By Sandra Blakeslee

There comes a point in every great mystery when a confusing set of clues begins to narrow. For scientists who study autism, that moment may be near, thanks to a combination of new tools for examining brain anatomy and of old-fashioned keen observation.
 
Within the last year, several laboratories have reported finding
important new clues about the mysterious syndrome that derails normal
childhood brain development.
 
For the first time, they say, a coherent picture is emerging.
 
In autism, subtle brain abnormalities are present from birth. Infants
and toddlers move their bodies differently. From 6 months to 2 years, their
heads grow much too fast. Parts of their brain have too many connections,
while other parts are underconnected.
 
Moreover, their brains show signs of chronic inflammation in the same
areas that show excessive growth. The inflammation appears to last a
lifetime.
 
"Autism is still a confusing disorder, but one thing is now clear,"
said Dr. Pat R. Levitt, a neuroscientist who is the director of the Kennedy
Center for Research on Human Development at Vanderbilt University in
Nashville. "There is a specific disruption of circuitry in brain
development. We can really dig in and begin to explain the splintered brains
of autistic children."
 
To that end, Dr. Levitt and two dozen leading brain researchers held a
three-day "autism summit" in Malibu, Calif., sponsored by the Cure Autism
Now Foundation, to discuss this emerging view and to plan collaborative
studies. The meeting ended Sunday.
 
"Up to now, there was no theory to link one anatomical study to the
next," said Dr. William T. Greenough of the University of Illinois, an
expert on brain development. "We now have a theoretical framework that can generate predictions to test."
 
People with autism have great difficulty with social interaction. Some
cannot speak. Many are clumsy. A common trait is obsessive attention to
certain details. Symptoms can be severe to mild.
 
Diagnoses of the disorder have increased in recent years, although no
one knows why. One child in 166 born today may fall on the autism spectrum.
Researchers agree that an unknown number of genes interact with
unidentified environmental factors to produce the disorder. The new clues
focus on brain development and circuitry, and especially on the brain's
white matter. White matter contains fibers that connect neurons in separate
areas of the brain, whereas gray matter contains the neurons themselves.
 
"You can think of this distinction as analogous to that between cables, or
white matter, and circuit boards, or gray matter, inside a computer," said
Dr. Matthew Belmonte, an autism researcher at the University of Cambridge in England. "Even though each individual circuit board may be intact, if the
cables are disrupted then the computer can't function."
 
Using a new technique called morphometric analysis, in which
post-mortem brain tissue is divided into tiny parcels and examined, Dr.
Martha Herbert, a pediatric neurologist at Harvard Medical School, found an
anomaly in the white matter of autistic brains - it is asymmetrical.
 
In autism, white matter grows normally until 9 months, Dr. Herbert
said. Then it goes haywire. By 2 years, excessive white matter is found in
the frontal lobes, the cerebellum and association areas, where higher-order
processing occurs.
 
The right side of the brain, the nonverbal hemisphere, is especially
encased in white matter. The two sides of the brain are poorly connected.
Moreover, small functional regions in each hemisphere tend to be prematurely
insulated by excess white matter.
 
Another clue was reported last year by Dr. Eric Courchesne, a
neuroscientist at the University of California, San Diego. Using a simple
tape measure, he found that newborns who later developed autism had smaller head circumferences than average. From 1 to 2 months of age, their brains suddenly begin to grow rapidly. Another spurt occurs between 6 months and 2 years, giving rise to exceptionally large heads.
 
At age 3, one child could wear his father's baseball cap, Dr. Courchesne said. The rate of brain growth gradually slows from 2 to 4 years, reaching a peak a year later. A 5-year-old with autism has the same size brain as a normal 13-year-old.
 
But by midadolescence, when normally developing children catch up, the
autistic child's brain is again comparatively smaller.
 
Advertisement Dr. Ruth Carper, who works with Dr. Courchesne, went
on to show that the frontal lobes, the slowest and latest brain region to
develop, have the biggest size increase of all. But the nerve cells in this
region, which is responsible for social reasoning and decision making, are
actually much smaller than normal and "underpowered," Dr. Carper said.
 
A third clue, from the laboratory of Dr. Marcel A. Just, a neuroscientist at Carnegie Mellon University in Pittsburgh, reaffirms the odd circuitry in autism. In a study published in November, he found that people with autism remembered letters of the alphabet in a part of the brain that ordinarily processes shapes. That is, the subjects used a basic sensory region to deal with higher-level concepts.
 
"Autism results from a failure of various parts of the brain to work
together," Dr. Just said. "Distinct brain areas work independently. People
with autism are good at details but bad at conceiving the whole."
Local networks are overconnected, he said. Long-range networks are
underconnected.
 
Skewed brain wiring could explain a fourth clue: clumsiness. Dr.
Philip Teitelbaum, an expert on human movement patterns at the University of Florida, studies how babies with autism learn to roll over, sit up, crawl
and walk. By looking at videotapes of their early months, before their
disorders are diagnosed, he finds that autistic children use unusual
strategies for locomotion. It is as if the parts of their brains that
control movements are not properly connected.
 
A fifth clue, also reported in November, may turn out to be a major
piece of the puzzle. Dr. Carlos Pardo-Villamizar, an assistant professor of
neurology and pathology at Johns Hopkins, studied the brain tissue of 11
people with autism who died at ages 5 to 44. He found a pattern of
inflammation in the same regions that appear to have excess white matter.
The brain has an innate immune system separate from the body's immune
system, Dr. Pardo said. A sentinel cell type, called microglia, is always on
the lookout for trouble. When activated, the cells elicit inflammation and
growth factors.
 
Another cell type, astroglia, helps pattern the brain in fetal
development and is later involved in synaptic activity. The astroglia were
also elevated in the 11 brains.
 
Dr. Pardo then examined spinal fluid in six living children with
autism. He found evidence of activated microglia, hence inflammation, along
with astroglia.
 
It is not yet clear whether the inflammation is protective or
destructive, Dr. Pardo said. In either case, inflammation is most marked in
the same areas highlighted in all the other studies showing the same
abnormal circuitry.
 
Other researchers have begun studies to find out whether genes
involved in innate immunity and prenatal wiring are involved in the
disorder.
 
A crucial question is why does the brain grow rapidly and then stop
growing, Dr. Courchesne said. What accounts for the timing of the defect?
Dr. Herbert and others wonder whether the white matter is really larger.
"We don't know what is inside those enlarged areas," she said. "It
could be more axons, more white matter, or more glial cells and astrocytes."
Dr. Robert Miller, a white matter expert at Case Western Reserve
University School of Medicine in Cleveland, said he planned to examine white
matter from autistic brains to see what gives them their "odd architecture"
and perhaps discover the cause of the overgrowth.
 
Studies are under way to dissect the white matter in greater detail.
While these new clues are exciting, they do not lead to immediate
treatments. Parents should not, for example, rush to give their autistic
children anti-inflammatory medications at this time, Dr. Pardo said, because
the link between autism and inflammation is still preliminary, and in any
case, the drugs do not affect the type of inflammation particular to the
brain.
 
On the other hand, once autism is diagnosed, often around age 2 or 3,
when the frontal lobes fail to activate properly, therapies might focus on
activating multiple brain areas at the same time. This would not cure the
disorder, Dr. Herbert said, but could theoretically lead to improvement.
Meanwhile, other clues remain elusive. "Parents will tell you that
when their child spikes a high fever, the child becomes lucid and
communicative," said Dr. Levitt, of Vanderbilt. "A fever is a
neuroinflammatory response. That suggests the circuit defects could be
reversible. We just don't know."