Anyone who has spent even a little time with an autistic boy or girl soon becomes familiar with the behaviors that set these children apart: lack of eye contact, trouble verbalizing, overreacting or underreacting to activities around them, difficulty in expressing their feelings and in understanding the emotions of others. But how do parents and doctors know if a baby, who is too immature to be gauged on any of these traits, has autism? Early diagnosis has proved difficult.

Inability to detect autism until a child is two or three years old is a terrific disadvantage. It "eliminates a valuable window of treatment opportunity, when the brain is undergoing tremendous development," says David G. Amaral, professor of neurobiology and psychiatry at the University of California, Davis.

Amaral and researchers at other institutions, however, are closing in on techniques that could detect autism in babies as young as six months and perhaps even at birth. The results of these new tests--some controversial--are expanding the understanding of autism and raising hopes for much earlier, specialized care that could improve a toddler's chances for a more normal life as a child, teenager and adult.

A Simple Blood Test?
Autism affects a wide variety of developmental traits. Some young autistic children speak; others do not. Some possess almost average intellectual abilities; others are severely limited. As they grow older, certain autistic individuals display incredible talents in very specific domains. Known as savants, they can memorize an entire book in hours or solve complex math problems faster than people using a calculator. The 1988 movie Rain Man dramatized these abilities in a character named Raymond Babbitt, played by Dustin Hoffman, who won an Oscar for the role. Babbitt was based on a real savant named Kim Peek, who continues to astonish today.

It is no wonder, then, that determining whether a young child is autistic is fraught with uncertainty. Diagnosis typically involves rating a child's behaviors against a set of standards. The exercise usually is not conclusive until at least the child's second birthday. That is why scientists are seeking an earlier and more accurate test, and they are getting closer. At the International Meeting for Autism Research in Boston in May 2005, Amaral presented the initial results of a landmark study. His team compared blood samples from 70 autistic children ages four to six with samples from 35 randomly selected subjects in the same age group. The autistic children had a higher proportion of two basic immune system cells known as B cells and T cells. Significant differences also became apparent in more than 100 proteins and small molecules commonly found in the bloodstream.

After further analysis, the team decided that the pilot study results were strong enough to launch a full-scale investigation. In March 2006 Amaral announced that U.C. Davis's Medical Investigation of Neurodevelopmental Disorders Institute, which he heads, was starting the Autism Phenome Project. It will enroll 900 children with autism plus 450 more who have developmental delays and 450 who are developing normally. Researchers will analyze the children's blood proteins, immune systems, brain structures and functions, genetics and environmental exposures. The participants will be two to four years old at the outset and will be followed for several years. Amaral thinks it is probable that telltale genetic markers will be found. But it will take several years before the project is finished and analyzed and longer still before a routine test for autism could be administered at a doctor's office.

If the blood profiles prove to be reliable, the screening could occur just after a baby is born. But the validity of detection that early in life requires more scrutiny. Amaral says there is a growing view among experts that not all individuals who have autism are "doomed at birth," as has been commonly believed. "It may be that some children have a vulnerability, such as a genetic abnormality," he says, "and that something they encounter after being born, perhaps in their environment, triggers the disorder."

Environment is suspected in part because the incidence of autism is fairly high in American children. The disorder affects one in every 150 eight-year-olds, according to the latest estimates from the U.S. Centers for Disease Control and Prevention. The unexplained preponderance has frustrated scientists trying to find answers. Furthermore, tremendous variation exists among symptoms, "which leads us to believe that autism is a group of disorders rather than a single disorder--several autisms versus one," Amaral says. The blood work could possibly define distinct subtypes. Behavioral experts are reaching the same conclusion, many preferring the term "autism spectrum disorder" rather than simply "autism."

Earlier Treatment Is Key
An early diagnosis is so important because it would allow treatment to begin sooner, while the brain is still significantly strengthening and pruning neural networks. A paradigm shift is taking place on this issue, too. For a long time, scientists believed that functional deficits in certain brain regions caused autism--complications in brain structure that no change in wiring among neural networks would fix. Now they think symptoms arise because of communications problems between brain regions--problems that rewiring could solve if babies received specific therapy.

