Subject-specific EEG channel selection using non-negative matrix factorization for lower-limb motor imagery recognition

Objective- This study aims to propose and validate a subject-specific approach to recognize two different cognitive neural states (relax and pedaling motor imagery (MI)) by selecting the relevant electroencephalogram (EEG) channels. The main aims of the proposed work are: (i) to reduce the computational complexity of the BCI systems during MI detection by selecting the relevant EEG channels, (ii) to reduce the amount of data overfitting that may arise due to unnecessary channels and redundant features, and (iii) to reduce the classification time for real-time BCI applications. Approach- The proposed method selects subject-specific EEG channels and features based on their MI. In this work, we make use of Non-Negative Matrix Factorization (NMF) to extract the weight of the EEG channels based on their contribution to MI detection. Further, the neighborhood component analysis is used for subject-specific feature selection. Main results- We executed the experiments using EEG signals recorded for MI where 10 healthy subjects performed MI movement of the lower limb to generate motor commands. An average accuracy of 96.66%, average true positive rate of 97.77%, average false positives rate of 4.44%, and average Kappa of 93.33% were obtained. The proposed subject-specific EEG channel selection based MI recognition system provides 13.20% improvement in detection accuracy, and 27% improvement in Kappa value with less number of EEG channels compared to the results obtained using all EEG channels. Significance- The proposed subject-specific BCI system has been found significantly advantageous compared to the typical approach of using a fixed channel configuration. This work shows that less EEG channels not only reduce computational complexity and processing time (2 times faster) but also improves the MI detection performance. The proposed method selects EEG locations related to the foot movement, which may be relevant for neuro-rehabilitation using lower-limb movements that may provide a real-time and more natural interface between patient and robotic device.

[1]  Wang Haiping,et al.  Classification of Motor Imagery EEG Based on Sparsification and Non-negative Matrix Factorization , 2018 .

[2]  Andrzej Cichocki,et al.  Kernel nonnegative matrix factorization for spectral EEG feature extraction , 2009, Neurocomputing.

[3]  Wei Yang,et al.  Neighborhood Component Feature Selection for High-Dimensional Data , 2012, J. Comput..

[4]  Robert Riener,et al.  Control strategies for active lower extremity prosthetics and orthotics: a review , 2015, Journal of NeuroEngineering and Rehabilitation.

[5]  K.-R. Muller,et al.  Optimizing Spatial filters for Robust EEG Single-Trial Analysis , 2008, IEEE Signal Processing Magazine.

[6]  Fathi E. Abd El-Samie,et al.  A review of channel selection algorithms for EEG signal processing , 2015, EURASIP Journal on Advances in Signal Processing.

[7]  Josef Kittler,et al.  Floating search methods in feature selection , 1994, Pattern Recognit. Lett..

[8]  Jin Zhou,et al.  A Fast, Open EEG Classification Framework Based on Feature Compression and Channel Ranking , 2018, Front. Neurosci..

[9]  Klaus-Robert Müller,et al.  The Berlin Brain-Computer Interface: Accurate performance from first-session in BCI-naive subjects , 2008, IEEE Transactions on Biomedical Engineering.

[10]  Feiping Nie,et al.  Efficient and Robust Feature Selection via Joint ℓ2, 1-Norms Minimization , 2010, NIPS.

[11]  M. K. M. Rahman,et al.  A Review on the Components of EEG-based Motor Imagery Classification with Quantitative Comparison , 2017 .

[12]  V. Feigin,et al.  Global Burden of Stroke. , 2017, Circulation research.

[13]  J. Eggert,et al.  Sparse coding and NMF , 2004, 2004 IEEE International Joint Conference on Neural Networks (IEEE Cat. No.04CH37541).

[14]  Teodiano Freire Bastos Filho,et al.  Knee Impedance Modulation to Control an Active Orthosis Using Insole Sensors , 2017, Sensors.

[15]  M.N.S. Swamy,et al.  Nonnegative Matrix Factorization , 2014 .

[16]  ChengXiang Zhai,et al.  Robust Unsupervised Feature Selection , 2013, IJCAI.

[17]  Adam Prügel-Bennett,et al.  Rank Selection in Nonnegative Matrix Factorization using Minimum Description Length , 2017, Neural Computation.

[18]  H. Sebastian Seung,et al.  Algorithms for Non-negative Matrix Factorization , 2000, NIPS.

[19]  Na Lu,et al.  Motor imagery classification via combinatory decomposition of ERP and ERSP using sparse nonnegative matrix factorization , 2015, Journal of Neuroscience Methods.

[20]  Isabelle Bloch,et al.  Subject-Specific Channel Selection Using Time Information for Motor Imagery Brain–Computer Interfaces , 2016, Cognitive Computation.

[21]  Chun Kwang Tan,et al.  Lateral Symmetry of Synergies in Lower Limb Muscles of Acute Post-stroke Patients After Robotic Intervention , 2018, Front. Neurosci..

[22]  G. Pfurtscheller,et al.  Brain-Computer Interfaces for Communication and Control. , 2011, Communications of the ACM.

