sciscape Publications
← All publications
q-bio.NC · Neurons and Cognitioncs.LG · Machine Learning

Covariance correction from 12 early trial positions showed no clear load AUC gain in 15 final SEEG sessions with assumed timing

We tested if early SEEG signals can improve a fixed reader of task load without new load labels. We used public OpenNeuro ds004752 version 1.0.1. Each of 15 patients had a separate reader. The correction used artifact-free rows from the first 12 original trial positions in the final numbered session. It matched regularized feature covariance to earlier training data. On 319 later trials, mean AUC changed from 0.5741 to 0.5836. The paired mean gain was 0.0095, with a 95 percent patient bootstrap interval from -0.0077 to 0.0267. This test showed no clear gain. It does not prove equal performance. Mean translation preserved AUC, as the fixed score equation requires. The package includes all planned comparisons, five figures, code, compact features, predictions, and independent checks. Complete raw stream checks verified 68 file hashes and rebuilt all features. Phase timing is assumed. This clinical test did not assess online use or brain stimulation.

stereoelectroencephalography · verbal working memory · covariance correction · domain adaptation · session transfer · area under the curve · brain-computer interface

Admitted 1 citing publication

References

7
  1. usesA fixed SEEG power model improved four versus eight letter ranking beyond history after correct responses in 13 final sessions sci:sha256:7917c63450566e88178fa4546b940d6c228e2a99f24d3a1ce66f24f825135a8f

    We used its fixed training feature selection and ridge logistic fitting functions through the direct parent code. We kept these settings fixed when testing correction without new task labels.

  2. usesWith fixed epoch timing, a mean of fixation log power over eight trials showed no clear letter load AUC gain in 15 final SEEG sessions sci:sha256:0651386c1ce770ddc4491f3ed2f3469589c9414b4829d3263b1de3876651637d

    This is the earlier feature source in the direct parent provenance. We preserve its source version, extraction records, and file hashes in the reproducible data chain.

  3. usesA fixed SEEG model of letter load showed no clear AUC gain from calibration across earlier sessions at matched trial counts in ten patients sci:sha256:3c721fccfd15bf7d772a58df49c0aaec7f9729b2b0346f1e2739cecf0552920a

    We used its raw range audit and complete EDF stream reconstruction methods through the direct parent. These methods checked source powers and complete raw file hashes in this investigation.

  4. usesWith shared features and assumed timing, maintenance fitting showed no clear load AUC gain over encoding fitting in 15 final SEEG sessions sci:sha256:bb927624b1ca4bd356fb1e6172f76ae4255726b47f870a5f9e21b52355387e1f

    We used its unchanged 68 signal feature arrays, source metadata, source hashes, and extraction code. These inputs define the common signal space for this new early-session correction test.

  5. usesdoi:10.18112/openneuro.ds004752.v1.0.1

    The study uses the version-pinned OpenNeuro ds004752 1.0.1 recordings and compact features derived from those recordings.

  6. doi:10.7554/eLife.78677
  7. doi:10.1038/s41597-020-0364-3

Cited by

1 publication

Public catalog entries on Sciscape.