Nature Mental Health

Substance use disorders (SUD) are chronic conditions with devastating effects on brain health, functioning, and survival. In this study, we compared brain morphometry of 2,782 individuals with SUD to 1,951 controls and assessed the topographic overlap of these differences with brain connectivity and receptor architecture. Across SUD, we identified a morphometric signature involving frontal, parietal, temporal and limbic systems that overlapped with cortical hub regions and harbored cortical and subcortical disease epicenters. Findings were highly consistent across six substances and numerous robustness and generalizability analyses. Transdiagnostic comparisons showed high spatial overlap of SUD epicenters with those of schizophrenia and bipolar disorder, suggesting shared network-constrained cortical differences. Finally, multivariate mapping revealed that SUD brain differences aligned with two neurotransmitter axes contrasting cannabinoid-opioid and dopaminergic systems. These findings indicate that addiction-related brain differences are shaped by connectome and neurotransmitter architecture, positioning brain network and neurochemical organization as key principles of SUD-related brain alterations.

Mechanistic understanding and biomarkers of gonadotropin-releasing hormone antagonist treatment effect in paedophilic disorder are absent but may enhance outcomes and reduce sexual-offending risk. 52 help-seeking self-referred Swedish men with paedophilic disorder enrolled in a double-blinded, placebo-controlled, randomized clinical trial. Participants underwent task-based fMRI before, and two weeks after, subcutaneous injection of 120mg of degarelix or equal volume of placebo. fMRI blood-oxygen-level-dependent activation was compared between child and adult (child>adult) stimuli in task-derived regions of interest. Primary outcome was within region-of-interest child>adult activation change, whereas secondary outcomes correlated region-of-interest child>adult activation change to change in clinical measurements of risk, paedophilic interest, sexual preoccupation, hyper- and hyposexuality. 19 degarelix and 22 placebo participants had sufficient fMRI data quality. Reductions in paedophilic interest were strongly correlated with increased child>adult cerebellar (vermis) region-of-interest activation following degarelix (r=-0.740, p<0.001) but not placebo (r=0.183, p=0.41; between-group correlation coefficient z=3.347, p<0.001). Treatment did not significantly change child>adult region-of-interest activity. Post hoc analysis indicated that baseline autism symptoms correlated with degarelix-induced changes in paedophilic interest (r=0.717, p<0.001; between-group correlation coefficient z=2.958, p=0.003) and cerebellar activation (r=-0.581, p=0.01; between-group correlation coefficient z=-1.930, p=0.05). Increased child>adult cerebellar activation was associated with degarelix-induced reductions of paedophilic interest, suggesting cerebellar activity as mechanistically important to, and a prospective biomarker of, degarelix treatment effect. Additionally, autism symptoms may inform treatment prediction. Together, these findings have mechanistic and clinical implications for degarelix treatment of paedophilic disorder. EU clinical trials register identifier: 2014-000647-32 https://www.clinicaltrialsregister.eu/ctr-search/trial/2014-000647-32/SE, registered on 05/06/2014.

Arousal and valence are fundamental dimensions of affective experience signifying levels of activation and pleasantness, respectively. These dimensions play a crucial role in shaping emotional responses and behaviors, with significant implications for psychopathology. Previous machine learning studies had some success decoding these states from brain activation patterns observed during task-based functional magnetic resonance imaging (fMRI), but the results have varied across studies. Moreover, prior studies have often been limited by small sample sizes, weak decoding performance, and non-whole-brain analyses, leaving the neural representations of arousal and valence largely unresolved. Here we successfully decoded arousal and valence from whole-brain task-fMRI data collected from 132 participants during exposure to 300 unique emotional stimuli, including 150 movie clips and 150 text scenarios that reliably induced a wide range of arousal and valence states. Mass univariate general linear models identified block-level activation (emotion stimuli > washout) from all gray matter voxels. Multivariate regression analysis predicted arousal and valence ratings based on these gray matter activations. Patterns in the fMRI data underlying arousal and valence were robust, as they were successfully decoded across both induction modalities using five different linear multivariate regression models. Although significant, decoding from scenarios was less successful than from movies, likely due to their more imaginative nature. In particular, decoding arousal from scenarios only showed low predictive utility. Representations of arousal and valence were widespread throughout the brain, and we reveal cerebellar and brainstem contributions that have largely been absent in past fMRI decoding studies. These findings clarify the distributed neural basis of arousal and valence and provide a foundation for future clinical research on the role of these constructs in affective dysregulation.

