Elizabeth Torres, a computational neuroscientist at Rutgers University, thinks many experts are making a mistake when they focus only on what's malfunctioning in the brains of people with autism.

She sees autism as a condition of the whole body in which information from all sorts of sensory channels - movement, touch, pain, vision, temperature - is not reaching the brain properly while messages from the brain that tell the body what to do also are not getting through.

"The whole loop is disrupted," she said as she explained two studies published last month in Frontiers in Neuroscience that lay out her theories on the importance of movement as a form of sensation and perception in autism. That loop, she said, plays a huge role in how normal people make sense of their environment and anticipate what's coming.

The lack of integration between sensory-motor input and the brain may explain why people with autism can become so narrowly focused. "They must live in a very uncertain world," she said, "so the moment they get an anchor in that world, they try to hold on to it as much as possible."

Torres, who started out in computer science and math and then began working in cognitive science, has a particular interest in analyzing the connection between movement and learning. "I saw what people had done with movement in autism, and I thought it was very pedestrian," she said.

One of her new studies, with Jorge Jose, a theoretical physicist and computational neuroscientist at Indiana University, measured intentional and spontaneous movements and found differences between autistic and normal children. The other mapped how those movements changed as autistic children learned a new skill. Torres thinks these movement differences ultimately can be harnessed to diagnose autism earlier and track the effectiveness of teaching techniques.

In an e-mail to Rutgers, Anne Donnellan, director of the University of San Diego's Autism Institute and co-editor of the Frontiers in Neuroscience issue, called Torres' work truly groundbreaking and said it was "bound to have a broad impact on multiple disciplines of brain science." She did not respond to a request for comment from The Inquirer.

Ben Yerys, an autism expert at Children's Hospital of Philadelphia, called Torres "very innovative in the way she's measuring movement." He said that learning how information comes in from the senses in autism is understudied, but that others are also interested in it.

He said Torres still has many questions to answer. One of the more important ones is whether her approach can help improve the integration between senses and the brain.

Torres' team used sensors to measure intentional movements and those that are random and spontaneous. We'd all be amazed, she said, at "how many facial expressions that you make that you're not even aware of, how many gestures you're not even aware of." The pattern of movements was different in the 3-to-24-year-old autistic people the team studied than in people who had developed normally. The differences were consistent across age, ability, and gender.

The movements of those with autism were more random than those of "typical" people, a sign that they were not learning from what they had sensed in the past or anticipating the future. "Most of them must have to live in the here and now," Torres said.

She thinks scientists may eventually be able to use facial movement, which matures earlier than limb movement, to identify babies with autism.

In the other study, autistic children played a game that let them choose what they wanted to see on a computer screen. Their movements in space triggered cartoons or video of themselves to appear. Eventually they discovered for themselves how to choose what they wanted to see. They were also able to remember how the game worked over time.

The pattern of their movements changed as they learned. "We were able to shift the pattern from noisy and random to predictable and reliable," Torres said.

She thinks this kind of self-discovery is a better way to teach autistic children than the rote, highly controlled learning that is common today. She also thinks instructors should worry less about odd repetitive physical behaviors that are common in autism. Those may be a way people with autism have found to gather sensory input that may be helpful.

Matt Tincani, an associate professor of special education and applied behavior analysis at Temple University, said a teaching strategy called applied behavior analysis currently has the most scientific support. It does involve a lot of repetition and overt instruction. "There's quite a bit of research that shows that that's how kids with autism learn best," he said. Many children are able to move to a more typical learning pattern as they gain skills, he said.

James Connell, clinical director of Drexel University's A.J. Drexel Autism Institute, said discrete trial training, which relies on repetition and reward, had produced enormous gains. It is often combined now, he said, with play-based approaches that are driven by children's interests.

Contact Stacey Burling at 215-854-4944 or sburling@phillynews.com.