{"id":"e2786765-e896-41b7-ab81-6dd6c727e329","arxiv_id":"2412.04071","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"AtLAST [CII] line-intensity mapping, combined with CMB-S4 and DESI BAO forecasts, could reach sigma(sum m_nu) ~ 18 meV, and ~11.5 meV with a 21-cm tau prior.","lead":"Forecasts show that a future [CII] line-intensity mapping survey like AtLAST could measure the neutrino mass sum to about 50 meV by itself, and about 18 meV when combined with CMB-S4 and DESI BAO forecasts. This paper maps which future instruments and velocity-reconstruction methods would actually move neutrino mass measurements forward.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed >5σ neutrino-mass detection rests entirely on a self-cited 21-cm τ prior; without it the forecast gives σΣmν≈18 meV, only ~3.3σ, so the headline sensitivity is prior-dominated rather than LIM-dominated.","rationale":"The paper is a careful Fisher-forecast study of [CII] LIM with velocity tomography, and the machinery (Eqs. 8–42, public GitHub code, CLASS-based growth rates) is largely explicit and reproducible. The strongest claim, however, is the >5σ detection of Σmν = 60 meV in Sec. IV C. The reader's weakest-assumption call points exactly where the argument is least secure: the τ prior from Ref. [50] is a forecast by the same first and last authors, not a measured constraint. Quantitatively, the paper's own no-τ-prior combined forecast is 18 meV, i.e., 60/18 ≈ 3.3σ; the additional factor that brings the claim to >5σ comes from that projected prior. This is not a flaw in the Fisher code or in the LIM model, but it makes the headline contingent on an external program's success. A concrete, decisive check is to rerun the combined forecast with currently available τ constraints (Planck or CMB-S4 expectations) and with a ±50% variation of the Ref. [50] prior width; if σΣmν then exceeds ~12 meV, the 5σ statement should be reworded as conditional on 21-cm achieving its forecast. This supports the reader's CONDITIONAL verdict, so no change in verdict is needed.","tokens_in":28182,"tokens_out":4333,"duration_ms":45400,"concrete_test":"Recompute the Fig. 8 Fisher pipeline with the Ref. [50] τ prior replaced by (i) no prior, (ii) the current Planck 2018 τ measurement with σ(τ)=0.007, and (iii) the CMB-S4 conservative prior from Ref. [68]. If σΣmν rises above 12 meV in cases (ii) or (iii), the >5σ claim should be explicitly conditioned on the 21-cm forecast being realized; also run a sensitivity test varying the assumed Ref. [50] prior width by ±50% to check whether the detection status flips.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central numerical claim in Sec. IV C and Fig. 8 is that combining AtLAST [CII] LIM, CMB-S4, and DESI-BAO with a τ prior 'derived using 21-cm forecasted observations' from Ref. [50] yields σΣmν ≈ 11.5 meV and hence a >5σ detection of Σmν = 60 meV. This prior is not a measured quantity; Ref. [50] is a forecast by the same first and last authors, projecting what HERA-like 21-cm observations might deliver. The paper's own no-τ-prior result is σΣmν ≈ 18 meV, which corresponds to 60/18 ≈ 3.3σ, so the difference between a 3σ bound and the advertised 5σ detection is supplied almost entirely by this external projected prior. If the 21-cm experiment does not reach the assumed prior width, or if the forecast has unaccounted systematic uncertainty, the 5σ statement fails even though the LIM analysis itself is unchanged. The issue is not that a prior is used; it is that the headline sensitivity is not robust to the one external input that is a self-projection rather than data.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents Fisher-matrix forecasts for the sum of neutrino masses, Σmν, using [CII] line-intensity mapping, alone and combined with kinetic Sunyaev-Zel'dovich and moving-lens velocity tomography, CMB-S4, and DESI BAO forecasts. The authors report that AtLAST-like [CII] LIM alone reaches σΣmν ≈ 50 meV, that velocity tomography improves this by a few percent, and that the combination of AtLAST + CMB-S4 + DESI-BAO reaches σΣmν ≈ 18 meV without a τ prior. With a τ prior derived from forecasted 21-cm observations (Ref. [50]), they report σΣmν ≈ 11.5 meV, which they translate into a ≳5σ detection of the normal-hierarchy minimum Σmν = 60 meV. The analysis is clearly specified, the code is public, and the parameter set, noise models, and window functions are described in detail.","tokens_in":28434,"tokens_out":6004,"duration_ms":66160,"significance":"If the