According to new Yale research, 2-year-olds with autism lack a key social mechanism that normally allows non-autistic children to recognize human movement.

Instead, autistic children focus on the physical aspects of motion and fail to pick up on important social information, according to the study, which was published in the March 29 online edition of Nature. The findings provide new clues about the origins of autism that researchers hope will help with early diagnosis and treatment.

While non-autistic children have innate tendencies to pick up on social cues that help them maneuver social situations, autistic children struggle with this, said Ami Klin, director of the Autism Program at the Yale Child Study Center.

“Infants are born very fragile,” he said. “It makes sense that over the course of many thousands of years, there are social adaptive mechanisms that children are born with — predispositions that push babies toward their caregivers.”

Klin, who co-authored the study with Yale research scientist Warren Jones and three other colleagues, said autistic children do not show any signs of preferential attention to human motion. While non-autistic children pay careful attention to socially motivated movements such as changes in facial expressions, autistic children focus on the physical — namely, motions that are synchronized with sounds.

“They are interacting with people but paying attention to physicality of the mouth moving rather than the social content of eyes,” Klin said. “They watch the face without experiencing those faces as people.”

Klin said he serendipitously observed this effect when a 15-month-old autistic patient showed no preference for any of the light displays, except for when pat-a-cake was playing. After some thought, Klin and Jones realized that in pat-a-cake, the clapping noise coincided with a collision of points of light. The audio-visual synchronization intrigued 90 percent of the autistic 2-year-olds tested.

The study employed video-game technology to animate points of light in sequences of situations babies encounter frequently, such as someone playing peek-a-boo or pat-a-cake. Researchers showed 2-year-olds a model of the situation using animated spots of light to represent human figures, and allowed the sound to play normally. Alongside this, a transposed version of the sequence was shown in which the spots of light were inverted, no longer resembling human motion. Using eye-tracking technology, researchers determined if toddlers held a preference for either image by measuring how often the toddlers looked at either display.

As hypothesized, 2-year-olds with autism had no preference, while other groups — both typically developed 2-year-olds and those with cognitive impairments other than autism — showed strong preferences for the social light display, Klin said.

“They are exposed to people but rather than seeking the social nature of what you’re experiencing,” he said, “they experience physical contingencies.”

Klin said he hopes these results will begin to answer lingering questions about the neurobiology of autism, adding that it could lead to earlier diagnoses and hence, the probability of effective treatments.

“Autism is a neural developmental disorder, the most strongly genetic of them all,” Klin said. “We hope to capitalize on early neuroplasticity and actually make amends into natural course of the condition, optimizing the child’s outcome.”

The study was funded by the National Institutes of Health’s National Institute of Mental Health.