{"id":"23d9654e-3c16-4224-bb18-7edc03af5d1b","arxiv_id":"2508.18195","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A matched-filter search of 115 gamma-ray pulsars finds no confirmed flux state changes from minutes to years and limits most flux variations to under 10%.","lead":"This paper searches for sudden brightness changes in 115 gamma-ray pulsars using a new statistical method sensitive to timescales of minutes, and finds that their emission is remarkably stable. It matters because it suggests pulsar magnetospheres rarely switch between different power-output states, which constrains models of how these spinning neutron stars accelerate particles.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Headline claim is conditional on the quasiperiodic two-state model family; aperiodic or unbounded-residence-time state switching is explicitly outside the search (end of §7), so the 'any sort' limits and 'single configuration' conclusion are not established for those processes.","rationale":"The reader's weakest_assumption pinpoints the same load-bearing concern: the paper's own end-of-§7 caveat that state switching with unbounded maximum state residence time would evade detection. My stress-test pass confirms that this is not a peripheral technicality. The matched-filter statistic in Eq. 5 is designed around harmonic templates of a square-wave process with bounded, uniform state-duration randomness; the approximate templates in Appendix A likewise cover only four asymmetry values and a finite range of spectral widths. Any real variability process lacking a characteristic quasi-period—such as a two-state process with exponential or power-law residence times—will have its power spread over a broad continuum, partly at frequencies removed by the slow-variability filtering step, and will not produce the harmonic structure the search targets. Since the paper's central claim is specifically about gamma-ray pulsar magnetospheres maintaining a single configuration, the strength of that claim is capped by the model family searched. The concrete test I propose would settle whether this concern is material by injecting such processes into the actual pipeline and measuring the false-negative rate at M = 0.1 and M = 1. I do not see a reason to move the reader's conditional verdict: the concern is real but explicitly acknowledged, and the paper's sensitivity statements for the quasiperiodic model family appear carefully derived and internally validated. The secondary issue of unconfirmed >7σ candidates (J1658−5324, J0613−0200) also supports keeping the verdict conditional, since the abstract's 'no new instances' is slightly stronger than the body's 'not confirmed' language, but it is less load-bearing than the model-family limitation.","tokens_in":19740,"tokens_out":5016,"duration_ms":71283,"concrete_test":"Simulate 15.6 yr of LAT photon-weight data for a Geminga-like bright pulsar whose true flux is a two-state process with exponentially or Pareto-distributed state durations (no bounded maximum state residence time), for modulation factors M = 0.1 and M = 1 and mean state lifetimes spanning minutes to months. Pass these simulations through the full pipeline used in the paper: Bayesian-blocks slow-variability filtering (§5), then the C_D matched-filter search over the same A, W, and f grid (§6). Measure the fraction of M = 1 realizations that exceed the paper's 7σ threshold, and compare the recovered upper limits with the injected M. If most high-M realizations are not detected, the reported <10% limits cannot be extended to aperiodic or unbounded-residence-time switching, and the conclusion should be explicitly restricted to quasiperiodic two-state processes.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The decisive limitation is stated in §7: 'a state switching process with an unbounded maximum state residence time would evade detection.' The detection statistic C_D (Eq. 5) is a matched filter built from harmonic templates at a trial frequency f, and the template family (§2, Appendix A) covers only bounded uniform state durations, W_f ≤ T_f and W_b ≤ T_b, at four discrete asymmetry values |A| ≤ 0.94. The abstract's 'wide range of possible state changing models' and 'best cases limit variations of any sort to ≤1%' therefore apply only to processes with a characteristic quasi-period and bounded residence times. Real state-changing phenomena cited in §1—mode changing, nulling, intermittent pulsars—are often aperiodic; for example, exponential or heavy-tailed residence-time distributions have no preferred harmonic stack, and much of their power is broadband or at low frequencies that the Bayesian-blocks slow-variability filtering removes before the fast search. The headline 'mostly stable' thus remains untested for such processes. This is not a hidden inconsistency; the paper states the caveat, but the caveat is load-bearing because the central conclusion that magnetospheres maintain a single configuration is extrapolated from a search that can only see quasiperiodic two-state switching. A secondary concern is that the abstract's 'no new instances' sits uneasily with the ≥7σ candidate excesses in §6.1 (J1658−5324 at 8.5σ, J0613−0200 at 7.3σ with no obvious background); these are not confirmed, but they mean the claim of uniqueness of J2021+4026 is only as strong as the background/systematic vetting of those candidates.