Modeling nonlinear errors in surface electromyography due to baseline noise: A new methodology

The surface electromyographic (EMG) signal is often contaminated by some degree of baseline noise. It is customary for scientists to subtract baseline noise from the measured EMG signal prior to further analyses based on the assumption that baseline noise adds linearly to the observed EMG signal. The stochastic nature of both the baseline and EMG signal, however, may invalidate this assumption. Alternately, “true” EMG signals may be either minimally or nonlinearly affected by baseline noise. This information is particularly relevant at low contraction intensities when signal-to-noise ratios (SNR) may be lowest. Thus, the purpose of this simulation study was to investigate the influence of varying levels of baseline noise (approximately 2–40% maximum EMG amplitude) on mean EMG burst amplitude and to assess the best means to account for signal noise. The simulations indicated baseline noise had minimal effects on mean EMG activity for maximum contractions, but increased nonlinearly with increasing noise levels and decreasing signal amplitudes. Thus, the simple baseline noise subtraction resulted in substantial error when estimating mean activity during low intensity EMG bursts. Conversely, correcting EMG signal as a nonlinear function of both baseline and measured signal amplitude provided highly accurate estimates of EMG amplitude. This novel nonlinear error modeling approach has potential implications for EMG signal processing, particularly when assessing co-activation of antagonist muscles or small amplitude contractions where the SNR can be low.     

On the errors in assessment of severity of involuntary movements using surface EMG.

Surface electromyogram (SEMG) is a useful tool to depict involuntary movements, but evaluation of the intensity of such movements with SEMG over multiple recording instances requires awareness of factors influencing quantified SEMG signals. We investigated the differences in the amplitude of SEMGs due to electrode displacement in isometric voluntary contraction in the upper arm, forearm and lower leg in 8 healthy men. The SEMGs of gross muscle activity simultaneously recorded with 4 electrode pairs from the agonist and antagonist sides in 3 displacement conditions with respect to parallel position, interelectrode distance, and rotation were compared. The amount of EMG integration (equivalent to the average SEMG amplitude) of each electrode pair was compared to the reference electrode pair with interelectrode distance of 40 mm placed on the center of the tested muscles. The average EMG difference ratios ranged 1.1-2.2%/mm in parallel shift, 1.0-1.9%/mm in distance shift, and 0.3-0.6%/degree in rotation shift. Displacement error of electrodes in separate recording instances should be reduced using anatomical landmarks, when SEMG is applied as a quantitative method to evaluate change in the states of involuntary movements.     

Rectification of SEMG as a tool to demonstrate synchronous motor unit activity during vibration

The use of surface electromyography (SEMG) in vibration studies is problematic since motion artifacts occupy the same frequency band with the SEMG signal containing information on synchronous motor unit activity. We hypothesize that using a harsher, 80–500 Hz band-pass filter and using rectification can help eliminate motion artifacts and provide a way to observe synchronous motor unit activity that is phase locked to vibration using SEMG recordings only. Multi Motor Unit (MMU) action potentials using intramuscular electrodes along with SEMG were recorded from the gastrocnemius medialis (GM) of six healthy male volunteers. Data were collected during whole body vibration, using vibration frequencies of 30 Hz, 35 Hz, 40 Hz or 50 Hz. A computer simulation was used to investigate the efficacy of filtering under different scenarios: with or without artifacts and/or motor unit synchronization. Our findings indicate that motor unit synchronization took place during WBV as verified by MMU recordings. A harsh filtering regimen along with rectification proved successful in demonstrating motor unit synchronization in SEMG recordings. Our findings were further supported by the results from the computer simulation, which indicated that filtering and rectification was efficient in discriminating motion artifacts from motor unit synchronization. We suggest that the proposed signal processing technique may provide a new methodology to evaluate the effects of vibration treatments using only SEMG. This is a major advantage, as this non-intrusive method is able to overcome movement artifacts and also indicate the synchronization of underlying motor units.     

