Mapping the structure and microstructure of the adolescent brain: Insights from the ABCD study

The Adolescent Brain Cognitive Development (ABCD) Study provides an unprecedented opportunity to characterize adolescent brain maturation and its correlates. This review synthesizes structural and diffusion magnetic resonance imaging (MRI) findings from ABCD, emphasizing methodological advances and implications for mental health, obesity, and environmental risk. We first summarize ABCD's harmonized multi-site acquisition, high completion rates, and longitudinal coverage, highlighting the impact of quality control and scanner harmonization. Structural MRI work converges on largely linear cortical thinning from late childhood through early adolescence, surface area (SA) peaks in early adolescence followed by modest contraction, and regionally patterned but small associations with cognition, puberty, sex, and psychopathology. Genetic and twin analyses show that cortical thickness (CT) and SA are highly heritable yet partly dissociable, and that stable person-specific variance is not captured entirely by genetic influences. We then review diffusion imaging, focusing on Restriction Spectrum Imaging (RSI), which decomposes the diffusion signal into restricted, hindered, and free-water compartments and provides neurite- and cell-body-sensitive indices in both white and grey matter. ABCD RSI studies reveal widespread age-related increases in restricted diffusion in cortico-striatal-thalamic circuits and demonstrate that these microstructural signals are highly heritable in some brain regions. RSI also sharpens inferences about environmental and metabolic exposures: air pollution, central adiposity, and weight gain show associations with altered intracellular diffusion, particularly in reward circuitry. Finally, we describe FEMA, a fast mixed-effects framework that enables vertex- and voxelwise modeling at ABCD scale, and outline opportunities to integrate structural, diffusion, and genomic data into developmentally informed brain-behavior models.