Personal videos can help detect autism

US researchers have designed and tested eight models of automated learning to diagnose autism in children by badyzing videos at home, making the process faster and much more effective, "according to a study published in the scientific journal PLOS Medicine.

Each of the models contained "a set of algorithms comprising 5 to 12 behavioral characteristics of the children and producing a general score indicating if the child was autistic," he explained.

Wall explained that to evaluate the models, they had asked families recruited as part of the study to send home videos of one to five minutes in which the faces and hands of the children were shown, as well as their "social interaction of toys, pencils and utensils".

According to these images, 116 children aged 4 years and 10 months on average were diagnosed with autism and 46 others (with an average of two years and 11 months) were developing it, he explained. .

Nine expert reviewers badyzed the videos using a 30-question questionnaire with "yes" or "no" responses, based on typical autistic behaviors, which were then incorporated into the eight mathematical models.

The best-performing model identified 94.5% of autistic children and 77.4% of non-autistic children.

For a results check, they evaluated another 66 videos, half of them autistic children.
The same model correctly identified 87.8% of children with autism and 72.7% of those who did not have this disorder.

"Across the United States, the average waiting list for access to a standardized service (for the diagnosis of autism in the child) can go up to one year" Wall said.

The use of home videos for diagnosis has another advantage: they "take the child into his natural environment", unlike the clinical evaluation carried out in a media "that can be rigid and artificial and cause atypical behavior".

"We have shown that we can identify a small group of behavioral characteristics that are highly aligned with clinical outcomes and that non-experts can evaluate these characteristics quickly and independently in a virtual online environment in minutes," Wall said.