Detection and Discrimination of SSVEP by Multiple Stimuli Using Blind Spots

The brain-computer interface (BCI) using steady-state visual evoked potentials (SSVEPs) is realized by detecting the frequency components of the electroencephalogram (EEG) which are identical to the flashing frequency of gazed stimuli. SSVEP-BCI in which the user gazes at the space between the multiple adjacent stimuli can detect multiple selections with a small number of stimuli; however, due to contamination of the frequency component evoked by nontarget stimuli, a drop in detection accuracy is often seen. To address this issue, we have proposed a novel approach that masks nontarget stimuli adjacent to target stimuli by a physiological blind spot. Using this approach may allow for the suppression of SSVEP responses to stimuli masked by physiological blind spots and the improvement of BCI performance. In this study, we first conducted a psychophysical experiment to investigate the location and size of the physiological blind spot in five subjects. Furthermore, we conducted an EEG measurement experiment with the SSVEP task, in which two stimuli are presented so that one stimulus is hidden by the blind spot, based on the information on the location and size of the blind spot obtained in the previous experiment for each subject. As a result, it was shown that for all subjects, the classification accuracy of the gaze point was higher when the stimulus that was not being gazed at was masked by the physiological blind spot than when it was not masked. In addition, the detection accuracy improved by 28.9% for one subject. These results demonstrate the importance of physiological blind spots, which have not been considered in BCI using visual stimuli to date, and suggest that the functional use of blind spots is effective in improving BCI performance.