Step time asymmetry but not step length asymmetry is adapted to optimize energy cost of split-belt treadmill walking

Key points The relationship between spatiotemporal gait asymmetry and walking energetics is currently under debate. The split-belt treadmill paradigm has been used to study adaptation of spatiotemporal gait parameters in relation to energetics, but it remains unclear why people reduce asymmetry in step lengths, but prefer asymmetry in step times. In this study we characterized the effects of step time asymmetry and step length asymmetry on energy cost during steady-state walking on a split-belt treadmill at increasing speed-differences. Both the optimal and preferred step time asymmetry increased with greater speed differences, while preferred step lengths remained constant and nearly symmetric. Preferred asymmetric step times were energetically optimal across all speed-difference conditions, while preferred step length asymmetry was not optimal. The findings show that humans will adopt an asymmetric gait that is associated with an energy reduction and suggest that step time asymmetry plays a dominant role in shaping the energetic cost of gait asymmetry. Healthy human walking is symmetric and economical; hemiparetic and amputee gait is often asymmetric and requires more energy. Consequently, asymmetry has been attributed to account for the added energy cost of pathological gait. But it is also possible that asymmetric gait may be adopted if it is energetically optimal under certain biomechanical and neurological constraints of the locomotor system. Here, we assessed how preferred asymmetry in step times and step lengths of healthy human gait is adapted during split-belt treadmill walking and tested the hypothesis that asymmetry is adapted to optimize metabolic energy cost. Ten healthy, young participants walked on a split-belt treadmill in three conditions in which the average belt speed was always 1.25 m s(-1)and the speed difference between the belts was 0.5 m s(-1), 1.0 m s(-1)and 1.5 m s(-1)while a range of values of step time asymmetry and step length asymmetry were enforced. We found that preferred step time asymmetry increased with greater speed differences while preferred step length asymmetry remained constant and nearly symmetric. With increasing speed differences participants increased their preferred value of step time asymmetry to coincide with the lowest energy cost. However, our results show that preferred step length asymmetry was not optimal even with extensive experience of split-belt treadmill walking. Overall, our results indicate that humans will adopt an asymmetric gait that is associated with an energy reduction and suggest that step time asymmetry plays a dominant role in shaping the energetic cost of gait asymmetry.