基于弹性变分模态分解的癫痫脑电信号分类方法

景鹏; 张学军; 孙知信

doi:10.7498/aps.70.20200904

摘要

癫痫脑电信号分类对于癫痫诊治具有重要意义. 为了实现病灶性与非病灶性癫痫脑电信号的分类, 本文利用弹性网回归重构变分模态分解算法, 提出弹性变分模态分解算法并将其应用到所提癫痫脑电信号分类方法中. 该方法先将原信号分割成多个子信号, 并对各子信号进行弹性变分模态分解, 然后从分解后的不同变分模态函数中提取精细复合多尺度散布熵作为特征, 最后利用支持向量机进行分类. 针对癫痫脑电的公共数据集, 最终的实验结果表明, 准确率、灵敏度和特异度三个性能指标分别达到92.54%, 93.22%和91.86%.

关键词:

Abstract

Epilepsy is an extensive nervous system disease nowadays. Electroencephalogram (EEG) can capture the abnormal discharge of nerves in the brain duration of seizure and provide a non-invasive way to identify epileptogenic sites in the brain. In order to distinguish between focal epilepsy EEG signal and non-focal epilepsy EEG signal, in this paper we propose an automated epileptic EEG detection method based on the elastic variational mode decomposition (EVMD). The proposed EVMD algorithm is a method of analyzing the signals and also a processing method in time-frequency domain, in which the elastic net regression is used to reconstruct a constrained variational model in variational mode decomposition (VMD). Used in the VMD algorithm is the Tikhonov regularization that is also statistically called ridge regression as a solution of recovering the unknown signal and assessing the bandwidth of a mode, namely the variational equation constructed by VMD only has L2 norm. However, the ridge regression cannot select variables when the equation has multiple variables. Another regression method, called lasso regression, only has L1 norm and can select a more accurate model from multiple variables, but it has worse performance when variables have group effect or co-linearity. The elastic net regression has advantages of ridge regression and lasso regression, in other word, the variational equation constructed by EVMD has both L1 regularization item and L2 regularization item, so in this paper we propose the EVMD by elastic net regression. In addition, in this paper the EVMD is used to distinguish between focal epilepsy EEG signal and non-focal epilepsy EEG signal. Firstly, the original EEG signals are divided into several sub-signals where the test signals are divided into sub-signals with shorter durations by time series and a reasonable time overlap is kept between successive sub-signals. After that each sub-signal is decomposed into intrinsic mode functions by using the EVMD. Furthermore, the refined composite multiscale dispersion entropy (RCMDE) as feature is extracted from each intrinsic mode function where a Student’s t-test is used to assess the statistical differences between RCMDEs extracted from focal and non-focal EEG signals respectively. Finally, the support vector machine (SVM) is used to classify their features. For an epilepsy EEG signalspublic data set, the final experimental results show that the performance indices of accuracy, sensitivity, and specificity can reach 92.54%, 93.22% and 91.86% respectively.

Keywords:

elastic variational mode decomposition /
refined composite multiscale dispersion entropy /
epileptic electroencephalogram

作者及机构信息

1.
南京邮电大学电子与光学工程学院, 微电子学院, 南京　210023

2.
南京邮电大学, 射频集成与微组装技术国家地方联合工程实验室, 南京　210023

3.
南京邮电大学, 江苏省邮政大数据技术与应用工程研究中心, 南京　210003

4.
南京邮电大学, 国家邮政局邮政行业技术研发中心(物联网技术), 南京　210003)

通信作者: 张学军, xjzhang@njupt.edu.cn

基金项目: 国家自然科学基金(批准号: 61972208, 61672299)资助的课题

Authors and contacts

1.
College of Electronic and Optical Engineering & College of Microelectronics, Nanjing University of Posts and Telecommunications, Nanjing 210023, China

2.
National and Local Joint Engineering Laboratory of RF Integration and Micro-Assembly Technology, Nanjing University of Posts and Telecommunications, Nanjing 210023, China

3.
Post Big Data Technology and Application Engineering Research Center of Jiangsu Province, Nanjing University of Posts and Telecommunications, Nanjing 210003, China

4.
Post Industry Technology Research and Development Center of the State Posts Bureau (Internet of Things Technology), Nanjing University of Posts and Telecommunications, Nanjing 210003, China

Corresponding author: Zhang Xue-Jun, xjzhang@njupt.edu.cn

Funds: Project supported by the National Natural Science Foundation of China (Grant Nos. 61972208, 61672299)

