Remote SEEG site distances at 20–100 and 100–500 ms had mean rank correlation 0.394 between pulse groups in five sessions

We asked whether distances between stimulation sites had a similar rank order at early and late SEEG response times. We used OpenNeuro ds004457 version 1.0.2 and five clinical sessions. The main set had 138 remote sites, 1981 site pairs. Each site had four pulses in each of two separate groups. We removed an estimate of pulse noise from squared voltage distances. This correction assumes stable site means and independent pulse errors. The mean cross-time rank correlation was 0.394 after a pulse and -0.075 before it. The after-minus-before contrast was 0.468. It was positive in four patients. Its two-sided patient sign-flip p-value was 0.125. Descriptive whole-patient bootstrap limits were 0.205 to 0.695; they do not establish a population effect. Early and late repeat correlations were 0.744 and 0.609. Neighboring-site cross-time correlation was 0.218. The package adds a checked benchmark for response-time comparisons. It has data, all distance tables, code, three figures, and a verified raw-data recipe. These are reused patients and offline known-site tests. The result does not establish neural origin, reader transfer, or a stimulation benefit.

stereoelectroencephalography · electrical stimulation · response distances · rank correlation · brain-computer interface · reproducibility