Motor Unit Identification From High-Density Surface Electromyograms in Repeated Dynamic Muscle Contractions

Area of Science:

• Biomedical Engineering
• Neuroscience
• Physiology

Background:

• Accurate identification of motor unit (MU) firings is crucial for understanding muscle control.
• Dynamic muscle contractions present challenges due to changing motor unit action potentials (MUAPs).
• Existing methods may struggle with the variability of MUAPs during dynamic contractions.

Purpose of the Study:

• To present a novel method for identifying MU firings from dynamic high-density surface electromyograms (hdEMG).
• To introduce a new convolutive data model and a pulse-to-noise ratio (PNR) metric for assessing MU identification accuracy.
• To analyze the impact of dynamic MUAP changes on MU identification.

Main Methods:

• Development of a new convolutive data model for dynamic hdEMG signals.
• Implementation of a pulse-to-noise ratio (PNR) metric for accuracy assessment.
• Testing the methodology on synthetic and experimental hdEMG data from biceps brachii, vastus lateralis, and rectus femoris muscles during dynamic contractions.

Main Results:

• The method successfully identified MUs in synthetic signals with high sensitivity and precision (>90%) and PNR >30dB.
• In experimental signals, 9.4±1.9 and 7.8±1.4 MUs were identified in vastus lateralis and rectus femoris muscles, respectively, with high PNR values.
• The new method demonstrated superior ability to track dynamic MUAP changes compared to the Convolution Kernel Compensation method, even in fast contractions.

Conclusions:

• The presented method provides accurate and reliable motor unit identification from dynamic hdEMG signals.
• The novel convolutive model and PNR metric are effective tools for analyzing dynamic muscle contractions.
• This approach advances the understanding of neuromuscular control during varying contraction speeds and forces.