Earlier MNI HFO duration gave no clear gain in later duration estimates under patient resampling on 259 SEEG channels from 18 sleep records

We tested whether earlier MNI HFO duration improves later duration estimates in public SEEG data. We used two half-hour samples from 18 sleep records in OpenNeuro ds005398 version 1.1.3. The primary test had 259 bipolar channels with at least five marks per sample. Comparison channels on the same shaft shared no contact with the target. The baseline used earlier comparison durations, event count, and contact order. Each test patient was excluded from model fitting. Adding earlier target duration reduced mean error from 0.095345 to 0.084801 squared log-duration units. The mean gain was 0.010544, or 11.06 percent. The patient bootstrap 95 percent interval was [-0.075379, 0.082961]. This approximate interval covers harm and benefit. Population use requires independent, exchangeable patients. Several fixed secondary checks had positive intervals but often changed the target set. The count rule uses later support, so this is a retrospective test. The package includes exact source bytes, code, all refits, figures, and successful independent and repeat checks. It does not test clinical benefit.

high-frequency oscillations · stereoelectroencephalography · event duration · patient bootstrap · OpenNeuro