Slow Versus Fast Robot-Assisted Locomotor Training After Severe Stroke: A Randomized Controlled Trial

Background and Purpose: Robot-assisted locomotor training on a bodyweight-supported treadmill is a rehabilitation intervention that compels repetitive practice of gait movements. Standard treadmill speed may elicit rhythmic movements generated primarily by spinal circuits. Slower-than-standard treadmill speed may elicit discrete movements, which are more complex than rhythmic movements and involve cortical areas. Objective: Compare effects of fast (i.e., rhythmic) versus slow (i.e., discrete) robot-assisted locomotor training on a bodyweight-supported treadmill in subjects with chronic, severe gait deficit after stroke. Methods: Subjects (N = 18) were randomized to receive 30 sessions (5 d/wk) of either fast or slow robot-assisted locomotor training on a bodyweight-supported treadmill in an inpatient setting. Functional ambulation category, time up and go, 6-min walk test, 10-m walk test, Berg Balance Scale, and Fugl-Meyer Assessment were administered at baseline and postintervention. Results: The slow group had statistically significant improvement on functional ambulation category (first quartile–third quartile, P = 0.004), 6-min walk test (95% confidence interval [CI] = 1.8 to 49.0, P = 0.040), Berg Balance Scale (95% CI = 7.4 to 14.8, P < 0.0001), time up and go (95% CI = −79.1 to 5.0, P < 0.0030), and Fugl-Meyer Assessment (95% CI = 24.1 to 45.1, P < 0.0001). The fast group had statistically significant improvement on Berg Balance Scale (95% CI = 1.5 to 10.5, P = 0.02). Conclusions: In initial stages of robot-assisted locomotor training on a bodyweight-supported treadmill after severe stroke, slow training targeting discrete movement may yield greater benefit than fast training.