Structural and functional connectivity uniquely contribute to explained variance in episodic memory performance in cognitively normal older adults

Abstract

AbstractBackgroundGiven mounting evidence that AD‐related neuropathological changes may be present in cognitively normal individuals, there is a need to examine brain biomarkers related to variability in cognition that may precede cognitive decline. Although prior studies demonstrate age‐related changes in structural connectivity (SC) and functional connectivity (FC) separately (Contreras et al., 2015), few have adopted a multimodal approach to test the potential interplay between SC and FC, and their relative associations with cognition. Here, we examined the independent contributions of SC and FC to individual differences in episodic memory performance in cognitively normal older adults.MethodThis cross‐sectional study included 107 cognitively normal participants from the BIOCARD study. SC was indexed using radial diffusivity (RD) and mean diffusivity (MD) of two limbic white matter tracts (fornix and hippocampal cingulum), as measured by diffusion‐weighted MRI. FC within five large‐scale brain networks (control, default, limbic, dorsal attention, and salience/ventral attention) was derived using resting state fMRI. Hierarchical linear regression was utilized to directly compare the relative contributions of SC and FC to individual variability in a composite delayed episodic memory score, while also accounting for age, sex, cerebrospinal fluid (CSF) Aß42 and total tau, and gray matter volumes of the entorhinal cortex and hippocampus.ResultsWhen all five FC networks were included in the regression, only the salience network emerged as a significant FC predictor. Therefore, we performed a follow‐up analysis with salience network connectivity as the only FC measure of interest. In this model, the addition of both SC and FC measures (but not CSF or gray matter volumes) significantly increased the proportion of explained variance in memory performance, and all three measures (fornix/hippocampal cingulum microstructure, and salience network connectivity) were significant independent predictors of memory performance.ConclusionThese findings demonstrate that both connectivity modalities make unique contributions to episodic memory performance, but CSF and gray mater volumes do not. Furthermore, the highest proportion of explained variance was achieved when both modalities were included in the model, suggesting that the combination of SC and FC markers best accounts for individual variability in episodic memory in cognitively normal older adults.