{"id":"59a63a80-b41f-4245-b592-462a60971c1d","arxiv_id":"2602.10208","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A stacking search of 14 southern Seyferts with IceCube starting tracks yields a 3.0σ excess of 6.7 neutrinos, supporting the disk-corona model.","lead":"IceCube found a 3-sigma excess of high-energy neutrinos from a group of 14 X-ray-bright Seyfert galaxies in the southern sky. This is new, independent evidence that these black-hole-powered galaxies contribute to the cosmic neutrino flux.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Stacking significance depends on disk-corona model weights; without a model-independent test (e.g., binomial), the 3.0σ excess may not robustly establish Seyfert population emission.","rationale":"The paper is a well-executed IceCube stacking search with a modest (3.0σ) excess. The background model is validated with multiple checks, and the pre-trial/post-trial individual-source significances are properly reported. The reader's CONDITIONAL verdict correctly identifies the disk-corona model calibration as the weakest assumption. My reading agrees: the stacking weights are derived from a model that is calibrated to NGC 1068 and extrapolated to 13 Southern Seyferts via X-ray luminosity. If this scaling is wrong, the stacking's significance is not necessarily invalid (the p-value is computed under the background-only hypothesis), but the interpretation of the excess as evidence for a Seyfert population becomes model-dependent. The paper lacks a model-independent collective test for the Southern sample, unlike the Northern-sky analysis which used a binomial test. The consistency between observed and predicted event counts is too weak to independently validate the scaling. A binomial or equal-weight stacking test would directly probe whether the spatial excess persists without the model weights. Since the concern is addressable and the result is already framed as 'evidence' at 3.0σ, the reader's CONDITIONAL verdict is appropriate; my critique does not change it. No fatal flaw or internal inconsistency was found; the non-public nature of the data and code limits independent verification but is not a technical flaw in the argument.","tokens_in":16620,"tokens_out":12404,"duration_ms":141665,"concrete_test":"Perform a model-independent binomial test on the 13 stacked sources (excluding Centaurus A): for each source, count events within a fixed angular radius (e.g., 3°) and compare the total to the background expectation determined from right-ascension-scrambled events at the same declinations, without using disk-corona weights or spectral shapes. If the resulting p-value is substantially larger than 1.3×10⁻³ (e.g., >0.01), the reported 3.0σ excess is not robust to the model weighting. Alternatively, rerun the stacking likelihood with equal source weights and a generic E⁻² spectrum; if the p-value rises above 0.01, the central claim depends critically on the assumed disk-corona scaling.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that Seyfert galaxies contribute to the extragalactic neutrino flux rests on the stacking analysis in Sec. 5, which achieves a 3.0σ excess (p=1.3×10⁻³) using a signal PDF with each source weighted by the expected event count n_exp from the disk-corona model calibrated to NGC 1068 (Sec. 2). This is the only collective test for the Southern sample; no model-independent check (analogous to the binomial test used in the Northern-sky analysis, Abbasi et al. 2025a) is reported. If the assumed scaling between 2–10 keV X-ray luminosity and neutrino flux (or the assumed spectral shape) is incorrect, the likelihood ratio test remains a valid rejection of background, but the interpretation of the excess as emission from the selected Seyfert galaxies is compromised: the significant TS could be dominated by the model's emphasis on a few nearby sources (Circinus Galaxy, ESO 138-1, NGC 7582). The claimed consistency between the best-fit 6.7 events and the predicted 4.7 events is weak evidence because the 68% interval (3.5–10.7) is wide. Thus the evidence for a Seyfert population is contingent on an unvalidated X-ray-to-neutrino scaling.