The Expectation Based Eye-Brain-Computer Interface: An Attempt of Online Test

Abstract

In this preliminary study we tested online a new Eye-Brain-Computer Interface (EBCI) for selection of positions on a screen with a combination of gaze based control and a passive brain-computer interface (BCI). This hybrid BCI was trained offline to recognize the electroencephalogram (EEG) patterns recorded during gaze dwells intentionally used to make moves in a computer game. The patterns were presumably related to expectation of the interface feedback. In the online test, 500 ms gaze dwells led to actions each time the BCI classified them as intentional. When the BCI made a miss, a participant could still communicate the intention by prolonging the dwell up to 1000 ms. Also playing the game was possible, it was found that defining the ground truth for such an online system is not trivial and that further efforts will be needed to evaluate the performance of the expectation based EBCI reliably.