18632
Who Believes in (Bayesian) Ghosts? a Study of Interpersonal Predictive Coding in High-Functioning Autism

Friday, May 15, 2015: 5:30 PM-7:00 PM
Imperial Ballroom (Grand America Hotel)
T. von der Luehe1, V. Manera2, I. Barisic3, C. Becchio4, K. Vogeley5 and L. Schilbach5, (1)Department of Psychiatry, University Hospital Cologne, Cologne, Germany, (2)EA CoBTeK Cognition Behaviour Technology, University of Nice Sophia Antipolis,, Nice, France, (3)Cognitive Science Department, ETH Zürich, Zürich, Switzerland, (4)Department of Psychology, University of Turin, Torino, Italy, (5)Department of Psychiatry and Psychotherapy, University Hospital Cologne, Cologne, Germany
Background: Communicative actions reveal information about an actor, but they also help to predict the presence and behavior of a second agent. This has been referred to as “interpersonal predictive coding”. High-functioning autism (HFA) is characterized by impairments in social interaction, which are assumed to be related to an inability of automatically responding to social cues, while more explicit capacities to relate to others  may be intact. 

Objectives: This study was conducted to examine the presence of interpersonal predictive coding in HFA.

Methods: 16 adults with HFA and 16 matched controls (HC) observed point-light displays on an eye-tracker monitor. Stimuli belonging to the communicative (COM) condition showed an agent (A) performing a communicative gesture toward a second agent (B) who responded according to it. In the individual (IND) condition, agent A´s communicative actions were replaced by non-communicative actions. Using a simultaneous masking detection task, we asked participants to report the presence of the second agent B.  

Results: A mixed repeated measure ANOVA revealed a significant main effect of condition with higher sensitivity in the COM than in the IND condition and a significant interaction effect between condition and group. A simple effect analysis demonstrated a significant effect of condition in the HC group with higher sensitivity in the COM condition, while no such effect was observed in the HFA group. Analysis of the gaze data confirmed that both HC and HFA always looked at both agents. 

Conclusions: The present study replicates and extends previous findings of a “second agent effect”, which demonstrates that communicative actions of one agent may help to predict the presence and behavior of a second agent. This has been related to prior expectations that are built up through participation in social interactions. Furthermore, our results demonstrate that individuals with HFA are able to recognize and label communicative actions presented by means of point- light displays, but do not automatically respond to them in such a way that it would facilitate the detection of a reaction of a possible respondent to a given communicative gesture.

See more of: Social Cognition and Social Behavior
See more of: Social Cognition and Social Behavior