A spatially robust ICA algorithm for multiple fMRI data sets

In this paper we derive an independent-component analysis (ICA) method for analyzing two or more data sets simultaneously. Our model extracts independent components common to all data sets and independent data-set-specific components. We use time-delayed autocorrelations to obtain independent signal components and base our algorithm on prediction analysis. We applied this method to functional brain mapping using functional magnetic resonance imaging (fMRI). The results of our 3-subject analysis demonstrate the robustness of the algorithm to the spatial misalignment intrinsic in multiple-subject fMRI data sets.