Structure Guided Resting State Networks Estimation via Copula linked parallel ICA
Poster No:
1582
Submission Type:
Abstract Submission
Authors:
Oktay Agcaoglu1, Rogers Silva2, Vince Calhoun3
Institutions:
1Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Atlanta, GA, 2TReNDS Center, Atlanta, GA, 3GSU/GATech/Emory, Atlanta, GA
First Author:
Oktay Agcaoglu, PhD
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Co-Author(s):
Introduction:
fMRI is a powerful yet noisy and high-dimensional tool for investigating brain connectivity. Group independent component analysis (gICA) is commonly employed to reduce dimensionality and isolate functional networks. However, gICA decomposes fMRI data into components that can reflect contributions from gray matter (GM), white matter (WM), or artifacts, with components overlapping with WM or motion often being discarded (Allen,2011). Building on prior research demonstrating that covarying networks identified via sMRI-GM (Luo,2020) show correspondence to those observed in resting-state fMRI networks, we designed a novel multimodal framework that leverages structural data to guide resting state networks (RSN). This approach aims to encourage components from fMRI to align with structural GM maps while simultaneously linking extraneous variance to WM, thereby reducing noise contamination in the RSNs. Specifically, extraneous components related to WM are isolated to minimize their interference with GM-related networks while still allowing informative variance to be captured by the model.
Methods:
Traditional ICA methods (e.g., Infomax, FastICA) rely on fixed learning rules, limiting their ability to handle complex loss functions required for multimodal imaging. Modern computational tools, particularly deep learning frameworks (DL-F) with automatic differentiation, now enable the optimization of such functions. We introduced copula-linked parallel ICA (CLiP-ICA) (Agcaoglu, 2024), a novel method integrating DL-F, copulas, and ICA. CLiP-ICA was applied to estimate structure-guided RSNs from 4D fMRI data using 3D structural data to enhance component integrity. Data from 864 ADNI subjects included preprocessed fMRI and T1-weighted sMRI segmented into GM and WM probability maps using SPM. We analyzed two settings. First, we linked fMRI and GM with a model order of 75 using Gaussian copula dependencies ranging from 0.95 to 0.5. Second, we linked 65 fMRI components with GM and 10 fMRI components with WM to capture "extraneous components," using similar dependency values. All datasets underwent PCA, and GM/WM data were resampled to match fMRI resolution.
Results:
Results demonstrate that structural guidance reveals a greater number of meaningful components and fewer artifact components, addressing the longstanding issue of the optimal model order in ICA. Specifically, CLiP-ICA identified 50 RSNs and 45 intrinsic structural networks (ISNs) out of 75 components. Additionally, CLiP-ICA detected complex functional network connectivity (FNC) patterns across different stages of cognitive decline. It also provided complementary insights by leveraging the intricate patterns in the loading parameters of ISNs. Furthermore, introducing WM as an additional constraint further refined the estimated connectivity and enhanced sensitivity to group differences, as shown in Figure 1. The components highly linked to WM effectively captured the variance associated with WM-linked components, confining extraneous variance to them. These extraneous components are displayed in Figure 2b.
Conclusions:
This study introduces a novel multimodal fusion method that guides RSNs using both GM and WM structural data while preserving the full 4D features of fMRI. CLiP-ICA uncovered highly linked and modality-unique components, highlighting distinct FNC patterns and subtle variations in FNC associated with cognitive decline. Adding WM as a constraint further improved performance by confining extraneous variance to WM-linked components and refining connectivity estimates, uncovering additional group differences. Structural-functional coupling is particularly important in Alzheimer's disease, where rapid functional and structural changes are imminent. By enhancing connectivity analysis, CLiP-ICA offers critical insights into neurodegenerative processes and holds promise for broader multimodal neuroimaging applications.
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s)
Modeling and Analysis Methods:
Bayesian Modeling
fMRI Connectivity and Network Modeling 2
Methods Development 1
Task-Independent and Resting-State Analysis
Keywords:
Degenerative Disease
Design and Analysis
FUNCTIONAL MRI
Machine Learning
Modeling
Statistical Methods
STRUCTURAL MRI
Structures
White Matter
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
Allen, E. A., Erhardt, E. B., Damaraju, E., Gruner, W., Segall, J. M., Silva, R. F., Havlicek, M., Rachakonda, S., Fries, J., Kalyanam, R., Michael, A. M., Caprihan, A., Turner, J. A., Eichele, T., Adelsheim, S., Bryan, A. D., Bustillo, J., Clark, V. P., Feldstein Ewing, S. W., . . . Calhoun, V. D. (2011). A baseline for the multivariate comparison of resting-state networks. Front Syst Neurosci, 5, 2. https://doi.org/10.3389/fnsys.2011.00002
Luo, N., Sui, J., Abrol, A., Chen, J., Turner, J. A., Damaraju, E., Fu, Z., Fan, L., Lin, D., Zhuo, C., Xu, Y., Glahn, D. C., Rodrigue, A. L., Banich, M. T., Pearlson, G. D., & Calhoun, V. D. (2020). Structural Brain Architectures Match Intrinsic Functional Networks and Vary across Domains: A Study from 15 000+ Individuals. Cereb Cortex, 30(10), 5460-5470. https://doi.org/10.1093/cercor/bhaa127