cs.LGAug 25, 2026

MSR-IVA: Masked Structural Residual Independent Vector Analysis for State-Aware Fusion of Structural MRI and Dynamic Functional Network Connectivity

Authors: Victor Solomon, Zening Fu, Rafal Angryk, Vince D. Calhoun, Jingyu Liu

Organizations: Department of Computer Science, Georgia State University · Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State University, Georgia Institute of Technology, and Emory University

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Multimodal fusion of structural MRI (sMRI) and dynamic functional network connectivity (dFNC) can reveal how brain structure relates to changing functional states. When the same structural latent representation is coupled with multiple states, applying independent vector analysis (IVA) separately to each state can produce unrelated structural decompositions, while forcing identical decompositions may suppress state-specific relationships. In addition, not every subject expresses every dynamic state. We propose masked structural residual IVA (MSR-IVA), a state-aware framework that combines a shared structural representation with state-specific residual adaptations and masks for incomplete state expression. On an Alzheimer's Disease Neuroimaging Initiative cohort, MSR-IVA improved matched source coupling by 6.5% and reduced unmatched dependence by 15.7% relative to the independent pairwise IVA baseline. Among subjects expressing both states, mean absolute cross-state structural source correlation was 0.9177 for MSR-IVA versus 0.2978 for no sharing, demonstrating controlled structural sharing that preserves source correspondence while allowing state-specific adaptation.

Figures & tables

Fig. 1 : Overview of structural and dynamic functional connectivity processing.
Fig. 2 : Sensitivity of MSR-IVA to the structural residual weight α\alpha . (a) Within-SCV coupling across dynamic states, where higher values indicate stronger matched source coupling. (b) Cross-SCV dependence, where lower values indicate weaker dependence between unmatched sources.
Fig. 3 : Repeated-run performance across five initialization seeds. (a–c) Performance averaged across the two dynamic states for within-SCV coupling, cross-SCV dependence, and coupling contrast. (d–f) Corresponding state-specific results for State 1 and State 2.
MethodStructural similarity
No sharing0.2978±0.00750.2978\pm 0.0075
MSR-IVA0.9177±0.00810.9177\pm 0.0081
Hard sharing1.0000±0.00001.0000\pm 0.0000
Table 1: Cross-state structural source similarity on the 124 common subjects (mean ±\pm SD across five seeds).
StateMethodCwithin↑\mathbf{C_{\mathrm{within}}\uparrow}Ccross↓\mathbf{C_{\mathrm{cross}}\downarrow}Ccontrast↑\mathbf{C_{\mathrm{contrast}}\uparrow}
State 1IP-IVA0.85890.06040.7985
State 1MSR-IVA0.93890.04900.8899
State 2IP-IVA0.93430.06330.8710
State 2MSR-IVA0.97800.04800.9300
Table 2: Source coupling metrics for the seed-42 comparison.

Explore similar work

CardsList

  1. NeuroAlign: Hierarchical Multimodal Fusion of Dynamic and Structural Neuroimaging for MCI Analysis

    May 31, 2026Xiongri Shen, Zhenxi Song, Jiaqi wang +13Feature AttributionFunctional Connectivity

    Department of Computer Science and Technology, Harbin Institute of Technology (Shenzhen), Shenzhen, 518055, China · School of Intelligence Science and Engineering, College of Artificial Intelligence, Harbin Institute of Technology, Shenzhen, 518055, China · School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China +5

  2. Learning Structural-Functional Brain Representations through Multi-Scale Adaptive Graph Attention for Cognitive Insight

    Mar 31, 2026Badhan Mazumder, Sir-Lord Wiafe, Aline Kotoski +2Brain Network AnalysisMultimodal Graph Learning

    Department of Computer Science, Georgia State University · Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State University, Georgia Institute of Technology, and Emory University · Neuroscience Institute, Georgia State University

  3. MVMGNN;Multi-View Masked Graph Neural Network for Alzheimer's Disease Diagnosis using Structural MRI

    Jul 8, 2026Ni Yao, Zhenxu Wang, Danyang Sun +6Graph Neural NetworksMulti-View Learning

    School of Computer Science and Technology, Zhengzhou University of Light Industry, Zhengzhou 450002, Henan, China · Department of Computer Science, Kennesaw State University Marietta, GA, USA · Department of Applied Computing, Michigan Technological University, Houghton, MI, USA +1