The 10 Hz EEG block product had a positive mean with separate power weights in 88 processed ds004504 v1.0.9 records

We made a signed block product for imaginary EEG cross spectra. A third block set the power weights. A proof shows that the score has no self-product noise term when block means are equal and block errors are uncorrelated conditional on calibration. Separate time blocks do not establish these conditions. Estimated power weights change the target. The score is not ordinary multivariate interaction measure. We tested all 88 processed eyes-closed records in OpenNeuro ds004504 version 1.0.9. Each of two score blocks had 50 seconds of accepted data. Calibration used the first 100 accepted seconds. At 10 Hz, the equal-person mean was 0.06431, with a 95 percent person-bootstrap interval of 0.04241 to 0.08956. Fifteen scores were negative. All fixed time and calibration checks had positive means. The 6 Hz and 20 Hz intervals included zero. In 44,000 independent zero-coupling model draws, the mean block product was 0.000087 and the mean self product was 0.08381. A model with dependent block errors gave a positive block product with zero true coupling. An independent trace formula checked 528 scores. A fresh source download and direct DFT checked one record. A full replay matched all numerical arrays. The package preserves code, Fourier data, source hashes, all score draws, proof, results, and instructions. The EEG result does not establish independent errors, a neural source, or a causal effect.

electroencephalography · imaginary cross spectrum · signed block product · cross-validation · sensor-space contrasts · EEG connectivity