Abstracts

Rationale: Electric source imaging (ESI) of interictal epileptiform activity is gaining interest as one technique for localizing the epileptic focus in patients with pharmaco resistant partial epilepsy. Recent studies have demonstrated the importance of a sufficient number of recording electrodes, and also the importance of selecting the correct time point for the analysis in order to avoid involvement of propagated areas. Furthermore, spike averaging entails several problems, and a technique involving statistical analysis of individual spikes would be more accurate. The purpose of this study was to suggest an analysis protocol that would incorporate these different considerations, and to evaluate the feasibility and clinical yield of this protocol for source imaging of interictal epileptic activity. Methods: 50-100 interictal spikes are marked in the routine Video EEG long term monitoring (24-48 hours), recorded from 128/256 electrodes, and visually categorized. All spikes of a category are processed using two different paths: (1) Spatio-temporal segmentation of averaged spikes using k-means clustering (Cartool software) is used to identify periods in the spike-wave complex with a quasi-stable voltage field (microstates). Two source models (LAURA/LORETA, implemented in Geosource, (Electrical Geodesic Inc)), constrained to the gray matter, are applied to the average map of each microstate. (2) The same two source models are applied to all time frames (spike peak +/- 500 ms) of each individual spike. The source reconstruction results are analyzed statistically, so that each timeframe during the spike-wave complex is compared to a baseline period. The results from the segmentation procedure (path 1) can be used to get an estimation of the propagation patterns of the spike, whereas the statistical evaluation (path 2) gives a more detailed estimation of the time and area of spike onset. Results: In a first pilot study interictal spikes were recorded with 256 electrodes in 6 patients who were subsequently subject to intracranial recordings. The ESI results were compared to the ictal onset zone as determined form the intracranial recordings. In 4/5 cases in whom a single spike category was found, the ESI results indicated early activation of the ictal onset area, as subsequently demonstrated in the intracranial recordings. In one patient with a single spike category the ESI indicated a spike focus clearly different from the ictal onset area. In the last patient two different spike foci were found, one of which corresponded to the ictal onset area. Conclusions: We found a good correspondence between interictal ESI results and results from intracranial recordings. We think that the current protocol for interictal source imaging (high density EEG, segmentation and statistical evaluation) will be an improvement compared to source imaging of single timepoints of averaged spikes recorded with a standard electrode montage.