Abstracts

Rationale: Hippocampal sclerosis (HS) is the major neuropathological abnormality in patients with temporal lobe epilepsy (TLE), and TLE in the language dominant hemisphere may impair language function. We used functional magnetic resonance imaging (fMRI) to study the language network, as well as functional connectivity of the hippocampus in patients with left-sided HS and right-sided HS, and how it may be disrupted due to the underlying disease. Methods: Analysis was based on fMRI data (3T, verb-generation task) of a homogeneous group of 24 patients with TLE due to MR positive HS (12 left) and 12 healthy controls. Subject specific hippocampal segmentations were obtained with FreeSurfer. Data was preprocessed with SPM12, and functional connectivity (FC) between 360 cortical regions of interests (ROIs) and the hippocampal ROIs was quantified with Pearson’s correlation, using only task active timepoints. We performed seed based FC analysis of language related ROIs and of each hippocampus. Functional dissociation of hippocampal regions was based on an analysis of similarity in connectivity patterns (cosine similarity). For each group, the dissociation between anterior and posterior hippocampus was quantified with a paired t-test. Results: FC analysis of the language network showed significant (p < 0.05) impairment only for lTLE. FC analysis from the hippocampal seeds revealed widespread cortical anticorrelations from the diseased hippocampus, and cross-correlation analysis with a lag of 5 time points hinted at a delayed connectivity. Analysis of the dissociation between anterior and posterior parts of the hippocampus revealed a shift towards the posterior part for the diseased hippocampus. Controls and rTLE showed similar dissociation patterns for the left hippocampus. The anterior part was more similar to motor and insular regions, as well as regions associated with the default mode network. The posterior hippocampus showed a higher similarity to frontal, temporal and parietal regions. Patients with lTLE showed a decreased similarity of the left anterior hippocampus, while increased similarity of the posterior hippocampus hinted at a shift in connectivity. For the right hippocampus, in controls, mainly similar connectivity patterns of the anterior part were found. For lTLE and rTLE, similarities to the posterior part of the hippocampus were increased, hinting at a shift in connectivity for both groups. Conclusions: Our analysis revealed significant impairment of the language network only in patients with lTLE. During language forming, the diseased hippocampus exhibited widespread cortical anticorrelations, and cross-correlation analysis hinted at a delayed connectivity from the diseased hippocampus. Analysis of the dissociation between the anterior and posterior parts of the hippocampus revealed more similar connectivity patterns in the posterior part of the affected hippocampus. In lTLE this was also observed in the right hippocampus, while rTLE showed similar connectivity patterns as controls in the left hippocampus. Funding: -