31615
Predictive Gaze Shifts during the Observation of Actions in Toddlers and Preschoolers with and without Autism Spectrum Disorders
The large variety in autism phenotypes complicates the search for understanding underlying mechanisms. A few years ago a new theory was proposed, which suggests that a broad range of autistic symptoms may be manifestations of an underlying impairment in predictive abilities (Van de Cruys et al., 2014; Pellicano & Burr, 2012; Sinha et al., 2014). While there is some initial empirical support for this theory, exactly how impairments in prediction are linked to autism severity and the onset of autism spectrum disorders (ASD) remain unclear. An important ability in social situations is to predict what someone else is going to do, which can be assessed by measuring predictive gaze shifts, eye movements towards the outcome of an observed action before the action is completed.
Objectives:
The current study aimed to compare predictive gaze shifts in typically developing (TD) young children and young children with ASD prior to the onset of goal-directed actions. The second aim was to explore whether the ability to predict others’ actions is related to the severity of autism symptoms.
Methods:
Data has been processed for 18 TD children (ages 25-59 months; M = 44.6, SD = 2.5) and 14 children with ASD (ages 27-58 months; M = 39.9, SD = 3.1). ASD diagnosis was confirmed by administrating the Autism Diagnostic Observation Schedule (ADOS-2 Toddler module, Module 1 or 2). Participants observed 30 video clips in which an actor performed a grasping and placing action or pointing action (Figure 1). They were able to predict that the actor would first move towards the toy before moving toward the container, since the sequence was the same in all videos and the participants were first familiarized with the sequence. During the observation of the videos, eye movements were recorded using a Tobii Pro X2-60 eye-tracking monitor. Predictive gaze shifts were defined as eye movements toward the toy before action onset.
Results:
There was a significant difference in predictive gaze shifts between the two groups, F(1, 30) = 5.83, p = .02, ηp² = .16 (Figure 2). The TD children look more towards the toy prior to the onset of the action (M = .33, SD = .10) compared to the group with ASD (M = .22, SD = .16). In the group with ASD, there was a negative correlation between the total score on the ADOS, indicating the severity of autism symptoms, and the proportion of looking at the toy prior to action onset relative to looking anywhere else on the screen, r(14) = -.68, p = .01.
Conclusions:
Our findings suggest that prior to the onset of an observed action, young children that are typically developing predict what another person is going to do, while young children with ASD make these predictions less frequently. Especially, children with more severe autism symptoms showed fewer predictive gaze shifts, supporting a link between autism symptoms and impairments in prediction abilities. Therefore, the current study provides new evidence for the theory that individuals with ASD have impairments in predictive abilities.