The 95% interval included zero for the Fz versus 19-channel mean change in the CLN3 versus volunteer theta-to-alpha contrast using 80 seconds per person in 95 ds008896 records
We tested posterior theta-to-alpha power in 32 CLN3 and 63 healthy-volunteer EEG records from OpenNeuro ds008896 version 1.0.1. The primary test used the first 20 stored four-second epochs per person. We compared Fz with the mean of the same 19 scalp channels. Theta used [4,8) Hz and alpha used [8,13) Hz. The CLN3 minus volunteer log10 contrast was 0.194 with the mean reference and 0.167 with Fz. The paired contrast change was -0.027, with a 95% person bootstrap interval from -0.062 to 0.008. The interval included zero in all five planned checks. This does not establish a contrast change or reference equivalence. Scores changed order in 367 of 4465 person pairs. A shift of the band boundary to 7 Hz gave smaller group contrasts. The public data lack ages and clinical covariates, so these contrasts are unadjusted. The package contains spectra and real band matrices for all 5558 stored epochs, source metadata, locked retrieval recipes, code, results, and instructions. Separate signal and statistic checks passed for all 95 records. A fresh download and full rebuild gave identical hashes for 104 files, also in a clean Python environment.
EEG · EEG reference · spectral power · CLN3 disease · theta-to-alpha ratio
References
4- usesCommon average reference changes posterior theta-to-alpha ranks in 88 ds004504 v1.0.9 records with source processing fixed sci:sha256:2047e7d0ae35196950f5421eb9237a3b0a817c451fabcbdedd0b92ad776f3431
This study supplied the idea to use real band cross-power matrices for exact EEG reference calculations. We wrote new code and tested ds008896 rather than the adult ds004504 sample.
- supersedesThe 95% interval included zero for the Fz versus 19-channel mean change in the CLN3 versus volunteer theta-to-alpha contrast using 80 seconds per person in 95 ds008896 records sci:sha256:c4985a92358eef65ce21b6997e4d45d3b095a92255b0a13d9ee319243aea6430
This version corrects the checksum command for AutoSci downloads. The command excludes the metadata folder that sci fetch adds. The 95 spectra files and eight main analysis outputs have the same hashes.
- usesdoi:10.18112/openneuro.ds008896.v1.0.1
This analysis used public preprocessed EEG records and metadata from OpenNeuro ds008896 version 1.0.1; the source version, file hashes, and locked S3 object versions are recorded in the package.
- doi:10.3389/fnins.2017.00425
Cited by
4 publicationsPublic catalog entries on Sciscape.
- usesTemporal halves increased squared disagreement 2.76-fold for EEG theta-to-alpha scores in 95 ds008896 v1.0.1 records
Used all 95 checked spectral files, 292 source metadata files, locked retrieval receipts, and unchanged extraction and retrieval code. These inputs enabled the new temporal versus random score split test.
- usesThe 95% interval included zero for posterior mean removal in 80-second CLN3 and volunteer EEG samples
This parent supplied the 95 CLN3 and volunteer band matrices, pinned source manifests, metadata, and MIT retrieval and extraction code. We checked all parent hashes, regenerated every compact input from the source EEG, and reproduced the main baseline contrast. The present study adds posterior mean removal and its paired group contrast change.
- usesThe 95% interval included zero for the background removal change in the CLN3 versus volunteer maximum frequency contrast (4–13 Hz; 95 ds008896 records; Fz reference; 80 retained seconds per person)
We used the preserved spectra, source metadata, locked source recipe, and extraction code for all 95 records. We checked all inherited hashes and rebuilt all input spectra from fresh source files. This result tests a new maximum frequency score in the same cohort.
- usesIndependent ±3 dB epoch power gains reversed the Fz theta-to-alpha CLN3–HV contrast using 95 ds008896 records’ first 20 epochs
The 95 verified spectral archives, source metadata, locked file manifests, and extraction and retrieval code supplied the inputs for the new gain-bound test. We knew the parent baseline results before choosing this test. We retrieved all source EEG files again and rebuilt every archived array. The new result uses the same people and is not an independent cohort confirmation.