Principal component analysis (PCA) of complex spike activity recorded simultaneously from multiple Purkinje cells
Qin Hu, Eric J. Lang, Izumi Sugihara · 2003
A matrix of microelectrodes is inserted into the layer of Purkinje cells of the cerebellum cortex of a rat. Each wavelet of the cell activity is a burst of positive and negative pulses (complex spike), but interest is only in the onset instants, which are recognized by software from the recorded pulses. Four formats are designed to store and analyze the onset sequences: time-principal-component-analysis (PCA), space-PCA, and unbalanced forward and backward PCAs. Good results are obtained by analyzing the count/bin spikes. Clustering of cells on the plane of eigenvectors 1 and 2 and PCA-invariably are found.>