Discriminant analysis of functional connectivity patterns on Grassmann manifold

The functional brain networks, extracted from fMRI images using independent component analysis, have been demonstrated informative for distinguishing brain states of cognitive function and brain disorders. Rather than analyzing each network encoded by a spatial independent component separately, we propose a novel algorithm for discriminant analysis of functional brain networks jointly at an individual level. The functional brain networks of each individual are used as bases for a linear subspace, referred to as a functional connectivity pattern, which facilitates a comprehensive characterization of fMRI data. The functional connectivity patterns of different individuals are analyzed on the Grassmann manifold by adopting a principal angle based Riemannian distance. In conjunction with a support vector machine classifier, a forward component selection technique is proposed to select independent components for constructing the most discriminative functional connectivity pattern. The discriminant analysis method has been applied to an fMRI based schizophrenia study with 31 schizophrenia patients and 31 healthy individuals. The experimental results demonstrate that the proposed method not only achieves a promising classification performance for distinguishing schizophrenia patients from healthy controls, but also identifies discriminative functional brain networks that are informative for schizophrenia diagnosis.