Slow visual attention shifting in autism: could it explain social & behavioral symptoms?

It's common knowledge that kids with ASD don't like to look people in the eye, and tend not to follow others' gaze. A lot of researchers1 have assumed that kids with ASD have a specific face processing deficit and try to trace it back to the brain, particularly the fusiform face area (a part of the network for visual object recognition that is particularly active for most people when recognizing faces). But what if a more general visual processing problem causes them to have trouble following gazes--along with some of the other visual and motor differences seen in autism?

A study by Eric Courchesne illustrates the sort of visual processing problems found in autism. Subjects had to look at a computer screen containing two empty square boxes, with an X between them. Subjects were told to focus on the X and to press a button as soon as they saw that a light had been turned on in one of the boxes. Autistic patients took much longer to register the light than normal subjects. These results suggest that people with autism are slower to notice and react to new visual stimuli.

John Ratey, who describes this study in his book "A User's Guide to the Brain," thinks that problems with the cerebellum might explain autistic people's problems with visual attention. The cerebellum does not just coordinate smooth movements and balance. It also seems to help us put our attention where we want it to be without having to think about it. Intriguingly, in Courchesne's study, patients with damage to the cerebellum, like autistic patients, took much longer to register the light than did normal subjects.

Such problems placing visual attention can lead to problems with reading faces, seeing wholes rather than parts, and gaze following. Eric Courchesne found that while a typically developing baby can shift its attention from the parent's nose to an eye or to the mouth in a fraction of a second, an autistic baby may need as many as five to six seconds to make such a shift.2 If it takes a child five or six seconds to shift his gaze from his mother's nose to her eyes, it will be difficult for him to form a coherent image of his mother's face. He will, instead, store it as a bunch of face-pieces. Thus, face recognition may be impaired--as has been observed in some studies.

The length of time required for autistic people to shift their gaze may also explain why they tend to focus on parts of an object rather than the whole. (Or as Tim Page, a man with Asperger's, put it, "not only did I not see the forest for the trees, I was so intensely distracted that I missed the trees for the species of lichen on their bark.") One can only take in a small amount of information--a part of an object--without moving one's visual attention.

If all one sees is parts, one's concepts may also be highly part-driven and analytical rather than holistic (since we typically abstract concepts based on what we see). Perhaps autistic people's constant need to make some sort of larger sense out of tiny pieces of percepts and concepts explains their tendency to systematize?

It's easy to understand why slow gaze shifts might impair face reading. Facial expressions are holistic, involving most facial features, and they shift very quickly--much faster than an autistic person can move his or her gaze. Thus, even with the utmost attention and care, most of the information in emotional expressions will simply pass an autistic person by. As John Ratey puts it: "Social information, the look on a mother's face...is fleeting; it happens in a moment and is gone. The autistic baby, locked into whatever stimulus has captured his gaze, cannot move his eyes up to his mother's face quickly enough. If the baby is staring at a puppy and the mother smiles, he will miss her smile. By the time he can attend to her face, her expression has changed."

It's not surprising that slow gaze shifts would impair autistic's gaze following abilities. Suppose a mother says to her autistic baby, "look at the dog." A typically developing baby would notice that his mother just spoke, look at her face, notice that she is looking towards the dog, and look at the dog, too, all within a second or two. An autistic baby will look at her mother's mouth, since after all the sound came from there, and a few seconds later will look at her mother's eyes. By then her mother might no longer be looking at the dog but at the child, frustrated that her child didn't immediately respond by following her gaze as expected. Or, when the baby turns to look at her mother, she will see only her mother's ear, because her mother has already turned to look at the dog. Either way, the baby and her mother have not shared joint attention. Not surprisingly, autistic toddlers are unlikely to ask their mothers to look at something they are, since as far as they are concerned, such behavior has never been modeled for them. Unfortunately, gaze following is a vital milestone for social understanding. It teaches young children to recognize, and seek out, common ground--the awareness that you and another person are looking at, and referring to, the same thing. The knowledge that you and another person are looking at the same thing are sometimes thought to be the foundation for more complex understandings of other's knowledge that develop later. More advanced "theory of mind" might not develop in some people with ASD without the foundation laid by gaze following in infancy (or so such prominent researchers as Simon Baron-Cohen and Uta Frith believe).

