The neuroprotective effects of hypothermia following cardiac arrest and acute stroke have been demonstrated in experimental models and clinical trials. Experimental studies indicate that improved efficacy and broadened indications can be achieved with moderate to deep hypothermia. However, current techniques require systemic cooling, and are unable to cool rapidly and deeply without serious detrimental effects. A new catheter-based system and technique to rapidly, deeply andselectively cool the brain by selective perfusion of cold blood to the carotid artery was investigated. In vivo studies in large pigs demonstrated cerebral cooling to as low as 15°C with no significant systemic cooling. Initial cooling rates of 1.8°C/min were attained, and were dependent on flow rate and temperature of the perfused blood. Passive rewarming did not result in rebound hyperthermia. No adverse events were observed. In a randomized porcine stroke reperfusion model, a significant reduction in stroke volume was observed by selective cerebral cooling to 26°C compared to a normothermic control group. In initial human experience, selective cerebral cooling to 26°C resulted in excellent outcomes with no neurological deficits in the settings of neurosurgery and out of hospital cardiac arrest. This new catheter-based system and technique shows promise in providing rapid, selective, deep cerebral hypothermia, and may offer an improved method for neuroprotection during neurosurgery, cardiac arrest, acute stroke and other ischemic insult.