'A cognitive neuroscience approach to the early identification of autism'

March 19, 2014


Charles Nelson, PhD
Research Director, Division of Developmental Medicine, Boston Children's Hospital

There is wide support for the notion that the outcome of children diagnosed with autism is superior among those who receive early intervention. However, early intervention is predicated on early identification, and at the present time, the average age of diagnosis is 3+ years. For a number of years now an international consortium of more than 18 sites has been studying infants at high risk for developing autism by virtue of having an older sibling with the disorder (placing the infant’s chances of developing an ASD at 1:5 vs. 1:88). Despite great promise, there are few behavioral signs of the disorder that are predictive in infants less than 12-18 months. One reason for the failure to identify signs in younger infants is how limited the behavioral repertoire is in infants less than 12 or so months. In the program of research I will talk about, we hypothesize that examining the brain directly, bypassing behavior, may prove more efficacious, as changes in behavioral development typically follow changes in brain development. In this context I will describe select findings from a longitudinal study focused on tracking high risk infants from 3-36 months of life, using a battery of EEG, ERP, and fNIRS tasks. Two clear patterns appear to be emerging from our data thus far: that errors in neural circuitry, inferred from EEG and fNIRS data, appear by 6 months of life. Second, preliminary findings suggest that there are neural “signatures” that appear by 6 months of age that distinguish infants who will subsequently go on to develop autism from those who do not.

Key Words: Autism, Early Identification, Connectopathy, EEG