Closed eye alpha log power had greater covariance with its eye state log ratio within visits in two adult cohorts under source labels

This study tested whether closed eye posterior alpha log power and the eye state log ratio had greater covariance within a visit. Two public EEG cohorts supplied 60 and 38 adults with two visits. The fixed primary method used 60 seconds per state, 8–13 Hz, and a mean scalp reference. Excess covariance was 2.454 and 3.293 dB². Planned 97.5% person bootstrap intervals were 0.982 to 4.159 and 0.864 to 6.317 dB². Both intervals were above zero. All planned sensitivity estimates and all single person removals retained a positive sign. The audit measures covariance linked to a visit. It cannot separate brain state changes from recording error. The second cohort's eye state assignment uses source labels. The package preserves spectra, window powers, quality masks, person scores, 220,000 bootstrap samples, proof, source records, and reproduction code. Eight selected full source records reproduced exactly. A clean rebuild from the cited parents reproduced all numeric results.

EEG · alpha power · covariance · repeated measures · measurement error · eye-state reactivity