Abstracts

RATIONALE: The accuracy of interictal scalp dipole source localization may be influenced by multiple generators located outside of the epileptogenic zone (EZ). Methohexital (MHX) may augment dipole source modeling by improving the signal-to-noise ratio (SNR) eliminating fast propagation outside of the EZ. At low doses, MHX is known to activate epileptiform activity. At high doses MHX suppresses epileptiform discharges and the propagation to distant sources. Autonomous epileptogenic lesions remain more resistant to MHX and are the last to be suppressed. These discharges are the first to return during drug elimination. The reader will understand the potential utility of augmenting dipole source modeling with the MHX suppression test (MHXST) enhancing accuracy of depth, location and complexity of the sources of scalp interictal activity.
METHODS: Retrospective data were analyzed from 3 subjects with Landau-Kleffner syndrome. An initial MHXST was performed during Phase I scalp recordings. Scalp EEG was digitally recorded using 26 channels with single density electrode placement. EEG data were imported into BESA2000 (MEGIS, Munich, Germany). Spatio-temporal dipole modeling (STDM) was performed on a 50 msec segment before and after each interictal discharge maximum. The source models were defined by: 1) number of dipoles, 2) temporal delay of dipoles, 3) dipole locations defined by the x, y, z coordinates for the fitted center, 4) dipole orientations, and, 4) global field power (GFP) defined by the goodness of fit (GoF), and residual variance (RV). The GFP is a measure of signal strength against the background noise, and, therefore, represents the SNR.
RESULTS: The scalp MHXST revealed multiple dipole components of discharge morphology not seen during baseline scalp recordings in subjects 1 & 2. Subject 1: 2 baseline surface dipoles modeled the waveform morphology. Dipole 1 was radially oriented, localized in the superficial left sylvian fissure (LSF). Dipole 2 represented rapid propagation to the contralateral hemisphere. During MHX suppression, the dipole model placed the initial radial dipole deeper within the LSF. The first interictal discharge to reappear during drug elimination was modeled by 3 ipsilateral dipoles. The initial radial dipole was localized deepest in the LSF when compared to MHX suppression and baseline interictal data. Subject 2: a single tangential dipole was seen deep within the LSF during baseline scalp recording (GoF=90.630%). Both MHX suppression and initial return of interictal activity were best fit to 3 dipole components. Dipole 1 for each condition was located deeper in the LSF than the modeled baseline scalp dipole. Subject 3: a single tangential dipole was modeled within the LSF during baseline scalp recording (GoF=90.862%). Both MHX suppression and initial return of interictal activity concurred (GoF:93.09% & 87.88%) but with a deeper dipole localization.
CONCLUSIONS: MHX may reveal complex dipoles improving the depth and location of the epileptic focus. Furthermore, MHX at high doses does not activate regions outside of the EZ as suggested by subject 3. The MHXST can be used as a tool to refine scalp dipole modeling by improving the overall SNR of the EZ.