Prior observational studies have reported conflicting results regarding whether antidepressant treatment reduces long-term dementia risk, likely due to confounding by indication and reverse causation. We aimed to investigate the association between baseline antidepressant use and incident dementia, incorporating cognitive and neuroimaging outcomes.

We conducted a prospective cohort study using UK Biobank participants free of dementia at baseline. Antidepressant use was self-reported at baseline (2006–2010). Incident dementia was identified through linked electronic health records until December 19, 2022. Cox proportional hazards models estimated hazard ratios (HRs) for all-cause dementia, Alzheimer’s disease (AD), and vascular dementia (VD), adjusting for sociodemographic, lifestyle, health-related, antidepressant indication factors, and co-medication of other anticholinergics. In subsamples, cognitive performance (n = 57,330) and structural brain imaging (n = 42,276) were examined as intermediate outcomes.

Among 461,464 participants, 33,721 (7.3%) reported baseline antidepressant use. Over a mean follow-up of 13.4 years, 7,922 (1.7%) developed incident dementia. Baseline antidepressant use was associated with higher risks of all-cause dementia (adjusted HR: 1.47, 95% CI 1.36–1.60), AD (1.53, 1.36–1.73), and VD (1.44, 1.23–1.70). Users performed worse on fluid intelligence and prospective memory tasks and showed lower total and gray matter volume, regional reductions in the hippocampal gray matter and basal nucleus, and greater white matter hyperintensity volume.

Baseline antidepressant use was linked to a higher risk of dementia, poorer cognitive performance, and adverse brain structural changes. These findings underscore the importance of judicious prescribing, regular cognitive monitoring, and consideration of non-pharmacological approaches in clinical care.

Youth at familial risk for mood disorders often report disturbed sleep, yet it is unclear which aspects of sleep distinguish them from low-risk peers or when differences emerge developmentally.

A total of 176 youth at familial risk for mood disorders (mean age = 14.6 years) and 119 control youth (mean age = 13.8 years), aged 8 to 25, wore wrist actigraphy continuously for 2 weeks. Total sleep time, sleep midpoint timing, and sleep regularity were derived from actigraphy data. Generalized additive mixed models estimated nonlinear developmental trajectories across age. Confidence bands identified age intervals at which groups diverged.

High- and low-risk youth followed different developmental patterns of total sleep time (F = 6.35, adjusted p < 0.001), sleep midpoint timing (F = 11.3, adjusted p < 0.001), and sleep regularity (F = 3.96, adjusted p = 0.028). High-risk youth had lower total sleep time in middle childhood (ages 8–8.8) and higher total sleep time in young adulthood (ages 18.2–23.9). They had a later sleep midpoint in late childhood (ages 10.2–14.1) and an earlier sleep midpoint from late adolescence to young adulthood (ages 17.3–25.0). High-risk youth slept more regularly from late adolescence to young adulthood (ages 17.8–21.7).

High-risk youth showed poorer sleep than controls in early adolescence but better sleep in late adolescence through young adulthood, possibly reflecting shifting sleep priorities across development, though the mechanisms underlying these differences remain unclear.

Mindfulness-based cognitive therapy (MBCT) was developed for relapse prevention in people with remitted depression but is increasingly used for those with difficult-to-treat depression (DTD). A key question regarding this broader application is whether ongoing depressive symptoms constrain therapeutic responsiveness or disrupt MBCT’s proposed mechanism, decentering. We explored whether baseline depressive severity moderates clinical outcomes, whether changes in decentering mediate treatment effects, and whether this mediation varies by baseline severity.

Secondary moderation, mediation, and moderated mediation analyses were conducted using data from the RESPOND randomized trial (N = 234), comparing MBCT plus treatment as usual (TAU) with TAU alone in adults not remitted after high-intensity psychological therapy. Depressive symptoms (PHQ-9) and decentering (Experiences Questionnaire) were assessed at baseline, post-treatment (10 weeks), and follow-up (34 weeks). Analyses were conducted using structural equation modelling.

