Sibling Studies Shed New Light on Autism

While many parents of children with autism want to know more about the possible role of environmental factors in the development of the condition, scientific studies show that perhaps as much as 90 percent of the risk comes from genes. Autism is highly heritable, an observation underscored by the fact that, among the younger siblings of kids with autism, roughly 10 percent will also develop the disorder. But what about the other 90 percent? "They're the ones that interest us," says Karen Dobkins, professor of psychology at the University of California, San Diego. Their quirky behavior is not enough to qualify for a diagnosis of autism, but "they still show atypical patterns of social interaction and communication," she says. What's going on in their brains? And can it yield any clues that might help lead to earlier diagnosis for those with the full-blown autism?

This week, Dobkins and her colleague Leslie Carver unveiled two joint studies at the International Meeting for Autism Research that, in different ways, help to chip away at those questions. In one of the studies, the pair looked at social behaviors; in the other, at sensory perceptions. But both are highly revealing.

The first of the studies examined a behavior known as "social referencing"—or the tendency to look to others to help read meaning into an unfamiliar event. Dobkins explains it this way. "Say, you've never flown in an airplane before and it starts bouncing around," she says. You're not sure if this means you've hit a rough patch or the plane is about to nosedive. "So you look around at your fellow passengers," she says. "If they're screaming, you think, 'This is bad. I have to start screaming, too'."  This behavior is very common. Children usually begin doing it around 12 months of age, checking for mom's smile of approval before investigating a new caterpillar in the park, for example.

But in a study of 20 "high-risk" toddlers (18-month-old children with an autistic older sibling), Carver and Dobkins showed that the social-referencing behavior of these children was very different from that of a control group of 16 children the same age with no autism in the family. They put each child into a room with a parent, a researcher (who was a stranger) and a series of novel mechanical toys—a spider with eyes that light up and flash, a raccoon that pushes a ball around the room with its nose, and a dinosaur that walks in circles and beeps. Both groups stared intently at the toys for a good 30 seconds or so before looking up at the parent for cues as to how to interpret these puzzling creatures. But the children with no autism in the family looked at the stranger 2.5 times more often than the high-risk children did. Why? Carver speculates: "The high-risk toddlers might be comfortable with their moms, but less comfortable with a complete stranger."

Even more intriguing was their response to the cues from the adults. When the stranger registered delight or disgust or showed no particular emotion at all, the children in the control group responded accordingly, reacting more positively to a toy "tagged" with a positive emotion, more negatively to one that drew negative responses from the adult stranger. But the high-risk toddlers simply ignored the cues. "There was no match of emotions," says Carver. "Again, we think that one of the things in autism is being able to read emotions, know what they mean and apply them to social situations."

That's not all the study revealed. The toddlers all wore caps with tiny electrodes embedded in them to record brain waves—and those, too, showed differences. The control-group children had three different types of brain-wave responses, depending on the emotions the researcher displayed—positive, negative or neutral. But the high-risk children showed identical patterns of brain activity in response to all three. This is particularly interesting, says Carver, because these are not children that have been identified as quirky or problematic or anything but normal. Yet there are clearly differences there. "We often note that adult family members of kids with autism don't like big crowds or avoid parties or prefer to work with data than people," she says. None of these behaviors constitutes a diagnosis. "But it shows social inhibition"—and that may run in families.

At this stage, Carver and Dobkins are unable to interpret the data finely enough to predict which of these high-risk children will be in the unlucky 10 percent who develop autism. But the hope is that by continuing to follow these children out to an age where a diagnosis is possible, they will then be able to look back at the earlier data and see if certain characteristic patterns were indicators of trouble ahead. Currently, a child can be diagnosed by the age of 3—although many are not diagnosed until about 5. This line of inquiry could conceivably lead to the ability to diagnose children as early as 18 months.

There are other potential early clues beyond social referencing. In the second of the studies, Dobkins and Carver looked at responses to one type of sensory information—"low-level visual sensitivity," or the ability simply to detect contrast between light and dark. "Children with autism are often more sensitive to sensory stimuli," says Dobkins. "Noises are too loud. Things they touch feel weird. Visual stimuli are too great. We began thinking, maybe we can find traits that are simpler than social interaction and communication, but still closely tied to the genes underlying the disorder." These might be even easier to detect earlier.

Dobkins and Carver took 13 high-risk infants at 6 months of age and 26 control infants from families free of autism and showed them a split computer screen. One side of the screen was blank. The other showed a series of blurry stripes against a background providing low, medium or high contrast. Because babies prefer to look at an image rather than a blank screen, the experimenter would follow the child's gaze to determine the threshold at which he or she could begin to detect the emerging contrast levels. The high-risk infants were nearly twice as sensitive as control infants. Since the brain structures used for low-level visual perception feed into higher-level structures used for seeing and recognizing faces, these early, low-level perceptual differences may be indicative of changes that will come later in higher-level processing. The results, say Dobkins and Carver, are further evidence that "autism is a biological-developmental disorder" and that "to understand it, we will need to understand the development of the brain across time."

Ultimately, there is no single brain region that will tell the whole story, and other researchers at the conference are focusing on different areas. Mirella Dapretto of UCLA presented a study on the response of so-called "mirror neurons"—those brain cells that help you mimic another person's body movements and facial expressions and, in doing so, feel empathy by figuratively putting yourself in that person's shoes. Mirror neurons help explain why one crying toddler can set off others nearby, starting one loud, communal wail. Dapretto began with interviews and questionnaires to gauge how readily 12 children aged 9 to 16 with high-functioning autism imitated others and empathized with them. Using sophisticated brain imaging, she then measured activity in the mirror-neuron brain areas of these children as they looked at pictures of faces that were happy, sad, angry, fearful or neutral. Those with the lowest empathy scores also had the least activity in the mirror-neuron areas. That doesn't mean these children are helpless in social situations. Because they're high-functioning, Dapretto says, they have learned to distinguish different expressions—though they may do so through a series of deliberate calculations rather than the automatic way a normal child would. The interesting part is that the difference shows up as a different pattern of brain activity, even though the end result is the same—being able to recognize the expressions.

Taken together, says Carver, these studies show that "if you want a solid picture, you will need to look at the development of the whole brain and interaction between affected areas." When researchers are able to do that, they will be much closer to a full understanding of this tragic disorder.

Join the Discussion