"The neuronal networks apparently do not coordinate very well," explains Fritz Poustka, director of child and adolescent psychiatry at Goethe University in Frankfurt, Germany. Poustka says regions that get too little input from other parts of the brain do not develop well. This effect is well known among children who were neglected when they were young, some isolated from almost all human contact. A child who develops this way shares some similar consequences, such as poor use of language and difficulty in making social connections. "A quick diagnosis of autism would enable us to stimulate the networks very early in life by deliberately providing the right inputs," Poustka says. He cannot say if such interventions would cure the disorder, but he believes that intensive behavioral training could make the symptoms milder.

Although Poustka doubts that markers in the blood would permit early diagnosis, he favors attempts to try to define telltale traits as young as possible to maximize the success of treatment. In speech development, for example, the best results are achieved when deliberate exercises are instituted before the child's second birthday. By the time a boy or girl is three or four, deficits can still be reduced, but fundamental changes are no longer possible, because the critical period during which speech develops has passed by.

Behaviors Untangled
Whether or not Amaral's project leads to common blood tests, it could prove beneficial to behavioral approaches as well because it includes developmentally delayed children. The standardized checklists that doctors now use for diagnosis, such as the autism diagnostic observation schedule, are adequate only for children who are at least one and a half to two and a half years old. And then, usually only for the so-called high functionals--autistic children with IQs over 80. The tests are inconclusive for many of the other suspected individuals because children who are delayed in their intellectual development often score similarly to children who truly have autism. It is difficult to determine whether cognitive problems are being misdiagnosed as symptoms of autism, Poustka says. Delay, or a completely different disorder, can prompt what appear to be autismlike patterns.

A Canadian research team is trying to clarify this overlap. Led by Lonnie Zwaigenbaum, now at the University of Alberta in Edmonton, the group devised a 16-point observational checklist called the Autism Observation Scale for Infants and used it to evaluate 65 one-year-old children, all of whom had older siblings with autism and therefore had an above-average chance of developing the disorder themselves. The researchers also assessed another 23 babies with no familial ties to or signs of autism.

Zwaigenbaum's group reappraised the children when they were two, this time using traditional tests. They found that almost all the children who were diagnosed as autistic at age two had seven or more distinguishing traits when they were only one. "The predictive power of these markers is remarkable," Zwaigenbaum says.

Even among children just six months old, certain behavioral patterns forecast the onset of the disorder, notably a passive temperament and low physical activity levels. By their first birthdays, the children who later turned out to be autistic were easily irritated, had problems with visual tracking, tended to focus on a very few objects, failed to look around for a speaker who said their name, and barely interacted with others. They also tended to have certain obsessive motions, such as stroking surfaces, yet made very few gestures toward other people. And they understood less spoken language than their age-mates who were later identified as nonautistic.

As Amaral acknowledged about his first blood-profile exploration, Zwaigenbaum notes that further studies must include children who are at risk for other developmental disorders to help distinguish which symptoms are specific to autism. He is also open to the possibility of environmental influences in triggering or at least exacerbating autism. He says it is hard to know if the traits his group identified are early manifestations of the disorder or if they contribute to a pattern of development that may lead to autism.

Either way, his investigation, Amaral's and those of others are all improving our understanding of when autism starts, providing hope for earlier diagnosis and more effective treatment. The goal, of course, is to offer toddlers a greater chance at a more fruitful childhood, which in turn raises their chances for more satisfying years as teenagers and adults. The many challenges that autistic individuals face as they mature--learning, communicating with others, making and keeping friends, building life skills, securing a job, finding love--will be less daunting if they can get off to an earlier, better start.

(The Author)
ULRICH KRAFT is a contributor to Gehirn & Geist. He wrote about disorders of the brain's timekeeping mechanism in the June/July 2007 issue of Scientific American Mind.

(Further Reading)

  • Behavioral Manifestations of Autism in the First Year of Life. Lonnie Zwaigenbaum et al. in International Journal of Developmental Neuroscience, Vol. 23, Nos. 23, pages 143152; AprilMay 2005.
  • Autistic Brains Out of Synch? Ingrid Wickelgren in Science, Vol. 308, pages 18561858; June 24, 2005.