Offline phase error limits set before seizures had lower mean coverage during seizures for a fixed reader in 30 SEEG pairs
This study tested whether phase error limits set before seizures kept their coverage during seizures. Data came from 30 public SEEG records in OpenNeuro ds004100, version 1.1.3. We used one fixed bipolar pair per patient outside shafts with clinical onset labels. A fixed ridge reader used past voltage to forecast a nominal 4 to 8 Hz phase 50 ms ahead. The offline phase reference used a centered filter and future voltage. Equal-patient mean coverage fell from 90.19% before seizures to 87.95% during seizures. The mean change was -2.24 percentage points. The descriptive 95% patient bootstrap interval was -3.45 to -1.10 points. Coverage fell in 23 of 30 patients. Mean absolute phase error rose from 56.63 to 61.68 degrees. The mean calibration error radius was wide, at 131.54 degrees. A fixed 6 Hz comparator retained about 91% coverage with mean error near 90 degrees and wider limits. Thus, coverage must be inspected with limit width and phase accuracy. The finding applies to this reader, one selected pair per patient, and the first ictal record. Seizure state and elapsed time cannot be separated here. No stimulation was delivered. The package preserves exact digital samples, source hashes, frozen metadata, code, models, forecasts, patient results, three figures, and repeat instructions.
intracranial EEG · phase forecasting · epilepsy · seizure state · calibration · coverage · brain-computer interface · signal processing
References
5- usesComplex voltage targets reduced mean 50 ms offline theta phase error across files from 36 selected SEEG patients sci:sha256:7bb26e9262cd7cf318aacf4eb9c26e67d5a3c1bf62ddd55c19ff2723bdfab146
The complex voltage target and nominal 50 ms forecast horizon came from this earlier phase forecast study. This study uses different inputs, fitting intervals, and filter settings. Its phase scores are not a direct performance baseline here.
- usesPower-ratio deviations exceed early limits more often during seizures outside onset-labeled contact groups in 30 SEEG records sci:sha256:46ab84bca2ba1c6a358b6eaf9312db59b5ed68f3f01bca4eb378539fd4ff1a65
We used the fixed 30-patient cohort, exact selected digital samples, frozen source metadata, source file hashes, and EDF reader. We added a test of phase error limits across seizure states. The patient cohort is reused.
- usesdoi:10.18112/openneuro.ds004100.v1.1.3
Version 1.1.3 of the HUP iEEG Epilepsy Dataset supplied the source records and labels used in this study.
- usesdoi:10.21105/joss.01896
The source dataset metadata states that MNE-BIDS was used to convert the dataset to BIDS and cites this software paper.
- usesdoi:10.1038/s41597-019-0105-7
The source dataset metadata cites the iEEG-BIDS specification used for the intracranial EEG data format.
Cited by
4 publicationsPublic catalog entries on Sciscape.
- usesDelayed rolling error limits reduced mean theta phase arc scores during seizures in 30 public SEEG pairs
We used its 30 fixed SEEG pairs, exact digital signals, frozen metadata, fitted theta phase readers, and reference forecasts. All 510 model and forecast arrays matched a fresh rebuild. We added a continuous replay and delayed rolling error limits. The patient cohort and readers are reused.
- usesDelayed rolling error limits reduced mean theta phase arc scores during seizures in 30 public SEEG pairs
We used its 30 fixed SEEG pairs, exact digital signals, frozen metadata, fitted theta phase readers, and reference forecasts. All 510 model and forecast arrays matched a fresh rebuild. We added a continuous replay and delayed rolling error limits. The patient cohort and readers are reused.
- usesVoltage normalization reduced fixed gate retention during seizures in 30 SEEG pairs with one offline phase reference
This study uses the fixed 30-person data, contact rule, source manifests, and phase reader. It reproduces the archived reader outputs.
- usesVoltage normalization reduced fixed gate retention during seizures in 30 SEEG pairs with one offline phase reference
This study uses the fixed 30-person data, contact rule, source manifests, and phase reader. It reproduces the archived reader outputs.