Psychosis is increasingly understood as a disorder of disrupted cortical excitation-inhibition balance, yet robust non-invasive translational biomarkers remain lacking. The resting-state fMRI Hurst exponent (HE) and EEG aperiodic spectral exponent are promising complementary biomarkers, with lower values in each proposed to reflect a shift towards cortical hyperexcitability, but they have not been jointly examined in psychosis, and the spatial and molecular architecture of HE alterations remains poorly defined. We therefore tested for convergent systems-level signatures across independent cohorts and modalities, using resting-state fMRI (107 patients, 53 controls) and EEG (547 patients, 363 controls). Whole-brain and regional HE were estimated using wavelet methods, and EEG aperiodic exponents were quantified using spectral parameterisation. Compared with healthy controls, individuals with psychosis showed reduced whole-brain HE and widespread regional reductions. Regional HE case-control differences were associated with cortical gene-expression patterns, with enrichment for potassium channel and GABA receptor pathways, and correlated with noradrenergic, muscarinic, serotonergic, glutamatergic and dopaminergic receptor density maps, but not with cortical thickness or symptom or cognitive measures. In the independent EEG cohort, psychosis was similarly associated with a reduced aperiodic spectral exponent. Together, these findings provide cross-modal evidence for altered cortical resting-state dynamics in psychosis, consistent with a shift towards cortical hyperexcitability. Integration with receptor-density and transcriptomic maps implicates biologically plausible molecular pathways and supports HE and EEG aperiodic activity as scalable translational biomarkers in psychosis.

Internet gaming disorder (IGD) presents a significant public health challenge, yet its complex biopsychosocial mechanisms and dynamic risk trajectories remain poorly understood due to a scarcity of comprehensive longitudinal and multimodal cohorts. To address this critical gap, we established the Chinese College Student Gamers Cohort (CCSGC), a prospective, multimodal longitudinal study of 793 first-year undergraduates primarily playing Honor of Kings from 2022 Sept. The CCSGC integrates semi-annual psychosocial questionnaires, annual neuroimaging (EEG/fMRI), and biospecimen collection over multiple years. Baseline data revealed individuals with IGD (n=211) exhibited significantly higher gaming craving, psychological distress (depression, anxiety), impulsivity, and maladaptive motivational features compared to non-IGD gamers (regular players (RP) n=400; casual players (CP) n=182). Longitudinal analyses across four waves indicated bidirectional temporal associations between IGD severity and mental symptoms, and a stabilization of IGD incidence after an initial decrease. Furthermore, specific neurophysiological (e.g., N400 amplitude to game cues) and neuroimaging (e.g., superior parietal activation) markers were identified that correlated with IGD severity and predicted one-year outcomes in gaming disorder or social functioning. The CCSGC provides an invaluable resource for dissecting the heterogeneity, comorbidity, and intricate biopsychosocial mechanisms of IGD, holding significant potential to advance risk prediction, early identification, and targeted intervention strategies.

Early-childhood temperament is associated with mental health outcomes decades later. Temperament reflects early-emerging individual differences in emotional and behavioral tendencies. These differences are relatively stable across development and shaped by both genetic and environmental influences. However, the consequences of departures from expected developmental trajectories remain largely unexplored. Using data from more than 50,000 children in the Norwegian Mother, Father and Child Cohort Study, we modeled longitudinal temperament trajectories at 1.5, 3, and 5 years of age and quantified deviations from expected development. Multivariate pattern analysis revealed latent dimensions linking these deviations to clinical diagnoses, with ADHD as the most prominent outcome. Time-to-event analysis showed that these dimensions were associated with a higher hazard of ADHD diagnosis across childhood and adolescence. Finally, genetic analyses identified loci jointly associated with temperament trajectories and ADHD, revealing age-dependent genetic effects. Together, these findings show that deviations from temperament trajectories in early childhood capture transdiagnostic vulnerability across development. Early temperament monitoring may thus serve as an indicator of later mental health risk.

