一种基于脑电波突变分析的癫痫疾病检测方法  

An epilepsy detection method based on brain wave mutation analysis

作　　者：李娜[1] 齐金鹏[1] 邹俊晨 朱厚杰 刘佳伦 LI Na;QI Jin-peng;ZOU Jun-chen;ZHU Hou-jie;LIU Jia-lun(School of Information Science and Technology,Donghua University,Shanghai 201620,China)

机构地区：[1]东华大学信息科学与技术学院

出　　处：《信息技术》2019年第12期41-46,共6页Information Technology

基　　金：国家自然科学基金资助项目(61305081,61104154)

摘　　要：癫痫是一种复杂的难治愈性疾病,单纯靠医生肉眼分析脑电图诊断存在很大的误差。文中以脑电时序信号为基础,分别应用不同算法进行基于突变点分析的癫痫疾病检测研究。针对合成与实时脑电波序列设计不同组实验检测,实验结果表明,TSTKS(Ternary Search Trees and modi-fied Kolmogorov-Smirnov)统计方法在误差率、准确率和时间复杂度等分析指标来看都比其它方法更加适合癫痫疾病的检测,从而对癫痫疾病的分析和诊断具有辅助和指导的意义。Epilepsy is a complex and refractory disease.There are great errors in the diagnosis of EEG by naked eye analysis.In this paper,based on EEG time series signals,different algorithms are applied to detect epilepsy based on mutation point analysis.The experimental results show that TSTKS(Ternary Search Trees and modified Kolmogorov-Smirnov)statistical method is more suitable for the detection of epilepsy than other methods in terms of error rate,time complexity and accuracy,which is helpful and instructive for the analysis and diagnosis of epilepsy.

关 键 词：癫痫 脑电信号 时序信号 TSTKS 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]