Longitudinal changes of ADHD symptoms in association with white matter microstructure: A tract-specific fixel-based analysis

Abstract

Background: Variation in the longitudinal course of childhood attention deficit/hyperactivity disorder (ADHD) coincides with neurodevelopmental maturation of brain structure and function. Prior work has attempted to determine how alterations in white matter (WM) relate to changes in symptom severity, but much of that work has been done in smaller cross-sectional samples using voxel-based analyses. Using standard diffusion-weighted imaging (DWI) methods, we previously showed WM alterations were associated with ADHD symptom remission over time in a longitudinal sample of probands, siblings, and unaffected individuals. Here, we extend this work by further assessing the nature of these changes in WM microstructure by including an additional follow-up measurement (aged 18 - 34 years), and using the more physiologically informative fixel-based analysis (FBA).

Methods: Data were obtained from 139 participants over 3 clinical and 2 follow-up DWI waves, and analyzed using FBA in regions-of-interest based on prior findings. We replicated previously reported significant models and extended them by adding another time-point, testing whether changes in combined ADHD and hyperactivity-impulsivity (HI) continuous symptom scores are associated with fixel metrics at follow-up.

Results: Clinical improvement in HI symptoms over time was associated with more fiber density at follow-up in the left corticospinal tract (lCST) (t_max = 1.092, standardized effect[SE] = 0.044, p_FWE = 0.016). Improvement in combined ADHD symptoms over time was associated with more fiber cross-section at follow-up in the lCST (t_max = 3.775, SE = 0.051, p_FWE = 0.019).

Conclusions: Aberrant white matter development involves both lCST micro- and macrostructural alterations, and its path may be moderated by preceding symptom trajectory.

Keywords: Attention deficit hyperactivity disorder; Diffusion imaging; Magnetic resonance imaging; Microstructure; White matter.

Fig. 1

Schematic of how this study chronologically relates to previous studies, the samples included in each, relevant clinical and neuroimaging measurements, study sample age ranges, mean years (standard deviation) in between each acquisition wave, and the analysis methods used. The present study is a fixel-based analysis of W1 to W2 and W2 to W3, using only the models in which we found significant effects in a previous voxel-wise tract-based spatial statistical analysis of W1 to W2.

Fig. 2

Change in participant characteristics from Wave 1, to Wave 2, to Wave 3. Longitudinal data points are connected by a line. Note that participants at W3 were selected on the basis of their history of ADHD diagnosis, so W3 tends to differ quite markedly from the other two waves, which also include never-affected controls. This conceals the typical pattern of average symptom remission that would be expected in a follow-up study without this selection criterion.

Fig. 3

Tract-specific region-of-interest masks of the left corticospinal tract for Wave 2 (top) and Wave 3 (bottom) colored by direction (red: left–right, green: anterior-posterior, blue: inferior-superior). (A) Fixel mask overlaid on a single representative coronal slice of the study-specific white matter fiber orientation distribution template. Template contrast was adjusted and fixels have been thickened for visualization. (B) Coronal, (C) sagittal, and (D) axial views of the tract reconstruction from TractSeg (applied to each fiber orientation distribution template) and displayed in glass brains for visualization. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Symptom change may precede lCST WM microstructure plasticity. Top: Improvement of HI score is associated with more follow-up fiber density (FD). Bottom: Improvement of combined score is associated with more follow-up fiber cross-section (FC). Streamline segments have been cropped from the template tractogram to include only streamline points that correspond to significant fixels for this tract (FWE-corrected p-value < 0.05). Significant streamlines are colored by the standard effect size of ‘Δscore’ on ‘FD at W2′ and ‘log(FC) at W3′ and displayed across coronal and sagittal slices of the study-specific white matter fiber orientation distribution templates.