CUPA-T2*: Covariance-Aware Uncertainty Propagation and Alignment for T2* Mapping in Accelerated MRI

Christine Preibisch; Dirk H. J. Poot; Gideon N. L. Rouwendaal; Hannah Eichhorn; Julia A. Schnabel; Natascha Niessen

REVIEW 5 major objections 5 minor 24 references

CUPA-T2: Covariance-Aware Uncertainty Propagation and Alignment for T2 Mapping in Accelerated MRI

T0 review · 5 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read CUPA-T2* propagates the full voxel-wise inter-echo covariance from accelerated MRI reconstructions into downstream T2* fitting, improving white-matter error at R=3 and R=4 while aligning predicted variance with reconstruction uncertainty.

desk verdict Covariance-aware uncertainty propagation for T2* mapping is a genuine idea, but the reported alignment gain is partly circular and the 100-sample covariance estimate needs validation. read the letter →

arxiv 2608.08693 v1 pith:HIDBGSYC submitted 2026-08-09 cs.CV

Gideon N. L. Rouwendaal , Natascha Niessen , Hannah Eichhorn , Dirk H. J. Poot , Christine Preibisch , Julia A. Schnabel This is my paper · ORCID

classification cs.CV

keywords T2*mappingquantitativeMRIuncertaintypropagationMonteCarlodropoutcovariance-awaresamplingheteroscedasticregressionacceleratedalignment

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Accelerated T2* mapping is clinically useful only if reconstruction artifacts do not corrupt the fitted decay parameter, and this paper's claim is that the fix is to carry the full inter-echo covariance forward rather than a per-voxel scalar variance. CUPA-T2* does this in three connected stages: a Monte Carlo dropout reconstruction produces many stochastic versions of the images, a per-voxel covariance matrix over the 12 echoes is factored by Cholesky decomposition and used to sample uncertainty-aware inputs, and a heteroscedastic network trained with a tissue-stratified correlation regularizer predicts both T2* and its variance. The experiments show the framework stays competitive overall on brain data while achieving the lowest white-matter error at accelerations R=3 and R=4, and it makes predicted T2* variance track reconstruction uncertainty far more closely than a heteroscedastic baseline (Pearson correlation 0.65-0.77 versus 0.19-0.28). The same results expose a trade-off: calibration error and area under the risk-coverage curve worsen as alignment increases. If the result holds, accelerated T2* protocols could ship with voxel-wise uncertainty maps that point at the regions most likely to be wrong.

What carries the argument

The central mechanism is a per-voxel inter-echo covariance matrix $\Sigma_v$, estimated from $S=100$ Monte Carlo dropout reconstructions over the $2E$ real and imaginary components of $E=12$ echoes and factorized as $\Sigma_v \approx L_v L_v^T$ via Cholesky decomposition. It enters the regression through the sampling rule $\mathbf{x}_v^{(s)} = \mu_v + \alpha_v L_v \mathbf{z}^{(s)}$ with $\mathbf{z}^{(s)} \sim \mathcal{N}(0, I)$, where the scalar $\alpha_v$ grows with the RMS marginal standard deviation $u_v = \sqrt{\frac{1}{2E}\mathrm{tr}(\Sigma_v)}$, so clean voxels receive almost no injected noise while noisy voxels receive correlated echo-to-echo variation. The second half of the machinery is a correlation-based regularizer: the Pearson correlation between predicted log-variance and $\log u_v$, averaged with equal weights over white matter, gray matter, and CSF, is added to a $\beta$-NLL fitting loss as $\mathcal{L} = \mathcal{L}_{\beta\text{-NLL}} + \lambda(1-\rho)$. This regularizer is what forces the predicted aleatoric variance to line up with the upstream reconstruction uncertainty, and the covariance-aware samples are what feed that alignment meaningful structure.

