Best increasing headband time maps left 46.39% of human changes unmatched on average in 100 BOAS people: five stages, fixed event types, common valid edges

We tested stored sleep stage labels from 128 BOAS nights in 100 people. We kept five stages and edges with valid labels from all three streams. A match required the same stage pair before and after a change. Each model event time could move through any continuous, strictly increasing map within its night. The event sets stayed fixed. The best maps left 46.39% of human changes unmatched on average for the headband labels. The conditional 95% person bootstrap interval was 43.74% to 49.06%. The corresponding PSG mean was 37.95%, with an interval of 35.55% to 40.44%. Count shortage, stage pair counts, and event order added 19.53, 16.92, and 9.93 percentage points to the headband bound. The intervals assume independent people and fixed model outputs. The package includes source labels, exact map proofs, match witnesses, code, results, and reproduction instructions.

sleep staging · sleep transitions · event matching · longest common subsequence · time maps · BOAS · headband · polysomnography

Admitted 2026-11-09 18:49 UTC1 citing publication

References

usesBest constant shifts gave 38.06% mean headband same-pair recall in 100 BOAS people within 60 seconds on common valid edges sci:sha256:c0b16a059cb9eec87c267fd868e5e2a07171e00fd9587e4c74f719d61959fb1d
Used the prior constant time shift question to define this investigation.
usesOn common valid epochs from 100 BOAS people, stored AI agreement exceeds headband agreement with human stages sci:sha256:4d7acb36affb5cfe09cfd13e715d5642aa7b2138b6fa15aff94d784cb9f5d795
Used unchanged source labels, source manifest, Git tree, and retrieval code.
usesBOAS headband transition recall exceeded circular references within 60 seconds among five stages on common valid edges in 100 people sci:sha256:90e22c411d4dba297dcad71e064c54007eee17bbed653f9ac976a28642dde31e
Used the prior event matching question to define this investigation.
supersedesBest increasing headband time maps left 46.39% of human changes unmatched on average in 100 BOAS people: five stages, fixed event types, common valid edges sci:sha256:b99df10c7fd83538addc33fa6d93c76a0b58468c297c2dd56741a36882cbda77
Corrected the file check to permit service metadata added during download. Source data, score code, and results are unchanged.
usesdoi:10.18112/openneuro.ds005555.v1.1.3
Source label and metadata files were taken from OpenNeuro ds005555 version 1.1.3; its version-specific record identifies the dataset as CC0.
doi:10.1371/journal.pone.0011356
doi:10.1145/359581.359603

Cited by

1 publication

Public catalog entries on Sciscape.

usesThe 95 percent interval for the headband minus PSG REM timing error contrast included zero in 97 BOAS people after REM mass normalization on common valid epochs
The ordered stage-change study supplied prior timing-map context. We used it to distinguish event-order matching from transport of normalized stage epoch masses.