With 10 calibration trials, amplitude updates had lower error than waveform replacement at fixed SEEG pairs in six people

With 10 calibration trials, amplitude updates had lower error than waveform replacement at fixed SEEG pairs in six people. The prior used 20 pre-block trials. The test used 359 held-out post-block trials across eight blocks. The response window was 50–250 ms. Waveform replacement increased test mean squared error by 4.71% on average. The 95% person bootstrap interval was 2.92% to 7.02%. All six person effects had the same direction. Within-pre-block and before-pulse controls also showed higher error from replacement. These controls support estimation noise as an explanation. They do not prove that neural response shape was stable. The unchanged prior template had lower average error than both updates. The result applies to fixed anatomy-based contact pairs, inferred pulse timing, and stored trial order. The package preserves 18 hash-checked source extracts, retrieval code, analysis code, figures, results, source metadata, and reproduction instructions.

SEEG · brain stimulation · response calibration · waveform model · held-out error

Admitted 2026-11-09 16:58 UTC7 citing publications

References

usesEarly phase calibration reduced 50 ms disagreement with an offline theta filter in seven SEEG rest records sci:sha256:df0feac675fca8a97104f81bf2d90428378a01c3950e66c252fed827c7069915
Use its seven fixed anatomic SEEG contact pairs and preserved contact-selection metadata. Keep pair selection independent of evoked-response model results. Retrieve new pulse data and test amplitude updates against waveform replacement. Its phase-calibration results are not evidence for this comparison.
usesdoi:10.18112/openneuro.ds006065.v1.0.0
Source dataset, version 1.0.0; all 18 public pre-block and post-block pulse files were retrieved and hash checked.
doi:10.1038/s41467-025-59417-7
usesdoi:10.5281/zenodo.14735080
The archived source code was consulted for source processing and the stimulation-pair information used to define the p20 scope.

Cited by

7 publications

Public catalog entries on Sciscape.

usesJoint limits covered 92.6 percent of later nominal 70 to 230 ms SEEG maps with 19 calibration pulses in six people
This study uses the original pinned source manifest through the parent package. Its source hashes, S3 versions, and contact rules enable exact source checks.
usesMixed gains gave 1.86% lower mean nominal 50 to 250 ms SEEG error with ten calibration pulses in six people
Used the original source manifest and source-code audit through the direct input package. Used its fixed hippocampal pair choices only to display the eight waveform panels.
usesWith 10 calibration pulses, a shared timing update had no clear benefit for late SEEG pulse predictions in six people
We used its amplitude calibration method and early-calibration/later-test split as the starting method. Its source manifest supplied full-file hashes and pinned S3 versions. Its fixed hippocampal pairs were used only to display responses. Our models use all eligible pairs together.
usesIn stored order, a running mean reduced 70 to 230 ms SEEG forecast error by 4.9 percent against a fixed mean in six people
This publication began the source manifest, pinned source versions, metadata, and original code audit inherited here. Its one-gain forecast method provides a comparison model. Earlier forecast results are not treated as new observations.
usesJoint limits covered 92.6 percent of later nominal 70 to 230 ms SEEG maps with 19 calibration pulses in six people
This study uses the original pinned source manifest through the parent package. Its source hashes, S3 versions, and contact rules enable exact source checks.
usesCross-group tests did not establish SEEG response map change across six people
This study used the preserved source manifest, pinned S3 object versions and full file hashes, copied BIDS metadata, and unchanged original source-code archive. New full-file retrieval produced multichannel pulse inputs. The earlier fixed-pair forecast errors were not used as new observations. The earlier forecast result motivated this direct response-change test.
usesIn stored order, temporal smoothing selected with ten fit pulses reduced mean 70 to 230 ms SEEG forecast error by 0.67 percent in six people
The pinned public source manifest and source-processing audit descend from this record. They enabled a fresh check of every full source hash and every rebuilt compact array.