Predictive musculoskeletal simulations reveal the mechanistic link between speed, posture and energetics among extant mammals

In this work we created 3D computational models of the human musculoskeletal system across body masses ranging from 0.1 kg to 2000 kg (Fig. 1c). Musculoskeletal properties in each model were scaled with geometric similarity such that segment masses scaled as m1.0 , linear dimensions of bones and muscles scaled as m0.33 , and muscle maximum isometric forces scaled as m0.67 (i.e., proportional to cross-sectional area). We generated predictive muscle-driven simulations in these models across a range of gait speeds (Movie S1) and examined how maximum speed and cost of transport vary across this 20,000-fold size range. We show that the speed-mass relationship displays a hyperbolic pattern whereby locomotor speed is maximised at an intermediate body mass, and that the simulated gaits at different body masses can reproduce the variation in posture and cost of transport (COT) seen among the phylogenetically broad group of extant species for which this information is available.