Individual limb measures of split-belt adaptation are correlated with shifts in lower limb position sense

Stable gait in humans relies on both precise estimates of lower limb position and the ability to robustly adapt walking movements across different environments. However, the specific relationship between sensorimotor adaptation and the perception of lower limb position during gait remains unclear. In this study, we tested a specific theoretical model of how sensorimotor adaptation during gait could cause a shift in the perception of lower limb position. To test this theory, we probed lower limb position sense estimates using a valid and reliable two-alternative forced choice task before and after a long bout of split-belt treadmill walking. The theoretical model assumes that a shift in the perception of limb position is caused by the optimal integration between the predicted and actual sensory feedback of step lengths. Using this model, we generated hypotheses relating to the direction of the shift in limb position sense estimates as well as the relationship between the magnitude of these sensory shifts and the plateau of adaptation when measured using the canonical step length asymmetry index. Like prior studies, participants robustly adapted their step length asymmetry during split-belt treadmill walking. However, we did not observe reliable changes in limb position sense after this adaptation relative to baseline. Furthermore, the changes in limb position sense were not related to the plateau of sensorimotor adaptation. However, post-hoc analysis of other gait parameters, most notably double limb support time, revealed a strong relationship between its adaptation plateau and sensory recalibration. These results suggest that while the specific theorical model we proposed was not supported, sensory recalibration could play an important role in split-belt adaptation and raise questions regarding the error signal driving split-belt adaptation.