Pith. sign in

REVIEW 3 major objections 5 minor 182 references

This paper claims that a self-supervised neural reconstruction that couples multi-echo volumes through a Bloch-equation T2-decay regularizer yields the first submillimeter fetal brain T2 maps at 0.55 T and cuts acquisition from 15 to 5 minu

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 10:39 UTC pith:442LNDI7

load-bearing objection Useful multi-echo SVR with real in-vivo data, but T2 'accuracy' claims are self-referential and need an independent standard. the 3 major comments →

arxiv 2607.20136 v1 pith:442LNDI7 submitted 2026-07-22 physics.med-ph cs.AI

PRIME-SVR: Physics-infoRmed Implicit Multi-Echo Slice-to-Volume Reconstruction for Fetal T2 mapping

Busra Bulut , Maik Dannecker , Thomas Sanchez , Sara Neves Silva , Steven Jia , Jean-Baptiste Ledoux , Leo Pomar , Joanna Sichitiu
show 9 more authors
Yvan Gomez Meriam Koob Vincent Dunet Maria Deprez Guillaume Auzias Francois Rousseau Jana Hutter Daniel Rueckert Meritxell Bach Cuadra
This is my paper
classification physics.med-ph cs.AI
keywords fetal brain MRIslice-to-volume reconstructionimplicit neural representationT2 mappingmulti-echo MRIBloch equation regularizationlow-field MRIself-supervised learning
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper aims to make quantitative fetal brain T2 mapping clinically practical by reconstructing high-resolution volumes jointly across multiple echo times instead of reconstructing each echo independently. It argues that a single continuous neural volume shared across TEs, plus a physics-based regularizer that forces every voxel's log-signal to lie in the mono-exponential T2-decay subspace, gives sharper, more anatomically accurate reconstructions at late echo times and low field strength, where standard slice-to-volume reconstruction fails. A sympathetic reader would care because T2 is a protocol- and center-independent biomarker of fetal brain maturation, and the method is self-supervised, so it works across scanners, field strengths, and non-clinical echo times without retraining. The paper reports the first 0.8 mm isotropic T2 maps of the fetal brain at 0.55 T, and it shows the acquisition can be shortened from 15 to 10 or 5 minutes while keeping T2 errors within 1.7–2.3%.

Core claim

The central claim is that slice-to-volume reconstruction can be done jointly across echo times by representing the high-resolution volume as a continuous function of 3D coordinates that outputs intensities for all TEs, and by regularizing that function so that, at every point, the log-signal across TEs lies in the plane spanned by a constant and the echo time (the mono-exponential Bloch decay). With a second network estimating per-slice motion, intensity, and outlier weights, this fully self-supervised model reconstructs late-echo volumes that single-echo methods cannot handle and produces T2 maps directly from the decay fit. On 13 fetuses at two centers, 1.5 T and 0.55 T, the paper reports

What carries the argument

The load-bearing device is a subspace projection derived from the Bloch equations. For three echo times, the vector of log-intensities y(x) = [log V_1(x), log V_2(x), log V_3(x)] should lie in the 2D column space of D = [[1, TE_1], [1, TE_2], [1, TE_3]] if each voxel decays as M0 exp(-TE/T2). The regularizer computes ||(P - I)y(x)||^2, with P the orthogonal projector P = D(D^T D)^{-1}D^T, and penalizes any deviation, so the shared volume network is pulled toward physically consistent T2 decay. The same P - I matrix is precomputed, so the cost is negligible. An adaptive weight raises this coupling for stacks flagged as low quality, strengthening the prior exactly when data are most corrupted.

Load-bearing premise

The central assumption is that the signal at each reconstructed voxel decays as a single exponential with one T2 value, unaffected by the way the fast MRI sequence excites each slice; if that is wrong, the cross-echo coupling will pull the reconstruction toward a biased T2.

What would settle it

Use a phantom with known T2 values spanning the fetal range (roughly 200–400 ms), scanned with the same single-shot fast spin-echo multi-echo protocol at 0.55 T and 1.5 T, with and without simulated slice motion. If PRIME-SVR's T2 values drift by more than the quoted 1.7–2.3% when slice thickness, refocusing angles, or the number of stacks are changed, or if they deviate systematically from a reference multi-echo spin-echo measurement, the mono-exponential subspace assumption and its adaptive weighting are the cause.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Joint multi-echo reconstruction succeeds at late echo times and at 0.55 T, where single-echo SVR and T2 fitting previously failed, producing the first 0.8 mm isotropic T2 maps of the fetal brain at 0.55 T.
  • Because the volume network is shared across TEs, data from all TEs contribute to every reconstructed volume, so fewer stacks per TE are needed: T2 error stays within 1.7% in white and deep gray matter with two stacks per TE, and within 2.3% with one stack per TE for high-quality acquisitions.
  • The approach cuts the multi-echo acquisition from 15 minutes to 10 or 5 minutes, reducing a major practical barrier to routine quantitative fetal MRI.
  • Cross-TE structural consistency improves by 14% and reconstruction sharpness by 47% relative to standard single-TE slice-to-volume reconstruction, with lower residual T2 fit error.
  • The method is fully self-supervised and does not require training data at specific echo times, so it can be applied at non-clinical TEs and different field strengths without retraining.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the same subspace-projection regularizer could be extended to a multi-exponential or extended-phase-graph signal model, which would address the paper's own stated limitation that the simple mono-exponential fit overestimates T2 because slice-profile effects are unmodeled.
  • Editorial inference: the quality-adaptive weighting scheme suggests a general recipe for physics-informed self-supervised reconstruction—use the physical prior more aggressively when the data are degraded and back off when the data are clean—which could transfer to T1 or T2* mapping and to other motion-corrupted quantitative imaging.
  • Editorial inference: the ability to reconstruct with one stack per TE implies acquisition protocols could be redesigned around fewer, faster stacks spread across echo times rather than many stacks at a single contrast; a prospective study could test whether 5-minute protocols preserve the maturational T2 trajectories the paper expects in white matter.
  • Editorial inference: because the volume representation is continuous and resolution-agnostic, the same framework should apply to other moving organs with quantitative mapping, such as musculoskeletal T2 or cardiac T1 mapping, with the caveat that non-rigid motion would need to be modeled.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. PRIME-SVR proposes a self-supervised implicit neural representation (INR) framework for joint multi-echo slice-to-volume reconstruction of fetal brain MRI, with a second network estimating slice-specific degradations and a Bloch-equation-derived regularization coupling the reconstructed log-signal across echo times. The method is evaluated on 39 in vivo acquisitions (13 subjects × 3 TEs) from two centers, two vendors, and two field strengths, and compared against NeSVoR and SVRTK. The authors report improved sharpness, anatomical consistency, and cross-TE coherence, enable reconstruction at late TEs and at 0.55 T, produce T2 maps where per-TE SVR fails, and claim that reduced input data (2 or 1 stacks per TE) yields T2 errors within 1.7–2.3% of the full-data reconstruction, corresponding to acquisition time reductions from 15 to 5–10 minutes.