Higher baseline severity predicted greater symptom improvement across both groups. Treatment-related increases in decentering partially mediated the effect of MBCT on depressive symptoms at follow-up. Although baseline severity did not moderate the treatment effect, it moderated the indirect effect, with decentering more strongly associated with symptom reduction among those with higher baseline depression. Severity did not moderate the acquisition of decentering skills.

Concerns that more severe depressive symptoms limit the effectiveness of MBCT were not supported. MBCT’s core mechanism remained operative under substantial symptom burden, with clinical impact amplified at higher severity. These findings reduce key uncertainties regarding the application of MBCT in DTD and support its use across a broad range of symptom severity.

Alcohol, nicotine, and cannabis use problems have a poorer prognosis when they co-occur with internalizing distress (INT). One hypothesis to account for the association between INT and alcohol and substance use problems (SUP) is that people engage in heavy substance use to reduce aversive affective states characteristic of INT (e.g. affective reinforcement). Whether INT operates as a cause or a shared liability with SUP (e.g. shared genetic and environmental risk) over and above an alternative cause (e.g. externalizing problems, EXT) remains unclear.

We tested INT as a developmental pathway for SUP by applying a prospective, co-twin control approach to the Minnesota Twin and Family Study sample (n = 1881 twin pairs). A multilevel modeling framework was used to estimate the between-twin pair effect and the within-twin pair effect of INT over and above EXT on separate SUP outcomes (i.e. alcohol, nicotine, and cannabis use problems) from adolescence (age 14) to young adulthood (age 29).

We failed to detect any non-familial causal effects of INT on SUPs. Instead, their associations were due to common genetic and shared environmental influences. In contrast, EXT had consistent non-familial causal influences on alcohol and cannabis use in adolescence and young adulthood.

The link between INT and SUPs is primarily due to shared, familial influences while EXT likely serves as a unique determinant, with elevated EXT increasing selection into situations that pose SUP risk (e.g. peer environments that normalize and sustain heavy use) over and above common genetic influences.

Post-stroke depression (PSD) is the most common neuropsychological complication in stroke survivors. While psychological interventions are promising non-pharmacological treatments, the most effective intervention type, as well as their mid- to long-term efficacy and effectiveness across PSD severity levels, remains unclear. This study aims to compare the efficacy of psychological interventions for PSD, focusing on overall effectiveness, mid- to long-term efficacy, and effectiveness across PSD severity levels.

We systematically searched seven databases for randomized controlled trials (RCTs) of psychological interventions targeting depression in stroke survivors. The primary outcome was depression symptom reduction. Post-intervention, mid- to long-term data were analyzed. A frequentist network meta-analysis (NMA) with random effects was conducted in Stata 17.0. Subgroup analyses by PSD severity were performed.

Thirty-eight RCTs involving 3106 participants were included. Post-intervention, third-wave therapies (Hedges’ g = −1.08, 95% CI: −1.62 to −0.55) significantly improved PSD compared to treatment as usual. At mid- to long-term follow-up, third-wave therapies (Hedges’ g = −0.82, 95% CI: −1.25 to −0.39) maintained significant effects. Relaxation therapy significantly ameliorated mild PSD (Hedges’ g = −0.77, 95% CI: −1.27 to −0.27); for moderate-to-severe PSD, no significant effects were observed.

Third-wave therapies are potentially effective for PSD, showing sustained mid- to long-term benefits. Relaxation therapy seems to be most effective for mild PSD, while no significant effects were observed for moderate-to-severe PSD. However, the NMA showed high heterogeneity and generally low to very low evidence certainty for most comparisons, the findings should be interpreted with caution.

Accurate psychiatric assessment requires understanding a person’s unique experience within their psychosocial context. Clinical interviews have been the gold standard for assessment as the only methods capable of this complex task, but they are time and resource-intensive. Consequently, psychiatric assessment typically relies on patient report surveys that are decontextualized and narrow in scope. This comprehensiveness-scalability tradeoff is a major bottleneck in studying and treating psychopathology. We propose using large language models (LLMs) to score psychopathology from brief personal narratives as a low-burden, context-sensitive solution.

