People presenting with first-episode psychosis (FEP) have heterogenous outcomes. More than 40% fail to achieve symptomatic remission. Accurate prediction of individual outcome in FEP could facilitate early intervention to change the clinical trajectory and improve prognosis.

We aim to systematically review evidence for prediction models developed for predicting poor outcome in FEP.

A protocol for this study was published on the International Prospective Register of Systematic Reviews, registration number CRD42019156897. Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidance, we systematically searched six databases from inception to 28 January 2021. We used the Checklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies and the Prediction Model Risk of Bias Assessment Tool to extract and appraise the outcome prediction models. We considered study characteristics, methodology and model performance.

Thirteen studies reporting 31 prediction models across a range of clinical outcomes met criteria for inclusion. Eleven studies used logistic regression with clinical and sociodemographic predictor variables. Just two studies were found to be at low risk of bias. Methodological limitations identified included a lack of appropriate validation, small sample sizes, poor handling of missing data and inadequate reporting of calibration and discrimination measures. To date, no model has been applied to clinical practice.

Future prediction studies in psychosis should prioritise methodological rigour and external validation in larger samples. The potential for prediction modelling in FEP is yet to be realised.

About 30% of patients drop out of cognitive–behavioural therapy (CBT), which has implications for psychiatric and psychological treatment. Findings concerning drop out remain heterogeneous.

This paper aims to compare different machine-learning algorithms using nested cross-validation, evaluate their benefit in naturalistic settings, and identify the best model as well as the most important variables.

The data-set consisted of 2543 out-patients treated with CBT. Assessment took place before session one. Twenty-one algorithms and ensembles were compared. Two parameters (Brier score, area under the curve (AUC)) were used for evaluation.

The best model was an ensemble that used Random Forest and nearest-neighbour modelling. During the training process, it was significantly better than generalised linear modelling (GLM) (Brier score: d = –2.93, 95% CI (−3.95, −1.90)); AUC: d = 0.59, 95% CI (0.11 to 1.06)). In the holdout sample, the ensemble was able to correctly identify 63.4% of cases of patients, whereas the GLM only identified 46.2% correctly. The most important predictors were lower education, lower scores on the Personality Style and Disorder Inventory (PSSI) compulsive scale, younger age, higher scores on the PSSI negativistic and PSSI antisocial scale as well as on the Brief Symptom Inventory (BSI) additional scale (mean of the four additional items) and BSI overall scale.

Machine learning improves drop-out predictions. However, not all algorithms are suited to naturalistic data-sets and binary events. Tree-based and boosted algorithms including a variable selection process seem well-suited, whereas more advanced algorithms such as neural networks do not.

Autism spectrum disorder (ASD) is a highly heterogeneous disorder that affects nearly 1 in 189 females and 1 in 42 males. However, the neurobiological basis of gender differences in ASD is poorly understood, as most studies have neglected females and used methods ill-suited to capture such differences.

To identify robust functional brain organisation markers that distinguish between females and males with ASD and predict symptom severity.

We leveraged multiple neuroimaging cohorts (ASD n = 773) and developed a novel spatiotemporal deep neural network (stDNN), which uses spatiotemporal convolution on functional magnetic resonance imaging data to distinguish between groups.

stDNN achieved consistently high classification accuracy in distinguishing between females and males with ASD. Notably, stDNN trained to distinguish between females and males with ASD could not distinguish between neurotypical females and males, suggesting that there are gender differences in the functional brain organisation in ASD that differ from normative gender differences. Brain features associated with motor, language and visuospatial attentional systems reliably distinguished between females and males with ASD. Crucially, these results were observed in a large multisite cohort and replicated in a fully independent cohort. Furthermore, brain features associated with the motor network's primary motor cortex node predicted the severity of restricted/repetitive behaviours in females but not in males with ASD.

Our replicable findings reveal that the brains of females and males with ASD are functionally organised differently, contributing to their clinical symptoms in distinct ways. They inform the development of gender-specific diagnoses and treatment strategies for ASD, and ultimately advance precision psychiatry.

Despite efforts to predict suicide risk in children, the ability to reliably identify who will engage in suicide thoughts or behaviours has remained unsuccessful.

We apply a novel machine-learning approach and examine whether children with suicide thoughts or behaviours could be differentiated from children without suicide thoughts or behaviours based on a combination of traditional (sociodemographic, physical health, social–environmental, clinical psychiatric) risk factors, but also more novel risk factors (cognitive, neuroimaging and genetic characteristics).

