2.3.3. EEG preprocessing
The EEG preprocessing was based on the procedures described in Kotowski
et al. (2023). The code was written in Python 3.9.7 using the MNE 0.24.0
package (Gramfort et al., 2013) and will be publicly available after publication. In short, any manually identified
corrupted channel (various individual channels in 17 participants) were
interpolated using spherical splines (Perrin et al., 1989). The EEG
channels were recalculated to common-average reference and band-pass
filtered between 2 and 40 Hz using a third-order zero-phase
forward-backward digital Butterworth filter (Gustafsson, 1996). ERP
epochs between -100 to 500 ms relative to stimulus onset in correct
trials were extracted. Ocular artifacts were corrected by removing
independent components (ICs) correlated with eye blinks in EOG signal by
more than 2 SD s higher than others. We manually removed some
additional artifactual ICs for 16 participants (on average 1.4 ICs per
participant, as reported in the code provided above) that were not
detected by the algorithm due to weak EOG signals or due to artifacts of
non-ocular origin. Finally, the ERPs were corrected with respect to the
100 ms pre-stimulus baseline.