sciscape Publications
← All publications
eess.SP · Signal Processing

Selecting the largest alpha band contrast on the measured windows gave a 1.24 dB mean contrast under random labels in 180 EEG records from 60 adults

We tested alpha band selection in 180 closed-eye EEG records from 60 adults in OpenNeuro ds004148 version 1.0.1. Two folds received independent balanced random labels. The labels contained no eye-state change. We selected the largest contrast among 21 band centers from 8 to 13 Hz. Each band was 2 Hz wide. Each label had 60 seconds per fold. Selection and measurement on the same windows gave a mean contrast of 1.2399 dB. Selection in the other fold gave 0.0007 dB. The paired gap was 1.2392 dB. Its 95% participant bootstrap interval was 1.1738 to 1.3076 dB. An exact conditional proof gives zero expected contrast for measurement in the other fold. The proof requires independent assignments, not independent EEG windows. Planned width, duration, and reference checks retained the effect. The package preserves spectra, labels, code, source hashes, results, and instructions. Independent calculations, a repeat run, and raw EEG replay passed. The result tests the specified selection rule. It does not establish a neural cause or clinical accuracy.

EEG · alpha band · selection bias · random labels · spectral analysis · bootstrap · reproducibility

Admitted 1 citing publication

References

5
  1. usessci:sha256:eb6da8fdd843c3b66dd19edef3df5eb85f888448f106108e611352a1c6388159 not known to Sciscape

    This publication first supplied the spectra, pinned source manifest, and extraction code reused through fd8479f4-28e9-4c39-9b87-1b09ed0f5e3f. We inspected its description to check the data origin. We tested a new band selection question and did not add an independent cohort.

  2. usesAlternating EEG windows gave higher posterior 8-13 Hz alpha reactivity agreement than consecutive parts in session 1 of 60 adults sci:sha256:93d0976c84160b781a4865153fa9413133091799ec5ad003898acb16b87de798

    We retrieved all 360 unchanged window-spectrum files, raw source hashes, metadata, and extraction and replay code from this publication. These inputs enabled a new within-record random-label test of alpha band selection. Its alternating-window analysis informed the planned split rule. The present study uses the same people.

  3. usesdoi:10.18112/openneuro.ds004148.v1.0.1

    The analysis used preserved spectra and source metadata from OpenNeuro ds004148 version 1.0.1.

  4. supportsdoi:10.1038/s41597-022-01607-9

    This paper supplies the reported source protocol, session timing, eye-state order, and open-eye ICA/ICLabel processing.

  5. doi:10.1038/nn.2303

Cited by

1 publication

Public catalog entries on Sciscape.