forecasts are robust, the paper makes a useful and timely contribution by showing that a next-generation [CII] LIM experiment can match or exceed current galaxy-survey forecasts for neutrino mass constraints, and by quantifying the marginal role of velocity tomography. The strengths include a clear Fisher framework with a public code, explicit treatment of the velocity-reconstruction noise from kSZ and moving-lens effects, and an honest presentation of the no-τ-prior result of 18 meV. The main caveat is that the headline 5σ statement depends on an external, projected τ prior rather than on the LIM data themselves, so the significance of the paper's central claim is sensitive to how that prior is framed and validated.","major_comments":[{"comment":"The advertised ≳5σ sensitivity is not a LIM+CMB+BAO result but a LIM+CMB+BAO+τ(21-cm forecast) result. Without the τ prior the same combination gives σΣmν ≈ 18 meV, which for Σmν = 60 meV is only ≈3.3σ. The difference between this 3σ bound and the claimed 5σ detection is supplied almost entirely by a projected prior from Ref. [50], a forecast by the same first and last authors rather than a measured value. The paper should present the no-prior 18 meV constraint as the primary LIM-based forecast, clearly separate the 'with forecast τ prior' scenario in the abstract and conclusions, and show how σΣmν depends on the assumed prior width (e.g., a curve over σ(τ) ≈ 0.001–0.02) so the robustness of the 5σ claim can be judged.","section":"Sec. IV C, Fig. 8, Abstract"},{"comment":"The CMB-S4 noise model in Eq. (2) omits foregrounds, and the paper states only that including them 'will amount to a more conservative approach, as shown in Ref. [94].' Because the combined constraints in Fig. 8 rely on CMB-S4's large-scale sensitivity to break the τ–Σmν degeneracy, the forecast should include a quantitative foreground treatment or at least a demonstration that a reasonable foreground-residual model does not shift the 18 meV and 11.5 meV numbers substantially. Citing Ref. [94] for this point is not a substitute for showing the sensitivity in the present analysis.","section":"Sec. II B and Sec. IV C"},{"comment":"The parameter bRSD, which multiplies the f μ² term in c_X, is fixed to 1, whereas the companion galaxy-survey analysis in Ref. [94] marginalizes over it. Since the RSD term carries part of the neutrino-mass information, fixing bRSD can artificially tighten the forecasted constraints. The authors should marginalize over bRSD with a broad prior, or demonstrate explicitly that the quoted values (σΣmν ≈ 50 meV, 18 meV, and 11.5 meV) are insensitive to this choice.","section":"Sec. III C, Eq. (41), footnote 7"}],"minor_comments":[{"comment":"Reference [94] is cited with the incomplete arXiv identifier 'arXiv:24xx.xxxx'; this placeholder must be replaced with a full reference or a preprint number before publication.","section":"References, Ref. [94]"},{"comment":"The abstract says velocity tomography improves the AtLAST constraint by 4%, but Table III shows 51.0 meV → 50.0 meV with velocities alone (≈2%) and 51.0 meV → 49.0 meV with the additional 1% b∥ prior. The stated percentage should be reconciled with the table.","section":"Abstract and Table III"},{"comment":"The text refers to the 'Thompson cross-section'; the correct name is Thomson cross-section.","section":"Sec. III B 1, Eq. (24)"},{"comment":"The phrase 'kB the Boltzmann constant' lacks a space and a verb; it should read 'k_B is the Boltzmann constant' for clarity.","section":"Sec. III A, Eq. (3)"}],"recommendation":"major_revision","confidential_remarks":"The central forecasting machinery is sound and the paper is generally transparent about its assumptions. However, the headline 5σ sensitivity depends on a τ prior that comes from a forecast by the same first and last authors (Ref. [50]), and the no-prior result is only ≈3.3σ. This is a matter of framing and robustness rather than an internal inconsistency, but it is important enough that the abstract and conclusions should be adjusted and the prior dependence quantified. I also note that the companion paper Ref. [94] is cited with a placeholder identifier; given that it is from the same group and is used to justify the foreground simplification, the editor may wish to verify its status before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a serious Fisher forecast, not a toy. The new numbers are the AtLAST [CII] limits: 51 meV from the LIM auto-spectrum alone, a modest 4% improvement from velocity tomography, and 18 meV when combined with CMB-S4 and DESI-BAO. The machinery is laid out carefully, the code is public, and the authors are explicit about what they marginalize over and what they fix. That is real work, and the 18 meV combined forecast is the takeaway I would trust as a plausible projection.