Removal of 118–122 Hz gave positive mean and negative median excess edge turnover in 19 bipolar 80–150 Hz SEEG records

SEEG graphs can change when frequencies near a line harmonic are removed. We tested OpenNeuro ds003876, version 1.0.2. The fixed sample had 19 patient records, 1,773 bipolar nodes, and 82,162 eligible edges. We removed [118,122) Hz from squared band coherency graphs in the [80,150) Hz band. We compared this removal with five intervals of equal width. Mean edge turnover was 16.01 percent for the primary removal and 7.34 percent for controls. The mean excess was 8.67 percentage points. The median excess was minus 0.91 points. Seven records had positive effects. Twelve had negative effects. The exact two-sided sign-flip test gave p = 0.10760. It requires independent patients and null sign symmetry. The nominal patient bootstrap interval was 0.26 to 19.81 points. The planned test did not establish a positive group effect at the 5 percent level. Some individual graphs changed substantially. The package preserves source code, metadata, pinned retrieval recipes, spectral statistics, all patient results, raw examples, and validation. The work tests graph sensitivity. It does not prove that removed signals were electrical artifacts. It does not establish clinical value.

stereoelectroencephalography · SEEG · intracranial EEG · high gamma · line-frequency harmonic · squared band coherency · graph edge turnover · frequency omission · signal processing