1–40 Hz normalization lowered occipital alpha rank agreement between eye states in 45 screened ds004362 v1.0.0 pairs under average reference
We tested whether power normalization changes participant order between eyes-open and eyes-closed EEG records. We used all 218 baseline files from OpenNeuro ds004362 v1.0.0. The planned voltage screen retained 45 of 109 participant pairs. Under average reference, absolute 8–13 Hz alpha power had a rank correlation of 0.428 between states. Alpha divided by 1–40 Hz power had a correlation of 0.076. The absolute-minus-relative difference was 0.352, with a paired bootstrap 95% interval of 0.143 to 0.561. A 4–40 Hz denominator gave a difference of 0.057, with an interval that included zero. The finding depends on the tested denominator, reference, and selected sample. It does not establish causal eye-state effects or neural trait validity. The package includes source hashes, retrieval code, window spectra, all planned checks, and reproduction instructions. Independent decoding checked every source file. A fresh run reproduced all numerical outputs exactly.
EEG · alpha power · relative power · rank agreement · eye state · spectral power · normalization
References
5- usesRelative power gave a smaller geometric mean alpha ratio in 25 screened ds005420 v1.0.0 pairs with the source reference sci:sha256:5d1697dfa2fe69df771b4791d53c490cc96eba26197a337d585cb9d1b02713f8
The report used this result as context for power normalization after the numerical analysis was complete. It tested alpha change ratios in ds005420. The present analysis tested agreement of power ranks between eye states in ds004362.
- usesdoi:10.18112/openneuro.ds004362.v1.0.0
Used the 218 baseline EDF files in this fixed OpenNeuro snapshot. Every file passed the recorded size and SHA-256 checks.
- usesdoi:10.13026/C28G6P
Original PhysioNet dataset source for the baseline EEG records; source attribution and terms are preserved in LICENSES.md.
- doi:10.1109/TBME.2004.827072
- doi:10.1038/s44360-026-00096-z
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesSingle reference electrode removal changed occipital alpha contrast by less than 1 dB in ds004362 v1.0.0 on fixed windows
The earlier alpha study supplied the fixed source manifest, retrieval code, metadata, and archived base spectra. We checked its manifest against the fixed OpenNeuro snapshot. We reproduced its base powers before the new reference test. We then tested all 61 single electrode removals on identical windows and saved their measurement and screen effects.