Supporting Therapists’ Assessment in Parent-Mediated Training through Autonomous Data Collection

Parental Training is a methodology where therapists teach caregivers how to train their kids in specific behaviors practicing. It can be used in verbal behavior acquisition for children with autism mediated by the parents. In this paper, we are presenting an innovative software that aids the assessment of children’s tact acquisition where therapists can design interactive activities to support the remote measurement of the parental training progress. The developed software stores in its server the child's answers and autonomously compute, in run time, the child's face deviation through an Artificial Intelligence algorithm. The therapists have access to such data and can assess children’s performance by that without watching the interaction. The performed experiment presents initial validation of the proposed system utilization by one therapist and one dyad of a child with autism and the mother. The system was applied to the child’s evaluation phase of an entire parental training cycle in the Brazilian Portuguese language context. Through the interviews and questionnaires, all users claimed they considered our solution adequate and robust for its purposes. After comparing the child's performance in a baseline activity and an activity after the training session, the software was able to provide to the therapist the data to confirm an enhancement in the child’s tact acquisition.