Development and validation of a system for the prediction of challenging behaviors of people with autism spectrum disorder based on a smart wearable shirt: A mixed-methods design.

Background: Most people with autism spectrum disorder (ASD) present at least one form of challenging behavior (CB), causing reduced life quality, social interactions, and community-based service inclusion. Objectives: The current study had two objectives: (1) to assess the differences in physiological reaction to stressful stimuli between adults with and without high-functioning ASD; (2) to develop a system able to predict the incoming occurrence of a challenging behaviors (CBs) in real time and inform the caregiver that a CB is about to occur; (3) to evaluate the acceptability and usefulness of the developed system for users with ASD and their caregivers. Methods: Comparison between physiological parameters will be conducted by enrolling two groups of 20 participants with and without ASD monitored while watching a relaxing and disturbing video. To understand the variations of the parameters that occur before the CB takes place, 10 participants with ASD who have aggressive or disruptive CBs will be monitored for 7 days. Then, an ML algorithm capable of predicting immediate CB occurrence based on physiological parameter variations is about to be developed. After developing the application-based algorithm, an efficient proof of concept (POC) will be carried out on one participant with ASD and CB. A focus group, including health professionals, will test the POC to identify the strengths and weaknesses of the developed system. Results: Higher stress level is anticipated in the group of people with ASD looking at the disturbing video than in the typically developed peers. From the obtained data, the developed algorithm is used to predict CBs that are about to occur in the upcoming 1 min. A high level of satisfaction with the proposed technology and useful consideration for further developments are expected to emerge from the focus group. Clinical trial registration: [https://clinicaltrials.gov/], identifier [NCT05340608].