基于EEG去趋势波动分析和极限学习机的癫痫发作自动检测与分类识别

Automatic Detection and Classification of Epileptic EEG Based on Detrended Fluctuation Analysis and Extreme Learning Machine

癫痫是一种常发的中枢神经失调疾病.基于脑电(EEG)的癫痫发作自动检测与准确识别在临床诊断和治疗上具有重要意义.本文首先采用经验模态分解(EMD)将被试者脑电信号分解成多个固有模态函数(IMF),然后计算低尺度IMF的去趋势波动指数、均值和标准差并组成特征向量,再由极限学习机(ELM)进行自动分类.经使用波恩大学和波士顿儿童医院的脑电数据集(含健康志愿者与癫痫患者)检测验证,结果表明本文所提出的自动检测与快速识别方法仅需较少训练样本即可达到较高的癫痫发作准确识别率(≥95%),具有较好临床应用价值.

Epilepsy is one of the most common neurological diseases.Automatic detection and accurate identification of epileptic seizure based on electroencephalogram （ EEG） plays an important role in the dia gnosis and treatment of epileptic seizures.In this paper, EEG signals were decomposed into a number of intrinsic mode functions （ IMFs） by empirical mode decomposition （ EMD） , and then the detrended fluc-tuation index, mean and standard deviation （ SD） of IMFs of lower scales were calculated.The three pa-rameters were combined into a feature vector and fed into an extreme learning machine （ ELM） classifier. The proposed method was validated on the EEG data sets from Bonn University and Boston Children＇s Hospital, involving healthy subjects and epileptics.Results show that the proposed method of automatic detection and rapid identification requires fewer training samples while achieving a higher recognition rate （≥95%）,indicating that it is a promising tool for automatic detection and classification of epileptic sei-zures.