A novel extracellular spike detection algorithm based on sparse representation

Abstract

Identification of spikes in the extracellular recording signals is a technical challenge because of large amounts of background noise and contributions of many neurons to recorded signals. In this paper, a novel method based on sparse representation is proposed for high accuracy and robust spike detection. Considering the diversity of spikes, a universal dictionary is first learned for giving a sparse representation to various spike signals. In addition, in order to improve the robustness to noise, we propose to use sparse coefficients as features for the discrimination of spikes. Finally, the number and locations of spike events in the recorded signal are determined through a thresholding process. Experimental results on both synthesized extracellular neural recordings and real data demonstrate that the proposed method performs much better than the existing methods in terms of both robustness and flexibility.