Abstracts

Rationale: The success of temporal lobe epilepsy (TLE) surgery relies on accurate localization of the presumed epileptogenic zone (EZ). In patients with inconclusive non-invasive studies extended electroencephalogram evaluation with intracranial electrodes is required (phase II evaluation). Previous diffusion tensor imaging (DTI) studies reported white matter degeneration beyond the identified EZ, which may be helpful in epilepsy surgery planning. Interictal high frequency oscillation (HFO) including ripples (80-150 Hz), can detect the EZ with better sensitivity and specificity than conventional EEG spikes. We hypothesized that white matter changes in TLE patients will be the most severe in the EZ defined by increased rates of HFO ripples. The purpose of this study is to evaluate the predictive power of the combination of HFO measurements and DTI parameter changes of intracranial macroelectrodes for EZ localization in TLE patients with phase II evaluation. Methods: Fourteen adult patients (mean age 41.0y SD+/-12.1y, 8F/6M), with refractory bilateral or unilateral TLE without hippocampal sclerosis were selected. Six patients had lesional abnormalities on conventional brain MRI. Fourteen individually age and gender matched neurologically and psychiatrically healthy controls with normal DTI were included. Eight to twelve intracranial electrodes were implanted in each patient bilaterally into mesial and lateral temporal lobes by MR guided stereotactic procedure. Image postprocessing included coregistration of MRI to the CT with intracranial electrodes. Two region of interests (ROIs) were selected for each macroelectrode. Mean fractional anisotropy (FA) of tracts seeded by ROIs was measured on the DTI images (Figure 1). Patients' ROIs were coregistered to controls' DTI images. Ripples were acquired for each macroelectrode during interictal sleep over 60min using a Hilbert detector and visual inspection. Mean ripple rate/min was calculated for each macroelectrode-ROIs. Results: We found significant negative correlation (strongest in the hippocampus) between mean tract FA and ripple rates in 140 electrodes using mixed effect model (p=0.04). Age, gender, epilepsy duration and lesions had no confounding effects on the correlation. There was no correlation between controls' mean tract FA and patients' ripple rates. Conclusions: This is the first DTI study measuring white matter integrity at intracranial macroelectrode sites and correlating these measurements with neurophysiological signals of HFOs. We demonstrate that reduced tract FA is associated with increased ripple rate at intracranial electrode sites. Combining DTI and HFO measurements may result in improved precision for localizing epileptogenic brain regions. Funding: Dr. Orosz is supported by the Otfrid-Foerster grant of the German Epilepsy Society. Dr. Weiss is supported by the Epilepsy Foundation Award: Research and Training Fellowship. Dr. Engel is supported by the NINDS grant of NS033310, Dr. Staba by NS071048 and Dr. Fried by NS033221.