BOSTON -- Autism begins when certain brain cells fail to properly mature while the fetus is within the womb, according to scientists who say the scattered pattern of the undeveloped cells may explain the wide range of symptoms seen in children.
The researchers, studying post-mortem brain tissue taken from kids, found that those diagnosed with autism missed key genetic markers for brain cells that are supposed to develop prior to birth. The defects occurred in regions that control emotion, communication, language and social comprehension, all functions impaired in autism, the researchers wrote.
The finding, reported today in the New England Journal of Medicine, may one day offer a way to diagnose the disorder earlier, when the potential for treatment is greatest. The cause of autism and its related disorders is unknown. Affecting 1 in 50 U.S. children, it is mostly diagnosed from the ages of 3 to 5, based on behavioral changes that can occur.
"We found a novel aspect of cortical development never seen before that provides clues to the potential cause of autism and when it began," said Eric Courchesne, director of the University of California at San Diego's Autism Center of Excellence. "The type of defect we found points directly and clearly to autism beginning during pregnancy."
What triggers the brain-cell disruptions remains a mystery, Mr. Courchesne said in a telephone interview. The cause could be genetic, environmental or a combination, he said.
While most autism studies have been done using imaging technology, such as an MRI, a handful have been based on an analysis of brain tissue, mostly taken from adults.
The latest research, by scientists at the Allen Institute for Brain Science in Seattle and the University of California at San Diego School of Medicine, examines samples from 22 children, ages 2 to 15, half of whom were diagnosed with autism before they died. Samples from the 11 children who didn't have autism were used as a control group in the study.
Even with the small sample, the study-- involving painstaking analysis of more than 12,000 slides -- was as large or larger than previous postmortem studies, the researchers said. To visualize the work, the study's lead author, Rich Stoner of the UC San Diego Autism Center of Excellence, created the first three-dimensional model showing brain locations where patches of cortex failed to develop normally.
The findings also suggested that the disruption of cell development in the brain probably occurred in the second and third trimesters, Mr. Courchesne said in a telephone interview.
"These patches of disorganized cortex are patches where cell types have failed to develop," he said. "This indicates this must have happened in fetal life when the brain is setting up neural cell types, neural connections and neural layers."
By the second trimester, fetal brain cells are making complex connections, the scientists said. At this point, the cortex develops into six layers, each with its own specific types of cells, assembly patterns and connections that perform unique roles in processing information.
As each cell develops, it leaves behind an observable genetic marker that the researchers used to find differences in the brains of children with the disease. The undeveloped cells they found, meanwhile, occurred in disorganized patches, possibly showing why the disease takes on so many forms.
They found the disorganized patches of defective cells in brain tissue from 10 of 11 children with autism and only 1 of 11 unaffected children.
"The next step is to investigate the cause of these abnormalities and if anything can be done about it," said Ed Lein, an investigator on the study from the Allen Institute for Brain Science.
Because the defects weren't uniform, the findings may show why the illness has such a wide range of symptoms and can occur with varying severity, Mr. Courchesne said. Some children diagnosed with autism also improve over time, which may have to do with brain cells rewiring themselves to compensate for defects.
"The broader implication is now we have more evidence that brain changes early in life are very important for understanding autism, and it helps make the case that we have to start earlier to understand causative factors," said Dan Smith, senior director for discovery and research at Autism Speaks, an advocacy group based in New York.
"We need to expand on findings like this study to understand whether there are biological markers that we can detect before we detect behavioral changes," Smith said in a telephone interview.
There is no cure for autism. Treatments for children with the disorder involve behavioral therapies designed to improve learning, communication and social skills, as well as medications or a combination of the two.
Bigger studies that include more subjects and genetic markers, as well as more extensive brain mapping will be needed to confirm the findings or to find potential triggers that cause the brain differences, Mr. Courchesne said.
"Our panel had only 25" gene markers, "the largest panel ever used, but it's not enough," he said. "Even though we had 12,000 slides, that still isn't sufficient."