Participants (N = 108) completed brief (~1 minute), freeform audio diaries daily for 2 weeks. We used six LLMs to score wide-ranging psychopathology (Internalizing, Detachment, Disinhibition, Antagonism, Anankastia) from the diary transcripts. Leveraging an array of self-report and clinical interview measures, we tested the convergent, discriminant, concurrent, and clinical validity of LLM ratings for between-person differences and within-person fluctuations in psychopathology.

Supporting convergent and discriminant validity, LLM ratings correlated most strongly with corresponding self-report domains at the between (average convergent r = .42) and within-person (r = .28) levels. LLM and self-report ratings had similar patterns of associations with external variables, except for Anankastia and Antagonism. Further, every LLM-rated domain related to psychopathology ascertained by clinical interview.

Across multiple forms of validity, we showed that LLMs can assess most major forms of psychopathology from mere minutes of audio. These results support scoring open-ended narratives with LLMs as a scalable, portable method to translate idiographic diagnostic data into standardized psychiatric assessments.

Among military veterans, loneliness is common and associated with poorer mental and physical health and functioning. Epidemiological studies have revealed bidirectional, longitudinal associations between loneliness and posttraumatic stress disorder (PTSD). Little is known about loneliness among treatment-seeking veterans, how loneliness impacts PTSD symptom change during treatment, and whether loneliness has differential associations with specific PTSD symptom clusters.

Data were derived from the Veterans Outcome Assessment, an annual, repeated-measures (baseline and 3-month follow-up) telephone survey of patients beginning new episodes of mental health care in the Department of Veterans Affairs. Participants were 1,155 veterans (23% women) entering PTSD specialty care. We tested whether baseline loneliness predicted change in total PTSD symptom severity and according to the five-factor Dysphoric Arousal model of PTSD using hierarchical regressions in which we statistically adjusted for demographics and mental health encounters.

Loneliness decreased between baseline and 3-month follow-up (d = 0.39). Baseline loneliness predicted change in PTSD for the dysphoric arousal and negative alterations in cognitions and mood symptom clusters, with lonelier veterans demonstrating more PTSD change between baseline and follow-up. However, lonelier veterans also had more severe PTSD symptoms at both timepoints. This pattern of results held even after accounting for different types of treatment encounters between timepoints.

While lonelier veterans have more severe PTSD symptoms, they also make greater improvements in symptoms salient to interpersonal functioning over time. Assessment of loneliness should be incorporated into PTSD treatment planning and considered as a treatment target that could be addressed to promote recovery.

Attention-deficit hyperactivity disorder (ADHD) is associated with impairments in real-life decisions, many involving balancing costs and benefits. Increasing dopamine and norepinephrine with stimulant medication improves decision-making in ADHD. However, it is unclear whether stimulant medication affects sensitivity to costs and/or benefits or alters the speed-accuracy tradeoff. Here, we applied the drift diffusion model (DDM) to assess how the stimulant medication methylphenidate (MPH) affects subprocesses of cost-benefit decision-making in ADHD.

Sixty-one adult participants with ADHD and 63 healthy controls completed a cost-benefit decision-making task in a randomized double-blind placebo-controlled design. ADHD participants were tested once on and once off MPH. Control participants performed the task in two sessions without medication. The task entailed deciding whether to accept or reject a stimulus based on learned cost and benefit value associations. The effect of ADHD and medication on subprocesses of decision-making was disentangled by estimating task performance with DDM parameters drift rate, starting point bias, and decision threshold, representing evidence accumulation, a priori accept/reject bias, and the speed-accuracy tradeoff, respectively. Primary analyses used a modified intention-to-treat sample, including participants with main-phase data from both sessions.

The DDM analysis identified impaired evidence accumulation combined with premature responding in participants with ADHD, indicated by reduced sensitivity to costs and benefits and narrower decision thresholds. MPH altered the cost-benefit balance by improving sensitivity to costs. Lastly, lower cost sensitivity was associated with greater symptom severity and higher prescribed dosage in ADHD.

Methylphenidate alters cost-benefit decision-making in ADHD by increasing sensitivity to costs.

