In neonates, a definite method of monitoring brain temperature during brain hypothermia has not been established. To estimate the optimal body temperature monitoring during brain hypothermia, we investigated the changes in cephalic vein blood temperature (CVBT), and statistically compared CVBT with the other temperatures such as the tympanic membrane (TMT), nasopharyngeal (NPT), esophageal (EPT) and rectal temperatures (RT) in severe asphyxiated infants with hypoxic-ischemic encephalopathy (HIE). We performed brain hypothermia in four severely asphyxiated infants with HIE. Selective head cooling was performed in two patients, and general body cooling using extracorporeal membrane oxygenation (ECMO) in the other two. In all cases, CVBT, TMT, NPT, EPT, and RT were continuously monitored. Cephalic vein blood temperature during brain hypothermia was 35.0degrees +/- 0.7degreesC (mean SD), TMT 33.9degrees +/- 0.7degreesC, NPT 34.4degrees +/- 0.7degreesC, EPT 34.6degrees +/- 0.6degreesC, and RT 34.5degrees +/- 0.6degreesC. Cephalic vein blood temperature was significantly higher than the other temperatures (P < 0.0001). The differences of CVBT from TMT, NPT, EPT and RT during brain hypothermia were 1.1degrees +/- 0.5degreesC, 0.6degrees +/- 0.4degreesC, 0.5degrees +/- 0.5degreesC, and 0.4degrees +/- 0.4degreesC, respectively. This study suggested that during brain hypothermia, CVBT could be regarded as the optimal body temperature, since brain temperatures seemed to be underestimated when TMT, NPT, EPT, and RT were the markers of body temperatures.