A comparison of minimum norm and MUSIC for a combined MEG/EEG sensor array

Abstract

Abstract. Many different algorithms for imaging neuronal activity with magnetoencephalography (MEG) or electroencephalography (EEG) have been developed so far. We validate the result of other authors that a combined MEG/EEG sensor array provides smaller source localisation errors than a single MEG or single EEG sensor array for the same total number of sensors. We show that Multiple Signal Classification (MUSIC) provides smaller localisation errors than an unweighted minimum norm method for activity located in the cortical sulcus regions. This is important for many medical applications, e.g. the localisation of the origin of epileptic seizures (focal epilepsy) that can be located very deep in the cortical sulcus.