Adjacent contact subtraction did not consistently reduce 70 to 110 Hz power threshold crossings near labeled discharges in 22 SEEG sleep records

We tested whether adjacent contact subtraction reduces power threshold crossings near labeled discharges. We used public sleep SEEG from Falach and colleagues, Figshare version 3. The primary test included 22 patients and 441 events. We used 70 to 110 Hz power and separate 95th percentile thresholds from first-minute background data. Test events occurred later. The selected pairs were on shafts absent from each event's channel list. Mean crossing excess above background was 1.96 percentage points for recorded contact mean power. It was 2.88 points for adjacent difference power. The paired change was +0.92 percentage points. The descriptive 95% patient bootstrap interval was -0.44 to 2.47 points. Crossings decreased in 10 patients and increased in 12 patients. All five fixed sensitivity tests also had positive mean changes. The evidence does not show a consistent reduction under these settings. Sleep labels do not establish unwanted BCI commands or the absence of activity on other shafts. The package preserves code, source hashes, compact data, three figures, all results, and passed verification checks.

stereo EEG · adjacent contact subtraction · high frequency power · interictal discharges · brain-computer interfaces · signal processing