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a two-stage search for flux variability in 115 bright gamma-ray pulsars using Fermi-LAT photon weights. Slow variability (timescales >= 2 weeks) is modeled and filtered with Bayesian blocks; fast variability (down to ~10 min) is searched with matched-filter statistics built from simulated response functions for a two-state quasiperiodic square-wave process with bounded random state durations. The author reports no confirmed new state changes, gives upper limits on modulation amplitude for the model family (typically <10%, and <=1% for the best cases), and interprets the null results as evidence that gamma-ray pulsar magnetospheres maintain a nearly constant single configuration. The paper is careful about exposure systematics (Appendix B), trial accounting, and sensitivity calibration, and it explicitly acknowledges in Section 7 that a state-switching process with unbounded residence times would evade the search.","tokens_in":20154,"tokens_out":4816,"duration_ms":61179,"significance":"If the central claim holds, this is a valuable ensemble result: it places the first sub-hour flux-variability constraints on a large sample of gamma-ray pulsars and substantially extends the parameter space probed by earlier slow-variability and single-pulse searches. The method itself is a useful contribution: the matched-filter aggregation of leaked power (Eq. 5) and the explicit response-function construction are sound, and the validation with injected signals (Figure 3) and the percent-level exposure agreement (Appendix B) are strengths. The upper limits on two-state quasiperiodic switching, especially the exclusion of strong nulling over most of the sample, are a concrete falsifiable result. The paper's astrophysical interpretation, however, is broader than the model family actually searched, and the abstract's 'no new instances' claim sits uneasily with the >7-sigma candidate excesses reported in Section 6.1.","major_comments":[{"comment":"The caveat stated at the end of Section 7 — 'a state switching process with an unbounded maximum state residence time would evade detection' — is load-bearing for the headline conclusion. The model in Section 2 (Eqs. 1–4) has bounded uniform state durations (0≤W_f≤T_f, 0≤W_b≤T_b), and the search templates (Section 6 and Appendix A) cover only four asymmetry values |A|≤0.94 and harmonically structured spectra. Real state-changing phenomena cited in Section 1, such as mode changing, nulling, and intermittent pulsars, are often aperiodic and can have heavy-tailed or exponential residence times, whose power is mostly broadband and partially removed by the slow-variability filtering. Thus the abstract's 'wide range of possible state changing models' and 'variations of any sort to ≤1%' are not established for that model family. Please restrict the abstract and conclusions to the bounded quasip","section":"§7 and abstract"},{"comment":"Several candidates exceed the adopted 7σ threshold: J1658−5324 at 8.5σ, J0218+4232 at 7.6σ, and J0613−0200 and J1816+4510 at 7.3σ. For J0613−0200 the text states there is 'no obvious background contamination,' and for J1658−5324 the excess is very broad. Calling the result 'no new instances of state changes' in the abstract is therefore overstated unless 'instances' means 'confirmed state changes after a more stringent criterion.' The paper should state the expected number of false positives under the quoted trial estimate, and either report these as candidate state changes or justify their assignment to background more explicitly, especially for J0613−0200 and J1658−5324.","section":"§6.1, §6, and abstract"}],"minor_comments":[{"comment":"Typos: 'methodlogy' (Section 4), 'suparass' (Section 5.3), 'impossible to difficult' and 'pulars' (Section 7), and 'Cygnux X-3' (Figure 5 caption).","section":"§4, §5.3, §7, Figure 5"},{"comment":"The slow-variability sensitivity estimate uses a generic σ_v for the modulation and then labels the limit 'M,' the same symbol as the two-state modulation factor in Eq. 1. Rename or explicitly distinguish this quantity to avoid conflating the slow and fast model families.","section":"§5.3 and Figure 6"},{"comment":"The sentence 'We select the known γ-ray pulsars those that have a 4FGL-DR4...' is a grammatical fragment; please rephrase.","section":"§4"},{"comment":"The phrase 'low (infinite) variance in the A→1 (A→−1) limits' is confusingly worded; clarify that the variance diverges as A→−1 and vanishes as A→1.","section":"§2, after Eq. 3"},{"comment":"The figure includes Q<1 cases labeled 'not meaningful.' Consider removing them or adding an explicit note that they are outside the physical range, so readers do not interpret them as valid templates.","section":"Figure 9"}],"recommendation":"major_revision","confidential_remarks":"The paper is built on the author's previous methodology and is largely self-contained. The two major comments concern scope matching between the model family searched and the astrophysical conclusion, and the treatment of candidate excesses relative to the 'no new instances' summary. These are fixable with text revisions and a more explicit treatment of the aperiodic/unbounded-residence-time case, but they are central to what the paper claims."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth reading. Kerr extends K19's power spectrum estimator with a matched filter that aggregates spectral leakage (Eq. 5) and uses it to search for stochastic two-state flux variability in 115 bright Fermi pulsars. That is genuinely new; the sensitivity curves in Fig. 8 are a useful resource. The pipeline is well documented: two-stage Bayesian-blocks filtering of slow variability, response functions built from simulations, trial correction, and comparison of P0/P1 spectra. Appendix B shows exposure residuals at the percent level for Vela. Credit where due: the method is clever, the validation is honest, and the null results on the bright pulsars are convincing for the model family considered.