Effect of cognitive reserve on amnestic mild cognitive impairment due to Alzheimer's disease defined by fluorodeoxyglucose-positron emission tomography

Abstract

This study aimed to investigate the effect of cognitive reserve (CR) on the rate of cognitive decline and cerebral glucose metabolism in amnestic mild cognitive impairment (MCI) using the Study on Diagnosis of Early Alzheimer's Disease-Japan (SEAD-J) dataset. The patients in SEAD-J underwent cognitive tests and fluorodeoxyglucose-positron emission tomography (FDG-PET). MCI to be studied was classified as amnestic MCI due to Alzheimer's disease (AD) with neurodegeneration. A total of 57 patients were visually interpreted as having an AD pattern (P1 pattern, Silverman's classification). The 57 individuals showing the P1 pattern were divided into a high-education group (years of school education ≥13, N = 18) and a low-education group (years of school education ≤12, N = 39). Voxel-based statistical parametric mapping revealed more severe hypometabolism in the high-education group than in the low-education group. Glucose metabolism in the hippocampus and temporoparietal area was inversely associated with the years of school education in the high- and low-education groups (N = 57). General linear mixed model analyses demonstrated that cognitive decline was more rapid in the high-education group during 3-year follow-up. These results suggest that the cerebral glucose metabolism is lower and cognitive function declines faster in patients with high CR of amnestic MCI due to AD defined by FDG-PET.

Keywords: Alzheimer’s disease (AD); cerebral glucose metabolism; cognitive reserve; education; mild cognitive impairment.

FIGURE 1

Regions with statistically significant decrease (p < 0.01, extent threshold >400) in glucose metabolism PET after adjusting for Alzheimer’s Disease Assessment Scale-Cognitive Component-Japanese version in the high-education group (years of school education ≥13, N = 18) compared with the low-education group (years of school education ≤12, N = 39) on (A) the brain surface projection and (B) multi-planar cross-sectional views. The red-yellow scale indicates the level of statistical significance. The blue crossed lines indicate the maximum peak voxel in the left hippocampus [(–32, –11, –23), T = 5.40].

FIGURE 2

Regions where regional glucose metabolism was inversely associated with years of school education upon controlling for the Alzheimer’s Disease Assessment Scale-Cognitive Component-Japanese version score in a combined group (N = 57) of high- and low-education group. Statistically significant areas (p < 0.01, extent threshold >400) are displayed on (A) the brain surface projection and (B) multiplanar cross-sectional views. The red-yellow scale indicates the level of statistical significance. The blue crossed lines indicate the maximum peak voxel in the left hippocampus [(–32, –11, –24), T = 4.20].

FIGURE 3

A scatterplot representing a significant inverse relationship (R² = 0.189, p < 0.001) between the years of education (x-axis) and regional cerebral glucose metabolism in the left hippocampal cluster detected in the regression analysis (y-axis). Glucose metabolism values were normalized with the mean value in individual images as 50.

FIGURE 4

The results of general mixed linear models, which estimated values at baseline and visit of 3-year follow-up for (A) Mini-Mental State Examination (MMSE), (B) Alzheimer’s Disease Assessment Scale-Cognitive Component-Japanese version (ADAS), (C) Wechsler Memory Scale-Revised, logical memory I (LM1), and (D) Wechsler Memory Scale-Revised, logical memory II (LM2). Displayed p-values are significant for the interaction of time and education (high or low cognitive reserve). The solid and dashed lines connect the mean and 95% confidence interval at baseline and 3 years, respectively. Blue lines: low-education group (years of school education ≤12); Red lines: high-education group (years of school education ≥13).