Estimation of evoked potentials using wavelet transform based time-frequency adaptive filtering

A time-frequency domain adaptive filtering method is presented to estimate evoked potentials. The wavelet transform is used to represent the original responses in the time-frequency domain with a discrete set of wavelet coefficients. Each coefficient, which is related to the time extent and the frequency extent in the time-frequency plane, is processed by an adaptive signal enhancer (ASE) to enhance the signal components. The processed coefficients are then used to reconstruct the evoked potential signals with the inverse wavelet transform. Visual evoked potentials (VEPs) from human subjects are estimated, and good results are obtained by this method