BrainWorld: A Structural-Prior-Conditioned Generative Model for Whole-Brain 4D fMRI Dynamics

2026-06-16

AI Summary

This work addresses the limitation of existing fMRI foundation models, which primarily focus on representation learning and prediction but struggle to generate structurally consistent whole-brain 4D dynamic signals. To overcome this, the authors propose BrainWorld, a novel framework that, for the first time, deeply integrates individual structural MRI (sMRI) as an anatomical prior into the diffusion denoising process, enabling structure-conditioned generation of long-sequence 4D fMRI data. The method supports stable synthesis of trajectories up to 400 frames and demonstrates significant improvements over current baselines across 22 datasets in terms of generation quality, downstream task performance, and cross-modal representational transferability.

Abstract

Whole-brain 4D fMRI generation is valuable for modeling functional brain dynamics, yet existing fMRI foundation models mainly target representation learning and downstream prediction rather than conditional predictive generation. We introduce BrainWorld, a structural-prior-conditioned generative model for whole-brain 4D fMRI dynamics. BrainWorld uses sMRI as subject-level anatomical context to guide future fMRI generation, integrating structural information into the denoising process rather than treating it as a parallel modality. Evaluated on 22 datasets spanning diverse cohorts and brain states, BrainWorld generates stable 4D fMRI trajectories up to 400 frames, improves downstream performance through generated-example augmentation, and learns transferable multimodal representations that outperform baselines. Together, these results establish BrainWorld as a condition-aware generative framework for long-horizon brain dynamics modeling and multimodal representation learning.

Problem

• 4D fMRI generation
• functional brain dynamics
• conditional generative modeling
• whole-brain modeling
• multimodal representation

Innovation

• structural-prior-conditioned
• 4D fMRI generation
• whole-brain dynamics
• multimodal representation learning
• denoising diffusion