pith. sign in

arxiv: 2512.10657 · v3 · submitted 2025-12-11 · 📡 eess.SY · cs.SY

Estimating Hormone Concentrations in the Pituitary-Thyroid Feedback Loop from Irregularly Sampled Measurements

Seth Siriya , Tobias M. Wolff , Isabelle Krauss , Victor G. Lopez , Matthias A. M\"uller This is my paper

Pith reviewed 2026-05-16 23:25 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords hormone estimationpituitary-thyroid feedback loopmoving horizon estimationirregular samplingthyroid disordersstate estimationmodel-based control
0
0 comments X

The pith

Moving horizon estimation recovers unmeasurable hormone concentrations from irregularly sampled blood measurements in pituitary-thyroid models.

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

The paper establishes that a sample-based moving horizon estimator can accurately track both measured and internal hormone levels in models of the pituitary-thyroid feedback loop for patients with hypo- and hyperthyroidism. It first verifies a sample-based detectability property that accounts for irregular sampling times. The estimator is then tested on virtual patients across different sampling schemes, showing stable performance and reduced errors with more frequent sampling even under model uncertainty and dosage errors. This approach addresses the challenge of estimating internal states needed for precise medication dosing when only sporadic blood samples are available.

Core claim

Sample-based moving horizon estimation applied to two pituitary-thyroid loop models recovers the internal hormone concentrations from irregularly sampled measurable concentrations, with the estimator exhibiting robust stability across all tested scenarios and improved accuracy at higher sampling rates despite model uncertainty and misreported dosages.

What carries the argument

Sample-based moving horizon estimation incorporating a detectability condition tailored to irregular sampling of the measurable hormone levels.

If this is right

  • The estimator remains stable for both hypo- and hyperthyroid virtual patient models under irregular sampling.
  • More frequent sampling reduces estimation error in the presence of model uncertainty.
  • The method handles misreported dosages without loss of stability.
  • Internal states can be estimated sufficiently well to support model-based dosing recommendations.

Where Pith is reading between the lines

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

  • This estimation technique could support automated dosing systems that rely on estimated internal concentrations rather than only external measurements.
  • The same framework might extend to other biological feedback loops where sampling is irregular and some states are unmeasurable.
  • Validation on real patient data with actual irregular sampling patterns would provide further confirmation.
  • Optimizing sampling schedules to balance estimation accuracy and patient convenience becomes feasible.

Load-bearing premise

The virtual patient models accurately represent the dynamics of hypo- and hyperthyroidism.

What would settle it

If the estimator shows instability or large persistent errors when applied to data from real patients with irregular sampling, the claims of robust stability would be falsified.

Figures

Figures reproduced from arXiv: 2512.10657 by Isabelle Krauss, Matthias A. M\"uller, Seth Siriya, Tobias M. Wolff, Victor G. Lopez.

Figure 1
Figure 1. Figure 1: Simple diagram of PT loop from Wolff et al. (2025). [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: LHS and RHS of (3) for two sampled trajectory [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Measured hormone concentrations of simulated [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Unmeasured hormone concentrations of simulated [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: SAE of estimates for simulated patients with [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Unmeasured hormone concentrations of simulated [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
read the original abstract

Model-based control techniques have recently been investigated for the recommendation of medication dosages to address thyroid diseases. These techniques often rely on knowledge of internal hormone concentrations that cannot be measured from blood samples. Moreover, the measurable concentrations are typically only obtainable at irregular sampling times. In this work, we empirically verify a notion of sample-based detectability that accounts for irregular sampling of the measurable concentrations on two pituitary-thyroid loop models representing patients with hypo- and hyperthyroidism, respectively, and include the internal concentrations as states. We then implement sample-based moving horizon estimation for the models, and test its performance on virtual patients across a range of sampling schemes. Our study shows robust stability of the estimator across all scenarios, and that more frequent sampling leads to less estimation error in the presence of model uncertainty and misreported dosages.

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 manuscript empirically verifies a sample-based detectability property for irregularly sampled measurable hormone concentrations in two pituitary-thyroid feedback loop ODE models (one each for hypo- and hyperthyroidism), augments the states with internal concentrations, implements a sample-based moving horizon estimator (MHE), and reports robust stability of the estimator together with reduced estimation error under higher sampling rates when model uncertainty and misreported dosages are present. All claims are supported exclusively by simulation on virtual patients whose trajectories are generated from the same two ODE models.

Significance. If the virtual-patient dynamics faithfully represent real hypo- and hyperthyroidism physiology, the work would supply a practical route to reconstruct unmeasurable states from sparse blood samples, thereby supporting model-based dosage recommendation. The reported robustness to sampling irregularity and model mismatch is a useful empirical observation for moving-horizon estimation in biomedical systems; however, the absence of any external clinical data or independent parameter identification substantially limits the immediate translational value.