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper searches for high-energy neutrino emission from 14 X-ray bright Seyfert galaxies in the Southern Hemisphere using roughly ten years of IceCube Enhanced Starting Track Events (ESTES). Individual-source searches, assuming either a power law or the disk-corona spectrum, find no source that survives trial corrections; the best individual candidate, the Circinus Galaxy, has a post-trial significance of only 1.8σ. The main result is a stacking analysis in which each source is weighted by the number of events predicted by a disk-corona model normalized to NGC 1068. Excluding Centaurus A, the stacking analysis finds a best-fit excess of 6.7 signal events over background, with a single-sided p-value of 1.3×10⁻³ (3.0σ), which the authors interpret as new independent evidence that Seyfert galaxies contribute to the extragalactic neutrino flux.","tokens_in":16961,"tokens_out":5803,"duration_ms":70319,"significance":"If the result holds, it is a valuable addition to the growing case for Seyfert galaxies as neutrino emitters, extending the NGC 1068 discovery to a population of southern X-ray bright AGNs and demonstrating the power of the ESTES starting-track sample. The paper has real strengths: the likelihood formalism is described in detail, the background model combines data scrambling with Monte Carlo modeling of the Galactic plane, the analysis was validated before unblinding, and the authors show that a more traditional background treatment yields a reduced but still nonzero significance of 2.8σ. The main weakness is that the quoted significance is tied to a specific model weighting whose normalization is calibrated to NGC 1068, and no model-independent collective test is provided for the southern sample.","major_comments":[{"comment":"The 3.0σ stacking significance is computed for a signal hypothesis whose source weights are the disk-corona model predictions, with the model normalization fixed to the NGC 1068 flux (Sec. 1). The p-value therefore rejects background for this particular weighted combination, but it does not by itself establish that the selected Seyfert galaxies are neutrino sources if the X-ray-to-neutrino scaling is incorrect. The paper reports no model-independent collective test for the southern sample, unlike the binomial test used in the northern-sky analysis (Abbasi et al. 2025a). I recommend adding such a test (e.g., an unweighted stack or a binomial test of excess directions) or explicitly limiting the abstract's claim to the model-weighted stacking result.","section":"Secs. 4–5, Table 2"},{"comment":"The statement that the best-fit 6.7 events are 'consistent with the model prediction of 4.7 events' is presented as support for the model, but this consistency is weak and partly by construction: the model normalization was calibrated to NGC 1068, and the quoted 68% interval (3.5–10.7) is broad. Please quantify the systematic uncertainty in the predicted event rate arising from the assumed X-ray–luminosity scaling and avoid presenting the 4.7-event prediction as an independent confirmation.","section":"Sec. 5"},{"comment":"The source list is selected from the Swift-BAT/BASS catalog using a declination cut and a threshold of 0.1 predicted signal events per source, both based on the same disk-corona model. The stacking p-value is derived for this fixed list, but the paper does not state whether the threshold and source selection were fixed before unblinding or how they affect the trials. If the list or threshold was chosen after inspecting model predictions, a trials correction may be needed; if it was fixed a priori, this should be stated explicitly.","section":"Secs. 2 and 5"}],"minor_comments":[{"comment":"The p-value entries appear garbled: for example, '−0.003' should read '0.003' (and similarly for other rows), and 'ES 138-1' in the table note is a typo for 'ESO 138-1'.","section":"Table 2"},{"comment":"The declination cut 'δ < −5°' and the phrase 'the horizontal region' would benefit from a brief explanation of the geometric/background motivation, since the latter term is not defined in the text.","section":"Sec. 2"},{"comment":"The text says the stacking excess is 'driven by five galaxies–led by three nearby AGNs,' but the individual best-fit counts in Table 4 for the three leading sources (3.6, 1.7, 1.1) already sum to 6.4, and adding the two additional sources gives more than the stacked best fit. Clarify whether Table 4 lists individual-source fits rather than components of the stacking fit.","section":"Secs. 5–6"},{"comment":"The abstract states the result as 'new independent evidence that Seyfert galaxies contribute to the extragalactic flux of high-energy neutrinos.' Given the model-dependent weighting, I suggest adding a brief qualifier such as 'under disk-corona model assumptions' to accurately represent what is tested.