Effects of muscle mass on muscle force predictions in human movement

This study demonstrates that while incorporating muscle mass into Hill-type models significantly improves force predictions for larger-than-human muscles or high-cadence movements, the difference between mass-enhanced and traditional massless models remains negligible (<1%) for standard human-sized locomotion.

The Gist

The Big Question: Does Muscle Weight Matter?

Imagine you are swinging a light, fluffy feather versus a heavy, dense bowling ball. Even if you try to move them with the exact same "muscle power," the bowling ball is much harder to start moving and much harder to stop.

For decades, scientists have built computer models to simulate how human muscles work. Most of these models treat muscles like invisible, weightless strings. They assume that whether you are a tiny mouse or a giant gorilla, the muscle's "engine" works the same way, ignoring the fact that real muscle has weight (mass).

This paper asks a simple question: Does the actual weight of the muscle matter when we move? Specifically, does the muscle's own weight slow it down or change the force it produces, especially when we move fast or when the muscle is very large?

The Experiment: The "Muscle Scale"

To answer this, the researchers didn't just look at real people; they used a clever computer trick.

1. The Real Data: They recorded 20 real people walking, running, hopping, sitting down, and cycling. They measured how their muscles activated and how their limbs moved.
2. The "Magic Scale": They took this real data and ran it through two different computer models:
   • Model A (The Ghost Muscle): A traditional model where the muscle has zero weight.
   • Model B (The Heavy Muscle): A new model where the muscle has real weight (inertia).
3. The Shrink and Grow: They then "shrank" and "grew" the muscles in the computer. They simulated muscles that were 1/10th the size of a human, normal human size, and up to 10 times larger than a human (like a giant).

The Findings: When Weight Counts

Here is what they discovered, broken down by scenario:

1. The "Feather" Scenario (Small Muscles & Slow Movements)

When the muscles were small (like a mouse) or the movement was slow (like walking or sitting down), the weight didn't matter at all.

• Analogy: Imagine trying to shake a wet noodle. It's so light that its own weight doesn't fight against your hand. The "Ghost Muscle" and the "Heavy Muscle" models gave almost identical results.
• Result: For normal humans doing daily tasks, ignoring muscle weight is actually a pretty good shortcut. The error is less than 1%.

2. The "Bowling Ball" Scenario (Big Muscles & Fast Movements)

When they made the muscles huge (10x human size) or made them move very fast (like sprinting or cycling at high speed), the weight started to matter a lot.

• Analogy: Now imagine trying to shake a giant, heavy wet towel. The towel is so heavy that it fights back. You have to spend energy just to get the towel moving, and it takes longer to stop.
• Result: In these "Giant" scenarios, the heavy muscle model showed that the muscle had to work harder just to accelerate its own weight. This reduced the power available to do actual work (like moving a bike pedal). The difference between the two models grew to about 7%.

3. The "Cadence" Factor (How Fast You Move)

The researchers found that speed was a bigger trigger for these weight effects than strength.

• Analogy: Think of a car. If you drive slowly, the weight of the car doesn't feel that different. But if you try to accelerate from 0 to 100 mph instantly, the heavy engine and chassis make it much harder to get moving.
• Result: High-speed movements (like hopping or fast cycling) caused the "heavy muscle" to lag behind the "light muscle" more than slow movements did.

Why This Matters

For the average person: You can breathe a sigh of relief. If you are a normal-sized human walking or jogging, your muscles are light enough that their own weight doesn't really slow you down. The old, simple computer models are fine for studying us.

For the giants and the speedsters: If you are designing a robot with giant muscles, or studying a very fast animal, or simulating extreme athletic performance, you must account for muscle weight. Ignoring it is like trying to calculate how fast a rocket goes without accounting for the weight of the fuel tank.

The Catch (Limitations)

The authors admit their computer models were a bit simplified. They treated muscles like 1D strings (like a rope). Real muscles are 3D blobs that bulge and squish in all directions.

• The Analogy: They studied the weight of a rope, but real muscles are more like a water balloon. When you shake a water balloon, the water sloshes around in 3D, which might change the physics even more.

The Bottom Line

Muscle mass is like a hidden tax on your movement.

• Small size + Slow speed: The tax is zero.
• Big size + Fast speed: The tax gets expensive.

For most human daily life, the tax is so small we can ignore it. But if we want to understand the limits of human performance or simulate giant creatures, we have to start counting the weight of the muscle itself.

Technical Summary

1. Problem Statement

Traditional musculoskeletal simulations typically utilize 1D massless Hill-type muscle models. These models assume that muscle tissue has no inertia, treating large muscles and small fibers as having identical intrinsic properties regardless of size. However, skeletal muscle possesses mass, and accelerating this mass requires energy (inertial cost), which detracts from the energy available for external work.

The core problem addressed is whether neglecting muscle mass significantly impacts the prediction of muscle force, power, and work during dynamic, submaximal human movements. Specifically, the authors investigate:

• How muscle inertia influences contractile dynamics.
• Whether the discrepancy between massless and mass-enhanced models scales with muscle size (geometric scaling).
• Whether these discrepancies are significant enough to warrant the computational complexity of including mass in standard human locomotion simulations.

