On the use of clustering and local singular spectrum analysis to remove ocular artefacts from electroencephalograms

Abstract

In this work we present a method based on singular spectrum analysis to remove ocular artifacts (EOG) from an Electroencephalogram (EEG). After embedding the EEG signals in a feature space of time-delayed coordinates, the data are clustered and the principal components (PCs) of each cluster are computed. We assume that the EOG artifact is associated with the PCs belonging to largest eigenvalues. We incorporate a Minimum Description Length (MDL) criterion to estimate the number of eigenvalues which correspond to the EOG artifact. The extracted EOG signal is subtracted from the original EEG signal to obtain the corrected EEG signal we are interested in.