Morphine sulfate effects (30 mg, intramuscularly) on cerebral glucose utilization and subjective self-reports were examined in 12 polydrug abusers by positron emission tomography and [fluorine 18]fluorodeoxyglucose in a double-blind placebo-controlled crossover study. During testing, subjects sat with eyes covered, listening to white noise and "beep" prompts. Morphine significantly reduced glucose utilization by 10% in whole brain and by about 5% to 15% in telencephalic areas and the cerebellar cortex, assuming no contribution of hypercapnia. When the contribution of PaCO2 (45 minutes after morphine was administered) was partialled out, significant morphine-induced reductions persisted in whole brain and six cortical areas. Irrespective of morphine, left-greater-than-right asymmetry occurred in the temporal cortex, and an interaction between hemisphere and drug was noted in the postcentral gyrus. In most cases, effects on glucose utilization were not significantly related to measures of euphoria.