Changing the EEG screen reference changed 6.4% of trial decisions at 150 µV in 39 runs from ds003061 v1.1.2

We tested a reference change in an EEG amplitude screen. We used all 39 runs from 13 adults in OpenNeuro ds003061 v1.1.2. We filtered 64 channels at 0.1–20 Hz. The screen required every channel range to be at most 150 µV. We changed the reference from the P9/P10 mean to the mean of 64 EEG channels. This changed 546 of 8,509 complete target and noise trial decisions (6.4%). We compared target-minus-noise Pz-minus-Fz scores in 12 matched adults and 34 common runs. The mean score change was -0.229 µV. Its 95% t interval was -0.590 to 0.133 µV. This interval includes zero. The mean absolute change across participants was 0.508 µV. A fixed trial set gave equal scores under all three references. This result tests sensitivity to processing choices in one dataset. The data do not show which screen gives a more accurate brain response. The package includes code, all trial ranges, results, source hashes, and a trial selection check.

EEG · reference choice · artifact screening · trial selection · auditory oddball · event-related potential · signal processing