Towards Cognitive Mirroring: Identifying stimuli and features to investigate hypo-prior in Autism Spectrum Disorder
Autism Spectrum Disorder (ASD) is characterized not only by problems with social interaction and stereotyped repetitive behaviours, but studies have shown that people with ASD also have problems with perception and catego- rization. Following the hypo-prior account by Pellicano and Burr, I suggest that problems in ASD are caused by a hypo-prior, and propose an experiment in which a computational neural network with an adjustable prior-in uence imitates participants in a human-robot interaction to investigate this idea. In this thesis, I conduct an experiment in which typically developed (TD) par- ticipants draw trajectories they have observed one step at a time, to explore how they respond to the task and the di erent stimuli. The aim is to identify stimuli and features on which TD participants show generalization behaviour. Any identi ed stimuli and features can be used in further experiments, and the behaviour by TD participants can be used as a baseline to compare with behaviour by participants with ASD, to investigate the possible presence of a hypo-prior in ASD. In short, if participants with ASD show less generalization behaviour and more accurate replication on trajectories that TD participants do show generalization behaviour on, that would indicate a hypo-prior in ASD. I discuss the results from the experiment and make several recommendations for further experiments investigating hypo-prior in ASD.
Faculteit der Sociale Wetenschappen