Pith. sign in

REVIEW 5 major objections 5 minor 56 references

Nonlinear reconstruction of 21cm global signal from 21cm power spectrum with artificial neural networks

T0 review · 5 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A neural network trained on 400 simulated universes reconstructs the 21cm global signal from the 21cm power spectrum alone, with mean correlation 0.95 on held-out models and 0.83 under SKA-1 thermal noise.

desk verdict Honest feasibility study of ANN-based 21cm signal reconstruction that is overstated because the 'accurate recovery' claim leans entirely on correlation coefficients, with no absolute error metric or baseline. read the letter →

arxiv 2412.20862 v3 pith:LTLCWM3L submitted 2024-12-30 astro-ph.CO

Hayato Shimabukuro This is my paper · ORCID
classification astro-ph.CO
keywords 21cmcosmologyglobalsignalpowerspectrumartificialneuralnetworksEpochofReionizationcosmicdawnSKA-1thermalnoisesemi-numericalsimulations
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

Hydrogen's 21cm line can be measured in two complementary ways: single-dish telescopes record the sky-averaged global signal, while interferometers record spatial fluctuations summarized by the power spectrum. This paper argues that a neural network can bridge the two, reconstructing the global signal from power-spectrum measurements even though no direct mathematical relation exists between them. The connection is indirect—both observables respond to the same astrophysical and cosmological parameters—and the network learns that mapping from 400 semi-numerical simulations. On 100 held-out models, the reconstructed global signal matches the truth with mean correlation 0.95 when the input is the power spectrum at $k=0.1\,h\,\mathrm{Mpc}^{-1}$, and with mean correlation 0.83 when realistic SKA-1 thermal noise is added. If correct, this gives interferometric experiments an independent handle on the global signal and a way to cross-check single-dish results without an extra telescope.

What carries the argument

The central object is a feedforward artificial neural network with one hidden ReLU layer, trained by backpropagation with the Adam optimizer for 20,000 iterations on 400 semi-numerical 21cm models. Its input is the 21cm power spectrum at a fixed wavenumber (or, in one experiment, a grid of wavenumbers over EoR redshifts), sampled over 76 redshift bins from $z=7.5$ to 35, and its output is the global brightness-temperature history over the same redshift range. The network serves as a nonlinear regression that captures the implicit, model-dependent relationship between the two observables: because both are produced by the same underlying astrophysical parameters, sufficiently many training models let the network interpolate new combinations. The scale-dependence result is diagnostic of what the network actually uses: larger-scale modes are correlated with the astrophysical processes that shape the sky-average, whereas small-scale modes mostly track local structure.

What would settle it

Retrain the identical single-hidden-layer network on a random 300-model subset, keep 100 models for validation and 100 for testing, and watch the validation error during training; if the validation error rises while the training error falls, or if the held-out correlation at $k=0.1\,\mathrm{Mpc}^{-1}$ lands well below 0.95, then the claimed generalization is not supported.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the sky-averaged 21cm brightness temperature—the global signal, which is mathematically the $k=0$ Fourier mode—can be predicted from the 21cm power spectrum at finite wavenumbers using a single-hidden-layer neural network. The prediction succeeds not because small-scale fluctuations average out, but because the ANN learns the model-dependent correlations set by the underlying astrophysical parameters. Quantitatively, using the power spectrum at $k=0.1\,\mathrm{Mpc}^{-1}$ recovers the global signal across $z=7.5$ to 35 with a mean correlation coefficient of 0.95 on the 100-model test set; adding SKA-1-level thermal noise lowers the mean to 0.83, and even 1000-times-stronger noise still permits adequate recovery. Recovery quality depends strongly on scale: $k=0.5\,\mathrm{Mpc}^{-1}$ gives a mean correlation of 0.88, while $k=1.0\,\mathrm{Mpc}^{-1}$ gives only 0.18, because small-scale power spectra lack the three-peaked structure that encodes the astrophysical history.

Load-bearing premise

The accuracy claim rests on the assumption that a network trained on 400 simulated models, with no separate validation set, generalizes to the 100 held-out models; the paper infers this from the training error flattening out, not from checking the error on untouched data.

