Eyes-closed brain computer interface using modulation of steady-state visually evoked potential and auditory steady-state response

This paper proposes a binary brain computer interface (BCI) which is independent of eye-gaze and eye-movement. The proposed approach focuses on modulations of steady-state responses in the EEG to flicker (visual stimuli) and amplitude-modulated tone (auditory stimuli) independently. The modulations of the steady-state response could be elicited by performing a mental task under the flicker stimuli and paying attention to a target tone burst, respectively. For the BCI based on the modulation of steady-state visual response, we recruited eighteen normal subjects aged 21-24 with normal or corrected-to-normal vision. A linear discriminant classifier was used to discriminate between task-engaging and relaxed states. Classification performance depended on the subjects, electrode locations, mental tasks and flickering frequencies and the mean classification accuracy reached 78~88 % under the flickering frequency of 10 Hz and 61~88 % under that of 14 Hz across the subjects at the optimal electrode site, respectively. For another BCI based on the auditory steady-state response, the present study exploited a stochastic resonance effect and obtained the mean classification accuracy of 77 % for noise-added condition for the four electrode sites along one of the two lateral hemispheres across nine normal subjects aged 21-24 using a support vector machine. It is feasible to develop the eye gaze- and movement-independent BCI by optimizing the parameters such as stimulation type and electrode sites.