The study included 5885 unrelated children (50% female, 67% White, 9–11 years of age) from the Adolescent Brain Cognitive Development (ABCD) study. We performed penalised logistic regression analysis to distinguish between: (a) children with current or past suicide thoughts or behaviours; (b) children with a mental illness but no suicide thoughts or behaviours (clinical controls); and (c) healthy control children (no suicide thoughts or behaviours and no history of mental illness). The model was subsequently validated with data from seven independent sites involved in the ABCD study (n = 1712).

Our results showed that we were able to distinguish the suicide thoughts or behaviours group from healthy controls (area under the receiver operating characteristics curve: 0.80 child-report, 0.81 for parent-report) and clinical controls (0.71 child-report and 0.76–0.77 parent-report). However, we could not distinguish children with suicidal ideation from those who attempted suicide (AUROC: 0.55–0.58 child-report; 0.49–0.53 parent-report). The factors that differentiated the suicide thoughts or behaviours group from the clinical control group included family conflict, prodromal psychosis symptoms, impulsivity, depression severity and history of mental health treatment.

This work highlights that mostly clinical psychiatric factors were able to distinguish children with suicide thoughts or behaviours from children without suicide thoughts or behaviours. Future research is needed to determine if these variables prospectively predict subsequent suicidal behaviour.

Response to lithium in patients with bipolar disorder is associated with clinical and transdiagnostic genetic factors. The predictive combination of these variables might help clinicians better predict which patients will respond to lithium treatment.

To use a combination of transdiagnostic genetic and clinical factors to predict lithium response in patients with bipolar disorder.

This study utilised genetic and clinical data (n = 1034) collected as part of the International Consortium on Lithium Genetics (ConLi+Gen) project. Polygenic risk scores (PRS) were computed for schizophrenia and major depressive disorder, and then combined with clinical variables using a cross-validated machine-learning regression approach. Unimodal, multimodal and genetically stratified models were trained and validated using ridge, elastic net and random forest regression on 692 patients with bipolar disorder from ten study sites using leave-site-out cross-validation. All models were then tested on an independent test set of 342 patients. The best performing models were then tested in a classification framework.

The best performing linear model explained 5.1% (P = 0.0001) of variance in lithium response and was composed of clinical variables, PRS variables and interaction terms between them. The best performing non-linear model used only clinical variables and explained 8.1% (P = 0.0001) of variance in lithium response. A priori genomic stratification improved non-linear model performance to 13.7% (P = 0.0001) and improved the binary classification of lithium response. This model stratified patients based on their meta-polygenic loadings for major depressive disorder and schizophrenia and was then trained using clinical data.

Using PRS to first stratify patients genetically and then train machine-learning models with clinical predictors led to large improvements in lithium response prediction. When used with other PRS and biological markers in the future this approach may help inform which patients are most likely to respond to lithium treatment.

Clinical high-risk states for psychosis (CHR) are associated with functional impairments and depressive disorders. A previous PRONIA study predicted social functioning in CHR and recent-onset depression (ROD) based on structural magnetic resonance imaging (sMRI) and clinical data. However, the combination of these domains did not lead to accurate role functioning prediction, calling for the investigation of additional risk dimensions. Role functioning may be more strongly associated with environmental adverse events than social functioning.

We aimed to predict role functioning in CHR, ROD and transdiagnostically, by adding environmental adverse events-related variables to clinical and sMRI data domains within the PRONIA sample.

Baseline clinical, environmental and sMRI data collected in 92 CHR and 95 ROD samples were trained to predict lower versus higher follow-up role functioning, using support vector classification and mixed k-fold/leave-site-out cross-validation. We built separate predictions for each domain, created multimodal predictions and validated them in independent cohorts (74 CHR, 66 ROD).

Models combining clinical and environmental data predicted role outcome in discovery and replication samples of CHR (balanced accuracies: 65.4% and 67.7%, respectively), ROD (balanced accuracies: 58.9% and 62.5%, respectively), and transdiagnostically (balanced accuracies: 62.4% and 68.2%, respectively). The most reliable environmental features for role outcome prediction were adult environmental adjustment, childhood trauma in CHR and childhood environmental adjustment in ROD.

Findings support the hypothesis that environmental variables inform role outcome prediction, highlight the existence of both transdiagnostic and syndrome-specific predictive environmental adverse events, and emphasise the importance of implementing real-world models by measuring multiple risk dimensions.