National Institute of Mental Health 5R01MH123610-05

Data-driven solutions for temporal, spatial, and spatiotemporal dynamic functional connectivity

Award to Georgia State University for developing advanced spatio-temporal models using deep learning to study dynamic brain connectivity in mental illnesses, with validation through simulations and large datasets.

Description

This project aims to develop flexible spatio-temporal models leveraging joint blind source separation and deep learning to study dynamic brain connectivity, particularly in mental illnesses such as schizophrenia, bipolar disorder, and autism spectrum disorder. The models will characterize nonlinear spatio-temporal dynamics, relax assumptions of fixed spatial nodes, and produce metrics for in-depth analysis. The project will validate approaches through simulations and large normative datasets, and apply tools to understand disruptions in brain dynamics associated with mental health conditions.

View original record