What would settle it

On the T2*-MOVE test set at R=4, randomly permute the off-diagonal entries of every voxel's inter-echo covariance matrix before Cholesky factorization; if white-matter NRMSE stays at 0.168 rather than moving toward the 0.173 of the diagonal baseline, the specific covariance structure is not carrying the reported benefit.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that structured inter-echo uncertainty from undersampled MRI reconstruction can be propagated into T2* regression without collapsing it to a scalar, and that this changes where the fitted parameter is accurate. The load-bearing result is that covariance-aware sampling improves white-matter T2* accuracy at R=3 and R=4, while the RMS correlation regularizer raises the Pearson correlation between reconstruction-derived uncertainty and predicted T2* variance from roughly 0.19-0.28 with a heteroscedastic baseline to 0.65-0.77 with CUPA-T2*. The paper attributes the tissue-dependence to the reliability of the estimated covariance: white matter is homogeneous with fewer physiological fluctuations, whereas high-uncertainty CSF has less coherent inter-echo structure. It also reports an explicit trade-off, since the alignment objective that maximizes correlation degrades ECE and AURC, and the ablation at R=4 shows that the covariance structure and the regularizer each contribute separately.

Load-bearing premise

The load-bearing premise is that a 24-by-24 per-voxel covariance matrix estimated from only 100 stochastic reconstructions reflects genuine inter-echo correlation structure rather than estimation noise, so Cholesky sampling injects useful variability instead of corrupting the fit.

Editorial extensions

If this is right

White-matter T2* mapping, which matters for iron-related lesion assessment, becomes accurate at R=3 and R=4 while total scan time falls to roughly 55 seconds at R=4.
Predicted T2* variance now highlights the same voxels where reconstruction artifacts concentrate, so the uncertainty map can serve as a per-voxel trust signal when reading accelerated maps.
Deployment must treat calibration and selective prediction as competing objectives: with the alignment weight at 0.05, the correlation gain comes at the cost of higher ECE and AURC, so applications that need calibrated confidence should lower $\lambda$.
The same covariance-aware propagation and alignment recipe should transfer to other quantitative parameters whose echoes or contrasts share undersampling-induced correlations, as the paper states as future work.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A testable extension the paper does not run: if shared undersampling is what creates the inter-echo covariance, then independently drawn sampling masks per echo should shrink or eliminate CUPA-T2*'s white-matter advantage; this can be checked by simulation before any clinical study.
Because the regularizer uses per-tissue Pearson correlation on log-variances, it rewards monotone association but not calibrated scale; a rank-based or quantile-matching alignment might retain most of the rho gain with less ECE penalty.
The dependence on $S=100$ Monte Carlo samples suggests a sensitivity check: repeating the R=4 experiment with $S=20$ and $S=200$ would reveal whether the benefit comes from covariance structure itself or simply from more ensemble averaging.
CSF performance degrades at R=4 under the current uncertainty-based alpha schedule, so explicitly conditioning $\alpha_v$ on tissue type in addition to $u_v$ is a natural way to recover high-uncertainty regions without losing the white-matter gain.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things you should know. First, this is a real methodological step: nobody else, as far as I can tell, has propagated voxel-wise inter-echo covariance from stochastic reconstructions into downstream T2* fitting, and the tissue-stratified RMS correlation regularizer is a reasonable way to avoid variance-dominated alignment. The authors release code and are honest about the calibration/selective prediction trade-off. Second, the paper's strongest-sounding result—that CUPA-T2* substantially increases alignment ρ between reconstruction uncertainty and predicted variance—is largely circular, because the training loss directly maximizes that Pearson correlation. The ablation shows λ=0 gives ρ=0.250, λ=0.01 gives 0.370, so it's doing what it's told. That's not a discovery; it's a check on the optimizer.

The independently interesting results are the tissue-specific ones: WM NRMSE improves at R=3 and R=4 (0.135 vs 0.137, 0.168 vs 0.174), while overall NRMSE is slightly worse than the simpler PUQ baseline. The effect sizes are small, but the authors use proper statistical tests and don't oversell. The covariance ablation (ΣI vs full Σ) also suggests the off-diagonal structure matters, which is the core claim.

