{"id":"7167df53-fbb1-42b3-98a8-322aff71e15b","arxiv_id":"2501.08792","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A compact detector plus 6LiPE pinhole imaged neutrons scattered from a plastic cube at the ILL beam with peak-to-background about 15, and simulations predict a coded-aperture mask improves efficiency by roughly 15x over a 5 mm pinhole.","lead":"The GN-Vision prototype, a compact dual neutron-gamma imager, has produced its first experimental neutron images by pairing a position-sensitive CLYC detector with a lithium-polyethylene pinhole. If the concept holds up, a single handheld device could simultaneously locate gamma and neutron sources for nuclear security, waste inspection, and proton-therapy monitoring.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The coded-aperture imaging test in Sec. 4.3 is circular: the test patterns are generated as linear combinations of the same response matrix used for deconvolution, so the claimed imaging capability is not independently validated.","rationale":"The reader's weakest_assumption focused on the fidelity of the Gaussian position-response model and the deferral of experimental details to the companion paper. Those are legitimate concerns, but the more immediate and self-contained flaw is the circular validation of the coded-mask reconstruction: using the same response matrix to generate and invert the test data cannot expose model error. This does not overturn the pinhole proof-of-concept, which is an experimental result reported in the paper (though detailed in [33]), nor does it invalidate the counting-based efficiency ratio, which follows from open fraction. It does mean the simulation-based 'optimization study' for the coded aperture is weaker than presented. Since the reader already assigned a conditional verdict, and this concern is consistent with that conditionality, no change to the verdict is needed. The proposed concrete test—independent generation of test patterns outside the response-matrix basis—would directly settle whether the coded-mask imaging claim survives non-circular testing.","tokens_in":15533,"tokens_out":5943,"duration_ms":59980,"concrete_test":"Generate a test neutron source pattern (e.g., the 'G' shape) using an independent forward simulation with source positions not included in the 10x10 response-matrix grid, or with a different Monte Carlo code, then reconstruct it with the original Geant4-based response matrix and the Sec. 4.3 algorithms. If the reconstructed image fails to reproduce the source pattern with contrast comparable to the pinhole PBR of ~15, or if source centroids are biased by more than one resolution element, the coded-aperture imaging claim is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central optimization claim—that a Rank-5 MURA coded mask with unfolding algorithms can image thermal neutron sources and provide a 15.6x efficiency gain over a 5 mm pinhole—rests on a closed-loop simulation test. In Sec. 4.3, the response matrix is built from 100 simulated point-source responses at the 10x10 grid positions. The test patterns (e.g., the 'G' shape) are then created by linearly combining those same responses via Eq. (2): R_j = sum_i D_ij v_i. The Agostini and Maximum Entropy algorithms reconstruct the images using exactly this same matrix D. This demonstrates only that the algorithms can invert their own forward model, not that the Geant4 response model correctly represents the physical detector. Any systematic error in the simulated position response, mask geometry, or neutron transport is invisible to this test. The efficiency ratio itself is mostly determined by the open-fraction difference (0.48 vs. pi/100), so the 15.6x factor is plausible as a counting-rate statement, but the qualitative demonstration that coded-aperture neutron imaging 'works' with this device is unsupported. Additionally, the experimental proof-of-concept numbers (peak-to-background ~15, angular resolution <5 degrees) are deferred to the companion paper [33], and the simulation model uses a 5 mm FWHM Gaussian response with edge compression that is only verified over the central 30 mm of the crystal. If the actual spatial response is more non-linear, or if [33] reports different resolution, the predicted coded-mask and pinhole imaging performance would need revision.