Parents were able to identify those trajectories, indicating impending ASD, with prospective reports, "but not as well retrospectively" with the Autism Diagnostic Interview-Revised (ADI-R), which was accurate in only 30% of cases," said Meghan Miller, Ph.D., a postdoctoral fellow at the University of California, Davis, MIND Institute. These results suggest that while parental reports of regression on the ADI-R can be accurate, they might only pick up "the tip of the iceberg," she noted.
The findings, based on a small sample, raise concerns about the use of the ADI-R and other retrospective measures in research, but do provide "some hope for early screening using parent report data," Dr. Miller pointed out. Such reports would be more feasible than neuroimaging, eye tracking, and other sophisticated testing that could be used over several well child visits, to detect declines and help determine which children should be further evaluated further or referred for interventions, she added. Phase III of the study is currently underway.
In other sibling studies, Dr. Saulnier discussed the use of eye-tracking testing in studies that prospectively followed high-risk infants whose siblings have been diagnosed with ASD and low-risk infants with typically developing siblings.
Under eye-tracking methodology, the child watches prerecorded movies on a monitor of women acting as caregivers, singing, and talking to an infant. It is well established that when viewing a social scene, typically developing adults focus on the eye region of the face, but those with autism look less at the eyes and significantly more at the mouth, body, and objects in the social scene. Those findings have been replicated in school-aged children, adolescents, and toddlers down to aged 2 years, she noted.
In infant studies, eye-tracking testing is performed at least 10 times from birth to age 3, with vocal recordings and clinical assessments at different periods.
Summarizing the findings, Dr. Saulnier said infants at low risk who have a typical outcome show a steady fixation of eye gaze from birth. But infants who are at high risk for ASD and "develop ASD by age 3 show a rapid decline in eye fixation between 2 and 6 months of age that is more predictive of ASD than any clinical measure."
Autism cannot be diagnosed with this test alone, she said. (Most of this work was conducted at the Yale Child Study Center, before she moved to Emory, she said.)
She provided examples of a 5-month-old low-risk infant, who developed typically and focused at the eye region of the face, contrasting with another 5-month old infant, who went on to develop autism by agd 3, who was not looking at the eye region of the face, focusing on the mouth, body, and objects.
Another example she showed was an infant who had normal results at 3 and 6 months, but by 9 months, the researchers started to see a shift, which she described as the "unfolding of autism." This child was diagnosed with autism at 12 months. She added, however, that it is rare to see that symptomatology so early and pronounced.
While it would be assumed that the children who go on to develop autism manifest the lack of eye fixation from birth, another finding from eye fixation studies is that the opposite is true, and "babies who went on to develop autism actually had significantly more eye fixation in the first 2 months of life," Dr. Saulnier said.
Typically developing infants maintain this focus on the eyes, "but whatever is happening in autism to make social monitoring not as salient, they’re trailing off, and you’re seeing this derailment in this eye fixation."
Other results of eye-tracking sibling studies might provide information about a window of opportunity – at about age 9-12 months – for an intervention "that could capitalize on the potential for resilience, she noted.
In infants whose older siblings have ASD and are either unaffected or have "shadow symptoms" of autism at age 3 years, eye-tracking results indicate that after reduced eye fixation, they exhibit a "course correction" in increased eye fixation that appears to start around 9 months of age.
Why this occurs is not understood, but this observation sheds some light on a possible window of time during which something could be done, before autism fully unfolds, such as coaching strategies for parents on keeping their child engaged "to produce a course correction if it wasn’t naturally going to occur," Dr. Saulnier said.
She and her associates are now conducting a randomized, controlled study of infants at risk at 12 months, evaluating a potential way to correct the course.