Sensitivity of Estimated Muscle Force in Forward Simulation of Normal Walking

Generic muscle parameters are often used in muscle-driven simulations of human movement to estimate individual muscle forces and function. The results may not be valid since muscle properties vary from subject to subject. This study investigated the effect of using generic muscle parameters in a muscle-driven forward simulation on muscle force estimation. We generated a normal walking simulation in OpenSim and examined the sensitivity of individual muscle forces to perturbations in muscle parameters, including the number of muscles, maximum isometric force, optimal fiber length, and tendon slack length. We found that when changing the number of muscles included in the model, only magnitude of the estimated muscle forces was affected. Our results also suggest it is especially important to use accurate values of tendon slack length and optimal fiber length for ankle plantar flexors and knee extensors. Changes in force production by one muscle were typically compensated for by changes in force production by muscles in the same functional muscle group, or the antagonistic muscle group. Conclusions regarding muscle function based on simulations with generic musculoskeletal parameters should be interpreted with caution.