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents the first experimental neutron images from the GN-Vision prototype, a compact dual neutron-gamma imager based on a monolithic CLYC-6 crystal with SiPM readout and a 6LiPE pinhole collimator, and then reports Geant4 simulation studies of the final design. The experimental section (Sec. 3) summarizes the detector characterization and shows reconstructed images of a polyethylene cube irradiated by a thermal neutron beam, claiming a peak-to-background contrast of about 15 and an angular resolution below 5 degrees, with details deferred to the companion paper [33]. The simulation section (Sec. 4) evaluates the impact of experimental resolutions on Compton and neutron imaging, compares natural and enriched LiPE collimator materials, and investigates a Rank-5 MURA coded-aperture mask with deconvolution algorithms, reporting a 15.6x efficiency gain over a 5 mm pinhole and a 60-degree field of view.","tokens_in":15789,"tokens_out":9661,"duration_ms":84629,"significance":"If the experimental proof-of-concept holds, GN-Vision represents a meaningful step toward compact handheld dual neutron-gamma imaging with applications in nuclear security and hadron therapy. The simulation study provides useful guidance for collimator upgrades, particularly the natural LiPE option for cost reduction and the order-of-magnitude efficiency gain from a coded aperture. However, the experimental evidence is largely delegated to a companion paper, and the coded-aperture imaging demonstration in Sec. 4.3 is a self-consistency check rather than an independent validation. The efficiency gain claim is plausible but contains an internal inconsistency in the pinhole geometry that must be resolved before the headline number can be accepted.","major_comments":[{"comment":"The efficiency comparison is internally inconsistent. The text states \"the aperture radius was D = 5 mm, resulting in an open fraction rho = pi/100 ~ 0.031\" and later refers to a \"5 mm diameter pinhole\". For a 50x50 mm detector, rho = pi/100 corresponds to a pinhole radius of 5 mm (diameter 10 mm), not a 5 mm diameter; a 5 mm diameter pinhole would have rho ~ 0.00785. The claimed 15.6(2) efficiency ratio therefore refers to a 10 mm diameter pinhole, not the stated 5 mm diameter. If the 5 mm diameter is intended, the efficiency gain would be about 61. Please correct the geometry and recompute the ratio.","section":"Sec. 4.3"},{"comment":"The \"G\" pattern test is constructed by Eq. (2) as a linear combination of the same response matrix D used by the deconvolution algorithms. Reconstructing this pattern demonstrates only that the algorithms can invert their own forward model; it does not validate that the Geant4 response model correctly represents the physical detector. The sentence \"Upon successful reconstruction of the first thermal neutron images using the coded mask and unfolding algorithms\" overstates the result, as no experimental coded-mask data are presented. Please reframe this as an algorithmic self-consistency check, and if the coded-aperture imaging capability is to be claimed, provide an independent test (e.g., a simulated source pattern not built from D, or an experimental measurement).","section":"Sec. 4.3"},{"comment":"The paper's title promises first experimental results, but the quantitative experimental claims—PSD figures of merit (3.8 and 2.9), position resolution of 5 mm, linear response over the central 30 mm, and angular resolution below 5 degrees—are deferred to Ref. [33] without presenting the underlying data or analysis. The image reconstruction in Sec. 3.2 also lacks a description of event selection, background subtraction, and the uncertainty on the ~15 peak-to-background ratio. Please add the essential details or explicitly state which results come from the companion paper and how they were obtained.","section":"Sec. 3"},{"comment":"The simulation predictions use a CLYC spatial response of 5 mm FWHM Gaussian with edge compression, verified only over the central 30 mm of the 50 mm crystal (Sec. 4.1). The coded-aperture study in Sec. 4.3 uses a 2 mm sigma Gaussian and omits the edge compression entirely, despite the 60-degree FOV illuminating a large peripheral area. The robustness of the predicted efficiency and image quality to the spatial-response model should be tested (e.g., by varying the blur width and including the edge compression) or explicitly stated as a limitation.","section":"Sec. 4.1 and 4.3"}],"minor_comments":[{"comment":"The introduction contains an incomplete sentence: \"the gamma-imaging capability in GN-Vision is already at a very high technology readiness level (TRL) following .\"","section":"Sec. 1"},{"comment":"Citation placeholders remain in the text: \"[47? , 48]\" and \"[45, 24, 46, 47 ? , 48]\"; these need to be resolved before publication.","section":"Sec. 1 and Sec. 5"},{"comment":"The phrase \"utter GN-Vision prototype\" should likely be \"ultimate GN-Vision prototype\".","section":"Sec. 4.1"},{"comment":"The spatial resolution notation is inconsistent: Sec. 4.1 uses \"5 mm (FHWM)\" while Sec. 4.3 uses \"2 mm (sigma)\"; please use a single convention and correct the FWHM typo.","section":"Sec. 4.1 and 4.3"},{"comment":"The angular resolution statement \"8 mm (sigma) ... corresponds to an angular resolution of ~5 degrees\" should specify whether the 5 degrees is FWHM or sigma; the conversion implies FWHM.","section":"Sec. 4.1"},{"comment":"The caption contains a typo: \"top pannel\" should be \"top panel\".","section":"Fig. 10 caption"}],"recommendation":"major_revision","confidential_remarks":"The manuscript relies heavily on companion paper [33] for the experimental results that constitute the paper's central claim. If the journal requires experimental papers to be self-contained, this is a scope concern. The reference list also contains unresolved placeholders, suggesting the manuscript was submitted in an incomplete state."