In 11 patients, distance-adjusted SOZ CCEP N1 degree did not pass Holm correction

This pre-registered analysis tested whether clinical seizure-onset-zone (SOZ) stimulation sites send or recording contacts receive more single-pulse electrical stimulation CCEP N1 responses than non-SOZ units, after adjustment for distance and same-lead effects. It used OpenNeuro ds005448 STReEF files for 11 patients with SOZ labels (5 ECoG, 6 SEEG); a response was a finite n1_peak_sample. In the primary cohort, out-degree had mean AUC 0.569 (8/11 patients above 0.5; exact sign-flip p=0.222; Holm p=0.222), and in-degree had mean AUC 0.616 (8/11; p=0.028; Holm p=0.057). Neither pre-registered hypothesis passed Holm correction at alpha 0.05. Distance and same-lead adjustment reduced the raw AUCs (out-degree 0.615 to 0.569; in-degree 0.647 to 0.616), indicating that electrode geometry accounts for part of the raw SOZ effect. The in-degree trend depended on response definition: requiring checked==1 yielded AUC 0.560 (p=0.325), and the source files do not document the meaning of checked. The small cohort limits power; clinical SOZ labels were used without outcome or resection data, regional response rate was not controlled, and stimulation current and pulse count were not modeled. The publication package preserves the plan and its pre-analysis hash, analysis code, source file hashes, and per-patient results.

seizure-onset zone · CCEP · N1 response · single-pulse electrical stimulation · SOZ · epilepsy · ECoG · SEEG · distance adjustment · OpenNeuro