Editorial extensions

If this is right

  • SKA-1's power-spectrum measurement at $k=0.1\,\mathrm{Mpc}^{-1}$ could be converted into an estimate of the global 21cm signal, giving interferometers an indirect route to a single-dish observable.
  • Reconstruction quality falls rapidly with wavenumber, so practical applications should restrict inputs to larger-scale modes; including $k=1.0\,\mathrm{Mpc}^{-1}$ data, with its mean correlation of 0.18, would degrade the result.
  • The method remains useful under realistic noise: with SKA-1 thermal noise the mean correlation is 0.83, and with 1000-times-stronger noise the recovered signal is still adequate, which the paper takes as evidence the approach can work for MWA or LOFAR.
  • The ANN is model-dependent: if the true global signal contains physics absent from the training simulations, the network will interpret it within the learned parameter space rather than flagging it as new physics, so this is a cross-check of standard-model cosmic history, not an anomaly detector.

Reading between the lines

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

  • The paper's acknowledged omission of a validation set means the headline 0.95 and 0.83 correlations should be read as provisional until test-side loss curves are reported; checking those curves is the cheapest way to test the claim before real data arrive.
  • Because the best results come from low-k input, a natural next experiment the paper does not run is to feed the full multi-k, full-redshift power spectrum into the same network; the EoR-only multi-k case already shows mean correlation 0.774, so a complete multi-k input could improve on the single-k numbers.
  • The wide spread of correlation coefficients (some test models below 0.6 even without noise) points toward a per-model reliability diagnostic: if the power spectrum lacks the usual peak structure, the recovered global signal should be flagged as untrustworthy rather than averaged into a headline number.
  • If paired with Bayesian or ensemble uncertainty estimates, the network could output a confidence interval per redshift, turning the method from a point predictor into a tool for identifying which cosmic histories leave recoverable imprints.
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

5 major / 5 minor

Summary. The manuscript proposes a method to reconstruct the sky-averaged 21cm global signal from the 21cm power spectrum using a single-hidden-layer artificial neural network. The training and test data are 500 semi-numerical models from the Cohen et al. simulation code, with 400 models used for training and 100 held out for testing. The ANN takes the 21cm power spectrum at fixed wavenumbers (k = 0.1, 0.5, 1.0 Mpc^-1) or over a range of k and redshifts, and outputs the global brightness temperature as a function of redshift. The authors report mean correlation coefficients of 0.95 for the noise-free k=0.1 Mpc^-1 case, 0.83 with SKA-1 thermal noise, and 0.18 at k=1.0 Mpc^-1, concluding that the method accurately recovers the global signal and is robust to realistic noise. The paper acknowledges the model-dependence of the approach and lists several limitations, including the absence of a validation set, lack of uncertainty quantification, and no treatment of foregrounds.

Significance. If the reported accuracy is reliable, this work would provide a computationally inexpensive way to cross-validate 21cm global-signal measurements from single-dish experiments (e.g., EDGES) against power-spectrum measurements from interferometers (e.g., SKA-1), within the framework of standard astrophysical models. The study benefits from using a physically motivated simulation suite, testing multiple wavenumbers, and including a realistic SKA-1 noise model with a clear statement of the inherent model-dependence. However, the quantitative claims as presented are not yet fully supported: the primary accuracy metric (correlation coefficient) is insensitive to amplitude and offset biases, no absolute error metric is reported over the test set, the 1000x-noise robustness claim is based on a single example, and no comparison with a simpler baseline is made. The proof-of-concept is interesting, but the evidence as presented is insufficient to establish the stated accuracy in a way that would be useful for practical cross-validation.

