pith. sign in

arxiv: 1907.04854 · v1 · pith:BYMPOXIOnew · submitted 2019-07-09 · ⚛️ physics.comp-ph

Towards Accurate Modeling of the Multidimensional Magnetic Particle Imaging Physics

Tobias Kluth , Patryk Szwargulski , Tobias Knopp This is my paper

Pith reviewed 2026-05-24 23:44 UTC · model grok-4.3

classification ⚛️ physics.comp-ph
keywords magnetic particle imagingNéel rotationsystem matriximage reconstructionnanoparticlesmagnetization dynamicstomographic imagingmodel calibration
0
0 comments X

The pith

A Néel rotation model with fitted parameters matches measured 2D MPI data more closely than prior models and produces system matrices that reduce reconstruction artifacts.

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

Magnetic particle imaging solves a linear inverse problem that needs an accurate operator mapping particle positions to measured signals. Standard practice measures this operator in a slow calibration step because available physical models fail to capture the nanoparticles' magnetization response under realistic fields. The work replaces that measurement with a simulation based on Néel rotation dynamics for particle ensembles. Parameters of the model are adjusted until the simulated signals match experimental 2D MPI recordings at higher precision than existing models achieve. Phantom tests then confirm that reconstructions performed with the simulated operator contain fewer artifacts than those obtained from measured operators or simpler models.

Core claim

A physical model based on Néel rotation for large particle ensembles, using only a small number of fitted parameters, describes measured 2D MPI data with much higher precision than state-of-the-art MPI models. Phantom experiments demonstrate that the resulting simulated system matrix can be substituted for a measured matrix in image reconstruction and reduces artifacts caused by model mismatch considerably.

What carries the argument

The Néel rotation model for magnetization dynamics of nanoparticle ensembles under multidimensional excitation fields, with parameters fitted to data.

If this is right

  • Calibration measurements can be replaced by simulation once parameters are known for a given particle type and scanner.
  • Reduced model mismatch improves the conditioning of the linear inverse problem solved during reconstruction.
  • The same fitted model can be reused across multiple imaging sessions without repeating the calibration scan.
  • Higher-fidelity operators become feasible for 2D and potentially 3D excitation patterns without added measurement noise.

Where Pith is reading between the lines

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

  • If the fitted parameters prove stable across similar particle batches, the approach could support on-the-fly modeling for new scanner geometries.
  • The model may be tested by checking whether its predictions hold for excitation sequences outside the 2D training set.
  • Parameter values extracted during fitting could be compared against independent physical measurements of the same particles to check consistency.

Load-bearing premise

The Néel rotation model with a small number of fitted parameters captures the dominant magnetization dynamics of the specific nanoparticle ensemble and the 2D excitation fields used in the experiments.

What would settle it

Reconstruction of a new phantom using the simulated matrix produces artifact levels no lower than those obtained with a standard analytic model or an unmatched measured matrix.

Figures

Figures reproduced from arXiv: 1907.04854 by Patryk Szwargulski, Tobias Kluth, Tobias Knopp.

Figure 1
Figure 1. Figure 1: Imperfections of the equilibrium model compared to measured data. Absolute of two selected [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Selected frequency components of modeled system matrices using the N´eel model with anisotropy [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Selected frequency components of modeled system matrices using the N´eel model B3 with [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Comparison of measured and modeled system matrices for the equilibrium model A and the N´eel [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Error of the modeled system matrices (A and B3) shown in Fig. 4. The error is shown in two [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Reconstruction results of measured phantom data using three different system matrices: The [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
read the original abstract

The image reconstruction problem of the tomographic imaging technique magnetic particle imaging (MPI) requires the solution of a linear inverse problem. One prerequisite for this task is that the imaging operator that describes the mapping between the tomographic image and the measured signal is accurately known. For 2D and 3D excitation patterns, it is common to measure the system matrix in a calibration procedure, that is both, very time consuming and adds noise to the operator. The need for measuring the system matrix is due to the lack of an accurate physical model that is capable of describing the nanoparticles' magnetization behavior. Within this work we introduce a physical model that is based on N\'{e}el rotation for large particle ensembles and we find model parameters that describe measured 2D MPI data with much higher precision than state of the art MPI models. With phantom experiments we show that the simulated system matrix can be used for image reconstruction and reduces artifacts due to model-mismatch considerably.

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.

Referee Report

2 major / 2 minor

Summary. The paper introduces a physical model based on Néel rotation for large particle ensembles to describe the magnetization behavior in multidimensional MPI. Model parameters are identified that describe measured 2D MPI data with much higher precision than state-of-the-art models; phantom experiments then show that the resulting simulated system matrix can be used for reconstruction and reduces artifacts due to model mismatch considerably.

