Pith. sign in

REVIEW 3 major objections 5 minor 27 references

Generative MR Multitasking with complex-harmonic cardiac encoding: Bridging the gap between gated imaging and real-time imaging

T0 review · 3 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read A single free-breathing MRI scan can now produce both gated cine and real-time beat-resolved images, plus sharper T1/T2 maps.

desk verdict Novel CVAE-based cardiac representation with an honest proof-of-concept, but the T1/T2 improvement claim rests on precision only, not accuracy. read the letter →

arxiv 2511.17847 v3 pith:7R7RHEOJ submitted 2025-11-22 eess.IV

Xinguo Fang , Anthony G. Christodoulou This is my paper
classification eess.IV
keywords CardiacMRIreal-timeimaginggatedMRMultitaskingimplicitneuralrepresentationconditionalvariationalautoencoderT1/T2mappingcomplexharmonicencoding
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

The paper establishes that the longstanding split between gated cardiac MRI (averaging many beats into one phase-resolved movie) and real-time MRI (resolving individual beats but with lower quality) can be bridged by representing cardiac motion as a complex harmonic whose phase tracks timing and whose amplitude tracks beat-to-beat variability. The method, Generative Multitasking, uses a conditional variational autoencoder (CVAE) trained per scan to learn an interpretable latent space of cardiac and respiratory motion and to generate temporal basis functions at arbitrary motion/contrast coordinates. With this, the same free-breathing, non-ECG-gated acquisition can be reconstructed as either a cardiac-phase-resolved 'gated-like' cine (an archetypal beat) or a time-resolved 'real-time-like' series retaining beat-to-beat differences. In healthy-volunteer experiments, this improved T1/T2 map SNR (lower intraseptal coefficient of variation by roughly twofold) and selectively removed eddy-current artifacts tied to the golden-angle sampling trajectory.

What carries the argument

The central object is the complex-harmonic cardiac latent coordinate: after band-pass filtering the cardiac latent mean and applying a Hilbert transform, each time point gets a complex value whose angle is the cardiac phase and whose magnitude is a time-varying latent amplitude capturing beat-to-beat differences. This coordinate, together with a low-pass-filtered respiratory coordinate, forms the structured latent space of a conditional variational autoencoder (CVAE) whose encoder learns motion identification (replacing binning) and whose decoder acts as an implicit neural representation of the temporal subspace (replacing the low-rank tensor). Decoder queries at arbitrary coordinates genera

What would settle it

In a patient or animal model with a ground-truth rhythm (e.g., ECG-triggered pressure or volume measurements), check whether the latent amplitude's time course correlates with beat-to-beat stroke volume or contraction strength. If the amplitude is roughly constant across beats of very different strength, or jumps discontinuously with small phase changes, the central bridging claim fails.

Watch

Extended reading notes

Core claim

The central claim is that cardiac motion, normally binned into discrete phases under the assumption that all beats are functionally equivalent, can be represented instead as a continuous complex harmonic coordinate—phase encodes timing within the cycle, while a learned latent amplitude encodes functional variation from beat to beat. A CVAE decoder, trained on each scan, maps any point in this interpretable latent space plus pulse-sequence timings to a temporal basis vector; the encoder simultaneously replaces conventional self-gating/binning. Because the decoder is continuous, it can be queried at the original time points (real-time-like) or along an averaged constant-amplitude circle (arche

Load-bearing premise

The learned cardiac latent amplitude is assumed to encode real beat-to-beat functional changes rather than noise or reconstruction artifacts, and the latent space is assumed smooth enough that interpolating between arbitrary coordinates produces physically valid motion states.

Editorial extensions

If this is right

  • The same free-breathing, non-ECG-gated acquisition can be reconstructed offline either as a conventional phase-resolved cine or as a beat-resolved real-time series, moving the gated/real-time choice from scan prescription to reconstruction.
  • Conditioning the decoder on the previous k-space readout angle and replacing that term with the self-gating line constant removes eddy-current-induced high-frequency fluctuations synchronized with the golden-angle schedule, while preserving other high-frequency motion and contrast content.
  • T1 and T2 mapping from dual-flip-angle data achieved roughly twofold lower intraseptal coefficient of variation (0.13 vs 0.31 for T1; 0.12 vs 0.32 for T2; p<0.001) without spatial regularization, indicating a better-conditioned temporal basis.
  • The continuous latent space permits interpolation of intermediate motion states, reducing temporal discontinuities compared with discrete cardiac-phase binning (20 phases, 6 respiratory bins).

