Seizure-offset slowing did not replicate at the primary threshold in 56 SEEG seizures from 17 patients

This independent-center replication tested whether discharge intervals lengthen in the final 20 s of focal seizures, as reported in HUP data, in OpenNeuro ds007445 (UCLA; 56 sessions from 17 patients, ages 1.9–28 years). At the pre-specified primary threshold (K = 5 robust standard deviations), 33 seizures from 15 patients met the minimum event count. The patient-level ISI trend was positive but did not meet the criterion (rho = +0.16, p = 0.082); the final-minus-control log2 ISI ratio also failed ( +0.66, p = 0.13, 9 patients), and amplitude decreased (rho = -0.22, p = 0.036), failing all three pre-specified criteria. Median rate changed from 1.29 to 1.23 Hz, compared with 3.66 to 2.01 Hz in HUP. Slowing appeared at K = 3 and on the second-ranked channel, but not at K = 8. The direction of the ISI trend agreed with HUP in most runs, but the result was sensitive to threshold and channel in this mostly pediatric cohort. Ten sessions were not fetched; two patients were absent from the analysis. Age and center are confounded, offset times are clinical annotations, and the detector was not validated against manual marks.

seizure · stereoelectroencephalography · seizure offset · interspike interval · replication · OpenNeuro · independent-center replication