Now the soft spots. The inter-echo covariance is estimated per voxel from only S=100 MC dropout reconstructions, giving a 24×24 sample covariance with no shrinkage, eigenvalue clipping, or spatial pooling. Hundred samples in 24 dimensions from a highly correlated process will produce a poorly conditioned matrix, and the Cholesky factor will amplify directions of small estimated variance. The paper even attributes the WM benefit to 'more reliable inter-echo covariance estimates' in white matter, which makes this assumption load-bearing. If the covariance is noisy, the ΣI vs full Σ comparison could reflect estimation noise rather than true inter-echo structure. This needs a sensitivity analysis—e.g., varying S, adding shrinkage, or comparing against a larger-sample reference.

Also missing: values for α_min, α_max, and γ in the sampling schedule are never given, which makes the method hard to reproduce. The λ and β are reported, so the omission is fixable.

Bottom line: worth a serious referee. The idea is new, the writing is clear, the limitations are acknowledged, and the code is out. But the circular alignment metric and the unvalidated covariance estimate need to be addressed before I'd trust the quantitative claims. Recommend major revision, not rejection.

Referee Report

5 major / 5 minor

Summary. The paper introduces CUPA-T2*, a two-stage framework for T2* mapping from accelerated multi-echo MRI. It estimates a per-voxel inter-echo covariance matrix from S=100 Monte Carlo dropout reconstructions, factorizes it via Cholesky decomposition, and generates uncertainty-aware complex-valued samples whose magnitudes feed a heteroscedastic MLP. The MLP is trained with a beta-NLL loss and an RMS-based Pearson correlation regularizer that aligns predicted log-variance with reconstruction-derived RMS uncertainty. Experiments on the public T2*-MOVE dataset show statistically significant white-matter NRMSE improvements at R=3 and R=4 relative to two baselines, along with a large increase in the correlation between reconstruction uncertainty and predicted variance, at the cost of worse ECE and AURC. An ablation at R=4 isolates the contributions of covariance-aware sampling, full versus diagonal covariance, and the alignment weight.

Significance. If the technical claims hold, CUPA-T2* is a meaningful step beyond scalar uncertainty propagation in accelerated quantitative MRI, explicitly carrying inter-echo covariance structure into downstream parameter fitting. The paper's strengths include public code and dataset, a clear ablation design, and statistical tests on tissue-specific metrics. The alignment regularizer and the WM improvements are potentially useful for biomarker studies if the covariance estimation is reliable. However, the central headline result—the increased alignment ρ—is to a large extent an optimization artifact because ρ is exactly the objective being maximized, and the reliability of the per-voxel covariance estimate remains a load-bearing but unexamined assumption.

major comments (5)

[§2.3, Table 3] The reported increase in ρ is expected by construction, because ρ is the exact objective being optimized through L = L_β-NLL + λ(1−ρ). Reporting this as a finding is circular; it should be framed as a sanity check that the regularizer works as intended. The independent evidence for the method lies in the WM NRMSE improvements and the AURC/ECE trade-off, and these should be foregrounded in the abstract and conclusions.
[§2.2, §3.3] The per-voxel 24×24 empirical covariance Σ_v is estimated from only S=100 Monte Carlo samples, with no shrinkage, eigenvalue clipping, or spatial pooling described. The paper's own explanation for the WM benefit is that white matter yields 'more reliable inter-echo covariance estimates,' which makes the conditioning of Σ_v load-bearing for the central claim. Please report the condition number or eigenvalue spectrum of Σ_v, or introduce covariance regularization, and demonstrate that the Cholesky-based sampling is not dominated by estimation noise rather than true inter-echo structure.
[§2.3] The test-time protocol is ambiguous. The method section states that the magnitude of the sampled representation x̃_v serves as input to the fitting network, but later states 'During validation and testing, the MC mean and standard deviation are used.' If sampling is applied only during training as a data-augmentation strategy and the test-time input is deterministic, then the claim of explicitly propagating covariance to downstream fitting at inference is not supported. Please clarify whether test-time inputs are stochastic, and, if not, explain how the covariance structure is preserved at inference.
[§2.2, Eq. (1)] The sampling coefficient α_v is a function of u_v, which is itself derived from the trace of the same covariance matrix that produces L_v. This double-counts the uncertainty magnitude and may confound the effect of the off-diagonal covariance structure with the noise schedule. An ablation with a constant α (or with α based on an independent uncertainty estimate) would help isolate the contribution of the covariance structure per se.
[§3.3, Table 2] The reported WM improvements are small (0.135 vs 0.137 at R=3; 0.168 vs 0.174 at R=4). Since the paper performs many tissue × acceleration Wilcoxon signed-rank tests at p=0.001, please state whether any multiple-comparison correction was applied and report the number of independent units (subjects or slices) entering each test. Without this, the statistical superiority claims are not fully supported.

