South-Central Minzu University,a. College of Biomedical Engineering; b. Key Laboratory of Cognitive Science of State Ethnic Affairs Commission; c. Hubei Key Laboratory of Medical Information Analysis and Tumor Diagnosis & Treatment,Wuhan 430074,China
In order to identify dorsiflexion and plantar flexion of ankle, EEG and EMG signals were collected from 5 subjects. The 9-channel EEG and 2-channel EMG signals were selected for pre-processing and feature extraction, and the Event Related Desynchronization (ERD) features of the EEG signals were calculated. The Mean Absolute Value (MAV) and Slope Sign Changes (SSC) characteristics of EMG were calculated. The features were sent to XGBoost (Extreme Gradient Boosting) and Support Vector Machine (SVM) classifiers individually or in combination to compare the recognition effects of different features and classifiers. It is concluded that the recognition effect of EEG signal fusion based on XGBoost classifier is the best.
在真实运动和运动想象中,感觉运动区的EEG频率存在着事件相关去同步化(Event Related Desynchronization,ERD)和事件相关同步化(Event Related Synchronization,ERS)特征[17].PFURTSCHELLER等人证实:执行单侧手任务时,α频段和β频段上对侧EEG振幅降低,同侧EEG振幅增加,这种降低和增加的现象称为ERD和ERS[26].TANG等人使用运动执行或运动想象的28通道EEG信号的ERD/ERS特征来研究左手与右手、左手与双脚的分类,平均分类精度达到85%以上[27].YU等在脚部运动想象任务中,利用7通道EEG信号的ERD特征进行两分类模式识别,准确率达到78%[28],分类性能与实际运动执行非常接近.上述的研究为基于少通道EEG的ERD/ERS特征的脚步运动识别提供了可行依据.因此,本文提取ERD作为EEG信号特征,具体步骤如下[29]:
WENTINKE C, SCHUTV G H, PRINSENE C, et al. Detection of the onset of gait initiation using kinematic sensors and EMG in transfemoral amputees[J]. Gait & Posture, 2014, 39(1): 391-396.
LEES W, YIT, JUNGJ W, et al. Design of a gait phase recognition system that can cope with EMG electrode location variation[J]. IEEE Transactions on Automation Science and Engineering, 2017, 14(3): 1429-1439.
[5]
HUANGH, LIT, BRUSCHINIC, et al. EMG pattern recognition using decomposition techniques for constructing multiclass classifiers[C]//2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob). UTown: ACM, 2016: 1296-1301.
[6]
KUANGY, WUQ, SHAOJ, et al. Extreme learning machine classification method for lower limb movement recognition[J]. Cluster Computing, 2017, 20(4): 3051-3059.
[7]
NIAZII K, JIANGN, TIBERGHIENO, et al. Detection of movement intention from single-trial movement-related cortical potentials[J]. Journal of Neural Engineering, 2011, 8(6): 066009.
[8]
NIAZII K, MRACHACZ-KERSTINGN, JIANGN, et al. Peripheral electrical stimulation triggered by self-paced detection of motor intention enhances motor evoked potentials[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2012, 20(4): 595-604.
[9]
KEEY J, WOND O, LEES W. Classification of left and right foot movement intention based on steady-state somatosensory evoked potentials[C]//2017 5th International Winter Conference on Brain-Computer Interface (BCI). Gangwon Province: IEEE, 2017:106-108.
[10]
HUANGY, ENGLEHARTK B, HUDGINSB, et al. A Gaussian mixture model based classification scheme for myoelectric control of powered upper limb prostheses[J]. IEEE Transactions on Bio-Medical Engineering, 2005, 52(11): 1801-1811.
[11]
LEEBR, SAGHAH, CHAVARRIAGAR, et al. A hybrid brain-computer interface based on the fusion of electroencephalographic and electromyographic activities[J]. Journal of Neural Engineering, 2011, 8(2): 025011.
HERMENSH J, FRERIKSB, DISSELHORST-KLUGC, et al. Development of recommendations for SEMG sensors and sensor placement procedures[J]. Journal of Electromyography and Kinesiology, 2000, 10(5): 361-374.
[14]
HASHIMOTOY, USHIBAJ, KIMURAA, et al. Correlation between EEG-EMG coherence during isometric contraction and its imaginary execution[J]. Acta Neurobiologiae Experimentalis, 2010, 70(1): 76-85.
[15]
WITHAMC L, RIDDLEC N, BAKERM R, et al. Contributions of descending and ascending pathways to corticomuscular coherence in humans[J]. The Journal of Physiology, 2011, 589(Pt 15): 3789-3800.
[16]
HUANGH, KUIKENT A, LIPSCHUTZR D. A strategy for identifying locomotion modes using surface electromyography[J]. IEEE Transactions on Bio-Medical Engineering, 2009, 56(1): 65-73.
[17]
FARRELLT R, WEIRR F. The optimal controller delay for myoelectric prostheses[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2007, 15(1): 111-118.
[18]
乔晓艳,顼春元,王艳景.基于小波变换方法的运动想象脑电特征提取[C]//Proceedings of the 2011 International Conference on Future Computer Science and Application(FCSA 2011 V2).Kota Kinabalu: International Industrial Electronic Center, 2011: 317-320.
[19]
LIM, CHENW. FFT-based deep feature learning method for EEG classification[J]. Biomedical Signal Processing and Control, 2021, 66: 102492.
BRARIZ, BELGHITHS. A novel Machine Learning approach for epilepsy diagnosis using EEG signals based on Correlation Dimension[J]. IFAC-PapersOnLine, 2021, 54(17): 7-11.
PFURTSCHELLERG, LOPES DA SILVAF H. Event-related EEG/MEG synchronization and desynchronization: Basic principles[J]. Clinical Neurophysiology, 1999, 110(11): 1842-1857.
[27]
TANGZ, SUNS, ZHANGS, et al. A brain-machine interface based on ERD/ERS for an upper-limb exoskeleton control[J]. Sensors, 2016, 16(12): 2050.
[28]
GORDLEEVAS Y, LOBOVS A, GRIGOREVN A, et al. Real-time EEG-EMG human-machine interface-based control system for a lower-limb exoskeleton[J]. IEEE Access, 2020, 8: 84070-84081.
[29]
GRAIMANNB, HUGGINSJ E, LEVINES P, et al. Visualization of significant ERD/ERS patterns in multichannel EEG and ECoG data[J]. Clinical Neurophysiology, 2002, 113(1): 43-47.
WEIP N, ZHANGJ H, TIANF F, et al. A comparison of neural networks algorithms for EEG and sEMG features based gait phases recognition[J]. Biomedical Signal Processing and Control, 2021, 68: 102587.
[32]
PENGF, PENGW, ZHANGC, et al. IoT assisted kernel linear discriminant analysis based gait phase detection algorithm for walking with cognitive tasks[J]. IEEE Access, 2019, 7: 68240-68249.
[33]
ALTANE, PEHLIVANK, KAPLANOĞLUE. Comparison of EMG based finger motion classification algorithms[C]//2019 27th Signal Processing and Communications Applications Conference (SIU). Sivas: IEEE, 2019: 1-4.