Physiologically based simulation of clinical EMG signals

Evaluating Muscle Synergies With EMG Data and Physics Simulation in the Neurorobotics Platform

Although we can measure muscle activity and analyze their activation patterns, we understand little about how individual muscles affect the joint torque generated. It is known that they are controlled by circuits in the spinal cord, a system much less well-understood than the cortex. Knowing the contribution of the muscles toward a joint torque would improve our understanding of human limb control. We present a novel framework to examine the control of biomechanics using physics simulations informed by electromyography (EMG) data. These signals drive a virtual musculoskeletal model in the Neurorobotics Platform (NRP), which we then use to evaluate resulting joint torques. We use our framework to analyze raw EMG data collected during an isometric knee extension study to identify synergies that drive a musculoskeletal lower limb model. The resulting knee torques are used as a reference for genetic algorithms (GA) to generate new simulated activation patterns. On the platform the GA finds solutions that generate torques matching those observed. Possible solutions include synergies that are similar to those extracted from the human study. In addition, the GA finds activation patterns that are different from the biological ones while still producing the same knee torque. The NRP forms a highly modular integrated simulation platform allowing these in silico experiments. We argue that our framework allows for research of the neurobiomechanical control of muscles during tasks, which would otherwise not be possible.

Development of an Adaptive Artifact Subspace Reconstruction Based on Hebbian/Anti-Hebbian Learning Networks for Enhancing BCI Performance

Brain-computer interface (BCI) actively translates the brain signals into executable actions by establishing direct communication between the human brain and external devices. Recording brain activity through electroencephalography (EEG) is generally contaminated with both physiological and nonphysiological artifacts, which significantly hinders the BCI performance. Artifact subspace reconstruction (ASR) is a well-known statistical technique that automatically removes artifact components by determining the rejection threshold based on the initial reference EEG segment in multichannel EEG recordings. In real-world applications, the fixed threshold may limit the efficacy of the artifact correction, especially when the quality of the reference data is poor. This study proposes an adaptive online ASR technique by integrating the Hebbian/anti-Hebbian neural networks into the ASR algorithm, namely, principle subspace projection ASR (PSP-ASR) and principal subspace whitening ASR (PSW-ASR) that segmentwise self-organize the artifact subspace by updating the synaptic weights according to the Hebbian and anti-Hebbian learning rules. The effectiveness of the proposed algorithm is compared to the conventional ASR approaches on benchmark EEG dataset and three BCI frameworks, including steady-state visual evoked potential (SSVEP), rapid serial visual presentation (RSVP), and motor imagery (MI) by evaluating the root-mean-square error (RMSE), the signal-to-noise ratio (SNR), the Pearson correlation, and classification accuracy. The results demonstrated that the PSW-ASR algorithm effectively removed the EEG artifacts and retained the activity-specific brain signals compared to the PSP-ASR, standard ASR (Init-ASR), and moving-window ASR (MW-ASR) methods, thereby enhancing the SSVEP, RSVP, and MI BCI performances. Finally, our empirical results from the PSW-ASR algorithm suggested the choice of an aggressive cutoff range of c = 1-10 for activity-specific BCI applications and a moderate range of for the benchmark dataset and general BCI applications.

Analytical Modelling of Surface EMG Signals Generated by Curvilinear Fibers with Approximate Conductivity Tensor

OBJECTIVE

Mathematical modelling of surface electromyographic (EMG) signals has been proven a valuable tool to interpret experimental data and to validate signal processing techniques. Most analytical EMG models only consider muscle fibers with specific and fixed arrangements. However, the fiber orientation and curvature may change along the fiber paths and may differ from fiber to fiber. Here we propose a subject-specific EMG model that simulates the fiber trajectories in muscles of the upper arm and analytically derives the action potentials assuming an approximate conductivity tensor.

METHODS

Magnetic Resonance (MR) images were acquired to identify and generate muscle fiber paths and to determine the muscle locations in a cylindrical volume conductor. While the propagation of the action potentials followed the identified curvilinear fiber paths, the conductivity tensor was not adapted to the fiber direction but approximated along the longitudinal axis of the cylindrical volume conductor. Single fiber action potentials (SFAPs) were computed by simulating the generation, propagation, and extinction of membrane current sources. To validate the assumption of the approximate conductivity tensor, two numerical models were implemented for comparison with the analytical solution. The first numerical model reproduced the analytical model and therefore included an approximation for the conductivity tensor. The second numerical model included the exact conductivity tensor derived from the fiber curvatures.

RESULTS

The motor unit action potentials generated by the proposed analytical model and the two numerical models were highly similar (cross-correlation >0.98, normalized root mean square error, nRMSE 0.04, relative error in the median frequency of the simulated waveforms of approximately 3%). The proposed analytical model was also evaluated by comparing simulated and experimentally recorded compound muscle action potentials (CMAPs). The CMAPs simulated with the proposed model better matched the experimental data (cross-correlation >0.90 and nRMSE <0.25 for the majority of the channels) than a model with straight fibers. Finally, the proposed model was representatively used to test the accuracy of an EMG decomposition algorithm, providing a realistic benchmark.

