Separate delay fits reduced mean offline theta phase error by 1.45 degrees at five known delays in seven SEEG rest records after stimulation
A phase reader must account for a known delay when it selects a pulse time. We compared one 50 ms fit with learned phase rotations and six separate delay fits. We fit on public SEEG rest voltage before stimulation. We scored 28 fixed contact pairs in seven rest records after stimulation. The primary score used five other delays: 0, 20, 80, 100, and 150 ms. Mean absolute error was 61.92 degrees for rotation and 60.47 degrees for separate fits. The paired difference was 1.45 degrees, with a descriptive person bootstrap interval of 0.96 to 1.95 degrees. All seven person differences were positive. Rotation used 127 numeric constants per pair, compared with 432 for the bank. It changed 16.96 percent of individual forecasts by more than 45 degrees. The phase target used a centered two-second reference and future voltage. This reused cohort gives an exploratory storage and accuracy result. No device timing or stimulation effect was tested. The package includes voltage, source metadata, code, models, forecasts, complete scores, three figures, and independent checks.
theta phase · phase forecasting · stereoelectroencephalography · SEEG · known delay · ridge regression · brain stimulation · signal processing · offline phase reference
References
5- usesFive recent voltage means reduced mean 50 ms offline theta phase error by 6.73 degrees after causal input filtering in seven SEEG rest records sci:sha256:7759bb4d76a110b5edc5d4732f6405386081aa44a0cecc61abb5248c0f3fd736
This publication supplied the hybrid40 causal input rule, centered phase reference, and ridge fit rule. We adapted those rules to compare delay rotation with separate delay fits on later post-rest voltage.
- usesUpdates using 56 seconds of past labels gave no clear gain in 50 ms offline theta phase forecasts in seven SEEG rest records after stimulation sci:sha256:1e483ef450e6774308a365f4037aa19fae430da6b3acd25dc843b9aa07025344
This publication supplied all 14 compact pre- and post-rest voltage files, the fixed pair list, source metadata, and retrieval code. We preserved the inputs and checked fresh source ranges across all records. We also checked five full source hashes and voltage arrays.
- usesTwo shared values added 0.50 degrees of mean offline theta phase error across 0 to 100 ms forecasts in seven SEEG rest records sci:sha256:b93d00e9cd43be0337e4db12c261cb8411900b98124474e7c7cb19b30211b06f
This study of rank-two delay-bank compression motivated our model storage question. We tested a different rule: one 50 ms hybrid reader and constant rotations. We used post-rest transfer and delays through 150 ms. We did not reuse its fitted models or claim to repeat its test.
- usesdoi:10.18112/openneuro.ds006065.v1.0.0
The analysis uses the seven participants’ pre- and post-rest SEEG records from OpenNeuro ds006065 version 1.0.0.
- doi:10.1038/s41467-025-59417-7
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.