An early linear model gave 5.20 degrees less error than a calibrated oscillator for a future filter angle at a 50 ms delay in seven SEEG rest records
We tested a future filter angle in seven public SEEG rest records. A complex sum used voltage from a time 50 ms after the input cutoff. A fixed early mean was removed first. The learned model used 500 ms of past voltage. Early data set all parameters. Later data supplied the test. The learned model gave 77.09 degrees mean angle error. A calibrated oscillator gave 82.30 degrees. Error decreased in all seven people. The mean decrease was 5.20 degrees. A descriptive person bootstrap interval was 2.67 to 8.41 degrees. The constant direction gave 84.48 degrees. Mean future voltage MSE skill was 6.94 percent at 50 ms and 0.92 percent at 200 ms. Independent Gaussian noise gave 90.12 degrees error for the future angle. The same noise gave 68.48 degrees for the centered reference. These targets measure different quantities. The future filter has a broad frequency response. Its angle is not a validated biological theta phase target. The package preserves selected data, source records, models, forecasts, code, and figures. All 105 source ranges matched their hashes. Every derived voltage sample matched the bundled input. A second analysis gave identical main output hashes. This is a reused cohort and test interval. The study does not measure stimulation benefit or live timing.
stereoelectroencephalography · SEEG · phase forecasting · future filter angle · ridge regression · brain interfaces · independent noise control
References
5- usesA fixed learned model raised the share of commands within 45 degrees of target phase from 39.79 to 51.28 percent in offline 50 ms SEEG replay of seven rest records sci:sha256:f95e54eb370bc2e0e49af090e06dc6d39029b2a4542255ccea2520190080d0a9
We used its seven selected voltage files, source metadata, and range hashes. We used its input and oscillator designs. We fitted new models for the future target and all comparisons.
- usesA 500 ms past voltage sum gave a 63.26 degree error for centered theta phase 50 ms ahead in seven SEEG rest records sci:sha256:678ac0e1b4e8d00b5e72eb22a8e9bbbfc9c4bd67c224a77925cb4cba677fc26f
We used its centered-label support split and independent noise control as the method basis. We moved the first future target sample from the input cutoff to delivery. We fitted angle models for this changed target.
- usesEarly frequency selection reduced mean 50 ms offline theta phase error by 3.55 degrees in seven SEEG rest recordings sci:sha256:5c3ae075960de0774e0fcc176ee07a5e3c9a353d86238f1e49023a23aa4d2bfa
We retained its early pair selection and source snapshot through the direct parent. We preserved the selection function, fixed pairs, candidate ranks, and source records.
- usesdoi:10.18112/openneuro.ds006065.v1.0.0
The analysis used seven selected voltage products from OpenNeuro ds006065 version 1.0.0; the pinned source commit, range hashes, and derived voltage checks are preserved.
- doi:10.1038/s41467-025-59417-7
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesPhase and future signal scores chose different SEEG pairs in 17 of 36 patients with a 50 ms phase target
Used its future filter target distinction to separate future recorded signal prediction from the centered phase reference.