脑电信号预处理方法研究综述

doi:10.16180/j.cnki.issn1007-7820.2023.04.005

摘要/Abstract

摘要：

脑电信号是一种复杂且重要的生物信号,被广泛应用于类脑智能技术和脑机接口领域的研究。文中介绍了干扰正常脑电信号的常见非生理性伪迹和生理性伪迹的类型及特点,并对生理性伪迹的产生原因进行了详细分析。通过对各种脑电信号去除伪迹方法的回顾以及应用现状的分析,比较并总结了传统去除伪迹方法和新型去除伪迹方法的研究进展,并进一步分析去除伪迹方法的优缺点。部分方法已经成功应用于处理脑电信号中的眼电、心电和肌电等伪迹中。文中还针对目前脑电信号去除伪迹的需求及所面临的问题给出了应对策略,并对未来的研究方向进行了分析和展望。

关键词: 脑电信号, 预处理, 去伪迹, 盲源分离, 小波变换, EMD, 人工神经网络, 深度学习

Abstract:

EEG signal is a complex and important biological signal, which is widely used in the research of brain-like intelligence technology and brain-computer interface. In this study, the types and characteristics of common non-physiological artifacts and physiological artifacts that interfere with normal EEG signals are introduced, and the causes of physiological artifacts are analyzed in detail. Through the review of various EEG artifact removal methods and the analysis of the application status, the research progress of traditional artifact removal methods and new artifact removal methods is compared and summarized, and the advantages and disadvantages of artifact removal methods are further analyzed. Some methods have been successfully applied to the processing of electrocardiogram, ECG and EMG artifacts in EEG signals.The current demand for artifact removal from EEG signals and the problems faced are also given in the present study, and the future research directions are analyzed and prospected.

Key words: EEG signal, preprocessing, artifacts removal, BSS, wavelet transform, EMD, artificial neural network, deep learning

中图分类号:

TP391

骆睿鹏,冯铭科,黄鑫,邹任玲,李丹. 脑电信号预处理方法研究综述[J]. 电子科技, 2023, 36(4): 36-43.

LUO Ruipeng,FENG Mingke,HUANG Xin,ZOU Renling,LI Dan. A Review of Research on EEG Signal Preprocessing Methods[J]. Electronic Science and Technology, 2023, 36(4): 36-43.

