomarker for cognitive outcomes and transdiagnostic psychopathology at the individual level. By integrating developmental neuroimaging with postmortem mitochondrial profiles, our findings put forward a mechanistic framework linking energy-dependent white matter development to adolescent cognitive trajectories and psychiatric vulnerability. These results highlight potential neurobiological targets for early identification and intervention in youth at risk for psychiatric disorders.

Keywords: Adolescence, White-matter tract, Psychopathology, Psychiatric diagnosis

Disclosure: Nothing to disclose.

P495. Single-dose target engagement of the novel BK channel modulator SPG601 in fragile X syndrome

Craig Erickson, Peter Vanderklish, Stella Sarraf, Sharron Gargosky, Sarah Richter, Meredith Nelson, Grace Westercamp, Christina Gross, Ernest Pedapati

Cincinnati Children’s Hospital Medical Center and CinciNeuro, Cincinnati, Ohio, United States

Background: Fragile X syndrome (FXS) is the most common inherited cause of intellectual disability and autism spectrum disorder. Despite extensive research, no targeted treatments exist for the core symptoms of FXS. SPG601 represents the first large conductance (BK) potassium channel activator to enter clinical testing for FXS, designed to restore synaptic function by correcting specific ion channel dysfunction downstream of fragile X messenger ribonucleoprotein protein (FMRP) loss. Individuals with FXS, in particular males, exhibit consistent EEG abnormalities that relate to the severity of clinical phenotype while also being translationally relevant and present in the FMR1 KO mouse model. Specifically, excessive high frequency resting gamma band activity has been repeatedly demonstrated across species in FXS and related to the severity of cognitive deficits and behavioral concerns in FXS. Additionally, reduced alpha power in FXS has additionally been noted in humans with FXS and correlates with the severity of cognitive impairment.

Methods: We conducted a randomized, double-blind, placebo-controlled, 2-period balanced crossover study in 10 adult men with genetically confirmed full-mutation fragile X syndrome. Participants received single doses of SPG601 800 mg and placebo separated by a 1-week washout period. Endpoints included resting-state EEG power spectral density analysis across predefined frequency bands and auditory-evoked gamma oscillations combined with cognitive assessments using validated instruments and clinical global impressions scales.

Results: SPG601 demonstrated significant modulation of resting-state EEG power spectral density compared to placebo. Significant treatment effects were observed for gamma power (F = 5.20, p = 0.023), alpha2 power (F = 17.43, p < 0.001), and theta power (F = 7.06, p = 0.008). SPG601 also significantly modulated aperiodic EEG slope (F = 5.28, p = 0.022), indicating alterations in the broadband 1/f component reflecting excitation-inhibition balance. Cognitive improvement was observed in the Flanker Inhibitory Control and Attention Test across multiple scoring metrics (p = 0.027). No significant effects were observed on auditory-evoked gamma measures. SPG601 was well tolerated with a favorable safety profile.

Conclusions: This study provides the first clinical evidence that BK channel activation produces significant neurophysiological changes in fragile X syndrome, accompanied by a cognitive enhancement in executive function. This establishes proof-of-concept for BK channel modulation as a therapeutic strategy in FXS.

Keywords: fragile X syndrome, EEG biomarkers and target engagement, Touchscreen cognitive testing

Disclosure: Spinogenx, Advisory Board, Self

P496. Exercise intervention linked to altered social cognitive brain responses in autistic adolescents

Kelly Cosgrove, Anika Guha, Keith Dodd, Maureen McHugo, Judy Reaven, Robin Gabriels, Susan Hepburn, Andrew Novick, Fulvia Castelli, Jason Tregellas, Kristina Legget

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

Background: Autism is characterized by social cognitive challenges, including difficulties with theory of mind (ToM), or the ability to understand the mental states of others. Such challenges can contribute to reduced quality of life. Functional magnetic resonance imaging (fMRI) studies indicate that autistic individuals often exhibit atypical function in multiple brain networks across development, including the default mode, salience, and executive control networks. These systems support social understanding, detection of socially relevant cues, and regulation of goal-directed responses in social contexts, respectively. Exercise is a promising intervention for improving social cognitive function, as it may alter neuronal activity in the aforementioned networks and improve prosocial behaviors and executive function. The effects of exercise on social cognition and associated neurobiology in autistic adolescents have not yet been examined, however. To address this knowledge gap, this pilot study determined if a 10-week exercise intervention, compared to a gaming control condition, was associated with changes in neuronal responses during a ToM task.

Methods: Thirty-eight sedentary autistic adolescents (15.8 ± 2.1 years old; 24% female; physical activity < 60 minutes per week; without co-occurring intellectual disability) were randomized to a 10-week exercise program (n = 26) or a time- and attention-matched gaming control condition (n = 12; clinicaltrials.gov identifier: NCT02455193). Participants in the exercise group engaged in preferred aerobic activities at a moderate intensity (50–70% maximum heart rate) for at least 30 minutes, three times per week. They completed one exercise session in the lab with research staff approximately every two weeks during intervention. Those in the gaming group came into the lab at the same frequency, but played non-exerting games with staff. At baseline and post-intervention, all participants completed fMRI during the New False Belief Animation Task (Castelli et al., 2000; Castelli et al., 2002; Isernia et al., 2023), in which they viewed short video clips of two triangles either moving around the screen randomly (“random”) or interacting in ways that implied social intent (e.g., trying to trick or surprise each other; “sneaky”). After each clip, participants indicated which interaction type was being depicted (e.g., random, sneaky). The task involved one ~11-minute session and included 14 sneaky and 14 random trials lasting ~12 seconds each. Group differences (exercise vs. gaming) in intervention-associated effects on neuronal response for sneaky > random trials were assessed using linear mixed-effects models within a probabilistic grey matter mask. Behavioral performance on the ToM task was indexed by mentalizing scores (proportion of correctly categorized clips).

Results: A significant group*session interaction effect was observed in the right dorsolateral prefrontal cortex (dlPFC; cluster-corrected p < 0.005). Post hoc analyses of estimated marginal means indicated significantly increased right dlPFC response from baseline to post-intervention in the exercise group (t(28) = 3.50, Tukey-adjusted p = 0.008, Cohen’s d = 0.65), with decreased response in the control group (t(29) = −3.50, Tukey-adjusted p = 0.008, Cohen’s d = −0.64). Similar subthreshold patterns were observed in the left dlPFC, superior frontal gyrus, putamen, and anterior insula (uncorrected ps < 0.005). No significant effects were observed for mentalizing scores (ps > 0.320).

Conclusions: The 10-week moderate-intensity exercise intervention was associated with increased dlPFC response during a ToM task in autistic adolescents, whereas the gaming control group showed decreased response. The exercise-associated increases may reflect enhanced engagement of executive control systems that support social-cognitive processing. Although behavioral ToM performance did not change for either group, these preliminary findings suggest that exercise may serve as an accessible, low-cost approach for modulating social cognitive neurocircuitry in autistic youth. Longer or more intensive interventions may be required to translate these neurobiological effects into measurable behavioral improvements.

Keywords: Autism, Physical exercise, Functional MRI (fMRI), Social cognition, Dorsolateral prefrontal cortex (DLPFC)

Disclosure: Nothing to disclose.

P497. Retinoic acid pathway controls hippocampal subfield specification

Zhiping Shao, Faith De Kuyper, Panos Roussos, Deepak Kaji

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

Background: Hippocampal dysfunction plays a critical role in neurodegenerative diseases like Alzheimer’s disease and schizophrenia, where the CA1 region, in particular, exhibits unique vulnerabilities in these disease states. Organoids derived from human induced pluripotent stem cells (hiPSCs) have emerged as powerful tools for modeling both development and disease in the brain. Unfortunately, these models fail to capture the complexity of the hippocampus, especially the diverse cellular composition and organization of its subfields (i.e. the DG, CA3, CA2, and CA1). Current hippocampal organoid protocols generate DG and CA3-like cells in a disorganized manner, with a notable absence of CA1 and CA2 cells. These limitations suggest an incomplete understanding of the signaling pathways that regulate hippocampal development and subfield specification. By combining developmental insights from the mouse, mining human sequencing atlases of the adult and fetal hippocampus, and mechanistic experiments with human organoids, our team discovered that the retinoic acid pathway controls critical fate specification not only between the hippocampal subfields but also between the hippocampus and the neighboring neocortex.

Methods: hiPSC and embryonic stem cells (ESC) from 4 donors (called ReproCELL 802-3G, MSN25, MSN8, and H9) were obtained from the ReproCELL vendor or the Mount Sinai Stem Cell Core. Pluripotent stem cels were differentiated towards the hippocampal fate by using previously described protocols for generating dorsal forebrain and modified to include Wnt and BMP agonism from day 18 to d21 as per the original hippocampal organoid differentiation. However, newer protocols for organoids such as our dorsal forebrain protocol include media formulation improvements that improve organoid health. Organoids were then exposed to all-trans retinoic acid (ATRA), AGN193189 (an inverse agonist of the retinoic acid pathway), or our base condition from d18 until d90 when the differentiation ends.

Results: We find that the DG markers PROX1, SGCZ, and SEMA5A as well as the CA3 markers GRIK4 (KA1) and CFAP299 exhibit significant upregulation in the ATRA condition relative to the base or AGN193189 conditions at day 60 by qPCR (p < 0.05). Interestingly, at day 90, DG marker expression is almost completely abolished in the ATRA condition but CA3 markers remain upregulated in the ATRA condition compared to the AGN193189 condition. Conversely, the CA1 marker POU3F1 is 10-fold higher in the AGN193189 condition compared to the ATRA condition at days 60 and d90 (p < 0.05). Interestingly, TBR1, a layer 6 neocortical marker, was significantly upregulated by ~1000 fold, 3000 fold, and 6000 fold at days 30, day 60, and day 90 respectively in our AGN193189 condition relative to our ATRA condition (p < 0.05).

Conclusions: Retinol, retinaldehyde, and retinal acetate containing serums are a hallmark of most organoid differentiations and can be converted to all-trans retinoic acid. Our data suggests that ATRA may be a potent inducer of a DG/CA3 bipotent progenitor. Interestingly, prolonged use of ATRA resulted in a loss of DG regulators between day 60 and day 90, suggesting that manipulation of RA may require carefully timed intervention that may involve the alternating use of retinoic acid agonism and inverse agonism. The inverse agonist was successfully in generating higher levels of CA1 and neocortical markers. This may suggest that the CA1 region and neocortex come from a similar developmental origin. Somewhat paradoxically, retinoic acid is known as a prefrontalizer of the neocortex. Again, our data suggests that during timepoints that correspond with radial glia specification, inverse agonism encourages neocortical specification and leaves the possibility for agonism to play a role in prefrontalization of neurons at later time points. Future experiments will attempt to disentangle the way that the Wnt and RA pathways interact at various points in time and on unique cell types in the cortical developmental tree.

Keywords: Brain organoids, Induced pluripotent stem cells (iPSCs), Hippocampus subfields

Disclosure: Nothing to disclose.

P498. Multicohort genome-wide association studies of resting state functional connectivity phenotypes in early infancy

Ann Alex, Haitao Chen, Nicholas Panchy, Jacob Zieba, Prarthna Mudumala, Claudia Buss, Elysia Davis, Kirsty Donald, David Edwards, John Gilmore, Jessica Girault, Benjamin Hankin, Pilyoung Kim, Joseph Piven, Jerod M. Rasmussen, Dan Stein, Paul Thompson, Martin Styner, Wei Gao, Gustavo de los Campos, Rebecca Knickmeyer, on behalf of ENIGMA-ORIGINs working group

Michigan State University, East Lansing, Michigan, United States

Background: Information transfer within and between brain networks is essential to cognition and behavior and may be disrupted in psychiatric conditions. rsfMRI (resting-state functional MRI) can provide information about brain connectivity. Recently, genome-wide association studies (GWAS) revealed over 600 significant locus–trait associations for rsfMRI traits in adults. However, we do not know when in development these associations arise. While adult-like functional network features are apparent in the neonatal brain, no GWAS of rsfMRI traits has been conducted in infancy. The current study addresses this gap.

Methods: This multi-cohort, mega-analytic genetic association study included individuals between 0 and 3 months of age in the ORIGINs dataset with high quality genetic and rsfMRI data as of 05/21/2025 (N = 1000, 503 males, 497 females). Univariate GWAS were run for 102 traits including two phenotypes derived from graph theory (global efficiency and modularity), measures of functional connectivity within and between 13 brain networks, and nine summary measures based on the following clusters: primary cortical networks, socioemotional networks, and cognitive networks.

Results: The mean heritability estimate across all 102 traits was 0.05 with a range of 0 to 0.48. Global efficiency was the most heritable phenotype (p = 0.000118 for GCTA). Further, the left frontal-parietal network was involved in several of the more heritable phenotypes. The univariate GWAS detected one locus including significant variants at p  <  4.9 × 10−10 for connectivity between the auditory and right frontal-parietal networks. Follow-up analyses in FUMA suggest the genes EXT2 and ACCS may play a role in this association. We also detected 8 additional loci including variants above the threshold for suggestive associations (5 × 10⁻⁸).

Conclusions: Our results suggest common variants contribute to intrinsic brain connectivity in early infancy. Similar to findings in adults, phenotypes including frontal-parietal connectivity, which is related to executive function, were moderately heritable. However, individual variants were most related to phenotypes including primary motor and sensory networks. Future analyses will determine if resting-state-associated traits in infancy correlate with neurodevelopmental conditions, psychiatric disorders, and/or cognitive/behavioral traits.

Keywords: Infant fMRI, GWAS, Resting State Functional Connectivity

Disclosure: Nothing to disclose.

P499. Behavioral and molecular effects of hormone dysregulation in early development in wild-type mice

Pravda Quinones-Labernik, Emily Hagan, Anthony Lange, Danielle Preuschl, Charlotte Tesar, Ted Abel, Sarah Ferri

University of Iowa Carver College of Medicine, Iowa City, Iowa, United States

Background: Neurodevelopmental disorders (NDDs) are complex conditions that affect 1 in 6 children and exhibit sex differences in onset, prevalence, and symptom presentation. One early developmental factor likely contributing to these differences is the male gonadal testosterone surge. This critical period of development occurs in mid-gestation in humans and around birth in mice and results in sex-specific neural circuit organization. Exposure to increased testosterone occurs in a number of conditions associated with increased risk of NDD diagnosis in offspring, including preeclampsia, maternal stress, PCOS, gestational diabetes, and hypoxia. We have recently shown that a single exposure to excess testosterone during this critical period of brain development, but not shortly after, induces social and fear memory deficits in juvenile and adult wild-type male mice. Our current experiments are investigating the mechanisms underlying sex-specific developmental vulnerability to hormone dysregulation-induced neurodevelopmental deficits.

Methods: We administered a brain-masculinizing dose of testosterone (100 µg. s.c.) to wild-type C57/BL6 mice on their day of birth (P0) to investigate behavioral and molecular effects at multiple timepoints. Behavioral tests included assays of physical and neuromuscular reflex developmental milestones including growth, eye opening, fur development, negative geotaxis, cliff avoidance, grasp reflex, ultrasonic vocalizations during dam separation, markers of onset of puberty, and marble burying, rotarod, and object location memory in adults. Molecularly, we assayed protein and gene expression of synaptic markers, glutamatergic signaling, and glia using immunohistochemistry, RNAScope, and RNA-Seq in limbic brain areas.

Results: We have found that both male and female pups treated neonatally with testosterone have delayed eye opening, fur development, pinnae detachment, cliff avoidance, and grasp reflex. Excess neonatal testosterone also caused male-specific deficits in ultrasonic vocalizations at PN5 and in rotarod and object location memory in adults. Juvenile males exposed to excess neonatal testosterone exhibited decreased microglia, increased reactive astrocytes, and increased PSD-95 in the dorsal hippocampus and decreased expression of genes related to glutamatergic signaling in the basolateral amygdala compared to controls.

Conclusions: Hormone dysregulation during an early critical period of brain development induces sex-specific and long-lasting effects on behavior and limbic brain regions.

Keywords: Neurodevelopmental Disorders, Sex differences, Gonadal Hormones

Disclosure: Nothing to disclose.

P500. Pharmacological screens identify novel suppressors of autism gene-associated behavioral phenotypes in zebrafish

Priyanka Jamadagni, Yi Dai, Yunqing Liu, Hellen Weinschutz Mendes, April Pruitt, Gang Xu, Yihan Liu, Yulia Surovtseva, Kristen Brennand, Zuoheng Wang, Ellen Hoffman

Yale University, New Haven, Connecticut, United States

Background: Over 200 genes are strongly associated with autism spectrum disorders (ASDs), yet advancing from large-effect ASD genes to actionable biological mechanisms and pharmacological candidates remains a central challenge. This is due in part to our limited understanding of the underlying biology of ASDs as well as their considerable clinical and genetic heterogeneity. For this reason, many new drugs under investigation for ASDs fail to demonstrate efficacy in clinical trials, underscoring the need to further parse out the heterogeneity across ASDs. One approach to address this heterogeneity involves stratifying ASD genes based on shared phenotypes and identifying pharmacological targets relevant to subgroups of ASD genes, consistent with a precision or personalized medicine approach. In addition, ASDs encompass a range of clinical manifestations beyond social and communication deficits and restrictive, repetitive behaviors, including sleep and sensory processing deficits, yet there are no FDA-approved drugs that target these basic arousal or sensory sensitivities in ASDs. To address these challenges, we are in need of novel drug screening pipelines that allow us to: (i) predict pharmacological pathways relevant to multiple ASD genes and gene subgroups, informed by precision medicine; (ii) interrogate the effects of compounds at the whole-organism level, obtaining a readout of drug effects on basic arousal and sensory processing behaviors; and (iii) uncover novel pharmacological suppressors relevant to specific ASD genes. Zebrafish offer multiple advantages as a highly scalable, in vivo vertebrate system that has already been leveraged by our group and others to identify novel pharmacological candidates relevant to ASDs, epilepsy, and Alzheimer’s disease.

Methods: We perform pharmaco-behavioral screens of 774 U.S. FDA-approved drugs using large-scale automated assays of visual-startle and sleep-wake behaviors in larval zebrafish at 5–7 days post fertilization. These assays provide a readout of drug effects on basic sensory processing and arousal circuits and have been used by our group and others to identify behavioral phenotypes in multiple zebrafish ASD gene mutants. We selected a compound library that spans a range of clinical indications, including anti-neoplastic, anti-diabetic, and antibiotic agents, as well as psychotropic compounds. Individual larvae were exposed to each drug at 10 μM dissolved in DMSO or DMSO alone in a 96-well plate, after which their responses to visual-startle (lights-off or lights-on) stimuli and their locomotor activity during a 14h:10h light:dark cycle were measured, generating a behavioral fingerprint of 24 visual-startle and sleep-wake parameters for each drug. In total, we generated more than 15,000 behavioral profiles across all drugs, representing short- and long-term effects of drug exposure on visual-startle and sleep-wake behaviors, respectively. Using this pharmaco-behavioral database, we identified compounds with behavioral profiles that correlate or anti-correlate with the behavioral phenotypes of zebrafish mutants of nine large-effect ASD genes. To identify suppressors, we conducted targeted screens of anti-correlating drugs in zebrafish mutants of two ASD genes, SCN1A/SCN2A and DYRK1A, which display robust visual-startle and sleep-wake phenotypes.

Results: We establish a pharmaco-behavioral database and searchable website of 774 U.S. FDA-approved drugs on basic arousal and sensory processing behaviors in larval zebrafish. Using automated assays, we identify unique behavioral fingerprints for 520 of the 774 compounds that have significant effects on at least two parameters and are non-toxic, and demonstrate that drugs with shared molecular targets display highly correlated and conserved behaviors in zebrafish and mammals. By mapping the behavioral profiles of zebrafish mutants of nine large-effect ASD genes onto the drug database, we identify significant enrichment of pharmacological mechanisms that anti-correlate with each ASD gene and subgroups of ASD genes with shared behavioral phenotypes informed by a precision medicine approach. In targeted screens in zebrafish scn1lab and dyrk1a mutants, we show that both mutants display the most differential responses to compounds targeting serotonergic, dopaminergic, and retinoic acid signaling, revealing convergent dysregulated pathways. By screening compounds that anti-correlate with each mutant behavioral profile, we identify three top suppressors of the behavioral phenotypes of scn1lab and dyrk1a mutants, nominating new pharmacological candidates and molecular pathways for further investigation.

Conclusions: This study establishes a novel pharmaco-behavioral database and searchable website of 774 U.S. FDA-approved compounds with translational relevance for drug discovery. By integrating large-scale genetic and pharmacological screens, we identify dysregulated pathways associated with specific ASD genes. Furthermore, we uncover novel suppressors of the behavioral phenotypes of two ASD gene mutants, nominating new pharmacological targets for further evaluation.

Keywords: Autism Spectrum Disorder (ASD), Human Genetics, Zebrafish, Drug Discovery—new approaches, Pharmacology

Disclosure: Nothing to disclose.

P501. Developmental changes in the expression of Teneurin-4 in mice brain

Tomoya Kaigawa, Jun Yokose, Takako Kikkawa, Yusuke Yano, Shin-ichi Muramatsu, Noriko Osumi, Atsumi Nitta

Faculty of Pharmaceutical Sciences, University of Toyama, Toyama, Japan, Toyama, Japan

Background: In recent years, the number of patients with psychiatric disorders has markedly increased, exceeding one billion worldwide. With the progression of super-aged and stress-laden societies, the prevalence of psychiatric disorders such as dementia and depression is expected to rise further. Various genomic analyses in clinical studies have identified numerous genetic variants that increase the risk of developing psychiatric disorders such as schizophrenia, autism spectrum disorder, and bipolar disorder. Genome-wide association studies (GWAS) in bipolar disorder have reported mutations in Odz4. The gene Odz4 encodes Teneurin-4 (Tenm4), which is distributed in the central nervous system. Tenm4 has been reported to promote oligodendrocyte differentiation and myelination through enhancing cell adhesion. We have shown that knockdown of Tenm4 in the adult mouse brain induces cognitive impairment and depression-like behavior. Although the function of Tenm4 in an adult brain has been increasingly elucidated, its role during brain development and in early life remains unclear. Therefore, in this study, I investigated the expression levels and developmental distribution of Tenm4 during embryonic and early postnatal stages in mice brain.

Methods: We measured Tenm4 mRNA levels during development using quantitative real-time RT-PCR. The measurements were performed at embryonic day (E) 14 and E18, as well as postnatal day (P) 4, P7, P10, P14, P21, P28, and P56. In addition, we analyzed the distribution of Tenm4 mRNA by in situ hybridization in the brains of C57BL/6J mice at P56 and E14. Probes were generated from a mouse brain cDNA library by cloning a partial sequence of Odz4 cDNA and digesting with restriction enzymes, followed by synthesis of Tenm4-specific antisense probes and sense probes as negative controls.

Results: Real-time RT-PCR revealed that Tenm4 mRNA expression decreased from E14 to P7 and thereafter remained constant. In situ hybridization methods demonstrated that Tenm4 mRNA was broadly expressed throughout the brain in both adult and embryonic stages, with particularly was found strong signal in the hippocampus. Currently, based on the expression changes identified by RT-PCR, we are knocking down Tenm4 at postnatal day 2 and evaluating its effects using behavioral tests.

Conclusions: This study is the first to characterize the developmental expression levels and spatial distribution of Tenm4 mRNA. Based on these findings, we are generating early-life Tenm4 knockdown mice and plan to evaluate the functional roles of Tenm4 during development using behavioral assays.

Keywords: Teneurin-4, Brain development, Myelination

Disclosure: Nothing to disclose.

P502. Effects of a novel social conditioning paradigm in BTBR T+ Itpr3tf/J mice, modeling early intensive behavioral intervention in a mouse model of ASD

David Beversdorf, Brianna Becher, Felicia Padilla, Nick Ahmed, Olivia Smith, Todd Schachtman, Matthew Will

University of Missouri, Columbia, Missouri, United States

Background: Autism Spectrum Disorder (ASD) is characterized by difficulties with social interaction, as well as repetitive and restricted patterns of behavior. Applied Behavior Analysis (ABA) therapy, which uses principles of reinforcement to shape behavior, remains one of the most empirically supported approaches for improving functional independence in individuals with ASD. However, despite its widespread use, responses to ABA-based interventions vary widely, suggesting that individual factors—such as ASD etiology—may influence treatment effectiveness. Therefore, greater understanding of the effects of ABA-based interventions needs to be examined across multiple known ASD etiologies to understand this effect. As an initial step towards this, our study aimed to model an ABA-style intervention in one well characterized ASD mouse model to allow for further exploration of which etiologies, and therefore, which individuals might benefit most.

Methods: We developed a novel social conditioning paradigm designed to model one of the most common techniques in ABA therapy: pairing a desirable behavior with a rewarding stimulus. This paradigm was tested in BTBR T+ Itpr3tf/J (BTBR) mice, a well-established idiopathic mouse model of ASD known for its limited social interaction. We hypothesized that our conditioning paradigm would increase sociability in BTBR mice. Additionally, we wanted to examine any potential sex differences in response to the intervention, given the well-documented variability in ASD presentation between males and females. Beginning at postnatal day 21, mice were divided into four groups by sex and treatment status: male control, female control, male conditioned, and female conditioned. All mice underwent a two-day habituation period, and conditioning began on postnatal day 23. For ten days, conditioned mice were exposed to a novel conspecific and immediately received a palatable food reward (chocolate-flavored Ensure), while control mice experienced the same social interaction but received the reward two hours later, preventing an association between the stimuli. Sociability was assessed using the 3-chamber sociability assay following the 10-day conditioning period, starting with a social interest phase, assessing interest in a mouse in a chamber compared to an empty chamber, followed by a social novelty phase, assessing interest in a novel mouse in a chamber compared to a familiar mouse in a chamber.

Results: Social interaction levels during the conditioning period did not show any consistent patterns or learning curves across the 10 days. In the 3-chamber assay, no significant group differences emerged during the social interest phase. However, during the social novelty phase, conditioned female mice spent more time interacting with a novel conspecific than female controls. No differences were observed between male groups.

Conclusions: These findings suggest that an ABA-style conditioning approach may enhance social novelty-seeking in female BTBR mice, highlighting the importance of sex-specific considerations when evaluating treatment outcomes in ASD and calling for replication in other ASD models. While response to ABA does not appear to differ between males and females in the clinical population, it will be important to determine whether sex specificity of effects varies across different ASD subtypes. This preclinical modeling approach will also begin to allow better understanding of the biological mechanisms of ABA-style conditioning across a range of models going forward, allowing development of personalized medicine approaches for therapeutic interventions.

Keywords: Autism Spectrum Disorder (ASD), Behavioral Model, Social Interactions

Disclosure: Nothing to disclose.

P503. Reduced placental production of IGF-1 causes persistent brain and behavior changes in mice relevant to neurodevelopmental disorders

Annemarie Carver, Faith Fairbairn, Robert Taylor, Njenga Kamau, Amrita Gajmer, Hanna Stevens

University of Iowa Carver College of Medicine, Iowa City, Iowa, United States

Background: Fetal brain growth and development depends on placental functioning, but how the provision of specific hormones by the placenta influence these critical stages and persistently influence brain and behavior is poorly understood. Insulin-like growth factor 1 (IGF-1) is primarily produced in the placenta in utero and is required for fetal neurogenesis and embryonic neural cell differentiation. In disorders of placental dysfunction, the higher risk of neurodevelopmental disorders (NDDs) in offspring may be induced by insufficient IGF-1. Therefore, we sought to model this reduced placental production of IGF-1 and examine fetal brain processes, subsequent NDD-relevant behaviors, and postnatal brain structure to determine the critical nature of this placenta-brain molecular communication. In our approach, we also sought to provide advancements for how mammalian placenta could be targeted with CRISPR technology.

Methods: To specifically change placental hormone production, we created a partial knockout of placental Igf-1 using a placental-targeted CRISPR manipulation (knock-out or control plasmid) on embryonic day 12 (E12) in pregnant CD-1 mouse dams. With this method, we created over 60 each knock-out and control offspring independently manipulated from over 25 pregnant dams, with about equal distribution across male and female offspring, evaluated for outcomes at different embryonic and postnatal ages. We utilized qPCR to assess placenta at E13, immunohistochemistry to evaluate the brain at E14, and RNA sequencing to assess the brain at E18. For some litters which dams gave birth to and raised, we evaluated NDD-relevant behaviors including open field, rotarod, water T-maze, and stereotypies at juvenile and adult time points and assessed adult brain structure and cell populations. Statistical analyses of brain structure and behavior were performed separately by sex with logistic regression controlling for litter in R. RNA sequencing analysis was performed using DESeq2, with false discovery rate adjustment to address multiple comparisons. Differentially expressed genes were additionally assessed for enriched pathways and gene sets relevant to NDDs.

Results: With an approximate 30% reduction of placental Igf-1 in both males and females (n = 7,9,11,7; p < 0.05, p < 0.05), preliminary data showed a decrease in E14 forebrain volume in only placental Igf1KO females (n = 6,9,8,10; p < 0.05). RNA sequencing data (n = 6 per condition per sex) from E18 ventral forebrain demonstrated sex-specific functional changes, with additional analyses ongoing.

In postnatal offspring (n = 12–13 per condition per sex), both male and female juveniles displayed increased locomotor activity (p = 0.06, p < 0.05) with placental Igf1KO but only male Igf1KO juveniles showed impaired motor learning (p < 0.01). Adult males displayed increased stereotypies (p < 0.05) after placental Igf1KO while females displayed decreased stereotypies (p < 0.05) and deficits in procedural learning (p < 0.05) which males did not. NDD-relevant brain morphological analysis (n = 9–10 per condition per sex) determined that no volume changes were apparent across offspring groups but that males specifically had a decreased total neuronal population in the striatum in Igf1KO offspring compared to controls (p < 0.05).

Conclusions: Overall, these findings demonstrate that reduced placental Igf-1 results in brain and behavioral changes in mice relevant to NDDs which are highly sex specific. Clinically, reduced placental sourcing of IGF-1 may occur with preterm birth or placental insufficiency, suggesting that the shifts in striatal development shown here may inform the increased NDD etiology that arises in these contexts.

Keywords: IGF-1, Placenta, Mouse, CRISPR/Cas9

Disclosure: Nothing to disclose.

P504. Prenatal choline supplementation rescues learning and memory deficits in mice exposed to elevated kynurenic acid during neurodevelopment

Ana Pocivavsek, Sarah Beggiato, P. Leon Brown, Corbin Witt, Snezana Milosavljevic, Robert Schwarcz

Maryland Psychiatric Research Center, Baltimore, Maryland, United States

Background: Elevated brain kynurenic acid (KYNA) levels are implicated in the pathophysiology of schizophrenia (SZ). As an antagonist of cholinergic and glutamatergic neurotransmission, elevated KYNA levels may be causally related to cognitive dysfunctions in SZ. We developed an embryonic kynurenine model (“EKyn”; Beggiato et al., 2018) in mice to mimic elevated KYNA levels during neurodevelopment and study cognitive function in adulthood. We presently investigated if perinatal supplementation of choline (Freedman et al., 2022), an essential nutrient and agonist of α7nACh receptors, would alleviate adverse learning outcomes in EKyn adult offspring.

Methods: Pregnant C57Bl/6J mice were fed a standard diet (0.1% choline chloride) or choline-supplemented diet (0.5% choline chloride) from embryonic day (ED) 11 to postnatal day (PD) 21. From ED 11 to ED 18, each diet remained control (ECon) or was laced with the KYNA precursor kynurenine (50 mg/day; EKyn). Upon weaning, all offspring received control rodent chow until experimental testing in adulthood (PD 56–85). Separate cohorts of adult male and female offspring were then assessed in biochemical (in vivo microdialysis in prefrontal cortex; N = 10 per group), electrophysiology (ex vivo recordings of interhemispheric transmission; N = 17–22 per group), and behavioral studies (Barnes maze; N = 13–30 per group).

Results: EKyn male and female offspring had significantly higher KYNA levels in the prefrontal cortex in adulthood (64% increase; t = 3.483, Df = 18, P = 0.03) and significantly delated interhemispheric transmission (F1,74 = 16.11, P = 0.0001). EKyn exposure significantly impacted spatial learning such that latency to find the escape box (males: F1,69 = 5.613, P = 0.02; females: F1,67 = 3.785, P = 0.05) and errors (males: F1,70 = 6.958, P = 0.01; females: F1,67 = 6.591, P = 0.01) were increased in the Barnes maze across days. On the last day, latency was significantly higher in EKyn males (P = 0.001) and females (P = 0.002), and this impairment was reversed in offspring of mothers that received choline-supplemented diet with EKyn exposure (males: P = 0.005;, P = 0.02; females: P = 0.05). Taken together, EKyn impaired Barnes maze performance in both sexes, but choline reversed these deficits, suggesting rescue of hippocampal-dependent learning. Analysis of Barnes maze behavior revealed significant difference in the use of random, serial, or direct search strategies across learning days between both sexes of ECon and EKyn (P < 0.05).

Conclusions: Physiological and behavioral abnormalities are seen in adult EKyn mice with concurrent elevations in brain KYNA levels, indicating that prolonged treatment with kynurenine during the embryonic period causes permanent, functionally significant changes in brain development in these experimental animals. Understanding the role of choline supplementation, which mitigated cognitive impairments in EKyn offspring, as a potential intervention for alleviating prenatal risk factors associated with cognitive dysfunction and SZ helps to inform future treatment strategies.

Keywords: Prenatal, Kynurenine, Kynurenic acid, Choline, Cognitive deficits

Disclosure: Kynexis BV, Advisory Board, Self, Kynexis BV, Founder, Self

P505. Deciphering persistent prefrontal cortex kynurenic acid alterations and molecular shifts in adult mice following adolescent THC exposure

Francesca Frescura, Edoardo Tiziani, Luca Ferraro, Robert Schwarcz, Sarah Beggiato

University of Ferrara, Ferrara, Italy

Background: Adolescent THC use represents a risk factor for cognitive deficits in psychiatric disorders emerging in adulthood among individuals initiating cannabis consumption during adolescence (D’Souza et al., 2020). Importantly, adolescent THC exposure induces severe persistent cognitive impairments and disrupts glutamatergic and GABAergic neurotransmission in the rat prefrontal cortex (PFC; Zamberletti et al., 2014). These neurochemical changes correlate with a sustained long-term elevation of kynurenic acid (KYNA) in the PFC, a key tryptophan degradation metabolite via the kynurenine pathway (KP), believed to be implicated in the developmental pathogenesis of several psychiatric disorders, including schizophrenia (SZ). Because not all cannabis users display the harmful THC-induced impairments, it is likely that certain individuals possess underlying pre-existing genetic susceptibilities susceptible to THC exposure (Levine et al., 2017; Hasin, 2018). Mice with a targeted knockout of kynurenine 3-monooxygenase (KMO), a KP enzyme indirectly regulating KYNA synthesis (Giorgini et al., 2013), exhibit elevated KYNA concentrations relative to wild-type (WT) counterparts, thereby offering a potent developmental model to investigate the specific contributions of PFC KYNA signaling in the dysfunction induced by the exposure to THC during adolescence.

Methods: Male and female C57BL6/J (a well characterized mouse strain with respect to the behavioral effects of THC; Wise et al., 2011) WT (i.e. Kmo + /+; N = 12, 6 males, 6 females) and HET (i.e. Kmo + /-; N = 12, 6 males, 6 females) were chronically treated with vehicle or ascending doses of THC (2.5 mg/kg; 5 mg/kg; 10 mg/kg i.p.), twice a day from postnatal day (PD) 35 to 45 (see Zamberletti et al., 2014 for protocol validation). Extracellular KYNA and GABA levels have been measured by in vivo microdialysis in the PFC of adult (PND 75) mice. On day 1 after the surgery, after three stable basal values were obtained, the effects of a challenge of L-kynurenine (i.e. the biological precursor of KYNA (10 mg/kg i.p) mimicking a translationally relevant “second hit”, on extracellular KYNA and GABA levels were measured.

Results: In vivo microdialysis showed that basal PFC KYNA levels were higher in both male and female HET mice (P < 0.05), while basal PFC GABA was elevated only in female HET (P < 0.05) compared to WT animals. Adolescent THC produced a sustained rise in PFC KYNA in male WT mice (P < 0.05), without affecting GABA levels. L-kynurenine (10 mg/kg, i.p.) significantly increased PFC KYNA levels in all groups (P < 0.01), with a smaller increase in HET than WT (P < 0.05) and was not affected by adolescent THC treatment. L-kynurenine did not alter PFC GABA, although THC-treated HET showed a trend to a decrease. In addition, prepulse inhibition is being assessed in HET and WT mice exposed to adolescent THC, molecular profiling via RT-PCR is under way, and preliminary proximity ligation assays are being conducted to probe potential protein interactions.

Conclusions: The present study revealed that the use of cannabinoids early in life can induce a malfunction of the KP in male, but not female, WT mice. This observation mirrors earlier findings reported in male rat models. Moreover, elevated cerebral KYNA following targeted KMO downregulation correlates with distinct altered sensitivity to THC-driven shifts in prefrontal cortex KYNA and GABA concentrations across HET and WT mice. Elaboration of the mechanisms underlying these changes can be expected to provide useful information regarding the etiology of adolescent THC-induced risk of neuropsychiatric illnesses in adulthood.

Keywords: Cannabinoids, Kynurenine pathway, GABA

Disclosure: Nothing to disclose.

P506. Differential utilization of hippocampal learning strategies as a source of individual differences within wild-type rats and those with a mutation to an ASD risk gene

Christina Lee, Nicole Yokotoa, Cristofer Holobetz, Greer Williams, Mihika Mehra, Ali Chaudhry, Jane Ton, Peter Dayan, David Kastner

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

Background: Learning is a dynamic process that varies widely across individuals, often involving distinct strategies. To understand this variability, we conducted two analyses of the behavior of rats performing a task requiring spatial alternation among specific triplets of arms in an environment containing six arms in a row.

Methods: We measure large-scale behavior in rats with bilateral hippocampal lesions and utilize a data-driven behavioral analysis method called the choice-wide behavioral association study (CBAS). CBAS breaks behavior down into all sequences of choices of the animals and identifies those sequences that differ between groups. We also measure large-scale behavior in a separate population of wild-type (WT) and Scn2a haploinsufficient (Scn2a + /-) rats. Scn2a expresses a sodium channel throughout the brain and is a high confidence, large risk, ASD gene.

Results: In the first analysis, we compared rats with bilateral hippocampal lesions (n = 50) to rats that underwent control surgery (n = 55). Using CBAS, we identify learning strategies that were distinguished by sequences of arm visits that differed between the two groups. Control rats more frequently moved in consistent directions across the arms of track, progressively narrowing down their choices to visit just the correct arms.

This group-level contrast defines partial strategies of control- and lesion-associated sequences of arm visits. However, individual learning trajectories were very variable. To test the hypothesis that individual WT animals might rely to different degrees on their hippocampus to perform the task, we conducted a second analysis on a larger cohort of WT rats (n > 200). Consistent with the hypothesis, individuals varied in their use of control- and lesion-associated sequences, with some WT rats exclusively favoring lesion-associated choices, despite having an intact hippocampus.

If differential hippocampal utilization is a source of inter-animal variability, it could also be modulated by psychiatric risk genes as a source of pathogenicity. We found that rats haploinsufficient for Scn2a (n > 200), a high-confidence autism spectrum disorder (ASD) risk gene, are more biased toward lesion-associated strategies than WT rats, and therefore potentially less reliant on the hippocampus.

Conclusions: Together, these findings suggest that the hippocampus supports one or more specific learning strategies, but that individual animals can vary in their reliance on these strategies. Characterizing learning at the individual animal level may yield a better understanding of the neural basis of flexible and impaired learning.

Keywords: Behavioral Phenotyping, ASD, Individual differences, Memory and Learning, Hippocampus

Disclosure: Nothing to disclose.

P507. Characterizing longitudinal patterns of central and peripheral insulin resistance

Audrey Evers, Fahim Abbasi, Katie Watson Lin, Erez Eitan, Natalie Rasgon

Icahn School of Medicine at Mount Sinai, Rockefeller University, New York, New York, United States

Background: Metabolic dysfunction of insulin resistance (IR) may compromise brain function decades before overt disease. We assessed biomarkers of peripheral and central IR for three years in 125 cognitively intact adults without diabetes or depression (23–61 y) to clarify their relative contributions to trajectories of metabolic and cognitive change.

Methods: Peripheral IR was quantified by measuring steady-state plasma glucose (SSPG) concentration during the insulin-suppression test and ancillary metabolic indices; central IR was indexed by phosphorylated insulin-receptor-substrate-1 (p-IRS1) in neuron-derived extracellular vesicles (NDEs) isolated from plasma.

