Comparison of independent component analysis algorithms for EEG-fMRI data fusion

Fusion of EEG and fMRI data helps researchers to provide a more comprehensive understanding of neural basis for the functional behavior in human brain. EEG and fMRI Joint analysis for cognitive tasks indicates plausible results to obtain a better spatiotemporal resolution of event related responses in the brain. Joint-ICA as a multivariate data analysis method, assumes more than two features type (modalities) have common mixing data and it tries to maximizes independency among joint components. Here, we study the performance of five ICA algorithms when applied to joint analysis of EEG/fMRI data. We use the visualization and computational tools to quantitatively analyze the performance of different ICA algorithms for EEG/fMRI fusion and discuss the results for the simulation and real data.