Nested Control Systems — a pre-registered falsification program (NHANES)

STATUS: pre-registered protocol, frozen 2026-03-04 — NOT yet executed. No verdict is claimed; this seed records the TEST, not a result. The nested-control-systems framework (the autonomic spine) is bound to a real, frozen, falsifiable validation against NHANES 2011–2018 — with operational definitions pinned to specific NHANES variables and explicit falsifiers set BEFORE the data are touched: • H1 — the upstream manifold predicts 10-year mortality. Falsifier: AUC < 0.60 OR ΔAUC ≤ 0 vs age+sex. • H2 — hyperglycemia is mechanistically heterogeneous (clusters). Falsifier: best_k < 2 OR medication χ² p > 0.05 OR bootstrap ARI < 0.40. • H3 — Layer 5 (systemic) adds predictive lift for diabetes. Falsifier: ΔAUC_L5 < 0.015 pooled OR negative in any adequately-powered cycle; clinical bar ΔAUC ≥ 0.02 to be SUPPORTED. The rubric forbids self-declaration: no hypothesis may be labeled SUPPORTED without surviving its full truth-suite (primary + stability + sensitivity). Three outcomes only — SUPPORTED / FALSIFIED / INCONCLUSIVE. Acknowledged limit: NHANES has no direct L1 (cellular-stress) markers — the L1 layer uses a downstream proxy and does NOT test the framework's mechanistic L1 claims; a molecular cohort (UK Biobank Olink proteomics) is the honest next step. This is the FLOOR and BROTHERS gates made concrete: the claim is exposed to physical reality with falsifiers named in advance, and to external witness. Defining what would break it — before looking — is the rigor, whether or not it has yet been run. Tamper-evident: the frozen package is fingerprinted (SHA-256 below); re-hashing the same files reproduces it exactly.