arxiv: 2604.22713 · v1 · submitted 2026-04-24 · ⚛️ physics.med-ph

Recognition: unknown

OpenMRF: A Modular, Vendor-Neutral Open-Source Framework for Reproducible Magnetic Resonance Fingerprinting using Pulseq

Ivaylo Angelov, Jannik Stebani, Jesse I Hamilton, Jon-Fredrik Nielsen, Martin Blaimer, Maximilian Gram, Maxim Zaitsev, Nicole Seiberlich, Peter Martin, Peter Nordbeck, Petra Albert, Qi Liu, Qingping Chen, Sydney Kaplan, Tobias Wech, Tom Griesler, Xiang Wang, Zhibo Zhu

Pith reviewed 2026-05-08 08:54 UTC · model grok-4.3

classification ⚛️ physics.med-ph

keywords magnetic resonance fingerprintingopen-source frameworkPulseqreproducible MRIquantitative imagingmulti-site validationlow-rank reconstructionT1 T2 mapping

0 comments

The pith

OpenMRF supplies a modular Pulseq-based framework for consistent magnetic resonance fingerprinting across MRI vendors and field strengths.

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

The paper presents OpenMRF as an open-source end-to-end system that combines Pulseq sequence design, Bloch dictionary simulation directly from .seq files, and low-rank subspace reconstruction. It seeks to replace fragmented tools with a single platform that supports reproducible sequence implementation and parameter mapping. Validation covers digital phantom simulations with sub-2% deviations, a multi-site ISMRM/NIST phantom study on Siemens, GE, and United Imaging scanners at 0.55 T, 1.5 T, and 3 T showing mean deviations near 1-8%, and in vivo scans in liver, myocardium, and brain that yield usable maps. The work argues this setup removes vendor lock-in and enables direct method comparison and multi-site harmonization in quantitative MRI.

Core claim

OpenMRF integrates modular Pulseq-based sequence design, Bloch-simulation-based dictionary creation directly from .seq files, and iterative low-rank subspace reconstruction to enable consistent, reproducible, and transferable MRF across vendors, sites, and field strengths, as demonstrated by accurate simulations and multi-site phantom and in vivo results.

What carries the argument

The OpenMRF framework, which uses Pulseq for sequence implementation, performs physics-accurate Bloch simulations from .seq files for dictionary generation, and applies iterative low-rank subspace reconstruction for parameter mapping.

If this is right

Digital phantom simulations produce T1 deviations of 0.03+/-0.32% and T2 deviations of 0.12+/-1.94%.
Multi-site phantom measurements remain consistent with references at all tested field strengths and vendors, with mean deviations of -0.1+/-2.9% for T1, -1.5+/-8.7% for T2, and -4.0+/-7.2% for T1rho.
In vivo liver, myocardium, and brain acquisitions generate high-quality parameter maps on platforms spanning 0.55 T to 3 T.
The unified platform supports direct method development, comparison, and multi-site validation for quantitative MRI.

Where Pith is reading between the lines

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

Groups running multi-center studies could adopt OpenMRF to reduce protocol variability across sites.
The modular structure may support quick extension to new MRF variants or related quantitative techniques such as diffusion mapping.
Wider testing on additional scanner models would further test the claim of full vendor neutrality.

Load-bearing premise

The modular Pulseq-based design and simulations will yield accurate and transferable results on real scanners without needing substantial vendor-specific adjustments or calibrations.

What would settle it

Repeating the multi-site phantom protocol on an untested MRI vendor or field strength and finding T1 or T2 deviations exceeding 10% from reference values would undermine the transferability claim.

Figures

Figures reproduced from arXiv: 2604.22713 by Ivaylo Angelov, Jannik Stebani, Jesse I Hamilton, Jon-Fredrik Nielsen, Martin Blaimer, Maximilian Gram, Maxim Zaitsev, Nicole Seiberlich, Peter Martin, Peter Nordbeck, Petra Albert, Qi Liu, Qingping Chen, Sydney Kaplan, Tobias Wech, Tom Griesler, Xiang Wang, Zhibo Zhu.

