Spoken and imagined task SEEG patterns matched for 19 words in one participant under exchangeable word labels
We tested whether spoken SEEG word patterns also appear in mimed and imagined tasks. We used public VocalMind data from one participant. We kept 19 words with six trials in each mode. The test used 342 trials, 110 contacts, and 1,980 log power features. It used pairs with different repetition numbers. The word mapping tests gave Holm adjusted p values of 0.0004 for spoken–mimed patterns and 0.0195 for spoken–imagined patterns. The imagined within-mode test did not establish reliable patterns across repetitions (p = 0.1323). The 95% word item interval for the spoken–imagined product includes zero. Thus, the result shows a weak match in the tested word set. It does not establish a useful imagined speech decoder. All modes used word cues. Visual cues, word order, and artifacts can explain a match. A signal interpretation requires independent repetition noise. Permutation inference requires exchangeable word labels. We provide figures, code, features, source metadata, raw hashes, and retrieval instructions. All 357 trial features were rebuilt from public raw files with zero difference. The complete reproduction script passed in a separate folder.
stereotactic EEG · imagined speech · speech brain-computer interfaces · word pattern alignment · permutation test · VocalMind
References
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
We used its VocalMind feature array, trial metadata, raw retrieval manifest, and source readers. We rebuilt all trial features from the pinned public NEMAR raw files. The new analysis tests word pattern products across different repetition numbers. It measures alignment rather than classifier accuracy.
- usesdoi:10.5281/zenodo.14696348
Source VocalMind recordings were used through the pinned NEMAR conversion; this is the original public record.
- usesdoi:10.82901/nemar.nm000251
The analyzed source files came from NEMAR nm000251, version 1.0.0, pinned to commit b4ea3a6ce1360e8990ecf1243e522ceb2bcfef88.
- supportsdoi:10.1038/s41597-025-04741-2
This data paper documents the VocalMind protocol and task design described here.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesEarly spoken SEEG gave 13.2% mimed word accuracy in one participant without mimed training trials
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.
- usesContact mean removal retained spoken and mimed word label association for a fixed classifier in one SEEG participant under exchangeable labels
Its contact pattern analysis helped define the question. This study tests word classification and gain invariance in the same public participant sample.
- usesSix SEEG time windows gave higher spoken and mimed word accuracy than their average in one participant under repetition splits
We used its public VocalMind feature arrays, trial labels, raw audit arrays, pinned source metadata, and full raw validation record. The new analysis compares separate time windows with their average under nested repetition splits. We checked three public raw trials again in this run.