Scalable Diffusion Transformer for Conditional 4D fMRI Synthesis

TL;DR

This paper introduces a scalable diffusion transformer model for generating 4D fMRI sequences conditioned on cognitive tasks, effectively capturing brain activation patterns and outperforming baseline models.

Contribution

It presents the first voxelwise 4D fMRI conditional generation model combining diffusion, transformer, and latent compression techniques with strong task conditioning.

Findings

Reproduces task-evoked activation maps with high correlation (0.83).

Preserves inter-task representational structure (RSA 0.98).

Outperforms U-Net baseline on all metrics.

Abstract

Generating whole-brain 4D fMRI sequences conditioned on cognitive tasks remains challenging due to the high-dimensional, heterogeneous BOLD dynamics across subjects/acquisitions and the lack of neuroscience-grounded validation. We introduce the first diffusion transformer for voxelwise 4D fMRI conditional generation, combining 3D VQ-GAN latent compression with a CNN-Transformer backbone and strong task conditioning via AdaLN-Zero and cross-attention. On HCP task fMRI, our model reproduces task-evoked activation maps, preserves the inter-task representational structure observed in real data (RSA), achieves perfect condition specificity, and aligns ROI time-courses with canonical hemodynamic responses. Performance improves predictably with scale, reaching task-evoked map correlation of 0.83 and RSA of 0.98, consistently surpassing a U-Net baseline on all metrics. By coupling latent diffusion with a scalable backbone and strong conditioning, this work establishes a practical path to conditional 4D fMRI synthesis, paving the way for future applications such as virtual experiments, cross-site harmonization, and principled augmentation for downstream neuroimaging models.