| Hi-index | 0.00 |
It's essential to classify neurons firings to study the dynamic mechanism of neural spikes. Spike feature extraction is a key step in automatic spike sorting. Spikes cann't be sorted correctly without proper features. This paper introduces a new method for extracting spike features. This method combines Hilbert-Huang transformation(HHT), which is an adpative time-frequency analysis technique to acquire the features of spikes. 1334 spikes form 100s recordings can be sorted into three clusters by our method. The number of spikes in each cluster is 355, 650 and 329 respectively. The new method is capable of sorting spikes from different neurons.