Functional network modularity and efficiency supports episodic memory in older adults with amyloid‐beta pathology

Abstract

AbstractBackgroundEpisodic memory is one of the first cognitive domains to decline in preclinical Alzheimer’s disease. We investigated whether graph metrics of functional networks, specifically modularity and local efficiency, are related to episodic memory and amyloid‐β (Aβ) positivity in cognitively normal older adults.MethodSixty‐five cognitively normal older adults (mean age 72.3 ± 6.2 years; 62% female) underwent high‐resolution resting state fMRI (1.8mm3, partial acquisition) to measure functional connectivity, 18F‐Florbetapir‐PET to measure Aβ pathology, and cognitive assessment. Resting state fMRI data was processed with CONN. ROIs (n=135) from the Brainnetome Atlas contained within rsfMRI coverage were used to generate ROI‐to‐ROI functional connectivity matrices. Graph measures of modularity (Q) and local efficiency were computed with the Brain Connectivity Toolbox using binary graphs at costs ranging from 5%‐50% of network sparsity (5% step size). Aβ status was determined using a >1.11 SUVR threshold of a composite region spanning the association cortex, resulting in 32% of the sample being classified as Aβ+. Episodic memory was measured using the immediate recall measure of the Rey Auditory Verbal Learning Test (RAVLT A5).ResultAβ+ participants had significantly increased modularity compared to Aβ‐ (Fig1A, p<0.05 for costs 25 and 50), but no differences in local efficiency (Fig1B). There was no significant effect of Aβ positivity on RAVLT A5 performance (t(63)=‐0.67, p = 0.51). However, there were significant interactions between Aβ status and RAVLT A5 with both modularity and local efficiency at most cost functions (Fig2). Better RAVLT A5 performance was associated with higher modularity (Fig2A; ex. cost 25: r = 0.54, p=0.01) and higher local efficiency (Fig2B; ex. cost 25: r = 0.63, p=0.002) within Aβ+ participants, but not Aβ‐ participants (ps>0.36).ConclusionOur findings suggest that increases in functional network modularity and local efficiency within Aβ+ cognitively normal older adults may act as a compensatory mechanism to maintain equivalent levels of episodic memory performance.