On YouTube and Zoom:
Beyond group averages: using machine learning and big data to chart individual variability in mental disorders
Neuroscience has truly transitioned into the era of big data, yet most analytical approaches still focus principally on detecting group level differences between clinical cohorts. In this talk I will describe a series of machine learning techniques we have developed to move the field beyond this impasse, including normative modelling or ‘brain growth charting’ techniques that allow us to chart variability at the level of each individual person, providing a platform for precision stratification of mental disorders. I will illustrate this discussion by showing applications in schizophrenia, autism, ADHD and bipolar disorder which show that the prevalent approach is overly simplistic and disguises considerable variability in most mental disorders. The methodological approaches we have developed provide a principled approach to parse the heterogeneity in mental disorders and a workable route towards precision medicine in psychiatry.