sciscape Publications
← All publications
eess.SP · Signal Processing

Shared contact cells certified 58.6 percent of later SEEG decisions for constant contact gains within five percent in eight recordings

We tested a continuous gain certificate for eight fixed public SEEG readers and 1348 later windows. Sixteen cells per contact certified 58.61 percent as an equal-person mean, compared with 39.35 percent for the earlier whole-box bound. The same reader and gains in [0.95,1.05] were used. Minimum-sum messages on the contact forest enforce one gain cell at every use of a contact. The matched separate-pair cell control certified 56.68 percent. Outward arithmetic and a finite-series log enclosure support the guarantee. The package preserves the proof, code, data, source range records, figures, per-window bounds, and independent checks. Stability does not prove correctness: 163 of 814 pooled certificates had a wrong task-label decision. This is an offline method study on reused epilepsy recordings. It does not test time-varying gains, device faults, online commands, or brain stimulation.

stereoelectroencephalography · brain-computer interfaces · signal processing · robust classification · interval arithmetic · contact gain · decision stability

Admitted 0 citing publications

References

4
  1. usesQuadratic bounds certified 39.4 percent of later SEEG decisions for simultaneous constant contact gain changes of up to five percent in eight recordings sci:sha256:84281693cc487ddfcf3c2068318e3a80ccf9052c5526b3cc3992b8438e3deb78

    Uses its exact spectral moments, candidate features, frozen effective model values, chronological splits, whole-box certificates, exact allowed change witnesses, and public source range recipes. The new bounds enforce shared contact cells.

  2. usesEight SEEG pairs classify pre-response and rest windows at 70.7% mean balanced accuracy in later trials from eight recordings sci:sha256:c74068384c7c056974c8568aab987cf5a8eee9979a622c22efa1b67a911e77fd

    Uses the original rest versus pre-response fixed-reader design, eight training-selected adjacent pairs, candidate feature rules, and later-window split through the direct parent.

  3. usesdoi:10.18112/openneuro.ds004473.v1.0.2

    Source SEEG recordings and dataset metadata for OpenNeuro ds004473 version 1.0.2.

  4. doi:10.1088/1741-2552%2Facae0a

    Source study describing the SEEG recordings and movement-related oscillations.

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.