{"id":"221a2ed7-7244-404f-8047-a6206b8bbc41","arxiv_id":"2505.04789","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Sweat on the forehead progresses from cyclic pore-level droplets to spreading puddles then to films, and leftover salt makes re-sweating spread as a film more quickly.","lead":"Using imaging and sensors on the foreheads of six subjects, this paper maps how sweat appears, spreads, and evaporates at the pore scale during heating, cooling, and reheating. It links these micro-patterns to standard sweat-rate, skin-conductance, and hydration readings, and proposes that salt left on the skin changes how sweat spreads in a second round of sweating.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Macroscale SR/GSC/HYD are measured on skin outside the imaged patch and under local perturbation; without evidence of synchronized onset and mode timing, the microscale-to-macroscale temporal story may be a site-mismatch artifact.","rationale":"Read in good faith, this is a carefully executed observational study with genuine multimodal imaging, and the porewise, transition, and filmwise modes are visually plausible. The central structure of the argument, however, is comparative: it explains the temporal behavior of macroscopic signals by imaging a different skin site. The reader flagged the same-state assumption as weakest, and I agree that it is load-bearing. My concern is slightly broader: even if sweat onset were spatially uniform, the ventilated capsule and GSC electrode gel perturb local sweating, so the macroscopic instruments are not passive observers of the same state. The paper contains no validation that the 2–4 cm separation and local perturbations are inconsequential, and its own air-jet data (Figure 2d, Section 3.1) show that capsule airflow can materially change local sweat dynamics. This does not invalidate the descriptive imaging findings, but it means the headline temporal correlations should be treated as hypothesis-generating rather than demonstrated. The conditional verdict remains appropriate, with the condition made explicit: co-location or synchronization validation of the macroscopic and microscopic measurements is needed before the relative-dynamics claim can be accepted.","tokens_in":20074,"tokens_out":8167,"duration_ms":95572,"concrete_test":"Repeat the protocol on at least three subjects with two synchronized MWIR cameras: one over the current central patch and one viewing skin through an IR-transparent window in the ventilated capsule (the authors' earlier wind-tunnel capsule, ref. 36, provides this capability), with GSC electrodes placed in or immediately adjacent to the imaged field. If lateral and central sweat onset times differ by more than the reported 1.7±0.8 min lead, or if porewise-to-transition and transition-to-filmwise transition times differ by more than the stated 5-minute binning, then the macroscale-microscale time alignment and the mode-correlated interpretation in Figures 3 and 5 and Table 1 are not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central temporal claim—that GSC rises 1.7±0.8 min before MWIR-visible pore sweating and that SR and HYD rise 1.4±0.7 and 2.7±1.8 min afterward, with mode transitions aligning with inflections in SR, GSC, and HYD (Figures 3 and 5; Table 1)—requires that the lateral GSC electrodes, the ventilated capsule, and the 2–4 cm displaced HYD probe observe the same sweating state as the central MWIR/OCT patch at the same times. The manuscript does not test this. The threat is not only natural spatial heterogeneity of onset: the ventilated capsule imposes forced dry airflow over the skin beneath it, cooling and drying that site. The paper itself demonstrates this effect for the air-jet capsule (Figure 2d) and describes the cylindrical capsule's apparent saturation; GSC electrodes sit under isotonic gel and occlude the skin, and the HYD probe applies 1.5–2 N. Thus the macroscopic signals may reflect a different, locally perturbed sweating state, and the observed ordering and mode-correlated changes cannot be attributed to the imaged porewise-to-transition-to-filmwise dynamics without a co-location check. Table 1 uses MWIR onset as time zero for all instruments, thereby assuming the very equivalence that is at issue.