Reading between the lines

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

  • If the latent amplitude faithfully tracks true beat-to-beat variability in arrhythmias, this representation could enable beat-type-specific cines (e.g., separating dominant sinus beats from ectopic beats) from a single acquisition—a direct clinical test the paper proposes but did not perform.
  • The same complex-harmonic encoding could be inserted into other cardiac-resolved reconstruction frameworks to provide a continuous, variable-amplitude cardiac coordinate without necessarily using a CVAE.
  • A falsifiable test: compare the latent amplitude trajectory against ECG or invasive pressure waveforms across a range of rhythms; if the amplitude mostly reflects noise or model error rather than stroke-volume or functional changes, the bridging claim would need revision.
  • The eddy-current correction mechanism suggests a broader principle: neural implicit temporal models can encode system-specific trajectory memory, potentially applying to other trajectory-dependent artifacts beyond cardiac imaging.
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.

Referee Report

3 major / 5 minor

Summary. The paper proposes 'Generative Multitasking,' a scan-specific CVAE-based reconstruction framework for free-breathing, non-ECG-gated cardiac MRI. Cardiac motion is represented as a complex harmonic in latent space, with phase encoding timing and a learned time-varying amplitude intended to capture beat-to-beat variability. The decoder acts as an implicit neural temporal basis, allowing the same acquisition to be rendered either as a cardiac-phase-resolved 'gated-like' series or as a time-resolved 'real-time-like' series. The method is applied to three protocols: steady-state cine, T2prep/inversion-recovery multicontrast imaging, and dual-flip-angle T1/T2 mapping. The central quantitative claim is that T1/T2 mapping is improved relative to conventional Multitasking, based on reduced intraseptal coefficient of variation (T1 CoV 0.13 vs. 0.31; T2 CoV 0.12 vs. 0.32; p<0.001). Qualitative results also suggest smoother motion representation and removal of golden-angle-synchronized eddy-current artifacts.

Significance. If the claims hold, the framework is a conceptually valuable step toward unifying gated and real-time CMR, with a potentially elegant treatment of beat-to-beat variability through a continuous, interpretable latent cardiac coordinate. The eddy-current correction by conditioning on the previous k-space angle and replacing it at inference is a clever, mechanism-aware approach. However, the paper's principal quantitative evidence for improved T1/T2 mapping rests on a single internal-consistency metric (intraseptal CoV) with no reference standard or accuracy validation. The method is trained and evaluated on the same data, so the reported SNR gain could partly reflect bias introduced by the learned temporal basis rather than genuine noise reduction. The conceptual novelty is real, but the current evidence does not yet support the strong conclusion of improved quantitative imaging.

