19792
Ensemble Perception of Emotions in Children with Autism
Ensemble perception, the ability to rapidly and automatically assess the summary or ‘gist’ of large amounts of information presented in visual scenes, is crucial for navigating an inherently complex world (Haberman & Whitney, 2011). Given the processing limitations of the human brain, it is more efficient to lose representations of individual elements of visual scenes in favour of a concise summary representation. Ensemble perception is employed for low-level visual information (e.g. size, texture), as well as high-level social stimuli (e.g., emotion) and is available early in development.
A recent theoretical account proposed that individuals with autism present limitations in their ability to maintain summary statistics for the recent history of sensory input, which might explain their unique perceptual experience. We therefore predicted that ensemble perception in children with autism might be compromised compared to typically developing children.
Objectives:
We investigated ensemble perception of emotions in children with autism and typically developing children. We also employed eye-tracking methodologies to investigate cognitive mechanisms underlying ensemble perception of emotions.
Methods:
We have so far tested 17 children with autism and 17 typically developing children, aged between 6 and 14 years, and individually matched on age and on verbal and reasoning abilities. Each child received three child-friendly tasks (see Figure A): 1) an average emotion discrimination task, assessing their ability to judge the average happiness of a set of ‘morphs’ (variations of the same face) drawn from a continuum of emotional faces, from happy to sad; 2) a non-average happiness discrimination task, evaluating baseline emotion discrimination; and 3) a face identification task, estimating children’s ability to identify morphs previously presented to them.
We predicted that children with autism should present difficulties in the average emotion discrimination task, where performance relies upon the efficient use of summary statistics for facial emotions, but no difficulties in non-average emotion discrimination. We also predicted that children with autism, who rely upon detailed representations of individual items rather than condensed summary statistics representations, might present better performance than typical children in the face identification task.
Results:
Analyses showed that children with autism were just as precise and as fast as typical children in the average and non-average emotion discrimination tasks. Children with autism, however, presented lower accuracy than typical children in the face identification task, suggestive of more general weaknesses in facial emotion processing. On eye-movement variables, the two groups were indistinguishable in terms of the average number of fixations, the number of morphs sampled per trial, and their preferences in their first fixation. In their last fixation, however, children with autism showed a looking preference for happier rather than sadder morphs.
Conclusions:
We show here that, contrary to our predictions, autistic children’s ability for ensemble perception of emotions is similar to that of typically developing children. We further demonstrate that, while children with autism show no difficulties in non-average emotion discrimination, they present certain weaknesses in face identification, as well as a pronounced preference for sampling the happier rather than the sadder morphs in their last fixation.