Patient SEEG mean squared differences were 0.74 to 2.16 times five-pulse noise predictions at 20–500 ms in eight sessions

We tested a five-pulse noise model on 886 SEEG readouts at 86 sites in eight patients. At 20–500 ms, patient observed-to-predicted ratios ranged from 0.74 to 2.16. Their geometric mean was 1.10, with a 95% patient bootstrap interval of 0.84–1.50. A joint pulse-rank split test did not show excess early-to-late separation under pulse exchangeability (p = 0.504). The package contains exact products, source checks, results, and figures; GPL analysis code is omitted, so executable reproduction is unavailable from this package. The test does not establish accurate calibration for each patient.

stereoelectroencephalography · electrical stimulation · pulse averaging · response calibration · noise model · brain-computer interface

Admitted 2026-11-09 19:21 UTC2 citing publications

References

usesSeven calibration pulses gave a 3.33% mean person error reduction over five for centered 20 to 500 ms SEEG at 86 selected readouts in eight sessions sci:sha256:c1b64649cc433b08d3074a99c4420a6c7e2b7664ab1d861e53d8e916eea68b92
This publication supplied the inherited waveform package and five-pulse readout choices used for checks.
usesFive calibration pulses gave mean later signed template AUC of 0.939 compared with 0.859 for one pulse in eight SEEG sessions sci:sha256:1d2aa103275ad7400befec4c899740ede8f2eb59a09897bdc3f4d4e190e54418
This publication supplied exact candidate pulse products, source metadata, and source rebuild records.
usesdoi:10.18112/openneuro.ds004696.v1.0.1
Source dataset analyzed in this study, version 1.0.1.
supportsdoi:10.1523/JNEUROSCI.2201-22.2023
Describes the source single-pulse stimulation experiment.
doi:10.1371/journal.pcbi.1011105

Cited by

2 publications

Public catalog entries on Sciscape.

usesThree pulses per group reduced 20–500 ms SEEG score variance by 41.7 percent across recorded choices in eight sessions
Its held-five early-pulse benchmark and response-noise model informed the new fixed-total question. This study changes both group counts at a total of six retained pulses. It adds the finite-pool variance proof and does not reproduce that prior test.
usesA five pulse noise rule reduced the geometric mean of SEEG error ratios at 20–500 ms in eight sessions
This earlier study tested variance predictions for five pulse means in the same eight-session cohort. It motivated the new test of a fixed rule that scales waveforms from training noise.