A running maximum keeps input times within recorded limits in ten Dutch speech SEEG files
All ten public Dutch speech SEEG files contain backward time-stamp steps. The complete audit found 27,738 negative intervals across 3,059,812 samples. A fixed test used 40,795 decisions and input windows of 256 samples. Raw binary search passed the recorded decision time in 329 windows. A nominal 1024 Hz clock passed it in 2,608 windows. A running maximum kept all tested windows within the recorded limit and kept source sample order. The package provides a proof, exact time arrays, source receipts, code, results, figures, and independent checks. The bound concerns recorded time; physical acquisition time and decoder performance were not tested.
SEEG · brain-computer interfaces · time stamps · sample order · running maximum · recorded-time bound
References
4- usesAn archive preserves complete time arrays from ten public speech SEEG files sci:sha256:37ebf8bc325f256a36402388a8738f5ccb05ab1e91310393a0eb09a4d53e4bc7
This result uses the exact SEEG time arrays and source receipts to check sample order and compare input rules.
- doi:10.1038/s41597-022-01542-9
- usesdoi:10.17605/OSF.IO%2FNRGX6
Source data record for the ten Dutch speech SEEG files whose stored time arrays were audited.
- usesdoi:10.82901/nemar.nm000371
NEMAR mirror used to retrieve the source NWB metadata and time arrays.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.