Differential effects of aging and Alzheimer's disease on medial temporal lobe cortical thickness and surface area

Abstract

The volume of parcellated cortical regions is a composite measure related to both thickness and surface area. It is not clear whether volumetric decreases in medial temporal lobe (MTL) cortical regions in aging and Alzheimer's disease (AD) are due to thinning, loss of surface area, or both, nor is it clear whether aging and AD differ in their effects on these properties. Participants included 28 Younger Normals, 47 Older Normals, and 29 patients with mild AD. T1-weighted MRI data were analyzed using a novel semi-automated protocol (presented in a companion article) to delineate the boundaries of entorhinal (ERC), perirhinal (PRC), and posterior parahippocampal (PPHC) cortical regions and calculate their mean thickness, surface area, and volume. Compared to Younger Normals, Older Normals demonstrated moderately reduced ERC and PPHC volumes, which were due primarily to reduced surface area. In contrast, the expected AD-related reduction in ERC volume was produced by a large reduction in thickness with minimal additional effect (beyond that of aging) on surface area. PRC and PPHC also showed large AD-related reductions in thickness. Of all these MTL morphometric measures, ERC and PRC thinning were the best predictors of poorer episodic memory performance in AD. Although the volumes of MTL cortical regions may decrease with both aging and AD, thickness is relatively preserved in normal aging, while even in its mild clinical stage, AD is associated with a large degree of thinning of MTL cortex. These differential morphometric effects of aging and AD may reflect distinct biologic processes and ultimately may provide insights into the anatomic substrates of change in memory-related functions of MTL cortex.

Fig. 1

Illustration of novel semi-automated medial temporal cortical morphometric measurement protocol. Slices of greyscale MRI volume data are used to manually identify landmarks for rostrocaudal boundary definitions (left). Surface reconstruction models are used to manually identify sulcal and gyral landmarks for mediolateral boundary definitions (right). Grey–white and pial surface boundaries generated using automated processing (yellow and red lines, respectively, in left image) are checked for accuracy and then used to measure surface area, mean thickness, and volume of cortical regions of interest. Dark purple = ERC, yellow = PRC, light purple = PPHC. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Fig. 2

Bar graph illustrating morphometric differences of ERC volume, surface area, and cortical thickness between Normal Younger, Normal Older, and Mild AD groups. Measurements displayed were obtained from the average of both hemispheres. Error bars represent standard error. *p < 0.05 compared to Normal Younger group; **p < 0.05 compared to Normal Older group.

Fig. 3

Bar graph illustrating morphometric differences of PRC volume, surface area, and cortical thickness between Normal Younger, Normal Older, and Mild AD groups. Measurements displayed were obtained from the average of both hemispheres. Error bars represent standard error. The legend used is identical to the one used in Fig. 2.

Fig. 4

Bar graph illustrating morphometric differences of PPHC volume, surface area, and cortical thickness between Normal Younger, Normal Older, and Mild AD groups. Measurements displayed were obtained from the average of both hemispheres. Error bars represent standard error. The legend used is identical to the one used in Fig. 2.

Fig. 5

Scatterplots illustrating relationships between WMS Logical Memory II performance and entorhinal and perirhinal thickness for Mild AD and Normal Older subjects. Thickness measurements for each subject were calculated as an average of both hemispheres. Regression lines are displayed for Mild AD (*p < 0.05) in bold and for Normal Older subjects as dashed lines.