[23]  Andrzej Cichocki,et al.  Nonnegative Matrix Factorization for Motor Imagery EEG Classification , 2006, ICANN.

[24]  G. Ferrigno,et al.  Cycling Induced by Electrical Stimulation Improves Muscle Activation and Symmetry During Pedaling in Hemiparetic Patients , 2012, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[25]  Gary E. Birch,et al.  Sparse spatial filter optimization for EEG channel reduction in brain-computer interface , 2008, 2008 IEEE International Conference on Acoustics, Speech and Signal Processing.

[26]  Mingwei Yu,et al.  Feature selection method based on mutual information and class separability for dimension reduction in multidimensional time series for clinical data , 2015, Biomed. Signal Process. Control..

[27]  Yi-Hung Liu,et al.  Motor Imagery EEG Classification for Patients with Amyotrophic Lateral Sclerosis Using Fractal Dimension and Fisher’s Criterion-Based Channel Selection , 2017, Sensors.

[28]  Rui Zhang,et al.  Local Temporal Correlation Common Spatial Patterns for Single Trial EEG Classification during Motor Imagery , 2013, Comput. Math. Methods Medicine.

[29]  D. Farina,et al.  A brain–computer interface for single-trial detection of gait initiation from movement related cortical potentials , 2015, Clinical Neurophysiology.

[30]  Dong Liu,et al.  An EEG-based brain-computer interface for gait training , 2017, 2017 29th Chinese Control And Decision Conference (CCDC).

[31]  Lee Shepstone,et al.  Effects of Lower Limb Reciprocal Pedalling Exercise on Motor Function after Stroke: A Systematic Review of Randomized and Nonrandomized Studies , 2012, International journal of stroke : official journal of the International Stroke Society.

[32]  Klaus-Robert Müller,et al.  The non-invasive Berlin Brain–Computer Interface: Fast acquisition of effective performance in untrained subjects , 2007, NeuroImage.

[33]  H. Sebastian Seung,et al.  Learning the parts of objects by non-negative matrix factorization , 1999, Nature.

[34]  A. Ishiyama,et al.  Nonnegative matrix factorization common spatial pattern in brain machine interface , 2015, The 3rd International Winter Conference on Brain-Computer Interface.

[35]  M. Goffredo,et al.  Predictors of activities of daily living outcomes after upper limb robot-assisted therapy in subacute stroke patients , 2018, PloS one.

[36]  Christian Jutten,et al.  Classification of covariance matrices using a Riemannian-based kernel for BCI applications , 2013, Neurocomputing.

[37]  Klaus-Robert Müller,et al.  Spatio-spectral filters for improving the classification of single trial EEG , 2005, IEEE Transactions on Biomedical Engineering.

[38]  Zhengya Sun,et al.  L0-norm Based Structural Sparse Least Square Regression for Feature Selection , 2015, Pattern Recognit..

[39]  Richard R Neptune,et al.  Coordination of Hemiparetic Locomotion after Stroke Rehabilitation , 2005, Neurorehabilitation and neural repair.

[40]  José Maria Azorín,et al.  Effects of tDCS on Real-Time BCI Detection of Pedaling Motor Imagery , 2018, Sensors.

[41]  Teodiano Freire Bastos Filho,et al.  Admittance Controller with Spatial Modulation for Assisted Locomotion using a Smart Walker , 2018, Journal of Intelligent & Robotic Systems.

[42]  Jianhua Wang,et al.  Detection of lower-limb movement intention from EEG signals , 2017, 2017 12th IEEE Conference on Industrial Electronics and Applications (ICIEA).

[43]  Benjamin Blankertz,et al.  Robustifying EEG data analysis by removing outliers , 2009 .

[44]  Dong Liu,et al.  EEG-Based Lower-Limb Movement Onset Decoding: Continuous Classification and Asynchronous Detection , 2018, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[45]  Xingyu Wang,et al.  Improved SFFS method for channel selection in motor imagery based BCI , 2016, Neurocomputing.

[46]  Ning Jiang,et al.  Peripheral Electrical Stimulation Triggered by Self-Paced Detection of Motor Intention Enhances Motor Evoked Potentials , 2012, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[47]  Hernando Ombao,et al.  An exploratory data analysis of electroencephalograms using the functional boxplots approach , 2015, Front. Neurosci..

[48]  Zuren Feng,et al.  Structure constrained semi-nonnegative matrix factorization for EEG-based motor imagery classification , 2015, Comput. Biol. Medicine.

[49]  Klaus-Robert Müller,et al.  Combined Optimization of Spatial and Temporal Filters for Improving Brain-Computer Interfacing , 2006, IEEE Transactions on Biomedical Engineering.

[50]  Andri Mirzal,et al.  NMF versus ICA for blind source separation , 2017, Adv. Data Anal. Classif..

[51]  Xingyu Wang,et al.  An Optimized Channel Selection Method Based on Multifrequency CSP-Rank for Motor Imagery-Based BCI System , 2019, Comput. Intell. Neurosci..