Causal map of perimenopause onset — the −2 bit is a σ threshold; the first licit node, if it exists, is not read in the classifier nor in primordial freeze

1. Executive scientific abstract

The operational onset of perimenopause under STRAW+10 is not a location shift of mean cycle length nor an AMH threshold. It is the first persistent episode of |L_t − L_{t−1}| ≥ 7 days (recurrence ≤10 cycles): a consecutive-variance operator. The classical endocrine machine —fewer preantral/small-antral granulosa, less dimeric inhibin B, earlier and higher FSH via TGFBR3L, advanced selection with FSH→LH growth indistinguishable 10.7 vs 11.0 days— manufactures STRAW −3a (shorter cycles) and does not fire that operator if shortening is uniform (M1-1-MS-C1). Huang, Elliott and Harlow 2014 (PMID 24729638; TREMIN n=617 women, 95 246 segments) already estimated, on the same bleeding process, that the population variance change point (42.84 y) precedes the mean change point (46.23 y) by 3.39 y. That is not an ovarian node. It is the classifier operating on L_t.

Adversarial review split the hypothesis-generation conjunction. Clauses (i)–(iii) of M1-1-HG-H1 v1 survive; clause (vi) —“therefore there is no first node of the generating process”— is a non sequitur and is withdrawn. A threshold of SD(L) does not prove that the process lacks an ovarian generator. ReSTAGE (PMID 17681300) isolates the persistent >6 d difference as the only early marker that does not collapse toward late (TREMIN median 41.0 vs 48.8 y) and attributes it to ovarian activity. Vanden Brink 2013 cannot vote first: in LRA 36–44 y (n=10) LPDF kinetics are indistinguishable from MRA; the ARA 45–55 phenotype arrives late, mixed and ovulatory-only. The operational competitor is selection jitter of the already-growing cohort in −3 (M1-1-HG-H2 v2), not LPDF prevalence (already NS) nor ARA morphokinetics.

Experimental design added a gate this report cannot skip. The “SD(FP) anchored to menses” index is not exogenous to SD(L) until within-woman Var(LP)/Var(L) is shown to be material in −3. This turn closes a gap that EXPERIMENT_DESIGN left in between-woman tables (Klein, Reame) and does not close U2. PubMed eutils 2026-10-19: hormonal change-point vs first non-circular −2 = 0. Henry, Shirin, Goshtasebi and Prior 2024 (PMID 39320898, Hum Reprod; n=53, 21–41 y, mean 33.9, 676 ovulatory cycles) finally deliver within-woman variances of cycle/FP/LP —medians 3.1, 5.2 and 3.0 days— with FP > LP (P<0.001); cycle dispersion (3.1) is smaller than FP (5.2), so SD(L) is not a rotation of SD(FP) (negative covariance). Polarity to the crossing: I (mid-reproductive stage, prescreened ovulatory). Visby Madsen, Adornetto and Kirkegaard 2025 (PMID 39718569, Menopause; n=29, mean 44.5 y, commercial saliva) do not support that a short luteal phase predicts irregularity, long cycles or anovulation after adjusting the number of observations. U2 remains open in −3. D1 stays closed (n≈85/122, does not recruit). Primordial freeze (Liu 2026: cessation 32→36 weeks, fewer pups) and do(TGFBR3L) (Brûlé: larger litter) remain illicit. G8 closed. No prescription.

Scientific delta of this report (M1-1-REPORT-EN-C1): the English twin freezes the bilingual factual base that REPORT_ES locked. The three operational hypotheses are M1-1-HG-H1 v2, M1-1-HG-H2 v2 and M1-1-HG-HT1 v2 (not the HG v1 conjunction nor SCOPE-H1 mean as the −2 clock). PubMed 2026-10-19 reconfirms 0 hormonal change-point vs first −2 and resolves Huang 24729638 and Liu 42169137. Henry 2024 kills the global identity SD(L) ≡ SD(FP) as a mid-reproductive assumption and does not estimate ρ in −3; Visby Madsen 2025 does not elevate the luteal residual as the generator of the ReSTAGE operator. D0-IDENTITY remains the first desk cut; D0-OPERATOR stays specified and not executable without Huang intercepts (they are not invented); D1 gated.

