Covariate-Conditioned One-Step Diffusion for Multi-Horizon 3D Brain MRI Prediction in Aging and Neurodegeneration
Organizations: A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland · University of Petroleum and Energy Studies (UPES), Dehradun, India
Abstract
Predicting longitudinal brain changes from magnetic resonance imaging (MRI) can help characterize structural progression associated with aging and neurodegeneration. However, diffusion-based forecasting typically requires iterative denoising, limiting inference efficiency. We propose a covariate-conditioned one-step diffusion framework for multi-horizon prediction of 3D gray-matter density (GMD) maps. Rather than directly generating future anatomy, the model predicts longitudinal GMD changes through residual prediction in a single denoising step. The predicted changes are added to the baseline GMD map to reconstruct future anatomy. The model is conditioned on baseline anatomy, age, diagnosis, APOE4 allele count, and continuous follow-up time. Training combines mean squared error with a Pearson correlation loss to improve the spatial agreement between predicted and observed GMD changes. A unified model supports prediction horizons from 2 to 10 years. On the held-out ADNI test set, our method achieves a -Pearson of 0.478, mean squared error (MSE) of , and structural similarity index measure (SSIM) of 0.979, with -Pearson reaching 0.582 at 10 years. Without retraining, external evaluation on AIBL yields a -Pearson of 0.450, MSE of , and SSIM of 0.985. Ablation studies further demonstrate the benefits of residual prediction, covariate conditioning, and correlation-based supervision. Code: https://github.com/natenaile/OneStepBrainMRI.
Figures & tables
| Interval | Pairs | Participants | Age (years) |
|---|---|---|---|
| ADNI (Internal Dataset) | |||
| 2 y | 3440 | 1434 | |
| 3 y | 1942 | 858 | |
| 4 y | 1568 | 676 | |
| 5 y | 1044 | 448 | |
| 6 y | 871 | 402 | |
| Setting | Configuration |
|---|---|
| Data split | Participant-level: 1,028 train / 220 validation / 220 test; nine follow-up intervals |
| Optimiser | AdamW, learning rate linearly annealed to 0, gradient clipping 1.0 |
| Training | Batch size 4, FP16, EMA 0.9999, 30,000 steps |
| Augmentation | Random left–right flip (training only) |
| Hardware | NVIDIA GH200, 50 h training |
| Whole-Brain Change | Reconstructed GMD | Regional -Pearson | ||||||
| Horizon | -Pearson | MSE ( ) | MAE ( ) | Pearson | PSNR (dB) | SSIM | Hippocampus | Entorhinal Cortex |
| ADNI (Main Test Set) | ||||||||
| 2 y | 0.936 | 26.124 | 0.982 | 0.455 | 0.479 | |||
| 3 y | 0.927 | 25.543 | 0.980 | 0.494 | 0.503 | |||
| 4 y | 0.923 | 25.367 | 0.978 | 0.513 | 0.515 | |||
| 5 y | 0.914 | 24.972 | 0.976 | 0.531 | 0.536 | |||
Longitudinal Medical Image Analysis
Longitudinal medical image analysis studies how scans change over time to track disease, predict outcomes, and guide clinical decisions.
See all 104 papers