Cerebral glucose utilization measured with fluorine-18-fluoro-2-deoxy-D-glucose is characterized by considerable variability both among different persons and for the same person examined on different occasions. The goal of this study was to explore whether some regions of the brain were more variable than others with respect to glucose utilization and whether there was a pattern in their covariance. The global and regional cerebral utilization of glucose was measured in 12 healthy young volunteers on 3 or 4 occasions. In all, 24 regions were examined. The interrelation of the glucose utilization rates of the brain regions was investigated by factor analysis of the metabolic rates. Some 70% of the total variance was attributable to only 1 factor, while 80% of the total variance could be attributed to 2 factors. Regions making up the first factor were the frontal and temporal cortex, cingulate gyrus, caudate nucleus, thalamus and putamen. These regions are functionally related to the limbic system. Regions of the second factor were the parietal cortex, occipital cortex and cerebellum, regions more clearly related to sensory and motor functions. The 2-factor pattern was highly reproducible, being found with different algorithms for factor extraction and rotation. Under resting conditions, the variance of cerebral metabolism seems to be primarily related to regions which are closely involved with the limbic system. Cortical regions involved primarily in motor and sensory functions have less influence on the variance.