This slow gaze shifting theory, perhaps most elaborately argued by Dinakhar Wadhwa, might even help explain autistic people's visual strengths. To compensate for slow gaze shifting, autistic people might become highly skilled at visual search. After all, if large amounts of information are passing you by while you try to shift your gaze, it pays to minimize the time needed for search and maximize the accuracy. If you don't look to the right place the first time, shifting focus again will put you even further behind. Indeed, as predicted, autistic people are unusually good at visual search.

Unable to shift their gaze quickly enough to process a whole, autistic people might see a world of pieces. You can imagine how strange and confusing a world of small pieces must seem. For example, if a baby pulls a puppy's tail, "he will miss his mother's frown. Her 'No!' and his tail-pulling will not form a coherent whole in his memory. They are separate pieces of reality, disparate fragments in a life that does not add up."3 People will seem to be yelling at him all the time, for no reason. No wonder autistic people engage in self-soothing behavior to calm down ("stimming"), or develop obsessive routines to make sense of the world and assert some control over it. No wonder some autistic people are so overwhelmed that they go into fight or flight mode and fly into rages where they seem to lose all self-control.

In short, if at least some people with autism have abnormalities in the cerebellum that make them take longer to shift their visual attention, it could explain the variety of social and perceptual symptoms they experience.

1 i.e., the following:
D. Hubl, MD, S. Bolte, PhD, S. Feineis-Matthews, H. Lanfermann, MD, A. Federspiel, PhD, W. Strik, MD, F. Poustka, MD and T. Dierks, MD (2003). Functional imbalance of visual pathways indicates alternative face processing strategies in autism. Neurology, 61: 1232-1237

Kim Dalton, Brendon M Nacewicz, Tom Johnstone, Hillary S Schaefer, Morton Ann Gernsbacher, H H Goldsmith, Andrew L Alexander, and Richard J Davidson (2005). Gaze fixation and the neural circuitry of face processing in autism. Nature Neuroscience, 8:4 519-526

Geraldine Dawson, Sara Jane Webb, and James McPartland (2005). Understanding the nature of face processing impairment in autism: Insights from behavioral and electrophysiological studies. Developmental Neuropsychology, 27(3): 403-424. Can be found here

2 Some of you will ask: but don't we all constantly make minute movements of our eyes in order to "refresh" our visual image, since we literally cease to see things if they don't change? I don't think the Courchesne study refers to this sort of movement, which barely shifts the focus of gaze, if at all. Presumably, an autistic baby's eyes still randomly wobble a bit from central fixation, as with the rest of us. Unlike the rest of us, they are slow to make purposeful, attention-driven gaze shifts.

3 John Ratey, A User's Guide to the Brain


  1. Very interesting, and plausible. btw, this week, Good Morning America reported on a new study on infants and gaze, but it was focused on the binary (follow mother? yes/no), rather than the progressive (what kind of gaze, how long to follow mother's gaze). The goal of that study is to screen for autism in infants.

  2. Hmm, I'll have to check out that study! Do you know who the researchers were?

    I wonder if they look at more complex gaze measurements too, but it just wasn't emphasized...?

    Also...do you know of any exercises/therapies out there to help kids direct their visual attention better, or what they might be?

  3. I can't remember now. Direct visual attention - http://www.therapro.com/Highlighter-Strips-P7109.aspx That's all I can think of.

  4. Cool, looks like a great study skills tool! Thanks for the link.

  5. This is my theory on autism spectrum condition, namely Attentional Shift Theory.


  6. Thank you for sharing this incredibly detailed and elaborate theory! I will add that this is your theory as well.