Gray Matter and White Matter Abnormalities in Temporal Lobe Epilepsy Patients with and without Hippocampal Sclerosis

Abstract

The presentation and distribution of gray matter (GM) and white matter (WM) abnormalities in temporal lobe epilepsy (TLE) have been widely studied. Here, we investigated the GM and WM abnormalities in TLE patients with and without hippocampal sclerosis (HS) in five groups of participants: healthy controls (HCs) (n = 28), right TLE patients with HS (n = 26), right TLE patients without HS (n = 30), left TLE patients with HS (n = 25), and left TLE patients without HS (n = 27). We performed a flexible factorial statistical test in a whole-brain voxel-based morphometry analysis to identify significant GM and WM abnormalities and analysis of variance of hippocampal and amygdala regions among the five groups using the FreeSurfer procedure. Furthermore, we conducted multiple regression analysis to assess regional GM and WM changes with disease duration. We observed significant ipsilateral mesiotemporal GM and WM volume reductions in TLE patients with HS compared with HCs. We also observed a slight GM amygdala swelling in right TLE patients without HS. The regression analysis revealed significant negative GM and WM changes with disease duration specifically in left TLE patients with HS. The observed GM and WM abnormalities may contribute to our understanding of the root of epilepsy mechanisms.

Keywords: FreeSurfer; hippocampal sclerosis; magnetic resonance imaging; temporal lobe epilepsy; voxel-based morphometry.

Figure 1

Significant gray matter volume alterations in the five groups by voxel-based morphometry analysis using SPM12 (family-wise error-corrected at p < 0.05 and an extent threshold K of 50). (A) F-test among the five groups, (B) HC > LTLE-HS, (C) HC > RTLE-HS, and (D) RTLE-no > HC. HC, healthy control; LTLE-HS, left temporal lobe epilepsy with hippocampal sclerosis; RTLE-HS, right temporal lobe epilepsy with hippocampal sclerosis; RTLE-no, right temporal lobe epilepsy without hippocampal sclerosis.

Figure 2

Significant white matter volume alterations in the five subject groups shown by the voxel-based morphometry analysis using SPM12 (family-wise error-corrected at p < 0.05 and an extent threshold K of 50). (A) F-test among the five groups, (B) HC > LTLE-HS, and (C) HC > RTLE-HS. HC, healthy control; LTLE-HS, left temporal lobe epilepsy with hippocampal sclerosis; RTLE-HS, right temporal lobe epilepsy with hippocampal sclerosis.

Figure 3

Distribution of left and right volumes among the five subject groups, obtained from the FreeSurfer results: (A) hippocampal and (B) amygdala. HC, healthy control; LTLE-HS, left temporal lobe epilepsy with hippocampal sclerosis; LTLE-no, left temporal lobe epilepsy without hippocampal sclerosis; RTLE-HS, right temporal lobe epilepsy with hippocampal sclerosis; RTLE-no, right temporal lobe epilepsy without hippocampal sclerosis.

Figure 4

Z score distributions of hippocampal and amygdala volumes among the subject groups. (A) Right hippocampal, (B) left hippocampal, (C) right amygdala, and (D) left amygdala. HC, healthy control; LTLE-HS, left temporal lobe epilepsy with hippocampal sclerosis; LTLE-no, left temporal lobe epilepsy without hippocampal sclerosis; RTLE-HS, right temporal lobe epilepsy with hippocampal sclerosis; RTLE-no, right temporal lobe epilepsy without hippocampal sclerosis.

Figure 5

The range of hippocampal and amygdala volumes among the subject groups. (A) Right hippocampal, (B) left hippocampal, (C) right amygdala, and (D) left amygdala. HC, healthy control; LTLE-HS, left temporal lobe epilepsy with hippocampal sclerosis; LTLE-no, left temporal lobe epilepsy without hippocampal sclerosis; RTLE-HS, right temporal lobe epilepsy with hippocampal sclerosis; RTLE-no, right temporal lobe epilepsy without hippocampal sclerosis.

Figure 6

The result of a negative correlation of gray matter (GM) and white matter (WM) volumes with disease duration among left temporal lobe epilepsy patients with hippocampal sclerosis shown by multiple regression analysis using statistical parametric mapping software (uncorrected at p < 0.001 and an extent threshold K of 250). (A) GM and (B) WM.