Edge Participation Coefficient Unveiling the Development of Functional Connectome during Infancy
Poster No:
1213
Submission Type:
Abstract Submission
Authors:
Tianyu Fang1, Yue Wang2, Qiuyun Fan3, Yuanyuan Chen4
Institutions:
1Medical School of Tianjin University, Tianjin University, Tianjin, China, 2Tianjin University, Tinajin, China, 3Tianjin University, Tianjin, Tianjin, 4Tianjin University, Tianjin, China
First Author:
Co-Author(s):
Introduction:
Previous research has delved into the intricacies of the brain's community structure in adults by employing methodologies that segment co-fluctuating edges into non-overlapping communities. However, our understanding of the brain's community structure during infancy, a critical period marked by vital refinements in functional systems, remains relatively limited. In this study, we aim to bridge this gap by investigating how different functional connections within the infant brain interact and evolve as the brain undergoes development based on edge-centric functional analyses.
Methods:
All participants in this study were recruited as part of the developing Human Connectome Project, approved by the UK National Research Ethics Authority. we selected 781 structural-functional scans from the third release, including scans acquired at 37–44.5 weeks postmenstrual age (PMA), at and before term-equivalent age (TEA). All scans were collected in the Evelina Newborn Imaging Centre, Evelina London Children's Hospital, using a 3T Philips Achieva system during natural sleep without sedation. fifteen minutes of high temporal resolution rs-fMRI optimized for neonates was acquired using a multislice gradient-echo echo planar imaging sequence with multiband excitation.
The rs-fMRI data were preprocessed by dHCP minimal processing pipeline and then followed with a nuisance regression, bandpass filtering and registration to template space, which was the UNC-neonate brain template and AAL atlas. Edge-centric functional connectivity (eFC) matrix was calculated between co-fluctuations of paired reginal signals. Then, edge time series were clustered into 10 communities utilizing a standard k-means algorithm with Euclidean distance. The representative edge community assignments were used as the initial centroids to cluster the edge time series of each single subject. We subsequently calculated edge participation coefficient (edge PC) of a given node within the modular eFC networks.
The rs-fMRI data were preprocessed by dHCP minimal processing pipeline and then followed with a nuisance regression, bandpass filtering and registration to template space, which was the UNC-neonate brain template and AAL atlas. Edge-centric functional connectivity (eFC) matrix was calculated between co-fluctuations of paired reginal signals. Then, edge time series were clustered into 10 communities utilizing a standard k-means algorithm with Euclidean distance. The representative edge community assignments were used as the initial centroids to cluster the edge time series of each single subject. We subsequently calculated edge participation coefficient (edge PC) of a given node within the modular eFC networks.
Results:
Fig. 1 illustrated representative communities of brain connections obtained in term-born infants. Regions within sensorimotor system exhibited the greatest levels of within-system similarity. The overlapping community structure of infant cerebral cortex was displayed in Fig. 1e and f. In line with previous study on adults, regions within sensorimotor system were simultaneously involved in many communities, thereby connecting regions from multiple other systems. Fg. 2 illustrated changes in edge participant coefficient during infant brain maturation. Notably, edge PC values were significantly higher in preterm born infants compared to term born infants. When visualizing the association between edge PC and PMA at scan by node, significant negative correlation was observed with 89 regions surviving FWE correction. Further investigation at both interconnected network and connection level revealed that PC values of edges connecting the visual system with other networks and within "higher-order" networks displayed robust associations with the early brain maturation of normal neonates.
Conclusions:
The overlapping community structure of human cerebral cortex becomes established in the early stage of development. Edge-centric approach demonstrates a high level of participant specificity, capturing unique features of individuals. Decrease in functional diversity suggests the specialization of functional systems as the infant brain matures.
Lifespan Development:
Early life, Adolescence, Aging 1
Normal Brain Development: Fetus to Adolescence
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural)
fMRI Connectivity and Network Modeling 2
Keywords:
Data analysis
Development
Open Data
Other - Dynamic organization; Edge-centric approach; Functional diversity; Neonate cortex;
1|2Indicates the priority used for review
Provide references using author date format
2. Faskowitz, J. (2020). 'Edge-centric functional network representations of human cerebral cortex reveal overlapping system-level architecture', Nature neuroscience, vol. 23, no. 12, pp. 1644-1654.
3. Hughes, E. J. (2017). 'A dedicated neonatal brain imaging system', Magnetic resonance in medicine, vol. 78, no. 2, pp. 794-804.