Discrimination between idle and work states in BCI based on SSVEP

We present a novel method for idle and work states classification in brain computer interface (BCI) based on steady-state visual evoked potentials (SSVEP). Canonical correlation analysis (CCA) and maximum contrast combination (MCC) are used to extract features of electroencephalogram (EEG) signals. The correlation coefficients from CCA and SNR from MCC were classified by a linear classifier. Then an extra strategy of excluding alpha wave interference helped improve the classification accuracy. This method had a good performance in real EEG signals.