Computerized rejection of artefacts was performed before averaging to discard epochs in which eye movements, blinks, excessive muscle potentials, or amplifier blocking occurred. The artefact's rejection criterion was a peak-to-peak amplitude exceeding ± 50 μV and the rejection rate was about 5%. ERPs were averaged offline from 100 ms before until 1000 ms after presentation of the final word. ERP trials associated with an incorrect behavioural response were excluded from further analysis. For each subject, distinct ERP averages were obtained according to infant's facial expression. ERP components were identified and measured with reference to the baseline voltage averages over the interval from -100 ms to 0 ms. P1 and N1 peak amplitude and latency values were measured at lateral occipital sites (OL, OR), where both components reached their maximum amplitude, in the time window between 90–140 ms and 145–175 ms. The McCarthy-Wood correction, sometimes used to normalize ERP amplitudes, was not applied to our data, in line with recent findings in the literature [60].
