Peter the Great Saint Petersburg Polytechnic University

Abstract. We offer an approach to the inter-channel connectivity analysis of EEG using correlation ratios which can objectively estimate characteristics of interactions in all their models. It is shown that the values of the correlation ratios can serve as an indicator of nonlinear correlation. The ability to analyze multiple interactions is shown on the basis of the proposed correlation ratio matrix. The concept of dominant channels is introduced. Algorithms and analysis software are discussed.

Key words: inter-channel electroencephalogram analysis,  EEG, correlation ratio, correlation analysis, nonlinear connectivity.