A new method based on wavelet and greedy pusuit analysis for neuro-spike detection

This paper introduces a novel greedy model for spike detection system aimed to neural signal obtained from the brain such as EEG, ECOG. This Presented algorithm combines wavelet transform and simultaneous orthogonal matching pursuit (SOMP). The proposed method is capable of data compression for spike detection. The spike detection experiment demonstrates the TPR in ROC curve is up to 98.76% and the compact factor reaches 42.52 when the length of window we used is 500. Moreover, our method is simple for low power system design and real-time hardware realization.