APOE-ε4 selectively modulates posteromedial cortex activity during scene perception and short-term memory in young healthy adults

Abstract

Apolipoprotein E (APOE) ε4 is a major genetic risk factor for Alzheimer's disease (AD), yet the mechanisms by which APOE-ε4 influences early-life brain function, and hence, in turn, risk for later-life AD, are poorly understood. Here, we report a novel, and selective, pattern of functional brain activity alteration in healthy young adult human APOE-ε4 carriers. Our findings suggest that APOE-ε4 may influence vulnerability to poorer later life cognitive health via its effect on posteromedial cortex (PMC), a hub region within a brain network involved in spatial processing, and necessary for episodic memory. In two neuroimaging tasks, APOE-ε4 carriers showed an inability to effectively modulate PMC during scene, but not face and object, working memory and perception. This striking pattern overlaps both functionally and topographically, with the earliest cognitive deficits seen in clinical AD, as well as reported alterations in the default network in amyloid-positive individuals at increased risk of AD.

Figure 1

(a) A schematic of the one-back visual working memory task used in Task A. This comprised four conditions, scenes, faces, objects, and scrambled objects, and required participants to indicate when there was an immediate repeat of an item (indicated here by ‘R’). Activity for each separate stimulus condition was contrasted between APOE-ε4 carriers and APOE-ε4 non-carriers to identify regions within posteromedial cortex (PMC) sensitive to APOE-ε4 status. Scenes were created using Deus Ex (Ion Storm L.P., Austin, TX, USA, with software development package Deus Ex Software Development Kit v1112f); faces were generated with FaceGen Modeller 3.3 (Singular Inversions Inc., Toronto, ON, Canada); the objects shown in the figure are examples of the types of trials used in the task and are taken from the Hemera Photo-Objects 50,000, Volumes 1–3. (b) Examples of scene, face, object and squared blocks (baseline) odd-one-out trials from Task B (asterisks indicate the odd item). Scene photographs were taken by J.P.S.; faces are part of the Psychological Image Collection at Stirling (PICS. http://pics.stir.ac.uk/); objects are from the Hemera Photo-Objects 50,000, Volumes 1–3.

Figure 2. In the one-back visual working memory task (Task A), a selective group difference between APOE-ε4 carriers versus non-carriers in the scene condition (constrained to an a priori PMC ROI) was evident in (a) posterior cingulate cortex (PCC), (b) precuneus, and (c) cingulate.

For the analyses undertaken in Task B, these significant clusters were binarised to generate independent regions-of-interest (ROIs). For information, in Task B, there were no clusters that survived the whole-brain threshold used in Task A (Z > 3.1, p > 0.05). All statistical maps are presented on the MNI152 template image.

Figure 3. In the independently defined PCC ROI (left figure), extracting percentage signal change values for each odd-one-out condition separately compared to the squared blocks odd-one-out baseline revealed significantly greater activity (i.e., a failure to deactivate PCC) in APOE-ε4 carriers relative to APOE-ε4 non-carriers for scenes only (right figure).

*p = 0.001 (Bonferroni-corrected critical  = 0.017).

Figure 4. Percentage signal change values extracted from (a) the precuneus and (b) the cingulate ROIs in the scene, face and object conditions of the odd-one-out paradigm (compared with the squared blocks odd-one-out baseline).

⁺p = 0.03 (Bonferroni-corrected critical  = 0.017).