Functional activation stimulates CMRglc more than CMRO2 and raises lactate levels in brain. This has been interpreted as evidence that brain work is supported mainly by energy derived from anaerobic glycolysis. To determine if lactate production accounts for the "excess" glucose consumption, cerebral arteriovenous differences were measured in conscious rats before, during, and 15 minutes after sensory stimulation; the brains were rapidly frozen in situ immediately after completion of blood sampling and assayed for metabolite levels. The molar O2/glucose uptake ratio fell from 6.1+/-1.1 (mean+/-SD) before stimulation to 5.0+/-1.1 during activation (P<0.01); lactate efflux from brain to blood was detectable at rest but not during stimulation. By 15 minutes after activation, O2 and lactate arteriovenous differences normalized, whereas that for glucose fell, causing the O2/glucose ratio to rise above preactivation levels to 7.7+/-2.6 (P<0.01). Brain glucose levels remained stable through all stages of activity. Brain lactate levels nearly doubled during stimulation but normalized within 15 minutes of recovery. Brain glycogen content fell during activation and declined further during recovery. These results indicate that brain glucose metabolism is not in a steady state during and shortly after activation. Furthermore, efflux from and increased content of lactate in the brain tissue accounted for less than 54% of the "excess" glucose used during stimulation, indicating that a shift to anaerobic glycolysis does not fully explain the disproportionately greater increases in CMRglc above that of CMRO2 in functionally activated brain. These results also suggest that the apparent dissociation between glucose utilization and O2 consumption during functional activation reflects only a temporal displacement; during activation, glycolysis increases more than oxidative metabolism, leading to accumulation of products in intermediary metabolic pools that are subsequently consumed and oxidized during recovery.