About Harvard Health Publications

Subscription/Purchasing Information

Contact Us

Harvard Health Publications Home Page

 

Harvard Mental Health Letter
January 2003

 

How autism looks

Psychologists at Yale using new technology have found a way to scrutinize human faces through the eyes of autistic people. The results promise to provide insight into the way autism cripples the capacity to grasp the meaning of emotional reactions and social situations.

The experiments were inspired by the familiar observation that autistic people don’t make eye contact. To monitor the eye movements of their subjects, the researchers used an ingenious two-camera tracking device originally developed for fighter pilots. Young autistic men described as "high-functioning” — they had IQs in the normal range and adequate language — watched brief clips from the 1966 movie, "Who’s Afraid of Virginia Woolf?” starring Richard Burton and Elizabeth Taylor. This film was chosen because its drama is concentrated in the talk and facial expressions of four characters, with little distracting background activity or scene shifting. The autistic men’s eye movements were compared with those of controls matched for age, sex, and verbal IQ.

The autistic view

The results vividly showed how little of the emotional and social context of the drama the autistic men were absorbing. The controls looked mainly at the region around the actors’ eyes, where most of us direct our gaze when we’re trying to judge the mood, thoughts, and intentions of speakers. The autistic men looked mostly at mouths and bodies and at objects in the room. On average, they directed their attention at eyes 2.5 times less, at mouths twice as long, and at bodies and inanimate objects 2.5 times longer. There was almost no overlap; no control subjects looked less at eyes or more at mouths than any of the autistic men. As a corollary, the controls were much more likely to look at (and presumably notice the reactions of) characters who were not speaking.

The way the viewers’ gaze changed direction from moment to moment was also revealing. In one scene, for example, a character points to a picture hanging on the wall behind him and asks who painted it. Control subjects followed the pointing finger with their eyes; autistic men often did not, but responded only to the words that came a few seconds later. As the conversation continued, the autistic viewers’ eyes tended to move from one picture to another along the wall, as though they were not sure which one the characters were talking about. They usually had no trouble afterward saying what the pointing finger meant, but they hadn’t been able to use that knowledge to comprehend the situation presented in the film.

A different approach

Advanced technology was not the only novelty in this study. It also differed from most earlier experimental research in its approach to understanding the social competence of autistic people. Experiments testing their ability to understand and respond to social cues usually involve set problems with definite solutions, often stated in words. Intelligent autistic persons sometimes do fairly well on these tests; for example, they can work out the meaning of facial expressions if they’re told what to look for. What sets them apart is the way they perform the task — the process rather than the result. In unusual situations, their abnormal processing can even give them an advantage; they sometimes read expressions equally well whether a face is upside down or right side up.

But in the complicated social situations of real life, we’re not told in advance what the problem is. We have to extract meanings for ourselves, deciding what’s relevant among many rapidly passing cues, most of which are not verbal. That’s more like the situation of viewers watching a movie.

Researchers who try to explain the lack of capacity for empathy and communication among people with autism have often looked for an underlying neuropsychological deficiency in attention, perception, or language. They may speak of poor "central coherence” — difficulty in integrating bits of information into a meaningful whole — or lack of executive function, the skill that allows us to make plans and adjust them to meet goals. They may say that autistic people are missing a "theory of mind” — the ability to sense intuitively, without using explicit rules or making a conscious effort, the beliefs, desires, and purposes of others, and to use those intuitions in understanding behavior. But these deficiencies are not confined to autism, and attempts at experimental measurement of the capacity for central coherence or a theory of mind have not thrown much light on the social capacities of autistic persons.

Unexpected finding

To their surprise, the Yale experimenters found that among the autistic subjects, higher IQ and better social adjustment (as measured by a standard questionnaire) were correlated not with more time looking at the actors’ eyes but with more time looking at their mouths, where the words were coming from. (The longer an autistic man looked at objects, the lower his social competence.) The psychologists suggest that the autistic men may have been unable to learn anything from the actors’ eyes, not simply because of a neuropsychological deficiency that made the understanding of facial expressions difficult, but because they had rarely looked at people’s eyes and lacked the experience necessary to interpret their expressions. They were comparatively expert at reading mouth movements and understanding the content of speech.

An implication is that the relationship between neuropsychological deficiencies and social development in autistic people may go both ways. Failures in social learning may be more significant for their adaptation than most researchers have thought. Early social engagement could facilitate the development of the characteristics described by the terms central coherence, executive function, and theory of mind.

In the future, the researchers mean to study eye tracking in autistic children and in autistic people who are either retarded or have milder conditions like Asperger’s syndrome. They are interested in the relationship between poor social competence and the tendency to look at objects rather than faces. They also want to see how brain scan images change as a person’s gaze changes its focus. And they hope that some day similar experiments with infants will help in the study of autism at a stage of life where it’s now often difficult to detect and describe.