Social interactions are vital for healthy brain development. Burgeoning behavioural evidence indicates that a caregiver who provides contingently timed vocal responses to infant vocalisations provides key support for early language development. Understanding how contingently timed vocal responses relate to neurodevelopment in early infancy is lacking. This study compares event-related potentials (ERPs) to contingent and non-contingently timed vocalisations in 6- and 9-month-old infants (n = 36), and adults (n = 24). ERPs were recorded from each age group while listening to a naturalistic 21-minute recording of a mother playing and conversing with her baby. At 6-months, infants showed a significant positive ERP response to contingent vocalisations by the mother and infant. At 9-months infants showed negative ERP response to the mother’s contingent speech. Adults showed no differences in ERPs between contingent and non-contingent speech regardless of the talker. We interpret the increased positivity in response to contingent speech as suggesting that infants show an attentional response at 6-months, and the increased negativity at 9-months relates to lexical-semantic processing. Further work is necessary to confirm the development of distinct ERPs shown in response to natural speech.