CONCLUSIONS AND SIGNIFICANCE

The proposed analytical model generates action potentials that reflect the spatial distributions of muscle fibers with curvilinear paths. The simulated signals are more realistic than signals generated by analytical models with straight fibers and can therefore be applied for testing EMG processing algorithms with a trade-off between simulation accuracy and computational speed.

Gait analysis in neurological populations: Progression in the use of wearables

Gait assessment is an essential tool for clinical applications not only to diagnose different neurological conditions but also to monitor disease progression as it contributes to the understanding of underlying deficits. There are established methods and models for data collection and interpretation of gait assessment within different pathologies. This narrative review aims to depict the evolution of gait assessment from observation and rating scales to wearable sensors and laboratory technologies and provide limitations and possible future directions in the field of gait assessment. In this context, we first present an extensive review of current clinical outcomes and gait models. Then, we demonstrate commercially available wearable technologies with their technical capabilities along with their use in gait assessment studies for various neurological conditions. In the next sections, a descriptive knowledge for existing inertial and EMG based algorithms and a sign based guide that shows the outcomes of previous neurological gait assessment studies are presented. Finally, we state a discussion for the use of wearables in gait assessment and speculate the possible research directions by revealing the limitations and knowledge gaps in the literature.

Feasibility of Data-driven EMG Signal Generation using a Deep Generative Model

Despite recent advancements in the field of pattern recognition-based myoelectric control, the collection of a high quality training set remains a challenge limiting its adoption. This paper proposes a framework for a possible solution by augmenting short training protocols with subject-specific synthetic electromyography (EMG) data generated using a deep generative network, known as SinGAN. The aim of this work is to produce high quality synthetic data that could improve classification accuracy when combined with a limited training protocol. SinGAN was used to generate 1000 synthetic windows of EMG data from a single window of six different motions, and results were evaluated qualitatively, quantitatively, and in a classification task. Qualitative assessment of synthetic data was conducted via visual inspection of principal component analysis projections of real and synthetic feature space. Quantitative assessment of synthetic data revealed 11 of 32 synthetic features had similar location and scale to real features (using univariate two-sample Lepage tests); whereas multivariate distributions were found to be statistically different (p <0.05). Finally, the addition of these synthetic data to a brief training set of real data significantly improved classification accuracy in a cross-validation testing scheme by 5.4% (p <0.001).

Parkinson's Disease EMG Data Augmentation and Simulation with DCGANs and Style Transfer

This paper proposes two new data augmentation approaches based on Deep Convolutional Generative Adversarial Networks (DCGANs) and Style Transfer for augmenting Parkinson’s Disease (PD) electromyography (EMG) signals. The experimental results indicate that the proposed models can adapt to different frequencies and amplitudes of tremor, simulating each patient’s tremor patterns and extending them to different sets of movement protocols. Therefore, one could use these models for extending the existing patient dataset and generating tremor simulations for validating treatment approaches on different movement scenarios.

Can Wavelet Denoising Improve Motor Unit Potential Template Estimation?

Background:

Electromyographic (EMG) signals obtained from a contracted muscle contain valuable information on its activity and health status. Much of this information lies in motor unit potentials (MUPs) of its motor units (MUs), collected during the muscle contraction. Hence, accurate estimation of a MUP template for each MU is crucial.

Objective:

To investigate the possibility of improving MUP template estimation using the wavelet denoising technique.

Material and Methods:

In this analytical study, several MUP template estimators were developed by combining conventional estimation methods and wavelet denoising techniques. A MUP template was initially estimated using conventional methods such as mean, median, median-trimmed mean, or mode. Thereafter, it was post-processed using the wavelet denoising technique.

Results:

Evaluation results of the studied estimators using 40 simulated EMG signals with a true template for each constituent MUP train showed that augmented wavelet- based template estimation methods are more reliable than conventional methods. However, on average, wavelet denoising was not much effective. Around 40 MUPs of a MU is sufficient to estimate its MUP template.

Conclusions:

Although wavelet techniques are effective in EMG signal analysis, here wavelet denoising did not practically improve MUP template estimation. Considering computational simplicity and estimation error, the two methods median and median-trimmed mean are practical estimators that can provide a good estimation of a MUP template for a MU when approximately 40 MUPs are available. Nevertheless, the baseline noise level in the MUP templates estimated using the median-trimmed mean method is slightly lower than that in the templates estimated using the median method.

