Walking and running are essential tasks people take for granted every day. However, these are highly complex tasks that require significant neural control. This is complicated by the inherent redundancy of the nervous system and by physiological noise. Humans may adopt different control strategies to achieve different goals (environmental or task specific). More specifically, walking/running on a treadmill only requires that one not walk off the treadmill. Of the many possible strategies that can achieve this goal, humans attempt to maintain a constant speed from each stride to the next (Dingwell, John et al. 2010). However, how humans alter the stride-to-stride regulation of their gait when the task goals change (e.g., by maintaining stride length and/or time, during running, or during a predicted walk to run transition speed) has not yet been demonstrated. In the first two of three experiments conducted, healthy adults either walked or ran on a motorized treadmill at a comfortable speed under the following conditions:​ constant speed, constant speed with the stride length goal (targets on the treadmill), constant speed with the stride time goal (metronome), or constant speed with both stride length and stride time goals. In a third experiment, subjects walked and/or ran at a comfortable speed and also at their predicted theoretical walk to run transition speed. Goal functions derived from the task specifications yielded new variables that defined fluctuations either directly relevant to, or irrelevant to, achieving each goal. The magnitude of the variability, as well as the stride-to-stride temporal fluctuations in these variables, were calculated.

During walking, subjects exploited different redundancy relationships in different ways to prioritize certain task goals (maintain stride speed) over others (maintain stride length or stride time) in each different context. In general, subjects made rapid corrections of those stride-to-stride deviations that were most directly relevant to the different task goals adopted in each walking condition. Thus, the central nervous system readily adapts to achieve multiple goals simultaneously.

During running, subjects exhibited similar adaptations to walking, but over-corrected to prioritize maintaining stride speed even more strongly. This suggests that stepping control strategies adapt to the level of perceived risk. This purposeful adaptability of these stride-to-stride control strategies could be exploited to developing more effective rehabilitation interventions for patients with locomotor impairments.

During the predicted walk-to-run speeds, subjects were able to largely exploit the redundancy within task goal, and effectively operated at “uncomfortable” speeds. These results suggest that the stride speed control is robust even with additional novel tasks and uncomfortable, abnormal speeds of locomotion.