Tutorial: Analysis of motor unit discharge characteristics from high-density surface EMG signals

Recent work demonstrated that it is possible to identify motor unit discharge times from high-density surface EMG (HDEMG) decomposition. Since then, the number of studies that use HDEMG decomposition for motor unit investigations has increased considerably. Although HDEMG decomposition is a semi-automatic process, the analysis and interpretation of the motor unit pulse trains requires a thorough inspection of the output of the decomposition result. Here, we report guidelines to perform an accurate extraction of motor unit discharge times and interpretation of the signals. This tutorial includes a discussion of the differences between the extraction of global EMG signal features versus the identification of motor unit activity for physiological investigations followed by a comprehensive guide on how to acquire, inspect, and decompose HDEMG signals, and robust extraction of motor unit discharge characteristics.     

Analysis of motor units with high-density surface electromyography

Although the behaviour of individual motor units is classically studied with intramuscular EMG, recently developed techniques allow its analysis also from EMG recorded in multiple locations over the skin surface (high-density surface EMG). The analysis of motor units from the surface EMG is useful when the insertion of needles is not desirable or not possible. Moreover, surface EMG allows the measure of motor unit properties which are difficult to assess with invasive technology (e.g., muscle fiber conduction velocity or location of innervation zones) and may increase the number of detectable motor units with respect to selective intramuscular recordings. Although some limitations remain, both the discharge pattern and muscle fiber properties of individual motor units can currently be analyzed non-invasively. This review presents the conditions and methodologies which allow the investigation of motor units with surface EMG.     

Sex differences in the detection of motor unit action potentials identified using high-density surface electromyography

Sex-related disparities in force production of humans have been widely observed. Previous literature has attributed differences in peripheral traits, such as muscle size, to explain these disparities. However, less is known about potential sex-related differences in central neuromuscular traits and many comparable studies, not exploring sex-related differences, exhibit a selection-bias in the recruitment of subjects making the generalization of their findings difficult. Utilizing high-density electromyography arrays and motor unit (MU) decomposition, the aim of the current study is to compare MU yield and discharge properties of the tibialis anterior between male and female humans. Twenty-four subjects (10 females) performed two submaximal (20%) isometric dorsiflexion contractions. On average, males yielded nearly twice the amount of MUs as females. Further, females had significantly higher MU discharge rate, lower MU action potential amplitude, and lower MU action potential frequency content than males despite similar levels of torque and MU discharge variability. These findings suggest differences in central neuromuscular control of force production between sexes; however, it is unclear how lower yield counts affect the accuracy of these results.     

A method for editing motor unit potential trains obtained by decomposition of surface electromyographic signals

Rather than discarding motor unit potential trains (MUPTs) because they do not meet 100% validity criteria, we describe and evaluate a novel editing routine that preserves valid discharge times, based on decreasing shape variability (variance ratio, VR) within a MUPT. The error filtered estimation (EFE) algorithm is then applied to the remaining ‘high confidence’ discharge times to estimate inter-discharge interval (IDI) statistics. Decomposed surface EMG data from the flexor carpi radialis recorded from 20 participants during 60% MVC wrist flexion was used. There were two levels of denoising criteria (relaxed and strict) criteria for removing MUPs to decrease the VR and increase the signal-to-noise ratio (SNR) of a MUPT. In total, VR decreased 24.88% and SNR increased 6.0% (p’s < 0.05). The MUP template peak-to-peak (P-P) amplitude and P-P duration were dependent on the level of denoising (p’s < 0.05). The standard error of the estimate (SEE) of the mean IDI before and after editing using the relaxed criteria (3.2% versus 3.69%), was very similar (p > 0.05). The same was true for the SEE between denoising criteria, which increased only to 5.14% for the strict criteria (p > 0.05). Editing the MUPTs resulted in a significant decrease in MUP shape variability and in the measures extracted from the MUP templates, with trivial differences between the SEE of the mean IDI between the edited and unedited MUPTs.     

