It has been suggested that insulin acts centrally by altering brain glucose uptake. Previous studies of the effect of insulin on brain glucose metabolism have been difficult to interpret due to lack of steady state conditions for glucose and/or insulin. To determine whether insulin per se alters brain glucose metabolism, we measured glucose utilization rates, using the deoxyglucose method, in the medial basal hypothalamus, locus coeruleus, and motor cortex of conscious, unrestrained rats undergoing 2-h euglycemic clamps (blood glucose, 4.1 +/- 0.1 mmol/liter) at increasing insulin infusion rates. Plateau insulin levels were 29 +/- 5 mU/liter (controls) and 48 +/- 4, 146 +/- 8, 670 +/- 40, and 7560 +/- 410 mU/liter (clamped). Glucose utilization rates in the medial basal hypothalamus fell significantly from 60 +/- 4 mumol/100 g X min (controls) to 46 +/- 3, 39 +/- 2, 35 +/- 2, and 39 +/- 3 mumol/100 g X min in respective insulin-infused animals (P less than 0.01 vs. controls). Similar falls of up to 39% and 48% were seen in the locus coeruleus and motor cortex, respectively. A significant inverse correlation was found between the glucose utilization rate in each brain region and the log plasma insulin level. The reduction in glucose utilization rate was associated with marked increases in serum corticosterone levels (995 +/- 157 vs. 91 +/- 31 nmol/liter in controls, P less than 0.001). Serum potassium was significantly lower in clamped animals (5.2 +/- 0.3 to 5.9 +/- 0.3 mmol/liter) than in controls (7.0 +/- 0.4 mmol/liter, P less than 0.01). However, the inverse correlation between regional brain glucose utilization and log plasma insulin was independent of changes in serum potassium, while there was no independent correlation with serum potassium. The data reveal reduced glucose utilization in specific brain regions in the presence of insulin levels both equal to and above those found in the postabsorptive state and support a direct effect of insulin in suppressing regional brain glucose utilization.