Surface electromyography as a tool to assess the responses of car passengers to lateral accelerations. Part II: Objective comparison of vehicles

The purpose of this study was to objectively assess the response of car passengers to lateral accelerations. Surface EMG signals were collected bilaterally from the cervical erector spinae (CES), latissimus dorsi (LD), erector spinae (ES), external oblique (EO), and vastus lateralis (VL) muscles of 10 subjects. Lateral acceleration was also recorded. Three chassis-seat configurations AA, BA and BB were tested, with the first letter denoting the chassis and the second the seat. SEMG signals were often contaminated by noise, and were, therefore, denoised using the methods explained in part I. Reciprocal phasic activity was observed for all muscles except for the EO, and the reaction of passengers to lateral accelerations was interpreted as a bust torsion. The RMS of EMG segments was used as an indication of muscle activity. Muscle activation of VL and ES were significantly affected by the configuration tested (p < 0.05), with greater activation levels observed for the chassis A than for the chassis B. Such a finding implies that greater roll requires greater muscle activity, thus resulting in less comfortable vehicles. Therefore, SEMG can be used to provide an objective measure of discomfort in passengers subjected to lateral accelerations in a car seat.     

Agonist-antagonist common drive during fatiguing knee extension efforts using surface electromyography.

AIM: This study examined the electromyographic (EMG) activity of knee extensor agonists and a knee extensor antagonist muscle during fatiguing isometric extensions across a range of force levels. METHODS: Five female subjects performed isometric knee extensions at 25%, 50%, 75% and 100% of their maximal voluntary contraction (MVC) with the knee flexed to 75 degrees. Surface EMG (SEMG) was recorded with bipolar electrodes from the vastus lateralis (VL), vastus medialis (VM), rectus femoris (RF) and biceps femoris (BF) and the root-mean-squared (RMS) amplitude and the percentage frequency compression of these recordings were calculated. Commonality and cross talk between recordings were also examined. RESULTS: Cross talk between recordings was deemed negligible despite significant levels of commonality between the agonist and antagonist SEMG, which was attributed to common drive. SEMG RMS amplitude increased significantly for all muscles during the 25%, 50%, 75% MVC knee extensions until task failure, and decreased significantly for 100% MVC. The frequency spectrum of the SEMG compressed significantly for all muscles and % MVC levels. The VM, VL and BF SEMG recordings responded similarly to fatigue. The RF's frequency spectrum compressed to a significantly higher degree. CONCLUSIONS: The VM, VL, RF, and BF fatigue in parallel, with high similarity between VM, VL and BF, giving support to the concept of a shared agonist-antagonist motoneuron pool.     

Methodologies for evaluating electromyographic field data in ergonomics.

Surface electromyography (SEMG) is an important tool for work load assessment in ergonomics. Several different approaches using amplitude as well as frequency parameters give fruitful information depending on question at issue in the laboratory as well as in field studies. One basic factor determining the choice of analysis method is whether the SEMG is interpreted as an indicator of forces/torques or pure muscular activation. Two methods for occupational SEMG data reduction representing two different approaches to SEMG applications in ergonomics, Exposure Variation Analysis (EVA), and Joint Analysis of EMG Spectrum and Amplitude (JASA), applied on the same SEMG recording from three muscles during urology surgeon work, have been compared. The EVA method categorised the three muscle recordings as too static with no EMG gaps while the JASA method identifies fatigue in two of the three recordings. The practical relevance of these findings is discussed.     

Multi-scale AM–FM analysis for the classification of surface electromyographic signals

In this work, multi-scale amplitude modulation–frequency modulation (AM–FM) features are extracted from surface electromyographic (SEMG) signals and they are used for the classification of neuromuscular disorders. The method is validated on SEMG signals recorded from a total of 40 subjects: 20 normal and 20 abnormal cases (11 myopathy, and 9 neuropathy cases), at 10%, 30%, 50%, 70% and 100% of maximum voluntary contraction (MVC), from the biceps brachii muscle. For the classification, three classifiers are used: (i) the statistical K-nearest neighbor (KNN), (ii) the self-organizing map (SOM) and (iii) the support vector machine (SVM). For all classifiers, the leave-one-out methodology is used to validate the classification of the SEMG signals into normal or abnormal (myopathy or neuropathy). A classification success rate of 78% for the AM–FM features and SVM models was achieved. These results also show that SEMG can be used as a non-invasive alternative to needle EMG for differentiating between normal and abnormal (myopathy, or neuropathy) cases.     

