Stroke is a major cause of death and disability in North America. Major efforts have been placed into blocking mechanisms of excitotoxicity following stroke. Immediately after the onset of stroke, a plethora of glutamate is released in the affected area, and activation of NMDA-receptor by glutamate has been attributed to be the main cause of neuronal damage during stroke. Nevertheless, clinical studies of stroke patients given NMDA-receptor blockers have proven to be disappointing. It has been suggested that NMDA-receptor blockers are efficacious only when given prior and/or soon after stroke onset. Once the receptors and downstream death signaling cascades are activated, blocking NMDA-receptor is no longer useful. Since stroke patients often reach the hospital and receive their diagnosis several hours after stroke onset, the conventional therapeutic strategy of blocking NMDA-mediated cell death is not clinically useful. Surprisingly, our preliminary study suggested that only the NR2B subunit-containing NMDA-receptors (NR2BR) promote cell death, and paradoxically, the NR2A subunit-containing NMDA-receptors promote cell survival (NR2AR). We propose that selective activation of NR2AR-dependent cell survival may be a more effective stroke treatment strategy than blocking NR2BR-dependent cell death. In addition, because we are proposing to promote cell survival and not blocking cell death, our treatment should be effective even when administered along time after stroke onset.