An Adaptive Filtering Approach to the Processing of Single Sweep Event Related Potentials Data

We present an adaptive filtering method applied to the single sweep event related potentials data. The method is based on time-sequenced version of the typical adaptive signal enhancers. Two different signals are used at the reference input and the results are compared. The evaluation is made using both simulated and real measurement data. It is found that the use of a reference input composed of the eigenvectors of the desired target signals, improved the estimation error while minimizing the distortion for the non target sweeps. It also enables a better tracking of the sweep to sweep changes of the ERP waveform. KeywordsEEG, ERP, adaptive filter, single sweep, principal eigenvector