{"id":"a4b0dd51-3a37-4d97-a750-0729d79e2014","arxiv_id":"2506.17476","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"In the BRAHMA simulations, the most lenient heavy-seed model keeps the black hole mass-velocity dispersion relation roughly fixed from z=5 to z=0, while more restrictive seed models shift it upward at late times, a difference traceable to merger- versus accretion-driven black hole growth.","lead":"Using the BRAHMA cosmological simulations, this paper predicts that different ways of seeding supermassive black holes should leave visible marks in the relation between black hole mass and galaxy velocity dispersion, especially at high redshift. The result gives observers a new way to test how the first black holes formed using JWST measurements.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The M_bullet-sigma 'predictions' rest on a JWST-tuned Jcrit=10 J21 (Section 2.2), 100x below canonical DCBH values; if the canonical flux is required, BII-BIV seed densities collapse and BI's claimed JWST consistency becomes circular rather than predictive.","rationale":"The reader's weakest assumption is essentially the same Jcrit issue, and I agree that it is the most load-bearing concern. The paper is transparent about adopting Jcrit=10 to match JWST, so the internal comparisons among BI-BIV are reproducible and not fraudulently presented; the problem is epistemic. The claimed 'natural explanation' of JWST observations is not a prediction but a consistency check, because the seed model was tuned in prior work to produce overmassive BHs. This concern does not overturn the qualitative statement that different seeding prescriptions leave different imprints on M_bullet-sigma at fixed sigma—that follows from the abundance ladder—but it does weaken the paper's strongest observational claim and its title-level 'predictions' framing. The authors list a number of honest caveats (small volume, optimistic BH merging, post-hoc seed-mass cut in the slope fit), which further support a CONDITIONAL rather than a REJECT or ACCEPT verdict. Given the public analysis code and the falsifiable nature of the proposed Jcrit test, the reader's CONDITIONAL verdict remains appropriate; no change is recommended, though the final published version should explicitly label the JWST comparison as a consistency check of a tuned model until the Jcrit sensitivity is assessed.","tokens_in":24681,"tokens_out":9843,"duration_ms":105843,"concrete_test":"Post-process the BRAHMA halo catalogs at z=5,7,10 with the same dense, metal-poor, spin, and environment seeding criteria but replace Jcrit=10 J21 with Jcrit=1000 J21 (the canonical threshold from Shang et al. 2010; Sugimura et al. 2014). Count seeded halos in each of the four models. If the BII-BIV seed counts drop by more than two orders of magnitude or fall to O(1) objects in the [18 Mpc]^3 volume, the seed-abundance ladder—and hence the claimed z>2 M_bullet-sigma discriminants—is not representative under standard DCBH formation conditions, and the BI-JWST agreement must be reframed as a tuned consistency check rather than a prediction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central observational-facing claim—that the lenient BI model naturally explains JWST AGN being overmassive on M_bullet-M* yet on the local M_bullet-sigma—depends on a seeding abundance ladder built on Jcrit=10 J21. Section 2.2 states this value was adopted explicitly 'to produce overmassive BHs consistent with JWST,' citing dynamically heated major mergers (Regan et al. 2020; Wise et al. 2019). Canonical DCBH models require Jcrit~1000 J21 (Shang et al. 2010; Sugimura et al. 2014); the paper itself notes canonical fluxes lower the expected seed abundance by >2 orders of magnitude (Section 4.1). If the true critical flux is near canonical, BII-BIV produce almost no seeds in an [18 Mpc]^3 box, and the claimed different normalizations at z>2 across all sigma—while true in the simulations—are not a probe of any plausible seed model. More importantly, the JWST agreement is not an independent prediction: the seed model was tuned to yield overmassive BHs at z~4-7 in Bhowmick et al. (2024c). Calling the resulting match 'natural' (Section 4.1) is circular. This does not make the simulation results internally wrong, but it does mean the paper's headline conclusion overstates the evidential weight of the JWST comparison.