图/表 5

表1

图1

图2

图3

表2

参考文献 50

[1]	王韬, 柯余峰, 王宁慈, 等. 空间滤波方法在脑-机接口中的应用及研究进展[J]. 中国生物医学工程学报, 2019, 38(5):599-608.
	Wang Tao, Ke Yufeng, Wang Ningci, et al. Application and research development of spatial filtering method in braincomputer interfaces[J]. Chinese Journal of Biomedical Engineering, 2019, 38(5):599-608.
[2]	Hartmann M M, Schindler K, Gebbink T A, et al. PureE-EG:Automatic EEG artifact removal for epilepsy monitoring[J]. Neurophysiologie Clinique/Clinical Neurophysiology, 2014, 44(5):479-490.
[3]	郑展鹏, 尹钟. 基于集成SDAE和EEG的跨被试认知工作负荷识别[J]. 电子科技, 2021, 34(3):48-52.
	Zheng Zhanpeng, Yin Zhong. Inter-subject recognition of cognitive workload based on ensemble SDAE and EEG[J]. Electronic Science and Technology, 2021, 34(3):48-52.
[4]	Jiang X, Bian G B, Tian Z A. Removal of artifacts from EEG signals:A review[J]. Sensors, 2019, 19(5):987-1004. doi: 10.3390/s19050987
[5]	Croft R J, Barry R J. Removal of ocular artifact from the EEG:A review[J]. Neurophysiologie Clinique/Clinical Neurophysiology, 2000, 30(1):5-19.
[6]	Gratton G. Dealing with artifacts:The EOG contamination of the event-related brain potential[J]. Behavior Research Methods,Instruments & Computers, 1998, 30(1):44-53. doi: 10.3758/BF03209415
[7]	Chen X, Liu A, Chiang J, et al. Removing muscle artifacts from EEG data:Multichannel or single-channel techniques?[J]. IEEE Sensors Journal, 2016, 16(7):1986-1997. doi: 10.1109/JSEN.2015.2506982
[8]	Singh A, Hussain A A, Lal S, et al. A comprehensive re-view on critical issues and possible solutions of motor imagery based electroencephalography brain-computer interface[J]. Sensors, 2021, 21(6):2173-2207. doi: 10.3390/s21062173
[9]	Widmann A, Schröger E, Maess B. Digital filter design for electrophysiological data-a practical approach[J]. Journal of Neuroscience Methods, 2015, 250(7):34-46. doi: 10.1016/j.jneumeth.2014.08.002
[10]	Motamedi-Fakhr S, Moshrefi-Torbati M, Hill M, et al. Si-gnal processing techniques applied to human sleep EEG signals:A review[J]. Biomedical Signal Processing and Control, 2014, 10(1):21-33. doi: 10.1016/j.bspc.2013.12.003
[11]	Jung T P, Makeig S, Humphries C, et al. Removing electroencephalographic artifacts by blind source separate-on[J]. Psychophysiology, 2000, 37(2):163-178. pmid: 10731767
[12]	Sejnowski T J. Independent component analysis of electroencephalographic data[J]. Advances in Neural Information processing Systems, 1996, 8(8):1548-1551.
[13]	Ablin P, Cardoso J F, Gramfort A. Spectral independent component analysis with noise modeling for M/EEG source separation[J]. Journal of Neuroscience Methods, 2021, 356(1):9-21.
[14]	Albera L, Kachenoura A, Comon P, et al. ICA-based EEG denoising: A comparative analysis of fifteen methods[J]. Bulletin of the Polish Academy of Sciences,Technical Sciences, 2012, 60(3):407-418. doi: 10.2478/v10175-012-0052-3
[15]	Bell A J, Sejnowski T J. An information-maximization approach to blind separation and blind deconvolution[J]. Neural Computation, 1995, 7(6):1129-1159. doi: 10.1162/neco.1995.7.6.1129 pmid: 7584893
[16]	Hyvärinen A, Oja E. A fast fixed-point algorithm for i-ndependent component analysis[J]. Neural Computation, 1997, 9(7):1483-1492. doi: 10.1162/neco.1997.9.7.1483
[17]	王建雄, 张立民, 钟兆根. 基于FastICA算法的盲源分离[J]. 计算机技术与发展, 2011, 21(12):93-96.
	Wang Jianxiong, Zhang Limin, Zhong Zhaogen. Blind s-ource separation based on FastICA algorithm[J]. Computer Technology and Development, 2011, 21(12):93-96.
[18]	Pontifex M B, Miskovic V, Laszlo S. Evaluating the efficacy of fully automated approaches for the selection of eyeblink ICA components[J]. Psychophysiology, 2017, 54(5):780-791. doi: 10.1111/psyp.12827 pmid: 28191627
[19]	Hazarika N, Chen J Z, Tsoi A C, et al. Wavelet transform[J]. Signal Processing, 1997, 59(1):61-72. doi: 10.1016/S0165-1684(97)00038-8
[20]	Sharma R K. EEG signal denoising based on wavelet transform[C]. Coimbatore: Proceedings of the International Conference of Electronics,Communication and Aerospace Technology, 2017.
[21]	胡春海, 信思旭, 刘斌, 等. 基于小波变换和盲源分离的P300识别算法研究[J]. 计量学报, 2017, 38(2):242-246.
	Hu Chunhai, Xin Sixu, Liu Bin, et al. Study on recognition algorithm of P300 based on wavelet transform and blind source separation[J]. Acta Metrologica Sinica, 2017, 38(2):242-246.
[22]	Donoho D L. De-noising by soft-thresholding[J]. IEEE Transactions on Information Theory, 1995, 41(3):613-627. doi: 10.1109/18.382009
[23]	于向洋, 罗志增. 基于小波系数非线性连续函数衰减的脑电信号去噪[J]. 计量学报, 2017, 38(6):754-757.
	Yu Xiangyang, Luo Zhizeng. EEG signal denoising based on a wavelet nonlinear continuous function[J]. Acta Metrologica Sinica, 2017, 38(6):754-757.
[24]	孙铭阳, 谢子殿, 韩龙, 等. 自适应阈值函数小波算法的电机振动信号去噪[J]. 电子科技, 2020, 33(1):63-67.
	Sun Mingyang, Xie Zidian, Han Long, et al. Motor vibration signal denoising of adaptive threshold function wavelet algorithm[J]. Electronic Science and Technology, 2020, 33(1):63-67.
[25]	Huang N E, Shen Z, Long S R, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceedings Mathematical Physical & Engineering Sciences, 1998, 454(1971):903-995.