**Figure 1.** Figure 1: Schematic illustration of the OpenMRF framework for developing, simulating, and view at source ↗

**Figure 2.** Figure 2: Sequence parameters and Bloch-simulated signal evolutions for the IR-FISP MRF view at source ↗

**Figure 3.** Figure 3: Preparation modules and Bloch-simulated signal evolutions for the view at source ↗

**Figure 4.** Figure 4: Digital phantom-based comparison of SVD and low-rank reconstruction for the IR view at source ↗

**Figure 5.** Figure 5: Multi-field-strength validation of the IR-FISP MRF sequence in the ISMRM/NIST view at source ↗

**Figure 6.** Figure 6: Multi-field-strength validation of the T1-T2-T1ρ cardiac MRF (cMRF) sequence in the ISMRM/NIST system phantom, acquired using Siemens Free.Max (0.55 T), Sola (1.5 T), and Vida (3 T) scanners. All datasets were reconstructed using the low-rank subspace method. The dictionary was generated with T1 values ranging from 10 ms to 4 s in 5% steps, and T2 and T1ρ values ranging from 1 ms to 3 s in 5% steps. Only p… view at source ↗

**Figure 7.** Figure 7: Multi-vendor comparison of quantitative parameter maps obtained with IR-FISP view at source ↗

**Figure 8.** Figure 8: Multi-site comparison of T1, T2, and T1ρ estimates obtained with three MRF protocols (IR-FISP MRF, T1-T2 cMRF, and T1-T2-T1ρ cMRF) at 1.5 T in the ISMRM/NIST system phantom. Data were acquired at three sites (Würzburg, Freiburg, and Michigan) using Siemens Avanto, Aera, and Sola scanners. A–C) Estimated relaxation times plotted against the gold-standard reference values for T1, T2, and T1ρ. D–F) Percenta… view at source ↗

**Figure 9.** Figure 9: Multi-site comparison of T1, T2, and T1ρ estimates obtained with three MRF protocols (IR-FISP MRF, T1-T2 cMRF, and T1-T2-T1ρ cMRF) at 3 T in the ISMRM/NIST system phantom. Data were acquired at four sites (Würzburg, Freiburg, Michigan, Houston) using Siemens Prisma, Cima.X and Vida, as well as GE Signa and United Imaging uMR790 systems. A–C) Estimated relaxation times plotted against the gold-standard ref… view at source ↗

**Figure 10.** Figure 10: In vivo demonstration of the OpenMRF framework across multiple field strengths. A) Liver imaging at 0.55 T using the T1-T2 cMRF sequence with a fixed segment duration of 1 s. B) Cardiac short-axis imaging at 1.5 T using the T1-T2-T1ρ cMRF sequence with prospective R-wave triggering. The trigger timings were automatically parsed to the dictionary simulation, and the cardiac acquisition window was adapted u… view at source ↗

read the original abstract

Purpose: Widespread adoption and methodological advancement of Magnetic Resonance Fingerprinting (MRF) are limited by the lack of unified, reproducible implementation frameworks and fragmented open-source tools. To address these barriers, we introduce OpenMRF - a comprehensive Pulseq-based solution - designed to enable consistent, reproducible, and transferable MRF research across vendors, sites, and field strengths. Methods: OpenMRF integrates modular Pulseq-based sequence design, Bloch-simulation-based dictionary creation directly from .seq files, and iterative low-rank subspace reconstruction. The framework was evaluated through digital phantom simulations, a multi-site ISMRM/NIST phantom study on Siemens MRI systems at 0.55 T, 1.5 T, and 3 T as well as GE and United Imaging 3 T platforms, and representative in vivo acquisitions in the liver (0.55 T), myocardium (1.5 T), and brain (3 T). Results: Simulations demonstrated high mapping accuracy in an ISMRM/NIST-like digital phantom, with low-rank reconstruction yielding deviations of 0.03+/-0.32 % (T1) and 0.12+/-1.94 % (T2). The multi-site phantom study yielded relaxation times consistent with reference values at all field strengths, with mean deviations of -0.1+/-2.9 % (T1), -1.5+/-8.7 % (T2), and -4.0+/-7.2 % (T1rho). In vivo acquisitions produced high-quality parameter maps across platforms and field strengths. Conclusion: OpenMRF provides a robust, open-source, end-to-end Pulseq-based solution for MRF that enables reproducible sequence implementation, physics-accurate dictionary simulation, and advanced reconstruction across vendors and field strengths. By providing a unified platform for method development, comparison, and multi-site validation, OpenMRF aims to accelerate reproducible and harmonized quantitative MRI research within the community.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

OpenMRF ties Pulseq sequence design to direct Bloch simulation and low-rank recon in one open-source package, backed by multi-vendor phantom data that tracks reference values within a few percent.

