Quantification and rejection of ocular artifacts in auditory evoked fields in schizophrenics

In a magnetoencephalographic investigation of the auditory evoked field (AEF) in 17 schizophrenics and 17 controls, 37% of the schizophrenics and 12% of the controls showed eye artifacts in every second trial or even more frequently. In the uncorrected average fields, the ratio between the power of artifacts and the power of the magnetoencephalogram (MEG) exceeded the value of 0.1 for 48% of the schizophrenics and for 29% of the controls. Ocular artifacts biased the locations of equivalent current dipoles of the M100 component towards deeper positions. A regression algorithm for the correction of ocular artifacts in raw data and an identification technique of ocular artifacts based on the topography of transmission coefficients is described.

A linear dependence of ocular artifacts in AEF on the electrooculogram (EOG) was confirmed. Possible errors introduced by the correction are discussed. Transmission coefficients should be calculated for several individual trials with the same type of artifact. Errors due to evoked potentials in the EOG were found to be comparable in amplitude to noise in the AEF. Examples of transmission coefficients from the EOG to the MEG are given.