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a multi-method study of human forehead sweating in six supine subjects undergoing passive heating, cooling, and reheating. The authors combine macroscopic measurements (ventilated-capsule sweat rate, galvanic skin conductance, and dielectric stratum corneum hydration) with three microscale imaging modalities (visible-light macro photography, midwave infrared, and optical coherence tomography). They describe a cyclic porewise mode, a transition mode with shallow spreading puddles, and a filmwise mode, and interpret the temporal relationships among SR, GSC, and HYD in terms of these imaged dynamics. They also compare a novel air-jet ventilated capsule with a standard cylindrical capsule and report that the former over-cools the skin and depresses measured sweat rate. The paper further proposes that salts deposited during the first heating stage alter the mode of sweating in the second heating stage, and that sweat at pores forms shallow menisci rather than the hemispherical droplets often drawn in the literature.","tokens_in":20276,"tokens_out":4886,"duration_ms":49615,"significance":"If the main claims hold, the paper offers a valuable advance in linking microscale sweat dynamics to macroscopic physiological signals. The simultaneous application of SR, GSC, HYD, MWIR, and OCT on non-glabrous skin is original and technically demanding; the air-jet capsule comparison is a useful practical result; and the reported staging of porewise, transition, and filmwise modes, with supporting images and movies, is internally consistent and generally agrees with prior literature. The authors are also candid about several limitations, including the inability to resolve salt deposits and the exploratory nature of the fingertip comparison. However, the central temporal claim, that GSC rises before and SR/HYD rise after MWIR-visible pore sweating, depends on an assumption of synchronized onset across a spatially distributed instrument layout that the manuscript does not validate, and the salt-driven mechanism is asserted rather than directly supported. These issues are load-bearing and require attention.","major_comments":[{"comment":"The temporal comparisons between macroscopic signals and imaged pore dynamics assume that the GSC electrodes, the ventilated capsule, and the HYD probe observe the same sweating state as the central MWIR/OCT patch. As described in Section 2.1 and Figure 1b, the GSC electrodes and capsule are attached to the sides of the forehead, and SM S1.3 places the HYD probe 2 to 4 cm from the imaged center. Table 1 then sets the MWIR onset to 10 minutes for all instruments. This is exactly the equivalence that is at issue. The ventilated capsule imposes forced dry airflow that locally cools and dries the skin (the paper itself demonstrates the stronger air-jet version depresses SR in Figure 2d), and the GSC electrodes sit under isotonic gel, which hydrates and occludes the skin. Without a co-location check or other evidence that onset and mode timing are synchronized across the roughly 2- to 4-cm instrumented area, the reported ordering (GSC rises 1.7 +/- 0.8 min before MWIR onset; SR and HYD rise 1.4 +/- 0.7 and 2.7 +/- 1.8 min afterward) is not established. This assumption is load-bearing for the paper's central explanation and should be addressed by co-locating measurements, by reanalyzing the timing with an explicit synchronization check, or by substantially softening the causal interpretation.","section":"Sections 2.1, 3.3, Table 1, SM S1.3"},{"comment":"The salt-deposition mechanism is presented as a key factor explaining the faster filmwise second-stage sweating, but the authors state that the salt deposits were 'undetectable using any of the employed imaging techniques'. This is an acknowledged missing support. The observed differences between first and second heating stages are also consistent with elevated core temperature, residual stratum corneum hydration, or residual sweat under the GSC electrodes, which the authors mention only as contributing to faster onset. As written, point 5 of the Conclusions asserts a mechanistic claim that the paper's own evidence cannot distinguish. The salt-facilitated film-spreading mechanism should be explicitly framed as a hypothesis, and, if retained as a conclusion, should be supported by direct salt visualization or by a control condition such as washing the skin between heating stages.","section":"Section 4.2, Figure 8, Conclusions point 5"},{"comment":"The mode onset times and the claimed correspondences between mode transitions and inflections in SR, GSC, and HYD rely on visual classification of MWIR and OCT images. No quantitative criterion is given for distinguishing cyclic porewise, transition, and filmwise modes, and no inter-rater reliability or sensitivity analysis is reported. Because the alignment of macroscopic inflections with mode transitions is part of the central narrative, a reproducible rule (for example, based on wet-area fraction, MWIR contrast, or OCT-derived film thickness) or at least a sensitivity check on the visual thresholds should be provided.","section":"Section 3.3, Table 1"},{"comment":"The mean lead/lag values (GSC -1.7 +/- 0.8 min, SR +1.4 +/- 0.7 min, HYD +2.7 +/- 1.8 min) are presented without statistical tests. With n = 6 for the first stage and