major comments (5)
  1. [Section 4.1, Eq. (8)] The primary quantitative accuracy metric is the correlation coefficient (CC) defined in Eq. (8). CC is invariant under affine transformations (a*y_true + b) of the predicted signal, so it cannot detect amplitude bias or offset errors. The reported mean CC of 0.95 therefore does not by itself establish accurate recovery of the global signal in an absolute sense. The manuscript reports no absolute error statistic (e.g., RMSE, bias, or scatter) over the 100-model test set; the examples in Figs. 2 and 8 are individual cases, and Fig. 3, while showing trough-depth scatter, lacks a quantitative error summary. Please add absolute error metrics (e.g., RMSE as a function of redshift, mean bias, and scatter) and, ideally, error bars on the recovered light curve (e.g., via ensemble or bootstrap over noise realizations).
  2. [Section 3.3] No validation set is used. The assertion that the training MSE plateau at 20,000 iterations "suggests that our training dataset is sufficiently large to avoid severe over- or under-fitting" is not an adequate substitute for monitoring validation or test loss during training. Because the network architecture and number of iterations are fixed choices, this may not be fatal, but the claim of generalization to the test set would be strengthened by reporting the test MSE as a function of training iterations, or by using a train/validation/test split. Please provide evidence that the network is not overfitting, e.g., a test-loss curve or comparison of train and test performance.
  3. [Section 4.2, Fig. 8] The claim that the global signal can be "adequately recovered" when the thermal noise is 1000 times larger than SKA-1 is based on a single example shown in Fig. 8. No aggregate performance metric (e.g., CC distribution or RMSE) is reported for the 1000x-noise scenario, so the robustness claim is not supported. Please provide a distribution of CC or error metrics over the test set for this case.
  4. [Section 4.1, Fig. 4] The CC distribution shows a non-negligible tail with CC below 0.6 (Fig. 4). The manuscript attributes this to anomalous power spectra but does not quantify the absolute errors for these cases. Since the mean CC may be dominated by high-CC models, the practical accuracy for the worst cases is unknown. Please report the RMSE or bias for the low-CC tail and discuss whether these correspond to identifiable regions of parameter space.
  5. [Sections 3.2 and 4.1] The paper claims the ANN captures a nonlinear mapping between the power spectrum and the global signal, but no comparison is made with a simpler baseline, such as a linear regression from the same power-spectrum input to the global signal. Without such a baseline, it is unclear whether the high CC reflects the ANN's nonlinear modeling or simply the linear correlation already present between these observables. Please include a baseline comparison (e.g., linear regression or linear regression with polynomial features) to support the "nonlinear reconstruction" claim.
minor comments (5)
  1. [Section 3.3] The phrase "we show the MSE as a function of the number of iterations" is repeated twice in two consecutive sentences; please remove the duplication.
  2. [Section 3.2] The text below Eq. (5) uses "wights" instead of "weights"; please correct the typo.
  3. [Section 4.2, Eq. (9)] Equation (9) is typeset in a garbled manner (e.g., "d2 Ay Ω t T2 sys ¯n(Lk⊥)Ae") and the terms are not clearly defined. Please rewrite the equation with clear notation and define each symbol.
  4. [Section 4.2] The text states "at redshifts z = 7.5 − −15" and "z = 7.5 − −35" with double minus signs; these should be written as ranges, e.g., "z = 7.5–15" and "z = 7.5–35".
  5. [Section 4.2] The paper states that cosmic variance "contributes significantly at large scales" at the beginning of the section, but later claims "For SKA-1, the cosmic variance is negligible." Please clarify which scales and signal levels are meant.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: ANN recovery is a held-out regression with model-dependence explicitly acknowledged.

full rationale

The paper's derivation chain is a supervised regression: 500 Cohen et al. semi-numerical models produce (power spectrum, global signal) pairs; 400 train the ANN and 100 held-out models evaluate it. The claimed 'recovery' is the trained network's output on inputs not seen during training. There is no step in which the target quantity is defined in terms of the input, no parameter fitted to the test data and then called a prediction, and no load-bearing uniqueness or ansatz imported through self-citation. The fact that training and test data come from the same simulation code means the evaluation is in-distribution and the method is model-dependent; the paper states this limitation explicitly ('fundamentally model-dependent... cannot provide a genuinely independent validation of anomalous signals'). The reliance on CC and the absence of a validation set are legitimate statistical/generalization criticisms, but they are not circularity by the definitions used here: they do not reduce a prediction to an input by construction. The self-citations in the introduction are contextual and not load-bearing.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the training data generated by an external simulation code, on the unvalidated ANN setup, and on the noise model. There are no new physical entities, but there are several hand-chosen parameters (learning rate, batch size, iteration count, wavenumbers) and an undisclosed hidden-layer width. The most consequential assumption is that the ANN generalizes beyond its training distribution, which the paper is only partially equipped to support.