Significance. If the model captures the underlying physics rather than merely fitting the calibration data, replacing measured system matrices with simulated ones would be significant for MPI: it would eliminate lengthy calibration scans, reduce noise in the imaging operator, and improve reconstruction fidelity across a range of field settings.

major comments (2)
  1. [§4] §4 (parameter identification procedure): the Néel-model parameters are optimized directly to the same 2D measured signals later used to claim higher precision; because the simulated matrix is partly defined by the benchmark data, the reported improvement does not yet demonstrate that the model reproduces the magnetization dynamics for reasons other than its fitting flexibility.
  2. [§5] §5 (phantom reconstruction experiments): the artifact reduction is demonstrated on phantoms acquired with the same particle batch and 2D excitation fields used for fitting; without an independent test (held-out field amplitudes, different particle concentration, or 3D excitation) the claim that the simulated matrix reduces model-mismatch artifacts in general remains unproven and is load-bearing for the central contribution.
minor comments (2)
  1. The abstract states 'much higher precision' without supplying quantitative metrics (e.g., normalized RMS error or correlation coefficients); adding these numbers in the results section would make the comparison concrete.
  2. [Notation] Notation for the effective anisotropy field and the ensemble averaging in the Néel model should be defined explicitly once, with a reference to the governing equation, to aid reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the positive assessment of our work's potential significance and for the constructive major comments. We provide point-by-point responses below and indicate where revisions will be made to the manuscript.

read point-by-point responses

  1. Referee: §4 (parameter identification procedure): the Néel-model parameters are optimized directly to the same 2D measured signals later used to claim higher precision; because the simulated matrix is partly defined by the benchmark data, the reported improvement does not yet demonstrate that the model reproduces the magnetization dynamics for reasons other than its fitting flexibility.

    Authors: We agree that the parameters were fitted to the measured 2D signals used for the precision comparison. However, the Néel rotation model is derived from physical principles governing the particle magnetization, with parameters corresponding to physical quantities such as magnetic anisotropy and relaxation times. This physical grounding constrains the model and distinguishes it from arbitrary fitting functions. The superior match compared to the Langevin-based model, which is also commonly fitted, indicates that the improvement stems from better physical modeling. To strengthen the presentation, we will add a paragraph in §4 discussing the number of free parameters and their physical meaning, as well as a note on the validation approach. revision: partial

  2. Referee: §5 (phantom reconstruction experiments): the artifact reduction is demonstrated on phantoms acquired with the same particle batch and 2D excitation fields used for fitting; without an independent test (held-out field amplitudes, different particle concentration, or 3D excitation) the claim that the simulated matrix reduces model-mismatch artifacts in general remains unproven and is load-bearing for the central contribution.

    Authors: The phantom experiments were performed using the same particle batch and excitation fields as the fitting data. We note, however, that the fitting procedure optimizes the model to match the system matrix signals, which are typically measured for a grid of positions, while the phantom data consists of signals from a distributed object under the same fields but representing a different inverse problem. This serves as a test of the model's utility in reconstruction. We acknowledge the value of additional independent tests such as varying field amplitudes or using 3D excitations. As new experimental data cannot be acquired for this revision, we will revise the discussion in §5 to explicitly state the scope of the current validation and suggest directions for future generalization tests. revision: partial

Circularity Check

1 steps flagged

Parameters fitted to measured 2D data; simulated matrix performance on phantoms reduces to fit by construction

specific steps
  1. fitted input called prediction [Abstract]
    "we find model parameters that describe measured 2D MPI data with much higher precision than state of the art MPI models. With phantom experiments we show that the simulated system matrix can be used for image reconstruction and reduces artifacts due to model-mismatch considerably."

    Model parameters are optimized directly to the measured 2D data; the simulated matrix is therefore defined in part by that fit. Claiming higher precision on the data and reduced artifacts on phantom experiments (when those experiments use or overlap the calibration data) is then forced by the fitting procedure rather than an independent test of the Néel model.

full rationale

The paper determines model parameters by fitting to measured 2D MPI data, then presents the resulting simulated system matrix as achieving higher precision and reduced reconstruction artifacts on phantom experiments. This matches the fitted_input_called_prediction pattern: the 'prediction' (simulated matrix) is constructed from the same data used for fitting, so improved match and artifact reduction on that data (or closely related phantoms) is expected by construction rather than independent physical validation. No held-out prediction or external benchmark is described in the provided text that would falsify an overfitting account. The central claim therefore reduces to the fitting step.

