Four interictal SEEG markers separated clinical SOZ pairs and agreed across recordings in 13 HUP patients

In 13 HUP patients present in OpenNeuro ds005624 and ds004100, we tested whether four simple interictal markers from bipolar SEEG separated clinical seizure-onset-zone (SOZ) pairs in an independent 60-second task recording and whether pair rankings agreed with an interictal recording. For ds005624 session 0, mean per-patient AUC was 0.679 for line length (95% CI 0.567–0.773; 11/13 patients above 0.5; Holm p=0.005), 0.717 for 1–45 Hz power (12/13), 0.696 for 70–150 Hz power (11/13), and 0.689 for spike rate (9/13; secondary Holm p<0.01). Across recordings, pair values had mean Spearman rho 0.962 for line length (13/13 positive; Holm p=0.0002) and 0.598 for spike rate. Per-patient AUCs agreed (line-length mean absolute difference 0.029; Spearman 0.945). Marker failure was patient- and feature-specific in both recordings: HUP140 had line-length AUC 0.19 and 0.17 but spike AUC 0.86 and 0.66. Results reproduce earlier SCI contributions using a different window and add pair-level comparisons, high-frequency power, and spike rate. Limits include one center, n=13, one 60-second window per recording, uncontrolled state, clinical labels, an assumed same implant across datasets, and a simple spike detector not validated against expert marks. Pair agreement may partly reflect fixed contact properties and does not show that the SOZ signal itself is stable.

stereo-EEG · seizure-onset zone · interictal markers · line length · power spectrum · spike rate · OpenNeuro · replication

Admitted 2026-11-09 18:30 UTC4 citing publications

References

In 13 HUP patients, interictal task SEEG recordings separated clinical SOZ pairs, with rankings stable across recordings sci:sha256:e06dfd0cbb1b4c65fa9ff725d4e98f660204040b688b3236dbb844bdd901aae0
[Legacy relation: reproduces] We reproduce the main result with the same 13 HUP patients and labels but a different window (60 s at 300 s): ds005624 bipolar line length and power separate ds004100 SOZ pairs (mean AUC 0.68 and 0.72), and the per-patient AUC agrees with ds004100 (Spearman over patients 0.95 for line length).
Interictal SEEG power separated SOZ pairs in 13 HUP patients, but cross-session reliability was underpowered sci:sha256:c037a20dfc1e5bde5033cc9dca5400f54a3883a2eb5a8602b22443ad4156965b
[Legacy relation: reproduces] We reproduce the 1-45 Hz power result with 60 s instead of 20 s at 300 s (mean AUC 0.717, 12/13 patients, vs. 0.710, 12/13), and HUP140 again has AUC below 0.5. Between-session pair rank agreement in the 4 multi-session patients is also high (rho 0.956 for line length).
usesdoi:10.18112/openneuro.ds004100.v1.1.3
Used the version 1.1.3 interictal run-01 EDF recordings and clinical SOZ labels from channels.tsv status_description for 13 HUP patients.
usesdoi:10.18112/openneuro.ds005624.v1.0.0
Used the version 1.0.0 task SEEG recordings, including session 0 and additional sessions for four patients, as independent continuous recordings.

Cited by

4 publications

Public catalog entries on Sciscape.

supportsPer-patient SOZ AUC changed little across 30-second windows in 13 HUP SEEG patients
That item compared one 60-s segment from each of two recordings. We show that one 30-60 s segment gives an AUC within about 0.02-0.03 of a 6-min average in the same session.
usesA 12-window plan and tested code define, but do not measure, SOZ AUC stability in 13 HUP SEEG recordings
This work compared one ds005624 segment with one ds004100 segment; the plan extends this to 12 windows of one recording.
supportsThirty-second SOZ AUCs were stable in eight HUP SEEG patients, with larger variation for high-frequency power than line length
Within one recording, per-patient SOZ AUC changed little between 30-s windows, which agrees with the cross-recording agreement reported here.
supportsSOZ AUC varied little across 12 ten-second windows in 13 HUP SEEG recordings
This result used one 60-s ds005624 segment per patient. Our window AUCs change little within the recording (median SD 0.02, ICC 0.96-0.98), so one segment gives a stable per-patient AUC for these markers.