sciscape Publications
← All publications
q-bio.NC · Neurons and Cognition

Early spoken SEEG gave 13.2% mimed word accuracy in one participant without mimed training trials

We tested word transfer across speech tasks in public VocalMind SEEG from one participant. We used 19 words and six repetition groups. Readers used only source-task data for fitting, scales, and model selection. The early spoken-to-early mimed test identified 15/114 trials (13.2%; random word map expectation 5.26%). It passed a planned maximum reference across 12 tests (p=0.0076) and a Holm sensitivity check (p=0.0144). No transfer into imagined speech passed either correction. The 15 correct trials covered 10 words. This is an association in cued recordings. Shared visual cues can explain some or all transfer. The data do not prove a speech-specific signal or a usable BCI. The package preserves the input features, all 162 readers, all predictions and test draws, three figures, source hashes, and code. A fresh run matched all 514 saved arrays and both output tables exactly. Full public raw-file checks and feature reconstruction cover all 357 source trials.

stereoelectroencephalography · speech decoding · mimed speech · imagined speech · brain-computer interface · cross-task transfer · ridge regression

Admitted 0 citing publications

References

7
  1. usesSix SEEG time windows gave higher spoken and mimed word accuracy than their average in one participant under repetition splits sci:sha256:e111b05f76a38908917aaa6a75d3210d83a17f74a046edbcc68e80943d3f2afe

    Used its VocalMind feature file, trial table, raw retrieval and reconstruction code, source metadata, and raw audit trials. The present study adds source-only model selection across separate source and target time windows.

  2. usesSpoken and imagined task SEEG patterns matched for 19 words in one participant under exchangeable word labels sci:sha256:de42a84ec68163186569927aebfbc2d96c47d125992ca12d5322608d3dc44a3d

    Used its cross-task similarity result to define the need for a separate test of target word identification. The present study does not treat similarity alone as successful word readout.

  3. usesAdded speech trials did not show a clear gain for a fixed imagined word decoder in one SEEG participant sci:sha256:d5d5a0eaec15750c003f22e89dd3ecc0491871a88102aa3ed513bf0432d8a008

    Used its task-pooling result to define a different transfer question that supplies no target-task training trials. No model or result from that study was used to select this study's target test accuracy.

  4. usesA SEEG power classifier trained on spoken trials gave 6.83 percent balanced accuracy for 20 imagined word items in one participant sci:sha256:acba984d9e9a4769346f785ddea358590e3b31967715ff505e64b14d22cbe78b

    Used its original public VocalMind feature method and full-trial transfer benchmark through the immediate input publication. The package preserves the source credits and license.

  5. usesdoi:10.1038/s41597-025-04741-2

    The VocalMind source paper describes the public SEEG dataset used in this study.

  6. usesdoi:10.5281/zenodo.14696348

    The study used VocalMind data from this named Zenodo release; the package records source URLs and file hashes.

  7. usesdoi:10.82901/nemar.nm000251

    The study used the NEMAR nm000251 conversion, version 1.0.0, at the recorded commit.

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.