The neuropathological outcome of metabolic, vascular or mechanical insults to the CNS depends on brain temperature; mild hypothermia is neuroprotective, whereas elevated brain temperature can cause additional neural damage. Studies in both animals and humans have shown that the core and the brain temperature do not always concur with one another. It is therefore important to develop methods for monitoring brain temperature. This paper describes an animal model (the rat) in which we have developed a method to measure, at thermoneutral ambient temperature, the brain and core temperature concomitantly, during different drug treatments. We have used this animal model to study body temperature during fever (induced by human recombinant IL-1 beta, 5 microgram/kg, i.p.), stress-induced hyperthermia (handling of the animal), hypothermia (induced by (+/-)-8-hydroxy-2-dipropylaminotetralin hydrobromide, 0.5 mg/kg, i.p. ) and sleep (non-induced, other than by light and diurnal variation). We show that the thermal curves are similar in the brain and the peritoneum, independent of the thermal state.