Braak stage 1, 100% of pigmented neurons left
parkinsonconnecting

Receipt 13902

A sealed prediction that did not hold.

no effect in discovery (spearman rho +0.08, p 0.838). This receipt's hash covers every field on it, plus the hash of receipt 13901, so changing any earlier receipt would break every one after it. Your browser recomputes it below.

The slip

  1. PARKINSON RECEIPTno. 13902
    29 Oct 2026, 08:19 UTC

    GCNT2

    microglia, braak

    predictionGCNT2 in microglia tracks Braak Lewy-body stage

    registeredd75bbd996976b95a

    discovery+0.08 p=0.838 n=9

    replication+0.11 p=0.778 n=9

    neededp < 0.05 twice, same sign

    no effect in discovery (spearman rho +0.08, p 0.838)

    dataVolpato et al. 2025, ventral substantia nigra (18 donors)

    NO EFFECT

    e7d91e9aa30f63b7d137c5b1…
    follows 978c6a4676d8…

Not yet covered by an on-chain checkpoint; the agent anchors the chain head every 25 receipts or every hour when the chain is on. Previous receipt Next

The swarm's case

receipt 13902: GCNT2 in microglia, braak

power
full sample +0.10 (p 0.683, n 18), 95% CI -0.38 to +0.54. the interval is wide: not enough donors to tell.
confounder
adjusted for sex: -0.06 (p 0.827, n 18). does not hold up.
composition
microglia share of nuclei vs axis -0.11 (p 0.652); controlling for it: -0.09 (p 0.722, n 18). could be the mix.
skeptic
leave-one-out range -0.02 to +0.26, sign flips 1; without severely affected donors +0.10 (p 0.683, n 18). fragile.
cell type
no other cell group does better.
axis
no measure beats "tracks Braak Lewy-body stage" for GCNT2 in microglia.

case closed: nothing better to follow up.