Improved MUP Template Estimation Using Local Time Warping and Kernel Weighted Averaging

A motor unit potential (MUP) template, which represents the shapes of the MUPs within a MUP train, provides information related to the morphology and physiology of the sampled motor unit. This work presents an improved MUP template estimation technique that uses local time warping and kernel weighted ensemble averaging. An analysis of the algorithm, and a description of the improvements compared with spike triggered averaging is given. MUP template estimates were evaluated using simulated EMG signals with a known gold standard template for each motor unit potential train. Statistically significant reduction in template estimation error is shown, both within the baseline and duration portions of a MUP.

Decomposing single-channel intramuscular electromyography signal sampled at a low frequency into its motor unit action potential trains with a generative adversarial network

Conventional methods decompose single-channel intramuscular electromyography (iEMG) signals into their constituent motor unit action potential trains (MUAPTs) by detecting and clustering individual motor unit action potentials (MUAPs). However, these methods are not applicable for iEMG signals recorded by electrodes with a large sensing areas or iEMG signals sampled at a low frequency, in which detecting and clustering individual MUAPs are difficult due to superimpositions of the MUAPs and the loss of MUAP morphological characteristics. In this study, we propose an approach based on a generative adversarial network to decompose iEMG signals, which does not depend on detecting and clustering individual MUAPs from the iEMG signal. The proposed approach decomposes the iEMG signal into its MUAPTs based on Bayes' law and a Wasserstein generative adversarial network with gradient penalty (WGAN-GP). MUAPTs generated by the WGAN-GP were used to decompose the iEMG signal to maximize the posterior probability of the generated MUAPTs given the iEMG signal. The accuracy of the proposed approach is analysed directly by decomposing the simulated iEMG signal with seven gold-standard motor units. The results showed that the proposed approach achieved a 53% accuracy in capturing the firing regularities of the MUs, while the conventional method achieved a 37% accuracy on the same task.

Cross Comparison of Motor Unit Potential Features Used in EMG Signal Decomposition

Feature extraction is an important step of resolving an electromyographic (EMG) signal into its component motor unit potential trains, commonly known as EMG decomposition. Until now, different features have been used to represent motor unit potentials (MUPs) and improve decomposition processing time and accuracy, but a major limitation is that no systematic comparison of these features exists. In an EMG decomposition system, like any pattern recognition system, the features used for representing MUPs play an important role in the overall performance of the system. A cross comparison of the feature extraction methods used in EMG signal decomposition can assist in choosing the best features for representing MUPs and ultimately may improve EMG decomposition results. This paper presents a survey and cross comparison of these feature extraction methods. Decomposability index, classification accuracy of a -nearest neighbors classifier, and class-feature mutual information were employed for evaluating the discriminative power of various feature extraction techniques commonly used in the literature including time domain, morphological, frequency domain, and discrete wavelets. In terms of data, 45 simulated and 82 real EMG signals were used. Results showed that among time domain features, the first derivative of time samples exhibit the best separability. For morphological features, slope analysis provided the most discriminative power. Discrete Fourier transform coefficients offered the best separability among frequency domain features. However, neither morphological nor frequency domain techniques outperformed time domain features. The detail 4 coefficients in a discrete wavelets decomposition exceeded in evaluation measures when compared with other feature extraction techniques. Using principal component analysis slightly improved the results, but it is time consuming. Overall, considering computation time and discriminative ability, the first derivative of time samples might be efficient in representing MUPs in EMG decomposition and there is no need for sophisticated feature extraction methods.

Comparative Analysis of Wavelet-based Feature Extraction for Intramuscular EMG Signal Decomposition

BACKGROUND

Electromyographic (EMG) signal decomposition is the process by which an EMG signal is decomposed into its constituent motor unit potential trains (MUPTs). A major step in EMG decomposition is feature extraction in which each detected motor unit potential (MUP) is represented by a feature vector. As with any other pattern recognition system, feature extraction has a significant impact on the performance of a decomposition system. EMG decomposition has been studied well and several systems were proposed, but feature extraction step has not been investigated in detail.

OBJECTIVE

Several EMG signals were generated using a physiologically-based EMG signal simulation algorithm. For each signal, the firing patterns of motor units (MUs) provided by the simulator were used to extract MUPs of each MU. For feature extraction, different wavelet families including Daubechies (db), Symlets, Coiflets, bi-orthogonal, reverse bi-orthogonal and discrete Meyer were investigated. Moreover, the possibility of reducing the dimensionality of MUP feature vector is explored in this work. The MUPs represented using wavelet-domain features are transformed into a new coordinate system using Principal Component Analysis (PCA). The features were evaluated regarding their capability in discriminating MUPs of individual MUs.

RESULTS

Extensive studies on different mother wavelet functions revealed that db2, coif1, sym5, bior2.2, bior4.4, and rbior2.2 are the best ones in differentiating MUPs of different MUs. The best results were achieved at the 4th detail coefficient. Overall, rbior2.2 outperformed all wavelet functions studied; nevertheless for EMG signals composed of more than 12 MUPTs, syms5 wavelet function is the best function. Applying PCA slightly enhanced the results.

