Time varying filters for estimation of evoked potentials

Presents a method for the estimation of evoked potentials (EPs) using an a posteriori time varying filter based on the discrete wavelet transform (DWT). The proposed filter uses alternate subensemble averaging (ASA) to arrive at the signal and noise scalogram estimates. This filter, when applied on the ensemble average (EA), leads to an improvement in the SNR when applied on the individual trials, the ongoing EEG is significantly eliminated, revealing the underlying EP. The proposed filter has been tested on simulated data.