\n\nThe soft spot is exactly the one flagged in the stress-test note: the advertised >5-sigma detection of 60 meV hinges on a tau prior from Ref. [50], a forecast by the same first and last authors, not a measured value. Without that prior the combined forecast is 18 meV, which is about 3.3-sigma. The paper is transparent about this, and they show both numbers, but the abstract and conclusions lead with the 5-sigma statement. If the 21-cm experiment does not deliver that prior, the headline fails even though the LIM analysis is untouched. That is worth a revision, not a rejection.\n\nOther soft spots are minor and mostly acknowledged. The CMB noise model has no foregrounds; the paper says including them would be more conservative, and that is probably right, but it means the CMB-S4 part of the combined forecast is optimistic. Fixing bRSD to 1 is a simplification, and the companion galaxy velocity paper is cited only as an arXiv placeholder, which makes the comparison harder to check. The velocity-tomography improvement is small, so those assumptions do not carry the central result.\n\nThe citation pattern is fair. They cite the prior LIM neutrino forecasts and the kSZ-with-LIM work, and the self-citation in the tau prior is the only place where the circularity burden is real. Their treatment of it is honest, which counts for something.\n\nWho is this for? Anyone working on LIM forecasts, neutrino mass projections, or multi-tracer CMB-LSS combinations. It deserves a serious referee and, with a reframing that demotes the 5-sigma claim to a conditional projection, it should be publishable. I would send it out rather than desk reject.","headline":"A clean, well-documented Fisher forecast for neutrino masses with LIM; the headline 5-sigma claim leans on a self-cited 21-cm tau prior, but the no-prior 18 meV result is the solid core.","tokens_in":29032,"tokens_out":1371,"would_cite":true,"duration_ms":16331,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["98.80.Es","14.60.Pq"],"model":"deepseek-v4-flash","headline":"A multi-tracer forecast combining [CII] line-intensity maps with CMB-S4 and DESI-BAO reaches 18 meV sensitivity on the summed neutrino mass, and with a 21-cm reionization prior it crosses the 5σ threshold for the 60 meV normal-hierarchy…","keywords":["neutrino mass sum","line-intensity mapping","[CII] 158 micron line","Fisher forecast","velocity tomography","kinetic Sunyaev-Zeldovich effect","moving lens effect","optical depth to reionization"],"falsifier":"Re-run the same Fisher pipeline with $\\tau$ assigned its current measured level — Planck-era $\\sigma(\\tau)$ near 0.007–0.02, or the Planck low-$\\ell$ EE-polarization covariance the paper says behaves like a 0.0075 prior — instead of the forecasted 21-cm prior; if $\\sigma(\\sum m_\\nu)$ then lands above roughly 12 meV, the claimed $\\gtrsim 5\\sigma$ measurement of $\\sum m_\\nu = 60$ meV does not follow. A complementary check is to add a foreground-residual noise term to the CMB-S4 forecast on multipoles $\\ell \\lesssim 30$, where the current model is foreground-free.","tokens_in":27954,"feed_emoji":"🔭","tokens_out":20666,"duration_ms":169011,"temperature":0.7,"pith_summary":"The paper asks whether line-intensity mapping — collecting the total spectral-line glow of unresolved galaxies instead of cataloguing individual ones — can move cosmology toward a measurement of the absolute neutrino-mass scale. Its central claim is a forecasted sensitivity ladder: the future AtLAST [CII] survey alone reaches $\\sigma(\\sum m_\\nu) \\approx 50$ meV from its clustering power spectrum, and combining that signal with velocity tomography, CMB-S4, and DESI-BAO forecasts yields $\\sigma(\\sum m_\\nu) \\approx 18$ meV without any external prior on the reionization optical depth $\\tau$. Adding a $\\tau$ prior derived from forecasted 21-cm observations tightens the sensitivity to $\\sim 11.5$ meV, which would be a $\\gtrsim 5\\sigma$ detection of the 60 meV mass sum that is the minimum of the normal neutrino hierarchy. If the forecast holds, line-intensity mapping becomes a standalone route to the neutrino-mass measurement at redshifts ($z \\approx 1$–$8$) largely beyond the reach of galaxy surveys, while a secondary result is