Aggregate features in multisample classification problems

This paper evaluates the classification of multisample problems, such as electromyographic (EMG) data, by making aggregate features available to a per-sample classifier. It is found that the accuracy of this approach is superior to that of traditional methods such as majority vote for this problem. The classification improvements of this method, in conjunction with a confidence measure expressing the per-sample probability of classification failure (i.e., a hazard function) is described and measured. Results are expected to be of interest in clinical decision support system development.

Detecting neuropathy using measures of motor unit activation extracted from standard concentric needle electromyographic signals

OBJECTIVE

Motor unit loss associated with neuropathic disorders affects motor unit activation. Quantitative electromyographic (EMG) features of motor unit activation estimated from the sequences of motor unit potentials (MUPs) created by concurrently active motor units can support the detection of neuropathic disorders. Interpretation of most motor unit activation feature values are, however, confounded by uncertainty regarding the level of muscle activation during EMG signal detection. A set of new features circumventing these limitations are proposed, and their utility in detecting neuropathy is investigated using simulated and clinical EMG signals.

METHODS

The firing sequence of a motor neuron was simulated using a compartmentalized Hodgkin-Huxley based model. A pool of motor neurons was modelled such that each motor neuron was subjected to a common level of activation. The detection of the firing sequence of a motor neuron using a clinically detected EMG signal was simulated using a model of muscle anatomy combined with a model representing muscle fiber electrophysiology and the voltage detection properties of a concentric needle electrode.

SIGNIFICANCE

Findings are based on simulated EMG data representing 30 normal and 30 neuropathic muscles as well as clinical EMG data collected from the tibialis anterior muscle of 48 control subjects and 30 subjects with neuropathic disorders. These results demonstrate the possibility of detecting neuropathy using motor unit recruitment and mean firing rate feature values estimated from standard concentric needle detected EMG signals.

EMG signal decomposition using motor unit potential train validity

A system to resolve an intramuscular electromyographic (EMG) signal into its component motor unit potential trains (MUPTs) is presented. The system is intended mainly for clinical applications where several physiological parameters of motor units (MUs), such as their motor unit potential (MUP) templates and mean firing rates, are of interest. The system filters an EMG signal, detects MUPs, and clusters and classifies the detected MUPs into MUPTs. Clustering is partially based on the K-means algorithm, and the supervised classification is implemented using a certainty-based algorithm. Both clustering and supervised classification algorithms use MUP shape and MU firing pattern information along with signal dependent assignment criteria to obtain robust performance across a variety of EMG signals. During classification, the validity of extracted MUPTs are determined using several supervised classifiers; invalid trains are corrected and the assignment threshold for each train is adjusted based on the estimated validity (i.e., adaptive classification). Performance of the developed system in terms of accuracy (A(c)), assignment rate (A(r)), correct classification rate (CC(r)) , and the error in estimating the number of MUPTs represented in the set of detected MUPs (E(NMUPTs)) was evaluated using 32 simulated and 30 real EMG signals comprised of 3-11 and 3-15 MUPTs, respectively. The developed system, with average CC(r) of 86.4% for simulated and 96.4% for real data, outperformed a previously developed EMG decomposition system, with average CC(r) of 71.6% and 89.7% for simulated and real data, by 14.7% and 6.7%, respectively. In terms of E(NMUPTs), the new system, with average E(NMUPTs) of 0.3 and 0.2 for simulated and real data respectively, was better able to estimate the number of MUPTs represented in a set of detected MUPs than the previous system, with average E(NMUPTs) of 2.2 and 0.8 for simulated and real data respectively. For both the simulated and real data used, variations in A(c), A(r), and E(NMUPTs) for the newly developed system were lower than for the previous system, which demonstrates that the new system can successfully adjust the assignment criteria based on the characteristics of a given signal to achieve robust performance across a wide variety of EMG signals, which is of paramount importance for successfully promoting the clinical application of EMG signal decomposition techniques.

Augmenting the decomposition of EMG signals using supervised feature extraction techniques

Electromyographic (EMG) signal decomposition is the process of resolving an EMG signal into its constituent motor unit potential trains (MUPTs). In this work, the possibility of improving the decomposing results using two supervised feature extraction methods, i.e., Fisher discriminant analysis (FDA) and supervised principal component analysis (SPCA), is explored. Using the MUP labels provided by a decomposition-based quantitative EMG system as a training data for FDA and SPCA, the MUPs are transformed into a new feature space such that the MUPs of a single MU become as close as possible to each other while those created by different MUs become as far as possible. The MUPs are then reclassified using a certainty-based classification algorithm. Evaluation results using 10 simulated EMG signals comprised of 3-11 MUPTs demonstrate that FDA and SPCA on average improve the decomposition accuracy by 6%. The improvement for the most difficult-to-decompose signal is about 12%, which shows the proposed approach is most beneficial in the decomposition of more complex signals.

