White Matter microstructure effect in ADHD: a two-sample mendelian randomization study

Introduction Genome Wide Association Studies (GWAS) revealed the highly polygenic architecture of Attention-Deficit/Hyperactivity Disorder (ADHD) and highlighted the contribution of common variants related to brain development and function. In parallel, several imaging studies attempted to discover disorder-related brain structures, with some significant findings concerning white matter. Two-sample mendelian randomization (2SMR) is a powerful tool to evaluate causality between two phenotypes using summary statistics data. We aimed to investigate a possible causal relationship between white matter genetically predicted variation and ADHD diagnosis through 2SMR. Methods A unidirectional two-sample MR analysis was performed based on summary statistics of GWAS between 22 different white matter (WM) mean fractional anisotropy measures and ADHD. We used 4 different MR approaches, considering IVW random effects as the main analysis, followed by several sensitivity analyses. Linkage Disequilibrium Score Regression (LDSC) was evaluated in the same set of samples to corroborate the direction of associations. Results and Discussion Our most consistent finding across MR and LDSC approach, following the sensitivity analyses, indicate that the decreased WM microstructure integrity of the fornix stria terminalis (FXST ivw beta:-0.266 SE:0.083 p FDR: 0.021) genetic liability has a causal influence on ADHD diagnosis. The FXST is formed by connection fibers inside the limbic system, which is crucial to emotional processing, learning, and memory, functions usually impaired in ADHD. Therefore, this study increases knowledge concerning ADHD neurobiology and provides novel evidence of the causal effect of WM integrity in the limbic system, which could contribute to the advances in additional diagnostic tools as well as pharmacological brain structure targets.