Automation of Error Recognition in Therapies Executions Based on ECogFun-VR

Abstract

Attention deficit hyperactivity disorder (ADHD) is a neurodevelopmental disorder commonly diagnosed in children and adolescents that can impair academic performance and daily living skills. Therapies based on activities of daily living (ADLs) aim to promote patient independence in self-care, including communication and mobility. ECogFun-VR is a virtual reality tool that allows children with ADHD to perform therapies in a controlled environment. This application is based on the multiple errands test paradigm and consists of four instrumental cognitive activities. It includes five scenarios 1) Room: the starting and ending point of the tasks, connected to all other scenarios; 2) Bedroom: a room where the child must prepare the backpack and store the necessary material 3) Kitchen: a place where they must prepare a snack 4) Street: a scenario that recreates a city with traffic and pedestrians, through which the child must navigate to reach the bookstore 5) Bookstore: a place where they must choose and pay for books and school supplies.

Evaluations of the patient’s performance are typically performed manually by a professional who is present during the session or by later reviewing a video recording of the session for errors. In this context, we propose a tool that allows professionals to define the tasks that patients must perform and to collect and analyze patient interactions during the session. The tool can compare patient interactions to a reference set of interactions to identify errors, using the same categorization currently used by professionals. This automation of the evaluation process can save time and improve the accuracy and consistency of evaluations.