Allocation to two leading sites showed no clear gain in later 20–500 ms SEEG voltage scores at an equal pulse budget in five offline sessions

This study compared two pulse allocation rules for one specified SEEG recording target in five offline sessions. Both rules used 3K calibration pulses for K candidate sites. One screened with two pulses per site, gave extra pulses to two leaders, and chose between them. The even rule used three pulses per site and could choose any site. Separate final-five pulses supplied the 20–500 ms test score. OpenNeuro ds004696 release 1.0.1 supplied 39 sites, 11 groups, and 124 target panels in five patients. Choices changed in 29 panels. The equal-patient mean difference was -0.12% of the initial voltage-energy scale A; the descriptive 95% patient bootstrap interval was -4.21% to +2.45%, and the exact sign-flip p value was 1.000. The result did not establish gain or equivalence. The two-leader rule scored higher than an exact random concentration reference, a secondary comparison. Source pulses occurred in site blocks. Tests were separate at each site, but some preceded calibration at other sites. Live interface benefit remains untested. Fresh hash-pinned public downloads reproduced all five included patients exactly. The package retains compact input data, outputs, provenance, and figures. GPL analysis and extraction scripts are omitted, so the exact analysis implementation is unavailable in the public package.

brain-computer interface · electrical stimulation · SEEG · pulse allocation · site selection · voltage score