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses four cosmological hydrodynamical simulations from the BRAHMA suite (comoving [18 Mpc]^3 boxes, TNG-based galaxy formation, heavy ~1.5e5 Msun seeds) with progressively more restrictive seeding criteria (BI, BII, BIII, BIV) to study the M_bullet-sigma relation from z=5 to z=0. It reports that the normalization of M_bullet-sigma differs across seed models at z>2 for all probed sigma, and at z=0 for sigma ~50-80 km/s; the most lenient model (BI) shows negligible redshift evolution, while more restrictive models evolve upward from z~2 to 0, mostly at high sigma ~100 km/s. The paper attributes these trends to merger-dominated BH growth in low-mass galaxies versus accretion-dominated growth in high-mass galaxies, and uses a time-derivative decomposition (Eq. 2) validated against the data (Fig. 9). It also compares to JWST broad-line AGN and argues that the BI model naturally explains the JWST overmassive population on the M_bullet-M* plane while remaining consistent with the local M_bullet-sigma relation.","tokens_in":24889,"tokens_out":5547,"duration_ms":53613,"significance":"If the claims hold, this would be one of the first cosmological simulation-based predictions that the z>2 M_bullet-sigma relation discriminates between BH seed models, and it would bound heavy-seed merger timescales. The paper's strengths include public analysis code, an explicit numerical check of the decomposition against the simulation data (Fig. 9), and bootstrapped confidence intervals in Figures 5-7. However, the central JWST comparison is not an independent test: the lenient seeding model was tuned to match JWST BH masses via the adopted low critical LW flux (Section 2.2), so the 'natural explanation' wording overstates the evidential weight of the JWST comparison. The small box volume and optimistic merger treatment also limit the robustness of the quantitative predictions.","major_comments":[{"comment":"The paper adopts Jcrit=10 J21 (Section 2.2) explicitly 'to produce overmassive BHs consistent with JWST', citing Bhowmick et al. (2024c), and Section 4.1 then states that the BI model 'offers a natural explanation' for the same JWST population, with the Conclusions repeating that current observations 'appear to require the abundant formation of heavy seeds'. This is a circular validation: the JWST M_bullet-M* agreement was built into the seed abundance. The manuscript should either (a) present the JWST comparison as a consistency check with the calibration caveat stated prominently, or (b) run an additional simulation with Jcrit at a canonical value (~1000 J21, Shang et al. 2010; Sugimura et al. 2014) to show what M_bullet-sigma would look like if the seed density were two orders of magnitude lower, as the paper itself notes in Section 4.1. Without this, the headline claim that M_bullet-sigma discriminates between seed models is not a falsifiable prediction for the JWST-era comparison.","section":"2.2, 4.1"},{"comment":"The slope-fit mass cut in Appendix B ('we perform a simple linear regression on the M_bullet-M* data at each redshift for systems with BH masses greater than 5 times the seed mass') is applied post hoc to remove the 'artificial flattening' that affects the restrictive models. This cut directly determines Component 2 of Eq. (2), and Figure 7b shows that the inversion of the slope ordering across seed models is the key driver of the M_bullet-sigma evolution. Because the cut removes the very population of ungrown seeds that the paper identifies as responsible for the increased scatter in restrictive models (Section 3.3), the decomposition may be systematically biased. Please show the sensitivity of Figures 7b and 9 to the 5x mass cut, either by repeating the analysis with a 3x or no cut, or by explicitly modeling the flattening instead of removing it.","section":"Appendix B"},{"comment":"The merger treatment is explicitly optimistic: BHs are repositioned to the nearest potential minimum and 'every halo merger wherein both halos have been seeded will also result in a prompt BH merger' (Section 2.1). The paper's central mechanism—merger-dominated BH growth in low-mass galaxies—therefore rests on the assumption of zero merger delay. Section 4.2 mentions that Bhowmick et al. (2024c) found delays <750 Myr are needed to reproduce the JWST observations, but the paper does not show how the M_bullet-sigma predictions in Figures 3-5 respond to such delays. If realistic dynamical-friction delays are longer, the separation between BI and BIV could shrink or disappear at the low-sigma end, undermining the z=0 low-sigma discriminant. Please provide a quantitative sensitivity test or clearly state that all predictions are conditional on the prompt-merger assumption.","section":"2.1, 4.2"},{"comment":"The small [18 Mpc]^3 volume leaves few seeded subhalos at the high-sigma end for the restrictive models at z>2 and, for BIV at sigma=10^1.5 km/s, the top right panel of Fig. 9 has no data while the Fig. 6 error bars grow large. The abstract's claim of 'different normalizations at higher redshifts across all sigma' is therefore based on very sparse bins for