Stereotyped movement recognition in children with ASD

Autism Spectrum Disorders (ASD) manifest in different behaviors, being one of them body rocking, mouthing, and complex hand and finger movements [1]. The traditional methods for recording the number of occurrences and duration of stereotypes are inadequate and time consuming. Therefore, it was used a commercial system with accelerometers sensors that records the movement of wrist. The collected data is sent through a wireless network to the computer, where an application was developed to distinguish the movement made by the children as a stereotyped gesture, hand flapping movement, from a normal gesture. The system was previously tested in laboratory environment and during the intervention sessions. Statistical methods were used to characterize the signal acquired from a previously expressed stereotype recorded from a child with ASD. The parameters that were analyzed are: Root Mean Square (RMS), Standard Variation, Peaks and Valleys. At the end, the proposed methodology facilitates to identify behavioral patterns special relevant when studying interaction skills in children with ASD.