Significance. If the quantitative claims were supported, this would be a substantial contribution: it would make fetal T2 mapping practical at low field, remove the dependence on per-TE SVR pipelines, and reduce acquisition time. The paper's reconstruction-quality evidence is generally plausible: the multi-echo INR with learned slice parameters is a sensible extension of NeSVoR, the two-center, two-field-strength dataset is a genuine strength, and the reported improvements in SSIM, sharpness, and cross-TE edge consistency are internally consistent with qualitative examples. However, the central quantitative T2 accuracy claims rest on a self-referential evaluation: the reduced-data 'accuracy' is measured against the method's own full-data reconstruction, and the Bloch regularizer actively shapes the T2 values being evaluated. The paper also openly acknowledges that the mono-exponential model overestimates T2 because slice-profile effects are not modeled. These issues do not invalidate the reconstruction method, but they do invalidate the paper's present wording that PRIME-SVR 'preserves T2 accuracy within 1.7%' and 'enables quantitative T2 mapping' as an established measurement. The manuscript ne

major comments (3)
  1. [Section 6.3, Table 4] The central quantitative claim — T2 error within 1.7% for 2 stacks/TE and 2.3% for 1 stack/TE — is computed as the mean absolute difference between T2 maps reconstructed from reduced data and T2 maps reconstructed from the full three stacks per TE using the same PRIME-SVR pipeline. This is a consistency measure, not an accuracy measure. Because the full-data reference itself contains the same network, the same Bloch regularizer, and the same T2 fitting procedure, the errors cannot capture systematic bias. The abstract and Section 7.3 therefore overstate the result by calling it 'T2 accuracy'. The authors should either validate against an independent T2 standard (phantom, adult reference, EPG-based dictionary fitting, or an independent reconstruction pipeline without the Bloch prior) or explicitly restrict the claim to self-consistency under data reduction.
  2. [Section 3.4, Eq. (5); Table 6] The Bloch regularizer projects each voxel's log-signal onto the two-dimensional column space of D = [[1, TE_i]], which is the same mono-exponential model later used to estimate T2 in Section 3.5. The regularizer therefore does not independently verify T2; it actively enforces the assumed decay shape. Table 6 shows the effect is numerically large: changing the fixed regularization weight from 0.1 to 20 shifts mean WM T2 from 353.8 ms to 377.4 ms and DGM T2 from 260.2 ms to 281.8 ms. The adaptive weighting keeps values close to the alpha=0.1 case, but this does not remove the circularity: the scale of the reported T2 values is partly determined by the prior rather than by measured signal. The authors should quantify the residual influence of the Bloch prior on the final T2 maps, for example by reporting the difference between T2 from the full PRIME-SVR reconstruction and T2 from a version
  3. [Section 7.4; Section 7.2] The authors concede that the simple mono-exponential model overestimates T2 because the SST2w slice profile is not modeled, and that EPG-based dictionary fitting would be required to correct the bias. This concession is in direct tension with the abstract and conclusion presenting the produced maps as quantitatively accurate T2 maps. The discussion in Section 7.2 also invokes the low residual fitting error as evidence that reconstruction quality is the main bottleneck; however, a low residual is expected when both the regularizer and the fitting procedure assume the same mono-exponential model. The paper should either correct the known slice-profile bias before presenting T2 values, or clearly frame the reported T2 values as demonstrations of feasibility and internal consistency, not as validated quantitative biomarker measurements.
minor comments (5)
  1. [Abstract and Section 7.3] The 15-minute acquisition time and the reduction to 10 or 5 minutes are quoted in the abstract, but the mapping from stacks per TE to acquisition minutes is never explicitly derived in the methods. Please specify the per-stack acquisition time or otherwise justify the time figures.
  2. [Section 3.2, Eq. (2)] The notation NTE is used inconsistently (N_TE vs NTE). Also, the line 'For clarity of exposition, more information are in the appendix' is a grammatical error and should be rephrased.
  3. [Section 6.3, Table 4 caption] The caption says subjects with q̄ < 0.9 are discarded, but the sample sizes differ across rows (6 vs 5, 5 vs 4). Please explicitly state the exclusion counts and whether the same subjects are used for the 2-stack and 1-stack conditions at each field strength.
  4. [Section 5.2, Experiment 2] The comparison of T2 values against SVRTK is useful as a method comparison, but SVRTK is not a ground truth. Please make clear in the text that Table 3 reports agreement between two reconstruction pipelines, not accuracy of either one.
  5. [Section 4.2] Training is stopped at epoch 100 with no explicit stopping criterion. Given that the Bloch regularization and adaptive weighting influence the result, a sentence justifying this fixed epoch choice and its stability across runs would improve reproducibility.

