Mapping the dm³ operator chain across four biological domains disrupted by preterm birth
The starting point is what is actually kernel-checked. From WP-31C and WP-31D, the following holds without sorry in AXLE/AutophagyDm3.lean:
AutophagyDm3_v2.lean — 24 theorems, no sorry, no True conclusions
kernel-verifiedThe following are proved without sorry or vacuous True conclusions, as of 2026-08-25. The file's own header documents what each obligation would have to assert and why the previous True-stub declarations were removed rather than kept.
contactCoeff_neg). Necessary condition for α∧dα ≠ 0;
not the full manifold statement.V_is_morse_at_one), with the non-vacuity guard
not_isMorseCritical_const confirming the definition rejects constant maps.
This is the algebraic half of Ob. 2 — kinase data and Mather finite determinacy
remain open.gronwall_radius, basin_asymmetry).dm3_basin_compact, dm3_basin_nonempty) — Heine-Borel
on a closed interval. The file is explicit: naming this interval "the dm³ basin"
does not make the theorem about the dm³ flow until the flow itself is defined.Three obligations remain open and are stated in prose in the file, not stubbed as True: full contact non-degeneracy on Xauto (Ob. 1, needs exterior derivative on manifold); Whitney A₁ from actual mTORC1 kinase data (Ob. 2, needs Mather + data); limit cycle via Poincaré–Bendixson (Ob. 3, needs the dm³ flow defined as a semiflow before any limit cycle theorem has content).
CRNT deficiency δ = 2 for the AMPK–mTORC1–ULK1 network was established in WP-31C by explicit graph construction (n=9, l=4, s=3), verified in SymPy. This is a different tool from Lean 4 and a different result from those above. It is necessary, not sufficient, for fold behavior, and the rate constant estimation did not clear (WP-31C: "Estimate: did not clear").
The withdrawn claim: μ_max ≈ −0.41 s⁻¹ was retracted in WP-30 — two inconsistent time constants, neither traceable to literature, real ULK1/mTORC1 kinetics hours-scale not seconds-scale. This paper does not use that parameter.
The working hypothesis of this paper is that four biological systems disrupted by preterm birth share the same fold structure — a Whitney A₁ topology at the relevant threshold — as the autophagic switch. The kernel-verified anchor is the algebraic fact that V(q) = q³ − 3q has a genuine Morse critical point, and the SymPy-verified anchor is that the AMPK–mTOR–ULK1 network has the deficiency required for fold behavior. Everything downstream is model or lean-needed until checked independently.
Pulmonary surfactant — a lipoprotein complex secreted by type II alveolar pneumocytes — reduces alveolar surface tension and prevents collapse on expiration. Its synthesis depends on the transcription factor TTF-1 (NKX2-1), which drives expression of surfactant proteins SP-B and SP-C. documented
SP-B synthesis is functionally absent before approximately 24–26 weeks gestation and reaches adequate levels around 34–36 weeks. The transition is not gradual: below a gestational threshold, respiratory distress syndrome (RDS) is near-universal in the absence of exogenous surfactant; above it, RDS incidence drops sharply. documented (Jobe & Ikegami 1987; Whitsett & Weaver 2015, New England Journal of Medicine)
In dm³ terms: the TTF-1/SP-B/SP-C synthesis cascade is a candidate for the same fold structure as the autophagy switch. The C operator is the hormonal and mechanical compression of fetal breathing movements and glucocorticoid signaling; K is the TTF-1 activation threshold; F is the fold at which lamellar body formation commits; U is sustained surfactant secretion. model
Antenatal corticosteroids (betamethasone/dexamethasone) accelerate surfactant maturation by acting upstream of the TTF-1 threshold — a clinical intervention that effectively shifts the fold threshold earlier. documented (Roberts & Dalziel 2006, Cochrane Database). This is consistent with the dm³ reading: the C operator (compression) is being augmented pharmacologically.
Chapter 15 of Vol VI (Cardiac Rhythms — Wenckebach, Lyapunov, and T* = 2π) establishes that cardiac conduction ratios (3:2 Wenckebach) map onto the dm³ operator chain, with the honest caveat that only the 3:2 ratio is a genuine Fibonacci convergent.
Preterm infants show high rates of cardiac arrhythmia, including bradycardia of prematurity — the most common arrhythmia in this population. The underlying mechanism is incomplete maturation of the autonomic nervous system's cardiac control, specifically the parasympathetic (vagal) tone that stabilizes sinus rhythm. documented (Poets et al. 1994, Pediatrics; Fairchild et al. 2016, Frontiers in Physiology)
In dm³ terms: the Lyapunov attractor basin for stable sinus rhythm (the limit cycle Γ of Ch 15) has not fully formed in the preterm autonomic system. The fold at K* — the threshold between stable rhythm and bradycardic episode — is crossed repeatedly because the attractor basin is shallow. model
This is falsifiable in principle: HRV (heart rate variability) data from preterm infants should show lower Lyapunov stability than term infants of matched corrected age, with convergence toward term values as corrected age increases. Such data exist in the literature (Lake et al. 2002, American Journal of Physiology) but have not been fitted to the dm³ parameters. open
Chapter 16 (Hopfield Networks) establishes the energy function E = −½ΣWᵢⱼSᵢSⱼ as a genuine Lyapunov function (Hopfield 1982, kernel-verified). Chapter 17 covers gamma oscillations and Communication-Through-Coherence (Fries 2015, cited as an active hypothesis, not settled fact).
