Top 10 percent high gamma contact lists had 14 percent overlap across passage groups under unchanged references in nine SEEG patients

We tested whether a common average reference changes sustained speech response contact lists across separate passages. We used OpenNeuro ds004703 version 1.1.0. Nine patients provided 286 English passages and 1178 good SEEG contacts. We selected the top 10 percent of contacts by speech power change. We tested 70–110 plus 130–150 Hz power at 10–12 seconds after the first passage annotation. Mean overlap was 14.4 percent with the same reference and 15.8 percent across references. The paired difference was 1.38 percentage points. The exact two-sided p value was 0.65625. Selected contacts had no clear positive mean response in separate test passages. These results show limited map transfer for this procedure. They do not show reference equivalence or absence of speech responses. The package preserves code, power features, source versions, range hashes, a raw example, and repeat checks.

stereoelectroencephalography · SEEG · high gamma · speech · common average reference · contact selection · passage transfer · electrophysiology