that kinetic Sunyaev–Zel'dovich and moving-lens velocity reconstruction improves the LIM-only number by only a few percent and adds almost nothing once CMB and BAO data are included.","feed_headline":"Line-intensity maps could weigh neutrinos at 5 sigma","feed_subtitle":"Forecast combining [CII] maps, CMB-S4 and DESI reaches 18 meV, and a 21-cm reionization prior pulls that to 11.5 meV.","key_machinery":"The argument runs on a Fisher-matrix forecast (Eq. 37) whose signal covariance couples three observables: the [CII] intensity power spectrum built from a halo model with a luminosity–mass relation (Eqs. 8–15), and the radial and transverse components of a reconstructed three-dimensional velocity field. The radial velocity comes from kinetic Sunyaev–Zel'dovich tomography — the CMB temperature shift from photons scattering off the bulk motion of free electrons — whose reconstruction noise follows from a squeezed-bispectrum estimator, a three-point statistic dominated by one long-wavelength mode (Eq. 27). The transverse velocity comes from the moving-lens effect, the temperature pattern imprinted by time-varying gravitational potentials of moving structures, with noise propagated from a reconstructed gravitational-potential field (Eq. 31). Massive neutrinos enter through the scale-dependent growth rate $f(k,z)$ in the linear velocity–density relation $v(k) \\propto (f a H/k)\\,\\delta(k)$ and through free-streaming suppression of the matter power spectrum on small scales. The degeneracy that blocks a CMB-only measurement — between $\\tau$, the amplitude $A_s$, and $\\sum m_\\nu$ — is broken in the combination by the LIM auto-power spectrum and, for the strongest claim, by the external $\\tau$ prior drawn from forecasted 21-cm observations.","core_discovery":"On its own terms the paper establishes a sensitivity ladder for [CII] line-intensity mapping. Current-generation instruments of the CCAT-prime type cannot meaningfully constrain the mass sum ($\\sigma \\approx 8$ eV); an intermediate reduced-AtLAST configuration forecasts $\\sigma \\approx 166$ meV, comparable to a ten-year VRO10 photometric survey; and the full AtLAST design reaches $\\sigma \\approx 50$ meV from the [CII] auto-power spectrum alone, improving by about 4% when the three-dimensional velocity field reconstructed from kSZ and moving-lens cross-correlations with CMB-S4 is added. Summing the Fisher matrices of AtLAST, CMB-S4, and DESI-BAO gives the paper's headline numbers: $\\sigma(\\sum m_\\nu) \\approx 18$ meV with no $\\tau$ prior, and $\\sigma \\approx 11.5$ meV once the 21-cm-forecast $\\tau$ prior is imposed — the configuration the authors associate with a $\\gtrsim 5\\sigma$ measurement of $\\sum m_\\nu = 60$ meV under the normal hierarchy. The paper further shows that velocity tomography inherits the CMB's $\\tau$–$\\sum m_\\nu$ degeneracy: it helps the LIM-only forecast (up to about 11% with a 1% prior on the radial velocity bias $b_\\parallel$) but contributes negligibly once CMB-S4 and BAO are combined in.","pith_inferences":["The same machinery transfers to other star-formation lines such as CO and [OIII]; because the constraining power rides on clustering and redshift coverage more than on the line's specific astrophysics, a multi-line Fisher combination is a natural extension that could push below the single-line 50 meV number — a step this paper does not take.","The $\\gtrsim 5\\sigma$ outcome may not need 21-cm cosmology: the paper notes that low-$\\ell$ EE polarization gives constraints comparable to a 0.0075 prior on $\\tau$, so whether an all-CMB route reaches the 11 meV regime depends on how tight a genuinely measured EE-based prior turns out to be.","The forecast restricts large-scale modes to $k \\le 0.1\\,\\mathrm{Mpc}^{-1}$ and the velocity channel weakens if pixel noise grows, so a testable stress case is to push $k$ beyond 0.1 and lengthen observing time per voxel; this would reveal whether velocity tomography can ever become the leading term rather than a few-percent accessory.","The [CII] luminosity–halo mass relation is held fixed apart from its per-bin normalization, so varying its slope and scatter, or folding in line-formation uncertainties, is the most direct stress test of the 18 meV and 11.5 meV figures."],"forward_implications":["An AtLAST-like [CII] survey alone constrains $\\sum m_\\nu$ at $\\sigma \\approx 50$ meV, enough to reject the inverted hierarchy at more than $2\\sigma$ from LIM maps alone, while probing redshifts beyond the range of optical galaxy