Results: Mixed-effects models controlling for age, sex, and BMI showed that higher baseline SSPG concentration robustly predicted subsequent increases in HOMA-IR and fasting insulin but was unrelated to cognitive outcomes. In contrast, higher baseline NDE p-IRS1 predicted better global cognition (MMSE) over time, an effect strongest in younger participants, yet showed no association with peripheral metabolic change. Participants with concordantly low fasting insulin and C-peptide maintained roughly two-fold higher p-IRS1 than those in the highest tertiles, despite similar trajectories across visits. Neither peripheral nor central IR predicted telomere attrition. Cross-sectionally, p-IRS1 correlated inversely with SSPG concentration, BMI, and leptin, suggesting compensatory brain–periphery coupling.

Conclusions: These findings indicate that NDE-based markers capture a dimension of brain metabolic vulnerability distinct from classical peripheral measures.

Keywords: Insulin resistance, Neuron-derived extracellular vesicles, Cognitive aging, Phosphorylated IRS1

Disclosure: Nothing to disclose.

P508. Effect of starting on oral hormonal contraception on serotonin 4 brain receptor binding and verbal memory: a randomized, placebo-controlled molecular brain imaging trial in healthy women

Søren V. Larsen, Anna Kauffmann, Mette C. Hochheim, Sofie Hvitved, Annika L. Rasmussen, Mette Clausen, Sif C. Olsen, Kristian R. Jensen, Sune Keller, Claus Svarer, Brice Ozenne, Øjvind Lidegaard, Annamaria Giraldi, Vibeke H. Dam, Vibe Frokjaer

Neurobiology Research Unit, University Hospital Rigshospitalet, Copenhagen, Denmark

Background: Oral contraceptive (OC) use has been associated with an increased risk of developing depression [Skovlund CW, et. al. JAMA Psychiatry 2016]. This may be related to a compromised serotonergic neurotransmission as suggested by our previous cross-sectional work; OC use was associated with 9–12% lower cerebral serotonin 4 receptor (5-HT4R) non-displaceable binding potential (BPND) as measured with Positron Emission Tomography (PET) [Larsen SV, et. al. Acta Psychiatr Scand. 2020]. Similarly, we have observed 7% lower 5-HT4R BPND in depressed individuals, which was associated with impaired verbal memory—a common symptom in depression [Köhler-Forsberg K, et. al. JAMA Psychiatry 2023].

The primary aim was to determine if our cross-sectional finding on lower 5-HT4R BPND in OC users replicated in a longitudinal design by comparing the change in cerebral 5-HT4R BPND after three months of OC use vs. placebo condition. The secondary objective was to determine if OC use affects verbal memory performance and whether such an effect is mediated by a change in 5-HT4R BPND.

Methods: In this single-blind, randomized, placebo-controlled trial (NCT05212389), 39 healthy women were randomized 1:1 to three months of OC use (150 μg levonorgestrel, 30 μg ethinylestradiol) or placebo. Women were PET scanned using the [11C]-SB207145 tracer and tested with the 24-item Verbal Affective Memory Test before allocation and after three months of intervention. We used latent variable models to evaluate the effect of OC use on 1) global 5-HT4R BPND defined as mediated through the shared covariance between the change in BPND across neostriatum, hippocampus, and neocortex adjusted for baseline BPND and change in injected tracer mass per kg bodyweight, and 2) global verbal memory performance defined as mediated through the shared covariance between the change in immediate, short-term and delayed memory recall adjusted for baseline recall scores. Based on the latent variable models, we tested whether a change in 5-HT4R BPND mediated any effect of OC use on verbal memory performance.

Results: OC use resulted in −9.9% (95 CI, −14.8%;−4.8%, p = 0.0002) change in global 5-HT4R BPND from baseline to follow-up compared to placebo. The region-specific changes were −9.9% (−14.8%;−4.8%, p = 0.0002) in neostriatum, −11.0% (−16.3%;−5.3%, p = 0.0002) in hippocampus, and −6.5% (−10.3%;−2.5%, p = 0.002) in neocortex. OC use resulted in a global change of −1.1 (−1.9;−0.3, p = 0.0096) words recalled compared to placebo. This was −1.1 (−1.9;−0.3, p = 0.0096) for immediate, −2.6 (−4.3;−0.8, p = 0.005) for short-term, and −1.9 (−3.4;−0.4, p = 0.011) for delayed recall conditions. The worse memory was explained by the absence of a learning effect in the OC users. The proportion mediated by the change in 5-HT4R BPND was 16.5% (−17.5%;50.4%, p = 0.33), i.e., we can only reject that more than 50.4% was mediated by the change in 5-HT4R BPND, but a larger sample size is required to reject a smaller mediation path.

Conclusions: This study provides evidence that OC use decreases cerebral 5-HT4R BPND and impairs verbal memory function in healthy women. We found no strong evidence supporting that the change in 5-HT4R BPND was the major pathway of action of OC on verbal memory function, but further research is warranted to establish the role of brain 5-HT4R on memory function and on depression pathology in women using OC.

Keywords: Sex Hormones, Serotonin 4 receptor, Molecular imaging, RCT, Depression

Disclosure: Nothing to disclose.

P509. Adults with a post-ingestive increase in FGF21 display striatal dopamine release in response to a high fat high sugar milkshake

Valerie Darcey, Juen Guo, Mary Walter, Kevin Hall

National Institutes of Health, Bethesda, Maryland, United States

Background: Fibroblast growth factor 21 (FGF21) is a potent metabolic hepatokine that responds to a variety of nutritional and metabolic conditions. The post-ingestive increase in FGF21 is thought to provide negative feedback on subsequent appetite for alcohol and sucrose via hypothalamic-striatal signaling. Appetitive behavior is also conditioned by striatal dopamine response. We explored whether post-ingestive dopamine (DA) response to a palatable liquid meal containing sucrose (milkshake) and appetitive behavior differed by FGF21 response profile in humans.

Methods: DA responses to a milkshake high in fat and sugar were measured as a part of a controlled inpatient feeding study using positron emission tomography (PET) (n = 56 [36F/20M]; 31.6 ± 7.0 years old; 51.9% African American; mean BMI 29.6 ± 7.7 kg/m2, range 20.3 - 44.8 kg/m2) [NCT03648892]. Healthy volunteers stabilized on a provided eucaloric diet (30% fat, 50% carb., 20% protein) for 6.8 ± 1.1 days prior to an [11C]raclopride (20mCi bolus) PET scan in the overnight fasted state. Participants then completed a repeat scan 30 minutes post-milkshake (8 fl. oz.; 418 kcal; 60% fat, 33% carb. including 18g sugar). Decreased DA D2 receptor binding potential (D2BP) in striatal regions of interest measured following the milkshake versus the fasted state was reflective of DA “response”. FGF21 was measured in EDTA plasma by ELISA at fasting, 30- and 90-minutes post-milkshake. Incremental area under the FGF21 response curve was split by median response for group comparisons (i.e., elevated vs. suppressed). Appetitive behavior was assessed via ad libitum lunch meal test with a variety of foods after a similar overnight fast prior to discharge. Alpha set at 0.05. Means and standard error indicated. Given the non-normality in data distribution, differences in two group means were evaluated using Wilcoxon rank-sum test.

Results: Median split of incremental FGF21 response to milkshake produced groups displaying distinct post-prandial exposure profiles: elevated FGF21 (26.0 [7.0] p g/mL) and suppressed FGF21 (−17.7 [4.2] pg/mL). Groups were similar in age (p = 0.98), BMI (p = 0.45), resting energy expenditure (p = 0.51), fasting and post-ingestive glucose response (p’s > 0.34), post-ingestive insulin response (p > 0.13), fasting FGF21 (p = 0.59), and indices of liver function (bilirubin total, ALP, ASP, LDH, p’s > 0.14). The elevated FGF21 response group had comparatively greater fasting insulin (18.2 [3.4] µU/mL vs 10.2 [1.0] µU/mL; p = 0.021), impairments in insulin sensitivity (HOMA-IR; 4.7 [1.1] vs 2.3 [0.2]; p = 0.023), and tended to have greater percent body fat (37.0% [2.2%] vs 31.4% [2.1%]; p = 0.059). Groups displayed similar regional D2BP while fasting (p’s > 0.08). However participants with elevated milkshake-induce FGF21 response had significantly greater post-ingestive DA release in the right putamen (−3.24% [4.1] vs. −0.2% [1.9%]; p = 0.017) than participants with milkshake-induced FGF 21 suppression, despite similar explicit ratings of milkshake liking (p = 0.48) and wanting (p = 0.89). Finally, while ad libitum intake of total energy, proportion of macronutrients, and total dietary sugar was similar between groups (p’s > 0.25), those with an elevated FGF21 postprandial response to milkshake consumed greater amount of energy from foods not specifically classified as ‘high fat, high sugar’.

Conclusions: Healthy adults display heterogeneity in FGF21 response to a palatable mixture of macronutrients. These exploratory analyses suggest that an elevated FGF21 post-ingestive response may reflect a mechanism by which metabolic state influences central dopaminergic signaling to guide food choice. How metabolic perturbations observed in those with an elevated FGF21 post-ingestive response (elevated fasting insulin, impaired insulin sensitivity) may influence dopamine signaling remains to be determined.

Keywords: Dopamine, FGF-21, PET Imaging Study, Nutrition, Raclopride

Disclosure: Nothing to disclose.

P510. Hypothalamic Urocortin-3 expressing neurons project to the pituitary gland and signal to the periphery

Ilaria Carta, Giovanni Podda, Tushar Arora, Anita Autry

Albert Einstein College of Medicine and Montefiore Medical Center, Bronx, New York, United States

Background: Urocortin-3 (Ucn3) expressing neurons located in the perifornical area of the hypothalamus (PeFA) are important for many functions, including social behavior and stress. Previous work from our lab revealed that these neurons send projections to the median eminence (ME), the fiber bundle that connects the brain and the pituitary gland. Ucn3 signaling via its high affinity receptor corticotropin releasing factor receptor 2 (CRFR-2) in the hypothalamus has previously been shown to have impacts on the periphery including increased blood pressure, increased heart rate, as well as decreased gut motility and food intake. Taken together, we hypothesize that PeFA Ucn3+ neurons coordinate the peripheral stress response on target organs through neuroendocrine signaling.

Methods: We used adult male and female C57BL/6 mice, Ucn3-Cre mice, and Ucn3-Cre::Ai9 mice in this project according to our Albert Einstein College of Medicine IACUC approved protocol. We used footshock, airpuff, white noise, and tail suspension as acute stressors in conjunction with fiber photometry based in vivo calcium imaging (n = 8 males, 8 females) and Fos histology (n = 4 control, 3 stress for male and female groups). Next, we used fluorogold tracing (n = 5 male mice) and anterograde synaptophysin-GFP tracing (n = 4 males, 4 females) to demonstrate that Ucn3 neurons contact the anterior and posterior pituitary. We additionally imaged Ucn3 and oxytocin puncta using confocal microscopy in Ucn3-Cre::Ai9 mice with or without intracerebroventricular colchicine treatment (n = 3 vehicle, n = 7 colchicine). We used DISCO clearing followed by antibody staining to identify Ucn3 puncta in C57BL/6 pituitaries (n = 3 males, 3 females). We collected plasma from animals 0 min, 30 min, 1 hr, 2 hr, 6 hr, or 24 hr after acute footshock stress to quantify circulating Ucn3 using a sandwich ELISA (n = 7 males, 7 females per timepoint). We also quantified circulating Ucn3 in plasma after optogenetic stimulation of PeFA Ucn3 neurons (n = 6 males, 5 females). Finally, we quantified Ucn3 at baseline and after footshock in males and females with or without ablation of Ucn3 neurons using a conditional caspase-3 virus (n = 6–9 controls, n = 8–9 ablated in male and female groups).

Results: We found that acute stressors of all types significantly increased calcium activity in PeFA Ucn3 neurons by around 50% over baseline in both male and female mice, which was also observed in Fos colocalization with Ucn3 in the PeFA. We observed 6 times more PeFA Ucn3 neurons than expected by chance colocalized with i.p. injected fluorogold. We observed Ucn3+ puncta in both the posterior and anterior pituitary of male and female mice, with puncta close to blood vessels in the posterior pituitary and in a similar pattern to the neuropeptide oxytocin in both C57BL6/J mice and Ucn3-Cre mice injected with synaptophysin GFP in the PeFA. These results were similar between confocal imaging of brain sections and lightsheet imaging of the intact, cleared pituitary. Footshock significantly increased circulating Ucn3 in male (2-fold) and female (0.5-fold) mice between 30 min and 2 hrs post-stress. We were able to mimic this increase in circulating Ucn3 by optogenetically stimulating PeFA Ucn3 neurons. Finally, we were able to block footshock-induced increase in circulating Ucn3 after PeFA Ucn3 neuron ablation.

Conclusions: Taken together, our preliminary findings suggest that PeFA Ucn3+ neurons may regulate the stress response by signaling to peripheral organs, either by interacting with the HPA (hypothalamic-pituitary-adrenal) axis or via an independent neuroendocrine circuit. Results from this project will allow us to understand the role of PeFA Ucn3+ neuronal signaling in peripheral responses to stress, such as changes in cardiovascular function, gut motility, and secretion of insulin from the pancreas.

Keywords: hypothalamus, Acute Stress, HPA axis, Neuropeptides

Disclosure: Nothing to disclose.

P511. Mechanisms of sex-differences in strategies underlying food-related decision making

Michaela Cooke, Priyanka Das, Rachel Ross

Albert Einstein College of Medicine, Bronx, New York, United States

Background: Changes in eating behavior are commonly associated with psychiatric and metabolic illness. Studies in rodents have shown that the medial prefrontal cortex (mPFC) regulates food-seeking and meal termination which can underlie hyperphagia. In models of addiction and aversion, discrete mPFC subregions are thought to have specialized, potentially opposing roles in different domains of cognitive function, but most feeding studies have not distinguished between subregions of the mPFC. The feeding-related melanocortin 4 receptor (MC4R), expressed primarily in the brain, is involved in regulation of energy metabolism, motivation to feed, and meal-patterning, but most studies have focused on its expression in the hypothalamus. We have discovered a subpopulation of glutamatergic MC4R neurons within the mPFC, that seem to respond to food-related information and are important for food-related decision making associated with metabolic homeostasis.

Methods: We use genetically modified mice (n = 4–14 per group) of both sexes. We infused genetically modified virus into the prelimbic (PL) or infralimbic (IL) region of each mouse to delete the MC4R bilaterally from each region. Viruses lacking the active cre-component were infused into mice of the same genotype for a control group. Food intake, body weight, energy expenditure, was obtained for each mouse, and averaged over time. Operant and exploratory behavior tests were performed once. All data was analyzed using 2-way ANOVA within sex, and when appropriate, 3-way ANOVA.

Results: We have found that deletion of MC4R from the ventral, infralimbic (IL) region leads to an increase in food intake and body weight in male mice, as well as increased latency to access food in a novel environment. Suprisingly, manipulation of MC4R expressing neurons in the dorsal, prelimbic (PL) region, may blunt food intake and body weight gain in male mice. This implies that in male mice, these two subpopulations play opposing roles in feeding-related decisions. Female mice seem to employ a different strategy to interpret food and metabolic information. Deletion of MC4R from the IL in female mice does not affect food intake or body weight, but still has an impact on metabolic homeostasis; in the fasted state females without MC4R expression in the IL paradoxically increase their energy expenditure, which is not seen in females with intact MC4R, nor in the fed state. This is also not seen in male mice.

Conclusions: These differences, which are based in the function of a single molecular marker that is genetically associated with disorders of feeding across the weight spectrum (anorexia and obesity), may help us

Keywords: sex differences, Decision Making, Eating disorders, Neural circuits

Disclosure: Nothing to disclose.

P512. 3-Dimensional mapping of paraventricular oxytocin neurons reveals sex- and region-specific differences in two genetic mouse models of autism

Aishwarya Patwardhan, Siyao Li, Jessica Chen, Katrina Choe

McMaster University, Hamilton, Canada

Background: Oxytocin (OXT) neurons within the paraventricular nucleus of the hypothalamus (PVN) play a key role in coordinating various forms of social behaviours and have been implicated in the lower sociability phenotype observed in multiple genetic models of autism (Patwardhan and Choe, 2025). Yet, the spatial organization of PVN OXT neurons remains incompletely characterized, limited by the spatial heterogeneity across subregions of PVN and technical constraints in conventional histology.

Methods: To address this, we applied a high-throughput pipeline integrating SHIELD tissue clearing, whole-PVN OXT immunolabeling, light-sheet microscopy, and three-dimensional (3D) machine learning-based soma detection analysis to quantify OXT-immunopositive (OXT+) neurons across the PVN of male and female Cntnap2 knockout (KO) and Fmr1 KO mice, two genetic mouse models of autism (n = 4–6 mice/group). We validated our pipeline against conventional section-based immunohistochemistry and confocal imaging, followed by manual cell counting.

Results: Section-based and whole-PVN volumetric analyses both revealed consistent genotype-dependent differences in OXT+ neuron counts within the rostro-middle and middle PVN subregions, with several alterations shared across the two KO models. The whole-PVN approach further uncovered additional changes in rostral and caudal regions that were not detected by section-based methods. Notably, the two models displayed divergent sex effects: Cntnap2 KO males and Fmr1 KO females exhibited more OXT+ neurons in the rostro-middle subregion (p < 0.0001 and p = 0.005, respectively, Monte Carlo permutation, 10,000 iterations), whereas Cntnap2 KO females exhibited reduced counts in the middle subregion (p = 0.038 Monte Carlo permutation, 10,000 iterations).

Conclusions: Our whole-PVN OXT+ volumetric analyses revealed subregion- and sex-specific differences in paraventricular OXT+ neuron counts in two genetic models of autism. Our results underscore the utility of intact-tissue 3D quantification for not only measuring specific cell populations, but also uncovering their distinct spatial organization within complex brain regions.

Keywords: oxytocin, Autism, Whole-Brain Rodent Imaging

Disclosure: Nothing to disclose.

P513. Estrogen receptor beta localized on ventral tegmental area dopamine neurons regulates nicotine self-administration acquisition in ovary-intact female rats

Ashley White, Kathleen McNealy, Shailesh Khatri, Sally Pauss, Brandon Henderson, Heather Bimonte-Nelson, Terry Hinds, Jr., Cassandra Gipson

University of Kentucky, College of Medicine, Lexington, Kentucky, United States

Background: Women exhibit greater nicotine use vulnerability than men. High estradiol (E2) exacerbates nicotine use outcomes in women, effects which have been modeled in preclinical nicotine self-administration (SA) studies. Nicotine SA is maintained by dopamine (DA) release from the ventral tegmental area (VTA) to the nucleus accumbens (NA). E2 exerts its effects by binding to estrogen receptors (ER), including ERα, ERβ, and G-protein coupled ER-1 (GPER-1)s. E2 action at ERs specifically has been shown to potentiate DA neuronal excitability within the VTA. Despite clear evidence of mechanistic relationships between E2, ERs, DA, and nicotine, no studies to date have functionally evaluated the specific role of ERβ located on VTA DA cells in driving nicotine consumption during SA in females. Further, there are currently no tools that allow for evaluations of relationships between nicotine neurobiology and ERs with cell-type specificity, as ERβ is also localized on other (non-DA) cell types within the VTA. We determined the influence of ERβ involvement in dopaminergic function of VTA using a systematic approach. We hypothesize that ERβ knockdown in dopaminergic neurons will decrease nicotine consumption in female rats compared to control female rats through decreasing dopamine cell firing within the VTA.

Methods: A total of 22 ovary intact female rats and 3 ovary intact female C57BL/6J mice were used in this project. First to determine if ERβ pharmacological antagonism alters DA neurophysiology, we first patched DA neurons within the VTA of 3 ovary-intact female C57BL/6J mice and bath applied an ERβ antagonist, 4-[2-Phenyl-5,7-bis(trifluoromethyl)pyrazolo[1,5-a]pyrimidin-3-yl]phenol (PHTPP). To validate ERβ knockdown construct, 6 ovary intact female rats were used for whole patch clamp electrophysiology. Whole-cell recordings were made from the soma of TH + GFP-expressing cells using recording pipettes (4–7 mΩ) after establishing a giga-ohm seal (resistance range: 1–10 GΩ). Upon membrane rupture, the cell membrane potential was held at −70mV. For current-clamp recordings, action potentials were evoked by depolarizing current injections (200 ms, 700 pA) through the patch pipette from a hyperpolarized baseline membrane potential (−70 mV). Cell firing frequency was assessed before and after bath application of an ERβ agonist, diarylpropionitrile (DPN). After validation of the viral construct, 16 ovary intact female rats underwent jugular vein catheterization (JVC) surgery then stereotactic craniotomies for viral infusions of ERβ-shRNA AAV5 or scramble control AAV5. After 7-day recovery, rats completed a minimum of 11 sessions of nicotine (0.06 mg/kg/infusion) self-administration (SA) sessions on a fixed-ratio (FR)-1 schedule. All rat and mouse use practices were approved by the Institutional Animal Care and Use Committee of University of Kentucky (UK; Protocol # 2020-3438) and Marshall University (Protocol #664).

Results: An unpaired t-test was conducted on firing frequency and firing amplitude between PHTPP (+) and non PHTPP (-) slices to determine if pharmacological antagonism alters DA neurophysiology. While there were no effects of PHTPP on firing frequency (p = 0.610), slices with bath applied PHTPP exhibited significantly lower firing amplitude (p < 0.05). We determined the AAV construct successfully infects TH+ neurons within the VTA. To determine if ERβ viral knockdown alters dopamine neurophysiology, an unpaired t-test was conducted on percent change from baseline (before DPN application) for DA cell firing frequency in slices applied with DPN between control and ERβ Knockdown rats. Control rats exhibited significantly greater increases from baseline DA firing frequency relative to ERβ Knockdown rats, t(4) = 2.973, p = 0.410. To determine if ERβ knockdown decreases nicotine consumption, we used linear mixed effects and found a significant effect of session β = 0.134, t(1,14) = 2.354, indicating that control rats increased their nicotine consumption by 0.134 mg/kg per session. We also found a significant Session by Virus interaction, β = −0.199, t(1,14) = −2.471, p = .027, suggesting that the rate of increase in consumption over sessions was attenuated in ERβ knockdown rats by 0.199 mg/kg per session. Notably, consumption did not significantly change over time for the ERβ knockdown rats, β = –0.065 mg/kg, t(1,14) = −1.141, p = 0.273. Additional post-hoc testing examining pairwise differences in consumption between control and ERβ knockdown rats at each session determined that consumption was estimated to be significantly greater in control rats relative to ERβ knockdown rats for sessions 9–11 (ps≤.029).

Conclusions: Here we show that akin to OVX, ERβ knockdown in VTA DA neurons results in significant decreases in nicotine consumption during acquisition of SA in ovary-intact female rats. These results are important because they are the first to demonstrate reductions in nicotine use due to a cell-type specific mechanistic interaction between estrogenic signaling and the mesolimbic DA system, which can recapitulate effects on nicotine SA found when endogenous sex hormone production and release is abruptly halted via OVX.

Keywords: Dopamine, nicotine, Estrogen receptor beta, Drug self-administration, Ventral Tegmental Area (VTA)

Disclosure: Nothing to disclose.

P514. Cultural image inductions using the American Indian Multimedia Stimulus (AIMS) set increase neural correlates of positive affective functions

Evan White, Nicole Baughman, Danielle Bethel, Xi Ren, Gabe Cochran, Andrea Wiglesworth, Makiah Torres, Wesley Vaught, Rayus Kuplicki, Martin Paulus

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

Background: Cultural connectedness is a protective factor for mental health among American Indian (AI) peoples, yet the neurocognitive mechanisms by which cultural information supports positive affect and motivation remain underspecified. We tested whether AI cultural images engage neural systems indexing motivated attention and positive affect and how these responses relate to individual differences in cultural engagement and acculturation.

Methods: AI adults (N = 55; 80% female; Mage = 37.8) completed a passive picture-viewing task during simultaneous EEG–fMRI. Stimuli were blocks of American Indian cultural images (AIMS set) and closely matched non-cultural comparators (48 images/condition). EEG outcomes were LPP (early: 400–700 ms; late: 700–1000 ms at Pz) and alpha asymmetry (AA; frontal F8–F7/F4–F3; parietal P8–P7/P4–P3). fMRI outcomes included a priori ROIs (amygdala, anterior cingulate cortex [ACC], nucleus accumbens/ventral striatum [NAcc/VS], anterior insula) and exploratory whole-brain contrasts (cultural–comparator). Individual differences were assessed with well established measures of cultural engagement.

Results: EEG—LPP. Cultural images elicited larger LPPs than comparators: early LPP, t(39) = 2.75, p = .009, d = 0.43 (ΔM = 0.74 µV); late LPP, t(39) = 2.48, p = .020, d = 0.39 (ΔM = 0.66 µV).

EEG—Alpha asymmetry. Frontal AA did not differ by condition (p = .981; cultural M = 0.159, SD = 0.479; comparator M = 0.160, SD = 0.507; ΔM = −0.001). Parietal AA showed greater relative right-parietal alpha during cultural vs comparator viewing, F(1,41) = 4.35, p = .043, ηp² = .096 (ΔM = −0.043; d = 0.32).

fMRI—ROIs. Amygdala activation was greater for cultural images, t(54) = 2.97, p < .001, d = 0.40. ACC showed deactivation in both conditions with marginally greater deactivation to comparators, t(54) = 2.49, p = .080, d = 0.24. No condition differences emerged in NAcc/VS or anterior insula (contrasts near zero).

fMRI—Whole brain. Thirteen clusters differed between conditions, including reduced percent-signal change in bilateral inferior frontal gyrus and right middle frontal gyrus, and increased activation in inferior temporal, lateral occipital, right fusiform, left superior parietal lobule, and right medioventral occipital cortex.

Cross-level associations. Early LPP (cultural) correlated with amygdala activation to cultural images, ρ = .464, p = .003. Greater right-parietal AA during cultural viewing related to larger late-LPP cultural–comparator differences, r = −.361, p = .039. Individual differences modulated neural responses: enculturation (AIES) correlated positively with a temporal–occipital whole-brain cluster (e.g., ρ≈.305, p = .023); acculturation (NAAS) correlated negatively with several frontal cluster contrasts (|ρ|≈.286–.330, p≤.035); Native spirituality (NASS) correlated negatively with selected frontal clusters (|ρ|≈.267–.280, p≤.049). More frequent historical loss thoughts (HLS) related to smaller early LPP cultural–comparator differences, ρ = −.330, p = .037.

Conclusions: AI cultural images robustly increased neural markers of motivated attention (larger LPPs) and positive-affect circuitry (greater amygdala activation), with convergent right-parietal AA effects. Whole-brain results in frontal control and occipito-temporal perceptual regions suggest culture-relevant cues reallocate attentional resources while enhancing visual–affective processing. Neural responses scaled with cultural engagement, acculturation, spirituality, and historical loss, delineating testable mechanisms by which cultural connectedness may confer protection for mental health and guiding culturally grounded prevention and intervention. Limitations include heterogeneous tribal representation and non-tribe-specific stimuli; future work should incorporate community-matched images and longitudinal designs.

Keywords: Multimodal Neuroimaging, Protective factors, American Indian

Disclosure: Nothing to disclose.

P515. People with psychosis have dissimilar functional connectomes from each other—a trend that disappears with increasing age

George Ling, Mengchen Ding, Peyton Myers, Eric Reavis, Scott Sponheim, Gerhard Hellemann, Junghee Lee

University of Alabama At Birmingham, Birmingham, Alabama, United States

Background: Psychotic disorders are thought to exist on a spectrum along dimensions such as cognition, which are commonly extended out to first-degree relatives and community controls. This study aims to evaluate whether this pattern is observed for the brain connectome, and for which dimensions such similarities exist.

Methods: This study included Psychosis Human Connectome Project (PHCP) resting-state functional magnetic resonance imaging (rs-fMRI) data from 115 people with psychosis (PwP), 83 first-degree relatives to PwP, and 47 controls. The Euclidean distance was calculated between subject dyads’ vectorized whole-brain functional connectomes, then inverted to find intersubject similarity scores (ISS). Cross mixed effects linear model was performed to account for dyad interdependency. Power analysis indicates that with a total sample of 253 participants with 6 dyad groups, 4 measurements per participant, power 80%, and alpha of 0.05, a repeated-measures ANOVA can detect a between-factor effect size of 0.18 and within-between interaction effect size of 0.074.

Results: PwP were found to be least similar to each other (B = −0.0008, p = 0.002), as well as between PwP and controls (B = −0.0004, p = 0.002). Relatives were found to be more significantly similar to each other compared to the sample mean (B = 0.0008, p = 0.002), as well as those between relatives and controls (B = 0.0004, p = 0.001). No significant similarities compared to the sample mean were found for control-to-control (B = 0.00003, p = .94) and PwP-to-relative dyads (B = −0.00003. p = 0.87). A three-way interaction between dyad average age, dyad age similarity, and group similarity was found to be significant (F(5, 63237) = 9.9, p < 1E-8). Examination of the three-way interaction plot found that PwP-to-PwP dyads became more similar compared to other group dyads as dyad age similarity increased and dyad average age increased. Next, we examined whether a similar pattern exist in each functiona subnetwork by examining intrinsic functional connectome network similarity. We found significant three-way interaction involving group, dyad age similarity and dyad average age in the auditory (F(5, 63237) = 9.9, p < 1E-8), cingulo-opercular (F(5, 63269) = 10.9, p < 1E-9), cingulo-parietal (F(5, 63263) = 15.4, p < 1E-14), retro-splenial temporal (F(5, 63257) = 6.5, p < 1E-5), salience (F(5, 63259 = 19.4, p < 1E-18), somatomotor hand (F(5, 63252) = 4.0, p = 0.001), and visual networks (F(5, 63263) = 7.8, p < 1E-6). Examination of these network effects found that wider age gaps moderate network similarity on different group dyads with increasing age.

Conclusions: Intersubject functional connectome similarities (ISS) were strongest in relative-to-relative and relative-to-control dyads, and weakest in PwP-to-PwP and PwP-to-control dyads. PwP dyad functional connectome dissimilarity indicates that psychotic disorders are both heterogeneous and more distinct from un-afflicted relatives and controls than the literature suggests. This effect is most pronounced in younger adults, but with increasing age and age similarity, these functional connectomes converge in similarity.

Keywords: Resting-state fMRI, Schizophrenia and related disorders (SRD), Age

Disclosure: Nothing to disclose.

P516. The role of phenotypic heterogeneity in volumetric brain associations with autistic traits in preadolescence

Hexin Luo, Andrew Kim, M. Catalina Camacho, Deanna Barch, Chad Sylvester, Rebecca Schwarzlose

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

Background: Brain volume is positively associated with autism in infancy and toddlerhood, however volumetry studies from later childhood have yielded mixed and contradictory results. Large sample sizes afford greater statistical power to resolve effects while accounting for crucial covariates. Using a large sample to clarify the relationship between autistic traits and brain volume in later childhood could provide valuable information about neurodevelopment and context for interpreting the consistent positive relationships observed in early childhood.

Methods: We analyzed total brain volume (TBV) and regional brain volumes derived from structural MRI (sMRI) in a large community sample of preadolescents from the Adolescent Brain Cognitive Development (ABCD) Study. Participants (N = 9961 after exclusions) were 9–10 years of age at scan. Autistic traits and psychiatric symptoms were assessed with the Short Social Responsiveness Scale (S-SRS) and Child Behavior Checklist, respectively. Associations between TBV and either total S-SRS score or individual S-SRS item values were computed using hierarchical mixed-effects models nested by scanner that included sex, age, socioeconomic disadvantage, internalizing and externalizing symptoms as covariates. Itemwise analyses based on the same mixed-effects models were computed using individual item values instead of total S-SRS scores and FDR-corrected using the Benjamini-Hochberg method. Follow-up models to identify specific regional volumetric associations with the sensory over-responsivity (SOR) item included intracranial volume as an additional covariate.

Results: Across the full sample, total autistic traits were not significantly associated with total brain volume (p = .83). However, itemwise analyses revealed distinct brain-behavior associations. Of the eleven items, three related to cognitive difficulties (reduced awareness of global features, unusual eye contact, and inability to understand social cues) were negatively associated with TBV (all p < .005), whereas five related to sensory sensitivity and problems making and maintaining relationships were positively associated with TBV (all p < 0.05). Targeted analyses focused on the SOR item revealed significant and specific positive associations between SOR and the volumes of right and left ventral diencephalon and left thalamus (all p < 0.001).

Conclusions: Relations between brain volumetry and autistic traits in later childhood appear to vary based on the specific autistic trait considered. Analyses based on diagnostic status or cumulative measures of autistic traits may obscure brain-behavior relationships. Moreover, targeted analyses for single traits (i.e., SOR) can identify specific volumetric relationships and inform our understanding of underlying neural differences.

Keywords: Structural MRI, Autism Spectrum Disorder (ASD), ABCD study, Autism Spectrum Disorders

Disclosure: Harrow, Honoraria, Spouse/Partner, Regeneron, Advisory Board, Spouse/Partner, Bausch and Lomb, Consultant, Spouse/Partner, Genentech, Advisory Board, Spouse/Partner, Apellis, Advisory Board, Spouse/Partner

P517. Lower dietary fiber intake and increased saturated fat and sugar consumption may exacerbate the neuronal response to food cues and negatively impact mood

Kristina Legget, Maureen McHugo, Lauren Sarabia, Marc-Andre Cornier, Andrew Novick, Allison Hild, Jason Tregellas

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

Background: Obesity rates continue to rise, with substantial impacts on comorbid health conditions and quality of life. Consuming a typical “Western” diet (i.e., low in fiber, high in saturated fats and sugar) is associated with increased obesity risk as well as increased risk for depression, stress, and anxiety. The rapid proliferation of ultra-processed foods (UPFs) in these diets likely plays a key role in rising obesity rates and mood-related eating behaviors. UPFs, characterized by high palatability, energy density, and convenience (e.g., sugar-sweetened beverages, potato chips, breakfast cereals), are a central part of the modern Western diet, accounting for more than half of total caloric intake in the U.S. Though nutritional content varies, UPFs are associated with overall higher saturated fat and sugar intake and lower fiber intake. As greater UPF consumption has been associated not only with increased weight, but also higher rates of stress and depression, it is critical to improve our understanding of the neurobiological effects of these dietary patterns. To begin to address this, the current study used functional magnetic resonance imaging (fMRI) to examine how UPF-associated nutrient intake relates to the neuronal response to high-calorie food cues, as well as to mood and appetitive behaviors.

Methods: The study included 44 adults with overweight/obesity (23F/21M; mean[SD]: age = 38.2[7.2], BMI = 28.4[5.9]), with imaging data for N = 30 (15F/15M; age = 38.3[7.6], BMI = 28.7[4.6]). Participants completed a one-day, macronutrient-controlled, eucaloric run-in diet, then arrived the morning after an overnight fast for fMRI. During fMRI, participants completed a visual food cues task (primary contrast of interest: high-calorie foods > objects). Daily food intake records were completed for 3 days and analyzed for nutrient composition (Nutrition Data System for Research). Given relevance to UPFs, the following nutrients were assessed (average daily intake in g): fiber, sugar, saturated fatty acids (SFA), and sugar/fiber ratio. Participants also completed measures of perceived stress (Perceived Stress Scale [PSS]) and depression (Center for Epidemiologic Studies Depression Scale-Revised [CESD-R]) in the past week and trait-based eating behaviors (Three Factor Eating Questionnaire [TFEQ]: restraint, disinhibition, and hunger). Correlations between nutrient intake (fiber, sugar, SFA, sugar/fiber ratio) and both (a) neuronal response to high-calorie food cues (vs. objects) and (b) behavioral outcomes (PSS, CESD-R, TFEQ) were assessed. Imaging data were preprocessed using fMRIPrep, with parameter estimates from first-level models entered into second-level analyses in SPM12 to investigate correlations. Correlations between nutrient intake and behavioral outcomes were performed in R.

Results: Less fiber intake was associated with greater response to high-calorie food cues in brain regions in the dorsal attention network, including middle frontal gyrus and intraparietal sulcus (p < 0.005). Lower fiber intake was also associated with greater perceived stress (r = −0.33, p = 0.031). Greater SFA intake was associated with increased trait-based hunger (r = 0.41, p = 0.005) and disinhibition (r = 0.31, p = 0.038), with higher sugar/fiber ratio associated with increased depression (r = 0.44, p = 0.003). No significant relationships were observed between the brain response to high-calorie food cues and sugar, SFA, or sugar/fiber ratio.

Conclusions: Lower fiber intake was associated with increased response to high-calorie food cues in the dorsal attention network. This relationship may support a link between UPF consumption and heightened food cue responsivity, a mechanism thought to contribute to obesity. The observed behavioral effects, including links between lower fiber consumption and higher perceived stress, as well as higher sugar/fiber ratio being associated with higher depression scores, may additionally indicate a negative effect of UPF consumption on emotional health. These links and others, such as the observed relationship between SFAs and trait-based hunger and disinhibition, are likely complex and may also be bidirectional, with emotional effects additionally affecting eating behaviors and vice versa. Overall, findings suggest that future work investigating prospective UPF effects on eating-related neurobiology and mood is needed.

Keywords: Functional MRI (fMRI), Obesity, Ultra-processed foods, Eating behavior

Disclosure: Nothing to disclose.

P518. Multivariate neural representations of distress during self-viewing: laying the groundwork for decoded neurofeedback interventions for body dysmorphic disorder

Kaixi Zhang, Joel Diaz, Hayden Peel, Jamie Feusner

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

Background: Body dysmorphic disorder (BDD) is a psychiatric condition characterized by distorted self-appearance perception. Previous neuroimaging studies implicate abnormal activation and connectivity in those with BDD across the dorsal visual stream, ventral visual stream, orbitofrontal–striatal circuits, caudate, and amygdala–visual system connections. Decoded neurofeedback (DecNef) is a novel approach that uses real-time fMRI and multivariate pattern analysis to modify specific neural representations without requiring participants’ explicit awareness. Such an approach may be necessary for those with BDD who cannot engage in exposure-based psychotherapy due to shame or inability to tolerate distress. While DecNef has been explored in perceptual and affective domains, its potential for clinical psychiatric disorders remains largely untested. Distress related to self-appearance perception in BDD patients is a promising target for intervention, although the neural representation is likely distributed across multiple systems including visual, limbic, and fronto-striatal systems. In this study, we determined the multivariate neural patterns associated with distress levels during self-face viewing in those with BDD and non-clinical participants, in the interest of developing a decoder for DecNef. We analyzed fMRI data from a self-faces viewing task using searchlight-representational similarity analysis (RSA), a multivariate approach that measures trial-by-trial consistency of neural responses.

Methods: We recruited 130 participants, including 41 healthy controls, 24 individuals with subclinical BDD, and 65 clinical BDD participants. During MRI scanning, participants were instructed to passively view images of their own faces presented from four viewing angles under two lighting contrast conditions. Each participant underwent two MRI sessions and, after the second MRI scan session, rated their subjective distress in response to self-face stimuli on a visual analog scale ranging from 0 (no distress) to 100 (extreme distress).

We applied fMRI-based representational analyses to investigate the similarity of BDD participants’ brain representations of naturalistic self-face viewing compared to the neural activation pattern of nonclinical participants (including healthy control and subclinical groups). Functional data were realigned to the first image and segmented into tissue classes, which were then forwarded to transform back into participants’ native space. Single-trial estimates were then generated using the FBR-separate model within a 6-second time window to compare activations. Ultimately, we focused on the activation delay at 2 seconds post-stimulus onset as our main reference point. Based on previous studies, we selected bilateral dorsal visual stream, ventral visual stream, amygdala, caudate and orbitofrontal cortex as regions of interest. All regions were merged into a single region of interest mask in MNI space and then forwarded to be transformed in each participants’ native space.

We applied a searchlight-RSA analysis, systematically moving a 4 mm spherical region across the brain to assess local pattern similarity, and regressed the results on participants’ self-face viewing distress rating scores, separately in the clinical and subclinical groups. This allowed us to identify multivariate regions whose consistency of activation is associated with higher or low distress.

Results: The clinical group showed substantially higher distress scores (M = 62.27, SD = 24.92) compared with the non-clinical group (M = 30.79, SD = 26.36). Welch’s t-test indicated a highly significant difference, t (~127) = 7.16, p < .0001, with a large effect size (Hedges’ g = 1.22).

ROI-based searchlight analysis revealed distinct patterns of covariation between individual self-face viewing distress rating scores and neural representational consistency. In the nonclinical group, associations were sparse and localized primarily in occipital regions (Zmax = 3.36). The clinical BDD group exhibited stronger effects (Zmax = 3.96), with clusters emerging in the occipital cortex and right inferior parietal lobule. No voxel survived after the FDR correction.