Evolution of surface electromyography: From muscle electrophysiology towards neural recording and interfacing

Surface electromyography (EMG) comprises a recording of electrical activity from the body surface generated by muscle fibres during muscle contractions. Its characteristics depend on the fibre membrane potentials and the neural activation signal sent from the motor neurons to the muscles. EMG has been classically used as the primary investigation tool in kinesiology studies in a variety of applications. More recently, surface EMG techniques have evolved from single-channel methods to high-density systems with hundreds of electrodes. High-density EMG recordings can be deconvolved to estimate the discharge times of spinal motor neurons innervating the recorded muscles, with algorithms that have been developed and validated in the last two decades. Within limits and with some variability across muscles, these techniques provide a non-invasive method to study relatively large populations of motor neurons in humans. Surface EMG is thus evolving from a peripheral measure of muscle electrical activity towards a neural recording and neural interfacing signal. These advances in technology have had a major impact on our fundamental understanding of the neural control of movement and have exposed new perspectives in neurotechnologies. Here we provide an overview and perspective of modern EMG technology, as derived from past achievements, and its impact in neurophysiology and neural engineering.     

Inter-operator agreement in decomposition of motor unit firings from high-density surface EMG.

High-density surface EMG can be used to obtain a spatially selective representation of several motor unit action potentials. Recently, a decomposition of the signal into the underlying motor neuron firing patterns has been described. The reliability of the algorithm has not yet been tested. Eleven healthy subjects participated. High-density surface EMG was recorded from the vastus lateralis muscle during an isometric knee extension. Two independent operators analyzed the signals. After operator-supervised cluster analysis of spikes, motor unit action potential templates were constructed and an automatic template matching was performed. The decomposition was adjusted by hand. Agreement between operators was calculated for the number of coincident firings. Bland-Altman plots of peak-to-peak amplitude were constructed and limits of agreement were calculated. For completely decomposed motor unit action potential trains the between-operator agreement of firing events was very high. The peak-to-peak amplitude of monopolar motor unit action potentials was 115microV (SD 74microV). The agreement was within 3microV and independent of amplitude. With partial decomposition agreement within 26microV was achieved. For bipolarly derived motor unit action potentials the peak-to-peak amplitude was 54microV (SD 49microV), the agreement was within 3microV. Only for recordings obtained from a force level below 5% of the maximum voluntary contraction full decomposition was possible. It was concluded that when full decomposition is achieved, two independent operators are likely to arrive at nearly identical firing patterns.     

Analysis of point process signals applied to motor unit firing patterns: I. Superposition of independent spike trains

The aim of the present study was to examine whether statistical methods common for the analysis of point process signals could be applied to the electromyogram, in order to extract information concerning the physiological mechanisms involved. This was carried out on the assumption that the electromyogram can be treated as the superposition result of a number of point process signals, each representing the firing pattern of one motor unit. No correlated activity between the different spike trains was assumed at this stage. A digital model for the superposition of event sequences was constructed, assigning to the individual sequences a Gaussian interval distribution. The effects of varying the number of spike trains participating in the superposition process, and changing the mean rates of firing were explored. The statistical methods used in the analysis were serial correlation, event autocorrelation, and power spectrum studies. It has been found that serial correlograms of the superimposed processes may be helpful in detecting the number of spike trains involved in the superposition, whereas power spectrum studies are useful in determining the mean rates of firing of the individual sequences.     

Analysis of point process signals applied to motor unit firing patterns: II. Superposition of phase-locked spike trains

The different possibilities of cross-constraints between the firing patterns of a number of motor units are laid out. Correlated phasic activity is defined, and the effect of phase locking on the superposition of event sequences is being investigated by the simulation model. For superposition of four spike trains, the two cases of phase locking investigated by the model, φ=0.25 and φ=0.0 may represent an asynchroneous and synchroneous motor unit activity, respectively. A filtering method for estimation of the phase, in cases of phase-locked activity, is described.     

Assessing the validity of surface electromyography for recording muscle activation patterns from serratus anterior

PurposeNo direct evidence exists to support the validity of using surface electrodes to record muscle activity from serratus anterior, an important and commonly investigated shoulder muscle. The aims of this study were to determine the validity of examining muscle activation patterns in serratus anterior using surface electromyography and to determine whether intramuscular electromyography is representative of serratus anterior muscle activity.MethodsSeven asymptomatic subjects performed dynamic and isometric shoulder flexion, extension, abduction, adduction and dynamic bench press plus tests. Surface electrodes were placed over serratus anterior and around intramuscular electrodes in serratus anterior. Load was ramped during isometric tests from 0% to 100% maximum load and dynamic tests were performed at 70% maximum load. EMG signals were normalised using five standard maximum voluntary contraction tests.ResultsSurface electrodes significantly underestimated serratus anterior muscle activity compared with the intramuscular electrodes during dynamic flexion, dynamic abduction, isometric flexion, isometric abduction and bench press plus tests. All other test conditions showed no significant differences including the flexion normalisation test where maximum activation was recorded from both electrode types. Low correlation between signals was recorded using surface and intramuscular electrodes during concentric phases of dynamic abduction and flexion.ConclusionsIt is not valid to use surface electromyography to assess muscle activation levels in serratus anterior during isometric exercises where the electrodes are not placed at the angle of testing and dynamic exercises. Intramuscular electrodes are as representative of the serratus anterior muscle activity as surface electrodes.     

