{"id":"0b260082-26f8-46ba-a192-65dfe7bd8350","arxiv_id":"2608.02562","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A curvature-based inflection-point criterion classifies circadian waveforms as harmonic; circadian data and the Goodwin model meet this criterion across analyzed parameters.","lead":"The paper defines an oscillation as 'harmonic' if its phase-plane trajectory has no inflection point, and reports that circadian bioluminescence data and the Goodwin oscillator satisfy this criterion. The approach offers a model-free way to characterize circadian waveform shape beyond period and phase.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Post-hoc Gaussian smoothing makes the empirical harmonicity claim fragile; σ=0.1 can annihilate weak periodic inflection points, so the real-data conclusion is not independently established.","rationale":"I agree with the reader's weakest assumption. The central claim has two pillars: the Goodwin model and the experimental recordings. The Goodwin pillar is supported by extensive numerics and a semi-analytic piecewise-linear argument, though the 'throughout' statement ultimately rests on a finite scan and a numerically located passage point. The experimental pillar is weaker: the only way to compute curvature from 48-point-per-cycle digitized data is strong smoothing, and the choice of σ is justified by a model-specific synthetic test. Since the paper's abstract and title present the empirical harmonicity as a main result, this unresolved dependence on the smoothing choice is the most load-bearing concern. A positive outcome of the proposed σ-sweep and injection test would make the empirical claim solid; a negative outcome would require either a principled denoising criterion or a softening of the conclusion. Thus the reader's conditional verdict remains appropriate; no further adjustment is needed.","tokens_in":13912,"tokens_out":8024,"duration_ms":75941,"concrete_test":"Take the digitized cyanobacteria and SCN series and scan σ from 0.02 to 0.3 (normalized period units), detecting curvature sign changes and classifying inflections as periodic if they recur at the circadian frequency over at least three cycles. Then inject a known non-harmonic feature, e.g. add ε sin(3ωt+φ) with ε chosen so the unsmoothed curve has periodic inflections, and measure the smallest σ at which these injected inflections vanish. If this annihilation threshold is below 0.1, then σ=0.1 cannot be relied on to preserve essential inflections and the empirical harmonicity may be an artifact. Cross-check with a model-free denoising approach (penalized spline or AIC-selected Fourier series) and report whether any inflections remain.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central empirical claim (Fig. 2) is that bioluminescence recordings from cyanobacteria and mouse SCN have no inflection point in the (x, ẋ) plane. Because κ requires a third derivative, the analysis is possible only after Gaussian smoothing; the paper selects σ=0.1 after inspecting the data (Methods, 'Noise removal from experimental data') and validates the choice on FitzHugh–Nagumo, where the essential inflection points are large, spike-like features. This validation does not transfer to circadian recordings: a true waveform that is only mildly non-harmonic has weak periodic inflections (curvature near zero), and the same smoothing that removes aperiodic noise will also annihilate them before they can be counted as 'periodic' over the short (≈6-cycle) records. The paper itself concedes that 'the choice of smoothing strength can affect the detection of inflection points' and that statistical methods for extracting essential inflection points should be developed. Therefore the experimental harmonicity is at least partly a consequence of the denoising pipeline, not a demonstrated robust waveform property. This is the load-bearing weakness in the claim that circadian waveforms are harmonic.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper introduces a geometric criterion for classifying oscillatory waveforms: a periodic time series is called 'harmonic' when its trajectory in the (x, ẋ) plane has no inflection point (curvature sign change). The authors derive the curvature condition, illustrate it on Stuart–Landau (harmonic) and FitzHugh–Nagumo (non-harmonic), and apply the criterion to experimental bioluminescence data from cyanobacteria and the mouse SCN, to two published circadian models (Sasai stochastic KaiC, Kim–Forger), and to the Goodwin model. For the Goodwin model they report a numerical parameter scan (n–λ) with no inflections wherever a limit cycle exists, and they give a partial semi-analytic confirmation using a piecewise-linear n→∞ version, checking a sufficient condition for the passage