Performance Evaluation of a "Switch-To-Target" Based Asynchronous SSVEP BCI Paradigm

Because of the strong advantages such as anti-interference capability, less electrode montage, higher information transfer rate and no need of specific training for participants, steady-state visual evoked potentials (SSVEP) based brain-computer interface (BCI) has attracted more and more attentions. In order to achieve high accuracy and maintain considerable stability, this paper proposed an eye tracking technique based asynchronous SSVEP BCI method by directly localizing asynchronous eye-tracking-based switch to desired stimulation target to accelerate the BCI process. And real-time visual gaze feedback was also provided in necessity when participants could not focus their gaze to achieve satisfactory target identification results. By combining the heterogeneous signals of eye gaze position with a conventional asynchronous BCI paradigm, the proposed method reduced the trial duration while considerably high identification accuracy was maintained. Experiments were carried out on four participants with an average accuracy above 93% when the trial duration was 3s, while considerable performance can also be achieved when the trial length was shorten to 2s.