The Joker has stolen the world's biggest diamond, and it's up to Batman and Robin to get it back. On a video monitor, hands move toy action figures through the paces of the story, as an off-camera voice speaks the dialogue.

John, age 6, watches the monitor with rapt attention. He is autistic, and this is a technique called "video modeling," used by educators at the New England Center for Children (NECC) in Southborough, Mass., where John and some 200 other autistic kids attend school.

When the video ends, Jen, his teacher, affectionately ruffles his hair and directs him to a table that holds the same Batman toys seen in the video. He is supposed to play with them in exactly the same way, saying the same lines, as he has just been shown.

Normally developing children play by imagining scenarios and acting them out with toys. Kids with autism do not. They have to be taught how to play this way. The goal is for them to understand the concept well enough to expand on their play, using their own imaginations.

Learning New Behaviors, Changing Harmful Ones

Teaching play with video modeling is something new that the NECC is studying. On the staff working with the kids are not only educators, but also researchers, and they report on what happens at the NECC to the scientific community. Video modeling is just one small part of the NECC's whole approach, called "applied behavioral analysis," widely regarded as the gold standard in autism treatment.

In the simplest terms, applied behavior analysis offers carrots, with no stick. Children are engaged in learning new behaviors and helped to change disruptive or harmful ones, by constant positive reinforcement. The curriculum at the NECC ranges from the most basic language and social skills to academics and vocational training. "We plug them into that wherever they're at," says Rebecca MacDonald, PhD, director of the preschool program, which includes kids aged 3 to 7.

Another new area she is studying is what's called "joint attention," an important early step in relating to others. If you were to turn and look at something, a typical child observing you would probably turn to look at it, too. An autistic child would not notice. "It's one of the hallmarks of autism," MacDonald tells WebMD. "They tend not to care what you are looking at or thinking."

To encourage them to care, she will make something interesting or rewarding happen if the child follows her gaze. For example, she'll activate an electronic toy by remote control if the child looks at it when she does. Improving joint attention behavior won't make all other social skills fall easily into place. Like video modeling, it's just one tool used in the intensive work that the NECC does. Kids who attend the school go for 30 hours a week, all year long.

Starting Early

The earlier this kind of work can begin, the better the outcomes tend to be for autistic children. Autism can usually be diagnosed by 18 months of age, but some scientists hope that in the future, a blood test at birth might detect it.

In May 2005, researchers at the University of California, Davis MIND Institute announced that they had found remarkable differences in blood tests of autistic and nonautistic children. The children had different levels of certain proteins in the blood and more of some kinds of immune cells.

"The idea for early detection is not only that you can intervene early, which is beneficial, but there's the notion that not all children who ultimately have autism are doomed to it at birth," David Amaral, PhD, research director at the MIND Institute, tells WebMD.

Scientists have speculated that maybe something in the environment makes children who are susceptible to autism develop the disorder. If researchers could identify the trigger, avoiding it might prevent autism.

"In some cases the information might allow full-blown prevention, and in other cases more tailored treatment," MIND Institute researcher Blythe Corbett, PhD, tells WebMD.

It's too early, however, to say for sure what the differences seen in the study mean. "We don't know whether our findings indicate a cause or an effect," Amaral says.

It may be that the immune system plays a role in some children's autism, but "there simply is not going to be a single cause," he says. "In fact, we think of autism not as autism, but as autisms."

What's more, the differences may not be specific to autistic children. "You have to show, for example, that it differentiates kids with autism from kids with obsessive-compulsive disorder or attention deficit disorder," says Eric Hollander, MD, director of the Seaver and New York Autism Center of Excellence at the Mount Sinai School of Medicine in New York City.

Hollander's own research has shown that a particular trait in the immune systems of autistic kids relates to the severity of repetitive behaviors, a common autism symptom. This same trait has been linked to Tourette's syndrome and obsessive-compulsive disorder.

Emerging Treatments

Recently, Hollander studied the use of Prozac for treating repetitive behavior in children with autism. Those who took low doses of the drug in liquid form showed better improvement than those who took a placebo. But selective serotonin reuptake inhibitors (SSRIs) like Prozac are not stand-alone treatments for autism.

"The treatment of choice for most individuals is an integrated approach," Hollander tells WebMD.

At the University of California San Francisco, professor Michael Merzenich, PhD, is working on a computer program to teach language skills to autistic kids through what is called "neural retraining." It may sound like science fiction, but it's not all that speculative.

Scientists have come to understand that the brain is not hardwired, but very flexible, or plastic. There are software programs, such as one called Fast ForWord, that can train the brains of kids with impaired language ability to process speech better.

"We have very strong documentation that this kind of brain-plasticity-based training can have an effect," Merzenich tells WebMD.

But programs that exist now are too complex for many autistic kids to use. "The ways that these programs have been designed for nonautistic children just don't apply to most autistic children," Merzenich says.

Once Merzenich and his team finish building their program, they will have to put it through years of rigorous testing, which he says they hope to begin later in 2005.

Although its methods are state of the art, the New England Center for Children prides itself on only applying treatment that is backed up by solid research.

"People are faced with a raft of alternative treatments that have no merit," Vincent Strully Jr., the NECC's founder and executive director, tells WebMD. He counts special diets, secretin treatment, and mercury detoxification among those.

"We're not claiming any cure," he says, but the center's approach makes a difference. "It's advancing the lives of these kids dramatically."

By Martin F. Downs, reviewed by Michael W. Smith, MD

SOURCES: Rebecca MacDonald, PhD, program director, New England Center for Children. Vincent Strully Jr., executive director, New England Center for Children. David Amaral, PhD, research director, MIND Institute, University of California, Davis. Blythe Corbett, PhD, assistant professor, department of psychiatry and behavioral health, University of California, Davis. Eric Hollander, MD, director, Seaver and New York Autism Center of Excellence, Mount Sinai School of Medicine, New York. Michael Merzenich, PhD, professor, University of California, San Francisco. WebMD Medical News: "Research Offers Hope for Autism Blood Test." Hollander, E. Neuropsychopharmacology, March 2005; pp 582-589. Dube, W. The Behavior Analyst, Fall 2004: pp 197-207. Hollander, E. American Journal of Psychiatry, February 1999: pp 317-320.