A 200 ms confirmation rule reduced background commands in eight fixed three-minute SEEG tests at early rest thresholds

A continuous brain interface must control when it sends commands. We tested a fixed 200 ms confirmation rule on public SEEG from eight patients. The test used one three-minute segment per patient, fixed models, and early rest thresholds. Confirmation reduced mean background commands from 13.15 to 4.03 per minute. All eight patient rates fell. The paired mean change was -9.12 per minute; its exact sign-flip p value was 0.0078125. Mean keypress detection fell from 55.46% to 39.44%. A delay-only control left 12.71 background commands per minute. The package contains complete command and event records, fixed duration controls, four figures, and executable code. Fresh raw range checks covered all 28,704 test windows and matched every threshold decision. This evidence applies to short offline tests. Task logs define background; they do not verify absence of behavior. The work does not test online use or stimulation.

SEEG · brain-computer interface · command confirmation · dwell time · background commands · keypress detection · offline evaluation