Implicit application of head orientation in brain-computer interface to improve egocentric movement control in virtual reality

Abstract

Brain-computer interfaces (BCIs)-based control applications that utilize motor imagery (MI) often suffer from the inaccurate classification of user's intention. While one common approach to this problem is to use hybrid BCI, the issue with most hybrid BCIs is that the added paradigm or modalities usually requires the user to demonstrate additional intentions. Inspired by earlier studies on the relationship between the direction of head orientation and locomotion, we propose a hybrid BCI system that jointly considers EEG and head orientation to predict the user's intention in a task that controls the movement of a first-person-view virtual drone. The proposed system does not ask the user to consciously express the intention through head movement as the idea of using head orientation is based on the results that head movement is likely to implicitly include information about the intended direction of locomotion. In the experiment that compared the virtual drone control performance of the MI-BCI control system and the proposed system in a virtual reality environment, eighteen out of twenty participants were able to show better performance in completing the task with the proposed system. Given that participants were not informed about the involvement of head orientation in the experiment, the proposed method presents a novel hybrid BCI system that implicitly uses head orientation which allows the user to concentrate solely on exhibiting MI signals while improving the control performance.