Segregation and integration of the functional connectome in neurodevelopmentally 'at risk' children

Abstract

Functional connectivity within and between Intrinsic Connectivity Networks (ICNs) transforms over development and is thought to support high order cognitive functions. But how variable is this process, and does it diverge with altered cognitive development? We investigated age-related changes in integration and segregation within and between ICNs in neurodevelopmentally 'at-risk' children, identified by practitioners as experiencing cognitive difficulties in attention, learning, language, or memory. In our analysis we used performance on a battery of 10 cognitive tasks alongside resting-state functional magnetic resonance imaging in 175 at-risk children and 62 comparison children aged 5-16. We observed significant age-by-group interactions in functional connectivity between two network pairs. Integration between the ventral attention and visual networks and segregation of the limbic and fronto-parietal networks increased with age in our comparison sample, relative to at-risk children. Furthermore, functional connectivity between the ventral attention and visual networks in comparison children significantly mediated age-related improvements in executive function, compared to at-risk children. We conclude that integration between ICNs show divergent neurodevelopmental trends in the broad population of children experiencing cognitive difficulties, and that these differences in functional brain organisation may partly explain the pervasive cognitive difficulties within this group over childhood and adolescence.

Keywords: cognitive development; executive function; fMRI; functional connectivity; intrinsic connectivity networks; neurodevelopment.

Figure 1. Loadings of cognitive variables on the two rotated principal components

Note. Loadings lower than 0.3 are suppressed for visualisation purposes. Dot (Dot Matrix), Matrices (Matrix Reasoning), Search (Hector Cancellation/Balloon Hunt), Back_Digit (Backwards Digit Recall), Instructions (Following Instructions), Nonword_Rep (Children’s test of Nonword Repetition), Digit (Digit Recall).

Figure 2. Age-by-Group interactions on functional connectivity between ICNs

Note. Positive edges between the visual (purple) and ventral attention networks (orange; a) and the limbic (blue) and fronto-parietal networks (green, b) at 25% cost threshold. Age associations with positive functional connectivity between the visual and ventral attention networks (c) and the limbic and fronto-parietal networks (d).

Figure 3. Moderated mediation of visual-ventral attention network functional connectivity on age-related changes in executive function.

Note. (a) The moderated mediation model examining group-moderated mediation effect of positive functional connectivity (FC) between the visual and ventral attention networks on age-related changes executive function (Exec). Beta weights are shown for the control group. (b) 1000 bootstrapped estimates of the moderated mediation effect across proportional thresholds compared to mean permuted point estimates when group labels were shuffled 10,000 times. Error bars denote 95% confidence interval. Thresholds 15-17 are not displayed because no edges were present. *p<0.05, ***p<0.001