Evaluation of white matter microstructure in childhood ADHD is crucial for examining the stability and replicability of white matter microstructural abnormalities in ADHD across development, in order to promote the identification of early biomarkers of the disease. To that end, our results confirm that microstructure in long-range white matter pathways shows abnormality, even prior to adolescence, in a sample that is largely medication naïve and free of major comorbidity (except ODD, which did not account for results). Specifically, childhood ADHD was associated with reduced FA in bilateral temporo-occipital, and corona radiata, left cerebellar, and right fronto-parietal white matter pathways, clusters of which were broadly associated with symptoms of both inattention and hyperactivity, with many relating most significantly to symptoms of inattention. In addition, with the youngest DTI study of ADHD to date, we also confirmed that pre-adolescent children with ADHD possess abnormal white matter microstructure in later developing bilateral frontal-limbic projection fibers. Notably, several of these white matter regions are anatomically close to those that have been identified previously as being atypical in older ADHD youth,21, 24, 27
and thus, our findings provide converging evidence for widespread, rather than localized, atypical white matter microstructure in ADHD. As higher FA reflects greater coherence in white matter, given restriction of diffusion in dense, organized, myelinated white matter pathways,15, 43
these findings suggest diminished white matter organization or integrity in children with ADHD.
Although white matter perturbations were widespread, the tracts involved are interesting in relation to the behavioral correlates of ADHD, as the cortical regions connected by these tracts have been implicated in ADHD and associated cognitive dysfunction (for review, see 11, 12
). Temporo-occipital and fronto-parietal white matter are similar areas to which increased FA has been correlated with aspects of attention-related cognition including response control,44
and mental arithmetic. Given this, as well as work suggesting a positive relationship between white matter integrity and brain activation,46, 47
and the fact that long-range functional connections between brain regions strengthen across development,48
our findings suggest that ADHD involves altered maturation in these networks. These findings converge with prior structural and functional imaging data suggesting that both cortical and subcortical structures are affected in ADHD and may help to explain prior findings of atypical brain activation,49-51
and functional connectivity in ADHD.52, 53
Future studies will be instrumental in clarifying these relationships.
Consistent with previous studies,21, 54
we also found significantly reduced FA in cerebellar white matter among ADHD youth. The cerebellum’s role in executive functioning is well-accepted,55
based on its bidirectional connectivity with the cerebral cortex. Cerebellar circuits have been central in neurobiological theories of ADHD,14
and the present findings confirm that connections between the cerebellum and associated structures are different in ADHD, not merely the cerebellum itself. 56, 57
The reproducibility of this finding across various age-ranges and the fact that the cerebellum is thought to begin myelination early in life (for review, see 58
), suggests that cerebellar white matter abnormalities may be a robust and stable feature of ADHD.
A novel result of this study was atypical FA and MD in fronto-limbic white matter, which was largely accounted for by significantly greater radial diffusion among ADHD youth. Given that radial diffusion reflects restriction perpendicular to myelin bundles and decreases with myelination,13
these findings suggest aberrant or delayed myelin development in ADHD. This finding is striking in light of the increasing appreciation for the role of emotional dysregulation in ADHD, 59, 60
and the fact that frontolimbic white matter is among the latest to mature. 18, 19
Thus, it may be that these pathways show an altered developmental trajectory in ADHD that is no longer detectable in older children or adults with the disorder. These findings highlight the importance of examining the ADHD brain prior to adolescence, as continued experience and maturation may confound the early developmental picture.
In addition to increased MD in frontolimbic white matter, we identified two clusters of reduced MD in ADHD. In both clusters (the superior longitudinal fasciculus and posterior limb of internal capsule), axial and radial diffusivity was lower in ADHD than controls. Silk and colleagues29
also showed reduced radial diffusion in white matter of youth with ADHD, but in association with increased (not reduced) axial diffusion. Additionally, a DTI study of youth with phenylketonuria (PKU), a childhood disease associated with ADHD-type symptoms and executive dysfunction, also reported reduced MD in patients relative to controls.61
While atypical myelination may contribute to the current findings, another potential explanation is that differences in the distribution of axonal fiber orientations exist, such that diffusivity is more restricted when averaged over all orientations within ADHD white matter than is the case for controls. Future analyses that account for heterogeneity in microscopic structure in individual voxels may be of utility.62
Although a strength of the current study is that nearly all children with ADHD were medication naïve, effects of stimulant medication on brain development remain unclear.3, 63
Second, although this study is slightly larger than nearly all prior DTI studies of children with ADHD, power remains low. Relatedly, for the current study, we pooled ADHD subtypes and did not compare them. This was justified on grounds that it is unclear whether DSM-IV subtypes are stable or biologically defensible in children.41
Nonetheless, subtypes may contain important biological information, a point we intend to investigate with larger samples. Likewise, although all results presented here are independent of gender, it remains necessary to determine if different developmental white matter trajectories are present in boys and girls with ADHD.
We conclude that, even by 7-9 years of age, ADHD is associated with widespread alterations in white matter microstructure in a complementary, developmentally important manner. Abnormal microstructure in early-maturing cerebellar and cortical-cortical pathways has now been replicated in children, adolescents, and adults and thus appears as a stable marker of ADHD. Conversely, alterations in late-maturing frontolimbic pathways, seen in the current study but not in older samples, may indicate an early, dynamic marker that can provide additional clues to the pathophysiology of ADHD. The current findings suggest the possibility that ADHD entails an altered developmental trajectory in the structural connectivity of the brain with neuroantomical biomarkers specific to developmental stage.