The VHA estimates that over 15,000 Veterans incur a stroke each year. As the population of older Veterans grows, stroke will become an increasingly important problem to the VHA. Recovery of walking function is the most common goal of stroke survivors. The proposed study will test whether training with accurate walking tasks to engage the damaged supraspinal motor pathways is more effective than training with steady state walking. The investigators expect that training accurate tasks will be more effective, thereby improving walking function of Veteran stroke survivors and reducing the burden of care placed on families and on the VHA. Furthermore, this rehabilitation approach can be accomplished at comparable cost to existing rehabilitation approaches, which is important given that the VHA invests $88 million per year toward outpatient care, including physical rehabilitation, in the first six months after stroke. This research also has implications for rehabilitation of other neurologically injured populations, including traumatic brain injury and incomplete spinal cord injury.

Current approaches for rehabilitation of walking following stroke do not sufficiently restore mobility function. For instance, fewer than 50% of individuals with stroke-induced walking dysfunction recover the ability to walk independently in the community. New breakthroughs in rehabilitation are needed that will target the motor impairments responsible for poor walking function in individuals post-stroke. Functional recovery can occur in response to task-specific neuroplasticity of damaged brain circuitry. The corticospinal tract is an important target for neuroplasticity because it plays an important role for control of walking in humans. Research has shown that, compared to steady state walking, accurate gait modification (ACC) tasks are a potent behavioral stimulus for activating the corticospinal tract. Therefore, the investigators propose that training with ACC tasks (e.g., obstacle crossing/avoidance, accurate foot placement, etc.) may be superior to training with steady state walking (SS) for eliciting corticospinal neuroplasticity and recovery of walking function. Most rehabilitation paradigms have previously focused on SS training. This is largely because therapists consider it premature to progress to ACC tasks when persistent deficits of steady state walking still remain. However, this reasoning might be counter-productive, because training only steady state walking may not sufficiently stimulate neuro-plasticity of the damaged corticospinal pathway. In contrast, ACC training is specifically designed to stimulate corticospinal neuroplasticity. Importantly, since ACC training targets a central mechanism, its benefits are expected to generalize across walking conditions. Furthermore, it is expected to benefit most stroke survivors who possess at least a minimal residual capability to activate the corticospinal tract. ACC training also provides an opportunity to practice tasks that are analogous to challenges encountered in the home and community environments. Accordingly, there is strong mechanistic and practical rationale for ACC training.

A number of earlier studies have cumulatively established exciting preliminary evidence showing that walking function is enhanced by training with ACC tasks. However, no prior study has been specifically designed and sufficiently powered to determine the extent to which the "accurate gait modification" ingredient is crucial for recovery of walking function. Also not known is the extent to which ACC training reduces the neural impairments underlying poor walking function. The central hypothesis of this study is that ACC training will be superior to SS training for increasing walking function and for reducing underlying neural control of the paretic leg in adults with post-stroke hemiparesis. Each intervention will involve twelve weeks of training, 3 days per week (36 sessions total), and will emphasize the motor learning principles of high intensity, repetition and task-specificity. Assessments will be conducted immediately pre-intervention, immediately post-intervention and at a follow-up session 3 months later. Walking function will be measured in the lab and in the "real world". Neural impairment measures will include electromyography-based measures of inter-muscular coordination and corticospinal drive.

The investigators expect that the benefits of ACC training will justify larger randomized controlled trials to optimize the use of ACC training, including timing relative to stroke, combination with other therapeutic approaches, and identifying individuals who are most likely to benefit from this approach. This research is expected to enhance walking function in stroke survivors, including for the 15,000 Veterans who suffer a stroke each year.

Inclusion Criteria:

• occurrence of a single unilateral stroke within the previous 6-18 months (verified by MRI or CT from medical record)
• living in the community and able to travel to training and assessment sites
• approval of participation by primary care physician
• age 18 - 80 years
• lower extremity paresis indicated by Fugl-Meyer Assessment Score < 30
• deficit in at least one "synergy" subcategory (II - IV) of the Fugl-Meyer Assessment
• self-selected 10m gait speed of 0.4 - 0.8 m/s (with or without an ankle/foot orthosis or cane)
• able to provide informed consent
• willingness to be randomized to either intervention group
• written approval by primary care physician

Exclusion Criteria:

neurological disorder or injury (other than stroke) such as Parkinson' s disease or spinal cord injury

• severe arthritis, such as awaiting joint replacement, that would interfere with study participation
• cardiovascular disease (congestive heart failure, significant valvular disease, history of cardiac arrest, presence of an implantable defibrillator, uncontrolled angina)
• myocardial infarction or major heart surgery in the previous year
• cancer requiring treatment in the past three years, except for nonmelanoma skin cancers and other cancers having an excellent prognosis (e.g., early stage breast or prostate cancer)
• lung disease requiring use of corticosteroids or supplemental oxygen
• renal disease requiring dialysis
• current diagnosis of schizophrenia, other psychotic disorders, or bipolar disorder
• Mini-Mental State Examination (MMSE) score <23
• major depression (Patient Health Questionnaire score > 10)
• severe obesity (body mass index > 35)
• uncontrolled hypertension (systolic > 200 mmHg and/or diastolic > 110 mmHg)
• uncontrolled diabetes with recent diabetic coma or frequent hypoglycemia
• bone fracture or joint replacement in the previous six months
• diagnosis of a terminal illness
• current participation in physical therapy or cardiopulmonary rehabilitation
• significant visual impairment affecting capability to gauge movement accuracy
• previous enrollment in a clinical trial for recovery of walking function
• current enrollment in any clinical trial
• planning to relocate out of the greater Gainesville FL area during the study period
• unable to communicate sufficiently with study personnel
• clinical judgment regarding safety or noncompliance

Adverse events were monitored over a period of approximately 6 months, from enrollment in the study until completion of the 3-month follow up assessment.