The first Autism Research Matrix (IACC, 2003) listed the identification of behavioural and biological markers of risk for autism as a top priority. This emphasis was based on the hypothesis that intervention with infants at-risk, at an early age when the brain is developing and before core autism symptoms have emerged, could significantly alter the developmental trajectory of children at risk for the disorder and impact long-range outcome. Research has provided support for specific models of early autism intervention (e.g., Early Start Denver Model) for improving outcomes in young children with autism, based on both behavioural and brain activity measures. Although great strides have been made in ability to identify risk markers for autism in younger infant/toddler samples, how and when to intervene during the prodromal state remains a critical question. Emerging evidence suggests that abnormal brain circuitry in autism precedes altered social behaviours; thus, an intervention designed to promote early social engagement and reciprocity potentially could steer brain development back toward the normal trajectory and remit or reduce the expression of symptoms.