point.","tokens_in":14198,"tokens_out":10476,"duration_ms":101049,"significance":"Should the result hold, the paper offers a useful model-free descriptor of waveform shape, with potential for classifying circadian oscillators and for probing links between network architecture and waveform. Strengths: the numerical scan in Fig. 4D is carefully executed; arbitrary-precision arithmetic is used where ordinary double precision fails; the sufficient condition (Eq. 16) is derived correctly; and code is made available on GitHub. The principal weakness is the load-bearing choice of Gaussian smoothing width for the experimental data: σ=0.1 is selected after inspection and validated only on FitzHugh–Nagumo-type large inflections, so the conclusion that bioluminescence rhythms are harmonic may be an artifact of the denoising pipeline. The semi-analytic Goodwin result is also narrower than the abstract suggests, applying to the z–ż plane for a piecewise-linear limit and relying on numerically located passage points.","major_comments":[{"comment":"The experimental harmonicity claim in Fig. 2 is not robustly established because the Gaussian smoothing width σ=0.1 is selected after inspecting the data ('applying σ=0.1 eliminated the aperiodic inflection points'), and the validation on FitzHugh–Nagumo does not transfer to circadian waveforms: the essential inflections there are large and spike-like, whereas mildly non-harmonic circadian rhythms would have weak inflections that the same smoothing can annihilate. The records are short (~6 cycles), and the paper concedes that smoothing strength can affect detection. Please report a sensitivity analysis of the inflection-point outcome over σ for the real data and validate the protocol on synthetic circadian-like signals with weak periodic inflections; otherwise the empirical conclusion may be an artifact of the pipeline.","section":"Methods, 'Noise removal from experimental data'; Fig. 2 and Fig. S2"},{"comment":"The semi-analytical confirmation is more limited than the abstract's phrase 'confirmed this numerical trend semi-analytically' implies. Condition (16) is a sufficient condition for the absence of inflections for all t, but the passage point of the limit cycle is obtained by numerically solving the simultaneous equations (17)–(18), and the check is performed only for several values of λ. In addition, the proof concerns the z–ż plane of the n→∞ piecewise-linear model, whereas the numerical claim in Fig. 4D concerns the x–ẋ plane of the finite-n Goodwin model. Thus the analytical result does not by itself establish harmonicity throughout the n–λ region. Please state these restrictions explicitly in the abstract and in the section's summary, and avoid the suggestion that the full numerical scan has been proved analytically.","section":"Harmonicity of the piecewise-linearized Goodwin model; Methods, Eq. (16)–(20)"}],"minor_comments":[{"comment":"The sentence 'the Python packagempmath' is missing a space: it should read 'the Python package mpmath'.","section":"Methods, 'Numerical computation of the curvature'"},{"comment":"The gene name 'Bmals' is non-standard; for the mouse, the usual name is 'Bmal1'. Please revise or clarify.","section":"Results, 'Classification of waveforms from circadian clock models'"},{"comment":"The experimental time series were 'collected manually from the published figures' without specifying which figures in the cited papers, and the extracted series are not deposited. Please include the digitized data in the GitHub repository so that Fig. 2 can be reproduced.","section":"Data availability"},{"comment":"The term 'harmonic' is already used in Fourier analysis; consider a short remark distinguishing the new geometric sense from the spectral sense to avoid ambiguity.","section":"Introduction"}],"recommendation":"major_revision","confidential_remarks":"The paper is worth pursuing. The main revision should focus on the smoothing sensitivity and on explicitly stating the limitations of the semi-analytic proof. With those addressed, I would support publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things you should know. First, the paper defines a clean, model-free waveform class: an oscillation is 'harmonic' if its phase-plane trajectory has no inflection point, and this is genuinely new as a tool for classifying circadian waveforms. Second, the empirical headline — that real bioluminescence data from cyanobacteria and mouse SCN are harmonic — is not independently established, because the Gaussian smoothing width σ=0.1 was chosen after inspecting the data and was validated only on FitzHugh–Nagumo, where essential inflections are large. Weak periodic inflections in real data could be destroyed by that same smoothing.