Voltage 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
We used public SEEG data from seven people. A linear reader forecast theta phase 50 ms ahead from the last 500 ms of voltage. It used the mean and autocovariance of 56 s of early voltage for calibration. Current calibration used no phase or waveform labels. We scored 94,080 later forecasts against a centered 2 s complex filter reference. Mean absolute phase error was 55.24 degrees. A ridge fit with phase labels every 100 ms gave 57.25 degrees. The paired mean difference was 2.01 degrees, with a person bootstrap 95 percent interval of 1.61 to 2.45 degrees. A control added after the main result used phase labels every 10 ms. It gave 54.42 degrees. Label spacing thus changed the order of the methods. Voltage statistics can calibrate this reader for a stated offline phase reference. The results do not show a general accuracy gain over fits with phase labels. This was an offline rest test in seven reused recordings. It did not test stimulation or transfer to another session. The package contains voltage, model coefficients, forecasts, figures, code, source checks, and reproduction instructions.
theta phase · stereoelectroencephalography · SEEG · autocovariance · phase forecasting · neural signal processing · brain interface
References
4- usesLearning reduced mean 50 ms offline theta phase error by 11.25 degrees in spectrum controls from seven SEEG rest recordings sci:sha256:be46eec0bf0062fd4ea3fded570f754676cc2c3a9e4df6be657676ba2ae8d0c5
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.
- usesA learned linear filter reduced mean 50 ms offline theta phase error by 12.54 degrees in seven SEEG rest recordings sci:sha256:681793db790925fefcdd246f20a125cf7ec0a4d0c7ac415b89cd84c8e64b5d35
This work reuses the exact selected voltage arrays, contact choices, original phase and sinusoid readers, and source retrieval code. It adds a reader calibrated from voltage mean and autocovariance.
- usesdoi:10.18112/openneuro.ds006065.v1.0.0
This analysis uses the seven selected SEEG recordings from OpenNeuro ds006065 version 1.0.0. The source dataset metadata records this version-specific DOI and CC0 license.
- doi:10.1038/s41467-025-59417-7
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesAn early amplitude gate increased prediction skill for the future part of a centered theta label at 50 ms in seven SEEG rest records
We used its seven 590 s voltage arrays, metadata, pinned source versions, and full retrieval and extraction code. Fresh downloads and extraction verified the inputs.
- usesNominal 90 percent Gaussian voltage arcs gave a full circle in 95.20 percent of 50 ms theta forecasts against an offline reference in seven SEEG rest recordings
We used its selected public SEEG voltage, fixed covariance reader, offline theta reference, and source retrieval code. We added and tested full residual covariance and phase arcs.
- usesMidpoint targets reduced the larger error across two theta phase references for 50 ms forecasts in seven SEEG rest records
Used its half second voltage history design and its dense phase-label control to set 10 ms training steps. This study compares single-reference and two-reference targets with matched label spacing.