Podcast: Play in new window
BOB HIRSHON (host):
Infant autism prediction. I’m Bob Hirshon and this is Science Update.
Early intervention can make a big difference in kids with autism, but symptoms don’t usually emerge until the age of two. Now, researchers report in Science Translational Medicine that they’ve developed mathematical algorithms that correctly predict whether 6-month-old-infants at high risk will go on to develop the disorder.
ROBERT EMERSON (University of North Carolina, Chapel Hill):
Usimng machine learning, we were able to identify 82% of the infants who went on to have autism, and that was 9 out of 11 of the infants in our study, and we were able to correctly identify all of the infants who did not develop autism.
HIRSHON:
That’s University of North Carolina, Chapel Hill cognitive neuroscientist Robert Emerson. He says the technique looks at connection patterns between different brain regions. It could be used to screen infants for interventional clinical trials after larger studies replicate the findings. I’m Bob Hirshon, for AAAS, the science society.
Story by Susanne Bard