Mean duration direction scores from start intervals were positive in time symmetric MNI HFO controls from 18 SEEG sleep records

We tested duration and time direction in released HFO event times from OpenNeuro ds005398 version 1.1.3. The primary setting used MNI episodes joined across gaps of at most 50 ms. It used the second half hour. It used complete 60 s blocks with at least five episodes. It retained 6,105 blocks on 362 channels in 18 SEEG sleep records. Each exact control gave equal weight to a block and its time reverse. The control start interval score was positive in all 18 patient means. Its mean was 0.01222. Its descriptive 95 percent patient bootstrap interval was 0.00811 to 0.01660. The control silent gap score was zero exactly. The original silent gap score was 0.01366, with an interval of -0.01116 to 0.03712. It did not establish a common direction. A proof shows the duration term inside start intervals. A separate circular proof gives expected normalized start covariance score n/(n-1) under independent duration permutation. This expected score is positive. These results limit time direction claims based on start intervals. They do not establish neural time irreversibility or a clinical marker. The package preserves public input hashes, version pins, source tables, code, scores, exact checks, figures, and a full repeat-run receipt.

high-frequency oscillations · SEEG · time reflection · silent gaps · rank correlation · duration · interval statistics