Abstract: Analysis of electroencephalography（EEG） is an important approach for epilepsy study. Scalp EEG data collected clinical is used as object of study. Permutation mutual information between different electrodes is computed. The results show mutual information of epileptic patients is obviously higher than that of healthy objects. So permutation mutual information can be used as an important feature for epilepsy diagnosis. Synchronism analysis is conducted on EEG based on permutation mutual information. The results show the exchange of information in left brain region, right region and left-right region of epilepsy patients is obviously more than that of healthy objects. That is the synchronism of epilepsy patients is obviously enhanced. Mutual information and synchronism analysis can also be used to analyze EEG before a seizure and that during a seizure to predict epileptic seizure.

Key words: epilepsy, electroencephalography（EEG）, recurrence plot, mutual information, synchronism, phase space reconstruction

摘要： 脑电（EEG）分析是研究癫痫的一个重要手段。以临床采集的健康对象和癫痫患者的头皮EEG为研究对象，计算不同导联EEG数据之间的排序互信息，结果表明癫痫患者不同导联之间的互信息明显高于健康对象，因此，排序互信息可以作为癫痫疾病诊断的重要特征。以排序互信息为依据，对癫痫脑电进行了同步性的分析，结果表明癫痫患者左脑区域内、右脑区域内及左右脑区之间的信息交流明显增强，即其同步性强于健康对象。互信息和同步性的分析方法还可对癫痫发作前期和发作阶段的EEG进行分析，对癫痫发作作出预测。

关键词: 癫痫, 脑电图, 递归图, 互信息, 同步性, 相空间重构