JIN Xingyu,SHENG Zhiwei,DU Jie.A EEG Noise Reduction Network based on Parallel Multiple Attention Mechanisms[J].Journal of Chengdu University of Information Technology,2026,41(01):7-16.[doi:10.16836/j.cnki.jcuit.2026.01.002]
基于并行多重注意力机制的脑电降噪网络
- Title:
- A EEG Noise Reduction Network based on Parallel Multiple Attention Mechanisms
- 文章编号:
- 2096-1618(2026)01-0007-10
- Keywords:
- deep neural network; EEG; EEG noise reduction; EEG artifact removal
- 分类号:
- TN911.7
- 文献标志码:
- A
- 摘要:
- 脑电图作为实现脑机接口的必须工具,随着脑机接口热度不断升高,对脑电信号的预处理也成为脑机接口研究中不可或缺的一部分。然而脑电在实际采样过程中,容易受到环境因素和被采样人生理因素产生的运动伪影的影响。因而,削弱或去除采样脑电信号中的噪声成为当前重要课题。近几年,基于深度学习的脑电信号降噪方案有了长足的进步,并获得较好的降噪效果。在脑电信号降噪过程中,数据的自相似性特征被广泛用于EEG降噪环节,但在当前基于深度学习的脑电降噪研究中,一般只片面地关注局部特征或只关注了全局特征。所以为优化对于脑电时序依赖性的降噪处理,提出一种基于并行的LSH Attention和Multi-head Self-attention的脑电降噪模型DuoAttentionNet,模型结合LSH Attention擅长处理局部相关性强的数据和Multi-head Self-attention擅长捕捉全局的长距离依赖关系的优点。实验表明,DuoAttentionNet在无须进行数据增强的条件下,取得当前基于深度学习的脑电降噪模型的最好成果。
- Abstract:
- Electroencephalography(EEG)as an essential tool for realizing brain-computer interfaces(BCIs)has become increasingly important as the popularity of BCIs continues to rise. The pre-processing of brain electrical signals has become an indispensable part of BCI research. However, EEG signals are prone to being affected by environmental and physiological factors of the subjects during the sampling process, leading to the generation of movement artifacts. Therefore, reducing or eliminating noise in the sampled EEG signals has become an important topic. In recent years, solutions for EEG signal denoising based on deep learning have made significant progress and achieved good denoising results. During the EEG signal denoising process, the self-similarity feature of the data is widely used in the EEG denoising stage. However, current research on deep learning based EEG denoising usually focuses on either local features or global features, neglecting a comprehensive approach. To optimize the denoising processing of EEG signals that depend on temporal relationships, this paper proposes a brain EEG denoising model called DuoAttentionNet, which combines the advantages of LSH(Locality Sensitive Hashing)Attention, which is good at processing data with strong local correlations, and Multi-head Self-attention, which is good at capturing long-range dependencies at the global level. Extensive experiments have shown that DuoAttentionNet achieves the best results among current deep learning-based EEG denoising models without the need for data augmentation.
参考文献/References:
[1] Borhani S,Kilmarx J,Saffo D,et al.Optimizing prediction model for a noninvasive brain-computer interface platform using channel selection,classification,and regression[J].IEEE journal of biomedical and health informatics,2019,23(6):2475-2482.
[2] Radüntz T,Scouten J,Hochmuth O,et al.Automated EEG artifact elimination by applying machine learning algorithms to ICA-based features[J].Journal of neural engineering,2017,14(4):046004.
[3] Urigüen J A,Garcia-Zapirain B.EEG artifact removal——state-of-the-art and guidelines[J].Journal of neural engineering,2015,12(3):031001.
[4] Arpaia P,Moccaldi N,Prevete R,et al.A wearable EEG instrument for real-time frontal asymmetry monitoring in worker stress analysis[J].IEEE Transactions on Instrumentation and Measurement,2020,69(10):8335-8343.
[5] Liu A,Liu Q,Zhang X,et al.Muscle artifact removal toward mobile SSVEP-based BCI:A comparative study[J].IEEE Transactions on Instrumentation and Measurement,2021,70:1-12.
[6] Klados M A,Papadelis C,Braun C,et al.REG-ICA:a hybrid methodology combining blind source separation and regression techniques for the rejection of ocular artifacts[J].Biomedical Signal Processing and Control,2011,6(3):291-300.
[7] Sweeney K T,Ward T E,McLoone S F.Artifact removal in physiological signals——Practices and possibilities[J].IEEE transactions on information technology in biomedicine,2012,16(3):488-500.
[8] Marque C,Bisch C,Dantas R,et al.Adaptive filtering for ECG rejection from surface EMG recordings[J].Journal of electromyography and kinesiology,2005,15(3):310-315.
[9] Somers B,Francart T,Bertrand A.A generic EEG artifact removal algorithm based on the multi-channel Wiener filter[J].Journal of neural engineering,2018,15(3):036007.
[10] Lakshmi K G A,Surling S N N,Sheeba O.A novel approach for the removal of artifacts in EEG signals[C].2017 International Conference on Wireless Communications,Signal Processing and Networking(WiSPNET).IEEE,2017:2595-2599.
[11] Somers B,Bertrand A.Removal of eye blink artifacts in wireless EEG sensor networks using reduced-bandwidth canonical correlation analysis[J].Journal of neural engineering,2016,13(6):066008.
