The “holy grail” of autism research is to be able to diagnose it many months before its symptoms are clear, giving doctors time to change its course before it has seeped into and resculpted the brain of an affected child. Researchers disagree over whether this will ever be possible, but Klin has little doubt. “There is no question we will be able to attenuate autism,” he says. He is interested in the social mind — social cognition — and how it is disrupted in children with autism. One path toward possible early diagnosis of autism is to use eye-tracking studies to see and measure social engagement. How often, for example, does an infant look at the face of an interviewer? Some children have abnormalities in visual scanning in normal, social settings. Such studies could reveal vulnerabilities for autism in the first months of life, a year or two before the disease begins to become apparent. They might also help predict the degree of impairment and social disengagement that at-risk children can expect.