基于EEG的癫痫发作预测当前研究与进展
|
摘要
癫痫(Epilepsy)是神经系统常见疾病,具有反复发作和难以治愈的特征,据世界卫生组织统计,其患病率为0.4%～1.4%。在过去的30年中,国内外在基于脑电图(Electroencephalogram,EEG)的癫痫发作诊断方面取得了巨大的进展。近些年来随着机器学习等计算机技术的不断发展,癫痫研究的前沿已经转向了癫痫预测方面。虽然抗癫痫药物可以解决大部分癫痫患者的发作问题,但是将近三分之一的癫痫患者患有医学难治性癫痫,药物可以减少癫痫发作的频率但不能控制其发作。如果能够通过机器学习方法在其临床表现之前预测癫痫的发作,那么就可以通过包括自我给药,植入药物或者电刺激植入装置等顿挫疗法给予治疗。介绍了国内外期刊关于癫痫发作预测的研究进展,并对其中常使用的机器学习方法进行汇总,分析了现有研究存在的问题,并对未来研究方向提出了一种新思路。
Epilepsy is a common disease of nervous system, which has the characteristics of recurrent seizures and difficult to cure.According to the statistics of the World Health Organization, the prevalence of epilepsy is 0.4% ～ 1.4%. In the past 30 years, great progress has been made in the diagnosis of epileptic seizure based on EEG at home and abroad. In recent years, with the continuous development of computer technology such as machine learning, the frontier of epilepsy research has turned to epilepsy prediction.Although antiepileptic drugs can solve the seizure problem of most epileptic patients, nearly one third of epileptic patients suffer from medically refractory epilepsy. Drugs can reduce the frequency of seizures but can not control the onset of epilepsy. If epilepsy can be predicted by machine learning before its clinical manifestations, it can be treated by contusion therapy including self-administration,implantation of drugs or electrical stimulation implantation devices. This paper introduces the research progress of epileptic seizure prediction in domestic and foreign journals, and summarizes the machine learning methods commonly used. Finally, the paper analyses the existing problems and puts forward a new idea for future research directions.
Epilepsy is a common disease of nervous system, which has the characteristics of recurrent seizures and difficult to cure.According to the statistics of the World Health Organization, the prevalence of epilepsy is 0.4% ～ 1.4%. In the past 30 years, great progress has been made in the diagnosis of epileptic seizure based on EEG at home and abroad. In recent years, with the continuous development of computer technology such as machine learning, the frontier of epilepsy research has turned to epilepsy prediction.Although antiepileptic drugs can solve the seizure problem of most epileptic patients, nearly one third of epileptic patients suffer from medically refractory epilepsy. Drugs can reduce the frequency of seizures but can not control the onset of epilepsy. If epilepsy can be predicted by machine learning before its clinical manifestations, it can be treated by contusion therapy including self-administration,implantation of drugs or electrical stimulation implantation devices. This paper introduces the research progress of epileptic seizure prediction in domestic and foreign journals, and summarizes the machine learning methods commonly used. Finally, the paper analyses the existing problems and puts forward a new idea for future research directions.
引文
[1]刘珑,李胜,王轶卿.基于小波包变换的脑电波信号降噪及特征提取[J].计算机工程与科学,2015,37(4):790-795.
[2]Heydari E,Shahbakhti M.Adaptive wavelet technique for EEG de-noising[C]//20158th Biomedical Engineering Internationa Conference(BMEi CON).Pattaya,2015:1-4.
[3]陈宏铭,王远大,程玉华.基于结合小波变换与Fast ICA算法的脑电信号降噪(英文)[J].生物医学工程学进展,2014,35(3):138-145.
[4]Acharya UR,Yanti R,Zheng JW,et al.Automated diagnosis of epilepsy using CWT,HOS and texture parameters[J].Int JNeural Syst,2013,23(3):1350009.
