{"id":"1f8297cd-04e5-4722-9798-a9cae5f2e19a","arxiv_id":"2506.18112","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A 4.8 GHz VLBA survey of the JWST North Ecliptic Pole field detects 12 compact parsec-scale AGN cores among 106 faint radio sources, showing black hole driven emission is common in radio-quiet AGN.","lead":"Astronomers used very long baseline radio telescopes to zoom into 106 faint radio sources in a JWST survey field and found 12 with tiny, bright cores powered by supermassive black holes, not just by star formation. This shows that even 'radio quiet' galaxies often harbor compact AGN emission, and JWST data help measure how much of the radio glow comes from the black hole.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"SF-fraction claim rests on unverified VLA flux completeness; an independent SFR-based check would settle it","rationale":"The reader's weakest_assumption correctly identifies that the VLBA/VLA ratio is being used to infer the SF fraction, and that this assumes the non-VLBA VLA flux is star formation. However, the reader's stated alternative—diffuse AGN emission in the missing flux—would actually lower the inferred SF fraction, strengthening rather than weakening the AGN-dominance claim. The load-bearing risk is the opposite one: the VLA observations may have resolved out extended star-forming emission, so the total flux (and hence the SF fraction) is underestimated. This concern is partially mitigated by the selection of unresolved VLA sources and is independently addressed by the JWST SFRs, which confirm low SF for most sources with measurements. Because the claim is likely correct but is supported by an untested flux-completeness assumption, the conditional verdict is appropriate and no change is needed. A direct SFR-based test would verify the claim without relying on the VLBA/VLA ratio assumption.","tokens_in":28616,"tokens_out":23845,"duration_ms":213855,"concrete_test":"For each of the 8 VLBA detections with JWST/NIRCam counterparts, convert the JWST SFR (Table 6) to an expected 1.4 GHz radio luminosity using the Murphy et al. (2011) relation, extrapolate to 4.8 GHz assuming α = -0.7, and compare with the observed VLA 3 GHz luminosity (extrapolated to 4.8 GHz) and the VLBA 4.8 GHz luminosity. If the SFR-predicted radio luminosity is less than 50% of the VLA 4.8 GHz luminosity for at least 5 of the 7 sources with reliable SFRs (excluding PC 47, whose SFR appears inconsistent with its radio luminosity), the SF<50% claim is independently confirmed. For the four sources without JWST SFRs, apply the same test using WISE 22 micron flux (with an AGN decomposition) or SCUBA-2 SFRs as an SFR proxy.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim—that star formation contributes less than 50% of the radio emission in most of the 12 VLBA detections—is inferred from the VLBA/VLA flux density ratio. This inference requires two assumptions: (1) the VLA 3 GHz flux represents the true total radio flux, i.e., the 0.7-arcsecond VLA observations have not resolved out any significant extended star-forming emission; and (2) the VLA flux not recovered by the VLBA is dominated by star formation. Assumption (2) is conservative in the sense that if the missing flux were diffuse AGN emission, the SF fraction would be even lower, so it does not threaten the claim. The real risk is assumption (1): if a substantial fraction of star-forming emission is on scales larger than 0.7 arcsecond, it would be missing from both the VLA and VLBA measurements, making the VLBA/VLA ratio artificially high and overestimating the AGN-dominated fraction. The paper's selection of unresolved VLA sources mitigates this risk but does not eliminate it, and no test of VLA flux completeness is presented. The independent JWST SFRs (Table 6) actually support the claim for the 7 sources with measurements (SFR-based radio luminosities are well below 50% of the observed radio luminosity for at least 5 of them), so the conclusion is probably correct, but the headline statement is not directly established by the ratio argument alone.