Market Brief

New

Daily market recaps with key events, stock movements, and global influences

Listen to this article
Economics
Business

New screening tool uses AI to detect autism in toddlers, but raises concerns about accuracy and overdiagnosis

Authors
  • Nature Medicine
  • Medical Xpress
  • Benjamin Ryan

6 posts • GPT (3)

Published

Researchers have developed a screening tool that uses computer vision and machine learning to analyze autism-related behaviors, potentially aiding in the detection of autism. The tool, known as digital phenotyping, was created to identify autism in children as young as 17-18.5 months old. However, the app incorrectly identified 34 typical children as autistic, while correctly identifying 5 autistic children. Additionally, a study found that nearly 40% of children diagnosed with autism spectrum disorders as toddlers no longer met the criteria for diagnosis by age 6.
Autistic individuals were also found to have an increased risk of chronic physical health conditions. Furthermore, a link between a plastic additive and autism and ADHD was discovered. The study also revealed that a quarter of teens with autism go undiagnosed. However, the quality of evidence in autism research has been criticized, particularly in relation to Applied Behavior Analysis therapy.

More breaking stories on DeepNewz — updated live.

Sources

    Michelle Dawson
    Michelle Dawson
    @autismcrisis

    Among N=9 autistic adults with "low financial wellbeing", N=2 (22%) have autistic children; among N=12 autistic adults with "high financial wellbeing," N=5 (42%) have autistic children--total number of autistic children is N=3 & N=7, respectively https://t.co/a0z3PmTYHW free

    Nature Medicine
    Nature MedicineTwitter verified badge
    @NatureMedicine

    Digital phenotyping could help detect #autism: Researchers developed a screening tool that uses computer vision and machine learning to analyze autism-related behaviors. Read the News and Views from Catherine Lord & Rujuta Wilson @dgsomucla https://t.co/J5Ndv8zCZA

    Michelle Dawson
    Michelle Dawson
    @autismcrisis

    An app for early autism screening? For the youngest age range (17-18.5 months; N=164), N=5 autistics were correctly identified as autistic, vs N=34 typical children incorrectly identified as autistic?--see Table 3 https://t.co/cOTSoJbxeW big COIs, free

Similar Stories