Note: Most Internet Explorer 8 users encounter issues playing the presentation videos. Please update your browser or use a different one if available.

Numerosity Estimation in Autism: A MEG Study

Thursday, 2 May 2013: 09:00-13:00
Banquet Hall (Kursaal Centre)
E. Meaux1,2, M. J. Taylor1, E. W. Pang1 and M. Batty2, (1)Department of Diagnostic Imaging, Hospital for Sick Children, Toronto, ON, Canada, (2)Pedopsychiatrie, INSERM U930, Tours, France

Our ability to manipulate numerical quantities and especially to estimate and compare them is a fundamental skill, by which we make sense of the external world. In addition to the core triad of symptoms that characterize Autism Spectrum Disorders (ASD), exceptional visual skills are frequently reported, including superior abilities in numerosity estimation. This phenomenon has been addressed by two theories which argue that people with ASD perceive the word differently, the weak central coherence hypothesis (WCC) and the enhanced perceptual function hypothesis (EPF). These hypotheses attribute the superior performance on numerosity estimation displayed by individuals with ASD to their specific processing of local features, rather than focusing on the whole Gestalt. However, superior skills for numerosity estimation could also result from superiority in numerosity process itself.


The current study examined the neural substrates of numerosity estimation, in young adults with and without autism, to disentangle perceptual atypicalities from aspects of numerosity processing itself.


Fourteen adults with ASD diagnosed according to the DSM IV criteria, without mental retardation, were matched with 14 typically developing (TD) adults. Participants were asked to estimate the number of dots (between 80 and 150) arranged either randomly (local information) or in a meaningful pattern (global information) while brain activity was recorded by magnetoencephalography (MEG). Source analyses were performed within the time windows determined by analysis of the sensor data using an event-related beamformer algorithm, first, for all stimuli and then according to the meaningfulness of the stimuli. Only statistically significant source differences are described. 


Behavioural results showed no enhanced numerosity processing abilities in ASD participants. However, analyses of MEG data revealed significant differences in brain activation during this process between ASD and TD subjects. At an early stage of numerosity estimation (80-120ms), significant differences in source amplitude of activation (TD > ASD) were found in the visual areas, in the lingual gyrus and cuneus.  From 120 to 400ms, group differences extended to first temporal and then parietal regions, showing greater activation in TD subjects in the precuneus, middle temporal gyrus, superior parietal lobule and inferior parietal lobule. Finally, after 400ms, a source was found in the superior frontal gyrus in TD but not in ASD participants. Moreover, in TD participants, the activation in the temporal areas between 120-290ms was sensitive to the global arrangement of the dots (global>local). This effect was not observed in ASD participants for whom an inverted effect (local>global) appeared later in this brain region (after 290ms).


Brain responses during numerosity estimation are atypical in autism.  We suggest that the early occipital differences could be explained by sensory abnormalities already observed in ASD.  Differences in temporal regions could be linked to perceptual atypicalities, specifically, differences in sensitivity to global perception. Interestingly, our MEG data demonstrated that parietal and frontal activations known to be involved in numerosity processes also are affected in ASD. Taken together, these results suggest that the atypical number estimation ability seen in ASD may be due to “overlapping impairments” in sensory, perceptual and numerosity processing.

| More