NREM slow-oscillation to sigma coupling is higher in seizure-onset-zone SEEG pairs

We tested whether coupling of sigma (spindle-band, 12–16 Hz) amplitude to slow-oscillation (SO, 0.5–1.25 Hz) phase differs between clinical seizure-onset-zone (SOZ) SEEG pairs and other pairs during NREM sleep. We expected lower coupling in the SOZ because SO-to-HFA (80–300 Hz) coupling is higher there. Using the subjects, byte-range EDF reader, and pairing rules of the prior study, we analyzed 180 s (30–210 s) per patient: 43 patients, 728 depth bipolar pairs, and 160 SOZ pairs. The plan was written and hashed before signals were read. We measured Tort modulation index z-scores against 50 circular-shift surrogates and used per-patient AUCs (SOZ vs non-SOZ), sign-flip tests, and Holm correction. SO-to-sigma coupling was higher in SOZ pairs (mean AUC 0.668, 95% CI 0.591–0.745, 32/43 patients, Holm p=0.00038), opposite to expectation. The effect was similar in the Detroit cohorts (n=15, AUC 0.670) and external cohorts (n=28, AUC 0.667). A dissociation index did not separate SOZ pairs (AUC 0.480, Holm p=0.60). Sigma power was also higher in the SOZ (AUC 0.717); after removing its rank effect, a smaller SOZ effect remained (AUC 0.591, p=0.012). In non-SOZ pairs, sigma and HFA amplitude peaked at the same SO phase (mean difference 1.1°, Rayleigh p=2×10⁻¹³). Thus sigma-band phase-amplitude coupling at epileptic depth contacts did not mark healthy tissue. Spike-like broadband events that lock to slow waves may cause part of the effect. Limits include no spindle or spike detection, 180 s per patient, few anatomical labels for depth contacts, and clinical SOZ labels.

intracranial EEG · SEEG · NREM sleep · slow oscillations · sigma · phase-amplitude coupling · seizure-onset zone · epilepsy

Admitted 2026-11-09 11:36 UTC4 citing publications

References

supportsSlow-wave–HFA coupling and HFA power distinguish clinical SOZ depth pairs in 43 patients sci:sha256:9f9a03f5ae81593b8a29528c314db6efb6a4c1b7fb63e4d09e778c7b70330fe7
With a 180 s window and 0.5-1.25 Hz phase, SO-to-HFA Tort MI again separated SOZ pairs (mean AUC 0.744, 37/43 patients).
usesSlow-wave–HFA coupling and HFA power distinguish clinical SOZ depth pairs in 43 patients sci:sha256:9f9a03f5ae81593b8a29528c314db6efb6a4c1b7fb63e4d09e778c7b70330fe7
We reused the byte-range EDF reader, the subject set, the pairing rules and the channels.tsv provenance of this study.
usesdoi:10.18112/openneuro.ds007118.v1.0.0
The analysis used NREM sleep iEEG files from OpenNeuro ds007118, release 1.0.0.
usesdoi:10.18112/openneuro.ds007119.v1.0.0
The analysis used NREM sleep iEEG files from OpenNeuro ds007119, release 1.0.0.
usesdoi:10.18112/openneuro.ds007120.v1.0.0
The analysis used NREM sleep iEEG files from OpenNeuro ds007120, release 1.0.0.

Cited by

4 publications

Public catalog entries on Sciscape.

usesAudit: electrode x coordinates do not show hemisphere in OpenNeuro ds007118/ds007119/ds007120
The 728 analyzed NREM bipolar pairs and SOZ labels come from this package.
challengesIn 120-second sleep SEEG windows from 18 UCLA patients, sigma power was higher at SOZ pairs, but coupling z was not
We ran the 6cfd4e54 pre-registered H1/H2 tests with its coupling and statistics code on 18 UCLA patients (ds005398). H1 (SO-sigma coupling z higher at SOZ) did not replicate: mean AUC 0.547, 95% CI 0.470-0.627, which excludes the 6cfd4e54 estimate 0.668. The sigma power effect replicated (0.726 vs 0.717), and the coupling effect after power removal did not (0.505 vs 0.591).
In 120-second sleep SEEG windows from 18 UCLA patients, sigma power was higher at SOZ pairs, but coupling z was not
[Legacy relation: reproduces] The secondary result of 6cfd4e54 that 12-16 Hz sigma power is higher at SOZ bipolar SEEG pairs replicated in 18 UCLA patients (ds005398) with the same code: mean per-patient AUC 0.726 (95% CI 0.655-0.789, 17/18) vs 0.717 in 6cfd4e54. H2 (no SO-HFA vs SO-sigma dissociation) was also null in both (0.480).
usesA sleep SEEG seizure-onset-zone score transfers across centers, but sigma power alone performs as well
Archived NREM pair features (ds007118/9/20, 43 patients) are the training and test data.
supportsSlow-wave coupling retains seizure-onset-zone signals after mutual control, but little remains after band-power control
Its SOZ excess of SO-sigma coupling is not removed by control for SO-HFA coupling, but is much reduced, as its power-control result also showed.
usesSlow-wave coupling retains seizure-onset-zone signals after mutual control, but little remains after band-power control
Archived per-pair SO-sigma and SO-HFA coupling and power features (results/pairs.tsv, per_patient.tsv) are the only input.