A Multi-Branch 3D Convolutional Neural Network for EEG-Based Motor Imagery Classification

被引:217
作者
Zhao, Xinqiao [1 ]
Zhang, Hongmiao [1 ]
Zhu, Guilin [1 ]
You, Fengxiang [1 ]
Kuang, Shaolong [1 ]
Sun, Lining [1 ]
机构
[1] Soochow Univ, Coll Mech & Elect Engn, Suzhou 215000, Peoples R China
关键词
Electroencephalogram (EEG); motor imagery (MI); 3D convolutional neural network (3D CNN); multi-branch structure; FEATURE-EXTRACTION; WAVELET TRANSFORM; BRAIN; POTENTIALS;
D O I
10.1109/TNSRE.2019.2938295
中图分类号
R318 [生物医学工程];
学科分类号
0831 ;
摘要
One of the challenges in motor imagery (MI) classification tasks is finding an easy-handled electroencephalogram (EEG) representation method which can preserve not only temporal features but also spatial ones. To fully utilize the features on various dimensions of EEG, a novel MI classification framework is first introduced in this paper, including a new 3D representation of EEG, a multi-branch 3D convolutional neural network (3D CNN) and the corresponding classification strategy. The 3D representation is generated by transforming EEG signals into a sequence of 2D array which preserves spatial distribution of sampling electrodes. The multi-branch 3D CNN and classification strategy are designed accordingly for the 3D representation. Experimental evaluation reveals that the proposed framework reaches state-of-the-art classification kappa value level and significantly outperforms other algorithms by 50% decrease in standard deviation of different subjects, which shows good performance and excellent robustnesson different subjects. The framework also shows great performance with only nine sampling electrodes, which can significantly enhance its practicality. Moreover, the multi-branch structure exhibits its low latency and a strong ability in mitigating overfitting issues which often occur in MI classification because of the small training dataset.
引用
收藏
页码:2164 / 2177
页数:14
相关论文
共 48 条
[1]   Global cortical activity predicts shape of hand during grasping [J].
Agashe, Harshavardhan A. ;
Paek, Andrew Y. ;
Zhang, Yuhang ;
Contreras-Vidal, Jose L. .
FRONTIERS IN NEUROSCIENCE, 2015, 9
[2]  
Ang KK, 2008, IEEE IJCNN, P2390, DOI 10.1109/IJCNN.2008.4634130
[3]  
[Anonymous], P 2008 INT C EL DES
[4]  
Bashivan P, 2015, Learning Representations from EEG with Deep Recurrent-Convolutional Neural Networks
[5]   Random Forest and Filter Bank Common Spatial Patterns for EEG-Based Motor Imagery Classification [J].
Bentlemsan, Maouia ;
Zemouri, ET-Tahir ;
Bouchaffra, Djamel ;
Yahya-Zoubir, Bahia ;
Ferroudji, Karim .
PROCEEDINGS FIFTH INTERNATIONAL CONFERENCE ON INTELLIGENT SYSTEMS, MODELLING AND SIMULATION, 2014, :235-238
[6]   BCI competition 2003 - Data sets Ib and IIb: Feature extraction from event-related brain potentials with the continuous wavelet transform and the t-value scalogram [J].
Bostanov, V .
IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, 2004, 51 (06) :1057-1061
[7]   Sitting and standing intention can be decoded from scalp EEG recorded prior to movement execution [J].
Bulea, Thomas C. ;
Prasad, Saurabh ;
Kilicarslan, Atilla ;
Contreras-Vidal, Jose L. .
FRONTIERS IN NEUROSCIENCE, 2014, 8
[8]   A parametric feature extraction and classification strategy for brain-computer interfacing [J].
Burke, DR ;
Kelly, SR ;
de Chazal, P ;
Reilly, RB ;
Finucane, C .
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING, 2005, 13 (01) :12-17
[9]   SUPPORT-VECTOR NETWORKS [J].
CORTES, C ;
VAPNIK, V .
MACHINE LEARNING, 1995, 20 (03) :273-297
[10]   NEAREST NEIGHBOR PATTERN CLASSIFICATION [J].
COVER, TM ;
HART, PE .
IEEE TRANSACTIONS ON INFORMATION THEORY, 1967, 13 (01) :21-+