Dynamic time segment selection in steady state visual evoked potential detection

The research in non-invasive Brain-Computer Interface (BCI) has led to significant improvements in the recent years. However, the user experience and the BCI illiteracy problem remain key issues to address for obtaining robust and resilient applications. In this paper, we address the choice of the time segment for the detection of steady state visual evoked potential (SSVEP) detection. The choice of this parameter is typically fixed and has a direct influence on the accuracy of detection, and therefore the information transfer rate. We propose to shift the problem of the time segment to the choice of the threshold for determining if a response has been properly detected. We consider an open-dataset of 10 participants to validate the rationale of the approach. The results support the conclusion that an adaptive time segment can lead to a better ITR on average across participants compared to a fixed time segment equal to the average of the mean adapted time segment for each subject. The ITR increases from 68.87 to 75.39 bpm with 12 targets, and from 54.20 to 72.66 bpm with 6 targets, highlighting the need of adaptive methods.