Ultra-High Order Independent Component Analysis for Intrinsic Connectivity Networks in Resting-State Functional Magnetic Resonance Imaging Data

Spatial group independent component analysis (sgr-ICA) has become a crucial method to understand brain function in functional magnetic resonance imaging (fMRI) research, especially in resting-state fMRI (rs-fMRI) studies. Early studies using low-order sgr-ICA identified large-scale brain networks, b...

Full description

Saved in:
Bibliographic Details
Published inProceedings (International Symposium on Biomedical Imaging) pp. 1 - 4
Main Authors Mirzaeian, Shiva, Jensen, Kyle M., Ballem, Adithya Ram, Calhoun, Vince D., Iraji, Armin
Format Conference Proceeding
LanguageEnglish
Published IEEE 14.04.2025
Subjects
Biomedical imaging
Brain modeling
brain parcellation
functional connectivity
Functional magnetic resonance imaging
high-order ICA
Independent component analysis
Indexes
Organizations
Quality control
Reliability
resting-state fMRI
Schizophrenia
Stability criteria
Online AccessGet full text

Cover

More Information
Summary:Spatial group independent component analysis (sgr-ICA) has become a crucial method to understand brain function in functional magnetic resonance imaging (fMRI) research, especially in resting-state fMRI (rs-fMRI) studies. Early studies using low-order sgr-ICA identified large-scale brain networks, but higher-order sgr-ICA (75-150 components) revealed more refined intrinsic connectivity networks (ICNs). While previous studies employing high-order sgr-ICA models (300 or even 1000 components) were often limited by small datasets, we applied an ultra-high-order ICA model with 500 components to a large rs-fMRI dataset of over 100,000 individuals' selecting approximately 58,000 after quality control. This comprehensive dataset enabled us to delineate the brain into fine-grained intrinsic connectivity networks (ICNs), offering a robust and detailed representation of these networks. We also evaluate functional connectivity differences between healthy people and individuals with schizophrenia. Our results demonstrate that ultra-high-order sgr-ICA provides reliable, precise estimations of ICNs and captures significant group differences between healthy people and schizophrenia patients. These findings highlight the potential of ultra-high-order sgr-ICA for improving rs-fMRI's clinical applications and single-subject analyses.
ISSN:1945-8452
DOI:10.1109/ISBI60581.2025.10981312