{"id":"936dde87-f979-4545-8cd3-c1be4893e7e4","arxiv_id":"2607.23020","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"PhotoIFU maps stellar mass, star-formation rate, dust, and metallicity pixel-by-pixel from JWST imaging, and finds gas-outflow regions are distinct from host galaxies in three z≈1.3–3.7 systems.","lead":"This paper turns JWST/NIRCam images into low-resolution maps of galaxy properties, pixel by pixel, to compare gas outflow regions with the rest of their host galaxies. It applies the method to three distant galaxies and finds that outflow-related regions stand out in dust, recent star formation, and metallicity.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The reported distinctness of gas-selected regions is not yet isolated from two measurement artifacts: line-sensitive bands used to select the regions also enter the SED fits, and the one neutral-outflow aperture sits in the PSF/light-concentration radial pattern the authors themselves flag.","rationale":"The reader's conditional verdict is appropriate; my stress test sharpens the reason. The weakest point is not merely that neighboring pixels are correlated, which the paper explicitly concedes, but that the region definition and the property estimation share the same photometric channels. For the Cosmic Cigar and Cosmic Rose, the selection filters are direct inputs to Prospector; for the Cosmic Ember, the selected aperture overlaps the PSF/light-concentration artifact cited by the authors. Neither the GA-NIFS benchmark nor the GMM consistency check addresses this: the benchmark is on a different source without a gas-selected region comparison, and the GMM uses the same fitted properties, so it cannot detect a shared-band bias. The proposed refit/control test is a single, decisive way to determine whether the central assertion is an artifact. The paper remains a useful pilot, and the limitations are unusually explicit, but the central claim should not be regarded as fully established until this check is run. I therefore keep the reader's conditional verdict rather than moving to acceptance or rejection.","tokens_in":24258,"tokens_out":12879,"duration_ms":146990,"concrete_test":"For each target, rerun the exact PhotoIFU fit on the same pixels with the selection-coupled filters removed (Cosmic Cigar: drop F210M; Cosmic Rose: drop F300M; retain F090W-F200W and F277W-F444W continuum bands), and recompute the pooled A-region versus global offsets in Dust2, log sSFR, log Zgas, and log(SFR0/SFR1). For the Cosmic Ember, additionally compare the A aperture with a control aperture at the same radius and matched S/N but not associated with the gas signature, to separate PSF/central-concentration effects. If the offsets persist at more than half their reported magnitude, the concern is resolved; if they vanish or change sign, the reported distinctness is dominated by the shared photometric bands or the PSF artifact.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The core claim requires that the SED-property differences of gas-selected regions be intrinsic to the host galaxy rather than produced by the measurement/analysis. For the two extended-emission systems this is not yet established: the detection/selection filters are F210M (Cosmic Cigar) and F300M (Cosmic Rose), the same line-sensitive bands used to identify the extended emission (Section 3.1; Zhu et al. 2025), and Section 3.2 fits those bands with a Prospector model whose nebular-emission component maps line excess into Dust2, Zgas, SFR0, and SFR1. A pixel selected because F210M/F300M is elevated will therefore be driven toward lower dust and more recent star formation even with an identical continuum population; the reported A-region offsets (Delta Dust2 = -0.24, -0.07) point in exactly this direction. For the Cosmic Ember, the sole neutral-outflow case, the A region is at 0.4 kpc from the center, and the authors state that the F444W-matched PSF and central light concentration produce a weak radial pattern (Figure 2 caption); the reported Delta Dust2 = +1.04 and Delta log sSFR = +0.68 are in the region where that artifact is largest. No system in the sample provides a region comparison that is clean of both effects. The Appendix A NIRSpec-IFU benchmark validates broad spatial structure but does not test the gas-selected-region comparison or the shared-band circularity.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents PhotoIFU, a workflow that treats PSF-matched JWST/NIRCam multi-band imaging as a low-resolution photometric integral field unit. Each spatial pixel is fit independently with Prospector (accelerated by the Parrot emulator), producing resolved maps of stellar-mass surface density, sSFR, Dust2, gas-phase metallicity, and the recent-to-past star-formation ratio. The method is applied to three z=1.3–3.7 JADES galaxies with known gas signatures: two systems with extended ionized emission (Cosmic Cigar and Cosmic Rose) and one post-starburst galaxy with a strong NaD neutral outflow (Cosmic Ember). The central empirical claim is that manually selected regions associated with the gas signatures occupy distinct parts of the resolved SED-property distribution compared with the full host. A Gaussian-mixture clustering analysis is included as a descriptive check, and an external benchmark against