The Feno program - Fenocosm record v56: card 03: Ivanov et al. 1999, multifractality in human heartbeat dynamics + the CHF loss (P-11.7)

Record v56 — card 03: Ivanov et al. 1999, multifractality in human heartbeat dynamics + the CHF loss (P-11.7). A sealed, append-only results record of the Feno program (repository arm: Fenocosm). The audit campaign's fourth executed card and the first on the founding positive of the biosignal multifractality literature (audit/PROTOCOL.md; audit/cards/03-ivanov-1999-heartbeat.md). The target is Ivanov, Amaral, Goldberger, Havlin, Rosenblum, Struzik & Stanley 1999 (Nature 399:461, DOI 10.1038/20924), whose audited claim is, quoted from its abstract: "We report on evidence for multifractality in a biological dynamical system --- the healthy human heartbeat. Further, we show that the multifractal character and nonlinear properties of the healthy heart rate are encoded in the Fourier phases. We uncover a loss of multifractality for a life-threatening condition, congestive heart failure." Two readings, scored separately (protocol amendment 8): the mechanism — the healthy interbeat series is multifractal and its multifractality is nonlinear in origin, destroyed by phase randomisation — and the biomarker — CHF shows a "loss of multifractality" and the paper's three per-subject quantities (the degree of multifractality Δh = max h − min h, τ(3), the SD of the intervals) discriminate healthy from CHF. The pipeline is reconstructed as published, every "the paper says" checked against the paper's own text with a locator (amendment 1): the WTMM singularity spectrum with the 3rd derivative of the Gaussian, scales a = 2·1.15^i (i = 0..41), q = −5..5, τ(q) fitted over 16 < a < 700 (Fig. 2 caption), on the NN-interval series of the daytime (12:00–18:00) records of the 18 healthy subjects (PhysioNet nsrdb) and the CHF patients (chfdb), with the paper's own stated preprocessing (Methods: normal-to-normal intervals only; the 5-beat 2× outlier filter; the 3σ increment-pair filter). Every unstated convention was pinned and pushed before any target cell was scored (commits 75954a5 and the dated amendment 7b1f61a; amendment 6) — above all Muzy, Bacry & Arneodo 1994's maxima-line supremum in Z_q(a), which the paper cites but does not state and without which its literal sentence reads a monofractal as multifractal (sealed in the validation certificate). The P4 instruments are the campaign's, unchanged: the scattering omnibus (v22), the wavelet-leader c₂ arm (v39), the Chhabra–Jensen t_MF (v38); the biomarker machinery is v41's. Register ALLIED, emphatically. No personal material; the audit is of a pipeline. The paper's healthy reading reproduces to its published envelope and survives the marginal-and-spectrum-preserving null; its CHF interpretation — "reduction of nonlinearities with pathology" — survives on the modern instrument even where its statistic fails; the discrimination it built is real. Those survivals are stated first and with the prominence a reversal would get. What the audit prices is the attribution step (an FT-surrogate comparison, which cannot separate nonlinearity from a heavy-tailed marginal) and the raw width as a CHF marker (which does not reproduce on these data). Regeneration ```bash python3 -m venv venv && venv/bin/pip install -r requirements-dev.txt ./verify.sh # apparatus gate (incl. smoke_heartbeat) venv/bin/python -m fenocosm.p56_heartbeat # -> out/p56_heartbeat.* (~2.1 h, 4 workers) cp out/p56_heartbeat* out/v56/ ``` The only external inputs are the PhysioNet nsrdb (.atr) and chfdb (.ecg) beat annotations and headers, fetched to data/heartbeat/ (git-ignored) at run time from https://physionet.org/files/{nsrdb,chfdb}/1.0.0 (66 files, 4.3 MB). They are static archival products; every cached file was verified against PhysioNet's own SHA256SUMS.txt, and the record pins the combined manifest SHA-256 prefix 0c936107a76a4f9d. Everything else is synthetic and seeded. Sealing-run wall clock: 122.8 min at 4 workers. Deposit note. All seven Feno-program repositories are private and remain private, so this deposit is standalone: it carries its own reproduction materials, its metadata contains no repository link, and sibling work is referenced by DOI rather than by repository path. Part of the Feno program. The source repository is private; the reproduction materials for this record are contained in the record itself.