Both donor rules chose a penalty of 1000 for nine SEEG sound readers with two calibration passages and a 400 ms delay
We tested two ways to choose a SEEG reader penalty from other people. Each of nine recipients used only two local calibration passages. We excluded the recipient and every donor passage with an exact recipient test sound ID. Both donor rules chose penalty 1000 for all recipients. Their predictions matched the fixed-1000 control exactly. On 100 later passage records, the mean relative error decrease from local two-passage tuning was 2.65%. Four people improved, two worsened, and three had identical predictions. Descriptive ranges over people and shared sounds included zero. This reused clinical sample does not establish a new-person gain or a benefit over fixed 1000. The SCI package retains choices, models, predictions, source checks, figures, and replay code, but omits source-derived inputs because the saved dataset terms conflict. Reproduction requires upstream data under its terms. The scope is heard digital sound amplitude with a 400 ms algorithm delay. No live interface or stimulation was tested.
stereo-electroencephalography · brain-computer interface · ridge regression · cross-validation · speech envelope · calibration · signal processing
References
4- usesSix calibration passages gave a mean 6.5% lower later squared error than two in nine SEEG sound readers at a 400 ms delay sci:sha256:f5145388b8b3b2413f77551421cbd7d8b1f82e06b2852f358c8cd5d4eb6a0e6a
We used the compact public inputs, passage choices, and exact reader functions. The local two- and six-passage predictions reproduced this source result. We added person and exact sound exclusions to test donor penalty selection.
- usesFixed SEEG readers had higher mean sound amplitude correlation at 400 ms than at zero delay on 100 unused later passage records in nine people sci:sha256:3c8b962f91454c5d1b32469fb10a6ecb9534841b65f6b7ab946bc56a062e2126
The inherited later-test record rule and corrected clock mapping define the calibration and test boundary. We preserved this source chain and used its later records without new person exclusions.
- usesdoi:10.18112/openneuro.ds004703.v1.1.0
The study used SEEG recordings, heard sounds, and task records from OpenNeuro ds004703 version 1.1.0. Source-derived inputs are omitted from this SCI package because the saved source README and metadata state conflicting terms.
- doi:10.1038/s41467-024-44844-9
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.