NIRSpec IFU maps of GS5001 is presented in Appendix A. An illustrative MACER simulation is used to argue that a fossil outflow and compact central rejuvenation can coexist, providing a qualitative scenario for the Cosmic Ember.","tokens_in":24701,"tokens_out":3161,"duration_ms":32739,"significance":"If the central differential claim survives scrutiny, the paper delivers a practical, broadly applicable method for connecting feedback-related gas signatures to resolved stellar-population and ISM properties in samples far larger than those accessible to NIRSpec IFU. The manuscript is commendably explicit about several limitations: the pixels are PSF-correlated, the p-values are descriptive, and the maps are differential rather than deconvolved. The external benchmark in Appendix A and the planned public release of the workflow are strengths. However, the main astrophysical conclusion — that gas-selected regions are physically distinct in dust, recent star formation, and metallicity — is not yet isolated from two measurement channels: the line-sensitive bands used for region selection also enter the SED fits, and the sole neutral-outflow case has its key aperture in the PSF/light-concentration pattern the authors themselves flag. The benchmark validates broad spatial structure but does not test either artifact channel for the region-comparison analysis.","major_comments":[{"comment":"Circularity between region selection and SED fitting is load-bearing for the 'less dusty along extended emission' claim. The detection filters for the Cosmic Cigar and Cosmic Rose are F210M and F300M, respectively — the same line-sensitive bands used to identify the extended emission in Zhu et al. (2025). These bands enter the Prospector fits, whose nebular-emission component maps line excess into Dust2, Zgas, SFR0, and SFR1. A pixel selected because F210M/F300M is elevated is therefore driven toward lower dust and more recent star formation even with an identical continuum population; the reported A-region offsets (ΔDust2 = −0.24 and −0.07) point in exactly that direction. Please demonstrate with a control — e.g., refitting with the detection/line-sensitive bands excluded, or injecting a line-excess-only signal into identical continuum SEDs and showing the region offsets disappear — or","section":"§3.1, §3.2, §4.4"},{"comment":"The Cosmic Ember's A region is the only neutral-outflow constraint and its centroid is at 0.4 kpc from the adopted center. The paper states that the F444W-matched PSF and strong central light concentration produce a weak radial pattern and that compact gradients are upper limits. The reported A-region offsets (ΔDust2 = +1.04, Δlog sSFR = +0.68) lie in the region where this artifact is largest. A PSF-convolved null test is needed: e.g., fit a smooth radial profile to the galaxy, PSF-match it to F444W, run the same pixel SED fits, and compare an aperture-matched radial baseline to the observed A-region offsets. Without such a test, the Cosmic Ember result cannot yet be distinguished from a central light-concentration artifact.","section":"§3.2, Figure 2 caption, §4.3"},{"comment":"The Mann–Whitney U p-values are explicitly descriptive and the text correctly notes that PSF-correlated pixels are not independent. However, the paper's central claim is that regions 'occupy distinct parts' of the distribution, and the only quantitative support offered is these p-values plus median offsets. Please add a statistic that accounts for the PSF correlation — e.g., a spatial block bootstrap, a test on independent resolution elements after binning to the F444W beam, or a covariance-aware effect size. This is needed to distinguish a genuine distributional offset from oversampling of a few correlated structures.","section":"§4.4"},{"comment":"The GS5001 benchmark demonstrates that PhotoIFU reproduces broad spatial structure seen by NIRSpec IFU (dust lane, SFR peaks, metallicity gradients). It does not test the gas-selected-region comparison, the manual apertures, or the shared-band circularity described above. Please state this limitation explicitly in the main text, and consider adding a benchmark region comparison on GS5001's known IFU features if possible. The current wording ('broad spatial consistency') is accurate but may lead readers to over-interpret the benchmark as validating the paper's central differential claim.","section":"Appendix A"}],"minor_comments":[{"comment":"The title has a typo ('F eedback') and there are missing spaces in the text ('Prospectorto', 'Parrotemulator'). A full proofreading pass is needed.","section":"Title / throughout"},{"comment":"The priors for the Prospector parameters are not stated. Since the paper emphasizes that the same model and priors are applied to every pixel, please list or reference the prior definitions (e.g., the Prospector-α style priors) so that the differential interpretation is reproducible.","section":"§3.2"},{"comment":"The GMM uses a fixed K=6 for all targets. A brief sensitivity check (K=4–8) or a statement of why K=6 is robust would strengthen the descriptive clustering claim.","section":"Figure 3 / §4.2"},{"comment":"The manual apertures are shown in figures but their exact boundaries and sizes are not tabulated. For reproducibility, please provide an aperture coordinate/size table or