Abstracts

LOCALIZING EPILEPTIC FOCI USING VARIABLE RESOLUTION ELECTROMAGNETIC TOMOGRAPHY (VARETA)

Abstract number : 1.298
Submission category :
Year : 2004
Submission ID : 4326
Source : www.aesnet.org
Presentation date : 12/2/2004 12:00:00 AM
Published date : Dec 1, 2004, 06:00 AM

Authors :
1Kenneth Alper, 2Manoj Raghavan, 3Leslie Prichep, 3Robert Isenhart, 3Bryant Howard, and 3Roy John

Source-analysis techniques have been applied to localize cortical generators of scalp-recorded epileptic spikes and rhythmic ictal potentials. However, source analysis of the EEG background itself has seldom been attempted in partial epilepsy. Variable Resolution Electromagnetic Tomography (VARETA) allows frequency-domain source-localization of background EEG potentials. We sought to determine the value of VARETA in localizing epileptic foci by applying this technique to scalp-recorded EEG from patients with partial epilepsy. We obtained samples of standard digital EEG recordings from 6 adult patients who later underwent invasive intracranial EEG localization of their epileptic foci. We selected two-minute-long samples of artifact-free EEG for analysis. Our analysis first transforms the time-domain EEG data into the frequency-domain. The source generators of the EEG at each frequency is then computed using VARETA and registered in MNI brain coordinates. The magnitude of the generator at each voxel is compared to values based upon a normative EEG database of 305 normal adult subjects and Z-score deviations from norm are computed and displayed. The analysis is repeated at several different EEG frequencies. We applied this method blinded to clinical data from each patient, and compared the results of VARETA analysis to localization based on invasive intracranial EEG, visual analysis of the EEG sample, and 2-D topographic spectral maps generated by conventional Q-EEG techniques. Lateralization and localization of seizure foci to a lobe based on VARETA analysis in the delta and theta range was concordant with the results of intracranial recordings in all of the 6 patients studied. The presence of bi-temporal dysfunction in one instance, and bi-hemispheric extra-temporal dysfunction were correctly predicted by VARETA analysis. By comparison, visual inspection of the EEG samples correctly localized dysfunction to a hemisphere in 5 patients, but was able to localize epileptic dysfunction to a lobe only in 1 of the 6 patients. Visual inspection of the 2-D topographic spectral maps lateralized the focus in 2 patients, while localization to a lobe was possible only in 1. Source localization of background EEG activity using VARETA can potentially lateralize and localize epileptic dysfunction to a lobe. The EEG signal frequencies at which VARETA analysis yielded the best results were in the theta and delta range. At these frequencies, VARETA may provide a novel technique for imaging epileptic foci even in the absence of population-based normative EEG data. (Supported by F.A.C.E.S Foundation, New York, U.S.A.)