Abstracts

Rationale: As a key relay station of neuronal information flow between subcortical structures and cortical networks, the thalamus plays a significant role in seizure propagation in temporal lobe epilepsy (TLE).¹ A common treatment to control drug-resistant epilepsy is the resection of the epileptic focus through procedures such as selective amygdalohippocampectomy (SAH). Prior work has used diffusion-weighted MRI (DWI) to assess white matter microstructural abnormalities in TLE patients before surgery with fixel-based analysis², which quantifies multiple distinct fiber populations within a voxel. Here, we aim to measure changes in the microstructure of the thalamus in TLE patients from pre- to post-SAH surgery.

Methods: This study included 30 healthy controls (35.5 ± 14.2 yrs, 17 males) and 21 patients with right TLE (42.0 ± 10.3 yrs, 11 males). All subjects underwent a T1-weighted scan (1x1x1 mm3) and DWI (2.5x2.5x2.5 mm3, 92 directions, b = 1600 s/mm2) on a 3T MRI. Controls were scanned once, and patients were scanned before and after (mean 27 months) undergoing right SAH. Thalami were segmented from the T1-weighted scan.

Single-tissue constrained spherical deconvolution estimated the fiber orientation distribution (FOD) at each voxel of the DWI data using a group average response function within the thalamus. Controls were used to create a study-specific FOD template to which the patient FODs were registered (Fig 1, top). The characteristic “peaks” of the FODs were then segmented into individual fixels that quantify the magnitude and direction of specific fiber populations within a voxel (Fig 1, bottom).³ This enables quantitative analysis of fiber-specific properties when multiple crossing fiber populations exist within a voxel. A combined measure of fiber density and cross-section (FDC) was computed at each fixel to represent the total axonal integrity in each fiber population. A general linear model was used to evaluate FDC changes from pre- to post-surgery at each fixel using MRtrix3.⁴

Results: We found pre- to post-surgical decreases in FDC of lateral-oriented fixels in regions of the thalamus overlapping with the ventral and pulvinar nuclei ipsilateral to surgery (p < 0.05, Fig 2). We did not find any increases in thalamic FDC after surgery. No changes in FDC were found in the contralateral thalamus.