We evaluated the role of positron emission tomography (PET) with [18F]deoxyglucose (FDG) (FDG-PET) for planning surgery in 53 patients who had temporal lobectomy for uncontrolled seizures at National Institutes of Health from 1981 to 1990. Investigators blinded to PET data used results of telemetered video-electroencephalographic ictal monitoring and other standard criteria to decide whether subdural electrodes (22 patients, i.e., the "invasive" group) should be implanted or surgery performed. PET scans were analyzed using a standard regional template. Mean lateral but not mesial temporal asymmetry was significantly higher in patients who became seizure free (p < 0.03). Patients with > or = 15% hypometabolism were significantly more likely to be seizure free in the entire study population and the invasive subgroup. Visual identification of hypometabolism was less accurate. When a clear temporal ictal surface electroencephalographic focus was present, FDG-PET provided less additional information. FDG-PET may be particularly valuable if the surface electroencephalographic scan is nonlocalizing. In addition to helping to identify the seizure focus, it may allow limitation of invasive electrode placement to those necessary for functional mapping. When PET is used to identify epileptic foci, quantitative measurements of asymmetry should be made.