Classifying neuronal spikes from multiunit recording by using a multisite electrode

A method of classifying neuronal spikes was proposed and achieved by multiunit recording using a multisite electrode and by considering spatial damping vectors. The method consists of three procedures, i.e., separation of overlapping spikes, treatment of burst spikes, and clustering by a multi-dimensional statistical test. The first and second procedures clarify the border between clusters and the last procedure validates the classification statistically. By using this method, multiple neuronal spikes were automatically classified and the waveforms of the spikes in each class were well superimposed. The authors concluded that this method is superior to other methods of classifying neuronal spikes