Five pulses preserved more later SEEG utility than random sets of the same size under uniform weights in seven offline sessions

This offline test used public SEEG data from OpenNeuro ds004696 version 1.0.1. It included seven patients, 57 stimulation sites, 20 groups, and 216 target contexts. A Pareto filter used the first five good pulses at each site. It kept sites with a target and off-target response tradeoff that another site did not dominate. The final five distinct pulses supplied evaluation scores. The test integrated the best later linear utility across uniform target weights from zero to one. Random reference sets had the same size as each retained set. The filter retained 74.3% of sites. Its mean utility advantage was 0.0340 normalized score units. The 95% patient bootstrap interval was 0.0163 to 0.0538. All seven patient means were positive. The filter still lost 0.0329 units against all later choices. Random sets lost 0.0670 units. No group placed all fit pulses before all evaluation pulses. This result supports an offline site-set benchmark. Online benefit remains untested. The package preserves data products, code, figures, an exact score method, source hashes, a fresh source audit, and reproduction instructions.

stereoelectroencephalography · electrical stimulation · Pareto filter · site selection · cross-pulse energy · offline benchmark