Children with vs. without ADHD may have modest structural brain differences

Children with ADHD exhibited modest differences on structural brain measures compared with peers without ADHD, according to results of a cross-sectional population-based study published in The Lancet Psychiatry.

“The development of large datasets, such as the Adolescent Brain and Cognitive Development (ABCD) Study, offers new opportunities to explore the structural correlates of ADHD,” Joel Bernanke, MD, of New York State Psychiatric Institute, and colleagues wrote. “The ABCD Study is the largest longitudinal study of brain development in the U.S. The recruitment strategy of this study, including the locations of the 22 participating sites, was designed to obtain a locally random cohort of children aged 9–10 years, representative of the sociodemographic features of the population in the U.S.”

Using the ABCD sample, Bernanke and colleagues compared structural brain measures among children with (n = 949) and without (n = 9,787) ADHD recruited between Sept. 1, 2016, and Aug. 31, 2018, with ADHD diagnosed via parent report of symptoms. Specifically, they examined participants’ baseline demographic, clinical and neuroimaging data. Researchers used linear mixed effects models to estimate Cohen d values linked to ADHD for 79 brains measures of cortical thickness, cortical area and subcortical volume. Further, they employed a novel simulation strategy to examine the ability to pinpoint significant effects despite possible diagnostic misclassification.

Results from the full model that included potential confounding variables selected a priori showed 11 significant differences across 79 measures after correcting for false discovery rate. These differences indicated reductions in brain measures among those with ADHD. Bernanke and colleagues reported small Cohen d values that ranged from –0.11 to –0.06. Use of a more restrictive comparison group or alternative diagnostic methods did not significantly change these values. Even when significant diagnostic misclassification was present, simulations still had enough statistical power to detect differences.

“Our results suggest that structural brain differences in a general sample of youth in the U.S. are smaller in scope and magnitude than previously reported in clinical populations,” Bernanke and colleagues wrote. “Future studies might need to incorporate other MRI modalities, novel statistical approaches or alternative diagnostic classifications, particularly for research aimed at developing ADHD diagnostic biomarkers.”

In a related editorial, Jeffrey M. Halperin, PhD, of the department of psychology at Queens College in New York, noted a potential limitation of this research.

“Large epidemiological samples like this one often represent a strength in research,” Halperin wrote. “However, for ADHD, in-depth clinical assessments are difficult to accomplish, and the acquisition of teacher reports, which are highly recommended in clinical settings, is limited (here <30% had teacher ratings). As such (and acknowledged by the authors), the degree to which the findings can be generalized to patients with ADHD is hard to determine. Fortunately, the authors provide data to allow for some inferences on this topic.”