Abstracts

Rationale: Frontal lobe epilepsy (FLE) is a common form of epilepsy (Lawson et al., Neurology, 2002, 58:723-729) either idiopathic or more often secondary to cortical dysplasia. This leads to great variability in the brain characteristics of patients, raising issues in identifying homogeneous samples for neuroimaging studies. However, patients with non lesional FLE (nlFLE, where there is no clearly identifiable abnormality on magnetic resonance imaging (MRI)) may represent an ideal sample for the study of the epileptic syndrome itself, regardless of the nature and location of the epileptogenic focus. In this study, we applied graph-analysis to diffusion MRI data of nlFLE patients and healthy controls, in order to analyze network, rather than local, properties that may be altered due to the disease. Methods: Twenty-two patients with nlFLE (7 female, mean age 37.0 +- 15.5) and 22 age- and sex-matched healthy controls underwent the same 3 Tesla MRI protocol including whole-brain, 3D T1-weighted, spoiled gradient recall echo (TE/TR = 3.7/ 9.2 ms, flip angle 12°, voxel size= 1×1×1 mm3) and diffusion-weighted MRI (b=1000 s/mm2; diffusion-weighting along 27 non-collinear gradient directions; matrix size 128 × 128; 80 axial slice; number of b0 images = 4; NEX = 2; voxel size = 2 × 2 × 2 mm3). The structural connectome (Bonilha et al., Neurology, 2013, 81:1704-1710) was computed as follows. The AAL90 atlas was used to identify cortical and subcortical regions to be used as nodes of the network. Probabilistic tractography in network-mode was used to obtain connectivity matrices, in which each entry represented the number of probabilistic fibers connecting regions i and j. Entries of the matrices were normalized by the total number of generated fibers, and the following graph-based measures were computed at different network densities [ten values from 0.001 to 0.01]: clustering coefficient, measuring how strongly inter-connected neighboring nodes are; shortest path length and global efficiency, related to the global integration of the network; the nodal efficiencies, related to the importance of a node in the network. All measures were compared through analysis of covariance, with age, sex and intracranial volume as covariates. Results: Patients with nlFLE showed abnormalities in network properties at both global and nodal level. In particular, across all network densities, patients showed significantly lower values of global efficiency (p=0.003) and clustering coefficient (p=0.001), whereas the shortest path length was lower in the control group (Figure 1). At the nodal level, we observed that several regions, not limited to the frontal lobe (Figure 2), had decreased nodal efficiency in patients compared to healthy controls (p < 0.05 corrected). Conclusions: The absence of focal lesions allowed us to explore the characteristics of a uniform sample of patients with nlFLE. This syndrome seems to involve the brain as a network that goes beyond the frontal lobe, rather than to affect a specific region. Funding: No funding was received in support of this abstract