The search for evidence-based treatments has resulted in an exciting new era for neurorehabilitation intervention strategies for stroke. Although stroke rehabilitation research poses many methodologic challenges, evaluation of stroke rehabilitation interventions is clearly moving beyond descriptive and observational studies toward well designed randomized clinical trials. The goals of this article are to summarize issues of trial design for stroke rehabilitation, to discuss promising stroke rehabilitation treatments currently undergoing rigorous evaluation, and to present treatments that may be candidates for randomized clinical trials in the future on the basis of promising preliminary data. Several examples of new developments in neuroscience research that are leading to possible rehabilitation interventions will be discussed. New modalities to evaluate the response of neural networks to rehabilitation interventions are also reviewed.