Craving and maladaptive choices are intertwined across addictive disorders, yet the specific computational mechanisms mediating their interactions remain elusive. Here we tested a hypothesis that momentary craving and reinforcement learning influence each other during substance-related decision-making. Two substance-using groups with moderate to high addiction risk levels (alcohol drinkers and cannabis users; total n = 132) performed a decision-making task in which they received a group-specific addictive cue or monetary outcomes and reported moment-to-moment subjective craving. Computational modeling revealed that momentary craving biased substance-specific learning rate in both groups, but in opposite directions. In addition, expected values and outcomes jointly influenced elicited craving across groups and decision contexts. Finally, regressions incorporating model-derived parameters best predicted alcohol, but not cannabis, addiction risk scores, supporting the selective utility of using these model-based parameters in making clinical predictions. Together, these findings provide a computational framework that accounts for the interaction between craving and maladaptive choices across addictive domains.

Social relationships are best understood not as isolated interactions but as dynamic representations shaped through evaluative processes, interaction-driven learning, motivational value, and memory-based structure. This review synthesizes research across these domains to trace how relationships develop from initial impressions through repeated interactions to stable social bonds and networks. We organize the literature by relationship stage and highlight how overlapping neural systems are differentially engaged as relationships evolve over time. By treating relationships as unfolding processes, we reveal computational principles that become apparent only across time. Building on cognitive map theory, we propose that social relationships are represented as structured relational spaces supporting updating and generalization. We further highlight emerging computational and methodological approaches-including map-based models, naturalistic paradigms, and tools for analyzing dynamic social behavior-that enable the study of relationships across their temporal arc. This framework offers new ways to study the formation, maintenance, and adaptation of social bonds.

Social anxiety (SA) frequently co-occurs with autism spectrum disorder (ASD), but it remains unclear whether SA in ASD reflects greater symptom severity or a distinct subtype. Across two complementary studies, we characterized the behavioral and neural correlates of SA in ASD using a naturalistic social navigation task measuring affiliative and power-related behaviors. In Study 1 (ASD = 575; controls = 357), autistic individuals with SA exhibited significantly reduced task-derived power behaviors (e.g., compliance with others’ requests) compared to those without SA—a pattern not observed in non-autistic individuals. In Study 2 (ASD = 72; controls = 72), individuals with both ASD and SA showed enlarged amygdala volumes relative to both ASD without SA and non-autistic individuals. These behavioral and neural differences found in Study 1 and Study 2 were not observed in ASD with co-occurring depression or generalized anxiety disorder. Moreover, in Study 2, larger amygdala volume in the ASD group was associated with reduced task-derived power behaviors, indicating a neural signature linking social-behavioral differences and anxiety-related features. Together, these findings point to a distinct ASD phenotype marked by social anxiety, altered interpersonal dynamics, and amygdala enlargement, with implications for diagnosis and tailored intervention.

Social difficulties in autism are often framed as reduced motivation, yet this account does not explain when and why autistic individuals affiliate. We show that autism selectively alters the architecture-not the presence-of similarity-based social behavior. Across two independent samples (online, n = 714; in-person, n = 225), autistic and neurotypical adults exhibited comparable context-dependent (selective) preference for relatively similar others. In contrast, autistic individuals showed a markedly stronger global coupling between perceived similarity and affiliative behavior, such that low perceived similarity was associated with sharply reduced affiliation. This effect was strongest among those with lower trait empathy. Structural MRI revealed dissociable contributions of hippocampal and posterior cingulate cortex volumes to this coupling across groups. These findings demonstrate that autism preserves similarity attraction while amplifying its role as a stable heuristic for social engagement, supporting a model in which social motivation is restructured toward similarity-dependent engagement rather than diminished.

Clinical decision-making in psychiatry has traditionally relied on rating scales and clinical impressions documented in the electronic health record (EHR). Yet, clinical interviews contain rich behavioral signals that remain underutilized in psychiatric care. Recent advances in artificial intelligence (AI) now enable quantification of these signals, and prior work demonstrates that computational measures of speech, language, and facial expression can inform diagnosis and estimate symptom severity. Despite this progress, most prediction efforts remain confined to single modalities and individual diagnoses and focus on diagnostic classification rather than clinically actionable outcomes such as treatment discontinuation or the need for crisis care. Here, we contextualize advances in behavioral quantification and multimodal data fusion, and present the Phenotypes REimagined to Define Clinical Treatment and Outcome Research (PREDiCTOR) study, a prospective cohort study of 2100 patients entering outpatient mental health care. PREDiCTOR is designed to develop and validate dynamic, multimodal prediction signatures that predict treatment discontinuation, emergency department visits, and hospitalizations over a one-year follow-up period. The study audiovisual recordings of clinical encounters, EHR data, cognitive assessments, smartphone passive sensing, therapeutic alliance measures, and audio/text diaries within a Contextual Bandit framework that continuously updates individualized outcome estimates as new data become available. Both interpretable features and learned embeddings are leveraged, with large language models serving as feature extractors rather than clinical decision-makers. We describe the study design, data collection, and processing pipelines, hybrid predictive modeling approach, and prospective validation strategy, and discuss the potential for translating dynamic behavioral quantification into individualized clinical prognostics.

