Biopsychosocial and Demographic Predictors of Functional Brain Network Specialization and Segregation Across the Adult Lifespan

Abstract

The properties of functional brain networks are an important determinant of cognitive function in aging and dementia. Despite this, few studies have comprehensively examined demographic and biopsychosocial predictors of functional brain networks, and none have attempted to do so across the adult lifespan while accounting for collinearity among these predictors. The current study used data from 525 individuals between the ages of 35 and 100 years from the Human Connectome Project 2.0 Lifespan Release, which includes task-based functional neuroimaging, physical and emotional health, and demographic information. Two functional brain network properties previously identified as moderators of cognitive decline, entropy and modularity, were used as outcome metrics in four elastic net regression models that identified and ranked predictors of these metrics as well as their age-interaction terms. We identified biological sex, sleep duration, instrumental support, visual acuity, education, social isolation, diastolic blood pressure, and vigorous physical activity as the strongest and most consistent predictors of entropy and modularity beyond age. Importantly, these predictors differed from ranked correlational results, suggesting many predictors share large amounts of overlapping statistical variance. Additionally, we found that biological sex exhibited a significant moderation effect such that males demonstrated greater age-related decreases to network resilience with increasing age compared to females. In the current study, we ranked biopsychosocial health determinants of network properties in an adult lifespan sample. Given previous research implicating modularity and entropy as possible measures of cognitive reserve, these results may inform our understanding of resilience to cognitive decline in aging.