Learning reduced mean 50 ms offline theta phase error by 11.25 degrees in spectrum controls from seven SEEG rest recordings
We tested learned theta phase forecasts for brain interfaces that can time stimulation. We used 28 selected SEEG pairs from seven public rest recordings in OpenNeuro ds006065 v1.0.0. Both methods used early fitting and later testing. We made 199 Fourier phase controls per person. Each control kept the power and cross spectra. Both methods were fitted again for each control. Learning reduced mean 50 ms offline phase error by 11.77 degrees in real voltage and 11.25 degrees in controls. The extra real gain was 0.52 degrees. Its person-bootstrap interval was -0.29 to 1.28 degrees; the exact person sign-flip p-value was 0.25. A separate conditional rank for these fixed records gave p = 0.02 under exchangeable Fourier phases. The shorter-block check gave a similar person effect. Thus, a large gain over a sinusoid can remain after Fourier phase changes. This does not establish model equality or a neural mechanism. The target used future voltage. No live device or stimulation outcome was tested. The package preserves exact inputs, source records, code, models, checks, tables, and three figures.
theta phase · stereoelectroencephalography · SEEG · Fourier phase controls · phase forecasting · brain interfaces · signal processing
References
6- usesEarly linear fitting reduced mean 50 ms offline theta phase error from 69.52 to 57.75 degrees in seven SEEG rest recordings sci:sha256:6e0d73f3cfa87fcbac3f6dd9697937e67dc539c31c7ddffb19140640d26f1405
This study used its exact seven voltage inputs, source records, retrieval code, and early learned phase method. The real-record mean errors reproduce its 50 ms, two-second-reference results. We added independent control fitting and tested the real learning gain against the control learning gain.
- usesSpectrum controls had a mean 50 ms offline phase error of 68.59 degrees in seven SEEG rest recordings with fixed early readouts and frequencies sci:sha256:4c870610d86ccd8aec8e65f86e3418d58b3d158694b2c5dc16a3b96d8e15ba64
This study used its question and shared Fourier phase control construction. That earlier study tested a fixed sinusoid method. The new study fits the sinusoid and learned method again for every control, then compares the gains from learning.
- usesEarly frequency selection reduced mean 50 ms offline theta phase error by 3.55 degrees in seven SEEG rest recordings sci:sha256:5c3ae075960de0774e0fcc176ee07a5e3c9a353d86238f1e49023a23aa4d2bfa
This was the original source of the selected bipolar voltage, source manifests, and source retrieval programs reused through publication 72953d07-e966-47a3-b1b0-8cf4686946ed. Its selection and public-source snapshot remain fixed.
- usesdoi:10.18112/openneuro.ds006065.v1.0.0
This is the recorded version-specific DOI for the OpenNeuro dataset used for the seven rest recordings.
- doi:10.1038/s41467-025-59417-7
- supportsdoi:10.1103/PhysRevLett.73.951
This paper describes shared Fourier phase changes that preserve linear relations among simultaneously measured signals, the control construction used here.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesVoltage statistics gave a mean absolute theta phase error of 55.24 degrees at a 50 ms forecast delay against a centered offline reference in seven SEEG rest recordings
The prior finding of similar phase-reader gains in spectrum controls motivated a direct test of calibration from voltage lag statistics. It does not prove that all physiological information is absent.
- usesRefitted spectrum controls gave 20.82 percent forecast skill at 50 ms for a future theta filter term in seven SEEG rest records
We use its shared Fourier-phase control and model refit method as context. The present targets exclude all pre-cutoff voltage. We use new later test blocks and a shorter-block check.
- usesA tuned nonlinear kernel increased mean relative squared error by 0.18% for future-only theta targets in seven SEEG rest records
Uses its finding that spectrum controls explain much of learned phase performance to motivate full-pipeline Fourier phase controls for a nonlinear future-only forecast comparison.