Study of removal of artifacts in MEG using PCA and ICA

Principal Component Analysis (PCA) and Independent Component Analysis (ICA) are two representative algorithms in Blind Source Separation. In this paper, a novel method for removal of artifacts in Magnetoencephalography (MEG) by combining PCA and ICA is presented. The basic concepts and algorithms of PCA and ICA are introduced firstly, MEG data are decomposed by PCA method in order to reduce the dimension of the original signals and take the redundancies out for getting the main components of data. Then the de-dimensioned data are further processed by using the adaptive Infomax algorithm of ICA. The study shows that the various artifacts can be separated from the MEG successfully and that removal of artifacts can be realized by signal reconstruction.