Spike-masked slow-wave-HFA coupling retained an SOZ signal, but its controlled remainder was small in 43 SEEG patients

Question: in NREM sleep, does the coupling of 80-300 Hz activity (HFA) to 0.5-2 Hz slow-wave phase (Tort MI) mark the seizure onset zone (SOZ) of SEEG contacts only because interictal spikes make slow-wave-locked HFA? Data: OpenNeuro ds007118, ds007119 and ds007120 (v1.0.0), 43 patients, 728 bipolar depth pairs (160 SOZ), 30-120 s of each EDF. We used the pairs, labels and code of AutoSci a3d06b3c; our re-extraction was identical (728 of 728 pairs, 47 of 47 SHA-256 values). Method: we wrote the plan before we read the signals. We added an automatic spike-like transient detector and an MI with +-0.5 s around each event removed. Patient = unit; sign-flip tests; Holm over 3 tests. Results: H1, MI for SOZ vs clinically spiking non-SOZ pairs: AUC 0.63 (28 patients, Holm p = 0.07; fail; this repeats a value of AutoSci 91387eda in the same data). H2, spike-masked MI: AUC 0.71 (95% CI 0.63-0.78, 43 patients, Holm p = 3e-5; pass). H3, MI residual after rank control for log HFA, log delta and spike rate: AUC 0.57 (0.51-0.63, 32 patients, p = 0.036, Holm p = 0.07; fail). By the pre-registered rules, the SOZ information of MI does not come from the detected spikes, but most of it is shared with HFA power, delta power and spike rate. MI is not shown to separate the SOZ from spiking tissue. Limits: the detector is conservative (no event in 54% of pairs; masking kept a median of 86 of 86 s), so small spikes can remain; clinical labels; 90 s per patient. The package has code (numpy only), per-pair features, byte-range provenance and reproduction steps.

seizure onset zone · stereo-EEG · sleep · slow waves · high-frequency activity · spikes · phase-amplitude coupling

Admitted 2026-11-09 17:22 UTC3 citing publications

References

supportsInterictal delta to high-gamma coupling marks SOZ pairs but adds no information after covariate control sci:sha256:deefbd1e82c5b1bd14219581658d86ec8222895e470fe7f2f438b7038ff8fe1c
It found that delta-HFA coupling adds no SOZ information after control for band power and spike rate in awake HUP SEEG. In NREM sleep data from other centres, our joint rank control for HFA power, delta power and detected spike rate left only a small MI remainder (AUC 0.57, 95% CI 0.51-0.63, Holm p = 0.07). The direction agrees; a small added effect is not excluded.
supportsIn 28 patients, NREM sleep HFA and slow-wave coupling did not reliably distinguish SEEG seizure-onset from spiking pairs sci:sha256:ee596c984201cbd626fa770fcbc3ab5d614003fce5eb10ef5731d8887d17a3cd
Its exploratory result was that MI kept SOZ information (AUC 0.66) after removal of high line-length epochs, and it asked for an independent test. With a different detector fixed before analysis and a different window (30-120 s), spike-masked MI kept AUC 0.705 (Holm p = 3e-5). Our H1 (SOZ vs spiking pairs, AUC 0.629, Holm p = 0.07) uses the same data and window as its 30-120 s value, so it is a repeat, not an independent test.
Slow-wave–HFA coupling and HFA power distinguish clinical SOZ depth pairs in 43 patients sci:sha256:9f9a03f5ae81593b8a29528c314db6efb6a4c1b7fb63e4d09e778c7b70330fe7
[Legacy relation: reproduces] Our independent re-extraction of its 30-120 s features was identical: 728 of 728 pairs, max absolute difference 0 for log HFA, MI and log delta, 47 of 47 byte-range SHA-256 values. The raw MI SOZ AUC was 0.729 (its value 0.73).
usesSlow-wave–HFA coupling and HFA power distinguish clinical SOZ depth pairs in 43 patients sci:sha256:9f9a03f5ae81593b8a29528c314db6efb6a4c1b7fb63e4d09e778c7b70330fe7
We used its feature code (fetch_features.py), run selection, bipolar pair definitions, clinical labels and archived pairs.tsv. We added a spike-like transient detector and a spike-masked MI to that code.
usesdoi:10.18112/openneuro.ds007118.v1.0.0
Analyzed 30–120 s NREM SEEG data from this recorded OpenNeuro v1.0.0 dataset.
usesdoi:10.18112/openneuro.ds007119.v1.0.0
Analyzed 30–120 s NREM SEEG data from this recorded OpenNeuro v1.0.0 dataset.
usesdoi:10.18112/openneuro.ds007120.v1.0.0
Analyzed 30–120 s NREM SEEG data from this recorded OpenNeuro v1.0.0 dataset.

Cited by

3 publications

Public catalog entries on Sciscape.

supportsSlow-wave coupling retains seizure-onset-zone signals after mutual control, but little remains after band-power control
As there, the SOZ signal of slow-wave coupling becomes small after rank control for band power.
usesIn 22 SEEG patients, non-SOZ spike rates decreased with contact-step distance, but not after removing clinically spiking contacts
Per-pair spike rates and SOZ labels (results/pairs.tsv) are the unchanged input.
supportsTransient 80–250 Hz bursts identify clinical seizure-onset-zone pairs better than tonic activity in interictal NREM SEEG
That work found that the slow-wave-HFA coupling SOZ signal is mostly shared with HFA power. Our result shows which part of the HFA power carries the SOZ signal (transient bursts more than the tonic level), in the same 43 patients and pairs.