Combined high resolution EEG and MEG data for linear inverse estimate of human event-related cortical activity

A new spatial deblurring method for the modeling of human event-related cortical activity from electroencephalography (EEG) and magnetoencephalography (MEG) data is proposed. This method includes high surface sampling of EEG-MEG data (128-50 sensors), realistic magnetic resonance-constructed subject's multi-compartment (scalp, skull, dura mater, cortex) head model, multi-dipole source model, and regularized linear inverse estimate based on boundary element mathematics. As a novelty, linear inverse estimates are regularized not assuming that covariance of background electromagnetic noise between sensors was zero. EEG and MEG data were recorded (separate sessions) while two normal subjects executed voluntary right one-digit movements. Linear inverse estimates of movement-related cortical activity from the combined EEG and MEG data showed higher spatial information content than those obtained from the MEG and EEG data considered separately. In conclusion, the new spatial deblurring method represents a powerful multi-modal neuroimaging approach to the noninvasive study of human brain functions.