Common errors in textbook descriptions of muscle fiber size in nontrained humans

Exercise science and human anatomy and physiology textbooks commonly report that type IIB muscle fibers have the largest cross-sectional area of the three fiber types. These descriptions of muscle fiber sizes do not match with the research literature examining muscle fibers in young adult nontrained humans. For men, most commonly type IIA fibers were significantly larger than other fiber types (six out of 10 cases across six different muscles). For women, either type I, or both I and IIA muscle fibers were usually significantly the largest (five out of six cases across four different muscles). In none of these reports were type IIB fibers significantly larger than both other fiber types. In 27 studies that did not include statistical comparisons of mean fiber sizes across fiber types, in no cases were type IIB or fast glycolytic fibers larger than both type I and IIA, or slow oxidative and fast oxidative glycolytic fibers. The likely reason for mistakes in textbook descriptions of human muscle fiber sizes is that animal data were presented without being labeled as such, and without any warning that there are interspecies differences in muscle fiber properties. Correct knowledge of muscle fiber sizes may facilitate interpreting training and aging adaptations.

Quantification of motor unit action potential energy

OBJECTIVE

Motor unit action potentials (MUAPs) recorded by needle electrode reflect the functional state of the motor unit and its force-generating capacity, and are usually described morphologically (e.g. amplitude, duration). However, since the purpose of motor unit activation is force generation, MUAP energy seems a more physically meaningful measurement.

METHODS

MUAPs were obtained by multi-MUAP decomposition of real interference patterns taken from human patients with neurological diseases. The energy content of each MUAP was measured from a time-frequency representation (TFR), specifically the Choi-Williams distribution, and compared with the standard MUAP morphological measure, the Size Index. The sample included normal, neurogenic, and myopathic MUAPs, from 11 patients.

RESULTS

There is an exponential distribution of energy within a sample of MUAPs and a strong exponential relationship between the Size Index and MUAP energy was observed.

CONCLUSIONS

The energy content of a MUAP can be quantified and corresponds very well with the current quantitative standard. Energy is a possible addition to MUAP quantification.

SIGNIFICANCE

MUAPs could be classified as having normal, large (neurogenic), or low (myopathic) energy. MUAP energy has direct physical and physiological meaning that reflects the force-generating capacity of the motor unit. Time-frequency analysis could also be used to study the specific frequency content of MUAPs and the energy of MUAPs within an interference pattern, without the need for decomposition.

Application of time-varying analysis to diagnostic needle electromyography

Quantification in clinical, diagnostic electromyography (EMG) currently includes motor unit action potential (MUAP) analysis and interference pattern analysis. Early efforts to examine the frequency/power spectra of the interference pattern showed modest value but the technique was not developed further. This paper re-examines spectral analysis, extending it into the time-varying domain, which has never been studied in diagnostic needle EMG. Time-frequency and time-scale analysis employing wavelet and non-wavelet techniques were applied to short trains of MUAPs. The results show that time-varying analysis produces clear visual representations of the energy content of individual MUAPs within an interference pattern. The time frequency representations allow easy, qualitative distinction between normal and neurogenic MUAPs. Furthermore, the quantified MUAP energy correlates well with the current morphological standard and the quantification process is substantially faster. Time-varying analysis links classical power spectral analysis in the realm of interference patterns with quantitative MUAP analysis. In addition to morphological classification, MUAPs might also be classified by energy content, which more closely reflects the physical and physiological consequences of neuromuscular pathology on the motor unit.

Adaptive motor unit potential train validation using MUP shape information

A decomposed electromyographic (EMG) signal provides information that can be used clinically or for physiological investigation. However, in all instances the validity of the extracted motor unit potential trains (MUPTs) must first be determined because, as with all pattern recognition applications, errors will occur during decomposition. Moreover, detecting invalid MUPTs during EMG signal decomposition can enhance decompositions results. Eight methods to validate an extracted MUPT using its motor unit potential (MUP) shape information were studied. These MUPT validation methods are based on existing cluster analysis algorithms, four were newly developed adaptive methods and four were classical cluster validation methods. The methods evaluate the shapes of the MUPs of a MUPT to determine whether the MUPT represents the activity of a single motor unit (i.e. it is a valid MUPT) or not. Evaluation results using both simulated and real data show that the newly developed adaptive methods are sufficiently fast and accurate to be used during or after the decomposition of EMG signals. The adaptive gap-based Duda and Hart (AGDH) method had significantly better accuracies in correctly categorizing the MUPTs extracted during decomposition (91.3% and 94.7% for simulated and real data, respectively; assuming 12.7% of the extracted MUPTs are on average invalid). The accuracy with which invalid MUPTs can be detected is dependent on the similarity of the MUP templates of the MUPTs merged to create the invalid train and suggests the need, in some cases, for the combined use of motor unit firing pattern and MUP shape information.

