Soon after Noah turned a year old, his parents, Leslie and Paul, noticed something was not quite right with their son. At 10 months Noah had learned to say “Mama” and “Dada,” but at 14 months he no longer uttered any discernible words. Music had a powerful and strange effect on Noah: when he heard it, he would stop what he was doing and “zone out,” according to Leslie.

Four months later Noah's parents brought up their concerns about their son with his pediatrician. The doctor recommended they wait until his second birthday to see if he would catch up with his peers. The advice did little to allay Leslie and Paul's worries.

Absence of language, heightened sensitivity to sound or other sensory stimuli, and difficulty shifting focus from such stimuli, including music, raise the specter of autism in very young children. Autism, a complex disorder of brain development, is now estimated to affect one in 88 kids. It is characterized by communication deficits, impaired social interaction, repetitive motor behaviors, and, sometimes, intellectual disability or physical health problems. Because autism is defined by complex behaviors, obvious signs do not emerge until children start stringing words together and engaging in play with their parents, at about two years old. Children with autism may, for example, continue to play alone far longer than is normal.

The concept that autism is treatable is controversial and new. Newer still is the idea that if children receive therapy very early in life, they are more likely to overcome their deficits. Now, however, many experts believe that delivering therapy to children as young as age one or two—instead of four, as is more typical—can garner greater improvements in IQ, language and social skills. Officials at the American Academy of Pediatrics and the Centers for Disease Control and Prevention, among others, now recommend early detection. To identify autism (or autism risk) in younger children, researchers have had to develop novel screening tools. In addition, one new intervention designed for children as young as a year old has been shown to significantly improve social communication skills.

Missing Milestones
During the first year of life, trillions of new neural connections form in the brain. As a child gains experience with the world between age one and adolescence, these links are pruned—some eliminated and others fortified. One theory posits that in autism this crucial shaping goes awry. Indeed, in studies published in 2012 psychiatrist Joseph Piven of the University of North Carolina at Chapel Hill and his colleagues used magnetic resonance imaging to visualize neural connections in the brains of 92 infants who were at risk for autism because they had siblings with the disorder. They discovered that at six months, the 28 infants who were later diagnosed with autism had a higher density of nerve fibers connecting certain brain regions than did the children who developed normally. At 24 months, however, researchers found that the typically developing toddlers now had greater fiber density in these same regions and that these toddlers also showed a steady strengthening, or thickening, of some initially weak, nonspecific connections. This neural differentiation was far less pronounced in the kids who ended up on the spectrum.

Many experts believe that intelligently guiding the formation, and elimination, of neural connections early in life could sculpt a more functional brain. This belief forms the theoretical basis for early detection. In an initial foray into this area, psychologist Diana Robins of Georgia State University and her colleagues developed the Modified Checklist for Autism in Toddlers (M-CHAT) in 1999. The checklist, designed for children between the ages of 16 and 30 months, probes for precursors of the disorder's main deficits. Is the child using his index finger to point at something he wants or that interests him? If not, his communication skills may not be advancing properly. When Mom or Dad smiles, does the child smile back? If not, she may be lacking another sign of normal social development. A toddler who shows three or more of the 23 behaviors listed on the M-CHAT will be rated “at risk” for autism spectrum disorder.

Leslie and Paul had filled out the M-CHAT during Noah's last visit to the pediatrician. Noah had passed, but at 22 months he was still missing milestones. His language had not improved, and Leslie and Paul noted he was not making eye contact with them when they spoke to him. He showed more interest in objects than in other children. At a family gathering, Noah focused intently on solitary activities rather than engaging in the goings-on around him.

Noah's parents found an M-CHAT online. This time they answered yes to new items: Noah was not pointing, was not responding consistently to his name and was not bringing his parents objects that he wanted to show them. Noah now scored “at risk.” Leslie and Paul brought him to a neuropsychologist, who watched him play and tested him for behaviors that are considered red flags. His parents completed additional rating scales. These assessments indicated that Noah indeed had autism. His was a moderate case; his cognitive abilities outpaced his social skills.

Catching It Early
That diagnosis arrived too late for Leslie and Paul's liking, however. “We saw that for neuroplasticity every month counts,” Paul says. “And we had given up months just waiting.”

