The overall goal of this study is to determine whether robotic-assisted gait training is better than standard physical therapy treatments for improving walking ability in hemiparetic stroke patients.

Subjects enrolled into the study will be randomly assigned to one of 2 groups. The first group will receive one hour of conventional gait training, consisting of lower extremity strengthening exercises, stretching, and full weight bearing walking as tolerated, with appropriate physical assistance from a therapist. The second group will receive walking therapy using a Lokomat, which is a special treadmill that works in conjunction with 2 light-weight robotic arms that assists the subject to move their legs while they try to walk on the treadmill. Some of the subject's body-weight will be supported using a harness. During Lokomat training sessions, subjects will receive feedback of their walking performance on a computer monitor to help them walk. Both groups will be trained for 8-10 weeks, 3 times per week, for 1 hour training sessions (24 total sessions). Resources for transportation to the National Rehabilitation Hospital will be provided to all study participants.

Inclusion:

• Unilateral brain lesion.
• Within 6 months post-stroke.
• Receiving no other therapy targeting function of lower limb.
• Demonstration of hemiparesis (motor dysfunction in lower limb).
• Able to walk 5 meters without therapist assistance (walking device only).
• Able to follow commands and protocol.

Exclusion:

• Significant cognitive or communication impairments.
• Uncontrolled hypertension.
• Uncontrolled diabetes.
• Clinical depression.