Circularity Check

2 steps flagged

Quantitative T2 claims are self-referential: reduced-data 'T2 accuracy' is measured against PRIME-SVR's own full-data output, and the Bloch regularizer imposes the same mono-exponential subspace later used for T2 fitting; reconstruction-quality comparisons remain independent.

specific steps
  1. fitted input called prediction [Section 6.3, Table 4; Abstract]
    "Table 4 summarizes the reconstruction similarity metrics and the errors in T2 estimation within WM and DGM obtained using reduced numbers of input stacks per TE, with the HR volumes reconstructed from 3 stacks per TE (full data) serving as the reference."

    The reduced-data T2 error is computed against the full-data output of the same PRIME-SVR pipeline, not against an independent T2 standard. The abstract's 'T2 accuracy within 1.7%' is therefore a self-consistency statistic between two runs of the same network with the same Bloch regularizer, not a validated accuracy bound. This is the only evidence for the central acceleration claim (15 to 10 or 5 minutes), so that claim reduces to internal reproducibility rather than quantitative correctness.

  2. self definitional [Section 3.4 Eq. (5) and Section 3.5]
    "The Bloch regularization computed on a set X of coordinates becomes R_T2 (V 1, . . . , V NT E ) = 1/|X| Σ x∈X ∥(P−I)y(x)∥ 2 2 (5) ... We use the same closed-form approximation than in Bloch regularization to initialize the T2 and M0."

    The regularizer projects each voxel's log-signal onto the column space of D=[1, TE_i], and the T2 estimator then uses exactly that same subspace and closed-form approximation. Thus the reported T2 maps are forced by construction to agree with the mono-exponential prior; the prior cannot be falsified by the data. Table 6 makes this concrete: raising the regularization weight from 0.1 to 20 shifts WM T2 from 353.8 to 377.4 ms and DGM T2 from 260.2 to 281.8 ms, showing the prior directly sets the T2 scale. Presenting these as quantitative measurements, when Section 7.4 concedes the mono-exponential model overestimates T2 due to unmodeled slice-profile effects, is circular in the accuracy claim rather than in the reconstruction algorithm itself.

full rationale

The core INR reconstruction (shared SIREN, slice acquisition model, loss Eq. 6) is self-contained and its reconstruction-quality improvements (SSIM, AES, NMI, Edge Dice) are evaluated against independent baselines, so those claims are not circular. However, the quantitative T2 story has two self-referential links. First, Experiment 3 defines 'T2 error' as the difference between reduced-data PRIME-SVR T2 maps and full-data PRIME-SVR T2 maps; no phantom, EPG dictionary, or other independent T2 reference is used, so the abstract's 1.7-2.3% 'accuracy' figure is internal consistency. Second, the Bloch regularizer (Eq. 5) constrains log-signal to the same [1, TE] subspace from which T2 is later fitted (Section 3.5), and the weight ablation shows T2 values move monotonically with regularization strength. Therefore the quantitative values are partially produced by the prior, not independently measured. The paper's own limitations (Section 7.4) acknowledge the mono-exponential model overestimates T2 because the SST2w slice profile is unmodeled and EPG dictionary fitting would be required. This does not invalidate the reconstruction contribution, but it means the headline quantitative-mapping and acceleration-accuracy claims are not supported as stated. No load-bearing self-citation circularity was found; the prior workshop paper is cited only as preliminary evidence.

Axiom & Free-Parameter Ledger

2 free parameters · 5 axioms · 0 invented entities

The central T2 maps are co-determined by the hand-set Bloch weight and the assumed monoexponential model; no new physical entities are introduced. The paper relies on standard INR capacity assumptions and a fixed Gaussian PSF from prior literature.

free parameters (2)
  • alpha(q_bar) adaptive Bloch regularization weight = 10–30 if q_bar<0.9 or <3 stacks/TE; 0–1 if q_bar>=0.9
    Hand-set schedule in Section 3.4.1. Table 6 shows T2 estimates shift monotonically with alpha (WM 353.8→377.4 ms, DGM 260.2→281.8 ms), so the central T2 values depend on this choice.
  • gamma (log-barrier weight on outlier coefficients) = not specified (stated only as 0<gamma<1)
    Controls the strength of the outlier-weight barrier in Eq. (6); no numeric value or sensitivity analysis is given.
axioms (5)
  • domain assumption Mono-exponential Bloch decay V_i(x)=M0(x)exp(-TE_i/T2(x))
    Used in Eq. (2), the Bloch regularizer Eq. (5), and the T2 fitting. The paper acknowledges in Section 7.4 that this simplified model overestimates T2 because it ignores the SST2w slice profile.
  • domain assumption Anisotropic Gaussian PSF with covariance fixed from Rousseau et al. 2006
    Supplementary Section 9.1 defines the PSF covariance with fixed factors (1.2r/2.355 etc.); the PSF is assumed known rather than estimated per acquisition.
  • domain assumption SIREN is sufficiently expressive to approximate the continuous volume
    Section 3.1 explicitly states: 'We assume that this architecture is sufficiently expressive to approximate V with high accuracy.'
  • domain assumption Slice motion is rigid and slice-specific; bias/scale are the only other degradations
    Eq. (1) models only rigid T_j, global scaling C_j, and bias field; no non-rigid motion or through-plane deformation is considered.
  • domain assumption FetMRQC quality score q captures motion artifacts and signal dropouts
    The adaptive alpha schedule in Section 3.4.1 is keyed to q_bar, so if FetMRQC is miscalibrated on multi-TE low-field data, the physics coupling is applied incorrectly.