\n\nCredit where due. The Goodwin-model part is the strongest. The numerical scan over n and λ (Fig. 4D) is careful, and the piecewise-linear sufficient condition (Eq. 16) is derived correctly. The phase-space picture of the inflection surface (Fig. 6) is a nice geometric addition. The paper is also honest: the Discussion concedes that smoothing strength affects inflection detection and that statistical methods are needed.\n\nSoft spots, in order of severity. The smoothing issue is load-bearing. Manual digitization of published figures at 0.5 h gives only 48 points per cycle; a third derivative from that is extremely noisy, so the result is essentially determined by the smoothing pipeline. The validation on synthetic data with spike-like inflections does not transfer to mild non-harmonicity. This alone means the empirical conclusion should be read as provisional. Second, the 'semi-analytic' confirmation of the Goodwin model covers only the piecewise-linear z–ż plane for a few numerically obtained passage points; the claim of parameter-independent harmonicity is supported mainly by the numerical scan, not by proof. That is fine if stated as a numerical result, but the word 'throughout' overstates it. Both issues are fixable.\n\nWho is this for? Chronobiologists who care about waveform, and any dynamical-systems person interested in a threshold-free curvature-based classification. It deserves a serious referee; the idea is simple, checkable, and likely to be cited if the empirical part is tightened. My recommendation: engage with it, but send it back for revision focused on a principled smoothing-robustness analysis and softer wording of the empirical claim.","headline":"A genuinely new waveform criterion and a careful Goodwin-model analysis, but the empirical 'harmonic circadian waveforms' claim rests on a post-hoc smoothing choice that was not independently validated.","tokens_in":14640,"tokens_out":2397,"would_cite":false,"duration_ms":24006,"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":"The paper defines a circadian waveform as harmonic when its phase-plane trajectory has no inflection point, and reports that cyanobacterial and mouse-SCN rhythms, most modeled clock variables, and every Goodwin-model limit cycle examined pa","keywords":["circadian rhythm","waveform","inflection point","curvature","harmonic oscillation","Goodwin model","phase-plane trajectory","Gaussian smoothing"],"falsifier":"Record a high-resolution circadian bioluminescence time series from cyanobacteria or the mouse SCN, apply curvature analysis either without smoothing or with a data-driven smoothing width, and check whether a sign change in κ recurs in every cycle; a single periodically recurring inflection point in the real data would falsify the central empirical claim. Likewise, solving the piecewise-linear limit-cycle equations to high precision and finding a λ for which the passage point violates inequality (16) would overturn the semi-analytic confirmation.","tokens_in":13788,"feed_emoji":"🕐","tokens_out":7092,"duration_ms":69048,"temperature":0.7,"pith_summary":"The paper aims to establish that circadian oscillations are geometrically smooth—harmonic in a phase-plane sense—rather than pulse-like. It defines a waveform as harmonic when the curve traced by (x, ẋ) never has an inflection point, and shows that real bioluminescence recordings from cyanobacteria and the mouse suprachiasmatic nucleus, the core variables of several circadian clock models, and the limit cycles of the minimal negative-feedback Goodwin model all satisfy this condition. The claim matters because waveform shape has been linked to clock function—photoperiod sensing, fat-cell differentiation, and the duration of rest—yet most analyses have focused on period and phase. The proposed criterion is model-free, requiring only a time series of one variable, so it offers a new way to classify biological oscillations generally.","feed_headline":"Circadian clocks run on curves with zero inflection points","feed_subtitle":"A curvature-based check of cyanobacteria and mouse SCN rhythms finds no inflection points, linking smooth waveforms to negative-feedback clo","key_machinery":"The central object is the curvature κ(t) = (ẋ ...x − ẍ²)/(ẋ² + ẍ²)^(3/2) along the phase-plane trajectory Γ(t) = (x(t), ẋ(t)). A waveform is classified harmonic when the curvature never changes sign, i.e. the trajectory never touches the condition ẋ d³x/dt³ = ẍ², which corresponds geometrically to an inflection point and locally to exponential-like behavior. The method detects inflection points as sign changes of κ, and uses Gaussian smoothing at a width chosen so that only aperiodic, noise-induced inflection pairs are annihilated while period-recurring ones survive.","core_discovery":"Using the curvature of the trajectory in the plane of a variable and its time derivative, the authors