Refining Schizophrenia Risk Assessment: Machine Learning Delineates a Brain Signature of Cognitive Basic Symptoms

Research Square (Research Square), Год журнала: 2025, Номер unknown

Опубликована: Апрель 1, 2025

Biological risk signatures could aid the early detection of schizophrenia, but their precision likely depends on clinical definitions they are derived from. Using machine learning, we analyzed structural MRI data from 1,425 patients and 907 healthy individuals in a multi-site multi-diagnostic database to detect validate different syndromes―Cognitive Disturbances (COGDIS), Ultra-High-Risk (UHR) or COGDIS + UHR―compared schizophrenia. Patients with not UHR-related syndromes were detectable using MRI. schizophrenia brain highly correlated based shared prefronto-parieto-perisylvian volume reductions. The expressions these intensified through affective disorders non-schizophrenic psychoses be predicted up 21% variance explained by genetic, neurocognitive, phenotypic factors. signature poor functional outcomes after two years. These results emphasize need refine tools integrating cognitive basic symptoms neural correlates into assessment.

Язык: Английский

---

Disrupted Hierarchical Functional Brain Organization in Affective and Psychotic Disorders: Insights from Functional Brain Gradients

Опубликована: Апрель 17, 2025

Patients with psychosis and depression show widespread alterations in brain resting-state functional connectivity (rs-FC), affecting both sensory higher-order regions. In this study, we investigate disruptions the hierarchical organization of networks patients psychotic affective disorders. We derived gradients, low dimensional representations rs-FC that capture cortical hierarchy, a large patient sample including clinical high-risk for (CHR-P) patients, recent-onset (ROP) (ROD) healthy controls (HC). examined regional alterations, network-level differentiation their relationship to symptoms. addition, linked case-control differences receptor expression maps explore underlying neurobiological mechanisms. All groups exhibited visual-to-sensorimotor gradient, while only ROP showed association-to-sensory gradient. CHR-P lower values ventral attention network. Additionally, combined higher somatomotor network, reduced gradient range altered between-network dispersion. ROD within-network dispersion attentional range. Correlational analysis revealed weak associations measures functioning, visual dysfunctions cognition. Furthermore maps, suggesting involvement neurotransmitter systems these disruptions. Our findings reveal transdiagnostic disease-specific organization. These indicate deficits integration across psychiatric diseases, highlighting role disease processes.

Язык: Английский

---

Adolescent neurodevelopment and psychopathology: The interplay between adversity exposure and genetic risk for accelerated brain ageing

Developmental Cognitive Neuroscience, Год журнала: 2023, Номер 60, С. 101229 - 101229

Опубликована: Март 15, 2023

In adulthood, stress exposure and genetic risk heighten psychological vulnerability by accelerating neurobiological senescence. To investigate whether molecular brain network maturation processes play a similar role in adolescence, we analysed genetic, as well longitudinal task neuroimaging (inhibitory control, incentive processing) early life adversity (i.e., material deprivation, violence) data from the Adolescent Brain Cognitive Development study (N = 980, age range: 9–13 years). Genetic was estimated separately for Major Depressive Disorder (MDD) Alzheimer's Disease (AD), two pathologies linked to allegedly sharing causal connection (MDD-to-AD). Adversity MDD/AD jointly predicted functional segregation patterns suggestive of accelerated (GABA-linked) visual/attentional, but delayed (dopamine [D2]/glutamate [GLU5R]-linked) somatomotor/association system development. A positive relationship between psychopathology emerged only among less vulnerable adolescents, thereby implying that normatively maladaptive neurodevelopmental alterations could foster adjustment more exposed genetically susceptible youths. Transcriptomic analyses suggested sensitivity may underpin joint effect MDD/AD, line with proposed negative emotionality precursor AD, likely account alleged impact MDD on dementia onset.