minor comments (5)

[§2.3, Eq. (2)] The definition of the β-NLL loss in Eq. (2) is difficult to parse because of the nested exp/log structure; please add a short derivation or reformulate it to make the stop-gradient and β-exponent roles clearer.
[§3.2] The abbreviation PUQ is used for the baseline from Sun et al. but the text calls it a 'non-heteroscedastic MLP (PUQ)'; please define the acronym explicitly in the experiments section to avoid confusion with the phase-wise uncertainty method in the cited work.
[Table 4] In the ablation, λ=0.0 achieves the lowest WM NRMSE (0.167) while the full method (λ=0.05) yields 0.168; the discussion should explicitly address why λ=0.05 was chosen despite this, beyond the statement that it gave the best validation performance in most settings.
[§2.2] The text says 'a covariance factor L_v is calculated such that Σ_v ≈ L_v L_v^T' but does not state how non-positive-definite empirical covariances are handled; please mention whether eigenvalues are clamped or a diagonal jitter is added.
[§3.1] The paper does not report the computational overhead of generating S=100 MC reconstructions and computing per-voxel covariance matrices; a brief runtime or complexity note would help readers assess practical feasibility.

Circularity Check

1 steps flagged · score 6.0 of 10

The reported ρ improvement is the regularizer being optimized, so that headline result reduces to the loss definition; the WM NRMSE and AURC/ECE results remain independent.

self definitional [Section 2.3 (loss definition), Abstract, Table 3]
"we introduce an RMS-based correlation regularizer. Specifically, we align the predicted log-variance with the reconstruction-derived RMS uncertainty u_v by adding ρ = Pearson(ˆℓ^{T2*}_v, log u_v) to L = L_β-NLL + λ(1−ρ). ... Compared with a heteroscedastic baseline, the proposed framework substantially increases alignment between reconstruction uncertainty and predicted T2* variance."

The alignment metric ρ is both the training objective and the headline result. The loss is explicitly defined as L = L_β-NLL + λ(1−ρ), so maximizing ρ is the optimization target. Table 3 then reports that CUPA-T2* substantially raises ρ (e.g., 0.766 vs 0.280 at R=2), which is the optimizer doing exactly what the loss asks. The reported 'increase in alignment' is therefore contained in the loss definition by construction, not an independent empirical discovery. The WM NRMSE improvement and the AURC/ECE trade-off are not directly optimized by this term and provide independent content, so the circularity is partial rather than total.

full rationale

CUPA-T2*'s core covariance-aware sampling pipeline (Eq. 1) is not circular: the inter-echo covariance is estimated from MC-dropout reconstructions, and the downstream T2* fitting accuracy is evaluated with NRMSE/SSIM, which are not the training objective beyond the fitting loss. The tissue-specific WM improvement at R=3 and R=4 is an independent empirical finding. However, the paper's prominent claim of 'substantially increases alignment' is circular in the specific, quotable sense that the Pearson correlation ρ between predicted log-variance and reconstruction RMS uncertainty is the exact quantity being optimized by the loss L = L_β-NLL + λ(1−ρ). Reporting a large ρ gain over the non-regularized Het baseline is reporting success on the training objective, not a free prediction. The AURC/ECE trade-off and tissue-dependent accuracy remain non-circular and give the paper independent grounding. The statistical robustness concern about estimating a 24×24 covariance from 100 MC samples is a validity issue, not a circularity issue, and does not affect the score.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central machinery rests on four assumptions: MC dropout as a Bayesian approximation, the mono-exponential T2* model, the reliability of a 100-sample covariance estimate for a 24-dimensional quantity, and the conceptual mapping from reconstruction epistemic uncertainty to aleatoric input variability. The first two are standard in the field; the last two are specific to this method and are not independently validated.

free parameters (4)

