Sample Entropy of High Frequency Oscillations for Epileptogenic Zone Localization

For epileptic patients whose seizures are poorly controlled with medication, removing the brain region responsible for seizure onset is a treatment option. This requires the epileptogenic zone (EZ) to be accurately delineated. In this paper, a two-stage approach for EZ delineation is proposed. The algorithm starts by detecting events of high-frequency oscillations (HFOs) directly from the multi-channel intracranial electroencephalograms (iEEGs). The sample entropy is then computed for each of their channels that will be used for determining the channels correlated with the EZ. The performance of our proposed method is evaluated using the receiver operating characteristic curve analysis, and the results indicate that our proposed approach can provide an accurate estimation of the EZ.