free parameters (3)
  • ANN hyperparameters: learning rate, batch size, iteration count = learning_rate=1e-3, batch_size=20, n_iter=20000
    Hand-chosen values; 20,000 iterations were selected because the training MSE plateaus, with no validation set to justify the choice.
  • Hidden layer width (number of neurons) = not reported
    The architecture description in Section 3.2 does not state how many hidden neurons are used, so the network capacity is an undocumented free choice.
  • Wavenumber k selections = 0.1, 0.5, 1.0 h/Mpc
    Chosen as representative of SKA-1 sensitivity; recovery accuracy strongly depends on this choice, as shown in Section 4.1.
assumptions (5)
  • domain assumption Flat Lambda-CDM cosmology with Planck 2018 parameters
    Stated in Section 2; all simulations and noise calculations assume this cosmology.
  • domain assumption Cohen et al. 2017/2018 semi-numerical code produces reliable 21cm power spectra and global signals
    Sections 2 and 3.1; the entire training and test data come from this code, not from independent measurements.
  • ad hoc to paper A single-hidden-layer ReLU ANN trained on 400 models generalizes without a validation set
    Section 3.3; the paper infers adequate generalization from the training-loss plateau alone, which is not a sufficient check.
  • domain assumption SKA-1 thermal noise model (224 stations, 1000 hours, stated beam/baseline parameters) is correct
    Section 4.2; the noise-robustness results depend on this assumed noise model.
  • ad hoc to paper The test set drawn from the same simulation suite is representative of future real data
    Section 3.3; the paper itself acknowledges the model dependence and the limitation in Section 5.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Nonlinear reconstruction of 21cm global signal from 21cm power spectrum with artificial neural networks." pith.science (2026). https://pith.science/paper/LTLCWM3L

@misc{pith2026241220862,
  author       = {Pith},
  title        = {Pith review of: Nonlinear reconstruction of 21cm global signal from 21cm power spectrum with artificial neural networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LTLCWM3L}},
  note         = {Machine review of arXiv:2412.20862}
}
read the original abstract

In this paper, we propose a novel method to recover the 21cm global signal from the 21cm power spectrum using artificial neural networks (ANNs). The 21cm global signal is crucial for understanding cosmic evolution from the Dark Ages through the Epoch of Reionization (EoR). While interferometers like LOFAR, MWA, HERA, and SKA focus on detecting the 21cm power spectrum, single-dish experiments such as EDGES target the global signal. Our method utilizes ANNs to establish a connection between these two observables, providing a means to cross-validate independent 21cm line observations. This capability is significant as it allows different observational approaches to verify each other's results, ensuring greater reliability in 21cm cosmology. We demonstrate that our ANN-based approach can accurately recover the 21cm global signal across a wide redshift range (z=7.5-35) from simulated data, even when realistic thermal noise levels, such as those expected from SKA-1, are considered. This cross-validation process strengthens the robustness of 21cm signal analysis, offering a more comprehensive understanding of the early universe.

Figures

Figures reproduced from arXiv: 2412.20862 by the authors.

