Mixed Effects Model for Functional Surfaces with Applications to Cortical Surface Task fMRI
share
Motivated by cortical surface task fMRI, we propose a framework for jointly modeling the geometry and functionality in high-dimensional functional surfaces. The proposed model characterizes effects of subject-specific covariates and exogenous stimuli on functional surfaces while accounting for the mutual-influence of their geometry and the functionality. This is accomplished through a computationally efficient estimation method developed for the proposed mixed effects model, incorporating regularized estimation of the precision matrix of random effects. We apply the proposed approach to cortical surface task fMRI data of the Human Connectome Project and discover geometric shapes of cortical surface and activated regions in the fMRI associated with demographics and task stimuli. In particular, new modes of correspondence between the shapes and activation relevant to emotion processing are revealed.
READ FULL TEXT