α_min, α_max, γ (sampling schedule parameters) = not reported
Control the noise injection strength α_v = α_min + (α_max - α_min) * û_v^γ in Eq. (1); presumably chosen by hand or validation, exact values omitted.
λ (alignment weight) = 0.05
Regularizer weight in L = L_β-NLL + λ(1−ρ), chosen as best validation performance across acceleration rates (Section 2.3).
β (beta-NLL exponent) = 0.5
Taken from Seitzer et al. [10]; not tuned here, but a constant in the loss that affects training.
Dropout probability p = 0.3
Inserted in reconstruction network denoiser layers (Section 2.1); hyperparameter, not reported as tuned.

assumptions (4)

domain assumption MC dropout with p=0.3 approximates Bayesian model uncertainty for the reconstruction network
Section 2.1 follows Gal & Ghahramani [6]; the entire epistemic uncertainty estimate relies on this approximation being valid.
domain assumption T2* signal follows a mono-exponential decay model y = S0 * exp(-TE/T2*)
Section 2.3 uses this model to define reference targets via non-linear least squares; if the true signal deviates, all targets and downstream training are biased.
ad hoc to paper The multi-echo reconstruction uncertainty at each voxel is well represented by a full-rank Gaussian covariance estimated from S=100 samples
Section 2.2 assumes Σ_v ≈ L_v L_v^T via Cholesky; no validation that 100 samples suffice for a 24-dimensional covariance (E=12 echoes, real/imaginary).
ad hoc to paper Reconstruction-derived epistemic uncertainty acts as effective input-dependent aleatoric variability for the regression
Section 2.3 states this conceptual mapping; the RMS correlation regularizer is built on it, but the paper does not test the assumption directly.

how reviews work

0 comments

Cite this review

Pith. "Pith review of CUPA-T2*: Covariance-Aware Uncertainty Propagation and Alignment for T2* Mapping in Accelerated MRI." pith.science (2026). https://pith.science/paper/HIDBGSYC

@misc{pith2026260808693,
  author       = {Pith},
  title        = {Pith review of: CUPA-T2*: Covariance-Aware Uncertainty Propagation and Alignment for T2* Mapping in Accelerated MRI},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HIDBGSYC}},
  note         = {Machine review of arXiv:2608.08693}
}

read the original abstract

Quantitative T2* maps have strong potential for biomarker discovery but are limited by long scan times, rendering them impractical in clinical settings. Significant acceleration can be achieved through undersampling in k-space combined with learning-based reconstruction. However, reconstruction artifacts and noise can propagate into downstream T2* fitting, degrading its accuracy. We introduce CUPA-T2*, a framework that explicitly propagates voxel-wise inter-echo uncertainty from stochastic Monte Carlo dropout reconstructions to downstream T2* fitting via covariance-aware sampling. T2* fitting is performed with a heteroscedastic MLP and a correlation-based regularizer that encourages alignment between predicted variance and reconstruction uncertainty. Experiments on accelerated brain MRI data show tissue-dependent behavior: CUPA-T2* achieves competitive overall T2* fitting performance and improves white-matter performance at higher accelerations. Compared with a heteroscedastic baseline, the proposed framework substantially increases alignment between reconstruction uncertainty and predicted T2* variance, while also revealing a trade-off with calibration (ECE) and selective prediction performance (AURC). CUPA-T2* enables reconstruction uncertainty-aware T2* fitting and delivers voxel-wise uncertainty maps to support the interpretation of quantitative T2* estimates.

Figures

Figures reproduced from arXiv: 2608.08693 by the authors.

