Diagnosing autism spectrum disorder in children is difficult, but that info can give families a wealth of new support options, not to mention a helpful new perspective.
Thanks to some infant-focused AI, doctors may soon be able to reach that diagnosis for their little patients at a very early age.
Researchers at the University of North Carolina, Chapel Hill have developed a high-tech method for detecting signs of autism spectrum disorder in children as young as six months.
Based on decades-old research that connects brain volume to autism, the technology analyzes scans of babies' brains to spot physical clues of this condition during development and has demonstrated a high rate of success with its predictions.
As IEEE Spectrum explains, the team trained its algorithm on scans of infant brains, which can reveal telltale increases in the brain's surface area when children are under a year old, and which often precedes an overgrowth in brain volume that scientists have previously linked to autism spectrum disorder.
Researchers then deployed the AI as part of the Infant Brain Imaging Study, a collaborative project on autism and early brain development funded by the U.S. National Institutes of Health [PDF].