In Alzheimer's disease, hypometabolism in low-amyloid brain regions may be a functional consequence of pathologies in connected brain regions

In patients with Alzheimer's disease (AD), prominent hypometabolism has been observed in brain regions with minor amyloid load. These hypometabolism-only (HO) areas cannot be explained merely as a consequence of local amyloid toxicity. The aim of this multimodal imaging study was to explore whether such HO phenomenon may be related to pathologies in functionally connected, remote brain regions. Nineteen AD patients and 15 matched controls underwent examinations with [(11)C]PiB-PET and [(18)F]FDG-PET. Voxel-based statistical group comparisons were performed to obtain maps of significantly elevated amyloid burden and reduced cerebral glucose metabolism, respectively, in patients. An HO area was identified by subtraction of equally thresholded result maps (hypometabolism minus amyloid burden). To identify the network typically functionally connected to this HO area, it was used as a seed region for a functional connectivity analysis in resting-state functional magnetic resonance imaging data of 17 elderly healthy controls. The resulting intrinsic connectivity network (HO-ICN) was retransferred into the brains of AD patients to be able to analyze pathologies within this network in the positron emission tomography (PET) datasets. The most prominent HO area was detected in the left middle frontal gyrus of AD patients. The HO-ICN in healthy controls showed a major overlap with brain areas significantly affected by both amyloid deposition and hypometabolism in patients. This association was substantiated by the results of region-of-interest-based and voxel-wise correlation analyses, which revealed strong correlations between the degree of hypometabolism within the HO region and within the HO-ICN. These results support the notion that hypometabolism in brain regions not strongly affected by locoregional amyloid pathology may be related to ongoing pathologies in remote but functionally connected regions, that is, by reduced neuronal input from these regions.