Abstract: The aim of this work is to introduce correlation test to detect artifacts in visual evoked potentials. If an artifact occurs a sample signal deviates from the ensemble average. In this paper signals with less correlation coefficient are considered as artifacts and are removed from the further analysis. An attempt has been made to apply these techniques to 14 -channel visual evoked potentials (VEPs) obtained from different subjects.

Keywords: evoked potentials, ensemble average, and correlation.