基于矩阵灰建模的单次P300检测新方法
|
摘要
针对少导联P300单次提取识别率较低的问题,提出了一种基于矩阵灰建模的参数模型法提取特征的方法,提高了P300单次识别率.首先对脑电信号进行预处理,然后选择导联组合,接着对每个Epoch进行建模,将模型参数作为特征向量输入SVM分类识别.结果表明,单次P300的平均识别率为91.43%,叠加平均3次正确率可高达97.87%.
Aiming at the drawback of lowidentification accuracy in single trial P300 feature extraction and classification,a parameter model method based on Matrix Grey Modeling to extract P300 feature was proposed to improve the recognition accuracy of the visual evoked potential P300 in single trial classification. Firstly,EEG signal was preprocessed,and then channel set selection was applied. After that,the model parameters of Matrix Grey Modelling for each epoch was connected as the feature vector and were input to the SVMclassifier. The experimental results showthat the average accuracy of single trial P300 across all the subjects is 91. 43%,and the accuracy can be up to 97. 87% if 3 times averaging is used.
Aiming at the drawback of lowidentification accuracy in single trial P300 feature extraction and classification,a parameter model method based on Matrix Grey Modeling to extract P300 feature was proposed to improve the recognition accuracy of the visual evoked potential P300 in single trial classification. Firstly,EEG signal was preprocessed,and then channel set selection was applied. After that,the model parameters of Matrix Grey Modelling for each epoch was connected as the feature vector and were input to the SVMclassifier. The experimental results showthat the average accuracy of single trial P300 across all the subjects is 91. 43%,and the accuracy can be up to 97. 87% if 3 times averaging is used.
引文
[1]HOCHBERG L R,SERRUYA MD,FRIEHS G M,et al.Neuronal ensemble control of prosthetic devices by a human with tetraplegia[J].Nature,2006,442(13):164-171.
[2]ABOOTALEBI V,MORADI MH,KHALILZADEH MA.A comparison of methods for ERP assessment in a P300-based GKT[J].International Journal of Psychophysiology,2006,62(2):309-320.
[3]闫蓓,刘莎,李建华,袁海文,丁锋锋.单通道视觉诱发脑电的单次提取方法研究[J].仪器仪表学报,2012,33(4):905-910.YAN Bei,LIU Sha,LI Jianhua,YUAN Haiwen,DING Fengfeng.Study on single-trial feature extraction method of single-channel visual evoked potential[J].Chinese Journal of Scientific Instrument,2012,33(4):905-910.(in Chinese)
[4]张宇,张建华,王行愚,金晶.基于Fast ICA的P300电位快速提取方法[J].华东理工大学学报(自然科学版),2009,35(5):750-755.ZHANG Yu,ZHANG Jian-hua,WANG Xing-yu,JIN Jing.A fast ICA-based approach to extracting P300 potential[J].Journal of East China University of Science and Technology(Natural Science Edition),2009,35(5):750-755.(in Chinese)
[5]万柏坤,朱欣,杨春梅,高扬.ICA去除EEG中眼动伪差和工频干扰方法研究[J].电子学报,2003,31(10):1571-1574.WAN Bai-kun,ZHU Xin,YANG Chun-mei,GAO Yang.Study on applying independent component analysis to remove blink artifacts and power noise in EEG[J].Acta Electronica Sinica,2003,31(10):1571-1574.(in Chinese)
[6]AVIYENTE S,BERNAT E M,MALONE S M,et al.Time-frequency data reduction for event related potentials:Combining principal component analysis and matching pursuit[J].EURASIP Journal on Advances in Signal Processing,2010,11(55):289-571.
[7]王攀,沈继忠,施锦河.基于小波变换和时域能量熵的P300特征提取算法[J].仪器仪表学报,2011,32(6):1284-1289.WANG Pan,SHEN Jizhong,SHI Jinhe.P300 feature extraction algorithm based on wavelet transform and temporal energy entropy[J].Chinese Journal of Scientific Instrument,2011,32(6):1284-1289.(in Chinese)
[8]LI K,SANKAR R,ARBEL Y,DONCHIN E.Single trial independent component analysis for P300 BCI system[A].Proceedings of the 31st Annual International Conference of the IEEE EMBS[C].USA:IEEE Press,2009.4035-4038.
[9]陈洪波,李蓓蕾,陈真诚.基于ICA的脑电信号P300少次自动提取[J].电子学报,2012,40(6):1257-1262.CHEN Hong-bo,LI Bei-lei,CHEN Zhen-cheng.Automatically extract P300 within several trials from EEG based on ICA[J].Acta Electronica Sinica,2012,40(6):1257-1262.(in Chinese)
[10]LI Kun,et al.A newsingle trial P300 classification method[J].International Journal of E-Health and Medical Communications,2012,3(4):31-41.
[11]XIE S,WU Y,ZHANG Y,et al.Single channel single trial P300 detection using extreme learning machine:Compared with BPNN and SVM[A].Proceedings of the International Joint Conference on Neural Networks(IJCNN)[C].China:IEEE Press,2014.544-549.
