{"id":"07cd30f2-2d8d-459b-84a1-8bd1a54508b9","arxiv_id":"2502.09342","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"PSR J1935+1616 shows 127 microsecond micropulses in 9.7% of FAST single pulses, a 232 microsecond quasi-period, and a preference for one of four pulse-shape modes.","lead":"An analysis of archived FAST observations of pulsar PSR J1935+1616 finds that roughly one in ten single pulses contains short micro-bursts, about half of those with a regular 232 microsecond rhythm. The micro-bursts appear most often in pulses with only a central bright component, adding a new observational constraint on pulsar emission physics.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed micropulse fractions, 127.63 µs widths, and 231.77 µs periods rest on a fifth-order-polynomial residual pipeline never validated against noise or synthetic signals; at 49.512 µs sampling these are near the resolution limit, so a processing artifact cannot be excluded.","rationale":"The reader's conditional verdict and weakest-assumption identification are on target. The most load-bearing premise is that the 49.512 µs sampling and the fifth-order-polynomial smoothing/residual procedure resolve genuine intrinsic micropulse structure without creating or destroying it. The paper presents no end-to-end validation of this premise: no noise-only null test, no synthetic injection-recovery, no explicit accounting for dispersion smearing, scattering, or the sampling window. The reported timescales are only a few bins wide, making the analysis especially sensitive to such artifacts. If the residual pipeline generates false microstructure, then the 9.69% MP fraction, the 231.77 µs quasi-period, the energy double-Gaussian claim, and the mode-A correlation all lose their observational basis. I do not see this as grounds for rejection: the pulsar has prior reports of micropulses, the example pulse shows plausible ACF/PSD features, and the issue is directly testable with synthetic data. However, because the central numbers are unvalidated at the resolution limit, the conditional verdict should stand until the injection-recovery and null tests are performed. No ad hominem is intended; the critique is purely on the measurement chain.","tokens_in":14710,"tokens_out":4705,"duration_ms":53214,"concrete_test":"Run the exact §3.1.1 pipeline on (i) synthetic noise-only single pulses with no injected microstructure and (ii) synthetic pulses with injected Gaussian micropulses of known widths/periods spanning 50–300 µs and 150–350 µs, using the measured off-pulse noise and the average profile. If the false-positive MP/QMP fraction is comparable to 9.69%/5.20%, or if recovered widths are pinned near 127 µs regardless of injection, the detections and quoted timescales are pipeline artifacts; if false positives are below ~1% and injected widths/periods are recovered within errors, the central claim is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.1.1 defines microstructure from the residual of a fifth-order polynomial fit to each single pulse; widths come from Eq. (1)'s second central moment and quasi-periods from ACF/PSD peaks. The native sampling time is 49.512 µs (§2), so the reported width of 127.63 µs is only ~2.6 bins and the period of 231.77 µs is ~4.7 bins. At this resolution the residual method can create apparent short-timescale structure: a low-order polynomial fit to a three-component pulse has a characteristic curvature scale, and its residuals against noise can contain quasi-oscillatory features at the exact timescales claimed. The paper gives no null test on noise-only pulses and no injection-recovery test with synthetic micropulses of known width and period. It also does not state the expected dispersion-smearing and scattering timescales or deconvolve the 49.512 µs sampling window, so the measured width is not demonstrated to be intrinsic. Because MP/QMP classifications feed every downstream claim (9.69% fraction, width, period, energy double-Gaussian, mode-A coupling), an artifact in this step would invalidate the central conclusion. Agreement with Popov's 150 µs is suggestive but does not validate the method.