Date of Graduation


Document Type


Degree Type



School of Medicine


Physiology, Pharmacology & Neuroscience

Committee Chair

Sergiy Yakovenko

Committee Co-Chair

Valeriya Gritsenko

Committee Member

Jean McCrory


The ability to move in the environment is crucial to the survival of all animals. Neural pathways that control locomotion can be described as a hierarchy, with multiple levels of control, and those ultimately converge on spinal pattern generators. Neural pathways controlling locomotion are hierarchical, highly integrated, and well characterized anatomically, but functional explanations are lacking. Previous computational modeling of the CPG has proposed that they essential signal driving these spinal networks are expressed in the modality of desired velocity. To date, no published research has empirically tested velocity as being the control signal of locomotion. The purpose of this study was to evaluate human ability to discriminate inter-limb velocity on a split-belt treadmill. If the modality of locomotor control signal is indeed velocity then, according to the classical control theory, limb velocity should also be accurately sensed. We tested this hypothesis by probing human ability to detect minute changes in the velocity of each leg. Healthy volunteers with no previous history of neurological conditions or serious musculoskeletal damage to the lower extremities were recruited to walk on a split-belt treadmill with separately controlled belt speeds. Subjects were exposed to randomized asymmetric speeds of left and right legs for approximately 3 steps. A high-pitch cue instructed subjects to report the fastest leg. In addition, we tested velocity discrimination skills in two other conditions when subjects were either supported or loaded by 10% of their body weight. The perception threshold defined as the velocity detected with better than chance probability (above 50%) was 1.02+/-0.43% m/s, with no significant differences between body weight conditions. Variance of step cycle was found to significantly impact probability detection at the differential speed of 0.01 m/s, which is equivalent to the 1% detection level. The accurate velocity discrimination ability supports the idea that the velocity signal is represented within the locomotor control pathways. We propose that errors in this velocity signal are ultimately used to tune heading direction. Solving for the signal controlling locomotion has positive clinical implications, as it could be used in therapies such as locomotor rehabilitation following neurological injury.