{"id":"50f20618-ab12-4730-abf1-7a99a285c6f1","arxiv_id":"2411.19518","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Stacking of 919 little red dots yields no 1.4 or 3 GHz radio emission above 3 sigma, placing them among the most radio-quiet high-redshift sources.","lead":"Astronomers stacked radio images of 919 'little red dots' and found no emission down to about 10 to 18 microjanskys. This suggests these puzzling high-redshift objects host weaker or no radio-loud active galaxies.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified; the empty-field selection does not exclude LRD fields, so the reader's concern is mitigated.","rationale":"The reader identified the empty-field criterion as the weakest assumption, but a close reading of Section 3 shows that the paper's own numbers indicate all 919 LRD positions have empty-field VLASS maps in the first two epochs. Therefore no brightest LRD fields are discarded by the SNR>6 threshold, and the stacked upper limits apply to the full sample. The central claim of non-detection is supported by the data: both mean and median stacks give SNR below 3, and the upper limits are sensible. The comparison to the radio-quiet AGN sample is interpretative but properly hedged. The paper has standard stacking methodology and public data. No load-bearing methodological flaw emerges, so the verdict remains unchanged.","tokens_in":11577,"tokens_out":20608,"duration_ms":183687,"concrete_test":"Verify the empty-field counts by re-running the VLASS query for all 919 LRD coordinates in epochs 1 and 2 without the SNR>6 filter, and confirm that every LRD position has at least one empty-field cutout (i.e., no LRD is excluded). If any exclusion is found, recompute the stacked 3σ upper limit including all cutouts with a clipped mean to handle outliers; if the upper limit changes by more than ~20%, the selection bias would be real.","verdict_should_be":"UNCHANGED","load_bearing_attack":"No load-bearing objection identified. The reader's concern is that the empty-field criterion (SNR>6 in the central 15″×15″ region) excludes bright LRD fields and biases the stacked upper limits. This concern is mitigated by the paper's own reported counts: it finds 1015, 1015, and 959 empty-field VLASS maps centred on 919, 919, and 872 individual LRD positions for epochs 1, 2, and 3, respectively. Since all 919 LRDs have empty-field maps in epochs 1 and 2, no LRD field was discarded by the SNR>6 criterion; the full sample is included in those epochs. The epoch-3 shortfall of 47 positions is attributed to incomplete survey coverage, not to non-empty fields. The independent cross-match finding zero radio counterparts in the VLASS/FIRST catalogues further confirms that no LRD has SNR>6 in the relevant images. Thus the derived upper limits apply to the complete LRD sample as claimed. The only residual ambiguity is that the paper does not explicitly state the number of fields rejected by the empty-field criterion, which is a clarity point rather than a correctness flaw.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper compiles a sample of 919 little red dots (LRDs) from 17 literature sources, cross-matches their positions with the VLASS (3 GHz) and FIRST (1.4 GHz) radio surveys, and finds no radio counterparts. The authors then perform mean and median image stacking of empty-field cutouts around the LRD positions. No radio emission is detected in the stacked images, yielding 3σ upper limits of approximately 11 μJy beam^-1 at 3 GHz and 18 μJy beam^-1 at 1.4 GHz (from the median stacking). These translate to characteristic monochromatic power upper limits of roughly 1.8 × 10^24 W Hz^-1 and 2.9 × 10^24 W Hz^-1 at the median redshift of the sample (z ≈ 7.1). The non-detections are compared with a stacked sample of high-redshift radio-quiet AGNs that shows ~30 μJy flux densities, suggesting that LRDs are either very radio-quiet or have sub-dominant AGN emission. The authors also translate the 1.4 GHz upper limit to star formation rate limits of ~350–650 M_sun/yr.","tokens_in":11708,"tokens_out":14368,"duration_ms":118339,"significance":"The result is an important empirical constraint on the radio properties of the recently discovered LRD population. If the upper limits are robust, they indicate that LRDs are among the most radio-quiet high-redshift sources known, which has implications for the AGN content and star formation in these objects. The paper makes use of public survey data and a standard stacking methodology; the source catalog is provided as machine-readable supplementary material, and the analysis is reproducible. The comparison with the earlier stacking of radio-quiet AGNs provides a useful benchmark, though the interpretation is appropriately cautious.","major_comments":[{"comment":"The paper compiles 919 LRDs from 17 papers but