surveys.","The combined forecast reaches $\\sigma(\\sum m_\\nu) \\approx 18$ meV with no $\\tau$ prior, well below the 0.072 eV 95% CL upper bound that the paper cites from Planck, ACT lensing, and DESI.","A $\\tau$ prior from 21-cm observations converts the 18 meV constraint into an 11.5 meV one, crossing the $\\gtrsim 5\\sigma$ threshold on $\\sum m_\\nu = 60$ meV and making reionization-era 21-cm cosmology a necessary partner for the strongest neutrino-mass claim.","Velocity tomography is a small correction rather than the engine: about 4% at full AtLAST and up to about 11% at reduced-AtLAST with a 1% prior on $b_\\parallel$, and nearly zero once CMB-S4 and BAO are included, so the kSZ and moving-lens channels mainly serve as cross-checks on the LIM auto-spectrum."],"supporting_citations":[{"why":"Supplies the squeezed-bispectrum formalism and the radial-velocity reconstruction noise (Eq. 27) that underlies the kSZ tomography forecasts.","marker":"[70]"},{"why":"Supplies the moving-lens velocity-reconstruction machinery, including the gravitational-potential noise expression (Eq. 31) and the transverse velocity bias.","marker":"[93]"},{"why":"The galaxy-survey velocity-tomography forecast of which this paper is the LIM analogue; provides the Fisher setup and comparison baseline.","marker":"[94]"},{"why":"First proposed kSZ tomography with line-intensity mapping as the large-scale-structure tracer, the starting point this paper generalises to neutrino masses.","marker":"[95]"},{"why":"Provides the forecasted 21-cm measurement of τ that drives the 11.5 meV, >5-sigma result in Sec. IV C.","marker":"[50]"},{"why":"CMB-S4 Science Book; sets the CMB experiment's noise, angular resolution, and the conservative τ prior used in Fig. 8.","marker":"[68]"},{"why":"Supplies the survey specifications (noise, sky area, observing time, redshift bins) for the CCAT-prime, AtLAST, and reduced-AtLAST configurations in Tab. I.","marker":"[32]"},{"why":"Motivates the 1% prior on the radial velocity bias b∥ via fast radio bursts, used in the LIM-plus-velocity forecasts.","marker":"[126]"},{"why":"DESI BAO observations whose Fisher forecast enters the combined Fig. 8 constraints.","marker":"[17]"}],"fun_headline_variants":["Neutrino mass forecasts from [CII] maps reach 11.5 meV","Multi-tracer LIM could hit 5 sigma on neutrino mass","AtLAST + CMB-S4 + DESI forecast neutrino mass at 18 meV","Neutrino mass detection at 5 sigma with line-intensity mapping","21-cm prior sharpens neutrino mass forecast to 11.5 meV"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The $\\gtrsim 5\\sigma$ claim rests on a $\\tau$ prior that is itself a forecast from future 21-cm observations rather than a measurement (the analysis relies on it in Sec. IV C), and the CMB forecast in the analysis is computed without foregrounds, which the paper itself notes would act conservatively.","fun_headline_variants_meta":{"raw":{"variants":["Neutrino mass forecasts from [CII] maps reach 11.5 meV","Multi-tracer LIM could hit 5 sigma on neutrino mass","AtLAST + CMB-S4 + DESI forecast neutrino mass at 18 meV","Neutrino mass detection at 5 sigma with line-intensity mapping","21-cm prior sharpens neutrino mass forecast to 11.5 meV"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00054,"raw_usage":{"total_tokens":2685,"prompt_tokens":1135,"completion_tokens":1550,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":751,"completion_tokens_details":{"reasoning_tokens":1447}},"tokens_in":751,"tokens_out":1550,"duration_ms":10786,"temperature":1.0,"reasoning_tokens":1447,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:47:46.994906+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the same Fisher pipeline with $\\tau$ assigned its current measured level — Planck-era $\\sigma(\\tau)$ near 0.007–0.02, or the Planck low-$\\ell$ EE-polarization covariance the paper says behaves like a 0.0075 prior — instead of the forecasted 21-cm prior; if $\\sigma(\\sum m_\\nu)$ then lands above roughly 12 meV, the claimed $\\gtrsim 5\\sigma$ measurement of $\\sum m_\\nu = 60$ meV does not follow. A complementary check is to add a foreground-residual noise term to the CMB-S4 forecast on multipoles $\\ell \\lesssim 30$, where the current model is foreground-free.","supporting_citations":[{"cited_title":"Optical depth to reionization from perturbative 21cm clustering","cited_arxiv_id":"2205.11504","evidence_quote":"Provides the forecasted 21-cm measurement of τ that drives the 11.5 meV, >5-sigma result in Sec. IV C."}],"review_version":1}