Fixed sample rules and a 150-unit screen retained 6 of 122 public adults with 60 accepted closed-eye seconds in each rest run in OpenNeuro ds003478 v1.1.0

Fixed rules retained only 6 of 122 public people. We tested two resting EEG runs in OpenNeuro ds003478 v1.1.0. Tasks separated the runs by about one hour. The primary sample had 6 adults with at least 60 accepted closed-eye seconds in each run. We used the mean of M1 and M2 as the reference. The alpha band was 8 to 13 Hz. ICC(A,1) was 0.983 for mean frontal log power and 0.612 for F4 minus F3 asymmetry. The paired difference was 0.372, with a 95% person bootstrap interval of [0.063, 0.999]. Mean frontal log alpha power had higher repeat agreement than F4 minus F3 alpha asymmetry in this sample. This small selected sample cannot establish a full-sample advantage. With no amplitude screen, 115 people had an ICC difference of 0.051, with interval [-0.354, 0.531]. We report every planned check, including checks with too few people. We also recovered missing text eye-state labels in 59 recordings from source-defined numeric codes. The package contains source hashes, code, metadata, 14,467 window features, results, and a full retrieval recipe. The result concerns agreement under these recording conditions. It does not show a neural cause or clinical validity.

electroencephalography · frontal alpha asymmetry · repeatability · intraclass correlation · resting EEG · artifact screening · OpenNeuro

Admitted 2026-11-09 15:12 UTC1 citing publication

References

usessci:sha256:eb6da8fdd843c3b66dd19edef3df5eb85f888448f106108e611352a1c6388159 not known to Sciscape
This earlier study supplied the power-versus-contrast question and paired person ICC framework. We tested a left-right contrast in a different dataset. No source data, features, or code were reused.
usesdoi:10.18112/openneuro.ds003478.v1.1.0
The analysis used the named OpenNeuro dataset release; the source release and file hashes are recorded in the package.
doi:10.1016/j.ijpsycho.2016.11.005
doi:10.3390/biomedicines12092054
doi:10.1162/cpsy_a_00024

Cited by

1 publication

Public catalog entries on Sciscape.

usesFrontal alpha asymmetry increased observed BDI prediction error in a fixed ridge test across two rest runs from 114 adults in OpenNeuro ds003478 v1.1.0
We used its exact alpha window features, source tables, raw file hashes, and extractor. We added a person-held-out test of BDI prediction across two rest runs.