An exact check reduced the largest background rate at stricter cutoffs in one of eight fixed SEEG replays with commands at positive run starts

A command at the start of a positive score run can have a low count when almost all scores are positive. A larger cutoff can split these runs and increase the count. We tested an exact early check at every stricter cutoff. We used public SEEG data from eight people in OpenNeuro ds004473 version 1.0.2. The first 40% of retained trials fit the reader. The next 20% set the cutoff. The final 40% gave the later test. The check changed one cutoff. In that record, the largest later background rate at any stricter cutoff fell from 40.12 to 6.37 commands/min. The nominal rate rose from 2.63 to 6.37 commands/min. Response hits rose from 1/120 to 6/120. The other seven cutoffs and results stayed the same. Four of eight later records exceeded 5 commands/min with the new check. The proof gives a count bound only on a fixed calibration stream, including every constant score decrease. It gives no bound on a new stream. The later worst-rate inequality follows from cutoff order. The package contains code, compact data, source versions and hashes, exact curves, validation receipts, figures, and retrieval instructions. This is an exploratory reuse of one small cohort. It does not test live BCI use or brain stimulation.

stereoelectroencephalography · brain-computer interface · command rate · threshold selection · fixed-stream validation · positive run starts