Research on Mutual Information Analysis in EEG Based on Different Entropy Measures

This paper mainly uses three different entropy estimation methods: linear estimation, kernel estimation, and k-nearest-neighbor estimation to calculate mutual information, respectively, and then uses mutual information to process three types of EEG signals (healthy EEG, interictal EEG and EEG signals during epileptic seizures), and extracts the digital features of each type of EEG signal. By comparison, the discrimination effect of the mutual information under the three entropy estimates in the processing of EEG signals is measured.