If the efficacy of hypothermia and barbiturates in ameliorating ischemic brain injury lies in reducing the cerebral metabolic rate of oxygen (CMRO2), the greater efficacy of mild hypothermia (34°C) compared with barbiturates is inconsistent with the 15–20% reduction of CMRO2caused by mild hypothermia compared with 50% caused by barbiturates. This paradox, we hypothesized, derives from the fact that whereas barbiturates lower CMRO2associated with EEG activity or thiopental (TP)-suppressible CMRO2, not essential for cellular viability, hypothermia lowers CMRO2associated with providing energy, i.e., adenosine triphosphate, to maintain transmembrane ion gradients or TP-nonsuppressible CMRO2, essential for neuronal viability. To test this hypothesis, we measured whole brain cerebral blood flow (CBF) and CMRO, in two groups of rats mechanically ventilated with 70% N2O/30% O2before and after TP-induced isoelectric EEG. In the normothermic group (n = 7), measurements were made at a brain temperature (Tb) of 38°C, while in the hypothermie group (n = 7), they were made at 34°C. In the normothermic group, TP-induced isoelectric EEG reduced CMRO2by 50%, from 7.92 ± 1.05 to 3.95 ± 0.70 ml 100 g−1min−1(X ± SD). Thus, at 38°C, TP-suppressible and TP-nonsuppressible CMRO2were both 50 ± 4% of total CMRO2. In the hypothermie group, decreasing Tbfrom 38 to 34°C caused a 17% decline in CMRO2, from 7.62 ± 1.92 to 6.28 ± 1.22 ml 100 g−1min−1(p > 0.05). AT 34°C, TP infusion lowered CMRO2to 2.15 ± 0.46 ml 100 g−1min−1. At 34°C, TP-suppressible and TP-nonsuppressible CMRO2values were 64 ± 7% and 36 ± 8% of total CMRO2, respectively. TP lowered CBF by 50% at both 38 and 34°C. In conclusion, mild hypothermia selectively lowers TP-nonsuppressible CMRO2associated with the maintenance of viability rather than EEG-associated or TP-suppressible CMRO2.