Imaging Science & Engineering seminar | Janine D. Bijsterbosch

About this event

Population Neuroimaging for Depression Subtyping and Classification: What Works, What Doesn't, and What's Next?

Research into the neural mechanisms of psychopathology is traditionally performed in dedicated patient-control study samples with robust clinical diagnostics. The recent availability of large-scale population neuroimaging data offers a new lens on the brain basis of mental health. In this talk, I will discuss what population data can and cannot tell us about depression, drawing on several recent large-scale studies from our lab. I will cover advantages of population data, including improved power, reliability, and transdiagnostic classification. I will also discuss key challenges raised by population data, including considerations for data harmonization, lack of convergence across subtyping approaches, and broader challenges in modeling depression heterogeneity. I will end with a discussion on next steps we’re pursuing, including teasers on individualized resting state network topography and the development of dedicated statistical approaches for population data.