Motor Differences in Children with Autism Engaged in Ipad Gameplay

Background:  

One of the early markers of autism spectrum disorder (ASD) are abnormalities in the development of intentional movements. However, very few studies directly addressed the question of whether children suffering from ASD can be distinguished from typically developing ones solely on the basis of analysing how they make intentional movements. Individuals diagnosed with ASD should display abnormalities in prospective planning and execution of movements, thus we hypothesised that their movement patterns while using mobile device should be different from patterns found in typically developing individuals.

Objectives:  

The aims of the study were the following: (1)  To determine whether or not motor information could differentiate children with autism from children developing typically; (2)  To determine the kinds of movements responsible for differentiating between children with autism and children developing typically. 

Methods:  

45 children aged 3 to 6 years old and clinically diagnosed with Childhood Autism (n=42) were included in the ‘Autism’ group of the study.  Of these, 12 were female.  45 children age-matched and gender matched typically developing children were included in the ‘Control’ group.  All the participants had normal or corrected-to-normal vision and no other sensory or motor deficits that could make engaging in tablet gameplay difficult. For the purpose of investigating movement patterns, iOS tablets (iPad minis) were used.  The children were asked to use two educational applications: (1) ‘Sharing’ where the main gameplay was to divide a piece of food (e.g. an apple) and distribute it evenly among four children present on the screen. (2) Creativity where the game open and unstructured and involved colouring pictures of toys and animals.  During gameplay, data from tablets' inertial sensors (tri-axial accelerometer, gyroscope and magnetometer) and data from its touch screen were collected.  262 features were selected and subsequently analysed using the Regularized Greedy Forest (RGF; Johnson, Zhang, 2014) machine learning algorithm with 10 repetitions of 10-fold cross-validation with pre-selected features. 

Results:  

The RGF algorithm with classified movement patterns as related to ASD with up to 83% sensitivity and 85% specificity based solely on motor features, suggesting that movements of children diagnosed with ASD can be distinguished from those displayed by typically developing ones.  In the Sharing game, 4 out of 10 of the most salient (greatest difference between ASD and control group) features were those derived from the inertial sensors (accelerometer and gyroscope), while the other 6 features were measures of the finger swipe kinematic.  In the Creativity game, 10/10 of the most salient features were derived from the inertial sensors. 

Conclusions:  

The analysis of movements could add significant value to the process of early diagnosis of ASD and could be used as a biobehavioural marker of the disorder.  The results indicate that it is the increased force of impact onto the screen made during a touch, or maintained during a swipe, that is specific to children with ASD.  Finger swipe kinematics are another distinctive feature of the disorder.  Altogether, motor assessment made during iPad gameplay proves to be an accessible and enjoyable paradigm for autism research and assessment.