Edges with shared contacts are more common in the strongest 5% than in all 1–40 Hz bipolar edges from 12 patients with interictal SEEG

This audit used public interictal SEEG from 12 patients in OpenNeuro ds004100, version 1.1.3. Each patient supplied two fixed 60-second samples. The networks used consecutive-contact bipolar signals and absolute Pearson correlation at 1–40 Hz. Edges with a shared contact formed 33.7% of the strongest 5% of edges. They formed 2.73% of all candidate edges. Mean patient enrichment was 12.72 (95% patient bootstrap interval 11.34 to 13.92). Exclusion of these edges removed a mean of 21.9% of the original top-decile nodes from that set. Independent contact phases retained the shared subtraction terms. Independent phases after bipolar formation removed their dependence. The two nulls test a known method effect. They do not identify false observed edges or validate clinical targets. Slow transfers required the first 12 eligible patient identifiers as a convenience subset. The original 36-patient plan and the recorded change are preserved. The package includes endpoint audit code, independent decoder checks, source identities, sufficient statistics, tables, figures, and exact retrieval instructions.

stereo-EEG · bipolar signals · shared contacts · functional connectivity · Pearson correlation · epilepsy · phase surrogates · network analysis · signal processing