V3. Idiopathic generalized epilepsy shows widespread increased network connectivity

Abstract

Introduction Idiopathic generalized epilepsy (IGE) is a subtype of epilepsy presumed to have a genetic etiology and characterized by generalized seizures, i.e. seizures starting and rapidly engaging distributed networks. Although routine magnetic resonance imaging is apparently normal, many studies have reported structural alterations in idiopathic generalized epilepsy patients using diffusion tensor imaging and voxel based morphometry. Changes were also reported in functional networks during generalized spike wave discharges. In this study we wanted to investigate the IGE brain networks during the discharge-free intervals using Magnetoencephalography (MEG). To study the characteristics of the network, we used a graph theoretical approach. Methods We recorded resting state MEG data from 13 (9 females, mean age 38.6 years) IGE patients and 19 (11 females, mean age 38.5 years) healthy controls. Epileptic discharges were identified and marked by a neurologist and excluded from the analysis. Source localization of the resting state activity was performed using dynamic imaging of coherent sources (DICS) beamforming. Connectivity was estimated between all sources in the gray matter using the imaginary part of coherence. Another low-resolution network was generated by averaging connections between sources grouped by anatomical labeling. Graph measures were calculated on both the high resolution and low resolution networks. Results Network connectivity was significantly increased in IGE patients in the high-resolution networks in beta1 ( p = 0.003) and beta2 ( p = 0.0003) bands as compared to controls. In the low-resolution networks, connectivity in beta1 and beta2 was significantly higher in patients than in control subjects ( p = 0.005 and p = 0.0003 respectively). The further analysis of nodal strength using cluster-based statistics yielded significant clusters in the beta1 and beta2 bands. In the beta1 band, three significant clusters were found with higher nodal strength in IGE patients than in controls ( p = 0.004, p = 0.011 and p = 0.029 respectively). In the beta2 band, four significant clusters were found ( p = 0.0004, p = 0.002, p = 0.005 and p = 0.034 respectively) (see Fig. 1 ). We found no significant clusters of higher connectivity in controls compared to IGE patients in any of the frequency bands. Using Network Based Statistics, we found significant subnetworks with higher connectivity in IGE patients. The differences were found in alpha, beta1 and beta2 bands (see Fig. 2 ). We did not find any significant correlation between global connectivity and the age of participants. Conclusions Using graph theoretical network analysis, we found a widespread increase in connectivity in IGE patients compared to controls. These changes were more pronounced in the motor network and the mesio-frontal cortex. We did not, however, find any significant difference in clustering coefficient and characteristic path length which are measures that represent architectural differences in the network. Our findings suggest that an increased resting state connectivity in idiopathic generalized epilepsy could be a mechanism of seizure generation and represents a functional brain imaging endophenotype of the disease.