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a straightforward status report from a group that knows what it is doing. The genuinely new piece is the first neutron image from the GN-Vision front plane — a 50x50 mm CLYC-6 crystal with Anger-logic SiPM readout behind a 6LiPE pinhole — and it shows a clean ~15:1 peak-to-background spot that shifts as expected. The experimental details (PSD figure of merit, position resolution, the sub-5 degree angular resolution claim) live in the companion paper [33], which is on arXiv, so the deferral is annoying but not a black hole.\n\nThe simulation work is mixed. The natLiPE vs 6LiPE collimator comparison is straightforward Geant4 with sensible inputs and gives a clear design message: natural LiPE is fine for thermal-neutron applications, enriched is needed for epithermal. The Rank-5 MURA coded-aperture study is where I part ways with the authors' framing. The efficiency gain of 15.6x is basically the open-fraction ratio (0.48 vs pi/100 ≈ 0.031) and the Monte Carlo just confirms the counting-rate expectation — that part is fine. But the demonstration that coded-aperture imaging 'works' is circular: the test 'G' pattern is synthesized as a linear combination of the same response-matrix columns that are then used for deconvolution. That shows the unfolding algorithms can invert their own forward model, and it cannot catch any systematic error in the simulated spatial response, mask geometry, or neutron transport. The authors do include a sensitivity check with mismatched resolution in the response matrix, but the source patterns are still generated from the same model. So the qualitative imaging capability claim is unsupported; the efficiency statement stands.\n\nThe paper also has citation glitches and typos, but nothing that undermines the main content. I do not see a load-bearing flaw in the experimental proof of concept — the image in Fig. 4 is real, and the claim that it is a pinhole image is credible. The main weakness is that the paper asks the reader to take too much on faith from [33] and from a self-consistency simulation.\n\nFor whom: detector developers and people in nuclear security or hadron therapy who want a compact dual imager. It deserves a serious referee, but the referee should push for (a) a non-circular test of the coded-aperture algorithm, e.g. using a measured or at least independently generated pattern, and (b) either moving the key experimental characterization from [33] into this paper or making the dependency explicit. I would accept it for review with moderate revisions expected.","headline":"Solid prototype status report; the experimental pinhole image is real, but the coded-aperture demonstration is circular and key characterization numbers are deferred to a companion paper.","tokens_in":16467,"tokens_out":2444,"would_cite":false,"duration_ms":24938,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["29.40.Mc"],"model":"deepseek-v4-flash","headline":"A hand-portable imager that sees gamma rays and slow neutrons at once has produced its first true neutron images, and a coded-mask upgrade is projected to multiply neutron efficiency by 15.6.","keywords":["dual neutron-gamma imaging","hadron therapy","nuclear inspections","neutron collimator","coded-aperture mask","CLYC detector","pinhole camera","MURA"],"falsifier":"Measure the same thermal-neutron source with the pinhole and with the rank-5 MURA mask under identical beam conditions: if the count-rate ratio deviates strongly from 15.6, or if off-axis reconstructed centroids drift beyond the few millimeters the model predicts, the simplified spatial-response model fails.","tokens_in":15293,"feed_emoji":"🔬","tokens_out":6094,"duration_ms":58940,"temperature":0.7,"pith_summary":"The paper reports the first experimental demonstration that GN-Vision, a compact dual-modality imager, can localize thermal neutron sources. Using a position-sensitive CLYC crystal behind a lithium-polyethylene pinhole, it reconstructed images of a neutron-scattering target with 11 mm FWHM, peak-to-background contrast of about 15, and angular resolution below 5 degrees. It then uses Geant4 simulations, calibrated to those measurements, to argue that a rank-5 MURA coded aperture would increase detection efficiency by a factor of 15.6 over the 5 mm pinhole and widen the field of view from 32 to 60 degrees. The same simulations indicate that natural-lithium polyethylene could replace