PP-PCA eigenvalues analysis of cerebral cortex local field potentials

In this article, the main eigenvalues of the non-linear method of principle component analysis based on projection pursuit (PP-PCA) are used to analyze the multi-channels local field potentials (LFPs) in rat prefrontal cortex under the different states during Propofol anesthesia, and the results of PP-PCA and linear principal component analysis (PCA) are compared. The principles of 16-channels LFPs data are calculated with 100-ms window sliding and 40% overlap. Meanwhile, the changes of the PP-PCA eigenvalues during the states of wake and anesthesia are applied to analyzing. The result shows that, the first eigenvalues of PP-PCA are much larger than the others both during the periods of wake and anesthesia, whereas during anesthesia, the eigenvalues´ values decrease rapidly, by contrast, they are characterized by a slower reducing during waking time. These prove that the method of PP-PCA effectively represents the characterization of better synchronization in LFPs during anesthesia.