Functional Connectivity Network Disruption Underlies Domain-Specific Impairments in Attention for Children Born Very Preterm

Abstract

Attention problems are common in school-age children born very preterm (VPT; < 32 weeks gestational age), but the contribution of aberrant functional brain connectivity to these problems is not known. As part of a prospective longitudinal study, brain functional connectivity (fc) was assessed alongside behavioral measures of selective, sustained, and executive attention in 58 VPT and 65 full-term (FT) born children at corrected-age 12 years. VPT children had poorer sustained, shifting, and divided attention than FT children. Within the VPT group, poorer attention scores were associated with between-network connectivity in ventral attention, visual, and subcortical networks, whereas between-network connectivity in the frontoparietal, cingulo-opercular, dorsal attention, salience and motor networks was associated with attention functioning in FT children. Network-level differences were also evident between VPT and FT children in specific attention domains. Findings contribute to our understanding of fc networks that potentially underlie typical attention development and suggest an alternative network architecture may help support attention in VPT children.

Keywords: attention; functional connectivity; networks; outcome; preterm.

Figure 1

Schematic of network level analysis workflow. After preprocessing and motion correction, BOLD signal time series were extracted from 300 spherical regions of interest. Nodes of the same color belong to the same brain network community. A whole brain connectome was constructed by averaging over all r-z transformed individual matrices in each group. The nonparametric correlation between each attention subtest and whole brain connectome was examined separately for each group, in addition to comparing connectome-behavior relationships between groups. Significance was established by randomly permuting the attention subtest scores 10 000 times for each subtest and measuring the permuted connectome behavior relationship. Mot, Motor; AUD, auditory; Mem, memory network; Vis, visual; MT, middle temporal; Rew, reward; BG, basal ganglia; Thal, thalamus; CB, cerebellum.

Figure 2

Attention and connectome-attention associations within group. A) Distribution of attention subtest scores for term and VPT children. The VPT group demonstrated reduced executive shifting (P < 0.005) and divided (P < 0.001) attention scores compared to term children. B) Brain network pairs significantly correlated with the corresponding attention subtest scores (top). The blue network-pairs are associated with attention in VPT, orange are associated with attention in term, and green network-pairs are associated with attention in both groups. All connectome-attention associations corrected at P < 0.05 per subtest. Mot, Motor; AUD, auditory; Mem, memory network; Vis, visual; MT, middle temporal; Rew, reward; BG, basal ganglia; Thal, thalamus; CB, cerebellum.

Figure 3

Conjunction of connectome-behavior associations across the five attention domains in term and VPT groups (P < 0.05 per subtest). Network pairs were implicated in multiple domains of attention. However, networks that supported multiple attention processes differed between term and VPT children. Attention was supported by subcortical and bottom-up processing within the VAN, visual, basal ganglia (BG), and thalamus (Thal) in VPT children. In contrast, attention was supported by cortical networks including FPN, CO, DAN, and SN networks in term children.

Figure 4

Between-groups differences in brain network connectivity supporting domain-specific attention skills. Split color in network pairs indicates both sustained and selective attention were associated with functional connectivity differences between term and VPT children. All between group network level associations with attention corrected at P < 0.01 per subtest (Family Wise Error P < 0.05). The blue lines on glass brains indicate a negative correlation with attention, while the red lines indicate a positive correlation. Lines represent fc-attention correlations (P < 0.05) used in the network-level analysis. For selective and divided attention tasks, higher scores indicated worse performance, thus a negative fc-behavior correlation (blue lines) would indicate higer fc was associated with better task performance, while a positive correlation would indicate higher fc was associated with worse task performance (red lines). Mot, Motor; AUD, auditory; Vis, visual; Thal, thalamus.