SVM-Based Validation of Motor Unit Potential Trains Extracted by EMG Signal Decomposition

Motor unit potential trains (MUPTs) extracted via electromyographic (EMG) signal decomposition can aid in the diagnosis of neuromuscular disorders and the study of the neural control of movement, but only if they are valid. In this paper, support vector machine (SVM)-based supervised classifiers are proposed to estimate the validity of extracted MUPTs. The classifiers use either the MU firing pattern or the MUP shape consistency of an MUPT, or both, to estimate its validity. The developed classifiers estimate the class label of an MUPT (i.e., valid/invalid) and a degree of support for the decision being made. A single SVM that estimates the validity of a given MUPT using extracted MU firing pattern and MUP shape features was investigated. In addition, the effectiveness of multiclassifier techniques which estimate the overall validity of a train by fusing the MU firing pattern and MUP shape validity of a given MUPT, determined separately by two distinct SVMs, was also investigated. Training based only on simulated data showed robust classification performance of the several multiclassifier methods when tested using both simulated and real test data. Of the methods studied, the multiclassifier constructed using trainable logistic regression to aggregate base classifier outputs had the best performance. Assuming 12.7% of extracted MUPTs are on average invalid, the estimated accuracy for this method in correctly categorizing MUPTs extracted during decomposition was 99.4% and 98.8% for simulated and real data, respectively.

[1]  C. D. De Luca,et al.  Control scheme governing concurrently active human motor units during voluntary contractions , 1982, The Journal of physiology.

[2]  Mohamed S. Kamel,et al.  Diversity-based combination of non-parametric classifiers for EMG signal decomposition , 2008, Pattern Analysis and Applications.

[3]  D W Stashuk,et al.  Motor unit potential characterization using "pattern discovery". , 2008, Medical engineering & physics.

[4]  Kevin C. McGill,et al.  Automatic Decomposition of the Clinical Electromyogram , 1985, IEEE Transactions on Biomedical Engineering.

[5]  John L. Semmlow,et al.  Biosignal and biomedical image processing : MATLAB-based applications , 2004 .

[6]  Kevin C. McGill,et al.  EMGLAB: An interactive EMG decomposition program , 2005, Journal of Neuroscience Methods.

[7]  Carlo J De Luca,et al.  Decomposition of indwelling EMG signals. , 2008, Journal of applied physiology.

[8]  David G. Stork,et al.  Pattern Classification (2nd ed.) , 1999 .

[9]  Anders Fuglsang-Frederiksen,et al.  The role of different EMG methods in evaluating myopathy , 2006, Clinical Neurophysiology.

[10]  John Platt,et al.  Probabilistic Outputs for Support vector Machines and Comparisons to Regularized Likelihood Methods , 1999 .

[11]  Hossein Parsaei,et al.  MUP shape-based validation of a motor unit potential train , 2009, 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[12]  K C McGill,et al.  Automatic decomposition of multichannel intramuscular EMG signals. , 2009, Journal of electromyography and kinesiology : official journal of the International Society of Electrophysiological Kinesiology.

[13]  K C McGill,et al.  Rigorous a Posteriori Assessment of Accuracy in EMG Decomposition , 2011, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[14]  Mohamed S. Kamel,et al.  Integrating Heterogeneous Classifier Ensembles for EMG Signal Decomposition Based on Classifier Agreement , 2010, IEEE Transactions on Information Technology in Biomedicine.

[15]  Mohamed S. Kamel,et al.  Fusion of multiple classifiers for motor unit potential sorting , 2008, Biomed. Signal Process. Control..

[16]  H. Clamann.  Statistical analysis of motor unit firing patterns in a human skeletal muscle. , 1969, Biophysical journal.

[17]  D.W. Stashuk,et al.  Probabilistic inference-based classification applied to myoelectric signal decomposition , 1992, IEEE Transactions on Biomedical Engineering.

[18]  Christos D. Katsis,et al.  A novel method for automated EMG decomposition and MUAP classification , 2006, Artif. Intell. Medicine.

[19]  Mohamed Kamel,et al.  Adaptive fuzzy k-NN classifier for EMG signal decomposition. , 2006, Medical engineering & physics.

[20]  Daniel W Stashuk,et al.  Decomposition‐based quantitative electromyography: Effect of force on motor unit potentials and motor unit number estimates , 2005, Muscle & nerve.

[21]  D. Hosmer,et al.  Applied Logistic Regression , 1991 .

[22]  Mohamed S. Kamel,et al.  A Hybrid Classifier Fusion Approach for Motor Unit Potential Classification During EMG Signal Decomposition , 2007, IEEE Transactions on Biomedical Engineering.

