Eye-Movement-Control-Independent Brain Computer Interface Using Modulation of Steady-State Responses in Visual Evoked Potentials

Abstract

Brain computer interfaces based on steady-state visually evoked potential (SSVEP-BCI) have been developed substantially in recent years, but these are not available to patients with severe paralysis or eye-impairment who lost the control ability of eye-movement. This paper proposes an eye-movement-independent SSVEP-BCI available in the eyes-closed state based on the modulation of SSVEP elicited by performing a mental task. Although performance of the proposed BCI depended on subjects, electrode locations, mental tasks employed, and flickering frequencies, the mean precision and recall, which were obtained from the confusion matrix, reached 72% to 95% using the support vector machine classifier across 18 normal subjects under the stimulus frequency of 10Hz or 14Hz. Results from simulated information transfer rate and its inter-individual difference suggest that it is adequate to set an inter-trial interval at 2s to 3s for better performance of the proposed BCI. It is consequently feasible to develop a practical eye-movement-control-independent BCI by optimizing the parameters such as the stimulation frequency and electrode sites each user.

Keywords

• brain computer interface
• brain machine interface
• steady-state visually evoked potential