Abstracts

Rationale: Patients with medically refractory epilepsy may be helped with resective surgery, in which the Epileptogenic Zone (EZ) is completely removed. The effectiveness of the procedure depends on proper definition of the EZ. Recent studies suggest that not a single focus is driving the epileptogenesis, but an epileptic network. Removing or disconnecting the parts of this network, is supposed to lead to seizure freedom. The high frequency band has recently received special interest, because high frequency oscillations are thought to arise from misconnected epileptogenic tissue. The purpose of this study is to find a method based on the graph theory techniques in the gamma band and fast ripple band, that is able to localize the EZ intra-operatively. Methods: We selected patients with superficial lesions and seizure freedom during the follow up period, with a minimum duration of 1 year to enable optimal distinction between channels that were likely to record over epileptogenic tissue en channels covering healthy brain tissue. Patients with hippocampal or mesiotemporal pathologies were excluded. We selected four epochs of two seconds from the ECoG data. The Phase Lag Index (PLI) was calculated for every epoch and used to compute local network measures: Strength and Eigenvector Centrality (EC). The mean strength and EC and their standard deviations for resected and non-resected electrodes were compared using paired t-tests with a significance level of 95%. Results: Intraoperative electrocorticography (ECoG) measurements of twelve patients were analyzed. All patients had focal cortical dysplasias or brain tumors WHO grade I or II. Preliminary comparisons showed that the EC and Strength of resected electrodes in the gamma band was higher than the EC of the non-resected electrodes (EC: p=0,006, Strength: p=0,018). The standard deviation of the EC and Strength of the fast ripple band were lower in the resected electrodes (EC: p=0,009, Strength: p=0,026). Conclusions: The EC and Strength and their standard deviations appear to be promising measures for intraoperative characterization of the epileptogenic network. The measures indicate increased connectivity of the epileptogenic tissue in the gamma band. The networks in the fast ripple band seem to be more stable in the epileptogenic tissue than in the surrounding tissue. We aim to expend the research, investigating more connectivity and network measures. Funding: none