Conclusions: This study provides preliminary evidence for distributed patterns of multivariate contributions to distress when viewing one’s face. Moreover, (although not surviving conservative thresholding), this was most strongly represented in visual systems, both in clinical participants with BDD and nonclinical participants with low-to-moderate levels of distress. These shared and distinct representational patterns may serve as targets for developing DecNef interventions to reduce distress-related self-face representations by training patients to shift their neural activation patterns toward those observed in individuals with lower self-face viewing distress. Future work will incorporate hyperalignment approaches to optimize decoder construction and enhance the generalizability of DecNef training across individuals.

Keywords: body dysmorphic disorder, multivariate pattern analysis, Representational similarity analysis, Functional MRI (fMRI)

Disclosure: Nothing to disclose.

P519. 1Hz repetitive transcranial magnetic stimulation to the right frontal pole increases connectivity in cortio-striatal-thalamo-cortical circuits in healthy controls

Meghan Kulak, Jennifer Barredo, Nicole C. R. McLaughlin, Linda L. Carpenter, W. Luke Acuff, Mascha van’t Wout-Frank, Sarah L. Garnaat

Alpert Medical School of Brown University, Providence, Rhode Island, United States

Background: Transcranial magnetic stimulation (TMS) was approved for the treatment of obsessive-compulsive disorder (OCD) in 2019 based on a randomized controlled trial of 20Hz stimulation for the dorsomedial prefrontal cortex. Fewer than 50% of patients respond to this protocol, fueling a growing interest in alternative cortical targets. OCD pathology emerges from dysfunction within separate, parallel anatomical loops connecting the orbitofrontal cortex (OFC), anterior cingulate cortex (ACC), and dorsolateral prefrontal cortex (DLPFC) with the striatum, projecting through the thalamus back to frontal cortex. These cortiocostriatal-thalamocortical (CSTC) loops integrate valuation, motivation, and executive control into complex signals that support adaptive behavior. The frontopolar cortex (FPC), an area adjacent to the OFC, is of particular interest in OCD due to its role in behavioral flexibility and shared connections with the OFC, striatum, and thalamus. The FPC is also anatomically superficial relative to other CSTC nodes, making it a feasible TMS target. We explored the effects of 1 Hz TMS to the right FPC (rFPC) on resting state functional connectivity (RSFC) within CSTC loops in healthy controls.

Methods: Two 9-minute runs of resting state functional magnetic resonance imaging (fMRI) data were obtained from 23 healthy, right-handed adults aged 18–55 without any with any psychiatric diagnosis or medication use using a Siemens 3 Tesla Prisma system and a 64-channel head coil. Participants were randomized to undergo either one (n = 11) or three (n = 12) sessions (one session per day) of neuronavigated 1Hz TMS (1800 pulses, 80% of the resting motor threshold) to the rPFC using the Nexstim NBT device. fMRI was repeated within 24 hours following the final TMS session. We used the fMRIPrep Toolbox for image preprocessing and Nilearn for regression-based connectivity. For each participant and session, we extracted and cross correlated Blood Oxygen Level Dependent timescourses from ten, bilateral seed regions of interest (ROIs) located in frontal pole, superior ventral striatum, inferior ventral striatum, dorsal striatum, and perigenual anterior cingulate cortices. Changes in RSFC between the rFPC and all ROIs was analyzed in all participants before and after TMS using a paired T-test. We then compared RSFC change scores (Pre-post TMS) between participants who received 1 versus 3 TMS sessions for each ROI pair using independent two-sample t-tests and the two-tailed Bonferroni significance threshold of p < 0.0063. Additional control analyses examining connectivity change in functionally distinct cognitive control regions, and primary sensory regions were conducted to assess specificity. We report significant findings with and without correction due to our limited sample size.

Results: When all participants were grouped together, 1Hz TMS to the rFPC resulted in increased RSFC between the rFPC and the left FPC (lFPC) (t = 3.42, p = 0.0024), left dorsal striatum (t = 3.01, p = 0.0064), and bilateral perigenual ACC (right: t = 3.57, p = 0.0017; left t = 3.65, p = 0.0014). These findings remained significant with Bonferroni correction. There were no changes in RSFC between the rFPC and control regions. Between conditions (1 vs. 3 TMS sessions) differences in change in connectivity were significant at uncorrected thresholds between the right inferior ventral striatum and left FPC (t = 2.214, p = 0.038), with a mean change of (M = + 0.049), with stronger connectivity change observed in the higher dose condition (M = –0.128). Group differences in connectivity change between left superior ventral striatum and right FPC were more robust (t = −2.8, p = 0.011), with connectivity increasing in the three sessions condition (M = +0.095) and decreasing in the single dose condition (M = –0.016). However, group differences did not remain significant after Bonferroni correction.

Conclusions: 1Hz TMS to the rFPC was overall effective at modulating key nodes of CSTC loops, with significant changes in connectivity between the rFPC and the lFPC, left dorsal striatum and bilateral perigenual ACCs. Changes in RSFC between the FPC and striatum differed significantly between participants who received 1 versus 3 TMS sessions, with some suggestion that higher doses of TMS resulted in stronger connectivity changes. However, the sample was small and the group differences did not withstand correction for multiple comparisons. These findings suggest that stimulation to the FPC may be effective at modulating the CSTC loops that underlie neuropathological models of OCD and supports further exploration of a dose-response relationship.

Keywords: Obsessive Compulsive Disorder, Functional MRI (fMRI), Cortico-striatal-thalamo-cortical Circuits, Healthy controls

Disclosure: Nothing to disclose.

P520. Intergenerational neuroimaging markers: a preliminary study finding maternal functional brain network organization predicts the emergence of psychopathology in young children

Hajer Nakua, Jennifer Warmingham, Nicholas Bustos, Grace Smotrich, Diana More, Paul Curtin, Elena Arduin, Hope Hendry, Dani Dumitriu, Seonjoo Lee, Rachel Marsh

Columbia University, Mississauga, Canada

Background: Psychiatric illnesses typically onset during childhood and adolescence and are often predicted by elevated internalizing and externalizing symptoms which begin to present in early childhood. We know that child developmental outcomes are influenced by mother-to-child interactions during the pre- and peri-natal periods and that maternal mental health problems confer risk for childhood psychopathology. Less understood is how maternal brain function influences such transmission pathways. Identifying intergenerational neuroimaging markers that influence the transmission of early symptoms can inform long-term risk pathways implicated in the development of mental illness. To address this gap, we examined whether and how indices of maternal global brain organization predict the emergence of internalizing and externalizing symptoms over early childhood.

Methods: We used available neuroimaging and clinical data from the COVID-19 Mother-Baby Outcomes (COMBO) Initiative (n = 70). Perinatal maternal brain organization was indexed by resting-state functional magnetic resonance imaging (rs-fMRI)-derived global graph theory metrics, small world propensity and modularity. Maternal depression and stress symptoms 18 months postpartum were indexed by the Patient Health Questionnaire (PHQ-9) and Perceived Stress Scale (PSS), respectively. Child internalizing and externalizing symptoms were measured using the Child Behavioral Checklist (CBCL) at 24, 36, and 48 months postpartum. First, we fit separate zero-inflated poisson linear mixed-effect models to assess whether maternal brain indices predicted trajectories of child internalizing or externalizing behaviors over a two-year period. Fixed effect covariates included child sex, mother age at scan, and maternal education. Random effects included participant ID. Second, we performed separate mediation analyses using structural equation modelling to determine whether maternal depression or stress mediates the relationship between maternal brain indices and child internalizing and externalizing symptoms at 36 months. Covariates included child sex and maternal education. Finally, we tested the cross-sectional interactive effects of maternal brain indices and maternal psychopathology on child internalizing and externalizing symptoms at 36 months. Covariates included child sex, mother age at scan, and maternal education. Sensitivity analyses were conducted to assess robustness of findings.

Results: Greater values of perinatal maternal modularity and small world propensity (reflecting more optimal organization) predicted increases in externalizing (βunstandardized = 0.5–0.77, p < 0.01), but not internalizing, symptoms over time. Maternal depression and stress symptoms at 18 months postpartum strongly predicted child internalizing and externalizing symptoms at 36 months (βunstandardized = 0.49–1.089, puncor = 0.05- < 0.001). Maternal psychopathology did not mediate the relationship between maternal brain indices and child psychopathology. Finally, more greater values of maternal modularity and small world propensity modified the relationship between maternal and child psychopathology symptoms (βunstandardized = −1.41–−3.83, puncor = 0.02–0.04). Results were consistent across various sensitivity analyses suggesting robustness of findings.

Conclusions: These novel and preliminary findings support the predictive capacity of intergenerational neuroimaging markers on early childhood psychopathology outcomes. Although maternal brain functional organization predicted externalizing symptoms over time, maternal brain network organization did not influence early child psychopathology at 36 months through maternal depression or stress symptoms. However, the degree of maternal brain organization buffered the relationship between perinatal maternal psychopathology and child psychopathology at 36 months. Understanding child developmental periods and risk factors within the framework of intergenerational transmission provides a key avenue towards identifying long-term pathways of illness that could be targeted to mitigate risk.

Keywords: Mother-child dyad, Intergenerational neuroimaging, Developmental psychopathology

Disclosure: Nothing to disclose.

P521. Moment-to-moment neural variability tracks symptom reduction during cognitive behavioral therapy for social anxiety disorder

Konstantinos Tsikonofilos, Douglas D. Garrett, Amirhossein Manzouri, Ke Bo, Tomas Furmark, Tor Wager, Kristoffer Månsson

Karolinska Institutet, Stockholm, Sweden

Background: Neural changes associated with psychiatric treatment response remain poorly understood. Neural variability, once regarded as noise, is recognized as a promising biomarker for psychiatry (see review Tsikonofilos et al., 2025, Biol Psychiatry). Here, we examined neural variability dynamics during cognitive behavioral therapy (CBT) for social anxiety disorder (SAD). We hypothesized that neural variability will change during therapy, and that this change will be associated with symptom reduction.

Methods: Patients with SAD (n = 45) underwent 4 functional magnetic resonance imaging sessions (totaling 180 scans): session 1 (S1) and S2 (pre-treatment baselines), S3 (mid-therapy; right before exposure based therapy components were introduced), and S4 (post-treatment), while patients were performing a common emotional face recognition task. Neural variability was quantified as the standard deviation (SD) of the blood-oxygen-level-dependent (BOLD) signal. Symptom severity was quantified using the self-reported Liebowitz Social Anxiety Scale (LSAS-SR). Test-retest reliability of neural variability was quantified using the intraclass correlation coefficient (ICC). Voxel-wise whole-brain signatures of neural variability changes across sessions were extracted using partial least squares (PLS) models, and the effect of session on LSAS-SR was assessed using linear models. Spatial similarity between salience maps was quantified using cosine similarity, and we compared salience maps to external reference atlases including PET-derived neurotransmitter binding maps, to assess the neurochemical relevance of our findings.

Results: First, neural variability was reliable between baseline sessions (median voxelwise ICC = 0.53; 95% CI [0.532, 0.536]). Second, we identified a relationship between neural variability and session within a multivariate pattern (p < 0.001) with stable clusters encompassing the default-mode network. Specifically, increases in neural variability was found in the early therapy phase (S2 vs. S3: p < 0.05) and persisted until the end of therapy (S2 vs. S4: p < 0.05). Third, LSAS-SR decreased from pre- to post-treatment (p < 0.001) and interacted with neural variability increases (p < 0.05), suggesting that the relationship between neural variability and LSAS-SR varies by treatment phase. Specifically, higher neural variability at post-treatment was associated with lower social anxiety symptoms. Fourth, neural variability changes across treatment mapped most strongly to serotonergic and excitation–inhibition–related systems (e.g., 5HT1b2, 5HT2a3, mGluR5, GABAa1; r = 0.25–0.26).

Conclusions: Neural variability changes during CBT for SAD and correlates with symptom reduction at post-treatment suggesting a neural substrate of treatment efficacy. Our findings highlight the potential of neural variability as a biomarker of treatment response in psychiatric populations.

Keywords: BOLD Signal Variability, Anxeity, CBT, Neural flexibility

Disclosure: Nothing to disclose.

P522. Investigating development of the dopamine system in children and adolescents with ADHD using neuromelanin-sensitive MRI

Clifford Cassidy, Rami Al-Haddad, Philippe Robaey

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

Background: In recent years neuromelanin-sensitive MRI (NM-MRI) has revealed alterations in dopamine system function in neuropsychiatric disorders. This method can complement prior PET imaging findings on dopamine system abnormalities in ADHD but has the advantage of being brief and non-invasive, facilitating imaging studies in children. NM-MRI is particularly relevant to study neurodevelopmental disorders because the NM-MRI signal shows developmental change due to gradual accumulation of NM with age as a result of dopamine system activity. We investigated whether youth with ADHD would show reduced NM-MRI signal, suggesting abnormality in dopamine system development.

Methods: Preliminary neuroimaging and cognitive data were collected on a sample of youth aged 6–17 including a group with ADHD (n = 29, 15 male, mean age 13.3 ± 2.7) and a neurotypical group (n = 41, 20 male, mean age 12.4 ± 3.6). Most ADHD cases were recruited from outpatient psychiatry services of the Children’s Hospital of Eastern Ontario and the rest through community outreach. A parent or physician provided documentation of evaluation and diagnosis of ADHD. Most ADHD cases (n = 20) were currently treated with stimulant medications. Neurotypical participants had no psychiatric history and no treatment with psychiatric medication.

NM-MRI was acquired with a turbo spin echo sequence. General childhood psychopathology was assessed using the Child Behavior Checklist (CBCL) and ADHD psychopathology with the NICHQ Vanderbilt Assessment Scale (both completed by a parent). Cognitive data was collected with the NIH Toolbox using the List Sorting Working Memory Test. NM-MRI data was processed using standard methods with ANTs software and custom Matlab scripts. Briefly, SN signal was extracted from NM-MRI images after warping these images to standardized space. NM-MRI contrast to noise ratio (CNR) was measured relative to a crus cerebri reference region. NM-MRI signal was extracted from a subregional SN ROI, SNdev, consisting of voxels showing marked increase in signal with age, consistent with developmental change in the dopamine system linked to NM accumulation (the top 20% of age-increasing SN voxels, ΔNM-MRI CNR/year > 0.36).

Linear regression and partial correlation analyses to related NM-MRI signal from the SNdev ROI to clinical and cognitive measures while controlling for age and sex. Statistical significance for all tests was set at α = 0.05.

Results: Participants with ADHD had significantly decreased NM-MRI signal in the SNdev ROI relative to n = 41 neurotypical controls (t66 = −3.42, p = 0.001, Cohen’s d = −0.84, linear regression controlling for age and sex). Low NM-MRI signal in SNdev was associated with severity of parental report of symptoms of ADHD according to the Subscale Score from the CBCL (t21 = −3.03, p = 0.006, partial correlation r = −0.57, n = 24) and NICHQ Vanderbilt Assessment Scale (inattention: t17 = −2.52, p = 0.022, partial correlation r = −0.43; hyperactivity: t17 = −2.13, p = 0.048, partial correlation r = −0.34, n = 21). Low NM-MRI signal in the SNdev ROI correlated with poorer working memory performance measured with the List Sorting Working Memory Task, both in the combined sample of NT and ADHD (r = 0.29, p = 0.030, partial correlation, n = 59) and in the ADHD group alone (r = 0.45, p = 0.030 partial correlation, n = 26).

Conclusions: Our results are consistent with ADHD pathology being linked to hypodopaminergia in the nigrostriatal pathway. Future longitudinal work can better reveal the development and persistence of this condition throughout early life. NM-MRI also has potential as a biomarker in ADHD, for instance to support more precise dosing of psychostimulant medication.

Keywords: ADHD, Dopamine, Neuromelanin-sensitive MRI, Working memory

Disclosure: Terran Biosciences, Patent, Self

P523. A Head-To-Head Assessment of VMAT2 Target Occupancy of Therapeutic Doses of INGREZZA Compared to Therapeutic Doses of AUSTEDO Extended Release

Daniel Albrecht, Satjit Brar, Eduardo Dunayevich, Hui Zhang, Gena Riggs, Shannon Ingersoll, Jessica Maynard-Scott, Ryan Terry-Lorenzo, Christine Sandiego, Cristian Constantinescu, Eugenii A. Rabiner, David S. Russell, Dietrich Haubenberger

Neurocrine Biosciences, Inc., San Diego, California, United States

Background: The therapeutic effects of vesicular monoamine transporter 2 (VMAT2) inhibitors in the treatment of hyperkinetic movement disorders are thought to be directly related to their occupancy (TO) of VMAT2, which disrupts the packaging of dopamine (and other monoamines) into synaptic vesicles, reducing dopamine transmission. INGREZZA (valbenazine tosylate [VBZ]) and AUSTEDO XR (deutetrabenazine extended release [DTBZ]) are both inhibitors of VMAT2 approved for the treatment of tardive dyskinesia (TD) and chorea associated with Huntington disease in adults, but the VMAT2 TO resulting from therapeutic doses of each compound has not been directly tested. This crossover study was conducted to measure VMAT2 TO, using [18F]AV-133 PET, after single doses indicated as therapeutic for VBZ and DTBZ.

Methods: Imaging, PK, and safety data were collected and analyzed in 8 healthy male volunteers. The general study design was as follows: Baseline PET Imaging Visit 1, Dosing Visit 1, Baseline PET Imaging Visit 2 (after 14–21 days of washout), and Dosing Visit 2. For VBZ dosing scans, participants received a single oral dose of 40mg or 80mg VBZ 6–8 hours before [18F]AV-133 administration, with PET imaging occurring around the reported Tmax of (+)-α-dihydrotetrabenazine (HTBZ), the active metabolite of VBZ. For DTBZ dosing scans, participants received a single oral dose of 24mg or 48mg DTBZ 2–4 hours before [18F]AV-133 administration, with PET imaging occurring around the reported Tmax of DTBZ metabolites. The sequence and order of drug administration were fixed so that alternating participants received either VBZ in period 1 followed by DTBZ in period 2, or DTBZ in period 1 followed by VBZ in period 2. The dosing level was fixed so that half of participants received 80mg VBZ and 48mg DTBZ, and half received 40mg VBZ and 24mg DTBZ. Volume of distribution (VT) in predefined regions of interest was estimated using a metabolite-corrected arterial input function and Logan graphical methods. VMAT2 TO was determined using Lassen plot analysis. The mean plasma concentration (Cave) of active metabolites (HTBZ for VBZ and a summation of (+)-α-dihydrodeutetrabenazine (dHTBZ) and (+)-β-dHTBZ for DTBZ) was matched to TO for each participant and assessed with an Emax model to provide an exposure-response curve. The primary analysis of TO was conducted with a linear mixed effects model (LMEM), with sequence, period, treatment (VBZ and DTBZ) and dose (medium and high) as fixed effects, and participant as a random effect.

Results: Eight participants completed all four PET visits. All doses of VBZ and DTBZ were generally well-tolerated and consistent with the known safety profile of each compound. Test-retest values for [18F]AV-133 VT across both baseline scan visits were < 10% for striatal regions and < 15% for extrastriatal regions. Both VMAT2 inhibitors caused reductions in [18F]AV-133 binding consistent with occupancy of VMAT2. The LMEM showed least squares (LS) mean TO of 76.5% for VBZ and 38.3% for DTBZ, with a between-group LS mean difference of 38.2% (95% confidence interval of [27.1%, 49.4%], p value = 0.0002). Preliminary exposure-occupancy assessments for both compounds followed sigmoidal curve fits. When modeling the relationship between TO and concentration of active metabolites for each compound, an approximately 60% lower EC50 value was observed for VBZ than DTBZ. We leveraged PK data from the current study and known (VBZ) or inferred (DTBZ) accumulation ratios to estimate Cavg at steady-state. Using EC50 values from the Emax model fits, these steady-state concentration estimates were converted to estimated steady-state VMAT2 TO for VBZ (80mg—88%; 40mg—82%) and DTBZ (48mg—47%; 24mg—40%).

Conclusions: Single doses of both VBZ and DTBZ exhibited displacement of [18F]AV-133 consistent with engagement of VMAT2, considered to be the primary mechanism underlying efficacy of these medications in hyperkinetic movement disorders. However, VBZ resulted in significantly higher occupancy at VMAT2 than DTBZ in the same participants. Exposure-occupancy analyses also revealed a lower EC50 (i.e., higher potency) for VBZ compared to DTBZ. This result may likely be due to the presence of multiple Austedo DTBZ metabolites with a lower affinity for VMAT2 than VBZ’s single high affinity HTBZ metabolite. At the starting dose of VBZ (40mg), observed VMAT2 TO has been linked to significant reductions of TD severity (Terry-Lorenzo et al., 2025, PMID: 39757283; Hauser et al, 2017, PMID: 28320223), which is in line with the VMAT2 TO observed for VBZ in the current study.

Keywords: Positron Emission Tomography (PET) Imaging, Movement disorders, Valbenazine, Deutetrabenazine

Disclosure: Neurocrine Biosciences, Inc., Employee, Self

P524. Free water correction identifies altered white matter microstructure in anorexia nervosa

Caitlin Lloyd, Alexandra Muratore, Bemsi Nkuo, Xinwei Han, Karin Foerde, Marta Peña-Gonzalez, Clara Moreau, Stefan Ehrlich, Paul Thompson, Jonathan Posner, Joanna Steinglass

Columbia University Irving Medical Center and New York State Psychiatric Institute, New York, New York, United States

Background: Anorexia nervosa (AN) is a serious eating disorder characterized by abnormal attitudes and behavior relating to eating and weight, and the maintenance of a low body weight. White matter connections underlie the integration of information across distributed regions of the brain, supporting neuronal communication and co-ordinating complex cognition and behavior. As a method to examine white matter microstructure in AN, diffusion tensor imaging (DTI) may identify patterns of dysconnectivity that may contribute to illness symptoms. However, the interpretation of DTI metrics is complicated by the presence of ‘free water’, which refers to water molecules that are not restricted by cellular structures. Free water is likely elevated in AN, in part due to the loss of gray matter and ventricular enlargement that occurs with severely low weight, as well as possible neuropathology (for example, neuroinflammation). ‘Free water correction’ is a modeling approach that separates the isotropic diffusion signal from extracellular free water making it easier to interpret alterations in white matter metrics. This study examines the impact of free-water correction on inferences concerning white matter microstructure in AN, and characterizes the extent of free-water elevation throughout white matter tissue in AN.

Methods: Diffusion-weighted brain imaging data from 92 females with AN and 87 female healthy controls (HC), aged between 14 and 40 years, were combined across four studies. Data were processed with and without free water correction prior to diffusion tensor estimation, calculation of fractional anisotropy (FA) and other tensor metrics (radial and axial diffusivity), and between-group (AN versus healthy controls (HC)) comparisons of average tensor metrics in 25 core bilateral white matter regions. Effect sizes for between-group differences were compared between standard and free water corrected analyses using bootstrapping, and results were corrected for multiple comparisons using the False Discovery Rate (FDR) approach. Exploratory analyses compared white matter free water between groups, and estimated associations between body mass index (BMI) and tensor metrics in free water-corrected and standard analyses.

Results: Free water correction increased sensitivity to between-group (AN versus HC) differences in regional FA, though effect sizes did not differ between standard and corrected analyses when FDR correction was applied. In free water-corrected analyses, individuals with AN had increased FA in the inferior cerebellar peduncle and fornix, and reduced FA in the posterior limb of the internal capsule, relative to HC (pFDR < 0.05). Outcomes of group comparisons of FA were consistent with outcomes of group comparisons of axial and radial diffusivity in free water-corrected analyses. Fornix free water was elevated in the AN group compared to HC (d = 0.42, p < 0.001), and among individuals with AN low BMI was related to elevated fornix free water (d = −0.66, 95% CI: [−1.09, −0.23], p = 0.003). In standard analyses, there was a positive association between FA in the fornix and BMI (0.63, [0.20,1.06], p = 0.004) that was not observed in free water corrected analyses: d = −0.21, [−0.63,0.21], p = 0.330 (Figure 3). Bootstrapping indicated that associations with BMI differed between standard and free water-corrected analyses with respect to FA of the fornix (d = −0.86, [−1.56,−0.14], p = 0.02).

Conclusions: Free water correction promotes robust inferences surrounding white matter microstructure in AN and should be implemented in future research seeking to elucidate the neural causes and consequences of eating pathology. Elevated FA in the fornix has not been reported previously in AN, likely owing to free water impacting diffusion tensor estimation and sensitivity to true group differences in DTI metrics. The behavioral effects of increased fornix FA in AN should be a focus of future research given the established role of the fornix in various cognitive functions. Free water was specifically elevated in the fornix in AN suggesting this elevation resulted from loss of gray matter tissue volume that is secondary to weight loss, rather than neuroinflammatory processes that have been reported in other psychiatric illnesses. It is unclear whether white matter microstructural differences resolve with weight restoration in AN, highlighting an additional direction for future research.

Keywords: Diffusion Tensor Imaging (DTI), Free water imaging, Anorexia nervosa

Disclosure: Nothing to disclose.

P525. Inter-individual heterogeneity in cortical thickness deviations across schizophrenia, schizoaffective disorder, and bipolar disorder

Natalie Remiszewski, Maria Stanica, Scott Sponheim, Adrienne Lahti, Gerhard Hellemann, Junghee Lee, Nina Kraguljac

The Ohio State University Wexner Medical, Columbus, Ohio, United States

Background: Inter-individual variations are disregarded as ‘noise’ in group-comparison neuroimaging studies, even though they could contain clinically relevant information. Normative modeling, inspired by ‘pediatric growth curve charting’, characterizes normal biological variations in brain features in sex and age across the lifespan, which allows a more precise definition of both ‘normal’ and ‘abnormal’. Here, we used the normative modeling framework to assess inter-individual heterogeneity in cortical thickness abnormalities across three diagnostic categories: Schizophrenia, Schizoaffective disorder, and bipolar disorder with Psychosis to test the hypothesis that group-level findings placing these disorders on a severity spectrum may not accurately reflect individual pathology.

Methods: We included the P-HCP, EP-HCP, and an in-house dataset in medication-naïve first episode psychosis patients (n = 337) in the analyses. Normative modeling was performed using the Predictive Clinical Neuroscience toolkit and the braincharts pipeline. Cross-site data harmonization was done in a Bayesian framework. One-way ANOVAs were run to compare positive and negative cortical thickness deviation loads across groups. Person-Based Similarity Index (PBSI) were calculated to further investigate the degree of cortical similarity between individuals of the same diagnosis.

Results: We did not find significant differences in positive (F = 2.861, p = .06) or negative (F = 2.792, p = .06) cortical thickness deviation loads or distributions between the schizophrenia, schizoaffective, and bipolar disorder with psychosis groups. Distributions in deviation loads did not differ either. Anatomical loci of deviations were heterogeneous, with no more than 14% of participants sharing deviations in the same brain region. Person based similarity scores did not differ between the three diagnostic groups (F = 1.053, p = .350).

Conclusions: Our findings suggest that cortical pathology is highly heterogeneous across individuals, but do not support the hypothesis that cortical alterations are more severe in schizophrenia compared to schizoaffective disorder and bipolar disorder. These findings underscore that brain pathology should be characterized at the individual level.

Keywords: Psychosis, Normative Modeling, Diagnostic Categories

Disclosure: Nothing to disclose.

P526. Charting imaging-based cranial biometry for neurodevelopmental risk assessment

Jakob Seidlitz, Philip Mattisson, Ayan Mandal, Margaret Gardner, Elizabeth Levitis, Anna Zapaishchykova, Gareth Ball, Richard Bethlehem, Joshua Bernstock, Susan Sotardi, Barbara Chaiyachati, Benjamin Kann, Raquel Gur, Aaron Alexander-Bloch

Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, United States

Background: Abnormal growth of the head and body is a common feature of many neurodevelopmental disorders. Rare variants, such as 16p11.2 or 22q11.2 deletions. are frequently associated with deviations in head circumference, height, and weight, while polygenic risk for conditions like autism, schizophrenia, and ADHD has also been linked to altered craniofacial and brain morphology. These early-emerging, measurable traits offer a noninvasive window into underlying neurodevelopmental disruption and may serve as clinically useful biomarkers of psychiatric risk.

Head circumference (HC) is a critical pediatric growth metric but is typically measured via tape-based methods that are prone to variability and limited in anatomical detail. Imaging-derived HC offers a precise, reproducible, and retrospective alternative, allowing extraction from routine MRI scans using the true outer cranial boundary. Beyond overall size, imaging enables quantification of cranial shape features, such as asymmetry and flattening, that improve diagnostic specificity for neurodevelopmental and syndromic conditions.

When linked to normative growth models, these metrics allow centile scoring analogous to height and weight charts. Here, we present a deep-learning pipeline for automated HC and cranial shape measurement from MRI, paired with growth reference charts. This framework enables scalable, anatomically grounded cranial biometry for both clinical and research applications.

Methods: We processed 10,870 MRI scans from research and clinical pediatric populations (0–25 years) using an end-to-end pipeline. Preprocessing included age-specific registration, normalization, and bias correction. A DenseNet model identified optimal axial slices for head contour segmentation, followed by iterative measurement of circumference, area, linear diameters, and regional quadrant-based asymmetry metrics. Output included shape-derived indices such as cephalic index (CI), cranial vault asymmetry index (CVAI), anterior–posterior asymmetry (APA), and cranial roundness index (CRI).

Head circumference growth was modeled within the Generalized Additive Models for Location, Scale, and Shape framework. using the Box-Cox Power Exponential (BCPE) distribution and penalized B-splines on log-transformed age, with sex-specific splines for location (μ) and scale (σ), a general spline for skewness (ν), and a fixed intercept for kurtosis (τ). Model selection was guided by both quantitative (AIC) and qualitative (wormplots) metrics.

MRI-derived HC centiles were compared to established tape-based growth charts, and inter-method agreement was evaluated via ICC for matched scans. In clinical cohorts, z-scores were generated from MRI-based centiles.

Results: MRI-derived head circumference (HC) followed expected developmental trajectories, characterized by rapid growth during infancy and early childhood, followed by gradual stabilization through adolescence. Median HC increased from approximately 33 cm at birth to over 50 cm by age two, plateauing near adult values (~55–56 cm) in late childhood. Sex-stratified centile curves revealed consistently larger HC in males across all percentiles, with the greatest divergence observed at the upper extremes (e.g., ~1 cm difference at the 97th percentile). Growth velocity peaked during the first year of life, underscoring the importance of high-resolution measurement during this critical developmental window. Centile curves derived from MRI closely mirrored those from standard tape-based references (CDC and UKWHO), with intraclass correlation coefficients (ICCs) ranging from 0.88 to 0.94 for matched percentiles across age groups. Group-level analysis demonstrated trends of macrocephaly in patients with 16p11.2-deletion (N = 380, median z = 0.73, 8.2% > +2 SD), and microcephaly in patients with 16p11.2-duplication (N = 390, median z = –0.85, 23.8% < –2 SD) and 22q11.2-deletion (N = 72, median z = –0.72, 8.3% < –2 SD).

Cranial shape and asymmetry metrics demonstrated distinct age-related and anatomical patterns. Global asymmetry measures, including the asymmetry coefficient (AC) and cranial vault asymmetry index (CVAI), declined with age, particularly during early childhood, reflecting increasing hemispheric and regional symmetry. Anterior–posterior asymmetry (APA) and flatness index (FI) were highest in infancy and declined thereafter, while the cranial roundness index (CRI) increased and plateaued in later childhood. Shape variation was also significantly associated with cephalic index group: dolichocephalic individuals exhibited greater anterior-dominant asymmetry (APA: β = 1.152, p < 0.001) and reduced roundness (CRI: β = –0.0003, p < 0.001), whereas brachycephalic individuals showed greater posterior flattening (FI: β = 0.002, p = 0.001), reduced posterior symmetry (PHA: β = –0.180, p < 0.001), and increased leftward hemisphere dominance (LHR: β = 0.002, p = 0.006). These patterns remained robust after adjusting for age, sex, and HC.

Conclusions: Our pipeline enables high-throughput extraction of cranial biometrics from routine MRI. Normative modeling revealed substantial developmental and interindividual variation, including expected deviations in clinical populations with micro- or macrocephaly. Integrating head biometry with brain imaging supports scalable screening for neurodevelopmental disorders and lays a foundation for longitudinal monitoring and precision diagnostics.

Keywords: Structural MRI, Neurodevelopmental Disorders, Deep learning

Disclosure: Centile Bioscience, Founder, Self

P527. A potential metabolic basis for brain activity changes after transcranial photobiomodulation in Alzheimer’s disease

Naomi Gaggi, Xianfeng Shi, Ryan Brown, Katherine Collins, Francisco Castellanos, Perry Renshaw, Ricardo Osorio, Dan Iosifescu

Nathan S Kline Institute/New York University School of Medicine, New York, New York, United States

Background: Transcranial photobiomodulation (t-PBM) is an emerging neuromodulation strategy that may enhance cerebral metabolism and cognition by targeting mitochondrial electron transport. Previous work has demonstrated that t-PBM increases intrinsic brain activity, measured as fractional amplitude of low-frequency fluctuations (fALFF), in patients with early Alzheimer’s disease (AD). Since fALFF correlates with cerebral glucose metabolism, baseline metabolic state may predict responsiveness to t-PBM. Using phosphorus magnetic resonance spectroscopy (31P-MRS), we evaluated whether markers of energy metabolism and oxidative stress, phosphocreatine/inorganic phosphate ratio (PCr/Pi) and pH, predict fALFF changes after t-PBM in our preliminary study.

Methods: Participants with MCI/AD (N = 11; 74% White and 79% non-Hispanic/Latinx; male n = 6; female n = 5) underwent baseline 31P-MRS using a 3T Siemens Prisma MRI system equipped with a dual-tuned 31P/1H 8-channel transmit/receive array coil. MRS data were acquired with chemical shift imaging (TR/TE = 2000/2.3 ms; FOV 24 × 24 × 24 cm³; matrix 8 × 8 × 8; voxel size 3 × 3 × 3 cm³; 18 averages). Multichannel coil data were reconstructed with sum-of-squares to optimize signal-to-noise ratio, interpolated to a 16 × 16 × 16 matrix (15 mm isotropic voxels), and apodized with a 75% Hamming filter before Fourier transformation. To account for anatomical variability, frontal ROIs defined in MNI space (covering F3, F4, FP1, FP2) were transformed into subject space and mapped onto the spectroscopic grid. A weighted averaging scheme (distance-based weighting to voxel center) was applied to generate ROI-specific spectra.

Spectral fitting was performed in the time domain using AMARES in MATLAB, modeling PCr and Pi as damped sinusoids. Brain tissue pH was derived from the chemical shift difference between Pi and PCr using a modified Henderson–Hasselbalch equation. PCr/Pi was calculated as an index of energy reserve relative to metabolic demand.

On a separate day, participants underwent MRI scanning with three rs-fMRI runs: baseline (pre-stimulation), during t-PBM irradiation (frontal ROIs above), and post-stimulation. MRI data were acquired and processed as described in our previous publication (Gaggi et al., 2023 Brain Stim.). fALFF was computed within stimulated ROIs, and pre-stimulation to post-simulation changes were correlated with baseline 31P-MRS indices within the ROIs using Spearman rank correlation. We calculated confidence intervals (CI) using a nonparametric bootstrap resampling approach (2,000 iterations), which sampled the data with replacement to generate a distribution of Spearman’s rho and derive 95% percentile-based confidence intervals. All analyses were conducted in Python.

Results: Baseline 31P-MRS indicated pH below the physiological level of 7.4 across participants (mean = 7.1) within each brain region, consistent with mitochondrial dysfunction. fALFF changes from pre- to post-t-PBM negatively correlated with baseline PCr/Pi in F4 (rho = −0.833, p = 0.010, CI [−1.00–−0.32]) and FP2 (rho = −0.786, p = 0.036, CI [−1.00–0.06]). These right-sided regions exhibited both lower PCr/Pi at baseline and greater increases in fALFF following t-PBM. However, this relationship was not seen in FP1 (rho = −0.21, p = 0.610, CI [−0.92–0.80]), while a positive association was documented in F3 (rho = .762, p = 0.028, CI [0.14–1.00]). fALFF change also correlated with pH in FP1 (rho = 0.810, p = 0.015, CI [0.24–1.00]) and FP2 (rho = 0.893, p = 0.007, CI [0.34–1.00]), regions vulnerable to AD-related neurodegeneration. Those with less impaired pH, closer to 7.4, showed stronger fALFF responses. Whereas, fALFF was not correlated with pH either F3 (rho = −0.24, p = 0.570, CI [−0.92–0.54]) or F4 (rho = 0.64, p = .086, CI [−0.16–1.00]).

Conclusions: These preliminary findings suggest that baseline cerebral metabolism may predict neural responsiveness to t-PBM in MCI/AD. Reduced oxidative metabolism, as indexed by pH and PCr/Pi, was associated with changes in intrinsic brain activity in some regions after local stimulation. Results suggest 31P-MRS as a promising biomarker for identifying individuals most likely to benefit from t-PBM. Moreover, fALFF may provide an accessible functional correlate of metabolic status and treatment engagement. Limitations of this preliminary study include small sample size and correlations not corrected for multiple comparisons. In addition, these results hint towards changes that may serve as predictors, but individual changes are consistent across these small ROIs only in 10 of 11 subjects. Together, this work supports t-PBM as a potential therapeutic approach in early AD and justifies further investigation to clarify the heterogeneous nature of these data in larger, longitudinal studies assessing clinical outcomes.

Keywords: Transcranial photobiomodulation, Resting-state fMRI, 31P Magnetic Resonance Spectroscopy

Disclosure: Nothing to disclose.

P528. Mapping affective symptoms to brain circuits: a symptomic SV2A pet imaging study

Ruth Asch, Karina Moisieienko, Ryan Cool, Mia Weed, Sophia D’Ambrosi, Richard Carson, Robert Pietrzak, Esterlis Irina

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

Background: Depression is a clinically heterogeneous condition, encompassing a wide range of affective, cognitive, somatic, and behavioral symptoms that vary substantially across individuals. Traditional categorical diagnostic frameworks often do not fully capture this complexity, leading to growing interest in network-based approaches. Such “symptomic” methodologies conceptualize psychiatric disorders as dynamic constellations of interrelated symptoms, biological factors, and psychosocial features, and offer a more nuanced alternative to reductionist diagnostic classifications. However, to date there has not been a concerted research effort towards integrating symptomics analyses into neuroimaging studies, an approach that could provide neurobiological mechanistic insight into how different symptom network profiles map onto brain structure and/or function.

Notably, preclinical and postmortem studies of depressive symptomatology converge on synaptic deficits as underlying depression-related disorders. Advances in molecular imaging over the past decade now make it possible to measure presynaptic density in the living brain via quantification of synaptic vesicle glycoprotein 2A (SV2A) with positron emission tomography (PET). As such, SV2A PET has emerged as a powerful tool for investigating synaptic integrity in the context of neuropsychiatric disorders. Here, we take a novel approach combining symptomics with SV2A PET to test if synaptic fidelity within functional brain networks is associated with symptom-level phenotypes.

Methods: A total of 38 (n = 21 female) middle and older aged adults (56.9 ± 12.2 yrs) with depressive symptoms, including 13 individuals with sub-threshold symptoms but no psychiatric diagnosis history and 25 individuals with a history of any mood disorder, participated PET imaging with [18F]SynVesT-1. The primary PET outcome measure, volume of distribution (VT) was defined as the ratio of the [18F]SynVesT-1 concentration in tissue relative to the metabolite-corrected concentration in venous plasma at equilibrium (90–120 min). The mean VT, representing presynaptic density, was calculated across four canonical functional brain networks: default mode (DMN), salience (SN), frontoparietal-executive control (FPN), and basil ganglia-motor control (BGMN) networks. Relative Importance Analyses (RIA), a statistical method which decomposes total variance explained in regression models into proportional contributions, were used to examine relationships between network-level presynaptic density and the expression of depressive and anxiety symptoms as assessed by the Hamilton Depression Rating Scale (17 item version, HAMD) and the Hamilton Anxiety Rating Scale (HAMA) factor scores, respectively.

Results: Across all four networks examined, lower measures of presynaptic density (i.e., SynVesT-1 VT) were associated with worse overall depressive symptoms (r’s: −0.34 to −0.73; p-values: 0.037 to 0.021). The RIA indicated that presynaptic density in the DMN was the strongest predictor of depressive symptoms, followed by the SN, explaining 52.7% and 20.9% of variance in HAMD scores, respectively. On the other hand, only presynaptic density in the SN was significantly associated with anxiety symptoms (r = −0.33, p = 0.046), and presynaptic density in the SN was the greatest predictor of overall anxiety, explaining 29.8% of variance in HAMA scores.

