There are a lot of myths and misunderstandings about autism. Simply put, it is a physical disorder associated with brain biology and chemistry that impairs an individual’s social and communications skills. The exact cause is unknown, but it is generally believed that genetic factors play a large role.
Early diagnosis is extremely important. As there is no ‘cure,’ treatments are behavioral and run over the course of years. Starting therapy early is a huge benefit to the patient’s quality of life. That said, it is a difficult and expensive process to evaluate a toddler for autism, one that requires expert monitoring over a long period of time. But what if that could be automated?
That’s where Jordan Hashemi and his team at the University of Minnesota come in. Using video footage of at-risk toddler behavior, which families can record at home or at a clinic, they have developed a computer vision system that analyses the movements of children as young as two years old and identifies behaviors linked with autism.
By tracking position and movement of the child’s head, torso, arms and legs, the program can recognize the restricted, repetitive behaviors that are indicative of autism. For example, autistic children tend to walk on their toes, hold their limbs in asymmetrical positions while moving, and show delays in tracking objects in their field of vision.
Hashemi’s system looks for these behaviors automatically and has performed well in tests. The research team videoed 15 children who were identified as at risk for autism and showed the footage to a specialist in child autism, a child psychiatrist, two psychology students and their system. Taking the specialist’s opinion as the standard, the computer system did better than both the psychiatrist and the students in that it agreed more often with the specialist’s assessment. Pretty impressive for a first generation technology.
Of course the system requires lots of additional testing. But that shouldn’t be too hard for Hashemi and his team; they can use archival footage of autistic children whose condition is already known to calibrate the system’s assessments. Once perfected, this technology could greatly reduce the time required to diagnose autism and other developmental disorders. Of course a human expert would need to review the conclusions, but the time saved in diagnosis would lead to earlier interventions and improved quality of life for patients.
In more common circumstances, a future system could monitor any young child who has missed developmental milestones. Because doctors can’t watch an infant for significant periods of time and parents might not be able to expertly judge their baby’s behavior, a pediatrician might recommend that a baby with ‘red flag’ behavior (like not smiling or reacting to a parent’s face at three months) have a webcam installed on their crib. This would send video to a center with an advanced version of Hashemi’s system that would monitor and compare the baby’s behavior with normal developmental benchmarks. The parents could interact with this monitoring as well, moving objects across the baby’s field of vision to see if it can track them or if it reacts to human speech with excited gestures, as babies of a certain age should.
There would be privacy issues involved, but when the early intervention of developmentally delayed children can prevent lifelong struggles and bring them ‘up to speed’ with their peers, the potential benefits would likely outweigh any of the associated stigmas.
PSFK has partnered with Boehringer Ingelheim to bring you a steady stream of inspiring news and ideas in the health and wellness space. Once each week, we will be posting an article on PSFK.com. If you would like to gain access to the full stream of content, please check out Boehringer Ingelheim’s Facebook page, where they are publishing a regular stream of inspiring and informative content.