Is Face Recognition Selectively Impaired In Children with ASD?

Background: Impairments in face memory are widely thought to be at the core of autism. However, only a few studies have compared face recognition performance in autism with performance on visual recognition of other non-face visual objects (e.g. cars). Thus, it is still an open question if the recognition impairment is specific for faces, or reflects a more general deficit in object memory per se. Further, few if any studies have asked whether any deficit in face recognition in autism arises because of a difficulty in perceptual discrimination of faces, or a difficulty with encoding faces into memory.

Objectives: In the present study, we asked if face recognition is impaired in those with autism compared to control subjects, and more importantly i) whether any such deficit is specific to faces, and ii) whether any such deficit reflects a problem in perceptual discrimination, memory encoding, or both .

Methods: So far we have tested 51 typically developing children (age range: 5 to 10 years), 11 children with ASD (age range: 5 to 10 years) and 32 healthy adults on a newly developed task battery that measures both perceptual discrimination thresholds and memory in each of four different stimulus classes: faces, bodies, cars, and scenes/houses. Crucially, the same stimuli were used for the perception and memory tasks so that any differences in perceptual versus memory performance cannot be due to stimulus differences. In the memory task, subjects studied ten items from each category, then chose which of two items was the previously studied one. For the discrimination task, pairs of similar faces, cars, bodies (without heads), and houses/scenes were morphed with each other to create a continuum of morph levels.  Participants saw a single un-morphed exemplar and then immediately thereafter chose which of two stimuli they had just seen: an exact match of the sample or a stimulus at a particular distance from it along the morph continuum. Using the QUEST staircase procedure, morph levels were gradually adjusted to determine the threshold stimulus distance for which participants could discriminate between the two pictures correctly 75% of the time.  Discrimination thresholds were ascertained for each stimulus class separately.

Results: Our preliminary results suggest that children with ASD perform worse than typically developing children in both face memory and face discrimination tasks, while performance for the car, body, and scene memory and discrimination tests is similar for children with ASD and typically developing children.

Conclusions: We provide preliminary evidence that both face discimination and face memory are selectively impaired in children with ASD.