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes 9,998 single pulses of PSR J1935+1616 from FAST archival data. It reports that 969 pulses (9.69%) contain micropulses with characteristic width 127.63(+70.74/-46.25) microsec, that 520 of these show quasi-periodic micropulses with period 231.77 +/- 9.90 microsec, and that micropulse occurrence is highest in morphological mode A and lowest in mode D. Additional results concern circular-polarization quasi-periodicity (244.70 microsec), single-pulse energy distributions described by single or double Gaussians, waiting-time statistics, and polarization fractions. The data reduction uses DSPSR/PSRCHIVE; the microstructure analysis is based on a fifth-order polynomial residual method with ACF/PSD analysis, as described in Sections 2 and 3.1.1.","tokens_in":15006,"tokens_out":5913,"duration_ms":57179,"significance":"If the timescale measurements are reliable, the paper would provide one of the few large-sample measurements of micropulse microstructure in a normal pulsar, plus a possible coupling between microstructure and single-pulse morphology. Strengths include the use of high-sensitivity FAST archival data, the large number of single pulses, the explicit data-processing description, and the presentation of quantitative distributions for widths, periods, energies, and polarization. The main caveat is that the short-timescale quantities are near the 49.512 microsec sampling limit and are not validated by noise-only or injection-recovery tests, so the central claim of intrinsic micropulse timescales is not yet fully supported.","major_comments":[{"comment":"The central claim that the observed residuals correspond to intrinsic micropulses is not validated. At 49.512 microsec sampling, the reported width of 127.63 microsec spans only ~2.6 samples and the reported quasi-period of 231.77 microsec spans ~4.7 samples. The fifth-order polynomial residual method can produce quasi-oscillatory residuals when fitting a smooth envelope to a three-component profile, and the paper gives no test on noise-only pulses, no injection-recovery of synthetic micropulses with known width and period, and no estimate of dispersion-smearing or scattering timescales at 1250 MHz. Without such controls, the measured width and period cannot be distinguished from an artifact of the smoothing procedure. Please add a null test using off-pulse noise and an injection-recovery test, and quantify the smearing timescales.","section":"Section 3.1.1, Eq. (1)"},{"comment":"The criterion for classifying a single pulse as containing micropulses is not defined quantitatively. The statement that 'The ACF curve of the residuals shows a change in slope at 106.5 microsec, indicating the presence of microstructures' is not an operational detection rule. The fraction 969/9998 is therefore not reproducible without the specific code and thresholds. Please specify the algorithm, including how a 'change in slope' is detected and how the 'significant periodic components' are identified in the PSD and ADP.","section":"Section 3.1.1"},{"comment":"The statement that normal pulses (NP) follow a single Gaussian energy distribution is inconsistent with Table 3, which lists S1 parameters for NP (A=0.0017, mu=0.4000, sigma=0.1200). Since the single- versus double-Gaussian distinction is a central result, the paper must report a formal model comparison (e.g., likelihood-ratio test or BIC) showing that the S1 component is insignificant for NP and significant for MP and mode A. Similarly, the phrase 'significantly different' for mode A versus other modes is not supported by any statistical test.","section":"Section 3.2 / Table 3"},{"comment":"The K-S test results are reported incorrectly. The text states 'Dn = 0.59 > 0.05 and P = 0 < 0.05', which compares the test statistic to the significance level rather than to the proper critical value, and the same issue occurs in the second K-S test. The test statistic should be compared to critical values that depend on the sample sizes, and p-values should be reported with a finite nonzero value or a clear threshold statement. Please redo these tests and state the sample sizes used.","section":"Section 3.1.1"},{"comment":"The claimed correlation between micropulse occurrence and morphological mode may be influenced by pulse brightness, because micropulse detection requires resolving short-timescale fluctuations and is easier in higher-S/N pulses. Mode A has a relatively strong central component and mode D has the weakest central component, so the trend in MP fractions could partly reflect S/N rather than a physical coupling. To support conclusion 4, please test whether the MP fraction depends on mode after matching single-pulse energy or by including energy as a covariate.","section":"Section 3.2 / Table 2"}],"minor_comments":[{"comment":"The word 'tatol' should be 'total'.","section":"Section 3.1.3"},{"comment":"'Fllowing' should be 'Following'; 'Usingthe first minimumminimal' should be rewritten as a complete phrase; and 'M_i is the the amplitude' contains