Multimodal prediction of psychotic-like experiences using elastic net modeling: external validation in a clinical...

Area of Science:

• Neuroscience
• Psychiatry
• Genetics

Background:

• Psychotic-like experiences (PLEs) exist on a continuum with psychosis.
• Multimodal factors contribute to PLEs, but comprehensive research is lacking.
• This study models multiple factors predicting PLEs in a large youth sample.

Purpose of the Study:

• To develop and validate a model predicting psychotic-like experiences (PLEs) using multimodal factors.
• To examine if these predictors generalize to individuals with psychosis.
• To identify intermediate phenotypes across the psychosis spectrum.

Main Methods:

• Variable selection from 741 factors (environmental, cognitive, clinical, brain connectome).
• Elastic Net classification model predicting high/low PLEs in 396 healthy adolescents and young adults (14-24 years).
• External validation in a clinical cohort (first-episode psychosis patients, siblings, controls).

Main Results:

• Eleven factors (environmental, cognitive appraisals) and 16 structural brain network properties predicted PLEs.
• The model achieved moderate accuracy (75%) and AUC (0.750) for predicting PLEs.
• High specificity (84.2%) was observed in distinguishing psychosis patients from their siblings.

Conclusions:

• Multimodal factors including environmental burden, cognitive schemas, and brain network alterations predict PLEs.
• These predictors partially generalize to clinical psychosis, suggesting shared mechanisms.
• Identified variables may represent intermediate phenotypes relevant to psychosis vulnerability and resilience.