0034
Variations in tract-specific fibre density and morphology with puberty and behaviour across childhood1Department of Paediatrics, University of Melbourne, Parkville, Australia, 2Developmental Imaging, Murdoch Children's Research Institute, Parkville, Australia, 3Department of Medicine, University of Melbourne, Parkville, Australia, 4School of Pschology, Deakin University, Geelong, Australia
Synopsis
Microstructural development is dynamic throughout childhood and adolescence. Modelling these profiles requires statistical flexibility to understand complex interplays between related phenotypes such as sex, pubertal stage, and age. We leverage the fixel-based analysis framework and compute fibre density and morphology metrics in selective white matter tracts, to enable longitudinal mixed-effects modelling of multiple phenotypes. We show that longitudinal development of white matter fibre properties in children aged 9–14 dominates in posterior fibres. Increases in fibre density are associated with increases in pubertal stage and attention dysfunction, and protracted increases in fibre density are associated with greater internalising behaviours.
Introduction
The pubertal period involves dynamic microstructural development across childhood and adolescence. Posterior white matter fibres complete maturation during earlier stages of pubertal onset, and through pubertal progression rapid fibre development extends to association fibres. The pubertal period also corresponds with increased risk of developing behavioural difficulties, such as the emergence of internalising behaviours [1]. Characterising the complex interplay between related phenotypes such as age, sex, puberty, and behaviour requires statistical flexibility appropriate for longitudinal neuroimaging data.
The aim of this study was to investigate longitudinal relationships between fibre properties and phenotypic variables, using a linear mixed-effects modelling approach. To do this, we leverage the fixel-based analysis (FBA) framework [2] to derive measures of fibre density and morphology in fixels (fibre direction per voxel) traversing specific white matter tracts. We compute metrics describing fibre density (FD); and fibre morphology: using fibre cross-section (FC) and fibre density & cross-section (FDC), to evaluate how specific fibre properties vary longitudinally, and as a function of specific physical and behavioural phenotypes.
Methods
This study reports on a community-based sample of children aged 9-14 (n=130, 47 female). Parents of enrolled children completed a survey at both time-points which assessed pubertal stage, attention-deficit/hyperactivity disorder (ADHD) symptoms, and internalising behaviours using the strengths and difficulties (SDQ) questionnaire.
Image acquisition and processing
Diffusion-weighted imaging (DWI) data were acquired on a 3.0 T Siemens Tim Trio (b=2800 s/mm2, 60 directions, 2.4mm isotropic voxel size, TE/TR=110/3200 ms). Data were acquired at two time-points approximately 16 months apart: time 1 (M = 10.4, SD = .44 years old), time 2 (M = 11.7, SD = .51 years old). DWI data were processed using MRtrix3 (v3.0rc1) using a recommended pipeline [2], including data denoising and motion, eddy, and susceptibility-induced distortion correction. An unbiased longitudinal fibre orientation distribution template was generated across the two time-points, as previously described [3].
Tract identification
We manually identified 17 white matter tracts for further investigation (Fig 1). To identify each tract of interest in population template space, we: (a) registered a white matter atlas to our population template; (b) identified overlap between whole-brain tractogram and fixel mask; (c) used a defined protocol for ROI placement in regions which overlap the atlas and tractogram[4] (Fig 2). We then cropped the whole-brain tractogram map using the ROIs, and visually inspected for anatomical correctness. We converted each tractography map to a fixel map and calculated mean FD, FC, and FDC values in each of the fixel masks for each participant at each time-point.
Statistical analyses
Individual tracts were subjected to linear mixed-effects modelling to investigate the relationship between fibre properties with various phenotypic variables (Fig 3) using the package lme4 in R (v1.1.423). We set time interval between scans and subject ID as random effects [5]. We computed the 95% confidence intervals (CIs) and FDR-corrected p-values [6], where statistical significance was defined at pFDR < .05.
Results
We focus on reporting the FDC results, as this metric reflects alterations to the ‘capacity for information transfer’ across the brain. The results for the mixed-effects analysis for FDC are represented in Fig 3. The regions exhibiting significant increases in FDC are: bilateral cingulum cingulate gyrus (CCG), bilateral corticospinal tract (CST), forceps major, left inferior fronto-occipital fasciculus (IFOF), and bilateral superior longitudinal fasciculus (SLF) (Fig 4). We additionally observed a positive relationship between intra-cranial volume (ICV) and fibre morphology. Pubertal stage was positively associated with fibre properties in the right SLF. There was additional evidence for an association between fibre density and behavioural difficulties (attention dysfunction and internalising behaviours) in the right uncinate fasciculus (UF).Discussion
The longitudinal results shed light on regional development of fibre properties during the transition from childhood to adolescence. We observed protracted development of pathways connecting the frontal lobe, such as the right IFOF, bilateral uncinate fasciculus (UF), and bilateral inferior longitudinal fasciculus (ILF). Intra-cranial volume predicted fibre morphology across most regions studied, highlighting the importance for accounting for variation in ICV on fibre morphology (i.e. due to sex differences). The SLF is also known to actively develop across childhood [7], therefore our results may additionally link pubertal processes with this dynamic development.Conclusion
The development of fibre density and morphology during the early pubertal period predominantly involves the expansion of key white matter tracts, precluding fronto- and occipito-temporal regions known to have protracted development over adolescence. Additionally, our findings show that fibre properties in the SLF and UF are influenced by pubertal stage and behavioural difficulties. Future work should profile developmental trajectories across additional time-points using methods such as generalised additive models.Acknowledgements
We would like to thank the families and children enrolled in the NICAP study for their time and participation. We would also like to thank the contributions of Michael Kean for implementing advances in imaging acquisition at The Royal Children's Hospital, Parkville, Australia.References
1. Shankman, S.A., et al., Subthreshold conditions as precursors for full syndrome disorders: a 15-year longitudinal study of multiple diagnostic classes. J Child Psychol Psychiatry, 2009. 50(12): p. 1485-94.
2. Raffelt, D.A., et al., Investigating white matter fibre density and morphology using fixel-based analysis. Neuroimage, 2017. 144(Pt A): p. 58-73.
3. Genc, S., et al., Development of white matter fibre density and morphology over childhood: A longitudinal fixel-based analysis. NeuroImage, 2018.
4. Wakana, S., et al., Reproducibility of quantitative tractography methods applied to cerebral white matter. Neuroimage, 2007. 36(3): p. 630-44.
5. Morrell, C.H., L.J. Brant, and L. Ferrucci, Model Choice Can Obscure Results in Longitudinal Studies. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 2009. 64A(2): p. 215-222.
6. Benjamini, Y. and Y. Hochberg, Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society. Series B (Methodological), 1995. 57(1): p. 289-300.
7. Sawiak, S. J., Y. Shiba, L. Oikonomidis, C. P. Windle, A. M. Santangelo, H. Grydeland, G. Cockcroft, E. T. Bullmore and A. C. Roberts (2018). Trajectories and Milestones of Cortical and Subcortical Development of the Marmoset Brain From Infancy to Adulthood. Cerebral Cortex: bhy256-bhy256.
Figures
