{"id":"0f3cf628-5873-4650-aa9c-ef41fbbda0a2","arxiv_id":"1908.06408","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Using the Method of Critical Fluctuations, the authors report 100% non-criticality in 10 myocardial infarction ECGs and 88% criticality in 25 healthy-control ECGs.","lead":"This paper applies a physics-based test for critical fluctuations to 35 human electrocardiograms from a public database. The test labeled all 10 heart-attack records as non-critical and 22 of 25 healthy records as critical, which the authors interpret as possible hidden misdiagnoses.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The conclusion that the 3 discordant 'healthy' ECGs are misdiagnoses rests entirely on the untested equivalence criticality=health, which the paper's own closing disclaimer disavows.","rationale":"The reader's weakest-assumption analysis correctly identifies the core problem: the criticality-health equivalence is asserted from prior work and never validated here. My reading of the full text confirms this, and the paper's own final disclaimer makes the missing link explicit. The disagreements between MCF and clinical labels cannot bear the interpretive weight placed on them unless the method is independently validated as a diagnostic. I additionally note that the manual segment selection and threshold scanning provide ample room for confirmation bias, which reinforces the need for a pre-registered, automated replication. This does not change the reader's CONDITIONAL verdict: the reported percentages may be real, but the diagnostic interpretation cannot be accepted until the criticality-health link is tested against an independent clinical outcome and the MCF pipeline is made deterministic. The appropriate next step is a blinded, automated, outcome-labeled validation cohort, not a rejection of the exploratory method itself.","tokens_in":6661,"tokens_out":6370,"duration_ms":70909,"concrete_test":"Take a new ECG cohort (n≥100) with hard clinical endpoints (e.g., subsequent myocardial infarction or revascularization within 5 years), pre-register an automated MCF threshold rule (fixed V_L definition and deterministic V_U scan), and apply it blinded to endpoints. If MCF 'critical' status does not predict the endpoint with sensitivity and specificity significantly above chance, the 12% misdiagnosis conclusion is unsupported; if it does predict the endpoint, the criticality-health link would gain independent support.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing premise is the equivalence of physical criticality and cardiac health, stated as: 'When heart tissue is functioning properly (healthy state), then it is in a physical critical state.' This is imported from the authors' prior frog-heart study and is never independently tested on human data in this manuscript. The central conclusion that 'in approximately 1 out of 10 ECGs which presented the typical characteristics of a healthy ECG, the diagnosis may not be accurate' follows only if a non-critical MCF result is a validated marker of disease. Without that validation, the 3 discordant healthy-control records (p165, p245, p242) are just as plausibly MCF false negatives as misdiagnoses. The paper itself concedes: 'this work interprets the heart operation in terms of Physical Critical Phenomena and is not involved with Medical Physiology. However, a link between the two descriptions is necessary and is the subject of further work.' This admission directly undermines the diagnostic claim. In addition, the classification procedure has substantial manual freedom: stationary segments are chosen by eye, the fixed point V_L is located visually, V_U is scanned without a pre-specified rule, and 'p3≈0' has no stated cutoff. These researcher degrees of freedom can inflate the reported 88% and 100% agreement rates and make them hard to reproduce independently.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript applies the Method of Critical Fluctuations (MCF) to human ECG time-series from the PTB Diagnostic ECG Database. For each ECG, the authors manually select stationary segments, identify a fixed point V_L and scan an upper voltage V_U, then fit the laminar-length distribution with a power-law-with-exponential form to estimate exponents p2 and p3. They report that 22 of 25 healthy-control records (88%) satisfy their criticality condition (1 < p2 < 2, p3 approximately 0), while all 10 myocardial-infarction records are non-critical (100%). They interpret the three discordant healthy-control records as possible misdiagnoses and also present the autocorrelation function of the record p.121 as a candidate signature of optimal cardiac function.","tokens_in":6894,"tokens_out":4584,"duration_ms":46307,"significance":"If the equivalence between physical criticality and cardiac health were established on human data, the approach could offer a non-invasive, physiology-based screening marker, and the use of a public database with clearly reported agreement rates is a strength. The descriptive rates (88% and 100%) are a useful pilot observation. However, the central diagnostic conclusion depends on a criticality-health link imported from earlier frog-heart work, and the manual segmentation and threshold choices are not yet specified well enough for independent reproduction. As it stands, the paper is a promising but preliminary demonstration rather than a validated diagnostic claim.","major_comments":[{"comment":"The conclusion that 'in approximately 1 out of 10 ECGs which presented the typical characteristics of a healthy ECG, the diagnosis may not be