Michael McCormick, 11, was lying with his head inside a giant magnet, gazing at a computer monitor reflected in the mirror above him and trying hard not to fidget.

But, finally, he couldn't take it any longer. He cried out over the intercom: Could someone get rid of the toolbar on the bottom of the computer display?

A few feet away, in the control room of the Brain Imaging Research Center, his mother could only smile and shrug.

"You never know what's going to annoy him," said Linda McCormick, of Zelienople.

That's how it is with autism, a complex developmental disorder that can severely limit a person's ability to communicate and interact with others. Obsession about details such as computer displays is one aspect of the disorder McCormick has learned to accept.

The hour that Mike will spend lying still in the dark magnet room could help researchers shed some light on autism. Since the late 1990s, a succession of high-functioning autistic people have submitted to these brain-imaging studies by researchers at Carnegie Mellon University and the University of Pittsburgh.

The results are beginning to change their understanding of what autism is.

The first major research paper to emerge from the studies, published this month in the British journal Brain, suggests that different areas of the brains of autism patients don't work with each other in the coordinated manner necessary for most high-level thinking.

Individual brain areas function properly, but they have trouble communicating with each other, said Marcel Just, the Carnegie Mellon psychologist who has led the study and directs the Center for Cognitive Brain Imaging.

It's sort of like a kindergarten soccer team ---- all the players run to the ball instead of playing their positions.

Though this first study focuses on only how brains perform during tests of sentence comprehension, Just and his co-author, Pitt's Dr. Nancy Minshew, use the findings to propose a biological theory of autism they call underconnectivity.

A lack of connections between brain areas, or a surplus of connections, which results in inefficient communication, could explain the disordered thinking of autism patients, they said.

The work still needs to be buttressed with additional studies, but it is consistent with other developments in autism research and is generating a lot of interest in the autism research community.

"Dr. Just's work suggests the first plausible brain mechanism that may be able to explain all of the deficits seen in autism," said Dr. Janet Lainhart, an autism researcher at the University of Utah.

"I think that there's a new paradigm emerging in autism research," said Dr. Martha Herbert, a pediatric neurologist at Massachusetts General Hospital. Autism has long been defined by its behavioral symptoms, but now researchers are getting closer to understanding what goes haywire in the brain to cause those behaviors.

"It opens up a whole new universe of research," she added, "and I think people are beginning to realize this."

Gauging flow of blood

Minshew, a pediatric neurologist who heads the federally sponsored Collaborative Program of Excellence in Autism that includes Pitt, Carnegie Mellon and the University of Chicago, said the findings could lead researchers to the genes responsible for the underconnectivity. Other researchers suggest it also could guide the development of new therapies for the disorder, or perhaps lead to early interventions during infancy.

As exciting as the findings might be, Dr. Helen Tager-Flusberg, who heads a collaborative program at Boston University, counseled a wait-and-see approach.

"For the last 25 years, the field of autism has been filled with people who announce spectacular findings that other groups have been unable to duplicate," she said. One problem is that the severity and symptoms of autism can vary widely.

Mike, for instance, is a good student with a dry wit who, until recently, was in the gifted program of his school, McCormick said. A voracious reader, "He can read three pages in the time I take to read one." Yet he has trouble with writing, has difficulty coping with change and his social skills are so limited that he has no friends.

Most people with autism have more severe symptoms. Some never learn how to talk. Many avoid physical contact with others, avoid eye contact, and have difficulty reading other people's moods. It is easy for these people to become cut off from the rest of the world.

Many researchers have looked for structural brain abnormalities that might explain these characteristics. "That hasn't been all that fruitful," Mass General's Herbert said.

Just is among those who have taken a different approach, looking at how the brain works, not how it's built. Using a technique called functional magnetic resonance imaging, or fMRI, he and his colleagues can see which areas of the brain activate during different tasks by measuring the flow of blood to those areas.

Mike, for instance, was tested on his ability to comprehend ironic statements and to make judgments involving distance and time, high-level tasks that require several areas of the brain to work together, while the fMRI machine scanned his brain activity every second.

Though the scanner produces images, the research doesn't involve simply looking at them. Instead, Just and post-doctoral researchers Tim Keller and Rajesh Kana are interested in the timing of that activity and whether it is synchronized between the brain parts.

With each study involving huge data files on dozens of autistic subjects and normal control subjects, the analysis requires intensive computing. Just noted that the project has its own supercomputer, housed next to the fMRI imaging facility inside the McGowan Institute for Regenerative Medicine building on the South Side.

Such studies can be performed using only high-functioning autism subjects such as Mike, who has a form of autism known as Asperger syndrome. These high-functioning people make up about 15 percent of all people with autism. "It's very hard to find these people," Just said, so volunteers have been flown in from throughout the eastern United States to participate.

Underconnectivity theory

In the study published in Brain, the researchers tested sentence comprehension. People with autism often have a good understanding of individual words ---- "You don't need a dictionary when they're around," Minshew said ---- but have trouble understanding them in the context of a sentence.

"If you say, 'If we don't get this done, we're going to be in the doghouse with Mom,' they look around for the doghouse," Minshew said.

Not surprisingly, they found that 17 autistic adults showed more activation of Wernicke's area, the brain area responsible for word recognition, than 17 control subjects. And they had less activation of Broca's area, which is responsible for semantic understanding.

But they also found that the synchronization of activation in the areas was consistently lower in the autistic group than in the normal group.

They have since done similar studies of different high-level thinking tasks, such as problem solving and social cognition. "We're writing them up as fast as we can," Just said, to provide further evidence to bolster the theory of underconnectivity.

Underconnectivity is consistent with other findings, such as Lainhart's discovery that autistic people tend to have larger brains and Herbert's analysis of variations in the amount of white matter, the electrical cabling that connects brain cells.

Robert Schultz, a clinical psychologist at Yale University, said bigger brains and excess white matter necessarily results in less overall connectivity; at some point, it becomes impossible to interconnect so many extra brain cells.

Positron emission tomography, or PET, scan studies 20 years ago also had shown signs of this underconnectivity, he noted. Just's recent work "is the best data to date, but he's not the first person to suggest this."

Schultz is busy trying to replicate the Pittsburgh work, using the fMRI technique for studying face perception by autistic people, one of his specialties. "It's tricky. With fMRI, you have to keep in mind that we're measuring blood flow," and so the effects of the pumping heart have to be considered.

Uta Frith, deputy director for the Institute of Cognitive Neuroscience at the University College London and an influential player in autism research, said the findings of underconnectivity were consistent with her own theory of autism, called weak central coherence. But she said similar findings have been found in studies of dyslexia, a learning disorder that involves problems with word recognition, and so underconnectivity probably does not explain autism.

Lainhart, however, contends Frith's theory, which suggests that people with autism have a heightened focus on details, explains only some of the nonsocial aspects of autism. Underconnectivity, she said, appears to explain all autism symptoms.

Minshew expressed confidence that the evidence would eventually support the underconnectivity theory. "This is one of the major papers that will stand the test of time. They'll refer to this paper forever."

Mike McCormick will be happy if his contribution to the studies eventually helps people less fortunate than himself, his mother said. But he neither expects nor necessarily wants to benefit himself.

"He doesn't think he needs to be cured," she said. "He has the mind of a philosopher."