Detection and Sorting of Neural Spikes Using Wavelet Packets

Abstract
We propose a novel method for the detection and sorting of recorded neural spikes using wavelet packets. We employ the best basis via the Shannon's information cost function and local discriminant basis using mutual information. We demonstrate the efficiency of the method on data recorded in vitro from 2D neural networks. We show that our method is superior both in separation from noise and in identifying superimposed spikes.