read the letter

The main point is that this paper delivers an integrated open-source MRF framework built around Pulseq. It handles sequence design, generates dictionaries straight from the .seq files via Bloch simulation, and runs iterative low-rank subspace reconstruction, then tests the whole chain on digital phantoms, a multi-site ISMRM/NIST phantom study across Siemens, GE, and United Imaging systems at 0.55 T to 3 T, plus in-vivo scans in liver, myocardium, and brain. The reported deviations stay small: 0.03 % T1 in simulation and roughly -0.1 % T1 and -1.5 % T2 in the real phantom work. That cross-vendor consistency is the practical advance over the scattered tools that existed before. The modular structure also makes it easier for others to swap in new sequences or reconstruction steps without starting from scratch. The multi-site results give the strongest evidence that the pipeline can transfer without massive per-vendor fixes. The soft spot is the simulation fidelity claim. Standard Bloch solvers from .seq files usually skip slice-profile effects, gradient delays, and B1 inhomogeneity unless those are added by hand. The paper shows good end-point parameter accuracy, but it does not appear to include direct comparisons of simulated versus measured signal time courses on each scanner. Without those checks it is possible that the low-rank recon or fitting steps are compensating for any missing physics rather than the dictionary being fully accurate on its own. That issue is moderate, not central, because the final maps still line up with references. This work is aimed at quantitative MRI groups that need reproducible MRF implementations for method development or multi-center studies. Anyone trying to standardize relaxation mapping across sites will find the framework and the validation data directly useful. The paper shows clear engagement with the practical barriers in the field and supplies concrete cross-platform numbers, so it deserves a serious referee. I would send it out for peer review.

Referee Report

1 major / 3 minor

Summary. The paper introduces OpenMRF, a modular, vendor-neutral open-source framework for Magnetic Resonance Fingerprinting (MRF) using Pulseq. It integrates sequence design, Bloch-simulation-based dictionary creation directly from .seq files, and iterative low-rank subspace reconstruction. Validation consists of digital phantom simulations (deviations of 0.03±0.32% T1 and 0.12±1.94% T2), a multi-site ISMRM/NIST phantom study across Siemens (0.55 T, 1.5 T, 3 T), GE, and United Imaging 3 T systems (mean deviations -0.1±2.9% T1, -1.5±8.7% T2, -4.0±7.2% T1ρ), and in vivo acquisitions in liver (0.55 T), myocardium (1.5 T), and brain (3 T), with claims of high mapping accuracy and robustness.

Significance. If the results hold, this work offers a meaningful contribution to quantitative MRI by supplying a unified, reproducible platform that addresses tool fragmentation in MRF. The Pulseq-based modular design and open-source release could facilitate method development, cross-vendor comparisons, and multi-site harmonization, particularly valuable for advancing standardized quantitative imaging protocols.

major comments (1)

[§2.2] §2.2: The Bloch simulation pipeline from Pulseq .seq files is presented as enabling physics-accurate dictionary generation. Standard Bloch solvers typically omit scanner-specific effects including gradient delays, eddy currents, slice-profile imperfections, and B1 inhomogeneity unless explicitly modeled. Table 3 reports low mean deviations in the multi-site phantom study, but the manuscript provides no direct comparison of simulated versus measured signal evolutions on each platform or vendor, leaving open the possibility that reported consistency arises from post-acquisition scaling rather than intrinsic simulation fidelity.

minor comments (3)

[Abstract] Abstract: The purpose and conclusion paragraphs are repetitive; the conclusion could emphasize implications for community adoption and harmonization rather than restating framework features.
[Methods] Methods/Results: Additional details on error propagation, statistical tests, and any data exclusion criteria for the phantom and in vivo evaluations would strengthen assessment of the reported deviation metrics, even with code availability.
General: Figure captions for parameter maps should specify color scales, units, and any masking applied to improve interpretability across field strengths.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive comment and positive overall assessment of OpenMRF. We address the point raised in §2.2 below and will revise the manuscript accordingly.

read point-by-point responses

Referee: [§2.2] §2.2: The Bloch simulation pipeline from Pulseq .seq files is presented as enabling physics-accurate dictionary generation. Standard Bloch solvers typically omit scanner-specific effects including gradient delays, eddy currents, slice-profile imperfections, and B1 inhomogeneity unless explicitly modeled. Table 3 reports low mean deviations in the multi-site phantom study, but the manuscript provides no direct comparison of simulated versus measured signal evolutions on each platform or vendor, leaving open the possibility that reported consistency arises from post-acquisition scaling rather than intrinsic simulation fidelity.

Authors: We thank the referee for this observation. OpenMRF's dictionary generation applies a standard Bloch solver to the Pulseq .seq file without explicit modeling of scanner-specific effects (gradient delays, eddy currents, slice-profile imperfections, or B1 inhomogeneity). The multi-site phantom results show good agreement with reference values, but we agree that the absence of a direct simulated-versus-measured signal comparison leaves the intrinsic fidelity of the simulation less clearly demonstrated. In the revised manuscript we will add a direct comparison: we will extract representative signal evolutions from phantom acquisitions on one or more platforms and overlay them with the corresponding simulated signals generated by the OpenMRF Bloch pipeline. This addition will clarify the contribution of the simulation step to the observed mapping accuracy. revision: yes

