Combined HFO timing maps on contacts 1–2 showed positive mean agreement across dates in seven SEEG patients under fixed count, shaft, and hemisphere controls

We tested whether HFO timing maps repeat across dates in public SEEG data. We used 135 five-minute sleep records from seven patients in OpenNeuro ds003498 v1.1.1. The primary test used FRandR marks on bipolar contacts 1–2 and absolute delays from 0 through 10 ms. These marks indicate simultaneous ripple and fast ripple detection on one channel. We subtracted the exact expected pair count under uniform positions in fixed 1-second count bins. We removed fitted shaft and hemisphere terms. The equal-patient mean cosine between early and later maps was 0.483. Six patients had positive agreement and one had negative agreement. The descriptive 95% patient bootstrap interval was [0.046, 0.826]. A shaft label reference within each hemisphere gave two-sided p=0.00132, under conditional label exchangeability. All nine planned checks kept a positive mean. An added check after the primary results replaced times within the fixed bins and rebuilt the maps. Its mean was 0.015 and its two-sided p was 0.00095. That check assumes independent uniform conditional times. These results concern released detector marks in a selected cohort. They do not establish a causal connection or clinical value. The package preserves source annotations, provenance, plans, exact formulas, code, results, validation, figures, and reproduction instructions.

high frequency oscillations · stereoelectroencephalography · event timing · repeatability · epilepsy · sleep · timing maps