Mind wandering has been linked to a wide range of psychiatric conditions, yet most studies have examined these associations in isolation. Given the substantial comorbidity across the psychopathological spectrum, it remains unclear whether elevated mind wandering reflects a general marker of psychopathology or a more specific attentional-control deficit shared across symptom dimensions. To address this, we adopted a dimensional, transdiagnostic approach in a non-clinical sample (N = 376), simultaneously modeling seven symptom dimensions: ADHD, depression, obsessive-compulsive tendencies, schizotypy, autistic traits, hypomania, and eating disorder symptoms. At the bivariate level, mind wandering correlated positively with all symptom dimensions. However, when the substantial shared variance across dimensions was accounted for in both frequentist and Bayesian multivariate regression models, only ADHD symptoms emerged as a unique predictor ({beta} = 0.53; BF{square}{square} > 1000), with all remaining predictors yielding negligible unique contributions and Bayes factors supporting the null hypothesis. These findings suggest that previously reported associations between mind wandering and diverse psychopathological symptom dimensions largely reflect a shared liability with ADHD-related attentional dysregulation, rather than disorder-specific mechanisms. This positions mind wandering as a marker of attentional dysregulation more closely tied to ADHD symptomatology than to general psychopathological burden.

BackgroundLarge-scale T1-weighted MRI studies have established grey-matter abnormalities in bipolar disorder (BD), with our group contributing to consensus findings. However, structural connectivity, particularly within emotion- and reward-related circuits, remains poorly understood. Diffusion-weighted MRI (dMRI) enables investigation of white-matter pathways, yet prior work is constrained by small samples, methodological heterogeneity, and unclear medication effects. We conducted the largest dMRI network analysis in BD, relating symptom burden and polypharmacy to tractography-derived connectivity and graph-theoretic metrics. MethodsCross-sectional structural and diffusion MRI scans from 449 individuals with BD (35.7{+/-}12.6 years) and 510 controls (33.3{+/-}12.6 years), aged 18-65, were analyzed across 16 ENIGMA-BD sites. Standardized segmentation/parcellation and constrained spherical deconvolution tractography generated individual structural connectivity matrices. Graph-theoretic metrics of global and subnetwork organization were related to symptom severity and medications. ResultsBD showed widespread network alterations (lower density and efficiency, longer path length, and higher betweenness centrality), altered microstructural organization in a limbic-basal ganglia circuit, and abnormal streamline counts in a default-mode/salience/fronto-limbic-basal ganglia network. Longer illness duration, later onset, and psychosis history were associated with greater abnormalities in network architecture, whereas more manic episodes were associated with greater fronto-limbic connectivity. Antidepressant (particularly SSRI), anticonvulsant, and antipsychotic use related to poorer global and fronto-limbic connectivity; no clear lithium effects emerged. ConclusionsAs the largest structural connectivity study in BD, we reveal widespread disruption in reward and emotion-regulation networks influenced by illness severity and medication use. Results show that multisite harmonization is feasible and highlight ENIGMA-BD as a scalable framework for identifying reproducible neurobiological markers.

