Functional network structure supports resilience to memory deficits in cognitively normal older adults with amyloid-β pathology

Older adults may harbor large amounts of amyloid-beta (A beta) pathology, yet still perform at age-normal levels on memory assessments. We tested whether functional brain networks confer resilience or compensatory mechanisms to support memory in the face of A beta pathology. Sixty-five cognitively normal older adults received high-resolution resting state fMRI to assess functional networks, 18F-florbetapir-PET to measure A beta, and a memory assessment. We characterized functional networks with graph metrics of local efficiency (information transfer), modularity (specialization of functional modules), and small worldness (balance of integration and segregation). There was no difference in functional network measures between older adults with high A beta (A beta+) compared to those with no/low A beta (A beta-). However, in A beta+older adults, increased local efficiency, modularity, and small worldness were associated with better memory performance, while this relationship did not occur A beta- older adults. Further, the association between increased local efficiency and better memory performance in A beta+older adults was localized to local efficiency of the default mode network and hippocampus, regions vulnerable to A beta and involved in memory processing. Our results suggest functional networks with modular and efficient structures are associated with resilience to A beta pathology, providing a functional target for intervention.