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a 4.8 GHz VLBA survey of 106 compact VLA sources in the JWST North Ecliptic Pole Time-Domain Field, detecting 12 sources at roughly 3.3 uJy rms sensitivity and 4 mas resolution. The authors derive parsec-scale sizes, brightness temperatures, spectral indices, and VLBA/VLA flux ratios, and they match the detections to WISE, SCUBA-2, JWST/NIRCam, and SDSS counterparts. They conclude that most detections contain AGN-driven compact radio emission, that the VLBA/VLA flux ratio correlates with flat spectral indices, and that star formation contributes less than 50% of the radio emission in the majority of the sources. They also compare SCUBA-2-based and JWST-based star formation rates and discuss the resulting discrepancies.","tokens_in":28873,"tokens_out":13242,"duration_ms":127452,"significance":"If the central claims hold, this work demonstrates that deep, high-frequency VLBI observations can identify compact AGN components in a radio-quiet AGN population where star formation has often been assumed to dominate, and the JWST-based SFR comparison is a useful step toward AGN/SF decomposition. The survey is among the most sensitive VLBI deep fields at 4.8 GHz, and the detailed calibration and imaging description, together with the catalog in Tables 3 and A1, makes the observational result reproducible and valuable as a reference sample. The high brightness temperatures are a robust observational signature of non-thermal AGN activity. The main weaknesses are that the spectral-index/compactness correlation is not quantified statistically and that the SF-fraction claim rests on an unstated assumption about the origin of the VLA flux not recovered by the VLBA.","major_comments":[{"comment":"The claimed correlation between VLA spectral index and VLBA/VLA flux-density ratio, and the associated compactness trend, is not quantified. The figures show no error bars on the ratio even though the ratio is formed by extrapolating the 3 GHz VLA fluxes to 4.8 GHz using spectral indices whose uncertainties reach about ±0.33 in Table 3, and the sample contains only 12 sources. Before the sharp rise at α≳−0.5 is interpreted as evidence for two accretion regimes, the authors should report a rank correlation coefficient with a significance estimate (for example Spearman or Kendall with bootstrap). As presented, the visual trend is not established at the claimed level.","section":"§4.4, Figures 7–8"},{"comment":"The central statement that star formation contributes less than 50% of the radio emission in the majority of detections rests on interpreting the VLBA/VLA flux ratio as an AGN/SF decomposition. This assumes that the VLA emission missing from the 4 mas VLBA images is star formation resolved out at VLBA resolution; unresolved low-surface-brightness AGN jet or lobe emission would make the inferred SF fraction an upper limit rather than a measurement, and the 0.7-arcsecond VLA data could themselves miss extended star-forming emission. The independent JWST SFRs in Table 6 support the conclusion for at least five of the seven sources with measurements, but the abstract's claim covers all 12 detections, four of which lack JWST constraints, and the two high-SFR sources PC 46 and PC 47 complicate the picture. The ratio argument should be presented as suggestive, supported by the SFR check for the subset with JWST data, rather than as a direct measurement for the full sample.","section":"§5.2 and Abstract"},{"comment":"The use of the P_cross criterion to conclude that all VLBA detections are AGN-dominated appears to mix radio frequencies. The P_cross threshold of Magliocchetti et al. (2018) is built on 1.4 GHz radio luminosity functions, while Figure 12 and the text plot 4.8 GHz luminosities. Since L_4.8/L_1.4 = (4.8/1.4)^α, with α in the range roughly −1 to −0.5 for these sources, the threshold must be converted to 4.8 GHz before comparison. The paper does not state such a conversion, so the statement that all VLBI sources lie above this threshold is not currently established; a borderline source such as PC 41 could change classification.","section":"§5.1 and Figure 12"}],"minor_comments":[{"comment":"Table 2 lists a detection fraction of 20% for the NEP field, while the abstract and §4.1 report 12/106 ≈ 11%; these numbers should be reconciled.","section":"Table 2 vs. Abstract and §4.1"},{"comment":"The brightness temperature is computed with z=0 for all sources even though redshifts are known for eight sources in Table 6. Since the (1+z) factor only increases T_b, this is conservative, but Table 5 should either use the measured redshifts or explicitly state that all values are lower limits evaluated at z=0.","section":"§4.3, Equation (1)"},{"comment":"The table note says that for unresolved sources the deconvolved size is set to the beam size, but the listed deconvolved major/minor axes for PC 24, PC 64, PC 67, and PC 71 are smaller than or different from the 4.0×3.5 mas beam; please clarify whether these are fitted deconvolved values or beam-size lower limits.","section":"Table 5"},{"comment":"The text refers to 'flux densities above 200 mJy' for PC 6 and PC 33, but Table A1 gives 340 and 324 µJy; the unit should be µJy, not mJy.","section":"§4.4"},{"comment":"The descriptions of PC 3 