This prognostic study evaluates whether psychosis transition can be predicted in patients with clinical high-risk syndromes or recent-onset depression by multimodal machine learning that optimally integrates clinical and neurocognitive data, structural magnetic resonance imaging, and polygenic risk scores for schizophrenia.

APSS status was associated with greater suicidality and illness/psychophathology severity in this help-seeking cohort, emphasizing the need for clinical care. The age-related differences in the prevalence of APSS and the increasing proportion of APSS+/COGDIS+ may point to a higher proportion of non-specific/transient, rather than risk-specific attenuated positive symptoms in adolescents.

Performances on neurocognitive tests differed among our three at-risk groups and may therefore be useful in predicting psychosis. Overall, cognition had a profound effect on the extent of general functioning and satisfaction with life for subjects at risk of psychosis. Thus, this factor should become a treatment target in itself.

Our data provide no evidence for a progressive cognitive decline in individuals at risk of psychosis. In line with the neurodevelopmental model, our findings suggest that cognitive deficits are already present very early, before or during the prodromal stage of the illness.

No abstract