The RIA testing for relationships between circuit-level presynaptic density and symptom factor scores identified DMN as the most substantial contributor HAMD-insomnia (73.6%) and HAMD-anxiety (73.2%), with more modest contributions to HAMA-somatic (42.4%), HAMD-depression (39.0%), and HAMA-psychic factor scores. In contrast, presynaptic density in the SN (28.7%) and BGMN (26.2%) were the primary contributors to HAMD-somatic symptoms.

Conclusions: This study provides the first evidence that presynaptic density within large-scale functional brain networks maps onto specific depressive and anxiety symptom dimensions, linking circuit-level neurobiology with symptom expression. These findings highlight the potential of this novel approach– integrating molecular imaging with “symptomics”– to advance precision psychiatry efforts and point toward network- and symptom-targeted strategies for diagnosis and treatment of mood and anxiety disorders.

Keywords: Positron Emission Tomography (PET) Imaging, Synaptic Vesicle Glycoprotein 2A (SV2A), Symptom dimensions, functional brain networks

Disclosure: Nothing to disclose.

P529. Reward expectancy-related left ventrolateral and right dorsolateral prefrontal functional connectivity is differentially associated with depression and mania risk and a potential marker of treatment response in bipolar disorder

Manan Arora, Henry Chase, Michele Bertocci, Robert Raeder, Osasumwen Benjamin, Genna Bebko, Yiming Wang, Haris Aslam, Richelle Stiffler, Mary Phillips

Western Psychiatric Hospital, UPMC, Pittsburgh, Pennsylvania, United States

Background: Mania/hypomania is the pathognomonic feature of bipolar disorder (BD), and is characterized by elevated reward sensitivity, goal overvaluation, impulsivity and sensation seeking. Unfortunately, BD is often misdiagnosed as major depressive disorder (MDD). Replicable neural markers of mania/hypomania risk are needed for earlier BD diagnosis and pathophysiologically based treatment development. We previously showed that reward expectancy (RE) related left ventrolateral prefrontal cortex (L-vlPFC) activity is associated with higher mania risk, with this relationship being specific to mania, and not common to depression risk but it is unknown whether this pattern of elevated activity is accompanied by elevated functional connectivity (FC) with regions in the reward, executive and other neural networks. Using a robust measure of depression and mania risk, the MOODS-SR-Lifetime depressive and manic domain total score, we hypothesized that lower L-vlPFC and central executive network (CEN) FC would be associated with a higher risk of mania and lower depression risk. We further hypothesized that the magnitude of L-vlPFC and CEN FC in euthymic individuals with BD would be significantly associated with psychotropic medication load and/or number of medications.

Methods: Discovery sample: n = 171(mean age = 23.70 ± 3.38, women = 113) non-BD individuals completed a RE-task during 3TfMRI. 2 Wholebrain multiple regressions identified neural regions where RE-related FC with L-vlPFC seed was associated with depression risk and mania risk, covarying for mania and depression risk respectively. Age, education, mean framewise displacement (FWD-mean) and sex were added as covariates.

Test sample: n = 231 individuals (mean age = 21.69 ± 2.12, men/women = 165); using anatomical masks of the regions identified above, we examined these relationships in the Test sample using 2 separate Poisson loglinear regression models, with MOODS-SR-L depressive domain total score and MOODS-SR-L manic domain total score as the dependent variable, covarying for mania and depression risk respectively. Extracted parameters of RE-related FC were independent variables, and age, education, FWD-mean and sex were covariates in each model.

BD group n = 34, mean age = 25.07 ± 4.06, men/women = 7/27, Separate linear regression models to identify relationship of extracted FC parameters to measures of disease burden and treatment response. Age, education, FWD-mean, and sex were added as covariates.

Results: On wholebrain regression in the Discovery sample, only 1 cluster in the R-dlPFC survived cluster-level Pfwe correction < 0.05 (T(peak)=4.09,K(E) = 133,MNI(Coordiantes)=34,50,30), with RE-related FC with the L-vlPFC seed associated with depression risk. There were no significant clusters associated with mania risk; hence we extracted FC parameters using the R-dlPFC cluster associated with depression risk and used a Poisson loglinear regression, which showed a negative relationship with mania risk(β = −0.12,qFDR < 0.001).

Similar Poisson loglinear regression in the Test sample using extracted FC parameters from the R-dlPFC cluster replicated a positive relationship with depression risk (β = 0.03,qFDR = 0.012) and a negative relationship with mania risk (β = −0.04,qFDR = 0.012).

Extracted FC parameters in BD group had a positive relationship with the number of psychotropic medications(β = 0.34, P = 0.069) and total psychotropic medication load(β = 0.30, P = 0.069), specifically mood stabilizers(β = 1.20, P = 0.040).

Conclusions: We show RE-related L-vlPFC and R-dlPFC FC differentially predicts depression and mania risk in 2 independent samples. Notably a positive association with depression risk, while negative with mania risk. We also show that L-vlPFC and R-dlPFC FC is positively associated with the number of psychotropic medications and total medication load in euthymic individuals with BD. dlPFC is an integral hub in the CEN with a well-documented role in cognition control, working memory, executive attention, decision making and top-down control, with some evidence showing the involvement of left-dlPFC in primarily processing verbal, whereas R-dlPFC showing involvement with non-verbal and affect-related domains. Given the role of the L-vlPFC in reward valuation, our findings indicate that higher depression risk is characterized by a greater integration and top-down regulation of this key region by the R-dlPFC during decision making and reward valuation, and that mania risk is characterized by reduced integration. Our findings in the BD group show that psychotropic medications (specifically mood stabilizers) may be associated with increased integration of these key regulatory regions and could be a potential marker of treatment response in BD management.

Keywords: Bipolar Disorder, Major Depressive Disorder (MDD), Treatment-Response, Early identification of risk, new intervention targets

Disclosure: Nothing to disclose.

P530. Very preterm birth, cortical expansion, and cognitive and psychiatric outcomes at age 10 years

Lisa Gorham, Rachel Lean, Aidan Latham, Joshua Jackson, Berenice Anaya, Dimitrios Alexopoulos, David Loseille, Tara Smyser, Kara Garcia, Christopher Smyser, Cynthia Rogers

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

Background: Children born very preterm (VPT) are at increased risk for altered brain growth and adverse cognitive and psychiatric outcomes. Compared to full-term (FT) children, VPT children have smaller brain size at term equivalent age (TEA) and experience higher rates of executive functioning (EF) deficits and a 3-4-fold increased risk for psychiatric disorders, notably comorbid anxiety, autism, and ADHD (the preterm behavioral phenotype [PBP]). However, prospective associations between VPT birth, brain development, and PBP and EF outcomes remain poorly understood. In our previous work, we have shown that VPT children have smaller surface area than their FT peers at TEA. Over the first ten years of life, however, they exhibit increased expansion compared to FT children in the frontal, temporal, and parietal association cortices known to be important for cognitive and psychiatric functioning, resulting in no group difference in surface area by age 10. Importantly, VPT children also exhibit greater variability in patterns of cortical expansion. Given that smaller surface area is associated with worse cognition and psychiatric symptoms in other cohorts, it is possible that VPT children who experience less cortical expansion over the first decade of life are more likely to have PBP and EF challenges by age 10. However, whether individual differences in cortical expansion predict PBP and EF outcomes remains unknown.

Methods: To address this, we used existing structural MRI and behavioral outcome data from a unique, field-leading 10-year neuroimaging study of socioeconomically diverse VPT and FT children without brain injury (N = 93, 41 male/52 female). Previously, we calculated vertex-level maps of cortical expansion over the ten-year period for all individuals using a novel technique, anatomically constrained multimodal surface matching (aMSM). The Glasser parcellation was overlayed on each subject’s cortical expansion map to obtain an average expansion value for each of 360 parcels. To assess psychiatric functioning, we used a principal component analysis to create an overall measure of the PBP using data from existing parent-reported assessments of ADHD (Conners), Autism (Social Responsiveness Scale), and Anxiety (Screen for Child Anxiety Related Disorders) symptoms. Additionally, we used cognitive data from the NIH Toolbox and NEPSY-II to create an overall metric of EF. Next, we used Bayesian modeling to assess if the outcome of interest (EF or PBP score) was predicted by poverty at birth, post menstrual age at infant scan, the interaction between expansion and subject type (VPT vs. FT) in each parcel, and a random effect for expansion that allowed for differences in effect sizes across parcels. Sex was considered as a covariate but was unrelated to both our outcomes of interest (EF and PBP) and expansion, so it was not included as a covariate in the final models.

Results: EF and PBP metrics were first computed for the full, original cohort (N = 202). As predicted, VPT subjects had on average lower EF (t = 4.44, p < 0.001) compared to FT subjects. While the PBP metric did not show significant differences between groups (t = −1.25, p = 0.21), VPT subjects did report greater internalizing symptoms (t = −2.08, p = 0.04). Within the subsample of subjects with neuroimaging data, left hemisphere surface area at age 10 years was positively correlated with EF (r = 0.26, p = 0.01) and negatively correlated with PBP symptoms (r = −0.28, p = 0.01). Within our subsample of 93 individuals with cortical expansion maps, greater expansion in every parcel was associated with better executive functioning at age 10 years in VPT children (posterior mean of the estimate = +0.64, 95% CI 0.34 to 0.95), but worse executive functioning in FT children (posterior mean of the estimate = −0.45, 95% CI −0.59 to −0.31). Effect sizes were not significantly different between parcels for either group, indicating that the effect of expansion on EF was a cortex-wide effect. Additionally, greater expansion in every parcel was associated with fewer parent reported PBP symptoms in both VPT (posterior mean of the estimate = −2.13, 95% CI −2.49 to −1.77) and FT (posterior mean of the estimate = −0.78, 95% CI −0.94 to −0.61) children, with a stronger association in VPT children. Similarly, effect sizes were not significantly different between parcels for either group, indicating that the effect of expansion on PBP symptoms was a cortex-wide effect.

Conclusions: For VPT subjects, greater expansion in every parcel was associated with better executive functioning and fewer PBP symptoms. For FT subjects, greater expansion in every parcel was associated with fewer PBP symptoms but worse executive functioning. These results suggest that for VPT children, who in infancy have smaller brains than FT children, greater cortex-wide growth over the first decade of life may be protective against cognitive and psychiatric difficulties.

Keywords: preterm birth, executive function, child and adolescent psychiatry, structural neuroimaging, aMSM

Disclosure: Nothing to disclose.

P531. Transdiagnostic profiles of bold signal variability in autism and schizophrenia spectrum disorders: associations with cognition and functioning

Maria Secara, Zara Khan, Ayesha Rashidi, Lindsay Oliver, Ju-Chi Yu, George Foussias, Erin Dickie, Peter Szatmari, Pushpal Desarkar, Meng-Chuan Lai, Giulia Baracchini, Anil Malhotra, Robert Buchanan, Aristotle Voineskos, Stephanie Ameis, Colin Hawco

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

Background: Autism spectrum disorder (autism) and schizophrenia spectrum disorders (schizophrenia) exhibit overlapping social and neurocognitive impairment and considerable neurobiological heterogeneity. Functional magnetic resonance imaging (fMRI) based blood-oxygen-level-dependent (BOLD) signal variability, quantified using mean squared successive difference (MSSD), reflects moment-to-moment brain fluctuations and provides a sensitive marker of neural flexibility related to cognitive capacity. Studying BOLD signal variability provides a unique opportunity to examine the temporal unfolding of brain activity that was once deemed as noise, but is now recognized as essential to brain function. We aimed to characterize intra-regional BOLD variability during resting-state and social cognitive task (Empathic Accuracy; EA) fMRI across schizophrenia, autism, and typically developing controls (TDC), and to test associations with cognition, social functioning, and symptom severity.

Methods: Data were drawn from two harmonized NIMH-funded studies: Social Processes Initiative in the Neurobiology of Schizophrenia(s) (SPINS) and Social Processes Initiative in the Neurobiology of Autism-spectrum and Schizophrenia-spectrum Disorders (SPIN-ASD). The final sample included 176 individuals with schizophrenia (average age = 26 years; 28% female), 89 with autism (average age = 21 years; 45% female), and 149 TDC (average age = 26 years; 46% female), total N = 414 (38% female). fMRI preprocessing used fMRIPrep; participants with mean framewise displacement (FD) > 0.5 mm were excluded. Intra-regional MSSD was calculated for 392 regions (360 regions from the Human Connectome multimodal atlas and 32 subcortical regions) for both resting-state and EA task data, and mean-centered. Regional and network-level MSSD values were harmonized across scanners with neuroCombat and compared across groups using ANCOVAs controlling for age, sex, and FD, with multiple comparison correction by keeping false discovery rate at < 0.05. Tukey HSD tests quantified pairwise contrasts and effect sizes were evaluated (Cohen’s d). Dimensional associations between network MSSD and neurocognition, social cognition, social functioning, medication (chlorpromazine equivalents; CPZE) and symptom severity were examined.

Results: At rest, TDC showed significantly higher MSSD than schizophrenia in 96/392 regions (25%; pFDR < 0.01) and higher than autism in 46/392 regions (12%; pFDR < 0.05). Schizophrenia and autism differed in 18/392 regions (5%; pFDR < 0.05), primarily subcortical and visual areas. During EA, group differences were attenuated, with TDC > schizophrenia in 4/392 regions (1%; pFDR < 0.05), TDC > autism in 14/392 regions (4%; pFDR < 0.05), and schizophrenia and autism differences in 12/392 regions (3%; pFDR < 0.05). At the network level, significant resting-state group differences were detected in visual, somatomotor, auditory, and cingulo-opercular networks (pFDR < 0.05; Cohen’s d = 0.31 - 0.75); during EA, differences were limited to somatomotor and auditory networks (pFDR < 0.01; Cohen’s d = 0.40 - 0.68). Across groups, greater resting-state variability predicted higher neurocognition (pFDR < 0.02), social cognition (pFDR < 0.05), and social functioning (pFDR < 0.01). No significant associations emerged with symptom severity. Higher CPZE correlated with lower sensory network variability (pFDR < 0.05), suggesting medication effects. Task-based variability showed fewer associations, including a negative relationship between visual network variability and self-report social cognition (F = 7.31, pFDR = 0.029).

Conclusions: BOLD signal variability, especially at rest, was positively associated with neurocognition, social cognition, and social functioning across diagnostic groups, suggesting that variability impacts cognitive efficiency and behavior. These results suggest that BOLD signal variability may serve as a sensitive and dynamic neural marker of individual differences in cognitive capacity, complementing traditional mean-based metrics such as functional connectivity. Reduced variability in autism and schizophrenia may indicate similar patterns of neural rigidity among these related conditions, positioning BOLD variability as a potential biomarker for neural flexibility and a valuable target for future transdiagnostic clinical interventions.

Keywords: Schizophrenia Spectrum Disorders, Autism, Heterogeneity, BOLD Signal Variability, fMRI

Disclosure: Nothing to disclose.

P532. Single-session focused ultrasound modulation of amygdala emotional reactivity and resting state connectivity: a double-blind, sham-controlled study

Gregory Fonzo, Raquel Kosted, Lauren Enten, Annamarie DeMarco, Rachel Kline, Manoj Doss, Charles Nemeroff

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

Background: The amygdala plays a critical role in detection of and reactivity to emotional stimuli. Amygdala hyperactivity, as measured by functional magnetic resonance imaging (fMRI) blood oxygenation level-dependent (BOLD) responses, is implicated across mood, anxiety, and trauma-related disorders (MATRDs). Low-intensity transcranial focused ultrasound (tFUS) is a novel technique for non-invasive subcortical neuromodulation, affording the ability to directly modulate subcortical brain structures without reliance upon a cortical intermediary target. Here, we tested whether a single tFUS administration to the left amygdala can modulate task-based fMRI BOLD responses to emotional stimuli and intrinsic circuit connectivity in healthy individuals and in a transdiagnostic sample of patients with MATRDs.

Methods: MATRD patients (N = 29) and healthy comparison subjects (HCs; N = 23) underwent two double-blind counterbalanced tFUS sessions. In each, they completed task-based fMRI (with an emotional face processing task) and resting state fMRI prior to and following a 10-min, in-scanner, MRI-guided active or sham tFUS sonication to the left amygdala using the following tFUS parameters: 10 Hz pulse repetition frequency, 5% duty cycle, 30 sec. sonication duration with 30 sec rest periods repeated 10 times, 650 kHz fundamental frequency, de-rated spatial peak temporal average intensity of 719.9 mW/cm squared). Sessions were separated by at least a week to minimize carryover effects. Voxelwise linear mixed models tested the session (active vs. sham) x time (pre-tFUS vs. post-tFUS) and session x time x group (MATRD vs. HC) interaction effects for task and rest, and both effects in interaction with emotion condition for task fMRI. All statistics are derived from voxelwise linear mixed effects models with a random intercept and fixed effects of time (pre or post-FUS), session (active or sham), group (MATRD or HC), and all higher-order interactions. Task-based fMRI analysis additionally included a fixed effect for task condition (anger, fear, happy, neutral, and shape processing baseline) and all additional higher-order interactions. Effects of interest for task and resting fMRI were Type III tests of fixed effects for the time x session interaction and time x session x group interaction effects. For task fMRI only, effects of interest also included the condition x session x time interaction and the task condition x session x time x group interactions. Probabilistic threshold-free cluster enhancement in a bilateral amygdala, insula, and anterior cingulate cortex (ACC) region of interest mask was used for signal boosting and Type I error correction. Sample size of N = 40 affords power of 0.8 to detect a moderate effect size (partial eta squared = 0.06) for pre/post active vs. sham tFUS differential change, while N = 20 in each group affords power of 0.8 to detect group-related differences of a moderate-to-large effect size (partial eta squared = 0.11).

Results: Active (vs. sham) tFUS induced emotion non-specific attenuated activation in the left amygdala target, as well as right amygdala, bilateral anterior insula, and dorsal anterior cingulate (dACC) (time x session corrected p’s < 0.05). Greater emotion non-specific activation attenuations were observed for MATRD patients in more lateral portions of bilateral amygdala and bilateral mid/posterior insula (group x time x session corrected p’s < 0.05). Active (vs. sham) tFUS attenuated left amygdala intrinsic connectivity with the right amygdala (time x session corrected p’s < 0.05) and produced increased left amygdala connectivity with right dACC in the MATRD but not HC group (group x time x session corrected p’s < 0.05).

Conclusions: A single tFUS administration acutely attenuates amygdala fMRI BOLD responses to naturalistic emotional stimuli and modulates activity in an extended limbic circuit. Active vs. sham tFUS also attenuated interhemispheric intrinsic connectivity of the left amygdala with its contralateral homologue. More pronounced changes in MATRD patients suggest magnitude/extent/pattern of tFUS effects may depend upon baseline circuit phenotypic group or state variation. These findings further establish tFUS as a powerful tool for non-invasive subcortical neuromodulation with high therapeutic potential.

Keywords: focused ultrasound, Amygdala, PTSD, Anxiety and Depression, Functional MRI (fMRI)

Disclosure: Synapse Bio AI, Consultant, Self, Alto Neuroscience, Stock / Equity - Publicly Traded Company, Self

P533. Evaluating annotation of connectivity changes after ketamine with receptor density maps

Jennifer Evans, Carlos Zarate

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

Background: The rapid acting antidepressant ketamine is described as being a glutamatergic modulator [1] but its exact mechanism of action remains an active area of research. Recent developments in making neurotransmitter atlases available enables the investigation of the overlap of functional changes and receptor density. Robust changes in the blood oxygen level dependent (BOLD) signal have been measured using functional magnetic resonance imaging (fMRI) after a ketamine infusion [2] but most studies have been performed in healthy volunteers with drug doses that differ from that administered in clinical treatment. Here we investigate functional connectivity changes after ketamine administration in unmedicated patients with treatment resistant major depression in conjunction with receptor density maps.

Methods: Resting state data were drawn from a double-blind placebo cross-over trial in unmedicated patient with major depression. Participants received either a 40 minute infusion of racemic ketamine (0.5mg/kg) or placebo (saline) while they were being scanned in a 3T Siemens Skyra scanner with one week between each infusion. Complete data from nineteen patients (mean age 38 years, [19–59], 10 females) are used here. Briefly, a 15 minute eyes closed resting state scan was acquired prior to (baseline) and again about 25 minutes after the start of the infusion (post). Multi-echo fMRI data (3mm isotropic, TR: 2.08 s, TEs: 13,27,47 ms, FA: 75 degrees) was preprocessed with afni_proc which included the following steps: slice-timing correction, motion alignment, registration to MNI space, and spatial smoothing to 4mm and anaticor. Correlation matrices were generated for each scan using 3dNetcorr [3] using the Schaefer 100 parcellation [4] and node strength was calculated for each region at both baseline and post-infusion. Neuromaps was used to generate receptor density maps following Hansen et al [5]. Receptor densities were quantitatively compared against correlation changes using spin tests.

Results: The post-ketamine scan has widespread connectivity changes, which show regional specificity. In particular, the visual and the control network both have decreased within network connectivity. Comparing node strength against the density of glutamate receptors demonstrates a weak but non-significant correlation between (R2 :0.08 p: 0.4) for across the whole patient group. Restricting the comparison to the participants with a strong antidepressant response, greater than 10 point change on the MADRS scale, improves the correlation (R2:0.19, p = 0.04). Several other receptors also demonstrate strong relations notably serotonin, mu opioid, and cannabinoid. Interestingly, there was not a strong relation with functional changes and NMDA receptors which have also been implicated in ketamine’s mechanism of action. Comparable motion values were measured between paired ketamine and placebo scans suggesting that this result is not an effect of increased motion during drug administration.

Conclusions: The very early functional connectivity changes expressed by this depressed patient sample during a ketamine infusion reflect glutamatergic receptors involvement commensurate with ketamine’s expected mechanism of action. Integrating receptor density maps with functional connectivity data provides a more comprehensive understanding of how ketamine modulates neural circuits in a clinical population. This approach may enhance knowledge of ketamine’s therapeutic mechanisms and also has the potential to inform the development of more effective and targeted treatments for major depressive disorders.

References:

1. Zarate, CA, (2017) Molecular Psychiatry, volume 22, pages 324–327

2. McMillan, R. (2020) Reviews in the Neurosciences, vol. 31, no. 5

3. Taylor, PT (2013) Brain Connectivity 3:5;523-35.

4. Schaefer, A., (2018), Cereb Cortex, Sep 1;28(9):3095-3114

5. Hansen J.Y. Nat. Neurosci. 25, 1569–1581 (2022).

Keywords: Major Depression Disorder, Resting State Functional Connectivity, (R,S)-ketamine

Disclosure: Nothing to disclose.

P534. Brain Age as a Moderator in the Relationship Between Adverse Life Experiences and Early Adolescent Substance Use Initiation: Brain-Behavior-Environment Interactions in the Adolescent Brain Cognitive Development (ABCD) Study

Christian Schütz, Karina Aika Thiessen, Yuetong Yu, Laura Schmid, Alexis Brieant, Sophia Frangou1

University of British Columbia, Vancouver, Canada

Background: Substance use disorders (SUDs) are increasingly understood to emerge from interactions among biological, psychological, and social risk factors. Early substance use initiation during adolescence is a key modifiable predictor of later SUDs. Early adolescence is marked by rapid neurobiological maturation and heightened sensitivity to environmental influences.

Methods: Using data from the Adolescent Brain Cognitive Development Study (N = 5052), we investigated how the interaction between neurodevelopmental maturity—measured as brain-predicted age difference (brainPAD)—and adverse life experiences (ALEs) predicts substance use initiation by ages 13–14.

Results: BrainPAD represents the difference between chronological age and neurobiological age, derived from multimodal MRI. Regression analyses (χ² = 198.3, p < .001) revealed several significant predictors of early substance use initiation: age (β = 0.23, p < .001), sex (β = 0.14, p < .05), low SES/neighborhood safety (β = −0.41, p < .001), lack of primary (β = 0.14, p < .05) and secondary caregiver support (β = 0.23, p < .001), caregiver substance use or separation from a biological parent (β = 0.28, p < .001), and brainPAD (β = −0.05, p < .05). Notably, interaction effects showed that brainPAD moderated the relationship between specific ALEs and substance use initiation—suggesting differential vulnerability depending on neurodevelopmental maturity.

Conclusions: These findings emphasize the importance of brain-environment interactions in shaping early substance use behavior. BrainPAD may serve as a marker of differential susceptibility to environmental adversity, supporting the development of personalized, developmentally targeted prevention strategies. Brain network specific associations will be discussed.

Keywords: Substance Use Initiation, fMRI, Brain development

Disclosure: Nothing to disclose.

P535. Tolcapone effects on the neural substrates of spatial working memory among individuals with alcohol use disorder

Joseph Schacht, Drew Winters

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

Background: Executive function is frequently impaired among individuals with Alcohol Use Disorder (AUD). A core measure of executive function is spatial working memory (SWM), which is mediated, in part, by dopamine signaling in the prefrontal cortex (PFC). In the PFC, the principal method of dopamine inactivation is degradation by the enzyme catechol-O-methyltransferase (COMT). Tolcapone, a brain-penetrant, reversible COMT inhibitor that is FDA-approved for the treatment of Parkinson’s disease, potentiates extracellular dopamine release elicited by exogenous factors and enhances performance on executive function tasks. Thus, tolcapone might also improve SWM, and its neural substrates, among individuals with AUD.

Methods: We conducted a randomized, placebo-controlled trial of tolcapone (Schacht et al., 2022) in which non-treatment-seeking participants with AUD (N = 73 with usable neuroimaging data) received tolcapone (titrated to 200 mg t.i.d.) or placebo for 8 days. Participants completed an fMRI SWM task (Tapert et al., 2004), comprising working memory, vigilance, and rest trials, at baseline and following 7 days of study medication ingestion. We first tested the interaction between treatment group and time to identify brain regions in which activation for the contrast of working memory vs. vigilance trials differed between groups on Day 7 relative to baseline. We then tested whether activation to this contrast was associated with change in drinking during the medication period or peripheral COMT inhibition, as indexed by a normetanephrine ELISA assay.

Results: At a whole-brain voxelwise Bonferroni threshold of p < 0.01, tolcapone, relative to placebo and to baseline, increased activation of several regions, including the parietal cortex (MNI coordinates [16, −70, 62]), right insula [32, 26, −2], and right dorsolateral prefrontal cortex (dlPFC; [56, 18, 34]). Of these regions, only change in right dlPFC activation was associated with change in drinking, such increased activation for the working memory > vigilance contrast was associated with reduced drinks per day during the medication period (β = −0.11, p = 0.03). Among participants who received tolcapone, increased right dlPFC activation was also associated with greater peripheral COMT inhibition (β = 0.52, p = 0.006).

Conclusions: In this human laboratory study of non-treatment-seeking individuals with AUD, the COMT inhibitor tolcapone increased activation of the right dlPFC, a brain area canonically associated with executive function. This result is consistent with data from healthy controls suggesting that tolcapone executive function and working memory, as well as with data from the same participants showing that tolcapone increased dlPFC activation during response inhibition (Winters and Schacht, in press). The associations between increased dlPFC activation, reduced drinking, and peripheral COMT inhibition suggest that pharmacological interventions that increase cortical dopamine may yield promise for improving executive function among individuals with AUD.

Keywords: tolcapone, working memory, alcohol use disorder, executive function, Dopamine

Disclosure: Apollo Therapeutics, Advisory Board, Self, Source Bio, Consultant, Self, Alcohol Clinical Trials Initiative, Honoraria, Self, Bausch Health, Other Financial or Material Support, Self

P536. Precision functional brain imaging in psychotic disorders - preliminary findings relating symptoms to individualized functional connectivity

Colin Hawco, Lindsay Oliver, Alistair Flint, Javkhlan Byambadorj, Suddene Stone, Ghazaleh Mohammadalinejad, Mahavir Agarwal, George Foussias, Margaret Hahn, Kathleen Bingham, Benoit Mulsant, Aristotle Voineskos, Nicholas Neufeld

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

Background: Individual heterogeneity poses a great challenge in psychiatric neuroscience, contributing to poor reproducibility across studies. While functional magnetic resonance imaging (fMRI) has at times demonstrated limited reliability, functional connectivity (FC) has been shown to serve as a unique ‘fingerprint’, with longer scan durations producing more stable results. Precision fMRI (pfMRI) addresses this issue by using dense sampling to capture high-quality brain data within individuals, a novel alternative to large cross-sectional studies. Early work on non-psychiatric samples has demonstrated how pfMRI can uncover unique FC patterns, brain-behaviour relationships, and precise mapping of functional brain networks. Within a psychiatric context, variability in symptoms over time can be linked to changes in FC within individuals, potentially providing more replicable findings than cross-sectional approaches. This study (ongoing) applied pfMRI across individuals with schizophrenia spectrum (SCZ), remitted psychotic depression (PD), and controls (Con), to test feasibility of pfMRI in psychosis and explore brain-behaviour relationship.

Methods: All procedures were approved by the Research Ethics Board of the Centre for Addiction and Mental Health. SCZ, remitted PD, or Con were recruited into a precision fMRI design, in which participants completed up to six MRI visits, usually separated by two weeks. Visits included structural and functional MRI (up to five fMRI scans of 7 minutes) plus clinical assessments (collected for all participants) including the Brief Psychiatric Rating Scale (BPRS). For each participant, functional connectivity (FC) across 360 regions of the HCP multimodal atlas was calculated for each 7 minute fMRI scan (referred to as a ‘run’). FC across runs on a given scan day was averaged to obtain visit FC. First, we explored within-participant visit-to-visit variability (how much their FC changes across visits) by calculating the average correlation in FC across visits for each participant. This provides a measure of stability in FC across time, and was compared across groups via one-way ANOVA. To capture patterns of FC related to group and symptom severity (BPRS total score by visit), a dimensional reduction method, multidimensional scaling (MDS) was applied to FC at each visit across participants. MDS identified 10 ‘components’ (similar to a PCA), with one value per visit, capturing variance in FC. MDS factors were related to group (SCZ, PD, Con) and BPRS total score using a linear mixed-effects model, incorporating age as a covariate and participant as a random factor for repeated sessions.

Results: A total of 511 runs across 127 scanning visits and 25 participants (12 SSD, mean age 40, 5 males; 7 PD, mean age 67, 4 males; 6 Con, mean age 63, 2 males) had usable data and were included in the analysis. The SSD group showed higher average total BPRS scores (F(2,21) = 2.79, p = 0.039). Visit stability was marginally lower in SCZ (mean R = 0.76) compared to PD (mean R 0.80) and Con (mean R 0.78); one-way anova F(2,21) = 2.8, p = 0.08. The first two components of the MDS data reduction identified individuals. These two components also uniquely identified FC patterns related to symptoms (BPRS total score) and diagnostic group. MDS component 1 was related to higher FC in visual and auditory network; lower FC between the visual network and somatomotor, cingulo-opercular, language, and default mode networks, and higher FC between auditory to cingulo-opercular networks. Component 1 showed a significant relationship with BPRS score (t(123) = 2.94, p = 0.0038), but not group (t(123) = 0.31, p = 0.69). MDS component 2 was related to lower FC in visual and multimodal networks; higher between network FC from cingulo-opercular, dorsal attention, and fronto-parietal networks to sensory-motor networks. Component 2 showed a significant difference by group (t(123) = 3.88, p = 0.0002), but not BPRS score (t(123) = 0.62, p = 0.53). No other components showed significant effects after FDR correction.

Conclusions: We have demonstrated that repeated assessments across a smaller sample can identify significant brain-behavior relationships at the individual level. Precision neuroscience opens an opportunity to move beyond group-based diagnostic effects and link symptoms to brain function at an individual level. pfMRI may overcome challenges in heterogeneity and reproducibility found in cross-sectional designs. This study also demonstrates that individuals with psychosis, up to age 74, can complete precision fMRI designs with multiple scanning sessions. Data collection is ongoing, and future work will identify personalized brain networks associated with more specific symptom domains.

Keywords: Psychosis spectrum symptoms, fMRI, precision functional mapping, Brain-behaviour relationships, Longitudinal imaging

Disclosure: Nothing to disclose.

P537. Altered baseline brain activity (ALFF) in early abstinence from an alcohol use disorder

Nicole Zabik, Marshall Biven, Katelyn Kelley, Marisa Silveri, Jennifer Blackford

University of Nebraska Medical Center, Omaha, Nebraska, United States

Background: Early abstinence from an alcohol use disorder (AUD) is a vulnerable period of recovery. Preclinical models show that early abstinence is characterized by higher baseline brain activity, especially in brain regions associated with emotions and stress. Studies in early abstinence in humans have largely examined baseline brain connectivity, while baseline activity remains largely understudied. Amplitude of low frequency fluctuations (ALFF) during resting state provides a marker of baseline brain activity. To determine whether ALFF was altered in early abstinence, we people with an AUD during early abstinence (EA) and healthy controls (HC). Given the significant heterogeneity in both symptoms and neural responses during early abstinence, we also tested whether childhood trauma would predict variability in ALFF in the EA group based on preclinical evidence that early life stress alters baseline activity and that childhood trauma confers risk for AUD in humans. We hypothesized that ALFF values would be higher in EA compared to HC. Within the EA group, we also hypothesized that childhood trauma would be associated with higher ALFF values. We included sex in the models based on growing evidence for sex differences in early abstinence.

Methods: Participants were adults with AUD in early abstinence (n = 49, 67% men) and healthy controls (n = 34, 50% men) who completed resting-state functional imaging (rs-fMRI) and the Childhood Trauma Questionnaire (CTQ). RS-fMRI data were processed using the standard fMRIPrep pipeline; ALFF values were calculated using XCP-D. Whole brain regression analyses were performed for the test of group x sex in the full sample, and CTQ score x sex in the EA group (SPM12, voxel α=0.005, k = 127).

Results: For the test of group differences, we found a group x sex interaction in the ventrolateral prefrontal cortex (vlPFC), anterior insula, dorsal anterior cingulate cortex (dACC), and supplementary motor cortex. In these regions, EA women had higher ALFF scores, compared with HC women (p’s < 0.01) and EA men had lower ALFF scores compared with HC men (p’s < 0.05).

Within the EA group, we found CTQ score x sex interactions in multiple brain regions. In EA men, we found a positive correlation between CTQ scores and ALFF in the hippocampus and mid-temporal lobe (r’s > 0.5, p’s < 0.005); in EA women, CTQ scores were negatively correlated with ALFF values in the dACC and mid-temporal lobe (r’s < −0.5, p’s < 0.01). In EA women, we found a positive correlation between CTQ scores and ALFF values in the anterior and medial orbitofrontal cortex, vlPFC, and temporal lobe (r’s > 0.5, p’s < 0.03); EA men had a negative correlation between CTQ scores and ALFF in the temporal lobe (r = −0.43, p = 0.01).

Conclusions: These findings provide initial evidence that baseline brain activity is altered in early abstinence and that both sex and childhood trauma contribute to heterogeneity in brain activity. Group differences in baseline brain activity emerged in regions of the emotion processing and salience networks and were moderated by sex; baseline brain activity was significantly higher in EA women and significantly lower in EA men, compared to controls. Childhood trauma contributed to heterogeneity in AUD, with distinct patterns in EA women compared to EA men. These data highlight the importance of considering individual differences, like sex and trauma, in the development of precision medicine approaches to treating AUDs.

Keywords: Alcohol use disorder, ALFF, Childhood trauma, Sex differences

Disclosure: Nothing to disclose.

P538. The Neurobiological Cravings Signature (NCS) as a Predictive Neuromarker of Clinical Outcomes in Alcohol Use Disorder

Andreas Lofberg, Nicholas Harp, Irene Perini, Robin Kämpe, Hanna Karlsson, Michal Pietrzak, Hedy Kober, Markus Heilig

Linkoping University/Clinical and Experimental Medicine, Linköping, Sweden

Background: Craving is central to alcohol use disorder (AUD) and predicts relapse. The Neural Craving Signature (NCS), a machine-learning derived neuroimaging biomarker, correlates with subjective craving but has not been validated for predicting clinical outcomes. This study aimed to evaluate the NCS as a predictor or relapse and alcohol use in AUD.

Methods: We applied the NCS to cue-reactivity fMRI data from 39 patients with AUD in a randomized, placebo-controlled, double-blind clinical trial of repetitive transcranial magnetic stimulation (rTMS), which previously showed no significant treatment effect over a 12-week follow-up.

Results: NCS scores significantly correlated with self-reported craving (PACS, p < 0.0001) and predicted alcohol use measured by the biomarker PEth during follow up (p = 0.027). The effect on PEth was to 92.5% mediated by subjective craving (p = 0.037). A Cox regression showed that participants below median NCS response had a significantly lower hazard of relapse than those above median (HR = 0.35, p = 0.013). The NCS demonstrated moderate accuracy in classifying relapse (AUC = 0.79, p = 0.0001), achieving 66.7% sensitivity and 77.8% specificity at a cutoff of 3.27.

Conclusions: These findings strongly support the NCS as a predictor of clinical outcomes in AUD, and as a candidate biomarker in AUD research and treatment.

Keywords: Alcohol use disorder, fMRI, Multivariate pattern analysis, Biomarker Prediction

Disclosure: Nothing to disclose.

P539. Adverse childhood experiences and frontal cortical developmental trajectories across adolescence—ages 9 to 17 years

Priscila Goncalves, Tamara Sussman, Milenna van Dijk, Neo Gebru, Abby de Arellano, Silvia Martins, Matthew Albaugh, Ardesheer Talati

Columbia University, New York, New York, United States

Background: Adverse Childhood Experiences (ACEs) have been associated with short- and long-term consequences, including psychopathology and substance use in youth. The biological mechanisms through which these occur, such as frontal cortical brain development, are not clear. While exposure to ACEs has been linked to alterations in brain development (e.g., ACEs are cross-sectionally related to larger cortical surface area during early adolescence), few studies provide prospective assessments using large adolescent samples. Here, we investigated the associations between ACEs and frontal cortical development using data from a large longitudinal sample.

Methods: We analyzed data from 11,878 participants aged 9–17 in the Adolescent Brain Cognitive Development Study® (Data Releases 5.1 and 6.0). We categorized the exposure as having low ACEs (0–3 ACEs) or high ACEs (≥ 4ACEs), based on 21 possible experiences, including parental separation, child abuse/neglect, mental health, domestic violence, financial adversity, and sociocultural stressors, assessed at baseline and 1-year follow-up by parent and/or youth report. Outcome variables were standardized frontal cortical thickness ROIs based on the Desikan-Killiany cortical parcellation atlas (22 ROIs), based on quality-controlled imaging data collected at up to four time points from ages 9 to 17 (baseline, 2-year, 4-year, and 6-year follow-up).

We used linear mixed-effects models to examine how ACEs relate to frontal cortical development across time, adding an interaction term with age (as linear and quadratic) and ACEs. The models were adjusted for baseline sex assigned at birth, race/ethnicity, prenatal substance use exposure (tobacco, alcohol, cannabis, and other substances), and time-varying factors, such as puberty scale, substance use (i.e., nicotine, alcohol, cannabis, and other substances), and total intracranial volume. Family relatedness, MRI scanner, and participant ID were included as random intercepts. We applied FDR correction to account for multiple comparisons. Analyses were carried out using the ‘lme4’, ‘emmeans’, and ‘ggplot2’ packages in R version 5.4.0.

Results: A significant percentage of participants, 30.1%, experienced high ACEs at the beginning of the study (ages 9–11). At baseline, 52.2% were male; 64.3% white, 15.9% Black, 17.4% other or mixed, 2.4% Asian; and 20.6% reported being Hispanic. High ACEs were reported by 24.5% of white participants, 47.6% of Black participants, 39% of other/mixed participants, and 7.1% of Asian participants (p < 0.01), and 33.4% of Hispanic and 29.2% of non-Hispanic participants. Regarding substance use, 1.9% reported using any substance use (e.g., nicotine,1+ drink of alcohol, cannabis, and/or other substance) at baseline, and substance use increased as participants aged (i.e., 2-year follow-up: 3.1%, 4-year follow-up: 12.9%, 6-year follow-up: 29.2%).

ACEs showed significant associations with frontal cortical thickness development in the following ROIs: lateral orbitofrontal (Left Hemisphere [LH], at age 9, standardized estimated marginal means of 0.47 for low ACEs, 0.43 for high ACEs; at age 17, low ACEs = −0.87, high ACEs = −0.98), paracentral (LH: at age 9, low ACEs = 0.33, high ACEs = 0.32; at age 17, low ACEs = −0.87, high ACEs = −0.98. Right Hemisphere [RH]: at age 9, low ACEs = 0.31, high ACEs = 0.30; at age 17, low ACEs = −0.90, high ACEs = −0.97), pars orbitalis (LH: at age 9, low ACEs = 0.44, high ACEs = 0.46 at age 17, low ACEs = −0.71, high ACEs = −0.76; RH: at age 9, low ACEs = 0.45, high ACEs = 0.48; at age 17, low ACEs = −0.75, high ACEs = −0.77), pars triangularis (RH: at age 9, low ACEs = 0.47, high ACEs = 0.46; at age 17, low ACEs = −0.79, high ACEs = −0.86), rostral middle frontal (RH: at age 9, low ACEs = 0.49, high ACEs = 0.49; at age 17, low ACEs = −1.00, high ACEs = −1.08), superior frontal (RH: at age 9, low ACEs = 0.45, high ACEs = 0.46; at age 17, low ACEs = −0.78, high ACEs = −0.84).

