A same-day alpha interval contained 25 of 30 month repeats in one split of ds004148
We tested a prediction interval for closed-eye posterior alpha power in OpenNeuro ds004148 version 1.0.1. We used 30 people to set a same-day interval with a 90% target. We tested the same interval in 30 separate people. Its radius was 2.474 dB. It contained 29 of 30 same-day repeats and 25 of 30 month repeats. The paired difference was -13.3 percentage points. The exact paired p-value was 0.21875. The primary comparison did not establish a coverage difference or equal coverage. Exchanging the sets changed the radius to 1.648 dB. All nine planned settings are reported. The package supplies an interval tool, 180 source spectrum files, all decisions, bootstrap draws, source hashes, and raw retrieval code. All 900 power scores matched the prior analysis. Three fixed raw records reproduced the spectra exactly. Prior results from this cohort informed the question. The finite-rank guarantee requires exchangeable calibration and future scores. It gives no general guarantee across a changed repeat protocol.
EEG · alpha power · prediction intervals · split conformal prediction · test-retest reliability · posterior EEG · coverage · exchangeability · OpenNeuro ds004148
References
3- usessci:sha256:eb6da8fdd843c3b66dd19edef3df5eb85f888448f106108e611352a1c6388159 not known to Sciscape
We used the 180 closed-eye window spectra, source metadata, extraction code, and archived band-power results. These files supplied the scores for a new test of prediction intervals. Prior agreement results from the same cohort informed the question. We checked all new power scores against the archived scores and rebuilt three fixed raw records.
- usesdoi:10.18112/openneuro.ds004148.v1.0.1
This exact OpenNeuro release supplied the analyzed EEG records; the publication preserves source URLs and hashes.
- supportsdoi:10.1038/s41597-022-01607-9
The source dataset paper reports the cohort and matched visit timing used to describe the dataset.
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesRanges with a 90% target from ds004148 contained 32/39 ds007176 V1 alpha repeats at fixed width and 36/39 with scaling
This prior same-cohort interval result informed the transfer question. The new study uses different segments and calibration pairs, separate test people, and a fixed versus baseline-scaled radius comparison. It does not reproduce the prior counts.
- usesA joint rule narrowed alpha difference ranges in ds004148; 95% coverage requires exchangeable person records
Its single-value alpha repeat ranges informed the present question. This study extends the scope to three jointly predicted values and calibrates month repeats directly. It does not repeat the earlier same-day calibration experiment.
- usesSame-day ranges for posterior EEG alpha with closed eyes contained 57% of one-month pair differences in 30 test adults
We read this single-person repeat-range study during a scope check after our analysis. It supplies prior work context. It did not set our split, rules, or outcomes. Our range concerns differences between two distinct people.
- usesRanges with a 90% target from ds004148 contained 32/39 ds007176 V1 alpha repeats at fixed width and 36/39 with scaling
This prior same-cohort interval result informed the transfer question. The new study uses different segments and calibration pairs, separate test people, and a fixed versus baseline-scaled radius comparison. It does not reproduce the prior counts.
- usesSame-day calibration gave 88.4% mean month containment for nominal 90% frequency intervals from 280-second EEG in 60 adults
This prior alpha-power study tested same-day calibration for month prediction in one split of the same people. It informed the calibration question. We use frequency scores and exact averages over every uniform calibration group.