Permutation Conditional Mutual Information and Its Application to Epileptic EEG

The permutation conditional mutual information (PCMI) is proposed to estimate the directionality index between EEG signals. First, we demonstrate numerically that PCMI was a robust method to quantify the direction of information flow between time series from strongly and weakly coupled systems. Then, the PCMI method was applied to investigate the coupling direction between intracranial EEGs recorded from patients undergoing presurgical evaluation for drug resistant temporal lobe epilepsy. The propagation direction during the pre-seizure state could be elucidated through this coupling direction estimation method. The results suggest that the PCMI method has the potential of estimating directional coupling between epileptic EEGs.