Abstracts

Rationale: To investigate the validity of MSI to localize epileptogenic zone by 3D analysis using coregistration of preoperative MSI source onto post-operative MRIMethods: Forty-eight patients with medically intractable focal epilepsy had pre-operatively MSI and epilepsy surgery at The Children s Hospital of Alabama. Twenty-six were excluded because of either of the following:1) less than three spike dipole source estimates were captured, 2) lack of available raw data, or 3) lack of postoperative brain MRI. Twenty-two cases were analyzed. The median age of seizure onset and duration of epilepsy were 3 years (1 month-16 years) and 5 years (1-20 years), respectively. Seven had the visible lesions on the preoperative brain MRI. Sixteen underwent intracranial EEG monitoring. Resective surgery (n=19) and functional hemispherectomy (n=3) were done. The mean post-operative follow-up duration was 4.7 years (1-8 years). Nineteen (86 %) have a favorable outcome (15 Engel class I, 4 class II). The most common pathology was focal cortical dysplasia. MEG was recorded using a whole-head, 148-channel biomagnetometer system (4D Neuroimaging, San Diego, CA). Single equivalent current dipole modeling was used to estimate spike sources. Spike sources were superimposed onto postoperative MRI using 4D Neuroimaging software. The number of clusters of dipoles and proportion of dipoles in resection region calculated as number of dipoles in resection region divided by total number of dipoles were analyzed. Multiple regression analysis was done using postoperative seizure outcome as the dependent variable and number of spike dipole clusters and proportion of dipoles in resection region as the independent variables. To assess relation to postoperative seizure outcome, Fisher's exact test using 70% threshold for proportion of dipoles in resection region was employed. Results: Analysis revealed no significant relation between postoperative seizure outcome and either number of dipole clusters or proportion of dipoles in resection region. Sixteen cases had one (n=6) or two (n=10) clusters in the same lobe. Four cases had spike dipoles clustered in a different lobe or the contralateral hemisphere to surgery. Clustered dipoles with greater than 70% of proportion of dipoles in the resection region was not associated with seizure free outcome (p=0.62). Clustered dipoles, however, were helpful for localizing the epileptogenic zone in 10 of 15 cases with negative MRI and in 6 of 7 cases with positive MRI.Conclusions: Number and proportion of clustered spike dipole sources contained within the surgical resection region does not correlate with seizure-free outcome. This is not surprising since the dipole estimate of a spike source is a theoretical concept and often lies deep to actual cortex responsible for spike generation. Conversely, clustered spike dipole sources provided important localizing information for formulating hypothesis for epileptogenic zone localization, especially when MRI was not localizing.