333 - Altered Development of Network Hubs and Modules at Term Equivalent Age in Structural MRI Connectome of Preterm Brain

Poster Number: 333
Publication Number: 333.442

Yong Hun Jang, Hanyang University Graduate School of Biomedical Science and Engineering, Seoul, Seoul-t'ukpyolsi, Republic of Korea; Hyuna Kim, Hanyang University, Seoul, Seoul-t'ukpyolsi, Republic of Korea; hyun Ju Lee, Hanyang University Hospital, Hanyang University College of Medicine, Seoul, Seoul-t'ukpyolsi, Republic of Korea

Background: During the 3rd trimester, cortico-cortical neural circuits are formed in the human brain. Preterm birth disturbs brain network maturation at this critical period. We present hubs and modules which are network topologies that reorganize efficient brain communication.

Objective: After exploring development of global networks in preterm and full-term infants, we elucidate the influence of regional differences in the hub and modular organization of the two groups on brain development of preterm infants.

Design/Methods: Preterm (Nf134) and full-term infants (Nf76) underwent diffusion MRI on a 3Tscanner at term-equivalent age. Brain regions that satisfied degree (DC) and betweenness (BC) centrality > mean ± 1 standard deviation were identified as the hub. The global network measures under different sparsity thresholds were calculated. At the nodal level, the hubs were analyzed under 0.25 sparsity threshold by DC and BC to explore the differences in the development of the hubs between the two groups. Module detection was performed with Modified Greedy Optimization to classify the network into different modules. After we calculated inter (Cs,t) and intra-module (Cs) connectivity, within-module degree z-score (Z) were computed to quantify the role of the hubs in the inter and intra-modules. Statistical procedures were performed by ANCOVA with adjusted age at MRI and significant results were defined as p< .05 after false discovery rate correction.

Results: We identified 12 hubs in the two groups. All global network measures under all threshold are not significantly different between two groups. Compared to controls, preterm infants showed decreased DC in the superior frontal gyrus (p < .001). These metrics were increased (DC and BC; p< .05) in the fusiform gyrus in preterm infants (Table 1). In the modular level, the nodes were evenly distributed in the 4 modules in controls, whereas in preterm infants, the 66 nodes out of 90 were concentrated in the parieto-temporo-occipital lobe (Fig. 1). Cs,t and Cs showed notable differences between two groups in the left hemisphere (Table 2). In particular, the modules include the frontal lobe showed low Cs,t and Cs in preterm infants. These findings were identified by exhibiting that the Z was low in the superior frontal gyrus (p < .001) and high in the left fusiform gyrus (p=.034) in preterm infants.Conclusion(s): Preterm infants exhibited altered structural networks in the superior frontal gyrus and fusiform gyrus. These findings provide biomarkers for understanding the subtle reconstruction of neural circuits that may lead to neurodevelopmental delays in preterm infants.
Summary of differences in individual hubs at the nodal level between preterm and full-term infants.The class column indicates the brain lobes to which hub region belongs. Data are the mean ± standard deviation.
* ANCOVA with adjusted age at MRI and FDR corrected p-value.
Abbreviations: SD, standard deviation; L, left; R, right; DC, degree centrality; BC, betweenness centrality; SFGdor, superior frontal gyrus; PoCG, postcentral gyrus; ITG, inferior temporal gyrus; MTG, middle temporal gyrus; FFG, fusiform gyrus; PCUN, precuneus; MOG, middle occipital gyrus; PUT, putamen.
Modular organizations in preterm and full-term groups.A total of the 90 nodes are distributed in the 4 modules. Each colored line represents a connection between nodes within the intra-module. The gray line represents the connection of two nodes between inter-modules. Module I, red; Module II, yellow; Module III, green; Module IV, blue.