BACKGROUND Although selective brain hypothermia is expected to be a promising neuroprotective treatment, the thermal distribution under hypothermia is not fully investigated. We applied selective head cooling to seven newborn piglets under general anesthesia in order to investigate the mechanism of cooling. METHODS Seven healthy, large white piglets aged within 5 days after birth were studied. Temperatures were monitored at the superficial brain (0.5 cm), deep brain (2.0 cm), scalp skin, nasopharynx, tympanum, esophagus, and rectum. A radiant heater and a warmer blanket were used to maintain the normal rectal temperature (38.5-39 degrees C). For the first piglet, the coolant temperature was widely changed from 15 degree C to - 20 degree C in order to define the practical range. Subsequently, the coolant temperature was set at 10 degree C, 0 degree C, and - 10 degree C for the remaining six piglets. The target deep brain temperature was set at 35 degree C, as the same reduction of brain temperature might provide moderate brain hypothermia in the human neonate. RESULTS With 0 degree C coolant temperature, the deep brain temperature was cooled to 35 degree C; however, the scalp skin attached to the cooling cap became broadly blotchy and injured in all animals. When we induced minimal systemic hypothermia by 1C for a cohort of three piglets, the deep brain temperature decreased in parallel with the rectal temperature, which enabled us to achieve the target temperature with 10 degrees C coolant without injuring the scalp skin. The scalp skin and nasopharyngeal temperatures were good predictors of both superficial and deep-brain temperatures throughout the experiment. CONCLUSIONS Our results suggest that moderate brain hypothermia may be applied to newborn infants without inducing moderate systemic hypothermia.