Ranges from 60 EEG pairs one month apart contained 36 of 39 later 8–13 Hz grid maxima with 60 s spectra and average reference

We tested ranges for a posterior EEG frequency maximum at a later visit. We set the radius from 60 pairs in OpenNeuro ds004148 version 1.0.1. Their visits were one month apart. We tested the ranges on 39 CE pairs in ds007176 version 1.0.1. The primary measure used 60 s of spectral windows, the average scalp reference, and the 8–13 Hz grid. The radius was 1 Hz. The ranges contained 36 of 39 repeat maxima. Baseline window bootstrap bounds contained 28 of 39. Mean widths were 1.904 and 1.096 Hz. The mean interval score gain was 2.782 Hz. Its approximate joint bootstrap 95% limits were −1.609 to 5.833 Hz. The limits included zero. The supplied-reference and 20 s checks reversed the score result. Thus, higher observed containment did not establish a score gain in the target population. Cohort differences prevent a guaranteed transfer coverage claim. The result applies to the CE files; closed eyes is an assumed interpretation of that label. The package preserves 258 spectral records, all resample indices and draws, source provenance, code, figures, and reproduction commands. All nine settings had an identical full rerun. Fresh full EEG replay matched every stored array for one record from each cohort.

EEG · frequency estimation · prediction intervals · bootstrap · test-retest reliability · spectral analysis