does not describe any procedure for identifying and removing duplicate sources across the different literature samples. Several of the cited papers are based on overlapping JWST survey fields (e.g., COSMOS-Web, CEERS, UNCOVER), so the same LRD may be independently selected in more than one paper. If duplicates remain in the sample, the stacking is not an average over 919 independent LRDs, and the quoted upper limits would not be representative of the LRD population as a whole; the effective sample size could be smaller than stated. The authors should either confirm that they removed duplicates (e.g., by positional cross-matching within their own catalog) or quantify how many of the 919 entries are likely to be the same object, and discuss the impact on the results.","section":"Section 2"},{"comment":"The empty-field selection criterion (SNR less than 6 in the central 15 arcsec by 15 arcsec region) is a potential source of bias if some LRD fields were excluded because they contain radio emission. The reported counts of 1015, 1015, and 959 maps centred on 919, 919, and 872 positions suggest that all positions have at least one empty-field map in epochs 1 and 2, which mitigates the concern; however, the paper does not explicitly state the number of cutouts rejected by the criterion, nor the number of positions lacking empty-field maps in epoch 3 and the reason. Please provide these statistics to demonstrate that the upper limits apply to the full sample.","section":"Section 3"}],"minor_comments":[{"comment":"Please specify the assumed radio spectral index and the k-correction used when converting the flux density upper limits to the characteristic monochromatic powers quoted in Section 4, as these details are needed to reproduce the calculation.","section":"Section 3"},{"comment":"The abstract quotes 3σ upper limits of approximately 11 and 18 μJy beam^-1; these values correspond to the median-stacked images (Table 1). Please state explicitly in the text that the quoted limits are derived from the median stacking, and give the corresponding mean-stack limits for completeness.","section":"Abstract and Section 3"},{"comment":"There are several typos, for example 'applyed' in Section 4, 'identifed' in Section 4 and Appendix A, and 'different' appears in various places. A thorough proofread is recommended.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is suitable for A&A Letters in scope. The central stacking analysis is sound, and the main issue is the lack of explicit deduplication and empty-field rejection statistics. If the authors confirm that duplicates were removed and provide the rejection numbers, the paper would be acceptable. The comparison sample is from the same research group, which is acceptable as a benchmark."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline: this is the first systematic radio stacking constraint on little red dots, and the null result is real. I think you should read it if you work on high-redshift AGN or JWST-selected populations.\n\nWhat's new: 919 LRDs collected from the literature, cross-matched to VLASS and FIRST catalogues, no individual counterpart, then mean and median stacking of 2989 VLASS and 770 FIRST cutouts. The 3-sigma upper limits are ~11 and ~18 microJy/beam, translating to power limits below 1.8e24 and 2.9e24 W/Hz at the median redshift. The stacking methodology is standard and the noise estimates look right. The redshift-binned stacking in the appendix is a nice touch; it shows the non-detection is consistent across z.\n\nThe paper also does an honest job with the two apparent radio associations in FIRST: the appendix shows both are foreground sources, one a lensing system, one a cluster. That's the kind of care that makes me trust the cross-match.\n\nThe reader's main worry was the empty-field selection criterion (SNR>6 excluded). I looked at the counts and that concern doesn't land. Epochs 1 and 2 have empty-field maps for all 919 LRDs, so no field was actually discarded by that criterion. The epoch 3 shortfall (959 vs 919) is attributed to incomplete survey coverage, not to rejection. And the catalogue cross-match found zero LRD counterparts, consistent with nothing being removed. So the upper limits really do apply to the full sample as claimed.\n\nSoft spots: the comparison to the earlier stacked radio-quiet AGN sample (Perger et al. 2024) is used as a benchmark, and the conclusion about 'particularly radio-quiet' leans on that comparison. The two samples are selected differently and the AGN sample has spectroscopic redshifts, so treat that part as suggestive rather than definitive. The SFR upper limits are also not very constraining (hundreds of solar masses per year) — they don't rule out a starburst explanation. But those are minor; the central non-detection stands on its own.