Energy mean times reduced repeat gaps more before than after limbic stimulation in seven SEEG patients with fixed correction and 45 Hz filtering

We compared energy mean times with largest absolute peak times in public SEEG data. The source was OpenNeuro ds004696, version 1.0.1. We used 80 limbic stimulation groups from seven patients with epilepsy. The test used ten pulses per group, fixed source correction, and a 45 Hz filter. We compared two separate five-pulse means in all 126 distinct partitions. The energy mean reduced repeat gaps by 27.23 ms after stimulation and 81.38 ms before stimulation. The paired patient mean difference was −54.15 ms. Its conditional 95% patient bootstrap interval was −67.34 to −38.39 ms. All seven patient differences were negative. The exact two-sided sign-flip probability was 0.015625 under patient independence and null symmetry. Four fixed checks kept the same mean direction. An independent implementation checked all 1,528,128 primary timing values. Twelve numeric files matched byte for byte in a fresh process. One full pinned source file was retrieved and rebuilt with exact array matches. A later proof showed that independent Gaussian samples give the energy time 1/52 of peak-time variance on a 102-sample grid. This proof is not a model of filtered SEEG. Smaller repeat gaps alone did not establish more accurate neural timing. The package preserves code, voltage windows, source versions, results, figures, and reproduction instructions.

stereoelectroencephalography · SEEG · limbic stimulation · energy mean time · peak time · repeatability · Gaussian noise · signal processing