David Cotter

Background: Psychotic disorder represents a leading cause of disability worldwide, and relapse in psychosis is common. Artificial intelligence (AI) is increasingly recognized as a method that could aid clinical monitoring for individuals experiencing psychosis. Objective: This review aims to map the existing literature on AI-based approaches—including machine learning, deep learning, and natural language processing—used to detect relapse in individuals with psychotic disorders. Methods: A systematic search strategy was conducted on PubMed, PsycINFO, and Embase up to January 7, 2026. Observational studies, randomized controlled trials, and quasi-experimental studies that used AI methods to detect relapse in psychosis were eligible for inclusion. Screening and data extraction procedures were conducted by at least 2 reviewers working independently. Findings were extracted, charted, and described using narrative synthesis based on data extraction and consensus meetings with the research team. The scoping review was prospectively registered with the Open Science Framework. Results: Relevant studies identified (N=10) included the use of digital tools such as smartphone- and smartwatch-based monitoring, ecological momentary assessment tools, social media activity, and internet searches. Digital phenotyping via smartphones and wearables emerged as the most common method for data collection. The efficacy of AI models varied with sensitivity (or recall) ranging from 0.25 to 0.77 and specificity (or precision) ranging from 0.06 to 0.88. The reported area under the receiver operating characteristic curve for models ranged from 0.63 to 0.78. AI models were heterogeneous across studies, and most study findings were not replicated. Conclusions: This scoping review highlights both the promise and the current limitations of AI in psychosis relapse detection. Passive digital phenotyping research in the detection of psychosis relapse has progressed, and personalized approaches with individual-level modeling show promise; however, further studies need to include larger numbers of participants and should incorporate methods such as large language models. Future studies will require large collaborations aimed at delivering AI methods for use in real-world clinical practice.

Background: Psychotic experiences (PEs) in are associated with elevated risk for mental health difficulties. This study examined predictors of PEs, inclusive of the role of gender, ethnicity, and protective factors. Methods: Data were drawn from a 2021 Planet Youth survey of adolescents (n = 4,005). PEs were measured using the adolescent psychotic symptom screener. Effects of psychosocial predictors on PEs were measured by fitting multivariable logistic regression main effect and joint exposure models. Results: 29.8% reported PEs. Black/Asian/Other minorities had elevated odds (a OR = 1.59, 95% CI 1.26–2.02, p < .001). Increased odds in males, females, undisclosed gender and non-binary/transgender with elevated emotional/behavioural difficulties (a OR = 4.47, 95% CI 3.53–5.67, p < .001; a OR = 3.25, 95% CI 2.59–4.08, p < .001; a OR = 4.83, 95% CI 2.58–9.02, p < .001; a OR = 4.33, 95% CI 2.69–6.97, p < .001 respectively). High odds in undisclosed gender with low emotional/behavioural difficulties (a OR = 4.36, 95% CI 1.50–12.66, p = .007). Lower odds from perceived school/home safety (a OR = 0.69, 95% CI 0.58–0.83, p < .001 and (a OR = 0.81, 95% CI 0.66–0.99, p = .038, respectively). Elevated odds from recent adversities (a OR = 1.91, 95% CI 1.47–2.49, p = .011) attenuated by parental support (a OR = 1.76, 95% CI 1.17–2.65, p < .001). Each additional adversity (>12 months) increased odds (a OR = 1.12, 95% CI 1.07–1.17, p < .001). Conclusions: Findings highlight the interplay of risk and protective factors in adolescent PEs, with increased vulnerability among minoritized youth. Results support targeted interventions to reduce mental health disparities.

