A localized brain affection in a 10-year-old girl was closely followed for 2 1/2 years by means of EEG, nuclear brain scans, angiography and computed tomography. Owing to the absence of angiographic evidence of an arterial occlusion, the possibility of an infarct was abandoned and a tentative diagnosis of focal encephalitis, probably viral in origin, was made. The necessity of arriving at such a diagnosis by exclusion of other possibilities rather than by the positive demonstration of the causative agent is discussed. The importance of neuroradiological methods in stressed.