major comments (3)
  1. [§2.5 (Quantitative mapping evaluation) and Fig. 7] The only quantitative support for the central claim that generative Multitasking 'improves T1 and T2 mapping' is the intraseptal coefficient of variation, CoV=σ/μ, interpreted as inverse SNR. Lower CoV can result from reduced noise or from bias introduced by the reconstruction. Here, since Eq. (9) is deliberately unregularized for both methods, any smoothing must come from the learned generative temporal basis Φec. The paper provides no phantom with known T1/T2, no comparison to a validated mapping sequence (e.g., MOLLI), and no repeat-scan reproducibility. The ~2× CoV reduction could therefore be a bias-variance tradeoff rather than a true SNR improvement. An accuracy or reference-standard evaluation is needed to substantiate the 'improves T1/T2 mapping' claim.
  2. [§2.3.5 (Training loss) and §2.4.1 (Image calculation)] The CVAE is trained and evaluated on the same scan data: the encoder/decoder learn a temporal basis that minimizes reconstruction error on the very self-gating data that are later used for reconstruction, and CoV reduction is measured on the same acquisition. Because CoV is an internal consistency metric, a denoiser that smooths the temporal basis would trivially reduce it. The paper does not address this circularity concern, e.g., through cross-validation, a phantom experiment, or comparison to a quantitative reference standard. Without such evidence, the reported SNR gain is not an independent validation of the method.
  3. [§2.3.3 (Complex harmonic cardiac representation) and Fig. 3] The latent cardiac amplitude is claimed to encode beat-to-beat functional variability, but this is validated only qualitatively in a single healthy volunteer for cine. No quantitative metric ties the latent amplitude to cardiac function, and no comparison is made in arrhythmia patients where beat variability is clinically relevant. The bridging claim between 'gated-like' and 'real-time-like' imaging depends on the latent amplitude reflecting true physiological variability rather than noise or model error. As presented, this load-bearing interpretability claim is not yet supported by the experimental evidence.
minor comments (5)
  1. [General] There are several typographical and formatting issues: 'flexiblility' in the Conclusions, 'ANOV A' in §2.5, and inconsistent spacing in 'CV AE' throughout. These should be corrected.
  2. [§2.3.1, Eq. (4)] Equation (4) shows four rows (μresp, log σ²resp, μcard, log σ²card) but the text calls this a 'two-dimensional latent space.' Clarify that the latent variables are two-dimensional and the distribution parameters are what appear in Eq. (4).
  3. [§2.5] The statistical reporting of the ANOVA gives p-values but no effect sizes or confidence intervals. Reporting these would strengthen the quantitative comparison.
  4. [Fig. 4 and Video S1] The spectral analysis is qualitative. A quantitative measure of the suppression of the golden-angle harmonics, e.g., peak-to-background ratio at the expected frequencies, would make the eddy-current correction claim more rigorous.
  5. [References] Reference [14] is described as a previous CVAE study, but the text could benefit from a more explicit statement of how the current architecture differs from that prior work, beyond the complex-harmonic space.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the CVAE is trained as an autoencoder and the claimed flexibility follows from the architecture; the CoV-based evaluation lacks an accuracy anchor but is not a circular reduction.

full rationale

I walked the derivation chain. The core generative model is Eq. (3)/(6): decoder D maps interpretable latent motion coordinates and known pulse-sequence timings to a temporal factor; the CVAE is trained scan-specifically with the self-reconstruction loss Eq. (7). This is an unsupervised representation/reconstruction method, not a first-principles prediction of an external quantity. The gated-like and real-time-like views are both queries of the same trained decoder (Section 2.4.1), so their co-existence is by architectural design, not a circular derivation. The eddy-current correction is a counterfactual evaluation: replacing c_t's previous k-space angle with k_SG at inference removes that input channel; this is an ablation, not a hidden reuse of the target. The quantitative claim rests on intraseptal CoV (Section 2.5), which is a self-consistency precision metric without a reference standard; the paper explicitly says 'this study should be viewed as a proof-of-concept' and notes limited validation. That is an external-validity/correctness limitation, not a definitional reduction: CoV is not an Eq. (7) term, and lower CoV is not guaranteed by the loss. No load-bearing uniqueness theorem or self-citation chain is used; refs to the authors' own Multitasking work are baseline methodology. Finding: no significant circularity, score 0.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

The central claim depends on several hand-set hyperparameters (β, filter cutoffs, architecture) and domain assumptions about cardiac motion and artifact structure. No independently validated entities or external benchmarks are provided; the invented complex-harmonic latent has no falsifiable handle outside the paper.

free parameters (4)
  • KL regularization weight β = not reported
    Empirically selected by inspecting latent trajectories for one scan and fixed for all experiments (Methods 2.5.1).
  • Respiratory low-pass cutoff = 0.67 Hz
    Hand-set frequency threshold for disentangling respiratory motion (Methods 2.5.1).
  • Cardiac band-pass range = 0.67–2.5 Hz
    Hand-set frequency band for cardiac motion (Methods 2.5.1).
  • Network layer widths = 25-50-100-70 / 70-100-50-25
    Chosen architecture; not systematically optimized; affects capacity and denoising behavior (Methods 2.5.1).
assumptions (4)
  • domain assumption Cardiac motion in healthy subjects is well represented by a one-dimensional complex harmonic with slowly varying amplitude after band-pass filtering.
    Core to the latent representation; the authors acknowledge in the Discussion that arrhythmias may require relaxing this constraint.
  • domain assumption The initial SVD-based temporal features (Eq. 8) contain sufficient information for training a faithful decoder.
    The CVAE is trained on these features; if the subspace is corrupted by artifacts, the network may reproduce them.
  • ad hoc to paper Eddy-current artifacts from k-space jumps are determined by the previous k-space angle and can be removed by replacing this condition with a constant at inference.
    Specific to the proposed correction; supported by spectral alignment with the golden-angle sampling schedule but no independent validation.
  • domain assumption The learned latent space is smooth and interpolatable so that arbitrary latent coordinates correspond to valid motion states.
    Used to generate archetypal cycles and multidimensional images (Section 2.4.1).
invented entities (1)
  • Complex-harmonic cardiac latent coordinate (real/imaginary components) with time-varying amplitude
    purpose: Encodes cardiac phase (angle) and beat-to-beat functional variability (magnitude) in a 2D latent dimension
    No external falsifiable prediction is made; it is an internal modeling construct validated only on the same data used for training.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Generative MR Multitasking with complex-harmonic cardiac encoding: Bridging the gap between gated imaging and real-time imaging." pith.science (2026). https://pith.science/paper/7R7RHEOJ

