Model-Based Simulation of Surface Electromyography Signals and Its Analysis Under Fatiguing Conditions Using Tunable Wavelets
01 Jan 2021-pp 99-108
TL;DR: In this paper, the amplitude-based features of tunable Q-wavelet coefficients are used to identify the characteristic changes associated with varied fatiguing conditions, and the results show that synthetically generated signal is able to truly represent fatiguing and nonfatiguing conditions.
Abstract: Synthetic signals that represent fatiguing contractions of biceps brachii muscle are generated in this work using a comprehensive mathematical model. These signals are the biomarkers of muscle electrical activity that could be recorded non-invasively on the skin surface using Surface Electromyography (sEMG). The important components of the adopted synthetic sEMG model are current source, volume conductor, motor unit recruitment, and firing behavior functions. For this study, the amplitude (A) and scaling factor (λ) of the current source function is selected appropriate to fatiguing conditions. Further, tunable Q-wavelet method is applied to compute the frequency range associated with fatigue in the synthetic signal. The resultant wavelet coefficients are obtained using multirate filter bank where the scaling factors α and β are chosen so as to meet the anticipated Q-factor and the ranges of frequency bands. The results show that synthetically generated signal is able to truly represent fatiguing and nonfatiguing conditions. The amplitude-based features of tunable Q-wavelet coefficients are able to identify the characteristic changes associated with varied fatiguing conditions. Model generated frequency responses in fatiguing conditions are in agreement with the experimental results reported elsewhere. As fatigue is a temporary failure of skeletal muscles to maintain a required force for the accomplishment of a particular task, the model proposed here could be used as a validation of sEMG measurements in health and disease.
References
More filters
TL;DR: In this study, most complete and up-to-date thirty-seven time domain and frequency domain features have been proposed and it is indicated that most time domain features are superfluity and redundancy.
Abstract: Feature extraction is a significant method to extract the useful information which is hidden in surface electromyography (EMG) signal and to remove the unwanted part and interferences. To be successful in classification of the EMG signal, selection of a feature vector ought to be carefully considered. However, numerous studies of the EMG signal classification have used a feature set that have contained a number of redundant features. In this study, most complete and up-to-date thirty-seven time domain and frequency domain features have been proposed to be studied their properties. The results, which were verified by scatter plot of features, statistical analysis and classifier, indicated that most time domain features are superfluity and redundancy. They can be grouped according to mathematical property and information into four main types: energy and complexity, frequency, prediction model, and time-dependence. On the other hand, all frequency domain features are calculated based on statistical parameters of EMG power spectral density. Its performance in class separability viewpoint is not suitable for EMG recognition system. Recommendation of features to avoid the usage of redundant features for classifier in EMG signal classification applications is also proposed in this study.
1,151 citations
28 Jan 2005
TL;DR: This work focuses on the development of models for Surface EMG Signal Generation based on the principles of Structure--Based SEMG models, which were developed in the context of motor control and Muscle Contraction.
Abstract: Introduction. Contributors. 1 BASIC PHYSIOLOGY AND BIOPHYSICS OF EMG SIGNAL GENERATION (T. Moritani, D. Stegeman, R. Merletti). 1.1 Introduction. 1.2 Basic Physiology of Motor Control and Muscle Contraction. 1.3 Basic Electrophysiology of the Muscle Cell Membrane. References. 2 NEEDLE AND WIRE DETECTION TECHNIQUES (J. V. Trontelj, J. Jabre, M. Mihelin). 2.1 Anatomical and Physiological Background of Intramuscular Recording. 2.2 Recording Characteristics of Needle Electrodes. 2.3 Conventional Needle EMG. 2.4 Special Needle Recording Techniques. 2.5 Physical Characteristics of Needle EMG Signals. 2.6 Recording Equipment. References. 3 DECOMPOSITION OF INTRAMUSCULAR EMG SIGNALS (D. W. Stashuk, D. Farina, K. Sogaard). 3.1 Introduction. 3.2 Basic Steps for EMG Signal Decomposition. 3.3 Evaluation of Performance of EMG Signal Decomposition Algorithms. 3.4 Applications of Results of the Decomposition of an Intramuscular EMG Signal. 3.5 Conclusions. References. 4 BIOPHYSICS OF THE GENERATION OF EMG SIGNALS (D. Farina, R. Merletti, D. F. Stegeman). 4.1 Introduction. 4.2 EMG Signal Generation. 4.3 Crosstalk. 4.4 Relationships between Surface EMG Features and Developed Force. 