Combining Auditory and Visual Stimuli to Improve the Performance of Covert Attention-based BCIs

Abstract

Brain-Computer Interface (BCI) spellers allow their users to communicate with the outside world without moving or speaking. The ac- curacy of well-known BCI spellers such as the matrix speller and the Hex-o-Spell interface has been shown rely on eye gaze (overt attention): if directing gaze to attended options is prohibited, their accuracy drops signi cantly. In this paper, we perform a pilot experiment to test whether the performance of a visual covert attention- based BCI (in which eye movements are not allowed) can be improved by adding auditory stimuli. We tested two conditions (visual-only vs. visual and auditory stimuli) for a P300- based variant of this type of BCI. For the 4 subjects we tested, adding auditory stimuli did not have a large e ect on classi er performance.