Motion-onset visual evoked potentials for gaming: A pilot study

Abstract

This paper details a pilot study for a motion onset Visual Evoked Potential (mVEP) based Brain Computer Interface (BCI) controlled game. mVEP is a type of VEP that uses visual responses from the dorsal pathway of the visual system allowing elegant visual stimuli to elicit different brain patterns depending on the motion and position of the stimuli. The study here was conducted to determine the most appropriate methods, parameters and EEG setup to use in order to extract reliable information when classifying responses on up to five different stimuli. Initial offline results show that 80% accuracy can achieved by averaging stimuli over 5 seconds when discriminating target versus non target. This was achieved by the use of simple averaging techniques and support vector machines. The initial results are encouraging, showing that mVEP may be used as a control system within a computer game. Details of the proposed games are also included.