Measuring the accuracies of motor unit firing times and action potential waveforms derived from surface electromyographic decomposition

This study included spike trigger averaging (STA) procedures to examine the acceptability of the Precision Decomposition (PD) III derived motor unit action potential (MUAP) trains that met the >90% accuracy criteria from the reconstruct-and-test. MUs met the >90% accuracy criteria from the reconstruct-and-test with STA procedures then applied. Y-intercepts and slopes were calculated for the firing rate- and MUAP amplitude-recruitment threshold relationships. Gaussian noise (1% of the SD of the mean interspike interval) was added to the firing times with the changes in MUAPs quantified. A total of 455 MUs were decomposed with 155 MUs removed as a result of the reconstruct-and-test. Five additional MUs were excluded via the STA criteria. The MUAP waveforms deteriorated with the inclusion of Gaussian noise. There were differences in the derived action potentials amplitudes of higher-threshold MUs between the PD III algorithm and the STA procedure. There was excellent agreement among the slopes and y-intercepts between the relationships that included or excluded MUs that did not meet the STA criteria. There was good agreement between the MUAP amplitude-recruitment threshold relationships derived from the PD III and STA procedure. The addition of the STA procedures did not alter the MU-derived relationships.     

Comparison of methods for estimating motor unit firing rate time series from firing times

The central nervous system regulates recruitment and firing of motor units to modulate muscle tension. Estimation of the firing rate time series is typically performed by decomposing the electromyogram (EMG) into its constituent firing times, then lowpass filtering a constituent train of impulses. Little research has examined the performance of different estimation methods, particularly in the inevitable presence of decomposition errors. The study of electrocardiogram (ECG) and electroneurogram (ENG) firing rate time series presents a similar problem, and has applied novel simulation models and firing rate estimators. Herein, we adapted an ENG/ECG simulation model to generate realistic EMG firing times derived from known rates, and assessed various firing rate time series estimation methods. ENG/ECG-inspired rate estimation worked exceptionally well when EMG decomposition errors were absent, but degraded unacceptably with decomposition error rates of ⩾1%. Typical EMG decomposition error rates—even after expert manual review—are 3–5%. At realistic decomposition error rates, more traditional EMG smoothing approaches performed best, when optimal smoothing window durations were selected. This optimal window was often longer than the 400 ms duration that is commonly used in the literature. The optimal duration decreased as the modulation frequency of firing rate increased, average firing rate increased and decomposition errors decreased. Examples of these rate estimation methods on physiologic data are also provided, demonstrating their influence on measures computed from the firing rate estimate.     

Spatio-temporal processing of surface EMG signals from the sternocleidomastoideus muscle to assess effects of radiotherapy on motor unit conduction velocity and firing rate—A pilot study

Radiation therapy causes both muscle and nerve tissue damage. However, the evolution and mechanisms of these damages are not fully understood. Information on the state of active muscle fibres and motoneurons can be obtained by measuring sEMG signals and calculating the conduction velocity (CV) and firing rate of individual motor units, respectively. The aim of this pilot study was to evaluate if the multi-channel surface EMG (sEMG) technique could be applied to the sternocleidomastoideus muscle (SCM) of radiotherapy patients, and to assess if the CV and firing rate are altered as a consequence of the radiation.

Surface EMG signals were recorded from the radiated and healthy SCM muscles of 10 subjects, while subjects performed isometric rotation of the head. CV and firing rate were calculated using two recently proposed methods based on spatio-temporal processing of the sEMG signals. The multi-channel sEMG technique was successfully applied to the SCM muscle and CV and firing rates were obtained. The measurements were fast and simple and comfortable for the patients. Sufficient data quality was obtained from both sides of seven and four subjects for the CV and firing rate analysis, respectively. No differences in CV or firing rate were found between the radiated and non-radiated sides (p = 0.13 and p = 0.20, respectively). Firing rate and CV were also obtained from a myokymic discharge pattern. It was found that the CV decreased significantly (p = 0.01) during the bursts.