Our results remained consistent when we tested the models without adjusting for race/ethnicity and prenatal substance use exposures, using age as linear or quadratic. However, in models with age as a quadratic and adjusting for covariates related to race/ethnicity and prenatal substance use exposures, the pars orbitalis and pars triangularis were no longer significant after correcting for multiple comparisons.

We conducted sensitivity analyses stratifying the sample for participants without substance initiation at the time of their scan, using age as a quadratic term and all the covariates of the main model. Our results showed interactions of ACEs and age on the paracentral, rostral middle frontal, superior frontal, and precentral regions.

Conclusions: Participants with high ACEs showed a more rapid thinning of frontal cortical thickness from ages 9–17. Thinning in paracentral, rostral middle frontal, and superior frontal regions seems to be associated primarily with ACEs, whereas findings in the lateral orbitofrontal could also be influenced by substance use, as this latter association was not observed among those without substance use.

Our findings highlight the importance of exploring the relationship between ACEs, frontal cortical development, and substance use during adolescence. Future research should evaluate the effects of substance exposure over time and better understand how and when substance initiation influences frontal cortical development, particularly in individuals with high ACEs.

Keywords: Adverse childhood experiences (ACE), Brain development, Adolescent Brain and Cognitive Development Study, Substance Use Initiation

Disclosure: Nothing to disclose.

P540. Regional blood flow signatures of opioidergic modulation of ketamine in major depressive disorder

Luke Jelen, Owen O’Daly, Fernando Zelaya, James Stone, Allan Young, Mitul Mehta

King’s College London, London, United Kingdom

Background: Opioid mechanisms have been implicated in ketamine’s rapid antidepressant effects, but their influence on ketamine-evoked cerebral perfusion remains uncertain. We investigated whether naltrexone alters regional cerebral blood flow (CBF) responses to ketamine in major depressive disorder, whether these perfusion changes relate to acute subjective effects and next-day symptom change, and how they align with neuroreceptor distribution profiles.

Methods: Twenty-six adults with major depressive disorder completed a randomised, double-blind, crossover study. In two sessions, participants received oral placebo or naltrexone (50 mg), each followed by ketamine (0.5 mg/kg IV over 40 min). Whole-brain CBF was measured with 3D pseudo-continuous arterial spin labelling (3DpCASL) at baseline and during the final 10 min of infusion. 3DpCASL data were processed with the ASAP toolbox and SPM12. Analyses tested the effects of ketamine, pre-treatment condition, and their interaction on regional CBF changes, alongside associations with immediate subjective responses and day-one clinical outcomes. Subjective effects were assessed with the Clinician-Administered Dissociative States Scale (CADSS) and the Psychotomimetic States Inventory (PSI); clinical outcomes with the Montgomery–Åsberg Depression Rating Scale (MADRS) and the Quick Inventory of Depressive Symptomatology Self-Report (QIDS-SR).

To explore neurochemical underpinnings, we correlated spatial profiles of ketamine-induced CBF changes with receptor density maps. Unthresholded contrasts indexing ketamine’s direct effect (placebo pre-treatment: Ketamine > Pre-ketamine) and the ketamine-by-naltrexone interaction were masked and parcellated into 422 regions using a combined cortical–subcortical atlas. Volumetric PET receptor maps for MOR, KOR, NMDA, mGluR5, GABAA, and GABAAα5 from public datasets were similarly parcellated. Spearman correlations between CBF changes and receptor profiles were evaluated against BrainSMASH nulls to correct for spatial autocorrelation (pSA-corr < 0.05).

Results: Ketamine produced significant regional CBF increases in a large frontal cluster encompassing subgenual, pregenual, and dorsal anterior cingulate cortices (FWE-corrected p < 0.05), and these effects were not attenuated by naltrexone. Under placebo pretreatment, pregenual anterior cingulate cortex CBF changes correlated with the intensity of immediate subjective effects (PSI-delusional: r = 0.56, p = 0.004; PSI-perceptual distortion: r = 0.64, p < 0.001), and baseline subgenual anterior cingulate cortex CBF predicted day-one antidepressant improvement (MADRS: r = 0.6, p = 0.002; QIDS-SR: r = 0.67, p < 0.001). These predictive relationships were all disrupted by naltrexone pretreatment. Spatial correlations linked ketamine-induced CBF changes to MOR and mGluR5 distributions, while the interaction with naltrexone aligned with MOR and GABAAα5 profiles (pSA-corr < 0.05).

Conclusions: Ketamine acutely increases CBF in emotion-processing regions in major depressive disorder, a neural signature that persists despite opioid blockade. However, naltrexone disrupts the coupling between acute CBF changes, subjective effects, baseline perfusion, and next-day antidepressant response. Receptor-enriched mapping implicates coordinated glutamatergic, opioidergic, and GABAergic contributions to these CBF effects, providing preliminary mechanistic insight with implications for optimising ketamine-based treatments.

Keywords: Ketamine, Depression, Opioid system, Cerebral Blood Flow, ASL

Disclosure: Nothing to disclose.

P541. Neuromelanin sensitive magnetic resonance imaging in mood, personality, and nonsuicidal self-injury disorders: preliminary report of an exploratory study

Nathan Kolla, Fooroogh Raeisi-Makiani, Clifford Cassidy, Stefan Kloiber, Dorsa Rafiei, Michelle De Pol

Swinburne University of Technology, Melbourne, Australia

Background: Neuromelanin (NM), a dark pigment formed as a byproduct of dopamine metabolism, progressively accumulates within specific organelles in catecholaminergic neurons throughout an individual’s life. It is particularly concentrated in the dopaminergic neurons of the Substantia Nigra (SN). NM’s iron complexes possess paramagnetic properties, enabling their non-invasive detection through Neuromelanin Sensitive Magnetic Resonance Imaging (NM-MRI). The NM-MRI signal in the SN has been validated as an indirect measure of dopamine neuron function within the nigrostriatal pathway, with its intensity correlating with regional NM concentration, even in the absence of neurodegenerative conditions. The dopaminergic system in nonsuicidal self-injury remains unstudied directly. Animal studies show amphetamine-induced dopamine agonism increases self-biting in mice, and neonatal dopamine neuron destruction in rats causes self-mutilation under L-DOPA challenge in adulthood. The addictive model of NSSI posits that habitual self-harm may engage dopamine reward pathways

Previous NM-MRI research in mood disorders demonstrated a significantly lower neuromelanin contrast to noise ratio (CNR) in patients with unipolar depression and bipolar disorder compared to both healthy controls and patients with schizophrenia. Given the distinct clinical characteristics of Nonsuicidal Self-injury Disorder (NSSID) and its proposed neurobiological underpinnings, we hypothesized that NSSI-D might exhibit a lower NM-MRI in the SN, potentially reflecting altered dopamine system function. This exploratory study marks the first investigation into NM-MRI in Borderline Personality Disorder (BPD) and NSSID.

Methods: This exploratory study included 9 participants with NSSID, 12 participants with comorbid borderline personality disorder and depression (BPD + MDD), 14 participants with MDD, and 18 Healthy Controls (HCs). Both males and females were included. NSSID participants met the proposed DSM-5 criteria for NSSI-D, did not meet diagnostic criteria for BPD based on SCID-PD and remained below the 17-item Hamilton Depression Rating Scale (HDRS) cut-off of 17, unlike the participants with MDD and BPD + MDD. Participants with NSSID reported a minimum of five self-harm episodes in the preceding 12 months, and all participants were free from psychotropics and recreational substances at enrolment and throughout their participation in the study. HCs had no current or prior history of mental illnesses and no self-harm in the past four months according to deliberate self-harm inventory (DSHI). Structural T1 and NM-MRI scans of the SN area were acquired using a GE MEDICAL SYSTEMS DISCOVERY MR750 3T model. SN-CNR, represents the SN neuromelanin signal intensity relative to the crus cerebri. This was extracted using a custom MATLAB-based program based on Cassidy et al 2019. Multiple comparison one-way ANCOVA was used for statistical analysis, controlling for age and sex differences.

Results: A significant difference in SN-CNR was observed among the four included groups (p = 0.007, η²p = 0.291, F(3,23) = 9.03). While initial age and sex differences among the groups were non-significant, a significant CNR difference between male (n = 12) and female (n = 42) participants was identified (p = 0.03). Notably, the NSSI-D group exhibited the lowest SN-CNR (mean NSSI-D = 9.94, whole sample mean of 11.35). Specifically, NSSI-D participants demonstrated significantly decreased SN-CNR compared to HCs (p = 0.047, η²p = 0.166, F(3,23) = 4.4, representing a 14.6% mean difference compared to HCs). Our findings did not show significant differences in SN-CNR when comparing the MDD + BPD (p = 0.068) or MDD (p = 0.094) groups with HCs. Furthermore, no significant correlations were observed between SN-CNR and DSHI (p = 0.81) or HAM-D (p = 0.46) scores in the combined patient group pool.

Conclusions: This exploratory study reveals significantly lower substantia nigra neuromelanin levels in individuals with NSSI-D. Even though our sample of participants with depression addressed some limitations of previous studies; by having participants free of psychotropic medication and recreational substances, the results were similar of Shibata et al, 2008 and did not report any differences between the MDD group and the HCs. Given that the NM-MRI signal serves as a proxy for dopamine neuron function and accumulation, these results offer novel insight into the pathophysiology of NSSI-D, potentially indicating altered dopamine system function in the nigrostriatal pathway. These results could potentially implicate a lower dopaminergic activity baseline. Further research is necessary to fully elucidate the underlying mechanisms and clinical utility of these observed differences.

Keywords: Neuromelanin-sensitive MRI, Neuromelanin, Nonsuicidal self-injury disorder

Disclosure: Nothing to disclose.

P542. Relationship between cortical gyrification abnormalities and altered cortical and white matter microstructure in individuals with early psychosis

Daiki Sasabayashi, Johanna Seitz-Holland, Yogesh Rathi, Kang Cho, Dheshan Mohandass, Tashrif Billah, Suheyla Cetin-Karayumak, Grace Jacobs, Cornelius Berberich, Kathryn Lewandowski, Raquelle Mesholam-Gately, Joanne Wojcik, Shinsuke Koike, Yoji Hirano, Tsutomu Takahashi, Michio Suzuki, Ofer Pasternak, Matcheri Keshavan, Daphne Holt, Dost Öngür, Alan Breier, Michael Coleman, Sylvain Bouix, Martha Shenton, Marek Kubicki

University of Toyama, Toyama, Japan

Background: Previous neuroimaging studies have reported an altered brain gyrification index in individuals with psychosis. The deviated brain folding patterns observed in psychosis may reflect compromised neurodevelopmental processes, particularly cortical growth and white matter myelination or axon maturation. However, mechanisms that underlie the cortical folding disturbances and their relationship to clinical phenotypes in psychosis remain unclear. In the current study, we performed a multimodal neuroimaging study to explore the relationship between altered brain gyrification and cortical and white matter pathology in this psychiatric condition.

Methods: This cross-sectional study used structural and diffusion magnetic resonance imaging (MRI) in 203 individuals with early psychosis (136 non-affective and 67 affective psychosis) and 73 healthy individuals, obtained from the Human Connectome Project for Early Psychosis dataset. Using FreeSurfer (ver.7.1.0), we calculated local gyrification indices (LGI) as a metric of cortical folding complexity, which has the methodological advantage of accounting for the inherent three-dimensional nature of cortical surface. We also employed MRI techniques to assess the heterogeneity of free-water corrected fractional anisotropy (FA), a proxy for cortical cytoarchitecture. We performed whole-brain tractography with free-water model. First, we performed vertex-wise group comparisons of the LGI between individuals with early psychosis and healthy controls. Vertex-wise comparisons refer to analyses conducted at each cortical surface point individually in the surface-based model. Second, we used Spearman rank correlation coefficients to examine the relationship between altered LGI and changes in heterogeneity of FA in the brain regions showing LGI abnormalities. We also investigated the relationship between LGI abnormalities and the average, free-water corrected FA of the tracts connecting regions with LGI abnormalities to the frontal, temporal, parietal, and occipital cortex. Lastly, we conducted linear regression analyses to explore the associations between these neuroimaging measures and clinical and cognitive variables. Group comparisons were corrected for multiple comparisons using a Monte Carlo simulation with a vertex-wise threshold of p < 0.05 (two-tailed). Correlation and regression analyses used Bonferroni-corrected significance thresholds (p < 0.05 divided by the number of tests).

Results: Compared with healthy individuals, the early psychosis group exhibited a significantly lower LGI in the left temporo-parietal region (cluster-wise corrected p = 0.0271). The corresponding brain region encompassing these vertex-wise differences was identified, and their average values were computed. No significant LGI differences were found between the non-affective and affective subgroups. Average values of LGI were also significantly associated with those of heterogeneity of FA in this same region (r = 0.143, p = 0.044) only in the early psychosis group, driven by significant associations only for the non-affective psychosis subgroup (r = 0.274, p = 0.001). Moreover, lower LGI values in the early psychosis group were significantly correlated with lower FA values in the white matter tracts connecting the left temporo-parietal brain regions with the frontal lobe (r = 0.212, p = 0.005), driven by significant associations only for the non-affective psychosis subgroup (r = 0.261, p = 0.004). Finally, lower average LGI values in the left temporo-parietal region were significantly associated with more severe negative symptoms in the early psychosis group (Adjusted R2 = 0.081, Standardized β = −0.2340, p = 0.0061), driven by significant associations only for the non-affective psychosis subgroup (Adjusted R2 = 0.042, Standardized β = −0.2080, p = 0.0470).

Conclusions: Our results demonstrate that cortical folding alterations in early psychosis are localized to the left temporo-parietal region in the early psychosis individuals, compared with healthy controls. Such diminished complexity of brain folding may be underpinned by gray matter cytoarchitectural changes in the same region and decreased myelination or reduced coherence of axons in the ipsilateral long-range white matter tracts. These associations were more pronounced in the non-affective subgroup. The fronto-temporal dysconnectivity, possibly responsible for negative symptoms, can be developed as novel neuromodulation targets. By integrating advanced MRI techniques into our multimodal neuroimaging framework, we provide in vivo evidence suggestive of specific mechanisms underlying anomalous cortical folding in psychosis.

Keywords: Local gyrification index, Fractional anisotropy, Multimodal, Early psychosis, Free-water

Disclosure: Nothing to disclose.

P543. Neural processing of positive and negative outcomes as mechanisms of behavioral activation and exposure therapy

Robin Aupperle, Hannah Berg, Yujin Eun, Elisabeth Akeman, Christopher Martell, Kate Wolitzky-Taylor, Michelle Craske, James Abelson, Martin Paulus

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

Background: Identifying neural mechanisms of psychotherapy treatment response could inform augmentation strategies to optimize outcomes. Behavioral treatments often aim to enhance engagement in rewarding activities, reduce avoidance of perceived negative outcomes, and encourage new learning based on these experiences. Here, we examine how neural responses to positive and negative outcomes (monetary or affective) change with completion of behavioral activation (BA) and exposure-based therapy (EXP) for generalized anxiety disorder (GAD).

Methods: Adult participants were randomized to complete 10 sessions of BA or EXP. Individuals who completed ≥7 of 10 therapy sessions and had quality imaging data were included in analyses (N = 46; 27 BA, 19 EXP). Pre- and post-treatment, participants completed (a) the approach-avoidance conflict (AAC) task, involving conflict decisions and experience of associated positive and negative affective and reward outcomes, and (b) the monetary incentive delay (MID) task, involving anticipation and response to monetary gains and losses, during fMRI. Time by treatment effects on bilateral ventral and dorsal striatum and amygdala regions of interest were examined.

Results: For the AAC task, right amygdala activity to affective outcomes decreased (F[1,128.7] = 7.66, ηp² = 0.06, p = 0.006) and striatal activity to reward increased (F[1,132] = 11.27, ηp² = 0.08, p = 0.001) pre- to post-treatment. There was a trend for striatal activity to affective outcomes to increase over time (F[1,129] = 4.57, ηp² = 0.03, p = 0.034) and for left amygdala activation to affective outcomes to decreased more for EXP than BA (F[1,129] = 4.27, ηp² = 0.0.03, p = 0.041). For the MID task, right ventral striatal activity to loss increased across treatment arms (F(1,50) = 3.99, p = 0.051).

Conclusions: Results support reduction of amygdala reactivity as a mechanism underlying behavioral, particularly exposure-based, anxiety treatment. Striatal activation to positive and negative outcomes may also be enhanced by behavioral treatments, which could reflect enhancement of circuits underlying the learning or updating of action-outcome contingencies. Further research is warranted to examine whether neuromodulatory or behavioral augmentations targeting these mechanisms could be used to optimize treatment outcomes.

Keywords: Generalized anxiety disorder, Major Depression Disorder, Psychotherapy, Exposure therapy, Behavioral activation system

Disclosure: Nothing to disclose.

P544. A preliminary longitudinal fMRI study of brain response to cannabis cue in long-term, near-daily cannabis users and controls

Che Liu, Francesca Filbey

Center for BrainHealth, The University of Texas at Dallas, Dallas, Texas, United States

Background: Cannabis is one of the most widely used psychoactive substance worldwide. Craving, a key feature of Cannabis Use Disorder (CUD), drives continued use and can be triggered by cannabis-related visual and tactile cues. Neuroimaging studies have shown that cannabis users exhibit heightened responses in mesocorticolimbic regions during exposure to cannabis cues, which were associated with subjective craving and clinical symptoms. However, to date, long-term changes in brain response to cannabis cues have not yet been examined.

Methods: We measured changes in fMRI BOLD response to cannabis cues at baseline and 1-year follow-up in fourteen near-daily cannabis users (age 22.4 ± 3.9 years, 5 females, interval 367.1 ± 74.6 days) and sixteen non-using controls (age 22.2 ± 2.7 years, 10 females, interval 353.8 ± 85.7 days) recruited across two sites. Cannabis use patterns were determined using the past 30-day Timeline Follow-Back. The Mini International Neuropsychiatric Interview (MINI) and CUD Identification Test (CUDIT) were conducted to evaluate the presence of CUD. Participants also completed the Marijuana Craving Questionnaire (MCQ) to measure level of craving for cannabis, the Marijuana Withdrawal Checklist (MWC) to measure withdrawal symptoms prior to the MRI scan, and the Marijuana Problems Survey (MPS) to measure problems associated with cannabis use. BOLD response to cues were measured across two fMRI runs with pseudorandom order of 18 visual-tactile cues: cannabis cue (i.e., a joint or a pipe, depending on participant preference), natural reward cue (i.e., a favorite fruit) and neutral cue (i.e., a pen). Cues were placed in the participant’s left hand while a corresponding photo was displayed on the screen for ten seconds. Participants then rated their current urge to use cannabis on a scale from 0 (no urge) to 10 (extremely high urge). Cue-based fMRI data was preprocessed and analyzed in FSL. In the voxel-wise whole-brain analyses, FEAT FLAME 1 mixed effect models were used to examine group differences in changes between follow-up and baseline in brain reactivity for the cannabis > neutral contrast (repeated measures analysis of variance), and associations with cannabis measures (i.e., symptom count, grams per week, craving) within the cannabis users. Site was added as a covariate to adjust for the scanner difference. Multiple comparison correction was applied using a Z-threshold of 2.3 and a cluster p threshold of 0.05.

Results: None of the non-using controls became regular cannabis users at follow-up. Within the group of cannabis users, there was a significant reduction in CUDIT scores (p = 0.019) from baseline to follow-up. There were no changes in cannabis use pattern, MCQ and MPS scores between baseline and follow up. Cannabis users exhibited a significant one-year decrease in fMRI BOLD reactivity to cannabis cues versus neutral cues within the bilateral precuneus and posterior cingulate gyrus (PCG; Z ≥ 3.16, p = 0.002), while non-using controls showed no change. In the cannabis users, changes of MCQ scores were positively associated with changes in cue reactivity across four clusters, including (1) bilateral precuneus, PCG, lingual gyrus, intracalcarine cortex, left superior lateral occipital cortex (LOC), left occipital fusiform gyrus (OFG), and left supracalcarine cortex (Z ≥ 3.82, p < 0.0001); (2) bilateral frontal pole, middle frontal gyrus, superior frontal gyrus (Z ≥ 3.31, p = 0.001); (3) left superior LOC, superior parietal lobule, precuneus (Z ≥ 3.13 p = 0.007); (4) left amygdala, hippocampus and right thalamus (Z ≥ 3.23 p = 0.04). Among cannabis users, MPS scores were positively associated with three clusters, including (1) left paracingulate gyrus, anterior cingulate gyrus, PCG, left superior LOC, bilateral precuneus, precentral gyrus and postcentral gyrus (Z ≥ 3.85, p < 0.0001); (2) left paracingulate gyrus, inferior frontal gyrus, pars triangularis, frontal pole, frontal orbital cortex, middle frontal gyrus and subcallosal cortex (Z ≥ 3.22, p = 0.0002); (3) bilateral lingual gyrus, right precuneus, left posterior parahippocampal gyrus and thalamus (Z ≥ 3.22, p = 0.0002). MWC scores were positively associated with activity to cannabis cues in one cluster encompassing bilateral lingual gyrus, intracalcarine cortex, left OFG, right temporal occipital fusiform cortex, PCG and posterior parahippocampal gyrus (Z ≥ 3.17, p = 0.014).

Conclusions: Together, these findings suggest that over time, cannabis users exhibit habituation to cannabis cues, reflected in reduced reactivity in the precuneus and posterior cingulate despite stable use and craving levels. Altered cue reactivity across default mode, visual, frontal, and limbic networks was linked to subjective craving and cannabis-related problems, highlighting that longitudinal changes in brain responses track domain-specific aspects of cannabis experiences.

Keywords: Brain response to cannabis cue, Longitudinal change, Craving, Addiction, fMRI

Disclosure: Nothing to disclose.

P545. Characterizing Developmental Trajectories of Brain Functional Connectivity Linked to the Onset of Non-Suicidal Self-Injury (NSSI) in a Large Longitudinal Cohort of Children and Adolescents at High-Risk for Psychopathology

Jenna Traynor, Marcelo Branas, Marcos Croci, Julia Nolan, Giovanni Salum, Luis Augusto Paim Rohde, Euripedes Constantino Miguel, Pedro M. Pan, Lois Choi-Kain

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

Background: Non-suicidal self-injury (NSSI) is a powerful risk factor for suicide, observed in 17 percent of community sampled adolescents and between 35–80% of adolescents with psychopathology. The onset and peak frequency of NSSI occur during early adolescence, with a decline in prevalence by early adulthood. With a narrow window encompassing its onset, peak, and remission, relying on adolescent self-report precludes opportunities for early intervention to proactively address markers of NSSI risk. Resting-state fMRI is a powerful tool to identify neural trajectories of NSSI that may begin in early childhood. Although several theories have proposed widespread neural network disruptions linked to the development of NSSI, no study has followed children longitudinally to explore them. Using a large longitudinal sample of children at high risk for psychopathology, this study tracked resting-state functional connectivity (rsFC) over time to understand how patterns evolved in children who later went on to develop NSSI. We identified longitudinal differences in rsFC 1) between children who did and did not go on to develop NSSI and 2) between children with NSSI classified as experiencing low psychopathology vs. high psychopathology.

Methods: Data were derived from the Brazilian High-Risk Cohort for the Development of Childhood Psychiatric Disorders, a large longitudinal study of 2551 children at high risk for psychopathology based on family screening. This secondary analysis examined rsFC across three waves of fMRI collection in the imaging subsample of N = 802 children followed across seven years. The age range of participants was 6 to 14 years at Wave 1 (i.e., baseline, n = 802), 9 to 18 years at Wave 2 (n = 624), and 12 to 21 years at Wave 3 (n = 585). At each wave, participants completed psychodiagnostic and mental health screening. At Wave 3, participants completed the Deliberate Self-Harm Inventory (DSHI), capturing the total frequency of lifetime NSSI (e.g., cutting, overdose). Using a machine learning classifier, each subject was categorized into one of three clusters based on scores from the DSHI, the Strengths and Difficulties Questionnaire, and Mood and Feelings Questionnaire: 1) no NSSI, 2) NSSI with low psychopathology, and 3) NSSI with high psychopathology. fMRI data were analyzed using the CONN Toolbox. Subject and wave-specific pre-processing included: discarding the first 3 functional volumes; realignment and unwarping; co-registration; normalization to MNI space; segmentation, 2mm3 resampling; outlier detection; and 6mm smoothing. Participant connectivity maps were produced in first-level analysis after denoising with CompCor. Motion scrubbing, linear detrending, and a 0.008 – 0.09 Hz band pass filter were applied. To date, a preliminary, whole brain, second-level analysis examined between-group differences in rsFC strength across the three waves in > 11,000 ROI-to-ROI connections, controlling for age. Results were considered significant if they survived FDR correction (p < 0.05).

Results: Relative to children who did not develop NSSI, participants in the two NSSI+ clusters (collapsed across low and high psychopathology) had greater rsFC at Wave 1 between the left posterior superior temporal gyrus (pSTG) with the supracalcarine, intracalcarine, and visual cortices; p’s: 0.003 – 0.02); and decreasing rsFC from Wave 1 to Wave 2 between the bilateral pSTG and bilateral cuneal cortices (p’s: 0.004 - < 0.001). At Wave 2, NSSI+ participants showed decreased rsFC within key social cognitive regions (i.e., anterior STG and angular gyri, p’s: 0.04 – 0.004), and the STG and posterior parietal cortex (PPC, p = 0.03). From Wave 2 to Wave 3, participants in the NSSI+ clusters demonstrated increasing rsFC between bilateral anterior and posterior STG with the superior and middle frontal gyri, and the Heschl’s gyri (p’s: 0.02 - < 0.001). On the other hand, there were few differences in rsFC when restricting analyses to NSSI+ participants in the low vs. high psychopathology clusters, with significant differences emerging only at Wave 3 (i.e., late adolescence): relative to NSSI+ participants with low psychopathology, the high psychopathology cluster showed lower rsFC between the left Heschl’s gyrus and right PCC (T = −2.11, p = 0.003) and between the right Heschl’s gryus and right SFG (T = −2.03, p = 0.04). Analyses are underway to examine age as a covariate of interest (e.g., quadratic models) and sex differences.

Conclusions: Preliminary findings suggest that children who go on to develop NSSI can be differentiated from those who do not by trajectories of rsFC that evolve across development. In early childhood, the eventual onset of NSSI was linked to disrupted rsFC in neural regions underlying language and visual perception. By adolescence, NSSI-related disruptions shifted to concentrate in social cognitive regions, with social-cognitive network abnormalities further evolving by late adolescence/early adulthood to include higher-order executive functioning cortical areas. On the other hand, developmental trajectories were less differentiated between NSSI+ participants with low vs. high psychopathology, with differences emerging only by later adolescence, in auditory processing and social cognitive regions. Ultimately, these findings propose patterns of rsFC linked to NSSI risk that may be targeted with early interventions before its onset (e.g., emotion regulation skills training, transcranial magnetic stimulation).

Keywords: Non-Suicidal Self-injury (NSSI), Resting-state fMRI, Adolescents

Disclosure: Nothing to disclose.

P546. Reproducible brain charts: an open data resource for mapping brain development and its associations with mental health

Golia Shafiei, Nathalia Esper, Mauricio Scopel Hoffman, Andrew Chen, Sydney Covitz, Steven Giavasis, Connor lane, Kahini Mehta, Tyler Moore, Taylor Salo, Monica Calkins, Stanley Colcombe, Christos Davatzikos, Raquel Gur, Ruben Gur, Ariel Rokem, Russell T Shinohara, Nim Tottenham, Xi-Nian Zuo, Matthew Cieslak, Alexandre Rosa Franco, Gregory Kiar, Giovanni A. Salum, Michael Milham, Theodore Satterthwaite

University of Pennsylvania, Philadelphia, Pennsylvania, United States

Background: Major mental illnesses are increasingly understood as disorders of brain development [1]. Neuroimaging studies of brain development can help track healthy brain maturation and have the potential to identify deviations from normal development linked to psychopathology. However, large and diverse samples are required to capture reliable neurodevelopmental patterns on the population level [2]. While it is possible to aggregate data across multiple resources, data aggregation is not a straightforward process given the differences in neuroimaging and psychiatric phenotyping protocols used by independent studies. To this end, we introduce Reproducible Brain Charts (RBC), an open data resource that integrates several of the largest studies of brain development in youth.

Methods: RBC includes data from five prominent neurodevelopmental studies from three continents (total N = 6,346): BHRC (n = 610), CCNP (n = 195), HBN (n = 2611), NKI (n = 1329), and PNC (n = 1601). Confirmatory bifactor models were used to create harmonized measures of psychiatric phenotypes that capture major dimensions of psychopathology [3]. These factors include a general psychopathology factor (i.e., p-factor) as well as domain-specific factors such as internalizing and externalizing symptoms [4]. Neuroimaging data were curated with CuBIDS [5], summarizing heterogeneity in data acquisition and facilitating metadata-based configuration of processing pipelines. Structural Magnetic Resonance Imaging (MRI) data were processed using FreeSurfer and sMRIPrep. Functional MRI (fMRI) data (resting-state and task) were processed using C-PAC [6, 7]. The “FAIRly-big” framework [8] was adapted for reproducible image processing, ensuring that all processing steps were fully tracked via DataLad [9]. An extensive array of quality control (QC) metrics and specific inclusion criteria were generated for neuroimaging data to ensure consistent quality assurance procedures.

To illustrate the utility of the aggregated RBC data, we examined: (i) how imaging features were related to participant age and overall psychopathology; and (ii) whether implementing RBC’s QC guidelines and statistical harmonization of imaging data with CovBat-GAM [10] impacted the observed results. We used Generalized Additive Models (GAMs) to assess the relationship between imaging and phenotypic features, controlling for covariates such as sex and data quality. Specifically, we evaluated cortical thickness (CT), surface area (SA), and gray matter volume (GMV) from sMRI data and examined between- and within-network connectivity from fMRI data.

Results: All RBC data – including harmonized psychiatric phenotypes, unprocessed neuroimaging data, and fully processed imaging derivatives – are publicly shared without a data use agreement via the International Neuroimaging Data-sharing Initiative. The majority of participants included in RBC were 5 to 23 years old, with a relatively balanced sex distribution (45% female). Overall, approximately 90% of imaging data passed both structural and functional QC criteria. Our initial analysis of combined RBC data without QC or statistical harmonization identified developmental effects that varied markedly between studies. However, QC and statistical harmonization effectively removed variability among datasets while retaining heterogeneity due to interindividual variations. Aligning with prior work, we found evidence for a decline in structural features during development following QC and harmonization (p < 0.0001). The findings also suggested that reduced SA and GMV (but not CT) were linked to higher overall psychopathology (p < 0.0001). Our analyses of functional data demonstrated that between-network connectivity decreased while within-network connectivity increased during development. Additionally, greater functional connectivity between the default mode and frontoparietal networks was associated with higher psychopathology (FDR-corrected p = 0.0002), suggesting a loss of segregation between those two networks.

Conclusions: Taken together, RBC integrates data from five neurodevelopmental studies, facilitating large-scale, robust, and reproducible research in developmental and psychiatric neuroscience. By reducing barriers to large-scale data access, RBC promotes equity and collaboration in developmental and translational brain research.

Keywords: Human Neuroimaging, Brain development, Open science

Disclosure: Nothing to disclose.

P547. An ultra-high field magnetic resonance spectroscopy study of transcranial magnetic stimulation treatment in adolescents with depression

Cicek Bakir, Guglielmo Genovese, Paul A. Nakonezny, Irem Azamet, Julia Shekunov, Jennifer L. Vande Voort, Małgorzata Marjańska, Paul Croarkin

Mayo Clinic, Rochester, Minnesota, United States

Background: Major depressive disorder (MDD) is a leading cause of morbidity and mortality in adolescents. There are open questions regarding the mechanism of action for transcranial magnetic stimulation (TMS) in adolescents with depression. Prior preclinical and clinical studies suggest that TMS treatments address disruptions of excitatory–inhibitory (E/I) balance, particularly glutamatergic and γ-aminobutyric acid (GABA) signaling. Proton magnetic resonance spectroscopy (MRS) enables noninvasive, in vivo quantification of regional neurotransmitter concentrations. However, accurate measurement of low-concentration metabolites such as glutamine and GABA has been challenging at conventional field strengths (1.5T, 3T) due to overlapping resonance peaks. Ultra-high field (7T) MRS improves spectral resolution, signal-to-noise ratio, and chemical shift dispersion, allowing more precise metabolite separation and enhanced neurochemical specificity. Adult MRS studies report reduced cortical GABA and altered glutamatergic metabolites in MDD, but pediatric findings remain scarce and inconsistent. Adjunctive TMS is now FDA-cleared for the treatment of MDD with adolescents. The frequency-dependent effects of TMS on cortical excitability (1 Hz inhibitory, 10 Hz excitatory); and related neurochemical impact in youth remain largely unknown. With these considerations in mind, this study sought to evaluate baseline differences in glutamate and GABA between adolescents with MDD and healthy controls using 7T-MRS and evaluate neurometabolic changes following six weeks of 1 Hz or 10 Hz TMS treatment.

Methods: Adolescents with MDD (n = 48; ages 12–18) and age- and sex-matched healthy controls (n = 23) underwent 7T-MRS of the occipital cortex to quantify GABA, glutamate, and glutamine. Diagnostic status was confirmed with MINI/MINI-KID, and depression severity was assessed with the CDRS-R. All participants were psychotropic medication-free at baseline. MRS were acquired from the occipital lobe using a STEAM sequence, optimized for ultra-short echo times. Metabolite concentrations were quantified using the unsuppressed water signal and corrected by taking into consideration the gray matter, white matter, and CSF content, and T1 and T2 of water. Baseline metabolite concentrations were compared between groups using ANCOVA with age and sex as covariates.

Participants with MDD additionally completed intracortical facilitation (ICF) assessment, were stratified by baseline ICF, and randomized to receive either 1 Hz or 10 Hz TMS targeting the left dorsolateral prefrontal cortex (30 sessions; 2,400 pulses/session at 120% resting motor threshold). Longitudinal changes in metabolite concentrations were examined using linear mixed-effects models with fixed effects for treatment (1 Hz vs. 10 Hz), time (baseline, week 6), and treatment x time interaction, controlling for age, sex, and number of prior failed medications.

Results: Participants had a mean age of 15.4 ± 1.7 years; 61.97% were female, and 58% of those with MDD reported a family history of depression. At baseline, ANCOVA showed no significant differences between MDD and controls in occipital GABA, glutamate, or glutamine concentrations. In the MDD group, linear mixed-effects models revealed no significant main effects of TMS frequency (1 Hz vs. 10 Hz), time (baseline vs. week 6), or treatment × time interaction for any metabolite. Adjusted least squares means showed no significant metabolite changes from baseline to week 6 in either the 1 Hz or 10 Hz TMS groups.

Conclusions: This 7T-MRS study found no baseline differences in occipital glutamate, glutamine, or GABA between adolescents with MDD and healthy controls, and no changes in any of the neurometabolite concentrations following six weeks of TMS treatment. These findings highlight the challenges of identifying reliable neurochemical biomarkers of adolescent depression and TMS response. The findings should be interpreted with caution given study limitations, including the sensitivity of MRS to state-dependent factors (e.g., circadian rhythm) and the uncertain relevance of occipital neurometabolite profiles to MDD pathophysiology. Future studies incorporating additional brain regions are needed to further investigate glutamatergic and GABAergic function in adolescent MDD and TMS response.

Trial Registration: ClinicalTrials.gov, NCT03363919, registered on 30 November 2017

Keywords: Transcranial Magnetic Stimulation (TMS), Magnetic Resonance Spectroscopy (MRS), Major Depressive Disorder (MDD), Glutamate and GABA, Glutamine

Disclosure: Nothing to disclose.

P548. Neonatal functional brain connectivity is associated with observed empathy and callous-unemotional traits at ages 1–3 years

Rebecca Brady, Tara Smyser, Deanna Barch, Joan Luby, Christopher Smyser, Cynthia Rogers, Megan Rose Donohue

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

Background: Callous-unemotional (CU) traits, defined as low empathy, guilt, and prosociality, are a risk factor for severe and persistent conduct problems. There have been a number of studies examining the underlying neurobiology of CU traits, with one prior study examining early brain function. The prior study showed that neonatal functional connectivity between the cingulo-opercular network and medial prefrontal cortex (CO-mPFC) was associated with questionnaire measures of CU traits at age 3 years (Brady et al, 2024). However, this prior study used parent-report measures, which are inherently limited by the heritability of callous-unemotional traits. The current study extends prior work by examining whether neonatal CO-mPFC connectivity is associated with observed measures of emerging CU traits (i.e., low empathy, prosociality, and guilt) assessed even earlier in development via laboratory tasks.

Methods: We recruited 399 pregnant women and performed MRI scans on their healthy, non-sedated neonates using resting state fMRI (n = 319 with > 10 minutes of high quality, low-motion data). CO-mPFC seed-based connectivity from Brady et al., 2024 was used. Longitudinal measures included laboratory observations of empathy, prosociality, and guilt tasks at ages 1, 2, and 3 years. A confirmatory factor analysis was used to create composite measures of low empathy and low prosociality at all ages (n = 242, observations = 452) and emerging CU at ages 2 and 3 years (n = 208, observations = 299). Multilevel models tested longitudinal associations between CO-mPFC connectivity and observed measures.

Results: Neonatal CO-mPFC connectivity was associated with observed emerging CU traits (standardized beta = 0.15, CI = 0.01–0.30, p = .038), controlling for gestational age at birth, postmenstrual age at scan, sex, age at in-person visit, income-to-needs, and racial discrimination. In specificity analyses, neonatal CO-mPFC connectivity was associated with low empathy (s. beta = 0.14, CI = 0.03–0.25, p = .017), but not low prosociality (s. beta = 0.03, CI = −0.09–0.15, p = 0.598) or low guilt (s. beta = 0.08, CI = −0.06–0.22, p = 0.265). Specificity analyses also controlled for gestational age at birth, postmenstrual age at scan, sex, age at in-person visit, income-to-needs, and racial discrimination.

Conclusions: Neonatal CO-mPFC functional connectivity is associated with observed measures of emerging CU traits, specifically low empathy, extending parent report findings. The validity of neural correlates of CU traits evident at birth is further strengthened by relations with gold standard observational measures of empathy, prosociality, and guilt in the first years of life. This study also provides convergent validity that neural alterations precede the development of CU traits. Further understanding of the neural mechanisms that precede the development of CU traits could help identify at-risk children and facilitate development of novel early interventions.

Keywords: Callous-unemotional traits, Infant fMRI, Empathy, Child development

Disclosure: Nothing to disclose.

P549. Task-based functional MRI meta-analysis across risk for psychotic, mood, and anxiety disorders

Jacqueline Clauss, Zachary Millman, Debbie Burdinski, Joey Rodriguez, Dylan Cugley, Jennifer Blackford, Daphne Holt

Maryland Psychiatric Research Center, University of Maryland School of Medicine, Catonsville, Maryland, United States

Background: Risk factors for psychiatric disorders are both transdiagnostic and specific to individual disorders. For example, the clinical high-risk for psychosis syndrome was initially identified to predict future psychotic disorders; however, more recent research suggests that this syndrome also confers 3–4-fold increased risk for mood and anxiety disorders. Genetic risk for mood disorders is associated not only with risk for mood disorders, but also for psychosis. Functional magnetic resonance imaging (fMRI) can be used to identify common alterations in brain functioning across risk syndromes and identify specific patterns of neural responses associated with disease risk. We used coordinate-based meta-analysis across task-based fMRI studies to identify common and specific changes in brain activation in risk for anxiety disorders, mood disorders, and psychotic disorders.