2. Scientific question and relevance

The question is what determines when perimenopause begins and what is the first causal, measurable and manipulable node we should attempt to modify in order to delay it without shortening fertile function or displacing harm. Answering it requires separating objects that the title, the M1 north and the historical corpus fuse:

1. Is “when it begins” the first −2 cycle, a continuous endocrine change point, FMP, an age, an AMH threshold or a symptom?
2. Is the crossing criterion written without the candidate node (inhibin B, AMH, PdG, FSH, cycle)?
3. Is what is measured first cause, consequence, proxy or classifier?
4. Is delaying pool exhaustion extending competent ovulatory months?
5. Does AMH census primordials or read the growing pool (Weenen 2004)?
6. Does the first node live in primordial activation, the growing cohort, the corpus luteum, KNDy, DDR or an extraovarian accelerator?
7. Does a do(node) that delays −2 increase hormone-dependent cancer, conserve damaged oocytes or cut ovulations?
8. Did L7-4 (luteal loss inside −2/−1) or L10 (unidentified senescence) already answer this question? No.
9. Is an order of change points a do(A)? No (L7-5, L10-6-K2).
10. Is delaying peri lifespan? Healthspan remains H0. G8 closed.

The relevance is high because “delay peri” is a programme slogan and a cheap category error. A FOXO3/PTEN/AMH/VDR freeze that lengthens the calendar by cutting growing follicles fails the fertile clause (L6-5-K6). An AMH+FSH score that predicts FMP answers another title (Sowers, Broer, Depmann). A bleeding threshold is not an experimental stick. The correct unit is the woman in −3 approaching a non-circular crossing, with a coprimary of competent ovulatory months and organ-specific competing risks, not a product dataset nor a later FMP.

3. Scope, population and life stage

The primary biological estimand is the adult woman in natural reproductive aging, STRAW+10, not chronological age. The principal stratum is late reproductive (−3b/−3a) defined only by bleeding regularity: it is the only human calendar where the crossing toward −2 has not yet occurred. The outcome is the −3 → −2 crossing itself. Early transition (−2) is an arrival state and a negative-control of the map. −1 and +1 are non-targets: L7-4 already works −2/−1; Farr/KEEPS live in +1.

Inferential unit: the woman. Cycles, visits, follicles, Huang segments, nuclei and embeddings are not n. A Lua-user arm does not exist.

Preclinical material (Castrillon, Reddy, Durlinger, Visser, Liu, Brûlé) enters as a throughput and sign contrast, not as the n of the human protocol. Gonadotoxic injury, D-gal/VCD, cisplatin and POI are toxic clocks (L10-2-K1), out. This project does not gavage, does not dose, does not recruit.

Out of scope: pregnancy, lactation, PCOS mixed with natural transition, endometriosis, POI, ovarian cancer, oophorectomy, iatrogenic gonadal suppression; current Lua data, PHI, PII, private genomes; prescription of AMH, calcitriol, senolytics, NMN/NR or a non-indicated ovarian biopsy; FMP or a symptom as onset; p>0.05 as equivalence.

4. Background and mechanism map

4.1 Operational definitions

• Onset of perimenopause: the −3 → −2 crossing under STRAW+10: persistent difference ≥7 days in consecutive cycle lengths, persistence = recurrence ≤10 cycles (Harlow 2012, PMID 22344196, body). It is not −3a. It is not −1. It is not FMP. It is not “around 45.”
• −3a machine / H1-mean: smaller preantral–small-antral cohort → dimeric inhibin B ↓ → earlier and higher FSH → advanced selection, not faster growth → shorter FP.
• −2 generator / H1-σ: cycle-to-cycle variance of the selection moment, with ovulation still dominant.
• Inhibin B: INHA–INHBB heterodimer of preantral/small-antral granulosa. Reader of mass of that cohort. Not a product of FSH (Welt 2001). Not a primordial census.
• AMH: product of growing granulosa; brake on activation and on FSH sensitivity. Absent in human primordials (Weenen n=12). Not called a biomarker.
• Fertile function: months of competent ovulatory cycles (orthogonal ovulation + sufficient corpus luteum + oocyte competence). Not high AMH. Not late FMP.
• Displacing harm: organ-specific competing risks (hormone-dependent cancer, aneuploidy, bone, vessel). G8: delay is not called longevity.
• Observation ≠ do(node). Mapping an order does not authorize dosing.

4.2 Chain the literature would have to demonstrate

Preantral / small-antral cohort (not the primordial not encoded by AMH)
  → dimeric inhibin B ↓
  → gonadotrope reads via TGFBR3L (not betaglycan)
  → earlier and higher FSH
  → advanced selection (H1-mean = −3a)
  → selection jitter / SD(day | LH)  [H2 v2 candidate]
  → |ΔL|≥7 persistent = −2 bit         [classifier, H1 v2]
  → do(node) that lowers that σ
       WITHOUT cutting competent months    [G4]
       WITHOUT displacing organ harm       [G5, G8]

Today no human link anchors the jitter to the first non-circular −2. Primordial freeze attacks the inlet valve, not the −3a machine, and cuts throughput. Human TGFBR3L occupancy in −3 is absent; the female KO raises litter size.

