Heterogeneity and Classification of Recent Onset Psychosis and Depression: A Multimodal Machine Learning Approach

Abstract

Diagnostic heterogeneity within and across psychotic and affective disorders challenges accurate treatment selection, particularly in the early stages. Delineation of shared and distinct illness features at the phenotypic and brain levels may inform the development of more precise differential diagnostic tools. We aimed to identify prototypes of depression and psychosis to investigate their heterogeneity, with common, comorbid transdiagnostic symptoms. Analyzing clinical/neurocognitive and grey matter volume (GMV) data from the PRONIA database, we generated prototypic models of recent-onset depression (ROD) vs. recent-onset psychosis (ROP) by training support-vector machines to separate patients with ROD from patients with ROP, who were selected for absent comorbid features (pure groups). Then, models were applied to patients with comorbidity, ie, ROP with depressive symptoms (ROP+D) and ROD participants with sub-threshold psychosis-like features (ROD+P), to measure their positions within the affective-psychotic continuum. All models were independently validated in a replication sample. Comorbid patients were positioned between pure groups, with ROP+D patients being more frequently classified as ROD compared to pure ROP patients (clinical/neurocognitive model: χ2 = 14.874; P < .001; GMV model: χ2 = 4.933; P = .026). ROD+P patient classification did not differ from ROD (clinical/neurocognitive model: χ2 = 1.956; P = 0.162; GMV model: χ2 = 0.005; P = .943). Clinical/neurocognitive and neuroanatomical models demonstrated separability of prototypic depression from psychosis. The shift of comorbid patients toward the depression prototype, observed at the clinical and biological levels, suggests that psychosis with affective comorbidity aligns more strongly to depressive rather than psychotic disease processes. Future studies should assess how these quantitative measures of comorbidity predict outcomes and individual responses to stratified therapeutic interventions.

Keywords: MRI; comorbidity; depression; gray matter volume; machine learning; psychosis; transdiagnostic.

Fig. 1.

Classification performance of the pure and applied clinical and neurocognitive, GMV, and combined models. (A) Pure clinical and neurocognitive classification balanced accuracy 79.3%, sensitivity 76.3%, specificity 82.2%, AUC 0.86. Applied Clinical classification balanced accuracy 57.6%, sensitivity 39.1%, specificity 76%, AUC 0.71. (B) Classification performance of the pure GMV model and the applied GMV model. Pure GMV classification balanced accuracy 62.5%, sensitivity 39.5%, specificity 85.6%, AUC 0.70. Applied GMV classification balanced accuracy 50.3%, sensitivity 20.7%, specificity 80%, AUC 0.47. (C) Classification performance of the combined model and the applied combined model. Stacked classification balanced accuracy 79.5%, sensitivity 78.9%, specificity 80%, AUC 0.87. Applied stacked classification balanced accuracy 64.6%, sensitivity 53.3%, specificity 76%, AUC 0.71.

Fig. 2.

Feature weights and cross-validations ratios of the most significant features. (A) Feature weights. Derived from 1000 random permutations of the outcome labels and features. (B) Cross-validation ratio. Sum of the median weights across all CV1 folds divided by the standard error.

Fig. 3.

Significant regions in the imaging classification model. ROP GMV reductions in the thalamus and the cerebellum, ROD GMV reductions in orbitofrontal, limbic, and paralimbic regions.