Cerebral protection by hypothermia is commonly attributed to cerebral metabolic suppression. However, at temperatures below 28 degrees C, the relationship of temperature to cerebral metabolic rate of oxygen consumption (CMRO2) has not been well characterized. Accordingly, the relationship between brain temperature and CMRO2 was determined in eight dogs during cooling from 37 to 14 degrees C while the EEG was continuously monitored. Cardiopulmonary bypass was initiated and control measurements were made at 37 degrees C during anesthesia with nitrous oxide 50-60% inspired and morphine sulfate 2 mg.kg-1 intravenously (iv). Upon cooling to 27 degrees C, the nitrous oxide was discontinued and the morphine was antagonized with naloxone 2 mg iv. Measurements were repeated at 27, 22, 18, and 14 degrees C and in four dogs again at 37 degrees C after nitrous oxide 50-60% had been reestablished at 27 degrees C along with administration of morphine sulfate 2 mg.kg-1. For each temperature interval, the temperature coefficient (Q10) for CMRO2 was calculated (Q10 = CMRO2 at x degrees C divided by CMRO2 at [x - 10] degrees C). Between 37 and 27 degrees C the Q10 was 2.23, but between 27 and 14 degrees C the mean Q10 was doubled to 4.53. With rewarming to 37 degrees C, CBF and CMRO2 returned to control levels, and brain biopsies revealed a normal brain energy state. During cooling, the EEG developed burst suppression at or below 22 degrees C. With further cooling, the periods of suppression increased; however, burst activity continued in seven of eight dogs even at 14 degrees C.(ABSTRACT TRUNCATED AT 250 WORDS)