Energy centroids reduced repeat timing gaps more before than after stimulation in seven SEEG patients with fixed 400 ms windows

We compared energy centroid times with absolute peak times in public SEEG data from seven patients. The fixed cohort had 80 stimulation groups and 1,516 contact-group pairs. We used separate sets of five pulses, all 126 partitions, a 45 Hz filter, and fixed 400 ms windows. After stimulation, the repeat gap fell from 61.4 to 34.1 ms. Before stimulation, it fell from 131.1 to 49.2 ms. The paired reduction contrast was -54.7 ms. Its patient bootstrap 95 percent interval was -67.9 to -38.7 ms. The exact two-sided sign-flip probability was 0.015625. The test assumes independent patients and symmetric differences under a zero null. All seven contrasts had the same direction. Planned checks kept the cohort direction. A Gaussian-sample proof showed that centroids can agree without a shared signal. Repeat gaps do not establish accurate neural response times. The package preserves data, source identities, code, results, and repeat instructions.

stereoelectroencephalography · SEEG · energy centroid · repeatability · electrical stimulation · signal timing