Under patient independence and null sign symmetry, eight sleep SEEG patients showed no clear mean one-second FRandR occupancy deficit with 30-second count control

We tested public high frequency oscillation marks from OpenNeuro ds003498 version 1.0.1. FRandR means a fast ripple that overlaps a ripple. The analysis used 185 sleep SEEG records from eight patients with depth-only implants. It kept 31,176 marks in 55,710 seconds. We fixed each channel count within 30-second blocks. We compared occupied one-second bins with an exact independent uniform-time expectation. The equal-patient mean deficit was 1.605% of expected occupied bins. The 95% patient bootstrap interval was -0.199% to 3.516%. The exact two-sided sign-flip p value was 0.171875 under patient independence and null sign symmetry. Five patient effects were positive. A separate joint independent assignment model gave zero of 4,999 draws as extreme as the observed absolute mean. The add-one Monte Carlo p value was 0.0002, at simulation resolution. This joint model assumes independence across all marks, channels, blocks, and records, given each channel-block count. Bin width checks gave different signs. The result describes source detector marks and does not establish a neural mechanism or clinical benefit. The package preserves proofs, code, 570 source files, versions, hashes, all checks, and reproduction instructions.

high frequency oscillations · SEEG · sleep · event occupancy · conditional model · patient bootstrap