Linear contact number adjustment reduced the mean SOZ slope contrast within name groups in 30 to 100 Hz common average SEEG spectra from 29 patients

We tested a common linear contact number term in a spectral slope comparison within contact name groups. The source was public SEEG data from OpenNeuro ds004100 version 1.1.3. Both models used the same 619 contacts in 29 selected patients. The primary feature used common average reference and 30 to 100 Hz power, with 57 to 63 Hz omitted. The mean seizure onset zone (SOZ) coefficient decreased from 0.288 to 0.046 beta units after adjustment. The mean paired change was -0.243. Its 95 percent patient bootstrap interval was -0.462 to -0.035. The planned two-sided sign-flip test gave p = 0.03619. Several planned check intervals included zero. The result depends on the model and source contact selection. Contact number is not measured brain depth. This change does not prove anatomical confounding or absence of a remaining SOZ contrast. The package preserves 34 source spectrum files, both clinical label columns, the saved plan, code, results, figures, source metadata, and fixed raw data retrieval recipes. A raw source rebuild reproduced all 8,083,488 archived power values. Independent regressions and a fresh-copy reproduction passed.

stereoelectroencephalography · seizure onset zone · spectral slope · contact number · intracranial EEG · common average reference · patient bootstrap · sign-flip test