Diffusion brain MRI (dMRI) studies of substance use disorders have reported widespread but modest white matter (WM) microstructural alterations with limited anatomical specificity. Here, we applied segment-wise along-tract 3D tractometry to brain dMRI scans to localize fine-scale WM alterations associated with stimulant misuse using two complementary analytical frameworks: Bundle Analytics (BUAN) and Medial Tractography Analysis (MeTA). We analyzed 3D profiles of widely-used diffusion metrics across 33 major WM bundles in independent cohorts of cocaine (74 cases;58 controls) and amphetamine (22 cases;18 controls) users, testing the statistical associations with brain microstructure of pooled stimulant effects, substance-specific effects, and direct comparisons between stimulant classes. Segment-wise analyses revealed focal differences localized to specific tract segments rather than uniform differences along entire bundles. In pooled stimulant misuse, convergent findings across analysis pipelines were localized to hippocampal pathways and were consistent with altered microstructural organization. Amphetamines misuse showed a broader pattern of segment-wise differences across commissural, projection, and association pathways, involving altered axonal organization. No robust segment-wise differences were detected for cocaine misuse or between stimulant classes. These results show that WM alterations are spatially localized and reproducible across tractometry frameworks, highlighting the value of along-tract 3D mapping for improving anatomical specificity in addiction neuroimaging.

Craving is a hallmark feature of substance use disorders (SUDs) and a major risk factor for relapse, yet reliable biomarkers that enable individual-level prediction remain scarce. Here, we applied connectome-based predictive modeling (CPM) to resting-state functional magnetic resonance imaging (fMRI) data in a transdiagnostic sample of individuals with cannabis, opioid, or tobacco use disorder (n = 78). Using CPM, we identified a distributed functional brain network that reliably predicted self-reported craving. Computational lesion analyses revealed key contributions from the right medial orbitofrontal cortex, right dorsal posterior cingulate cortex, and left lateral medial frontal gyrus. Importantly, the craving network generalized across two independent datasets. In alcohol-dependent patients (n = 41), the identified craving network, along with its positive and negative subnetworks, predicted distinct cognitive and motivational components of craving. In a second external dataset of smokers (n = 28), the craving network predicted both nicotine craving after abstinence as well as intra-individual changes in craving between sated and craving states. Together, these findings provide evidence for a robust, transdiagnostic craving signature in SUDs. Future work should assess the networks predictive utility for longitudinal outcomes such as relapse risk and treatment response.

Schizophrenia spectrum disorders (SSDs) are clinically and neurobiologically heterogeneous. Normative modeling addresses heterogeneity of structural brain alterations by focusing on individual-level deviations, but their clinical relevance in SSDs remains controversial. We mapped the relationship between individual gray matter volume (GMV) deviations and schizophrenia diagnosis and symptoms. Normative models of GMV were established using cross-sectional, T1-weighted magnetic resonance imaging data from a large, multi-site, healthy reference cohort (N = 7957). Deviations were derived for SSD patients (n = 379) and healthy controls (n =149). Patients showed a significantly more negative average deviation compared to controls and regional deviations predicted diagnostic status with adequate performance (AUC = 0.79). A more negative deviation was associated with higher symptom severity and lower cognitive functioning in SSD. Negative deviations were scattered across the brain, with the largest alterations in the salience network. Our findings strengthen the potential of normative modeling to disentangle the heterogeneous underpinnings of SSD and provide further evidence for individualized structural deviations, particularly in the salience network, as promising markers of illness severity in SSDs.

Background. Major depressive disorder (MDD) is characterized by altered frontal alpha oscillations. Transcranial alternating current stimulation (tACS) can normalize aberrant oscillations in MDD, yet the daily dynamics of tACS target engagement of alpha oscillations in depression remain unclear. Methods. In a double-blind randomized controlled trial (NCT03994081), 20 participants with MDD received verum or sham 10 Hz tACS (40 min/day, 5 days) targeted to left and right dorsolateral prefrontal cortex (F3/F4). High-density EEG was collected pre/post-stimulation each day to quantify within-session and cumulative changes in alpha power and functional connectivity (wPLI). Results. Verum stimulation produced late-emerging, session-specific alpha power decreases compared to sham, with robust day (D)4 post-pre reductions at both IAF and 10 Hz across frontal and parietal regions (t=-2.42 to -3.82, p<0.05; parietal t=-3.82, pFDR<0.05). Whole-brain topographical analysis confirmed a distinct condition x D4 effect at left prefrontal cortex (t=2.9, pFWE<0.05, cluster permutation). Connectivity changes emerged earlier and more transiently, with D2 bilateral frontal wPLI reductions (t=-2.53, p<0.05). Cumulative analyses (change from D1) showed significant wPLI decreases on D2 and D3 (t=-2.65 and t=-2.46; p<0.05). Exploratory clinical correlations showed that the D4 IAF power decrease was associated with increased reward sensitivity (spearman rho= -0.6, p<0.05, cluster-corrected). Conclusions. Alpha-tACS produced a temporally distinct neural response: an early, transient decrease in functional connectivity on D2, which may have driven a later suppression of left prefrontal alpha power on D4, correlated with clinical and behavioral improvements. These results delineate target engagement and validation mechanisms in a multi-day tACS trial, supporting optimized dosing in future tACS interventions.