@misc{pith2026251117847,
  author       = {Pith},
  title        = {Pith review of: Generative MR Multitasking with complex-harmonic cardiac encoding: Bridging the gap between gated imaging and real-time imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7R7RHEOJ}},
  note         = {Machine review of arXiv:2511.17847}
}
read the original abstract

Purpose: To develop a unified image reconstruction framework that bridges real-time and gated cardiac MRI, including quantitative MRI. Methods: We introduce Generative Multitasking, which learns an implicit neural temporal basis from sequence timings and an interpretable latent space for cardiac and respiratory motion. Cardiac motion is modeled as a complex harmonic, with phase encoding timing and a latent amplitude capturing beat-to-beat functional variability, linking cardiac phase-resolved ("gated-like") and time-resolved ("real-time-like") views. We implemented the framework using a conditional variational autoencoder (CVAE) and evaluated it for free-breathing, non-ECG-gated radial GRE in three settings: steady-state cine imaging, multicontrast T2prep/IR imaging, and dual-flip-angle T1/T2 mapping, compared with conventional Multitasking. Results: Generative Multitasking provided flexible cardiac motion representation, enabling reconstruction of archetypal cardiac phase-resolved cines (like gating) as well as time-resolved series that reveal beat-to-beat variability (like real-time imaging). Conditioning on the previous k-space angle and modifying this term at inference removed eddy-current artifacts without globally smoothing high temporal frequencies. For quantitative mapping, Generative Multitasking reduced intraseptal T1 and T2 coefficients of variation compared with conventional Multitasking (T1: 0.13 vs. 0.31; T2: 0.12 vs. 0.32; p<0.001), indicating higher SNR. Conclusion: Generative Multitasking uses a CVAE with complex harmonic cardiac coordinates to unify gated and real-time CMR within a single free-breathing, non-ECG-gated acquisition. It allows flexible cardiac motion representation, suppresses trajectory-dependent artifacts, and improves T1 and T2 mapping, suggesting a path toward cine, multicontrast, and quantitative imaging without separate gated and real-time scans.

Figures

Figures reproduced from arXiv: 2511.17847 by the authors.

Figure 2
Figure 2. Diagram of the conditional variational autoencoder (CVAE) and conceptual illustration of the complex harmonic cardiac representation inside the latent space. (A) Time-resolved samples of an input distribution 𝜑 and conditional inputs 𝑐 containing pulse sequence timings are encoded into a latent space 𝑧. The latent space is enforced to be interpretable as motion states: a filter bank disentangles respiratory and card… view at source ↗
Figure 4
Figure 4. Time resolved images and their energy spectra. (A) Images from the initial SVD-based temporal subspace (input 𝚽) are noisy and show high frequency fluctuations (red arrow). These fluctuations match the energy spectrum of the golden-angle sampling schedule exp(±𝑖𝑛∠𝑘[𝑡]), n=1,2 (orange). (B) A reconstruction using the CVAE output without eddy-current correction (output 𝚽̂ ), improves SNR but retains high-frequency flu… view at source ↗
Figure 6
Figure 6. T1 and T2 maps of Multitasking and generative Multitasking from one healthy subject. Maps from generative Multitasking demonstrates visibly higher SNR comparing to conventional Multitasking while preserving details at the blood-myocardium interface [PITH_FULL_IMAGE:figures/full_fig_p017_6.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

27 extracted references · 2 linked inside Pith

  1. [1]

    real- time

    Introduction: Cardiovascular magnetic resonance imaging (CMR) is an essential diagnostic and management tool. However, compared to MRI of other body parts, CMR faces specific challenges from cardiac and respiratory motion. Broadly, there are two approaches to handling cardiac motion: cardiac-phase-resolved imaging (i.e., gated or segmented imaging) and ti...

