DMVFC: Deep Learning Based Functionally Consistent Tractography Fiber Clustering Using Multimodal Diffusion MRI and Functional MRI

2025-10-24

AI Summary

Existing dMRI tractography clustering methods rely solely on geometric trajectory features, neglecting functional and microstructural information, resulting in suboptimal functional and anatomical consistency. To address this, we propose DMVFC—a deep multi-view fiber clustering framework—that pioneers the integration of fMRI BOLD signals into white matter fiber clustering. DMVFC jointly models fiber spatial trajectories, diffusion-derived microstructural properties, and task- or resting-state functional activity. It comprises two core components: a multi-view pretraining module and a collaborative fine-tuning module, enabling cross-modal embedding learning and adaptive feature fusion. Evaluated on public datasets, DMVFC significantly outperforms state-of-the-art methods, achieving superior functional coherence, anatomical interpretability, and cross-subject stability. This work advances connectome-based neuroscience by providing biologically more meaningful white matter parcellations.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Data Mining & Knowledge Management: Mining of Visual, Multimedia & Multimodal Data
• Computer Vision: Multi-modal Vision

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Tractography fiber clustering using diffusion MRI (dMRI) is a crucial method for white matter (WM) parcellation to enable analysis of brains structural connectivity in health and disease. Current fiber clustering strategies primarily use the fiber geometric characteristics (i.e., the spatial trajectories) to group similar fibers into clusters, while neglecting the functional and microstructural information of the fiber tracts. There is increasing evidence that neural activity in the WM can be measured using functional MRI (fMRI), providing potentially valuable multimodal information for fiber clustering to enhance its functional coherence. Furthermore, microstructural features such as fractional anisotropy (FA) can be computed from dMRI as additional information to ensure the anatomical coherence of the clusters. In this paper, we develop a novel deep learning fiber clustering framework, namely Deep Multi-view Fiber Clustering (DMVFC), which uses joint multi-modal dMRI and fMRI data to enable functionally consistent WM parcellation. DMVFC can effectively integrate the geometric and microstructural characteristics of the WM fibers with the fMRI BOLD signals along the fiber tracts. DMVFC includes two major components: (1) a multi-view pretraining module to compute embedding features from each source of information separately, including fiber geometry, microstructure measures, and functional signals, and (2) a collaborative fine-tuning module to simultaneously refine the differences of embeddings. In the experiments, we compare DMVFC with two state-of-the-art fiber clustering methods and demonstrate superior performance in achieving functionally meaningful and consistent WM parcellation results.

Problem

• Integrating multimodal MRI data for fiber clustering
• Enhancing functional coherence in white matter parcellation
• Combining geometric, microstructural and functional fiber characteristics

Innovation

• Integrates geometric, microstructural, and functional data
• Uses multi-view pretraining for separate feature embeddings
• Collaborative fine-tuning refines multimodal embedding differences