Abstracts

Rationale: Currently, neuroimaging in focal epilepsy identifies substrates of seizure-generating brain regions, e.g., signal changes on MRI and dysmetabolic areas on radionuclide studies. EEG and MEG constitute the only direct ‘imaging' modalities that capture the electrical disturbance that is the defining feature of epilepsy. We report here on preliminary results obtained with a novel rapid MRI sequence that directly probes the local field potential slowing that is a virtually universal characteristic of epileptogenic cortex. Methods: Six patients with proven focal epilepsy underwent brain imaging on a Philips 3T MRI scanner. Six normal adult volunteers served as controls. Five patients had clear-cut unilateral lesions on previous structural MRI with concordant seizure semiology and/or video-EEG (VEEG) studies that unambiguously localized their epilepsies. A sixth patient's MRI-negative temporal lobe epilepsy had been confirmed by VEEG, neuropsychological evaluation and PET imaging. Standard anatomical sequences (volumetric T1, T2 and FLAIR) were obtained prior to single-slice EPI acquisitions in a sequential fashion through multiple coronal planes, including through lesions when present, with the following parameters: number of dynamics=3100; intersample time=36.12ms, TR=17.93ms; TE=10ms; slice thickness=10mm, FOV=256×256mm²; flip angle=20^o.Both magnitude and phase images were collected. The study was approved by the Institutional Review Board of UTHSC-Houston. Results: The MRI magnitude time-series signal showed prominent pulsatile vascular effects near major vessels and the ventricles, that on the raw images could be visualized in real-time as an angiographic or CSF-flow ‘movie'. Away from vascular structures, faster frequency activity was prominent. Phase MRI data also showed pulsation effects, with faster frequency activity less prominent than on magnitude images. In the five patients with epileptogenic lesions, comparison of phase data between lesional and contralateral homologous cortical volumes showed a clear tendency for waveforms to be ‘simpler' (relative deficit of higher frequency amplitudes) on the lesional side (Figure). This observation was also true when phase time-series of temporal lobe volumes were right-left compared in the single MRI-negative patient. Conclusions: Sequential rapid single-slice EPI yields a novel MR brain signal with rich detail at a resolution of 10 Hz or better. This MRI signal incorporates vascular, respiratory and CSF-flow effects, but also subtle effects at higher frequencies which may originate in the field-gradient effects of the macroscopic local field potential. Our results suggest that the excess of slow rhythms in epileptogenic tissue is reflected in the smoother profile of the MRI phase signal. Successful development of these preliminary ideas - true 3-D electrical imaging of the brain - will yield a novel disease-specific MRI modality of use in patients who are ‘MRI-negative' by current technology. Acknowledgement: GPK acknowledges funding from the National Institute of Neurological Disorders and Stroke (1K23NS079900-01).