n = 5 for the second stage, the paper should report the full distribution of individual differences and, ideally, a paired comparison or sign test to establish that the ordering is consistent across subjects rather than driven by one or two outliers. As it stands, the mean values and standard deviations do not by themselves support the claim that GSC systematically precedes MWIR-visible pore sweating.","section":"Section 3.3"}],"minor_comments":[{"comment":"The text says the hydration meter probes the outer '~50 mm' of skin; this should be '~50 µm'. The acquisition rate is also given as '2,000 kHz', which is presumably '2 kHz'.","section":"SM S1.3"},{"comment":"There are minor typographical errors: 'human trails' should be 'human trials', and 'separted' in the Figure 8 caption should be 'separated'.","section":"Section 3.2 and Figure 8 caption"},{"comment":"Table 1 is difficult to parse because the first- and second-stage columns are not visually separated and several entries are missing (for example, subject 5's F-mode row and some GSC values). Please reformat the table with clear stage headers and mark missing data explicitly.","section":"Table 1"},{"comment":"The captions state that GSC and HYD measurements are scaled 'as indicated in legend', but the specific scaling factors are not stated in the main text or captions. Please provide the scaling factors so that the reader can interpret the plotted magnitudes.","section":"Figures 3 and 5"},{"comment":"The data availability statement says that MWIR and OCT movies 'can be provided upon request'. Given the reproducibility value of the imaging data, please consider depositing the processed movies and a data summary in a permanent repository.","section":"Data Availability"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid observational contribution with a technically demanding experimental design, and it fits the interdisciplinary scope of the journal. The main concerns are the unvalidated spatial synchronization assumption underlying the temporal ordering claim and the speculative salt mechanism. These are fixable in principle, but they require either new co-location/control data or a careful reframing of the claims. The absence of public data sharing is an additional weakness given the paper's emphasis on imaging evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a careful observational study with two genuinely useful results—the first OCT imaging of thermal sweating on non-glabrous skin, and a clear demonstration that air-jet ventilated capsules locally suppress sweat rate by cooling. The descriptive staging of porewise → transition → filmwise modes is well supported by the images, and the observation that sweat forms shallow menisci rather than hemispherical droplets is a real challenge to a common textbook picture. The paper is worth refereeing, but the macro-micro temporal story has a site-mismatch problem that needs to be addressed.\n\nThe new material: OCT cross-sections of forehead sweating from onset to film, showing pore-confined sweat as flat or slightly convex menisci, then shallow puddles (contact angles around 20–40°), then thin films. The timing data (GSC leads MWIR onset by ~1.7 min; SR and HYD lag) and the salt-deposit interpretation of second-heating dynamics are more speculative, though the authors do flag that the salt deposits were not directly imaged. The air-jet capsule result is a clean, practically important negative: too much local cooling gives artificially low sweat rates.\n\nThe soft spots, in order of severity. First and most important: the GSC electrodes, the ventilated capsule, and the HYD probe are not on the imaged patch. They sit 2–4 cm away, under conditions that are locally perturbing—dry airflow in the capsule, isotonic gel under the electrodes, probe pressure. The paper uses MWIR onset in the central patch as time zero for all instruments and then interprets the ordering (GSC first, SR/HYD after) as a read on duct filling and stratum corneum hydration. That inference assumes the lateral sites sweat on the same schedule as the imaged site, and that the capsule itself does not delay local onset. The paper demonstrates exactly this delay for the air-jet capsule, and the cylindrical capsule also imposes dry airflow. The stress-test note lands; this is a load-bearing assumption, not a minor detail. Second, n=6, one skin site, supine posture, visually classified modes. For an observational study that is acceptable, but it limits the quantitative claims. Third, the salt mechanism is plausible but inferred; the authors say so, but the Discussion leans on it harder than the data warrant. Fourth, the movies are only 'available upon request' rather than archived. Minor: the air-jet pilot is n=1.