  2. [2]

    time dimensions

    Methods: 2.1. Conventional Multitasking Image Model Multitasking5 reconstructs a high-dimensional image 𝐴(𝐱, 𝛕) a function of spatial location 𝐱 and of N time-varying independent variables (“time dimensions”) collected as 𝛕 = [𝜏1 𝜏2 ⋯ 𝜏𝑁]𝑇 = [𝛕ℳ 𝑇 𝛕𝒞 𝑇]𝑇. Here 𝛕ℳ comprises motion-related time dimensions whose chronological timings 𝛕ℳ[𝑡] are unknown a prio...

  3. [3]

    This is typically performed by Bloch-equation aware clustering

    Identify the motion timings 𝛕ℳ[𝑡] from time-resolved self-gating data, in order to gate/bin the cardiac and respiratory data into the multidimensional space. This is typically performed by Bloch-equation aware clustering

  4. [4]

    Estimate the temporal factor 𝛗(𝛕) from those same self-gating data, typically by low- rank tensor completion of the binned self-gating data, followed by a high-order SVD

  5. [5]

    real-time-like

    Estimate the spatial factor 𝐮(𝐱), typically by regularized least-squares fitting of the estimated 𝛗(𝛕) to the full raw data. 2.2. Generative Multitasking Image Model Generative Multitasking (Fig. 1B) replaces the low-rank tensor model with an implicit neural representation6-8 of 𝛗(𝛕): 𝛗(𝛕) = 𝐷(𝛕ℳ, 𝛕𝒞) (3) where 𝐷(⋅) is, e.g. the decoder of a scan-specific...

  6. [6]

    real-time

    Results For cine imaging, Generative Multitasking replaces discrete cardiac phase bins with a continuous variable-amplitude cardiac trajectory, revealing beat-to-beat functional variability that gated reconstructions miss (Fig. 3). Discrete cardiac phase binning produces a piecewise constant signal over the cardiac cycle and assumes that each heartbeat is...

  7. [7]

    cardiac phase-resolved

    Discussion This work introduces a generative Multitasking framework that bridges conventional cardiac phase-gated imaging with real-time CMR by representing the cardiac cycle as a complex harmonic with both phase and a variable amplitude. Whereas cardiac phase continues to encode 18 timing, the latent amplitude dimension provides an avenue to capture beat...

  8. [8]

    Conclusions In conclusion, we presented a generative Multitasking framework that bridges conventional cardiac phase-gated imaging and real-time CMR. A complex harmonic cardiac representation with both phase and latent amplitude enables a single model to capture timing as well as beat-to- beat functional variability, effectively combining cine-like depicti...

Show all 27 references
  1. [9]

    & Laub, G

    Finn, J.P., Nael, K., Deshpande, V ., Ratib, O. & Laub, G. Cardiac MR imaging: state of the technology. Radiology 241, 338-354 (2006)

    2006

  2. [10]

    Imaging of arrhythmia: Real-time cardiac magnetic resonance imaging in atrial fibrillation

    Laubrock, K., et al. Imaging of arrhythmia: Real-time cardiac magnetic resonance imaging in atrial fibrillation. Eur J Radiol Open 9, 100404 (2022)

    2022

  3. [11]

    The future of cardiovascular magnetic resonance: All-in-one vs

    Christodoulou, A.G., et al. The future of cardiovascular magnetic resonance: All-in-one vs. real-time (Part 1). Journal of Cardiovascular Magnetic Resonance 26, 100997 (2024)

    2024

  4. [12]

    & Peters, D.C

    Contijoch, F., Rasche, V ., Seiberlich, N. & Peters, D.C. The future of CMR: All-in-one vs. real-time CMR (Part 2). J Cardiovasc Magn Reson 26, 100998 (2024)

    2024

  5. [13]

    Magnetic resonance multitasking for motion-resolved quantitative cardiovascular imaging

    Christodoulou, A.G., et al. Magnetic resonance multitasking for motion-resolved quantitative cardiovascular imaging. Nat Biomed Eng 2, 215-226 (2018)

