A fixed log background correction did not establish lower mean absolute one-month 8–13 Hz frequency change in 60 adults

We tested a fixed log spectral background correction in 60 healthy adults from OpenNeuro ds004148 v1.0.1. The primary rule used closed eyes, 280 planned seconds, eight posterior channels, average59, and an 8–13 Hz maximum. A line fitted in log power against log frequency used 2–7 and 14–30 Hz. Mean absolute one-month change was 0.3125 Hz before correction and 0.2833 Hz after correction. The paired reduction was 0.0292 Hz, with a 95% person bootstrap interval from -0.0042 to 0.0667 Hz. The interval included zero. The test did not establish less change or equal performance. Eight people had positive gain, 49 had zero gain, and three had negative gain. A proof shows that a decreasing fitted line cannot move the first band maximum down. Correction can increase or decrease session change. All seven planned checks are reported. The 7–13 Hz check had a positive interval; two 7 Hz boundary maxima account for its larger gain. That check is descriptive and has no multiple-test adjustment. The package preserves 180 spectra, source hashes, code, all random draws, results, and reproduction instructions. Independent calculations passed. Three fresh raw records rebuilt all arrays exactly. A complete repeat matched all six numerical output files. The result concerns this empirical line and selected cohort. It does not validate true neural frequency or test SpecParam.

EEG · alpha frequency · spectral background · test–retest reliability · power spectrum · frequency correction · bootstrap