ICA of fMRI data: Performance of three ICA algorithms and the importance of taking correlation information into account

Abstract

Independent component analysis (ICA) has proven useful for the analysis of functional magnetic resonance imaging (fMRI) data. In this paper, we compare the performance of three ICA algorithms and show the importance of taking sample correlation information into account. The three ICA algorithms are Infomax, the most widely used algorithm for fMRI analysis, entropy bound minimization (EBM) that adapts to a wide range of source distributions, and full blind source separation (FBSS) which has the ability to incorporate a flexible density model along with sample correlation information. We apply these three ICA algorithms to fMRI data from multiple subjects performing an auditory oddball task (AOD). We show that FBSS leads to significant improvement in the estimation of both the spatial activation and the time courses of several components. More importantly, by taking the correlation information into account, the default mode network (DMN) component, an important one in the study of brain function, is more consistently estimated using FBSS.