a duplicated article.","section":"Section 3.1.1"},{"comment":"The caption says 'PSR J1933+1616' but the paper consistently refers to PSR J1935+1616.","section":"Figure 6 caption"},{"comment":"The statement 'Approximately 5.20% of these single pulses with micropulses exhibit quasi-periodicity' is inconsistent with the numbers: 520 is 5.20% of the total 9,998 pulses, not of the 969 micropulse-containing pulses. The fraction of MP pulses with quasi-periodicity is about 53.7%.","section":"Section 5 / Conclusion 1"},{"comment":"The high-pass filtering used to remove red noise is mentioned only briefly, without specifying the filter type or cutoff. Please provide enough detail for reproducibility.","section":"Section 3.1.2"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses an interesting question and uses a good data set, but the central timescale measurement needs validation before it can support the main conclusions. If the authors provide the requested null tests, injection-recovery tests, and corrected statistical reporting, the paper could become publishable. The manuscript also needs careful editorial work for numerous typographical errors and internal inconsistencies."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a solid, workmanlike observational paper: it gives the first FAST-based look at microstructure in PSR J1935+1616, reports a new micropulse width (127.63 us) and quasi-period (231.77 us) in total intensity, finds quasi-periodicity in circular polarization in 208 pulses, and builds a four-mode morphology classification that correlates with micropulse occurrence. The writing is clear, the data description is adequate, and the comparisons to earlier work (Popov 150 us, Mitra 0.4 ms) are appropriate. Credit where it's due: this is exactly the kind of empirical characterization the pulsar microstructure field needs more of.\n\nThe main soft spot is real and worth taking seriously. The detection pipeline relies on subtracting a fifth-order polynomial fit from each single pulse and then analyzing the residuals with ACF/PSD. At the native 49.512 us sampling, the claimed 127 us width is only ~2.6 bins and the 232 us period is ~4.7 bins. The paper gives no noise-only null test and no injection-recovery of synthetic micropulses, so we cannot rule out that the polynomial residual method creates or biases short-timescale structure. That does not mean the results are wrong—the agreement with Popov's 150 us suggests they are not—but it means the headline numbers are not yet demonstrated to be intrinsic. The mode classification uses a data-defined 10-rms threshold, and the double-Gaussian energy decomposition is presented without model comparison (e.g., BIC), so those parts are also somewhat arbitrary. One smaller point: calling 231.77 us \"significantly shorter\" than Mitra's 0.4 ms is overstated, since Mitra's error bar is ±0.2 ms.\n\nThe paper is descriptive and does not overclaim mechanism; the correlation between micropulses and mode A is at least visually plausible from Table 2, though selection effects deserve a robustness check. All of these concerns are addressable with additional tests, not fatal flaws. The paper deserves a serious referee and, with revisions, likely publication.\n\nFor peer review: accept it, but require the authors to add null tests on noise-only pulses and injection-recovery of synthetic micropulses, to quantify scattering and dispersion smearing, and to discuss the resolution limit explicitly. Those additions would turn a promising but under-validated study into a properly grounded one.","headline":"A useful single-pulsar study with new FAST-era measurements, but the headline timescales rest on a detection pipeline that lacks null tests and sits near the resolution limit.","tokens_in":15642,"tokens_out":2596,"would_cite":true,"duration_ms":29457,"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":"Archival FAST observations of PSR J1935+1616 find that 9.69% of single pulses contain micropulses with a characteristic width of about 128 microseconds, that roughly half of those are quasi-periodic at about 232 microseconds, and that…","keywords":["single-pulse morphology","micropulses","quasi-periodic micropulses","PSR J1935+1616","pulsar mode changing","circular polarization","pulse energy distribution","FAST"],"falsifier":"Re-analyze the same 9,998 FAST single pulses without the fifth-order polynomial envelope subtraction, using an alternative high-pass filter and an independent de-dispersion and RFI-excision chain: if the ACF slope break near 106.5 microseconds and the PSD/ADP peaks near 3.9 kHz