In recent years, autism research has been a battleground. A vocal group of parents, advocates and a few scientists focused on vaccines containing traces of mercury as the lead suspects in the disorder. But most autism researchers were suspicious, arguing that the theory didn't fit the evidence.
Now, with a new Danish study offering the strongest evidence yet against the vaccine theory, the controversy may give way to a more baffling question: If vaccines aren't the culprit, then what is?
The range of theories underscores how little is known about autism, a developmental illness that afflicts as many as one in 200 American children and makes it difficult for them to connect with the outside world.
Researchers are looking at a variety of possible mechanisms: Do these children suffer from "extreme male" brains? Are they prenatal victims of their mothers' overaggressive immune systems? Or does faulty insulation between neurons fill their brains with maddening static? These are a few of many competing theories.
"There are a variety of tantalizing trails of evidence," says Diana Schendel, a Centers for Disease Control and Prevention epidemiologist who studies the disorder. "But none of them are conclusive, by any stretch of the imagination."
Last week's issue of Pediatrics published a report on autism cases in Denmark, which stopped using vaccines containing thimerosal, a common mercury-based preservative, in 1992. (In the past three years, U.S. manufacturers also have stopped or reduced thimerosal use in children's vaccines.)
The Pediatrics study found that Denmark's autism had risen sharply since 1992. "If thimerosal causes autism, then you'd expect rates to go down after it's banned," said the study's main author, Dr. Kreesten Madsen of the Danish Epidemiology Science Center.
Most autism experts agree with Madsen.
"This reaffirms that there is no evidence that thimerosal causes autism," said Dr. Neal Halsey, director of the Institute for Vaccine Safety at the Johns Hopkins University.
But some critics called Madsen's study deeply flawed. Because Denmark has such a low autism rate, the study is irrelevant to the United States, said Mark Blaxill, a director of the parent group Safe Minds. "We still consider thimerosal a very important suspect," he said.
Yet even scientists who suspect thimerosal minimize its role in the disorder. "If thimerosal was a huge contributor, we would have picked that up already," said University of California researcher Isaac Pessah, who is hunting for chemical triggers, including mercury and pesticides.
Most researchers suspect that autism is a family of disorders with several causes. While autistic people share a core of symptoms - they crave routine, have trouble communicating, and don't understand intuitive social rules - their behavior can vary greatly. While some have large vocabularies, others barely speak. Some seem riotously overstimulated; others seem locked in a world with little sensation.
"Autism is no one thing. It probably has multiple causes," said Dr. Andy Zimmerman of the Kennedy Krieger Institute's Center for Autism and Related Disorders.
A combination of abnormal genes - probably between four and 20 - likely lays the groundwork for the illness. Dozens of scientists are trying to find these oddities. Among these is a group at Vanderbilt University, which recently found that some autistics have a gene that may decrease brain serotonin levels. This neurotransmitter plays a role in several key autistic symptoms, including compulsiveness and anxiety.
But genes alone don't cause the illness. Most experts think genes simply create a vulnerability, and that environmental factors send some of these children over the edge.
Zimmerman, a pediatric neurologist, suspects that immune system abnormalities in mother and child may provide one push. Several studies have found that maternal infections during pregnancy can increase a child's autism risk. He theorizes that the illness is set in motion when a mother's aggressive immune response throws off the rhythm of fetal brain development.
In his most recent work, Zimmerman has found that in some autistics, the brain's immune cells - the microglia - are overactive. His next step: to find out whether these cells are fighting off an attack or doing harm themselves.
At Harvard Medical School, pediatric neurologist Margaret Bauman has recently found evidence of such subtle brain damage. In studying the brains of autistic cadavers, she found abnormalities in the "white matter," insulation that keeps neuronal messages free of static.
Many scientists think the illness may stem from an oversized brain. A study published this summer in the Journal of the American Medical Association found that autistic toddlers had larger than average heads. This "neuroproliferation" could explain why autistics often seem overwhelmed by ordinary stimuli.
"There are too many cells, and they could potentially create too much noise," said Rebecca Landa, director of the Center for Autism. "It's like sticking your finger in a socket. Their brains can't channel on the input."
She is working on a study to measure the levels of neurochemicals that may control brain size and structure. "Maybe the brain overbuilds itself," said Landa, who hopes to correlate neurochemical levels with specific autistic behavior.
Other scientists are focusing not on the size of the autistic brain, but its gender. Cambridge University psychiatrist Simon Baron-Cohen argues that those with the illness have "extreme male brains."
Men, he says, tend to see the world in terms of systems, while women view it through the prism of emotion and relationship. Autistics - 80 percent of whom are male - show many hyper-masculine traits, he argues. They focus on details while neglecting the whole, have trouble interpreting facial expressions and acquire language slowly. He has also found high rates of autism in families of physicists and engineers - professions that require systematizing skill.
Baron-Cohen is testing testosterone levels in the amniotic fluid of 3,000 children, some of them autistic. His hypothesis: high prenatal testosterone can produce an acutely "male" brain. The hormone may speed development of the right hemisphere, which probably controls these male characteristics.
If testosterone does play a role, prenatal treatments might help the disease. But Baron-Cohen is leery, arguing that "hormonal engineering" could radically alter personality.
"Would we want to intervene?" he asked. "You may make your child more sociable, but also maybe less focused on systems. There could be a cost for that."