Mean overlap of top 10 percent beta graphs was 0.126 above the uniform contact shuffle mean within shaft groups across adjacent minutes in 19 selected SEEG records

We tested whether beta graph overlap depends on contact identity within electrode shaft name groups. The fixed sample had 19 public SEEG records from OpenNeuro ds003876 version 1.0.2. Each record supplied two adjacent minutes. Each graph selected the top 10 percent of cross-group links by mean absolute imaginary coherency. Mean Dice overlap was 0.2835. Uniform contact shuffles within shaft groups gave an exact mean of 0.1572. The mean gap was 0.1263, with a descriptive 95 percent patient bootstrap interval of 0.0670 to 0.2002. All 19 gaps were positive. No mean from 9,999 control draws reached the observed mean. The corrected Monte Carlo value was 0.0001 under conditional invariance of the test graphs under the allowed contact maps. A finer control that also kept contact number parts gave a mean gap of 0.0672. The uniform edge baseline gave a mean gap of 0.1834. Keeping shaft structure reduced that gap by 31.1 percent. The result shows contact alignment under the specified controls. It does not establish neural causes or clinical accuracy. This is an informed follow-up in a reused cohort. All 4,222,800 Fourier values were rebuilt from fixed public EDF ranges. Separate calculations, full source precision, metadata checks, and two complete replays passed. The package preserves inputs, code, proof, source records, all result draws, figures, and repeat instructions.

SEEG · EEG · graph overlap · imaginary coherency · contact alignment · permutation test · epilepsy