Treadmill training with harness support is a promising, task-oriented approach to restoring locomotor function in individuals with post-stroke hemiparesis, but a scientific basis for the proper selection of training parameters is needed. The goal of this dissertation is to provide a biomechanical basis for the selection of training parameters (i.e., body weight support, treadmill speed, support stiffness, and handrail hold) to improve the gait pattern practiced by hemiparetic individuals during treadmill training.

By comparing gait characteristics of hemiparetic and non-disabled individuals at matched treadmill speeds, non-speed-related gait deviations associated with post-stroke hemiparesis were identified. In the hemiparetic subjects, leg kinetic energy at toe-off in the paretic limb was reduced, resulting in increased percentage swing time and reduced peak knee flexion during swing, consistent with inadequate leg propulsion by the plantarflexors or hip flexors during swing initiation. Energy cost associated with raising the trunk during pre-swing and swing of the paretic limb was exaggerated, consistent with compensatory pelvic hiking to clear the paretic limb with reduced knee flexion. Leg kinetic energy at toe-off in the non-paretic limb was exaggerated, resulting in reduced swing time, consistent with weakness or poor balance during single limb support on the paretic limb.

The adjustment of each training parameter during treadmill walking in the hemiparetic subjects was found to improve a specific set of the gait deviations identified. With increased body weight support or the addition of handrail hold, the exaggerated leg kinetic energy at toe-off and reduced swing time in the non-paretic limb were improved, resulting in increased single limb support time on the paretic limb. With increased treadmill speed, the reduced leg kinetic energy at toe-off in the paretic limb was improved but remained low relative to values in the non-paretic limb. With increased support stiffness, the exaggerated energy cost associated with raising the trunk was improved. We conclude that the proper selection of training parameters can improve the gait pattern practiced by individuals with post-stroke hemiparesis during treadmill training. These actions, ostensibly, may improve treatment outcome.