5. Evidence method

Unit: the woman. Classes: H human, A animal, V in vitro/ex vivo, C computational, I inferred, R review (does not vote causality). Polarity relative to the non-circular −3 → −2 crossing. p>0.05 is not equivalence. FMP is not −2. AMH is not a census. PdG is not ovulation. Pre/post is not STRAW. Observation is not do. Supplement hits, SEO, Farr/KEEPS as peri, senolytics, blood NAD, S1 indices and any user cohort do not vote.

Prior stages audited 43 PMIDs (2026-10-18), reread bodies of STRAW+10, Hale 2007, Weenen, Sowers 2010, Liu 2026, Huang 2014, ReSTAGE and Paramsothy, and skipped compute (decision: skip). This turn reconfirms eutils 2026-10-19 and adds Henry 2024 and Visby Madsen 2025, which EXPERIMENT_DESIGN did not have, on U2 (within-woman Var(LP)/Var(L)).

6. Evidence map

7. Contradictory evidence and null findings

8. Multiscale mechanistic synthesis

8.1 Molecular — two machines, not one chain

H1-mean (direct between women on approach, H0 for the crossing). Groome 1996: dimeric assay, early-FP peak. Welt 2001: preantral/small-antral source; FSH 30 ng/mL does not raise inhibin B (1730±354 → 2314±532, NS). Klein 2004: inhibin A and activin A do not mediate the intercycle FSH of 40–45 ovulatory women; inhibin A is higher (P=0.01). Lewis/Li: betaglycan reads inhibin A; Brûlé: TGFBR3L reads inhibin B. Klein 2002 / van Zonneveld 2003: short FP = advanced selection from the previous corpus luteum, not acceleration (the Klein 1996 title remains a kinetic error).

Closure: small-cohort granulosa ↓ → INHBB/inhibin B ↓ → less TGFBR3L–ACVR2–SMAD antagonism on FSHB → earlier/higher FSH → advanced selection, conserved growth rate → shorter FP = STRAW −3a.

Why it does not generate −2. If all cycles shorten 28→24, |ΔL|≈0. The operator requires a variance generator.

8.2 Cellular / follicular — jitter, not freeze

The cell that manufactures the H1-mean analyte (granulosa of antral ≤4 mm) is not biopsied. The primordial (nuclear FOXO3) is invisible in blood and is the valve that freeze moves. Weenen: AMH=0 there. Topology: H1-mean lives in the growing follicle already out; H1b lives at the inlet. Closing the inlet delays the end by cutting output (Liu).

Baerwald 2003: 2–3 waves per IOI already in mid-reproductive. Vanden Brink: the object is not born in peri; what changes in ARA is kinetics, and it arrives late. Hale 2009 LOOP is the same family already in transition. The candidate that survives is intermittency of selection in regular −3, without requiring the ARA phenotype.

8.3 System — classifier, calendar, uterus

STRAW −2 writes the bit on the endometrium. Paramsothy links long menses to fibroids; the consecutive-difference operator is more robust than duration, but the effector is still bleeding. Burger admits endometrial FMP and, at the same time, a follicular generator of erratic cycles. Huang quantifies σ-before-mean on TREMIN 1934–1939 with missingness 512/617. None observe inhibin B, ovulation or selection day relative to LH.

8.4 Direct versus inferred

9. Computational layer

Decision: skip. Conserved. Huang 2014 already parted E[L] from SD(L) inside bleeding and does not kill H6: STRAW −2 is a threshold on the ΔL process that Huang modelled. Re-estimating TREMIN without an exogenous is tautological; TREMIN is not deposited anonymously. GEO 2026-10-18: 0 peri×inhibin/AMH×omics series; 0 peri×pituitary; TGFBR3L = mouse n=4; LPDF = bovine IVF. BIMORA/ICPSR 4452: no inhibin B, AMH or ultrasound; cadence 6 months/year does not cover persistence to 10 cycles (L7-4-K4). BioNeMo does not observe |ΔL|, waves, occupancy or throughput.

D0-OPERATOR (ED) is a generative mapping on published moments, not a re-estimation nor a FASTQ. Discriminant: ΔAUROC of the −2 bit after SD_10(L)+age, margin 0.02. Intercepts α0/β0/Ω of Huang Tables 1–2 are not invented; without them the run stays specified and not executable. Reproducing σ-before-mean does not discriminate.

10. Primary hypothesis

M1-1-HG-H1 v2 — the −2 bit is a persistence threshold on SD(L); manipulating the classifier is tautology; the generating process remains open

Role: primary of object. Absorbs SCOPE-H6 as a bit threat. Cedes the generator. Not the v1 conjunction.

