Examining associations between FMRI and EEG data using canonical correlation analysis

Functional magnetic resonance imaging (fMRI) and electro-encephalography (EEG) provide complementary information about the brain function. We propose a novel scheme to examine associations between these two modalities using canonical correlation analysis (CCA). Our multimodal canonical correlation analysis (mCCA) scheme utilizes inter-subject covariations to quantitatively link the two modalities and to estimate the spatio-temporal areas of association. We evaluate the performance of mCCA using simulated fMRI- and EEG-like data and note its ability to effectively identify associations across modalities. Also, our experiments on actual data from an auditory oddball task reveals associations of the temporal and motor areas with the N2 and P3 peaks, a finding that is consistent with previous studies. Additionally, we compare the performance of mCCA to the recently introduced joint-ICA technique for estimating spatio-temporal connections from multimodal data and discuss the advantages and limitations of each.