基于近似熵的癫痫发作预测研究  

Prediction of Epileptic Seizure based on Approximate Entropy of EEG

作　　者：刘治远[1] 解玲丽[2] 陈子怡[3] 黄瑞梅[1] 李小江 周毅[1] 

机构地区：[1]中山大学中山医学院生物医学工程系,广州510080 [2]中山大学数学与计算科学学院,广州510275 [3]中山大学附属第一医院神经内科,广州510080

出　　处：《生物医学工程研究》2011年第1期20-23,共4页Journal Of Biomedical Engineering Research

基　　金：国家自然科学基金资助项目(10671213;10826068);广东省自然科学基金资助项目(8451008901000418)

摘　　要：临床采用诱发方法检测获得的失神发作患者EEG信号,研究其发作前脑电信号的动力学变化的规律,寻找预测癫痫失神发作一般规律和方法。我们选择合适的电极对,使用非线性动力学的方法,采用复杂度变化度量的近似熵指标,通过闪光刺激癫痫患者获得的EEG信号进行动力学特征研究,根据EEG信号表现出的同步情况实现对癫痫发作的预测。结果表明,本研究可以实现对临床光刺激进行诱发的癫痫失神患者的发作进行预测。研究中采用优化电极的方法优于采用固定电极对。In this study,we try to find out the general regularity and method to predict absences seizure by studying regularities of dynamics of spatiotemporal transitions of EEG during epileptic seizures.With nonlinear dynamical method,using Approximate Entropy(ApEn),we studied the dynamical features of EEG signal which was obtained from clinical data.The selection of critical cortical cites involved a model which could optimize the probability to be best.Before epileptic absence seizure,the T-index of those critical cites will change to some degree and will progressively converge.Then,a probably prediction can be given.The prediction scheme of optima electrode is better than using fixed electrodeduring two seizures is considerable.

关 键 词：癫痫 脑电图 T检验 近似熵 发作预测 

分 类 号：R318[医药卫生—生物医学工程]