This study investigated the influence of biological and technical factors on variations of global and regional cerebral metabolic rate of glucose (CMRglc) measured with 2-[18F]fluoro-2-deoxy-D-glucose ([18F]FDG). Twelve male volunteers (22-40 years) were investigated on three or four occasions for a total of 42 studies. We calculated the variance/covariance of the following parameters: CMRglc, six parameters of the blood clearance of [18F]FDG, hour of injection, peak time of blood radioactivity, and six components of the operational equation (nonradioactive blood glucose concentration, brain radioactivity, two integrals, numerator, and denominator). There was correlation among these six components, except for nonradioactive blood glucose. However, the correlation between the CMRglc and the individual components of the operational equation was poor. The inter- and intrapersonal CMRglc coefficients of variations were 13.8 and 7.1%, respectively. In contrast, coefficients of variations of the numerator and denominator of the operational equation were 34.6 and 32.6%, respectively, and were always in the same direction. No correlation was found between CMRglc and the technical factors in the numerator and denominator of the operational equation. Factor analysis disclosed that a single factor was responsible for 70% of the variance. This factor included caudate, putamen, thalamus, frontal cortex, temporal cortex, and cingulate gyrus. These structures are involved with multiple complex functions, from autonomic motor control to behavior and emotions. The intrinsic metabolic variability of these structures, along with the basal metabolic processes that are continuously going on in the brain, may be the best explanation for the variance encountered in our investigation.