A study on reducing training time of BCI system based on an SSVEP dynamic model

Abstract

In the field of steady-state visual evoked potential (SSVEP) based brain-computer interface (BCI), the lengthy training time was always an obstacle to practical application. In this paper, we explored a novel method to reduce the training cost by replacing the traditional sinusoidal template or signal template with a dynamic SSVEP model and conducting a sampling training strategy. To evaluate the method, the training time and the recognition accuracy under two conditions (sine/cosine template and dynamic model template) were compared on four different algorithms. The results showed that the dynamic model based template outstripped the sinusoidal template; and for signal template-based algorithms, our proposed method reduced the training time significantly while kept the decrease of performance within an insignificant range.