TL;DR
Researchers have developed a novel method using high-density surface electromyography (HDEMG) to reliably track individual motor units (MUs) across multiple experimental sessions, even when separated by weeks. This breakthrough allows for unprecedented longitudinal monitoring of MU properties, opening new avenues for understanding adaptations to training and the progression of neuromuscular disorders.
What Problem Does This Paper Address?
Studying how motor units change over time – whether due to training, injury, or disease – is crucial for understanding neuromuscular function. However, traditional methods for recording motor unit activity, like intramuscular fine-wire electrodes, are limited to tracking only a few MUs at a time and make it nearly impossible to identify the same MUs across different testing sessions. This lack of longitudinal tracking has severely limited our ability to gather robust evidence on how MU properties adapt or degrade over time. The challenge, therefore, is to develop a technique that can reliably identify and follow individual MUs across multiple experimental sessions, providing a more sensitive and accurate picture of neuromuscular changes.
Methods
The core innovation in this paper is a new processing method built upon existing convolutive blind source separation techniques for HDEMG. The key advancement is the introduction of a "semi-blind separation procedure" that specifically focuses on finding motor unit action potentials (MUAPs) that are highly similar across different recording sessions. This is achieved by maximizing the cross-correlation of MUAP profiles between sessions.
Here's a breakdown:
- HDEMG Data Acquisition: High-density surface EMG grids were used to capture electrical activity from muscles like the vastus medialis and vastus lateralis.
- Full Decomposition: Initially, a comprehensive blind decomposition was performed on the first recording session to identify all individual motor unit signals.
- Semi-Blind Tracking: For subsequent sessions, the algorithm was modified. Instead of a full blind separation, it prioritized identifying signals that strongly matched (high cross-correlation) the MUAPs already identified in the first session. This "matching" process allowed for the reliable identification of the same MUs across sessions.
- Similarity Measure: The normalized cross-correlation between MUAP profiles was used as the primary metric for determining if a MU was tracked across sessions. A threshold of 0.8 was used, with the highest correlation selected in cases of multiple potential matches.
- Reliability and Sensitivity Analysis: The effectiveness of the tracking method was evaluated by comparing the reliability of MU properties (like conduction velocity, discharge rate) for tracked MUs versus the entire population of identified MUs and a random subset of unmatched MUs. Sensitivity was tested by observing changes in MU properties after an endurance training intervention.
Datasets
The study utilized datasets collected from two experimental protocols involving human participants:
- Experiment I (Reliability): Ten healthy male participants performed isometric knee extensions at various force levels (10%, 30%, 50%, 70% MVC) across three sessions, each separated by one week. This dataset was used to assess the reliability of the tracking method.
- Experiment II (Sensitivity): Seven healthy male participants underwent a two-week endurance training program (cycling). They were tested before and after the intervention with isometric knee extensions (10% and 30% MVC). This dataset was used to evaluate the sensitivity of the tracking method to training-induced changes.
No new benchmark datasets were introduced; rather, existing data acquisition protocols were adapted for the purpose of longitudinal tracking.
Key Insights and Ideas
- The "Same" Motor Unit Across Time: The most significant insight is that it's possible to reliably identify and follow the exact same motor units across multiple experimental sessions, even when those sessions are separated by a week or more. This overcomes a major limitation of previous research.
- HDEMG as a Longitudinal Tool: The study demonstrates that HDEMG, combined with this novel decomposition and tracking algorithm, is a powerful tool for longitudinal studies, not just for single-session analysis.
- Enhanced Reliability and Sensitivity: Tracking individual MUs leads to more reliable estimates of their properties compared to analyzing the entire population of detected MUs. Crucially, this tracking method significantly increases the sensitivity to detect subtle changes in MU properties, such as those induced by training.
- Robustness to Force Variations: The method showed promise in tracking MUs even when force levels varied within a session, suggesting it can handle dynamic contractions and changes in activation.
Key Results
- Tracking Success Rates: In Experiment I, an average of 38.3% (vastus medialis) and 40.1% (vastus lateralis) of identified MUs could be tracked across two sessions separated by one week. For three sessions, this dropped to 21.0% and 16.3%, respectively.
- Improved Reliability: The intra-class correlation coefficients (ICCs) for tracked MUs were significantly higher (ranging from 0.63-0.99) compared to the full group of identified MUs (0.39-0.95), indicating more stable and consistent measurements.
- Training Sensitivity: In Experiment II, approximately 40% of MUs could be tracked pre- and post-training. The tracked MUs showed a larger effect size (ES = 2.1) for changes in motor unit conduction velocity after training compared to the entire group of identified MUs (ES = 1.5). This highlights the increased sensitivity of the tracking method.
- Consistent MUAP Shapes: Despite significant changes in conduction velocity after training (up to 10%), the MUAP shapes of tracked MUs remained highly consistent (cross-correlation > 90%), reinforcing the accuracy of the tracking.
Limitations and Future Work
- Time Gap: The current study focused on tracking across sessions separated by up to 2.5 weeks. The authors acknowledge that significant muscle structural changes (which can affect MUAP shapes) might occur over longer periods, potentially impacting tracking accuracy. Future work should evaluate the method's performance over longer intervention periods.
- Muscle Specificity: The study primarily focused on the vasti muscles. The authors note that the number of identified MUs can vary between muscles due to differences in muscle fiber architecture and spatial correlation of EMG signals. Further validation in muscles with lower spatial correlation is warranted.
- Pathological Conditions: While the method holds promise for studying neuromuscular disorders, its application in pathological conditions (e.g., ALS) over extended periods needs to be investigated.
- Multiple Matches: In a small percentage of trials (~15%), motor units showed multiple potential matches. While the algorithm selected the highest correlated source, this suggests that high spatial correlation in some muscles can lead to occasional similarities between different motor units.
- Future Directions: The authors suggest that this method can be applied to study adaptations to resistance training, rehabilitation interventions, and the progression of neuromuscular disorders. It could also provide precise measurements of recruitment thresholds without normalization, offering a more direct assessment of motor unit force capacity.
Summary updated