Working memory (WM) supports the temporary maintenance of goal-relevant information and is disrupted across many neuropsychiatric disorders. We examined whether scalp electroencephalography (EEG) data features beyond spectral power, including waveform shape, broadband spectral structure, and signal complexity, provide complementary information for predicting cognitive and clinical outcomes. EEG was recorded from 200 adults spanning a broad range of neuropsychiatric symptom severity while they completed three WM task paradigms: Sternberg spatial WM (SWM), delayed face recognition (DFR), and dot pattern expectancy (DPX). Separate machine learning models were trained on EEG features from the encoding, delay, and probe phase of each task to predict participants task accuracy, reaction time (RT) variability, WM capacity, and psychopathology scores (Brief Psychiatric Rating Scale). A split-half analytic framework was used, with cross-validated model development in an exploratory dataset (N=100) and evaluation of statistically significant models in a held-out validation dataset (N=100). In the exploratory dataset, SWM task data best predicted WM capacity, DPX task data predicted RT variability, and DFR task data predicted psychopathology, suggesting that these three WM paradigms engage distinct neural processes relevant to different outcomes. No models reliably predicted task accuracy. Models incorporating features beyond spectral power generally outperformed power-only models, and task-derived features outperformed resting-state-derived features. However, only those models predicting WM capacity and RT variability generalized to the validation dataset; models predicting psychopathology did not. These findings demonstrate functional heterogeneity across WM paradigms, show that complementary EEG features enhance predictive modeling, and highlight the importance of rigorous validation for identifying robust brain-behavior relationships.

Sustained attention is an important neurobiological process. Difficulties with attention play a key role in neurodevelopmental disorders, such as attention-deficit/hyperactivity disorder (ADHD) and autism. Here, we identified functional connections consistently associated with sustained attention across datasets, participant populations, and fMRI scan types. We interrogated five transdiagnostic, previously published connectome-based models predicting attention and autistic phenotypes. All models were related to sustained attention, including in samples comprising participants with autism. We found that model similarity was associated with participant characteristics, including age and clinical diagnosis, and predicted behavioral measure. As expected, models predicting attention phenotypes shared more similar features with each other than models predicting autism symptoms. Furthermore, predictive features overlapped more between datasets that included participants of similar ages (i.e., youth vs. adult) and diagnostic status (autism diagnosis vs. no diagnosis). This suggests that functional connectivity patterns predicting individual differences in behavior are phenotype-specific and may vary as a function of age and clinical diagnosis.

Psychosis prevention relies on early detection of individuals at clinical high risk for psychosis (CHR-P) remains limited, constraining preventive care. The effectiveness of the CHR-P state is constrained, in part due to clinical assessments requiring specialist interpretation of narrative interviews, limiting scalability. Here, we evaluate whether large language models (LLMs; deep learning models trained on large text corpora to process and generate language) can extract clinically meaningful information from such interviews to support psychosis risk assessment. We assessed 11 open-weight LLMs on 678 PSYCHS interview transcripts from 373 participants (77.7% CHR-P). Models inferred CHR-P status and estimated severity and frequency across 15 symptom domains, benchmarked against researcher-rated scores. Larger models achieved the strongest classification performance (Llama-3.3-70B: accuracy = 0.80, sensitivity = 0.93, specificity = 0.58). LLM-generated symptom scores showed good correlations with researcher-rated scores (ICCsev = 0.74, ICCfreq = 0.75). Performance disparities were minimal across most demographic groups but varied across sites. Generated summaries were largely faithful to source transcripts, with low rates of clinically relevant confabulation (3%). Errors primarily reflected over-pathologisation of non-clinical experiences. While accuracy scaled with model size, smaller models achieved competitive performance with substantially lower computational cost. These findings demonstrate that open-weight LLMs can assess psychosis risk from clinical interview transcripts, supporting scalable, human-in-the-loop approaches to early detection.

