北京生物医学工程

基于ＫＩＣＡ和ＭＡＭＡ?ＥＭＤ的眼电伪迹去除方法

An EOG artifact removal method based on KICA and MAMA-EMD

作者：张圆圆孙炎珺李明爱

单位：北京工业大学信息学部(北京 100124), <br />计算智能与智能系统北京市重点实验室(北京 100124）<br />,数字社区教育部工程研究中心(北京 100124), <br />通信作者：李明爱。E-mail: limingai@bjut.edu.cn

关键词：脑电信号; 眼电伪迹; 核独立分量分析; 模态分裂；掩膜最小弧长经验模态分解

分类号：R318.04  

出版年·卷·期（页码）：2022·41·4（360-367）

摘要：

目的为了准确地分离并去除脑电(EEG)信号中的尖峰状眼电(EOG)伪迹，本文提出一种基于核独立分量分析(kernel independent component analysis, KICA)和掩膜最小弧长经验模态分解(masking-aided minimum arclength empirical mode decomposition, MAMA-EMD)的眼电伪迹去除方法，即KICMME。方法首先，使用KICA将多通道受污染的EEG信号分离为多个独立分量(independent components, ICs)；然后，计算每个IC的峰度值，确定与EOG相关的IC，并利用MAMA-EMD算法对其进一步分解，得到一组固有模态函数(intrinsic mode function, IMF)；进而，通过计算各IMF的低频功率占比，识别并剔除与EOG相关性高的IMF；最后，基于MAMA-EMD和KICA的逆变换重构出“纯净”EEG信号。结果在半模拟和真实脑电两个数据集上进行实验研究，KICMME取得的均方误差和信噪比分别为0.82 〖μV〗^2和12.51 dB，获得的分类准确率和Kappa值分别为91%和0.83。结论 MAMA-EMD能够准确地分离出与EOG相关的IMF分量，使得KICMME可以在保留有用神经信息的同时，最大限度地去除EEG信号中的EOG伪迹，相对现有基于盲源分离的眼电伪迹去除方法具有明显优势。

Objective In order to isolate and remove the spike-like electrooculogram (EOG) artifacts from EEG signals accurately, this paper proposed a novel method based on kernel independent component analysis (KICA) and masking-aided minimum arclength empirical mode decomposition (MAMA-EMD), namely KICMME, for the removal of EOG artifacts. Methods Firstly, the multichannel contaminated EEG signals were separated by KICA into several independent components (ICs). Secondly, the kurtosis value of each IC was calculated to detect EOG-related IC, and MAMA-EMD algorithm was used to decompose it further to obtain a set of intrinsic mode functions (IMFs). Furthermore, the IMFs linked with EOG were identified and eliminated by calculating the low-frequency power proportion of each IMF. Finally, the ‘clean’ EEG was reconstructed by performing the inverse transform of MAMA-EMD and KICA. Results Based on the experimental research of semi-simulated and real EEG data, KICMME achieved a mean square error (MSE) of 0.82 〖μV〗^2and a signal-to-noise ratio (SNR) of 12.51 dB, and the classification accuracy and Kappa values were 91% and 0.83 respectively. Conclusions MAMA-EMD can accurately isolate the IMF component associated with EOG, and KICMME can remove EOG artifacts from EEG signals to the maximum extent while retaining useful neural information, which has great improvement compared with the existing EOG artifact removal methods based on blind source separation.

参考文献：

[1] Wolpaw JR, Birbaumer N, McFarland DJ, et al. Brain-computer interface for communication and control [J]. Clinical Neurophysiology, 2002, 113(6): 767-791.
[2] 张娜,孙炎珺,李明爱. 一种个性化动态脑功能网络的构建与特征提取方法[J]. 北京生物医学工程,2020, 39(6): 551-560.
Zhang N, Sun YJ, Li MA. A personalized dynamic brain functional network and feature extraction [J]. Beijing Biomedical Engineering, 2020, 39(6): 551-560.
[3] 李明爱, 刘帆. 脑电中眼电伪迹的自动识别与去除[J]. 北京生物医学工程, 2018, 37(6): 559-565.
Li MA, Liu F. The automatic identification and removal of ocular artifacts from EEG [J]. Beijing Biomedical Engineering, 2018, 37(6): 559-565.
[4] Islam K, Rastegarnia A, Yang Z. Methods for artifact detection and removal from scalp EEG: a review [J]. Neurophysiologie Clinique/Clinical Neurophysiology, 2016, 46(4-5): 287-305.
[5] De Clercq W, Vergult A, Vanrumste B, et al. Canonical correlation analysis applied to remove muscle artifacts from the electroencephalogram [J]. IEEE Transactions on Biomedical Engineering, 2006, 53(12): 2583-2587.
[6] Mognon A, Jovicich J, Bruzzone L, et al. ADJUST: an automatic EEG artifact detector based on the joint use of spatial and temporal features [J]. Psychophysiology, 2011, 48(2): 229-240.
[7] Bach F, Jordan M. Kernel independent component analysis [J]. Journal of Machine Learning Research, 2002, 3: 1-48.
[8] Li MA, Cui Y, Yang JF. Automatic removal of ocular artifact from EEG with DWT and ICA method [J]. Applied Mathematics and Information Sciences, 2013, 7(2): 809-816.
[9] Li MA, Guo SD, Zuo GY, et al. Removing ocular artifacts from mixed EEG signals with FastKICA and DWT [J]. Journal of Intelligent and Fuzzy Systems, 2015, 28(6): 2851-2861.
[10] Maddirala AK, Shaik RA. Removal of EOG artifacts from single channel EEG signals using combined singular spectrum analysis and adaptive noise canceler [J]. IEEE Sensors Journal, 2016, 16(23): 8279-8287.
[11] Gao C, Ma L, Li H. An ICA/HHT hybrid approach for automatic ocular artifact correction [J]. International Journal of Pattern Recognition and Artificial Intelligence, 2015, 29(2): 1558001.
[12] Yang BH, Zhang T, Zhang Y, et al. Removal of electrooculogram artifacts from electroencephalogram using canonical correlation analysis with ensemble empirical mode decomposition [J]. Cognitive Computation, 2017, 9(5): 626-633.
[13] Yang H, Jeng S, Young HV, et al. A minimum arclength method for removing spikes in empirical mode decomposition [J]. IEEE Access, 2019, 7: 13284-13294.
[14] Wang G, Teng C, Li K, et al. The removal of EOG artifacts from EEG signals using independent component analysis and multivariate empirical mode decomposition [J]. IEEE Journal of Biomedical and Health Informatics, 2016, 20(5): 1301-1308.
[15] Urigüen JA, Garcia-Zapirain B. EEG artifact removal-state-of-the-art and guidelines[J]. Journal of Neural Engineering, 2015, 12(3): 031001.
[16] Blankertz B, Dornhege G, Krauledat M, et al. The non-invasive Berlin brain-computer interface: fast acquisition of effective performance in untrained subjects[J]. Neuroimage, 2007, 37(2): 539-550.
[17] Mannan MMN, Kamran MA, Kang S, et al. Effect of EOG signal filtering on the removal of ocular artifacts and EEG-based brain-computer interface: a comprehensive study [J]. Complexity, 2018, 2018: 4853741.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录