the expensive enriched material for fully thermalized neutron spectra. If these results hold, a single handheld device could image gamma and neutron sources simultaneously for nuclear security, reactor inspection, and proton-therapy monitoring.","feed_headline":"First neutron images from hand-held dual imager","feed_subtitle":"Proof of concept spots thermal neutron sources with sub-5-degree resolution; a coded mask could cut acquisition time ~15x.","key_machinery":"The load-bearing components are the position-sensitive CLYC-6 detector and the Geant4 model. The detector couples a monolithic 50x50x13 mm CLYC-6 crystal, enriched to 95% 6Li, to an 8x8 SiPM array read out with Anger logic, giving a sub-pixel position resolution of 5 mm and pulse-shape-discrimination figures of merit of 2.9 with SiPM readout that separate neutrons from gamma rays. The pinhole imaging itself uses simple inversion with a scaling factor S = d/F. For the optimization, the Geant4 model (QGSP_INCLXX_HP physics list, G4NDL-4.6 data) implements the detector with a Gaussian 5 mm FWHM transverse blur plus edge compression, and the coded-aperture study builds a response matrix from 100 simulated point-source positions and reconstructs via the Agostini and Maximum-Entropy deconvolution algorithms. The MURA rank-5 mask, with open fraction 0.48, is the object that carries the efficiency gain.","core_discovery":"On the paper's own terms, the central discovery is that the first GN-Vision prototype, built from a 50x50x13 mm CLYC-6 crystal with SiPM Anger-logic readout and a 6LiPE pinhole collimator, produces genuine neutron images: the reconstructed source appears at the expected position with the expected size (11 mm FWHM), a peak-to-background ratio around 15, and an angular resolution below 5 degrees for thermal neutrons scattered by a 1 cm3 polyethylene cube. The simulations then show that replacing the pinhole with a rank-5 MURA coded mask raises efficiency by 15.6(2) times relative to the 5 mm pinhole (roughly four times more relative to the 2.5 mm pinhole used in the proof-of-concept experiment) while widening the field of view from 32 to 60 degrees, at the cost of coarser resolution that the authors expect to recover with better deconvolution or larger response matrices. The same simulations indicate that natural-lithium polyethylene gives acceptable contrast for fully thermalized spectra, making a cheaper collimator viable. For gamma rays, the realistic simulations show a two-fold broadening of the Compton image resolution from experimental effects, but the source centroids remain accurate.","pith_inferences":["If the 15.6x efficiency gain holds in hardware, the same acquisition time would yield substantially more counts, allowing real-time imaging of weak neutron sources; a natural next test is to measure the pinhole-versus-mask efficiency ratio directly on the ILL beam rather than in simulation.","The strong dependence of contrast on neutron energy (PBR falling from about 15 at thermal energies to roughly 2 to 6 at 1 eV even with enriched LiPE) implies that field deployments will need to know the source spectrum; in mixed thermal-epithermal environments the coded mask may need thicker or enriched material to keep useful contrast.","Because the coded-mask approach recovers image quality only when the response matrix includes the detector's measured blur, it effectively trades simple analytic imaging for a calibrated, computational imaging system; laboratory calibration of the spatial response may matter as much as the mask pattern itself.","For medical applications such as BNCT, where thermal neutrons dominate, the combination of natural LiPE and a coded mask could make a low-cost clinical dosimetry imager feasible, though that extrapolation goes beyond what the paper demonstrates."],"forward_implications":["A single handheld device can provide simultaneous gamma and neutron imaging, since the 6LiPE collimator is essentially transparent to gammas above about 100 keV while the CLYC layer separates the two particle types.","For thermal-neutron environments, collimator cost can be cut substantially by using natural LiPE rather than 95% enriched 6LiPE, with contrast still acceptable (PBR about 2.8 to 3.8) though the spatial resolution worsens by roughly 50 percent for thermal neutrons.","The coded aperture yields 15.6 times the efficiency of a 5 mm pinhole and a 60-degree field of view, enabling faster surveys or detection of weaker neutron sources; the resolution loss is expected to be mitigated by including detector blur in the response matrix and using more advanced deconvolution.","The full dual-imager integration would bring Compton gamma imaging at 511 keV with about two-fold resolution broadening from experimental effects but small centroid shifts, together with neutron imaging at roughly 5-degree angular resolution."],"supporting_citations":[{"why":"Provides the