machine-readable region files.","section":"§4.3 / Figures 4–5"},{"comment":"The MACER comparison is explicitly illustrative, but the choice of ν=0.55 is not discussed. A sentence on how the fiducial run was selected among the MACER viscosity suite would help place the qualitative match in context.","section":"§5.3"}],"recommendation":"major_revision","confidential_remarks":"The paper is a pilot study with a plausible and potentially useful methodology, and the authors are appropriately cautious in many places. The blocking issues are the two artifact channels identified in the major comments: shared-band selection/fitting circularity and the PSF/light-concentration pattern in the sole neutral-outflow case. Both are fixable within the manuscript's scope through control experiments or explicit reframing of the claims. The MACER section is somewhat tangential but can remain as an illustrative scenario. The citation pattern is heavy on the authors' own group's work, but that is typical for a JADES methodology paper and does not affect my assessment."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"PhotoIFU is a genuinely useful idea: treat deep NIRCam imaging as a resolved SED mapper for large samples where NIRSpec IFU coverage is impossible. The paper is transparent about its limitations, ships a public workflow, and the GS5001 comparison against NIRSpec IFU maps is a solid sanity check. The three case studies are well-illustrated and the region definitions are cleanly described.\n\nThe central result—gas-selected regions occupy distinct parts of the host SED-property distribution—is not yet isolated from two artifacts the paper itself flags. For the two extended-emission systems, the detection filters are the line-sensitive F210M and F300M, and those same bands enter the per-pixel SED fits. Since Prospector's nebular-emission component maps line excess into SFR and Dust2, a pixel selected because its line-sensitive band is elevated will tend to come out younger and less dusty even if the continuum is identical. The reported ΔDust2 of -0.24 and -0.07 go in exactly that direction. The authors never test what happens if you fit the SEDs without those bands or select on a continuum filter.\n\nFor the Cosmic Ember, the A region sits 0.4 kpc from the center, right where the F444W-matched PSF and central light concentration create the radial pattern the Figure 2 caption warns about. The large offsets (ΔDust2=+1.04, Δlog sSFR=+0.68) are in that same region. The caption says compact gradients should be interpreted as upper limits, but the region comparison still uses those PSF-matched pixels as if they were physical.\n\nThese are not fatal objections. The trend toward lower dust along ionized-emission channels could be real physics—you only see the line where the column is low. But the paper cannot distinguish that from a measurement selection effect, and that distinction is the whole point of the claim. The Mann-Whitney p-values are explicitly descriptive because of pixel correlation, and the sample is three hand-picked systems. The MACER simulation is illustrative, not a quantitative match.\n\nWho should read this? Anyone building resolved stellar-population tools for JWST imaging surveys; the method and code are worth studying. But the empirical headline should be treated as a hypothesis. I would send this to a serious referee rather than desk reject, with a request for mock SED tests or a fit without the line-sensitive bands to break the degeneracy.","headline":"A well-documented pilot of a useful idea, but the central claim is not yet clean of selection and PSF artifacts; worth refereeing with a request for robustness tests.","tokens_in":25289,"tokens_out":7231,"would_cite":false,"duration_ms":71275,"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":"Deep multi-band JWST images can be read as a low-resolution integral field unit: in three galaxies at z ≈ 1.3–3.7, regions tied to outflows and extended ionized emission occupy distinct parts of the resolved stellar-population and ISM-prope","keywords":["high-redshift galaxies","galactic feedback","galactic outflows","pixel SED fitting","resolved stellar populations","dust attenuation","post-starburst galaxies","JWST NIRCam"],"falsifier":"Construct a synthetic galaxy with a constant stellar population and dust-free ISM, put it through the same PSF matching, pixel SED fitting, and aperture comparison; if gas-shaped apertures still show the same offsets from the host distribution, the result is an artifact of the point-spread function rather than a physical association.","tokens_in":24183,"feed_emoji":"🔭","tokens_out":7048,"duration_ms":73937,"temperature":0.7,"pith_summary":"The paper tries to show that the gas signatures of galactic feedback — extended ionized emission and a powerful neutral outflow — are spatially associated with measurable differences in the host galaxy's resolved stellar populations and interstellar medium. To do this, it introduces PhotoIFU, a workflow that treats every resolved pixel of deep, point-spread-function-matched NIRCam imaging as a coarse spectral energy distribution and fits each pixel with a full stellar-population model. Applied to three galaxies, the workflow produces maps of stellar-mass surface density, specific star formation rate, dust