BIII and BIV. Please state the number of subhalos per sigma bin and redshift for each model (for instance, in a small table or by annotating Figures 2/3), so the reader can assess how many objects support each median and whether the differences are statistically robust. This would also clarify whether the empty bins are a volume effect or a genuine prediction.","section":"3.1, Figs. 6 and 9"}],"minor_comments":[{"comment":"The word 'nunmber' in the Introduction should be 'number'.","section":"Section 1"},{"comment":"The caption states 'very minimal redshift evolution for all seed models at sigma=10^1.5 km/s', but the panel behavior and the surrounding text indicate that the low-evolution bin is sigma=10^1.75 km/s; please correct the caption.","section":"Figure 5 caption"},{"comment":"The notation \\bar{M}_\\bullet is introduced in Eq. (2) without a definition; please state explicitly that it denotes the median logarithmic BH mass at fixed sigma.","section":"Equation (2)"},{"comment":"The phrase 'the extrapolated KH13 relation' is not defined; please specify how the extrapolation is performed and over what range of sigma it is intended to be valid.","section":"Section 4.1"},{"comment":"The kinematic-decomposition threshold of 1000 stars is applied only to TNG and not to BRAHMA; a short justification or a sensitivity test to this threshold would help the reader interpret the low-sigma end of the relations.","section":"Section 2.3"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a solid simulation study with a clear mechanistic decomposition and public analysis code, but the JWST comparison is partially circular and the 5x mass cut in Appendix B deserves close scrutiny. I recommend a major revision that requires the authors to reframe the JWST comparison as a consistency check (not an independent prediction) and to add sensitivity tests for the mass cut and merger delay. The central simulation result—different M_bullet-sigma normalizations across seed models—is likely correct within the model, but the observational-facing claims need to be re-quantified or softened."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nYou should know two things about arXiv:2506.17476. It is the first cosmological hydrodynamic simulation study I know of that isolates the effect of BH seeding prescriptions on the M_bullet-sigma relation, and the internal analysis is careful. But the headline claim that the most lenient seed model \"naturally explains\" JWST AGN on this plane is not an independent prediction: the critical Lyman-Werner flux in that model was tuned in prior work to produce exactly those overmassive BHs.\n\nThe new result is the systematic seed-model ladder. Four BRAHMA boxes with progressively stricter gas-based seeding criteria show that at z>2 the M-sigma normalization separates across all sigma, while at z=0 differences persist only at low sigma (50-80 km/s). The more restrictive models evolve bottom-up at high sigma, mostly at z<2. The paper's physical explanation—merger-driven BH growth in low-mass galaxies versus accretion-driven growth in massive ones—is plausible and internally consistent. They decompose the evolution via Eq. 2, check it against the data in Fig. 9, and provide bootstrap errors and public scripts. That is real work, and it advances the subfield.\n\nThe soft spots are real but not fatal. Jcrit=10 J21 is two orders of magnitude below the canonical ~1000 J21 for DCBH formation. The authors chose it to produce the JWST overmassive population, so the BI model's agreement with JWST on M-sigma is a consistency check, not a confirmation. They say as much in Section 2.2, but Section 4.1 still calls it \"strikingly well aligned\" and \"a natural explanation\"—that overstates the evidential weight. If the true critical flux is near canonical, the seed abundances in BII-BIV collapse, and the z>2 normalization differences are no longer a probe of any plausible seed model. The small [18 Mpc]^3 volume also gives poor statistics for the restrictive models; the bootstrapped errors are honest about that. The 5x seed mass cut in Appendix B is post hoc but justified as removing the artificial flattening from fixed seed mass, and it is secondary to the main trends. The merger prescription is optimistic; the authors flag it and are testing dynamical friction delays in follow-up work.