Abstract
Autism spectrum disorder (ASD) affects many aspects of development (e.g., social,
cognitive) throughout life, and altered neurodevelopment in early life is considered an
important feature of ASD pathophysiology. There is, however, considerable uncertainty
regarding the neuromaturational trajectory of ASD beyond early childhood—into the
transitional periods of adolescence and aging. White matter is necessary for the
development of specialized brain circuits. White matter tracts follow a protracted
developmental trajectory, peaking in the third to fourth decades of life, and, critically, their
maturation may be differentially altered in ASD. Diffusion MRI (dMRI) is currently the most
effective non-invasive technique for probing white matter in vivo. Studies of white matter
using dMRI in ASD in later childhood to adulthood, however, have provided inconsistent
findings. A major hindrance to reproducible white matter results in ASD is low statistical
power. Over the last decade, extensive shared multimodal ASD datasets have been
made available through large initiatives such as the Autism Brain Imaging Data
Exchange, and National Database for Autism Research. The dMRI data from these
resources, however, are currently underutilized. In addition, large consortia studies (e.g.,
the Human Connectome Project) have made high-quality dMRI data available in healthy
individuals across the lifespan. Recent methodological advances in data harmonization
of dMRI data allow mitigating differences across imaging sites and protocols. This project
aims to leverage available multi-site dMRI data (>2600 participants; >600 ASD, >2000
TD; from up to 17 sites) to perform joint large-scale analyses that will elucidate the
spatiotemporal pattern of white matter atypicalities in ASD across the lifespan (age 2-70
years). We propose to i) Aggregate, process, and quality control dMRI datasets of
individuals with ASD and TD controls from shared databases, and ii) Examine cross-
sectional lifespan age trajectories of microstructural measures of white matter tracts at
the group level, and multivariate patterns of white matter microstructure at the individual
level (using deep normative modeling), in ASD relative to TD controls, and link atypicality
of white matter patterns with ASD symptoms. Exploratory analyses on the effects of sex
and cognitive abilities will also be performed. This research may aid treatment and policy
development by identifying the location and timing of vulnerable neural pathways and
circuits in ASD, and, additionally, promote reproducible diffusion MRI research.
