Reliability of dipole models of epileptic spikes

Objective: In order to validate dipole-modeling results, we compared dipole localizations with the distribution of intracerebral potentials occurring simultaneously with scalp EEG paroxysms. Methods: Firstly, scalp EEGs were recorded from 11 patients. Dipole sources were estimated on averaged spikes and projected on 3D-MRIs. Secondly, ster eoelectroencephalography (SEEG) was recorded from implanted electrodes with direct identification onto MRI. Simultaneously with SEEG, control scalp el ectrodes were pasted where spikes peaked during the first session. SEEG was averaged, triggered by the main peak of scalp spikes. Results: SEEG activity during scalp spikes always involved several contacts . In 13 of 14 spikes, maximal fields occurred in neocortical regions. In 4 of 5 cases where intracerebral activity was simple, spikes could be modeled by one source. In all cases where intracerebral activity was complex, spik es had to be modeled by several sources. The main dipole source was 11 +/- 4.2 mm from the SEEG contact showing the maximal intracerebral potential. E arly and late dipole localization and SEEG fields were concordant in two th irds of cases. Conclusion: Results indicate that in our group of patients scalp spikes ref lect activity in large neocortical areas and never activity limited to mesi al structures. Dipole locations and time activation were confirmed most oft en and were more reliable for sources representing the main negative compon ent than for early or late sources. (C) 1999 Elsevier Science Ireland Ltd. All rights reserved.