[5]杨昌健,邓赵红,蒋亦樟,等.引入迁移学习的癫痫EEG信号自适应识别[J].计算机工程,2015,41(6):158-164.
[6]Wang HD,Shi WW,Choy CS.Integrating channel selection and feature selection in a real time epileptic seizure detection system[C]//Conf Proc IEEE Eng Med Bio Soc,2017:3206-3211.
[7]Hosseini MP,Pompili D,Elisevich K,et al.Random ensemble learning for EEGclassification[J].Artif Intell Med,2018,84:146-158.
[8]Buriro AB,Shoorangiz R,Weddell SJ,et al.Predicting Microsleep States Using EEG InterChannel Relationships[J].IEEE Trans Neural Syst Rehabil Eng,2018,26(12):2260-2269.
[9]Zandi AS,Tafreshi R,Javidan M,et al.Predicting epileptic seizures in scalp EEG based on a variational Bayesian Gaussian mixture model of zero-crossing intervals[J].IEEE Trans Biomed Eng,2013,60:1401-1413.
[10]Bandarabadi M,Teixeira CA,Rasekhi J,et al.Epileptic seizure prediction using relative spectral power features[J].Clin Neurophysiol,2015,126:237-248.
[11]刘伟楠,刘燕,佟宝同,等.基于功率谱的睡眠中癫痫发作预测[J].生物医学工程学杂志,2018,35(3):329-336.
[12]周梦妮,崔会芳,曹锐,等.基于排列熵和支持向量机的癫痫发作预测研究[J/OL].计算机应用研究,2019(6):1-3.
[13]韩凌,王宏.基于空频域特征分析方法的癫痫发作预测[J].仪器仪表学报,2014,35(11):2501-2507.
[14]李志萍.基于支持向量机的多通道癫痫发作预测[J].计算机工程,2014,40(2):199-202,207.
[15]李淑芳,周卫东,袁琦,等.基于脑电棘波频次的癫痫发作预测算法[J].中国生物医学工程学报,2011,30(6):829-833.
[16]朱天桥,黄力宇.单导癫痫脑电模糊特征提取的支持向量机发作预测[J].仪器仪表学报,2010,31(11):2434-2439.
[17]刘银霞.基于脑电棘波频次和AR模型的癫痫发作预测算法[D].济南:山东大学,2013.
[18]Fei K,Wang W,Yang Q,et al.Chaos feature study in fractional Fourier domain for preictal prediction of epileptic seizure[J].Neurocomputing,2017,249:290-298.
[19]Yuan S,Zhou W,Chen L.Epileptic seizure prediction using diffusion distance and Bayesian Linear Discriminate Analysis on Intracrania EEG[J].Int J Neural Syst,2018,28(1):1750043.
[20]Parvez MZ,Paul M.Epileptic seizure prediction by exploiting spatiotempora relationship of EEG signals using phase correlation[J].IEEE Trans Neural Syst Rehabi Eng 2016,24:158-168.
[21]Qin Y,Han C,Che Y,et al.Efficient epileptic seizure detection based on electroencephalography signal[C]//201736th Chinese Control Conference(CCC).Dalian,2017:5324-5327.
[22]Zhang Y,Zhou W,Yuan Q,et al.Alow computation cost method for seizure prediction[J].Epilepsy Res,2014,108:1357-1366.
[23]丁木涵.基于卷积神经网络的癫痫发作预测[D].济南:山东师范大学,2018.
[24]单绍杰,李汉军,王璐璐,等.基于LSTM模型的单导联脑电癫痫发作预测[J].计算机应用研究,2018,35(11):3251-3254.
[2]Heydari E,Shahbakhti M.Adaptive wavelet technique for EEG de-noising[C]//20158th Biomedical Engineering Internationa Conference(BMEi CON).Pattaya,2015:1-4.
[3]陈宏铭,王远大,程玉华.基于结合小波变换与Fast ICA算法的脑电信号降噪(英文)[J].生物医学工程学进展,2014,35(3):138-145.