Understanding depressive symptoms dynamics and their determinants is crucial for designing effective mental health support initiatives. This study compared two methods for describing youth depressive symptoms trajectories and investigated associations of early-life factors (maternal education, maternal perinatal depression, domestic violence, physical, emotional, or sexual abuse, bullying victimisation, psychiatric disorder) with trajectory features. Prospective data from 8,264 mostly White European participants (54% female), including self-reported Short Moods and Feelings Questionnaires on ten occasions between 10-25 years, were used. Trajectories were summarised using functional principal component analysis (FPCA) and P-splines linear mixed-effect (PLME) models. Estimated derivatives were used to obtain magnitude and age of peak symptoms and peak symptoms velocity. Both methods performed comparably, but PLME models tended to over-smooth trajectories. Peak symptoms and peak velocity were higher and occurred >1 year earlier in females than males. All early-life factors were associated with higher peak symptoms, and most associated with higher and earlier peak velocity. Abuse and bullying additionally associated with earlier age of peak symptoms. FPCA is a useful alternative for characterising depressive symptoms trajectories and informing time-sensitive preventative measures to reduce impact of depression before symptoms reach their peak. Early-life stressors may accelerate timeline and intensity of symptoms escalation during adolescence. Lay summaryUnderstanding development of depressive symptoms and factors shaping them is crucial for designing effective mental health support initiatives. This study used data from over 8,000 young people regularly followed up from before birth to compare two cutting-edge methods for describing depressive symptoms trajectories and examined how known risk factors for adulthood depression relate to the severity and rate of change of depressive symptoms in adolescence. We found that both methods performed well and that the peaks in depressive symptoms and their rate of change were, on average, higher and occurred over a year earlier in females than males. Our findings additionally suggest that early-life stressors (e.g., abuse, bullying) may accelerate the development of depression, highlighting the importance of early prevention.

BackgroundAdolescent major depressive disorder (AMDD) is a prevalent and heterogeneous psychiatric condition that emerges during a critical period of brain development. Neuroimaging-based biomarkers derived from resting-state functional magnetic resonance imaging (rs-fMRI) hold promise for objective diagnosis; however, pronounced inter-individual variability and limited sample sizes pose major challenges for robust model development. MethodsWe propose a memory-augmented Meta-Graph Convolutional Network (BrainMetaGCN) to classify AMDD using rs-fMRI functional connectivity. Individual functional connectivity matrices were constructed by parcellating rs-fMRI time series into cortical regions of interest and computing pairwise correlations. A meta-graph generator dynamically learned subject-specific graph structures, which were processed by lightweight graph convolutional layers. A memory neural network was incorporated to encode population-level prototypical connectivity patterns and generate individualized representations via attention-based retrieval. Model performance was evaluated across multiple independent datasets and compared with state-of-the-art deep learning approaches. Additionally, network interpretability was examined through cortical hierarchy analysis and functional enrichment of discriminative network components. ResultsThe proposed BrainMetaGCN consistently outperformed baseline models, including convolutional and transformer-based approaches, achieving higher accuracy, area under the receiver operating characteristic curve, sensitivity, and specificity. Memory-module-derived functional networks exhibited clear modular organization and showed a significant positive correlation with cortical functional hierarchy, supporting their neurobiological validity. Functional enrichment analyses implicated synaptic transmission, axon guidance, receptor tyrosine kinase signaling, and immune-related pathways, suggesting neurodevelopmental and neuroimmune mechanisms underlying AMDD. Ablation analyses confirmed that memory augmentation and dynamic meta-graph construction were critical for robust performance under small-sample conditions. ConclusionsThis study introduces a robust and interpretable memory-augmented graph learning framework for AMDD classification. By effectively balancing individual specificity and population-level generalization, BrainMetaGCN advances neuroimaging-based precision diagnosis and provides new insights into the neural and biological mechanisms of adolescent depression.