[26]	Gaur P, Pachori R B, Hui W, et al. An empirical mode decomposition based filtering method for classification of motor-imagery EEG signals for enhancing braincomputer interface[C]. Killarney: Proceedings of the International Joint Conference on Neural Networks, 2015.
[27]	郝欢, 王华力, 魏勤. 经验模态分解理论及其应用[J]. 高技术通讯, 2016, 26(1):67-80.
	Hao Huan, Wang Huali, Wei Qin. Theory of empirical mode decomposition and its application[J]. Chinese High Technology Letters, 2016, 26(1):67-80.
[28]	Wu Z H, Huang N E. Ensemble empirical mode decomposition:A noiseassisted data analysis method[J]. Advances in Adaptive Data Analysis, 2009, 1(1):1-41. doi: 10.1142/S1793536909000047
[29]	Yeh J R, Shieh J S, Huang N E. Complementary ensemble empirical mode decomposition:A novel noise enhanced data analysis method[J]. Advances in Adaptive Data Analysis, 2010, 2(2):135-156. doi: 10.1142/S1793536910000422
[30]	郑近德, 程军圣, 杨宇. 改进的EEMD算法及其应用研究[J]. 振动与冲击, 2013, 32(21):21-26.
	Zheng Jinde, Cheng Junsheng, Yang Yu. Modified EEMD algorithm and its applications[J]. Journal of Vibratiion and Shock, 2013, 32(21):21-26.
[31]	Mahajan R, Morshed B I. Unsupervised eye blink artifact denoising of EEG data with modified multiscale sample entropy,kurtosis and wavelet-ICA[J]. IEEE Journal of Biomedical & Health Informatics, 2014, 19(1):158-165.
[32]	Mammone N, Morabito F C. Enhanced automatic wavelet independent component analysis for electroencephalographic artifact removal[J]. Entropy, 2014, 16(12):6553-6572. doi: 10.3390/e16126553
[33]	Gao J F, Lin P, Yang Y, et al. Real-time removal of ocu-lar artifacts from EEG based on independent component analysis and manifold learning[J]. Neural Computing and Applications, 2010, 19(8):1217-1226. doi: 10.1007/s00521-010-0370-z
[34]	Chen X, Liu A, Peng H, et al. A preliminary study of muscular artifact cancellation in single-channel EEG[J]. Sensors, 2014, 14(10):18370-18389. doi: 10.3390/s141018370 pmid: 25275348
[35]	Bono V, Jamal W, Das S, et al. Artifact reduction in mu-ltichannel pervasive EEG using hybrid WPT-ICA and WPT-EMD signal decomposition techniques[C]. Florence: Proceedings of the IEEE International Conference on Acoustics,Speech and Signal Processing, 2014.
[36]	闫超, 孙占全, 田恩刚, 等. 基于深度学习的医学图像分割技术研究进展[J]. 电子科技, 2021, 34(2):7-11.
	Yan Chao, Sun Zhanquan, Tian Engang, et al. Research progress of medical image segmentation based on deep learning[J]. Electronic Science and Technology, 2021, 34(2):7-11.
[37]	孟昕. 基于深度学习的法律文书识别方法研究[J]. 电子科技, 2019, 32(12):84-86.
	Meng Xin. Research on recognition method of legal documents based on deep learning[J]. Electronic Science and Technology, 2019, 32(12):84-86.
[38]	Corley I A, Huang Y. Deep EEG super-resolution:Upsa-mpling EEG spatial resolution with generative adversarial networks[C]. Las Vegas: Proceedings of the IEEE EMBS International Conference on Biomedical & Health Informatics, 2018.
[39]	Fahimi F, Dosen S, Ang K K, et al. Generative adversarial networks-based data augmentation for brain-computer interface[J]. IEEE Transactions on Neural Networks and Learning Systems, 2020, 32(9):4039-4051. doi: 10.1109/TNNLS.2020.3016666
[40]	Hanrahan C. Noise reduction in EEG signals using convolutional autoencoding techniques[D]. Dublin: Technological University Dublin, 2019.
[41]	Zhang H, Zhao M, Wei C, et al. EEG denoiseNet:A benchmark dataset for deep learning solutions of EEG denoising[J]. Journal of Neural Engineering, 2021, 18(5):1-16.
[42]	Goodfellow I, Pouget-Abadie J, Mirza M, et al. Generative adversarial nets[C]. Montreal: Proceedings of the Twenty-seventh Conference on Neural Information Processing Systems, 2014.
[43]	Johnson J, Gupta A, Fei-Fei L. Image generation from scene graphs[C]. Salt Lake City: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018.
[44]	Yi Z, Zhang H, Tan P, et al. DualGAN:Unsupervised dual learning for image-to-image translation[C]. Venice: I-EEE International Conference on Computer Vision, 2017.
[45]	Lin G, Wu Q, Liang C, et al. Deep unsupervised learning for image super-resolution with generative adversarial network[J]. Signal Processing Image Communication, 2018, 68:88-100. doi: 10.1016/j.image.2018.07.003
[46]	姚粤汉. 脑电信号的伪迹去除与情绪识别研究[D]. 广州: 华南理工大学, 2020.
	Yao Yuehan. Study on EEG artifacts removal and emotion recognition[D]. Guangzhou: South China University of Technology, 2020.
[47]	Yang B, Duan K, Fan C, et al. Automatic ocular artifacts adversarial in EEG using deep learning[J]. Biomedical Signal Processing and Control, 2018, 43:148-158. doi: 10.1016/j.bspc.2018.02.021
[48]	Mclntosh J R, Yao J A, Hong L B, et al. Ballistocardiogram artifact reduction in simultaneous EEG-fMRI using deep learning[J]. IEEE Transactions on Biomedical Engineering, 2021, 68(1):78-89. doi: 10.1109/TBME.10
[49]	Zhang H M, Wei C, Zhao M Q, et al. A novel convolutional neural network model to remove muscle artifacts from EEG[C]. Toronto: Proceedings of the IEEE International Conference on Acoustics,Speech and Signal Processing, 2021.
[50]	Sun W Y, Su Y P, Wu X, et al. A novel end-to-end 1D-ResCNN model to remove artifact from EEG signals[J]. Neurocomputing, 2020, 404(9):108-121. doi: 10.1016/j.neucom.2020.04.029

