fMRI-Diffusion: Generating fMRI Time Series Via a Temporal Transformer Diffusion Model for Major Depressive Disorder Diagnosis

2026-05-22

AI Summary

This study addresses the challenge of scarce labeled fMRI data in clinical depression diagnosis, where existing data augmentation methods generate only static functional connectivity matrices and neglect critical temporal dynamics. To overcome this limitation, the authors propose fMRI-Diffusion, a novel framework that integrates a Temporal Transformer into a diffusion model to directly synthesize ROI-level fMRI time series, thereby preserving intrinsic temporal structures. The approach further incorporates task-oriented supervised pretraining to enhance the diagnostic utility of the generated data. Evaluated on the REST-meta-MDD dataset, the method consistently improves diagnostic accuracy across ten classifiers, six brain atlases, and three acquisition sites, achieving up to a 3.7 percentage point gain over the strongest baseline. Moreover, the synthetic data exhibit high fidelity to real data, with distributional discrepancies below 0.06.

Abstract

Diagnosing Major Depressive Disorder (MDD) from functional magnetic resonance imaging (fMRI) using functional connectivity (FC) analysis requires large amounts of labeled data that are scarce in clinical settings. Existing augmentation methods synthesize FC matrices, which compress fMRI recordings into static pairwise summaries and discard temporal information. We propose fMRI-Diffusion, a framework that synthesizes region-of-interest (ROI)-level fMRI time series rather than FC matrices. A Temporal Transformer serves as the denoising network within a denoising diffusion probabilistic model, treating each time point as a token to capture temporal dependencies through self-attention. A supervised pretraining strategy initializes the Transformer with task-relevant representations before diffusion training, and FC matrices are derived from the synthesized time series for classification. Experiments on the REST-meta-MDD dataset show that augmenting training data with synthetic time series consistently improves diagnostic accuracy across ten classifiers, six parcellation atlases, and three acquisition sites. The method outperforms five recent FC-based synthesis approaches, with accuracy gains of up to 3.7 percentage points over the strongest baseline. Ablation studies confirm the contributions of both the Transformer-based denoiser and the pretraining strategy. Distributional fidelity metrics remain below 0.06 across all conditions, indicating close agreement between real and synthetic distributions. These findings suggest that synthesizing fMRI time series before FC computation preserves temporal information lost in matrix-level augmentation and provides a practical strategy for MDD diagnosis under limited data.

Problem

• Major Depressive Disorder
• fMRI
• data scarcity
• functional connectivity
• temporal information

Innovation

• Temporal Transformer
• Diffusion Model
• fMRI Time Series Synthesis
• Functional Connectivity
• Data Augmentation