BrainFIBRE: A Foundation Model via Information Decomposition for Brain Microstructure

2026-07-01

AI Summary

This work addresses how to effectively integrate the unique, redundant, and synergistic information embedded in NODDI-based multimodal brain microstructural maps—specifically the neurite density index (NDI), orientation dispersion index (ODI), and free-water fraction (FWF)—to construct transferable and neurobiologically interpretable representations. The authors propose BrainFIBRE, the first foundation model for brain microstructure, trained on NODDI data from 55,592 participants in the UK Biobank. BrainFIBRE introduces partial information decomposition (PID) into self-supervised multimodal learning, leveraging a counterfactual candidate construction (CCC) strategy and a Mixture-of-Experts architecture to disentangle multimodal signals without labeled data. The model achieves state-of-the-art performance in predicting age, sex, cerebrovascular and neurodegenerative biomarkers, and cognitive outcomes across both Caucasian and Asian cohorts, while yielding representations with clear neurobiological interpretability.

Abstract

Diffusion MRI probes brain microstructure with particular sensitivity to early cerebrovascular and neurodegenerative changes. Neurite Orientation Dispersion and Density Imaging (NODDI) decomposes the diffusion signal into three biophysically interpretable maps: neurite density index (NDI), orientation dispersion index (ODI), and free water fraction (FWF), capturing neurite packing, fiber coherence, and extracellular fluid. These 3D maps offer a rich substrate for transferable microstructural representations, yet integrating them is challenging: standard representation learning struggles to disentangle the unique information in each map from their shared and synergistic interactions. We present BrainFIBRE, the first foundation model for brain microstructure, pretrained on NODDI-derived maps from 55,592 UK Biobank participants. We propose Self-supervised Partial Information Decomposition (SPID), which extends PID-guided multimodal learning to the self-supervised regime for the first time. A novel Counterfactual Candidate Construction (CCC) paradigm perturbs inter-modality alignment through modality dropping and swapping, providing the contrastive signal for a Mixture-of-Experts architecture to disentangle unique, synergistic, and redundant information without any downstream label. On both Caucasian and Asian cohorts, BrainFIBRE achieves state-of-the-art performance across diverse tasks predicting age, sex, cerebrovascular and neurodegenerative markers, and cognition, while yielding neurobiologically interpretable representations that reveal task- and cohort-specific interaction patterns. BrainFIBRE establishes a versatile foundation for neuroimaging analysis at the microstructural level.

Problem

• brain microstructure
• diffusion MRI
• NODDI
• information decomposition
• representation learning

Innovation

• Foundation Model
• Self-supervised Learning
• Partial Information Decomposition
• Multimodal Disentanglement
• NODDI