Axiom & Free-Parameter Ledger

1 free parameters · 0 axioms · 0 invented entities

The central claim rests on a physical model whose parameters are fitted to experimental data; no independent derivation or external benchmark is described in the abstract.

free parameters (1)
  • Néel model parameters
    A small set of parameters is adjusted to match measured 2D MPI signals; these values are not derived from first principles.

pith-pipeline@v0.9.0 · 5696 in / 1077 out tokens · 16368 ms · 2026-05-24T23:44:43.020363+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

32 extracted references · 32 canonical work pages · 1 internal anchor

  1. [1]

    R. J. Deissler, Y. Wu, and M. A. Martens. Dependence of Brownian and N´ eel relaxation times on magnetic field strength. Medical Physics, 41(1):012301, 1–12, 2014

    work page 2014

  2. [2]

    Enpuku, S

    K. Enpuku, S. Bai, A. Hirokawa, K. Tanabe, T. Sasayama, and T. Yoshida. The effect of neel relaxation on the properties of the third harmonic signal of magnetic nanoparticles for use in narrow-band magnetic nanoparticle imaging. Japanese Journal of Applied Physics , 53(10):103002, 2014

    work page 2014

  3. [3]

    Gleich and J

    B. Gleich and J. Weizenecker. Tomographic imaging using the nonlinear response of magnetic particles. Nature, 435(7046):1214–1217, 2005. 9

    work page 2005

  4. [4]

    P. W. Goodwill and S. M. Conolly. The x-space formulation of the magnetic particle imaging process: 1-D signal, resolution, bandwidth, SNR, SAR, and magnetostimulation. IEEE Transactions on Medical Imaging, 29(11):1851–1859, 2010

    work page 2010

  5. [5]

    Graeser, K

    M. Graeser, K. Bente, A. Neumann, and T. M. Buzug. Trajectory dependent particle response for anisotropic mono domain particles in magnetic particle imaging. Journal of Physics D: Applied Physics , 49(4):045007, 2015

    work page 2015

  6. [6]

    T. Kluth. Mathematical models for magnetic particle imaging. Inverse Problems, 34(8):083001, 2018

    work page 2018

  7. [7]

    Numerical Reconstruction in Magnetic Particle Imaging

    T. Kluth and B. Jin. Numerical reconstruction in magnetic particle imaging. arXiv preprint arXiv:1902.01199, 2019

    work page internal anchor Pith review Pith/arXiv arXiv 1902

  8. [8]

    Kluth and P

    T. Kluth and P. Maass. Model uncertainty in magnetic particle imaging: Nonlinear problem formulation and model-based sparse reconstruction. International Journal on Magnetic Particle Imaging , 3(2):ID 1707004, 10 pages, 2017

    work page 2017

  9. [9]

    Knopp, S

    T. Knopp, S. Biederer, T. Sattel, J. Weizenecker, B. Gleich, J. Borgert, and T. M. Buzug. Trajectory analysis for magnetic particle imaging. Physics in Medicine and Biology , 54(2):385–397, 2009

    work page 2009

  10. [10]

    Knopp, S

    T. Knopp, S. Biederer, T. F. Sattel, J. Rahmer, J. Weizenecker, B. Gleich, J. Borgert, and T. M. Buzug. 2D model-based reconstruction for magnetic particle imaging. Medical Physics, 37(2):485–491, 2010

    work page 2010

  11. [11]

    Knopp, N

    T. Knopp, N. Gdaniec, and M. M¨ oddel. Magnetic particle imaging: From proof of principle to preclinical applications. Physics in Medicine & Biology , 62(14):R124, 2017

    work page 2017

  12. [12]

    Knopp and M

    T. Knopp and M. Hofmann. Online reconstruction of 3D magnetic particle imaging data. Physics in medicine and biology , 61(11):N257, 2016

    work page 2016

  13. [13]

    Knopp, T

    T. Knopp, T. F. Sattel, S. Biederer, J. Rahmer, J. Weizenecker, B. Gleich, J. Borgert, and T. M. Buzug. Model-based reconstruction for magnetic particle imaging. IEEE Transactions on Medical Imaging, 29(1):12–18, 2010

    work page 2010

  14. [14]

    Ludwig, D

    F. Ludwig, D. Eberbeck, N. L¨ owa, U. Steinhoff, T. Wawrzik, M. Schilling, and L. Trahms. Character- ization of magnetic nanoparticle systems with respect to their magnetic particle imaging performance. Biomedizinische Technik/Biomedical Engineering, 58(6):535–545, 2013

    work page 2013

  15. [15]

    Martens, R

    M. Martens, R. Deissler, Y. Wu, L. Bauer, Z. Yao, R. Brown, and M. Griswold. Modeling the Brownian relaxation of nanoparticle ferrofluids: Comparison with experiment. Medical Physics , 40(2):022303, 2013

    work page 2013

  16. [16]

    M¨ oddel, C

    M. M¨ oddel, C. Meins, J. Dieckhoff, and T. Knopp. Viscosity quantification using multi-contrast magnetic particle imaging. New Journal of Physics , 20(8):083001, aug 2018

    work page 2018

  17. [17]

