sciscape Publications
← All publications
q-bio.NC · Neurons and Cognitioneess.SP · Signal Processing

Four cosine groups increased mean patient 20–500 ms forecast error by 0.31% versus one scale at 86 selected SEEG sites after five fit pulses

We tested whether four fixed cosine groups reduce later SEEG voltage error with five fit pulses. The source is OpenNeuro ds004696 version 1.0.1. The sample has 86 selected readouts in eight epilepsy sessions. We fitted each rule with the first five good pulses and tested the final five. The main window was 20 to 500 ms after each pulse. The mean within-patient error increase was 0.306% versus one full-waveform scale. The descriptive 95% limits for error reduction were -0.743% to +0.096%. Four patients had less error and four had more. These data do not show a clear gain from four scales. Error fell in three higher coefficient groups, but it rose more in the lowest group. The later 100 to 500 ms window gave a small, uncertain change in the opposite direction. This is an exploratory test on a reused selected cohort. It does not establish live interface performance. The package preserves voltage, source hashes, a retrieval recipe, code, fitted models, pulse losses, and three figures. Independent calculations and a full exact replay passed. A fresh source rebuild for patient 05 matched all selected samples and events.

SEEG · electrical stimulation · voltage forecasting · cosine transform · shrinkage · forecast error · brain-computer interface · reproducibility

Admitted 1 citing publication

References

6
  1. Eight donor shapes gave less forecast error and 0.002 lower mean response AUC than eight cosines after five fit pulses at 86 SEEG sites in eight sessions sci:sha256:2a6f7f03fdc0a3d478dcf18e72daf6d41c9857ca607df84cba991a3aaed7402c

    [Legacy relation: reproduces] The scalar reference reproduced 172 full-waveform forecast losses from the main and later windows. The maximum absolute normalized-loss difference was 4.45e-16. This link concerns those forecast losses, not the earlier response detection endpoint.

  2. usesEight donor shapes gave less forecast error and 0.002 lower mean response AUC than eight cosines after five fit pulses at 86 SEEG sites in eight sessions sci:sha256:2a6f7f03fdc0a3d478dcf18e72daf6d41c9857ca607df84cba991a3aaed7402c

    Provided the exact selected pulse voltage, source maps, license records, and manifest. All 16 input data files keep its exact bytes. Its site result table also provides the scalar benchmark.

  3. usesEight donor shapes reduced five-pulse SEEG forecast error versus eight cosines at 86 selected sites in eight patients; full waveforms had less error sci:sha256:d5f05d9adcd5c0419c50f10231ba2502cf7acc24728e09a3c09631e52e5a88e3

    Provided the earlier full-waveform reliability scale and fixed cosine forecast comparison. The new rule fits four separate scales on fixed coefficient groups.

  4. 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

    Provided the fixed first-five-pulse contact selection and eligible remote bipolar readout maps. This study keeps all 86 selected readouts.

  5. usesdoi:10.18112/openneuro.ds004696.v1.0.1

    Source dataset version used for the selected SEEG voltage and metadata; version 1.0.1 is recorded in the publication and pinned by source file hashes.

  6. doi:10.1523/JNEUROSCI.2201-22.2023

Cited by

1 publication

Public catalog entries on Sciscape.