pith-pipeline@v1.3.0-alltime-deepseek · 23655 in / 11568 out tokens · 111838 ms · 2026-08-01T10:39:40.542334+00:00 · methodology

0 comments
read the original abstract

Slice-to-volume reconstruction (SVR) is the standard method for obtaining high-resolution (HR) 3D fetal brain volumes from motion-corrupted 2D MRI slice stacks acquired in multiple orientations. Existing SVR methods are optimized and validated only for clinical-range echo times (TEs), limiting their use at non-clinical TEs and making them incompatible with quantitative T2 mapping, a protocol- and center-independent biomarker of fetal brain maturation requiring HR reconstructions across multiple TEs. We present PRIME-SVR, the first implicit neural representation (INR) framework for joint HR reconstruction from multi-echo MRI. A single fully connected network models a continuous function from spatial coordinates to signal intensities across TEs, while a second network estimates slice-specific acquisition degradations. Cross-TE coherence is enforced via a Bloch equation-derived regularization penalizing deviations from expected T2 decay, with adaptive weighting that strengthens coupling for degraded stacks. The method is fully self-supervised. We validate PRIME-SVR on 39 in vivo fetal acquisitions (13 subjects x 3 TEs) from two centers, two vendors, and two field strengths (1.5 T and 0.55 T). Compared to state-of-the-art SVR, PRIME-SVR improves reconstruction sharpness by 47%, anatomical accuracy by 30%, and cross-TE structural consistency by 14%. It enables reconstruction at late TEs previously inaccessible to SVR, yielding the first 0.8 mm isotropic T2 maps at 0.55 T and the first T2 maps derived from INR-based SVR. PRIME-SVR also accelerates quantitative imaging by reducing the data needed for multi-TE reconstruction, cutting acquisition from 15 to 10 minutes while keeping T2 accuracy within 1.7% in white and deep gray matter, or to 5 minutes with a mean T2 error of 2.3% for high-quality acquisitions.

Figures

Figures reproduced from arXiv: 2607.20136 by Busra Bulut, Daniel Rueckert, Francois Rousseau, Guillaume Auzias, Jana Hutter, Jean-Baptiste Ledoux, Joanna Sichitiu, Leo Pomar, Maik Dannecker, Maria Deprez, Meriam Koob, Meritxell Bach Cuadra, Sara Neves Silva, Steven Jia, Thomas Sanchez, Vincent Dunet, Yvan Gomez.

Figure 1
Figure 1. Figure 1: Overview of the proposed method PRIME-SVR. Multi-echo Low Resolu￾tion stacks are acquired across orthogonal orientations and TEs; a per-slice quality metric q¯ is estimated. (B) A shared SIREN Vθ maps spatial coordinates to HR intensities at each TE, while fSMθ predicts slice-specific acquisition parameters. Training combines a data-fidelity term with a Bloch regularization, adaptively weighted by q¯. (C) … view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the dataset Pre-processing. All stacks were denoised using the non-local means algorithm implemented in Advanced Normalization Tools (ANTs) (Avants et al., 2009), based on the method introduced in Manjón et al. (2010). Then, fetal brain masking was performed using FET-BET (Faghihpirayesh et al., 2024). In cases where the automated masking failed, manual corrections using ITK￾Snap (Yushkevich et… view at source ↗
Figure 3
Figure 3. Figure 3: Overview for 4 subjects, the HR reconstructions (0.8mm isotropic resolution) at [PITH_FULL_IMAGE:figures/full_fig_p025_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Overview of the HR reconstructions from PRIME-SVR across all subjects, along [PITH_FULL_IMAGE:figures/full_fig_p025_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: T2 distribution for each subject, obtained using PRIME SVR (blue) and SVRTK [PITH_FULL_IMAGE:figures/full_fig_p026_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Reconstructions obtained using 2 stacks/TE remain visually similar [PITH_FULL_IMAGE:figures/full_fig_p026_6.png] view at source ↗
Figure 6
Figure 6. Figure 6: HR reconstructions obtained with PRIME-SVR under reduced input data con [PITH_FULL_IMAGE:figures/full_fig_p027_6.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

182 extracted references · 111 canonical work pages · 20 internal anchors

  1. [1]

    Leslie Lamport, : a document preparation system, Addison Wesley, Massachusetts, 2nd edition, 1994

    1994

  2. [3]

    IEEE Journal of Biomedical and Health Informatics , author =

    Conditional. IEEE Journal of Biomedical and Health Informatics , author =. 2024 , pages =. doi:10.1109/JBHI.2024.3385504 , number =

    arXiv 2024

  3. [4]

    Glaszner, Lukas and Zach, Martin , year =. Bigger. doi:10.48550/ARXIV.2501.07376 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2501.07376

  4. [5]

    Pediatric Radiology , author =

    Biometry and volumetry in multi-centric fetal brain magnetic resonance imaging: assessing the bias of super-resolution reconstruction , volume =. Pediatric Radiology , author =. 2025 , pages =. doi:10.1007/s00247-025-06347-7 , abstract =

    work page doi:10.1007/s00247-025-06347-7 2025

  5. [6]

    FetalDiffusion: Pose-Controllable 3D Fetal MRI Synthesis with Conditional Diffusion Model

    Zhang, Molin and Golland, Polina and Grant, Patricia Ellen and Adalsteinsson, Elfar , year =. doi:10.48550/ARXIV.2404.00132 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2404.00132

