For the first time, researchers used machine learning on brain MRIs of six-month olds to correctly predict 81% of babies who would later meet criteria for autism at two years of age (positive predictive value 100%, negative predictive value 96.0%) published in Science Translational Medicine.

See the source article by following the link below:
For the first time, researchers used machine learning on brain MRIs of six-month olds to correctly predict 81% of babies who would later meet criteria for autism at two years of age (positive predictive value 100%, negative predictive value 96.0%) published in Science Translational Medicine.