Structural brain alterations before mild cognitive impairment in ADNI: validation of volume loss in a predefined antero-temporal region

Abstract

Volume losses in the medial temporal lobe, posterior cingulated, and orbitofrontal region have been observed in Alzheimer's disease (AD). Smaller reductions in similar regions have also been reported in amnestic mild cognitive impairment (aMCI), a canonical precursor to AD. We previously demonstrated that volume loss in bilateral anteromedial temporal lobe is present at baseline in longitudinally followed normal subjects who later developed MCI or AD. In this study we compared grey matter volumes within this predefined anteromedial temporal region (AMTR) at baseline between: 1) normal subjects enrolled in the Alzheimer's Disease Neuroimaging Initiative (ADNI) who subsequently developed cognitive complaints as reflected in a CDR memory box score of 0.5; and 2) normal subjects who remained normal over a median of 48 months of follow-up (CDR sum of boxes 0). We found significantly decreased volume within AMTR in the ADNI memory complainers. To relate AMTR results to those from conventional anatomy, we demonstrate that volumes extracted with the ICBM amygdala region had the best correspondence with AMTR volumes. In contrast, regions that have demonstrated volume loss in frank MCI and AD in ADNI, e.g., the posterior cingulate, did not show volume loss. These findings provide independent confirmation that volume changes preceding MCI occur in AMTR, a region of overlap between amygdala and anterior hippocampus.

Figure 1

A) AMTR Region. Illustration of AMTR showing ICBM single subject T1 axial image without (left) and with (right) overlay of AMTR for this slice level. The AMTR comprises mainly amygdala and hippocampus. B) Processing steps for image normalization. Grey matter segmented images from all 208 subjects were non-linearly co-registered to create a common template (image 2). Individual grey matter segmented images from each subject (image 1) were registered to the template and smoothed with an 8 mm kernel (image 3). Registration preserved grey matter volume in each voxel by incorporating the Jacobian determinant from each individual registration flow field.

Figure 2

A) Extraction of volumes from subject images. The AMTR mask region (image 1) was used to extract voxel values representing grey matter volume from each subject image (images 2 & 3) by isolating only the voxels underlying the mask (image 4). Total extracted volume was the sum of the voxel values scaled by the modulation factor of the image. B) A procedure similar to that used for AMTR (A; above) used masks of registered labeled regions from the ICBM atlas template to extract volumes. The amygdala region is shown in the same sequence of steps as A. This procedure was repeated for each region to calculate volumes (cf. Table 3).

Figure 3

Volume within AMTR is predictive of a later classification of MCI in baseline normal subjects in ADNI with 84% overall accuracy (Later MCI; MC_C). Accuracy was 71% for the entire MC group combined.