Output phase tests have a mean 70.89 percent rejection rate at nominal 5 percent for two step Laplacian beta pairs in eight SEEG models with independent contact phases

We tested a local Laplacian reference with public SEEG spectra from eight patients in OpenNeuro ds004473 version 1.0.2. Each output was a center contact minus half of each good immediate neighbor. Channels with centers two contact steps apart share one input. In independent contact phase models, tests with phases drawn after the reference rejected 70.89 percent of these edge targets. Tests with phases drawn before the reference rejected 4.95 percent. The nominal rate was 5 percent. The mean patient difference was 65.94 percentage points. Each patient supplied 200 targets per edge and 19 controls per target. A covariance proof explains the positive shared input effect. The result applies to this fixed spectral model. It does not measure brain coupling. The package supplies source spectra, pinned retrieval recipes, checks, trial results, and a general montage audit tool.

stereoelectroencephalography · Laplacian reference · phase randomization · correlation testing · shared inputs · rank calibration