and PC 7 report only 0.24% and 0.21% AGN contributions, which conflicts with §4.7's statement that the median fractional AGN contribution is approximately 0.21–0.25; if the latter is a fraction, the former should read 21–25%.","section":"§5.4"},{"comment":"The text says flat or inverted AGN spectra have α≳0.7, while Table 1 and the rest of the paper use α≳−0.5 as the flat/inverted boundary; this threshold should be made consistent.","section":"§5.1 vs. Table 1"},{"comment":"The statement that all VLBI sources lie above the P_cross threshold should be restricted to the eight sources with redshifts; the four sources without redshifts cannot be placed in Figure 12.","section":"§5.1 and Figure 12"}],"recommendation":"major_revision","confidential_remarks":"The observational data are valuable and the high brightness temperatures robustly support AGN activity, so I do not see grounds for rejection. The main issues are statistical and quantitative: the compactness correlation needs a significance test, the SF-fraction claim needs an explicit assumption or an independent check, and the P_cross comparison needs a frequency conversion. I would ask the authors to address these points before recommending acceptance, along with the smaller numerical inconsistencies noted in the minor comments."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Main takeaway: this is a worthwhile observational catalog, not a conceptual breakthrough. The 4.8 GHz VLBA survey of 106 VLA sources in the JWST NEP TDF is genuinely new data, and the 12 detections with brightness temperatures above 10^5 K are solid, high-S/N evidence for compact AGN cores in radio-quiet AGN. The detection rate, flux ratios, and spectral-index trends broadly echo earlier 1.4-1.6 GHz VLBI surveys, so the interpretive advance is incremental, but the measurements themselves are the most sensitive at this frequency in this field and will be a useful reference.\n\nWhat the paper does well: careful calibration and imaging (in-beam phase referencing, primary beam corrections, point vs. Gaussian fits), a clean detection criterion, and a serious multi-wavelength counterpart search (WISE, SCUBA-2, JWST/NIRCam, SDSS). The JWST SFR comparison, while preliminary, is a sensible way to confront AGN/SF decomposition, and the authors are appropriately cautious about SCUBA positional uncertainties and the reliability of radio-derived SFRs in AGN hosts. The self-citations to Hyun et al. and Willner et al. are natural in a coordinated field program, not circular.\n\nSoft spots, in order of real impact:\n\n1. Internal inconsistency on redshift in the Tb calculation. Section 4.3 states z=0 for all sources \"as we do not have information regarding the redshift,\" but Table 6 gives redshifts for eight of the twelve detections. Since the Tb formula includes (1+z), this is a straightforward error and makes the stated Tb values inconsistent with the rest of the paper. For z~1 sources the true Tb would be higher, so the AGN conclusion actually strengthens, but the calculation as presented is wrong.\n\n2. The claimed spectral-index/compactness correlation (Figures 7 and 8) is presented without error bars or any significance test. With 12 sources and no quoted scatter or Spearman/Pearson statistic, saying there is a \"sharp rise\" is qualitative. This is fixable, but it should not be stated as a robust correlation.\n\n3. The star-formation contribution claim. The paper equates high VLBA/VLA flux ratios with AGN dominance, assuming the VLA flux not recovered by the VLBA is star formation. The stress-test concern about missing extended SF in the 0.7\" VLA data is legitimate but partly mitigated by the unresolved-source selection. The independent JWST SFRs for seven sources actually support the claim (SFR-based radio luminosities are well below 50% of observed for most), but the paper does not connect those two pieces of evidence directly. So the headline \"SF contributes <50%\" is probably correct, but it is not as directly established by the ratio argument as the text suggests.\n\nAdditionally, the VLBA 4.8 GHz to VLA 3 GHz comparison relies on spectral-index extrapolation, and those errors are not propagated into the flux ratios; minor, but worth noting.\n\nWho this is for: radio and multi-wavelength AGN folks, especially anyone working on radio-quiet AGN cores, deep VLBI surveys, or the NEP TDF. This paper is citable as a catalog and a sensitivity benchmark.