[12] Borga M,Friman O,Lundberg P,et al.A canonical correlation approach to exploratory data analysis in fMRI[C].Proceedings of the ISMRM Annual Meeting,Honolulu,Hawaii.2002.
[13] Soomro M H,Badruddin N,Yusoff M Z,et al.Automatic eye-blink artifact removal method based on EMD-CCA[C].2013 ICME International Conference on Complex Medical Engineering.IEEE,2013:186-190.
[14] Lin J,Zhang A.Fault feature separation using wavelet-ICA filter[J].NDT & e International,2005,38(6):421-427.
[15] Soomro M H,Badruddin N,Yusoff M Z,et al.A method for automatic removal of eye blink artifacts from EEG based on EMD-ICA[C].2013 IEEE 9th international colloquium on signal processing and its applications.IEEE,2013:129-134.
[16] Yang Banghua,Duan Kaiwen,Fan Chengcheng,et al.Automatic ocular artifacts removal in EEG using deep learning[J].Biomedical Signal Processing and Control, 2018,43:148-158.
[17] Nagar S,Kumar A,Swamy M N S.Orthogonal features-based EEG signal denoising using fractionally compressed autoencoder[J].Signal Processing,2021, 188:108225.
[18] Sawangjai P,Trakulruangroj M,Boonnag C,et al.EEGANet:Removal of ocular artifacts from the EEG signal using generative adversarial networks[J].IEEE Journal of Biomedical and Health Informatics, 2022,26(10):4913-4924.
[19] Sun W,Su Y,Wu X,et al.A novel end-to-end 1D-ResCNN model to remove artifact from EEG signals[J].Neurocomputing,2020,404:108-121.
[20] Zhang H,Zhao M,Wei C,et al.EEGdenoiseNet:A benchmark dataset for deep learning solutions of EEG denoising[J].Journal of Neural Engineering,2021,18(5):056057.
[21] Zhang H,Wei C,Zhao M,et al.A novel convolutional neural network model to remove muscle artifacts from EEG[C].ICASSP 2021-2021 IEEE International Conference on Acoustics,Speech and Signal Processing(ICASSP).IEEE,2021:1265-1269.
[22] Wang W,Li B,Wang H.A novel end-to-end network based on a bidirectional GRU and a self-attention mechanism for denoising of electroencephalography signals[J].Neuroscience,2022,505:10-20.
[23] Pu X,Yi P,Chen K,et al.EEGDnet:Fusing non-local and local self-similarity for EEG signal denoising with transformer[J].Computers in Biology and Medicine,2022,151:106248.
[24] Gao T,Chen D,Tang Y,et al.EEG reconstruction with a dual-scale CNN-LSTM model for deep artifact removal[J].IEEE Journal of Biomedical and Health Informatics,2022,27(3):1283-1294.
[25] 尹瑾,刘爱萍,李畅,等.基于并行 CNN 和 Transformer 的脑电降噪网络[J].Journal of Signal Processing,2023,39(8).
[26] Yin J,Liu A,Li C,et al.A GAN guided parallel CNN and transformer network for EEG denoising[J].IEEE Journal of Biomedical and Health Informatics,2023.
[27] Kitaev N,Kaiser ,Levskaya A.Reformer:The efficient transformer[J].arXiv preprint arXiv,2020:2001.04451.
[28] Goldberger A L,Amaral L A N,Glass L,et al.PhysioBank,PhysioToolkit,and PhysioNet:components of a new research resource for complex physiologic signals[J].circulation,2000,101(23):215-220.
[29] Moody G B,Mark R G.The impact of the MIT-BIH arrhythmia database[J].IEEE engineering in medicine and biology magazine,2001,20(3):45-50.
[30] Klados M A,Bamidis P D.A semi-simulated EEG/EOG dataset for the comparison of EOG artifact rejection techniques[J].Data in brief,2016,8:1004-1006.
[31] Elbert T,Lutzenberger W,Rockstroh B,et al.Removal of ocular artifacts from the EEG——a biophysical approach to the EOG[J].Electroencephalography and clinical neurophysiology,1985,60(5):455-463.
[32] Saini M,Satija U,Upadhayay M D.Wavelet based waveform distortion measures for assessment of denoised EEG quality with reference to noise-free EEG signal[J].IEEE Signal Processing Letters,2020,27:1260-1264.
相似文献/References:
[1]王家珉,李田家,顾桃峰,等.一种基于深度卷积神经网络的电磁干扰识别与抑制方法[J].成都信息工程大学学报,2024,39(01):43.[doi:10.16836/j.cnki.jcuit.2024.01.008]
WANG Jiamin,LI Tianjia,GU Taofeng,et al.A Method of Electromagnetic Interference Identification and Suppression based on Deep Convolutional Neural Network[J].Journal of Chengdu University of Information Technology,2024,39(01):43.[doi:10.16836/j.cnki.jcuit.2024.01.008]
备注/Memo
收稿日期:2024-07-15
基金项目:国家重点研发计划“网络空间安全治理”重点专项课题(2022YFB3103103); 四川省重点研发计划项目(2022YFS0571、2021YFSY0012、2021JDRC0046、2020YFG03077)
通信作者:盛志伟.E-mail:7782988@qq.com