4.5 Conclusions. References. 5 DETECTION AND CONDITIONING OF THE SURFACE EMG SIGNAL (R. Merletti, H. Hermens). 5.1 Introduction. 5.2 Electrodes: Their Transfer Function. 5.3 Electrodes: Their Impedance, Noise, and dc Voltages. 5.4 Electrode Configuration, Distance, Location. 5.5 EMG Front--End Amplifiers. 5.6 EMG Filters: Specifications. 5.7 Sampling and A/D Conversion. 5.8 European Recommendations on Electrodes and Electrode Locations. References. 6 SINGLE--CHANNEL TECHNIQUES FOR INFORMATION EXTRACTION FROM THE SURFACE EMG SIGNAL (E. A. Clancy, D. Farina, G. Filligoi). 6.1 Introduction. 6.2 Spectral Estimation of Deterministic Signals and Stochastic Processes. 6.3 Basic Surface EMG Signal Models. 6.4 Surface EMG Amplitude Estimation. 6.5 Extraction of Information in Frequency Domain from Surface EMG Signals. 6.6 Joint Analysis of EMG Spectrum and Amplitude (JASA). 6.7 Recurrence Quantification Analysis of Surface EMG Signals. 6.8 Conclusions. References. 7 MULTI--CHANNEL TECHNIQUES FOR INFORMATION EXTRACTION FROM THE SURFACE EMG (D. Farina, R. Merletti, C. Disselhorst--Klug). 7.1 Introduction. 7.2 Spatial Filtering. 7.3 Spatial Sampling. 7.4 Estimation of Muscle--Fiber Conduction Velocity. 7.5 Conclusions. References. 8 EMG MODELING AND SIMULATION (D. F. Stegeman, R. Merletti, H. J. Hermens). 8.1 Introduction. 8.2 Phenomenological Models of EMG. 8.3 Elements of Structure--Based SEMG Models. 8.4 Basic Assumptions. 8.5 Elementary Sources of Bioelectric Muscle Activity. 8.6 Fiber Membrane Activity Profiles, Their Generation, Propagation, and Extinction. 8.7 Structure of the Motor Unit. 8.8 Volume Conduction. 8.9 Modeling EMG Detection Systems. 8.10 Modeling Motor Unit Recruitment and Firing Behavior. 8.11 Inverse Modeling. 8.12 Modeling of Muscle Fatigue. 8.13 Other Applications of Modeling. 8.14 Conclusions. References. 9 MYOELECTRIC MANIFESTATIONS OF MUSCLE FATIGUE (R. Merletti, A. Rainoldi, D. Farina). 9.1 Introduction. 9.2 Definitions and Sites of Neuromuscular Fatigue. 9.3 Assessment of Muscle Fatigue. 9.4 How Fatigue Is Reflected in Surface EMG Variables. 9.5 Myoelectric Manifestations of Muscle Fatigue in Isometric Voluntary Contractions. 9.6 Fiber Typing and Myoelectric Manifestations of Muscle Fatigue. 9.7 Factors Affecting Surface EMG Variable. 9.8 Repeatability of Estimates of EMG Variables and Fatigue Indexes. 9.9 Conclusions. References. 10 ADVANCED SIGNAL PROCESSING TECHNIQUES (D. Zazula, S. Karlsson, C. Doncarli). 10.1 Introduction. 10.2 Theoretical Background. 10.3 Decomposition of EMG Signals. 10.4 Applications to Monitoring Myoelectric Manifestations of Muscle Fatigue. 10.5 Conclusions. Acknowledgment. References. 11 SURFACE MECHANOMYOGRAM (C. Orizio). 11.1 The Mechanomyogram (MMG): General Aspects during Stimulated and Voluntary Contraction. 11.2 Detection Techniques and Sensors Comparison. 11.3 Comparison between Different Detectors. 11.4 Simulation. 11.5 MMG Versus Force: Joint and Adjunct Information Content. 11.6 MMG Versus EMG: Joint and Adjunct Information Content. 11.7 Area of Application. References. 12 SURFACE EMG APPLICATIONS IN NEUROLOGY (M. J. Zwarts, D. F. Stegeman, J. G. van Dijk). 12.1 Introduction. 12.2 Central Nervous System Disorders and SEMG. 12.3 Compound Muscle Action Potential and Motor Nerve Conduction. 12.4 CMAP Generation. 12.5 Clinical Applications. 12.6 Pathological Fatigue. 12.7 New Avenues: High--Density Multichannel Recording. 12.8 Conclusion. References. 13 APPLICATIONS IN ERGONOMICS (G. M. Hagg, B. Melin, R. Kadefors). 13.1 Historic Perspective. 13.2 Basic Workload Concepts in Ergonomics. 13.3 Basic Surface EMG Signal Processing. 13.4 Load Estimation and SEMG Normalization and Calibration. 13.5 Amplitude Data Reduction over Time. 13.6 Electromyographic Signal Alterations Indicating Muscle Fatigue in Ergonomics. 13.7 SEMG Biofeedback in Ergonomics. 13.8 Surface EMG and Musculoskeletal Disorders. 13.9 Psychological Effects on EMG. References. 14 APPLICATIONS IN EXERCISE PHYSIOLOGY (F. Felici). 14.1 Introduction. 14.2 A Few "Tips and Tricks". 14.3 Time and Frequency Domain Analysis of sEMG: What Are We Looking For? 14.4 Application of sEMG to the Study of Exercise. 14.5 Strength and Power Training. 14.6 Muscle Damage Studied by Means of sEMG. References. 15 APPLICATIONS IN MOVEMENT AND GAIT ANALYSIS (C. Frigo, R. Shiavi). 15.1 Relevance of Electromyography in Kinesiology. 15.2 Typical Acquisition Settings. 15.3 Study of Motor Control Strategies. 15.4 Investigation on the Mechanical Effect of Muscle Contraction. 15.5 Gait Analysis. 15.6 Identification of Pathophysiologic Factors. 15.7 Workload Assessment in Occupational Biomechanics. 15.8 Biofeedback. 15.9 The Linear Envelope. 15.10 Information Enhancement through Multifactorial Analysis. References. 16 APPLICATIONS IN REHABILITATION MEDICINE AND RELATED FIELDS (A. Rainoldi, R. Casale, P. Hodges, G. Jull). 16.1 Introduction. 16.2 Electromyography as a Tool in Back and Neck Pain. 16.3 EMG of the Pelvic Floor: A New Challenge in Neurological Rehabilitation. 16.4 Age--Related Effects on EMG Assessment of Muscle Physiology. 16.5 Surface EMG and Hypobaric Hipoxia. 16.6 Microgravity Effects on Neuromuscular System. References. 17 BIOFEEDBACK APPLICATIONS (J. R. Cram). 17.1 Introduction. 17.2 Biofeedback Application to Impairment Syndromes. 17.3 SEMG Biofeedback Techniques. 17.4 Summary. References. 18 CONTROL OF POWERED UPPER LIMB PROSTHESES (P. A. Parker, K. B. Englehart, B. S. Hudgins). 18.1 Introduction. 18.2 Myoelectric Signal as a Control Input. 18.3 Conventional Myoelectric Control. 18.4 Emerging MEC Strategies. 18.5 Summary. References. Index.
