Atypical Categorical Perception in Autism: Autonomy of Discrimination?

Abstract

A diminished top-down influence has been proposed in autism, to account for enhanced performance in low-level perceptual tasks. Applied to perceptual categorization, this hypothesis predicts a diminished influence of category on discrimination. In order to test this hypothesis, we compared categorical perception in 16 individuals with and 16 individuals without high-functioning autism. While participants with and without autism displayed a typical classification curve, there was no facilitation of discrimination near the category boundary in the autism group. The absence of influence of categorical knowledge on discrimination suggests an increased autonomy of low-level perceptual processes in autism, in the form of a reduced top-down influence from categories toward discrimination.

Is It a Bird? Is It a Plane? Category Use in Problem-solving in Children with Autism Spectrum Disorders

“…Research on basic learning of new categories has indicated potential differences in attention to perceptual similarities (Plaisted et al 1998) and generation of prototypes (Klinger and Dawson 2001) by individuals with ASD. However, subsequent findings have demonstrated largely intact learning of basic shape categories (Bott et al 2006;Soulieres et al 2007) and unimpaired recognition of prototype exemplars (Molesworth et al 2005(Molesworth et al , 2008. Similar results have arisen from studies examining basic sorting (Tager-Flusberg 1985;Ungerer and Sigman 1987), category recognition (Geurts et al 2004;Verté et al 2005) and free sequencing abilities (McGonigleChalmers and Alderson-Day 2010): where ASD participants are asked to sort or match according to basic and familiar categories, they perform as well as typicallydeveloping controls.…”

Cognitive versatility in autism cannot be reduced to a deficit

“…We propose instead that this illustrates autistics' greater autonomy of low-level perceptual processes from top-down influence, also found in perceptual categorization (Soulières, Mottron, Saumier, & Larochelle, 2007).…”

Free Classification as a Window on Executive Functioning in Autism Spectrum Disorders

“…The question posed at the start of this experiment was whether ASD participants of normal intelligence would show weaknesses in a task that is most effectively solved by the spontaneous imposition of a classification strategy, given previous claims that they show general weaknesses in planning (Pennington et al 1997), in top-down processes (Soulieres et al 2007) and spontaneity of categorical organization Minshew et al 2002). The main measures were effectiveness of exhaustive visual search of a complex changing visual array (sequential control) and sequential grouping of items according to a self-imposed rule (classification).…”

“…There is no suggestion from this task that the children with ASD could not use a top-down approach to this problem of perceptual search. Success on the task where it is based on a consistent classifying principle may not require planning of the full sequence in advance but it certainly could not be effective if there was ''a disconnection of the feedback from highlevel (category) to lower level visual processes'' (Soulieres et al 2007). One reason why this conclusion was reached by Soulieres et al may be that a discrimination task does not call for classification as an adaptive principle to the same extent as a search task with clear on-line executive demands.…”

“…Individuals at the high functioning end of the autistic spectrum have been found to have the have the necessary discriminatory sills to establish category boundaries based on shape or size when presented with unfamiliar stimuli (Bott et al 2006;Molesworth et al 2005), but some investigators have argued that this does not result in the same sort of category sensitivity as found in the normal population. Soulieres et al (2007) presented ellipses to individuals with and without high functioning autism and required them either to categorise them as either thin or wide or to discriminate between adjacent examples along this continuum. They found nearly identical classification curves in both groups, but whereas there was facilitation of discrimination at the category boundary in the non-autistic group there was no such effect for the participants with autism.…”