Automatic segmentation of surface EMG images: Improving the estimation of neuromuscular activity

Surface electromyograms (EMGs) recorded with a couple of electrodes are meant to comprise representative information of the whole muscle activation. Nonetheless, regional variations in neuromuscular activity seem to occur in numerous conditions, from standing to passive muscle stretching. In this study, we show how local activation of skeletal muscles can be automatically tracked from EMGs acquired with a bi-dimensional grid of surface electrodes (a grid of 8 rows and 15 columns was used). Grayscale images were created from simulated and experimental EMGs, filtered and segmented into clusters of activity with the watershed algorithm. The number of electrodes on each cluster and the mean level of neuromuscular activity were used to assess the accuracy of the segmentation of simulated signals. Regardless of the noise level, thickness of fat tissue and acquisition configuration (monopolar or single differential), the segmentation accuracy was above 60%. Accuracy values peaked close to 95% when pixels with intensity below ～70% of maximal EMG amplitude in each segmented cluster were excluded. When simulating opposite variations in the activity of two adjacent muscles, watershed segmentation produced clusters of activity consistently centered on each simulated portion of active muscle and with mean amplitude close to the simulated value. Finally, the segmentation algorithm was used to track spatial variations in the activity, within and between medial and lateral gastrocnemius muscles, during isometric plantar flexion contraction and in quiet standing position. In both cases, the regionalization of neuromuscular activity occurred and was consistently identified with the segmentation method.     

Amplification of background EMG activity affects the interpretation of H-reflex gain

In many H-reflex studies, the modulation of the H-reflex is usually compared relative to the normal EMG activity within the muscle. Such comparisons enable the investigators to infer whether the change in the amplitude of the H-reflex was independent of normally occurring muscle activity. This interpretation of the H-reflex is regarded as H-reflex gain, a popular dependent variable in human H-reflex studies. However, in many studies to date, the muscle activity level has been determined from the same EMG signal from which the H-reflex is recorded. This leads to an important methodological consideration: measuring the ongoing normal EMG activity from the same signal might result in an inaccurate measurement, since this EMG signal will need to be minimally amplified to capture the synchronous volley of the H-reflex amplitude. In this study we examined this possibility and found that comparing the EMG activity level from the seated position to standing position yields different results (on average 8.03% in the measurement of the increase of muscle activity). This difference was both dependent on the task and also on the EMG instrumentation used. To solve this problem we suggest the bifurcation of the EMG signal from the recording electrodes with differential amplification of the signal. With this method, both the naturally occurring muscle activity and the H-reflex signal are collected from the same area of the muscle and a more accurate measurement of the H-reflex gain will be yielded.     

Sampling, noise-reduction and amplitude estimation issues in surface electromyography.

This paper reviews data acquisition and signal processing issues relative to producing an amplitude estimate of surface EMG. The paper covers two principle areas. First, methods for reducing noise, artefact and interference in recorded EMG are described. Wherever possible noise should be reduced at the source via appropriate skin preparation, and the use of well designed active electrodes and signal recording instrumentation. Despite these efforts, some noise will always accompany the desired signal, thus signal processing techniques for noise reduction (e.g. band-pass filtering, adaptive noise cancellation filters and filters based on the wavelet transform) are discussed. Second, methods for estimating the amplitude of the EMG are reviewed. Most advanced, high-fidelity methods consist of six sequential stages: noise rejection/filtering, whitening, multiple-channel combination, amplitude demodulation, smoothing and relinearization. Theoretical and experimental research related to each of the above topics is reviewed and the current recommended practices are described.     

A segmentation approach to long duration surface EMG recordings.