\n\nThe important caveat is that the family is quasiperiodic two-state switching with bounded state durations. Section 7 says explicitly that a process with unbounded maximum residence time would evade detection. That caveat is load-bearing. Real mode-changing and nulling can be aperiodic with heavy-tailed residence times; those processes put power in broadband or low frequencies that the slow-variability filter removes before the fast search. So the abstract's 'wide range of possible state changing models' and 'limit variations of any sort to ≤1%' overreach. The limits are real for the grid of A, Q, and f they searched, not for 'any sort.' A second issue: the abstract says 'no new instances of state changes,' but §6.1 reports candidates at 7–8.5σ (J1658−5324, J0613−0200, J0218+4232). They are treated as unresolved and likely contaminated, which is fair, but the abstract should say 'no confirmed new instances' or soften the claim. That is a wording fix, not a fatal flaw.\n\nThe slow-variability results also include eight single steps at the ends of data spans; the paper handles them sensibly. The sensitivity estimates are transparent, and the caveat about fainter pulsars is stated.\n\nNet: a solid, useful contribution. The central conclusion should be stated as: within the quasiperiodic two-state model family, gamma-ray pulsars are stable at the 10% level or better, and PSR J2021+4026 remains unique. That is a real result. The broader claim about 'single configuration' magnetospheres is plausible but not established for aperiodic processes.\n\nI would cite this in variability work and would send it to a referee. The referee should push on the abstract-vs-body gap and on whether the unbounded-residence caveat can be relaxed, but the paper deserves referee time. The heavy reliance on K19 is fine; it is the same pipeline extended, and the self-citation is not a problem here.","headline":"A careful matched-filter search that rules out strong quasiperiodic switching in 115 gamma-ray pulsars; the 'mostly stable' headline is real but narrower than the abstract says.","tokens_in":20628,"tokens_out":2246,"would_cite":true,"duration_ms":28389,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Gamma-ray pulsars hold their brightness steady for years, a 115-pulsar study finds.","keywords":["gamma-ray pulsars","pulsar magnetospheres","flux variability","state switching","Fermi LAT","matched filter","Bayesian blocks","spindown power"],"falsifier":"Re-analyze the marginal candidate PSR J0613-0200 with pulse-phase weighting and full background modeling: if its roughly 33/day quasi-periodic excess survives above 7 sigma, the claim that fast state changes are absent in the sample is wrong for at least one pulsar. Separately, compute R(f) for a state-switching process with unbounded residence times; if it yields a flat spectrum, the paper's limits do not cover the most aperiodic switching models.","tokens_in":19631,"feed_emoji":"⚡","tokens_out":8285,"duration_ms":96249,"temperature":0.7,"pith_summary":"This paper tries to establish that the gamma-ray brightness of pulsars barely changes on any timescale from about ten minutes to years, and that most radio state changes are not mirrored by changes in gamma-ray flux. It develops a statistical search that can spot faint, quasi-periodic switches in flux even when individual photons cannot be cleanly assigned to the pulsar. Applied to the 115 brightest gamma-ray pulsars observed by Fermi LAT over 15.6 years, the search finds no new instances of state changes and limits flux swings to below 10 percent for most of the sample, with the best cases restricted to 1 percent or less. If correct, this means a gamma-ray pulsar's magnetosphere normally stays in a single configuration or a narrow range of configurations with nearly constant power output, and the known case of PSR J2021+4026 is unique. The result directly constrains theories of how pulsar magnetospheres switch between states.","feed_headline":"115 pulsars show gamma-ray flux stable to 10% or better","feed_subtitle":"A matched-filter search of 15 years of Fermi data finds no new state changes; J2021+4026 stays unique.","key_machinery":"The central object is the matched-filter statistic C^2_D, built from the response function R(f): the expected power spectrum that a candidate flux-modulation process would produce after convolution with Fermi's exposure window. R(f) is computed by simulating the proposed two-state process many times and averaging the resulting spectra; C^2_D then weights the observed power spectrum by R(f), collecting spectral leakage that a plain peak search would miss. The variability model is an asymmetric, quasi-periodic square wave parameterized by modulation M, asymmetry A, and randomness Q, whose power-spectrum templates are approximated for four values of |A| and a grid of spectral widths. This machi","core_discovery":"The central claim is that gamma-ray pulsar flux state changes are neither widespread nor strong. Using a two-state quasiperiodic square-wave model, parameterized by modulation strength M, asymmetry A, and randomness Q, the author builds matched filters that gather the power a true periodic signal would leak across Fermi's complicated exposure window. The search spans timescales from about 10 minutes