Circularity Check

0 steps flagged

No significant circularity; framework claims rest on independent experimental validations

full rationale

The paper describes OpenMRF as an integrated framework for sequence design, Bloch simulation from .seq files, and low-rank reconstruction, with all accuracy and reproducibility claims supported by direct experimental results: digital phantom simulations reporting specific deviation percentages, multi-site phantom studies on Siemens/GE/United Imaging systems at multiple field strengths showing mean deviations against reference values, and in vivo acquisitions. No load-bearing steps in the presented methods or results reduce by construction to self-definitions, fitted inputs renamed as predictions, or self-citation chains; the derivation chain is self-contained because outcomes are benchmarked externally rather than internally redefined.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard MRI physics and the assumption that Pulseq sequences can be faithfully executed and simulated across vendors. No new free parameters or invented physical entities are introduced in the abstract.

axioms (2)

domain assumption Bloch equations accurately model MR signal evolution when driven by Pulseq-defined RF and gradient waveforms
Used for dictionary creation directly from .seq files
domain assumption Iterative low-rank subspace reconstruction recovers accurate T1, T2, and T1rho maps from MRF data
Central to the reconstruction step described in the methods summary

pith-pipeline@v0.9.0 · 5740 in / 1558 out tokens · 72291 ms · 2026-05-08T08:54:50.914096+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

65 extracted references

[1]

Gulani and N

V. Gulani and N. Seiberlich. Quantitative mri: rationale and challenges. In N. Seiberlich, V. Gulani, F. Calamante, A. Campbell-Washburn, M. Doneva, H. H. Hu, and S. Sourbron, editors,Quantitative Magnetic Resonance Imaging, volume 1, pages xxxvii–li. Academic Press, 2020

2020
[2]

Hockings, N

P. Hockings, N. Saeed, R. Simms, N. Smith, M. G. Hall, J. C. Waterton, and S. Sour- bron. Mri biomarkers. In N. Seiberlich, V. Gulani, F. Calamante, A. Campbell-Washburn, M. Doneva, H. H. Hu, and S. Sourbron, editors,Quantitative Magnetic Resonance Imaging, volume 1, pages liii–lxxxvi. Academic Press, 2020

2020
[3]

K. E. Keenan, M. Ainslie, A. J. Barker, M. A. Boss, K. M. Cecil, C. Charles, T. L. Chen- evert, L. Clarke, J. L. Evelhoch, P. Finn, D. Gembris, J. L. Gunter, D. L. G. Hill, Jr. Jack, C. R., E. F. Jackson, G. Liu, S. E. Russek, S. D. Sharma, M. Steckner, K. F. Stupic, J. D. Trzasko, C. Yuan, and J. Zheng. Quantitative magnetic resonance imaging phantoms: A...

2018
[4]

D. R. Messroghli, J. C. Moon, V. M. Ferreira, L. Grosse-Wortmann, T. He, P. Kellman, J. Mascherbauer, R. Nezafat, M. Salerno, E. B. Schelbert, A. J. Taylor, R. Thompson, M. Ugander, R. B. van Heeswijk, and M. G. Friedrich. Clinical recommendations for cardiovascular magnetic resonance mapping of t1, t2, t2* and extracellular volume: A consensus statement ...

2017
[5]

Shukla-Dave, N

A. Shukla-Dave, N. A. Obuchowski, T. L. Chenevert, S. Jambawalikar, L. H. Schwartz, D. Malyarenko, W. Huang, S. M. Noworolski, R. J. Young, M. S. Shiroishi, H. Kim, C. Coolens, H. Laue, C. Chung, M. Rosen, M. Boss, and E. F. Jackson. Quantitative imag- ing biomarkers alliance (qiba) recommendations for improved precision of dwi and dce-mri derived biomark...

2019
[6]

D. Ma, V. Gulani, N. Seiberlich, K. Liu, J. L. Sunshine, J. L. Duerk, and M. A. Griswold. Magnetic resonance fingerprinting.Nature, 495(7440):187–192, March 2013

2013
[7]

S. Gaur, A. Panda, J. E. Fajardo, J. Hamilton, Y. Jiang, and V. Gulani. Magnetic resonance fingerprinting: A review of clinical applications.Invest Radiol, 58(8):561–577, August 2023

2023
[8]

K. J. Layton, S. Kroboth, F. Jia, S. Littin, H. Yu, J. Leupold, J. F. Nielsen, T. Stöcker, and M. Zaitsev. Pulseq: A rapid and hardware-independent pulse sequence prototyping framework.Magn Reson Med, 77(4):1544–1552, April 2017

2017
[9]

K. S. Ravi, S. Geethanath, and J. T. Vaughan. Pypulseq: A python package for mri pulse sequence design.J Open Source Softw, 4(42):1725, 2019

2019
[10]

