The invention discloses a brain-computer interface method based on steady-state asymmetric
visual evoked potentials, and the method comprises the steps: constructing an SSaVEP coding normal form withN instruction sets, enabling a visual field
fixation point to be at an original point, and enabling a visual stimulation
signal to appear around the original point in two different time periods; acquiring electroencephalogram
signal data of a subject by using a Neuroscan Synamps2
system, and performing preprocessing to extract a feature
signal; constructing a
spatial filter, and respectively obtaining a
training set template and a
test set template by utilizing the
spatial filter to realize feature enhancement of the SSaVEP; extracting correlation coefficients of the
training set template andthe
test set template by using a CCA
algorithm to maximize the potential correlation. In
mode matching, the relationship between the
training set template and the test signal is represented by a vector, a final
correlation coefficient decision value is further obtained, and a target category is obtained based on the decision value. According to the technology, the signal-to-
noise ratio of SSVEP can be remarkably improved, and the technology is suitable for a
low frequency band, a
medium frequency band and a
high frequency band; and meanwhile, visual fatigue can be reduced by utilizing a hiddenvisual fixation mode.