Figure 1
Figure 1. The mean square error (MSE) between obtained and target values in the training dataset. The MSE [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. True 21cm global signal(red solid) and recovered 21cm global signal from 21cm line power spectrum from EoR to cosmic dawn(z=7.5-35)(black dashed). 1.0, with a mean of 0.95 and a variance of 0.05. This quantitatively demonstrates that our artificial neural network (ANN) can successfully recover the 21cm global signal from the 21cm PS at k = 0.1Mpc−1 for most of the models we consider. However, for some models, the CC… view at source ↗
Figure 3
Figure 3. Comparison of depth of trough between obtained and true global signal values for all test datasets. [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Distribution of the correlation coefficient (CC) for all test datasets where the global signal is recov [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: (Top) The global 21cm signal recovered using the 21cm PS at fixed wavenumbers k = 0.1, 0.5, and 1.0Mpc−1 . (Bottom) The corresponding 21cm PS as functions of redshift at the same fixed wavenumbers. Note that the recovery deteriorates at k = 1.0Mpc−1 , where the PS exhi…
Figure 6
Figure 6. Figure 6: Distribution of the correlation coefficient (CC) for all test datasets where the global signal is re [PITH_FULL_IMAGE:figures/full_fig_p018_6.png]
Figure 7
Figure 7. Figure 7: Comparison of 21 cm power spectrum and thermal noise power spectrum. The 21 cm power spectrum [PITH_FULL_IMAGE:figures/full_fig_p019_7.png]
Figure 8
Figure 8. Figure 8: Recovered global signal from noisy 21cm line power spectrum. The target global signal ( [PITH_FULL_IMAGE:figures/full_fig_p020_8.png]
Figure 9
Figure 9. Figure 9: Distribution of the correlation coefficient (CC) for the recovery from the 21cm PS with thermal [PITH_FULL_IMAGE:figures/full_fig_p020_9.png]
Figure 10
Figure 10. Figure 10: Examples of the recovered 21cm global signal using the 21cm line power spectrum during the EoR [PITH_FULL_IMAGE:figures/full_fig_p021_10.png]
Figure 11
Figure 11. Figure 11: Distribution of the correlation coefficient (CC) for all test datasets when recovering the global signal [PITH_FULL_IMAGE:figures/full_fig_p022_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 43 canonical work pages

  1. [1]

    D., Bassett, N., Burns, J

    Bale, S. D., Bassett, N., Burns, J. O., et al. 2023, arXiv e-prints, arXiv:2301.10345

    arXiv 2023

  2. [2]

    2018, Nature, 555, 71

    Barkana, R. 2018, Nature, 555, 71

    work page 2018

  3. [3]

    2001, Phys

    Barkana, R., & Loeb, A. 2001, Phys. Rep., 349, 125

    2001

  4. [4]

    Bevins, H. T. J., Handley, W. J., Fialkov, A., de Lera Acedo, E., & Javid, K. 2021, MNRAS, 508, 2923

    work page 2021

  5. [5]

    K., Iliev, I

    Bianco, M., Giri, S. K., Iliev, I. T., & Mellema, G. 2021, MNRAS, 505, 3982

    work page 2021

  6. [6]

    S., La Plante, P., & Aguirre, J

    Billings, T. S., La Plante, P., & Aguirre, J. E. 2021, PASP, 133, 044001

    work page 2021

  7. [7]

    D., Rogers, A

    Bowman, J. D., Rogers, A. E. E., Monsalve, R. A., Mozdzen, T. J., & Mahesh, N. 2018, Nature, 555, 67

    2018

  8. [8]

    G., et al

    Breitman, D., Mesinger, A., Murray, S. G., et al. 2024, MNRAS, 527, 9833

    work page 2024

Show all 56 references
  1. [9]

    H., Portillo, S

    Bye, C. H., Portillo, S. K. N., & Fialkov, A. 2022, ApJ, 930, 79

    2022

  2. [10]

    2021, Philosophical Transactions of the Royal Society of London Series A, 379, 20190566

    Chen, X., Yan, J., Deng, L., et al. 2021, Philosophical Transactions of the Royal Society of London Series A, 379, 20190566

    2021

  3. [11]

    2024, arXiv e-prints, arXiv:2403.16409

    Chen, X., Gao, F., Wu, F., et al. 2024, arXiv e-prints, arXiv:2403.16409

    2024 arXiv

  4. [12]

    R., Paranjape, A., & Maity, B

    Choudhury, T. R., Paranjape, A., & Maity, B. 2024, J. Cosmol. Astropart. Phys., 2024, 027

    2024

  5. [13]

    2018, MNRAS, 478, 2193

    Cohen, A., Fialkov, A., & Barkana, R. 2018, MNRAS, 478, 2193

    2018

  6. [14]

    2017, MNRAS, 472, 1915

    Cohen, A., Fialkov, A., Barkana, R., & Lotem, M. 2017, MNRAS, 472, 1915

    2017

  7. [15]

