World Health Organization class I registered international clinical trial platform, ChiCTR-ROC-17012877.

Background: Interindividual variability is important in the evolution of adaptative profiles of children with ASD having benefited from an early intervention make up for deficits in communication, language and social interactions. Therefore, this paper aimed to determine the nature of factors influencing the efficacy variability of a particular intervention technique i.e., “Play-based communication and behavior intervention” (PCBI).Methods: The participants comprised 70 13–30-month-old toddlers with ASD enrolled in PCBI for 12 weeks. The Autism Treatment Evaluation Checklist (ATEC) was used to evaluate the efficacy of PCBI. Video recordings of 5 min of free-play before and after PCBI were used to examine behaviors of mothers and children and parent-child dyadic synchrony. Hierarchical multiple regression analyses and machine learning algorithms were performed to explore the effect of these potential predictors (mothers' factors, children's factors and videotaped mother-child interaction) of intervention efficacy.Results: The hierarchical regression analysis and the machine learning algorithms indicated that parenting stress, level of completion of training at home and mother-child dyadic synchrony were crucial factors in predicting and monitoring the efficacy of PCBI.Conclusions: In summary, the findings suggest that PCBI could be particularly beneficial to children with ASD who show a good performance in the mother-child dyadic synchrony evaluation. A better dyadic mother-child synchrony could enhance the PCBI efficacy through adapted emotional and behavioral responses of the mother and the child and has a beneficial influence on the child's psychological development.