Filling-in in Autism: a High-Density Electrical Mapping Study of Visual Object Binding Mechanisms

Background:

As well as the obvious clinical behavioral manifestations of Autism, it is now apparent that differences in very basic sensory-perceptual processing of environmental inputs may also be a core feature of Autistic Spectrum Disorder (ASD).  For instance, early visual processing of object boundaries in late teens and young adults with ASD appears to be atypical (Vandenbroucke et al., 2008).   In viewing occluded or partially fragmented objects, a typical adult brain can readily “fill-in” the missing information. That is, the brain is capable of binding disparate “local” elements into unified “global” wholes. One excellent means of studying this “filling-in” phenomenon involves the use of so-called illusory contour (IC) stimuli.   Pacman-shaped disks are oriented with their “mouths” open and pointed towards each other such that they induce the perception of the contours of an illusory shape, even though no such contours physically exist. Event-related potential studies (e.g. Murray et al., 2002, Foxe et al., 2005) have extensively investigated this phenomenon and established a set of robust dependent measures that represent binding processes within the ventral visual stream.  In particular, an early automatic contour-definition ERP component occurs between about 110-180 ms, and this has been termed the IC effect.

Objectives:

To investigate object-binding processes in persons on the autistic spectrum, using a well-established metric of ventral visual stream processing, the IC effect.  In particular, we manipulated the spatial disparity of the Pacman inducers on the premise that binding across greater spatial scales might be progressively more dysfunctional in ASD.

Methods:

ERPs to IC stimuli were compared between typical and ASD children, aged 7-15, and also in a normal healthy adult cohort. IC inducing stimuli were presented at three progressively greater retinal eccentricities (4, 7 and 10 degrees respectively).  The IC effect was measured for each eccentricity.  Behavioral measures of global/local processing styles (Children’s Embedded Figures Test (Witkin et al, 1971)) were also analyzed to assess any contributions these differences might make to group differences.

Results:

Preliminary results from typical adults replicate Murray et al’s (2002) IC effect, and show that this effect is evident at all three eccentricities. This pattern was replicated in typically developing children. Preliminary data from ASD children, however, suggest a less robust IC effect, even for the smallest degree of retinal eccentricity.

Conclusions:

If this pattern of results holds as our sample of ASD and TD children increases, it would point to a basic deficit in automatic object-binding mechanisms in ASD.  The interaction between retinal eccentricity of the stimuli and diagnosis could begin to reveal specific ways in which automatic processes of people on the spectrum are like and unlike their typical counterparts.