文章全文

参考文献

[1]	World Health Organization http://www.who.int/news-room/fact-sheets/detail/epilepsy/ [2019-6-20]
[2]	Andrzejak R G, Schindler K, Rummel C 2012 Phys. Rev. E 86 046206 Google Scholar
[3]	张瑞, 宋江玲, 胡文凤 2016 西北大学学报(自然科学版) 46 781 Zhang R, Song J L, Hu W F 2016 J. Northwest Univ. (Nat. Sci.) 46 781
[4]	Alam S, Bhuiyan M 2013 IEEE J. Biomed. Health Inf. 17 312 Google Scholar
[5]	Das A, Bhuiyan M, Alam S 2014 Signal Image Video Process. 10 259
[6]	Rahman M, Bhuiyan M, Das A 2019 Biomed. Signal Process. Control 50 72 Google Scholar
[7]	Sharma R, Pachori R, Acharya U 2015 Entropy 17 669 Google Scholar
[8]	Abhijit B, Ram B P 2017 Entropy 19 99 Google Scholar
[9]	谢平, 杨芳梅, 李欣欣, 杨勇, 陈晓玲, 张利泰 2016 物理学报 65 118701 Google Scholar Xie P, Yang F M, Li X X, Yang Y, Chen X L, Zhang L T 2016 Acta Phys. Sin. 65 118701 Google Scholar
[10]	王莹, 侯凤贞, 戴加飞, 刘新峰, 李锦, 王俊 2014 物理学报 63 218701 Google Scholar Wang Y, Hou F Z, Dai J F, Liu X F, Li J, Wang J 2014 Acta Phys. Sin. 63 218701 Google Scholar
[11]	Azami H, Rostaghi M, Abasolo D, Escudero J 2017 IEEE Trans. Biomed. Eng. 64 2872 Google Scholar
[12]	KiymiK M, Guler I, Dizibuyuk A, Akin M 2005 Comput. Biol. Med. 35 603 Google Scholar
[13]	张涛, 陈万忠, 李明阳 2016 物理学报 65 038703 Google Scholar Zhang T, Chen W Z, Li M Y 2016 Acta Phys. Sin. 65 038703 Google Scholar
[14]	Dragomiretskiy K, Zosso P 2014 IEEE Trans. Signal Process. 62 531
[15]	张哲, 梁冯珍 2013 哈尔滨商业大学学报 (自然科学版) 29 592 Zhang Z, Liang F Z 2013 J. Harbin Univ. Com. (Nat. Sci.) 29 592
[16]	Zou H, Hastie T 2005 J. R. Stat. Soc. 67 301 Google Scholar
[17]	Andrzejak R G http://www.dtic.upf.edu/~ralph/ [2019-3-20]
[18]	Andrzejak R G, Schindler K, Rummel C 2012 Physical Review E 86 046206
[19]	Chatterjee S, Pratiher S, Bose R 2017 IET Sci. Meas. Technol. 11 1014 Google Scholar
[20]	Abhijit B, Manish S, Ram B P, Pradip S Rajendra A 2018 Neural Comput. Appl. 29 47 Google Scholar
[21]	Li Z P, Chen J L, Zi Y Y, Pan J 2017 Mech. Syst. Signal Proc. 85 512 Google Scholar
[22]	Wang X B, Yang Z X, Y an, X A 2018 IEEE-ASME Trans. Mech. 23 68 Google Scholar
[23]	Wang Z J, He G F, Du W H, Zhou J, Han X F, Wang J T, He H H, Guo X M, Wang J Y, Kou Y F 2019 IEEE Access 7 44871 Google Scholar

施引文献

图 1 弹性变分模态分解处理癫痫脑电信号的流程图

Fig. 1. Architecture of processing epileptic EEG by EVMD.

下载: 全尺寸图片幻灯片

图 2 不同K值下中心频率随迭代次数的变化曲线

Fig. 2. The curves of the center frequency with the number of iterations under different K values.

下载: 全尺寸图片幻灯片

图 3 从各VMF中提取的RCMDE熵值的均值(± 标准差)随尺度因子变化曲线

Fig. 3. The curve of mean value(± SD) of RCMDE computed from VMF.

下载: 全尺寸图片幻灯片

图 4 10次5折交叉验证实验结果折线图

Fig. 4. The line chart of the results by 5-fold cross validation for 10 times.

下载: 全尺寸图片幻灯片

表 1 从各VMF中提取的RCMDE特征p值

Table 1. The p values of RCMDE computed from VMF.

	VMF1	VMF2	VMF3	VMF4
p值	1.17 × 10^–4	2.38 × 10^–2	4.42 × 10^–2	7.50 × 10^–3

下载: 导出CSV

表 2 EVMD与VMD实验结果对比

Table 2. Comparison of experimental result between EVMD and VMD.

