The assessment of two methods for removing eye movement artefact from the EEG

A quantitative assessment of both computerised correlation and analogue techniques for the removal of eye movement artefact from the electroencephalogram was undertaken. In both methods a fraction of the measured EOGs was subtracted from the measured EEG to leave the corrected EEG. In the correlation method the correction factors were computed from the cross-correlations of the EOGs and measured EEGs. In the analogue method the fraction subtracted was derived using a potential divider arrangement. The relative effectiveness of the two methods was determined by comparing the autocorrelation functions of the corrected EEGs at a lag of 2 sec. It was found that the computerised correlation technique was superior. A 3-channel EOG correlation correction procedure was found adequate in which two horizontal and one vertical EOGs were used. The analogue technique was very time consuming and possibly erroneous.