\n\nWho is this for? People working on sweat sensing, skin biophysics, thermal comfort, and wearable devices. It is a hypothesis-generating characterization, not a confirmed model. I would send it to review, with the co-location issue front and center. A revision that either co-locates macro and micro measurements on a smaller area, or explicitly softens the temporal-ordering claims, would make the paper solid. If the authors cannot co-locate, they should present the macro-micro alignment as suggestive rather than explanatory.","headline":"Careful, image-rich observational study with a genuinely new meniscus observation and a useful capsule-design caution, but the macro-micro temporal story rests on an unverified co-location assumption.","tokens_in":20899,"tokens_out":3809,"would_cite":true,"duration_ms":39909,"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":"Sweating on most skin proceeds through three microscale modes, from porewise cycling to a thin film, and the paper shows that macroscale sweat measurements follow from these imaged dynamics.","keywords":["sweat production","sweat evaporation","droplet dynamics","midwave infrared imaging","optical coherence tomography","galvanic skin conductance","stratum corneum hydration","salt deposits"],"falsifier":"Place the conductance electrodes and the ventilated capsule over the exact forehead patch that the MWIR camera images, or image a second patch adjacent to the sensors, and time pore activation pore-by-pore. If sweat first appears at the sensor-side pores more than about 1.7 minutes before or after the imaged patch, then the claimed ordering of GSC before MWIR before SR and HYD is a site-mismatch artifact rather than a physiological sequence.","tokens_in":19856,"feed_emoji":"💧","tokens_out":7452,"duration_ms":74585,"temperature":0.7,"pith_summary":"This paper sets out to show that visible sweating on non-glabrous skin is not a steady oozing of hemispherical droplets but a staged process: sweat first reaches and evaporates at individual pores in cycles, then spreads into shallow puddles, and finally merges into a thin film. By imaging the forehead with midwave infrared and optical coherence tomography while simultaneously measuring sweat rate, skin conductance, and stratum corneum hydration, the authors argue that the rise times and plateaus of the macroscopic signals line up with these microscale modes. If true, this changes how sweat onset, skin hydration, and sensor-skin contact are understood in thermoregulation, cosmetics, textiles, and wearable devices. It also implies that salts left behind after sweat dries reshape the next round of sweating, making it spread faster as a film.","feed_headline":"Imaging shows sweat spreads as flat films, not droplets","feed_subtitle":"Microscale imaging and skin sensors reveal a three-stage pore-to-film process that explains sweat-rate signals.","key_machinery":"The carrying mechanism is the coupled measurement stack: midwave infrared (MWIR, 3–5 µm) thermography, optical coherence tomography (OCT) cross-sectional imaging, and visible macrophotography, run alongside ventilated-capsule sweat rate, galvanic skin conductance, and dielectric stratum corneum hydration. MWIR sees thin water at individual pores, OCT resolves the cross-sectional shape of menisci, puddles, and films, GSC tracks sweat entering ducts before it reaches the surface, HYD tracks hydration of the outer roughly 50 µm of skin, and SR tracks evaporative flux. The mode sequence (cyclic porewise, transition, filmwise) is the organizing identity that links the microscopic images to the macroscopic curves.","core_discovery":"The paper's central claim is that thermal sweating on non-glabrous skin passes through three imaged modes—cyclic porewise, transition with spreading shallow puddles, and filmwise—and that the area-averaged physiological signals (galvanic skin conductance, sweat rate, and stratum corneum hydration) are explained by these modes. In the porewise mode, sweat periodically fills and evaporates at individual pores; OCT shows the liquid-air interface as a shallow convex or flat meniscus, not the hemispherical droplet of traditional textbook pictures. The transition mode consists of irregular puddles roughly 0.1 to 1 mm deep with contact angles of 20 to 40 degrees, and the filmwise mode is a thin film tens to hundreds of micrometers thick. GSC rises about 1.7 ± 0.8 minutes before MWIR-visible pore activity because it senses sweat filling the ducts, while sweat rate and hydration rise after onset as the number of active pores and local stratum corneum hydration grow. During a second heating after drying, salts deposited from evaporated sweat wick the new sweat into a rapidly spreading film, making mode shifts faster and more variable across subjects.","pith_inferences":["A controlled follow-up comparing second-heating sweat spreading