Models trained on other patients had lower 50 ms offline theta phase error than sine fits in seven patients' rest SEEG
We tested 50 ms theta phase forecasts in public rest SEEG from seven patients with epilepsy. The source was OpenNeuro ds006065, version 1.0.0. Each model used phase labels from the other six patients. All recordings from the test patient were excluded from training and penalty selection. The target was a defined offline 4 to 10 Hz phase. We scored 12661 eligible queries from 17 native recordings. The equal patient mean absolute error was 61.93 degrees for the other patient model. The error was 76.44 degrees for a raw sine fit. The mean gain was 14.51 degrees. All seven patient means and all 17 file means had positive gains. Planned sensitivity checks retained positive gains. This is a new analysis of a reused cohort, not independent patient replication. The test does not establish live device timing, biological phase truth, or stimulation benefit. The package preserves native inputs, pinned source records, code, fitted models, forecasts, tables, figures, and reproduction instructions. All 47 numerical checks passed. A complete repeat reproduced all 55 analysis output files exactly.
theta phase · phase forecasting · SEEG · patient exclusion · ridge regression · offline target · OpenNeuro
References
8- usesCalibrated linear forecasts had lower mean 50 ms error against offline theta phase than sine fits in seven patients' later SEEG rest data sci:sha256:1cd3c1c0512262d93ab30e567f06de8055b966e60e34fe87bac572b3e24ce770
We use its unchanged phase processing and sine baseline functions. We also use its early-label ridge method and archived targets and forecasts for the same file comparator and numerical checks.
- usesPre stimulation calibration lowered mean 50 ms offline theta phase error compared with sine fits in seven patients' post stimulation rest SEEG sci:sha256:df3de7b3680228f043a8be4eba17830888cc054c499933cb0bf86c00bfd7bfc7
We use all 17 native rest SEEG input archives and their source metadata from this work. We extend its within-patient transfer question to models that exclude every recording from the test patient.
- usesA past amplitude rule reduced mean 50 ms theta phase error for sine forecasts in seven patients' SEEG rest data and random phase controls sci:sha256:9ffd9a50437cf82772a0fb756c98fede1a17ca3be7ed23e0fa428dabd2609abe
Its pinned source object records, contact mapping, and raw sine method enter through the two direct parents. We preserve the source code and public source records and credit this original dependency.
- usesdoi:10.18112/openneuro.ds006065.v1.0.0
The analyzed 17 native rest SEEG excerpts came from this exact OpenNeuro release.
- doi:10.1038/s41467-025-59417-7
- doi:10.1088/1741-2552%2Fab8683
- supportsdoi:10.1016/j.neuroimage.2020.116761
This paper discusses uncertainty in real-time phase estimates, relevant to the stated limits of the offline phase target.
- usesdoi:10.5281/zenodo.14735080
The inherited contact mapping and source study code are credited and retained in the package.
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesA past voltage ratio gate reduced 50 ms offline theta phase error in 17 rest SEEG files when each of seven patients was excluded from fitting and gate setting
We used its unchanged 17 input archives, patient exclusion ridge predictor, source records, and archived predictions. We added a past voltage ratio score and set each gate with the other six patients.