With a fixed numeric pair rule, assigning mixed pairs to the onset class lowered mean 80 to 150 Hz power AUC by 0.048 within name groups in two 60 s clips from each of 30 SEEG patients

We tested pair label rules in public SEEG data from OpenNeuro ds004100, version 1.1.3. Each of 30 patients supplied two 60 s clips. We used fixed numeric pair selection and comparisons within contact name groups. A mixed pair had one contact with a seizure onset label. OR assigned mixed pairs positive. AND assigned them negative. The signals, scores, and 69 groups were fixed. Mean 80 to 150 Hz power AUC was 0.7700 under OR and 0.8176 under AND. The mean OR minus AND difference was -0.0476, with a 95% patient bootstrap interval of -0.0883 to -0.0096. The patient sign-flip test gave p = 0.02596. It assumes independent patients and symmetric patient differences under the null. The mean absolute change before averaging was 0.1148. Both clips and four fixed sensitivity checks had negative mean differences. The result shows dependence on the pair label rule. It does not establish the more accurate tissue label rule. The package preserves endpoint labels, window powers, source identities, code, resampling draws, independent checks, figures, and reproduction steps.

SEEG · stereo-electroencephalography · seizure onset zone · bipolar pairs · pair label rules · AUC · power · high-frequency activity · sensitivity analysis · intracranial EEG