**Figure 1.** Mean reconstructions across echo times (TE), computed over MC-dropout samples, with the corresponding voxel-wise reconstruction uncertainty map (summarized as the RMS marginal standard deviation) and an example voxel-wise inter-echo covariance matrix [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 2.** Overview of the proposed CUPA-T2* framework. (A) An unrolled multi-echo reconstruction network with MC dropout. (B) Voxel-wise inter-echo covariance is estimated from MC samples and factorized to enable covariance-aware sampling, modulated by RMS uncertainty. (C) A heteroscedastic T2* regressor predicts mean and variance, trained with β-NLL and RMS-based cross-stage uncertainty losses. components across echoes, fr… view at source ↗

**Figure 3.** Tissue-dependent uncertainty and sampling. (a) Tissue segmentation with example high- and low-u voxels. (b) Uncertainty distributions per tissue. (c) Sampling schedules α(u). (d) Mean ± 1 SD MC signal decay for high- and low-u voxels [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Qualitative comparison of T2* estimation and uncertainty. First column: reference and predicted T2* maps. Second column: absolute error maps. Columns 3–5: total predictive uncertainty, aleatoric uncertainty, and epistemic uncertainty, respectively [PITH_FULL_IMAGE:fi…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 13 canonical work pages

[1]

Magnetic Resonance Imaging23(1), 1–25 (2005)

Haacke, E.M., Cheng, N.Y.C., House, M.J., Liu, Q., Neelavalli, J., Ogg, R.J., Khan, A., Ayaz, M., Kirsch, W., Obenaus, A.: Imaging iron stores in the brain using magnetic resonance imaging. Magnetic Resonance Imaging23(1), 1–25 (2005)

work page 2005
[2]

Kendall, A., Gal, Y.: What uncertainties do we need in Bayesian deep learning for computer vision? Advances in Neural Information Processing Systems30(2017)

work page 2017
[3]

Brain Sciences15(9), 944 (2025)

Zhou, C.-H., Zhu, Y.-C.: Imaging of cerebral iron as an emerging marker for brain aging, neurodegeneration, and cerebrovascular diseases. Brain Sciences15(9), 944 (2025). https://doi.org/10.3390/brainsci15090944

work page doi:10.3390/brainsci15090944 2025
[4]

et al.: T2*-MOVE: An Open k-Space Dataset to Address Motion in T2* Quantification from Gradient Echo MRI

Eichhorn, H. et al.: T2*-MOVE: An Open k-Space Dataset to Address Motion in T2* Quantification from Gradient Echo MRI. Nefeli, Helmholtz Zentrum München (2026). https://nefeli.helmholtz-munich.de/records/9hqjb-tfw77

work page 2026
[5]

Magnetic Resonance in Medicine95, 346–362 (2026)

Eichhorn, H., Spieker, V., Hammernik, K., Saks, E., Felsner, L., Weiss, K., Preibisch, C., Schnabel, J.A.: Motion-robust T ∗ 2 quantification from low-resolution gradi- ent echo brain MRI with physics-informed deep learning. Magnetic Resonance in Medicine95, 346–362 (2026). https://doi.org/10.1002/mrm.70050

work page doi:10.1002/mrm.70050 2026
[6]

In: Proceedings of the 33rd International Conference on Machine Learning (ICML), PMLR 48, 1050–1059 (2016) 10 G

Gal,Y.,Ghahramani,Z.:DropoutasaBayesianapproximation:Representingmodel uncertainty in deep learning. In: Proceedings of the 33rd International Conference on Machine Learning (ICML), PMLR 48, 1050–1059 (2016) 10 G. Rouwendaal et al

work page 2016
[7]

In: International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), pp

Sun, H., Li, Z., Du, C., Li, H., Wang, Y., Chen, H.: Guiding quantitative MRI reconstruction with phase-wise uncertainty. In: International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), pp. 248–257. Springer (2025)

work page 2025
[8]

D. P. Kingma and J. Ba, Adam: A Method for Stochastic Optimization, inProceed- ings of the International Conference on Learning Representations (ICLR), 2015

work page 2015

Show all 24 references

[9]

Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, Image quality assess- ment: From error visibility to structural similarity,IEEE Transactions on Image Processing, vol. 13, no. 4, pp. 600–612, 2004. doi:10.1109/TIP.2003.819861

2004
[10]

arXiv preprint arXiv:2203.09168 (2022)

Seitzer, M., Tavakoli, A., Antic, D., Martius, G.: On the Pitfalls of Heteroscedas- tic Uncertainty Estimation with Probabilistic Neural Networks. arXiv preprint arXiv:2203.09168 (2022). https://doi.org/10.48550/arXiv.2203.09168

2022 doi arXiv
[11]

Frontiers in Neurology 11, 366 (2020)

Pyatigorskaya, N., Sanz-Morère, C.B., Gaurav, R., Biondetti, E., Valabregue, R., Santin, M., Yahia-Cherif, L., Lehéricy, S.: Iron imaging as a diagnostic tool for Parkinson’s disease: A systematic review and meta-analysis. Frontiers in Neurology 11, 366 (2020). https://doi.org...

