An expectation-based EEG marker for the selection of moving objects with gaze

Abstract

The use of an EEG expectation-related component, the expectancy wave (E-wave), in brainmachine interaction was proposed more than 50 years ago, but active exploration of this possibility has started only recently, in the context of developing passive brain-computer interfaces for the enhancement of gaze interaction. We report, for the first time, the results of a systematic experimental study that revealed an EEG marker for selecting intentionally an object among other moving objects using smooth pursuit eye movements. This marker appeared to have the same nature as the Ewave previously observed in the EEG accompanying the selection of static objects with gaze fixations. A convolutional neural network classified the intentional and spontaneous smooth pursuit eye movements with average ROC AUC 0.69±0.13 (M±SD). These results suggest that the E-wave might be robust enough to serve, after further improvement of the methodology, as the basis of hybrid eye-brain-computer interfaces applied for selection in dynamically changing visual environments.