A method for detecting and editing MUPTs contaminated by false classification errors during EMG signal decomposition

A robust method for detecting motor unit potential trains (MUPTs) contaminated with false classification errors (FCEs) during EMG signal decomposition and then removing the FCEs from a contaminated train is presented. Using motor unit (MU) firing pattern information provided by each MUPT, the developed algorithm first determines whether a given train is contaminated by high number of FCEs and needs to be edited. For contaminated MUPTs, the method uses both MU firing pattern and motor unit potential (MUP) shape information to detect MUPs that were erroneously assigned to the train (i.e., represent FCEs). For the simulated data used in this study contaminated MUPTs could be detected with 88.7% accuracy. For a given contaminated MUPT, the algorithm on average correctly detected 83.4% of the FCEs and left 93.4% of the correctly assigned MUPs. The accuracy of the MUPs classified to a MUPT was estimated to be 92.1% on average.

Evaluation of motor unit placement algorithms for EMG simulation

Motor unit layout algorithms have a significant effect on motor unit fibre densities recorded. Motor unit fibre densities are affected by both the method used to place the motor unit territories, and the mechanism by which muscle fibres are assigned to motor units. The first of these should emulate the process by which separate motor neurons create overlapping territories that cover the muscle cross section, while the second should have some relation to the processes involved with axonal arborization and development of the spatial dispersion of the neuro-muscular junctions. The success of an algorithm in creating physiologically realistic motor unit layouts may be evaluated, in part, by examining the distribution of the muscle fibres assigned to the motor units. This paper examines the motor unit fibre densities found in muscles created by two recent algorithms, and explores the degree to which the concepts used by these algorithms may be shared.

Multiplicative multi-fractal modeling of electromyography signals for discerning neuropathic conditions

In this paper, we present a new method for multi-scale analysis of electromyography signals based on an interesting fractal process known as multiplicative cascade multi-fractal. Using simulated needle electromyography signals, we show this method provides a means for discrimination of normal and neuropathic electromyography signals. We also present experimental results that show the new parameters, computed using multiplicative cascade multi-fractal modeling, are more robust than the conventional signal parameter, number of turns, in the presence of additive noise. Results of multiplicative cascade multi-fractal modeling are consistent with other multi-scale approaches; advantages and differences are high lighted.

A muscle architecture model offering control over motor unit fiber density distributions

The aim of this study was to develop a muscle architecture model able to account for the observed distributions of innervation ratios and fiber densities of different types of motor units in a muscle. A model algorithm is proposed and mathematically analyzed in order to obtain an inverse procedure that allows, by modification of input parameters, control over the output distributions of motor unit fiber densities. The model's performance was tested with independent data from a glycogen depletion study of the medial gastrocnemius of the rat. Results show that the model accurately reproduces the observed physiological distributions of innervation ratios and fiber densities and their relationships. The reliability and accuracy of the new muscle architecture model developed here can provide more accurate models for the simulation of different electromyographic signals.

MUP shape-based validation of a motor unit potential train

A method using the gap statistic is proposed to evaluate the validity of a motor unit potential train (MUPT) in terms of motor unit potential (MUP) shape consistency. This algorithm determines whether the MUPs of a given MUPT are homogeneous in terms of their shapes or not. It also checks if there are gaps in the inter-discharge interval (IDI) train of the given MUPT. If the MUPs are not homogeneous or if there is a temporal gap in the MUPT, the given MUPT is split into valid trains. To overcome MUP shape variability caused by jitter or needle movement during signal detection, similar MUPTs are merged if the resulting merged train is a valid train. Experimental results using simulated EMG signals show that the accuracy of the developed method in determining valid MUPTs and invalid MUPTs correctly is 97.58% and 99.33% on average, respectively. This performance encourages the use of this method for automated validation of MUPTs.

Design of Adaptive Filter Using Jordan/Elman Neural Network in a Typical EMG Signal Noise Removal

The bioelectric potentials associated with muscle activity constitute the electromyogram (EMG). These EMG signals are low-frequency and lower-magnitude signals. In this paper, it is presented that Jordan/Elman neural network can be effectively used for EMG signal noise removal, which is a typical nonlinear multivariable regression problem, as compared with other types of neural networks. Different neural network (NN) models with varying parameters were considered for the design of adaptive neural-network-based filter which is a typical SISO system. The performance parameters, that is, MSE, correlation coefficient, N/P, and t, are found to be in the expected range of values.

A comparison of three quantitative motor unit analysis algorithms

This study assessed the accuracy of three automatic motor unit analysis algorithms--multi-motor unit analysis, decomposition quantitative EMG, and EMGtools--on a set of real EMG signals whose true composition was determined by manual decomposition. All three algorithms correctly identified all the MUs in signals with up to 5 active MUs, and most of the MUs in signals with up to 10 active MUs. The algorithms accurately estimated MUAP amplitudes and firing rates, but they estimated duration less accurately because of baseline noise. These findings support the validity and utility of these algorithms.

