Fast ripple marks had 0.101 mean excess neighbors at 0.1–1 second under a uniform five-second count control in eight sleep SEEG patients

We tested within-channel fast ripple onset timing in OpenNeuro ds003498 version 1.0.1. The sample contained 185 sleep SEEG records from eight patients with depth electrodes. Complete five-second blocks retained 100,315 marks in 15.58 hours. The control held each channel block count fixed and placed distinct onsets uniformly within that block. At onset distances from 0.1 to 1 second, the sample mean excess was 0.101 neighbors per mark. Six of eight patient scores were positive. The exact two-sided sign reversal p value was 0.0703. This test requires independent patients and symmetric patient contrasts under the zero null. It did not reject zero at a 0.05 threshold. The 95 percent percentile patient bootstrap interval was 0.012 to 0.225. This interval is not an inversion of the sign reversal test. The mean excess changed with block size and was smaller at distances from 0.5 to 1 second. The uniform control does not preserve event duration or non-overlap. The result measures source detector marks. It does not establish a neural cause, clinical benefit, or equivalence to uniform timing. The package preserves source tables and hashes, all planned checks, code, an exact mean proof, an independent pair check, and repeat instructions. A separate run reproduced ten numerical files byte for byte.

fast ripples · SEEG · intracranial EEG · event timing · conditional interval jitter · epilepsy · high frequency oscillations