Fixed known labels left mean absolute headband error above 13.60 minutes for sustained REM time in 100 BOAS 1.1.3 people with a 10-minute bout rule
We tested time in uninterrupted REM bouts of at least 10 minutes in 128 released BOAS nights from 100 people. Each missing human or headband label could be REM or another stage. All known labels stayed fixed. Nights were averaged within each person. People had equal weights. The exact mean absolute night error was bounded by 13.60 and 26.22 minutes. Thus, missing-label choices could not remove this score error in the fixed sample. The signed mean error could be negative or positive. A finite-state program found every attainable score. It preserved gaps that an interval calculation would lose. A separate calculation checked all score sets, bounds, and primary assignments. A fresh download reproduced every analysis output byte for byte. The package contains source labels, fixed URLs, hashes, code, a local plan, exact proofs, assignments, results, figures, and reproduction instructions. The result concerns released labels and the human reference. It does not measure true REM physiology or new model performance.
rapid eye movement sleep · sleep staging · wearable EEG · label agreement · missing labels · exact bounds · finite-state dynamic programming
References
5- usesStored headband wake coverage was lower for 30 to 60 second than for at least 120 second internal human wake bouts in 87 BOAS people sci:sha256:189bb9f97d15f6eed6349973a60c2ec7eb75375ac6b2221a26edcbd1048fa1ce
Its duration-based Wake coverage results supplied context for testing REM continuity as a separate label measure. We read this work before saving the local plan. Its Wake results are not used as evidence for the new REM result.
- usesOn common valid epochs from 100 BOAS people, stored AI agreement exceeds headband agreement with human stages sci:sha256:4d7acb36affb5cfe09cfd13e715d5642aa7b2138b6fa15aff94d784cb9f5d795
We used its 773 fixed BOAS source files, source manifests, Git tree, retrieval code, and alignment checks. We downloaded all files again and reproduced the new analysis from that download. The new REM score and exact bounds are our work.
- usesWith known headband labels fixed, all completions left at least 7.90 min of mean absolute N3 time error after night averaging in 100 BOAS people sci:sha256:1eff2fc2e4ae5d64301b65f974e497e234687bbef81391e9532e907b10e96a83
Its total N3 time bounds supplied context for unrestricted missing-label assignments with known labels fixed. This study tests sustained REM run time and uses a new exact score-set program because attainable scores can have gaps.
- usesdoi:10.18112/openneuro.ds005555.v1.1.3
Source labels and metadata from BOAS version 1.1.3; every known label was held fixed and unknown labels were varied under the stated assignment model.
- doi:10.1523/JNEUROSCI.3220-14.2014
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesBOAS 1.1.3 headband labels had at least 28.49 minutes of mean absolute longest sleep bout error in 100 people when only missing labels could change
Its exact sustained REM bounds study motivated full possible score sets, completion witnesses, and separate feasibility checks. We read its description before the local plan. We developed a different interval solver for longest binary runs and did not reuse its scoring code.