The annual Society for Neuroscience (SfN) meeting is a bonanza of scientific achievement: famous keynote speakers, beautiful scientific results, and award ceremonies. This focus is exciting and invigorating but glosses over the many failures, mistakes, and rejections that typically lead to scientific success. Our goal has been to create a space within the annual SfN meeting for open conversation about scientific failure and, by doing so, increase transparency, resilience, and mental well-being within our community. In this article, we share the materials that we have used at SfN during the past 4 years (2021-2024) to promote discussions of scientific failure, including formal storytelling, individual and interactive games, and confessionals. For each activity, we provide the rationale and practical guidance regarding logistics and usage. We hope this will aid scientists interested in adapting the activities for their own communities or local events. We end with a call for scientific institutions to commit to providing space for open discussions of failure within their educational programs and conferences.

Background As we navigate changing social landscapes, maintaining maps of interpersonal dynamics can help guide our choices. Autism spectrum disorder (ASD) is associated with social challenges that may affect the accumulation or application of social information. However, little is known about social cognitive mapping in autistic adults. Methods Herein, we investigated differences in social navigation among 122 adults with ASD, typical development (TD), and misophonia (MIS) (included as a clinical comparison group) using a social interaction task during functional magnetic resonance imaging. Results Compared with other groups, adults with ASD behaved socially distant from task characters. Nevertheless, the groups displayed comparable neural tracking of social distances in regions previously identified in nonclinical samples, including the posterior cingulate cortex (PCC), as well as the parahippocampal place area, where tracking uniquely related to cross-diagnostic social avoidance symptoms. In contrast, the ASD group showed distinctive hypoactivity in the temporal pole (TP) during social decisions, associated with smaller real-world social networks and reduced insight into their external symptoms. Additionally, while the TD and MIS groups showed functional decoupling between the TP and PCC during social decisions, this was not detected in ASD. Conclusions Adults with ASD showed distinct behaviors and neural activity during deliberation in social interactions. However, brain systems supporting social mapping appear preserved across groups, consistent with previous findings, which have now been extended to a clinically diverse sample. These results highlight both shared and ASD-specific neural mechanisms of social navigation, thereby offering insight into potential neural differences in how social evidence guides choices in ASD.

Drug-related memories can hinder abstinence goals in drug addiction. Promoting nondrugmemories via ventromedial prefrontal cortex (vmPFC)-and amygdala-guided extinctionyields mixed success. Postretrieval extinction (RE) destabilizes and updates memoriesduring reconsolidation, improving extinction. Supplementing RE, we tested methylphenidate(MPH), a dopamine agonist that promotes PFC-dependent learning and memory in cocaineuse disorder (CUD). In a proof-of-concept double-blind randomized clinical trialusing a within-subjects design, participants received oral MPH (20 mg) or placebobefore the retrieval of some of the conditioned stimuli (CS) (i.e., reminded CS+ vs.nonreminded CS+) followed by extinction; lab-simulated drug-seeking was measuredthe following day. Lower vmPFC activity following nonreminded CS+ (standardextinction) under placebo replicated the putative impairments in CUD; separately, RE (trend) and MPH conditions recruited the vmPFC, and RE's vmPFC-reliance correlatedwith drug-seeking only under placebo. Crucially, MPH-combined RE normalized cortico-limbicprocessing, bypassing the vmPFC and its amygdala connectivity. Pharmacologically-enhanced drug memory modulation may inform intervention development for addictionrecovery.

Objective: This open-label clinical trial examined the preliminary efficacy of combining a course of 6 ketamine infusions with a brief, evidence-based exposure-based psychotherapy-written exposure therapy (WET)-in patients with chronic posttraumatic stress disorder (PTSD). Methods: The trial was conducted between June 2021 and October 2023. Patients with chronic PTSD and high-moderate to severe symptom levels received 6 intravenous ketamine infusions (0.5 mg/kg), 3 times a week for 2 consecutive weeks, plus 5 WET sessions over 2 weeks, beginning after the first 4 infusions and administered on different days than infusion days. The primary outcome was change in the Clinician Administered PTSD Scale for DSM-5 (CAPS-5) scores from baseline (before the first infusion) to 12 weeks from start of WET ("Week 12"). Results: Fourteen eligible patients began treatment, and 13 completed all infusions and WET. The combined treatment was associated with large-magnitude improvement in PTSD symptom severity from baseline (mean CAPS 5 = 41.6 [SD = 6.2]) to Week 12 (CAPS 5 = 20.8 [14.8], Cohen d [95% CI] = 1.9 [1.0-2.8], P < .001). Nine (69%) patients were treatment responders (≥30% improvement on the CAPS-5). Response was rapid and also durable in 8 (61.5%) patients, assessed up to 6 months from baseline. Conclusions: Preliminary findings from this open-label clinical trial suggest that the combined treatment may yield large magnitude and durable reductions in PTSD symptoms for patients with more severe chronic PTSD. Large-scale randomized controlled trials are needed to determine the efficacy and potential synergistic effect of this promising combined treatment in this patient population. Trial Registration: ClinicalTrials.gov identifier: NCT04889664.

