A high performance steady state visual evoked potential BCI system based on variational mode decomposition

Precise and efficient detection of steady state visually evoked potentials (SSVEP) in EEG for brain computer interface (BCI) programs is crucial. However canonical correlation analysis (CCA) have been employed profitably on SSVEP noises and artifacts which occur in the time of signal recording can abridge the detection rate, by extracting frequency sub-bands from the original signal the effect of artifacts and noises could be reduced. In this article a novel enhanced method for detection of SSVEP-related frequency components by variational mode decomposition (VMD) accompanied by CCA is suggested. Results claim that the suggested approach examined on EEG signals of 3 visual stimuli, demonstrates an increased performance in comparison with employing CCA on the original signal.