Supervised Deep Multimodal Matrix Factorization for Interpretable Brain Network Analysis

2026-05-13

AI Summary

This study addresses the challenges of interpretable supervised learning and cross-subject alignment in multimodal brain networks by proposing the SD3MF framework, which extends unsupervised single-graph clustering to supervised multimodal brain network analysis for the first time. SD3MF employs deep hierarchical symmetric non-negative matrix tri-factorization to jointly optimize graph reconstruction and classification objectives. It incorporates adaptive weights for data-driven multimodal fusion and constructs a shared latent representation to facilitate inter-subject alignment. Evaluated on multimodal connectomic data, SD3MF significantly outperforms strong baselines such as CNNs and GNNs, while yielding biologically interpretable, discriminative community interaction features.

Abstract

We present Supervised Deep Multimodal Matrix Factorization (SD3MF), an interpretable framework for integrative brain network analysis that generalizes Symmetric Nonnegative Matrix Tri-Factorization (SNMTF) from unsupervised single-graph clustering to supervised prediction over populations of multimodal graphs. SD3MF learns deep hierarchical factorizations for each modality together with a shared latent representation that aligns subjects across views. An encoder-decoder formulation jointly optimizes graph reconstruction and supervised prediction, while adaptive weights enable data-driven multimodal fusion. By representing each subject through community-level interaction matrices, the model yields interpretable and discriminative features. Experiments on multimodal connectome datasets show that SD3MF consistently outperforms strong deep learning baselines such as CNNs and GNNs, while enabling biologically interpretable insights. Code for reproducibility is available at: https://github.com/amjadseyedi/SD3MF.

Problem

• multimodal brain network
• supervised prediction
• interpretable analysis
• matrix factorization
• connectome

Innovation

• Supervised Deep Multimodal Matrix Factorization
• Interpretable Brain Network Analysis
• Symmetric Nonnegative Matrix Tri-Factorization
• Adaptive Multimodal Fusion
• Community-level Interaction