Falsifiable statement: in natural reproductive aging, (i) “when perimenopause begins” under STRAW+10 is the first persistent |L_t−L_{t−1}|≥7 (recurrence ≤10 cycles), a consecutive-variance operator; (ii) H1-mean produces −3a and does not fire −2 if shortening is uniform; (iii) Huang 2014 estimated CP_σ 42.84 y before CP_mean 46.23 y on TREMIN n=617 on the same L_t, as an alternative indicator of early/late MT, not as an ovarian node; (iv) a do validated against delaying the binary without an exogenous to bleeding is tautology; (v) ReSTAGE shows that this operator is the early most separable from late (TREMIN 41.0 vs 48.8 y) and that before 40 y it does not approximate FMP; (vi) the prohibition of every node of the generating process is withdrawn.

Mechanism: the bit is a function of consecutive differences. SD(L) is the second moment of the same process. Conditioning the bit on recent SD(L) is nearly identity — and that does not vote an ovary upstream.

Evidence for (H, definition and calendars): Harlow 2012; ReSTAGE 2008; Huang 2014; Klein 2002; Burger (1)–(2). Evidence against (vi): ReSTAGE names ovarian activity; Burger (3); Henry 2024 shows that SD(L) does not exhaust phase architecture.

Kill: an exogenous-to-bleeding node (selection day relative to LH, dimeric inhibin B, not SD(FP)-menses) whose CP precedes the first non-circular −2 after conditioning SD(L) and common cause (age, adiposity, uterus). Then this primary restricts to “the bit was a threshold” — which is exactly v2 — and H2 v2 wins the biological object.

Status: revised_primary. Maturity: H1 of literature for the operator and for CP_σ-before-mean on TREMIN. H0 for the prohibition (withdrawn). Confidence: 0.80 that the bit is a σ threshold; 0.28 that there is no ovarian node of the process.

11. Competing hypothesis

M1-1-HG-H2 v2 — the first biological generator of that σ is selection jitter of the already-growing cohort in regular −3, with stable ovulation and P4 AUC, without requiring LPDF-ARA

Role: competing. Contrary to reading Huang/ReSTAGE as closure of the generator. Does not rehabilitate H1-mean as the crossing nor H2-luteal as first (Hale).

Falsifiable statement: in women −3b/−3a defined only by bleeding regularity, (i) small-cohort mass already fell (dimeric inhibin B, −3a territory) before the first −2; (ii) what fires the operator is the variance of the selection moment (SD of the day the dominant ≥10 mm anchored to the LH surge, not to menses), not a uniform location shift and not LPDF prevalence (already NS); (iii) that CP precedes the first non-circular −2, with orthogonal ovulation and stable P4 AUC; (iv) LPDF-ARA is not (iii): in LRA it is indistinguishable from MRA; (v) conditioning SD(L) without the selection index does not absorb (iii); (vi) later, atypical LPDFs / LOOP feed L7-4 (arrival).

Metrology correction (ED + this report). SD(FP) from menses is forbidden as a D1 index until D0-IDENTITY, in −3, releases ρ = Var_w(LP)/Var_w(L). Henry 2024 suggests the global identity is false (cycle 3.1 < FP 5.2 within-woman) but in prescreened 21–41 y, not in the LRA window. H2 v2 survives only with an LH-relative index or with dimeric inhibin B (growing mass, Welt 2001), which are not functions of L_t.

M1-1-HG-H2-LPDF: demoted (weakened). Kill of first already fired in LRA. A second laboratory that shows the ARA phenotype inside regular −3 reopens it as a proxy, not as freeze.

Kill of H2 v2: the selection CP does not precede, or disappears after conditioning SD(L), or P4/ovulation fall first, or a uterine/thyroid index leads at equalized stock.

Status: revised_competing. Maturity: H1 of −3 architecture; H0 of anchor to the crossing. Confidence: 0.41 in growing-cohort jitter as an open generator; 0.12 in LPDF-ARA as first.

12. Translational hypothesis

M1-1-HG-HT1 v2 — primordial freeze and do(TGFBR3L) are not the stick of the title; the licit stick, if H2 v2 delivers a non-collinear index, is regularity of the already-growing cohort scored on competent months, not on −2 nor FMP

Role: translational negative with respect to the binary and the obvious levers; parked as a biomarker. HUMAN_QA_REQUIRED. Not H5. No prescription. Does not recruit.

Falsifiable statement: (i) a primordial-activation perturbation (AMH, PI3K–PTEN–FOXO3, pharmacological VD3) that delays cessation cuts growing follicles and does not increase output (Liu n=15, 32 vs 36 weeks) — fails G4; (ii) do(TGFBR3L) raises folliculogenesis and litter size (Brûlé) — opposite sign; human −3 occupancy absent; dKO infertile; (iii) FSH alone does not restore inhibin B (Welt 2001) — pushing gonadotropin advances selection = −3a, does not stabilize σ; (iv) an AMH-census + FSH + binary −2 or FMP parameter will not out-of-sample win competent months against age + recent SD(L); (v) if H2 v2 delivers an exogenous selection index, the research stick is reducing atresia/irregularity of the cohort already growing while conserving monthly recruitment, scored on competent months + SD(selection|LH), not on −2 nor FMP; (vi) G8 closed. No user n. Not called a biomarker.

