The N400 ERP component is a direct neural index of word meaning. Studies show
that the N400 component is already present in early infancy, albeit often delayed. Many researchers capitalize on this finding, using the N400 component to better understand how early language acquisition unfolds. However, variability in how researchers quantify the N400 makes it difficult to set clear predictions or build theory. Not much is known about how the N400 component develops in the first 2 years of life in terms of its latency and topographical distributions, nor do we know how task parameters affect its appearance. In the current paper we carry out a systematic review, comparing over 30 studies that report the N400 component as a proxy of semantic processing elicited in infants between 0 and 24 months old who listened to linguistic stimuli. Our main finding is that there is large heterogeneity across semantic-priming studies in reported
characteristics of the N400, both with respect to latency and to distributions. With age, the onset of the N400 insignificantly decreases, while its offset slightly increases. We also examined whether the N400 appears different for recently-acquired novel words vs. existing words: both situations reveal heterogeneity across studies. Finally, we inspected whether the N400 was modulated differently with studies using a between-subject design. In infants with more proficient language skills the N400 was more often present or showed itself here with earlier latency, compared to their peers; but no consistent patterns were observed for distribution characteristics of the N400. One limitation of the current review is that we compared studies that widely differed in choice of EEG recordings, pre-processing steps and quantification of the N400, all of which could affect the characteristics of the infant N400. The field is still missing research that systematically tests development of the N400 using the same paradigm across infancy.