AUG 4, 2026 · JOURNAL ARTICLE
Polygenic risk-informed white matter integrity improves deep learning-based prediction of youth depression
Communications Medicine
A deep learning model combining polygenic risk scores and neuroimaging achieves moderate discrimination (AUC 0.61–0.67) for predicting youth depression cross-sectionally and at 2-year follow-up, with cross-ethnic validation, but lacks comparison to clinical gold standards and requires external prospective validation.
Reported
Cross-sectional prediction AUC0.61 to 0.66
Two-year follow-up prediction AUC0.61 to 0.66
Independent Korean cohort AUC0.67