Decoding vulnerability to mental disorders: an artificial intelligence approach.

Lecture presented at the Covian Symposium (2020, online), organised by the Department of Physiology, Ribeirão Preto Medical School, University of São Paulo (FMRP-USP). The symposium honours Miguel Rolando Covian, a pioneer of modern physiology in Brazil, and brings together researchers working on advances across physiology and neuroscience. An interdisciplinary science of psychopathology is essential to understand the mechanisms of dysfunction and to open new possibilities for early detection, diagnosis, treatment and prognosis. This lecture presents advances in the diagnosis and prognosis of psychiatric disorders based on machine-learning analysis of brain-imaging data. It shows evidence that machine learning, a subfield of artificial intelligence, can predict the symptoms of mental disorders and personality traits from patterns of brain activity. This approach may reveal biomarkers that help identify, at the individual level, people at higher risk of mental disorders.