Childhood trauma (CT) is associated with cognitive impairment across major psychiatric disorders. We tested a novel transdiagnostic hypothesis that atypical connectivity of the default mode network (DMN) mediates the association between childhood trauma (CT) and cognitive impairment. The sample of 1851 individuals aged 18-25 included 433 patients with depression, eating disorders, alcohol use disorder, psychosis and ADHD) and were recruited as part of the ESTRA/STRATIFY/IMAGEN studies. CT was measured using the Childhood Trauma Questionnaire (CTQ). The CANTAB spatial working memory task was administered to assess cognition. Four a priori seeds of the default mode network (DMN) were measured during face processing, namely the medial prefrontal cortex (PFC), right lateral parietal (LP), left lateral parietal (LP) and posterior cingulate cortex (PCC), according to the Harvard-Oxford Cortical and Subcortical Atlas ( http://www.cma.mgh.harvard.edu/fsl_atlas.html ) as implemented in CONN. Patients had significantly reduced DMN connectivity between the four chosen DMN seeds and the rest of the brain. Reduced DMN connectivity mediated the association between higher CT and worse cognitive performance. Our findings are transdiagnostic in nature with stronger effects in some regions observed in depression, and suggest one transdiagnostic cortical network via which CT’s effects on cognition are transmitted.

Background: Most mental health difficulties (MHD) emerge during adolescence and early adulthood, placing young people at an increased risk for co-occurring physical and sexual health challenges. Shared models of care (SMOC) to connect specialist mental health care with physical and/or sexual health have been developed to address these health needs among young people with MHD. We aimed to identify and characterise SMOC that integrate physical and/or sexual healthcare for young people with MHD, and to synthesise SMOC implementation determinants using the Consolidated Framework for Implementation Research (CFIR) for policy makers, commissioners and practitioners seeking to strengthen youth-integrated service delivery. Methods: Five electronic databases and key grey literature sites were searched in October 2024. Studies were eligible for inclusion if they predominantly included young people (aged 10–25) with an MHD. SMOC had to address MHD as a primary concern or have parity with the physical and/or sexual health concern(s) being addressed. Key study details were extracted and were appraised using the mixed methods appraisal tool. Screening was conducted in duplicate, with extraction and appraisals conducted by one team member and verified by a second. Findings were thematically synthesised and mapped to CFIR domains to inform implementation planning in youth health systems. Results: Search results identified 25 relevant SMOC to include in the review. Almost all models ( n = 23/25) addressed shared care between mental and physical health, while nine addressed mental and sexual health and seven addressed mental, physical and sexual health needs. Reporting quality varied but most SMOC included referral pathways, assessment, treatment and external support components. Barriers frequently mapped to the inner and outer setting CFIR domains, with high staff turnover ( n = 9) and societal stigma towards mental health ( n = 7) common concerns. Enablers frequently mapped to the process and innovation constructs, including offering youth-specific care models ( n = 7) and clear communication between services ( n = 5). Conclusions: Despite evidence supporting the need for an integrated care approach, implementation remains limited by setting-specific barriers. Findings highlight the need for service planning and developing tailored, youth-specific models to ensure a holistic approach to care is available to young people experiencing MHD. Registration: Open Science Framework (osf.io/rj783).

Advances in proteomic assay methodologies and genomics have significantly improved our understanding of the blood proteome. Schizophrenia and psychosis risk are linked to polygenic scores for schizophrenia and other mental disorders, as well as to altered blood and saliva levels of biomarkers involved in hormonal signaling, redox balance, and chronic systemic inflammation. The Accelerating Medicines Partnership® Schizophrenia (AMP®SCZ) aims to ascertain biomarkers that both predict clinical outcomes and provide insights into the biological processes driving clinical outcomes in persons meeting CHR criteria. AMP®SCZ will follow almost 2000 CHR and 640 community study participants for two years, assessing biomarkers at baseline and two-month follow-up including the collection of blood and saliva samples. The following provides the rationale and methods for plans to utilize polygenic risk scores for schizophrenia and other disorders, salivary cortisol levels, and a discovery-based proteomic platform for plasma analyses. We also provide details about the standardized methods used to collect and store these biological samples, as well as the study participant metadata and quality control measures related to preanalytical factors that could influence the values of the biomarkers. Finally, we discuss our plans for analyzing the results of blood- and saliva-based biomarkers. Watch Dr. Perkins discuss their work and this article: https://vimeo.com/1062879582?share=copy#t=0.

