Perceptual sensitivity supports online control, while perceptual prediction errors drive motor memory formation during locomotor adaptation

To maintain stable locomotion, the nervous system must continually adapt, whether reacting to an unexpected perturbation, such as a trip on uneven terrain, or anticipating external demands, such as walking on snow, by forming and updating motor memories. Error-based learning is the dominant computational account of such adaptation, in which motor commands are updated to reduce prediction errors, that is, the mismatch between predicted and actual limb state. Yet this framework has been largely defined in reduced, single-effector paradigms, where the sensory consequences of movement are isolated and the prediction error is directly controlled by the experimenter. Whether the same principle governs whole-body, multi-segmental, multi-sensory behaviors such as walking has remained untested, owing in part to the challenge of identifying a behavioral proxy for prediction errors in this complex, dynamic setting. Here, we combined a split-belt treadmill paradigm with a novel method for quantifying perception of leg motion to address this open question. We found two perceptual contributions to locomotor adaptation: perceptual sensitivity predicted initial motor corrections during early adaptation, whereas perceptual prediction errors (i.e., the mismatch between the actual belt-speed difference and the difference perceived as symmetric) predicted the magnitude of motor aftereffects during post-adaptation. Together, these findings support error-based learning, long characterized in upper limb tasks, as a core computational principle underlying human locomotion.

Error-based learning is the dominant account of how the brain adapts movement, yet it has been defined almost entirely in reduced laboratory tasks, such as cursor rotations, planar reaching, and saccades, where prediction errors are directly controlled by the experimenter. Whether it governs whole-body behaviors such as walking has remained untested. Using a novel measure of leg-motion perception during split-belt walking, we show that two dimensions of perception contribute separably to adaptation. Perceptual sensitivity, a proxy for sensory uncertainty, relates to real-time motor corrections, consistent with Bayesian integration. Perceptual prediction error, a proxy for the prediction error that underlies error-based learning, relates to motor memory formation. These findings establish leg-motion perception as a behavioral window onto the computations driving locomotor adaptation, providing a framework for understanding individual and clinical variability in motor behavior.