sciscape Publications
← All publications
q-bio.NC · Neurons and Cognitioneess.SP · Signal Processing

Residual 80 to 150 Hz concentration gave an uncertain mean AUC gain of 0.021 in 19 NIH and JHU SEEG patients after HUP fitting

We tested whether residual high-frequency power concentration added information about clinical seizure-onset labels in public SEEG data. Residual concentration was observed concentration minus the mean of 32 Fourier phase controls. The controls preserved full auto spectra and cross spectra. We fitted fixed models on 34 HUP patients and tested them on 19 NIH and JHU patients. The baseline used beta power, 80 to 150 Hz power, mean control concentration, and contact number. Mean patient AUC was 0.735 for the baseline and 0.756 with residual concentration. The mean gain was 0.021. Its conditional 95% patient-bootstrap interval was -0.011 to 0.058. Nine patients improved and ten worsened. The exact two-sided sign-test p value was 1.000. The mean gain for comparisons within a contact-name prefix was -0.004. These results do not establish a reliable added benefit. The interval holds training data and simulated features fixed. Clinical labels defined the pair pool, so this is not a test on all unlabeled contacts. The archive contains 212 feature arrays, source metadata and range hashes, source readers, fitted models, predictions, all planned checks, figures, and repeat instructions. A second process reproduced all seven numerical result files exactly. Fresh source downloads for one patient in each dataset reproduced the study arrays. The source data and package use CC0.

stereoelectroencephalography · seizure-onset zone · high-frequency power · power concentration · Fourier phase controls · external validation · area under the curve

Admitted 1 citing publication

References

9
  1. usesAlpha excess selection gave more test window excess than alpha power selection in 19 SEEG records with an average reference sci:sha256:fbe93f72b2b5b6d83ae04bbdd3e55a7776af4f7b40c3da5101c3aeaa095ef7cf

    We used its metadata selection of 19 NIH and JHU records, pinned source range recipe, EDF headers, clinical labels, and source readers. We retrieved the signal ranges again and calculated new concentration features.

  2. usesIn 34 SEEG patients, 80 to 150 Hz concentration had mean AUC 0.706 compared with 0.649 for phase controls with equal spectra sci:sha256:b0fe9fe5e8eb7f7a92b5ef20c553c61f9afa61be6471983de8dd13f614a25e38

    We used its 34-patient HUP power and phase arrays, source metadata, source range receipts, and source code. They supplied the training data and the common-phase method.

  3. usesAdding time concentration of 80 to 150 Hz power increased mean AUC by 0.021 in 34 patient holdout tests of two 60 s SEEG clips. The 95 percent patient bootstrap interval included zero. sci:sha256:56007b420ccbb421d8b8e42cac1ebc3f6adab409768ca8e320226af28a6552da

    We used its HUP raw-concentration model result as related work. We read it after the local plan and made no change to the fixed rules. This publication adds a separate external residual-concentration test.

  4. usesWith independent patients and null sign symmetry, tests detected no HUP model SOZ AUC gain in 18 external SEEG patients sci:sha256:cacfcb73cf6624b3b7d324cad3700beeace003783bcff021dcba6e9769e783f7

    Its related HUP-to-external spectral model and contact-number comparison informed this question. We used no outcomes from that model as observations in the new analysis.

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

    Training source: HUP data from OpenNeuro ds004100, version 1.1.3.

  6. usesdoi:10.18112/openneuro.ds003876.v1.0.2

    External test source: NIH and JHU data from OpenNeuro ds003876, version 1.0.2.

  7. doi:10.1038/s41593-021-00901-w
  8. doi:10.1101/2021.10.15.464594
  9. doi:10.1103/PhysRevLett.73.951

Cited by

1 publication

Public catalog entries on Sciscape.