Biofeedback to Increase Propulsion During Walking After Stroke

Background/Purpose: Approximately 15,000 Veterans are hospitalized for stroke each year. Impairments of motor control and the subsequent functional limitations in ambulation are the most common manifestations and regaining the ability to walk is the number one stated goal of Veteran stroke survivors. Forward propulsion of the body's center of mass is a cardinal feature of gait that depends on the generation of appropriate anterior-posterior ground reaction forces. Decreased propulsive force generation by the paretic limb of stroke survivors has been identified through both simulation and cross-sectional studies as a major contributor to walking dysfunction. Extrinsic verbal feedback from a therapist is the standard approach used during gait retraining to improve propulsion generation. However, this key component of gait is not directly observable by therapists and patients are often unable to sense propulsion generation due to impaired intrinsic feedback, specifically deficits in somatosensation and proprioception, hindering recovery of paretic propulsion and compromising walking function. The objective of this study is to provide preliminary evidence that biofeedback as an adjuvant to therapists' verbal feedback will improve propulsion and enhance walking function for Veterans post-stroke. Subjects: Thirty individuals \> 6-months post-stroke will participate. Additional study criteria include: 1) Ambulation of household distances without physical assistance to advance or support the paretic leg; 2) Unilateral leg paresis confirmed by a score of \< 32 on the Fugl-Meyer Motor Assessment; 3) Step length asymmetry (paretic \> non-paretic step length); 4) Ambulation without an assistive or orthotic device. Methods: Participants will be randomized to either an experimental group that will train with propulsion biofeedback from commercially available pressure-sensitive insole sensors (Biofeedback group; n=15) or a control group that will train with standard therapist-provided verbal feedback alone (Standard group; n=15). The 12 session (3X/week for 4 weeks) gait training intervention will be delivered by a physical therapist-led team. For participants in the Biofeedback group, prior to the first intervention session, the baseline amount of pressure exerted by the paretic forefoot during late stance will be determined. The insole area underlying the forefoot will then be calibrated to produce a tone when pressure exceeds 5% of this baseline pressure. This threshold will be progressively increased at regular intervals throughout the intervention period to ensure participants are training at their challenge-point to improve propulsion of the paretic limb. An insole of similar thickness will be worn in the shoe of the non-paretic leg for symmetry and comfort but will not produce a tone during the intervention. Those in the Standard group will not wear insoles during intervention but will receive verbal feedback alone regarding propulsion of the paretic limb during gait training. Therapist-provided verbal feedback will be used to instruct participants on achieving and/or maintaining appropriate movement patterns that contribute to propulsion generation. For both groups, the therapist will choose from a standardized bank of gait activities, suitable to each participant's ability level. The goal for total walking time for each session will be 50 minutes: 5, 10-minute bouts with a 2-minute rest between each bout. Outcome Measures: Paretic limb propulsion is the primary outcome measure. Secondary measures include the Six Minute Walk Test, Functional Gait Assessment, Fall Self-Efficacy, temporal-distance gait measures and gait kinematics, all of which will be measured pre- and post-intervention. Data Analysis Plan: Descriptive statistics will be provided for all outcome measures. To identify the effect of the intervention, differences between the pre- and post-training assessment within each group (i.e. change scores) will be calculated. To test the hypotheses, the investigators will apply independent sample t-tests to the change scores of the Biofeedback and Standard group. Hypothesis testing will be conducted at a two-sided p \< 0.05 level.