Alternate-window maps had higher 80–150 Hz SEEG concentration rank agreement within than between recordings in 34 patients

We tested how a recording split changes SEEG concentration rank agreement. We used the first 60 seconds of two interictal files from each of 34 patients in OpenNeuro ds004100 version 1.1.3. The test used 3,325 fixed adjacent bipolar pairs. Power was 80 through 150 Hz, with 118 through 122 Hz excluded. Four maps each used 30 one-second windows. Concentration was the power in the three largest windows divided by power in all 30 windows. Mean patient rank correlation was 0.589 within recordings and 0.448 between recordings. The mean difference was 0.142, with a 95 percent patient bootstrap interval of [0.099, 0.189]. The difference was positive in 31 patients. The planned random sign test gave p = 0.000005, its simulation floor. A later exact calculation gave p = 0.0000000497. These tests require independent patients and symmetric null contrasts. All five planned checks gave positive mean differences. Top-set overlap was 57 percent within recordings and 37 percent between recordings. The result concerns these short clips and this reused cohort. It does not establish a neural cause or clinical benefit. The package preserves the plan, code, window powers, maps, random assignments, results, figures, and source recipe. All 72 source prefixes were retrieved again. All rebuilt feature arrays and quality records matched the archived inputs.

SEEG · intracranial EEG · power concentration · rank agreement · recording split · repeatability · OpenNeuro ds004100