[23]  Daniel W. Stashuk,et al.  Physiologically based simulation of clinical EMG signals , 2005, IEEE Transactions on Biomedical Engineering.

[24]  R. Suganya,et al.  Data Mining Concepts and Techniques , 2010 .

[25]  Armando Malanda-Trigueros,et al.  Automated decomposition of intramuscular electromyographic signals , 2006, IEEE Transactions on Biomedical Engineering.

[26]  Subhash C. Bagui,et al.  Combining Pattern Classifiers: Methods and Algorithms , 2005, Technometrics.

[27]  Roberto Merletti,et al.  Decomposition of Intramuscular EMG Signals , 2004 .

[28]  Alexander Adam,et al.  Motor unit control and force fluctuation during fatigue. , 2009, Journal of applied physiology.

[29]  D W Stashuk,et al.  Decomposition and quantitative analysis of clinical electromyographic signals. , 1999, Medical engineering & physics.

[30]  Hossein Parsaei,et al.  Adaptive motor unit potential train validation using MUP shape information. , 2011, Medical engineering & physics.

[31]  Hossein Parsaei,et al.  Validating motor unit firing patterns extracted by EMG signal decomposition , 2011, Medical & Biological Engineering & Computing.

[32]  C. Moritz,et al.  Discharge rate variability influences the variation in force fluctuations across the working range of a hand muscle. , 2005, Journal of neurophysiology.

[33]  Ludmila I. Kuncheva,et al.  Combining Pattern Classifiers: Methods and Algorithms , 2004 .

[34]  M.H. Hassoun,et al.  NNERVE: Neural Network Extraction of Repetitive Vectors for Electromyography. I. Algorithm , 1994, IEEE Transactions on Biomedical Engineering.

[35]  Vladimir Vapnik,et al.  Statistical learning theory , 1998 .

[36]  S. Podnar,et al.  Bayesian characterization of external anal sphincter muscles using quantitative electromyography , 2008, Clinical Neurophysiology.

[37]  Hossein Parsaei,et al.  Intramuscular EMG signal decomposition. , 2010, Critical reviews in biomedical engineering.

[38]  Dr. D. Stashuk,et al.  Robust supervised classification of motor unit action potentials , 2006, Medical and Biological Engineering and Computing.

[39]  E Stålberg,et al.  The role of electromyography in neurology. , 1997, Electroencephalography and clinical neurophysiology.

[40]  Kevin C McGill,et al.  Estimating motor-unit architectural properties by analyzing motor-unit action potential morphology , 2001, Clinical Neurophysiology.

[41]  Hossein Parsaei,et al.  A review of clinical quantitative electromyography. , 2010, Critical reviews in biomedical engineering.

[42]  S D Nandedkar,et al.  Multi‐motor unit action potential analysis (MMA) , 1995, Muscle & nerve.

[43]  Zeynep Erim,et al.  Decomposition of Intramuscular EMG Signals Using a Heuristic Fuzzy Expert System , 2008, IEEE Transactions on Biomedical Engineering.

[44]  R Dengler,et al.  The role of electromyography (EMG) in the diagnosis of ALS. , 2000, Amyotrophic lateral sclerosis and other motor neuron disorders : official publication of the World Federation of Neurology, Research Group on Motor Neuron Diseases.

[45]  Hossein Parsaei,et al.  Validation of motor unit potential trains using motor unit firing pattern information , 2009, 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[46]  Daniel W Stashuk,et al.  Physiological characteristics of motor units in the brachioradialis muscle across fatiguing low-level isometric contractions. , 2008, Journal of electromyography and kinesiology : official journal of the International Society of Electrophysiological Kinesiology.

[47]  Ronald S. Lefever,et al.  A Procedure for Decomposing the Myoelectric Signal Into Its Constituent Action Potentials - Part I: Technique, Theory, and Implementation , 1982, IEEE Transactions on Biomedical Engineering.

[48]  D. Stashuk,et al.  Automatic decomposition of selective needle-detected myoelectric signals , 1988, IEEE Transactions on Biomedical Engineering.

[49]  B Mambrito,et al.  A technique for the detection, decomposition and analysis of the EMG signal. , 1984, Electroencephalography and clinical neurophysiology.

[50]  Mohamed S. Kamel,et al.  Adaptive certainty-based classification for decomposition of EMG signals , 2006, Medical and Biological Engineering and Computing.

[51]  D Stashuk,et al.  EMG signal decomposition: how can it be accomplished and used? , 2001, Journal of electromyography and kinesiology : official journal of the International Society of Electrophysiological Kinesiology.

[52]  C. D. De Luca,et al.  Behaviour of human motor units in different muscles during linearly varying contractions , 1982, The Journal of physiology.