Motion calibration reduced later motion error in fNIRS models with one nearer channel for all 12 tested adults
We tested linear fNIRS models in 12 adults from OpenNeuro ds007420 version 1.0.5. Each model used signal changes from one nearer channel to predict a 30–40 mm target channel. One fit used the first 150 seconds of rest. The other fit used the first 15 motion events. Both fits predicted the same last 15 motion events. Motion calibration gave lower error in all 12 adults. The median paired gain was 1.5837 error-ratio units. The mean motion error ratio was 0.6830; zero prediction has a ratio of one. One adult accounted for 86.8% of the total gain, so the group mean was sensitive to that case. Rest calibration gave lower error during later rest in all 12 adults. The package includes signal samples, code, source hashes, all results, and a tested retrieval recipe. This test measures optical signal prediction. It does not measure the accuracy of brain activity estimates.
functional near-infrared spectroscopy · fNIRS · calibration transfer · motion artifact · optical signal prediction · ordinary least squares
References
2- usesdoi:10.18112/openneuro.ds007420.v1.0.5
This exact OpenNeuro release supplied the rest and motion records analyzed in the study.
- doi:10.1117/1.NPh.10.2.025007
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesCalibration that omitted the test movement code had higher later fNIRS prediction error in all 12 tested adults with equal sample counts
The parent supplied the event table, channel pairing rule, source manifest, and selected signals. All selected signals were checked by fresh extraction from the original sources. This study adds an equal-count comparison of calibration fits that retain or omit the later test code.
- usesSeparate event-code fits reduced later fNIRS difference error in 12 ds007420 adults with three-second windows and one nearer channel
This study used the archived motion differences, sorted event ledger, channel map, and source identities. The new analysis compared pooled fits with separate event-code fits. Fresh raw downloads verified the reused motion arrays and channel geometry.
- usesCalibration on matching codes lowered mean later fNIRS error in 12 adults with three-second windows and one nearer channel
This study uses the stored motion differences, channel pairs, event tables, and source identities from this publication. Full raw files verified the reused arrays. The new test compares matching and other single-code fits with equal training counts in the same cohort.