[12]杨立才,李金亮,姚玉翠.基于小波分解和支持向量机的P300识别算法[J].中国生物医学工程学报,2007,26(6):804-809.YANG Li-Cai,LI Jin-Liang,YAO Yu-Cui.P300 detection algorithm based on wavelet decomposition and support vector machines[J].Chinese Journal of Biomedical Engineering,2007,26(6):804-809.(in Chinese)
[13]刘歆.面向矩阵序列的灰建模研究[D].四川重庆:重庆大学,2012.
[14]CHEN Y W,LIN C J.Combining svms with various feature selection strategies[J].Studies in Fuzziness&Soft Computing,2005,207:315-324.
[15]丁世飞,齐丙娟,谭红艳.支持向量机理论与算法研究综述[J].电子科技大学学报,2011,40(1):2-10.DING Shi-fei,QI Bing-juan,TAN Hong-yan.An overviewon theory and algorithm of support vector machines[J].Journal of University of Electronic Science and Technology of China,2011,40(1):2-10.(in Chinese)
[2]ABOOTALEBI V,MORADI MH,KHALILZADEH MA.A comparison of methods for ERP assessment in a P300-based GKT[J].International Journal of Psychophysiology,2006,62(2):309-320.
[3]闫蓓,刘莎,李建华,袁海文,丁锋锋.单通道视觉诱发脑电的单次提取方法研究[J].仪器仪表学报,2012,33(4):905-910.YAN Bei,LIU Sha,LI Jianhua,YUAN Haiwen,DING Fengfeng.Study on single-trial feature extraction method of single-channel visual evoked potential[J].Chinese Journal of Scientific Instrument,2012,33(4):905-910.(in Chinese)
[4]张宇,张建华,王行愚,金晶.基于Fast ICA的P300电位快速提取方法[J].华东理工大学学报(自然科学版),2009,35(5):750-755.ZHANG Yu,ZHANG Jian-hua,WANG Xing-yu,JIN Jing.A fast ICA-based approach to extracting P300 potential[J].Journal of East China University of Science and Technology(Natural Science Edition),2009,35(5):750-755.(in Chinese)
[5]万柏坤,朱欣,杨春梅,高扬.ICA去除EEG中眼动伪差和工频干扰方法研究[J].电子学报,2003,31(10):1571-1574.WAN Bai-kun,ZHU Xin,YANG Chun-mei,GAO Yang.Study on applying independent component analysis to remove blink artifacts and power noise in EEG[J].Acta Electronica Sinica,2003,31(10):1571-1574.(in Chinese)
[6]AVIYENTE S,BERNAT E M,MALONE S M,et al.Time-frequency data reduction for event related potentials:Combining principal component analysis and matching pursuit[J].EURASIP Journal on Advances in Signal Processing,2010,11(55):289-571.
[7]王攀,沈继忠,施锦河.基于小波变换和时域能量熵的P300特征提取算法[J].仪器仪表学报,2011,32(6):1284-1289.WANG Pan,SHEN Jizhong,SHI Jinhe.P300 feature extraction algorithm based on wavelet transform and temporal energy entropy[J].Chinese Journal of Scientific Instrument,2011,32(6):1284-1289.(in Chinese)
[8]LI K,SANKAR R,ARBEL Y,DONCHIN E.Single trial independent component analysis for P300 BCI system[A].Proceedings of the 31st Annual International Conference of the IEEE EMBS[C].USA:IEEE Press,2009.4035-4038.
[9]陈洪波,李蓓蕾,陈真诚.基于ICA的脑电信号P300少次自动提取[J].电子学报,2012,40(6):1257-1262.CHEN Hong-bo,LI Bei-lei,CHEN Zhen-cheng.Automatically extract P300 within several trials from EEG based on ICA[J].Acta Electronica Sinica,2012,40(6):1257-1262.(in Chinese)
[10]LI Kun,et al.A newsingle trial P300 classification method[J].International Journal of E-Health and Medical Communications,2012,3(4):31-41.
[11]XIE S,WU Y,ZHANG Y,et al.Single channel single trial P300 detection using extreme learning machine:Compared with BPNN and SVM[A].Proceedings of the International Joint Conference on Neural Networks(IJCNN)[C].China:IEEE Press,2014.544-549.
[12]杨立才,李金亮,姚玉翠.基于小波分解和支持向量机的P300识别算法[J].中国生物医学工程学报,2007,26(6):804-809.YANG Li-Cai,LI Jin-Liang,YAO Yu-Cui.P300 detection algorithm based on wavelet decomposition and support vector machines[J].Chinese Journal of Biomedical Engineering,2007,26(6):804-809.(in Chinese)
[13]刘歆.面向矩阵序列的灰建模研究[D].四川重庆:重庆大学,2012.
[14]CHEN Y W,LIN C J.Combining svms with various feature selection strategies[J].Studies in Fuzziness&Soft Computing,2005,207:315-324.
[15]丁世飞,齐丙娟,谭红艳.支持向量机理论与算法研究综述[J].电子科技大学学报,2011,40(1):2-10.DING Shi-fei,QI Bing-juan,TAN Hong-yan.An overviewon theory and algorithm of support vector machines[J].Journal of University of Electronic Science and Technology of China,2011,40(1):2-10.(in Chinese)