The kinetics of the regional cerebral uptake of [11C]3-O-methyl-D-glucose ([11C]MeG), a competitive inhibitor of D-glucose transport, have been studied in normal human subjects and patients with cerebral tumours using positron emission tomography (PET). Concomitant measurement of regional cerebral blood volume and blood flow enabled corrections for the contribution of intravascular tracer signal in PET scans to be carried out and regional unidirectional cerebral [11C]MeG extractions to be determined. A three-compartment model containing an arterial plasma and two cerebral compartments was required to produce satisfactory fits to experimental regional cerebral [11C]MeG uptake data. Under fasting, resting conditions, normal controls had mean unidirectional whole-brain, cortical, and white matter [11C]MeG extractions of 14, 13, and 17%, respectively. Mean values of k1 and k2, first-order rate constants describing forward and back transport, respectively, of tracer into the first cerebral compartment, were similar for [11C]MeG and [18F]2-fluoro-2-deoxy-D-glucose (18FDG), a second competitive inhibitor of D-glucose transport. k3, a rate constant describing FDG phosphorylation, was 20 times higher for cortical FDG uptake than the k3 fitted for [11C]MeG cortical uptake. Glioma [11C]MeG extractions ranged from normal levels of 12% to raised levels of 30%. Transport of [11C]MeG in and out of contralateral cortical tissue was significantly depressed in patients with gliomas. It is concluded that under fasting, resting conditions, regional cerebral glucose extraction remains relatively uniform throughout normal brain tissue. Gliomas, however, may have raised levels of glucose extraction. The nature of the second cerebral compartment required to describe [11C]MeG uptake is unclear, but it could represent either a useless phosphorylation-dephosphorylation cycle or nonspecific tracer uptake by a cerebral subcompartment.