Top ten percent gamma lists had 42.7 percent mean agreement across two seizures with a rank rule over five starts and 46.2 percent with one window in 28 SEEG patients with common average reference

We tested whether a fixed rank rule over five window starts improves gamma contact list agreement with another seizure. We used public OpenNeuro ds004100 version 1.1.3 powers from 56 seizures in 28 SEEG patients. Primary signals used common average reference. The cohort had 2,820 eligible anchors, counted once per patient. Each list contained the top ten percent of that patient's eligible anchors. The new rule used each contact's worst rank across starts at -4, -2, 0, +2, and +4 seconds. Mean rank resolved ties. One seizure supplied the candidate list. The other supplied the nominal target list. We averaged both test directions within each patient. Mean agreement was 42.7 percent for the rank rule and 46.2 percent for one nominal window. The mean difference was -3.49 percentage points. Its 95 percent patient bootstrap interval was -7.79 to +0.44 percentage points, under independent patient sampling and fixed study selection. The result did not establish a mean benefit, equivalence, or a lower population mean. All seven fixed descriptive alternatives also had negative mean differences. This is a new test in an existing clinical cohort. It does not establish a treatment list. The package preserves the complete source archive, new code, decisions, random draws, source checks, and repeat instructions.

stereoelectroencephalography · SEEG · gamma power · contact ranking · seizure agreement · window selection · patient bootstrap · OpenNeuro ds004100