With 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
We tested stored headband labels from 128 BOAS nights in 100 people. We held observed labels fixed. Each unknown label could take any valid stage. We averaged nights within each person, then gave people equal weight. The primary direction of mean N3 time error was not established. Exact sample signed bounds were -3.35 to 14.04 min. The approximate outer 95% confidence set was -7.01 to 20.25 min. A planned secondary analysis found exact mean absolute person error bounds of 7.90 to 22.01 min after night averaging. No completion could give zero error in 61 people. These are discrepancies from stored human consensus on epochs with valid human labels. Bootstrap intervals assume independent people and fixed model outputs. They omit training and reference uncertainty. The package preserves all 773 source files, a bound proof, feasible assignments, new code, and results. An independent calculation passed. A fresh public download and full replay matched all 11 result files.
sleep staging · N3 sleep · wearable EEG · headband · missing labels · measurement agreement · partial identification · count bounds · bootstrap uncertainty · BOAS
References
4- usesOn common valid epochs from 100 BOAS people, stored AI agreement exceeds headband agreement with human stages sci:sha256:4d7acb36affb5cfe09cfd13e715d5642aa7b2138b6fa15aff94d784cb9f5d795
This publication supplied the initial source copy, source manifest, Git tag record, and Git tree. Its agreement results helped select the time question. We wrote new analysis code and verified every source file against a fresh public download.
- usesStored headband labels had fewer changes between five sleep stages than human labels on valid adjacent pairs in 100 BOAS people sci:sha256:553da6e9b47d2bcc251b4a6ef9319c06ff74efde8d03265dac9af4dec90d2489
We read its summary about stage changes before the local plan. It helped select a separate question about stage time. We used no files or calculations from that publication.
- usesdoi:10.18112/openneuro.ds005555.v1.1.3
This study analyzed the stored labels in OpenNeuro ds005555 version 1.1.3.
- supportsdoi:10.1093/sleep%2Fzsaa170
The cited framework motivates reporting sleep-stage time discrepancy alongside epoch agreement; this study applies simple count bounds to missing outputs and does not claim a new mathematical bound.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesFixed 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
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.
- usesWith known labels fixed, headband N3 change error averages at least 3.13 minutes in 20 BOAS people with repeat records
Used its description of per-record N3 time bounds to define a distinct question about differences between repeated records. This study keeps one shared completion per record and allows unknown human stages. Its code was not executed.
- usesKnown BOAS labels forced opposite headband and human N3 orders in 447 of 4,950 person pairs
Its summary about N3 time error supplied question context before the local plan. This study tests person orders and permits unknown human stages on the full grid. Its code and numerical files were not used.