accurate' rests on the premise, stated in the Introduction, that 'When heart tissue is functioning properly (healthy state), then it is in a physical critical state.' This equivalence is imported from the earlier frog-heart study [10] and is never independently validated on human data in this manuscript. The paper itself concedes in the Conclusions that 'this work interprets the heart operation in terms of Physical Critical Phenomena and is not involved with Medical Physiology.' Consequently, the three discordant healthy-control records (p165, p245, p242) are as plausibly MCF false negatives as they are misdiagnoses; the central diagnostic claim needs either an external validation of the criticality-health link or a clearly weakened formulation.","section":"Introduction and Conclusions"},{"comment":"The classification protocol has substantial manual freedom. The fixed point V_L is 'localized' visually, the upper boundary V_U is moved to a new position without a pre-specified rule, stationary segments are chosen by eye using 'the criterion of stationarity,' and the criticality condition 'p3 approximately 0' is never given a numerical threshold: in Fig.4a, p3=0.006 is treated as critical, while p3=0.29 is treated as non-critical, but no cutoff is defined. These degrees of freedom can inflate the reported 88% and 100% agreement rates and make independent reproduction difficult. Please specify an algorithmic protocol, including a p3 cutoff, a stopping rule for the V_U scan, and a stability criterion for segment selection, and apply it blinded to the diagnostic labels.","section":"MCF application steps (bulleted list) and Fig.4"},{"comment":"The headline rates are reported without statistical quantification. With 10 infarction cases, the 100% agreement rate has an exact 95% confidence interval of approximately [69%, 100%], so the claim of 'absolute agreement' is compatible with a wide range of true sensitivity. Please report confidence intervals or exact binomial tests for both rates, and ideally evaluate the procedure on a held-out or independent set rather than on the same 35 records used to illustrate the method.","section":"Results, 'We have analyzed 25 ECGs...' and 'We have analyzed 10 ECGs...'"}],"minor_comments":[{"comment":"The text refers to 'Eq. (11)' where it should refer to Eq. (1).","section":"Equation (1) and following sentence"},{"comment":"Reference [16], cited for the first MCF application to human ECG, is titled 'The Earth as a living planet: Human-type diseases in the earthquake preparation process'; please clarify whether this is the correct source or provide the correct prior human-ECG reference.","section":"Reference [16]"},{"comment":"The time-axis labels are inconsistent between Fig.2 ('t(1/1K s)') and Fig.3 ('t(10^{-3} s)'), which will confuse readers about the sampling scale.","section":"Figures 2 and 3"},{"comment":"The phrase 'In contrary' should be changed to 'In contrast'.","section":"Abstract"},{"comment":"The title of reference [5] contains 'mFractal'; this appears to be a typo for 'Fractal'.","section":"Reference [5]"},{"comment":"The autocorrelation analysis is presented without a quantitative criterion linking the observed 'symmetries' to criticality; please mark it as illustrative or add a quantitative measure.","section":"Fig.7 and autocorrelation discussion"},{"comment":"The condition 'p3 approximately 0' should be stated as a numerical criterion (for example, p3 below a specified tolerance) to avoid ambiguity in reproducing the method.","section":"Criticality condition, Eq. (3)"}],"recommendation":"major_revision","confidential_remarks":"The paper's methodological basis relies heavily on self-citations, and the health-criticality equivalence is inherited from earlier work rather than established here. I would not reject on that basis alone, but the authors should be pressed to either validate the link in this dataset or clearly limit the conclusions. The more actionable concern is the manual-selection and threshold ambiguity, which should be addressed with a concrete protocol and blinded evaluation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the quick read: the paper does add something new—a 35-record statistical extension of MCF to human ECGs—but its main conclusion, that roughly one in ten healthy-labeled ECGs may be misdiagnosed, rests on an equivalence that the authors themselves disavow in the final paragraph. Referee, but expect heavy revision.\n\nThe genuinely new piece is the sample: 25 healthy controls and 10 myocardial infarction records from PTB, with 22/25 critical and 10/10 non-critical. That is a checkable descriptive result and a fair extension of their earlier single-case demonstration. The method write-up is clear enough for someone else to attempt replication, and they acknowledge the two-lobe segmentation issue.\n\nThe soft spot is load-bearing. The claim that a non-critical MCF result means the 'healthy control' label is wrong presupposes that criticality equals cardiac health—an assumption inherited from frog-heart work and never tested on humans here. The paper's final disclaimer says this work is not involved with Medical Physiology and a link is necessary. You cannot in the abstract say 12% of healthy ECGs may be misdiagnosed and in the conclusion disclaim the physiological link. The three discordant cases (p165, p245, p242) are just as plausibly false negatives of MCF.