Contrast-Informed Augmentation and Domain-Adversarial Training for Adult-to-Neonatal MR Reconstruction Generalization

2026-06-11

AI Summary

This work addresses the poor generalization of MRI reconstruction models trained on adult data when applied to neonatal scans. To mitigate the domain shift between adult and neonatal MRI, the authors propose a cross-domain reconstruction method that integrates contrast-guided data augmentation with domain-adversarial training. For the first time, contrast information is incorporated into the data augmentation pipeline and jointly optimized with an end-to-end variational network (E2E-VarNet). Experimental results demonstrate that the proposed approach significantly outperforms baseline models trained solely on adult data on a neonatal test set, achieving SSIM of 0.924 and PSNR of 33.98 dB at acceleration factors of R=4 and R=8. Furthermore, t-SNE visualizations confirm improved consistency in cross-domain feature representations.

Abstract

Purpose: To investigate whether contrast-informed data augmentation and domain-adversarial training improve the adult-to-neonatal generalization of the E2E-VarNet. Methods: Three training regimes were investigated: (1) adult-only training with unaugmented adult data, (2) mixed training with paired unaugmented and neonatal-informed augmented adult data, and (3) mixed training with a domain-adversarial objective. Models were trained on retrospectively undersampled multi-coil adult T2-weighted brain MR data and evaluated on neonatal and adult test data at acceleration factors $R=4$ and $R=8$ using quantitative metrics and qualitative evaluation. Feature analyses assessed whether domain-adversarial training altered the latent representations of unaugmented adult, augmented adult, and neonatal test samples. Results: Mixed training (Mixed) and mixed domain-adversarial training (Mixed-DAT) outperformed unaugmented adult-only training (Unaug-Only) when evaluated on neonatal data. At R=4, Mixed-DAT achieved the best performance (SSIM = 0.924 +/- 0.027, PSNR = 33.98 +/- 1.15 dB). At R=8, Mixed-DAT performed best when measured using SSIM (0.848 +/- 0.031 vs. 0.766 +/- 0.037 for Unaug-Only and 0.814 +/- 0.035 for Mixed) and Mixed performed best when measured using PSNR (29.56 +/- 0.83 dB vs. 26.26 +/- 0.78 dB for Unaug-Only and 29.43 +/- 0.83 dB for Mixed-DAT). Qualitative assessment of t-SNE plots suggested that Mixed-DAT increased the overlap among the latent representations of the unaugmented adult, augmented adult, and neonatal test data. Conclusion: Contrast-informed augmentation and domain-adversarial training improved adult-to-neonatal generalization of deep learning-based MR reconstruction. These findings suggest that contrast-informed data augmentation combined with adversarial training may improve robustness to domain shift in undersampled neonatal MR reconstruction.

Problem

• domain generalization
• MR reconstruction
• adult-to-neonatal
• domain shift
• deep learning

Innovation

• contrast-informed augmentation
• domain-adversarial training
• adult-to-neonatal generalization
• MR reconstruction
• domain shift