Observational studies have suggested that brain imaging-derived phenotypes (IDPs) may serve as specific markers of pain-related phenotypes and severity. However, the shared genetic architecture between pain and brain IDPs and their potential causal relationships remains unclear.

We applied linkage disequilibrium score regression and Mendelian randomization (MR) analyses to uncover genetic correlations and potential causal links of brain structural (33,224 UK Biobank participants) and functional (47,276 UK Biobank participants) changes with site-specific pain phenotypes (approximately 500,000 Finngen participants). The scoping literature review was conducted to compare current findings with previous observational studies.

In this study, we identified 559 significant genetic correlations between 587 structural IDPs and 13 pain-related phenotypes. Using MR analyses, we found that genetic liability to headache, migraine, joint pain, and sciatica was causally associated with alterations in 15 structural IDPs. Additionally, changes in the surface area of three brain regions were linked to a lower risk of sciatica, low back pain, and overall pain. Among the six pain-related phenotypes associated with structural IDPs, further analyses demonstrated putative causal relationships between functional IDPs and these conditions. Notably, headache exhibited both significant structural and functional changes across three key brain regions: the superior frontal gyrus, lingual gyrus, and paracentral lobule.

These findings provide novel insights into the genetic correlations and genetically inferred associations between pain and neurobiological abnormalities from neuroimaging perspectives, with structural alterations as the primary findings and functional changes as complementary exploratory evidence, advancing the understanding of pain-related mechanisms.

Sleep disorder, cognitive impairment, and psychological distress usually co-occur in older adults, yet their dynamic relationships over time remain poorly understood, particularly in rural aging populations.

This study employed cross-lagged panel network (CLPN) analysis to examine temporal relationships among sleep quality, cognitive function, and psychological distress in 2,196 rural Chinese older adults across 4 years. We analyzed autoregressive and cross-lagged effects, centrality measures, and sex differences in network structures over two time intervals.

CLPN analysis revealed temporal associations among sleep quality, psychological distress, and cognitive function domains. Subjective sleep quality emerged as the most influential predictor node across both time periods. Cognitive components – particularly orientation, attention and calculation, and language – were identified as the most vulnerable target nodes. Sex-stratified analyses revealed moderate structural similarity between male and female networks across time periods.

This study provides new evidence for understanding the temporal relationships among sleep quality, psychological distress, and cognitive function in rural older adults. These findings may suggest that subjective sleep quality represents a central and modifiable target within the aging-related symptom network, with potential implications for early intervention across cognitive and psychological domains.

Post-traumatic sequelae are theorized to contribute to and maintain psychotic symptoms. The relative contribution of core posttraumatic stress disorder (PTSD) symptoms and those of complex PTSD are understudied. Network analysis was used to examine interactions between psychosis, core PTSD, and complex post-traumatic symptoms. Based on previous evidence, it was predicted that trauma-related beliefs would have the greatest predictability. It was also hypothesized that the association between voices and trauma-related beliefs would be greater than the association between voices and re-experiencing.

Baseline data from the Study of Trauma and Recovery (STAR), a randomized controlled therapy trial of people with schizophrenia-spectrum disorder who experience distressing delusions and/or hallucinations and comorbid PTSD (N = 305). A network was estimated using a Bayesian Gaussian graphical model, comprised of psychosis symptoms (delusions and voices), posttraumatic stress symptoms (re-experiencing, avoidance, and sense of threat), complex posttraumatic stress symptoms (affective dysregulation, negative self-concept, disturbed relationships), and trauma-related beliefs (negative cognitions about the self and world, self-blame). A direct acyclic graph was plotted to estimate the potential directionality of associations between symptoms.

The hypotheses were not supported. All symptoms had similar predictability scores. There was no evidence for conditionally dependent relations between re-experiencing and voices nor between trauma-related beliefs and voices. However, self-blame was associated with delusions, and affective dysregulation was found to predict voices.

Affective dysregulation and self-blame may maintain voices and delusions, respectively. Heterogeneity in trauma-related psychosis could explain inconsistent findings; future research should identify symptom presentation subgroups to improve network reliability.