attenuation, gas-phase metallicity, and recent star-formation history. In all three systems, regions selected from the gas signatures occupy distinct parts of the host's resolved SED-property distribution: the ionized-emission regions tend to be less dusty, while the post-starburst outflow system shows enhanced recent star formation. If this holds, feedback's spatial imprint can be mapped with imaging alone, extending resolved studies to samples far beyond what spectroscopic integral-field observations can cover.","feed_headline":"Gas outflows leave distinct signatures in pixel-level starlight maps","feed_subtitle":"Treating every NIRCam pixel's colors as a coarse spectrum links outflow regions to dust and star-formation differences.","key_machinery":"PhotoIFU is the central mechanism: a workflow that treats each 0.03-arcsecond pixel of PSF-matched multi-band NIRCam imaging as an independent coarse spectral energy distribution. Each pixel's fluxes in thirteen filters from F090W to F444W are fit with a Bayesian SED model that includes a flexible non-parametric star-formation history, nebular emission, dust attenuation, and gas-phase metallicity; the resulting per-pixel posterior medians become maps of physical properties. Because all filters are blurred to the same point-spread function before fitting, these maps are used only for differential comparisons within a target. The workflow completes with two comparisons: an unsupervised cluster","core_discovery":"The paper's central claim is that gas signatures of feedback are spatially associated with distinguishable stellar-population and ISM properties within their host galaxies. By fitting the coarse spectrum of every resolved pixel in point-spread-function-matched NIRCam images, it maps projected stellar-mass density, specific star formation rate, dust attenuation, gas-phase metallicity, and the ratio of star formation in the last 30 million years to the preceding 30–65 million years. In all three galaxies studied, pixels inside gas-selected apertures sit in distinct regions of this property distribution: the extended ionized-emission regions are generally less dusty, and the post-starburst gala","pith_inferences":["If dust-poor channels are what make extended line emission visible, then samples of galaxies with extended ionized emission are biased toward low-attenuation sight lines; population statistics of ionizing-photon escape would inherit that bias.","A natural next test is to run the same region-versus-host comparison on archival multi-band imaging of a larger galaxy sample; if offsets do not appear for most outflow hosts, the association may not be universal.","Forward modeling or deconvolution of the point-spread function, which the paper leaves to future work, would reveal whether the radial artifact seen in one target contributes to the measured region offsets."],"forward_implications":["Resolved stellar-population and ISM maps can be produced for entire galaxies from deep imaging alone, extending feedback studies to samples far larger than spectroscopic integral-field surveys can reach.","Extended ionized emission in high-redshift galaxies marks dust-poor, low-column-density channels, not simply the densest star-forming gas.","A post-starburst galaxy can show compact, recent star formation alongside a fossil neutral outflow, consistent with a rejuvenation episode after an earlier quenching event.","The same maps can identify where physical contrasts are strongest, providing a target list for follow-up spectroscopy.","The imaging-based maps reproduce the main dust, star-formation, and metallicity patterns of a galaxy that already has NIRSpec IFU maps."],"fun_headline_variants":["Pixel SEDs tie outflows to dust and starbirth","Outflow regions show distinct pixel-level star histories","Mapping feedback with per-pixel NIRCam SEDs","Pixel colors expose dust-poor outflow channels"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"After all filters are blurred to a common resolution, each pixel's remaining color differences reflect true variations in stars and interstellar matter rather than artifacts of that blurring — the paper itself notes one target shows a radial pattern tied to the point-spread function.","fun_headline_variants_meta":{"raw":{"variants":["Pixel SEDs tie outflows to dust and starbirth","Outflow regions show distinct pixel-level star histories","Mapping feedback with per-pixel NIRCam SEDs","Pixel colors expose dust-poor outflow channels"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000566,"raw_usage":{"total_tokens":2555,"prompt_tokens":813,"completion_tokens":1742,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":557,"completion_tokens_details":{"reasoning_tokens":1677}},"tokens_in":557,"tokens_out":1742,"duration_ms":13301,"temperature":1.0,"reasoning_tokens":1677,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T03:49:57.616092+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a synthetic galaxy with a constant stellar population and dust-free ISM, put it through the same PSF matching, pixel SED fitting, and aperture comparison; if gas-shaped apertures still show the same offsets from the host distribution, the result is an artifact of the point-spread function rather than a physical association.","supporting_citations":[],"review_version":1}