  6. [8]

    Scientific Reports , author =

    A. Scientific Reports , author =. 2022 , pages =. doi:10.1038/s41598-022-10335-4 , abstract =

    work page doi:10.1038/s41598-022-10335-4 2022

  7. [9]

    and Cheng, Yu-Chung N

    Brown, Robert W. and Cheng, Yu-Chung N. and Haacke, E. Mark and Thompson, Michael R. and Venkatesan, Ramesh , year =. Magnetic resonance imaging: physical principles and sequence design , isbn =. doi:10.1002/9781118633953 , abstract =

    work page doi:10.1002/9781118633953

  8. [10]

    T2 relaxation values in the developing preterm brain , volume =. AJNR. American journal of neuroradiology , author =. 2003 , pmid =

    2003

  9. [11]

    Simulation-

    De Dumast, Priscille and Sanchez, Thomas and Lajous, Hélène and Bach Cuadra, Meritxell , editor =. Simulation-. Medical. 2023 , doi =

    2023

  10. [12]

    Prior Mismatch and Adaptation in PnP-ADMM with a Nonconvex Convergence Analysis

    Shoushtari, Shirin and Liu, Jiaming and Chandler, Edward P. and Asif, M. Salman and Kamilov, Ulugbek S. , year =. Prior. doi:10.48550/ARXIV.2310.00133 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2310.00133

  11. [13]

    Magnetic Resonance in Medicine , author =

    Model‐based super‐resolution reconstruction of. Magnetic Resonance in Medicine , author =. 2020 , pages =. doi:10.1002/mrm.27981 , abstract =

    work page doi:10.1002/mrm.27981 2020

  12. [14]

    Magnetic Resonance Imaging , author =

    Joint reconstruction of multiecho. Magnetic Resonance Imaging , author =. 2011 , pages =. doi:10.1016/j.mri.2011.03.008 , language =

    work page doi:10.1016/j.mri.2011.03.008 2011

  13. [15]

    Computerized Medical Imaging and Graphics , author =

    Model-based super-resolution reconstruction with joint motion estimation for improved quantitative. Computerized Medical Imaging and Graphics , author =. 2022 , pages =. doi:10.1016/j.compmedimag.2022.102071 , language =

    arXiv 2022

  14. [16]

    Single-subject Multi-contrast MRI Super-resolution via Implicit Neural Representations

    McGinnis, Julian and Shit, Suprosanna and Li, Hongwei Bran and Sideri-Lampretsa, Vasiliki and Graf, Robert and Dannecker, Maik and Pan, Jiazhen and Ansó, Nil Stolt and Mühlau, Mark and Kirschke, Jan S. and Rueckert, Daniel and Wiestler, Benedikt , year =. Single-subject. doi:10.48550/ARXIV.2303.15065 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2303.15065

  15. [17]

    doi:10.48550/ARXIV.2308.04079 , abstract =

    Kerbl, Bernhard and Kopanas, Georgios and Leimkühler, Thomas and Drettakis, George , year =. doi:10.48550/ARXIV.2308.04079 , abstract =

    Pith/arXiv arXiv doi:10.48550/arxiv.2308.04079

  16. [19]

    ACM Transactions on Graphics , author =

    Instant neural graphics primitives with a multiresolution hash encoding , volume =. ACM Transactions on Graphics , author =. 2022 , pages =. doi:10.1145/3528223.3530127 , abstract =

    arXiv 2022

  17. [20]

    Wang, Yinsong and Fletcher, Thomas and Luo, Xinzhe and Dineen, Aine Travers and Cusack, Rhodri and Qin, Chen , year =. Self-. doi:10.48550/ARXIV.2601.22990 , abstract =

    Pith/arXiv arXiv doi:10.48550/arxiv.2601.22990

  18. [21]

    Fast and

    Dannecker, Maik and Jia, Steven and Stolt-Ansó, Nil and Girard, Nadine and Auzias, Guillaume and Rousseau, François and Rueckert, Daniel , year =. Fast and. doi:10.48550/ARXIV.2512.11624 , abstract =

    work page doi:10.48550/arxiv.2512.11624

  19. [22]

    Domain-Agnostic Stroke Lesion Segmentation Using Physics-Constrained Synthetic Data

    Chalcroft, Liam and Crinion, Jenny and Price, Cathy J. and Ashburner, John , year =. Domain-. doi:10.48550/ARXIV.2412.03318 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2412.03318

  20. [23]

    2024 , pages =

    Medical Image Analysis , author =. 2024 , pages =. doi:10.1016/j.media.2024.103282 , language =

    arXiv 2024

  21. [25]

    Pediatric Radiology , author =

    Fetal. Pediatric Radiology , author =. doi:10.1007/s00247-004-1246-0 , language =

    work page doi:10.1007/s00247-004-1246-0

  22. [27]

    Physical Review , author =

    The. Physical Review , author =. 1946 , pages =. doi:10.1103/PhysRev.70.474 , language =

    doi:10.1103/physrev.70.474 1946

  23. [28]

    Journal of Magnetic Resonance (1969) , author =

    The signal-to-noise ratio of the nuclear magnetic resonance experiment , volume =. Journal of Magnetic Resonance (1969) , author =. 1976 , pages =. doi:10.1016/0022-2364(76)90233-X , language =

    doi:10.1016/0022-2364(76)90233-x 1969

  24. [29]

    Jia, Steven and Mercier, Chloé and Pron, Alexandre and Girard, Nadine and Auzias, Guillaume and Rousseau, François , editor =. Joint. Perinatal,. 2025 , doi =

    2025

  25. [32]