    Rahmer, J

    J. Rahmer, J. Weizenecker, B. Gleich, and J. Borgert. Signal encoding in magnetic particle imaging. BMC Med Imaging , 9:4, 2009

    work page 2009

  18. [18]

    Rahmer, J

    J. Rahmer, J. Weizenecker, B. Gleich, and J. Borgert. Analysis of a 3-D system function measured for magnetic particle imaging. IEEE Transactions on Medical Imaging , 31(6):1289–1299, 2012

    work page 2012

  19. [19]

    D. B. Reeves and J. B. Weaver. Approaches for modeling magnetic nanoparticle dynamics. Critical Review in Biomedical Engineering , 42(1):85–93, 2014

    work page 2014

  20. [20]

    Rogge, M

    H. Rogge, M. Erbe, T. M. Buzug, and K. L¨ udtke-Buzug. Simulation of the magnetization dynamics of diluted ferrofluids in medical applications. Biomedizinische Technik/Biomedical Engineering, 58(6):601– 609, 2013

    work page 2013

  21. [21]

    S. A. Shah, D. B. Reeves, R. M. Ferguson, J. B. Weaver, and K. M. Krishnan. Mixed brownian alignment and n´ eel rotations in superparamagnetic iron oxide nanoparticle suspensions driven by an ac field.Phys. Rev. B, 92:094438, Sep 2015. 10

    work page 2015

  22. [22]

    Szwargulski, M

    P. Szwargulski, M. M¨ oddel, N. Gdaniec, and T. Knopp. Efficient joint image reconstruction of multi- patch data reusing a single system matrix in magnetic particle imaging. IEEE transactions on medical imaging, 2018

    work page 2018

  23. [23]

    Vaalma, J

    S. Vaalma, J. Rahmer, N. Panagiotopoulos, R. L. Duschka, J. Borgert, J. Barkhausen, F. M. Vogt, and J. Haegele. Magnetic particle imaging (mpi): Experimental quantification of vascular stenosis using stationary stenosis phantoms. PloS one , 12(1):e0168902, 2017

    work page 2017

  24. [24]

    Weizenecker

    J. Weizenecker. The Fokker-Planck equation for coupled Brown-N´ eel-rotation.Physics in Medicine and Biology, 63(3):035004, 2018

    work page 2018

  25. [25]

    Weizenecker, J

    J. Weizenecker, J. Borgert, and B. Gleich. A simulation study on the resolution and sensitivity of magnetic particle imaging. Physics in Medicine and Biology , 52:6363–6374, 2007

    work page 2007

  26. [26]

    Weizenecker, B

    J. Weizenecker, B. Gleich, J. Rahmer, and J. Borgert. Particle dynamics of mono-domain particles in magnetic particle imaging. In Magnetic Nanoparticles, pages 3–15. World Scientific, 2010

    work page 2010

  27. [27]

    Weizenecker, B

    J. Weizenecker, B. Gleich, J. Rahmer, and J. Borgert. Micro-magnetic simulation study on the magnetic particle imaging performance of anisotropic mono-domain particles. Physics in Medicine and Biology , 57(22):7317, 2012

    work page 2012

  28. [28]

    Yoshida and K

    T. Yoshida and K. Enpuku. Nonlinear behavior of magnetic fluid in brownian relaxation: Numerical simulation and derivation of empirical model. In T. M. Buzug and J. Borgert, editors, Magnetic Particle Imaging, pages 9–13, Berlin, Heidelberg, 2012. Springer Berlin Heidelberg

    work page 2012

  29. [29]

    Yoshida, K

    T. Yoshida, K. Enpuku, J. Dieckhoff, M. Schilling, and F. Ludwig. Magnetic fluid dynamics in a rotating magnetic field. Journal of Applied Physics , 111(5):053901, 2012

    work page 2012

  30. [30]

    Yoshida, Y

    T. Yoshida, Y. Matsugi, N. Tsujimura, T. Sasayama, K. Enpuku, T. Viereck, M. Schilling, and F. Lud- wig. Effect of alignment of easy axes on dynamic magnetization of immobilized magnetic nanoparticles. Journal of Magnetism and Magnetic Materials , 427:162 – 167, 2017

    work page 2017

  31. [31]

    Yoshida, N

    T. Yoshida, N. Othman, and K. Enpuku. Characterization of magnetically fractionated magnetic nanoparticles for magnetic particle imaging. Journal of Applied Physics , 114(17):173908, 2013

    work page 2013

  32. [32]

    E. Y. Yu, P. Chandrasekharan, R. Berzon, Z. W. Tay, X. Y. Zhou, A. P. Khandhar, R. M. Ferguson, S. J. Kemp, B. Zheng, P. W. Goodwill, et al. Magnetic particle imaging for highly sensitive, quantitative, and safe in vivo gut bleed detection in a murine model. ACS nano, 11(12):12067–12076, 2017. 11

    work page 2017