J. F. Nielsen and D. C. Noll. Toppe: A framework for rapid prototyping of mr pulse sequences.Magn Reson Med, 79(6):3128–3134, June 2018

2018
[11]

T. H. M. Roos, E. Versteeg, M. Gosselink, H. Hoogduin, K. M. Nam, N. Boulant, V. Gras, F. Mauconduit, D. W. J. Klomp, J. C. W. Siero, and J. P. Wijnen. ptx-pulseq in hybrid sequences: Accessible and advanced hybrid open-source mri sequences on philips scanners. Magn Reson Med, 94(5):1946–1962, November 2025. 23

1946
[12]

M.Gram, Z.Zhu, T.Griesler, S.Kaplan, P.M.Jakob, P.Nordbeck, N.Seiberlich, P.Martin, and Q. Liu. Initial tests of cardiac magnetic resonance fingerprinting at 5t on a whole-body mri scanner using pulseq. InProceedings of the ISMRM 2026 Annual Meeting, Capetown, May 2026. Abstract #02134

2026
[13]

A. S. Gaspar, N. A. Silva, A. N. Price, A. M. Ferreira, and R. G. Nunes. Open-source myocardial t1 mapping with simultaneous multi-slice acceleration: Combining an auto- calibrated blipped-bssfp readout with verse-mb pulses.Magn Reson Med, 90(2):539–551, August 2023

2023
[14]

Q. Liu, L. Ning, I. A. Shaik, C. Liao, B. Gagoski, B. Bilgic, W. Grissom, J. F. Nielsen, M.Zaitsev, andY.Rathi. Reducedcross-scannervariabilityusingvendor-agnosticsequences for single-shell diffusion mri.Magn Reson Med, 92(1):246–256, July 2024

2024
[15]

K. Herz, S. Mueller, O. Perlman, M. Zaitsev, L. Knutsson, P. Z. Sun, J. Zhou, P. van Zijl, K. Heinecke, P. Schuenke, C. T. Farrar, M. Schmidt, A. Dörfler, K. Scheffler, and M. Zaiss. Pulseq-cest: Towards multi-site multi-vendor compatibility and reproducibility of cest experiments using an open-source sequence standard.Magn Reson Med, 86(4):1845– 1858, Oc...

2021
[16]

D. Ma, S. Coppo, Y. Chen, D. F. McGivney, Y. Jiang, S. Pahwa, V. Gulani, and M. A. Griswold. Slice profile and b1 corrections in 2d magnetic resonance fingerprinting.Magn Reson Med, 78(5):1781–1789, November 2017

2017
[17]

Loktyushin, K

A. Loktyushin, K. Herz, N. Dang, F. Glang, A. Deshmane, S. Weinmüller, A. Doerfler, B. Schölkopf, K. Scheffler, and M. Zaiss. Mrzero - automated discovery of mri sequences using supervised learning.Magn Reson Med, 86(2):709–724, August 2021

2021
[18]

Stöcker, K

T. Stöcker, K. Vahedipour, D. Pflugfelder, and N. J. Shah. High-performance computing mri simulations.Magn Reson Med, 64(1):186–193, July 2010

2010
[19]

Castillo-Passi, R

C. Castillo-Passi, R. Coronado, G. Varela-Mattatall, C. Alberola-López, R. Botnar, and P. Irarrazaval. Komamri.jl: An open-source framework for general mri simulations with gpu acceleration.Magn Reson Med, 90(1):329–342, July 2023

2023
[20]

Schuenke, C

P. Schuenke, C. Redshaw Kranich, M. Lutz, J. Schattenfroh, M. Anders, P. Reisdorf, J. Schulz-Menger, I. Sack, J. Hamilton, N. Seiberlich, and C. Kolbitsch. Open-source cardiac magnetic resonance fingerprinting.MAGMA, 38(4):665–677, August 2025

2025
[21]

J. H. Lee, B. A. Hargreaves, B. S. Hu, and D. G. Nishimura. Fast 3d imaging using variable-density spiral trajectories with applications to limb perfusion.Magn Reson Med, 50(6):1276–1285, December 2003

2003
[22]

R. K. Robison, Z. Li, D. Wang, M. B. Ooi, and J. G. Pipe. Correction of b0 eddy current effects in spiral mri.Magn Reson Med, 81(4):2501–2513, April 2019

2019
[23]

Ong and M

F. Ong and M. Lustig. Sigpy: A python package for high performance iterative reconstruc- tion. ISMRM Workshop on Data Sampling & Image Reconstruction, 2019. Accessed from conference archive

2019
[24]

J. I. Hamilton, Y. Jiang, D. Ma, W. C. Lo, V. Gulani, M. Griswold, and N. Seiberlich. Investigating and reducing the effects of confounding factors for robust t1 and t2 mapping with cardiac mr fingerprinting.Magn Reson Imaging, 53:40–51, November 2018

