150 ms confirmation reduced mean SEEG background outputs in eight patients during two 3-minute tests with fixed thresholds

We tested a fixed 150 ms score confirmation rule for brain-computer interface outputs. The data came from eight patients with epilepsy in OpenNeuro ds004473 version 1.0.2. Each patient had a fixed decoder and early rest threshold. In the first 3-minute test, mean background outputs fell from 12.8 to 6.0 per minute. The mean paired change was -6.77 per minute. The exact patient sign-flip p value was 0.0078125. Mean keypress detection changed from 55.5% to 45.1%. Detection before keypress changed from 32.9% to 17.0%. A separate later 3-minute check gave 12.9 and 6.0 background outputs per minute. This check was added after the first result was known. It used the same patients and fixed rules. The package preserves two full decision grids, code, source versions and hashes, result tables, and four figures. All continuous inputs were checked from public raw ranges. Background is a defined task-log measure. The study does not test user intent, online control, a new day, or stimulation.

stereo-EEG · brain-computer interface · command confirmation · background outputs · event detection · signal processing · reproducibility