Abstract
Introduction: The accuracy of musculoskeletal models and simulations as methods for predicting muscle functional outputs is always improving. However, even the most complex models contain various assumptions and simplifications in how muscle force generation is simulated. One common example is the application of a generalised ("generic") force-velocity relationship, derived from a limited data set to each muscle within a model, uniformly across all muscles irrespective of whether those muscles have "fast" or "slow" contractile properties. Methods: Using a previously built and validated musculoskeletal model and simulation of trotting in the mouse hindlimb, this work examines the predicted functional impact of applying muscle-specific force-velocity properties to typically fast (extensor digitorum longus; EDL) and slow-contracting (soleus; SOL) muscles. Results: Using "real" data led to EDL producing more positive work and acting significantly more spring-like, and soleus producing more negative work and acting more brake-like in function compared to muscles modelled using "generic" force-velocity data. Extrapolating these force-velocity properties to other muscles considered "fast" or "slow" also substantially impacted their predicted function. Importantly, this also further impacted EDL and SOL function beyond that seen when changing only their properties alone, to a point where they show an improved match to ex vivo experimental data. Discussion: These data suggest that further improvements to how musculoskeletal models and simulations predict muscle function should include the use of different values defining their force-velocity relationship depending on their fibre-type composition.