We propose a study that uses challenging walking exercises as a research training program and compare balance and walking abilities against a non-challenging exercise program, in a group of 40 people with long standing (> 6 months) weakness that occurred after a stroke. Our main balance measure will be changes with the Berg Balance Score and Dynamic Gait Index, and our main walking measure will be walking speed over a 10 meter walkway and distance walked over a six minute period. Also, we will measure balance confidence, using scores on Activities-specific Balance Confidence Scale, changes in quality of life as measured by the Geriatric Depression Scale, SF-36 and Stroke Impact Scale. If challenging exercises are shown to be more effective than non-challenging exercises, then we will share this promising new approach with the community in hopes of improving people's lives after a stroke.

In 2009, the American Heart Association (AHA) reported that the incidence of hemiplegia in patients six months post-stroke and over 65 years of age was 50%. Stroke survivors with chronic hemiplegia are at an increased risk for falling due to poor motor control, muscle weakness, and balance problems. While over-ground walking training has been shown to improve muscle coordination and functional movement outcomes in stroke survivors, the physical challenges to balance during the training is limited due to safety concerns. The limited training does not reflect the individuals' natural environment, which studies have shown to contain hazards that put post-stroke individuals at greater risk of loss of balance and falls. Therefore, it is important for clinicians to safely implement challenging environmental-hazard tasks as a way to effect greater improvements in walking capability post-stroke.

The usage of body weight support (BWS) during treadmill training has been shown to improve walking speed, but without the context of real world hazards, individuals may not gain improvements in balance related tasks nor gain confidence in moving through hazardous environments. The purpose of this study is to introduce a novel, falls-based training (FBT) approach that will enable individuals to be challenged at high levels of balance and walking safely, using a new robotic device called the KineAssist®, and to determine its relative effects compared to traditional body weight support treadmill training (BWSTT). We hypothesize that like the standard BWSTT, the novel FBT will result in improved walking speed, greater 6 minute walking distance, and greater Berg Balance scores. We also hypothesize that FBT will result in greater gains than BWSTT after the training, with a greater difference 6 months after training. In addition to the walking tests and the Berg Balance scale, we will compare the outcomes of the Stroke Impact Scale (SIS), the Geriatric Depression Scale (GDS)SF-36 Health Survey, the Activities-Specific Balance Confidence (ABC) Scale, the K-9 task Balance test, and the Dynamic Gait Index (DGI).

Inclusion Criteria:

• Community dwelling unilateral stroke survivors, aged 19 years or older, 4 months to 5 years post incident, residual hemiplegia, who are able to ambulate at least 14m with an assistive device or the assistance of one person, with receptive and expressive communication capability, approval of physician, and voluntarily provided informed consent.

Exclusion Criteria:

• Significant and acute medical conditions, amputations, spasticity management that included phenol block injections within 12 months or botulinum toxin injections within 4 months of the study, any cognition involvement that impairs the ability to follow directions for, and plans to move out of the area within the next year or no transportation to the study area.