Abstracts

Rationale: MRI-based cortical thickness analyses have shown medial fronto-central cortical thinning in temporal lobe epilepsy (TLE) suggestive of damage secondary to seizure spread through thalamo-cortical routes (Bernhardt et al., 2008, Neuroimage; McDonald et al., 2008, Epilepsia). The purpose of this study was to map in vivo structural changes within the thalamocortical circuitry in TLE by: i) locating atrophy at a nuclear level using MRI-based surface-shape analysis and ii) correlating thalamic atrophy to fronto-central cortical thickness measurements. Methods: We studied 37 patients with pharmacoresistant TLE and 19 age- and sex-matched controls. The seizure focus was right-sided in 19 patients (RTLE) and left-sided in 18 (LTLE). T1-weighted high-resolution 1-mm isotropic images were acquired on a 1.5 Tesla scanner using. Thalami were segmented manually in all subjects. Thalamic labels were converted to surface meshes using spherical harmonic parametrization (SPHARM). Thalamic surfaces were rigidly aligned to a template constructed across all subjects. We calculated the signed surface normal component of the displacement vector (in mm) between each individual and the template to quantify local volume changes. Group differences were assessed using two-tailed t-tests. We extracted (Bernhardt et al., Neuroimage, 2008) and correlated mean cortical thickness of the medial fronto-central region with thalamic SPHARM data. Results: Compared to controls, both TLE groups had atrophy in all thalamic divisions, except the ventral in LTLE and the anterior in RTLE. In LTLE, thalamic atrophy was bilateral, whereas RTLE patients displayed atrophy only ipsilateral to the focus (Fig. 1). In controls, medial fronto-central cortical thickness was positively correlated with volume of medial and ventral thalamic divisions (Fig. 2B), whereas in TLE patients the same divisions showed decreased pattern of correlations (Fig. 2C). Conclusions: Although we found widespread thalamic atrophy in TLE, common damage was found in limbic divisions ipsilateral to the seizure focus. This pattern may reflect effects of mesiotemporal degeneration and seizure spread, as the pulvinar and mediodorsal nuclei participate in the generalization of limbic seizures (Rosenberg et al., Epilepsia, 2006; Bertram et al., Epilepsia, 2001). On the other hand, decreased morphometric coupling between medio-ventral divisions and the fronto-central neocortex suggests alternative pathological interactions between mesiotemporal and fronto-central regions.