We examined the course of caregiver motherese and the course of the infant’s response based on home movies from two single cases: a boy with typical development and a boy with autistic development. We first blindly assessed infant CG interaction using the Observer computer-based coding procedure, then analyzed speech CG production using a computerized algorithm. Finally we fused the two procedures and filtered for co-occurrence. In this exploratory study we found that the course of CG parentese differed based on gender (...) and child status. The course of an infant’s response to CG vocalization differed according to the type of speech and child status. Mothers spent more time interacting with infants, and fathers appeared to interact with their child preferentially between 12 and 18 months in the TD boy, but not in the AD boy. The TD boy responded equally well to motherese compared to other speech after 1 year of age. For the AD boy, the responses to both types of speech were lower than in the boy with TD and decreased from the second to the third semester. Keywords: Autism; motherese; early interaction; computational methods. (shrink)