2020
[12]

Magnetic Resonance in Medicine90(3), 1172–1208 (2023)

Shafieizargar, B., Byanju, R., Sijbers, J., Klein, S., den Dekker, A.J., Poot, D.H.J.: Systematic review of reconstruction techniques for accelerated quantitative MRI. Magnetic Resonance in Medicine90(3), 1172–1208 (2023). https://doi.org/10.1002/mrm.29721

2023 doi
[13]

IEEE Transactions on Medical Imaging40(1), 239–250 (2021)

Edupuganti, V., Mardani, M., Vasanawala, S., Pauly, J.: Uncertainty quantification in deep MRI reconstruction. IEEE Transactions on Medical Imaging40(1), 239–250 (2021). https://doi.org/10.1109/TMI.2020.3025065

2021
[14]

Journal of Imaging Informatics in Medicine38(4), 2071–2084 (2025)

Ekanayake, M., Pawar, K., Chen, Z., Egan, G., Chen, Z.: PixCUE: Joint uncer- tainty estimation and image reconstruction in MRI using deep pixel classification. Journal of Imaging Informatics in Medicine38(4), 2071–2084 (2025)

2025
[15]

arXiv preprint arXiv:1907.13418 (2019)

Tanno, R., Worrall, D., Kaden, E., Ghosh, A., Grussu, F., Bizzi, A., Sotiropou- los, S.N., Criminisi, A., Alexander, D.C.: Uncertainty quantification in deep learn- ing for safer neuroimage enhancement. arXiv preprint arXiv:1907.13418 (2019). https://doi.org/10.48550/arXiv.1907.13418

2019 doi arXiv
[16]

https://doi.org/10.1002/mrm.28117

Glang, F., Deshmane, A., Prokudin,S.,et al.: DeepCEST 3T:Robust MRI parame- ter determination and uncertainty quantification with neural networks—application to CEST imaging of the human brain at 3T Magnetic Resonance in Medicine84, 450–466 (2020). https://doi.org/10.1002/mrm.28117

2020 doi
[17]

MAGMA37(3), 335–368 (2024)

Heckel, R., Jacob, M., Chaudhari, A., Perlman, O., Shimron, E.: Deep learning for accelerated and robust MRI reconstruction. MAGMA37(3), 335–368 (2024). https://doi.org/10.1007/s10334-024-01173-8

2024 doi
[18]

In: 2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI), pp

Jeelani, H., Yang, Y., Zhou, R., Kramer, C.M., Salerno, M., Weller, D.S.: A my- ocardial T1-mapping framework with recurrent and U-Net convolutional neural net- works. In: 2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI), pp. 1941–1944 (2020). https://doi.or...

2020
[19]

Magnetic Reso- nance in Medicine82, 174–188 (2019)

Liu, F., Feng, L., Kijowski, R.: MANTIS: Model-augmented neural network with incoherent k-space sampling for efficient MR parameter mapping. Magnetic Reso- nance in Medicine82, 174–188 (2019). https://doi.org/10.1002/mrm.27707

2019 doi
[20]

Medical Image Analysis70, 102017 (2021)

Jun, Y., Shin, H., Eo, T., Kim, T., Hwang, D.: Deep model-based magnetic resonance parameter mapping network (DOPAMINE) for fast T1 mapping us- ing variable flip angle method. Medical Image Analysis70, 102017 (2021). https://doi.org/10.1016/j.media.2021.102017

2021
[21]

IEEE Transactions on Computational Imaging10, 628–639 (2024)

Zimmermann, F.F., Kolbitsch, C., Schuenke, P., Kofler, A.: PINQI: An end-to-end physics-informed approach to learned quantitative MRI recon- CUPA-T2* 11 struction. IEEE Transactions on Computational Imaging10, 628–639 (2024). https://doi.org/10.1109/TCI.2024.3388869