\n\nWho gets value: anyone working on BH seeding, high-z AGN populations, or scaling relations. The internal predictions for how M-sigma evolution depends on seed abundance are worth engaging with seriously. I would send it to peer review, with the expectation that the JWST-match language be reframed as a consistency check of a tuned model rather than a validation.","headline":"First cosmological hydro comparison of BH seeding on M-sigma, with a careful decomposition; the JWST 'match' is a consistency check of a JWST-tuned Jcrit, not an independent prediction.","tokens_in":25588,"tokens_out":2939,"would_cite":true,"duration_ms":28940,"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":"Black hole seeding reshapes the black-hole–sigma relation at high redshift, the paper claims.","keywords":["supermassive black hole seeding","M_bullet-sigma relation","BRAHMA simulations","direct collapse black holes","JWST high-redshift AGN","black hole-galaxy coevolution","merger-driven black hole growth","velocity dispersion scaling"],"falsifier":"A resolved radiation-hydrodynamic simulation of atomically cooled halos that finds Jcrit near 1000 J21 even when gas is dynamically heated during mergers would remove the foundation of the BI model, since BI's seed abundance rests on Jcrit = 10 J21; alternatively, direct stellar velocity dispersion measurements showing that z ~ 5 JWST AGN lie systematically above the local M_bullet–sigma relation would falsify the predicted non-evolution of the lenient model.","tokens_in":24332,"feed_emoji":"🔭","tokens_out":6143,"duration_ms":59504,"temperature":0.7,"pith_summary":"The paper asks whether the way supermassive black hole seeds form leaves a measurable imprint on the tight relation between black hole mass and host galaxy stellar velocity dispersion, the M_bullet–$\\sigma$ relation. Using four cosmological simulations that plant heavy ~$10^{5}$ solar-mass seeds under increasingly restrictive conditions, it finds that abundant-seed models produce high black hole masses at fixed $\\sigma$ at z > 2, while restrictive models start low and catch up mostly at z < 2 in massive galaxies. The most lenient model keeps the M_bullet–$\\sigma$ relation essentially unchanged from z = 5 to z = 0, which naturally explains JWST's early black holes that look overmassive on the M_bullet–M_star plane yet sit on the local M_bullet–$\\sigma$ relation. If correct, the high-redshift M_bullet–$\\sigma$ relation becomes a direct probe of black hole seeding physics.","feed_headline":"Black hole seeding reshapes the M-sigma relation at high z","feed_subtitle":"The most abundant seed model keeps the relation frozen from z=5 to today, matching JWST's overmassive early black holes.","key_machinery":"The load-bearing machinery is the BRAHMA simulation suite, which plants 1.5e5 M_sun seeds by cumulatively stacking four gas-based criteria: dense metal-poor gas, a Lyman-Werner flux above Jcrit = 10 J21, low gas spin below the Toomre instability threshold, and a rich merger environment. The paper's explanatory engine is a derivative identity that splits the time evolution of median black hole mass at fixed sigma into the slope of the M_bullet–M_star relation times the evolution of stellar mass at fixed sigma, plus the explicit time dependence of black hole mass at fixed M_star and sigma. This decomposition, together with the merger-versus-accretion growth regimes, converts seed abundance differences into concrete predictions for the normalization, slope, and scatter of the M_bullet–sigma relation across redshift.","core_discovery":"Across the four BRAHMA seed models, ranging from the most lenient (BI) to the most restrictive (BIV), the median M_bullet–$\\sigma$ relation separates cleanly at z > 2 over the full velocity dispersion range probed, and at z = 0 it still separates for low dispersions of roughly 50–80 km/s. The most lenient model shows negligible redshift evolution of the relation, while the restrictive models show a bottom-up evolution: at fixed $\\sigma$ near 100 km/s their black hole masses are well below the local relation at z > 2 and rise steeply between z = 2 and z = 0. The paper traces this behavior to the balance between merger-dominated black hole growth in low-mass galaxies below about $10^{9}$ M_sun and accretion-dominated growth in higher-mass galaxies, and it shows that the scatter at fixed $\\sigma$ grows for restrictive models because many seeds never grow far beyond their initial mass.","pith_inferences":["Editorial inference: if the true critical Lyman-Werner flux is near 1000 J21 rather than the adopted 10 J21, the BI model likely overproduces seeds, and its JWST agreement should be treated as an upper bound on black hole mass assembly rather than evidence for that specific seed channel.","Editorial inference: the paper's low-sigma regime at z = 0 (50–80 km/s) predicts a population of barely-grown seed remnants in low-dispersion galaxies; counting black holes in such dwarfs could directly measure the seed abundance ladder.","Editorial inference: the same derivative decomposition could be applied to other scaling relations, such as M_bullet versus halo mass or M_bullet versus galaxy size, to