1,078 citations
TL;DR: Isometric muscle force and the surface electromyogram were simulated from a model that predicted recruitment and firing times in a pool of 120 motor units under different levels of excitatory drive to determine which of the modeled schemes were plausible representations of the actual organization in motor-unit pools.
Abstract: 1. Isometric muscle force and the surface electromyogram (EMG) were simulated from a model that predicted recruitment and firing times in a pool of 120 motor units under different levels of excitatory drive. The EMG-force relationships that emerged from simulations using various schedules of recruitment and rate coding were compared with those observed experimentally to determine which of the modeled schemes were plausible representations of the actual organization in motor-unit pools. 2. The model was comprised of three elements: a motoneuron model, a motor-unit force model, and a model of the surface EMG. Input to the neuron model was an excitatory drive function representing the net synaptic input to motoneurons during voluntary muscle contractions. Recruitment thresholds were assigned such that many motoneurons had low thresholds and relatively few neurons had high thresholds. Motoneuron firing rate increased as a linear function of excitatory drive between recruitment threshold and peak firing rate levels. The sequence of discharge times for each motoneuron was simulated as a random renewal process. 3. Motor-unit twitch force was estimated as an impulse response of a critically damped, second-order system. Twitch amplitudes were assigned according to rank in the recruitment order, and twitch contraction times were inversely related to twitch amplitude. Nonlinear force-firing rate behavior was simulated by varying motor-unit force gain as a function of the instantaneous firing rate and the contraction time of the unit. The total force exerted by the muscle was computed as the sum of the motor-unit forces. 4. Motor-unit action potentials were simulated on the basis of estimates of the number and location of motor-unit muscle fibers and the propagation velocity of the fiber action potentials. The number of fibers innervated by each unit was assumed to be directly proportional to the twitch force. The area of muscle encompassing unit fibers was proportional to the number of fibers innervated, and the location of motor-unit territories were randomly assigned within the muscle cross section. Action-potential propagation velocities were estimated from an inverse function of contraction time. The train of discharge times predicted from the motoneuron model determined the occurrence of each motor-unit action potential. The surface EMG was synthesized as the sum of all motor-unit action-potential trains. 5. Two recruitment conditions were tested: narrow (limit of recruitment 70% maximum excitation).(ABSTRACT TRUNCATED AT 400 WORDS)
865 citations
TL;DR: Time domain, frequency domain, time-frequency and time-scale representations, and other methods such as fractal analysis and recurrence quantification analysis are described succinctly and are illustrated with their biomechanical applications, research or clinical alike.
694 citations
TL;DR: A discrete-time wavelet transform for which the Q-factor is easily specified and the transform can be tuned according to the oscillatory behavior of the signal to which it is applied, based on a real-valued scaling factor.
Abstract: This paper describes a discrete-time wavelet transform for which the Q-factor is easily specified. Hence, the transform can be tuned according to the oscillatory behavior of the signal to which it is applied. The transform is based on a real-valued scaling factor (dilation-factor) and is implemented using a perfect reconstruction over-sampled filter bank with real-valued sampling factors. Two forms of the transform are presented. The first form is defined for discrete-time signals defined on all of Z. The second form is defined for discrete-time signals of finite-length and can be implemented efficiently with FFTs. The transform is parameterized by its Q-factor and its oversampling rate (redundancy), with modest oversampling rates (e.g., three to four times overcomplete) being sufficient for the analysis/synthesis functions to be well localized.
500 citations