指标	准确度/%	灵敏度/%	特异度/%
EMVD	92.54	93.22	91.86
VMD	89.49	87.56	88.12

下载: 导出CSV

表 3 本文方法与其他方法对比

Table 3. Comparison between proposed method and previously published methods.

	提取方法	特征	分类器	灵敏度、特异度、准确度/%
文献[7]	EMD	平均香农熵、样本熵等	LSSVM	90, 84, 87
文献[8]	TQWT	模糊熵	LSSVM	83.86, 85.46, 84.67
文献[19]	MFDFA	多重分形	SVM	92.5, 92.7, 92.2
文献[20]	EWT	RPS, CTM	LSSVM	88, 92, 90
本文	EVMD	RCMDE	SVM	93.22, 91.86, 92.54

下载: 导出CSV

[1]	World Health Organization http://www.who.int/news-room/fact-sheets/detail/epilepsy/ [2019-6-20]
[2]	Andrzejak R G, Schindler K, Rummel C 2012 Phys. Rev. E 86 046206 Google Scholar
[3]	张瑞, 宋江玲, 胡文凤 2016 西北大学学报(自然科学版) 46 781 Zhang R, Song J L, Hu W F 2016 J. Northwest Univ. (Nat. Sci.) 46 781
[4]	Alam S, Bhuiyan M 2013 IEEE J. Biomed. Health Inf. 17 312 Google Scholar
[5]	Das A, Bhuiyan M, Alam S 2014 Signal Image Video Process. 10 259
[6]	Rahman M, Bhuiyan M, Das A 2019 Biomed. Signal Process. Control 50 72 Google Scholar
[7]	Sharma R, Pachori R, Acharya U 2015 Entropy 17 669 Google Scholar
[8]	Abhijit B, Ram B P 2017 Entropy 19 99 Google Scholar
[9]	谢平, 杨芳梅, 李欣欣, 杨勇, 陈晓玲, 张利泰 2016 物理学报 65 118701 Google Scholar Xie P, Yang F M, Li X X, Yang Y, Chen X L, Zhang L T 2016 Acta Phys. Sin. 65 118701 Google Scholar
[10]	王莹, 侯凤贞, 戴加飞, 刘新峰, 李锦, 王俊 2014 物理学报 63 218701 Google Scholar Wang Y, Hou F Z, Dai J F, Liu X F, Li J, Wang J 2014 Acta Phys. Sin. 63 218701 Google Scholar
[11]	Azami H, Rostaghi M, Abasolo D, Escudero J 2017 IEEE Trans. Biomed. Eng. 64 2872 Google Scholar
[12]	KiymiK M, Guler I, Dizibuyuk A, Akin M 2005 Comput. Biol. Med. 35 603 Google Scholar
[13]	张涛, 陈万忠, 李明阳 2016 物理学报 65 038703 Google Scholar Zhang T, Chen W Z, Li M Y 2016 Acta Phys. Sin. 65 038703 Google Scholar
[14]	Dragomiretskiy K, Zosso P 2014 IEEE Trans. Signal Process. 62 531
[15]	张哲, 梁冯珍 2013 哈尔滨商业大学学报 (自然科学版) 29 592 Zhang Z, Liang F Z 2013 J. Harbin Univ. Com. (Nat. Sci.) 29 592
[16]	Zou H, Hastie T 2005 J. R. Stat. Soc. 67 301 Google Scholar
[17]	Andrzejak R G http://www.dtic.upf.edu/~ralph/ [2019-3-20]
[18]	Andrzejak R G, Schindler K, Rummel C 2012 Physical Review E 86 046206
[19]	Chatterjee S, Pratiher S, Bose R 2017 IET Sci. Meas. Technol. 11 1014 Google Scholar
[20]	Abhijit B, Manish S, Ram B P, Pradip S Rajendra A 2018 Neural Comput. Appl. 29 47 Google Scholar
[21]	Li Z P, Chen J L, Zi Y Y, Pan J 2017 Mech. Syst. Signal Proc. 85 512 Google Scholar
[22]	Wang X B, Yang Z X, Y an, X A 2018 IEEE-ASME Trans. Mech. 23 68 Google Scholar
[23]	Wang Z J, He G F, Du W H, Zhou J, Han X F, Wang J T, He H H, Guo X M, Wang J Y, Kou Y F 2019 IEEE Access 7 44871 Google Scholar

