Primary Data On Using Sensors to Analyze Motor Aspects of Gesture Behavior in Children with Autistic Spectrum Disorders

Background: Within our motor repertoire gestures are specific motor acts proving to be very relevant from children’s first encounters with communicative contexts. Numerous studies indicate the presence of different impairments in gesture behaviour in children with autistic spectrum disorders (ASD) and highlight how these may play an important role in communication and social interaction deficits present in ASD. However, to date, analyses of gestures produced by children with ASD have relied heavily on traditional video annotation systems. The absence of different measurement tools has overshadowed the importance of analysing motor repertoires exploited by children with ASD during gesture behaviour and prevented comparisons between action and gesture production.

Objectives: In this study we report primary results from the broader TOUM (The Other Understanding in Movement) project. The project aims to develop a novel wearable sensorized platform for the analysis of motor parameters exploited during gestural behaviour in children with ASD and to assess the platform’s effectiveness alongside traditional annotation tools. This pilot study evaluates both motor (e.g. limb velocity, orientation, etc.) and formal (e.g. hand shape, palm orientation, etc.) parameters collected using respectively sensors and traditional annotation systems, during actions and gestures produced by children with ASD and matched typically developing (TD) controls.   

Methods: A total of 10 children participated to this pilot study: 5 children with ASD (chronological age 7.2; IQ 90.3) and 5 TD controls (chronological age 6.11; IQ 105.2), matched on chronological and mental age, evaluated using Leiter International Performance Scale Revised. Motor abilities were evaluated using Movement ABC, while linguistic level was evaluated using the Peabody Picture Vocabulary Test and the Boston Naming Test. Children observed a set of short video-clips of an adult performing either actions with objects (e.g. the adult grasps a glass to drink) or corresponding transitive gestures (e.g. the adult raises his hand to the mouth as if drinking) and asked to imitate the observed action/gesture while their movements were measured with the aid of wearable wireless sensors. Children’s behaviour during action/gesture production was also videotaped for later coding using traditional annotation systems (i.e. ELAN) and to provide independent measures of action/gesture expressive clarity.   

Results: All children were able to imitate the observed actions and to produce at least some gesture types. However closer analysis of both action sequences and gesture behavior of children with ASD showed specifically altered motor patterns, in comparison to children with TD. The different motor parameters exploited by the ASD group had a relevant effect on the quality of performed actions and of produced gestures, altering their expressive clarity. Furthermore, integrating data collected relying on the sensorized platform and using traditional annotation systems allowed to parse out the relevance of motor vs. formal parameters in evaluating gesture performance.

Conclusions: Initial data seem to indicate the possibility of employing novel measurement tools, originally dedicated to motor behavior, to measure gesture production in children with ASD. This new portable and ecological platform in fact allows to capture atypical action/gesture motor patterns, which influence non-verbal behavior and communicative capacities in children with ASD.