The stick does not wait for H6 to die as a classifier; it waits for a measurable generator. Today there is no named compound. Maturity H0 of a positive stick.

Kill: a freeze that conserves output + competence, without cytostasis, with named competing risks; TGFBR3L occupancy that lowers selection SD without cutting ovulations; wrong event (FMP, symptom, pleiotropy).

Status: revised_translational, parked. Maturity: H1 of vetoes; H0 of a positive stick. Confidence: 0.82 that freeze and TGFBR3L are not the node of the title; 0.18 that a growing-cohort perturbation exists today.

13. Falsifiable predictions and kill criteria

14. Discriminating experiment

No outcome of this project recruits. The de-risk sequence, frozen in EXPERIMENT_DESIGN and not reordered by Henry/Visby Madsen:

D0-IDENTITY     is Var_w(LP)/Var_w(L) material in −3?     NO → SD(FP)-menses illicit
       │         Henry 2024 does not substitute: wrong stage
       ↓ YES or INCONCLUSIVE (reading 2026-10-19: inconclusive in −3)
D0-OPERATOR     does the −2 bit add AUROC to rolling SD(L)  NO → H1 v2 wins the bit
       │         under Huang, margin 0.02?                 intercepts not invented → not yet executable
       ↓
D0-UTERUS       do fibroids/TSH associate with the          YES at equalized stage → H5/uterus rises
       │         persistent >6 d operator, not menses ≥10 d?
       ↓
D0-INVENTORY    is there a public file with a non-collinear NO → D1 CLOSED
       │         index + bleeding in −3?                    GEO/BIMORA = 0
       ↓
D1              six cycles, 36–44 y, −3 by regularity, TVUS q2 d anchored to LH,
                dimeric inhibin B, n≈85 completers / 122 enrolled. NOT OPEN.

D2-G4           PARALLEL. Liu 2026 is already the G4-fail of freeze. Coprimaries of any
                future perturbation: output NI 0.80–1.25 BEFORE cycle σ.

Frozen margins. Materiality of ρ: ≥0.20. Equivalence to identity: IC90% of ρ under 0.20. ΔAUROC absorption: increment < 0.02. TOST LPDF LRA≡MRA: ±2 d. p>0.05 is not equivalence. Lua n invalidates.

D1, if a human opened it (not this stage). Within-woman observational, six cycles, STRAW −3 by regularity alone, 36–44 y (LRA window, before ARA 45–55). Estimand: within-woman partial R² of SD(selection day | LH) predicting SD(|ΔL|), after SD(L) and SD(LP). Partial r 0.30, α 0.05, power 0.80 → n≈85 completers. One IOI is not n. Wrong event (FMP, symptom, AMH-census) relabels the study.

This design does not demonstrate extended human fertility nor longevity. It does not prescribe.

15. Biomarkers and stratification

Not called a biomarker: inhibin B, AMH, LPDF, SD(FP), SD(L), TGFBR3L or the −2 bit.

Prespecified stratification of a future D1: age 36–39 vs 40–44; BMI; fibroids yes/no; TSH tertiles. No fishing. No ethnicity as a mechanism.

16. Individual variability

Huang Fig. 2: women with early CPs have ~4–5 y between σ and mean; late ones ~1–2 y. The population “3.39 y” is not each woman’s crossing. ReSTAGE: 12.4% of TREMIN have persistent >6 d and late ≥60 d at <1 year; 87.6% do not. Firings <40 y do not approximate FMP.

Henry 2024: within-woman LP variance has range 0.7–7.9 days already in prescreened 21–41 y; 55% had >1 short luteal phase and 17% ≥1 anovulatory cycle despite two normal entry cycles. That underestimates the population (authors’ limitation) and does not transport to −3. Hale: anovulation 0/16 LRA vs 9/23 LMT — competence heterogeneity explodes after the crossing. Vollman (cited by Henry): FP and LP change across reproductive life; LP shorter in late peri. Polarity to U2 in −3: I, and it predicts that ρ may rise approaching −2, not that it is already measured.

Huang cohort: University of Minnesota women 1934–1939. Transporting 42.84 y as a universal clock is overclaim.

17. Pharma relevance and maturity

There is no regulatory use context, clinical utility, net benefit, external replication or partnering readiness. Not H5. HUMAN_QA_REQUIRED. Healthspan and longevity out (G8). No compound, dose, efficacy trial or companion diagnostic is proposed.

