Abstracts

Rationale: There is evidence that functional connectivity (FC) abnormalities in temporal lobe epilepsy (TLE) extend to the whole brain. Indeed, this apparent focal pathology is often referred to as a system disorder with disrupted neural networks. The degree to which FC alterations depend on factors such as age of seizure onset (SO) or the presence of mesial temporal sclerosis (MTS) is largely unknown in TLE.Methods: We explored functional organization in 50 TLE adult patients and 14 normal controls (NCs). Based on their structural MRI, patients were classified as either non-lesional (nTLE) or with MTS (mTLE). Based on the age of their SO, patients were classified as either early or late onset (EO or LO, respectively). To maintain balanced onset groups the threshold differed in the nTLE (25 y.o.; EO/LO, n s=14) and mTLE groups (20 y.o.; EO/LO, n s=11). All the participants underwent a 5-min resting-state fMRI scan. After standard preprocessing of the data, the brain was parcellated into 116 regions and correlations were individually computed between each pair of regions time-series. We investigated 3 graph-theory properties: modularity (M), clustering coefficient (CC) and global efficiency (Eg), at both whole-brain and regional levels. A bootstrapping strategy (5000 repetitions in each group) ensured these measures could reliably be distinguished from random network values.Results: At the whole-brain level, NCs showed the highest functional segregation (M and CC) and the lowest integration (Eg). The EO/LO difference was more pronounced in the mTLE compared to the nTLE group. Of the 4 groups, the most discrepant from NCs was the LO mTLE group, showing both the lowest segregation and highest integration. At the regional level, our data revealed that abnormal FC in the TLE groups was not solely located in the ictal temporal lobe, but involved multiple brain regions (Fig. 1). Each TLE group displayed strong reductions in CC evident in the medial frontal cortex relative to NCs. Both EO groups showed increased CCs in the non-ictal hemisphere relative to NCs. Regarding the integration measure (Eg), the largest differences between the groups were located in the ictal hemisphere, sparsely located through the whole brain (Fig. 2).Conclusions: Our results confirm disturbed FC in TLE, with reduced segregation and increased integration processes relative to NCs. We demonstrate that the impact of age of SO on FC depends on the presence of MTS. Our data support the hypothesis that patients with early SO are more prone to adaptive functional reorganization than patients with late SO, with this early group increasing segregation in the non-ictal hemisphere, perhaps to ensure access to healthy segregated networks remote from seizure activity. MTS reduces plasticity, particularly with seizures later in life, leaving networks more vulnerable, pointing to the importance of hippocampus to adaptive reorganization. Our data show that age of SO alters resting-state global and local FC, with the most negative effects associated with both late SO and the presence of MTS.