    Cohen, A., Fialkov, A., Barkana, R., & Monsalve, R. A. 2020, MNRAS, 495, 4845

    2020

  8. [16]

    R., Parsons, A

    DeBoer, D. R., Parsons, A. R., Aguirre, J. E., et al. 2017, PASP, 129, 045001

    2017

  9. [17]

    2019, MNRAS, 490, 371

    Doussot, A., Eames, E., & Semelin, B. 2019, MNRAS, 490, 371

    2019

  10. [18]

    2019, MNRAS, 486, 1763

    Fialkov, A., & Barkana, R. 2019, MNRAS, 486, 1763

    2019

  11. [19]

    2020, MNRAS, 491, 3108

    Fialkov, A., Barkana, R., & Jarvis, M. 2020, MNRAS, 491, 3108

    2020

  12. [20]

    2016, ApJ, 821, 59

    Fialkov, A., & Loeb, A. 2016, ApJ, 821, 59

    2016

  13. [21]

    R., Oh, S

    Furlanetto, S. R., Oh, S. P., & Briggs, F. H. 2006, Phys. Rep., 433, 181 Reconstructing 21cm global signal 15

    2006

  14. [22]

    2021, MNRAS, 504, 4716

    Gagnon-Hartman, S., Cui, Y ., Liu, A., & Ravanbakhsh, S. 2021, MNRAS, 504, 4716

    2021

  15. [23]

    2019, MNRAS, 484, 282

    Gillet, N., Mesinger, A., Greig, B., Liu, A., & Ucci, G. 2019, MNRAS, 484, 282

    2019

  16. [24]

    D., Watkinson, C

    Jennings, W. D., Watkinson, C. A., Abdalla, F. B., & McEwen, J. D. 2019, MNRAS, 483, 2907

    2019

  17. [25]

    C., Gagnon-Hartman, S., et al

    Kennedy, J., Carr, J. C., Gagnon-Hartman, S., et al. 2024, MNRAS, 529, 3684

    2024

  18. [26]

    S., Liu, A., Parsons, A

    Kern, N. S., Liu, A., Parsons, A. R., Mesinger, A., & Greig, B. 2017, ApJ, 848, 23

    2017

  19. [27]

    2015, Advancing Astrophysics with the Square Kilometre Array (AASKA14), 1

    Koopmans, L., Pritchard, J., Mellema, G., et al. 2015, Advancing Astrophysics with the Square Kilometre Array (AASKA14), 1

    2015

  20. [28]

    2019, MNRAS, 485, 2628

    Li, W., Xu, H., Ma, Z., et al. 2019, MNRAS, 485, 2628

    2019

  21. [29]

    R., Oh, S

    Lidz, A., Furlanetto, S. R., Oh, S. P., et al. 2011, ApJ, 741, 70

    2011

  22. [30]

    R., Allison, R., et al

    Liu, A., Pritchard, J. R., Allison, R., et al. 2016, Phys. Rev. D, 93, 043013

    2016

  23. [31]

    R., Tegmark, M., & Loeb, A

    Liu, A., Pritchard, J. R., Tegmark, M., & Loeb, A. 2013, Phys. Rev. D, 87, 043002

    2013

  24. [32]

    Madau, P., Meiksin, A., & Rees, M. J. 1997, ApJ, 475, 429

    1997

  25. [33]

    Mao, Y ., D’Aloisio, A., Zhang, J., & Shapiro, P. R. 2013, Phys. Rev. D, 88, 081303

    2013

  26. [34]

    2008, Phys

    Mao, Y ., Tegmark, M., McQuinn, M., Zaldarriaga, M., & Zahn, O. 2008, Phys. Rev., D78, 023529

    2008

  27. [35]

    M., Sokolowski, M., et al

    McKinley, B., Trott, C. M., Sokolowski, M., et al. 2020, MNRAS, 499, 52

    2020

  28. [36]

    McQuinn, M., Zahn, O., Zaldarriaga, M., Hernquist, L., & Furlanetto, S. R. 2006, Astrophys. J., 653, 815

    2006

  29. [37]

    Mellema, G., Koopmans, L. V . E., Abdalla, F. A., et al. 2013, Experimental Astronomy, 36, 235