[4]Acharya UR,Yanti R,Zheng JW,et al.Automated diagnosis of epilepsy using CWT,HOS and texture parameters[J].Int JNeural Syst,2013,23(3):1350009.
[5]杨昌健,邓赵红,蒋亦樟,等.引入迁移学习的癫痫EEG信号自适应识别[J].计算机工程,2015,41(6):158-164.
[6]Wang HD,Shi WW,Choy CS.Integrating channel selection and feature selection in a real time epileptic seizure detection system[C]//Conf Proc IEEE Eng Med Bio Soc,2017:3206-3211.
[7]Hosseini MP,Pompili D,Elisevich K,et al.Random ensemble learning for EEGclassification[J].Artif Intell Med,2018,84:146-158.
[8]Buriro AB,Shoorangiz R,Weddell SJ,et al.Predicting Microsleep States Using EEG InterChannel Relationships[J].IEEE Trans Neural Syst Rehabil Eng,2018,26(12):2260-2269.
[9]Zandi AS,Tafreshi R,Javidan M,et al.Predicting epileptic seizures in scalp EEG based on a variational Bayesian Gaussian mixture model of zero-crossing intervals[J].IEEE Trans Biomed Eng,2013,60:1401-1413.
[10]Bandarabadi M,Teixeira CA,Rasekhi J,et al.Epileptic seizure prediction using relative spectral power features[J].Clin Neurophysiol,2015,126:237-248.
[11]刘伟楠,刘燕,佟宝同,等.基于功率谱的睡眠中癫痫发作预测[J].生物医学工程学杂志,2018,35(3):329-336.
[12]周梦妮,崔会芳,曹锐,等.基于排列熵和支持向量机的癫痫发作预测研究[J/OL].计算机应用研究,2019(6):1-3.
[13]韩凌,王宏.基于空频域特征分析方法的癫痫发作预测[J].仪器仪表学报,2014,35(11):2501-2507.
[14]李志萍.基于支持向量机的多通道癫痫发作预测[J].计算机工程,2014,40(2):199-202,207.
[15]李淑芳,周卫东,袁琦,等.基于脑电棘波频次的癫痫发作预测算法[J].中国生物医学工程学报,2011,30(6):829-833.
[16]朱天桥,黄力宇.单导癫痫脑电模糊特征提取的支持向量机发作预测[J].仪器仪表学报,2010,31(11):2434-2439.
[17]刘银霞.基于脑电棘波频次和AR模型的癫痫发作预测算法[D].济南:山东大学,2013.
[18]Fei K,Wang W,Yang Q,et al.Chaos feature study in fractional Fourier domain for preictal prediction of epileptic seizure[J].Neurocomputing,2017,249:290-298.
[19]Yuan S,Zhou W,Chen L.Epileptic seizure prediction using diffusion distance and Bayesian Linear Discriminate Analysis on Intracrania EEG[J].Int J Neural Syst,2018,28(1):1750043.
[20]Parvez MZ,Paul M.Epileptic seizure prediction by exploiting spatiotempora relationship of EEG signals using phase correlation[J].IEEE Trans Neural Syst Rehabi Eng 2016,24:158-168.
[21]Qin Y,Han C,Che Y,et al.Efficient epileptic seizure detection based on electroencephalography signal[C]//201736th Chinese Control Conference(CCC).Dalian,2017:5324-5327.
[22]Zhang Y,Zhou W,Yuan Q,et al.Alow computation cost method for seizure prediction[J].Epilepsy Res,2014,108:1357-1366.
[23]丁木涵.基于卷积神经网络的癫痫发作预测[D].济南:山东师范大学,2018.
[24]单绍杰,李汉军,王璐璐,等.基于LSTM模型的单导联脑电癫痫发作预测[J].计算机应用研究,2018,35(11):3251-3254.