Decision support for QEMG

For clinicians to use quantitative electromyography (QEMG) to help determine the presence or absence of neuromuscular disease, they must manually interpret an exhaustive set of motor unit potential (MUP) or interference pattern statistics to formulate a clinically useful muscle characterization. A new method is presented for automatically categorizing a set of quantitative electromyographic (EMG) data as characteristic of data acquired from a muscle affected by a myopathic, normal or neuropathic disease process, based on discovering patterns of MUP feature values. From their numbers of occurrence in a set of training data, representative of each muscle category, discovered patterns of MUP feature values are expressed as conditional probabilities of detecting such MUPs in each category of muscle. The conditional probabilities of each MUP in a set of MUPs acquired from an examined muscle are combined using Bayes' rule to estimate conditional probabilities of the examined muscle being of each category type. Using simulated and clinical data, the ability of a "pattern discovery" based Bayesian (PD-based Bayesian) method to correctly categorize sets of test MUP data was compared to conventional methods which use data means and outliers. The simulated data were created by modeling the effects of myopathic and neuropathic diseases using a physiologically based EMG signal simulator. The clinical data was from controls and patients with known neuropathic disorders. PD-based Bayesian muscle characterization had an accuracy of 84.4% compared to 51.9% for the means and outlier based method when using all MUP features considered. PD-based Bayesian methods can accurately characterize a muscle. PD-based Bayesian muscle characterization automatically maximizes both sensitivity and specificity and provides transparent rationalizations for its characterizations. This leads to the expectation that clinicians using PD-based Bayesian muscle characterization will be provided with improved decision support compared to that provided by the status quo means and outlier based methods.

Motor unit potential characterization using “pattern discovery”

Typically in clinical practice, electromyographers use qualitative auditory and visual analysis of electromyographic (EMG) signals to help infer if a neuromuscular disorder is present and if it is neuropathic or myopathic. Quantitative EMG methods exist that can more accurately measure feature values but require qualitative interpretation of a large number of statistics. Electrophysiological characterization of a neuromuscular system can be improved through the quantitative interpretation of EMG statistics. The aim of the present study was to compare the accuracy of pattern discovery (PD) characterization of motor unit potentials (MUPs) to other classifiers commonly used in the medical field. In addition, a demonstration of PD's transparency is provided. The transparency of PD characterization is a result of observing statistically significant events known as patterns. Using clinical MUP data from normal subjects and patients with known neuropathic disorders, PD achieved an error rate of 30.3% versus 29.8% for a Naïve Bayes classifier, 30.1% for a Decision Tree and 29% for discriminant analysis. Similar results were found for simulated EMG data. PD characterization succeeded in interpreting the information extracted from MUPs and transforming it into knowledge that is consistent with the literature and that can be valuable for the capture and transparent expression of clinically useful knowledge.

Simulation analysis of interference EMG during fatiguing voluntary contractions. Part I: What do the intramuscular spike amplitude–frequency histograms reflect?

Decline in amplitude of EMG signals and in the rate of counts of intramuscularly recorded spikes during fatigue is often attributed to a progressive reduction of the neural drive only. As a rule, alterations in intracellular action potential (IAP) are not taken into account. To test correctness of the hypothesis, the effect of various discharge frequency patterns as well as changes in IAP shape and muscle fibre propagation velocity (MFPV) on the spike amplitude-frequency histogram of intramuscular interference EMG signals were simulated and analyzed. It was assumed that muscle was composed of four types of motor units (MUs): slow-twitch fatigue resistant, fast-twitch fatigue resistant, fast intermediate, and fast fatigable. MFPV and IAP duration at initial stage before fatigue as well as their changes differed for individual MU types. Fatigability of individual MU types in normal conditions as well as in the case of ischaemic or low oxygen conditions due to restricted blood flow was also taken into account. It was found that spike amplitude-frequency histogram is poorly sensitive to MU firing frequency, while it is highly sensitive to IAP profile lengthening. It is concluded that spike amplitude-frequency analysis can hardly provide a correct measure of MU rate-coding pattern during fatigue.

Using motor unit potential characterizations to estimate neuromuscular disorder level of involvement

Based on the analysis of electromyographic (EMG) data muscles are often characterized as normal or affected by a neuromuscular disorder. Motor unit potential (MUP) characterizations comprised of the conditional probabilities of a MUP being detected from a muscle of each of the following categories: myopathic, normal, and neuropathic, were estimated. The sets of MUP characterizations of a set of MUPs detected in a muscle were averaged to produce a set of muscle characterization measures related to the probability of the muscle belonging to each category conditioned on the set of MUPs detected. Using simulated EMG signals, the objective of this work was to evaluate the correlation between the muscle characterization measures produced by different MUP characterization methods and the level of involvement of a disorder. The results showed a correlation of 0.9 between myopathic and neuropathic muscle characterization measures and the actual level of involvement when using a Pattern Discovery (PD) method to estimate MUP characterizations. This work suggests that MUP characterizations can be used to assist clinicians in tracking the progress of a disease process.

