Abstract
Attention-deficit/hyperactivity disorder (ADHD) is the most common neurodevelopmental disorder in the U.S. Amphetamine (AMP), methylphenidate (MPH), and nonstimulants (NS) used to treat ADHD are among the most prescribed pediatric drugs. Although pharmacotherapy is the most effective treatment, its neurostructural effects remain poorly characterized. Leveraging the ABCD Study, we used a machine learning approach to identify neuroanatomical targets of medications (N = 1306), followed by linear mixed-effects modelling to estimate the effects of ADHD status and medication use (AMP, MPH, NS) on cortical thickness, surface area, and cortical and subcortical volumes (N = 8762). ADHD was not associated with significant differences in brain structure; effects were small and bidirectional. Nevertheless, a consistent pattern emerged in which AMP and MPH effects were opposite in direction to ADHD associations (i.e., "attenuation" toward control phenotype), while NS showed a weaker pattern. AMP and MPH use were both associated with significantly reduced right banks of the superior temporal sulcus and left posterior cingulate surface area. Additionally, AMP was associated with reduced left posterior cingulate volume and MPH with increased right temporal pole volume. Significant stimulant associations suggest partly independent structural effects in the cortex, with effects that extend beyond the ADHD phenotype.