Theory-driven computational psychiatry attempts to use cognitive models of computation to understand how mental health problems might relate to cognitive processes such as learning and decision-making. However, the potential applications and relevance of this approach are contingent on several (often implicit) assumptions, including that computational parameters 1) are recoverable, 2) are reliable over time, 3) reflect conceptually similar processes across different tasks, and 4) relate to symptoms.

To illustrate and test these assumptions for a selection of commonly used cognitive tasks, we recruited a large online sample of participants (n = 548), who completed seven mental health questionnaires and five tasks. A subset of n = 115 re-completed the tasks 14 days later. For each task, five models were fit, and the winning model was selected through Bayesian model comparison.

The parameters of the winning models showed excellent recovery and good-to-excellent test-retest reliability. Parameters that were theoretically similar to each other were, however, generally not related, with relationships only found within the class of ‘inverse temperature’-like parameters. Finally, only parameters from two of the five tasks (four-armed bandit and cognitive effort) related to symptoms, and these relationships were weak.

It seems that at least some of the implicit assumptions made when advocating for the use of computational models in psychiatry are not met for these popular tasks and computational models. Computational psychiatry researchers should carefully assess the assumptions and psychometric properties of their tasks and models as part of their early-phase development to ensure robustness of the field going forward.

Late-life depression often extends to the marital partner, yet mechanisms of dyadic interdependence and their cross-cultural generalizability remain unclear.

We analyzed harmonized longitudinal data from couples in the China Health and Retirement Longitudinal Study (CHARLS; n = 3,532; 5 waves, 2011–2020), the Health and Retirement Study (HRS; n = 2,332; 8 waves, 2006–2020), and the English Longitudinal Study of Ageing (ELSA; n = 1,895; 9 waves, 2002–2018). Depressive symptoms (CES-D) and activities of daily living (ADL)/instrumental ADL (IADL) limitations were measured at every wave. Complementary actor–partner interdependence models (APIM) and random-intercept cross-lagged panel models (RI-CLPM) were applied, with multiplicity adjustment, full-information maximum likelihood (FIML), and 20-imputation triangulation.

Within-couple correlations of depressive symptoms were positive in every cohort (r = 0.34–0.43 in CHARLS, 0.14–0.20 in HRS, 0.23–0.29 in ELSA). APIM partner effects were robust in CHARLS but sparse elsewhere. After partialling out stable between-couple similarity, RI-CLPM within-person cross-spouse associations were attenuated, reaching consistent significance under FIML and multiple imputation only in ELSA. Four-variable RI-CLPMs showed bidirectional depression–disability coupling, strongest in HRS and ELSA, with cross-spouse function-to-depression effects in Western cohorts but not in CHARLS. Sex distinguishability was not supported, and half-longitudinal mediation through ADL/IADL did not survive multiplicity correction.

Concordance in late-life depressive symptoms replicates across three major aging cohorts; in CHARLS, it primarily reflects stable between-couple similarity, whereas modest within-person prospective coupling persists in ELSA and in the multiple-imputation HRS estimates. Bidirectional depression–disability coupling in Western cohorts is the most replicable within-person signal, supporting combined depression management and functional rehabilitation for older couples.

Since the 1940s, the concept of autism has undergone marked changes, and autism spectrum disorder (ASD) is a wildly heterogenous condition. At the same time, the prevalence of ASD has increased massively.

This essay considers possible reasons for the changes and their consequences.

Among drivers for the increase in prevalence are cultural changes, such as the desire for inclusion, harm avoidance, reliance on self-report, and the acceptance of masking. These have all led to a lowering of the diagnostic threshold. Furthermore, the popularity of the concept of autism boosted numbers via self-diagnosis and a search for identity.

Marked differences now exist between those first diagnosed in childhood and others first diagnosed in adolescence or adulthood. The time has come to examine reasons for these conceptual changes and the consequences for clinical practice and research.

Posttraumatic stress disorder (PTSD) symptoms are among the most prevalent mental health sequelae in children and adolescents following earthquakes. A recently introduced metric under network theories of psychopathology, network temperature, quantifies the global stability of a symptom network.

