Eigenvectors and eigenfunctions in spatiotemporal EEG analysis.

Abstract

As a supplement to the article "An Eigenfunction Approach to the Inverse Problem of EEG" by Hjorth and Rodin in Brain Topography, 1988, 1 (2): 79-86, this paper discusses in greater detail the interrelations between the concepts of EEG sample vector, eigenvalue and eigenvector. It also describes how the method for assigning locations to uncorrelated EEG basic waveforms named eigenfunctions can be further developed by normalizing the EEG samples to unity global field power before computation of covariance. This enhances local persistence as a feature for revealing low-amplitude activity of possible diagnostic significance, even in the presence of more dominant activity.