Effects of gait speed on stability of walking revealed by simulated response to tripping perturbation.

The objective of this work was to study stability of walking over a range of gait speeds by means of muscle-driven simulations. Fast walking has previously been related to high likelihood of falling due to tripping. Various measures of stability have shown different relationships between walking speed and stability. These measures may not be associated with tripping, so it is unclear whether the increase in likelihood of falling is explicable by an increase in instability. Here, stability with respect to a constant tripping perturbation was quantified as the immediate passive response of torso to the perturbation. Subject-specific muscle-driven simulations of eight young healthy subjects walking at four speeds, created by combining a generic musculoskeletal model with gait data, were analyzed. In the simulations, short perturbations were performed several times throughout the swing-phase by applying a constant backward force to the swing-foot of the model. Maxima of changes in the torso (angular) velocity components during the swing-phase were studied. These changes in the velocity components correlated with the walking speed as follows: anterior-posterior r=0.37 (p<0.05), vertical r=0.41 (p<0.05), and medio-lateral r=-0.40 (p<0.05). Of the angular velocity components, only the vertical component correlated statistically significantly with speed, r=0.52 (p<0.01). The weak and varying speed effects suggest that fast walking is not necessarily more unstable than slow walking, in the sense of response to a constant perturbation.