Energy center times met a 10 ms repeatability target less often than peak times in eight limbic SEEG sessions with a 20–500 ms window and a shape check on the first five pulses

We compared two response time summaries in public stereoelectroencephalography (SEEG) data from eight patients. The data were from OpenNeuro ds004696, version 1.0.1. A correlation check between means of the first two and next three pulses selected 196 of 886 stimulation-site and readout pairs. We used a 20–500 ms window. We compared the first five and final five good pulses. The energy center used time weighted by squared centered voltage. The peak used the largest absolute centered voltage. The planned mean-change difference was 0.92 ms, with a descriptive 95% patient bootstrap interval of [-8.88, 9.33] ms. This result did not show a mean-change advantage for the energy center. The 10 ms target was met by 34.3% of energy centers and 77.0% of peaks, with equal patient weight. The energy center fraction was lower in every patient. Shorter windows gave different results. We derived and checked a finite-grid timing bound. We also saved trial products for a noise audit. The study reuses a known cohort and is exploratory. It tests time-summary repeatability within sessions. It does not test neural delay or live BCI control.

stereoelectroencephalography · SEEG · limbic stimulation · response timing · energy center · dominant peak · repeatability · brain-computer interface · finite-grid bound · trial noise