Autism spectrum disorder, experts know, is sometimes treatable. In one good-size study, about 9 percent of kids with a diagnosis at age 3 had shed it entirely at age 19, and they were more likely to have done so if they had received behavioral interventions geared at helping teach them certain skills and behaviors that come naturally to kids without an ASD diagnosis. And even when the disorder remains, treatment can make the difference between young adults with severe ASD being able to live relatively independent lives and having to be institutionalized. What remains something of a mystery, as The New York Times Magazine highlighted recently, is why treatment sometimes work and sometimes doesn’t.

One thing that’s clear, though, is that the earlier these sorts of interventions start, the better. The human brain gradually gets less plastic over time, and it becomes harder to change certain ingrained behavioral patterns. The challenge, though, is that it’s tough to diagnose autism from a very early age — there are no known biomarkers that can tip clinicians off to the presence of the disorder, forcing them to rely instead on behavioral signals that can be tough to decipher when it comes to very young infants.

 “The younger infants are, the more variation there is in development of speech and language,” said Sally Rogers, an autism researcher at UC Davis. Since it is hard to distinguish genuine symptoms from these natural variations at a young age, she said that most doctors will only provide an autism diagnosis — and a provisional one, at that — at 12 to 18 months, at the youngest.

In certain situations, this can lead to a “wait and see” approach that may delay behavioral interventions (for those parents with the resources to seek those interventions out). And that’s a problem, because for young kids with untreated autism, time is the enemy. “The longer children go in a lifestyle in which they're not interacting with others, in which they’re kind of removed from family interaction, the harder it is to bring them back into the circle of family social interaction and communications,” said Rogers.

Rogers and her colleagues wanted to see if it would be possible to identify at-risk infants at an earlier age, and if an earlier-onset treatment program would help blunt the disorder’s impact on them. So for a new study just published in the Journal of Autism and Developmental Disorders, they found seven kids between 6 and 15 months of age who were already showing significant signs of autism, had their parents participate in a “12-week, low-intensity treatment,” and then tracked their symptoms and development through 36 months, comparing the numbers between this group and a comparison group of symptomatic infants who didn’t get the treatment.

The treatment primarily involved coaching parents into how to best encourage their kids to engage in certain sorts of behaviors. Most kids will naturally pick up on skills like maintaining eye contact when they’re engaging with someone, for example, but parents of autistic kids are often stymied and frustrated by the fact that their children relate to them and the world differently. “Those are things that these babies don’t have a good sense of,” said Rogers.

With the proper coaching, though, many of them can eventually grasp and internalize these skills. “They learn the power of it, and then it becomes a very useful part of their repertoire as well,” said Rogers. At the most basic level, kids can realize that to get what they want, weather a cookie or a parents’ attention, it’s more effective to engage in a certain sort of behavior. The key is to teach and reinforce those behaviors.

Rogers and her colleagues focused on six “target symptoms” in the pilot study. To improve the problem of “Decreasing gaze, social interest and engagement,” for example, parents were coached to “Position self and child for maximal face to face orientation and provide object and social games that follow infant preferences, delivered to maximize infant attention and pause for infant turns.”

It appears to have worked. At 3 years old, the researchers write, the treatment group “had much lower rates” of autism symptoms than the comparison group; five of the seven, in fact, no longer qualified for an ASD diagnosis (or any other development disability) at all, while most of the kids in the comparison continued along the same path, maintaining their autism diagnoses at this age.

As the researchers acknowledge, this was a small pilot study rather than a large, randomized, controlled trial, and Rogers was quick to caution against over-interpreting this one set of results. “You don’t know what would have happened to these seven kids without treatment,” she said. “We know what happened with treatment, and it requires a well-done randomized controlled trial to understand what would have happened without it.” Rogers said she is seeking funding for a larger, better-controlled study.

Beyond this study, though, Rogers said that it’s encouraging that diagnostic procedures for autism are improving. Whether or not researchers can uncover biological markers of the disorder, she said, “I’m quite optimistic that we’ll be able to diagnose autism earlier.” And that could make a big difference, because “Intervening sooner, when symptoms are less severe, and before so much change has gone on [in a child’s brain] is liable to save money, pain, and prevent disability downstream.”