The nearest opportunity is, at most, a research-use (RUO) assay of regularity of the already-growing cohort — dimeric inhibin B + selection day relative to LH + coprimary of competent ovulatory months — conditioned on D0-IDENTITY and on a D1 that a human opens. Competing risks of late ANM (Ruth) are named, not “balanced” in a compound.

18. Limitations

• Zero within-woman change-point vs first non-circular −2 is not proof that none exists; it is published absence (eutils 2026-10-18 and 2026-10-19).
• Huang: hierarchical population CP, missingness 512/617, MAR, FMP anchor, 1934–1939 cohort. It is not each woman’s index cycle.
• Hale, Klein, Welt, Vanden Brink: cross-sectional or one IOI. Between women ≠ trajectory.
• Vanden Brink LRA n=10 could not declare LRA≡MRA equivalence; AR should not have read “indistinguishable” as TOST. ED converts it to TOST n≈50/group, secondary and closed.
• Henry 2024: QBT with mean delay 2.4±1.5 d versus the LH peak; “variances” reported in days (SD vs days² ambiguity); prescreened ovulatory; mean 33.9 y; underestimates population SOD. Does not close ρ in −3.
• Visby Madsen 2025: n=29, commercial device, COI, “presumed perimenopausal” by age >42, STRAW not written, no consecutive-difference operator. Underpowered null.
• Liu: exposure from 3 weeks of life, not human −3; n=15; “slightly smaller” without CI. Human crossing polarity: I.
• Brûlé: the mouse has no STRAW. Developmental KO ≠ adult occupancy.
• Sowers n=50 unique. Abel group n not consolidated in the abstract: it is not invented.
• D0-OPERATOR was not executed: intercepts not tabulated here.
• Lua data forbidden. This report does not recruit and does not prescribe.

19. Conclusions

The title asks for a first causal, measurable and manipulable node of perimenopause onset. What exists today is a σ classifier (STRAW −2 / ReSTAGE persistent >6 d / Huang CP_σ 42.84 y) on a continuum that staging itself already declared in motion at −3a, and a candidate ovarian generator —selection jitter of the already-growing cohort— that nobody anchored to the first non-circular crossing. The clause that forbade every node of the generating process is withdrawn. The lever that does move the end (primordial freeze) cuts throughput. The pituitary reader (TGFBR3L) has opposite animal sign. FSH does not restore the granulosa that manufactures inhibin B.

The smallest next step is not to recruit perimenopausal women nor to dose calcitriol. It is the D0-IDENTITY desk in the −3 window: extract, if they exist, repeated within-woman (L, FP, LP) pairs —Henry does not donate them for that stage— and, in parallel, not invent Huang intercepts. While U2 is open, SD(FP)-menses is not a node. D1 remains closed. G8 remains closed. Fertile function is not played by delaying a bleeding bit.