To examine the developmental trajectory of network temperature in children and adolescents after earthquakes, we established eight prospective cohorts (2–4 waves per cohort; total N = 5,794) after the Wenchuan, Yaan, and Jiuzhaigou earthquakes. We assessed PTSD symptoms across three postdisaster periods: early (≤2.5 years), middle (3.5–5.5 years), and late (8.5–10 years).

PTSD symptom network temperatures decreased across all eight cohorts. In the early post-Wenchuan period, adolescents declined more slowly than children. Across earthquakes, the early-period decline in children was faster after the Wenchuan earthquake than after the Yaan earthquake; for adolescents, the early-period decline rate followed the order Wenchuan < Jiuzhaigou < Yaan. Across post-earthquake periods, Wenchuan adolescents declined more slowly in the early period than in the middle/late period, whereas Yaan adolescents declined faster in the early period than in the middle period.

Across all earthquake types and post-earthquake periods, the PTSD symptom network system became increasingly stable and resistant to change. These findings underscore the necessity of early intervention alongside efforts to prevent deterioration. However, the rate of this decline varied by sample (children versus adolescents), earthquake, and post-earthquake period. Clinically, this calls for intervention strategies that are adjusted according to these factors.

Adults with mental disorders more frequently report getting into serious trouble with the police and the law than do adults with no mental health diagnosis. The mechanisms underlying this statistical association remain unclear.

The study used data from a US population survey to examine whether the previously recognized association between psychiatric diagnosis and getting into serious police or legal trouble is mediated by impaired psychosocial function or psychiatric symptoms. The predictor variable was psychiatric diagnosis. The mediator variables were two forms of psychosocial function, social and role-emotional, and the presence of recent symptoms. The outcome variable was police or legal trouble.

When known risk factors are included as covariates, the association between mental disorder and serious police or legal trouble over the past year is maintained and is significantly mediated (p < 0.05) both by social and role-emotional function (social, excess relative risk [ERR] 1.42 [95% CI: 0.96, 1.88]; role-emotional, ERR 1.45 [95% CI: 0.98, 1.91]) and by recent symptoms(ERR 1.63 [95% CI: 1.13, 2.13]). The contribution of mediation varies by diagnosis.

Previous research suggested that it is unusual for psychiatric symptoms to lead directly to violence. These results suggest that symptoms and signs of mental disorder nevertheless contribute indirectly to violence and other adverse outcomes. The identification of these mediators may, in turn, suggest new approaches to reducing the risk of those outcomes. Future research should further examine and describe the mechanisms that underlie these and other forms of mediation.

Psychological adaptability hinges on the dynamic balance between cognitive regulation (self-control) and emotional reactivity (impulsivity, reward/punishment sensitivity). However, traditional variable-centered approaches often fail to capture how these traits holistically co-occur, and the neural architectures predicting their longitudinal transitions in early adulthood remain under-explored.

This longitudinal study (N = 1,229 baseline; N = 432 2-year follow-up) integrated a person-centered behavioral approach with resting-state fMRI. We employed Latent Profile Analysis (LPA) to identify trait configurations based on self-control, impulsivity, and sensitivity to reward/punishment. Network-Based Statistic (NBS) analysis with family-wise error (FWE) correction was utilized to evaluate functional connectivity differences among groups. Finally, hierarchical regression and formal interaction models were conducted to test the prospective predictive validity of the identified neural circuits.

LPA delineated three distinct baseline profiles: Adaptive, Moderate, and Maladaptive. NBS analysis revealed that the ‘Adaptive’ profile is underpinned by robust cortico-striatal functional connectivity, integrating the medial prefrontal cortex and striatum. Longitudinally, initial regression models demonstrated that the baseline integrity of this circuit prospectively predicted behavioral adaptation at the 2-year follow-up. Furthermore, hierarchical regression and formal interaction analyses confirmed that this cortico-striatal circuit provided significant incremental predictive validity – above and beyond massive baseline behavioral stability – specifically in males.

These findings highlight cortico-striatal integration as a significant, gender-specific prospective predictor of longitudinal adaptation. By bridging person-centered profiling with network neuroscience, this study elucidates the neural correlates supporting future adaptive functioning and psychological resilience during the early adult transition.