Poster Session A, Saturday, March 23, 1:30 pm - 3:30 pm, Pacific Concourse
Within- and Between-Network Connectivity in Aging: How Correlation Direction Affects Discovery of Age Effects
Eleanna Varangis1, Christian G. Habeck1, Qolamreza Razlighi1, Yaakov Stern1; 1Columbia University
Studies examining functional connectivity have found a reduction in within-network connections, and an increase in between-network connections with aging. However, many of these studies focused only on positive correlations between regions/nodes. The present study examined the effect of age (N=245 adults age 20-80) on positive and negative correlations within and between predefined neural networks during 5 functional magnetic resonance imaging (fMRI) scan conditions: resting state and task-based scans in four domains (vocabulary, processing speed, fluid reasoning, and episodic memory). fMRI data were extracted from 264 regions of interest (ROIs) according to the Power et al. (2011) network taxonomy. Average positive and negative correlations were calculated within each network, and between networks for each of the five scan conditions. Results showed an interaction between age, scan, and correlation direction (within/between) for positive and negative correlations, such that advancing age was associated with lower positive within-network correlations and negative between-network correlations across most scan conditions, but had no effect on positive between-network correlations or negative within-network correlations. Further, a three-way interaction revealed that some networks exhibited an age-by-scan interaction, such that an effect of age on functional connectivity was only present during specific scan conditions. These results suggest that the effect of aging on functional connectivity may be driven by age-related reductions in positive correlations within networks, and reductions in the negativity of negative correlations between networks across resting and task-based scan conditions. Additionally, the effect of age may be more pronounced within certain cognitive networks, and under specific task-based scan conditions.
Topic Area: METHODS: Neuroimaging