Identification of multimodal mental health signatures in the young population using deep phenotyping

BackgroundMental health comprises emotional, psychological, and social dimensions, extending beyond the mere absence of illness. Shaped by a complex interplay of hereditary factors and life experiences, mental health can deteriorate into clinical conditions necessitating intervention. However, ambiguity between pathological and non-pathological states underscores the need for a dimensional approach to early risk detection and stratified psychiatry. MethodsWe analyzed multimodal data from approximately 300 young adults, including psychometric assessments, structural brain imaging, genomic data, and fasting-state plasma metabolomics. Using a psychometry-based soft-clustering approach (archetyping), we stratified participants based on cognitive, emotional, and behavioral traits. We evaluated associations between archetypes, biological features, and mental health outcomes both cross-sectionally and at five-year follow-up. ResultsWe identified five psychometric archetypes, capturing diverse psychological profiles that span a continuum from vulnerability to resilience. Archetypes at the vulnerable end, which were marked by emotional dysregulation and high neuroticism, were associated with elevated polygenic risk for psychiatric disorders, altered cortical structures in emotion-related regions, and metabolomic profiles previously linked to psychopathology. By contrast, resilient archetypes were characterized by emotional stability and adaptive functioning. Archetype scores were prospectively associated with symptom burden, and models guided by archetype-associated biological features outperformed agnostic models in predicting clinical outcomes, supporting their clinical relevance. ConclusionsThis study demonstrates that data-driven psychometric archetypes reflect biologically grounded variation in mental health and can inform prospective risk stratification. This approach offers a framework for understanding mental health heterogeneity and holds promise for advancing early screening and targeted interventions in the young population.