    The British Journal of Radiology , author =

    Retrospective motion correction in foetal. The British Journal of Radiology , author =. 2022 , pages =. doi:10.1259/bjr.20220071 , abstract =

    work page doi:10.1259/bjr.20220071 2022

  26. [33]

    SVoRT: Iterative Transformer for Slice-to-Volume Registration in Fetal Brain MRI

    Xu, Junshen and Moyer, Daniel and Grant, P. Ellen and Golland, Polina and Iglesias, Juan Eugenio and Adalsteinsson, Elfar , year =. doi:10.48550/ARXIV.2206.10802 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2206.10802

  27. [34]

    American Journal of Neuroradiology , author =

    Apparent. American Journal of Neuroradiology , author =. 2017 , pages =. doi:10.3174/ajnr.A5222 , language =

    work page doi:10.3174/ajnr.a5222 2017

  28. [35]

    American Journal of Neuroradiology , author =

    Normative. American Journal of Neuroradiology , author =. 2009 , pages =. doi:10.3174/ajnr.A1661 , language =

    work page doi:10.3174/ajnr.a1661 2009

  29. [36]

    Magnetic Resonance in Medicine , author =

    An efficient sequence for fetal brain imaging at. Magnetic Resonance in Medicine , author =. 2018 , pages =. doi:10.1002/mrm.27012 , abstract =

    work page doi:10.1002/mrm.27012 2018

  30. [37]

    Journal of Magnetic Resonance Imaging , author =

    Preliminary. Journal of Magnetic Resonance Imaging , author =. 2022 , pages =. doi:10.1002/jmri.28195 , abstract =

    work page doi:10.1002/jmri.28195 2022

  31. [39]

    2004 , pages =

    American Journal of Roentgenology , author =. 2004 , pages =. doi:10.2214/ajr.182.2.1820367 , language =

    work page doi:10.2214/ajr.182.2.1820367 2004

  32. [40]

    Prenatal Diagnosis , author =

    Fetal. Prenatal Diagnosis , author =. 2020 , pages =. doi:10.1002/pd.5579 , abstract =

    work page doi:10.1002/pd.5579 2020

  33. [41]

    Scientific Reports , author =

    T2* quantification using multi-echo gradient echo sequences: a comparative study of different readout gradients , volume =. Scientific Reports , author =. 2023 , pages =. doi:10.1038/s41598-023-28265-0 , abstract =

    work page doi:10.1038/s41598-023-28265-0 2023

  34. [42]

    Placenta , author =

    Exploring the role of a time-efficient. Placenta , author =. 2023 , pages =. doi:10.1016/j.placenta.2023.05.014 , language =

    work page doi:10.1016/j.placenta.2023.05.014 2023

  35. [43]

    Placenta , author =

    T2* weighted fetal. Placenta , author =. 2023 , pages =. doi:10.1016/j.placenta.2022.12.002 , language =

    work page doi:10.1016/j.placenta.2022.12.002 2023

  36. [44]

    Proceedings of the National Academy of Sciences , author =

    The. Proceedings of the National Academy of Sciences , author =. 1936 , pages =. doi:10.1073/pnas.22.4.210 , language =

    work page doi:10.1073/pnas.22.4.210 1936

  37. [46]

    2017 , pages =

    Ultrasound in Obstetrics & Gynecology , author =. 2017 , pages =. doi:10.1002/uog.17412 , language =

    work page doi:10.1002/uog.17412 2017

  38. [47]

    2014 , pages =

    BioMed Research International , author =. 2014 , pages =. doi:10.1155/2014/312142 , abstract =

    work page doi:10.1155/2014/312142 2014

  39. [48]

    American Journal of Roentgenology , author =

    Fetal anatomy revealed with fast. American Journal of Roentgenology , author =. 1996 , pages =. doi:10.2214/ajr.167.4.8819380 , language =

    work page doi:10.2214/ajr.167.4.8819380 1996

  40. [49]

    Clinical Nephrology , author =

    Treatment of renal hypertension , volume =. Clinical Nephrology , author =. 1975 , keywords =

    1975

  41. [50]

    NeuroImage , author =

    An efficient total variation algorithm for super-resolution in fetal brain. NeuroImage , author =. 2015 , pages =. doi:10.1016/j.neuroimage.2015.06.018 , language =

    work page doi:10.1016/j.neuroimage.2015.06.018 2015

  42. [52]

    Zalevskyi, Vladyslav and Sanchez, Thomas and Kaandorp, Misha and Roulet, Margaux and Fajardo-Rojas, Diego and Li, Liu and Hutter, Jana and Li, Hongwei Bran and Barkovich, Matthew and Ji, Hui and Wilhelmi, Luca and Dändliker, Aline and Steger, Céline and Koob, Mériam and Gomez, Yvan and Jakovčić, Anton and Klaić, Melita and Adžić, Ana and Marković, Pavel a...

    work page doi:10.48550/arxiv.2505.02784

  43. [53]

    Quantitative. AJNR. American journal of neuroradiology , author =. 2018 , keywords =. doi:10.3174/ajnr.A5668 , abstract =

    work page doi:10.3174/ajnr.a5668 2018

  44. [54]

    Magnetic Resonance in Medicine , author =

    Characterizing. Magnetic Resonance in Medicine , author =. 2024 , keywords =. doi:10.1002/mrm.30193 , abstract =

    work page doi:10.1002/mrm.30193 2024

  45. [55]

    Neuroradiology , author =

    The premature brain: imaging, anatomy and uncertain outcome , volume =. Neuroradiology , author =. 2013 , pages =. doi:10.1007/s00234-013-1244-8 , language =

    work page doi:10.1007/s00234-013-1244-8 2013

  46. [56]

