Jacobi Iteration for Spatially Constrained Independent Component Analysis

Electroencephalography (EEG) measures the potentials generated by the brain. Unfortunately, eye artifacts often contaminate EEG recordings. Independent Component Analysis (ICA) has been proposed to remove these artifacts. Standard implementations of ICA extract independent components from a set of measured signals in a fully blind way. In the case of eye artifacts, there is strong prior information about the mixing vector of this artifact component. We provide an algorithm which incorporates this knowledge into the calculation of the independent sources and use this constrained ICA technique to remove eye artifacts from the EEG. The advantage of this method is that the components do not have to be inspected afterwards in order to detect the artifactual ones.