A Transfer Entropy-based methodology to analyze information flow under eyes-open and eyes-closed conditions with a clinical perspective

Abstract

Studying brain dynamics under normal or pathological conditions has proven to be a challenging task, as there is no unified consensus on the best approach. In this article, we present a methodology based on Transfer Entropy to study the information flow between different brain hemispheres in healthy subjects during eyes-open (EO) and eyes-closed (EC) resting states. We used an experimental setup that mimics the technical conditions found in clinical settings and collected data sets from short records of 24 channels electroencephalogram (EEG) at a sampling rate of 65 Hz. Our methodology accounts for interhemispheric and intrahemispheric information flow analysis in both conditions and relies on 4 indexes calculated from the transfer entropy estimations between EEG channels. These indexes provide information on the number, strength, and directionality of active connections. Our results suggest an increase in information transfer in the EC condition for the alpha, beta1, and beta2 frequency bands, but no preferred direction of interhemispheric information movement under either condition. These results are consistent with previously reported studies conducted with denser EEG recordings sampled at a higher rate. In conclusion, our methodology shows a significant difference in the brains dynamics of information transfer between EO and EC resting states, which can also be applied to regular clinical sessions.