Studying normal infant development is a challenge for cognitive scientists in general and for neuroscientists in particular because: (1) physiological indices of infant cognition are generally noisy and technically difficult to obtain; and (2) interindividual variability and a paucity of established results make data interpretation very complex, even aside from the fact that one cannot overlook fundamental ethical issues on experimenting with infants. During the first year of life in particular, infants cannot follow instructions, speech is as yet virtually non-existent; and it is difficult to obtain overt behavioural responses that meet statistical validation criteria. For these reasons, studies of normal infant development have mainly capitalized on attentional capture paradigms which measure the infant's shifts of attention. Such paradigms include high amplitude sucking, preferential looking, and head turn protocols, in all of which overt behavioural measures are assumed to relate to the infant's perception and processing of stimuli. Electroencephalography and more specifically event-related potentials (ERPs), on the other hand, can offer an insight into brain implicit responses to external events, even when no external behaviour is observed. As a consequence, there is increasing interest in the use of ERPs for the study of early development. This review gives a concise, non-exhaustive overview of the type of ERP experiments that have been conducted with young infants and addresses the central theoretical issue of ERP component interpretation in infancy.