A new AI model can detect autism in children under the age of two.
In a study from the Karolinska Institutet, it turns out that the model has an accuracy of almost 80 percent.
– This can drastically change the conditions for early diagnosis and interventions, and in the long run improve the quality of life for many individuals and their families, says Shyam Rajagopalan, researcher at Karolinska Institutet.
The new machine learning model can detect autism in children early and this based on relatively limited information. In the study conducted by the Karolinska Institutet, it was found that the AI model could detect almost 80 percent of all people with autism.
– With an accuracy of almost 80 percent for children under the age of two, we hope that this will be a valuable tool for healthcare, says Kristiina Tammimies, docent at KIND, Department of Women’s and Children’s Health, Karolinska Institutet and last author of the study.
The AI model found 80 percent of the children with autism
Using 28 different parameters, four different models were developed. One of the models, called the AutMedAI model, succeeded in identifying about 80 percent of children with autism. In total, it analyzed 12,000 individuals. The work has been based on an American database of around 30,000 individuals, some of whom have autism spectrum conditions and some who do not.
According to the researchers, the advantage of being able to detect autism early in children is that interventions can be implemented in good time.
– Now we can talk to the parents about the fact that there is a combination of factors that make the child more likely to develop autism, says Kristiina Tammimies and adds:
– And then forwarded to the right care.
Hope to be able to use it in healthcare
The plan is now to improve the model so that it can eventually be used as an aid in healthcare.
– I think we can hopefully use it within the BVC and help to really identify the children who will go on to an investigation of autism, says Kristiina Tammimies.
At the same time, they want to emphasize that the AI model is not intended to replace the clinical investigation for autism, but rather should be seen as a valuable aid.