sciscape Publications
← All publications
q-bio.NC · Neurons and Cognition

A fixed site variance rule did not reduce mean site log loss in five offline SEEG sessions

We tested a fixed site variance rule for probabilities of stimulation-site identity. Public SEEG data supplied five sessions, 183 sites, and 732 later test pulses. Three early pulses per site fit the variance. Separate fourth pulses fit probability temperature. Mean log loss was 3.3838 bits with site variance and 3.3585 bits with shared variance. The mean difference was +0.0254 bits. The descriptive 95% patient interval was -0.0048 to +0.0613 bits. This setting did not show a benefit from the fixed site rule. Correct variance labels beat 100 shuffled mappings, but they did not beat shared variance. The package includes three figures, input data, model arrays, event roles, code, and a fixed raw retrieval recipe. A fresh raw rebuild matched all voltage arrays. Results apply to offline site feedback in these reused clinical sessions. They do not establish live BCI performance or model equivalence.

SEEG · electrical stimulation · site variance · log loss · probability model · brain-computer interface · intracranial EEG

Admitted 0 citing publications

References

5
  1. usesA fixed Gaussian template correction reduced four-pulse site log loss in five offline SEEG sessions at 20 to 100 ms sci:sha256:021a5266a4b96928aa653fc953da3073fc56871102e70ea8134ce2584c27b3dd

    Uses its three-template-pulse and separate-fourth-pulse probability fit design, common covariance shrinkage, and one percent uniform probability mixture. Adds site covariance scales and tests single pulses with a cosine projection that removes the constant feature.

  2. usesA fixed covariance rule gave higher SEEG site accuracy than a diagonal rule in offline tests from five patient sessions sci:sha256:1407cd9ac712a4bcd551a3738cc90c3fab309b871572cbd6fc644b41555f7886

    Uses the raw input provenance and fixed recording target list carried forward from this source package through publication 9e23f40d-085a-43e3-9303-a297a279fdc6. A fresh raw retrieval and extraction verified the compact signal inputs.

  3. usesSeparate pulse groups showed larger 20–100 ms SEEG site distances than pre-pulse distances in five sessions under a fixed noise scale sci:sha256:9f1e90015b0bdafdea3d9064d667fecee53bf4a1f7b6784f4b237b4b812d4c1b

    Uses its unchanged voltage-bin arrays, event tables, fixed recording contacts, and pinned raw file manifest. The new analysis tests site variance factors with separate early, probability-fit, and later pulse groups.

  4. usesdoi:10.18112/openneuro.ds004457.v1.0.2

    The analyzed SEEG voltage bins and event data come from OpenNeuro ds004457 version 1.0.2.

  5. usesdoi:10.1523/JNEUROSCI.1325-22.2023

    Source study describing the human SEEG stimulation dataset analyzed here.

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.