Specialists who share Leslie and Paul's concerns are developing diagnostic tools that provide preliminary answers sooner. One of these, a checklist called the First Year Inventory (FYI), is geared toward infants as young as 12 months. First proposed by occupational science professor Grace Baranek and her colleagues at U.N.C. Chapel Hill in 2003, the checklist looks for such risk factors as oversensitivity to touch—say, getting upset when hugged or refusing foods that are crunchy—that are precursors to later sensory regulation problems. Other early signs of trouble include a loose or floppy body when picked up, which signals motor deficits, and failing to turn to look at someone who calls the child by name.

In a study published in 2013 Baranek and her colleagues tested the checklist by giving it to the parents of 699 one-year-olds deemed at risk. They later followed up with parents to find out which of the children had received a diagnosis of autism at age three. At first glance, the results were not the most encouraging. The FYI had identified half of those who went on to develop autism and falsely cleared the other half, who showed no early warning signs. Further, approximately 70 percent of the children who received an at-risk score did not develop autism. Yet most of these kids turned out to have a learning deficit or a different developmental disability, such as attention-deficit hyperactivity disorder. The researchers concluded that though far from foolproof and in need of refinement, the test could serve as a useful initial indicator that a child needs evaluation.

Another experimental measure, the Geometric Preference Test for Autism, involves tracking a toddler's gaze. In previous work, adults with autism have shown a predilection for viewing patterns over people. (Most individuals show the opposite preference.) Some researchers thus reasoned that very young kids with and without autism might diverge in this respect as well. In a study published in 2011 neuroscientist Karen L. Pierce of the University of California, San Diego, and her colleagues showed 110 kids aged one to four a one-minute movie, in which half the screen displayed moving geometric patterns and the other half kids doing yoga, while they tracked the infants' eye movements. Then they administered a diagnostic test for autism. If the child was younger than three, the age at which autism becomes evident in virtually all cases, the researchers repeated the test every six months until he or she reached that age.

The researchers found that 40 percent of those who had, or went on to develop, autism spent more than 50 percent of their time looking at the side of the screen displaying geometric patterns. In contrast, 98 percent of the infants who turned out to be developing normally spent most of their time focused on the yoga video. This preference showed up in babies as young as 14 months. Thus, eye tracking as a gauge of interest in people can provide another early hint that a child is, or is not, at risk for autism. Still, this video test catches less than half of those who end up with a diagnosis.

The M-CHAT—the tool that Leslie and Paul used to assess Noah—also has its flaws. It identifies only about 85 percent of kids who go on to develop autism. About 40 percent of those who receive an at-risk score are false positives—they do not end up with autism—although the vast majority turn out to have another developmental disorder or delay, according to Robins.

Like many complex disorders, autism resists being defined by a single behavioral or genetic test. Although researchers have associated variants of certain genes with autism, having these variants is far from a sentence: 80 percent of infants with a genetic predisposition do not develop the disorder. Some experts say that combining different methods such as genetic screening, behavioral checklists and eye tracking in clinics could measure a child's risk more accurately.

Treatments for Tots
One major reason to start therapy as early as possible is that infants learn about the world and other people through social interaction. Because children with autism have difficulty communicating and relating to others, they miss out on critical learning opportunities. A child who focuses on objects rather than people will be slower to learn how to read facial expressions or convey emotions appropriately. As a result, problems in communication and behavior worsen with time. Children with autism may throw tantrums, act aggressively, and even injure themselves out of frustration and anxiety from being unable to express their needs. Therapy aims to avoid these troubles by teaching and rewarding more appropriate social behaviors.

In the standard autism therapy, applied behavior analysis (ABA), specific behaviors are introduced systematically. For example, a child might be instructed to imitate a particular gesture or to choose an object on command. A therapist then rewards a child with a lollipop or a favorite toy for doing something right. The intensive form of the treatment requires 25 to 40 hours a week of one-on-one therapy for one to three years, and parents often extend the therapy at home. Studies have shown that ABA improves children's communication and self-care skills, as well as their school performance. Many children with autism also receive speech therapy to help them use language in social settings, occupational therapy to help them get dressed, eat and use the bathroom on their own, and physical therapy to help them sit, walk and run to compensate for poor coordination and balance.

ABA did not work well for Noah, who did not like being told what to do or how to play, and he became upset when he was denied a toy for behaving incorrectly. He refused to do the tasks, threw tantrums, experienced night terrors—and engaged in more autism-typical behaviors such as meticulously lining up toy cars and screaming if anyone touched them.