isolate seeding signatures from feedback effects.","Editorial inference: JWST sigma values are currently inferred from gas with a correction factor near 1.3; direct stellar velocity dispersion measurements at z ~ 5 would cleanly test the predicted non-evolution of the lenient model."],"forward_implications":["At z > 2, the normalization of the M_bullet–sigma relation is a discriminant of black hole seed abundance across the full sigma range probed, so a single observed median relation can distinguish among seed models.","The most lenient seed model predicts negligible M_bullet–sigma evolution from z = 5 to z = 0, meaning high-redshift AGN should appear normal on this plane while still overmassive on the M_bullet–M_star plane.","Restrictive seed models predict that most of the rise in black hole mass at fixed sigma near 100 km/s occurs at z < 2 in galaxies above roughly 10^9 M_sun, implying rapid late-time accretion-driven assembly.","The scatter in M_bullet–sigma at fixed sigma near 100 km/s increases with seed restrictiveness, so precision scatter measurements at z = 0 can also constrain seeding.","Seed models have negligible effect on the M_star–sigma relation, so any seed-model variation in M_bullet–sigma must arise from the black hole side of the scaling relations rather than from galaxy structure changes."],"supporting_citations":[{"why":"Defines the BRAHMA simulations and the four seeding prescriptions whose seed number densities set the M_bullet–sigma normalizations studied here.","marker":"Bhowmick et al. 2024c"},{"why":"Establishes that black hole growth is merger-dominated in low-mass galaxies below ~1e9 M_sun and accretion-dominated above it, the physical split the paper uses to explain the relation's evolution.","marker":"Bhowmick et al. 2025"},{"why":"Supplies the local M_bullet–sigma and M_bullet–M_star relations against which the simulations' redshift evolution is compared.","marker":"Kormendy & Ho 2013"},{"why":"Provides the JWST high-redshift AGN black hole masses used to test which seed model matches observations on the M_bullet–sigma plane.","marker":"Maiolino et al. 2024b"},{"why":"Provides high-redshift AGN velocity dispersion estimates and shows they sit near the local M_bullet–sigma relation, the observational anchor for the BI model's non-evolution.","marker":"Juodzbalis et al. 2025"},{"why":"Details the implementation of the Lyman-Werner flux and gas spin seeding criteria that define the BII through BIV models.","marker":"Bhowmick et al. 2022a"},{"why":"Used to justify the low critical Lyman-Werner flux Jcrit = 10 J21 via dynamical heating during major mergers.","marker":"Regan et al. 2020b,a"},{"why":"Supplies the Toomre-stability spin criterion, lambda < lambda_max, that defines the low-spin seeding threshold for BIII and BIV.","marker":"Lodato & Natarajan 2006"}],"fun_headline_variants":["BRAHMA simulations tie BH seed models to M-sigma offsets","Seed model sets M-sigma normalization at high z","Restrictive BH seeds drive upward M-sigma evolution","JWST overmassive BHs echo lenient seed scenario","M-sigma relation splits by seeding efficiency at z>2"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that heavy ~$10^{5}$ M_sun seeds form abundantly when gas is dynamically heated during major mergers, so the critical Lyman-Werner flux is only 10 J21 rather than the canonical ~1000 J21; if the true threshold is much higher, the lenient seed model overproduces seeds and the claimed match to JWST collapses.","fun_headline_variants_meta":{"raw":{"variants":["BRAHMA simulations tie BH seed models to M-sigma offsets","Seed model sets M-sigma normalization at high z","Restrictive BH seeds drive upward M-sigma evolution","JWST overmassive BHs echo lenient seed scenario","M-sigma relation splits by seeding efficiency at z>2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000618,"raw_usage":{"total_tokens":2970,"prompt_tokens":1150,"completion_tokens":1820,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":766,"completion_tokens_details":{"reasoning_tokens":1736}},"tokens_in":766,"tokens_out":1820,"duration_ms":15737,"temperature":1.0,"reasoning_tokens":1736,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:09:30.472843+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A resolved radiation-hydrodynamic simulation of atomically cooled halos that finds Jcrit near 1000 J21 even when gas is dynamically heated during mergers would remove the foundation of the BI model, since BI's seed abundance rests on Jcrit = 10 J21; alternatively, direct stellar velocity dispersion measurements showing that z ~ 5 JWST AGN lie systematically above the local M_bullet–sigma relation would falsify the predicted non-evolution of the lenient model.","supporting_citations":[],"review_version":1}