    2013

  30. [38]

    2014, MNRAS, 439, 3262 Nambissan T., J., Subrahmanyan, R., Somashekar, R., et al

    Mesinger, A., Ewall-Wice, A., & Hewitt, J. 2014, MNRAS, 439, 3262 Nambissan T., J., Subrahmanyan, R., Somashekar, R., et al. 2021, arXiv e-prints, arXiv:2104.01756

    2014 arXiv

  31. [39]

    C., et al

    Philip, L., Abdurashidova, Z., Chiang, H. C., et al. 2019, Journal of Astronomical Instrumentation, 8, 1950004 Planck Collaboration, Aghanim, N., Akrami, Y ., et al. 2020, A&A, 641, A6

    2019

  32. [40]

    S., Burns, J

    Polidan, R. S., Burns, J. O., Ignatiev, A., et al. 2024, Advances in Space Research, 74, 528 Prelogovi´c, D., Mesinger, A., Murray, S., Fiameni, G., & Gillet, N. 2022, MNRAS, 509, 3852

    2024

  33. [41]

    E., Liu, A., & Parsons, A

    Presley, M. E., Liu, A., & Parsons, A. R. 2015, ApJ, 809, 18

    2015

  34. [42]

    C., Greenhill, L

    Price, D. C., Greenhill, L. J., Fialkov, A., et al. 2018, MNRAS, 478, 4193

    2018

  35. [43]

    R., & Loeb, A

    Pritchard, J. R., & Loeb, A. 2012, Reports on Progress in Physics, 75, 086901

    2012

  36. [44]

    E., Hinton, G

    Rumelhart, D. E., Hinton, G. E., & Williams, R. J. 1986, Nature, 323, 533

    1986

  37. [45]

    Sailer, N., Chen, S.-F., & White, M. 2022, J. Cosmol. Astropart. Phys., 2022, 007 Sathyanarayana Rao, M., Singh, S., K. S., S., et al. 2023, Experimental Astronomy, 56, 741

    2022

  38. [46]

    J., & Pritchard, J

    Schmit, C. J., & Pritchard, J. R. 2018, MNRAS, 475, 1213

    2018

  39. [47]

    Scott, D., & Rees, M. J. 1990, MNRAS, 247, 510

    1990

  40. [48]

    2023, PASJ, 75, S1

    Shimabukuro, H., Hasegawa, K., Kuchinomachi, A., Yajima, H., & Yoshiura, S. 2023, PASJ, 75, S1

    2023

  41. [49]

    2022, Research in Astronomy and Astrophysics, 22, 035027

    Shimabukuro, H., Mao, Y ., & Tan, J. 2022, Research in Astronomy and Astrophysics, 22, 035027

    2022

  42. [50]

    2017, MNRAS, 468, 3869

    Shimabukuro, H., & Semelin, B. 2017, MNRAS, 468, 3869

    2017

  43. [51]

    Shmueli, G., Sarkar, D., & Kovetz, E. D. 2023, Phys. Rev. D, 108, 083531

    2023

  44. [52]

    2024b, arXiv e-prints, arXiv:2410.04792

    Sikder, S., Fialkov, A., & Barkana, R. 2024b, arXiv e-prints, arXiv:2410.04792

  45. [53]

    2018, ApJ, 858, 54 16 H.Shimabukuro van Haarlem, M

    Singh, S., Subrahmanyan, R., Udaya Shankar, N., et al. 2018, ApJ, 858, 54 16 H.Shimabukuro van Haarlem, M. P., Wise, M. W., Gunst, A. W., et al. 2013, A&A, 556, A2

    2018

  46. [54]

    B., Tingay, S

    Wayth, R. B., Tingay, S. J., Trott, C. M., et al. 2018, PASA, 35, 33

    2018

  47. [55]

    2021, MNRAS, 506, 357

    Yoshiura, S., Shimabukuro, H., Hasegawa, K., & Takahashi, K. 2021, MNRAS, 506, 357

    2021

  48. [56]

    2023, ApJ, 945, 109

    Zhang, X., Yue, B., Shi, Y ., Wu, F., & Chen, X. 2023, ApJ, 945, 109

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.