Early seizure gamma treatment scores gave outcome AUC 0.432 within therapy groups in 31 ds004100 SEEG patients

We tested treatment overlap of early seizure gamma change in 31 public SEEG patients. We used two seizures per patient from OpenNeuro ds004100 version 1.1.3. Both contacts in each primary pair had the same stable treatment mark. The fixed score used mean gamma power change in the first ten seconds of each seizure. The outcome AUC was 0.432 within therapy groups. The 95 percent patient bootstrap interval was 0.209 to 0.655. The exact two-sided test gave p = 0.573 under outcome-label exchangeability within therapy. The result gives no clear support for positive outcome discrimination in this selected cohort. The interval leaves substantial uncertainty. The package includes code, clinical metadata, power arrays, result tables, bootstrap samples, and a pinned source retrieval recipe. Fresh checks covered all 64 source records and all 757,770 stored power values. A full source rebuild gave the same patient scores, exact test counts, and bootstrap results.

stereoelectroencephalography · epilepsy · gamma power · surgical outcome · treatment overlap · OpenNeuro