Motor unit tracking using blind source separation filters and waveform cross-correlations: reliability under physiological and pharmacological conditions

Recent advancements in the analysis of high-density surface electromyography (HDsEMG) have enabled the identification, and tracking, of motor units (MUs) to study muscle activation. This study aimed to evaluate the reliability of MU tracking using two common methods: blind source separation filters and two-dimensional waveform cross-correlation. An experiment design was developed to assess physiological reliability, and reliability for a drug intervention known to reduce the firing rate of motoneurones (cyproheptadine). HDsEMG signals were recorded from tibialis anterior during isometric dorsiflexions to 10%, 30%, 50% and 70% of maximal voluntary contraction. MUs were matched within session (2 hr) using the filter method, and between sessions (7 days) via the waveform method. Both tracking methods demonstrated similar reliability during physiological conditions (e.g., MU discharge: filter ICC 10% of MVC = 0.76, to 70% of MVC = 0.86; waveform ICC: 10% of MVC = 0.78, to 70% of MVC = 0.91). Although reliability slightly reduced after the pharmacological intervention, there were no discernible differences in tracking performance (e.g., MU disc filter ICC: 10% of MVC = 0.73, to 70% of MVC = 0.75; DR waveform ICC: 10% of MVC = 0.84, to 70% of MVC = 0.85). The poorest reliability typically occurred at higher contraction intensities, which aligned with the greatest variability in MU characteristics. This study confirms that tracking method may not impact the interpretation of MU data, provided that an appropriate experiment design is employed. However, caution should be used when tracking MUs during higher intensity isometric contractions.

Endurance-exercise training adaptations in spinal motoneurones: potential functional relevance to locomotor output and assessment in humans

It is clear from non-human animal work that spinal motoneurones undergo endurance training (chronic) and locomotor (acute) related changes in their electrical properties and thus their ability to fire action potentials in response to synaptic input. The functional implications of these changes, however, are speculative. In humans, data suggests that similar chronic and acute changes in motoneurone excitability may occur, though the work is limited due to technical constraints. To examine the potential influence of chronic changes in human motoneurone excitability on the acute changes that occur during locomotor output, we must develop more sophisticated recording techniques or adapt our current methods. In this review, we briefly discuss chronic and acute changes in motoneurone excitability arising from non-human and human work. We then discuss the potential interaction effects of chronic and acute changes in motoneurone excitability and the potential impact on locomotor output. Finally, we discuss the use of high-density surface electromyogram recordings to examine human motor unit firing patterns and thus, indirectly, motoneurone excitability. The assessment of single motor units from high-density recording is mainly limited to tonic motor outputs and minimally dynamic motor output such as postural sway. Adapting this technology for use during locomotor outputs would allow us to gain a better understanding of the potential functional implications of endurance training-induced changes in human motoneurone excitability on motor output.

Preprocessing surface EMG data removes voluntary muscle activity and enhances SPiQE fasciculation analysis

•

A novel preprocessing step removes the need for manual selection of relaxed surface EMG data.

•

SPiQE provides reliable fasciculation analysis from raw thirty-minute recordings in ALS.

•

This paves the way for clinical calibration of a potential novel biomarker of disease progression.

Objectives

Fasciculations are a clinical hallmark of amyotrophic lateral sclerosis (ALS). The Surface Potential Quantification Engine (SPiQE) is a novel analytical tool to identify fasciculation potentials from high-density surface electromyography (HDSEMG). This method was accurate on relaxed recordings amidst fluctuating noise levels. To avoid time-consuming manual exclusion of voluntary muscle activity, we developed a method capable of rapidly excluding voluntary potentials and integrating with the established SPiQE pipeline.

Methods

Six ALS patients, one patient with benign fasciculation syndrome and one patient with multifocal motor neuropathy underwent monthly thirty-minute HDSEMG from biceps and gastrocnemius. In MATLAB, we developed and compared the performance of four Active Voluntary IDentification (AVID) strategies, producing a decision aid for optimal selection.

Results

Assessment of 601 one-minute recordings permitted the development of sensitive, specific and screening strategies to exclude voluntary potentials. Exclusion times (0.2–13.1 minutes), processing times (10.7–49.5 seconds) and fasciculation frequencies (27.4–71.1 per minute) for 165 thirty-minute recordings were compared. The overall median fasciculation frequency was 40.5 per minute (10.6–79.4 IQR).

Conclusion

We hereby introduce AVID as a flexible, targeted approach to exclude voluntary muscle activity from HDSEMG recordings.

Significance

Longitudinal quantification of fasciculations in ALS could provide unique insight into motor neuron health.

SPiQE: An automated analytical tool for detecting and characterising fasciculations in amyotrophic lateral sclerosis

OBJECTIVES

Fasciculations are a clinical hallmark of amyotrophic lateral sclerosis (ALS). Compared to concentric needle EMG, high-density surface EMG (HDSEMG) is non-invasive and records fasciculation potentials (FPs) from greater muscle volumes over longer durations. To detect and characterise FPs from vast data sets generated by serial HDSEMG, we developed an automated analytical tool.

METHODS

Six ALS patients and two control patients (one with benign fasciculation syndrome and one with multifocal motor neuropathy) underwent 30-minute HDSEMG from biceps and gastrocnemius monthly. In MATLAB we developed a novel, innovative method to identify FPs amidst fluctuating noise levels. One hundred repeats of 5-fold cross validation estimated the model's predictive ability.

RESULTS

By applying this method, we identified 5,318 FPs from 80 minutes of recordings with a sensitivity of 83.6% (+/- 0.2 SEM), specificity of 91.6% (+/- 0.1 SEM) and classification accuracy of 87.9% (+/- 0.1 SEM). An amplitude exclusion threshold (100 μV) removed excessively noisy data without compromising sensitivity. The resulting automated FP counts were not significantly different to the manual counts (p = 0.394).

CONCLUSION

We have devised and internally validated an automated method to accurately identify FPs from HDSEMG, a technique we have named Surface Potential Quantification Engine (SPiQE).

SIGNIFICANCE

Longitudinal quantification of fasciculations in ALS could provide unique insight into motor neuron health.

High-density surface electromyography provides reliable estimates of motor unit behavior

OBJECTIVE

To assess the intra- and inter-session reliability of estimates of motor unit behavior and muscle fiber properties derived from high-density surface electromyography (HDEMG).

METHODS

Ten healthy subjects performed submaximal isometric knee extensions during three recording sessions (separate days) at 10%, 30%, 50% and 70% of their maximum voluntary effort. The discharge timings of motor units of the vastus lateralis and medialis muscles were automatically identified from HDEMG by a decomposition algorithm. We characterized the number of detected motor units, their discharge rates, the coefficient of variation of their inter-spike intervals (CoVisi), the action potential conduction velocity and peak-to-peak amplitude. Reliability was assessed for each motor unit characteristics by intra-class correlation coefficient (ICC). Additionally, a pulse-to-noise ratio (PNR) was calculated, to verify the accuracy of the decomposition.

RESULTS

Good to excellent reliability within and between sessions was found for all motor unit characteristics at all force levels (ICCs>0.8), with the exception of CoVisi that presented poor reliability (ICC<0.6). PNR was high and similar for both muscles with values ranging between 45.1 and 47.6dB (accuracy>95%).

CONCLUSION

Motor unit features can be assessed non-invasively and reliably within and across sessions over a wide range of force levels.

SIGNIFICANCE

These results suggest that it is possible to characterize motor units in longitudinal intervention studies.