Muscle fatigue modelling: Solving for fatigue and recovery parameter values using fewer maximum effort assessments

Abstract A three-compartment controller model (3CC) predicts muscle fatigue development. Determination of fatigue (F) and recovery (R) model parameters is critical for model accuracy. Numerical methods can be used to determine parameter values using maximum voluntary contractions (MVCs) as input. We tested the effects of using reduced MVC data on parameter solutions using twenty published datasets of intermittent, isometric contractions. The work here examines three sampling variations using approximately half of the MVCs: MVC measurements distributed equally (dMVC), split between the initial and final times (sMVC), and only during the first half (fMVC). Furthermore, solved F and R parameters were used to model fatigue development for three hypothetical task scenarios. Both model parameters and predictions were statistically insensitive to measured data reduction using dMVC, followed closely by sMVC. However, using the fMVC reduction frequently resulted in overestimated parameter values and produced significantly larger prediction errors. We conclude that parameter solutions are robust when using fewer MVCs as long as they are sampled in a manner that captures later fatigue behavior.