Unsupervised Spike Sorting Based on Discriminative Subspace Learning

TL;DR

This paper introduces two unsupervised spike sorting algorithms that learn discriminative subspaces for improved accuracy and robustness in neural data analysis, outperforming traditional methods like PCA and wavelet transforms.

Contribution

The paper presents novel unsupervised algorithms that simultaneously learn discriminative feature subspaces and perform clustering, enhancing spike sorting performance.

Findings

Higher accuracy in lower-dimensional feature spaces.

Robustness to noise surpassing existing methods.

Better cluster separability than PCA or wavelet-based approaches.

Abstract

Spike sorting is a fundamental preprocessing step for many neuroscience studies which rely on the analysis of spike trains. In this paper, we present two unsupervised spike sorting algorithms based on discriminative subspace learning. The first algorithm simultaneously learns the discriminative feature subspace and performs clustering. It uses histogram of features in the most discriminative projection to detect the number of neurons. The second algorithm performs hierarchical divisive clustering that learns a discriminative 1-dimensional subspace for clustering in each level of the hierarchy until achieving almost unimodal distribution in the subspace. The algorithms are tested on synthetic and in-vivo data, and are compared against two widely used spike sorting methods. The comparative results demonstrate that our spike sorting methods can achieve substantially higher accuracy in lower dimensional feature space, and they are highly robust to noise. Moreover, they provide significantly better cluster separability in the learned subspace than in the subspace obtained by principal component analysis or wavelet transform.