The shared genetic liability between cortical morphology and psychiatric disorders remains unclear. We aimed to identify whether the genetic loci shared between cortical morphology and six psychiatric disorders show regional or global effects. We identified substantial pairwise genetic overlaps of cortical surface area or thickness with psychiatric disorders; however, these loci lacked a uniform direction (~50% concordance). Cross-trait analyses revealed distinct architectures: internalizing disorders and schizophrenia/bipolar disorder shared more genetic loci with localized effects, whereas neurodevelopmental disorders shared fewer loci but more with widespread effects. We identified 17 genomic loci shared across all disorders, most of which had opposing directional effects across cortical regions. Only one locus (rs2431112) had region-specific and unidirectional effects (reduced primary visual and posterior cingulate surface area). This directional heterogeneity within and across pleiotropic loci reveals complex shared genetic architectures and likely limits the genetic predictive performance of brain morphology for psychiatric disorders.

Children exposed to severe childhood maltreatment often develop complex mental health disorders where standard treatments show limited efficacy. Current residential approaches combine psychopharmacological with behavioural interventions, yet the feasibility and clinical-neurobiological outcomes of intensive, medication-free psychotherapy have not been investigated in this population. Our prospective study followed severely traumatized children (aged 6-13 years) with documented histories of changes and failures in placement.They completed an intensive 6-8 months inpatient treatment program (5 individual psychotherapy and 3 group therapy sessions per week with high caregiver-patient ratio) grounded in a novel, multimodal, attachment-based therapeutic framework. Medication was discontinued prior to treatment. The intervention group was compared to healthy controls and waitlist controls receiving treatment as usual. Participants in the intervention group achieved high remission rates for dysregulated behaviour (Child Behaviour Checklist (CBCL) >60% post treatment, 50% on follow-up) and trauma-related symptoms (Parent Report of Post-traumatic Stress Symptoms (PROPS) >65% post treatment, >60% on follow-up). Within-group effect sizes for Total Problems Score, Externalising behaviour (both CBCL), Hyperactivity (Strengths and Difficulties Questionnaire) and trauma symptoms (PROPS) each exceeded Cohen's d = 1.0 and were maintained at 6-month follow-up. Resting-state fMRI identified significant functional reorganization in visual processing networks. Atypical correlation patterns between visual network activity and symptom severity resolved following treatment, yielding patterns comparable to those of healthy controls. These pilot findings provide initial evidence of the feasibility and effectiveness of intensive, medication-free, attachment-based inpatient treatment to promote clinical remission and neurobiological normalization in severely traumatized children.

Schizophrenia spectrum disorder (SSD) is a clinically and genetically heterogeneous condition, yet few studies have integrated real-world clinical data with both common and rare genetic variation to explore this complexity. In this study, we analyzed real-world data from 22,092 individuals in the Danish iPSYCH cohort (11,046 SSD cases and 11,046 matched population controls) leveraging nationwide registry data on diagnoses, hospitalizations, and parental history. Using a variational autoencoder (VAE), we compressed these features into a latent space and identified ten clinically distinct SSD subgroups that varied in comorbidity, parental diagnoses, hospital burden, and early-life adversity. Polygenic scores (PGSs) for five psychiatric disorders showed subgroup-specific enrichment, highlighting potential links between complex clinical profiles and common variant liability. In a subset with exome data (N=5,969), we assessed rare deleterious variant burden across SCZ-informed gene sets and Protein-Protein Interaction (PPI) networks, observing suggestive network-specific trends. This framework for integrating real world-based stratification with genetic evidence is scalable and transferable across cohorts, offering a path toward biologically informed patient classification.