AI detects autism in young children up to 81% accurate

Experts at the University of North Carolina (USA) have developed an artificial intelligence module (AI) that helps detect autism in babies with an accuracy of 81% compared to 50% of traditional methods. .

This result comes from the study of brain images of babies, which focus on understanding the brain development process of children with autism. By scanning the brain to identify three main factors: the surface area of ​​the brain, the child's brain volume and sex (boys have a higher risk of autism than girls). The follow-up period will be between 6 months, 1 year old and 2 years old.

For every 10 children with autism, there are 8 cases that this technology can detect.

Dr. Heather Hazlett - a psychologist and brain researcher said: "In previous research, we have found an increase in brain volume in autistic children, meaning that autism patients have a set of The brain is bigger than the average, so this time we decided to monitor brain changes in the first 2 years of life, usually the biggest change is between 1 and 2 years old. When the surface area increases again, it is 6-12 months old, the brain wrinkles will be visible at this stage ".

Autistic children have larger surface area and brain volume.

This study aims at early measures for the special needs of children. For every 10 children with autism, there are 8 cases that this technology can detect.

The immediate difficulty is that this technology is still expensive, can not be widely deployed and needs long time to monitor each change. But it is likely that this tool will soon be completed and expanded for more people to access.