Язык: Английский

---

Protocol for the Psychosis Immune Mechanism Stratified Medicine (PIMS) trial: a randomised double-blind placebo-controlled trial of single-dose tocilizumab in patients with psychosis

BMJ Open, Год журнала: 2023, Номер 13(3), С. e067944 - e067944

Опубликована: Март 1, 2023

Introduction Evidence suggests a potentially causal role of interleukin 6 (IL-6), pleiotropic cytokine that generally promotes inflammation, in the pathogenesis psychosis. However, no interventional studies patients with psychosis, stratified using inflammatory markers, have been conducted to assess therapeutic potential targeting IL-6 psychosis and elucidate mechanism effect. Tocilizumab is humanised monoclonal antibody receptor inhibit signalling, licensed UK for treatment rheumatoid arthritis. The primary objective this study test whether contributes first episode examine mechanisms by which affects psychotic symptoms. A secondary characteristics inflammation-associated Methods analysis proof-of-concept employing randomised, parallel-group, double-blind, placebo-controlled design testing effect inhibition on anhedonia Approximately 60 participants diagnosis schizophrenia related disorders (ICD-10 codes F20, F22, F25, F28, F29) evidence low-grade inflammation (IL-6≥0.7 pg/mL) will receive either one intravenous infusion tocilizumab (4.0 mg/kg; max 800 mg) or normal saline. Psychiatric measures blood samples be collected at baseline, 7, 14 28 days post infusion. Cognitive neuroimaging data baseline In addition, approximately 30 without (IL-6<0.7 matched healthy controls recruited complete identical assessments allow comparison characteristic features Ethics dissemination sponsored University Bristol has approved Cambridge East Research Committee (reference: 22/EE/0010; IRAS project ID: 301682). Study findings published peer-review journals. Findings also disseminated scientific presentation other means. Trial registration number ISRCTN23256704 .

Язык: Английский

---

Anhedonia as a Potential Transdiagnostic Phenotype With Immune-Related Changes in Recent-Onset Mental Health Disorders

Biological Psychiatry, Год журнала: 2024, Номер 96(7), С. 615 - 622

Опубликована: Май 30, 2024

BackgroundChronic low-grade inflammation is observed across mental disorders and associated with difficult-to-treat-symptoms of anhedonia functional brain changes – reflecting a potential transdiagnostic dimension. Previous investigations have focused on distinct illness categories in those enduring illness, few exploring inflammatory changes. We sought to identify an signal function underlying among young people recent onset psychosis (ROP) depression (ROD).MethodResting-state magnetic resonance imaging, markers, symptoms were collected from N=108 (M age=26.2[SD 6.2]years; Female =50) participants ROP (n=53) ROD (n=55) the EU-FP7-funded PRONIA study. Time-series extracted using Schaefer atlas, defining 100 cortical regions interest. Using advanced multimodal machine learning, marker model connectivity developed classify anhedonic group, compared normal hedonic group.ResultsA repeated nested cross-validation markers classified ROP/ROD groups balanced accuracy (BAC) 63.9%, area under curve (AUC) 0.61. The produced BAC 55.2% AUC 0.57. Anhedonic group assignment was driven by higher levels Interleukin-6, S100B, Interleukin-1 receptor antagonist, lower Interferon gamma, addition within precuneus posterior cingulate.ConclusionWe identified subtype that accounted for profile connectivity. Results implications as emerging target disorders.

Язык: Английский

---

Biomarker discovery using machine learning in the psychosis spectrum

Biomarkers in Neuropsychiatry, Год журнала: 2024, Номер 11, С. 100107 - 100107

Опубликована: Авг. 27, 2024

The past decade witnessed substantial discoveries related to the psychosis spectrum. Many of these resulted from pursuits objective and quantifiable biomarkers in tandem with application analytical tools such as machine learning. These approaches provided exciting new insights that significantly helped improve precision diagnosis, prognosis, treatment. This article provides an overview how learning has been employed recent biomarker discovery research spectrum, which includes schizophrenia, schizoaffective disorders, bipolar disorder psychosis, first episode clinical high risk for psychosis. It highlights both human animal model studies explores a varying range most impactful including cognition, neuroimaging, electrophysiology, digital markers. We specifically highlight applications opportunities impact noninvasive symptom monitoring, prediction future diagnosis treatment outcomes, integration methods traditional practice, personalized medicine approaches.

