Action Prediction in Infants at-Risk for Autism: Neural and Behavioral Findings

Saturday, May 14, 2016: 11:30 AM-1:30 PM
Hall A (Baltimore Convention Center)
R. Braukmann1,2, C. van den Boomen3,4, N. M. Munsters3,4, H. Bekkering2, J. K. Buitelaar1 and S. Hunnius2, (1)Radboud University Medical Centre, Nijmegen, Netherlands, (2)Radboud University Nijmegen, Donders Institute for Brain, Cognition and Behaviour, Nijmegen, Netherlands, (3)Utrecht University, Utrecht, UT, Netherlands, (4)University Medical Centre Utrecht, Utrecht, Netherlands

Recently, theoretical approaches on the underlying mechanisms of ASD suggest that abnormalities in forming and updating predictions about the environment might underlie several of the diverse deficits (van der Cruys et al.,2014; Pellicano & Burr,2012). The ability to predict others’ behavior is important for social functioning, and research has indeed demonstrated prediction difficulties in ASD. Not only do children with ASD show behavioral impairments in action prediction (Zalla et al.,2010; Fabbri-Destro et al.,2009), abnormalities of the neural representation of others’ actions in the motor system involved in forming predictions have also been reported (Cattaneo et al.,2007; Obermann et al.,2005).


Difficulties in predicting complex action sequences in older individuals with ASD might arise through abnormalities during the early development of action processing. In this study, we investigated whether deficits in processing and prediction can be observed already in infants at risk for ASD, by assessing action prediction abilities longitudinally in infants at high (HR) and low (LR) familial risk. From six months onwards, typically developing (TD) infants show more predictive eye movements towards the goal of usual compared to unusual everyday actions (Hunnius & Bekkering,2010). In addition, differential activation of the neural motor system during observation of usual compared to unusual actions in 12-month-old TD infants indicates that the infant motor system is already involved in action processing and forming predictions about observed actions (Stapel et al.,2010). In this study, we implement comparable experimental designs to study potential prediction abnormalities in HR infants.


Action prediction was studied using eye-tracking at 10 months and EEG at 14 months. In both experiments, infants were presented with videos displaying an actor using everyday objects either in usual or unusual ways (Figure 1). In Experiment1, we used eye-tracking to assess the participants’ ability to predict the action goal. In Experiment2, we investigated  the infants’ motor system response to the observed actions using EEG.


Preliminary eye-tracking data from 20 participants (15HR, Experiment1) suggests that HR infants perform similar to controls (Figure 2A). Both groups show predictive eye movements, anticipating more frequently to the correct goal location. Preliminary EEG data from 6 HR participants (Experiment2) suggests motor system activation during both action observation conditions (Figure 2B). Whereas previous research in TD infants suggests enhanced activation for unusual compared to usual actions (Stapel et al.,2010), in the HR group of the current study a conditional difference seems to be small or absent. Data collection is ongoing and results from a complete sample will be presented at the conference.


Our preliminary data suggest that HR infants show similar behavioral signatures of predicting goal-directed actions as controls. In addition, their motor system is activated during action observation, but differential activation for the unusual compared to usual actions currently seems to be small or absent. If these observations hold in the complete sample, the dissociation of behavior and neural processing implies that while HR infants are able to predict simple goal-directed actions behaviorally, the neural processes associated with the formation of predictions are impaired.