EXPLORING LONGITUDINAL BRAIN CONNECTIVITY DYNAMICS IN ADOLESCENT: A MULTIMODAL MRI ANALYSIS
Poster No:
1226
Submission Type:
Abstract Submission
Authors:
Rekha Saha1, Debbrata Kumar Saha2, Zening Fu3, Rogers Silva4, Vince Calhoun5
Institutions:
1Georgia State University, Atlanta, GA, 2Georgia Institute of Technology, Atlanta, GA, 3Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS) Georgi, Atlanta, GA, 4TReNDS Center, Atlanta, GA, 5GSU/GATech/Emory, Decatur, GA
First Author:
Co-Author(s):
Zening Fu
Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS) Georgi
Atlanta, GA
Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS) Georgi
Atlanta, GA
Introduction:
Functional and structural MRI play vital roles in brain development analysis. Our study explores brain development using fMRI/sMRI, merging modalities to capture nuanced brain changes during adolescence. Employing a novel symmetric multimodal fusion approach, we analyze FNC and GM data from the Adolescent Brain and Cognitive Development (ABCD) dataset, unveiling significant longitudinal change patterns. This approach reveals structured alterations in visual-sensorimotor connectivity and bilateral sensorimotor cortex. Through this analysis, we emphasize age-related multivariate brain changes, showcasing the dynamic nature of brain structure and connectivity during adolescence. This study emphasizes the dynamic nature of brain structure and connectivity during adolescence, showcasing the efficacy of our method in unraveling these changes across modalities.
Methods:
In our study, we utilized subject-specific fMRI and sMRI data collected at baseline and two-years later. To capture alterations in functional network connectivity (FNC) and gray matter (GM), we computed differences between the baseline and two-year datasets, creating ∆FNC and ∆GM matrices to represent their respective changes over time. Employing the mCCA+jICA (define) method, we deconstructed these matrices, extracting co-varying change patterns: functional change patterns (FCPs) and structural change patterns (SCPs). Utilizing an elbow criterion, we estimated five components for both GM and FNC data. Following the mCCA+jICA estimation, we analyzed loading parameters and the source matrix. To identify significant longitudinal changes, we conducted one-sample t-tests on the loading parameters aF and aG for both modalities, assessing their statistical significance at a 95% confidence level with corrections for multiple comparisons.
Results:
The study leverages the Neuromark template featuring 53 replicable networks grouped into 7 domains [1]. Figure 1 portrays the experimental findings, showcasing spatial maps illustrating the connections between multivariate FCPs and SCPs. These visualizations exhibit 2 FCP components (component 3 and 2) and their corresponding spatial maps of SCPs, represented with associated T-values indicating increased or decreased expression with age [2]. Notably, the upper aspects demonstrate associations of FCPs and SCPs components with increased age. Specifically, components 2 and 3 of FCPs reveal significant changes in brain functional connectivity with age, showing an increasing trend. Component 3 exhibits heightened connectivity between visual and sensorimotor domains in FNC data, complemented by decreasing changes in the bilateral sensorimotor cortex in sMRI data over two years. Additionally, FCP component 3 displays a decreasing trend with age in functional connectivity between visual and cerebellar domains, as well as between sensorimotor and cognitive control domains.
·Example of linked FNC and GM components. Here, component 3 from both functional and structural data exhibits highly structured change patterns including increased connectivity between VS and SM domain
Conclusions:
This study introduces an innovative method to explore links between brain functional and structural changes using FNC matrices and GM data. It investigates whole-brain structural and functional alterations over two years, revealing age-related trends and associations between these changes. Analyzing ABCD dataset data, the research identifies significant alterations in functional and structural change patterns within this period. These outcomes underscore the potential of our proposed methodology as a valuable approach for assessing holistic brain functional and structural alterations and their interplay in longitudinal investigations.
Lifespan Development:
Early life, Adolescence, Aging 1
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural)
Multivariate Approaches 2
Novel Imaging Acquisition Methods:
BOLD fMRI
Multi-Modal Imaging
Keywords:
Aging
Development
FUNCTIONAL MRI
Multivariate
STRUCTURAL MRI
1|2Indicates the priority used for review
Provide references using author date format
[2] R. Saha, D. K. Saha, M. A. Rahaman, Z. Fu, and V. D. Calhoun, “Longitudinal whole-brain functional network change patterns over a two-year period in the abcd data,” in 2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI). IEEE, 2022, pp. 1–4.