Average reference increases alpha edge counts below minus 0.5 in 88 ds004504 v1.0.9 records at normalized ridge 0.001
We tested ridge partial correlation after common average reference in 88 processed EEG records from OpenNeuro ds004504 v1.0.9. Each record had 19 channels. Alpha was 8 to less than 13 Hz. The ridge variance was 0.001 times the mean channel variance under each reference. The median fraction of 171 edges below minus 0.5 was 60.8% after average reference and 5.3% under the supplied reference. The fraction increased in all 88 records. A matched absolute ridge check gave the same direction in all records. The proof shows that every ridge partial correlation approaches minus one if average reference imposes the only null direction. Additional null directions change that limit. Fifty-two records had numerical rank below 18, so the package retains their full spectra. A sufficient finite bound confirmed all edges at or below minus 0.5 in two primary records. The package includes a checked audit tool, complete source band matrices, proofs, results, source hashes, metadata, licenses, and reproduction commands. Separate regression calculations checked 15,048 primary edges. Two fresh source extractions and a second offline run reproduced the saved values. The result describes these covariance models. It does not show neural inhibition or establish true neural connectivity.
EEG · average reference · ridge regularization · partial correlation · covariance · singular covariance · null space · numerical rank · signal processing
References
7- usesCommon average reference changes posterior theta-to-alpha ranks in 88 ds004504 v1.0.9 records with source processing fixed sci:sha256:2047e7d0ae35196950f5421eb9237a3b0a817c451fabcbdedd0b92ad776f3431
The study uses all 6377 archived real band cross-power matrices for all 88 ds004504 v1.0.9 records. It copies the source extraction code and pinned source metadata. Two fresh source extractions reproduce the corresponding archived matrices.
- usesA fixed EEG model predicts shuffled labels when training and test sets share the 88 ds004504 subjects sci:sha256:7ba0bee1f809898d24fefdeb839b4d53689a26bcea0ff0429014af5f5ef0c2a9
The copied retrieval and boundary code contains adaptations from this publication. Its code license and publication metadata are preserved. Its classification result is not used as evidence for the ridge effect.
- usesdoi:10.18112/openneuro.ds004504.v1.0.9
The study analyzes the archived, processed band matrices for all 88 records in OpenNeuro ds004504 v1.0.9.
- doi:10.1016/j.acags.2020.100026
- doi:10.3390/data8060095
- doi:10.1109/ACCESS.2023.3294618
- doi:10.1007/s11571-026-10464-w
Cited by
4 publicationsPublic catalog entries on Sciscape.
- usesA positive definite EEG contrast model permits every selected channel partial correlation between minus one and plus one under unrestricted common signal covariance
We reuse its unchanged EEG matrix archive and source records. Its average reference audit defined the starting question.
- usesA positive covariance on EEG contrasts permits either sensor partial correlation sign when there are at least three sensors and the common signal is unrestricted
Used the archive of 6,377 band matrices from 88 EEG records, source metadata, and the reference covariance audit to construct and check the compatible sensor covariance examples.
- usesIndependent common variance up to 1% allows both partial correlation signs in 88 ds004504 alpha models at contrast ridge 0.001
Uses the unchanged 88-record spectral matrix archive and its normalized ridge setting. The new audit adds exact common signal variance limits.
- usesA change from average reference to Cz gives a median alpha Riemannian distance of 11.29 in 88 ds004504 v1.0.9 records with a ridge fraction of 0.001 in all 19 channels
This study retrieved the archived band matrices, spectral code, source manifest, and extraction audit from this publication. It tests a different distance endpoint with the source processing fixed.