The purpose of this study was to develop an automatic segmentation method in order to identify postural surface EMG segments in long-duration recordings. Surface EMG signals were collected from the cervical erector spinae (CES), erector spinae (ES), external oblique (EO), and tibialis anterior (TA) muscles of 11 subjects using a bipolar electrode configuration. Subjects remained seated in a car seat over the 150-min data-collection period. The modified dynamic cumulative sum (MDCS) algorithm was used to automatically segment the surface EMG signals. Signals were rejected by comparison with an exponential mathematical model of the spectrum of a surface EMG signal. The average power ratio computed between two successive retained segments was used to classify segments as postural or surface EMG. The presence of a negative slope of a regression line fitted to the median frequency values of postural surface EMG segments was taken as an indication of fatigue. Alpha level was set at 0.05. The overall classification error rate was 8%, and could be performed in 25 min for a 150-min signal using a custom-built software program written in C (Borland Software Corporation, CA, USA). This error rate could be enhanced by concentrating on the rejection method, which caused most of the misclassification (6%). Furthermore, the elimination of non-postural surface EMG segments by the use of a segmentation approach enabled muscular fatigue to be identified in signals that contained no evidence of fatigue when analysed using traditional methods.     

Automatic selection of uncontaminated electromyogram as applied to respiratory muscle fatigue

An automatic procedure for detecting artifacts in the electromyogram (EMG) has been developed and applied to a study of respiratory muscle fatigue. Signal segments are characterized by a set of features, the normal variations of which have been estimated in a training session. From the features are calculated a classification variable, which expresses the degree of deviation from normal conditions. A deviation larger than a certain threshold value designates a segment as disturbed. The study deals with the choice of features, the selection of a suitable segment length, and the determination of an optimal classification threshold. The four features chosen include measures of amplitude symmetry, extreme excursions in the signal tracing, the signal-to-noise ratio, and the shape of the EMG power spectrum. Recordings from three subjects were used for the evaluation of the method. The results indicate that a segment length of 250 ms is appropriate. Accepting a 10% rate of false detections, the average rate of missed detections was 2.2%.     

Intra-Individual Variability of Surface Electromyography in Front Crawl Swimming

The variability of electromyographic (EMG) recordings between and within participants is a complex problem, rarely studied in swimming. The importance of signal normalization has long been recognized, but the method used might influence variability. The aims of this study were to: (i) assess the intra-individual variability of the EMG signal in highly skilled front crawl swimmers, (ii) determine the influence of two methods of both amplitude and time normalization of the EMG signal on intra-individual variability and of time normalization on muscle activity level and (iii) describe the muscle activity, normalized using MVIC, in relation to upper limb crawl stroke movements. Muscle activity of rectus abdominis and deltoideus medialis was recorded using wireless surface EMG in 15 adult male competitive swimmers during three trials of 12.5 m front crawl at maximal speed without breathing. Two full upper limb cycles were analyzed from each of the swimming trials, resulting in six full cycles used for the intra-individual variability assessment, quantified with the coefficient of variation (CV), coefficient of quartile variation (CQV) and the variance ratio (VR). The results of this study support previous findings on EMG patterns of deltoideus medialis and rectus abdominis as prime mover during the recovery (45% activity relative to MVIC), and stabilizer of the trunk during the pull (14.5% activity) respectively. The intra-individual variability was lower (VR of 0.34–0.47) when compared to other cyclic movements. No meaningful differences were found between variability measures CV or VR when applying either of the amplitude or the time normalization methods. In addition to reporting the mean amplitude and standard deviation, future EMG studies in swimming should also report the intra-individual variability, preferably using VR as it is independent of peak amplitude, provides a good measure of repeatability and is insensitive to mean EMG amplitude and the degree of smoothing applied.     

Adaptive filtering for ECG rejection from surface EMG recordings.

