Many neuroprotective agents that seemed promising in animal studies of ischemic brain injury prove to have no effect when tested in clinical trials, suggesting that fundamental elements of translational research require better definition. A number of modifications have led to improvements in preclinical and human studies since the earliest controlled trials failed to confirm hypotheses suggested by animal data. Continued re-evaluation and sharing of information derived from the laboratory bench or the patient's bedside should eventually lead to effective neuroprotection in acute stroke. Experimental data should be carefully studied to improve the quality of agents coming to clinical trials and to design trial phasing that effectively determines drug safety and efficacy. This article will examine preclinical modeling and its translation to prospective studies of acute stroke therapy and will focus on some potential solutions directed at clinical trial design.