We examined the cassette-recorded 24-hour ambulatory EEG findings in children who had either clinically definite seizures or episodic behavioral disturbances not regarded as epileptic on clinical grounds. Among 40 epileptic patients, 22 had one or more attacks during the 24-hour recording session. In 15 of these patients all clinical attacks had appropriate ictal electrographic accompaniments; in another 6 some (but not all) attacks did so. Among 55 children with nonepileptic spells clinically, the 24-hour recording was uninterpretable for technical reasons in one, and in 30 it provided no relevant information because there were no recorded clinical or electrographic attacks. In the remaining 24 patients, one or more clinical attacks were captured, and in no instance was there any accompanying electrographic change. Our findings indicate that the absence of ictal EEG changes during attacks cannot be used in isolation to make a diagnosis of pseudoseizures, but support such a diagnosis made on clinical grounds. The more important role of the ambulatory EEG is to exclude a diagnosis of nonepileptic attacks by demonstrating electrographic seizure activity accompanying typical clinical attacks. Whether the technique will have a useful role in the evaluation of patients when the nature of an episodic disturbance of cerebral function is unclear clinically remains to be established, but will require long-term follow-up to validate any conclusions reached by the electrophysiologic technique.