    Journal of Neuropathology & Experimental Neurology , author =

    Fundamentals of the. Journal of Neuropathology & Experimental Neurology , author =. 2021 , pages =. doi:10.1093/jnen/nlab024 , abstract =

    work page doi:10.1093/jnen/nlab024 2021

  47. [57]

    Progress in Neurobiology , author =

    The impact of (ab)normal maternal environment on cortical development , volume =. Progress in Neurobiology , author =. 2021 , pages =. doi:10.1016/j.pneurobio.2021.102054 , language =

    arXiv 2021

  48. [58]

    Mapping the Human Brain from the Prenatal Period to Infancy Using 3D Magnetic Resonance Imaging

    Cachia, Arnaud and Mangin, Jean-François and Dubois, Jessica , year =. Mapping the. doi:10.48550/ARXIV.2504.21406 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2504.21406

  49. [59]

    doi:10.48550/ARXIV.2511.14897 , abstract =

    Indrakanti, Pranav and Simpson, Ivor , year =. doi:10.48550/ARXIV.2511.14897 , abstract =

    work page doi:10.48550/arxiv.2511.14897

  50. [60]

    doi:10.48550/ARXIV.2511.13232 , abstract =

    Abed, Malek Al and Demir, Sebiha and Groteklaes, Anne and Germani, Elodie and Faghihroohi, Shahrooz and Sabir, Hemmen and Albarqouni, Shadi , year =. doi:10.48550/ARXIV.2511.13232 , abstract =

    doi:10.48550/arxiv.2511.13232

  51. [61]

    Cerebral

    Vanderweyen, Davy and Seifeldine, Kassem and Philibert, Maude and Wilk, Christine and Ouellet, Annie and Whittingstall, Kevin , month = oct, year =. Cerebral. doi:10.21203/rs.3.rs-7715526/v1 , abstract =

    work page doi:10.21203/rs.3.rs-7715526/v1

  52. [62]

    fastWDM3D: Fast and Accurate 3D Healthy Tissue Inpainting

    Durrer, Alicia and Bieder, Florentin and Friedrich, Paul and Menze, Bjoern and Cattin, Philippe C. and Kofler, Florian , year =. doi:10.48550/ARXIV.2507.13146 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2507.13146

  53. [63]

    WDM: 3D Wavelet Diffusion Models for High-Resolution Medical Image Synthesis

    Friedrich, Paul and Wolleb, Julia and Bieder, Florentin and Durrer, Alicia and Cattin, Philippe C. , year =. doi:10.48550/ARXIV.2402.19043 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2402.19043

  54. [64]

    and Puonti, Oula and Sorby-Adams, Annabel and Kimberly, W

    Aguila, Ana Lawry and Zemlyanker, Dina and Cheng, You and Das, Sudeshna and Alexander, Daniel C. and Puonti, Oula and Sorby-Adams, Annabel and Kimberly, W. Taylor and Iglesias, Juan Eugenio , year =. Deep generative priors for. doi:10.48550/ARXIV.2510.15119 , abstract =

    work page doi:10.48550/arxiv.2510.15119

  55. [65]

    Developmental Neuroscience , author =

    Ex vivo. Developmental Neuroscience , author =. 2024 , pages =. doi:10.1159/000542276 , abstract =

    work page doi:10.1159/000542276 2024

  56. [66]

    Multimodal imaging of human fetal brain development at the mesoscopic scale using 11.7

    Arcamone, Lucas and Poupon, Cyril and Adle-Biassette, Homa and Khung-Savatovsky, Suonavy and Alison, Marianne and Dubois, Jessica and Hertz-Pannier, Lucie and Leprince, Yann , month = sep, year =. Multimodal imaging of human fetal brain development at the mesoscopic scale using 11.7. doi:10.1101/2025.09.08.669657 , abstract =

    work page doi:10.1101/2025.09.08.669657 2025

  57. [67]

    NeuroImage , author =

    Conditional deep generative normative modeling for structural and developmental anomaly detection in the fetal brain , volume =. NeuroImage , author =. 2025 , pages =. doi:10.1016/j.neuroimage.2025.121442 , language =

    arXiv 2025

  58. [69]

    , year =

    Saragadam, Vishwanath and LeJeune, Daniel and Tan, Jasper and Balakrishnan, Guha and Veeraraghavan, Ashok and Baraniuk, Richard G. , year =. doi:10.48550/ARXIV.2301.05187 , abstract =

    Pith/arXiv arXiv doi:10.48550/arxiv.2301.05187

  59. [70]

    Essakine, Amer and Cheng, Yanqi and Cheng, Chun-Wun and Zhang, Lipei and Deng, Zhongying and Zhu, Lei and Schönlieb, Carola-Bibiane and Aviles-Rivero, Angelica I , year =. Where. doi:10.48550/ARXIV.2411.03688 , abstract =

    Pith/arXiv arXiv doi:10.48550/arxiv.2411.03688

  60. [71]

    doi:10.48550/ARXIV.2010.07492 , abstract =

    Zhang, Kai and Riegler, Gernot and Snavely, Noah and Koltun, Vladlen , year =. doi:10.48550/ARXIV.2010.07492 , abstract =

    Pith/arXiv arXiv doi:10.48550/arxiv.2010.07492 2010

  61. [72]

    2023 , pages =

    IEEE Transactions on Medical Imaging , author =. 2023 , pages =. doi:10.1109/TMI.2023.3236216 , abstract =

    arXiv 2023

  62. [73]

    IREM: High-Resolution Magnetic Resonance (MR) Image Reconstruction via Implicit Neural Representation