\n\nRecommendation: yes, this deserves serious peer review and is conditionally publishable. A referee should require the Tb redshift inconsistency to be fixed, a significance test for the correlation, and either a softening of the SF-fraction claim or a direct SFR-based check of it. None of these changes threaten the core detection result.","headline":"A solid VLBA catalog paper with robust AGN detections, but the SF-fraction headline is overreaching and one internal inconsistency (z=0 in Tb despite known redshifts) needs fixing.","tokens_in":29578,"tokens_out":2894,"would_cite":true,"duration_ms":31207,"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":"The paper reports that 4.8 GHz VLBA observations of 106 VLA radio sources in the JWST North Ecliptic Pole Time-Domain Field detect 12 compact parsec-scale AGN cores, and argues that in most of these radio-quiet AGN star formation…","keywords":["radio-quiet AGN","VLBA","very long baseline interferometry","compact radio cores","brightness temperature","star formation vs AGN","JWST North Ecliptic Pole Time-Domain Field","radio spectral index"],"falsifier":"Image the 12 VLBA-detected sources at a resolution between the VLA's 0.7 arcsecond and the VLBA's 4 mas (for example, at roughly 0.1 arcsecond) and measure the spectral index and morphology of the flux the VLBA resolves out; if that missing flux shows steep-spectrum extended lobes or a jet rather than a star-forming disk, the assumption that it is star formation is wrong and the inferred star-formation fractions would need revision.","tokens_in":28430,"feed_emoji":"📡","tokens_out":5190,"duration_ms":46618,"temperature":0.7,"pith_summary":"This paper presents the first results of a 4.8 GHz VLBA survey of the JWST North Ecliptic Pole Time-Domain Field, targeting 106 VLA-detected radio sources. At roughly 4 mas resolution and 3.3 microJy rms sensitivity, 12 sources were detected, and their parsec-scale sizes, brightness temperatures above $10^5$ K, and high VLBA/VLA flux ratios mark them as non-thermal AGN emission rather than star formation. The central claim is that in the majority of these radio-quiet AGN, star formation contributes less than half of the VLBA-scale radio emission, with several cases that are almost entirely AGN-driven. This matters because radio-quiet AGN are often assumed to be star-formation-dominated at radio wavelengths, and the result indicates a common black-hole-related component that arcsecond-scale surveys alone would overlook.","feed_headline":"Black-hole cores found in most detected radio-quiet AGN","feed_subtitle":"VLBA parsec-scale detections show star formation supplies under half the radio light in most cases.","key_machinery":"The central observational tool is the VLBA at 4.8 GHz with about 4 mas resolution and $\\sim3.3\\,\\mu$Jy/beam rms sensitivity, which separates compact nuclear emission from extended disk emission. The two quantitative handles are the brightness temperature $T_\\mathrm{B} > 10^5$ K, which excludes star formation as the dominant emission mechanism, and the VLBA/VLA flux density ratio, which measures how much of the arcsecond-scale VLA emission is recovered in the compact core. The VLA 3 GHz spectral index completes the picture: sources with flat spectra ($\\alpha \\gtrsim -0.5$) lie close to equality in the VLBA-versus-VLA plot, identifying them as self-absorbed AGN cores, while steeper-spectrum sources are more extended and retain a larger star-forming component.","core_discovery":"Most of the 12 VLBA detections harbor compact parsec-scale radio sources with brightness temperatures exceeding $10^5$ K and VLBA/VLA flux ratios that place the emission in the AGN-dominated regime. The paper concludes that star formation contributes less than 50% of the total VLBA radio emission in the majority of these sources, and in a few cases the emission is almost entirely AGN-driven. The compact emission is confined to regions smaller than about 40 pc, consistent with the base of a jet or the accretion-disk corona, and flatter-spectrum sources ($\\alpha \\gtrsim -0.5$) show higher VLBA/VLA ratios, indicating optically thick, self-absorbed synchrotron cores. Eight detections with JWST/NIRCam counterparts lie in early-type, bulge-dominated galaxies with low JWST-based star formation rates, and WISE colors of the detections are AGN-like or intermediate-disk rather than purely star-forming.","pith_inferences":["If compact AGN cores are common in radio-quiet AGN, the label 'radio-quiet' may describe low jet power or orientation rather than the absence of jet launching; a larger sample at similar sensitivity could test whether flat-spectrum cores appear in all such AGN.","The paper's star-formation fraction argument assumes the VLA flux that the VLBA does not recover is star formation resolved out at 4 mas resolution; imaging the same sources at intermediate resolution (about 0.1 arcsecond) would directly test whether any of