Classification of steady state visual evoked potential in EEG source space with application in brain-computer interfaces.

Abstract

: Brain computer interface any type of communication between the brain and an electronic system such as a computer that begins after brainwaves are recorded on them and usually results in a feedback path to the individual. This type of communication has different types and may be invasive or non-invasive. In this study, we have used a non-invasive communication that works using EEG brain signals. To implement a BCI system, there are various methods. One of these methods is the use of SSVEP, which works based on stable visual stimuli in the human brain. The BCI system based on the SSVEP has received more attention due to the use of a high information transfer rate, a small number of electrodes, and a short training time compared to other BCI systems. In this study, we use a method to train the combination of CCA and TRCA methods to improve the detection performance of stimulus frequencies in the channel and source space simultaneously. The obtained results show that the accurate detection and information transfer rate of the proposed method in SSVEP frequency detection is better than other methods. The obtained results show that the proposed method has improved the performance by 8% compared to the TRCA method in the time window of 0.4 seconds, which was the biggest difference in accuracy in this time window. Keywords: brain-computer interface (BCI); steady-state visual evoked potential (SSVEP); source reconstruction; canonical correlation analysis (CCA); Task related component analysis (TRCA)