\n\nWho it's for: anyone working on LRD demographics, radio-quiet AGN, or high-redshift stacking techniques. It's a short letter, well-suited to A&A. I'd send it to a referee; the method is sound and the result is the first of its kind. The reader's conditional verdict is fair if it just asks for a clarifying sentence about the empty-field counts, but I don't think there's a substantive flaw.","headline":"A clean, honest null result: the first radio stacking of 919 little red dots gives real upper limits, and the reader's empty-field worry doesn't survive contact with the paper's own counts.","tokens_in":12286,"tokens_out":2319,"would_cite":true,"duration_ms":19067,"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":"Stacking thousands of radio cutouts finds no emission from 919 little red dots, implying weak or absent radio AGN.","keywords":["little red dots","high-redshift galaxies","radio continuum","VLASS","FIRST","image stacking","active galactic nuclei","star formation"],"falsifier":"A deeper 3 GHz image of the same 919 fields reaching rms noise below about 1 microjansky per beam that shows a stacked peak above 3 sigma, or even a single LRD with a secure VLASS or FIRST counterpart at signal-to-noise above 5, would contradict the conclusion that LRDs are radio-silent.","tokens_in":11332,"feed_emoji":"📡","tokens_out":6108,"duration_ms":49168,"temperature":0.7,"pith_summary":"The paper tests whether little red dots (LRDs), a compact, red, high-redshift galaxy population of debated nature, host radio-emitting active galactic nuclei. It compiles 919 LRDs from the literature and searches for radio emission in the FIRST and VLASS surveys, both by cross-matching individual positions and by stacking empty-field image cutouts. Neither the individual cross-match nor the mean and median stacked images yields a detection above the 3σ noise level, placing 3-GHz and 1.4-GHz characteristic power upper limits below about $1.8\\times10^{24}$ and $2.9\\times10^{24}$ W Hz$^{-1}$ at the sample's median redshift of 7.1. The authors conclude that LRDs host weaker or no radio AGN, or are a composite population in which AGN activity is sub-dominant, and they translate the limits into star formation rates below roughly 350–650 solar masses per year.","feed_headline":"Little red dots show no radio emission in stacked maps","feed_subtitle":"Stacking 3,759 survey cutouts places 3-GHz power below 1.8e24 W/Hz, favoring weak or absent AGN.","key_machinery":"The central mechanism is pixel-by-pixel image stacking of empty-field radio cutouts. An empty field is defined as a $15''\\times15''$ region around the LRD whose peak signal-to-noise ratio is $\\lesssim6$ in the raw map; cutouts passing this filter are aligned on the LRD coordinates and combined by both mean and median averaging. Stacking roughly 3000 independent maps reduces the rms noise to about $3\\,\\mu$Jy beam$^{-1}$ at 3 GHz, which is what converts non-detections into meaningful upper limits. The complementary machinery is the cross-match with catalogued radio sources, used to show that no individual LRD has a secure counterpart and that the two apparent associations are foreground contamination.","core_discovery":"The central claim is that, as a population, little red dots are radio-silent to the limits of current wide-area surveys. After assembling 919 LRDs at declinations above $-40\\degree$, the authors find no VLASS or FIRST counterpart within $5''$ that can be securely associated with an LRD; the two radio sources found nearby belong to foreground lensing systems and a galaxy cluster. Stacking 2989 VLASS Quick Look cutouts and 770 FIRST cutouts, they detect no emission above $3\\sigma$ noise ($\\sim11$ and $\\sim18\\,\\mu$Jy beam$^{-1}$ for VLASS and FIRST respectively), yielding unresolved-source flux density upper limits of $8.8\\!-\\!10.8\\,\\mu$Jy at 3 GHz and $13.3\\!-\\!17.7\\,\\mu$Jy at 1.4 GHz. At the median redshift of 7.1 these correspond to characteristic monochromatic powers below $1.8\\times10^{24}$ W Hz$^{-1}$ and $2.9\\times10^{24}$ W Hz$^{-1}$, well below the $\\sim30\\,\\mu$Jy level recovered from stacked samples of high-redshift radio-quiet AGNs. Redshift-binned stacking likewise gives no detection, and the two apparently nearby FIRST sources are shown to be unrelated foreground objects.","pith_inferences":["If LRDs are heavily obscured AGNs, their radio silence pairs with the known X-ray deficit to suggest accretion that is intrinsically radio-weak or absorbed in a way that suppresses both bands, rather than merely being hidden by dust.","The non-detection may indicate that the LRD phase is short-lived relative to radio-loud AGN episodes, or that typical black hole masses in LRDs are low; deeper observations could test this by bounding radio luminosity per unit black hole mass.","A testable extension is