Thalamic structural connectivity in medial temporal lobe epilepsy

Abstract

The thalamus has been implicated in various stages of medial temporal lobe epilepsy (MTLE) seizure evolution. The relative density and functional significance (in epileptogenesis) of thalamic projections to MTL subregions, however, remains to be determined. This study used structural and diffusion magnetic resonance imaging (MRI) to evaluate thalamic connection density with distinct MTL subregions in terms of location and volume. Nineteen MTLE patients with unilateral hippocampal sclerosis (HS; 12 right; 10 female) were compared to 19 age-matched controls. Five regions of interest (ROIs) per hemisphere were created in native space: thalamus, amygdala, entorhinal cortex, hippocampus, and parahippocampus. Separate probabilistic tractography analyses were performed between the thalamus and each ipsilateral MTL subregion (four per hemisphere). Individual connectivity profiles and regional volumes were assessed. The medial pulvinar consistently showed the highest connection density with the hippocampus in healthy controls and in MTLE patients. Decreased thalamic connected volume was observed for thalamohippocampal pathways in patients with MTLE, and indicates pathway-specific deafferentation. Regional hippocampal and thalamic atrophy was also observed, indicating gray and white matter loss in the thalamohippocampal pathway. Consistent localization of dense medial pulvinar (PuM) connectivity with the hippocampus suggests chronic PuM stimulation could modulate the MTLE seizure network. Decreased thalamic connected volume is a promising biomarker for epileptogenesis that merits longitudinal validation. A PowerPoint slide summarizing this article is available for download in the Supporting Information section here.

Keywords: Diffusion-weighted imaging; Hippocampal sclerosis; Magnetic resonance imaging; Thalamic connectivity.

Figure 1

Structural and diffusion MRI image volumes were used to create ROIs for subsequent probabilistic tractography. (A) depicts an individual subject diffusion weighted scan with five regions per hemisphere: thalamus (green), hippocampus (yellow), parahippocampus (red), amygdala (blue), and entorhinal cortex (orange). Probabilistic tractography was performed in native space from thalamic seeds to each ipsilateral MTL target. (B) depicts individual subject center of gravity of maximum thalamic connection density with the hippocampus, located consistently within the medial pulvinar (B, outlined in blue) across all subject groups (center of gravity plotted in MNI152 space; control = blue; L-MTLE = green; R-MTLE = red). Although center of gravity did not vary between groups, individual variations were observed.

Figure 2

Reduced regional volume (above) and thalamic connected volume (below) in MTLE patients compared to controls. L, left; R, right; Amyg, amygdala; Ento, entorhinal cortex; Hipp, hippocampus; Para, parahippocampus; Thal, thalamus; *p < 0.05 with false discovery rate correction for multiple comparisons (Mann-Whitney test of specified MTLE group vs. controls, see also Table S1). Volume is reported as mm³.