disappear or shift by more than the quoted uncertainties, then the claimed micropulse width and quasi-period are artifacts of the processing chain. A faster-sampled follow-up observation, for example at 10-microsecond resolution, should resolve the same 127-microsecond widths and 232-microsecond periods if they are intrinsic.","tokens_in":14499,"feed_emoji":"📡","tokens_out":8346,"duration_ms":78795,"temperature":0.7,"pith_summary":"Archival FAST observations of PSR J1935+1616 reveal that 9.69% of 9,998 single pulses contain microstructure: 969 pulses have characteristic micropulse widths of $127.63^{+70.74}_{-46.25}$ microseconds, and 520 of those pulses show quasi-periodic micropulses with a period of $231.77 \\pm 9.90$ microseconds. The same short-timescale structure appears in circular polarization in 208 pulses, with a period of $244.70^{+45.66}_{-21.05}$ microseconds and a width of $106.52 \\pm 46.14$ microseconds. The pulses separate into four morphological modes (A, B, C, D) based on the strength of the leading and trailing profile components; micropulse-bearing pulses are most common in mode A (562 of 3893) and rarest in mode D (20 of 819). If correct, these numbers establish intrinsic emission timescales for this pulsar and tie the microstructure mechanism to the mechanism that switches single-pulse profile morphology.","feed_headline":"Micropulses appear in 9.7% of J1935+1616 pulses","feed_subtitle":"FAST archive data tie 127-microsecond micropulses and 232-microsecond quasi-periods to pulse-shape modes.","key_machinery":"The machinery is a time-series decomposition of each single pulse: a fifth-order polynomial fit defines the smooth sub-pulse envelope, and subtracting it leaves a residual that should contain any microstructure. The residual is then characterized three ways: an autocorrelation function (ACF), whose slope break and oscillation troughs identify width and quasi-period; the power spectrum of the residual (PSD); and the power spectrum of the ACF derivative (ADP), whose Gaussian peaks give the quasi-frequency. The ACF slope break near 106.5 microseconds and the PSD/ADP peaks near 3.9 kHz are the observable signatures that convert raw intensities into the claimed width and quasi-period. Micropulse widths are assigned with the second central moment of each fluctuation. This chain is what carries the argument from raw intensity time series to the quoted timescales.","core_discovery":"The paper's central discovery is that PSR J1935+1616, observed for one hour with FAST at 1250 MHz and 49.512-microsecond sampling, produces a population of single pulses whose sub-pulse envelopes contain short-timescale intensity fluctuations that are not noise. Using an established identification method, 969 of 9,998 pulses (9.69%) were classified as containing micropulses; their widths, measured by the second central moment of each fluctuation, follow a log-normal distribution with characteristic width $127.63^{+70.74}_{-46.25}$ microseconds. In 520 of these pulses the fluctuations are quasi-periodic, with a single-Gaussian period distribution centered at $231.77 \\pm 9.90$ microseconds, and in 208 pulses the circular-polarization fluctuations are also quasi-periodic at $244.70^{+45.66}_{-21.05}$ microseconds. The pulses classified into morphological modes A through D differ in energy distribution: mode A follows a double Gaussian, while modes B, C, and D follow a single Gaussian; the fraction of micropulse-bearing pulses differs across modes, from 562 of 3893 in mode A down to 20 of 819 in mode D. The paper concludes that micropulse emission is a real, quasi-periodic emission component and that its occurrence is coupled to the pulsar's single-pulse morphology.","pith_inferences":["Extension beyond the paper: because the quoted 127.63-microsecond width is only about 2.6 times the 49.512-microsecond sampling interval, the true intrinsic widths may be narrower; a faster-sampled observation of J1935+1616 would test whether 127 microseconds is a resolution-broadened upper limit.","Extension beyond the paper: the hint of phase-locked microstructure seen in the folded micropulse profile, if real, would be a new phenomenon; it could be tested by stacking many micropulse-selected pulses and checking whether the modulation phase is stable across rotation phases and observation epochs.","Extension beyond the paper: searching FAST archive data at other radio frequencies for the 231.77-microsecond quasi-period would discriminate between a geometric origin, which should be frequency-independent, and a plasma-propagation origin, which should