on the same subject with and without gently wiping the skin between cycles would isolate the salt-deposit mechanism from the confounding effect of already-elevated core temperature.","Because the cyclic porewise mode is driven by sympathetic nerve oscillations, MWIR-visible cycling frequency could become a non-invasive, spatially resolved proxy for sudomotor nerve activity in psychophysiology studies.","The finding that local cooling changes sweat rate implies that any ventilated capsule perturbs the quantity it measures; reporting a true local sweat rate may require extrapolation to zero cooling or concurrent skin-temperature measurement."],"forward_implications":["The textbook image of sweat as hemispherical droplets sitting on non-glabrous skin should be replaced by shallow menisci at pores, shallow puddles during transition, and thin films at high sweat rates.","Interpreting GSC and HYD as evidence of uniform stratum corneum hydration from beneath, before sweat reaches the surface, is not supported; hydration begins locally and cyclically at active pores.","Ventilated capsules that use impinging air jets can locally cool the skin and under-report sweat rate, so capsule flow geometry matters for absolute sweat rate values.","After a drying episode, salt residues on skin make subsequent sweat spread as a film rather than puddles, shortening and making more variable the time spent in porewise and transition modes.","Microscopic hairs on the forehead can wick sweat away from pores and participate in evaporation during all modes."],"supporting_citations":[{"why":"Supplies the GSC-plus-HYD measurement protocol and the comparison dataset against which the paper's macroscopic trends are plotted.","marker":"[15]"},{"why":"The authors' earlier wind-tunnel capsule and MWIR study established the cyclic dropwise-to-filmwise mode sequence and the imaging method this work extends.","marker":"[36]"},{"why":"Provides the roughly 5-minute full-hydration time for stratum corneum in constant water contact used to explain the ~14-minute porewise mode duration.","marker":"[33]"},{"why":"The OCT fingertip study that previously showed high-contact sweat droplets, the observation this paper contrasts with its shallow-meniscus finding.","marker":"[60]"},{"why":"Documents pulsatile sweating driven by sympathetic sudomotor activity, the physiological basis for the cyclic porewise mode.","marker":"[76]"},{"why":"Shows salts alter sweat droplet evaporation on artificial surfaces, supporting the paper's salt-deposit mechanism.","marker":"[69]"},{"why":"Demonstrates that evaporation conditions control salt crystal patterns, used to explain subject-to-subject variability in second-stage mode shifts.","marker":"[78]"},{"why":"Provides the physiology of sweat gland function and the wide individual variation in sweat salt concentration that underpins the salt-deposition argument.","marker":"[16]"}],"fun_headline_variants":["Sweat spreads as flat films, not droplets, imaging shows","Three-stage sweat: porewise, puddle, then film","Imaging reveals sweat forms films, not droplets","Sweat's micro journey: pores to puddles to films","Flat films, not droplets: imaging corrects sweat theory"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the patch of forehead being imaged and the nearby skin under the conductance electrodes and the sweat-rate capsule are sweating in the same state at the same times; if sweat onset is spatially patchy across the few centimeters separating them, the claimed ordering of signals collapses.","fun_headline_variants_meta":{"raw":{"variants":["Sweat spreads as flat films, not droplets, imaging shows","Three-stage sweat: porewise, puddle, then film","Imaging reveals sweat forms films, not droplets","Sweat's micro journey: pores to puddles to films","Flat films, not droplets: imaging corrects sweat theory"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000598,"raw_usage":{"total_tokens":2828,"prompt_tokens":1010,"completion_tokens":1818,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":626,"completion_tokens_details":{"reasoning_tokens":1733}},"tokens_in":626,"tokens_out":1818,"duration_ms":11808,"temperature":1.0,"reasoning_tokens":1733,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:20:54.186549+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Place the conductance electrodes and the ventilated capsule over the exact forehead patch that the MWIR camera images, or image a second patch adjacent to the sensors, and time pore activation pore-by-pore. If sweat first appears at the sensor-side pores more than about 1.7 minutes before or after the imaged patch, then the claimed ordering of GSC before MWIR before SR and HYD is a site-mismatch artifact rather than a physiological sequence.","supporting_citations":[],"review_version":1}