Surface electromyograms (EMG) of back muscles are often corrupted by electrocardiogram (ECG) signals. This noise in the EMG signals does not allow to appreciate correctly the spectral content of the EMG signals and to follow its evolution during, for example, a fatigue process. Several methods have been proposed to reject the ECG noise from EMG recordings, but seldom taking into account the eventual changes in ECG characteristics during the experiment. In this paper we propose an adaptive filtering algorithm specifically developed for the rejection of the electrocardiogram corrupting surface electromyograms (SEMG). The first step of the study was to choose the ECG electrode position in order to record the ECG with a shape similar to that found in the noised SEMGs. Then, the efficiency of different algorithms were tested on 28 erector spinae SEMG recordings. The best algorithm belongs to the fast recursive least square family (FRLS). More precisely, the best results were obtained with the simplified formulation of a FRLS algorithm. As an application of the adaptive filtering, the paper compares the evolutions of spectral parameters of noised or denoised (after adaptive filtering) surface EMGs recorded on erector spinae muscles during a trunk extension. The fatigue test was analyzed on 16 EMG recordings. After adaptive filtering, mean initial values of energy and of mean power frequency (MPF) were significantly lower and higher respectively. The differences corresponded to the removal of the ECG components. Furthermore, classical fatigue criteria (increase in energy and decrease in MPF values over time during the fatigue test) were better observed on the denoised EMGs. The mean values of the slopes of the energy-time and MPF-time linear relationships differed significantly when established before and after adaptive filtering. These results account for the efficacy of the adaptive filtering method proposed here to denoise electrophysiological signals.     

Influence of amplitude cancellation on the accuracy of determining the onset of muscle activity from the surface electromyogram

The purpose of the study was to quantify the influence of amplitude cancellation on the accuracy of detecting the onset of muscle activity based on an analysis of simulated surface electromyographic (EMG) signals. EMG activity of a generic lower limb muscle was simulated during the stance phase of human gait. Surface EMG signals were generated with and without amplitude cancellation by summing simulated motor unit potentials either before (cancellation EMG) or after (no-cancellation EMG) the potentials had been rectified. The two sets of EMG signals were compared at forces of 30% and 80% of maximum voluntary contraction (MVC) and with various low-pass filter cut-off frequencies. Onset time was determined both visually and by an algorithm that identified when the mean amplitude of the signal within a sliding window exceeded a specified standard deviation (SD) above the baseline mean. Onset error was greater for the no-cancellation conditions when determined automatically and by visual inspection. However, the differences in onset error between the two cancellation conditions appear to be clinically insignificant. Therefore, amplitude cancellation does not appear to limit the ability to detect the onset of muscle activity from the surface EMG.     

The differential effects of fatigue on reflex response timing and amplitude in males and females.

We examined the effects of fatigue on patellar tendon reflex responses in males and females. A spring-loaded reflex hammer elicited a standardized tendon tap with the knee positioned in an isokinetic dynamometer and flexed to 85 degrees. We recorded vastus lateralis activity (SEMG) and knee extension force production at the distal tibia (force transducer). Reflex trials were performed before and after (immediate, 2, 4, and 6 min) an isokinetic fatigue protocol to 50% MVC (90 degrees /s). For each event, pre-motor time (PMT), electromechanical delay (EMD), and total motor time (TMT) were obtained, as well as EMG amplitude (EMG(amp)), time to peak EMG (EMG(tpk)), peak force amplitude (F(amp)), time to peak force (F(tpk)), EMG:force ratio (E:F), and rate of force production (F(rate)=N/ms). TMT increased significantly in females following fatigue, while males showed no change. The increased TMT was due to an increased EMD with fatigue, while PMT was unaffected. EMG(amp) and F(amp) were somewhat diminished in females yet significantly augmented in males following fatigue, likely accounting for the differential changes in EMD noted. Results suggest males and females may respond differently to isokinetic fatigue, with males having a greater capacity to compensate for contraction force failure when responding to mechanical perturbations.     

