The bootstrap gain interval included zero when first-night N2 exit rates entered fixed ridge models with age, sex, and scorer codes for 75 Sleep-EDF pairs
We tested whether first-night N2 exit rates improve second-night estimates in 75 Sleep-EDF pairs. Both fixed ridge models used age, recorded sex, and both scorer codes. The personal model also used the first-night rate. Each test person was excluded from fitting. The measure counted N2 transitions to wake or N1 within eight hours after lights off. The mean squared errors were 6.362 and 6.486 squared percentage points. The gain was -0.125, with a 95% person bootstrap interval of -1.005 to 0.828. All seven planned check intervals also included zero. This does not establish zero gain or clinical accuracy. The package provides unchanged source files, all forecasts, 5,000 refit bootstrap draws, code, and run instructions. An independent check passed for 8,100 duplicate-row fits. A fresh download reproduced all 188 source, result, and figure files exactly. The result applies to this sample, source measure, and fixed models.
sleep staging · N2 transitions · prediction · ridge regression · test-retest · Sleep-EDF
References
3- usesScorer adjustment changes linear age estimates for N2 transitions to wake or N1 over eight hours in 153 Sleep-EDF nights sci:sha256:b6904d2cd1e0f440a4ddde5b0a38959518ff4b5710f6a107dd1f82813953cffc
This publication supplied the Sleep-EDF source annotation files, source hash pins, and eight-hour N2 transition definition. Its extraction code and count table supported independent count checks. We use these inputs for a new paired-night prediction test.
- usesdoi:10.13026/C2X676
Sleep-EDF Expanded supplied the source annotations and subject table used for this paired-night test.
- doi:10.1038/s44360-026-00096-z
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.