Язык: Английский

---

The challenges of using machine learning models in psychiatric research and clinical practice

European Neuropsychopharmacology, Год журнала: 2024, Номер 88, С. 53 - 65

Опубликована: Сен. 3, 2024

To understand the complex nature of heterogeneous psychiatric disorders, scientists and clinicians are required to employ a wide range clinical, endophenotypic, neuroimaging, genomic, environmental data biological mechanisms illness before this knowledge is applied into clinical setting. Machine learning (ML) an automated process that can detect patterns from large multidimensional datasets supersede conventional statistical methods as it both linear non-linear relationships. Due advantage, ML has potential enhance our understanding, improve diagnosis, prognosis treatment disorders. The current review provides in-depth examination of, offers practical guidance for, challenges encountered in application models research practice. These include curse dimensionality, quality, 'black box' problem, hyperparameter tuning, external validation, class imbalance, representativeness. particularly critical context psychiatry expected researchers will encounter them during stages model development deployment. We detail solutions best practices effectively mitigate outlined challenges. recommendations have reliability interpretability psychiatry.

Язык: Английский

---

Duration of the psychosis prodrome and its relationship to duration of untreated psychosis across all 12 DSM-IV psychotic diagnoses: Evidence for a trans-diagnostic process associated with resilience

European Neuropsychopharmacology, Год журнала: 2023, Номер 80, С. 5 - 13

Опубликована: Дек. 20, 2023

While duration of the psychosis prodrome (DPP) attracts attention in relation to developmental trajectory psychotic illness and service models, fundamental issues endure context dimensional-spectrum models psychosis. Among 205 epidemiologically representative subjects Cavan-Monaghan First Episode Psychosis Study, DPP was systematically quantified compared, for first time, across all 12 DSM-IV diagnoses. also compared with untreated (DUP) each then analysed premorbid features three age ranges: <12, 12-15 16-18 years. For diagnosis, medians both DUP were shorter than means, indicating common right-skewed distributions. Rank orders longest schizophrenia, intermediate other schizophrenia-spectrum psychoses, depression disorder not otherwise specified, shortest brief disorder, bipolar substance-induced though overlapping longer diagnoses except disorder. Across functional diagnoses, predicted by higher intelligence better adjustment during These findings indicate that, trans-diagnostically, share continuities, accordance a model illness, may reflect unitary process that has been dichotomized at subjective threshold along its trajectory. Better functioning years appears confer resilience delaying progression overt symptoms constitute particular target period psychosocial interventions.

Язык: Английский

---

Two multimodal neuroimaging subtypes of obsessive-compulsive disorder disclosed by semi-supervised machine learning

Journal of Affective Disorders, Год журнала: 2024, Номер 354, С. 293 - 301

Опубликована: Март 15, 2024

Язык: Английский

---

Brain texture as a marker of transdiagnostic clinical profiles in patients with recent-onset psychosis and depression

Nature Mental Health, Год журнала: 2024, Номер 2(1), С. 76 - 87

Опубликована: Янв. 2, 2024

Abstract The inter-relationships of voxels can be captured by the radiomics texture features across multiple spatial scales. Prediction models brain changes contrast feature in recent-onset psychosis (ROP) and depression (ROD) have recently been proposed, although validation these transdiagnostically at individual level investigation variability clinical profiles are lacking. Established prevention treatment approaches focus on specific diagnoses do not address heterogeneity manifold potential outcomes patients. Here we aimed to investigate cross-sectional utility for (1) identification psychopathological state (ROP ROD) (2) association individualized maps with symptom severity outcome profiles. We developed transdiagnostic based structural magnetic resonance imaging data 116 patients ROD, 122 ROP 197 healthy control participants from PRONIA (Personalized pROgNostic tools early psychosIs mAnagement) study applying a set frameworks explain classification decisions deep-learning algorithm (named explainable artificial intelligence) clustering analysis. investigated as key general state. discrimination power trained prediction model was >72% validated second independent age- sex-matched sample 137 ROP, 94 ROD 159 participants. Clustering analysis implemented map produced an intelligence algorithm, group fashion. explained grouped into eight homogeneous clusters. In group, between well Different patterns showed unique associations alterations outcomes, that is, age, positive, negative depressive symptoms, functionality. some clusters, mean values and/or contributed significantly decision predicted accurately PANSS (positive scale) scores, functionality change over time. conclusion, created clusters which predict profile

Язык: Английский

---