Assuming complete onset marks, baseline division lowered early gamma AUC within named shafts in 29 ds004100 SEEG patients

We tested whether division by baseline power improves early seizure gamma ranking against clinical onset marks within named SEEG shafts. The source was OpenNeuro ds004100, version 1.1.3. The fixed test used adjacent bipolar signals from 29 patients, 58 seizures, 62 shafts, and 472 paired anchors. Gamma covered 30–57 Hz and 63–80 Hz. Early power used the first ten seizure seconds. Baseline power used -30 to -10 seconds. Mean AUC fell from 0.904798 for absolute early power to 0.732833 for the ratio. The paired mean change was -0.171965. Its 95 percent patient bootstrap interval was -0.248235 to -0.098876. The ratio lowered mean AUC in 21 patients, raised it in four, and gave equal AUC in four. Baseline power alone had mean AUC 0.875901. All eleven descriptive checks had negative mean changes. The loss persisted when every comparison stayed within one shaft name. The test assumes complete source onset marks. Shaft names do not establish anatomy. The result does not establish treatment accuracy or a neural cause. The features came from an existing AutoSci cohort. Related all-contact studies use the same source and are dependent checks. Separate EDF decoding and spectrum checks passed for all 64 source files and 757,770 saved power values. The package preserves code, source features, raw examples, retrieval records, selection decisions, results, and reproduction instructions.

stereoelectroencephalography · seizure onset · gamma power · baseline division · area under the curve · epilepsy