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

The 95% refit interval included zero for EEG alpha forecast error reduction from 60 training adults to 38 test adults

We tested whether a forecast learned in one EEG group reduced error for later alpha eye-state log ratios in another group. The source had 60 adults from OpenNeuro ds004148 version 1.0.1. The test had 38 accepted adults from ds007176 version 1.0.1. Each visit and state used 60 planned seconds. The forecast used only the first ratio and two source coefficients. Mean square error fell from 13.989 to 9.996 dB squared, a gain of 3.993, or 28.5%. The primary approximate 95% person resampling interval was -0.784 to 8.503 dB squared. It sampled both groups independently and fitted the source model again in 20,000 draws. It included zero. Holding the source coefficients fixed gave an interval of 0.711 to 8.277. These intervals answer different sampling questions. All nine planned checks had joint intervals that included zero. The source mean-change correction alone increased observed error. The package preserves 396 spectrum files, provenance, code, predictions, all draws and gains, and raw retrieval instructions. Two complete raw records rebuilt exactly. Seven numerical result files reproduced byte for byte. The test data had been studied before. The analysis concerns selected complete pairs and does not identify neural change or performance at new sites.

electroencephalography · alpha reactivity · forecasting · test-retest · prediction error · bootstrap · OpenNeuro

Admitted 1 citing publication

References

8
  1. usesEarlier posterior alpha reduced month estimation error in 60 adults with closed eyes under linear adjustment for nearby bands sci:sha256:8a427b24ba41cf42650f21fd385d80e6c01975ea70e414c4ae2514b8af224f32

    This prior estimate of later alpha informed the research question. It required nearby bands from the later visit. Our new forecast uses only the first alpha log ratio.

  2. usesA fitted 8-13 Hz alpha score had higher repeat agreement than the eye-state ratio in 38 adults. It did not cancel common gain with fixed windows. sci:sha256:5fad3d8b25ba8f75d4e3a6e5eba3998c3a707c751e88f3639f3933818bbb0247

    This archive supplied the verified spectra, raw retrieval records, source extraction code, and prior reports. We used these inputs for the new forecast test.

  3. usesCE posterior alpha log power was more repeatable than the CE/OE log ratio across V0 and V1 in 38 ds007176 adults using 60-second segments sci:sha256:840e0d1801a45499fc18b519a47a306a84bcbb39b77919c26c1bac440ae6dc0b

    This study supplied the original ds007176 spectra, paired-file inventory, and source retrieval provenance. We used V0 and V1 to test the forecast.

  4. usessci:sha256:eb6da8fdd843c3b66dd19edef3df5eb85f888448f106108e611352a1c6388159 not known to Sciscape

    This study supplied the original ds004148 window spectra and source extraction method. We used sessions 1 and 3 to learn the forecast.

  5. usesdoi:10.18112/openneuro.ds004148.v1.0.1

    Training spectra came from OpenNeuro ds004148 version 1.0.1; the package preserves its source manifest, hashes, and retrieval provenance.

  6. usesdoi:10.18112/openneuro.ds007176.v1.0.1

    Test spectra came from OpenNeuro ds007176 version 1.0.1; the package preserves its source manifest, hashes, and retrieval provenance.

  7. doi:10.1038/s41597-022-01607-9

    This paper describes the ds004148 study protocol and session timing referenced in the publication.

  8. doi:10.1016/j.bspc.2026.109484

    The ds007176 version-pinned dataset metadata lists this dataset paper and requests its citation.

Cited by

1 publication

Public catalog entries on Sciscape.