Special Interests
Office Locations
Contact Jerry

The development of tools for earlier diagnosis of autism is moving quickly.
By Beth Azar Monitor staff

An article from the American Psychological Association’s Monthly, the APA Monitor:
VOLUME 29 , NUMBER 11 -November 1998

To the consternation of many parents, children with autism rarely receive a diagnosis before age 3 or 4. But behavioral researchers believe they are homing in on specific behaviors that should drop the age of first diagnosis down to as young as 18 months. Early diagnosis has become increasingly critical as treatments for the potentially devastating developmental disorder advance and research begins to show that the earlier the disorder is diagnosed, the better the prognosis.

Developmental researchers increasingly find that there are critical periods of child development after which certain systems—such as language, vision and motor skills—become less malleable. And the same is likely to be true for social behaviors and intellectual abilities—skills often affected in children with autism, says University of Washington psychologist Geraldine Dawson, PhD. Therefore, if researchers can learn how to diagnose autism in children at birth or several months after, they may be able to design interventions that derail the disorder before it develops, some researchers speculate.

The development of tools for earlier diagnosis of autism is moving quickly, say researchers. This summer, the National Institutes of Health (NIH) hosted a conference on autism diagnosis to discuss the state of the science and to begin to develop practice guidelines for diagnosing autism. And, as geneticists begin to locate potential 'autism genes' (see article on page 13), behavioral researchers are fast developing behavior-based diagnostic tools that will work at increasingly younger ages.


Autism is difficult to diagnose in very young children because several of the disorder’s main symptoms—such as the child’s relationship with peers—involve behaviors that don’t fully develop in children until later in childhood, says Fred Volkmar, MD, of the Yale Child Study Center. According to the Diagnostic and Statistical Manual of Mental Disorders-IV (DSM-IV), the essential features of autistic disorder include trouble interacting with others and a tendency to have narrowly focused and odd interests, as with the boy who was fascinated by watches to the exclusion of anything else. Most researchers admit that these traits are the hallmarks of autism, but they also say that they don’t help much when trying to diagnose autism in children younger than age 2 or 3.

Diagnosis of autism is also difficult because the best early indicators involve the absence of consistent social and communication behaviors rather than the presence of an abnormality. 'There could be lots of reasons you’re not seeing a behavior,' says autism researcher Wendy Stone, PhD, of Vanderbilt University School of Medicine. 'You could not be seeing it because it’s not developing, or it could be that the child is not showing it to you at that time.'

University of Chicago researcher Catherine Lord, PhD, has tried to address this issue by designing an observational diagnostic tool—the Autism Diagnostic Observation Schedule-Generic—which creates situations that attempt to elicit certain behaviors in children such as pointing or asking for something. Autistic children, researchers find, often fail to draw others’ attention to objects by pointing and tend not to ask for things in the same way as other children, says Lord. Her diagnostic technique also includes a parent interview, called the Autism Diagnostic Interview–Revised, which asks parents about how the child acts in typical situations.

Lord’s longitudinal work suggests that her tools can diagnose children as young as 18 months, she says. The diagnosis is somewhat inconclusive when children are this young, but becomes more stable at age 3 and again at age 5, she says.

Research by Lord and others finds that social and communication behaviors, such as imitation and use of gestures, are most indicative of autism in very young children, says Stone, who reviewed the research on early behavioral indicators of autism.

Based on her own research, Stone is developing a screening tool that she hopes will allow clinicians to check for autism in children age 2 and possibly younger.

In a study that followed a group of children from age 2 to 4, she identified three specific skill areas, all in the realm of social and communication skills, which appear to indicate autism:

• Play. Children with autism fail to engage in reciprocal play, where there is a back-and-forth between two people (as when rolling a ball) and functional play, where children play with toys such as dolls or cars in the way intended.

• Motor imitation. Children with autism copy others’ motor movements far less than other children.

• Joint attention. Unlike most children, children with autism don’t often use pointing or other techniques to direct another person’s attention.