conceptual design and Monte Carlo performance study of GN-Vision, including the pinhole collimator parameter choices and the baseline neutron imaging efficiency.","marker":"[25]"},{"why":"Companion paper that holds the experimental characterization of the CLYC-SiPM detector (PSD figures of merit, 5 mm position resolution, linear central region) and the detailed ILL campaign, including the sub-5-degree angular resolution.","marker":"[33]"},{"why":"Source of the MURA rank-5 binary array pattern and its correlation properties used for the coded aperture mask.","marker":"[58]"},{"why":"Supplies the Geant4 toolkit in which all the optimization simulations are implemented.","marker":"[53]"},{"why":"Defines the QGSP_INCLXX_HP physics list used for neutron transport below 20 MeV.","marker":"[54]"},{"why":"Provides the G4NDL-4.6 data library, based on JEFF-3.3, used for neutron-induced reaction cross sections in the simulations.","marker":"[55]"},{"why":"Establishes the CLYC-6 response to gamma rays, fast neutrons, and thermal neutrons that makes the first detection plane capable of particle discrimination.","marker":"[49]"},{"why":"Describes the predecessor i-TED Compton imager whose detection planes and gamma-imaging principle GN-Vision extends.","marker":"[35]"}],"fun_headline_variants":["First neutron images from hand-held GN-Vision imager","GN-Vision prototype images thermal neutrons with sub-5° resolution","Portable dual imager proves neutron vision in first tests","Coded mask upgrade speeds GN-Vision neutron imaging 15x","Optimized collimator widens GN-Vision field, boosts efficiency"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything projected for the final device depends on the detector's position response being close to a Gaussian blur of 5 mm FWHM with mild edge compression; if the true response is more nonlinear toward the crystal edges, the simulated angular resolution and the coded-mask efficiency gains will overstate what the hardware can deliver.","fun_headline_variants_meta":{"raw":{"variants":["First neutron images from hand-held GN-Vision imager","GN-Vision prototype images thermal neutrons with sub-5° resolution","Portable dual imager proves neutron vision in first tests","Coded mask upgrade speeds GN-Vision neutron imaging 15x","Optimized collimator widens GN-Vision field, boosts efficiency"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000259,"raw_usage":{"total_tokens":1586,"prompt_tokens":949,"completion_tokens":637,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":565,"completion_tokens_details":{"reasoning_tokens":549}},"tokens_in":565,"tokens_out":637,"duration_ms":6160,"temperature":1.0,"reasoning_tokens":549,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:18:14.201200+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the same thermal-neutron source with the pinhole and with the rank-5 MURA mask under identical beam conditions: if the count-rate ratio deviates strongly from 15.6, or if off-axis reconstructed centroids drift beyond the few millimeters the model predicts, the simplified spatial-response model fails.","supporting_citations":[{"cited_title":"Allison et al","cited_arxiv_id":null,"evidence_quote":"Supplies the Geant4 toolkit in which all the optimization simulations are implemented."},{"cited_title":"Lerendegui-Marco et al","cited_arxiv_id":null,"evidence_quote":"Provides the conceptual design and Monte Carlo performance study of GN-Vision, including the pinhole collimator parameter choices and the baseline neutron imaging efficiency."},{"cited_title":"Imaging neutrons with a position-sensitive monolithic CLYC detector","cited_arxiv_id":"2410.12533","evidence_quote":"Companion paper that holds the experimental characterization of the CLYC-SiPM detector (PSD figures of merit, 5 mm position resolution, linear central region) and the detailed ILL campaign, including the sub-5-degree angular resolution."},{"cited_title":"Gottesman and E","cited_arxiv_id":null,"evidence_quote":"Source of the MURA rank-5 binary array pattern and its correlation properties used for the coded aperture mask."},{"cited_title":"https://geant4.web.cern.ch/ node/155","cited_arxiv_id":null,"evidence_quote":"Defines the QGSP_INCLXX_HP physics list used for neutron transport below 20 MeV."},{"cited_title":"Mendoza et al","cited_arxiv_id":null,"evidence_quote":"Provides the G4NDL-4.6 data library, based on JEFF-3.3, used for neutron-induced reaction cross sections in the simulations."},{"cited_title":"Giaz et al","cited_arxiv_id":null,"evidence_quote":"Establishes the CLYC-6 response to gamma rays, fast neutrons, and thermal neutrons that makes the first detection plane capable of particle discrimination."},{"cited_title":"Domingo-Pardo","cited_arxiv_id":null,"evidence_quote":"Describes the predecessor i-TED Compton imager whose detection planes and gamma-imaging principle GN-Vision extends."}],"review_version":1}