ICA feature extraction for spike sorting of single-channel records

In neuroscience, an important class of signals are the extracellular actions potentials of neurons, which are called spikes. However, a single extracellular electrode can capture spikes from more then one cell. The process of sorting these spikes is typically made in some steps: detection, alignment, feature extraction and clustering. For the crucial feature extraction step, Principal Component Analysis (PCA) and Wavelet Transform are the most used methods. In this work we propose to use of Independent Component Analysis (ICA) for feature extraction associated with K-means, Fuzzy C-means (FCM) or Self Organizing Maps (SOM) in the clustering step. Our results demonstrate that using ICA as preprocessing gives better cluster of spikes separation than the other feature extraction methods, which yields a better final sorting accuracy using simulated data.