[1]	汪书潮, 苏秀琴, 朱文华, 陈松懋, 张振扬, 徐伟豪, 王定杰. 基于弹性变分模态提取的时间相关单光子计数信号去噪. 物理学报, 2021, 70(17): 174304. doi: 10.7498/aps.70.20210149
[2]	许子非, 岳敏楠, 李春. 优化递归变分模态分解及其在非线性信号处理中的应用. 物理学报, 2019, 68(23): 238401. doi: 10.7498/aps.68.20191005
[3]	刘备, 胡伟鹏, 邹孝, 丁亚军, 钱盛友. 基于变分模态分解与多尺度排列熵的生物组织变性识别. 物理学报, 2019, 68(2): 028702. doi: 10.7498/aps.68.20181772
[4]	杜义浩, 齐文靖, 邹策, 张晋铭, 谢博多, 谢平. 基于变分模态分解-相干分析的肌间耦合特性. 物理学报, 2017, 66(6): 068701. doi: 10.7498/aps.66.068701
[5]	郭家梁, 钟宁, 马小萌, 张明辉, 周海燕. 基于振幅-周期二维特征的脑电样本熵分析. 物理学报, 2016, 65(19): 190501. doi: 10.7498/aps.65.190501
[6]	雷敏, 孟光, 张文明, Nilanjan Sarkar. 基于虚拟开车环境的自闭症儿童脑电样本熵. 物理学报, 2016, 65(10): 108701. doi: 10.7498/aps.65.108701
[7]	谢平, 杨芳梅, 李欣欣, 杨勇, 陈晓玲, 张利泰. 基于变分模态分解-传递熵的脑肌电信号耦合分析. 物理学报, 2016, 65(11): 118701. doi: 10.7498/aps.65.118701
[8]	张涛, 陈万忠, 李明阳. 基于AdaBoost算法的癫痫脑电信号识别. 物理学报, 2015, 64(12): 128701. doi: 10.7498/aps.64.128701
[9]	谢平, 杨芳梅, 陈晓玲, 杜义浩, 吴晓光. 基于多尺度传递熵的脑肌电信号耦合分析. 物理学报, 2015, 64(24): 248702. doi: 10.7498/aps.64.248702
[10]	唐洁. 基于集合经验模态分解的类星体光变周期及其混沌特性分析. 物理学报, 2014, 63(4): 049701. doi: 10.7498/aps.63.049701
[11]	董泽芹, 侯凤贞, 戴加飞, 刘新峰, 李锦, 王俊. 基于Kendall改进的同步算法癫痫脑网络分析. 物理学报, 2014, 63(20): 208705. doi: 10.7498/aps.63.208705
[12]	黄晓林, 霍铖宇, 司峻峰, 刘红星. 等概率符号化样本熵应用于脑电分析. 物理学报, 2014, 63(10): 100503. doi: 10.7498/aps.63.100503
[13]	姚文坡, 刘铁兵, 戴加飞, 王俊. 脑电信号的多尺度排列熵分析. 物理学报, 2014, 63(7): 078704. doi: 10.7498/aps.63.078704
[14]	王莹, 侯凤贞, 戴加飞, 刘新峰, 李锦, 王俊. 改进的相对转移熵的癫痫脑电分析. 物理学报, 2014, 63(21): 218701. doi: 10.7498/aps.63.218701
[15]	孟庆芳, 陈珊珊, 陈月辉, 冯志全. 基于递归量化分析与支持向量机的癫痫脑电自动检测方法. 物理学报, 2014, 63(5): 050506. doi: 10.7498/aps.63.050506
[16]	薛春芳, 侯威, 赵俊虎, 王式功. 集合经验模态分解在区域降水变化多尺度分析及气候变化响应研究中的应用. 物理学报, 2013, 62(10): 109203. doi: 10.7498/aps.62.109203
[17]	孟庆芳, 周卫东, 陈月辉, 彭玉华. 基于非线性预测效果的癫痫脑电信号的特征提取方法. 物理学报, 2010, 59(1): 123-130. doi: 10.7498/aps.59.123
[18]	何亮, 杜磊, 庄奕琪, 李伟华, 陈建平. 金属互连电迁移噪声的多尺度熵复杂度分析. 物理学报, 2008, 57(10): 6545-6550. doi: 10.7498/aps.57.6545
[19]	谢勇, 徐健学, 康艳梅, 杨红军, 胡三觉. 皮层脑电的非线性降噪. 物理学报, 2003, 52(5): 1121-1126. doi: 10.7498/aps.52.1121
[20]	周世平, 徐克西. 负介电媒质观点下超导边值问题的变分解. 物理学报, 1996, 45(9): 1551-1561. doi: 10.7498/aps.45.1551

计量

文章访问数: 8728
PDF下载量: 132
被引次数: 0

姓名
邮箱
手机号码
标题
留言内容
验证码

搜索

留言板