20. References

1. Harlow SD, Gass M, Hall JE, et al. Executive summary of the Stages of Reproductive Aging Workshop + 10. J Clin Endocrinol Metab. 2012. PMID 22344196. PMC3319184.
2. Harlow SD, Mitchell ES, Crawford S, et al. The ReSTAGE Collaboration: defining optimal bleeding criteria for onset of early menopausal transition. Fertil Steril. 2008. PMID 17681300. PMC2225986.
3. Harlow SD, Cain K, Crawford S, et al. Evaluation of four proposed bleeding criteria for the onset of late menopausal transition. J Clin Endocrinol Metab. 2006. PMID 16772350.
4. Huang X, Elliott MR, Harlow SD. Modeling menstrual cycle length and variability at the approach of menopause using hierarchical change point models. J R Stat Soc Ser C Appl Stat. 2014. PMID 24729638. PMC3979630.
5. Hale GE, Zhao X, Hughes CL, et al. Endocrine features of menstrual cycles in middle and late reproductive age and the menopausal transition classified according to STRAW. J Clin Endocrinol Metab. 2007. PMID 17550960.
6. Hale GE, Hughes CL, Burger HG, Robertson DM, Fraser IS. Atypical estradiol secretion and ovulation patterns caused by luteal out-of-phase (LOOP) events. Menopause. 2009. PMID 18978637.
7. Klein NA, Battaglia DE, Fujimoto VY, et al. Reproductive aging: accelerated ovarian follicular development associated with a monotropic follicle-stimulating hormone rise in normal older women. J Clin Endocrinol Metab. 1996. PMID 8772573.
8. Klein NA, Harper AJ, Houmard BS, et al. Is the short follicular phase in older women merely a consequence of advanced recruitment? J Clin Endocrinol Metab. 2002. PMID 12466381.
9. Klein NA, Houmard BS, Hansen KR, et al. Age-related analysis of inhibin A, inhibin B, and activin A relative to the intercycle monotropic FSH rise. J Clin Endocrinol Metab. 2004. PMID 15181087.
10. Welt CK, McNicholl DJ, Taylor AE, Hall JE. Female reproductive aging is marked by decreased secretion of dimeric inhibin. J Clin Endocrinol Metab. 1999. PMID 9920069.
11. Welt CK, Smith ZA, Pauler DK, Hall JE. Differential regulation of inhibin A and inhibin B by luteinizing hormone, follicle-stimulating hormone, and stage of follicle development. J Clin Endocrinol Metab. 2001. PMID 11232020.
12. Groome NP, Illingworth PJ, O'Brien M, et al. Measurement of dimeric inhibin B throughout the human menstrual cycle. J Clin Endocrinol Metab. 1996. PMID 8636341.
13. Weenen C, Laven JS, Von Bergh AR, et al. Anti-Müllerian hormone expression pattern in the human ovary: lack of expression in primordial follicles. Mol Hum Reprod. 2004. PMID 14742691.
14. Sowers MR, Eyvazzadeh AD, McConnell D, et al. Anti-Müllerian hormone and inhibin B in the definition of ovarian aging and the menopause transition. J Clin Endocrinol Metab. 2008. PMID 18593767.
15. Sowers MR, McConnell D, Yosef M, et al. Relating smoking, obesity, insulin resistance, and ovarian biomarker changes to the final menstrual period. Ann N Y Acad Sci. 2010. PMID 20738279. PMC3296522.
16. Randolph JF Jr, Zheng H, Sowers MR, et al. Change in follicle-stimulating hormone and estradiol across the menopausal transition: SWAN. J Clin Endocrinol Metab. 2011. PMID 21159842.
17. Burger HG, Dudley EC, Hopper JL, et al. Prospectively measured levels of serum follicle-stimulating hormone, estradiol, and the dimeric inhibins during the menopausal transition. J Clin Endocrinol Metab. 1999. PMID 10566644.
18. Burger HG. The menopausal transition—endocrinology. Menopause. 2008. PMID 18574431.
19. Paramsothy P, Harlow SD, Greendale GA, et al. Bleeding patterns during the menopausal transition in the Study of Women's Health Across the Nation. BJOG. 2014. PMID 24735184. PMC4199918.
20. Vanden Brink H, Chizen D, Hale G, Baerwald A. Age-related changes in major ovarian follicular wave dynamics during the human menstrual cycle. Menopause. 2013. PMID 23571519.
21. Vanden Brink H, Robertson DM, Lim H, et al. Associations between antral follicle count and serum follicle-stimulating hormone, inhibin B, anti-Müllerian hormone, estradiol and luteal progesterone in older reproductive-aged women. Menopause. 2018. PMID 29112593. (n=9 vs 16, preliminar.)
22. Baerwald AR, Adams GP, Pierson RA. Characterization of ovarian follicular wave dynamics in women. Biol Reprod. 2003. PMID 12748128. Fertil Steril. 2003. PMID 12849812.
23. van Zonneveld P, Scheffer GJ, Broekmans FJ, et al. Do cycle disturbances explain the age-related decline of female fertility? Cycle characteristics of women aged over 40 years compared with a reference population of young women. Hum Reprod. 2003. PMID 12615813.
24. Reame NE, Kelche RP, Beitins IZ, et al. Age effects of FSH and pulsatile LH secretion across the menstrual cycle of premenopausal women. J Clin Endocrinol Metab. 1996. PMID 8636360.
25. Klein NA, Illingworth PJ, Groome NP, McNeilly AS, Battaglia DE, Soules MR. Decreased inhibin B secretion is associated with the monotropic FSH rise in older, ovulatory women. J Clin Endocrinol Metab. 1996. PMID 8675606. n=10 vs 13; hCG-day criteria 10.6±0.4 vs 14.5±1.0. (EXPERIMENT_DESIGN labelled it “Welt 1996”; the PMID is Klein.)
26. Castrillon DH, Miao L, Kollipara R, Horner JW, DePinho RA. Suppression of ovarian follicle activation in mice by the transcription factor Foxo3a. Science. 2003. PMID 12855809.
27. Reddy P, Liu L, Adhikari D, et al. Oocyte-specific deletion of Pten causes premature activation of the primordial follicle pool. Science. 2008. PMID 18239123.
28. Durlinger AL, Gruijters MJ, Kramer P, et al. Anti-Müllerian hormone attenuates the effects of FSH on follicle development in the mouse ovary. Endocrinology. 2001. PMID 11606457.
29. Visser JA, Durlinger AL, Peters IJ, et al. Increased oocyte degeneration and follicular atresia during the estrous cycle in anti-Müllerian hormone null mice. Endocrinology. 2007. PMID 17255205.
30. Liu S, et al. Calcitriol is involved in maintaining primordial follicle reserve through the inhibition of the PI3K/Akt signaling pathway. J Ovarian Res. 2026. PMID 42169137. DOI 10.1186/s13048-026-02135-5. PMC13386817.
31. Brûlé E, Wang Y, Li Y, et al. TGFBR3L is an inhibin B co-receptor that regulates female fertility. Sci Adv. 2021. PMID 34910520.
32. Lewis KA, Gray PC, Blount AL, et al. Betaglycan binds inhibin and can mediate functional antagonism of activin signalling. Nature. 2000. PMID 10746731.
33. Li Y, Fortin J, Ongaro L, et al. Betaglycan (TGFBR3) functions as an inhibin A, but not inhibin B, coreceptor in pituitary gonadotrope cells in mice. Endocrinology. 2018. PMID 30364975.
34. Abel TW, Rance NE. Stereologic study of the hypothalamic infundibular nucleus in young and older women. J Comp Neurol. 2000. PMID 10931489.
35. Ruth KS, Day FR, Hussain J, et al. Genetic insights into biological mechanisms governing human ovarian ageing. Nature. 2021. PMID 34349265.
36. Gold EB, Bromberger J, Crawford S, et al. Factors associated with age at natural menopause in a multiethnic sample of midlife women. Am J Epidemiol. 2001. PMID 11323317.
37. Freeman EW, Sammel MD, Lin H, Gracia CR. Anti-Müllerian hormone as a predictor of time to menopause in late reproductive age women. J Clin Endocrinol Metab. 2012. PMID 22378815.
38. Broer SL, Eijkemans MJ, Scheffer GJ, et al. Anti-Müllerian hormone predicts menopause: a long-term follow-up study in normoovulatory women. J Clin Endocrinol Metab. 2011. PMID 21613357.
39. Hsieh YT, Ho JYP. Thyroid autoimmunity is associated with higher risk of premature ovarian insufficiency. J Clin Endocrinol Metab. 2021. PMID 33569594.
40. Santoro N, Crawford SL, El Khoudary SR, et al. Daily salivary progesterone and estradiol across the menopausal transition: SWAN Daily Hormone Study. (O'Connor / Santoro series inherited from L7-4; anovulation/PdG change in late peri.) Santoro N et al. J Clin Endocrinol Metab. 2017 and related O'Connor KA et al.
41. Henry S, Shirin S, Goshtasebi A, Prior JC. Prospective 1-year assessment of within-woman variability of follicular and luteal phase lengths in healthy women prescreened to have normal menstrual cycle and luteal phase lengths. Hum Reprod. 2024;39(11):2565–2574. PMID 39320898. DOI 10.1093/humrep/deae215.
42. Visby Madsen C, Adornetto G, Kirkegaard K. The role of short luteal phases on cycle regularity during the perimenopausal transition. Menopause. 2025;32(3):258–265. PMID 39718569. DOI 10.1097/GME.0000000000002477.
43. Guthrie JR, Dennerstein L, Taffe JR, Lehert P, Burger HG. The menopausal transition: a 9-year prospective population-based study. The Melbourne Women's Midlife Health Project. Climacteric. 2004. PMID 15799609.
44. Wallace WH, Kelsey TW. Human ovarian reserve from conception to the menopause. PLoS One. 2010. PMID 20111701.

