Difficulty walking is common after a stroke. Although physical rehabilitation helps a little with the improvement of walking ability, recovery is usually incomplete. The purpose of this study is to explore how two different treadmill training approaches influence walking speed, symmetry, and balance in people with chronic severe stroke-related walking impairment. The two approaches involve either forward or backwards treadmill training. This study will look at changes in walking performance and balance, before and after training. This study may lead to more efficient methods for improving walking performance and balance after stroke.

Severe walking impairment after stroke is defined as the inability to walk faster than 0.4 meters per second and impacts nearly twenty-five percent of chronic stroke survivors. Walking speed is a key determiner of community independence, and stroke survivors classified as "severe" are more often symptomatically home-bound with limited mobility and are at higher risk of falls, fractures, and rehospitalizations. A recent study by the investigators suggests that backward locomotion treadmill training (BLTT) may be a promising rehabilitative approach in stroke survivors with severe walking impairment; however, the effect of training duration on behavioral outcomes is unknown.

The objective of this study is to obtain critical pilot data on the effects of extended BLTT on walking speed, symmetry, and balance (static and dynamic), in chronic stroke survivors with baseline severe walking impairment, relative to forward treadmill training controls (FLTT).

Aim 1. Determine the training-related effects of extended BLTT on overground walking speed (primary outcome) in survivors with severe walking impairment. To achieve this aim, we will compare the Pre-Post change in walking speed [10- meter walk test (10MWT)] between groups. The working hypothesis is that extended BLTT will increase walking speed to a clinically meaningful level (≥0.16m/s), compared to the control group (FLTT).

Aim 2. Determine the effects of extended BLTT on walking symmetry and balance. The Zeno Walkway Gait Analysis software will capture Pre-Post changes in temporal gait symmetry index during the 10-MWT. Working hypothesis 2a: BLTT will be associated with a favorable improvement in the temporal symmetry index score. Proprioception and spinovestibular function will be measured with the modified Sensorineural Integration Test (mSIT), and dynamic balance will be assessed with the completion time on the instrumented Timed Up & Go (i- TUG). Working hypothesis 2b and c: BLTT will be associated with a favorable improvement Pre-Post mSIT(a) and completion time on the i-TUG (b).

Inclusion Criteria:

• 18-80 years of age
• Walking speed less than or equal to 0.4 meters/second
• Ability to maintain greater than or equal to 0.3mph speed for 6-minute interval on the treadmill
• Able to walk independently (cane and hemi-walker acceptable)
• Ambulate >10 meters over ground with the Free Step Harness System (as a safety precaution)
• Discharged from formal rehabilitation

Exclusion Criteria:

• Unstable cardiac status which would preclude participation in a moderate-intensity exercise program.
• Significant language barrier which might prevent the participant from following instructions during training and testing.
• Adverse health condition that might affect walking capacity (severe arthritis, significant pulmonary disease significant ataxia, or severe hemi-neglect)
• Severe lower extremity spasticity (Ashworth >2)
• Depression (>10 on the Patient Health Questionnaire, if untreated).

Single (Outcomes Assessor)