Five-pulse SEEG means had higher 20–100 ms shape similarity than before-pulse controls in three recordings with an odd–even split

Separate five-pulse SEEG means had higher shape similarity 20–100 ms after surface stimulation than in a before-pulse control. The test used 240 pulses at 24 sites in three recordings from OpenNeuro ds004774 version 1.0.0. The fixed odd–even split gave mean similarity 0.262 after the pulse and -0.032 before it. All three patient changes were positive. The mean change was +0.294. The exact two-sided patient sign-flip probability was 0.25. A first-versus-last-five check lost the increase in one patient. Recorded contact order changed in that patient. The evidence is limited to offline repeatability. It does not establish neural origin, detector accuracy, or live BCI performance. A full source audit found 780 SEEG response label cells from secondary raters in only two recordings. The primary rater had no SEEG response labels. The package preserves native contact voltage, all source metadata and annotations used, range hashes, figures, result tables, source code, and reproduction instructions. A fresh download checked all 240 raw ranges. A separate decoder checked all 2,983,680 contact values. A separate calculation checked 288 similarity values. A source size limit in one recording and the selection change are recorded.

SEEG · evoked responses · shape similarity · repeatability · electrical stimulation · brain-computer interface · trial averaging