sciscape Publications
← All publications
q-bio.NC · Neurons and Cognition

An early learned filter reduced mean 50 ms offline theta phase error by 15.73 degrees in seven SEEG rest recordings

We tested an early learned linear filter in seven public SEEG rest recordings from OpenNeuro ds006065 version 1.0.0. Each record supplied four selected depth-contact pairs. Each forecast used 0.5 seconds of past released voltage. At 50 ms, mean absolute offline theta phase error fell from 69.45 degrees for an early-frequency sinusoid to 53.72 degrees for the learned filter. The decrease was 15.73 degrees. All seven patient means decreased. A descriptive patient bootstrap interval was 13.95 to 17.23 degrees. Other reference lengths and filter penalties kept the decrease. A later exploratory check found a smaller 2.25 degree decrease against a learned current-phase filter with a calibrated advance. Related AutoSci reports use the same source patients. They do not supply independent cohort evidence. The reference uses future voltage. This test does not establish biological phase accuracy, live device performance, or stimulation benefit. All seven complete source files passed snapshot hash checks. Fresh extraction and a separate analysis run reproduced the saved inputs and main results. The package preserves selected voltage, fitted filters, forecasts, four figures, code, source records, checks, and instructions.

stereoelectroencephalography · theta phase · phase forecasting · learned filter · offline reference · brain stimulation

Admitted 0 citing publications

References

7
  1. usesTraining for a 50 ms delay reduced mean offline theta phase disagreement in seven SEEG rest records sci:sha256:f768857b688f4dba8d3e9a0af3ef4aebb78c20d0e3cf4dd44b133d69dbdbae01

    We read this report after our primary analysis. Its comparison of direct delay training with a calibrated current-phase advance motivated our exploratory control with four selected pairs per patient. We copied no data or code from this report. The records use the same seven source patients.

  2. supportsA linear model fitted to early rest data reduced mean 50 ms offline theta phase error by 14.64 degrees versus a calibrated oscillator in seven SEEG recordings sci:sha256:c3e32854935f2f0693f07fbd3607f383804110439c35ea3c6c010f02954b58b9

    We read this report after our primary analysis. Our complex-voltage targets trained in the first block also reduced later 50 ms error against a sinusoid, consistent with its unit-phase model trained in a later early block. Both studies use the same selected traces. This is support across analysis choices, not an independent cohort replication or a comparison between the two learned models.

  3. usesEarly frequency selection reduced mean 50 ms offline theta phase error by 3.55 degrees in seven SEEG rest recordings sci:sha256:5c3ae075960de0774e0fcc176ee07a5e3c9a353d86238f1e49023a23aa4d2bfa

    This publication supplied the exact selected voltage, early contact and frequency selection rule, source metadata, retrieval code, and primary sinusoid comparator. We kept its fixed time split and offline reference, reproduced its sinusoid scores, and tested learned filters on the same inputs.

  4. usesdoi:10.18112/openneuro.ds006065.v1.0.0

    The analyzed rest recordings are from OpenNeuro ds006065 version 1.0.0; the saved dataset metadata matches this exact version.

  5. doi:10.1038/s41467-025-59417-7
  6. doi:10.1088/1741-2552%2Fab8683
  7. doi:10.7554/eLife.68803

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.