Using the [18F]fluorodeoxyglucose method and positron emission tomography, we studied cerebral glucose utilization during sleep and wakefulness in 11 young normal subjects. Each of them was studied at least thrice: during wakefulness, slow wave sleep (SWS) and rapid eye movement sleep (REMS), at 1 week intervals. Four stage 3-4 SWS and 4 REMS fulfilled the steady state conditions of the model. The control population consisted of 9 normal age-matched subjects studied twice during wakefulness at, at least, 1 week intervals. Under these conditions, the average difference between the first and the second cerebral glucose metabolic rates (CMRGlu was: -7.91 +/- 15.46%, which does not differ significantly from zero (P = 0.13). During SWS, a significant decrease in CMRGlu was observed as compared to wakefulness (mean difference: -43.80 +/- 14.10%, P less than 0.01). All brain regions were equally affected but thalamic nuclei had significantly lower glucose utilization than the average cortex. During REMS, the CMRGlu were as high as during wakefulness (mean difference: 4.30 +/- 7.40%, P = 0.35). The metabolic pattern during REMS appeared more heterogeneous than at wake. An activation of left temporal and occipital areas is suggested. It is hypothetized that energy requirements for maintaining membrane polarity are reduced during SWS because of a decreased rate of synaptic events. During REMS, cerebral glucose utilization is similar to that of wakefulness, presumably because of reactivated neurotransmission and increased need for ion gradients maintenance.