Factoring the Coherence Matrix: Patterning of the Frequency‐Specific Covariance in a Multichannel EEG

Current methods in research on the electrophysiology of human information processing include high speed quantification of EEG signals from multiple electrodes. The data from each electrode comprise a time series, from which the frequency information can be extracted through Fourier analysis. A problem in this research is appropriately handling the large amount of data in such a way that the topographic patterning of the electrical activity can be related to differential usage of brain regions during specific forms of cognition. We illustrate how factor analysis can be applied to describe the regional patterning of covariance in the spectral information of a multichannel EEG. Rotated factors delineate distinct patterns of electrophysiological activity across the head. Coherence factors show good replicability over repeated observations, and lateralized effects of cognitive activity are observed in the factor loadings. Further applications of this method are discussed, including some suggestions relevant to applying confirmatory, hypothesis testing factor analyses.