We report a detailed electroclinical analysis of 320 seizures recorded by foramen ovale electrodes in 77 potential candidates for selective temporal lobe surgery because of antiepileptic drug-resistant seizures. The exact localization of the origin of seizure discharges, the electroencephalographic (EEG) seizure onset patterns, transhemispheric propagation, propagation time, duration of discharge, laterality of discharge termination, postictal focal slowing, correspondence between foramen ovale recordings and the scalp EEG, and the influence of antiepileptic drug modifications were studied and correlated with the clinical seizure semiology and with postoperative outcome following selective amygdalohippocampectomy. In general, the foramen ovale electrode technique provided good neurophysiological information in candidates for selective amygdalohippocampectomy. The following ictal signs predicted a good surgical outcome: (a) unilateral and anterior mediobasal temporal lobe seizure onset, (b) short seizure duration, (c) no or infrequent contralateral seizure discharge propagation, and (d) if propagation to the contralateral mediobasal temporal lobe occurred, the postoperative outcome was better the later the contralateral mediobasal temporal lobe was affected. Postoperative outcome was also better the less frequently contralateral interictal spikes occurred. No direct predictive value could be attributed to the presence of an initial arrest reaction.