Each year 730,000 Americans experience a stroke. Forty percent are left with significant paralysis of one arm. Certain types of physical therapy, for example constraint induced movement therapy (CIMT), have been shown to be effective in improving arm function. However, for most subjects, improvement is modest. In this trial, we test two approaches that may increase the amount of improvement achieved: 1) distributing treatment over a greater amount of time; and 2) adding a drug, d-cycloserine, which theoretically enhances the molecular mechanisms of learning.

Each year, 730,000 Americans experience a stroke. Forty percent are left with persistent impairment of upper extremity function. Although scientifically vetted rehabilitation therapies for this impairment are starting to emerge, current treatment is generally unsatisfactory. Therapies that seek to engage neuroplastic mechanisms constitute one approach to this problem. A good example is constraint induced movement therapy (CIMT), a treatment that seeks, through extensive functional task practice, to overcome an acquired intentional predisposition to use the spared arm (learned non-use), and to improve motor function in the affected arm. CIMT has been tested in a host of trials, most recently a multicenter randomized controlled trial (RCT) - the EXCITE trial. These trials have generally demonstrated that on average, the treatment shows efficacy, and the results from the RCT indicate that it is more efficacious than "standard" therapies. However, problems with CIMT can be readily identified that pose research challenges: 1) on average, efficacy is limited; 2) only a fraction of subjects show substantial benefit. We propose to address these two problems in a pilot RCT of 20 subjects that will test two modifications of standard CIMT: 1) addition of a drug, d-cycloserine, that may enhance neuroplasticity by potentiating NMDA-glutamate receptor-mediated learning mechanisms; 2) delivery of a fixed amount of CIMT over a greater number of days, which according to learning research, may enhance long-term retention of gains.

All subjects in this trial will receive CIMT. Subjects will be randomized to one of 4 groups:

A. CIMT + d-cycloserine, more condensed treatment B. CIMT + d-cycloserine, less condensed treatment C. CIMT + placebo, more condensed treatment D. CIMT + placebo, less condensed treatment The primary outcome measure will be performance on the Wolf Motor Function Test (time) 3 months after completion of treatment.

Inclusion Criteria:

• Age 21-80,
• of either sex,
• diverse ethnic background,
• s/p a single unilateral hemispheric stroke 6 or more months prior,
• who meet upper extremity functional criteria for participation in constraint induced movement therapy.

Exclusion Criteria:

• History of more than minor head trauma,
• subarachnoid hemorrhage,
• dementia or other neurodegenerative disease,
• multiple sclerosis,
• lobar intracerebral hemorrhage,
• epilepsy,
• drug or alcohol abuse,
• serious medical illness,
• serum creatinine >1.5,
• schizophrenia,
• major refractory depression,
• insufficient cardiopulmonary function to participate in low-intensity,
• sustained upper extremity exercise,
• severe visual impairment,
• pregnancy,
• inability to understand the potential risks and benefits of the study,
• personally provide informed consent, and
• understand and cooperate with treatment.

Adverse event data were collected from trial entry through 3 months following treatment completion.