We performed 99mTc-hexamethylpropyleneamineoxime-single-photon emission computed tomography (SPECT) and 18F-fluorodeoxyglucose-positron emission tomography (PET) in 20 epileptic patients with well-lateralized temporal electroencephalographic focus, normal computed tomographic scan, and brain magnetic resonance imaging (MRI) either normal (n = 10) or showing nonspecific changes in the epileptogenic temporal lobe (n = 10). In patients with a normal MRI, PET exhibited focal hypometabolism in 80%, whereas SPECT showed corresponding hypoperfusion in only 20%. In patients with an abnormal MRI, PET and SPECT yielded 100% and 90% sensitivity, respectively. The metabolic and regional cerebral blood flow disturbances were topographically concordant with electroencephalographic and MRI findings in all these patients. Only patients with a large and pronounced hypometabolism on PET images exhibited hypoperfusion on SPECT. Spatial resolution appeared to be the critical factor responsible for the higher sensitivity of PET. However, this superiority of PET did not prove clinically useful in patients whose SPECT was abnormal, particularly when brain MRI showed nonspecific changes in the epileptogenic temporal lobe.