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The paper presents a useful and generally careful analysis, and the central excess is plausible. However, the main statistical claim is more model-dependent than the abstract conveys, and the absence of a model-independent collective test weakens the population-level interpretation. I think the manuscript can be made acceptable after the requested revisions, but the shading of the claim and the trial-factor clarification are important enough to warrant a major revision rather than a minor one."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this is a credible 3σ stacking excess from 13 southern Seyferts using the new ESTES starting-track sample, but the significance is tied to the disk-corona model weights. It is evidence, not a discovery, and the paper says so.\n\nWhat is actually new: the ESTES sample (10 years, starting tracks, good southern-sky sensitivity) applied to a BASS-selected list of 14 southern Seyferts. The stacking search is new; no one has done this with this sample. The analysis is careful: hybrid data+MC background to handle the Galactic plane, source masking to keep signal out of the background, and validation showing the result only drops to 2.8σ with the traditional scrambling method. The best-fit 6.7 events are consistent with the predicted 4.7, and the excess is driven by the three highest-predicted sources (Circinus, ESO 138-1, NGC 7582), with best-fit counts correlated with predictions. That is a good sign.\n\nSoft spots: the weights come from a disk-corona model normalized to NGC 1068, so the predicted event counts are not parameter-free. If that X-ray-to-neutrino scaling is wrong, the stacking loses power and the interpretation as Seyfert emission weakens. Unlike the northern-sky analysis, they do not include a model-independent binomial test; the 3σ is the only collective test in the south, so the population claim rests on the model scaling. The source-selection threshold (0.1 predicted events) is also hand-chosen, and the trials factor from that choice is not quantified — though the list was fixed pre-unblinding, so this is a minor issue. Individual post-trial significances are only ~1.8σ for Circinus; they are consistent with but not proof of the model.\n\nThe stress-test concern does not kill the paper. The excess is in the predicted directions, and Figure 4 shows the correlation between best-fit and expected event counts. But the missing model-free cross-check means the evidence is conditional, exactly as the abstract's \"evidence\" language admits.\n\nWho this is for: anyone working on AGN neutrino emission, IceCube analyses, or the origin of the diffuse neutrino flux. It deserves a serious referee. I would send it to review and ask for a model-independent test in the revision, plus a clearer statement about source-selection trials.\n\nRecommendation: engage with it, cite it, but treat the 3σ as model-dependent evidence.","headline":"A credible but model-dependent 3σ stacking excess from southern Seyferts — worth refereeing and citing, but not a discovery on its own.","tokens_in":19850,"tokens_out":2079,"would_cite":true,"duration_ms":24882,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper reports a 3.0σ excess of neutrino events from a stacking analysis of 13 southern Seyfert galaxies, providing new evidence that Seyferts are sources of high-energy neutrinos.","keywords":["neutrino astronomy","Seyfert galaxies","active galactic nuclei","IceCube","stacking analysis","disk-corona model","starting tracks","high-energy neutrinos"],"falsifier":"A concrete check: if the 6.7 excess events do not cluster around the three predicted-brightest sources (Circinus, NGC 7582, ESO 138-1) with the predicted spectral shape, but instead are spread across the sky, the background model is likely wrong. With doubled exposure, the excess should grow toward ~5σ if real; a decline would indicate a fluctuation.","tokens_in":16557,"feed_emoji":"🔭","tokens_out":3499,"duration_ms":38546,"temperature":0.7,"pith_summary":"The paper searches roughly ten years of IceCube starting-track data for neutrinos from 14 X-ray bright Seyfert galaxies in the Southern Hemisphere. Using the disk-corona model, calibrated to NGC 1068, to predict each source's neutrino flux from its 2–10 keV X-ray luminosity, a stacking analysis finds a best-fit excess of 6.7 signal events over background, inconsistent with background at the 3.0σ level. No single source is individually significant; the excess is driven by a few predicted-brightest galaxies, led by the Circinus Galaxy. The authors interpret this as