Methods: A meta-analysis of functional neuroimaging studies comparing a group at high-risk for developing an anxiety disorder, a mood disorder, or a psychotic disorder, to a control group or low risk group was performed. Risk was defined based on the presence of early or attenuated symptoms, familial risk, or temperament-based risk. The analysis was limited to studies which used task-based fMRI whole brain analysis, and with a mean age of <30 years old. Studies were excluded if they were not peer reviewed, not published in English, if the high-risk sample was selected based on a specific genetic profile (i.e., 22q11 syndrome), and if they did not include a comparison between a high-risk sample versus a control or low-risk sample. Analyses restricted to a priori regions of interest were excluded. Studies were identified through a search of manuscripts available on PubMed and Web of Science from January 2000 through July 2025. References from studies and known meta-analyses were examined to identify additional studies for inclusion. 7,392 studies were screened and discrepancies were determined by consensus. Coordinates of the peak voxel of significant between-group differences were extracted. Coordinates were entered into an activation likelihood meta-analysis using GingerALE software. ALE uses peak voxel coordinates and sample size to create a map of findings. Common functional activation differences across risk groups were identified by computing maps of differences between risk groups and control or low-risk groups. Contrasts were thresholded at p < 0.05 FWE cluster-corrected using 1000 permutations.

Results: One hundred and three studies were selected for inclusion, representing 5441 subjects (2715 high-risk and 2726 controls), including 12 studies of anxiety disorder risk, 25 studies of mood disorders risk, and 66 studies of psychosis risk. Broadly, across all risk samples, there were differences in functional brain activation in regions associated with emotional processing and regulation, including a cluster including the left anterior insula, dorsolateral prefrontal cortex (max ALE value = 0.034, Z = 5, p < 0.0001) and a second cluster including the right caudate, putamen, medial dorsal nucleus of the thalamus, and rostral anterior cingulate cortex (BA 24/32) (max ALE value = 0.0376, Z = 4.6, p < 0.0001). Mood disorder risk samples showed greater activation, compared to controls, in right hemisphere subcortical regions, including the globus pallidus, putamen, and amygdala, as well as the right parahippocampal gyrus and right inferior frontal gyrus (max ALE value = 0.014, Z = 3.45, p < 0.0001). Participants with risk for psychosis showed increased activation, compared to controls, in areas of the bilateral posterior default mode network, including the posterior cingulate and precuneus (max ALE value = 0.024, Z = 3.85, p < 0.0001), as well as the left dorsolateral prefrontal cortex and insula (max ALE value = 0.026, Z = 4.08, p < 0.0001). There were no regions showing significant differences in activation compared to controls in the anxiety risk group.

Conclusions: Common abnormalities across the salience network, default mode network, and central executive network were identified across risk samples. Psychosis risk maybe linked more specifically to changes in default mode network function and mood disorder risk to changes in subcortical functional activation. Broadly, these findings identify common underlying vulnerability patterns as well as disorder-specific patterns. Future studies using longitudinal designs will be necessary to confirm these patterns.

Keywords: Functional MRI (fMRI), Meta-analysis, High-risk, Affective disorders risk, Psychosis risk

Disclosure: MoonLake Therapeutics, Consultant, Self

P550. Normative development of insulo-cortical functional connectivity in youth aligns with insular microstructure

Elizabeth Flook, G Shafiei, AC Luo, DM Barch, M Cieslak, J Goldsmith, T Salo, LH Somerville, VJ Sydnor, RT Shinohara, LQ Uddin, TD Satterthwaite

Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, United States

Background: The insula plays a critical role in psychiatric disorders emerging in childhood and adolescence. Characterizing normative insula neurodevelopment is a foundational step towards understanding how trans-diagnostic psychopathology arises in youth. Here we define normative developmental patterns of insulo-cortical functional connectivity in youth using a large, cross-sectional neuroimaging dataset.

Methods: We used resting-state functional MRI data from the Human Connectome Project - Development (n = 562, age 6–22 years, 53% female) in this study. Data were processed using fMRIPrep and XCP-D. All available resting-state scans for each participant were concatenated; the final sample only included participants with > 6 minutes of data. We first evaluated the development of insula functional connectivity using parcellated data. Our atlas-based analysis used the HCP-MMP parcellation, which includes 12 insula regions and 348 non-insula cortical regions. The asymmetric matrix from the atlas-based approach included 4,176 edges. For each edge we evaluated development of insulo-cortical functional connectivity using Generalized Additive Models (GAMs) with penalized splines.

To identify major developmental patterns of connectivity we used a variant of a powerful multivariate data reduction method—population value decomposition (PVD)—to summarize patterns in high-dimensional data. In brief, PVD involves a multi-step PCA approach that allows for data reduction while retaining interpretability. For this study, the PVD identified distinct patterns of developmental trajectories which were then compared to both the cortical hierarchy defined by the sensorimotor-association axis (S-A axis) as well as canonical resting-state networks using spatial autocorrelation-preserving spin tests.

As a final step, to characterize developmental connectivity patterns more precisely within the insula, we also implemented a higher-resolution, vertex-based analysis that considered 4.3 million insulo-cortical edges. Based on the results of the PVD, vertex-wise developmental effects were evaluated using polynomial regression; these effects were summarized using principal component analysis (PCA). Given the insula’s complex histology, we examined whether the age effects identified in the high-resolution analysis aligned with established gradients of insular microstructure as defined by the BigBrainWarp toolbox. Due to the limited field of view of the insula, we used Moran randomization for statistical testing of alignment with insula microstructure.

Results: Significant age effects were observed in 45.3% of insulo-cortical functional connections (FDR-corrected p < 0.05). The first two PVD components grouped these age effects into both linear and quadratic associations with age. Notably, age effects captured by linear PVD component 1 showed opposite loadings at posterior and anterior insula – indicating dissociable developmental effects. In the posterior insula, age effects summarized by PVD component 1 were anticorrelated with the S-A axis (posterior 1: r = −0.313, p = 0.047; posterior 2: r = −0.386 p = 0.039; granular insula; r = −0.374 p = 0.042) such that functional connectivity with sensorimotor regions strengthened with age whereas functional connectivity with association cortices weakened with age. In contrast, age effects in anterior insula regions did not align with the S-A axis – but instead were enriched within specific canonical functional networks. For example, developmental increases in functional connectivity were enriched between anterior insula regions and the dorsal-attention network (agranular mean PC1 = 0.017 p = 0.006, anteroventral mean PC1 = 0.056 p = 0.010). Finally, the high-resolution vertex-wise analyses allowed us to relate patterns of insula development to insula microstructure from BigBrainWarp. This revealed that the anterior-posterior gradient of insula developmental functional connectivity significantly aligned with insula microstructure (r: −0.47, p < 0.05).

Conclusions: This study provides a large-scale characterization of normative insulo-cortical functional connectivity development across childhood and adolescence. By leveraging advanced data-driven methods, we identified dissociable profiles of anterior and posterior insula functional connectivity development. These findings highlight that insulo-cortical development is not homogeneous but instead organized along a consistent anterior-posterior axis that aligns with anatomy and microstructure. Together, these results offer a critical normative benchmark for future investigations into how insula connectivity supports—or deviates during—he emergence of mood and anxiety disorders that onset in youth.

Keywords: Brain development, Insula Connectivity, Resting State Functional Connectivity, Intrinsic connectivity (fMRI), Insula cortex

Disclosure: Nothing to disclose.

P551. Disrupted CSF-pupil associations in individuals with opioid use disorder (OUD)

Şükrü Barış Demiral, Faith Veenker, Kylee Miller, Michele-Vera Yonga, Sarah Abey, Peter Manza, Gene-Jack Wang, Nora Volkow

National Institute on Alcohol Abuse and Alcoholism, Bethesda, Maryland, United States

Background: Disrupted CSF-Pupil Associations in Individuals with Opioid Use Disorder (OUD)

Pupillometry has been used as a biomarker of the autonomic nervous system and has been linked with activity in the Locus Coeruleus (LC), a noradrenergic nucleus crucial for arousal. However, its dynamic associations with CSF and brain BOLD signals and the consequences of opioid use disorder (OUD) on these interactions have not been investigated. Through their respiratory depressant effects and their inhibition of the LC. opioids might impact glymphatic clearance and CSF flow in the brain. Here, we tested the hypothesis, that arousal changes indexed by pupil size would be associated with changes in BOLD activity in the LC and other Ascending Arousal Nuclei (AAN) in an opposing manner to the changes in CSF signals and that this association would be dampened in individuals with OUD.

Methods: Participants: Thirty-eight healthy controls (HC) and thirty-seven OUD participants (undergoing Methadone and Suboxone treatment) (ages 22–64) were recruited and gave informed consent under NIH IRB approval. In one session they were given Methylphenidate (MP, 60 mg oral), and in another a placebo (PL, sugar pill) 2h before the fMRI scans. Exclusion criteria included psychiatric, or substance use disorders (other than OUD for OUD group or nicotine dependence), neurological, and systemic medical disorders, use of psychoactive medication use (other than methadone or buprenorphine for OUD group), and major head trauma. fMRI: Participants completed structural and resting-state, food and drug video fMRI tasks in a 3T Siemens Prisma scanner. fMRI data were acquired using a multi-echo EPI sequence and preprocessed with fMRIPrep and xcp_d routines, including motion correction, normalization, and denoising using tedana. Respiratory variability (RV) was performed with deep learning methods applied on fMRI data (Bayrak et al. 2024). Eye tracking was conducted with ASL long-range system in the fmri scanner. Sessions with problematic eye-tracking data and eye closures > 70% were taken out of the analysis. Final analysis used 225 runs across two sessions and across HC and OUD participants. SPM analysis was conducted using pupil size and RV as regressors in −2s, 0s, and +2s lags. We compared HC and OUD groups with a cluster threshold selected as t > 1.9 (p = .05).

Results: We found that in HC pupil size and BOLD signals in the LC and in Ventral Tegmental Areas (VTA) were highly coupled at 0s lag (VTA cluster t = 2.83; LC cluster, t = 1.93) (Fig 1) whereas in OUD participants these associations were not significant (VTA Cluster t = .8, LC Cluster t = .6). Following the pupil-dependent BOLD signal increase in AAN, the BOLD signal increased in the cortical regions and this activity was accompanied with a CSF signal drop (Fig. 2) (CSF cluster in 4th ventricle t = −2.3, lateral ventricle, t = −2.8). However, this effect was not significant for the OUD group (4th ventricle t = −1.2, lateral ventricle t = −0.8). The pupil-LC and pupil-VTA correlations with MP showed a small non-significant increase compared to PL.

Conclusions: We showed that in HC but not in OUD participants changes in pupil size were linked with changes in BOLD activity in VTA and LC in opposition to the changes in T2* signal in the lateral ventricles (consistent with CSF flow). The effect of MP on this coupling was small and not significant. Agonist effects of methadone and buprenorphine in mu opioid receptors located in LC and VTA may dampen the function of these AAN regions. Also, chronic opioid use might impact respiratory regions, such as the Botzlinger area, which through its connections with the LC could interfere with its function. While limited sample size and the difficulty in collecting good quality eye-tracking data constraints the generalizability of our findings they provide preliminary data on disruptive effects of chronic opioid use on the brain’s glymphatic system.

Keywords: CSF, Pupil, OUD, LC, fMRI

Disclosure: Nothing to disclose.

P552. Methylphenidate compresses cortical hierarchy to enhance visual attention

Dardo Tomasi, Peter Manza, Şükrü Barış Demiral, Weizheng Yan, Kylee Miller, Faith Veenker, Joshua Zhao, Michele-Vera Yonga, Sarah Abey, Michaelene Vandine, Gene-Jack Wang, Nora Volkow

National Institute on Alcohol Abuse and Alcoholism, Bethesda, Maryland, United States

Background: Dopamine (DA) signaling in the striatum is vital for motor control, cognition, and motivation. Methylphenidate (MP), a DA transporter blocker used to treat ADHD, enhances attention and affects large-scale brain networks. This study tested whether MP might promote attention by modulating cortical functional hierarchy, using resting-state fMRI and diffusion embedding. Because MP increases catecholamine levels and catecholamines are thought to influence large-scale brain dynamics, we hypothesize that MP would compress the principal connectivity gradient, enhancing integration across cortical systems and shifting processing toward association areas.

Methods: Thirty-eight healthy adults (ages 22–64, 14 females) participated with informed consent under NIH IRB approval. Exclusion criteria included psychiatric, neurological, and systemic medical disorders, substance misuse, medication use, and head trauma. Participants completed structural and resting-state fMRI (rsfMRI) scans on a 3T Siemens Prisma scanner. Participants completed two sessions on separate days, receiving either placebo (PL) or 60 mg oral methylphenidate (MP) 120 minutes before testing, in a randomized order. A subset also performed a visual attention (VA) task involving covert tracking of moving targets. rsfMRI data were acquired using a multi-echo EPI sequence and preprocessed with fMRIPrep, including motion correction, normalization, and denoising using tedana. Functional connectomes were computed in CIFTI format using the 438-region iCSC atlas. Using diffusion embedding, we derived low-dimensional connectivity gradients based on cosine similarity of parcellated connectomes and selected the first two principal gradients representing the most dominant, reproducible axes of cortical functional organization. These gradients were aligned to group templates for cross-subject comparison. Linear mixed-effects models tested drug (MP vs. PL) effects, controlling for age, sex, BMI, IQ, and race.

Results: MP administration induced significant compression of the principal connectivity gradient (pFDR < 0.05), such that primary sensory/motor cortices showed increased gradient values, but higher-order association cortices showed decreased values (all |Cohen’s d’s| > 0.5); together these changes suggest that MP reduced functional segregation in unimodal and transmodal regions. MP induced similar changes in subcortical regions including thalamus, striatum, pallidum, and cerebellum. Effects of MP on the secondary gradient, which span sensory domains, were minimal. VA accuracy improved under MP (t = 2.6, P = 0.01), and these gains correlated with MP-induced increases in the principal gradient, particularly in bilateral parietal and attention-related regions.

Conclusions: Our findings show that MP compresses the brain’s principal functional gradient, reducing segregation (i.e., increasing integration) between sensory and association areas while preserving its overall hierarchy. The secondary gradient remained stable, suggesting selective modulation of the unimodal–transmodal axis. Notably, individual differences in MP-induced compression of the principal gradient were correlated with individual differences in MP-related improvements in VA task performance, highlighting a potential neural mechanism through which MP enhances cognition.

Keywords: Methylphenidate, Cortical Hierarchy, Visual Attention

Disclosure: Nothing to disclose.

P553. The SOAR Brain Health Study: Expanding Rural and Appalachian Representation in Large-Scale Neuroimaging Study Across the Life Course

Marybel Gonzalez, Nina Kraguljac, Anthony King, Stephanie Gorka, Grace Maynard-Wentzel, K. Luan Phan, Scott Langenecker, SOAR Coalition

Ohio State University, College of Medicine, Columbus, Ohio, United States

Background: The Adolescent Brain Cognitive Development (ABCD) Study has led the way in recruiting, enrolling, and following a large representative cohort of youth into emerging adulthood. However, rural communities are still not well represented in large neuroimaging studies. Including all demographics in the United States, including rural communities, is important for uncovering the root causes of mental health and substance use, since risks and experiences can differ between rural and urban settings.” The State of Ohio Adversity and Resilience (SOAR) Brain Health Study was designed to address the gap in recruitment of rural communities in neuroimaging studies through concentrated recruitment strategies and a mobile MRI unit to increase access for rural and Appalachian youth and families. We compared the demographic samples of the SOAR and ABCD studies to test the efficacy of recruitment of rural participants using novel mobile technology and concentrated local recruitment strategies.

Methods: The SOAR Brain Health Study is a statewide, family-level investigation of the bio-psycho-social factors of risk and resilience for mental health and substance use. From December 2023 to July 2025, SOAR enrolled N = 1648 families (N = 2433 participants; ages 12 – 84 years) from rural and urban sites across Ohio. A mobile unit was stationed in four rural/farmland/Appalachian locations everyin sequences of 3–4 months, where from March 2024-July 2025. There were also five brick and mortar urban sites that recruited from July 2024-July 2025. Participants completed multimodal assessments including: magnetic resonance imaging (MRI), electroencephalography (EEG), computerized neurocognitive testing, behavioral/psychological surveys, and biological samples. Recruitment was guided by a community engaged approach consisting of >100 consultations with community and public service groups, public health and policy groups, and healthcare, treatment and recovery providers, as well as attendance at local events throughout the state. This community engaged approach was important for recruiting a sample of participants representing all of Ohio, including rural and Appalachian communities. Chi-squared tests of independence were used to compare the demographic composition of the SOAR Brain Health Study sample to the publicly available dataset, the ABCD study, following over 11,000 youth across 10 years starting at ages 9–10 years of age.

Results: The distribution of rural versus urban participants differed significantly between SOAR and ABCD, χ²(1) = 397.82, p < 0.001. SOAR had a substantially larger proportion of rural participants (26.8%, n = 653) compared to ABCD (11.3%, n = 1336), and correspondingly fewer urban participants (68.9% vs. 83.0%). Standardized residuals indicated that SOAR had a greater share of rural participants, whereas ABCD had a greater share of urban participants. Within the White subsample, household income distributions continued to differ significantly between SOAR and ABCD, χ²(5) = 841.14, p < 0.001. SOAR had a substantially greater percentage of White families living at or below the poverty line (<$25,000; 17.2% vs. 4.3%) and in the $25,000–50,000 range (19.6% vs. 7.7%). In contrast, ABCD included a greater share of White families in higher income brackets (> $100,000: 54.1% vs. 29.4%). Standardized residuals showed that SOAR had a greater share of participants in the lowest two income categories, while ABCD was overrepresented in the highest two. A similar pattern was observed when comparing the SOAR youth sample (ages < 20 years) to the ABCD sample.

Conclusions: The SOAR Study demonstrates that intentional, community-engaged methods and the use of mobile neuroimaging technology can successfully expand representation in large-scale brain health research, particularly by reaching rural and farmland/Appalachian populations and families living at or below the poverty line. These recruitment strategies address critical gaps in existing cohorts such as ABCD, where urban and higher-income families are more prevalent. Together, the SOAR Brain Health and ABCD samples provide complementary and powerful resources to investigate the root causes of youth and young adult mental health and substance use, while advancing discoveries to promote health among all.

Keywords: Population Health, Alcohol and substance use disorders, Mental health outcomes, Psychological wellbeing, Human Neuroimaging

Disclosure: Nothing to disclose.

P554. Delayed cortical maturation, but not anticipatory reward activation, predicts addictive screen use in youth

Katie J. Paige, Mike Angstadt, Meghan E. Martz, Luke W. Hyde, Mary M. Heitzeg, Chandra Sripada, Omid Kardan

University of Michigan, Ann Arbor, Michigan, United States

Background: Addictive screen use refers to problematic patterns of screen use, such as social media use, videogaming, and mobile phone use, that include elements of addiction, such as mood modification, tolerance, withdrawal, conflict, relapse, and functional impairment. Addictive screen use has been linked to significant mental health risks, including internalizing and externalizing symptoms, increased risk for alcohol use initiation, eating disorder symptoms, and suicidal behaviors, yet little is known about the neural vulnerability factors that may predispose youth to addictive screen use. We examined two neurodevelopmental pathways highlighted in substance use research for predicting addictive screen use, namely the cognitive control and the reward neural pathways. We hypothesized that lower cortical maturation and lower Nucleus Accumbens (NAc) activation would both be associated with prospective addictive videogaming, addictive phone use, and addictive social media use.

Methods: We utilized resting-state and monetary incentive delay (MID) task fMRI data from the Adolescent Brain Cognitive Development (ABCD) Study® at baseline (Y0; ages 9–10) to predict addictive video gaming (N = 4522), addictive social media use (N = 3330), and addictive phone use (N = 4466) at year-two-follow-up (Y2; ages 11–12). Cortical connectomic maturation was operationalized as distance from early-life and proximity to adult functional networks in an individual connectome and used to index the development of cognitive control. Nucleus Accumbens (NAc) activation in anticipation of reward contrasted with neutral condition in the MID task was used to assess reward processing. All results are adjusted for data collection site, scanner, head motion, age, sex, family income, parental education, family history of alcohol and drug use, total screen time, and general factor of psychopathology. The reported p values are based on permutations and corrected for multiple tests.

Results: With respect to the cognitive control pathway, we found that, above and beyond general screen use and general factor of psychopathology, lower functional connectomic maturation in the cortex at Y0 was predictive of higher addictive video gaming at Y2 (N = 4522, r = −0.10, βadj = −0.06, p = 0.001). This association did not differ by sex despite addictive video gaming being much higher among male participants. Lower cortical maturation was also predictive of higher addictive phone use, but the association was significant above and beyond general screen use only in female participants (βadj = −0.06, p = 0.006). No association between cortical maturation score and addictive social media use was found. Regarding the reward anticipation pathway, we found no association between NAc anticipatory reward activation during MID task at Y0 and prospective addictive use in any of the three domains of screen use at Y2.

Conclusions: Delayed cortical networks maturation measured by resting-state fMRI in 9–10-year-old youth can predict future addictive videogaming and phone use in early adolescence.

Keywords: Addictive screen use, Cortical networks maturation, Anticipatory reward activation, Adolescence

Disclosure: Nothing to disclose.

P555. Shared and distinct brain structures in anxiety, depression, and neuroticism: insights from the UK Biobank

Chelsea Sawyers, Sarah Benstock, Brad Verhulst, Lisa Straub, John Hettema

Virginia Commonwealth University, Richmond, Virginia, United States

Background: Anxiety and depression are highly comorbid and share common clinical symptoms and etiological factors, particularly through overlapping genetic and environmental risk. Neuroticism strongly correlates with both anxiety and depression, and genetic studies suggest it accounts for a significant portion of the shared genetic risk between these disorders. Recent meta-analyses have primarily examined the association of brain morphometry and these disorders separately, with only a single study systematically examining the shared and specific brain regional differences among neuroticism, anxiety, and depression in a relatively small sample of young adults with subsyndromal to syndromal mood and anxiety symptoms. No imaging studies have analyzed the effects of the neuroticism subscales, anxious worry, and depressed affect on brain structure, nor their overlapping results with anxiety and depression.

Methods: This study used data from 46,541 participants from the UK Biobank with available imaging data including 4,355 with a lifetime anxiety disorder and 8,559 with major depressive disorder to investigate the neuroanatomical correlates of anxiety, depression, and neuroticism in men and women with a focus on identifying shared and distinct associated brain regions across these phenotypes. Using structural MRI data, we examined cortical thickness (CT), surface area (SA), and subcortical volume associations with both subscale-specific and full measures of neuroticism and these disorders while controlling for age, sex, intracranial volume, using linear regressions. Significance testing was corrected post-hoc for testing 140 regions per measure using a Bonferroni-adjusted p-value (p < 0.00036).

Results: Findings revealed significant associations between anxiety disorders and thinner CT in the left insula (p = 5.69E-06), posterior cingulate (p = 4.64E-06), and several frontal and temporal regions. An interesting pattern of left side lateralization emerged within the significant anxiety results. Depression was associated with increased right caudate volume (p = 2.23E-04), decreased SA in the pericalcarine (p = 5.74E-05) and cuneus (3.63E-04), and thinner lateral occipital (p = 1.00E-04) and posterior cingulate (p = 9.53E-05) CT. Neuroticism and its subscales showed more widespread brain involvement, with significant associations in areas including the bilateral caudate (p = 6.58E-05), medial orbitofrontal cortex (p = 8.44E-07), and precentral gyrus (p = 5.09E-05). Notably, neuroticism measures demonstrated stronger associations than diagnosis-specific outcomes, emphasizing the utility of dimensional approaches. Only one significant association (thinner left posterior cingulate) was shared between anxiety and depression (p = 4.64E-06 and p = 9.53E-05, respectively), but several trends were observed.

Conclusions: While we observed distinct patterns, such as left hemisphere lateralization in anxiety and associations with the ventral striatum in depression; these phenotypes also demonstrated overlapping findings. Neuroticism revealed more widespread associations than the clinical phenotypes, supporting the value of using dimensional measures to capture nuanced brain-behavior relationships. Our findings further support a partial neurobiological sharing between anxiety and depression with neuroticism providing a nuanced framework for their connections.

Keywords: Anxiety, Depression, Neuroimaging, Neuroticism

Disclosure: Nothing to disclose.

P556. Prior exposure determines effects of amphetamine on D1 receptor availability: a [11C]NNC112 PET study

Evan Gallagher, Irena Bass, Qikai Qin, Hongping Deng, Christin Sander

A. A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, Massachusetts, United States

Background: Dopamine D1 receptors (D1Rs) play a critical role in generating reward-related signals in the brain. Stimulant drugs trigger surges of dopamine that activate D1Rs, and repeated stimulant exposure is thought to result in both short- and long-term changes to D1R availability and behavior. However, the exact nature of these changes is currently unclear, and in vivo data describing the effects of repeated stimulant exposure on D1R availability is lacking. [11C]NNC112 is a positron emission tomography (PET) radioligand that primarily binds to dopamine D1 receptors (D1Rs) in vivo. Although it is well-established that [11C]NNC112 binding can be blocked by D1R antagonists and high concentrations of exogenous D1R agonists, it is currently not known whether high concentrations of endogenous dopamine (DA) can block or otherwise affect [11C]NNC112 binding. The goal of this study was to establish whether [11C]NNC112 uptake can be blocked by endogenous DA, and to characterize the effects of repeated stimulant exposure on D1R availability. Using a repeated amphetamine dose design, we hypothesized that each amphetamine dose would reduce D1R availability through competition with endogenous dopamine, with the second amphetamine dose producing larger blocking effects.

Methods: We performed a series of dynamic [11C]NNC112 PET imaging studies in anesthetized rhesus macaques (n = 4) before and after two doses of the stimulant amphetamine. Each animal received a set of 3 scans over two days: one baseline scan; and a set of two scans after intravenous administrations of 0.6 mg/kg amphetamine, administered 15 minutes before each scan, with a 4-hour gap in between scans. Each PET scan lasted 90 minutes, and reconstructed PET data were motion-corrected, and registered to standard atlas space, from which regions of interests were determined. D1R availability (i.e. non-displaceable binding potential, BPND) was quantified using kinetic modeling with a multi-linear reference tissue model (MRTM2) and cerebellum as a reference region.

Results: The effects of amphetamine on D1R availability were bidirectional, regionally distinct, and correlated with baseline receptor levels and stimulant exposure history. In the striatum, the first amphetamine dose decreased BPND by 24% in one animal, whereas BPND paradoxically increased by 9–32% in three others. Similarly, the second amphetamine dose produced mixed effects: a 4% reduction in striatal BPND in one animal vs. a 29–88% increase in the other animals. In the cortex, amphetamine-induced D1R blocking effects were more consistent: The first amphetamine dose reduced cortical BPND by 5–60% in three animals while only one animal showed an increase by 27%. The second amphetamine dose produced comparable patterns, with three animals showing a decrease in BPND (11–44%) and one animal showing an increase by 15%. Interestingly, the direction of BPND change correlated inversely with baseline D1R availability: Animals with higher baseline BPND values generally showed amphetamine-induced reductions in BPND, while those with lower baseline BPND values exhibited increases. Additionally, we noted that animals with previous stimulant exposure history showed BPND increases following acute amphetamine, whereas stimulant-naïve animals demonstrated reductions in BPND.

Conclusions: Our results suggest that amphetamine-induced changes in D1R availability are influenced by both short- and long-term stimulant exposure. Repeated acute amphetamine doses predominantly enhanced D1R availability in the striatum but reduced D1R availability in the cortex. However, response direction depended on baseline D1R availability: Animals with lower baseline levels of D1R availability showed amphetamine-induced increases, while those with higher baseline levels exhibited reductions. Long-term stimulant exposure history appeared to influence the response direction and magnitudes, as well as lower baseline D1R availability. These findings suggest a compensatory mechanism specific to D1 receptors whereby repeated stimulant exposure may reduce baseline D1R availability and predispose D1Rs to paradoxically upregulate during acute challenges. Additional studies are needed to confirm these relationships and explore other biological factors that may influence amphetamine-induced changes in D1R availability.

Keywords: Positron Emission Tomography (PET) Imaging, D1 dopamine receptors, Amphetamine, Non-human primates, Dopamine

Disclosure: Nothing to disclose.

P557. [C-11]ARMI: a novel radioligand for imaging alpha-1A adrenoceptors with positron emission tomography, and its evaluation in knockout mice and non-human primates

Garth Terry, Sridhar Narella, Jeih-San Liow, Sanjay Telu, Mudasir Maqbool, Matilah Pamie-George, John Grierson, Kenneth Stout, Sami Zoghbi, Murray Raskind, Victor Pike, Eric Petrie, Robert Innis

VA Puget Sound Health Care System and University of Washington, Seattle, Washington, United States

Background: Norepinephrine and the alpha-1 adrenoceptor (alpha1-AR) are implicated in several disorders of the central nervous system (CNS), including posttraumatic stress disorder (PTSD), Alzheimer’s disease, and postconcussive headache. Prazosin, a non-selective alpha1-AR antagonist, has demonstrated promise for treatment of each of these disorders in clinical trials. Of the three alpha1-AR subtypes, the alpha1A-AR has greatest expression in brain regions associated with fear conditioning and PTSD, such as the amygdala, hippocampus, and prefrontal cortex. We developed an alpha1A-AR-selective radioligand to probe its expression in living brain.

Methods: [11C]ARMI was imaged in mice lacking alpha1A-AR (adra1a-/-; KO; n = 6) and their wildtype (WT) controls (n = 5) to confirm receptor subtype selectivity. Differences between control and test conditions were compared by area under the curve (AUC) of resultant brain time-activity curves. [¹¹C]ARMI was also evaluated in monkeys (n = 5) at baseline and after treatment with prazosin (to assess engagement with α1-ARs), and elacridar (to screen for blood-brain barrier [(BBB)] efflux transporter sensitivity). Arterial blood was sampled concurrently to provide radiometabolite-corrected input functions. Brain region total distribution volume (VT) was derived from a two-tissue compartment model (2TCM) and evaluated for time stability.

Results: In mice, which have ~1/10th the alpha1A-AR Bmax of non-human primates, KO had 44% less brain uptake than WT controls (p = 0.001). [¹¹C]ARMI showed good uptake in monkey brain. A significant blocking effect was observed after prazosin 0.5 mg/kg i.v. (binding potential in whole brain: 5.16 ± 1.19; and occupancy: 75% ± 2.2%). [¹¹C]ARMI also showed time-stable specific signals from 70 to 120 minutes after injection. [¹¹C]ARMI was a weak substrate for efflux transport at the BBB in monkeys.

Conclusions: [11C]ARMI is a novel CNS penetrant radioligand with subnanomolar affinity for alpha1A-ARs. It is alpha-1 AR subtype selective as demonstrated in genetic KO mice. It can be reliably quantified and was sensitive to blockade by prazosin in non-human primate brain.

Keywords: Alpha-1A adrenoceptor, prazosin, CNS Radiotracer, Positron Emission Tomography (PET) Imaging, Alpha-Adrenergic

Disclosure: Merck and Co, Stock / Equity - Publicly Traded Company, Self

P558. Mesoscopic electrophysiology (MePhys): the missing functional link in neuropsychiatric drug discovery

Tobias Teichert

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Most neuropsychiatric drugs are delivered systemically and have the potential to alter neural activity anywhere in the brain. A comprehensive understanding of a drug thus depends on brain-wide measures of neural activity. However, available whole-brain neuroimaging methods either have either limited spatial resolution (electro- or magnetoencephalography) or depend on slow hemodynamic proxies (functional magnetic resonance imaging). To address this methodological need, we developed the first volumetric mesoscopic electrophysiology approach (MePhys) that combines the temporal resolution of electrophysiology with the large field of view and millimeter spatial resolution of fMRI. We showcase the broad utility of MePhys by describing the effect of ketamine on the monkey brain with unprecedented temporal and spatial resolution.

Methods: We developed a volumetric, mesoscopic recording approach (MePhys) by tesselating the volume of an entire monkey hemisphere (including cerebrum, cerebellum, diencephalon, and parts of the mesencephalon) with 992 electrode contacts. The contacts were distributed across 62 chronically implanted multi-electrode shafts arranged into 14 coronal slices with 2-to-7 shafts per slice and 16 contacts per shaft. The contacts formed an approximately regular three-dimensional grid with a pitch of 5 mm between shafts in the antero-posterior dimension and 4 mm in the medio-lateral dimension, and less then 3 mm along the shaft. The treatment of the monkey was in accordance with the guidelines set by the U.S. Department of Health and Human Services (National Institutes of Health) for the care and use of laboratory animals. All methods were approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh. Here we used this system to study the effects of ketamine both at a subanesthetic (1.9 mg/kg) and an anesthetic (7 mg/kg) dose. In addition, we measured the effects of a subanesthetic dose of midazolam (0.39 mg/kg).

Results: Preliminary recordings from our first prototype show that MePhys is feasible, safe and provides extremely rich recordings of local field potentials over a period of at least one year, with no noticeable degradation over time. First, we show that a bolus of an anesthetic dose of ketamine created a profound and long-lasting (> 40 minutes) increase of power in the gamma and delta band as well as a concurrent decrease in the beta-band. These long-lasting drug-induced changes were interrupted approximately every 10 seconds by a ~1-second-long microstate that was defined by a sudden and profound loss of gamma-band power in parietal cortex, as well as a brief and powerful burst of delta-band power in prefrontal cortex and cerebellum. The wide spatial distribution of the this microstate across many far-flung regions would have made it hard to detect and comprehensively describe with traditional microscopic electrophysiology. Second, taking advantage of the ability of MePhys to assess large-scale functional connectivity, we found that a subanesthetic dose of ketamine (but not midazolam) severely disrupted beta-envelope functional connectivity across the motor network, in particular between cerebellum and motor cortex. Furthermore, we found that ketamine (but not midazolam) abolished the emergence and recurrence of stable brain-wide states. Finally, we found that ketamine (but not midazolam) substantially increased the dimensionality of the data, suggesting that a subanesthetic dose of ketamine leads to a generalized state of functional disconnection.

Conclusions: Using ketamine and midazolam as examples, we show that MePhys can provide a comprehensive description of how a neuropsychiatric drug alters brain function and connectivity at the mesoscopic level. Such a mesoscopic functional fingerprint can complement and enrich the description of a drug in terms of receptor affinity, pharmacokinetics, and pharmacodynamics. Over the next years, we will compile an archive of mesoscopic functional fingerprints for a wide range of novel and established drugs across all classes of neuropsychiatric medications. We will use this archive to identify substances that are most likely to counter specific changes to the functional connectome that have been described in several neuropsychiatric conditions.

Keywords: Non-human primates, Novel methods, Resting State Functional Connectivity, Ketamine, Neuroimaging

Disclosure: Nothing to disclose.

P559. Longitudinal in vivo imaging of S-ketamine–induced synaptic plasticity in rats

Oscar Solís, Ingrid Schoenborn, William Dunne, Anna Tischer, Mike Michaelides

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

Background: Synaptic vesicle glycoprotein 2A (SV2A) positron emission tomography (PET) imaging with [¹⁸F]SynVesT-1 enables in vivo assessment of synaptic density. (S)-ketamine, a rapid-acting antidepressant, has been proposed to enhance synaptic function, but the temporal profile of its synaptic effects in vivo remains poorly defined. In this study, we aimed to assess the impact of (S)-ketamine on synaptic density at two post-treatment intervals using [¹⁸F]SynVesT-1 PET in rats.

Methods: Female Sprague–Dawley rats received a single intraperitoneal injection of (S)-ketamine (10 mg/kg; n = 9) or saline (n = 8). Animals were injected intravenously with 20 MBq of [¹⁸F]SynVesT-1, and 60-min dynamic PET scans were acquired at 24 h and 1 week after the treatment. Synaptic density was quantified using standardized uptake value ratio (SUVR) analysis with whole brain as reference. Region of interest analyses focused on cortical and limbic regions.

Results: No differences in [¹⁸F]SynVesT-1 uptake were detected between (S)-ketamine and saline groups at 24 h. At 1 week, significant increases were observed in the medial prefrontal cortex (p < 0.05) and entorhinal cortex (p < 0.01) of (S)-ketamine–treated rats compared to controls. No other regions showed significant changes.

Conclusions: A single dose of (S)-ketamine did not alter synaptic density at 24 h but increased [¹⁸F]SynVesT-1 uptake in medial prefrontal and entorhinal cortices one week later. These findings suggest that ketamine’s synaptic effects emerge in a delayed manner and persist beyond its acute pharmacological actions. Longitudinal [¹⁸F]SynVesT-1 PET provides a translational tool to characterize the time course of drug-induced synaptic plasticity. Future studies will extend this work to male cohorts and include histological confirmation to validate the PET findings.

Keywords: Esketamine, Synaptic Plasticity, F-18 PET Imaging

Disclosure: Nothing to disclose.

P560. A novel ELISA for semi-quantitative detection of IgG, IgA, and IgM autoantibodies targeting the N-terminal domain of the GluN1 subunit of NMDA receptors in neuropsychiatric and neurologic disorders

Shanni Yamaki, Roghiye Kazimi, Vallari Eastman, Andy Martinez, Amir Razai, Scott Snipas, Marty Jefson, Sankalp Gokhale, Peter Flynn, Cathrin Rohleder, Dagmar Koethe, Franz-Markus Leweke, Kotaro Hattori, Mitsuyuki Matsumoto

Arialys Therapeutics, Inc., La Jolla, California, United States

Background: Anti-NMDA receptor encephalitis (ANRE), the most common form of autoimmune encephalitis, is driven by IgG autoantibodies targeting the N-terminal domain of the GluN1 subunit (GluN1-NTD) of NMDA receptors (NMDAR). These patients present with high IgG titers in both blood and cerebrospinal fluid (CSF), along with prominent psychiatric and neurological symptoms.

Interestingly, anti-GluN1 autoantibodies, often of the IgA or IgM isotypes, have also been identified in individuals with schizophrenia, bipolar disorder, depression, dementia, and even in healthy controls, raising questions about their clinical relevance.

However, the current diagnostic standard, IgG-specific cell-based assays, lacks quantitative resolution, has limited sensitivity in blood samples, and is not amenable to high-throughput testing. To overcome these limitations, we developed a novel, semi-quantitative ELISA capable of detecting anti-GluN1-NTD autoantibodies of IgG, IgA, and IgM isotypes in blood and CSF.

Methods: We engineered a high-sensitivity ELISA using recombinant GluN1-NTD (aa 19–403), site-specifically biotinylated at the C-terminus for unidirectional immobilization on streptavidin-coated plates.

To validate assay performance, we synthesized anti-GluN1 IgG monoclonal antibodies based on published sequences from ANRE patients. IgA and IgM isotype variants were generated by grafting the variable regions of monoclonal antibody #003-102 onto human IgA and IgM constant domains. These were used as reference standards to assess assay sensitivity and estimate pseudo-concentrations of autoantibodies in patient blood and CSF samples. Their NMDAR internalization activity was evaluated using HEK293 cells co-expressing GluN1 and GluN2B.

Plasma samples were obtained from patients with schizophrenia, bipolar disorder, unipolar depression, epilepsy, mild cognitive impairment, Alzheimer’s disease, non-AD dementia, and Parkinson’s disease (N = 100 per group), and age-matched healthy controls (N = 200) in Japan via the National Center of Neurology and Psychiatry (NCNP) Biobank (ISO 20387-accredited). Additional serum samples were obtained from patients with schizophrenia (N = 385), bipolar disorder (N = 51), and unipolar depression (N = 81), and healthy controls (N = 117) in Germany (Drs. Rohleder, Koethe and Leweke).

Results: Unidirectional GluN1-NTD protein immobilization enhanced epitope exposure and enabled two-arm binding (avidity), dramatically improving ELISA signal detection, especially for low-affinity antibodies, compared to general random-immobilization ELISA.

IgG monoclonal antibodies #008-218 and #003-102 were selected as reference standards, and #003-102 was isotype-switched to IgA and IgM. Using these IgG, IgA, and IgM standards, semi-quantitative ELISAs for each isotype were established. We confirmed that all #003-102 IgG, IgA, and IgM isotypes induced NMDAR internalization, suggesting potential pathogenicity beyond IgG.

Across both Japanese and German cohorts, we detected variable GluN1-NTD binding activities in all IgG, IgA, and IgM isotypes. Overall, elevated IgG and IgA anti-GluN1-NTD binding was observed in multiple disease groups relative to healthy controls.

Conclusions: While IgG anti-GluN1 autoantibodies are established mediators of ANRE, our data reveal that IgA and IgM autoantibodies binding to GluN1-NTD are also present in diverse neuropsychiatric conditions and can potentially induce NMDAR internalization.

IgA is primarily produced in mucosal tissues, where it plays a critical role in defending against pathogens. However, recent discoveries of gut-educated, IgA-producing plasma cells in the dura mater, along with newly identified dura–brain connections through arachnoid cuff exit (ACE) points and perivascular spaces, suggest that IgA autoantibodies may have a pathogenic role, potentially triggered by molecular mimicry involving microbial antigens such as gut bacteria.

Our novel ELISA provides a robust platform for detecting low-affinity or low-abundance autoantibodies and will facilitate future research into the pathogenic roles of not only IgG but also IgA and IgM NMDAR autoantibodies in neuropsychiatric disorders—and their potential as therapeutic targets.

