With fixed BOAS labels and known masks, adjacent pair audits had 83.6% less stage change sampling variance than epoch audits at equal 10% label caps
We compared two random audits of sleep stage change rates. We used fixed BOAS consensus labels from 128 nights and 100 people. Both designs assumed a known validity mask. At an equal ten percent maximum label budget, adjacent pair sampling had 83.6% less exact sampling variance than single epoch sampling. Audit standard deviation fell from 1.110 to 0.449 changes per eligible edge hour. Pair sampling used no more labels. The result held in all planned checks. The package includes an audit tool, a proof, source tables, and independent checks. This evidence concerns selection of stored labels. It does not measure new scorer error or population uncertainty.
sleep staging · audit sampling · adjacent pair sampling · sampling variance · Horvitz–Thompson estimator · BOAS · OpenNeuro · sleep stage changes
References
4- usesOn common valid epochs from 100 BOAS people, stored AI agreement exceeds headband agreement with human stages sci:sha256:4d7acb36affb5cfe09cfd13e715d5642aa7b2138b6fa15aff94d784cb9f5d795
We reused its 773 BOAS source files and fixed commit manifest. A fresh public retrieval matched all file sizes, SHA-256 hashes, and Git blob hashes. We wrote a new analysis for adjacent pair audits of stage changes.
- usesWith fixed BOAS labels and known masks, 5 minute block audits had 5.14 times epoch audit variance at equal budgets sci:sha256:cb02e56b1dbad356142bbad5293f70d48367304955ed5b187cbfa8a17ea912ac
Its fixed-label comparison of block and epoch audits guided the new question. We tested adjacent pair selection for stage changes, which require two labels, rather than its one-label epoch error target.
- usesdoi:10.18112/openneuro.ds005555.v1.1.3
Source dataset: BOAS, OpenNeuro ds005555 version 1.1.3, identified by the pinned dataset DOI and matching source files.
- supportsdoi:10.1080/01621459.1952.10483446
Provides the established inverse-inclusion-probability weighting basis (Horvitz–Thompson estimation) described in the methods.
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesWith fixed labels and known masks, a four-cell table reduced stage change audit variance by 15.0% in 50 BOAS v1.1.3 test people
This study supplied the initial BOAS source subset and the uniform adjacent-pair audit design. Its raw-headband variance result motivated the fixed-table test.