Segregation of the Regional Radiomics Similarity Network Exhibited an Increase in Children
Poster No:
1234
Submission Type:
Abstract Submission
Authors:
Lei Chu1, Debin Zeng1, Yirong He2, Xiaoxi Dong2, Qiongling Li2, Shuyu Li2
Institutions:
1Beihang University, Beijing, China, 2Beijing Normal University, Beijing, China
First Author:
Co-Author(s):
Introduction:
Gaining a comprehensive understanding of the typical developmental trajectory of the human brain during late childhood to early adolescence is crucial for profound insights into its structure, function, and underlying mechanisms of developmental psychiatric and behavioral disorders(Blakemore, 2012; Vijayakumar et al., 2018). Using texture features of each brain region and calculating their similarities between nodes as edges to construct a network for each subject, R2SN has been well characterized with high reproducibility and significantly correlated with both genetic similarity networks and cognition in adults and been well applied to the diagnosis and staging of Alzheimer's disease(Zhao et al., 2022). Brain development exhibits increasing functional segregation to support cognition procession specialization within the framework of network neuroscience. Although coordinated variations in brain morphology have been extensively utilized to infer the macroscopic structural bases for developmental evolution of brain function, little is known about how microscopic features contribute to this process and their relationship with cognition. We hypothesized that alterations in the segregation of R2SN during development may be also associated with refinements in executive function (EF), a broad cognitive domain that encompasses multiple subdomains, including working memory, response inhibition, and set shifting, paralleling the human brain's protracted maturational time course.
Methods:
Participants. The longitudinal dataset of cognitively normal children was derived from the Children School Functions and Brain Development Project in China (CBD, Beijing Cohort) with a sample comprising 494 MR scans from 309 typically developing children ages 6.2-13 years at baseline (142 females and 167 males). 44 children were scanned at three separate times, 97 children managed to visit 2 with a scan interval of approximately one year. We uesed high-resolution anatomical images acquired with three dimensional 1mm3 isotropic T1-weighted magnetization-prepared rapid gradient echo (MPRAGE) sequence.
R2SN construction and statistical analysis. The anatomical MR images underwent Human Connectome Project (HCP) minimum preprocessing pipeline(Glasser et al., 2013) with several modifications for child's brain. we constructed individual regional radiomics similarity networks (R2SNs) based on similarities of radiomics features across 246 brain regions using CBD dataset.
To characterize R2SNs' segregation, we measured segregation indexes at both system and local levels. We corrected for multiple comparisons across all pairs of network metrics with false discovery rate (pFDR) < 0.05. To dimensionally delineate the longitudinal changes of the topology of similarity networks, we conducted several mixed-effect models to explore the developmental trajectories of each graph theoretical measures.
R2SN construction and statistical analysis. The anatomical MR images underwent Human Connectome Project (HCP) minimum preprocessing pipeline(Glasser et al., 2013) with several modifications for child's brain. we constructed individual regional radiomics similarity networks (R2SNs) based on similarities of radiomics features across 246 brain regions using CBD dataset.
To characterize R2SNs' segregation, we measured segregation indexes at both system and local levels. We corrected for multiple comparisons across all pairs of network metrics with false discovery rate (pFDR) < 0.05. To dimensionally delineate the longitudinal changes of the topology of similarity networks, we conducted several mixed-effect models to explore the developmental trajectories of each graph theoretical measures.
·Figure 1. Workflow of R2SN construction and network analysis.
Results:
Regarding fundamental properties of graph theory, we exclusively observed progressive linear increments in both the global clustering coefficient and local efficiency during the late childhood and early adolescence stages, indicating a heightened level of network segregation (Fig2A). Mixed-effect models revealed longitudinal increases in R2SNs' segregation, particularly evident in system-level segregation within subcortical regions, as well as decreasing segregation within the ventral attention network. (Fig2B-G)Furthermore, superior working memory and inhibitory control performance were associated with higher system-level segregation indexes in default and subcortical systems, along with lower local-level segregation indexes in several brain regions belonging to the visual network regardless of age(Fig2H-M).
·Figure 2. Network properties of R2SN and their developmental changes and association between the system and local segregation index and individual cognitive performance.
Conclusions:
Our findings offer novel insights into typical brain developmental changes as indicated by the segregation index computed using R2SN approaches, which can be a valuable tool for comprehending human brain structural and cognition maturation.
Higher Cognitive Functions:
Executive Function, Cognitive Control and Decision Making 2
Lifespan Development:
Early life, Adolescence, Aging 1
Normal Brain Development: Fetus to Adolescence
Keywords:
Cognition
Development
MRI
Other - brain network
1|2Indicates the priority used for review
Provide references using author date format
Glasser, M.F., et al., 2013. The minimal preprocessing pipelines for the Human Connectome Project. Neuroimage. 80, 105-24.
Vijayakumar, N., et al., 2018. Puberty and the human brain: Insights into adolescent development. Neurosci Biobehav Rev. 92, 417-436.
Zhao, K., et al., 2022. Regional Radiomics Similarity Networks Reveal Distinct Subtypes and Abnormality Patterns in Mild Cognitive Impairment. Adv Sci (Weinh). 9, e2104538.