Keywords: Autoimmune encephalitis, NMDA Receptor, Autoantibody, ELISA, Bipolar Disorder

Disclosure: Arialys Therapeutics, Inc., Employee, Self

P561. Biological and clinical stratification of treatment-resistant depression: the key role of the kynurenine pathway in ketamine response

Moritz Spangemacher, Bruno Pedraz-Petrozzi, Marta Marszalek-Grabska, Emilia Fornal, Maria Gilles, Martin Walter, Alexander Sartorius, Gerhard Gründer, Jonathan Reinwald

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

Background: Emerging evidence suggests that depression subtypes vary in their response to ketamine and its enantiomer esketamine. Stratification strategies are urgently needed to facilitate more personalized treatment approaches. One of the most promising biomarker-based approaches to patient profiling in depression involves the identification of the immuno-metabolic depression subtype. This is of special relevance for treatment with ketamine, which has been shown to exert immunomodulatory effects. One proposed mechanism by which ketamine may influence neuroinflammatory and metabolic processes involves modulation of the kynurenine pathway, the primary route of tryptophan catabolism. Additionally, there is growing evidence that features of the metabolic syndrome, such as increased body weight and systemic inflammation, are linked to dysregulations of these metabolite profiles. Consistently, body mass index (BMI) has been shown to moderate ketamine’s antidepressant effects. Building on these findings, we examined whether pre-treatment peripheral levels of kynurenine pathway metabolites predict response to intravenous ketamine in TRD.

Methods: Thirty-two adults with TRD received six intravenous ketamine infusions (0.5 mg/kg) over three weeks. Biological assessments included kynurenine pathway metabolites (tryptophan, kynurenine, kynurenic acid [KYNA], 3-hydroxykynurenine [3-HK], quinolinic acid [QUIN]). Depression severity (MADRS) and biomarkers were assessed at baseline (D1), after the first infusion (D3), and before the final infusion (D18). Body-Mass Index (BMI) was assessed only at baseline. The study was designed as a single-arm, open-label observational investigation, conducted at the Central Institute of Mental Health (CIMH) in Mannheim, Germany. The protocol was approved by the Ethics Committee of the Medical Faculty Mannheim at Heidelberg University (Registration number: 2021–902).

Results: Higher baseline KYNA predicted greater symptom improvement at D18, remaining significant after FDR correction for multiple comparisons (r = –0.567, p = 0.001). BMI (r = −0.375, p = 0.041) also correlated with D18 improvement. Mediation analysis showed that the association between BMI and symptom improvement at D18 was fully mediated by baseline log(KYNA). The indirect effect explained 70.06% of the total effect (estimate = −90.916, p = 0.026), whereas the direct effect of BMI was non-significant after adjusting for log(KYNA) (p = 0.520), indicating that KYNA mediates the primary influence of BMI on symptom change. In the comparison between responders and non-responders, KYNA achieved an area under the curve (AUC) of 0.759 (p = 0.0014), demonstrating a highly significant discriminatory capacity. Notably, the group differences in KYNA remained significant at D18 and showed no evidence of temporal change or interaction with treatment response over time. This stability suggests that neuroprotective KYNA-related markers represent trait-like predictors. In contrast, TRP was significantly elevated in responders only at baseline, but not at D18, and both TRP and KYN declined modestly over time in this group.

In an exploratory analysis, greater early improvement in hopelessness was significantly correlated with within-person changes in neurotoxic kynurenine pathway metabolites. Percentage changes in QUIN and 3-HK were associated with greater reductions in the MADRS hopelessness subscore (QUIN: r = 0.53, p = 0.003; 3-HK: r = 0.60, p < 0.001), with both associations surviving Bonferroni correction.

Conclusions: Baseline KYNA predicts antidepressant response to ketamine and mediates BMI-related effects. Early shifts in kynurenine metabolites may underlie improvements in hopelessness. KYNA remained stable across treatment, consistent with a trait-like predictor. By contrast, precursors such as TRP and KYN showed weaker or dynamic associations, underscoring that relative pathway balance was more informative than absolute precursor availability.

Given that this is an observational pilot study, our findings should be interpreted with caution and ought to be replicated in larger, randomized controlled cohorts to confirm their validity.

Keywords: IV ketamine, Kynurenic acid, Kynurenine pathway, Psychoneuroimmunology, Treatment Resistant Depression

Disclosure: Nothing to disclose.

P562. Inflammatory cytokines reduce serotonin neuron excitability and enhance in vivo serotonin clearance associate with sex-dependent transcriptome profiles of dorsal raphe Tph2+ neurons

Randy Blakely, Paula Gajewski, Hideki Iwamoto, Zachary Filliben, Alaina Tillman, Mathew Robson, Nicole Baganz, Ning Quan

FAU Stiles-Nicholson Brain Institute, Jupiter, Florida, United States

Background: Central nervous system inflammatory signaling has been associated with the manifestation of neuropsychiatric disorders for many years, though mechanistic aspects of this association continued to be revealed. Of particular interest has been a well-reported link between signaling of IL-1β and p38α MAPK signaling pathways and serotonin (5-HT) signaling. Unlike most other neurons, serotonergic neurons express receptors for IL-1β and TNFα which we hypothesize allows peripheral and central inflammatory states to influence the activity of an integrated set of behavioral state with activation producing changes in mood, cognition, eating, social behavior, reward and sleep. Previously, we reported that peripheral activation of the innate immune system rapidly increases the activity of the presynaptic 5-HT transporter (SERT), with behavioral effects demonstrated to be mediated by serotonergic p38α MAPK signaling. Identifying the cell autonomous effects of IL-1R1 activation on 5-HT neuron excitability and elucidating the projections and physiological consequences of 5-HT expressing IL-1R1 neurons may clarify how the immune system can influence behavior and offer insights into neuropsychiatric disorders and possible treatment strategies.

Methods: In order to visualize the expression of IL-1R1 on serotonergic neurons and their forebrain projecting fibers, we used transgenic mice that allow for both transcriptional and translational reporting of IL-1R1 expression, referred to as “IL-1R1 restore” mice. When crossed to cell-specific Cre expressing mice (e.g. ePet1-Cre for expression in serotonergic neurons), an HA-tagged IL-1R1 protein is expressed, as well as a tdTomato reporter driven by the endogenous Il1r1 promoter, both detected by immunocytochemical approaches. Using acute midbrain slices from these mice, we performed electrophysiological experiments to examine the impact of bath application of IL-1β and TNFα (10ng/mL) on 5-HT neurons from male and female mice. 5-HT neuron activity was monitored using whole cell patch clamp with current injection used to produce tonic firing. These studies were paralleled by recordings where endogenous IL-1R1 was selectively deleted in serotonergic neurons or with application of the p38α MAPK inhibitor MW-150. For in vivo studies, chronoamperometry was utilized to determine the effect of local IL-1β (2ng) on 5-HT clearance in the dorsal hippocampus of wild type male mice. In some studies, LPS (0.2mg/kg intraperitoneal) or saline was acutely administered to adult mice to trigger both peripheral and CNS expression of IL-1β. Brain tissue was collected after transcardial perfusion with 4% PFA three hours after treatment. Slices were immunolabeled with 5-HT and cFos and cell counts were obtained through automated counting using the Nikon NES Elements software. Spatial transcriptomic experiments to evaluate changes in gene expression with Tph2+ dorsal raphe neurons were performed after administration of 0.2 mg/kg LPS, assessed three hours later for transcriptional profiles.

Results: Immunocytochemical analysis of serotonergic HA-IL-1R1 reveals enrichment in the dorsal and dorsolateral subregions of the DRN. Projections of these neurons are evident in the medial prefrontal cortex, striatum, hippocampus, and amygdala among others. Nearly every SERT-positive fiber was found to express HA-IL-1R1 with a density of HA-IL-1R1/SERT co-labeling on each fiber ranged from 40–60% depending on the brain region examined. Investigation of the serotonergic IL-1R1 positive projections revealed dense innervation of the prefrontal cortex, lateral habenula, and hippocampus. Ex vivo slice electrophysiology demonstrated inhibition of 5-HT neuron firing by IL-1β or TNFα that was blocked by MW-150. In mice with a constitutive absence of IL-1R1 except for serotonergic expression retained IL-1β inhibition of firing, demonstrating a cell autonomous effect of the cytokine on 5-HT neuron excitability. In vivo chronoamperometry conducted in wild type mice with local application of IL-1β in the dorsal CA3 of the hippocampus revealed an increase in 5-HT clearance (t(9) = 4.055, p = 0.0037). cFos studies following peripheral LPS treatment reveals an IL-1R1-dependent decrease in serotonergic activation throughout the DRN, with specific subregions impacted in a sex-dependent manner, effects also seen in forebrain projection areas (males: F (1,28) = 5.204, p = 0.0303; females: F (1, 33) = 2.030, p = 0.0379). Finally, spatial transcriptomics analysis of 5-HT neurons the DRN provide evidence that male and female 5-HT neurons have distinct pathways activated in response to peripheral inflammation, with males showing a significantly larger number of differentially expressed genes in response to peripheral LPS compared to females.

Conclusions: Our studies are the first to reveal a sex-specificity of serotonergic IL-1R1 on the regulation of 5-HT neuron activity and transcriptional response to peripheral inflammation. Our findings indicate that IL-1β can modulate 5-HT neurotransmission both by suppression of 5-HT firing that is predicted to reduce 5-HT release, and via regulation of SERT-mediated 5-HT clearance. O These studies will allow us to clarify links between inflammatory and serotonergic signaling with the goal of targeting anatomically and functionally discrete aspects of immune-modulated 5-HT signaling to better define the mechanisms by which inflammatory cytokines regulate behavior and contribute to brain disorders.

Keywords: Serotonin, IL-1b, TNFa, Dorsal raphe, IL-1R1

Disclosure: Nothing to disclose.

P563. Extracellular vesicles as a maternal-fetal communication pathway altered by prenatal stress

Hannah Hazzard, Yesim Ates, Brianna Blaine, Mushroor Kamal, Robert Taylor, Serena Gumusoglu

University of Iowa Carver College of Medicine, Iowa City, Iowa, United States

Background: Maternal stress during pregnancy increases the risk for neurodevelopmental diagnoses in exposed children. This exposure can be modeled in mice using prenatal restraint stress (PS), which consistently produces anxious and avoidant offspring phenotypes. While stress is known to alter maternal and placental biology and modify fetal neurodevelopment, the molecular communication mechanisms underlying these effects remain unclear. Extracellular vesicles (EVs) represent one potential communication pathway between maternal and fetal systems. These lipid-bilayer bound nanoparticles are produced in particularly large quantities by the placenta and carry regulatory micro-RNAs (miRNAs) throughout maternal and fetal circulation. Our unpublished clinical cohort studies reveal that EV miRNA cargos are altered in pregnancies that result in children with neurodevelopmental disorders, with miRNA changes implicating neuroimmune targets. Our prior work also demonstrates that offspring neuroimmune mechanisms, particularly microglia, are altered in PS models. Microglia may aberrantly sculpt fetal brain circuits relevant to neurodevelopmental disorders in PS. Our objective is to determine whether fetal brain microglia interact with placental EVs in vivo, and whether EV impacts on microglia are altered by PS.

Methods: We utilized complementary in vitro and in vivo approaches to test the hypothesis that fetal brain microglia uptake placental EVs and that stress-derived EVs alter microglia transcription. For in vitro studies (n = 3 replicates per condition), BV2 murine microglia cells were exposed to PKH26-labeled EVs were isolated by size exclusion chromatography (IZON) from prenatal stress (PS) and non-stressed (NS) conditions for 12 hours. PS was delivered three times daily for the last week of gestation following established protocols. EV uptake was assessed using spectral flow cytometry and ImageStream analysis (Cytek Aurora5, Amnis Imagestream MkII). For in vivo studies (n = 2 litters per condition), female B6.FVB-Tg(Ada-Cre)5Xiay/J mice (expressing Cre under Ada promoter and chimeric Tpbpa/Ada placental trophoblast enhancer) were crossed with male B6;129S1-Gt(ROSA)26Sortm1(CAG-CD9/GFP)Dmfel/J mice (Cre-inducible His-tagged CD9/TurboGFP reporter targeted to EVs) to produce pregnancies with GFP-positive trophoblast-derived EVs. At late gestation, maternal and fetal tissues (brain, spleen, blood) from both male and female offspring were collected for flow cytometry. Cell populations were identified as macrophages (CD45+), microglia (CD11bhigh/CD45int), and monocytes (CD11bhigh/CD45high), with EV-positive events determined by comparing GFP fluorescent profiles to control tissues (0.5% false positive rate). Statistical analyses included t-tests for comparisons between conditions (significance p < 0.05).

Results: In vivo, fetal brain microglia exhibited preferential uptake of placentally-derived extracellular vesicles (2.47%) compared to other live brain cells (0.29%, p < 0.001) across all tested samples. This preferential EV uptake was specific to brain microglia, as splenic macrophages and circulating maternal macrophages showed no significant preference for EV uptake compared to other cells (1.1% vs 2.4% EV-positive cells, respectively, p = 0.14). In vitro, immorgalized microglia exposed to EVs from PS pregnancies demonstrated altered gene expression with a shift toward M2 phenotype. PS EVs also caused decreased overall EV uptake by microglia (p = 0.002) but increased fluorescence intensity when uptake occurred (p < 0.05).

Conclusions: In vivo studies demonstrate that placentally-derived EVs are preferentially taken up by fetal brain microglia compared to other brain cells or other macrophage types. Immportalized microglia exposed to prenatal stress-derived EVs in culture exhibited altered gene expression and modified uptake patterns. These findings provide evidence that placental EVs represent a targeted maternal-fetal brain communication mechanism that is altered by prenatal stress, suggesting a pathway through which maternal stress may influence fetal neurodevelopment via EV-microglia interactions.

Keywords: microglia, Extracellular vesicles, neurodevelopment, pregnancy, Prenatal Stress Model

Disclosure: Nothing to disclose.

P564. Neonatal Fc receptor-mediated maternal autoantibody transfer disrupts fetal neuroimmune homeostasis and drives sex-biased behavioral alterations

Indra Bishnoi, Izabella Bankowski, Haley Norris, Eva Shin, Margret Slamin, Evan Bordt

Massachusetts General Hospital, Harvard Medical School, Charlestown, Massachusetts, United States

Background: Maternal infection or autoimmunity during pregnancy increases offspring susceptibility to neuropsychiatric disorders. Rodent models of maternal immune activation (MIA) typically involve in utero immune activation using infectious agents such as bacteria or double-stranded RNA. Two of the most common models stimulate toll-like receptor 4 (TLR4) using the bacterial endotoxin lipopolysaccharide or TLR3 using the mimetic poly(I:C). However, some of the prenatal infectious agents that are most strongly linked to risk for neurodevelopmental disorders are single-stranded viruses such as rubella or influenza. Single-stranded mimetics that stimulate TLR7 are relatively underutilized in rodent MIA models. In our model, repeated subcutaneous injection of the TLR7 agonist imiquimod (IMQ) in pregnant mice induces systemic autoimmunity marked by elevated anti-nuclear autoantibodies. These autoantibodies preferentially transfer to male offspring, which show reduced communicative and increased perseverative behaviors. We hypothesize that these autoantibodies are transferred to the fetus via the neonatal Fc receptor (FcRn), altering microglial function and engendering behavioral alterations.

Methods: Pregnant mice are injected with the TLR7 agonist imiquimod (IMQ) or vehicle LAL water throughout mid-gestation. Maternal (4 hours following final injection) and offspring sera (postnatal days 10 or 30) are collected to quantify autoantibodies and cytokines. Fetal (embryonic day (E)16.75) and postnatal day brains are harvested for neuroimmune analyses. Flow cytometry is used to phenotype microglia and assess immune cell infiltration, while FACS enables single-cell RNA sequencing (scRNA-seq) for transcriptomic characterization. To test whether maternal antibody transfer is necessary for offspring neuroimmune/behavioral outcomes, we are 1) genetically preventing autoantibody production (MuMt dams), 2) evaluating genetic ablation of maternal FcRn (FcRn knockout dams), and 3) planning pharmacological blockade via anti-FcRn antibodies (treatment of wild type dams with anti-FcRn antibodies throughout gestation).

Results: IMQ robustly elicits robust maternal autoantibody and cytokine production (p < 0.05). Male offspring of IMQ-treated dams exhibit communication deficits and increased perseverative behaviors (p < 0.05), paralleling male-biased transfer of autoantibodies in offspring from IMQ-exposed dams. Flow cytometry reveals fetal brain immune cell infiltration in both sexes (p < 0.05). IMQ-treated B cell deficient dams (MuMt) produce cytokines without antibodies, while genetic FcRn ablation does not impact maternal antibody production (p < 0.05). Offspring microglial scRNA-seq to reveal distinct gene clusters associated with immune and antiviral signaling is under deeper evaluation. Pharmacologic FcRn blockage is under evaluation.

Conclusions: Our data show that maternal autoimmunity disrupts offspring brain immune homeostasis via FcRn-dependent antibody transfer, driving sex-biased neurobehavioral alterations. These findings identify FcRn as a promising early intervention target to potentially mitigate immune-mediated neurodevelopmental risks linked to psychiatric disorders.

Keywords: Microglia, Autoimmunity, TLR7, Placenta

Disclosure: Nothing to disclose.

P565. Brain immune signaling in response to binge-like drinking

Louis Nuñez, Joselinne Medrano, Angela Ozburn

Oregon Health and Science University, VA Portland Health Care System, Portland, Oregon, United States

Background: Binge drinking is associated with the genetic risk of developing alcohol use disorder (AUD) and is considered a strong predictor of AUD diagnosis. Chronic alcohol consumption leads to increased pro-inflammatory cytokines in brain (in both post-mortem tissue and animal models). These studies suggest immune signaling and inflammation may represent a target for reducing excessive drinking and its harmful consequences. Apremilast, a PDE4 (phosphodiesterase type 4) inhibitor with known anti-inflammatory actions has been shown to reduce binge-like ethanol intake in mouse models of genetic risk for drinking to intoxication (inbred high drinking in the dark mice, iHDID), as well as in non-treatment seeking individuals with AUD.

Methods: In experiment 1, we tested the effects of apremilast administration on early-stage (four days) and initial (one day) binge-like drinking on nucleus accumbens (NAc) cytokine levels. Male and female iHDID-1 mice (n = 7–9/factor) underwent a 4-day DID (drinking in the dark) with either 20% ethanol or water. Mice received apremilast (i.p. 40 mg/kg) or vehicle prior to day 4 DID. Following DID, peri-orbital blood was collected for determination of blood ethanol concentration (BEC). In experiment 2, we tested the effects of an initial exposure to ethanol (with or without apremilast treatment) using a 1-day DID and collected BECs (n = 7–9/factor). Brains were collected and NAc tissue was processed for multiplex cytokine immunoassay in both experiments.

Results: In the 4-day DID experiment, we found apremilast treatment reduced ethanol (p < 0.0001) and water intake (p < 0.01) compared to vehicle controls. We observed a significant main effect of fluid for 12/23 cytokines (IL-1α, IL-1β, IL-3, IL-6, IL-9, IL-17A, IFN-g, TNF-α, eotaxin, MIP-1α, RANTES and KC; all p-values < 0.05), where ethanol drinking mice exhibited higher expression levels. Additionally, a sex x fluid group x treatment interaction was observed for KC (p < 0.05). In the 1 day DID experiment, we found that apremilast decreased ethanol and water intake, with no observed effect on BEC. The multiplex immunoassay results revealed no significant effect of treatment or fluid type. For 18 of the 23 cytokines measured, we observed females have modestly higher NAc levels than males.

Conclusions: A single dose of apremilast was able to successfully reduce initial- and early-stage binge-like drinking. Only one of the 23 cytokines measured, KC (keratinocyte chemoattractant), was modulated by apremilast treatment, alcohol drinking, and sex. KC is a chemokine, a type of cytokine that attracts neutrophils and other immune cells to sites of inflammation.

Moreover, these data show that several pro-inflammatory cytokine levels in the brain are increased by alcohol drinking much earlier than previously thought (with early-stage, but not initial, binge-like drinking). Notably, IL-6 and TNFα were increased with early-stage binge-like drinking. IL-6 (interleukin 6) and TNFα (tumor necrosis factor alpha) have also been found to be increased in animal models of chronic alcohol and in post-mortem tissue of individuals with AUD. IL-6 can influence neuronal function, synaptic transmission, and contribute to neurodegenerative processes. TNFα can regulate neurogenesis, myelination, blood-brain barrier permeability and synaptic plasticity, as well as potentiate neuronal excitotoxicity and CNS inflammation. Thus, the induction of these cytokines in brain by excessive drinking may also contribute to neural dysfunction that promotes the development and persistence of AUD.

Together, these findings support neuroimmune signaling as an early marker and therapeutic target for reducing excessive alcohol consumption.

This work was supported by NIH (AA013519, AA010760, AA007468), US Department of Veterans Affairs (I01 BX0046990, I01 BX006570), John R. Andrews Family/OHSU Foundation, and OHSU Promising Scholars Fund.

Keywords: Apremilast, Neuroimmune, Cytokines, Binge drinking, Alcohol use disorder

Disclosure: Nothing to disclose.

P566. Lipocalin-2 drives persistent neuroinflammation and anxiety-like behavior after infection

Shany Yang, Jennifer Meng, Swapnil Gavade, Joanna Spencer-Segal

University of Michigan, Ann Arbor, Michigan, United States

Background: Post-sepsis survivors commonly face persistent neuropsychiatric complications, such as anxiety, depression, and cognitive dysfunction. Using a murine model of polymicrobial peritonitis, cecal ligation and puncture (CLP), our lab previously showed that CLP survivors develop persistent anxiety-like behaviors alongside chronic innate immune activation in the brain. Transcriptomic analysis identified lipocalin-2 (LCN2), a siderophore-binding protein linked to neuroinflammation and stress responses, as the top upregulated gene in the hippocampus of mice of both sexes. Subsequent RNAscope data from our lab revealed excess lcn2 mRNA in brain endothelial cells, suggesting a vascular contribution to persistent brain immune activation. This study investigates (1) the role of LCN2 in driving post-sepsis neuroinflammation and behavioral changes, and (2) its impact on hippocampal activity in vivo.

Methods: To investigate the role of LCN2 in post-sepsis behavior and neuroinflammation, we used adult male and female lcn2-loxP-CreERT2⁻ and CreERT2⁺ mice to induce global lcn2 knockout following SHAM or CLP surgery, without disrupting the innate immune response to acute infection. Tamoxifen (75 mg/kg) was administered daily for three days, and behavioral testing began ten days after the final injection. Anxiety-like behaviors were assessed using the elevated zero maze (EZM) and open field (OF) tests. Behavioral data were recorded with an overhead camera and analyzed with EthoVision XT 15.0. In the EZM, total distance traveled, time spent and entries into open areas were quantified; in the OF, total distance traveled, time spent and entries into the center zone were measured. Twenty-four hours after behavioral testing, mice were euthanized for molecular analyses. Whole blood was collected, and brain, liver, and spleen tissues were harvested following PBS perfusion. Lcn2 RNA expression was quantified via RT-qPCR, and plasma LCN2 levels were measured using an ELISA.

To assess hippocampal activity, we performed in vivo calcium imaging of dorsal CA1 (dCA1) pyramidal neurons using the nVista 3.0 miniature endoscope. Two adult male mice received stereotaxic injections of AAV5-CaMKIIα-GCaMP6f into the right dCA1 (AP: −2.05, ML: 1.75, DV: −1.3 mm), followed by GRIN lens implantation three weeks later. Mice were then administered either 0.9% saline or recombinant LCN2 (0.5 mg/kg), and neuronal activity was recorded for 10 minutes while the mice were allowed to explore a 40 × 40 cm arena one hour post-injection.

Results: First, we investigated whether our tamoxifen regimen had any effects on locomotion or anxiety-like behavior. Using a control cohort (n = 10 mice per group, 50% males), we found that our tamoxifen regimen had no effects on locomotion and anxiety-like behaviors in the EZM and OF. Next, we compared lcn2-loxP-CreERT2- and lcn2-loxP-CreERT2+ mice after SHAM or CLP surgeries and found that brain lcn2 RNA levels were associated with increased anxiety-like behaviors in the elevated zero maze (fewer entries in the open areas, model estimate = −1.35, p = 0.011; less time spent in open areas model estimate = −5.37, p = 0.0431; n = 8–15 per group). This was also true in the open field, where higher brain lcn2 levels were associated with anxiety-like behavior measured as indicated by reduced time in the center (model estimate = −6.52, p = 0.0483; n = 8–15 per group). Additionally, we found that in CLP mice of both genotypes, brain lcn2 RNA levels were positively correlated with brain RNA levels of two pro-inflammatory targets upregulated by CLP: leucine-rich alpha-2-glycoprotein 1 (Spearman r = 0.6429, p = 0.0040; n = 7–11 per genotype) and p-selectin (Spearman r = 0.5377, p = 0.0214; n = 7–11 per genotype).

To examine the role of excess LCN2 on hippocampal function, we used microendoscopic in vivo calcium imaging and assessed changes in pyramidal cell activity in the dorsal CA1 (dCA1) following acute administration of recombinant LCN2. As expected, circulating LCN2 protein levels were significantly increased in rLCN2-injected mice compared to those vehicle-injected ones (12.7-fold increase; t = 6.909, df = 18, p < 0.0001; n = 10 per group, 50% males). Furthermore, we found that excess LCN2 increased the number of active cells, and the average calcium amplitude in the dCA1 (linear regression model: slope = 0.6530, p = 0.041, n = 2 males, n = 92–760 cells).

Conclusions: Our findings demonstrate that LCN2 is a pleiotropic mediator that contributes to anxiety-like behavior and sustained neuroinflammation following infection, while also increasing dorsal hippocampal neuronal activity in vivo. Our findings highlight LCN2 as a promising target for interventions aimed at improving post-sepsis recovery.

As a next step, we aim to employ an endothelium-specific lcn2 knockout model to dissect the contribution of brain endothelial-derived LCN2 to behavioral dysfunction and chronic neuroimmune activation.

Keywords: Neuroinflammation, Post-sepsis, Behavior, Lipocalin-2, Infection

Disclosure: Nothing to disclose.

P567. Probing the social brain in autism using transcranial magnetic stimulation (tms): towards optimized treatment strategies

Michal Assaf, Christopher Hyatt, Dongyu Kang, Nicholas Balderston, Godfrey Pearlson, Vaughn Steele

Olin Neuropsychiatry Research Center, Institute of Living, Hartford Hospital, Hartford, Connecticut, United States

Background: Mentalizing, or theory of mind (ToM), is a high-order, multifaceted social cognitive process that allows individuals to build representations of others’ state of mind (e.g. emotions and motivation) and to adjust their own behaviors accordingly. As such, it is a crucial process in maintaining effective social interactions, which are fundamental to our survival as social creatures. Mentalizing differences are hypothesized to underline the core social communication differences characterizing autism spectrum disorder (AT) and have been targeted with behavioral interventions, with only partial success. Mentalizing processes are subserved by a specialized neural network, including the medial prefrontal cortex (MPFC), posterior cingulate cortex (PCC), Precuneus (PrC) and bilateral temporo-parietal junction (TPJ, including the posterior superior temporal sulcus, pSTS). We recently showed that the activity in the right TPJ/pSTS is specifically modulated by mentalizing processes, probed with the social-competitive fMRI Domino task, and that young adults diagnosed with AT show decreased mentalizing-related task (MTR) activity in this region compared to neurotypical (NT) controls, which is associated with real-life social communication skills, suggesting it can be an effective target for neuromodulation. Thus, the current ongoing study is designed to validate the right TPJ (rTPJ) as a neural target in AT to be modulated by repetitive transcranial magnetic stimulation (rTMS), to affect neural functions related to mentalizing. We assess the effects of a single putatively excitatory vs. inhibitory (vs. sham) rTMS sequence to the rTPJ on MTR functional connectivity within the mentalizing network.

Methods: We calculate changes in MTR functional connectivity (FC) following acute session of continuous or intermittent theta burst stimulation (cTBS, iTBS, 600 pulses) and sham, administered to the rTPJ in 8 AT and 16 NC adults ages 18–35 (IQ > 80). Participants were engaged in competitive intersocial Domino games while undergoing fMRI scans before and after each TMS session. Generalized Psychophysiological Interaction (gPPI) analysis with rTPJ as a seed region of interest (ROI) was used to assess MTR FC between rTPJ and the other areas of the mentalization network (MPFC, PCC, PrC and left TPJ) before and after each acute TMS session (iTBS, cTBS) as well as sham TMS session. We applied paired t-tests to assess FC changes pre-to-post each TMS session and independent t-tests to assess between-group differences in regions showing MTR FC changes. Cohen’s d effect size was used to assess group effects due to the preliminary nature of these analyses.

Results: Changes in MTR FC within the mentalizing network following TMS across all participants were seen post cTBS session only. We found increased rTPJ-PCC and rTPJ-PrC FC (Cohen’s d > 1.1) and decreased rTPJ-MPFC FC (d = 1.3) across participants. Group differences were found in rTPJ-PrC and rTPJ-PCC FC (d > 0.5), such that AT showed increased FC in the former, and NT showed increased FC in the latter.

Conclusions: Our preliminary results support using rTPJ as a neuromodulation target to engage the mentalizing network, including PCC, PrC and MPFC. They specifically point to the role of cTBS, a putative inhibitory rTMS sequence, but not iTBS (a putative excitatory sequence), in modulating this network. Results further demonstrate differentiation between AT and NT, with group-specific effects in FC between rTPJ and PCC vs. PrC. While we are currently engaged in increasing our sample size, these results support right TPJ-applied cTBS as a potential promising intervention in autistic adults to promote social functions.

Keywords: Autism, TMS, Right temporo-parietal junction, Mentalizing, Social Interactions

Disclosure: Nothing to disclose.

P568. Personalized Adaptive Cortical Electro-Stimulation (PACE) in Treatment-Resistant Depression

Ziad Nahas, Robert Hermosillo, Spencer Eiting, Tariq Hattab, Rachel Johnson, Maya Hazimeh, Seth Konig, Thomas Madison, Cash Differding, Kimberly Weldon, Butool Durani, Jeremy Chaikind, Eric Waldron, Alik Widge, Alexander Herman, Michael Park, David Darrow, Theoden Netoff, Damien Fair

University of Minnesota, Minneapolis, Minnesota, United States

Background: Treatment-resistant depression (TRD) is a leading cause of premature death. For decades investigators have assessed the clinical efficacy of direct brain stimulation for TRD. Outcomes have been inconsistent due to imprecise brain targeting. Minimally invasive Personalized Adaptive Cortical Electro-stimulation (PACE) uses fMRI-based precision functional mapping (PFM) to target patient-specific network anomalies with brain surface electrodes.

Methods: We utilized this novel technology in an ongoing study to treat 2 patients: a 44-year-old man (TRD-1) and a 54 year-old woman with both over 30 years of severe unipolar depression. A third patient is planned for implant October 2025. PFM revealed a 400% expansion in cortical area of the Salience Network (SN) and a 25% reduction in default mode (DMN) and frontoparietal (FP) networks compared to a group average controls. Stimulation paddles were implanted targeting patient-specific SN, DMN, and FP networks. Postoperative stimulation testing (within 24h) revealed immediate electrode-specific mood and cognition responses that matched the underlying functional networks. Stimulation parameters were iteratively optimized using ‘Bayes Tuning’ based on patient feedback.

Results: TRD-1’s suicidal ideations ceased within 7 weeks. Full remission of symptoms was achieved within 9 months and maintained at 34 months. TRD-2 in in her 11 weeks post-implant.

Conclusions: PACE therapy appears to be a scalable, safe, and cost-effective approach to treat TRD.

Keywords: Treatment Resistant Depression, Neuromodulation, precision functional mapping, Adaptive deep brain stimulation, Medial Prefrontal Cortex (mPFC)

Disclosure: Turing Medical, Consultant, Self, LivaNova, Consultant, Self, SigmaStim, Consultant, Self

P569. Poster Withdrawn

P570. The effects of real-time fMRI neurofeedback on emotion regulation in adults with and without anxiety: a pilot study

Stefanie Russman Block, Sophia Hovakimian, Kristin Mannella, Krisanne Litinas, Kate Fitzgerald, David Fresco, Scott Peltier, Steve Taylor

University of Michigan, Ann Arbor, Michigan, United States

Background: Research indicates that anxiety disorders are associated with ineffective emotion regulation. Real-time fMRI neurofeedback (rt-fMRI-NF) is a promising tool that may teach self-regulation and improve anxiety symptoms. Effective emotion regulation reliably engages a set of frontoparietal brain regions, particularly the posterior medial frontal cortex (pMFC) and the left ventrolateral prefrontal cortex (vlPFC). In the present pilot study, we investigated the effect of rt-fMRI-NF targeted at the pMFC and vlPFC on cognitive reappraisal ability (i.e., changing the meaning of a situation to alter emotional content) in healthy adults and adults with anxiety.

Methods: Twenty-nine healthy participants (aged 18–55, 65% Caucasian, 93% Non-Hispanic) and 23 participants with a primary anxiety disorder (social, generalized, panic, and illness; aged 18–52, 91% Caucasian and Non-Hispanic) completed this pilot study. All participants were first taught to use cognitive reappraisal in a 30–45 practice session out of the scanner. Within 1–5 days, participants returned for an fMRI scan with real time neurofeedback, during which they completed the Emotion Regulation Task, adapted from Keller et al. (2021). Instructions were to “look” at negative images without altering their emotional response or “reappraise” to generate fewer negative feelings. Participants completed 5 runs of the task. During runs 2–4, feedback was presented after each reappraise period as a sliding thermometer reflecting percent signal change in the targeted brain region during the most recent reappraise period compared to the most recent look period. No neurofeedback was presented during runs 1 (baseline) and 5 (transfer). After each run, participants rated their current affect on a visual analog scale.

Participants received veritable feedback from the vlPFC (n = 7, 28% anxiety), veritable feedback from the dmPFC (n = 13, 54% anxiety), yoked-sham feedback (n = 6, 100% anxiety) or veritable feedback from a non-reappraisal-related brain region due to an online script error (n = 26, 30% anxiety). Non-reappraisal-related brain regions included left Brodmann area 6, the medial partial cortex, the temporal parietal junction and the posterior insula. Participants with anxiety completed self-report measures of reappraisal use (ERQ) and anxiety symptoms (DASS-21) at baseline, 2-weeks post and 1-month post the fMRI scan.

Linear regressions explored the effects of training group, diagnosis and sex on neuromodulatory ability and neuromodulatory learning in the online targeted brain region (computed online), the vlPFC (computed offline), and the pMFC (computed offline). Neuromodulatory ability was defined as greater brain activity in the reappraise > look contrast, and learning was defined as the linear regression slope of brain activity from the reappraise > look contrast across runs 1 to 5.

Results: The acceptability of the neurofeedback training was high. On a scale of 1–10, participants rated the likelihood of repeating the training as a 7.65 (SD = 1.75, range 3–10), and the likelihood of recommending the training to a friend as a 7.18 (SD = 7.18, range = 2–10).

For online data, participants who received vlPFC and dmPFC feedback showed greater neuromodulatory ability than individuals receiving sham or alternate brain region groups (p < 0.001). Males showed marginally greater neuromodulatory ability than females (p = 0.09), but there was no difference associated with diagnostic status. Collectively, 64% of all participants evidenced a positive learning slope from run 1 to 5, but the magnitude of learning slope did not differ by training group, sex or diagnostic status. The magnitude of the learning slope was negatively correlated with baseline brain activity (r = −0.73), such that participants with lower activity in the target region at baseline showed larger learning slopes over the task.

Across all participants irrespective of training group, offline data showed that a larger vlPFC learning slope was related to a greater decrease in anxiety/nervousness over the scan (B = −1.37, p = 0.005). Similarly, across all participants, better vlPFC neuromodulation was related to a decrease in negative affect over the scan (B = −2.74, p = 0.007), while poorer neuromodulatory ability was related to an increase in negative affect over the scan.

In the subsample of participants with anxiety, reappraisal use increased over time in all training groups (B = 0.32, p = 0.01). Inconsistent with the findings from the scan day, poorer vlPFC neuromodulatory ability was associated with a larger decrease in anxiety severity over time (B = 5.58, p < 0.001), and a smaller vlPFC slope was marginally associated with a greater increase in reappraisal use over time (p = 0.06).

Conclusions: Results confirm the involvement of vlPFC and pMFC activation in emotion regulation and demonstrate that adults can learn neuromodulation, though there may be greater potential for individuals who evidence low target activity at baseline. The temporal relationship between neuromodulation of brain regions involved in emotion regulation and affect should be further explored in a larger sample.

Keywords: Real-time fMRI neurofeedback, Cognitive reappraisal, Anxiety

Disclosure: Nothing to disclose.

P571. Variability in the electric field induced by theta-burst stimulation of the dorsolateral prefrontal cortex in early-course schizophrenia

Francesco Donati, Sabine Janssen, Caitlin Moore, Chloe Huston, Omeed Chaichian, Allison Kim, Ahmad Mayeli, Fabio Ferrarelli

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Background: Repetitive transcranial magnetic stimulation (rTMS) protocols, including theta burst stimulation (TBS), are increasingly utilized as treatment interventions for several neuropsychiatric conditions, including major depression, obsessive-compulsive disorder, and migraine. In schizophrenia, a growing body of work has investigated the role of rTMS as a candidate intervention for ameliorating hallucinations, negative symptoms, and cognitive dysfunction of these patients. However, results have often been mixed, possibly reflecting the high biological heterogeneity of schizophrenia and the absence of optimized stimulation parameters. Indeed, a major source of variability in rTMS outcomes is represented by the intensity of stimulation at the targeted area (e.g., the dorsolateral prefrontal cortex, DLPFC). This intensity is usually chosen by applying a standard, one-size-fits-all correction to the resting motor threshold (RMT, e.g. 110% or 120% RMT). Recent studies employing cortical modeling tools have shown that, when using RMT-based methods for dosing, the electric field (E-field) induced by TMS on the surface of the brain can vary significantly across subjects. In subjects with schizophrenia, this variability may be further exacerbated by disease-related pathological processes, such as prefrontal cortical thinning. However, evidence directly characterizing the TMS-induced E-field in patients with schizophrenia remains limited.

Methods: In an initial sample of an ongoing triple blind, sham-controlled, cross-over, mechanistic clinical trial, we retrospectively calculated the E-field induced by TBS of the DLPFC in 20 subjects with early-course schizophrenia (age: 26.02 ± 5.3; M = 15; F = 5), defined as up to 5 years from their first psychotic episode. In the trial, all subjects received, on 2 separate days, a single session of intermittent TBS (iTBS) of the left DLPFC at 110% of the RMT and a single session of sham stimulation delivered at the same intensity using an active/sham coil. Given the blinded nature of the trial, here we included only subjects who reached the full intensity of 110% RMT for both conditions. Individual structural MRIs were utilized for TMS neuronavigation and E-field modeling. TMS-induced E-field values were calculated using the Simulation of Non-invasive Brain Stimulation (SimNIBS, v4) software employing five-tissue isotropic conductivities. Individual TMS targets defined in MNI space were nonlinearly warped to subject space. The region-of-interest (ROI) for the E-field estimate was set as a 10-mm sphere, centered 2 mm inward along the local grey matter normal. The TMS coil frame used Y along the handle, Z from the local scalp normal at the coil center, orthogonalized to Y, and oriented inward toward the head centroid. To account for skin and hair, a minimum along-normal coil–scalp spacing of 4 mm was enforced.

Results: The TBS intensity of 110% RMT was tolerated by all subjects, with no adverse events. The mean E-field induced by TBS of the DLPFC was 105.6 V/m (SD = 19.1 V/m). Half of the E-field values fell within an interquartile range of 26.1 V/m (Q1 = 93.0 V/m, Q3 = 119.1 V/m). However, there was a large inter-individual variability, with an overall range of 69.6–135.7 v/m. Furthermore, higher RMT values were associated with higher E-fields induced in the DLPFC (r = 0.63, p = 0.003).

Conclusions: Our findings suggest that, in patients with schizophrenia, TMS dosing based on the RMT may result in large interindividual variability of the intensity of cortical stimulation, possibly leading to underdosing and overdosing across subjects. This variability should be carefully considered not only when interpreting treatment effects or designing dose-normalization and power analyses, but also when evaluating safety in subjects receiving antipsychotic medications. Future rTMS trials should therefore consider adopting individualized E-field–guided dosimetry rather than RMT-based methods to ensure more uniform cortical dose delivery across participants and sites, improve interpretability and reproducibility of the findings, and enhance patient safety.

Keywords: TMS, Theta-burst stimulation, e-field modeling

Disclosure: Nothing to disclose.

