Gray Matter Deterioration Pattern During Alzheimer's Disease Progression: A Regions-of-Interest Based Surface Morphometry Study

Abstract

Accurate detection of the regions of Alzheimer's disease (AD) lesions is critical for early intervention to effectively slow down the progression of the disease. Although gray matter volumetric abnormalities are commonly detected in patients with mild cognition impairment (MCI) and patients with AD, the gray matter surface-based deterioration pattern associated with the progression of the disease from MCI to AD stages is largely unknown. To identify group differences in gray matter surface morphometry, including cortical thickness, the gyrification index (GI), and the sulcus depth, 80 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database were split into healthy controls (HCs; N = 20), early MCIs (EMCI; N = 20), late MCIs (LMCI; N = 20), and ADs (N = 20). Regions-of-interest (ROI)-based surface morphometry was subsequently studied and compared across the four stage groups to characterize the gray matter deterioration during AD progression. Co-alteration patterns (Spearman's correlation coefficient) across the whole brain were also examined. Results showed that patients with MCI and AD exhibited a significant reduction in cortical thickness (p < 0.001) mainly in the cingulate region (four subregions) and in the temporal (thirteen subregions), parietal (five subregions), and frontal (six subregions) lobes compared to HCs. The sulcus depth of the eight temporal, four frontal, four occipital, and eight parietal subregions were also significantly affected (p < 0.001) by the progression of AD. The GI was shown to be insensitive to AD progression (only three subregions were detected with a significant difference, p < 0.001). Moreover, Spearman's correlation analysis confirmed that the co-alteration pattern of the cortical thickness and sulcus depth indices is predominant during AD progression. The findings highlight the relevance between gray matter surface morphometry and the stages of AD, laying the foundation for in vivo tracking of AD progression. The co-alteration pattern of surface-based morphometry would improve the researchers' knowledge of the underlying pathologic mechanisms in AD.

Keywords: Alzheimer's disease; cognition impairment; gray matter; magnetic resonance imaging; surface morphometry.

Figure 1

Flowchart of the regions-of-interest (ROI)-based surface morphometry analysis. After brain extraction and segmentation (white matter, gray matter, and cerebrospinal fluid), spatial normalization was performed to correct the orientation and the size of the brain. Then, the surface of gray matter was resampled and smoothed. The ROI-based surface parameters were extracted according to the DKT40 parcellation atlas. Cortical thickness, gyrification index (GI), and sulcus depth were used to characterize the deterioration patterns of gray matter during Alzheimer's disease (AD) progression.

Figure 2

Visualization of the DKT40 cortical parcellation atlas, comprising 68 local regions. (A) Top view, (B) Bottom view, (C) Right view, and (D) Left view.

Figure 3

Whole-brain mapping of surface thickness, gyrification index (GI), and sulcus depth maps estimated using CAT12 toolbox. From left to right, each column represents a subject in control, early mild cognitive impairment (EMCI), late MCI (LMCI), and AD groups, respectively.

Figure 4

Comparison of cortical thickness across HC, EMCI, LMCI, and AD groups. (A) The nodal distribution (mean ± SD) of cortical thickness for each group. (B) The Kruskal–Wallis test was performed, and ROIs that exhibit significant difference across four groups were listed. The value of p of the Kruskal–Wallis test is reported in Table 3. Region indexes refer to Table 2. Red crosses denote outliers.

Figure 5

Comparison of the GI across HC, EMCI, LMCI, and AD groups. (A) The nodal distribution (mean ± SD) of the GI for each group. (B) The Kruskal–Wallis test was performed, and ROIs that exhibit significant difference across four groups were listed. The value of p of the Kruskal–Wallis test is reported in Table 3. Region indexes refer to Table 2. Red crosses denote outliers.

Figure 6

Comparison of the sulcus depth across HC, EMCI, LMCI, and AD groups. (A) The nodal distribution (mean ± SD) of the sulcus depth for each group. (B) The Kruskal–Wallis test was performed, and ROIs that exhibit significant difference across four groups were listed. The value of p of the Kruskal–Wallis test is reported in Table 3. Region indexes refer to Table 2. Red crosses denote outliers.

Figure 7

Surface morphometric co-alteration patterns among different cortical subregions. For cortical subregion indices, please refer to Table 2. (A) Binarized Spearman's correlation matrices estimated from the ROI-based surface morphometric metrics, including cortical thickness, GI, and sulcus depth. The correlation matrices were thresholded at the value of 1. The value of 1 indicates that the related subregions share the same decreased trend in surface morphometry (B) Binary networks correspond to the Spearman's correlation matrices in (A). (C) Degree of each cortical subregion estimated from (A).