Early identification and interventions for autism are important and scientists have identified certain behavior patterns in very young children that could signal autism. These signs are often missed by parents and doctors because the differences in young children with and without autism are very subtle. Testing infants and young children include measuring the time it takes to react to stimulus and the interaction between child and caregiver.
Engineers at Duke University have developed a software to help doctors, schools and parents identify potential problems. This software analyzes videotaped screening tests to pick up subtle signs of autism. For example, screening tests might include:
- Drawing a child’s attention to a toy held out to their right and then moving the toy to the left with the medical professional measuring how long it takes for the child to shift attention as the toy moves
- Moving a toy out of a child’s vision and looking for delays in the child tracking the toy
- Rolling a ball to the child and looking for eye contact to indicate interaction between the child and the caregiver
When screening for autism, the clinician performing the test must pay attention to the child’s reactions as well as measure the time it takes for the child to react. This can be difficult and is not always accurate.
The software program developed at Duke does the counting, measuring and analysis. While it doesn’t replace the clinician, it provides accurate and valuable information about whether a potential problem exists. The software can analyze eye gaze, walking patterns and motor behaviors.
Tests on the software have shown it to be as good as a clinician in spotting signs of autism and better than non-medical professionals, such as parents and schools in detecting subtle signs of autism.
Despite the success with the software program, the engineers are already moving on to create even more practical ways of screening for autism. They are working on creating an app that could be used on any computer or tablet. A child would need only sit in front of the device for a few minutes. The app would look for facial, physical and visual cues and provide a report of any potential problems. This program would make screening for autism available to parents, teachers and clinicians. The engineers are working with experts at Duke Medicine to come up with simple tasks that can be used on a computer and tablet.
This research was published in the journal Autism Research and Treatment, May 2014.