2018
[25]

C. R. Wyatt, T. M. Barbara, and A. R. Guimaraes. T1ρmagnetic resonance fingerprinting. NMR Biomed, 33(5):e4284, May 2020. 24

2020
[26]

Sharafi, K

A. Sharafi, K. Medina, M. W. V. Zibetti, S. Rao, M. A. Cloos, R. Brown, and R. R. Regatte. Simultaneous t1, t2, and t1ρrelaxation mapping of the lower leg muscle with mr fingerprinting.Magn Reson Med, 86(1):372–381, July 2021

2021
[27]

Simultaneous t1, t2, and t1ρcardiac magnetic resonance fingerprinting for contrast agent-free myocardial tissue characterization.Magn Reson Med, 87(4):1992–2002, April 2022

C.Velasco, G.Cruz, B.Lavin, A.Hua, A.Fotaki, R.M.Botnar, andC.Prieto. Simultaneous t1, t2, and t1ρcardiac magnetic resonance fingerprinting for contrast agent-free myocardial tissue characterization.Magn Reson Med, 87(4):1992–2002, April 2022

1992
[28]

Kaplan, G

S. Kaplan, G. L. da Cruz, C. Madamanchi, V. L. Murthy, S. Swanson, J. Hamilton, and N. Seiberlich. Simultaneous t1, t2, and t1ρmapping of the myocardium using cardiac mr fingerprinting with a deep image prior reconstruction.Magn Reson Med, 94(4):1500–1513, October 2025

2025
[29]

Schuenke, C

P. Schuenke, C. Koehler, A. Korzowski, J. Windschuh, P. Bachert, M. E. Ladd, S. Mundiyanapurath, D. Paech, S. Bickelhaupt, D. Bonekamp, H. P. Schlemmer, A. Rad- bruch, and M. Zaiss. Adiabatically prepared spin-lock approach for t1ρ-based dynamic glucose enhanced mri at ultrahigh fields.Magn Reson Med, 78(1):215–225, July 2017

2017
[30]

2nd Witschey, W. R., A. Borthakur, M. A. Elliott, E. Mellon, S. Niyogi, D. J. Wallman, C. Wang, and R. Reddy. Artifacts in t1ρ-weighted imaging: compensation for b(1) and b(0) field imperfections.J Magn Reson, 186(1):75–85, May 2007

2007
[31]

M. Gram, M. Seethaler, D. Gensler, J. Oberberger, P. M. Jakob, and P. Nordbeck. Balanced spin-lock preparation for b1-insensitive and b0-insensitive quantification of the rotating frame relaxation time t1ρ.Magn Reson Med, 85(5):2771–2780, May 2021

2021
[32]

C. Wan, M. Gram, W. He, Z. Xu, Q. Chen, S. Littin, T. Lange, and M. Zaitsev. Quadruple- refocused spin-locking: A robust method for high-amplitude t1ρimaging.Magn Reson Med, 94(5):2023–2037, November 2025

2023
[33]

Mangia, T

S. Mangia, T. Liimatainen, M. Garwood, and S. Michaeli. Rotating frame relaxation during adiabatic pulses vs. conventional spin lock: simulations and experimental results at 4 t. Magn Reson Imaging, 27(8):1074–1087, October 2009

2009
[34]

M. Gram, M. Christa, F. T. Gutjahr, P. Albertova, T. Williams, P. M. Jakob, W. R. Bauer, and P. Nordbeck. Quantification of the rotating frame relaxation time t2ρ: Comparison of balanced spin-lock and continuous-wave malcolm-levitt preparations.NMR Biomed, 37(11):e5199, November 2024

2024
[35]

D. F. McGivney, E. Pierre, D. Ma, Y. Jiang, H. Saybasili, V. Gulani, and M. A. Griswold. Svd compression for magnetic resonance fingerprinting in the time domain.IEEE Trans Med Imaging, 33(12):2311–2322, December 2014

2014
[36]

J. A. Fessler and B. P. Sutton. Nonuniform fast fourier transforms using min-max interpo- lation.IEEE Trans Signal Process, 51(2):560–574, February 2003

2003
[37]

J. I. Hamilton, Y. Jiang, D. Ma, Y. Chen, W. C. Lo, M. Griswold, and N. Seiberlich. Simul- taneous multislice cardiac magnetic resonance fingerprinting using low rank reconstruction. NMR Biomed, 32(2):e4041, February 2019

2019
[38]

D. Kim, S. F. Cauley, K. S. Nayak, R. M. Leahy, and J. P. Haldar. Region-optimized virtual (rovir) coils: Localization and/or suppression of spatial regions using sensor-domain beamforming.Magn Reson Med, 86(1):197–212, July 2021

