Predicting Autism from Cortical Surface Data: Insights from the Infant Brain Imaging Study (IBIS)
Autism spectrum disorder is a lifelong neurodevelopmental disorder that is typically diagnosed by 2-3 years of age. Despite the early age of clinical diagnosis, relatively few neuroimaging studies have focused on evaluating the neural basis of autism in very young infants and children. The identification of imaging markers of ASD that precede clinical diagnosis could have great impact in identifying infants at risk for ASD and initiating early interventions. The Infant Brain Imaging Study, an NIH funded Autism Center of Excellence network, aims to characterize that such early brain developmental trajectories in high-risk infant siblings who go on to develop ASD, those who do not and in typically developing controls, ages 3 to 36 months. I will present insights and findings from IBIS of predictive, atypical early cortical brain development in infants that go on to develop ASD. I will discuss the methods and tools development needed for these results, including the potential of the new brain imaging marker of extra-axial CSF, as well as of our novel convolutional deep learning approach on the cortical surface.