sciscape Publications
← All publications
cs.LG · Machine Learning

Six features added to a fixed SOZ model gave a 0.014 mean treated contact AUC gain in 17 SEEG patients with source outcome S

We tested if a model for seizure onset labels also ranked treated contacts. We used 31 public SEEG patients from OpenNeuro ds004100 version 1.1.3. Each test patient was excluded from model training. The primary estimate used the 17 patients in the source success group. Six spectral features raised mean treated contact AUC from 0.7063 to 0.7206 beyond a contact number model. The mean gain was 0.0143. Its 95% patient bootstrap interval was -0.0604 to 0.0836. We used 500 patient bootstrap draws and fit both models again within each draw. The interval includes zero. The same scores gave a 0.0926 mean gain for seizure onset labels in these patients. This is a new label test in an earlier selected cohort. It is not an independent patient validation. We checked all 62 public signal prefixes and all used clinical metadata. We preserved code, spectra, labels, predictions, bootstrap results, source hashes, figures, and repeat instructions. These data do not establish a precise positive gain for treatment labels or a clinical benefit.

stereoelectroencephalography · seizure onset zone · contact ranking · epilepsy surgery · spectral features · area under the curve · patient bootstrap

Admitted 0 citing publications

References

4
  1. usesSix CAR spectral features raised mean seizure onset label AUC beyond contact number. Fixed logistic tests excluded each of 31 SEEG patients. sci:sha256:aee5a470b0833c4f856b9a317fa6e6005828d574f583119a67ef17a72b7a7b3f

    This publication supplied the fixed models, features, archived predictions, 31-patient cohort, Newton solver, and preserved signal source files.

  2. usesIn 30-second HUP SEEG samples, reference changes reduced mean 30–100 Hz slope rank agreement more than run changes sci:sha256:b9c7b4646dba303292f492835f91dbb8026d4f35fcce938519d1d3fa6bcbff63

    This publication created the saved spectra, clinical metadata archive, and fixed public retrieval recipes preserved through the direct model source.

  3. usesA patient permutation test did not detect an outcome group difference in 80 to 150 Hz common average power ranks for treated contacts in 34 selected SEEG patients sci:sha256:30be157eef2724ecaea6b1494d620503ce5006dc3ab6f6c26df97c130e82f21d

    Its prior label audit supplied treatment-token and outcome definitions and known interpretation limits. Its high-gamma estimates are not reused as new results.

  4. usesdoi:10.18112/openneuro.ds004100.v1.1.3

    The analysis used 31 SEEG patients and selected contacts from OpenNeuro ds004100 version 1.1.3.

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.