to a few days, after slow variations are modeled and filtered with Bayesian blocks. No new state changes are found: slow flux variability is limited to the 10 percent level across the sample, fast variability to roughly 10-20 percent for nearly periodic switching, and full nulling is excluded in","pith_inferences":["The reported limits apply only to switching that is at least quasi-periodic; a magnetosphere that switched on unbounded, aperiodic residence times would leave a flat power spectrum and escape the matched filter, so the 'mostly stable' conclusion may not cover the most irregular state-changing process.","If flux variations track spindown power, gamma-ray-bright pulsars' spin-down is now constrained to be stable on minute-to-day timescales, a new restriction on models that locate state-changing mechanisms near the polar cap rather than in the outer magnetosphere.","The marginal fast-variability candidates, such as PSR J0613-0200 at about 33/day and PSR J1658-5324, are natural targets for the proposed pulse-phase-resolved analysis; if one survives full background modeling, the sample would not be perfectly stable.","The same two-stage matched-filter framework could be extended to other Fermi LAT source classes, such as blazars or binary systems, where exposure leakage similarly limits searches for quasi-periodic flux modulation."],"forward_implications":["If gamma-ray flux traces spindown power, gamma-ray pulsars spend essentially all of their time in a force-free magnetosphere state, with excursions limited to 1-10 percent.","PSR J2021+4026 remains the only known gamma-ray flux state changer, so any general theory of pulsar state changing must explain why this pulsar is unique among the brightest 115.","The 0.1-10 percent spindown variations seen in radio pulsar timing are not accompanied by comparable gamma-ray flux changes on short timescales, suggesting those variations either operate on longer timescales or occur in regions that do not affect the gamma-ray-emitting magnetosphere.","Substantial nulling of gamma-ray emission is excluded for nearly all of the sample, so gamma-ray pulsars do not null the way some radio pulsars do.","Pulse-phase weighting should improve sensitivity by more than a factor of 3, allowing the best limits to be pushed below 1 percent and enabling searches for pulse-profile shape variations."],"supporting_citations":[{"why":"Supplies the photon-weighting likelihood framework and the Leahy-normalized power spectrum P(f) that the matched-filter search extends.","marker":"M. Kerr (2019, hereafter K19)"},{"why":"Establishes that optimal photon weights are the probability that a photon originates from the source of interest.","marker":"P. Bickel et al. (2008)"},{"why":"Shows that constructing and applying such weights boosts the sensitivity of pulsation and variability searches.","marker":"M. Kerr (2011)"},{"why":"Provides the Bayesian blocks algorithm used to model and filter slow source and background variability.","marker":"J. D. Scargle et al. (2013)"},{"why":"Discovery of the only known gamma-ray flux state-changing pulsar, PSR J2021+4026, the benchmark the search aims to extend.","marker":"A. Allafort et al. (2013)"},{"why":"Quantifies the step-like correlated spindown and gamma-ray flux changes in PSR J2021+4026 on 2-3 year timescales.","marker":"A. Fiori et al. (2024)"},{"why":"Documents 0.1-10 percent spindown variations tied to radio profile changes, the comparison population for the gamma-ray stability limits.","marker":"A. Lyne et al. (2010)"},{"why":"The 4FGL catalog supplies the sky model used to compute photon weights and predicted exposures.","marker":"S. Abdollahi et al. (2020)"},{"why":"4FGL-DR4 provides the test-statistic threshold used to select the 115-pulsar sample.","marker":"J. Ballet et al. (2023)"}],"fun_headline_variants":["Gamma-ray pulsars stay remarkably stable over years","No new gamma-ray pulsar state changes in 15 years of Fermi data","Flux variability of 115 pulsars capped at 10%","Pulsar magnetospheres: stable configurations, limited variation"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The search assumes real state-switching would leave a non-flat power spectrum, so a process whose states last for unbounded, aperiodic times could hide completely and the reported stability limits would not apply.","fun_headline_variants_meta":{"raw":{"variants":["Gamma-ray pulsars stay remarkably stable over years","No new gamma-ray pulsar state changes in 15 years of Fermi data","Flux variability of 115 pulsars capped at 10%","Pulsar magnetospheres: stable configurations, limited variation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00023,"raw_usage":{"total_tokens":1321,"prompt_tokens":748,"completion_tokens":573,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":492,"completion_tokens_details":{"reasoning_tokens":510}},"tokens_in":492,"tokens_out":573,"duration_ms":7173,"temperature":1.0,"reasoning_tokens":510,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T16:32:12.403007+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-analyze the marginal candidate PSR J0613-0200 with pulse-phase weighting and full background modeling: if its roughly 33/day quasi-periodic excess survives above 7 sigma, the claim that fast state changes are absent in the sample is wrong for at least one pulsar. Separately, compute R(f) for a state-switching process with unbounded residence times; if it yields a flat spectrum, the paper's limits do not cover the most aperiodic switching models.","supporting_citations":[],"review_version":1}