Claims of this report: M1-1-REPORT-EN-C1 (delta: bilingual factual base + Henry/Visby Madsen on U2); M1-1-REPORT-EN-C2 (operational hypotheses = AR v2, not HG v1 nor SCOPE-H1 mean); M1-1-REPORT-EN-C3 (PubMed 2026-10-19: 0 hormonal CP vs first −2; Huang 24729638 and Liu 42169137 resolve); M1-1-REPORT-EN-C4 (Henry 2024 kills global identity SD(L)≡SD(FP) and does not estimate ρ in −3); M1-1-REPORT-EN-C5 (Visby Madsen 2025 does not elevate short LP as a generator of irregularity); M1-1-REPORT-EN-C6 (D1 closed; freeze/TGFBR3L illicit; G8 closed; HUMAN_QA_REQUIRED; not written into reportes/).

Hypotheses touched: M1-1-HG-H1 v2 (revised_primary); M1-1-HG-H2 v2 (revised_competing); M1-1-HG-H2-LPDF (weakened); M1-1-HG-HT1 v2 (revised_translational, parked); SCOPE-H1 not promoted; SCOPE-H6 absorbed as a bit threat; H1b intact; H3/H4 not elevated; H5/uterus as D0-UTERUS; SCOPE-HT1 parked.

Not written into memoria agentes/Lua Labs/reportes/. Laboratory memory is not consolidated. No prescription. Lua data are not used. Next stage: BILINGUAL_QA.