    Wu, Qing and Li, Yuwei and Xu, Lan and Feng, Ruiming and Wei, Hongjiang and Yang, Qing and Yu, Boliang and Liu, Xiaozhao and Yu, Jingyi and Zhang, Yuyao , year =. doi:10.48550/ARXIV.2106.15097 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2106.15097

  63. [74]

    Sitzmann, Vincent and Martel, Julien N. P. and Bergman, Alexander W. and Lindell, David B. and Wetzstein, Gordon , month = jun, year =. Implicit. doi:10.48550/arXiv.2006.09661 , abstract =

    Pith/arXiv arXiv doi:10.48550/arxiv.2006.09661 2006

  64. [75]

    Physics-Informed Joint Multi-TE Super-Resolution with Implicit Neural Representation for Robust Fetal T2 Mapping

    Bulut, Busra and Dannecker, Maik and Sanchez, Thomas and Silva, Sara Neves and Zalevskyi, Vladyslav and Jia, Steven and Ledoux, Jean-Baptiste and Auzias, Guillaume and Rousseau, François and Hutter, Jana and Rueckert, Daniel and Cuadra, Meritxell Bach , year =. Physics-. doi:10.48550/ARXIV.2508.10680 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2508.10680

  65. [76]

    Frontiers in Neuroscience , author =

    Harmonized. Frontiers in Neuroscience , author =. 2021 , pages =. doi:10.3389/fnins.2021.662005 , abstract =

    arXiv 2021

  66. [77]

    , volume =

    T2 relaxation values in the developing preterm brain. , volume =. AJNR. American journal of neuroradiology , author =. 2003 , pages =

    2003

  67. [78]

    2020 , pages =

    T2. 2020 , pages =. doi:10.1007/978-3-030-59713-9_12 , author =

    work page doi:10.1007/978-3-030-59713-9_12 2020

  68. [79]

    2021 , pages =

    Quantitative. 2021 , pages =. doi:10.1007/978-3-030-87615-9_2 , author =

    work page doi:10.1007/978-3-030-87615-9_2 2021

  69. [80]

    2022 , pages =

    A. 2022 , pages =. doi:10.1007/978-3-031-17117-8_2 , author =

    work page doi:10.1007/978-3-031-17117-8_2 2022

  70. [81]

    2023 , pages =

    Quantitative. 2023 , pages =. doi:10.1007/978-3-031-45544-5_4 , author =

    work page doi:10.1007/978-3-031-45544-5_4 2023

  71. [82]

    2024 , pages =

    Improving. 2024 , pages =. doi:10.1007/978-3-031-72378-0\_41 , author =

    work page doi:10.1007/978-3-031-72378-0 2024

  72. [83]

    Journal of Advanced Research , author =

    Fetal. Journal of Advanced Research , author =. 2014 , pages =. doi:10.1016/j.jare.2013.06.001 , number =

    work page doi:10.1016/j.jare.2013.06.001 2014

  73. [84]

    Radiology , author =

    Normal maturation of the neonatal and infant brain:. Radiology , author =. 1988 , pages =. doi:10.1148/radiology.166.1.3336675 , number =

    work page doi:10.1148/radiology.166.1.3336675 1988

  74. [85]

    Topics in Magnetic Resonance Imaging , author =

    Quantitative. Topics in Magnetic Resonance Imaging , author =. 2010 , pages =. doi:10.1097/RMR.0b013e31821e56d8 , number =

    work page doi:10.1097/rmr.0b013e31821e56d8 2010

  75. [86]

    and Sutin, Jason and Stout, Jeffrey N

    Vasung, Lana and Turk, Esra Abaci and Ferradal, Silvina L. and Sutin, Jason and Stout, Jeffrey N. and Ahtam, Banu and Lin, Pei-Yi and Grant, P. Ellen , month = feb, year =. Exploring early human brain development with structural and physiological neuroimaging , volume =. doi:10.1016/j.neuroimage.2018.07.041 , journal =

    work page doi:10.1016/j.neuroimage.2018.07.041 2018

  76. [87]

    Journal of Perinatal Medicine , author =

    Fetal magnetic resonance imaging and ultrasound , volume =. Journal of Perinatal Medicine , author =. doi:10.1515/jpm-2015-0226 , abstract =

    work page doi:10.1515/jpm-2015-0226 2015

  77. [88]

    Brain Structure and Function , author =

    Normative biometry of the fetal brain using magnetic resonance imaging , volume =. Brain Structure and Function , author =. 2017 , pages =. doi:10.1007/s00429-016-1342-6 , number =

    work page doi:10.1007/s00429-016-1342-6 2017

  78. [89]

    Delineation of early brain development from fetuses to infants with diffusion

    Ouyang, Minhui and Dubois, Jessica and Yu, Qinlin and Mukherjee, Pratik and Huang, Hao , month = jan, year =. Delineation of early brain development from fetuses to infants with diffusion. doi:10.1016/j.neuroimage.2018.04.017 , journal =

    work page doi:10.1016/j.neuroimage.2018.04.017 2018

  79. [90]

    The Lancet , author =

    Use of. The Lancet , author =. 2017 , pages =. doi:10.1016/S0140-6736(16)31723-8 , number =

    work page doi:10.1016/s0140-6736(16)31723-8 2017

  80. [91]

    and Dehaene-Lambertz, G

    Dubois, J. and Dehaene-Lambertz, G. and Kulikova, S. and Poupon, C. and Hüppi, P. S. and Hertz-Pannier, L. , month = sep, year =. The early development of brain white matter:. doi:10.1016/j.neuroscience.2013.12.044 , journal =

    work page doi:10.1016/j.neuroscience.2013.12.044 2013

Showing first 80 references.