Autism as a disorder of prediction: Empirical support for a new hypothesis
Margaret Kjelgaard, Ph.D., CCC-SLP
We have recently proposed a theory that several aspects of autism may be explained as arising from an impairment in predictive abilities (Sinha et al., 2014). We are now embarking on a multi-faceted program to empirically test the theory. I shall describe results from two sets of studies, the first focusing on sensory habituation and the second on interactions with dynamic objects.
Several studies have shown that we rapidly habituate to predictable stimuli. Endogenously compromised prediction would be expected to lead to reduced habituation. In a series of experiments, we have found that habituation indeed appears to be impaired in those with ASD for both peripheral (GSR) and central (MEG) measurements.
Another area of behavior that is anecdotally documented to be difficult in ASD, and that requires predictive skills, is the ability to interact with dynamic objects in the environment. I shall present results from a pilot study designed to test this issue systematically using computer vision methods to track anticipatory movements during ball catching.
Taken together, these results provide preliminary support to our theory. Though far from a definitive validation, they do indicate that further empirical testing of the theory may be a worthwhile pursuit.