An accumulation of research evidence has pointed to parent-implemented communication treatment as effective in reducing the severity of social communication deficits in preschool children with ASD. Despite even high-quality evidence, real-world translation to clinical practice remains challenging, especially for children from lower-income families, for two reasons. First, the treatment outcome is highly variable despite study-level efficacy data, most likely due to unique child and parent factors that make treatment response uneven across individual children. Second, the cost of intervention with the largest effect sizes remains high due to its one-on-one format. With the overarching goal to reduce cost and to increase treatment effectiveness at the individual-child level, this project will conduct a randomized controlled trial (RCT) to compare the effectiveness of two options for intervention to address two specific objectives. The investigators will first ascertain whether parent-implemented communication treatment taught by a speech therapist in an Individual (one-on-one) format is more effective than treatment taught in a Group format (up to 8 families learning together) at the study level. The Individual format is at least 4 times more expensive than the Group format; its relative treatment effect must be empirically ascertained to justify its cost. The investigators will then evaluate what combinations of parent and child behavioral and neural factors determine which format of intervention is likely to be more effective at the individual-child level. It is likely that not all families require the more costly Individual format of intervention. Machine-learning analytics with cross-validation will be used in constructing predictive models of treatment response, which will increase the likelihood of these models being generalizable to new patients. This study will be among the first examples of fulfilling the promise of Precision Medicine in providing guidance to patients and families with developmental disorders not about whether to receive intervention but which option for intervention to receive in the context of multiple options. This predict-to-prescribe approach of ASD intervention will likely lead to a paradigm shift in clinical practice and ultimately result in lowering the overall cost and increasing the effectiveness of intervention for children with ASD as individuals.