The increasing prevalence of Alzheimer's disease (AD) emphasizes the need for sensitive biomarkers. Memory, a core deficit in AD, involves the interaction of distributed brain networks. We propose that biomarkers should be sought at the level of disease-specific disturbances in large-scale neural networks instead of alterations in a single brain region. This is the first voxel-level quantitative meta-analysis of default mode connectivity and task-related activation in 1196 patients and 1255 controls to detect robust changes in components of large-scale neural networks. We show that with disease progression, specific components of networks are widespread altered. The medial parietal regions and the subcortical areas are differentially affected depending on the disease stage. Specific compensatory mechanisms are only seen in the earliest stages, before symptoms are evident, and could become a functional network biomarker or target for interventions. These results underline the need to further fine-grain these networks spatially and temporally across disease stages. To conclude, AD should indeed be considered as a syndrome involving neural network disruption before cognitive deficits are detectable. (C) 2013 Elsevier Ltd. All rights reserved.
View graph of relations
- Alzheimer's disease, Default mode network, Connectivity, Meta-analysis, Compensation, Biomarker, MILD COGNITIVE IMPAIRMENT, DEFAULT-MODE NETWORK, MONKEY RETROSPLENIAL CORTEX, WHITE-MATTER CHANGES, CINGULATE CORTEX, ASSOCIATION WORKGROUPS, DIAGNOSTIC GUIDELINES, NATIONAL INSTITUTE, BRAIN CONNECTIVITY, ENTORHINAL CORTEX