Swipe kinematics differ in different aged children with autism spectrum disorders during smart-tablet gameplay

Research output: Contribution to conferencePosterpeer-review


Background: Atypical movement patterns in people with autism spectrum disorders (ASD) have been reported (Dowd et al., 2012; Trevarthen and Delafield-Butt, 2013; Cook, 2016). In our previous observation, children with ASD moved slower than typical developing (TD) children in a goal-oriented food-sharing task during smart-tablet gameplay (Lu et al., 2019). However, contradictory results have been noticed in another cohort. Further, it is well known that visual-motor integration skills change dynamically in early childhood, improving remarkably at 4-5 years old (von Hofsten and Rösblad, 1988; Fang et al., 2017). Consequently, this study re-examines the swipe kinematics in different age groups. Objectives: This study compares the goal-oriented swipe kinematics of ASD and TD children in different age groups, and determines the feasibility of using smart-tablet devices to identify movement signatures in children with ASD. Methods: Gameplay data from two cohorts were analysed: Cohort 1, a previous study (Anzulewicz et al., 2016), in which 45 TD and 37 ASD children aged 25-79 months took part; in Cohort 2, 316 TD and 202 ASD children aged 30-90 months participated as part of a multi-site trial (Millar et al., 2019). The participants played a food-sharing game on a smart-tablet (iPad mini, Apple Inc.) and their on-screen touch trajectories were recorded. The data were analysed using customised MATLAB scripts. The food-to-target swipes data were mined to yield the duration, travelled distance, minimal distance (distance between the start and end points), distance difference (the difference between the travelled and minimal distances), and average speed (travelled distance divided by duration) of each swipe. The data were then categorised into two age groups: <60 months and ≥60 months, and Mann-Whitney U tests were used to determine kinematic differences between TD and ASD in each age group. Results: In Cohort 1, ASD moved significantly faster than TD at the younger age (median of 52.65 mm/s vs. 49.70 mm/s) while they moved significantly slower than TD at the older age (median of 46.82 mm/s vs. 81.72 mm/s). The same pattern was observed in Cohort 2, the median value of average speed being 57.81 mm/s in ASD vs. 49.65 mm/s in TD at the younger age, and 65.81 mm/s in ASD vs. 67.00 mm/s in TD at the older age. An analysis on the combined data from the two cohorts yielded the same patterns in the average speed (see details in the attached image). Conclusions: The kinematic differences between TD and ASD during goal-oriented swipe movements follow different profiles in different age groups; ASD moved faster than TD at the younger age (<60 months) while they moved slower than TD at the older age (≥60 months), suggesting different strategies should be used for different aged children with ASD to compensate the disruption to their sensorimotor control. This study also demonstrates the feasibility of using smart-tablet devices to support the early identification of ASD in the gameplay paradigms.
Original languageEnglish
Number of pages1
Publication statusPublished - 9 May 2020
EventInternational Society for Autism Research Virtual Annual Meeting -
Duration: 18 Aug 202019 Aug 2020


ConferenceInternational Society for Autism Research Virtual Annual Meeting


  • movement patterns
  • autism spectrum disorder (ASD)
  • visual motor integration

Cite this