Connectivity-Driven Registration

This project aims to use connectivity information from resting-state functional MRI and/or diffusion MRI tractography, to drive the registration (alignment) of the human cortex between individuals. Cortical brain regions are not clearly differentiated from each other on the basis of T1-weighted intensities or geometric folding, but have different functional and structural connectivity profiles.  However, most current registration methods base their alignment only on structural image intensity and geometry. Since resting-state fMRI and diffusion MRI sequences are commonly acquired (or easily added, with a few extra minutes of scanning) this rich source of information is readily available, and can be accurately registered within-subject using state-of-the-art distortion correction methods.  The between-subject registration combines this high-dimensional connectivity information with diffeomorphic spherical registration methods to perform cortical alignment.