In HUP SEEG, envelope synchronization rose at seizure onset, but the primary end-versus-start increase was not significant
We measured cross-channel co-fluctuation of 1–45 Hz line-length envelopes in 10-s epochs around annotated seizures in HUP SEEG. The analysis used archived envelopes for 116 seizures from 35 patients. The pre-registered primary contrast, late seizure minus early seizure, showed a small, non-significant increase in normalized leading-eigenvalue synchronization (S1 = +0.037, 95% CI [−0.005, +0.077], p = 0.089; 23/35 patients positive); the primary hypothesis was not supported. In contrast to the prediction of onset decorrelation, S1 rose from pre-seizure to early seizure (+0.051, Holm p = 0.002). The end-versus-start increase was absent when the late epoch ended 5 s before annotated offset and when only each patient’s first seizure was analyzed, and was significant among seizures lasting at least 30 s. Envelope amplitude also rose; adjustment reduced the H1 estimate to +0.026 (95% CI [−0.012, +0.064], p = 0.19). These findings concern slow line-length envelopes, not raw-signal zero-lag correlation. The 20-bin epochs, clinical event annotations, single-centre drug-resistant epilepsy sample, and incomplete run coverage limit interpretation.
epilepsy · intracranial EEG · SEEG · seizure · synchronization · line length · envelope correlation · HUP · OpenNeuro
References
2- usesSEEG seizure end times were less spread than start times, but sharper end crossings may explain part of the difference sci:sha256:15be8d43810ebd9584c8ced26dd440744daa9ea319a72ba90194ac2a6cae2eb4
We used its archived per-run 1-45 Hz line-length envelope files (results/features, 132 runs) as the input data. Input SHA-256 values are in provenance/input_features_sha256.tsv.
- usesdoi:10.18112/openneuro.ds004100.v1.1.3
The analyzed seizures were from OpenNeuro ds004100 (HUP iEEG), version 1.1.3; the archived per-run envelopes and their hashes are documented in provenance.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.