What is particularly striking about children with autism is how different they are from each other, say researchers such as Stone and Grace Baranek, PhD, at the University of North Carolina, Chapel Hill. In an attempt to delineate behaviors that distinguish children with autism from other children, Baranek examined early home videos taken when the children were 9 months to 12 months old—well before they were diagnosed with autism. When she compared the videos with home videos of children with other developmental disabilities and children with no disabilities, she found that the children with autism didn’t all have the same symptoms but instead showed a pattern of behaviors as a group.

Overall, the autistic children showed subtle problems in sensory attention and arousal. In particular, they oriented less to visual information in their environments; they put objects in their mouths more often; they needed more cues before they would look when someone called their names; and they pulled away from social touch slightly more than either of the other groups of children.

But 'you can’t tell these kids apart by glancing at their home movies—it takes a detailed analysis of their subtle behaviors,' adds Baranek. In fact, at the study’s end, the students who coded the specific behaviors from the videos were unable to reliably determine whether a child had autism, another developmental disability or no disability.

Even trained clinicians have a hard time seeing behavioral problems in very young children later diagnosed with autism, says Dawson of the University of Washington. She’s conducted three studies using home videos, and in each she asked experienced pediatricians to look at the tapes and make a qualitative judgment about which children had autism. When the children were 1 year old or older, the pediatricians did reasonably well. But when the children were 8 months to 10 months old, the pediatricians were unable to tell the difference.

In fact, even trained video coders had a hard time finding differences in behaviors of the youngest children. In Dawson’s first home video study, she examined videos of babies’ first birthday parties and, similar to what others find, identified four behaviors that distinguished the children with autism from the children with no developmental disability: failure to make eye contact with others, to point, to show objects to others and to orient to one’s name being called.

When she examined videos of children 8 months to 10 months old, the inability to make eye contact with others and to respond to one’s name continued to distinguish the children. But joint attention behaviors, such as the inability to point and to show, did not. This isn’t surprising, says Dawson, since such behaviors are not typical of children much younger than 1, but it limits the number of criteria clinicians will be able to use for diagnosis at such young ages, she adds.


If researchers develop behavioral tools for screening and diagnosing children with autism at age 2 and younger—and most researchers believe they’re close—they will face a seeming paradox. How can they prove their early diagnosis was right if the children they diagnose at 18 months enter into interventions that help treat their dysfunction?

'It’s a huge problem,' says Stone. 'You don’t know if you were wrong at age 2 or whether the intervention worked so well that your diagnosis is no longer valid.'

That’s why most researchers hope they will find biological or genetic markers for autism that could accurately diagnose autism at birth and would bolster the behavioral measures. Some teams are well on the way toward finding one or several autism genes, and others are using neuroimaging to search for specific anatomical or functional differences in the brains of people with autism as well as for biochemical markers. However, having something that is solid enough for diagnosing the disorder is still a long way off, agreed researchers at the NIH 'Working conference on the state of the science in autism: screening and diagnosis' held in June. That conference, sponsored by the NIH Autism Coordinating Committee, which represents several NIH institutes, and the Office of Behavioral and Social Sciences Research, brought together researchers, clinicians and representatives of parents’ groups to review the research on autism diagnosis and draft an outline for autism screening and diagnosis practice guidelines. Although NIH is not in the business of setting diagnostic criteria or practice guidelines, it tries to help professional organizations do so when appropriate, said National Institute of Child Health and Human Development Director Duane Alexander, MD, as he welcomed people to the conference.

And it’s time to do so for autism because it is so often mis-diagnosed or mistaken for other developmental disorders. Also, although the DSM-IV provides diagnostic criteria, it doesn’t provide clinicians with guidelines on how to perform initial screenings or on which tools to use to measure behavior.

Several professional organizations, including the American Academy of Neurology, the American Academy of Pediatrics, the Society for Research in Child Development and the American Academy of Child and Adolescent Psychiatry, fully backed the initial conference. And many others, including APA, were invited to participate. The same professional groups are invited to meet this winter to review a draft of the proposed guidelines and negotiate a final product they can all agree on and endorse, says Pauline Filipek, MD, co-chair with Marie Bristol-Power, PhD, of the first meeting, and chair of the next meeting.

'It is very important to get everyone on board and on the same wavelength,' says Filipek, a child neurologist at the University of California, Irvine. That will be the only way to ensure accurate and early diagnosis of autism.

Back to Special Interests | Back to Top

©2007 Go North Media