One early voltage scalar increased prediction error by a mean of 0.31% at 3–15 ms in 20 SEEG trains after 24 fit pulses

We tested whether one early voltage scalar improves a later voltage forecast during SEEG stimulation. The source was OpenNeuro ds005169 version 1.0.0. We fitted on 24 pulses and tested on the next 24 pulses in 20 first trains. The input used 16 channels. Outputs used other electrode shafts. The target was centered voltage at 3–15 ms. The reference used two pulse class means. The added scalar used a fixed slope factor of 0.5. Its mean error reduction was -0.31%, with a 95% record bootstrap interval from -0.72% to +0.11%. Seven records had lower error. Correct pulse pairing beat shifted pairing, but a pairing benefit also occurred before stimulation. This control did not establish a benefit over the class reference. All 20 raw EEG files passed source hash and array rebuild checks. A separate run reproduced all analysis files exactly. The package includes data, source metadata, code, results, three figures, and instructions. This is a reused cohort. The result concerns recorded voltage, not identified neural activity or an online controller.

stereoelectroencephalography · intracranial EEG · electrical stimulation · voltage prediction · forecasting · brain-computer interfaces

Admitted 2026-11-09 17:54 UTC3 citing publications

References

usesA linear pulse trend did not establish higher test SEEG voltage reduction at 3 to 15 ms in 18 records with 16 fit pulses sci:sha256:2f9ba734a570ec9ae4b5fa5b12d74a5f7507ee84d45e0763619d73902ca8fd70
Its pulse-trend analysis helped motivate the linear trend comparator. The present comparator uses first trains, 24 fit pulses, and the next 24 test pulses. No code or result arrays were used.
usesA scale from 24 early pulses reduced later centered SEEG power more than full subtraction at 3 to 15 ms in 20 stimulation trains sci:sha256:b9ad60de31173cc279fcc5e26c00fdec2c8b47de8b2e92e053c97f44012f15ff
Its fixed scale for pulse class subtraction helped motivate a test of an early voltage scalar from each pulse. No code, scale estimates, or result arrays were used. The present result tests voltage prediction rather than subtraction power.
usesRepeatable pulse class differences exceeded baseline at 3 to 15 ms after detected edges in 17 of 20 SEEG recordings sci:sha256:879b5223cbf14d14e8e6ed3642b3968e076bf8d799c75c3c86fea26e328c4ab9
Uses its 20 compact first-train voltage records, source metadata and hashes, and unchanged extraction code. The new analysis predicts later voltage on other electrode shafts from an early scalar. The same cohort is reused and is not independent validation.
usesdoi:10.18112/openneuro.ds005169.v1.0.0
Source dataset used for the 20 selected first-session SEEG records; exact version 1.0.0 is identified in saved dataset metadata.
doi:10.1002/hbm.25749

Cited by

3 publications

Public catalog entries on Sciscape.

usesA rolling 12-pair mean gave higher contrast prediction error than a fixed 12-pair mean at 3 to 15 ms in 14 of 20 early SEEG trains
We used its unchanged compact SEEG epochs and fixed source manifest. These files supplied the same 20 records for the equal-history forecast test.
usesRecent pulse templates reduced error by a mean of 2.51% relative to initial templates at 3 to 15 ms in 20 SEEG records
This publication supplied the inspected raw source pins, 120 sidecar files, and extraction code. We adapted its header parsing and median edge detector. We extracted new late pulse blocks from all 20 hash-checked raw EEG records. No saved voltage array or error score from that publication was used.
usesA fixed map gave 8.00% mean error reduction for simulated missing centered SEEG voltage at 3–15 ms in 20 calibrated stimulation trains
Its early-scalar forecast test motivated the simultaneous spatial repair question. We reused its fixed first-16 alphabetic input rule and separate-shaft output design. We copied no fitted model, score table, or code.