The abstract is read at the publisher; the summary is JClub's.
Marcenko-Pastur denoising under a simplified noise covariance model breaks under common diffusion MRI reconstructions across 28 public in vivo datasets, yielding weaker denoising than intended.
- Why it matters: The mismatch between assumed and actual noise statistics limits how well diffusion MRI can separate signal from noise, potentially degrading microstructure estimates.
- What they did: The authors estimate the noise covariance directly from reconstructed images to restore the intended Marcenko-Pastur behavior, validating the estimated model on repeated ex vivo and in vivo scans.
- The result: This reconstruction-derived covariance model improves denoising without extra measurements or reconstruction knowledge, enabling more reliable microstructure analysis in both existing and newly acquired diffusion MRI data.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.