The 95 percent interval for reference minus time score order changes included zero at 64 SEEG mark boundaries in 19 patients

We tested a fixed power score with public SEEG data from OpenNeuro ds003876, version 1.0.2. The study used 64 adjacent contact pairs with different onset marks in 19 patients. Each pair had four score orders: stored and common average references in the first and last 56 second samples of one 120 second record. Patients had equal weight. The mean time change rate was 39.6 percent. The mean reference change rate was 28.6 percent. The reference rate minus the time rate was -11.0 percentage points. Its 95 percent patient bootstrap interval was -25.1 to 0.8 points. The exact patient sign-flip p value was 0.133. The bootstrap assumes independent patients from a common population. The sign-flip test assumes independent differences with symmetric signs under the null. The data did not establish which change gave more order changes at the mark boundaries. All 1,773 adjacent pairs showed time and reference change rates of 39.2 and 31.8 percent in a fixed secondary check. The result applies only to the tested score, references, samples, and cohort. It does not identify the correct reference or provide a surgery rule. The package preserves all source features, one raw EDF prefix, object versions and hashes, code, results, uncertainty draws, and raw rebuild instructions. A fresh raw rebuild for one patient matched the source hash and feature arrays.

stereoelectroencephalography · intracranial EEG · epilepsy · reference sensitivity · score order stability · power score · patient bootstrap