Myelination of cortical and subcortical tracts proceeds through the third trimester and into the first postnatal years, but the most rapid phase occurs between 28 and 40 weeks gestation. In the preterm brain, this window occurs ex utero, in the NICU environment rather than in utero. documented (Counsell et al. 2006, Pediatrics; Ball et al. 2013, Brain)
Incomplete myelination reduces conduction velocity and synchrony between neural populations. In Hopfield network terms: the weight matrix W is sparser and less structured than in a term brain at equivalent developmental stage. The energy wells (attractor basins) are shallower, and the network capacity (~0.138N, Ch 16) is lower. model
The practical consequence: preterm infants at corrected term age show different EEG coherence patterns than term newborns, consistent with shallower attractor basins. documented (Stjerna et al. 2015, NeuroImage). Whether the dm³ parameters (ε₀, μ_max) can be fitted to this data is an open calibration problem. open
Auditory stimulation in the NICU — including the mother's recorded voice and selected music — has documented effects on preterm neurodevelopment. Randomized controlled trials show improvements in feeding behavior, physiological stability, and neurodevelopmental outcomes at follow-up. documented (Loewy et al. 2013, Pediatrics; Filippa et al. 2019, Acta Paediatrica)
In attractor terms: the mother's voice is a known input to an auditory cortex that is forming its weight matrix in real time. Consistent, meaningful auditory input during this window may contribute to attractor depth in the relevant neural circuits. This is a plausible mechanistic reading, not a verified one. open
Chapters 12–14 cover VDJ recombination, thymic selection (three-stage proofreading cascade, ~2% of thymocytes mature), and germinal center affinity maturation. These are the best-documented instances of the C∘K∘F∘U operator chain in immunology.
Thymic development is substantially complete by 28 weeks gestation, but the thymus continues to mature through term. Preterm infants have reduced naive T cell output, lower thymic index, and altered regulatory T cell proportions compared to term infants. documented (Gibbons et al. 2014, Journal of Allergy and Clinical Immunology; Rechavi et al. 2015, Science Translational Medicine)
In dm³ terms: the three-stage proofreading cascade of Ch 13 (positive selection at K*, negative selection window closing at K**) is compressed in the preterm thymus. The ~2% mature T cell output rate — which depends on the precision of two sequential threshold crossings — may be altered when the attractor basins for each stage are incompletely formed. model
| Domain | Biological system | dm³ operator | Preterm disruption | Evidence class | Book6 ref |
|---|---|---|---|---|---|
| Molecular | Autophagy (AMPK–mTOR–ULK1) | Contact form α = dz−ρ²dθ; Whitney A₁ fold at q=1; Gronwall radius 1/3; CRNT δ=2 | Neonatal autophagy flux altered in preterm gut and brain cells | kernel-verified fold algebra + contact coefficient (AutophagyDm3_v2.lean, 24 theorems, no sorry) · documented CRNT δ=2 (WP-31C, SymPy) · rate constants open (WP-31C estimate did not clear) · 3 Lean obligations still open (Ob.1 manifold, Ob.2 kinase data, Ob.3 semiflow) | Ch A · AutophagyDm3_v2.lean |
| Pulmonary | Surfactant synthesis (TTF-1/SP-B/SP-C) | C = glucocorticoid/mechanical; K = TTF-1 threshold; F = lamellar body commitment; U = secretion | SP-B absent before ~26wk; RDS near-universal below threshold | documented (threshold); dm³ mapping model; CRNT δ=2 lean-needed-1 | No current Vol VI chapter — this paper is the first mapping |
| Cardiac | Autonomic cardiac rhythm (vagal tone, sinus node) | Limit cycle Γ, Lyapunov attractor basin, K* = bradycardic threshold | Bradycardia of prematurity; shallow Lyapunov basin due to incomplete autonomic maturation | documented (clinical); attractor basin model; HRV fitting open | Ch 15 |
| Neural | Cortical attractor formation (myelination, Hopfield weight matrix) | W matrix formation; energy well depth ~ attractor basin; capacity ~0.138N | Incomplete myelination → shallower attractors → altered EEG coherence | documented (EEG coherence); W-matrix interpretation model; parameter fitting open | Ch 16 · Ch 17 |
| Immunological | Thymic selection (positive + negative selection cascade) | Two-threshold C∘K*∘K** structure; ~2% output rate | Reduced naive T cell output; lower thymic index in preterm infants | documented (thymic index); two-threshold dm³ mapping model; CRNT characterization lean-needed-2 | Ch 13 |
The Preemie Manual growth tracker implements the most immediate practical consequence of this framework: WHO Child Growth Standards are suppressed before corrected age zero and the Fenton 2013 preterm chart is cited instead. This is not merely a technical fix — it is the same principle as the fold argument: a reference standard built for term-equivalent biology does not apply before the relevant thresholds have been crossed.
The tracker is valid for any child, preterm or full-term, from term-equivalent age to 5 years (WHO standards), and for preterm infants from birth using the Fenton chart (22–50 weeks postmenstrual age).