    2018

  6. [14]

    & Setsompop, K

    Abraham, D., Nishimura, M., Cao, X., Liao, C. & Setsompop, K. Implicit Representation of GRAPPA Kernels for Fast MRI Reconstruction. arXiv preprint arXiv:2310.10823 (2023)

    2023 arXiv

  7. [15]

    PISCO: Self-supervised k-space regularization for improved neural implicit k-space representations of dynamic MRI

    Spieker, V ., et al. PISCO: Self-supervised k-space regularization for improved neural implicit k-space representations of dynamic MRI. arXiv preprint arXiv:2501.09403 (2025)

    2025 arXiv

  8. [16]

    Unsupervised reconstruction of accelerated cardiac cine MRI using neural fields

    Catalán, T., et al. Unsupervised reconstruction of accelerated cardiac cine MRI using neural fields. Computers in Biology and Medicine 185, 109467 (2025)

    2025

  9. [17]

    & Yan, X

    Sohn, K., Lee, H. & Yan, X. Learning structured output representation using deep conditional generative models. Advances in neural information processing systems 28(2015)

    2015

  10. [18]

    Self-gated MRI of multiple beat morphologies in the presence of arrhythmias

    Contijoch, F., et al. Self-gated MRI of multiple beat morphologies in the presence of arrhythmias. Magn Reson Med 78, 678-688 (2017)

    2017

  11. [19]

    Using 5D flow MRI to decode the effects of rhythm on left atrial 3D flow dynamics in patients with atrial fibrillation

    Ma, L., et al. Using 5D flow MRI to decode the effects of rhythm on left atrial 3D flow dynamics in patients with atrial fibrillation. Magn Reson Med 85, 3125-3139 (2021)

    2021

  12. [20]

    Dynamic Regularized Adaptive Cluster Optimization (DRACO) for Quantitative Cardiac Cine MRI in Complex Arrhythmias

    Ming, Z., et al. Dynamic Regularized Adaptive Cluster Optimization (DRACO) for Quantitative Cardiac Cine MRI in Complex Arrhythmias. J Magn Reson Imaging 61, 248-262 (2025)

    2025

  13. [21]

    ECG-free cine MRI with data-driven clustering of cardiac motion for quantification of ventricular function

    Ming, Z., et al. ECG-free cine MRI with data-driven clustering of cardiac motion for quantification of ventricular function. NMR Biomed 37, e5091 (2024)

    2024

  14. [22]

    & Jacob, M

    Zou, Q., Priya, S., Nagpal, P. & Jacob, M. Joint Cardiac T(1) Mapping and Cardiac Cine Using Manifold Modeling. Bioengineering (Basel) 10(2023)

    2023

  15. [23]

    XD-GRASP: Golden-angle radial MRI with reconstruction of extra motion-state dimensions using compressed sensing

    Feng, L., et al. XD-GRASP: Golden-angle radial MRI with reconstruction of extra motion-state dimensions using compressed sensing. Magn Reson Med 75, 775-788 (2016)

    2016

  16. [24]

    Free-running cardiac magnetic resonance fingerprinting: Joint T1/T2 map and Cine imaging

    Jaubert, O., et al. Free-running cardiac magnetic resonance fingerprinting: Joint T1/T2 map and Cine imaging. Magn Reson Imaging 68, 173-182 (2020)

    2020

  17. [25]

    & Hamilton, J.I

    Rashid, I., Cruz, G., Seiberlich, N. & Hamilton, J.I. Cardiac MR Fingerprinting: Overview, Technical Developments, and Applications. J Magn Reson Imaging 60, 1753- 1773 (2024)

    2024

  18. [26]

    Free-breathing myocardial T(1) mapping using inversion-recovery radial FLASH and motion-resolved model-based reconstruction

    Wang, X., et al. Free-breathing myocardial T(1) mapping using inversion-recovery radial FLASH and motion-resolved model-based reconstruction. Magn Reson Med 89, 1368- 1384 (2023). 22

    2023

  19. [27]

    & Yerly, J

    Di Sopra, L., Piccini, D., Coppo, S., Stuber, M. & Yerly, J. An automated approach to fully self-gated free-running cardiac and respiratory motion-resolved 5D whole-heart MRI. Magn Reson Med 82, 2118-2132 (2019)

    2019

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.