A software package for interactive motor unit potential classification using fuzzy k-NN classifier

We present an interactive software package for implementing the supervised classification task during electromyographic (EMG) signal decomposition process using a fuzzy k-NN classifier and utilizing the MATLAB high-level programming language and its interactive environment. The method employs an assertion-based classification that takes into account a combination of motor unit potential (MUP) shapes and two modes of use of motor unit firing pattern information: the passive and the active modes. The developed package consists of several graphical user interfaces used to detect individual MUP waveforms from a raw EMG signal, extract relevant features, and classify the MUPs into motor unit potential trains (MUPTs) using assertion-based classifiers.

A hybrid classifier fusion approach for motor unit potential classification during EMG signal decomposition

In this paper, we propose a hybrid classifier fusion scheme for motor unit potential classification during electromyographic (EMG) signal decomposition. The scheme uses an aggregator module consisting of two stages of classifier fusion: the first at the abstract level using class labels and the second at the measurement level using confidence values. Performance of the developed system was evaluated using one set of real signals and two sets of simulated signals and was compared with the performance of the constituent base classifiers and the performance of a one-stage classifier fusion approach. Across the EMG signal data sets used and relative to the performance of base classifiers, the hybrid approach had better average classification performance overall. For the set of simulated signals of varying intensity, the hybrid classifier fusion system had on average an improved correct classification rate (CCr) (6.1%) and reduced error rate (Er) (0.4%). For the set of simulated signals of varying amounts of shape and/or firing pattern variability, the hybrid classifier fusion system had on average an improved CCr (6.2%) and reduced Er (0.9%). For real signals, the hybrid classifier fusion system had on average an improved CCr (7.5%) and reduced Er (1.7%).

Effects of intramuscular needle position on motor unit action potential metrics

It is unclear whether there are clinically significant differences in amplitude, duration, and numbers of turns and phases if an electromyographic (EMG) study is performed near to, or far from, the end-plate zone. The effects of temporal dispersion of arriving muscle-fiber action potentials on quantitative motor unit action potential (MUAP) metrics were assessed in simulated and biologic muscles. Two muscle simulation models were studied with electrode recording positions near the motor end-plate zone and 50-75 mm away. When the electrode was moved away from the end-plate zone, averages of 20 MUAPs significantly decreased in amplitude and area, and increased in numbers of turns and phases, but there was no significant change in duration. In biologic muscles (both normal and pathologic), similar changes in average metrics were observed, but to lesser degrees; few were statistically significant. Zones of innervation in biologic muscles are broadly distributed and, during routine electrode studies, distances between random electrode placements and end-plate zones are therefore relatively short, leading to clinically insignificant changes in quantitative MUAP metrics with distance from the end-plate zone. Thus, electrode position within a muscle is unlikely to affect clinical MUAP interpretation.

Adaptive fuzzy k-NN classifier for EMG signal decomposition

An adaptive fuzzy k-nearest neighbour classifier (AFNNC) for EMG signal decomposition is presented and evaluated. The developed classifier uses an adaptive assertion-based classification approach for setting a minimum classification threshold. The similarity criterion used for grouping motor unit potentials (MUPs) is based on a combination of MUP shapes and two modes of use of motor unit firing pattern information: passive and active. The performance of the developed classifier was evaluated using synthetic signals with specific properties and experimental signals and compared with the performance of an adaptive template matching classifier, the adaptive certainty classifier (ACC). Across the sets of simulated and experimental EMG signals used for comparison, the AFNNC had better average classification performance overall, but due to the assignment of higher numbers of MUPs it made relatively more errors. Nonetheless, these increased error rates would still be acceptable for most clinical uses of decomposed EMG data. An independent and a related set of simulated signals were used for testing. For the independent simulated signals of varying intensity, the AFNNC had on average an improved correct classification rate (CCr) (8.1%) but an increased error rate (Er) (1.5%) compared to ACC. For the related simulated signals with varying amounts of shape and/or firing pattern variability, the AFNNC on average had an improved CCr (5%) but a slightly increased Er (0.3%) compared to ACC. For experimental signals, the AFNNC on average had improved CCr (6%) but an increased Er (2.1%) compared to ACC. The greatest gains in AFNNC performance relative to that of the ACC occurred when the variability of MUP shapes within motor unit potential trains was high suggesting that compared to a template matching assignment strategy the NN assignment paradigm is better able to ameliorate the classification problems caused by MUP instability.