Functional Segregation of Human Brain Networks Across the Lifespan: An Exploratory Analysis of Static and Dynamic Resting-State Functional Connectivity

Abstract

Prior research has shown that during development, there is increased segregation between, and increased integration within, prototypical resting-state functional brain networks. Functional networks are typically defined by static functional connectivity over extended periods of rest. However, little is known about how time-varying properties of functional networks change with age. Likewise, a comparison of standard approaches to functional connectivity may provide a nuanced view of how network integration and segregation are reflected across the lifespan. Therefore, this exploratory study evaluated common approaches to static and dynamic functional network connectivity in a publicly available dataset of subjects ranging from 8 to 75 years of age. Analyses evaluated relationships between age and static resting-state functional connectivity, variability (standard deviation) of connectivity, and mean dwell time of functional network states defined by recurring patterns of whole-brain connectivity. Results showed that older age was associated with decreased static connectivity between nodes of different canonical networks, particularly between the visual system and nodes in other networks. Age was not significantly related to variability of connectivity. Mean dwell time of a network state reflecting high connectivity between visual regions decreased with age, but older age was also associated with increased mean dwell time of a network state reflecting high connectivity within and between canonical sensorimotor and visual networks. Results support a model of increased network segregation over the lifespan and also highlight potential pathways of top-down regulation among networks.

Keywords: brain dynamics; brain imaging; development; functional connectivity; resting state networks.

FIGURE 1

Scatterplot of the relationship between Age and Framewise Displacement in the current sample.

FIGURE 2

Significant effects of Age on static rsFC controlling for Gender, WM Performance, and Framewise Displacement, with resting-state functional connectivity tending to decrease among ROIs in different canonical functional networks.

FIGURE 3

(A) Age differences in static rsFC between right inferior occipital gyrus and left dorsolateral prefrontal cortex [t(21) = −4.01, p = 0.00063]. Of note, similar patterns were observed for the effect of age on static rsFC between other ROIs in visual systems and ROIs in other canonical networks. (B) Age differences in static rsFC between left putamen and right anterior insula [t(21) = −3.92, p = 0.00075]. Similar age effects were detected between other ROIs in the canonical affective network and ROIs in the canonical salience network.

FIGURE 4

Each of the intrinsic functional connectivity states are organized into a correlation matrix, with ROIs grouped by functional domain for pre-task (left) and post-task (right) functional runs. Scatterplots depict the association between Age and Mean Dwell Time for each of the ICN states. State 2 (left) and State 3 (right) exhibited similar profiles of connectivity, with Mean Dwell Time tending to increase across the lifespan [t(20) = 3.780, p = 0.001; t(19) = 3.326, p = 0.003].