Dysregulation of inflammatory mediators and complement cascade proteins has been implicated in psychosis. In the current study, we aimed to investigate the relationship between complement cascade proteins and inflammatory cytokines in blood from people at clinical high risk (CHR) for psychosis and at first episode of psychosis (FEP). Baseline blood samples from two cohorts of CHR participants [NEURAPRO (n = 153) and STEP (n = 146)], and one cohort of FEP patients [OPTiMiSE (n = 226)] were included. The blood levels of three Inflammatory markers including Interleukin (IL)-6, Tumour necrosis factor-alpha (TNF-α) and C-reactive protein (CRP) along with about 30 complement proteins were considered for the analyses. First, we evaluated the interrelationship between the inflammatory markers and then using regression models, we investigated their association with complement proteins. We detected positive associations among all three inflammatory markers IL-6, TNF-α, and CRP in CHR individuals, whereas in FEP positive association was observed only between IL-6 and TNF-α. Regression models showed strong positive associations for complement proteins C3, C4A, C4B, C5, CFB and CFI with all three inflammatory markers in both CHR cohorts. This indicates the presence of a complement related pro-inflammatory tone at risk of developing psychosis. In contrast, in the FEP cohort, complement proteins C1QA, C3, C5, FCN-2, and MASP2 showed an inverse association with TNF-α, and no association found with IL-6 or CRP. These results suggest a switch in the immune activity in the peripheral circulation of FEP compared to CHR. These novel findings propose that complement protein-targeted anti-inflammatory therapy could be effective at CHR state and hence could be used for early intervention in psychosis.

Accumulating evidence suggests individuals with psychotic disorder show abnormalities in metabolic and inflammatory processes. Recently, several studies have employed blood-based predictors in models predicting transition to psychotic disorder in risk-enriched populations. A systematic review of the performance and methodology of prognostic models using blood-based biomarkers in the prediction of psychotic disorder from risk-enriched populations is warranted. Databases (PubMed, EMBASE and PsycINFO) were searched for eligible texts from 1998 to 15/05/2023, which detailed model development or validation studies. The checklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) was used to guide data extraction from eligible texts and the Prediction Model Risk of Bias Assessment Tool (PROBAST) was used to assess the risk of bias and applicability of the studies. A narrative synthesis of the included studies was performed. Seventeen eligible studies were identified: 16 eligible model development studies and one eligible model validation study. A wide range of biomarkers were assessed, including nucleic acids, proteins, metabolites, and lipids. The range of C-index (area under the curve) estimates reported for the models was 0.67-1.00. No studies assessed model calibration. According to PROBAST criteria, all studies were at high risk of bias in the analysis domain. While a wide range of potentially predictive biomarkers were identified in the included studies, most studies did not account for overfitting in model performance estimates, no studies assessed calibration, and all models were at high risk of bias according to PROBAST criteria. External validation of the models is needed to provide more accurate estimates of their performance. Future studies which follow the latest available methodological and reporting guidelines and adopt strategies to accommodate required sample sizes for model development or validation will clarify the value of including blood-based biomarkers in models predicting psychosis.

Striatal dysfunction has been implicated in the pathophysiology of schizophrenia, a disorder characterized by positive symptoms such as hallucinations and delusions. Haloperidol is a typical antipsychotic medication used in the treatment of schizophrenia that is known to antagonize dopamine D2 receptors, which are abundantly expressed in the striatum. However, haloperidol’s delayed therapeutic effect also suggests a mechanism of action that may go beyond the acute blocking of D2 receptors. Here, we performed proteomic analysis of striatum brain tissue and found more than 400 proteins significantly altered after 30 days of chronic haloperidol treatment in mice, namely proteins involved in glutamatergic and GABAergic synaptic transmission. Cell-type specific electrophysiological recordings further revealed that haloperidol not only reduces the excitability of striatal medium spiny neurons expressing dopamine D2 receptors (D2-MSNs) but also affects D1-MSNs by increasing the ratio of inhibitory/excitatory synaptic transmission (I/E ratio) specifically onto D1-MSNs but not D2-MSNs. Therefore, we propose the slow remodeling of D1-MSNs as a mechanism mediating the delayed therapeutic effect of haloperidol over striatum circuits. Understanding how haloperidol exactly contributes to treating schizophrenia symptoms may help to improve therapeutic outcomes and elucidate the molecular underpinnings of this disorder.