Motor Kinematic Differences in Children with ASD: Ecological Gameplay with a Sensorised Toy

Background:  

Trevarthen and Delafield-Butt (2013) proposed that one of the early markers of ASD are abnormalities in the development of intentional movements.  A growing body of literature demonstrates kinematic and action patterns differences in children and adults with ASD.  However, these experiments typically require laboratory-based optical motion tracking systems. Here, we developed, sensorised wooden cubes for motor assessment of children’s play and report on the kinematic and action pattern differences of the children with autism compared to children developing typically. 

Objectives:  

A description of ASD-specific action patterns and kinematics using sensorised toys. 

Methods:  

Participants.  Children 3 to 5 years with ASD (n = 15) recruited from the Scottish Centre for Autism, Glasgow, UK.  Children 3 to 5 years old developing typically recruited from nurseries in Glasgow, UK.  Adults 20 to 25 years old without ASD recruited from Glasgow, UK. In children with ASD, pre-screening with Vineland-II, AQ-Child and Leiter-R Brief IQ was performed.

Procedure.  The children were seated at a table and instructed to play two simple games that involved moving the cube from one position to another: a Serially Organized Action (SOA) game and a Single Repetitive Action (SRA) game.  The first required complex motor sequencing and engagement with the experimenter, while the second consisted of a simple repetitive movement. An electronic board inside the cubes was equipped with tri-axial magnetometer, gyroscope and accelerometer wirelessly transferred the cube’s motion data to a laptop.  The signal (raw motion data) was extracted through a Matlab-based platform and analysed.

Data Analysis.  Kinematic features of movement duration; maximum value of acceleration, velocity, and jerk during each movement; time to maximum value; % duration to maximum value; and the acceleration, velocity, and jerk action patterns profiles were calculated.

Results:   The jerk profile of children with ASD showed increased maximum jerk, reduced time to maximum value and duration to maximum value, and lower variability than typically developing children. Further, movement duration was shorter compared to age-matched typically developing children, and maximum velocity was significantly higher in children with ASD compared to children developing typically.

Conclusions:

The increased jerk values and onset times in the ASD group are a particularly interesting finding that support new data appearing by other groups.  It appeared, especially in the SRA game, that when moving the cube from one position to the next, the children with autism impacted on the surface of the table with greater velocity and typically included the resulting force immediately into to the next movement, giving it a greater jerk value in a shorter span of time that typically children.  Typically developing children, on the other hand, paused for a moment (>100 ms) before commencing the next movement. The repetitive simplicity of the SOA game and its resulting jerk profile appears to report on a particular behavioural motor feature distinct to ASD, namely stopping an action and starting a new one, while also describing an underlying motor difference that may contribute to it.