Paul scoured the Internet for other interventions and found the Early Start Denver Model (ESDM), a therapy for infants as young as a year old developed by Sally J. Rogers of the University of California, Davis, MIND Institute and Geraldine Dawson, now at Duke University Medical Center. Because no ESDM therapists lived near them—only 50 or so individuals are certified to perform it in the U.S.—Leslie and Paul used the manual as their guide. But eventually they realized they needed professional help and contacted Dawson. They began driving three hours each way to get Dawson's advice, as well as therapy for Noah.

The ESDM emphasizes interaction as the basis for learning, and instead of doling out explicit rewards, the therapist aims to make the activities themselves rewarding to the child. An adult first searches for something that engages the child—tickling, say, or driving toy cars. Then the parent or therapist coaxes social behaviors in that context. If a child likes a specific book, for example, the adult might encourage the child to point to the book as a way of requesting it. In Noah's home, toys are now in bins and on high shelves so that he has to point to get what he wants.

Collaborative play involving eye contact and sharing is taught next. For example, Noah's parents try to surprise their son by changing the tone or speed of their voice while reading, so that Noah will look up at them and make eye contact—which they reward with a huge smile. Social skills are also taught as part of everyday activities, such as during bath time or lunch. If Leslie wants Noah to choose between two types of drinks at a meal, she holds both beside her face so that he naturally looks up at her before asking for one of them.

Playing Catch (Up)
ESDM has shown significant promise in moderating the symptoms of autism. In a study of children diagnosed with autism published in 2012, Dawson and her colleagues enrolled 48 toddlers in either ESDM-based therapy or a more typical regimen consisting of speech therapy, teaching self-care habits such as brushing teeth and hair, and classroom-behavior training. The researchers also recorded the children's brain activity while they looked at images of objects and faces.

After two years of treatment, the kids who received the ESDM intervention showed fewer symptoms than those who had been given the standard therapy. In addition, the brain activity of those in the ESDM group mirrored that of typical children and diverged from those given standard care. Like normally developing toddlers, children treated with the Denver approach had brain patterns reflecting greater attention to the faces than the objects, suggesting the therapy had helped these children establish a normal preference for social information. Those who received traditional treatment, on the other hand, displayed brain activity indicating that they paid more attention to the objects than the faces.

Other data show that the Denver approach, despite its hefty price tag, saves money in the long run, primarily because it appears to work. In a study presented last year at the Autism Speaks Toddler Treatment Network conference, psychiatry researcher David Mandell of the University of Pennsylvania and his colleagues found that at four years posttreatment, the average cost of autism-related services for 18 toddlers with autism who had received ESDM for two years was about $1,000 less a month than for 21 similar toddlers who had received speech and occupational therapy for the same period. The ESDM recipients needed 100 fewer hours of services every month, and they were more likely to be placed in a regular classroom than an autism-specific one.

Although early treatment is recommended for all affected children, its practical benefits depend on the severity of a child's diagnosis. In a study published in 2013 clinical psychologist Catherine Lord of Weill Cornell Medical College and her colleagues followed 100 children diagnosed with autism before age three until they were 20 years old. They found that of the mildly affected children who began therapy before age three, about 40 percent ended up succeeding in college, making friends and showing no obvious symptoms. In contrast, all of a group of similarly affected kids who did not receive early treatment continued to show clear signs of autism. For more severely impaired children, Lord has shown that early treatment improves verbal and social skills but does not eliminate the need for special services in adulthood.

In Noah's case, early therapy seems to have produced remarkable results. Within one month of first seeing Dawson, at 28 months old, Noah started pointing spontaneously. About three months later he looked at his parents to request something. One month after that, in April 2013, Noah gestured as a way of sharing information with others: he pointed something out to his younger sister, Elina.

Now three years old, Noah engages in creative play of his own design, telling stories with action figurines and dressing up in costumes. His ability to speak and understand language about facts is equal to that of his peers. His social skills remain at the level of a two-year-old. But that level is enough for him to make friends with younger kids. He now also plays with Elina, throwing a ball for her to catch or chasing her around the house. “His world has opened up now,” Paul says. “He doesn't have that focused tunnel vision anymore.”

Moving kids out of their mental tunnels may hinge on noticing that they are in them while their budding brains are most amenable to change.