伪迹类型	产生原因及特点
眼动或眨眼	主要由眼动/眨眼所产生,振幅较大,在大脑头皮前部比较明显。
肌肉活动	由头部、肢体、下巴或者舌头等运动产生干扰的信号,其频率通常大于30 Hz,与脑电快波活动相比较,肌肉活动频率更快,波幅更高。
心电	由心脏跳动产生的干扰信号,影响较小。心电伪迹的间隙大多是相等的,但在心律不齐的患者中,这种间隙是不等的。
血管波	在头皮动脉附近的电极产生。如果能同时记录到心电伪迹,则心电伪迹以不变的周期位于血管波伪迹之前。
出汗	有缓慢的漂移,持续时间达数秒为出汗所致的伪迹。这种伪迹在额部最常见,并可影响到临近的几个电极导联。
舌咽部运动	如讲话、咀嚼、吮吸、吞咽、咳嗽、打呃能引起舌运动的伪迹。

当前伪迹去除技术	存在的问题与缺陷
低通、高通、带通等数字滤波方法	伪迹信号通常与脑电信号频谱重叠,只能滤除不在同一频带的噪声。
线性回归、自适应滤波、贝叶斯滤波方法	需要额外的参考通道且采集脑电实验前需进行适当的校准。
独立成分分析(ICA)等盲源分离方法	需人为观察并判别伪迹成分以去除,十分耗时耗力且结果不一定准确;算法效果比较依赖数据量的大小。
小波包分解等基于小波变换方法	分离效果依赖于小波基函数的选择及其与源信号的相似性。
经验模态分解及其不同变体形式	分解过程存在模态混叠等现象;末端效应会影响分解效果;对噪声的鲁棒性较差;目前局限于单通道脑电信号的使用上。
随机森林、聚类等机器学习方法与改进的深度学习方法	在多个伪迹成分同时存在的情况下,算法模型的稳定性和可靠性有待提高;依赖大量脑电数据训练模型。