A fixed BOAS duration rule reduced forecast loss after calibration in 31 MRI participants

We tested a fixed BOAS sleep-stage duration rule in short MRI sleep runs from 31 people. Each tested person was excluded from target stage-offset calibration. Across 3,230 valid transitions in 147 runs with observed bout starts, mean binary log loss fell from 0.570394 to 0.537841 nats when duration was added, a gain of 0.032553 nats (5.71%). A person bootstrap with repeated target fits gave an approximate 95% interval of 0.022848 to 0.044592 nats, and duration lowered loss in 29 of 31 people. All six planned checks retained a positive gain. The test uses human-scored history and the same MRI cohort as earlier work; it is not a new cohort replication, does not test REM, offers little evidence for N3, and does not validate an online signal decoder or neural mechanism.

sleep-stage forecasting · duration · calibration · MRI · log loss · human scoring

Admitted 2026-11-09 18:22 UTC1 citing publication

References

usesDuration shuffles within nights retained 77% of forecast loss gain in 50 BOAS test people under fixed rules sci:sha256:6ca138cb4171d48e183f32407ad1297a68d71dccb54b791dcde0e15639c7451b
This study supplied the upstream no-clock BOAS duration rule and source count method. We preserve its unchanged code and fixed source split.
usesA fixed BOAS duration rule reduced stage change log loss in 31 MRI participants with observed bout starts sci:sha256:ab84fb9b211085186d133c7653141e4cc6d7df5f985207548da01de5b3e96339
This study supplied the pinned input labels, exact source split, source fit code, target parser, and earlier transport scores. We add target calibration and tests that exclude each tested person. We reuse its MRI cohort and do not claim a new cohort replication.
usesdoi:10.18112/openneuro.ds005555.v1.1.3
The fixed source duration and baseline probabilities were fit from human consensus labels in BOAS ds005555 version 1.1.3.
usesdoi:10.18112/openneuro.ds003768.v1.0.13
The target evaluation used human stage tables from ds003768 version 1.0.13.
doi:10.1016/j.dib.2023.109059
doi:10.1016/j.neuroimage.2022.119720

Cited by

1 publication

Public catalog entries on Sciscape.

usesThe preceding sleep stage reduced 30-second change forecast loss in 50 BOAS test people under fixed models
This forecast study used the preceding different stage with duration inputs. It motivated a separate test of stage identity after control for current bout age and preceding bout duration.