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

Adding a nonnegative theta term to a quadratic model raised mean later session error for 2 to 20 Hz hippocampal spectrum changes in 12 SEEG patients

We tested whether a theta term improved prediction of memory phase spectrum changes in public hippocampal SEEG. We used 12 patients, 44 matched pairs, and 1,085 trials from OpenNeuro ds004752 version 1.0.1. We fitted models on session 1 and fixed all coefficients for session 2. The test used 2 to 20 Hz and a quadratic background in log frequency. The Gaussian theta term had an amplitude at least zero. It raised mean squared error from 18.402 to 18.856 dB squared. The mean error rise was 0.454 dB squared. Its pointwise 95 percent patient bootstrap interval was [0.175, 0.767]. Error rose in 10 of 12 patients. The two-sided patient sign reversal test gave Holm p = 0.02051. This test assumes independent patients and symmetric null contrasts. All seven planned checks had the same direction. A later check that removed digital bound trials also had this direction. All 2,334 source windows and all rebuilt Fourier files matched their inherited records. The package preserves code, coefficients, raw examples, fits, results, source pins, and reproduction instructions. This is an exploratory test in a reused epilepsy cohort. It does not show that theta oscillations are absent.

hippocampus · SEEG · verbal working memory · theta · spectral change · cross-session prediction · epilepsy

Admitted 0 citing publications

References

7
  1. usesMean hippocampal theta templates increased average relative error during encoding and late maintenance between two sessions in 12 SEEG patients sci:sha256:dbe0b33f69fdbc8b88eaffe73dfdb5e173ac6213e699229888d4e209e0410860

    This related mean theta waveform transfer study was read after the main outcomes and is used for scientific context. It predicts complex trial coefficients against zero. The present test predicts a residual log spectrum difference against a curved background. Reading it did not change the planned main analysis.

  2. usesAn affine fit showed a hippocampal theta excess rise across a three-second delay in 12 SEEG patients sci:sha256:dbf185849d88010778652b5112f654ebcade37bc6528194b481e78dbdfc39bcb

    The known difference between straight and curved spectral background contrasts motivated the quadratic background in this fixed cross-session prediction test. This work does not use the earlier within-delay contrast as its outcome.

  3. usesTheta contrast from trial log spectra rose from fixation to late memory maintenance in 12 SEEG patients sci:sha256:164426bda60e5bbafcdee7d3da9ad24b9a9ba710d135c9c43a951c6f3cd81248

    The known fixed theta spectral contrast result motivated a test of whether an added theta shape predicts the phase change in a later session. This work tests a different model and score in the same patients.

  4. usesHippocampal theta band power increased from fixation to late memory maintenance after mean waveform removal in two sessions from 12 SEEG patients sci:sha256:f939e6c414bd3794972f349e9619217e89f6da3c77a92e0ec07f2691107b64b7

    This work uses the parent sample, complex Fourier coefficients, raw examples, metadata, source pins, and MIT source interface. All copied file hashes matched. Fresh retrieval checked every exported phase window and rebuilt all 24 coefficient files exactly. The patients are reused; this is not an independent cohort replication.

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

    The analyzed public source is OpenNeuro ds004752 version 1.0.1; the retained dataset metadata declares this exact version and CC0.

  6. doi:10.7554/eLife.78677
  7. doi:10.1038/s41593-020-00744-x

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.