define an oscillation as harmonic when no inflection point appears on the limit cycle. They report that bioluminescence rhythms of cyanobacteria and the mouse SCN, the core mRNA variables of a detailed mammalian circadian clock model, the mean KaiC phosphorylation rhythm of a stochastic cyanobacterial model, and the Goodwin negative-feedback model are all harmonic. For the Goodwin model, harmonicity holds throughout the two-parameter region that produces a limit cycle; a piecewise-linearized version confirms this semi-analytically, reducing the absence of inflection points to an explicit ine","pith_inferences":["An implicit consequence is that comparing harmonic classes across variables within a clock network could expose low-pass filtering steps: the paper notes a simple cascade attenuates high frequencies, so downstream species may show smoother waveforms than upstream ones.","The hourglass-shaped surface of inflection points in the Goodwin phase space suggests a geometric explanation for harmonicity; testing whether limit cycles in other feedback oscillators avoid analogous surfaces would connect waveform shape to oscillator topology.","Because the empirical result depends on a chosen Gaussian smoothing width, a data-driven procedure for selecting that width would extend the method to noisier recordings, such as human actigraphy or tissue-level bioluminescence.","Applied beyond chronobiology, the criterion offers a model-free way to separate spike-like (relaxation) oscillations from smooth ones in cardiac or neural data, where the same waveform question arises."],"forward_implications":["Circadian waveforms can be characterized without fitting a dynamical model: one observed variable, its derivatives, and the sign of a single curvature expression suffice.","The minimal negative-feedback Goodwin model is harmonic for every parameter pair that yields a limit cycle, making harmonicity a generic property of that oscillator structure rather than a fine-tuned one.","Not every variable in a circadian model is harmonic—some species, such as one of the Cryptochrome mRNA forms, do exhibit inflection points—so harmonic class is a property of individual variables and may trace signal-processing stages.","In the piecewise-linear limit of the Goodwin model, the absence of inflection points is equivalent to an explicit inequality on the limit cycle's crossing point, giving a semi-analytic handle on waveform shape.","The same curvature criterion can be applied to any self-sustained biological rhythm, replacing a qualitative harmonic-versus-relaxation label with a computable, data-driven classification."],"fun_headline_variants":["Curvature shows circadian rhythms are harmonic","No inflection points found in circadian rhythms","Harmonic circadian rhythms: a curvature-based criterion","Smooth curves: circadian clocks are harmonic","Curvature test reveals harmonic circadian waveforms"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The empirical harmonicity claims rest on the assumption, stated in the Methods' noise-removal section, that Gaussian smoothing with width σ=0.1 removes exactly the aperiodic, noise-induced inflection points and preserves every real periodic one; the semi-analytic Goodwin result additionally relies, as noted in the piecewise-linear analysis section, on locating the limit cycle's crossing point numerically rather than proving that the inequality holds for all λ.","fun_headline_variants_meta":{"raw":{"variants":["Curvature shows circadian rhythms are harmonic","No inflection points found in circadian rhythms","Harmonic circadian rhythms: a curvature-based criterion","Smooth curves: circadian clocks are harmonic","Curvature test reveals harmonic circadian waveforms"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002318,"raw_usage":{"total_tokens":8759,"prompt_tokens":710,"completion_tokens":8049,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":454,"completion_tokens_details":{"reasoning_tokens":7997}},"tokens_in":454,"tokens_out":8049,"duration_ms":50410,"temperature":1.0,"reasoning_tokens":7997,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T04:38:02.312845+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Record a high-resolution circadian bioluminescence time series from cyanobacteria or the mouse SCN, apply curvature analysis either without smoothing or with a data-driven smoothing width, and check whether a sign change in κ recurs in every cycle; a single periodically recurring inflection point in the real data would falsify the central empirical claim. Likewise, solving the piecewise-linear limit-cycle equations to high precision and finding a λ for which the passage point violates inequality (16) would overturn the semi-analytic confirmation.","supporting_citations":[],"review_version":1}