not."],"forward_implications":["If the claim is correct, the 127.63-microsecond width and 231.77-microsecond quasi-period are intrinsic to the pulsar's emission process, so any viable micropulse mechanism must produce these timescales at 1250 MHz.","The measured quasi-period is shorter than the earlier $0.4 \\pm 0.2$ ms reported at 1.5 and 4.5 GHz, implying that either the quasi-period is frequency-dependent or the earlier measurements were limited by coarser sampling.","The waiting-time distribution of micropulse-bearing pulses is Weibull with shape $k = 0.66$, meaning micropulse production is clustered in time rather than following a simple Poisson process.","Pulse-energy distributions separate cleanly: mode A and micropulse-bearing pulses show double-Gaussian energy, while modes B, C, and D show single-Gaussian energy, confirming that the four-mode morphological scheme captures a real physical distinction.","The rarity of micropulses in mode D (20 of 819) and their prevalence in mode A (562 of 3893) implies that the same magnetospheric condition that suppresses the outer profile components also favors the production of microstructure."],"supporting_citations":[{"why":"Supplies the identification method and the second-central-moment formula used to measure micropulse widths.","marker":"Zhao et al. (2023)"},{"why":"Provides the earlier quasi-periodicity measurement for this pulsar at 1.5 and 4.5 GHz that the new 231.77-microsecond period is compared against.","marker":"Mitra et al. (2016)"},{"why":"Provides the earlier ~150-microsecond micropulse width measurement in PSR B1933+16 that the new 127.63-microsecond width is checked for consistency with.","marker":"Popov et al. (2002b)"},{"why":"Supplies the PSD and power-spectrum-of-ACF-derivative technique used to detect quasi-periodicity.","marker":"Lange et al. (1998)"},{"why":"Provides the de-dispersion and folding software that produced the single-pulse stacks from FAST data.","marker":"van Straten & Bailes (2011)"},{"why":"Provides the pulsar analysis tools used for radio-frequency-interference removal.","marker":"Hotan et al. (2004)"},{"why":"Supplies the ephemeris used to fold the single pulses.","marker":"Manchester et al. (2005)"},{"why":"Documents the FAST telescope whose archived observations are the data source.","marker":"Nan et al. (2011)"}],"fun_headline_variants":["9.7% of J1935+1616 pulses show micropulses","231-µs quasi-period found in J1935+1616 micropulses","Micropulse-bearing pulses follow double Gaussian energy","Circular polarization quasi-period mirrors micropulse rhythm","Pulsar morphology modes predict micropulse occurrence"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that after 49.512-microsecond sampling, de-dispersion, and radio-frequency-interference cleaning, the true short-timescale structure is not broadened or destroyed by scattering or smearing, and that the fifth-order polynomial smoothing plus ACF/PSD residual analysis neither creates nor suppresses the features.","fun_headline_variants_meta":{"raw":{"variants":["9.7% of J1935+1616 pulses show micropulses","231-µs quasi-period found in J1935+1616 micropulses","Micropulse-bearing pulses follow double Gaussian energy","Circular polarization quasi-period mirrors micropulse rhythm","Pulsar morphology modes predict micropulse occurrence"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001302,"raw_usage":{"total_tokens":5452,"prompt_tokens":1229,"completion_tokens":4223,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":845,"completion_tokens_details":{"reasoning_tokens":4137}},"tokens_in":845,"tokens_out":4223,"duration_ms":29943,"temperature":1.0,"reasoning_tokens":4137,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T21:47:49.742614+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-analyze the same 9,998 FAST single pulses without the fifth-order polynomial envelope subtraction, using an alternative high-pass filter and an independent de-dispersion and RFI-excision chain: if the ACF slope break near 106.5 microseconds and the PSD/ADP peaks near 3.9 kHz disappear or shift by more than the quoted uncertainties, then the claimed micropulse width and quasi-period are artifacts of the processing chain. A faster-sampled follow-up observation, for example at 10-microsecond resolution, should resolve the same 127-microsecond widths and 232-microsecond periods if they are intrinsic.","supporting_citations":[{"cited_title":"1998, , 332, 111","cited_arxiv_id":null,"evidence_quote":"Supplies the PSD and power-spectrum-of-ACF-derivative technique used to detect quasi-periodicity."}],"review_version":1}