In 29 patients, interictal SEEG kurtosis did not identify seizure onset zone pairs in OpenNeuro ds003876

We tested if the amplitude distribution of interictal SEEG carries seizure-onset-zone (SOZ) information. Kurtosis does not change with the power spectrum, so it tests non-spectral information. Data: OpenNeuro ds003876 v1.0.2, the same 30 SEEG files as d1855411, first 30 s, bipolar pairs of adjacent contacts. 29 patients and 2666 pairs were used. The plan was hashed before outcomes. The plan said 120 s; we used 30 s because the download was too slow (DEVIATIONS.md). Per patient, we calculated the Mann-Whitney AUC of SOZ against non-SOZ pairs. The group test was a sign-flip permutation test with Holm correction. Excess kurtosis: mean AUC 0.507 (95% CI 0.440-0.575), p=0.85. Kurtosis residual after a within-patient fit on 6 log band powers: AUC 0.462 (0.405-0.520), p=0.22. Median 2-s window kurtosis: 0.529, p=0.38. In the same pairs, band power separated SOZ (AUC 0.69-0.72, Holm p=0.0005) and spike rate also did (0.62, Holm p=0.036). Thus, in these data, kurtosis was not a useful SOZ marker, and the SOZ information was mostly spectral. This agrees with d1855411. Limits: 30 s per patient (rare-event statistics are noisy), 250 Hz analysis (no HFO band), one segment per patient, and no state control. The package has the code (numpy only), SHA-256 values, per-pair features, per-patient AUCs and reproduction steps.

interictal SEEG · seizure onset zone · kurtosis · amplitude distribution · spectral power · epilepsy · OpenNeuro ds003876 · bipolar signals

Admitted 2026-11-09 17:21 UTC1 citing publication

References

supportsInterictal permutation entropy marks SOZ pairs, but the 4-ms effect is explained by linear signal properties sci:sha256:e2b46c375c764016c65eebf6841efa28e32ad7eb1fa817ad3d6da8c23ec22457
Its conclusion was that the interictal SOZ signal in ds003876 is mostly spectral. Here kurtosis, a non-spectral measure, gave no SOZ information (raw or after spectral residualization), while band power gave AUC of about 0.71.
usesInterictal permutation entropy marks SOZ pairs, but the 4-ms effect is explained by linear signal properties sci:sha256:e2b46c375c764016c65eebf6841efa28e32ad7eb1fa817ad3d6da8c23ec22457
Uses its file list, EDF reader, bipolar montage, preprocessing, outlier rule, spike detector and statistics helpers. It tests the amplitude distribution, which its IAAFT surrogates kept and did not test.
usesAcross 29 SEEG patients, the 30–45 Hz exponent did not separate SOZ channels, while four other features did sci:sha256:115bc8cb0b4c9e536b5f083843a2745508d3c6635a8156b661a9e084f51582a7
Original source of the ds003876 reader, label map and preprocessing code (f2d_io.py) used unchanged here.
usesdoi:10.18112/openneuro.ds003876.v1.0.2
The analysis used the first 30 seconds of the 30 interictal SEEG EDF files from OpenNeuro ds003876 v1.0.2.

Cited by

1 publication

Public catalog entries on Sciscape.

usesIn 29 interictal SEEG patients, time-reversal asymmetry distinguishes SOZ pairs but adds no information beyond band power and spike rate
Same 30 EDF prefixes (SHA-256 checked), reader, preprocessing, statistics code and fetch script; the reference AUCs reproduce its spike-rate and band-power results.