Onset gamma rise and post-seizure gamma fall had mean cross-seizure rank correlation -0.033 within shaft and label groups in 29 SEEG patients

We tested whether local gamma rise at seizure onset had a positive spatial link to gamma fall after another seizure. We used public OpenNeuro ds004100 version 1.1.3 and reused 64 feature records from 32 SEEG patients. The main test had 29 patients, 353 shaft and clinical label groups, and 2,811 pair locations. Each group had at least four matched bipolar pairs. Gamma used 30–57 and 63–80 Hz. Onset was the first 10 seconds. Post-seizure time was 10–30 seconds after the source end mark. Mean two-direction cross-seizure Spearman correlation was -0.0325. Its 95 percent patient bootstrap interval was -0.1113 to 0.0364. The primary endpoint subtracted the corresponding relation to drift before onset. Its mean was -0.0723, with interval -0.1484 to 0.0020 and two-sided random sign-flip p = 0.075766. The tested positive relation was not supported. The test did not establish equivalence or biological absence. All seven fixed descriptive checks had non-positive mean paired contrasts. The sample mainly represented non-SOZ contact groups. Population inference assumes independent, exchangeable patients. The sign-flip null also assumes symmetric patient differences. Cohort reuse and selection limit the scope. The package preserves the local fixed plan, source code, all feature arrays, source credits, fixed public object versions, range hashes, results, uncertainty draws, and repeat instructions. Four fresh source rebuilds at four sample rates matched every array. Separate signal and rank checks passed. A complete repeat gave identical numerical results. This is a new endpoint in reused patients, not a new patient cohort.

SEEG · intracranial EEG · gamma power · seizure onset · post-seizure · cross-seizure rank correlation · epilepsy