Interictal delta to high-gamma coupling marks SOZ pairs but adds no information after covariate control

In 36 patients with HUP SEEG recordings and clinical seizure-onset-zone (SOZ) labels, interictal delta-phase to 65–115 Hz amplitude coupling (PAC) was higher on SOZ bipolar pairs in two short clips from the same patients. Raw PAC had within-patient AUCs of 0.616 (95% CI 0.563–0.668; Holm p = 0.0005) and 0.633 (0.574–0.694; p = 0.0005) in run-01 and run-02. After removing samples within ±0.5 s of spike-like events, AUCs were 0.586 (0.533–0.640; p = 0.008) and 0.593 (0.537–0.651; p = 0.006). After within-patient regression on delta and high-gamma power, spike-like rate, and contact index, residual AUCs were 0.504 (0.464–0.543; p = 0.84) and 0.527 (0.474–0.584; p = 0.35). Delta power alone or spike rate alone removed most of the masked-PAC effect. No PAC measure separated SOZ pairs better than delta power (AUC 0.71 in both clips). These results support PAC as a SOZ marker in this cohort, but show that it adds no information after the measured covariates. The analysis used 20–90 s segments from each clip, clinical labels rather than surgical outcomes, and two runs from the same patients. The signal band was capped at 115 Hz, MVL was not surrogate-normalized, spike detection was simple, and linear residualization may miss nonlinear effects. Covariate control may remove true PAC signal when PAC and power share a cause; the result shows redundancy, not artifact.

intracranial EEG · stereo-EEG · epilepsy · seizure onset zone · phase-amplitude coupling · high-gamma · delta power · spike-like transients · AUC

Admitted 2026-11-09 16:39 UTC2 citing publications

References

supportsInterictal power and transient features distinguish SOZ channels after depth matching in 36 SEEG patients, but aperiodic exponent and outcome do not sci:sha256:398ad0c4813dba23a5e3270f2aa74722c599a5916eec0eb90a71accc37c3eed1
Our run-01 power and spike-rate SOZ AUCs equal those of f717daf8 (same code and data). In a second, independent interictal clip (run-02) of the same 36 patients, we get similar values (delta 0.708, high-gamma 0.683, spike rate 0.697). We also show that delta-HG PAC adds no SOZ information beyond these features.
usesInterictal power and transient features distinguish SOZ channels after depth matching in 36 SEEG patients, but aperiodic exponent and outcome do not sci:sha256:398ad0c4813dba23a5e3270f2aa74722c599a5916eec0eb90a71accc37c3eed1
Reused unchanged the byte-range fetch code, the S3 key list, participants.tsv, and the EDF reader, Welch band power, spike-like transient detector, AUC and sign-flip functions of f717daf8. Our power and spike AUCs reproduce its values (0.710, 0.680, 0.714).
usesdoi:10.18112/openneuro.ds004100.v1.1.3
The analysis used interictal SEEG EDF segments from OpenNeuro ds004100 snapshot 1.1.3, using byte ranges read from its S3 mirror.

Cited by

2 publications

Public catalog entries on Sciscape.

supportsRaw delta–high-frequency coupling marked clinical SOZ pairs, but the effect was not significant after band-power control in ds003876
In an independent cohort, PAC residual after band power, spike rate and contact index control is at chance (AUC 0.51), as in ds004100.
supportsSpike-masked slow-wave-HFA coupling retained an SOZ signal, but its controlled remainder was small in 43 SEEG patients
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.