Three-dimensional localization of interictal epileptiform activity with dipole analysis: Comparison with intracranial recordings and SPECT findings

Seventeen patients with therapy-resistant partial epilepsy were studied in order to compare the localization of the epileptic focus provided by three-dimensional dipole analysis of scalp interictal epilepriform EEG activity to the localization of the ictal pacemaker zone obtained by intracranial recordings. The results were also compared with regional abnormalities in interictal ^99mTc-hexamethylpropy-lenenamine oxime single-photon emission computed tomography (SPECT). Dipole analysis was performed using a well-known commercial computer program based on a three-shell model. For 12 of the 17 patients, the dipoles for all the individual spikes were located in the same area, in two cases, two separate locations were found, and in three cases the dipoles were scattered over a large area. In 8 of the 12 patients with a single dipole location, intracranial ictal recordings revealed a seizure initiation zone corresponding to the dipole location. In another two cases, intraoperative corticography confirmed the dipole findings. In the two remaining patients, the intracranial recordings revealed a different pattern of seizure initiation than had been anticipated from the dipole analysis. In one of the two patients with two dipole locations, intracranial recordings revealed a seizure initiation zone corresponding to one of the two dipole location sites, whereas in the other patient the seizure onset pattern was more complex. In two of the three patients with widely scattered dipole areas, seizure onset was complex, whereas in the third patient a localized seizure onset was found. In 10 of the 12 patients with a single dipole location site, there was good correlation between this site and regions of low flow in the interictal SPECT measurements. In both patients with two dipole areas, there was a low flow area corresponding to at least one of the dipole localizations. In two of the three patients with widespread dipoles, there were major SPECT low-flow areas, whereas in the third patient interictal SPECT was normal. Although our investigations so far are tentative, we think the results indicate that dipole analysis of interictal epileptiform EEG activity may become a useful supplementary method for the localization of the epileptogenic region in patients who are under consideration for surgical therapy, especially if combined with SPECT or other independent localizing techniques.