that missing flux is diffuse AGN lobe emission.","Extending the same survey to 1.4 GHz would test how much synchrotron self-absorption shapes the 4.8 GHz detections, since self-absorbed cores are relatively brighter at higher frequencies and lower-frequency observations might reveal additional extended AGN components."],"forward_implications":["Arcsecond-scale radio surveys that classify radio-quiet AGN by spectral index alone may systematically miss the compact AGN component, so VLBI follow-up is needed to reveal it.","The detection rate rises from about 11% overall to roughly 35% for VLA sources brighter than 50 microJy at 3 GHz, implying that deeper or longer VLBA observations should recover many more compact cores in this field.","JWST-based star formation rates for the VLBA detections are lower than SCUBA-2 estimates, indicating that radio-inclusive spectral energy distribution fits can overestimate star formation when an AGN contributes to the radio flux.","The host galaxies of the VLBA detections are predominantly early-type, bulge-dominated systems, linking compact AGN cores to massive bulges rather than to actively star-forming disks.","The absence of kpc-scale radio emission from most of the compact AGN cores raises the open question of why these jets or coronae do not produce larger-scale radio structures."],"supporting_citations":[{"why":"Supplies the parent VLA 3 GHz catalog of 588 sources from which the 106 VLBA phase centers and their spectral indices are drawn.","marker":"Hyun et al. 2023"},{"why":"Provides the Lockman Hole VLBI comparison sample and the spectral-index trend that the TDF detections are compared against.","marker":"Middelberg et al. 2013"},{"why":"Defines the expected spectral-index, VLBI-morphology, and Eddington-ratio properties of proposed AGN emission mechanisms used to interpret the detections.","marker":"Panessa et al. 2019"},{"why":"Establishes the brightness-temperature regime in which star formation can produce radio emission, anchoring the $T_\\mathrm{B} > 10^5$ K AGN argument.","marker":"Condon 1992"},{"why":"Supplies the JWST/NIRCam counterparts and the host-galaxy SED fitting method used to derive low star formation rates for the detections.","marker":"Willner et al. 2023"},{"why":"Provides the catalog of JWST galaxies with point-source cores used to classify three VLBA detections and to compare infrared AGN signatures.","marker":"Ortiz et al. 2024"},{"why":"Defines the $P_\\mathrm{cross}$ radio-luminosity threshold separating AGN-dominated from star-formation-dominated radio emission, above which all VLBA detections lie.","marker":"Magliocchetti et al. 2018"},{"why":"Supplies the WISE color-color classification scheme used to separate AGN, intermediate-disk, and star-forming hosts.","marker":"Jarrett et al. 2017"}],"fun_headline_variants":["Star formation under half the radio light in most radio-quiet AGN","Parsec-scale AGN cores dominate radio emission in quiet AGN","VLBA reveals AGN-driven radio in most radio-quiet AGN","Most quiet AGN radio emission traced to black holes, not stars"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the arcsecond-scale VLA flux the VLBA does not recover is star formation resolved out at 4 mas resolution, not diffuse AGN jet or lobe emission, and if some of that missing flux is AGN-related the claim that star formation contributes less than half of the radio emission in most detections would overstate AGN dominance.","fun_headline_variants_meta":{"raw":{"variants":["Star formation under half the radio light in most radio-quiet AGN","Parsec-scale AGN cores dominate radio emission in quiet AGN","VLBA reveals AGN-driven radio in most radio-quiet AGN","Most quiet AGN radio emission traced to black holes, not stars"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000284,"raw_usage":{"total_tokens":1746,"prompt_tokens":1087,"completion_tokens":659,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":703,"completion_tokens_details":{"reasoning_tokens":582}},"tokens_in":703,"tokens_out":659,"duration_ms":6213,"temperature":1.0,"reasoning_tokens":582,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:54:28.094212+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Image the 12 VLBA-detected sources at a resolution between the VLA's 0.7 arcsecond and the VLBA's 4 mas (for example, at roughly 0.1 arcsecond) and measure the spectral index and morphology of the flux the VLBA resolves out; if that missing flux shows steep-spectrum extended lobes or a jet rather than a star-forming disk, the assumption that it is star formation is wrong and the inferred star-formation fractions would need revision.","supporting_citations":[],"review_version":2}