Outcomes, with their limits attached

The result: Sleep-paralysis and NDE source-coded corpora support a trigger-specific Layer-1 body-boundary signature: sleep paralysis is led by paralysis/presence (7.209% / 8.909%), while NDE is led by vestibular separation/presence (17.76% / 14.269%).

What it means: The two experiences open differently in the telling: sleep-paralysis stories lead with being frozen and sensing a presence, while near-death stories lead with floating free of the body plus a presence. If that holds under validated labels, the trigger shapes the body story.

What it is NOT: Not usable as evidence yet, because the percentage contrasts rest on machine labels that must first pass a blinded human check. Not a mechanism claim, because word patterns in stories cannot show what the body did.

Next test that could settle it: Independently validate paralysis, presence, and vestibular labels on blinded rows, then recompute the corpus contrast.

Open question: Do G-LOC, syncope, coma, anesthesia, sleep-paralysis, and NDE data preserve trigger-specific body-boundary patterns?

The result: NDE life-review rows behave like a memory/social-affiliation subtype rather than a generic light/OBE subtype: within coded NDE life-review rows, deceased-relative co-occurrence is 58.14%, while bright-light co-occurrence is 16.279%.

What it means: Inside near-death stories that contain a life review, deceased relatives appear far more often than the famous bright light does. The life review may belong to a memory-and-family cluster of the experience rather than to the light-and-tunnel cluster.

What it is NOT: Not confirmed, because the coded rows behind the two percentages must be recovered and blindly re-validated before the contrast counts. Not a claim about real deceased persons, because these are story elements.

Next test that could settle it: Recover the coded rows, validate life-review/deceased-relative/light labels blindly, and reproduce both denominators.

Open question: Is the reported life-review/deceased-relative subtype reproducible from validated NDE rows?