to apply the same stacking method to JWST-selected samples with more complete spectroscopic coverage, which would separate the radio upper limit in narrower redshift bins and could reveal a population that turns on at later cosmic epochs.","The two foreground radio sources near LRDs show that contamination by unrelated structure is a real risk at arcsecond scales, so simulations of confusion noise could set a floor on how deep this stacking approach can go in crowded fields."],"forward_implications":["LRDs are at most very weak radio AGN: their 3 GHz characteristic power is bounded below $1.8\\times10^{24}$ W Hz$^{-1}$ at $z\\approx7.1$, fainter than the typical stacked signal from high-redshift radio-quiet AGNs.","The radio-loud fraction among LRDs is far below the roughly 8% seen in high-redshift AGNs, since none of the 919 LRDs has a secure radio counterpart.","If dust-obscured star formation powers the red continuum, the radio limits cap the star formation rate below about 350–650 solar masses per year, excluding extreme starbursts of order $10^3\\,M_\\odot$ yr$^{-1}$ for the typical LRD.","Future larger LRD samples or deeper radio surveys could detect the underlying stacked emission, allowing radio properties to be mapped as a function of redshift and other physical parameters."],"supporting_citations":[{"why":"supplies the largest LRD sample and the photometric redshift range (2.4–11.4) used to characterize the stacked population","marker":"Kocevski et al. 2024"},{"why":"defines LRD selection and shows that many LRDs have broad H-alpha, the AGN signature this radio search is testing","marker":"Greene et al. 2024"},{"why":"provides the VLASS epoch 1 Quick Look catalog and images used in the cross-match and stacking","marker":"Gordon et al. 2021"},{"why":"defines the FIRST survey whose 1.4 GHz images and catalog are stacked and cross-matched","marker":"White et al. 1997"},{"why":"supplies the updated FIRST catalog used for the individual cross-match","marker":"Helfand et al. 2015"},{"why":"provides the comparison stacked sample of high-redshift radio-quiet AGNs against which the LRD upper limits are measured","marker":"Perger et al. 2024"},{"why":"gives the redshift-dependent radio power-to-star-formation-rate relation used to translate the upper limits into SFR bounds","marker":"Novak et al. 2017"},{"why":"justifies the 5 arcsecond search radius and quantifies the random association rate used to reject the two nearby FIRST sources","marker":"Ivezić et al. 2002"}],"fun_headline_variants":["Radio silence: little red dots vanish in stacked survey maps","No radio counterparts for 919 little red dots in VLASS and FIRST","Little red dots show no radio emission to 3-sigma in stacks","Deep stacks find little red dots are radio-quiet AGN candidates","Zero radio detections from 919 little red dots in wide surveys"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The stacking only uses fields that appear empty at first glance, assuming that discarding the few cutouts with a bright central source does not remove the very LRDs whose radio emission the stacking is meant to measure.","fun_headline_variants_meta":{"raw":{"variants":["Radio silence: little red dots vanish in stacked survey maps","No radio counterparts for 919 little red dots in VLASS and FIRST","Little red dots show no radio emission to 3-sigma in stacks","Deep stacks find little red dots are radio-quiet AGN candidates","Zero radio detections from 919 little red dots in wide surveys"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000572,"raw_usage":{"total_tokens":2753,"prompt_tokens":1045,"completion_tokens":1708,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":661,"completion_tokens_details":{"reasoning_tokens":1617}},"tokens_in":661,"tokens_out":1708,"duration_ms":10890,"temperature":1.0,"reasoning_tokens":1617,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:05:23.923703+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A deeper 3 GHz image of the same 919 fields reaching rms noise below about 1 microjansky per beam that shows a stacked peak above 3 sigma, or even a single LRD with a secure VLASS or FIRST counterpart at signal-to-noise above 5, would contradict the conclusion that LRDs are radio-silent.","supporting_citations":[{"cited_title":"E., Labb \\'e , I., Goulding , A","cited_arxiv_id":null,"evidence_quote":"defines LRD selection and shows that many LRDs have broad H-alpha, the AGN signature this radio search is testing"},{"cited_title":"A., Boyce , M","cited_arxiv_id":null,"evidence_quote":"provides the VLASS epoch 1 Quick Look catalog and images used in the cross-match and stacking"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the comparison stacked sample of high-redshift radio-quiet AGNs against which the LRD upper limits are measured"}],"review_version":1}