EMG spectral shift as an indicator of fatigability in an heterogeneous muscle group

Changes in electromyogram (EMG) power spectra were investigated in the triceps surae muscles of two classes of individuals (untrained subjects and athletes) maintaining a plantarflexion torque of 80% of maximal voluntary contraction until exhaustion. A set of 23 parameters describing changes in the frequency content and power of EMG was defined. For most experiments, classical changes were found, indicating a shift of the EMG spectra towards lower frequencies and an increase in the total power of the signals. In 12% of the experiments, alternations in activity between synergistic muscles were found, leading to a large variability in the spectral parameters. After the expression of each experiment in terms of a reduced data matrix and matrix to vector transformations, three methods of discrimination were used to classify subjects with respect to changes in the EMG signal during sustained contraction: (1) evaluation of the most discriminating parameter, (2) principal components analysis, (3) transformation maximizing differences between classes. Method (3) was found to be preferable since it led to good separation of the two classes in a reference group of subjects and a satisfactory projection of each individual from a group of unknowns into the appropriate class. These results suggest using a method such as this for ergonomic or athletic training purposes rather than the usual method of monitoring the frequency shift of the EMG.     

ECG Enhancement and QRS Detection Based on Sparse Derivatives

Electrocardiography (ECG) signals are often contaminated by various kinds of noise or artifacts, for example, morphological changes due to motion artifact, non-stationary noise due to muscular contraction (EMG), etc. Some of these contaminations severely affect the usefulness of ECG signals, especially when computer aided algorithms are utilized. In this paper, a novel ECG enhancement algorithm is proposed based on sparse derivatives. By solving a convex ?₁ optimization problem, artifacts are reduced by modeling the clean ECG signal as a sum of two signals whose second and third-order derivatives (differences) are sparse respectively. The algorithm is applied to a QRS detection system and validated using the MIT-BIH Arrhythmia database (109,452 anotations), resulting a sensitivity of Se = 99.87% and a positive prediction of +P = 99.88%.     

Influence of motoneuron firing synchronization on SEMG characteristics in dependence of electrode position.

The frequency content of the surface electromyography (SEMG) signal, expressed as median frequency (MF), is often assumed to reflect the decline of muscle fiber conduction velocity in fatigue. MF also decreases when motor unit firings synchronize, and we hypothesized that this effect can explain the electrode-dependent pattern in our previous recordings from the trapezius muscle. An existing motoneuron (MN) model describes the afterhyperpolarization following a spike as an exponential function on which membrane noise is superimposed. Splitting the noise into a common and an individual component extended the model to a MN pool with a tunable level of firing synchrony. An analytical volume conduction model was used to generate motor unit action potentials to simulate SEMG. A realistic level of synchrony decreased the MF of the simulated bipolar SEMG by approximately 30% midway between endplate position and tendon but not above the endplate. This is in accordance with experimental data from the biceps brachii muscle. It was concluded that the pattern of decrease of MF during sustained contractions indeed reflects MN synchronization.     

Influence of amplitude cancellation on the simulated surface electromyogram.

The purpose of the study was to quantify the influence of selected motor unit properties and patterns of activity on amplitude cancellation in the simulated surface electromyogram (EMG). The study involved computer simulations of a motor unit population with physiologically defined recruitment and rate coding characteristics that activated muscle fibers whose potentials were recorded on the skin over the muscle. Amplitude cancellation was quantified as the percent difference in signal amplitude when motor unit potentials were summed before and after rectification. The simulations involved varying the level of activation for the motor unit population, the recording configuration, the upper limit of motor unit recruitment, peak discharge rates, the amount of motor unit synchronization, muscle fiber length, the thickness of the subcutaneous tissue, and the motor unit properties that change with advancing age. The results confirmed a previous experimental report (Day SJ and Hulliger M, J Neurophysiol 86: 2144-2158, 2001) that amplitude cancellation in the surface EMG can reach 62% at maximal activation. A decrease in the range of amplitudes of the motor unit potentials, as can occur during fatiguing contractions, increased amplitude cancellation up to approximately 85%. Differences in the amount of amplitude cancellation were observed across all simulated conditions, and resulted in substantial changes in the absolute magnitude of the EMG signal. The most profound factors influencing amplitude cancellation were the number of active motor units and the duration of the action potentials. The effects of amplitude cancellation were minimal (<5%) when the EMG amplitude was normalized to maximal values, with the exception of variations in peak discharge rate and recruitment range, which resulted in differences up to 17% in the normalized EMG signal across conditions. These results indicate the amount of amplitude cancellation that can occur in various experimental conditions and its influence on absolute and relative measures of EMG amplitude.