The last eye-state label rule gave a mean closed fraction of 64.7% for known-state pulse marks in 85 ds006036 records
We tested all 88 raw event files in OpenNeuro ds006036 version 1.0.6. The latest eye-state label set each pulse state. Among 85 people with known-state pulses, the mean closed fraction was 64.7%. The 95% person bootstrap interval was 63.3% to 66.2%. Three people had no known-state pulses. The package preserves 49,342 pulse assignments, source row links, interval tables, code, and fixed source hashes. Separate code checked every assignment. Nine raw EEG files confirmed the event exports. The result concerns annotations under a last-label assumption. It does not verify eyelid position or neural response.
electroencephalography · eye-state annotations · photostimulation · event data · reproducible research · OpenNeuro
References
3- usesA fixed EEG model had mean random-label balanced accuracy near 33% across rest and light stimulation in 72 shared people sci:sha256:b94d8cfd8435168622cde3fad8b29b748f60fdfeb32f94a62f6e20426c4691f3
This publication supplied the archived ds006036 version 1.0.6 source metadata at study start. We checked all 269 files against the fixed Git commit. We computed all state assignments anew. We did not use its EEG features or prediction results.
- usesdoi:10.18112/openneuro.ds006036.v1.0.6
Source dataset used: all 88 raw event files from ds006036 version 1.0.6; the archived source metadata records CC0 and a fixed Git commit.
- supportsdoi:10.3390/data10050064
The dataset paper reports eyes-open instructions, eye-closure annotations, and preprocessing that removed bad data periods; this work measures the extent of those annotations in the specified source version.
Cited by
2 publicationsPublic catalog entries on Sciscape.
- usesPosterior EEG ratios had rest-light concordance of 0.187 in 72 selected people with equal sample time and fixed source processing
This work supports the interpretation limit for eye state and raw event times. The new test does not assign eye state to edited derivative windows.
- usesPosterior 9.5 to 10.5 Hz EEG power was higher with closed-eye labels in 75 ds006036 records. The test used paired 10 Hz train windows and the last eye-state label rule.
This study supplied 269 fixed source metadata files and the earlier last eye-state label audit. Its state and frequency rules informed the local plan. New raw EEG downloads supplied every amplitude result.