Abstract
To advance our understanding of the neural control of muscle, we decoded the firing activity of extensive samples of motor units in the Tibialis Anterior and the Vastus Lateralis during isometric contractions of up to 80% of maximal force. Using a unique dataset of approximately 130 motor units per participant, we characterized the rate coding of each unit as the relationship between its instantaneous firing rate and muscle force. This relationship was best described by a natural logarithm function comprising two phases: an initial steep acceleration of firing rate, followed by a linear increase. Contrary to traditional assumptions of uniform control, our results demonstrate that firing rates follow diverse trends across the motoneuron pool. These findings highlight how motor unit pools utilize gain control to transform limited bandwidth neural inputs into precise muscle force.
Introduction: Mapping the Neural Control of Movement
The execution of movement relies on motoneurons transforming synaptic inputs into activation signals that control muscle force. While the "size principle"—the orderly recruitment of motor units from smallest to largest—is well-established, the dynamics of "rate coding" (the frequency of neural firing) have remained less clear. Historically, research has been constrained by small sample sizes or limited force ranges, often below 30% of maximal voluntary contraction (MVC).
This limitation has fueled a long-standing debate: does the entire motoneuron pool follow a rigid "onion skin" pattern, where earlier-recruited units always maintain higher firing rates, or is the system more dynamic? This study by Avrillon et al. addresses these gaps by providing a comprehensive map of rate coding across the entire operating range of two major human leg muscles.
Methodology: High-Density Decoding
The researchers employed a sophisticated experimental setup to capture the activity of hundreds of motor units simultaneously. High-density surface electromyography (HDsEMG) was used, with 256 electrodes placed over the Tibialis Anterior (shin) and Vastus Lateralis (quad) muscles in 16 participants.
To extract individual motor unit activity from the complex interference of the EMG signal, the team applied Blind Source Separation (BSS) algorithms. This allowed them to "unmix" the electrical signals into discrete pulse trains for each motor unit. Crucially, they tracked the same unique units as participants performed ramp-up and ramp-down contractions from 10% to 80% MVC.
The accuracy of this decoding was validated by reconstructing the EMG signal from the identified motor unit firing patterns and comparing it to the original recording (as seen in Figure 4B and 4C). The study generated an unprecedented dataset, identifying an average of 129-130 units per participant, far exceeding previous benchmarks.
Key Insights: The Non-Linear Nature of Rate Coding
The analysis revealed that the relationship between neural input (proxied by muscle force) and motor unit firing rate is fundamentally non-linear. Approximately 70% of the decoded motor units were best described by a natural logarithm function rather than a linear model.
Two-Phase Firing Trajectory
The researchers identified two distinct phases in the rate coding process:
- Acceleration Phase: Immediately following recruitment, motor units exhibit a steep increase in firing rate. This acceleration is most pronounced in low-threshold units.
- Saturation/Linear Phase: As force continues to rise, the firing rate transitions to a slower, linear increase. High-threshold units show a higher "gain" in this phase, allowing their firing rates to approach or even surpass those of earlier-recruited units at high force levels.
Gain Control and Neuromodulation
These non-linearities suggest that the spinal cord acts as more than a simple relay station. The authors propose that these patterns are driven by Persistent Inward Currents (PICs)—intrinsic properties of motoneurons that amplify synaptic inputs. This "gain control" mechanism allows the nervous system to transform relatively narrow-bandwidth neural commands into the wide range of forces required for human movement.
Results and Muscle-Specific Differences
The study's extensive sampling revealed that different muscles are "tuned" for their specific roles. For instance, the Tibialis Anterior showed a higher initial acceleration of firing rates compared to the Vastus Lateralis. Furthermore, during the "ramp-down" phase of contractions, medium-threshold units exhibited prolonged firing (hysteresis), a hallmark of neuromodulatory self-sustained firing. This effect was less prominent in the lowest and highest threshold units, indicating a threshold-dependent distribution of neuromodulatory effects across the pool.
Limitations and Future Directions
While this work represents a technical milestone, it focused on isometric (static) contractions. Future research is needed to determine how these rate-coding strategies adapt during dynamic movements like walking. Additionally, while 80% MVC is a significant range, decoding at 100% MVC remains a challenge due to signal "cancellation" effects at maximal intensities.
Conclusion
By decoding the activity of entire motoneuron populations in vivo, this research provides a new framework for understanding the neural control of muscle. It moves beyond simplistic models of uniform rate coding, demonstrating instead a sophisticated, non-linear gain control system that optimizes force production across the full functional range of human muscles.
Summary updated