Abstracts

Rationale: There has been a growing interest in using interictal spikes as markers of the epileptiform activity. It has been shown in some studies that regions where interictal spikes originate are correlated to the epileptogenic areas. Previously in our lab, a causality measure was developed to study the connectivity among multiple interictal spike origins and identify the primary epileptic source. In this study, we investigate the concept of combining cortical source imaging and causality estimation to localize epilepsy sources in patients with intractable epilepsy. Methods: 76-channel Electroencephalography (EEG) signals prior to operations were recorded from 7 epilepsy patients, who were not responsive to medications and had to undergo surgical resections. These patients had significant reduction in seizures or became entirely seizure free after operation. Computational analysis was carried out using eConnectome (econnectome.umn.edu), an open-source software developed at the University of Minnesota. 10 interictal spikes from each patient were selected and averaged to improve the signal to noise ratio. The averaged waveforms were then fed into the eConnectome software to estimate cortical current density (CCD) distributions for the duration of a spike. If multiple regions of interest (ROI) were present, a causal analysis, namely directed transfer function (DTF), was applied to identify the primary cortical source. The results were compared to post-operative magnetic resonance imaging (MRI) data and clinical reports which indicate foci identified by the epileptologists. Results: In 6 out of 7 patients, the identified primary source locations are in accord with clinically-defined seizure onset zones. In 1 patient the identified source is localized within the same frontal lobe. Among the 6 patients where satisfactory results were achieved, 4 showed multiple active sites during interictal spikes. The primary sources that have the most information outflow are consistent with the epileptogenic foci identified by the epileptologists. Conclusions: Noninvasive EEG source imaging and connectivity analysis were employed to study cortical generators of interictal spikes. Using eConnectome software, the identified cortical foci of interictal spike activity are overlapping or in close proximity with surgical resection sites. These encouraging results suggest the feasibility of using of interictal spikes in localizing epileptic sources. Further investigation is needed when individual subject s MRI data are used to quantify the estimation error.