Memory reactivation renders consolidated memory fragile and thereby opens the window for memory updates, such as memory reconsolidation. However, whether memory retrieval facilitates update mechanisms other than memory reconsolidation remains unclear. We tested this hypothesis in three experiments with healthy human participants. First, we demonstrate that memory retrieval-extinction protocol prevents the return of fear expression shortly after extinction training, and this short-term effect is memory reactivation dependent (Study 1, N=57 adults). Furthermore, across different timescales, the memory retrieval-extinction paradigm triggers distinct types of fear amnesia in terms of cue specificity and cognitive control ability dependence, suggesting that the short-term fear amnesia might be caused by different mechanisms from the cue-specific amnesia at a longer and separable timescale (Study 2, N=79 adults). Finally, using continuous theta-burst stimulation (Study 3, N=75 adults), we directly manipulated brain activity in the dorsolateral prefrontal cortex and found that both memory reactivation and intact prefrontal cortex function were necessary for the short-term fear amnesia after the retrieval-extinction protocol. The differences in temporal scale, cue specificity, and cognitive control ability dependence between the short- and long-term amnesia suggest that memory retrieval and extinction training trigger distinct underlying memory update mechanisms. These findings suggest the potential involvement of coordinated memory modulation processes upon memory retrieval and may inform clinical approaches for treating persistent maladaptive memories.

While allowing for rapid recruitment of large samples, online research relies heavily on participants’ self-reports of neuropsychiatric traits, foregoing the clinical characterizations available in laboratory settings. Autism spectrum disorder (ASD) research is one example for which the clinical validity of such an approach remains elusive. Here we compared 56 adults with ASD recruited in person and evaluated by clinicians to matched samples of adults recruited through an online platform (Prolific; 56 with high autistic traits and 56 with low autistic traits) and evaluated via self-reported surveys. Despite having comparable self-reported autistic traits, the online high-trait group reported significantly more social anxiety and avoidant symptoms than in-person ASD participants. Within the in-person sample, there was no relationship between self-rated and clinician-rated autistic traits, suggesting they may capture different aspects of ASD. The groups also differed in their social tendencies during two decision-making tasks; the in-person ASD group was less perceptive of opportunities for social influence and acted less affiliative toward virtual characters. These findings highlight the need for a differentiation between clinically ascertained and trait-defined samples in autism research. Comparing clinically diagnosed adults with autism spectrum disorder (ASD) with online participants with self-reported high autistic traits, this study detected higher rates of social anxiety and avoidance symptoms in online participants, emphasizing the potential distinctions between clinically ascertained autism and self-reported trait-based samples.

BackgroundThe prevalence of depression is elevated in individuals with autism spectrum disorder (ASD) compared to the general population, yet the reasons for this disparity remain unclear. While social deficits central to ASD may contribute to depression, it is uncertain whether social interaction behavior themselves or individuals' introspection about their social behaviors are more impactful. Although the anterior cingulate cortex (ACC) is frequently implicated in ASD, depression, and social functioning, it is unknown if it explains differences between ASD adults with and without co-occurring depression.MethodsThe present study contrasted observed vs. subjective perception of autism symptoms and social interaction assessed with both standardized measures and a lab task, in 65 sex-balanced (52.24% male) autistic young adults. We also quantified ACC and amygdala volume with 7-Tesla structural neuroimaging to examine correlations with self-reported depression and social functioning.ResultsWe found that ASD individuals with self-reported depression exhibited differences in subjective evaluations including heightened self-awareness of ASD symptoms, lower subjective satisfaction with social relations, and less perceived affiliation during the social interaction task, yet no differences in corresponding observed measures, compared to those without depression. Larger ACC volume was related to depression, greater self-awareness of ASD symptoms, and worse subjective satisfaction with social relations. In contrast, amygdala volume, despite its association with clinician-rated ASD symptoms, was not related to depression.LimitationsDue to the cross-sectional nature of our study, we cannot determine the directionality of the observed relationships. Additionally, we included only individuals with an IQ over 60 to ensure participants could complete the social task. We also utilized self-reported depression indices instead of clinically diagnosed depression, which may limit the comprehensiveness of the findings.ConclusionsOur approach highlights the unique role of subjective perception of autism symptoms and social interactions, beyond the observable manifestation of social impairment in ASD, in contributing to self-reported depression, with the ACC playing a crucial role. These findings imply possible heterogeneity of ASD concerning co-occurring depression. Using neuroimaging, we were able to demarcate depressive phenotypes co-occurring alongside autistic phenotypes.