major comments (2)
  1. [Numerical experiments / virtual-patient study] All detectability verification and MHE performance results (robust stability across scenarios, monotonic improvement with sampling frequency) are obtained by simulating the estimator on trajectories generated from the identical two pituitary-thyroid ODE models used to define the estimator itself. This closed-loop simulation setup does not probe structural mismatch with real patient physiology (unmodeled nonlinearities, unaccounted disturbances, or different parameter distributions) and therefore cannot substantiate the transferability of the stability and error-reduction claims.
  2. [Results on model uncertainty and dosage misreporting] The parameterization of “model uncertainty” and “misreported dosages” is defined internally to the same model family; no independent clinical data or cross-validation against alternative model structures is reported. Consequently the robustness result remains conditional on the simulation choices rather than on a broader validation.
minor comments (2)
  1. [Abstract] The abstract states that “more frequent sampling leads to less estimation error” but supplies no quantitative metrics (RMSE, maximum error, or statistical tests). Adding at least one table or figure with concrete error values would make the practical benefit clearer.
  2. [Preliminaries / Problem formulation] Notation for the sample-based detectability notion and the precise definition of the MHE cost function should be introduced with explicit equation numbers to facilitate reproducibility.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for the constructive comments and for recognizing the potential utility of the approach for reconstructing unmeasurable states. We address each major comment below, agreeing where the observations are accurate and outlining revisions to better contextualize the simulation-based nature of the study.

read point-by-point responses

  1. Referee: [Numerical experiments / virtual-patient study] All detectability verification and MHE performance results (robust stability across scenarios, monotonic improvement with sampling frequency) are obtained by simulating the estimator on trajectories generated from the identical two pituitary-thyroid ODE models used to define the estimator itself. This closed-loop simulation setup does not probe structural mismatch with real patient physiology (unmodeled nonlinearities, unaccounted disturbances, or different parameter distributions) and therefore cannot substantiate the transferability of the stability and error-reduction claims.

    Authors: We agree that the experiments rely on closed-loop simulations using the same ODE models for trajectory generation and estimation. This design enables direct verification of the sample-based detectability property and the MHE's robust stability under irregular sampling, model uncertainty, and dosage misreporting within the defined virtual-patient framework. We do not claim that the results demonstrate transferability to real physiology. We will revise the Discussion and Conclusions sections to explicitly note this scope limitation and the need for future validation against real clinical measurements to probe structural mismatches. revision: yes

  2. Referee: [Results on model uncertainty and dosage misreporting] The parameterization of “model uncertainty” and “misreported dosages” is defined internally to the same model family; no independent clinical data or cross-validation against alternative model structures is reported. Consequently the robustness result remains conditional on the simulation choices rather than on a broader validation.

    Authors: We concur that the uncertainty and misreporting scenarios are defined within the model family and that no independent clinical data or cross-validation is reported. The robustness findings are therefore specific to these simulation choices, serving as an empirical check of estimator behavior under the considered conditions. We will revise the manuscript by adding a dedicated limitations paragraph in the Conclusions that states the conditional nature of the results and the requirement for broader validation with clinical datasets and alternative models. revision: yes

standing simulated objections not resolved
  • Absence of external clinical data or independent parameter identification from real patient sources

Circularity Check

0 steps flagged

No circularity; performance claims rest on independent simulation validation

full rationale

The paper empirically verifies sample-based detectability for irregular sampling on two pituitary-thyroid ODE models and then evaluates moving-horizon estimation performance on virtual patients whose trajectories are generated from those same models. This is standard model-consistent simulation testing rather than any derivation that reduces to its inputs by construction. No self-definitional steps, fitted-input predictions, load-bearing self-citations, or ansatz smuggling appear in the reported chain. The robustness and sampling-frequency results are obtained from separate simulation runs under added uncertainty, not forced by the model equations themselves. The absence of external clinical data affects generalizability but does not create circularity in the logical structure.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Review based on abstract only. The paper relies on standard pituitary-thyroid feedback loop models and the validity of a sample-based detectability notion for irregular sampling; no explicit free parameters, axioms, or invented entities are detailed in the provided text.

pith-pipeline@v0.9.0 · 5453 in / 1023 out tokens · 32601 ms · 2026-05-16T23:25:32.200155+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

4 extracted references · 4 canonical work pages

  1. [1]

    and Rawlings, J.B

    Allan, D.A. and Rawlings, J.B. (2021). Robust stability of f ull infor- mation estimation. SIAM Journal on Control and Optimization , 59(5), 3472–3497. Andersson, J.A., Gillis, J., Horn, G., Rawlings, J.B., and D iehl, M. (2019). Casadi: a software framework for nonlinear optimiz ation and optimal control. Mathematical Programming Computation , 11(1), 1–3...

    work page 2021

  2. [2]

    Dietrich, J.W., Landgrafe, G., and Fotiadou, E.H. (2012). T sh and thyrotropic agonists: key actors in thyroid homeostasis. Journal of thyroid research , 2012(1), 351864. Dietrich, J.W.C. (2002). Der Hypophysen-Schilddr¨ usen-Regelkreis: Entwicklung und klinische Anwendung eines nichtlinearen M od- ells. Logos-Verlag. Gardner, D.G. and Greenspan, F.S. (20...

    work page 2012

  3. [3]

    Ji, L., Rawlings, J.B., Hu, W., Wynn, A., and Diehl, M. (2015) . Robust stability of moving horizon estimation under bounde d disturbances. IEEE Transactions on Automatic Control , 61(11), 3509–3514. Jonklaas, J., Bianco, A.C., Bauer, A.J., Burman, K.D., Capp ola, A.R., Celi, F.S., Cooper, D.S., Kim, B.W., Peeters, R.P., Ro sen- thal, M.S., et al. (2014)....

    work page doi:10.1109/tac 2015

  4. [4]

    Stengel, R.F. (1994). Optimal control and estimation . Courier Corporation. Theiler-Schwetz, V., Benninger, T., Trummer, C., Pilz, S., and Re- ichhartinger, M. (2022). Mathematical modeling of free thy roxine concentrations during methimazole treatment for graves’ d isease: development and validation of a computer-aided thyroid tre at- ment method. Fronti...

    work page 1994