Reference mixtures fitted on early 10 Hz EEG gave an uncertain later mean wPLI gain in 88 ds004504 records
A reference mixture was fitted to maximize ordinary sample weighted phase lag index (wPLI). The test used all 88 processed eyes-closed records from OpenNeuro ds004504 version 1.0.9. Each record supplied 120 channel pairs at 10 Hz. Each mixture used constant nonnegative weights for mean19, Fz, Cz, and Pz. The weights summed to one. The fit used the first 100 accepted seconds. The test used the last 100 accepted seconds, with no shared time samples. The comparator selected the best separate reference on the same fit data. The mean fit gain was 0.01255 (0.01150 to 0.01365) wPLI units. The mean later gain was -0.00101 (-0.00342 to 0.00119). Parentheses give 95% whole-person bootstrap intervals from 10,000 draws. The later interval includes zero. The later mean gain is uncertain. The test does not establish equal scores. This tests sample scores within processed records, not neural accuracy or clinical benefit. The package preserves the Fourier input, fitted weights, results, bootstrap draws, code, and source hashes. A separate score calculation, alternative optimizer checks, two complete source replays, and a full numerical repeat passed.
EEG · weighted phase lag index · wPLI · reference mixtures · connectivity · time split
References
9- usesRaw triangle sums did not change after common reference subtraction in 88 ds004504 EEG records sci:sha256:a19e5a14a0cf0fe0b6e33b44c74607ab38e8a468cc87ebe7ce5112e8938f092e
The preserved Fourier archive supplied all 88 EEG records. Its source manifest and extraction code enabled pinned source checks and repeat instructions.
- usesConstant convex common reference mixtures of mean19, Fz, Cz, and Pz gave 10 Hz sample wPLI more than 0.01 above each vertex score in 22.3% of 10,560 ds004504 v1.0.9 cases sci:sha256:508bae8d920f1f1e0f1a1b36135cca7a0d4ecb5d447c9f27fe544fffb0ebda91
The preserved solver fitted constant reference weights to maximize ordinary sample wPLI on each fit block. The new test scored those weights on separate time samples.
- usesdoi:10.18112/openneuro.ds004504.v1.0.9
The analyzed records are from OpenNeuro ds004504 version 1.0.9.
- doi:10.3390/data8060095
- doi:10.1109/ACCESS.2023.3294618
- doi:10.1007/s11571-026-10464-w
- doi:10.1088/1741-2560%2F13%2F3%2F036016
- doi:10.1016/j.neuroimage.2011.01.055
- doi:10.1002/nav.3800090303
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesClipped debiased squared wPLI had a higher null mean in 10-second blocks than in pooled data from 88 EEG records under independent signs with equal probabilities
This audit uses its Fourier archive, source manifest, and two extraction code files. New source checks verified all selected data in 88 EEG records. New code calculates the signed scores and random-sign null means.