\n\nThere are also reproducibility concerns that are real but not fatal on their own: segment selection is by eye, V_L is located visually, V_U is scanned with no pre-specified rule, and 'p3≈0' has no numerical cutoff. These are researcher degrees of freedom that could shift the reported rates. No error bars or statistical tests, so the 88% and 100% are summary counts, not tested rates.\n\nThe citation pattern is mostly self-citations because MCF is their method—that alone isn't a flaw. The math of the intermittent map is standard critical phenomena, and the specific claim about the PTB subset is checkable if they release the record list and parameters.\n\nBottom line: the paper deserves a serious referee because the descriptive result is concrete and the method is a known, published approach. But the diagnostic interpretation should be pulled back or explicitly framed as a hypothesis until the criticality-health link is independently validated. A careful reviewer should ask for the data, parameters, and a pre-registered protocol.","headline":"The 35-record extension is real, but the diagnostic conclusion outruns the method's validation.","tokens_in":7461,"tokens_out":3281,"would_cite":false,"duration_ms":29067,"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":"A physics-based test of ECG criticality flags 100% of infarction records and 88% of healthy controls.","keywords":["electrocardiogram","Method of Critical Fluctuations","criticality","power-law fluctuations","laminar lengths","myocardial infarction","healthy control","autocorrelation function"],"falsifier":"Run MCF blinded on a larger ECG cohort with an independent gold-standard diagnosis; the claim fails if a confirmed myocardial-infarction record passes the criticality condition ($1<p_2<2$, $p_3\\approx 0$) or if a cohort of clinically verified healthy subjects systematically fails it. A direct check is to submit the three flagged control records (p165, p245, p242) to independent cardiological assessment to see whether they are in fact diseased.","tokens_in":6440,"feed_emoji":"❤️","tokens_out":8176,"duration_ms":70463,"temperature":0.7,"pith_summary":"This paper applies the Method of Critical Fluctuations (MCF) to human electrocardiograms to test whether cardiac health shows up as physical criticality in the ECG signal. MCF extracts the distribution of waiting times (laminar lengths) between voltage levels in the high-frequency 'grass' fluctuations, and healthy critical dynamics are identified by a power-law distribution with exponent between 1 and 2. Applied to 35 ECG records from a public diagnostic database, the method finds critical dynamics in 22 of 25 healthy-control ECGs (88%) and in none of the 10 myocardial-infarction records, which all sit far from criticality. The authors conclude that roughly one in ten ECGs labeled healthy may actually be pathological, and they propose criticality as a physical criterion that can reveal healthy and diseased states beyond ordinary ECG reading.","feed_headline":"Criticality test catches every heart-attack ECG","feed_subtitle":"In 35 ECG records, a physics criticality check matched all 10 infarction cases and 22 of 25 healthy labels.","key_machinery":"The load-bearing object is the Method of Critical Fluctuations (MCF): a procedure that detects critical dynamics in an experimentally recorded time series by locating a fixed point $V_L$ where high-frequency 'grass' fluctuations begin, scanning an upper level $V_U$, and computing the distribution $P(L)$ of laminar lengths $L$ spent between $V_L$ and $V_U$. Criticality is declared when the fitted exponent $p_2$ of the power-law part of $P(L)$ satisfies $1<p_2<2$ and the exponential exponent $p_3$ is close to zero; departure from criticality shows as $p_2$ falling and $p_3$ growing. A wider zone $\\Delta V_U$ over which these conditions hold is read as a more stable critical state. The autocorrelation function of the most critical record is used as a second, related measure.","core_discovery":"The central claim is that a healthy heart's electrical fluctuations are in a critical state, and this criticality is measurable in an ECG. For a segment of high-frequency fluctuations, the authors define laminar lengths as waiting times inside the zone between the fixed point $V_L$ and a varied level $V_U$; when the laminar-length distribution follows a power law with $p_2$ in (1,2) and $p_3\\approx 0$ over a range of $V_U$, the signal is critical. By this criterion, all 10 myocardial-infarction records are non-critical while 22 of 25 healthy-control records are critical, giving 100% and 88% agreement. The paper therefore claims that MCF can serve as a diagnostic indicator and that the three non-critical control records may be misdiagnosed. For the record with the strongest criticality (p121), the autocorrelation function shows characteristic symmetries that the authors propose as a measure of optimal heart functionality.","pith_inferences":["Going beyond the paper: the three healthy-labeled but non-critical records (p165, p245, p242) are the natural test cases—if independent cardiological assessment finds them diseased, MCF is a genuinely prospective diagnostic; if it finds them healthy, the 'criticality equals health' premise is incomplete.","Going beyond the paper: the same MCF procedure could be applied to other labeled ECG classes such as arrhythmia, ischemia, or cardiomyopathy to see whether non-criticality is specific to infarction or common to all pathology, which would sharpen or weaken the diagnostic claim.","Going beyond the paper: the autocorrelation symmetries reported for record p121 could be computed automatically for the other records; if the three non-critical controls lack those symmetries, the autocorrelation profile might replace manual $V_U$ scanning as a faster screening statistic."],"forward_implications":["If criticality is genuinely a marker of healthy cardiac tissue, MCF offers a complementary screening signal that does not rely on matching ECG morphology to a template.","The reported 12% non-critical healthy controls become a concrete claim: about one in ten patients with a normal-looking ECG may be misdiagnosed and could warrant further clinical investigation.","The gradation of $p_3$ and the width of the critical zone $\\Delta V_U$ provide a quantitative scale of cardiac health, from strongly critical (record p121) to pathology-like, rather than a binary label.","For myocardial infarction, the perfect 10/10 separation suggests the method could be tested as an aid in cases where ordinary ECG reading is ambiguous.","Because the paper explicitly stops short of medical physiology, a validated link between criticality and cardiac pathophysiology would need to be established before clinical use."],"supporting_citations":[{"why":"Introduces the intermittent critical map and the power-law distribution of laminar lengths on which MCF is built.","marker":"[7]"},{"why":"Establishes the core premise that healthy frog heart relaxation-phase fluctuations are critical, which this paper extends to human ECGs.","marker":"[10]"},{"why":"Presents the first MCF application to human ECG for a healthy case and a myocardial-infarction case, the direct precursor of this statistical study.","marker":"[16]"},{"why":"Supplies the diagnostic ECG database records used for the healthy-control and infarction analyses.","marker":"[17]"},{"why":"Provides the clinical characterizations of the ECG records drawn from the diagnostic database.","marker":"[18]"},{"why":"Supports the criterion that a wider critical zone $\\Delta V_U$ indicates a more stable critical state.","marker":"[20]"},{"why":"Gives the deterministic-chaos derivation of the laminar-length power law used to identify criticality.","marker":"[14]"}],"fun_headline_variants":["Criticality test catches every heart-attack ECG in study","Heart attack ECGs fail the criticality check","ECG criticality screening detects all heart attacks","Healthy hearts critical, attack hearts not—ECG tells","Physics of critical fluctuations spots heart attacks"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a healthy heart is in a physical critical state, so that detecting critical fluctuations in an ECG is treated as proof of health; the paper inherits this equation from frog-heart data and never independently verifies it in humans.","fun_headline_variants_meta":{"raw":{"variants":["Criticality test catches every heart-attack ECG in study","Heart attack ECGs fail the criticality check","ECG criticality screening detects all heart attacks","Healthy hearts critical, attack hearts not—ECG tells","Physics of critical fluctuations spots heart attacks"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000386,"raw_usage":{"total_tokens":1994,"prompt_tokens":858,"completion_tokens":1136,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":474,"completion_tokens_details":{"reasoning_tokens":1063}},"tokens_in":474,"tokens_out":1136,"duration_ms":11643,"temperature":1.0,"reasoning_tokens":1063,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:46:23.414493+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run MCF blinded on a larger ECG cohort with an independent gold-standard diagnosis; the claim fails if a confirmed myocardial-infarction record passes the criticality condition ($1<p_2<2$, $p_3\\approx 0$) or if a cohort of clinically verified healthy subjects systematically fails it. A direct check is to submit the three flagged control records (p165, p245, p242) to independent cardiological assessment to see whether they are in fact diseased.","supporting_citations":[{"cited_title":"Intermittent dynamics of critical fluctuations","cited_arxiv_id":null,"evidence_quote":"Introduces the intermittent critical map and the power-law distribution of laminar lengths on which MCF is built."},{"cited_title":"Criticality in the relaxation phase of the spontan eous contracting atria isolated from the heart of the frog (Rana ridibunda )","cited_arxiv_id":null,"evidence_quote":"Establishes the core premise that healthy frog heart relaxation-phase fluctuations are critical, which this paper extends to human ECGs."},{"cited_title":"The Earth as a living planet: Human-type diseases in the earthquake preparation p rocess","cited_arxiv_id":null,"evidence_quote":"Presents the first MCF application to human ECG for a healthy case and a myocardial-infarction case, the direct precursor of this statistical study."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the diagnostic ECG database records used for the healthy-control and infarction analyses."},{"cited_title":"and Bousseljot, R.: Automatisierte EKG-Auswertung Mit Hilfe der EKG-Signaldatenbank CARDIODAT der PTB , Biomedical Engineering/Biomedizinische Technik, 40, 319–320, 1995","cited_arxiv_id":null,"evidence_quote":"Provides the clinical characterizations of the ECG records drawn from the diagnostic database."},{"cited_title":"Intermittent criticality revealed in ULF magnetic fields prior to the 11 March 2011 Tohoku earthquake ( /g1839/g3050=9)","cited_arxiv_id":null,"evidence_quote":"Supports the criterion that a wider critical zone $\\Delta V_U$ indicates a more stable critical state."},{"cited_title":"Deterministic Chaos","cited_arxiv_id":null,"evidence_quote":"Gives the deterministic-chaos derivation of the laminar-length power law used to identify criticality."}],"review_version":1}