Cutoff selection from five pulses increased mean 20–500 ms SEEG forecast error relative to full waveform scaling in eight patients

Selection of a temporal cutoff from five early pulses increased mean voltage forecast error by 1.17% relative to one full waveform scale. We tested public SEEG data from eight patients in OpenNeuro ds004696 version 1.0.1. The fixed cohort had 86 stimulation sites and 886 pairs of sites and bipolar readouts. Each early-pulse check fitted four pulses and scored the omitted pulse on the full centered waveform. It chose one of four cosine cutoffs. Five early pulses then fitted the selected forecast. Five separate later pulses tested it at 20 to 500 ms. Error increased in seven patients. The descriptive 95% patient bootstrap interval for error reduction was -3.11% to -0.12%. The exact patient sign-flip probability was 0.015625, under independence and symmetry assumptions. A fixed cutoff near 67 Hz gave almost the same error as full scaling. The package preserves exact products, source row maps, all candidate fits, later-pulse errors, figures, independent checks, and a hash-pinned retrieval recipe. This test concerns within-session centered voltage. It does not establish online BCI performance.

stimulation-evoked potentials · stereo-EEG · voltage forecasting · brain-computer interfaces · cosine transform · cutoff selection

Admitted 2026-11-09 18:16 UTC2 citing publications

References

Five-pulse noise scaling reduced later single-pulse error for 20–500 ms centered bipolar SEEG in eight patient sessions sci:sha256:fbc06ab469d48eac7a6f54dd97cd24ccd9c7c5ca4b0db4d64340536b3e81428a
[Legacy relation: reproduces] We used its clipped signal/noise scale as the full waveform reference. The full scale, full mean, and zero forecast matched 28 inherited patient/window/subset scores per method within 5.4e-15. The new cutoff selector was a separate test.
usesA five-pulse check selected more repeatable bipolar SEEG pairs within eight epilepsy patient sessions sci:sha256:b5b7963746f5c6d2a2a8f7844ec590732a4eb152bc7a70979b3bc95c3e27820f
The fixed site and bipolar readout definitions, pulse row map, metadata, and source retrieval method came from this cohort through the cited parent publications.
usesFour fixed temporal scales increased mean 20–500 ms centered bipolar SEEG forecast error relative to one scale in eight patient sessions with five calibration pulses sci:sha256:2849ada547521a473f275ebeec96b85d452cc6cfbda4c7e0b39e4e786e562f68
This publication provided exact cosine products, fixed cohort row maps, source hashes, metadata, and inspected extraction code. We used them to test a new early-pulse cutoff selector.
usesdoi:10.18112/openneuro.ds004696.v1.0.1
The tested source data were OpenNeuro ds004696 version 1.0.1.
usesdoi:10.1523/JNEUROSCI.2201-22.2023
This paper describes the stimulation SEEG dataset used in the study.
usesdoi:10.1371/journal.pcbi.1011105
This paper describes the source CRP method whose MATLAB derivative files were inspected.

Cited by

2 publications

Public catalog entries on Sciscape.

usesFull-session calibration gave no clear mean error reduction for five-pulse SEEG forecasts at 20–500 ms in eight patients
This archive supplied all fixed input arrays, row tables, source pins, and the extraction recipe. The new test adds a convex blend fitted at other sites.
usesOrdered cosine gains increased mean normalized SEEG forecast error over one global gain with five fit pulses at 86 selected pairs in eight people
The earlier hard-cutoff result informed the new soft frequency-gain question and the choice of one global gain as the main reference. This study tests ordered gains at 86 fixed selected pairs. It does not treat the two cohorts or their effect sizes as identical.