2021
[39]

Uecker, P

M. Uecker, P. Lai, M. J. Murphy, P. Virtue, M. Elad, J. M. Pauly, S. S. Vasanawala, and M. Lustig. Espirit–an eigenvalue approach to autocalibrating parallel mri: where sense meets grappa.Magn Reson Med, 71(3):990–1001, March 2014. 25

2014
[40]

Jiang, D

Y. Jiang, D. Ma, N. Seiberlich, V. Gulani, and M. A. Griswold. Mr fingerprinting using fast imaging with steady state precession (fisp) with spiral readout.Magn Reson Med, 74(6):1621–1631, December 2015

2015
[41]

K. F. Stupic, M. Ainslie, M. A. Boss, C. Charles, A. M. Dienstfrey, J. L. Evelhoch, P. Finn, Z. Gimbutas, J. L. Gunter, D. L. G. Hill, C. R. Jack, E. F. Jackson, T. Karaulanov, K. E. Keenan, G. Liu, M. N. Martin, P. V. Prasad, N. S. Rentz, C. Yuan, and S. E. Russek. A standard system phantom for magnetic resonance imaging.Magn Reson Med, 86(3):1194– 1211,...

2021
[42]

Simultaneous mapping of water shift and b1 (wasabi)-application to field-inhomogeneity correction of cest mri data.Magn Reson Med, 77(2):571–580, February 2017

P.Schuenke, J.Windschuh, V.Roeloffs, M.E.Ladd, P.Bachert, andM.Zaiss. Simultaneous mapping of water shift and b1 (wasabi)-application to field-inhomogeneity correction of cest mri data.Magn Reson Med, 77(2):571–580, February 2017

2017
[43]

Cardiacmrfoptimization at 3t using rosette trajectories and mt modeling in openmrf

S.Kaplan, E.Cummings, T.Griesler, M.Gram, andN.Seiberlich. Cardiacmrfoptimization at 3t using rosette trajectories and mt modeling in openmrf. InProceedings of the ISMRM 2026 Annual Meeting, Capetown, May 2026. Abstract #00881

2026
[44]

Griesler, E

T. Griesler, E. Cummings, S. Kaplan, M. Gram, J. Hamilton, M. Davenport, V. Gulani, H. Hussain, N. Seiberlich, and G. Cruz. Kidney t1/t2/t2*/t1ρ/pdff mapping with openmrf: Initial results and adiabatic vs continuous wave spin locking. InProceedings of the ISMRM 2026 Annual Meeting, Capetown, May 2026. Abstract #00661

2026
[45]

Hernando, P

D. Hernando, P. Kellman, J. P. Haldar, and Z. P. Liang. Robust water/fat separation in the presence of large field inhomogeneities using a graph cut algorithm.Magn Reson Med, 63(1):79–90, January 2010

2010
[46]

Y. Liu, J. Hamilton, B. Eck, M. Griswold, and N. Seiberlich. Myocardial t1 and t2 quan- tification and water-fat separation using cardiac mr fingerprinting with rosette trajectories at 3t and 1.5t.Magn Reson Med, 85(1):103–119, January 2021

2021
[47]

Scholten, D

H. Scholten, D. Lohr, T. Wech, and H. Köstler. Fast measurement of the gradient system transfer function at 7 t.Magn Reson Med, 89(4):1644–1659, April 2023

2023
[48]

Coletti, A

C. Coletti, A. Fotaki, J. Tourais, Y. Zhao, C. van de Steeg-Henzen, M. Akçakaya, Q. Tao, C. Prieto, and S. Weingärtner. Robust cardiac t1ρmapping at 3t using adiabatic spin-lock preparations.Magn Reson Med, 90(4):1363–1379, October 2023

2023
[49]

D. Si, M. G. Crabb, S. J. Littlewood, K. P. Kunze, J. Varghese, K. Binzel, M. Khan, O. P. Simonetti, C. Prieto, and R. M. Botnar. Initial experience of cardiac t1ρmapping at 0.55 t: Continuous wave versus adiabatic spin-lock preparation pulses.Magn Reson Med, 94(4):1644–1653, October 2025

2025
[50]

M. Weigel. Extended phase graphs: dephasing, rf pulses, and echoes - pure and simple.J Magn Reson Imaging, 41(2):266–295, February 2015

2015
[51]

Kellman and E

P. Kellman and E. R. McVeigh. Image reconstruction in snr units: a general method for snr measurement.Magn Reson Med, 54(6):1439–1447, December 2005

2005
[52]