2024
[22]

In: International Workshop on Uncertainty for Safe Utilization of Machine Learning in Medical Imaging, pp

Fischer, P., Thomas, K., Baumgartner, C.F.: Uncertainty estimation and propaga- tion in accelerated MRI reconstruction. In: International Workshop on Uncertainty for Safe Utilization of Machine Learning in Medical Imaging, pp. 84–94. Springer, Cham (2023)

2023
[23]

In: International Workshop on Uncertainty for Safe Utilization of Machine Learning in Medical Imaging, pp

Feiner, L.F., Menten, M.J., Hammernik, K., Hager, P., Huang, W., Rueckert, D., Braren, R.F., Kaissis, G.: Propagation and attribution of uncertainty in medical imaging pipelines. In: International Workshop on Uncertainty for Safe Utilization of Machine Learning in Medical Imag...

2023
[24]

Frontiers in Immunology9, 255 (2018)

Gillen, K.M., Mubarak, M., Nguyen, T.D., Pitt, D.: Significance and In Vivo Detec- tion of Iron-Laden Microglia in White Matter Multiple Sclerosis Lesions. Frontiers in Immunology9, 255 (2018). https://doi.org/10.3389/fimmu.2018.00255

2018

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Magnetic Resonance Imaging23(1), 1–25 (2005)

Haacke, E.M., Cheng, N.Y.C., House, M.J., Liu, Q., Neelavalli, J., Ogg, R.J., Khan, A., Ayaz, M., Kirsch, W., Obenaus, A.: Imaging iron stores in the brain using magnetic resonance imaging. Magnetic Resonance Imaging23(1), 1–25 (2005)

work page 2005

[2] [2]

Kendall, A., Gal, Y.: What uncertainties do we need in Bayesian deep learning for computer vision? Advances in Neural Information Processing Systems30(2017)

work page 2017

[3] [3]

Brain Sciences15(9), 944 (2025)

Zhou, C.-H., Zhu, Y.-C.: Imaging of cerebral iron as an emerging marker for brain aging, neurodegeneration, and cerebrovascular diseases. Brain Sciences15(9), 944 (2025). https://doi.org/10.3390/brainsci15090944

work page doi:10.3390/brainsci15090944 2025

[4] [4]

et al.: T2*-MOVE: An Open k-Space Dataset to Address Motion in T2* Quantification from Gradient Echo MRI

Eichhorn, H. et al.: T2*-MOVE: An Open k-Space Dataset to Address Motion in T2* Quantification from Gradient Echo MRI. Nefeli, Helmholtz Zentrum München (2026). https://nefeli.helmholtz-munich.de/records/9hqjb-tfw77

work page 2026

[5] [5]

Magnetic Resonance in Medicine95, 346–362 (2026)

Eichhorn, H., Spieker, V., Hammernik, K., Saks, E., Felsner, L., Weiss, K., Preibisch, C., Schnabel, J.A.: Motion-robust T ∗ 2 quantification from low-resolution gradi- ent echo brain MRI with physics-informed deep learning. Magnetic Resonance in Medicine95, 346–362 (2026). https://doi.org/10.1002/mrm.70050

work page doi:10.1002/mrm.70050 2026

[6] [6]

In: Proceedings of the 33rd International Conference on Machine Learning (ICML), PMLR 48, 1050–1059 (2016) 10 G

Gal,Y.,Ghahramani,Z.:DropoutasaBayesianapproximation:Representingmodel uncertainty in deep learning. In: Proceedings of the 33rd International Conference on Machine Learning (ICML), PMLR 48, 1050–1059 (2016) 10 G. Rouwendaal et al

work page 2016

[7] [7]

In: International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), pp

Sun, H., Li, Z., Du, C., Li, H., Wang, Y., Chen, H.: Guiding quantitative MRI reconstruction with phase-wise uncertainty. In: International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), pp. 248–257. Springer (2025)

work page 2025

[8] [8]

D. P. Kingma and J. Ba, Adam: A Method for Stochastic Optimization, inProceed- ings of the International Conference on Learning Representations (ICLR), 2015

work page 2015