The 95th-percentile pause threshold reduces mean SEEG trigger rates and response-window detection in later data from eight task recordings

We tested a small SEEG decoder during later pauses in eight task recordings. Each decoder used eight bipolar pairs and three power bands. We trained on the first half of trials and set thresholds with separate earlier pauses. The test used 38.36 pause minutes and 674 response windows. A one-second output lockout was used. The mean trigger rate fell from 30.13 to 8.26 per pause minute. The paired decrease was 21.87 (15.83 to 26.78) per minute, with a descriptive 95 percent person-bootstrap interval. Mean response-window detection fell from 67.7 to 31.5 percent. The earlier 5 percent positive-window rule was exceeded in later pauses for 3 of eight people. The result applies to selected task pauses in one recording per person. The package has source code, raw signal checks, window data, all decisions, figures, and source retrieval records.

stereo-electroencephalography · brain-computer interfaces · threshold calibration · trigger rate · response detection · intracranial EEG