Substantial research effort is currently being directed at establishing the system of pathways that connect the functional regions of the normal human brain - the human connectome project. How is the human connectome altered in autism spectrum disorder (ASD)? Existing neuroimaging studies highlight changes to multiple brain structures in ASD. Because ASD is a highly heterogeneous disorder, different structures and connectional systems may be altered in different children. This project aims to better define the neural phenotypes of ASD using emerging methods for detecting atypical brain organization. Using diffusion-weighted and structural MRI on 279 children aged 2-4 (including 160 with ASD), this project will investigate the developing connectome in typical children and those with ASD. Graph theoretical analysis will quantify “communicability” and “rich clubness” as markers of distributed aspects of connectome organization, widely assumed to be atypical in ASD. Data-driven cluster analysis will also be used to subdivide ASD children based on their connectomic profiles, and the Network-Based Statistic will be used to isolate the prominent anomalies specific to each group. It is hypothesized that subgroups will show disruptions to unique brain systems implicated in the pathophysiology of the core and co-morbid features of ASD. This project offers a unique and powerful approach for using brain connectomics to bring order to the heterogeneity of ASD.