Neural Array Projection Oscillation Tomography: A Ten-Stage Developmental Genealogy and Falsifiable Reconstruction Framework

Neural Array Projection Oscillation Tomography (NAPOT) proposes that distributed neural populations receive partial measurements, transform them through state-dependent temporal structure, project consequences, and recurrently update an action-relevant estimate of body and world. Draft 1 recovered ten developmental stages and converted their joined mechanism into a falsifiable latent-state framework. Draft 2 froze a multisite Neuropixels benchmark with eight capacity-matched model families, eleven nulls and ablations, mouse-isolated splits, sequential neural and action outcomes, and explicit failure conditions. Draft 3 began implementation without opening biological outcomes: a deterministic five-seed fixture exercised all eight information-admission contracts, all eleven destruction routes, mouse-disjoint splits, leakage guards, and causal phase and phase-wave-differential operators. Draft 4 added a capacity-matched recurrent architecture suite under one 2,146-parameter contract. Draft 5 triangulated the written ladder against four public 2022 explainer transcripts: 1,534 of 1,534 deterministic segments were reviewed, yielding 89 bounded authorial atoms. Draft 6 freezes the videos' live metadata and archive custody and calibrates the previously inconclusive synthetic action endpoint across nine conditions and 20 independent seeds per condition. The original five-seed result reproduces exactly. The zero-signal negative control wins only 6/20 seeds and fails the descriptive stability rule. At the baseline generator setting, M7 wins 15/20 seeds with a median advantage of -0.00570 bits per trial, but its 90th-percentile delta is +0.00585, so it also fails the stability rule. None of the nine conditions passes. The result identifies an architecture or endpoint limitation rather than converting synthetic evidence into a positive claim. Public recordings strengthen chronology and source fidelity, not biology. Synthetic results are engineering evidence about recoverability, causality, model routing, and failure behavior. They are not evidence that brains implement NAPOT, that phase is generally superior to rate or power, or that a decoded state is conscious. The operators have not been bound to Allen schemas, the biological file manifest remains empty, and no Allen file, test identity, or outcome has been opened. Draft 6 therefore narrows the remaining work to schema-bound preprocessing, exact biological manifests and splits, confirmatory null refits, uncertainty analyses, complete outcome reporting, and independent source, code, leakage, neuroscience, and statistical review.