new, independent evidence that Seyfert galaxies contribute to the diffuse extragalactic neutrino flux.","feed_headline":"Seyfert galaxies show 3-sigma neutrino glow","feed_subtitle":"Stacking 13 X-ray bright Seyferts yields 6.7 excess events, supporting disk-corona models.","key_machinery":"The analysis uses the Enhanced Starting Track Event Selection (ESTES), a dataset of muon-neutrino events with interaction vertices inside the detector, which suppresses atmospheric muons and achieves ~1.4° angular resolution for Southern-sky sources. An unbinned likelihood ratio test weights each candidate source by its predicted neutrino flux from the disk-corona model, which assumes cosmic-ray acceleration in the AGN corona and is calibrated to match NGC 1068. A hybrid background—scrambled data plus Monte Carlo Galactic-plane events—accounts for recently detected Galactic neutrino emission and avoids overestimating background near bright sources.","core_discovery":"The central claim is that the collective neutrino signal from 13 southern Seyfert galaxies (excluding Centaurus A) is real, with a best-fit of 6.7 events (single-sided p = 1.3×10⁻³, corresponding to 3.0σ). This is the first stacking evidence from a Southern-Hemisphere starting-track sample and is consistent with the disk-corona model prediction of 4.7 events. The excess is driven mainly by Circinus, NGC 7582, and ESO 138-1, whose best-fit counts correlate with the model predictions.","pith_inferences":["The 3.0σ significance is below the standard 5σ discovery threshold; if the model's relative flux predictions are partly incorrect, the true source population could be more concentrated (e.g., only Circinus and ESO 138-1 contributing) than the stacking assumes.","A concrete next test is to add cascade-type events, which have complementary sensitivity in the Southern sky; a combined analysis would either strengthen the excess or reveal inconsistencies.","The apparent correlation between best-fit neutrino counts and disk-corona predictions for the three brightest sources could be a statistical coincidence; checking whether the excess grows proportionally with exposure is a decisive follow-up.","The exclusion of Centaurus A, justified by its jet ambiguity, leaves open whether jet-associated emission also produces neutrinos; a separate, jet-model-based stacking could test that."],"forward_implications":["If the excess is real, Seyfert galaxies collectively contribute to the observed diffuse neutrino flux, not just NGC 1068.","The agreement with the disk-corona prediction supports emission from an optically thick corona, explaining the missing gamma-ray counterpart.","The result motivates dedicated searches for individual Seyferts, which could reach discovery significance with additional exposure or future detectors.","Weighting sources by X-ray luminosity provides a general method for targeting other AGN populations in neutrino searches."],"fun_headline_variants":["Neutrino excess from southern Seyferts reaches 3-sigma","IceCube stacking finds 6.7 extra neutrinos from Seyferts","New neutrino signal from Seyfert galaxies at 3-sigma"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The analysis assumes the disk-corona model, with parameters fixed by NGC 1068, correctly predicts the relative neutrino fluxes of all 14 Seyferts from their 2–10 keV X-ray luminosities; if this scaling is wrong, the stacking weights are suboptimal and the 3.0σ excess may not indicate true Seyfert emission.","fun_headline_variants_meta":{"raw":{"variants":["Neutrino excess from southern Seyferts reaches 3-sigma","IceCube stacking finds 6.7 extra neutrinos from Seyferts","New neutrino signal from Seyfert galaxies at 3-sigma"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000414,"raw_usage":{"total_tokens":1973,"prompt_tokens":737,"completion_tokens":1236,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":481,"completion_tokens_details":{"reasoning_tokens":1175}},"tokens_in":481,"tokens_out":1236,"duration_ms":10786,"temperature":1.0,"reasoning_tokens":1175,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T01:13:03.200801+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check: if the 6.7 excess events do not cluster around the three predicted-brightest sources (Circinus, NGC 7582, ESO 138-1) with the predicted spectral shape, but instead are spread across the sky, the background model is likely wrong. With doubled exposure, the excess should grow toward ~5σ if real; a decline would indicate a fluctuation.","supporting_citations":[],"review_version":1}