Marked time gave no clear mean STE list agreement gain in 18 SEEG sleep records with independent uniform cutoff choices and equal patient weights

We tested whether marked time gives more stable high burden channel lists than event count. The source was OpenNeuro ds005398 version 1.1.3. We used 18 SEEG sleep records and compared two 30 minute periods. The primary test used STE marks and top-fifth lists. Cutoff choices were independent and uniform. Each patient had equal weight. Mean agreement was 74.41 percent for count and 75.53 percent for marked time. The mean change was +1.12 percentage points. The 95 percent patient bootstrap interval was -1.85 to +4.02 points under an independent-patient model. The exact sign-flip value was p = 0.47531 under its sign-symmetry model. This evidence does not establish a mean stability gain or equal stability. Duration weighting changed 9.42 percent of list membership mass on average. MNI secondary checks had lower agreement with duration. An added exploratory check gave exact bounds over all allowed cutoff choices. The package preserves source bytes, pinned retrieval recipes, code, all scores, and instructions. Fresh checks verified 94 source inputs. Independent calculations passed, and clean offline runs reproduced all 11 numerical files exactly. The result concerns released annotations. It does not test seizure localization or surgical benefit.

high-frequency oscillations · stereoelectroencephalography · event duration · event count · channel list agreement · repeatability · epilepsy