American researchers asked a group of families to film their children interact with objects and people. They tried eight automatic learning models to diagnose autism, which allows "to simplify the process and make it much more efficient," according to a study published in the scientific journal PLOS Medicine.
The study was developed by a team from the Stanford University School of Medicine and led by Dennis Wall, professor of Pediatrics and Biomedical Data Science from the city of California.
Each model contains "a set of algorithms including 5 to 12 behavioral characteristics of children and it generates a general score that shows whether the child has autism, "he explained.
How the video is treated
Wall said that to evaluate the model, they asked the families recruited for the study to send home videos for one to five minutes. where children's faces and hands are shown and "social interactions and the use of toys, pencils, and equipment" they are captured.. From these images 116 boys with an average age of 4 years and 10 months were diagnosed with autism and 46 others (with an average of two years and 11 months) developed it, he explained.
Nine expert reviewers analyzed the video using a 30 question questionnaires with "yes" or "no" answers, based on the typical behavior of autism, which is then included in eight mathematical models.
The model that offers the best results is that which identified 94.5% of cases of children with autism and 77.4% of children without autism. To verify the results they evaluated 66 other videos, half of children with autism. The same model correctly identified 87.8% of cases of children with autism and 72.7% of those who did not have this disorder.
Another advantage of using home videos is to diagnosis is that "they take children in their natural environment", unlike clinical judgments carried out in a medium "that can be rigid and artificial and cause unusual behavior". "We show that we can identify a small group of behavioral characteristics that are very aligned with clinical outcomes and that non-experts can assess these characteristics quickly and independently in a virtual online environment, in minutes," Wall said.