P572. Accelerated course of intermittent theta burst stimulation for methamphetamine use disorder: feasibility and EEG-derived functional connectivity correlates of treatment response

Nabila Haque, Amrita Ghose, Snoben Kuruvila, Russell Toll, Cherise Chin-Fatt, Abu Minhajuddin, Madhukar Trivedi, Manish Jha

The University of Texas Southwestern Medical Center, Dallas, Texas, United States

Background: Intermittent theta burst stimulation (iTBS) has been shown to reduce craving in individuals with methamphetamine use disorder (MUD). Low retention rates, due in part to housing and transportation challenges, is a common barrier which may be overcome using accelerated iTBS protocols. Concurrent electroencephalograhy (EEG) offers the potential to measure brain network-level effects of iTBS in MUD. Here, we evaluated the feasibility of an accelerated iTBS protocol in MUD and explored EEG-derived connectivity changes with iTBS.

Methods: Participants received 50 sessions of open-label iTBS of left dorsolateral prefrontal cortex (L-DLPFC) over a three-week-period (NCT06424184). Resting state EEG (eyes-closed) was recorded before and after the first iTBS session. EEG-derived power envelope source space connectivity was computed using the Schaefer 100-parcel atlas for default mode (DMN), visual (VIS), somatomotor (SMN), dorsal attention (DAN), salience (SN), limbic (LN), and frontoparietal (FPN) networks. Craving was assessed using a visual analog scale (VAS) before treatment and self-rated ability to control use (11-point Likert scale, 0 = no ability and 10 = complete ability) was assessed after first treatment each day. Depression severity was assessed with the 16-item Quick Inventory of Depressive Symptomatology Self-Report version.

Results: Of the 14 participants [4 females/ 10 males; mean (standard deviation, SD) age: 44.5 (8.6)], 12 completed at least 45 sessions. After three weeks of accelerated iTBS, there was significant reduction in craving [est. (95% CI): −18.79 (−34.20, −3.37)] and QIDS-SR [est. (95% CI): −3.36 (−4.74, −1.97)] and significant improvement in self-rated ability to control methamphetamine use [est. (95% CI): 1.36 (0.28, 2.43)]. A single iTBS session was associated with increased DMN-FPN connectivity (FDR-adjusted p < 0.05). Greater SMN connectivity with LN (rspearman = 0.71, p = 0.015), SN (rspearman = 0.72, p = 0.012) and DMN (rspearman = 0.69, p = 0.018) after the first session was associated with higher self-reported ability to control use.

Conclusions: These findings suggest accelerated protocols enable delivery of therapeutic doses of iTBS to individuals with MUD. iTBS targeting L-DLPFC may increase DMN-FPN connectivity, reduce craving and improve ability to control methamphetamine use.

Keywords: Accelerated intermittent theta burst stimulation (iTBS), EEG connectivity, stimulant use disorder

Disclosure: Nothing to disclose.

P573. Sternal vibration during mindfulness reduces frontoparietal theta and emotional numbness in dissociative trauma-exposed adults: an electroencephalography study

Timothy McDermott, Amanda Johnston, Rebecca Krawczak, Divya Jagadeesh, Christine Savoie, Travis Fulton, Aziz Elbasheir, Negar Fani, Greg Siegle

Florida State University, Tallahassee, Florida, United States

Background: Neurostimulation methods hold promise for rapid symptom relief in psychiatric populations, particularly when paired with beneficial behavioral practices such as meditation. Here, we examined a novel neurostimulation method, sternal vibration, during mindfulness meditation and examined effects on concurrent electroencephalography (EEG) recordings in a sample of trauma-exposed adults with elevated dissociation. Frontoparietal EEG activity was examined as the primary outcome of interest as prior work in dissociative populations has shown either over- or under-modulation of these cortical areas depending on the task context and response of interest. Further, as dissociation is linked to dampened emotional experience and interoceptive awareness, we examined relationships between frontoparietal EEG activity and self-reported ratings of emotional numbing during mindfulness. These data are part of an ongoing clinical trial (NCT04670640) studying the neural mechanisms of four mindfulness interventions with a cross-design of interoceptive (breath focus) and exteroceptive (sternal vibration) feedback. As EEG recordings took place during the mindfulness intervention itself, findings from these preliminary analyses will provide novel insights into mechanistic neural changes throughout the protocol.

Methods: 352 EEG recordings were collected from 47 trauma-exposed adults (mean age = 35.6 years; SD = 11. 7 years; 38.3% identified as men) who volunteered to participate in the mindfulness clinical trial. EEG data were collected using 64-channel HydroCel Geodesic Sensor Net 130 system [Electrical Geodesics Inc. (EGI)]. EEG recordings consisted of six 3-minute blocks of mindfulness practice (18 minutes per visit), and visual analog ratings of feeling “Numb” were collected before and after each block (7 ratings per visit). EEG recordings underwent standard data cleaning procedures and visual quality checking, and activity was extracted and averaged across all six blocks for canonical frequency bands (theta = 4–7 Hz; alpha = 8–12 Hz; beta = 15–25 Hz; gamma = 35–45 Hz). Electrodes of interest were located over the following cortical areas: left lateral frontal (EGI Electrodes: 10–14, 18, 19; included F3, F7, and Fp1), medial frontal (EGI Electrodes: 3, 6, 8, 9; included Fz), right lateral frontal (EGI Electrodes:2, 5, 56–60; included F4, F8, Fp2), left parietal (EGI Electrodes: 25–28, 30; included P3, T5), and right parietal (EGI Electrodes: 42, 44–46, 48; included P4, T6) areas. Activity during meditation was standardized based on a two-second pre-mediation baseline period, and activity was averaged across electrodes of interest for each frequency. A series of linear mixed-effects models were run to test the effects of visit (1–8), presence of sternal vibration (1/0), presence of breath focus (1/0), and their interactions on each set of electrodes for each frequency. These models used maximum likelihood estimation to provide estimations for missing data points and outliers. False discovery rate (FDR) correction was used for multiple comparisons across all models following the Benjamini and Hochberg procedure. “Numb” ratings were also averaged for each visit, and these were added into secondary linear mixed-effects models to test for associations with activity in each frequency band.

Results: Primary findings showed the presence of sternal vibration during mindfulness meditation led to a significant reduction in theta band activity across all frontoparietal electrodes of interest [main effect: F’s > 7.87; pFDR < .012; η_p² > 0.15]. Similar reductions driven by the presence of sternal vibration were observed in right parietal electrodes for both beta [main effect: F = 8.42; pFDR = .019; η_p² = 0.16] and gamma [main effect: F = 8.82; pFDR = .019; η_p² = 0.17] bands. There were no main effects of visit nor interactions between visit and presence of breath focus. However, there were several interactions between visit and presence of vibration that were significant at uncorrected p < .05. This included interactions on frontal beta activity (left lateral: uncorrected p = .0061; medial: uncorrected p = .044; right lateral: uncorrected p = .037) and left lateral frontal gamma activity (uncorrected p = .012). These interactions showed a decrease in the non-vibration conditions with a steady level of activity in the vibration conditions. There was also an association between “Numb” ratings and frontoparietal cortical activity that was significant at uncorrected p < .05, which was a positive association between “Numb” and right parietal theta (uncorrected p = .043; r = .12).

Conclusions: Our findings from this ongoing clinical trial of mindfulness training for dissociative trauma-exposed adults demonstrate that sternal vibration during mindfulness mediation has a marked effect on frontoparietal cortical activity, such that slower wave theta activity was lowered in those who received vibration. However, the long-term effects of this global reduction of theta activity were not clear as there were no interactions between vibration and change in activity across the intervention sessions after corrections. The uncorrected association between right parietal theta and feeling emotionally numb may indicate potential benefits of decreased theta during mindfulness (e.g., greater emotional engagement), indicating this frequency as a relevant mechanism for neurostimulation interventions with psychiatric populations.

Keywords: Electroencephalography (EEG), Mindfulness Meditation, Anhedonia, Neuronal Oscillations, Neurostimulation

Disclosure: Nothing to disclose.

P574. Poster Withdrawn

P575. Relative duration of the non-motor component of ECT-induced seizures as a predictor of treatment response in depression and schizophrenia/schizoaffective disorder: an initial analysis

Lucas Lebovitz, Michael Lazarus, Darby Le, MohammedMehdi Kafashan, Ben Julian Palanca, Nuri Farber

Washington University, St Louis, Missouri, United States

Background: In electroconvulsive therapy (ECT) duration of an induced seizure as measured by bedside EEG (central seizure) has been inconsistent in its ability to predict treatment response. The motor (peripheral) component of an induced seizure also can be also consistently visualized by either optimal dosing of the paralytic agent or use of the cuff-technique. The motor component is shorter than the central seizure by varying degrees, resulting in varying duration of the non-motor component. Here we explored whether the relative time spent in non-motor seizure would correlate with treatment efficacy in patients with either depression (Major or Bipolar) or psychosis (Schizophrenia or Schizoaffective Disorder).

Methods: ECT at our center is typically given as right unilateral with an ultra-brief pulse width (0.25–0.3 msec) at 6x seizure threshold but could be given at a higher charge or with bifrontal or bilateral placement on a Thymatron IV. For each treatment session clinicians record the duration of the central and peripheral seizure. For the following studies, the relative time spent having a non-motor seizure was calculated as (central duration – motor duration) / central duration for each treatment. The mean was then calculated for each subject and compared to the response measure.

In Experiment 1, we conducted a secondary analysis of two prospective observational studies of subjects (n = 29) with major or bipolar depression undergoing routine clinical ECT: CET-REM (NCT04451135) and DNS-ECT (NCT05905705). Treatment response in depression was assessed by QIDS-SR 16. A responder was defined as a patient with a 50% or greater reduction in QIDS-SR 16 by the end of the study period.

Experiment 2 is a retrospective cohort study of all patients at our center with a diagnosis of schizophrenia or schizoaffective disorder who received inpatient ECT since our current medical record system was adopted in 2018. For each identified subject we compared only the first course of ECT in the record during an inpatient psychiatric hospitalization. Diagnosis was confirmed by the discharge summary. Two outcomes of response were defined a priori: 1) length of stay from the beginning of ECT to discharge and 2) Clinical Global Impression Scale – Improvement (CGI-I). CGI-I was performed by a study clinician blinded to documented seizure duration data.

Results: Experiment 1: Overall median (IQR) treatment number was 11 (3). Mean (SD) percent response was 0.49 (0.24), with 55.2% (n = 16) of patients defined as responders within the study period. Percent time with non-motor seizures correlated with degree of response (n = 29, R2 = 0.268, p = 0.004). Non-motor fraction at initial titration session was higher in responders than non-responders (0.47 vs 0.23, t = 1.78, df = 27, p = 0.040).

Neither the absolute duration of the motor component (n = 29, R2 = 0.013, p = 0.56) nor the central component (n = 29, R2 < 0.001, p = 0.89) correlated with percent response. Mean succinylcholine dose across treatment did not correlate with either mean non-motor fraction (n = 29, R2 = 0.004) or percent response (n = 29, R2 = 0.011), indicating that the finding was not due to succinylcholine.

Results for Experiment 2 continue to be analyzed.

Conclusions: In this small prospective exploratory analysis, the relative duration of the non-motor seizure component predicted anti-depressant response. Greater non-motor fraction in the initial titration seizure also predicted eventual response at the end of the study. Absolute duration of the peripheral and central components did not predict response. Whether this relationship also exists for chronic psychotic disorders is currently being explored.

This correlation, if it holds true with further study, could have at least 2 explanations. The first possibility is that the mechanism underlying the termination of the motor component of the seizure is involved in treatment response. Once understood such a situation would suggest that enhancing that mechanism would enhance treatment response. The second possibility is that since the observation was seen with the initial treatment, these subjects are a distinct phenotype who are preferentially responsive to ECT. Efforts at enhancing this mechanism might not necessarily improve response in this case. Furthermore, predicting response to ECT early in the treatment course is a fruitful area of study as identifying likely non-responders early could also allow sparing of unnecessary exposure to anesthesia.

Keywords: electroconvulsive therapy, Major depressive disorder (MDD), Schizophrenia and related disorders (SRD)

Disclosure: Nothing to disclose.

P576. Intermittent theta burst stimulation for attention deficit hyperactivity disorder

Brian Kavanaugh, Megan Vigne, Christopher Legere, Gian DePamphilis, Eric Tirrell, Luke Acuff, Joshua Brown, Richard Jones, Tony Spirito, Linda Carpenter

E. P. Bradley Hospital/Brown University, Riverside, Rhode Island, United States

Background: Attention deficit hyperactivity disorder (ADHD) is the most diagnosed psychiatric disorder in childhood and often causes lifelong symptom burden. Transcranial magnetic stimulation (TMS) has been investigated in adult ADHD with encouraging findings, although work in pediatric samples remains limited and no ADHD studies have examined the utility of intermittent theta burst stimulation (iTBS).

Methods: Twenty-nine adolescents with ADHD and working memory (WM) symptoms were randomized into a sham-controlled, counter-balanced, double-blind, crossover clinical trial of iTBS for adolescent ADHD. Participants completed ten active iTBS sessions (600 pulses per session) and ten sham iTBS sessions to the left dorsolateral prefrontal cortex (DLPFC) at 80% resting motor threshold. Primary outcome variables included safety, feasibility, and change in parent-reported ADHD and WM symptoms. Secondary outcomes consisted of parent and participant-reported affective symptom changes.

Results: The protocol was feasible (82% completed all scheduled sessions) and safe (zero serious adverse events). A statistically significant improvement was seen in active versus sham iTBS in parent-reported overall ADHD symptoms, hyperactivity/impulsivity, working memory, anger, depressive symptoms, and anxiety symptoms.

Conclusions: iTBS holds promise as a potential future treatment for ADHD, and effects achieved when targeting the left DLPFC may be most robust for transdiagnostic cognitive and affective symptoms. Increasing the number of iTBS sessions per day with accelerated protocols may maximize efficacy and feasibility for teens with ADHD and their parents.

Keywords: iTBS, Accelerated iTBS, Working memory, ADHD, Adolescence

Disclosure: Nothing to disclose.

P577. Does entrainment of oscillations with rTMS follow a linear input/output function?

Juliana Corlier, Mohamad Shamas, Thomas E Valles, Andrew Leuchter

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

Background: It has been proposed that the mechanism of action of repetitive transcranial magnetic stimulation (rTMS) is linked to the entrainment of rhythmic brain activity (1, 2) and through that, the augmentation of neuroplasticity (3).

However, it is unclear if brain oscillations can be induced at any given frequency, including the stimulation frequencies commonly used for rTMS treatment, such as 1 Hz, 5 Hz, 10 Hz or 18 Hz. It also remains poorly understood how the stimulation frequencies interact with existing individual endogenous frequencies of the brain and if they should be taken into consideration when administering clinical rTMS treatment for psychiatric or neurological conditions.

We examined the brain’s response to rTMS stimulation in the range between 4 and 18 Hz using source localized scalp electroencephalogram (EEG) to determine a) whether it is possible to entrain any given frequency and b) if the entrainment of oscillations is frequency-specific, i.e. the highest oscillatory increase occurs at/around the frequency of stimulation. We hypothesized that the strongest entrainment of oscillations would occur in the matched frequency band and that the entrainment response profile would represent a linear relationship across all stimulation and response frequencies.

Methods: Eighty-two participants with Major Depressive Disorder underwent a TMS-EEG interrogation procedure in which rTMS was administered to the left dorsolateral prefrontal cortex (DLPFC) at frequencies between 4 to 18 Hz, in 1 Hz steps, in a randomized order. Each stimulation train included 40 pulses with 26 second inter-train-intervals, and each stimulation frequency repeated 5–8 times, totaling 70–90 trains per interrogation. The EEG was recorded throughout the entire procedure.

Changes in EEG oscillations were examined 1s pre to 1s post each train. Source localization was performed using sLORETA. DLPFC signals were extracted, and spectral power was computed in for the response frequencies of 2–20 Hz to determine the rTMS-induced changes for each stimulation frequency. We used bootstrapping to determine the significance of the induced oscillatory responses as compared to baseline, with significance defined as alpha < 0.05, after correction for multiple comparisons.

Results: On average, only 0.5% ± 0.67% of stimulation frequencies across all individual subjects elicited significantly higher power at the matched response frequency than other corresponding frequencies. No patients showed significantly higher responses in at least half of all matched stimulation-response frequencies. These findings suggest there is no evidence for a linear input-output function of rTMS-induced oscillations, and the lack of frequency-specific TMS entrainment response after DLPFC stimulation.

In contrast, we found that 4% of all stimulation frequencies elicited an unspecific broadband entrainment across at least a 5 Hz wide bandwidth (p < 0.05 after correction). These frequencies varied across individuals, with the strongest average effect at the stimulation frequency of 12 Hz, which was present in 12% of subjects.

In some participants, we also observed a robust oscillatory increase shared across stimulation frequencies (p < 0.05 after correction): 4.5% of patients showed a significant increase in one specific response frequency, irrespective of stimulation frequency. The most current response frequency occurring in 12% of subjects was the alpha band (8–12 Hz), which could reflect the dominant alpha rhythm.

Conclusions: These results suggest that rTMS to the DLPFC at any given frequency does not necessary result in an oscillatory entrainment at that same frequency and that entrainment does not follow a linear input/output function. Instead, the entrainment of induced responses may depend on the interaction between the stimulation frequencies and endogenous brain rhythms. Broadband responses can occur in response to several stimulation frequencies, indicating a possibility of multiple “preferred” resonant frequencies of the brain suggesting the possibility that mixed or multi-frequency stimulation may better engage intrinsic cortical networks than a single frequency.

References:

1. Paus T, Sipila PK, Strafella AP. Synchronization of neuronal activity in the human primary motor cortex by transcranial magnetic stimulation: an EEG study. J Neurophysiol. 2001 Oct;86(4):1983–90.

2. Fuggetta G, Pavone EF, Fiaschi A, Manganotti P. Acute modulation of cortical oscillatory activities during short trains of high-frequency repetitive transcranial magnetic stimulation of the human motor cortex: a combined EEG and TMS study. Hum Brain Mapp. 2008 Jan;29(1):1–13.

3. Jannati A, Oberman LM, Rotenberg A, Pascual-Leone A. Assessing the mechanisms of brain plasticity by transcranial magnetic stimulation. Neuropsychopharmacol. 2023 Jan;48(1):191–208.

Keywords: repetitive transcranial magnetic stimulation (rTMS), Quantitative Electroencephalography (qEEG), Neural oscillations

Disclosure: Nothing to disclose.

P578. Low intensity focused ultrasound neuromodulation of the dentato‑rubro‑thalamic tract in subjects with treatment‑resistant essential tremor

Andrea Boscutti, Leo Sugrue, Tommaso Di Ianni

UCSF, San Francisco, California, United States

Background: Low‑intensity focused ultrasound (LIFU) is an emerging neuromodulation technique that enables targeting of deep brain structures with high spatiotemporal precision. Although several groups are investigating LIFU for neuropsychiatric applications, the efficacy of LIFU neuromodulation and its optimal sonication parameters remain to be determined. In this study, we deliver LIFU to the dentato‑rubro‑thalamic tract (DRTT), a target for high‑intensity focused ultrasound (HIFU) thalamotomy, to assess its efficacy in reducing tremor and to characterize LIFU parameter space.

Methods: We performed LIFU of the DRTT in six subjects scheduled to undergo HIFU for treatment-resistant essential tremor. During a single ~2 hour session, each subject received a series of sonications in a block design format, during which various combinations of ultrasound intensity and temporal parameters were tested. Ultrasound was delivered at a frequency of 500 Hz using the NeuroFUS PRO transcranial system (BrainBox Ltd.) guided by neuronavigation (Brainsight, Rogue Research). Anatomical targeting relied on Fast Gray Matter Acquisition T1 Inversion Recovery (FGATIR) MR imaging; when considered of satisfactory quality, DRTT definition via deterministic tractography (FACT algorithm in Brainlab FiberTracking) also informed trajectory planning. Acoustic and thermal simulations were performed using the k-plan software (Brainbox Ltd.). Tremor intensity was recorded during spiral-drawing tasks using a wrist-mounted triaxial accelerometer. Raw data were filtered using a fourth-order Butterworth filter, and the dominant tremor component was extracted via principal component analysis (PCA). Power spectral density (PSD) between 2 and 20 Hz was computed and visualized as spectrograms for each block. Tremor power was normalized per subject by setting the maximum observed power to one. To evaluate the relationship between sonication order and tremor reduction, we fitted a linear mixed-effects model with block number as a fixed effect and a random intercept for each subject.

Results: Six of eight subjects showed a significant linear decrease in tremor power across sonication blocks. At the group level, this trend was confirmed (β = −0.0198, p < 0.01). Although some parameter combinations produced pre‑post block reductions in individual subjects, these effects were not confirmed at the group level.

Conclusions: In this preliminary cohort, LIFU of the DRTT produced a significant reduction in tremor intensity over time. Potential contributors include cumulative ultrasound effects, changes in participant arousal or psychological state, and learning effects from task repetition. We plan to perform sham‑controlled sessions to disentangle these factors and definitively attribute tremor reduction to specific neuromodulatory mechanisms.

Keywords: Neuromodulation, Low Intensity Focused Ultrasound, Essential Tremor

Disclosure: Nothing to disclose.

P579. Single-neuron encoding of reward and choice history in the human GPe during reversal learning

Amber Kiely, Kopell Brian, Zarghona Imtiaz, Helen Mayberg, Alexander Charney, Xiaosi Gu, Ki Sueng Choi, Martijn Figee, Vincenzo Fiore, Andrew Smith

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

Background: Obsessive-compulsive disorder (OCD) is characterized by repetitive thoughts and behaviors, often leading to an inability to adapt in response to changing environments. The basal ganglia, a network of subcortical nuclei, plays a crucial role in cognitive and motor processes, and imbalances between the direct and indirect pathways of this network are implicated in OCD. The indirect pathway may be hyperactive in OCD, leading to excessive suppression of behavioral flexibility, effectively ‘locking’ patients into rigid behavioral states. We sought to investigate the behavioral role of the external segment of the globus pallidus (GPe), a key node in the basal ganglia indirect pathway. By leveraging intracranial single-unit recordings in OCD patients undergoing deep brain stimulation (DBS) surgery, we can assess how neural activity in the GPe contributes to adaptive and maladaptive decision-making during a reversal learning task. Specifically, we predicted that GPe activity will differ based on outcome valence (‘win’ versus ‘loss’). Additionally, we hypothesized that neural activity in the GPe would be modulated during choice selection (‘switch’ versus ‘stay’). Our goal was to generate novel insights into GPe activity, which has never before been recorded in behaving humans.

Methods: GPe neurons were recorded intraoperatively during DBS surgery in patients with severe OCD. In depth evaluation by an expert interdisciplinary team of providers (psychiatrist, psychologist, neuropsychologist, and neurosurgeon) was conducted for each surgical candidate to ensure their capacity and likelihood to benefit from DBS. As part of standard clinical care, each participant undergoes two surgical device implantation operations: left hemisphere, then right hemisphere. The cognitive tasks are administered during each operation and temporally aligned with single unit recordings using a trigger box. Specifically, the participants engage in a probabilistic reversal learning task involving choosing between two slot machines, with reward contingencies reversing twice per block. At the beginning of each block, the ‘winning’ slot machine is randomly assigned at the start as either the left or right option, and it is programed to give the winning outcome 80% of the time and losing outcome 20% of the time. The remaining, ‘losing,’ slot machine is programmed to win only 20% of the time and lose 80% of the time. Neural data was capture from the anterior (limbic) GPe using an Alpha-Omega intraoperative microelectrode system. Spiking patterns were identified using the OSort software package. Then, using a sliding-window general linear model (GLM), we assessed firing rate modulation time-locked to task events.

Results: After processing data from n = 7 individuals, we obtained high quality spiking data for n = 14 distinct neurons. Four neurons exhibited significantly elevated firing 200–400ms after reward delivery when comparing winning versus losing trials (all GLM p-values < 0.05). Of these neurons, three also exhibited significantly increased firing during the 500ms preceding the choice to stay versus switch, suggesting integration of outcome and behavior. Interestingly, one additional neuron encoded choice history alone – firing more at the choice point on trials following wins versus losses – despite showing no relationship to present trial choice or outcome.

Conclusions: These early findings suggest that the GPe encodes executed actions in a multifaceted way. Neurons recorded in the anterior GPe encoded both reward outcomes and behavioral biases. By combining assays of reward-motivated choice behavior with high-resolution intracranial neural recordings, our approach sheds light on the role of GPe in basal ganglia circuitry dynamics, never previously possible in humans. Ongoing work will also incorporate computational modeling of key cognitive processes.

Keywords: deep brain stimulation, obsessive-compulsive disorder (OCD), Single-unit electrophysiology in vivo

Disclosure: Nothing to disclose.

P580. Low intensity focused ultrasound modulates basolateral amygdala task-based activation and functional connectivity in depressed patients

Amanda Arulpragasam, Mascha van ‘t Wout, Yosef Berlow, Emily Aiken, Julia Gilotti, Alison Gorbatov, Nicole McLaughlin, Ryan Van Patten, Stephen Correia, Benjamin Greenberg, Noah Philip

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

Background: Low intensity focused ultrasound (FUS) applies acoustic energy non-invasively to reversibly modulate cortical and subcortical structures with millimeter precision. We are conducting a first-in-human clinical trial of FUS to modulate the amygdala in depressed patients with and without comorbid anxiety (NCT 05147142). This report describes secondary neuroimaging outcomes from the first a priori defined cohort of ten patients.

Methods: Ten patients with depression (Mage = 47; SDage = 13; Nfemale = 2) received MRI-guided FUS targeting the amygdala or control region (S1), separated by ≥1-week in a randomized, single-blind, crossover design. FUS settings were fundamental frequency: 650 kHz, PRF: 10Hz, pulse-width: 5ms, 20 sonications, 30s on, 30s off, ISPTA.3: 719 mW/cm2. Imaging data were acquired on a Siemens 3T Prisma. Eight participants completed the Emotional Faces Task while undergoing MRI using a 20-channel phased array head coil immediately following amygdala sonication and S1 sonication. Ten participants underwent twelve minutes of resting state functional connectivity (rsFC) using a 64-channel phased array head coil at baseline, following amygdala sonication, and following S1 sonication. fMRI data were preprocessed using the fMRIprep standard preprocessing pipeline. Major steps include: 1) realignment, 2) slice time correction, 3) registration to MNI-152 volumetric and FreeSurfer spaces, and 4) spatial smoothing with a 5 mm full-width half-max (FWHM) Gaussian kernel using SPM25. Task data were analyzed using ROI-based analyses in SPM25. rsFC data were analyzed using CONN. Seed-to-voxel analyses were performed with the basolateral amygdala (BLA) as an a priori seed.

Results: The Emotional Faces task revealed robust amygdala activity for the Faces > Shapes contrast following both amygdala and S1 sonication. FUS to the amygdala was associated with decreased BLA activation for Faces > Shapes compared to S1 (left BLA: x = −20, y = −9, z = −28, t = 4.98, p < 0.001; right BLA: x = 25, y = −6, z = −8, t = 2.56, p = 0.019).

Compared to baseline, FUS to the amygdala was associated with decreased connectivity between the right BLA and left postcentral gyrus (x = −22, y = −28, z = 88, pFDR = 0.001) as well the left supplementary motor cortex (x = −6, y = −10, z = 56, pFDR = 0.026). There were no changes to connectivity between right BLA and any other brain regions following S1 sonication compared to baseline. FUS to the amygdala was also associated with decreased rsFC between right BLA and right precentral (x = 48, y = 0, z = 32, pFDR < 0.001) and postcentral gyrus (x = −48, y = −36, z = 62, pFDR < 0.001) compared to S1 sonication.

Conclusions: This is the first active-controlled data indicating that FUS can non-invasively modulate the amygdala in psychiatric patients. These preliminary outcomes suggest clinically relevant target engagement of a subcortical region beyond the reach of other non-invasive neuromodulation technologies. With further study of neural effects and relationship with clinical symptom change, FUS is poised to be a transformative technology for neuropsychiatric research and treatment.

Keywords: Focused ultrasound, Depression, Functional neuroimaging, Basolateral amygdala

Disclosure: Nothing to disclose.

P581. Examining the white matter effects of intermittent theta burst stimulation (iTBS) in people who smoke cigarettes

Donavyn Mosley, Pavan Anand, Dylan Ballard, Rajendra Morey, Seth Himelhoch, Gopalkumar Rakesh

University of Kentucky, College of Medicine, Lexington, Kentucky, United States

Background: Cigarette smoking imposes a significant economic burden and remains the leading cause of preventable deaths in the US. Of those who quit, 60 % relapse after a year. Long-term response to pharmacological smoking cessation interventions is suboptimal. Even varenicline, our most efficacious treatment, has an abstinence rate of only 26% at one year. Transcranial magnetic stimulation (TMS) can augment other smoking cessation interventions. Deep TMS is approved as a short-term smoking cessation aid for adults; it comprises 18 sessions of 10 Hz TMS spread out over six weeks.

Intermittent theta burst stimulation (iTBS) is a potent TMS paradigm that has demonstrated efficacy in reducing craving in people who smoke cigarettes. Optimizing iTBS for smoking cessation requires a precision medicine approach, and there has been limited success in identifying facets that predict treatment response.

Diffusion tensor imaging (DTI) could potentially personalize iTBS for craving reduction and smoking cessation by examining white matter (WM) microstructural changes associated with decreased craving. Previous DTI studies showed WM differences in people who smoked cigarettes compared with those who did not. These differences were quantified using two measurements: fractional anisotropy (FA) and diffusivity (comprising mean diffusivity [MD], axial diffusivity [AD], and radial diffusivity [RD]). The anterior limb of the internal capsule (ALIC) showed lower FA and higher MD and RD in people with tobacco use disorder than those without it. The posterior limb of the internal capsule (PLIC) demonstrated significant correlations with the severity of cigarette smoking.

While no previous studies have examined the effects of iTBS or other TMS paradigms on WM in people who smoke cigarettes, previous studies have shown TMS-related WM changes in healthy controls, MDD, and stroke. These studies have yielded mixed results, with 15 Hz rTMS increasing FA in MDD and decreasing FA in healthy controls. We have previously demonstrated the impact of one iTBS session and accelerated iTBS (four sessions in a day), encompassing 1800 pulses/session, on the left dorsolateral prefrontal cortex (L.dlPFC) compared to sham, on resting-state functional connectivity and cigarette craving in people who smoke cigarettes (NCT04936594, NCT05049460). Both these trials acquired DTI brain scans before and after iTBS/sham iTBS.

Methods: We conducted post-hoc analyses of DTI data acquired before and after iTBS/sham iTBS in two clinical trials - NCT04936594 (iTBS [n = 6] and sham iTBS [n = 5]) and NCT05049460 (iTBS [n = 7] and sham iTBS [n = 7]). All DTI scans were processed through a validated pipeline, details of which can be found at https://github.com/grakesh-lab/dti_processing.

We performed linear mixed-effects (LME) regression analyses using the FA/AD/RD/MD of four WM tracts (bilateral ALIC and bilateral PLIC) as the dependent variables. Fixed effects in these LME analyses comprised the treatment arm (iTBS/sham iTBS), number of iTBS/sham iTBS sessions delivered in a day (one versus four), cigarette craving, measured with the short form of the Tobacco Craving Questionnaire [TCQ-SF] before and after iTBS/sham iTBS, time (before and after iTBS/sham iTBS), and the following interactions: treatment arm × time, number of sessions × time, and treatment arm × number of sessions. The only random effect was the participant ID. We hypothesized that iTBS over L.dlPFC would increase FA and decrease MD, RD, and AD in left ALIC and left PLIC in people who smoke, compared with sham iTBS. We predicted that four iTBS sessions would demonstrate these changes in greater magnitude compared to one iTBS session.

Results: An LME model (LMEM) with FA of left ALIC as the dependent variable and predictor variables as mentioned above showed significant effects for time (t41 = 2.22, p = 0.03), number of sessions (t41 = 2.32, p = 0.03), and number of sessions × time interaction (t41 = 3.38, p = 0.002). LMEM with left ALIC AD as the dependent variable showed significant effects for number of sessions (t41 = 2.82, p = 0.01) and treatment arm × number of sessions interaction (t41 = 2.30, p = 0.03). Planned comparisons revealed that left ALIC FA and AD decreased following one iTBS session and increased after four iTBS sessions.

Both right ALIC and right PLIC showed significant effects for time (t41 = 2.34, p = 0.02) and number of sessions × time interaction (t41 = 2.31, p = 0.03), respectively, on FA. Right ALIC and right PLIC AD showed significant associations with the number of sessions (t41 = 2.18, p = 0.04) and treatment arm × number of sessions interaction, respectively (t41 = 2.00, p = 0.05).

The left PLIC did not show any significant associations between FA and predictor variables; its AD showed significant associations with the number of sessions (t41 = 2.30, p = 0.03).

Craving showed significant associations with the RD of right ALIC and left PLIC. MD did not show any significant associations with predictor variables for any of the tracts.

Conclusions: Increased anisotropy and axial diffusivity of left ALIC point to WM neuroplasticity induced by iTBS, and a decrease in both metrics could indicate WM damage. Four iTBS sessions decreased both metrics in LALIC, whereas one session increased both. Therefore, more sessions and dosing may not be optimal for WM in people who smoke cigarettes. These preliminary results need to be interpreted with caution, given the limited sample size and multiple comparisons. Well-powered studies are required to characterize the WM impact of iTBS.

Keywords: transcranial magnetic stimulation, Accelerated intermittent theta burst stimulation (iTBS), Diffusion Tensor Imaging (DTI), cigarette smoking

Disclosure: Nothing to disclose.

P582. Local theta power changes as a reliable biomarker of iTBS target engagement

Ethan Solomon, Christopher Cline, Umair Hassan, Jade Truong, Winn Hartford, Lily Forman, Corey Keller

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

Background: Transcranial magnetic stimulation (TMS) is now understood to be an effective method for treating several neuropsychiatric conditions, most commonly major depressive disorder (MDD), but clinical outcomes are highly variable and up to 50% of patients with treatment-resistant MDD do not improve with TMS. Discovering individualized, quantitative, and objective measures of patient responsiveness to TMS is therefore essential to refining the current one-size-fits-all approach to clinical TMS. Electrophysiological phenomena, most notably the local TMS-evoked potential (TEP) seen on electroencephalography (EEG), are hypothesized to reflect neuroplastic changes induced by TMS, but these measures are limited by significant signal-to-noise constraints. Rhythmic patterns of stimulation like intermittent theta burst (iTBS) are known to affect neural responses in the frequency domain – primarily sustained oscillations in the theta band – but these signals have not yet been robustly explored as a biomarker of iTBS-induced change in brain activity.

Methods: We recruited 32 healthy adults (all genders) to undergo 3 sessions of neuronavigated intermittent theta burst TMS, either to the dorsolateral prefrontal cortex (DLPFC; N = 24) or the posterior parietal cortex at a target of maximal resting functional connectivity to the hippocampus (PPC; N = 8). Before and after each session of iTBS, the target region was stimulated with repeated single pulses of TMS (spTMS; 1.5s inter-stimulus interval, jittered) spanning low to high intensities relative to motor threshold. 64-channel scalp EEG was collected during all stimulation events. Single-pulse TMS trials were analyzed with spectral methods (multitapers) to ask whether TMS-provoked oscillations change following iTBS sessions. Spectral analysis was conducted on the interval following the extinction of the usual TEP, from 400ms onwards until the next pulse. TMS-provoked oscillations were compared between pre-TBS and post-TBS trials in the stimulated region (local responses) as well as the whole brain (remote or network-level responses).

Results: iTBS caused an increase in spTMS-provoked theta power (4–8 Hz) observed locally in the stimulated region, but only for the first iTBS session in a series of 3 spaced 45 minutes apart (DLPFC: 1-sample t-test, t = 2.25, p = 0.034: PPC: t = 3.35, p = 0.02). In the DLPFC, the effect was strongest at the F3 electrode, which corresponds closest to the targeted area in most subjects. 16/24 (66%) subjects exhibited an iTBS-related increase in theta power at the DLPFC site and all subjects (100%) showed an increase in theta power at the PPC site. After FDR-correction for multiple comparisons, areas remote to the stimulated region did not exhibit a significant iTBS-related modulation of theta power. No significant effect was observed in the alpha (9–13 Hz), beta (15–25 Hz), or gamma (30–50 Hz) bands at either the DLPFC or PPC stimulation sites (all p > 0.05). Moreover, in the DLPFC, the theta enhancement effect was specific to the post-TEP interval, peaking at 400–1000ms post-stimulation; at the PPC site, significant iTBS-related theta modulation was observed at the same 400–1000ms interval as well as during the TEP. Finally, the theta enhancement effect was found to be driven by TMS-provoked power at the highest intensity stimulation; low-intensity stimulation had no significant local effect and a trend towards iTBS-related decreases in theta power (t = −1.82, p = 0.08).

Conclusions: Biomarkers of neural engagement by TMS are crucial for (1) developing optimized stimulation protocols and (2) understanding the biological processes underlying treatment response. Here, using TMS-EEG, we demonstrated that iTBS causes a consistent increase in the occurrence of provoked theta oscillations local to the stimulation site, at both the DLPFC and PPC. This effect was specific in space (occurring local to the target site), time (only following the first iTBS session and following extinction of the TEP), and frequency (observed only in the theta band). Moreover, the effect was highly reliable, having been observed in 72% of participants. Taken together, our findings suggest TMS-provoked theta power may be a specific and reliable biological signature of neural engagement by iTBS. Future work will explore the utility of theta oscillations as a means to fine-tune and individualize TMS stimulation parameters/protocols to optimize clinical response.

Keywords: Theta burst transcranial magnetic stimulation, EEG biomarkers, EEG biomarkers and target engagement, Theta band oscillatory measures

Disclosure: Nothing to disclose.

P583. Improvement in depression and anhedonia following continuous theta burst stimulation of the frontostriatal reward system in transitional-age youth varies with baseline differences in symptom severity and reward circuit activity

Kalyan Tripathy, Helmet Karim, Neal Ryan, Iris Chat, Zachary Brodnick, Ashley Pogue, Neil Jones, Rebecca Price, Jihui Diaz, Erika Forbes

University of Pittsburgh Medical Center, Western Psychiatric Hospital, Pittsburgh, Pennsylvania, United States

Background: An early onset of depression and high levels of anhedonia predict a severe course of illness with poor response to mainstream treatments. Extensive research has identified aberrant reward processing in anhedonia with increased dorsomedial prefrontal cortex activity dampening striatal responses to reward. We hypothesized that modulating the dysfunctional frontostriatal reward circuit using continuous theta burst stimulation (cTBS) to inhibit dorsomedial prefrontal cortex (dmPFC) could alleviate anhedonic depression in treatment-resistant youth. We further sought to explore how treatment response varied with baseline symptoms and neurophysiology.

Methods: Youth (N = 50, 78% female) aged 15–25 years with a depressive disorder diagnosis per the Structured Clinical Interview for DSM-5 (SCID-5), a Montgomery-Åsberg Depression Rating Scale (MADRS) score ≥12, and a history of at least one adequate but failed antidepressant trial (≥6 weeks, with ≥3 weeks at recommended adult dosage) were enrolled in the study. Participants received two sessions per day of cTBS for ten days with an individualized target in dmPFC based on functional connectivity to the nucleus accumbens and with ten integrated sessions of positive affect training. Participants completed questionnaires to assess psychopathology – including MADRS for depression and the Snaith-Hamilton Pleasure Scale (SHAPS) for anhedonia – as well as a functional magnetic resonance imaging (fMRI) monetary reward task to map neural reward responses both before and after treatment. T-tests and linear mixed effects models were used to compare symptoms between pre- and post-treatment time points. Pearson correlation analysis was applied to evaluate how symptom changes varied with baseline symptoms and reward circuit activity.

Results: Participants experienced decreased depression (p = 1e-7, d = −0.94) and anhedonia (p = 0.002, d = −0.48) following treatment. 42% of participants met our threshold for treatment response (>30% reduction in MADRS). Improvement in depression was greater for those with greater baseline depression severity (r = 0.57, p = 4e-5) and weaker initial nucleus accumbens response to reward (r = −0.4, p = 0.008). Improvement in anhedonia similarly increased with the baseline severity of both anhedonia itself (r = 0.47, p = 0.001) and depression (r = 0.38, p = 0.01).

Conclusions: More severe symptoms and reward circuit dysfunction at baseline were associated with greater improvements in depression and anhedonia following a novel treatment paradigm targeting the frontostriatal reward system. Interventions targeting dysfunctional reward circuitry offer particular promise for alleviating treatment-resistant anhedonic depression in youth, warranting further clinical investigation.

Keywords: Anhedonia, Youth depression, Transcranial magnetic stimulation, Frontostriatal circuitry, Neuroimaging

Disclosure: Nothing to disclose.