Spike Sorting Based on Low-Rank and Sparse Representation

As the first step to study the coding mechanism and synergistic behaviour of neurons, spike sorting plays an important role in the neurosciences research community. Despite many empirical successes in spike sorting models, there are still sufferings from the overlapping and noise corruption problems. To ease these situations, in this paper, we present an efficient and effective method with the help of optimization theory. Firstly, by introducing the low-rank strategy, the global structure underlying the spike data could be discovered. Secondly, by engaging the sparse coding to balance the noise, the proposed model is robust in the overlapping and noise spike sorting scenario. We have conducted experiments on the Wave_clus dataset compared with two state of the art models. The results verify the efficacy of our scheme and confirm the claims above.