Stebani, I

J. Stebani, I. Angelov, P. Dawood, T. Griesler, P. Albertova, T. Kampf, K. Rak, P. M. Jakob, M. Blaimer, and M. Gram. High-resolution quantitative imaging of the inner ear using 3d magnetic resonance fingerprinting. InProceedings of the ISMRM 2025 Annual Meeting, Honolulu, May 2025. Abstract #0926

2025
[53]

Jiang, D

Y. Jiang, D. Ma, K. E. Keenan, K. F. Stupic, V. Gulani, and M. A. Griswold. Repeatability of magnetic resonance fingerprinting t1 and t2 estimates assessed using the ismrm/nist mri system phantom.Magn Reson Med, 78(4):1452–1457, October 2017. 26

2017
[54]

W. C. Lo, L. K. Bittencourt, A. Panda, Y. Jiang, J. Tokuda, R. Seethamraju, C. Tempany- Afdhal, V. Obmann, K. Wright, M. Griswold, N. Seiberlich, and V. Gulani. Multicenter repeatability and reproducibility of mr fingerprinting in phantoms and in prostatic tissue. Magn Reson Med, 88(4):1818–1827, October 2022

2022
[55]

B. K. Statton, J. Smith, M. E. Finnegan, G. Koerzdoerfer, R. A. Quest, and M. Grech- Sollars. Temperature dependence, accuracy, and repeatability of t1 and t2 relaxation times for the ismrm/nist system phantom measured using mr fingerprinting.Magn Reson Med, 87(3):1446–1460, March 2022

2022
[56]

Notni, M

L. Notni, M. Gram, B. Fänger, D. Güllmar, M. Krämer, and J. R. Reichenbach. Quan- titative comparison of t1, t2, t1ρ, and adc for assessing age-related intervertebral disc degeneration.Z Med Phys, pages S0939–3889(25)00132–1, October 2025

2025
[57]

M. Gram, D. Gensler, P. Winter, M. Seethaler, P. A. Arias-Loza, J. Oberberger, P. M. Jakob, and P. Nordbeck. Fast myocardial t1ρmapping in mice using k-space weighted image contrast and a bloch simulation-optimized radial sampling pattern.MAGMA, 35(2):325– 340, April 2022

2022
[58]

M. Gram, D. Gensler, P. Albertova, F. T. Gutjahr, K. Lau, P. A. Arias-Loza, P. M. Jakob, and P. Nordbeck. Quantification correction for free-breathing myocardial t1ρmapping in mice using a recursively derived description of a t1ρ* relaxation pathway.J Cardiovasc Magn Reson, 24(1):30, May 2022

2022
[59]

M. Gram, P. Albertova, V. Schirmer, M. Blaimer, M. Gamer, M. J. Herrmann, P. Nordbeck, and P. M. Jakob. Towards robust in vivo quantification of oscillating biomagnetic fields using rotary excitation based mri.Sci Rep, 12(1):15375, September 2022

2022
[60]

Albertova, M

P. Albertova, M. Gram, M. Blaimer, W. R. Bauer, P. M. Jakob, and P. Nordbeck. Rotary excitation of non-sinusoidal pulsed magnetic fields: Towards non-invasive direct detection of cardiac conduction.Magn Reson Med, 92(5):1965–1979, November 2024

1965
[61]

Lustig, S

M. Lustig, S. J. Kim, and J. M. Pauly. A fast method for designing time-optimal gradient waveforms for arbitrary k-space trajectories.IEEE Trans Med Imaging, 27(6):866–873, June 2008

2008
[62]

Vaziri and M

S. Vaziri and M. Lustig. The fastest arbitrary k-space trajectories. InProceedings of the International Society for Magnetic Resonance in Medicine, volume 20, 2012

2012
[63]

X. Wang, Q. Chen, C. Liu, L. Wang, P. Wang, X. Zhou, Y. Pu, M. Gram, T. Griesler, Y. Ni, P. Cao, S. Littin, J. Cai, M. Zaitsev, and T. Li. Astrad: Acquisition-sequence and k-space trajectory co-design for accelerated mr fingerprinting. InProceedings of the ISMRM 2026 Annual Meeting, Capetown, May 2026. Abstract #01511

2026
[64]

J. I. Hamilton. A self-supervised deep learning reconstruction for shortening the breathhold andacquisitionwindowincardiacmagneticresonancefingerprinting.Front Cardiovasc Med, 9:928546, June 2022

2022
[65]

J. I. Hamilton, W. Truesdell, M. Galizia, N. Burris, P. Agarwal, and N. Seiberlich. A low-rank deep image prior reconstruction for free-breathing ungated spiral functional cmr at 0.55 t and 1.5 t.MAGMA, 36(3):451–464, July 2023. 27 Supplementary Material Supplementary Figure 1: Pulseq sequence diagram of the IR-FISP MRF sequence. The im- plemented method ...

2023