In Yale autism research, the eyes have it

Attaining a perplexing result in an experiment can be frustrating for scientists, but sometimes unexpected findings lead researchers to a place of far greater clarity. Such is the case with a recent Yale study of children with autism spectrum disorder (ASD), which was inspired by a puzzling result obtained in earlier experiments. Published in May in the journal Nature, the study provides researchers with rich new ground to explore in the quest to understand this mysterious and socially debilitating disorder.

Investigators at the School of Medicine’s Child Study Center (CSC) have long been in the vanguard of research on ASD, a developmental disability that emerges in early childhood marked by deficits in social interaction, problems with verbal and nonverbal communication, and repetitive, stereotyped behaviors.

Early behavioral and educational interventions are known to significantly improve the lives of children with ASD, but limitations of present diagnostic techniques mean that many children are not diagnosed until age 3 or after. For the past decade, Yale researchers led by Ami Klin, Ph.D., Harris Professor of Child Psychology and Psychiatry and director of the CSC’s Autism Program, have strived to change this picture, using innovative technologies to search for signs of ASD at its very earliest stages.

With support from the Simons Foundation, the National Institutes of Health, and the advocacy group Autism Speaks, Klin and Warren Jones, Ph.D., a CSC neuroscientist, have pioneered the use of eye-tracking technology, which allows researchers to monitor precisely where a person is looking at any given time, in autism research. They have developed a novel apparatus that allows them to track eye movements remotely, while concealed from research subjects, which is particularly useful when working with infants and toddlers.

Using eye-tracking, Klin and Jones have discovered that children and adults with autism view the world in quite different ways than typically developing subjects, often ignoring important sources of information that could help them build bridges to the social world.

One realm in which eye-tracking reveals sharp differences between children with ASD and other children is biological motion, a term scientists use to describe the distinctive manner in which living things move. A sensitivity to and preference for viewing biological motion over other types of movements (of machinery, for example) can be observed in a broad range of species, from newly hatched chicks to monkeys, and it can be demonstrated in human infants as young as 2 days old. It is believed that this inclination is widespread because it aids familial and social bonding, so Klin and Jones surmised that it may be impaired in children with ASD.

This has proved to be the case. Klin and Jones devised point-light animations—depictions in which joints and other important body parts are represented by single dots—of people playing children’s games such as peek-a-boo or pat-a-cake. When two versions of these animations, one upright and the other upside-down and backward, are presented side-by-side on a screen, typically developing children direct their gaze significantly more toward the properly presented animations, but children with ASD exhibit no preference.

However, as Klin and Jones reported in 2008 in the journal Developmental Science, one child they studied, a 15-month-old girl with ASD, unexpectedly exhibited a normative pattern when viewing the pat-a-cake animation, choosing to view the conventional upright animation over the inverted, backward version more than 90 percent of the time. “Then we sat back and thought we should be adventurous,” Klin told Nature’s news staff in an interview, “in order to learn the profound lesson this little girl was teaching us.”

All five of the animations Klin and Jones had created had sound tracks with human voices, but they soon realized that the pat-a-cake movie was unique, because it featured a distinctive sound—clapping—that occurred whenever dots representing the hands came together. They called this phenomenon audiovisual synchrony, or AVS.

In the new study published in Nature, Klin, Jones, and Associate Research Scientist Gordon J. Ramsay, Ph.D.—along with Philip C. Gorrindo and David J. Lin, now medical students at Harvard Medical School and Vanderbilt University, respectively—analyzed eye-tracking data for just the pat-a-cake animation, and discovered that, like the 15-month-old girl, children with ASD as a group showed a strong preference for the conventional, upright version of that animation.

Convinced of a strong correlation between ASD and a heightened sensitivity to synchronized sounds and images, the team performed finer-grained mathematical analyses of all five animations to identify all instances of AVS, and they found that over 90 percent of the changes in gaze direction among toddlers with ASD occurred in tandem with instances of AVS, even though most of these sound/motion correlations were far subtler than the hand-clapping seen in the pat-a-cake animation.

As a final test of the idea that AVS may be more salient to children with ASD than biological motion, the researchers designed two new point-light animations in which instances of AVS were plentiful. They then created a mathematical model that predicted, based on the presence of these occurences of AVS, where children with ASD would look if they were shown these new animations paired with upside-down versions as before.

When the experiment was carried out with a new group of 2-year-olds with ASD, the AVS model was spectacularly successful, accurately predicting shifts in gaze by these children over 90 percent of the time.

The new AVS findings dovetail nicely with other discoveries Klin and Jones have made in eye-tracking experiments. For example, they have found that when children and adults with ASD view films of people speaking, they tend to look mostly at the speaker’s mouth, whereas typically developing subjects focus mostly on the eyes. Since the mouth is the region of the face with the greatest correspondence between movement and sound, this behavior is consistent with the new animation data.

With their research on biological motion and other eye-tracking work, Klin and Jones believe they may be homing in on one of the most elusive goals of autism research: a test that could find behavioral signatures of autism at an early age, allowing parents and teachers to put interventions in place that can help children with ASD to more fully integrate into human social interaction. As Klin told Nature’s news staff, “We want to come up with a behavior assay that will predict vulnerabilities for autism in the first year, if not months, of life.”