Evidence of Energetic Optimization during Adaptation Differs for Metabolic, Mechanical, and Perceptual Estimates of Energetic Cost

The theory that the sensorimotor system minimizes energetic cost during locomotion has long been supported by both computational models and empirical studies. However, it has yet to be determined if the behavior to which people converge when exposed to a novel perturbation during locomotion is also energetically optimal. We address this issue in the context of adaptation to walking on a split-belt treadmill, which can impose a left-right asymmetry in step lengths. In response to this asymmetry, participants gradually adjust their foot placement to adopt steps of equal length. Here, we characterized metabolic, mechanical, and perceptual estimates of energetic cost associated with a range of asymmetries to determine whether symmetry is the energetically optimal strategy for walking on a split-belt treadmill. We found that taking steps of equal length did not minimize metabolic cost or mechanical cost. In addition, perceptual estimates of cost were not sensitive to changes in asymmetry. However, symmetry was identified as the optimal strategy when energetic cost was estimated from a composite metric that combined both metabolic and mechanical costs. These results suggest that adaptation may arise from optimization of a composite estimate of effort derived from feedback about the interaction between the body and environment.