2. Methodology

Experimental Data Collection:

• Participants: 20 individuals (10 male, 10 female) performed daily activities (walking, running, hopping, sit-to-stand) on a treadmill.
• Supplementary Data: 14 competitive cyclists provided data for high-frequency pedaling tasks (varying cadence and torque).
• Measurements: Kinematic data (motion capture), Ground Reaction Forces (force plates), and Electromyography (EMG) from lower-limb muscles (MG, LG, SOL, RF, VM, VL).
• Processing: EMG signals were converted to muscle activation levels. Kinematic data were used to estimate Muscle-Tendon Unit (MTU) lengths via inverse kinematics in OpenSim.

Modeling Approach:

• Two Models Compared:
  1. Massless Hill-type Model: The traditional standard (Zajac, 1989).
  2. Mass-Enhanced Hill-type Model: A 1D model where the muscle is divided into 16 serial segments, each containing a contractile element (CE), parallel elastic element (PEE), and a point mass.
• Geometric Scaling: To isolate mass effects, muscles were geometrically scaled by a factor ($s$) ranging from 0.1 to 10.
  • Length ($L$) scaled by $s$.
  • Cross-sectional Area ($A$) and Peak Isometric Force ($F_0$) scaled by $s^2$.
  • Volume and Mass ($m$) scaled by $s^3$.
• Simulation: Both models were driven by the same experimentally derived MTU lengths and activation patterns. This kinematic constraint ensured that any differences in output were solely due to inertial effects.

Analysis:

• Metrics: Root Mean Square Difference (RMSD) between force traces, Coefficient of Determination ($r^2$), and volume-specific net work/power.
• Statistical Analysis: General Linear Model ANOVAs were used to test the effects of scale, movement type (cadence), and muscle on the RMSD of force and power.

3. Key Contributions

• First Application to Human Locomotion: This is the first study to apply a 1D mass-enhanced Hill-type model to a broad range of in vivo human movements (walking, running, hopping, cycling) rather than just isolated muscle experiments or theoretical simulations.
• Quantification of Scale Effects: The study rigorously quantifies how the error in massless models grows non-linearly with muscle size due to the cubic scaling of mass versus the square scaling of force.
• Cadence vs. Load: The study disentangles the effects of movement frequency (cadence) and load, showing that inertial errors are driven primarily by acceleration (cadence) rather than force magnitude (load).

4. Key Results

Force Prediction Differences:

• Human Scale ($s=1$): The difference between massless and mass-enhanced models was minimal (<1%) for normal human-sized muscles during daily activities.
• Large Scale ($s=10$): As muscle size increased, the discrepancy grew significantly. At the largest scale ($s=10$) and high cadence, the normalized RMSD in force reached ~7% of $F_0$.
• Cadence Dependence: Differences were most pronounced in high-frequency tasks (running, hopping, high-cadence cycling). For example, in cycling at 140 rpm with a large muscle scale, errors could exceed 5%.
• Load Independence: Increasing crank torque (load) did not significantly increase the RMSD, whereas increasing cadence did. This confirms that acceleration, not force magnitude, drives inertial effects.

Work and Power:

• Volume-Specific Work: In the mass-enhanced model, as muscle size increased beyond human scale, the volume-specific positive work decreased while the magnitude of negative work increased.
• Net Work: The net mechanical work output per cycle decreased as muscle mass increased, due to energy being consumed to accelerate the internal muscle mass rather than performing external work.

Statistical Findings:

• Significant main effects were found for scale factor and movement type (cadence).
• Participant variability was significant in daily activities (due to natural variations in speed/cadence) but not in constrained cycling tasks.

5. Significance and Limitations

Significance:

• Validation of Simplified Models: For standard human-sized muscles performing daily activities (walking, slow running), traditional massless Hill-type models remain highly accurate. The computational cost of adding mass dynamics is likely not justified for these specific applications.
• Boundary Conditions Identified: Mass effects become non-negligible for large muscles (e.g., in larger animals or pathological hypertrophy) and high-velocity movements (sprinting, high-cadence cycling).
• Theoretical Insight: The study confirms the theoretical prediction that inertial penalties scale with the cube of length, eventually overwhelming the force generation capabilities (scaling with the square of length) in very large or fast-moving systems.

Limitations & Future Directions:

• 1D Constraint: The study used 1D models. Real muscles deform in 3D (bulging), meaning mass accelerates in multiple directions. Previous 3D studies suggest mass effects could be even larger (up to 34% work reduction) than the 1D predictions here.
• Kinematic Constraint: The models were driven by fixed kinematics. In a full dynamic simulation, muscle mass would alter joint kinematics, potentially amplifying or dampening these effects.
• Architectural Simplifications: The models assumed zero pennation angle and uniform scaling, ignoring complex fiber arrangements and aponeurosis behavior which affect internal stress and velocity.

Conclusion:
While muscle mass significantly influences contractile dynamics in theoretical extremes (very large muscles or very high speeds), its impact on force predictions for human-sized muscles during typical locomotion is negligible (<1%). However, for high-performance applications involving high cadences or simulations of non-human (larger) scales, incorporating muscle inertia is essential for accuracy.