Mean high gamma rank agreement was 0.970 after name group adjustment across adjacent minutes in 19 SEEG records with common average reference

We tested whether high gamma power ranks stay stable after removal of differences between contact name groups. We used 19 public SEEG records from OpenNeuro ds003876, version 1.0.2. Each record supplied two adjacent 60-second windows with a common average reference. The test used 2046 contacts in 217 name groups. Mean pooled, name adjusted, and random adjusted rank correlations were 0.975, 0.970, and 0.975. The mean random-minus-name difference was 0.00484. Its 95 percent patient bootstrap interval was -0.00047 to 0.01159. The interval includes zero. This supports high short-term rank agreement after group adjustment in these records. It does not establish equal agreement or long-term reliability. We rebuilt all source power arrays from pinned public signal ranges. All 1117800 power values matched exactly. Independent calculation checks and a full repeated analysis passed. The package preserves code, power arrays, signal examples, random controls, source pins, hashes, results, limits, and repeat instructions.

stereoelectroencephalography · high gamma · rank correlation · intracranial EEG · common average reference · group adjustment · short-term stability