Development of white matter fibre density and morphology over childhood: A longitudinal fixel-based analysis

Highlights

• We present the first longitudinal application of fixel-based analysis (FBA).

• White matter fibre development was investigated in children aged 9–13.

• Fibre density increased in specific commissural and association fibres.

• The rate of fibre development was not associated with sex or puberty.

• Confirmatory Bayesian statistical analyses support the main FBA findings.

Abstract

Purpose

White matter fibre development in childhood involves dynamic changes to microstructural organisation driven by increasing axon diameter, density, and myelination. However, there is a lack of longitudinal studies that have quantified advanced diffusion metrics to identify regions of accelerated fibre maturation, particularly across the early pubertal period. We applied a novel longitudinal fixel-based analysis (FBA) framework, in order to estimate microscopic and macroscopic white matter changes over time.

Methods

Diffusion-weighted imaging (DWI) data were acquired for 59 typically developing children (27 female) aged 9–13 years  at two time-points approximately 16 months apart (time-point 1: 10.4 ± 0.4 years, time-point 2: 11.7 ± 0.5 years). Whole brain FBA was performed using the connectivity-based fixel enhancement method, to assess longitudinal changes in fibre microscopic density and macroscopic morphological measures, and how these changes are related to sex, pubertal stage, and pubertal progression. Follow-up analyses were performed in sub-regions of the corpus callosum to confirm the main findings using a Bayesian repeated measures approach.

Results

There was a statistically significant increase in fibre density over time localised to medial and posterior commissural and association fibres, including the forceps major and bilateral superior longitudinal fasciculus. Increases in fibre cross-section were substantially more widespread. The rate of fibre development was not associated with age or sex. In addition, there was no significant relationship between pubertal stage or progression and longitudinal fibre development over time. Follow-up Bayesian analyses were performed to confirm the findings, which supported the null effect of the longitudinal pubertal comparison.

Conclusion

Using a novel longitudinal fixel-based analysis framework, we demonstrate that white matter fibre density and fibre cross-section increased within a 16-month scan rescan period in specific regions. The observed increases might reflect increasing axonal diameter or axon count. Pubertal stage or progression did not influence the rate of fibre development in the early stages of puberty. Future work should focus on quantifying these measures across a wider age range to capture the full spectrum of fibre development across the pubertal period.

Introduction

The understanding of brain development over childhood and adolescence has been revolutionised by magnetic resonance imaging (MRI) applications. The in vivo quantification of brain structure has revealed age-related macrostructural processes, dominated by grey matter pruning and white matter volume expansion (Mills et al., 2016). White matter volume, however, is a crude measure that can be influenced by several distinct neurobiological properties – namely axon diameter, density, and myelin thickness. Diffusion MRI techniques are more appropriate for understanding the finer detail of microstructural organisation and have been widely used to study developmental populations (see Tamnes et al. (2017) for a recent review).

Arguably the most significant biological event across childhood is puberty. The onset and progression of puberty encompasses a cascade of endocrine changes, subsequently leading to the phenotypic characteristics such as growth spurt, skin and voice changes, and gonadal development. Key endocrine events, such as the rise in circulating levels of dehydroepiandrosterone (DHEA) and its sulphated form DHEAS, begin to rise in circulating levels at approximately 6–9 years of age (Dorn et al., 2006). These adrenal hormones are known to influence white matter maturation, by virtue of neurogenesis (new axons forming) and neurite growth (increase in axon diameter) (Maninger et al., 2009), establishing the neurobiological link between pubertal onset and axonal properties.

Several cross-sectional studies have reported a link between pubertal stage and white matter microstructure with the use of diffusion tensor imaging (DTI). Greater organisation of the white matter, typically indicated by increases in fractional anisotropy (FA), is associated with maturation of physical characteristics of puberty and circulating hormone levels (Herting et al., 2012; Menzies et al., 2015). Despite being sensitive to a wide array of phenotypic traits such as age-relationships (Lebel et al., 2008) and sex differences (Schmithorst et al., 2008) over development, diffusion tensor metrics are non-specific, and therefore conclusions around relative change in such metrics cannot be attributed to any specific white matter property. Therefore, interpretation of change in such metrics must be done with care (Jones et al., 2013).

Whilst cross-sectional studies are useful in studying static links between brain structure and phenotypic traits, longitudinal studies are vital in understanding the dynamics of age-related white matter maturation (Mills and Tamnes, 2014). In particular, as the brain rapidly develops during childhood and adolescence, it is imperative to longitudinally assess differential trajectories of white matter development and how these relate to variations in behaviour, cognition, and puberty. To the best of our knowledge, only one longitudinal study has revealed the influence of pubertal processes on white matter microstructure as assessed with DTI (Herting et al., 2017). Notably, pubertal processes are thought to preferentially influence axon calibre, as opposed to myelin (Perrin et al., 2008), emphasising the importance of quantifying axonal properties independent of myelination (Paus, 2010).

Advances in diffusion modelling techniques allow quantification of white matter fibre properties in the presence of complex fibre geometry (Jeurissen et al., 2013; Raffelt et al., 2012). A recently developed whole-brain analysis framework, Fixel-Based Analysis (FBA) (Raffelt et al., 2017), allows the comprehensive statistical analysis of white matter quantitative measures in the presence of such complex fibre populations. This analysis framework offers advantages above existing voxel-wise approaches: rather than computing some scalar metric from the diffusion model in each image voxel, scalar quantitative measures are represented within fixels (specific fibre populations within voxels), enabling inference of fibre-specific properties in the white matter. The commonly investigated metrics within this framework are:

• “Fibre Density (FD)”: A microscopic estimate of the density of axons within a particular fibre population in a given voxel. In the context of spherical deconvolution (Tournier et al., 2004) and the Apparent Fibre Density (AFD) metric (Raffelt et al., 2012), this is an estimate of the intra-cellular volume of fibres oriented in a particular direction. An increase in FD could result from developmental processes, such as growth in axon diameter, or increase in the number of axons occupying a given space (Genc et al., 2018); whereas a decrease in fibre density could be due to loss of axons, such as in multiple sclerosis (Gajamange et al., 2018).

• “Fibre Cross-section (FC)”: A morphological measure of the macroscopic change in cross-sectional area perpendicular to a fibre bundle experienced during registration to a template image.

• “Fibre Density and Cross-section (FDC)”: A combined measure that incorporates both the microscopic and macroscopic effects described above, thus providing sensitivity to any differences related to the capacity of the white matter to transmit information.

Our recent work using FBA cross-sectionally in a developmental context revealed that pubertal children have greater fibre density than age-matched pre-pubertal children (Genc et al., 2017b). This finding supports previous theories around the influence of pubertal onset on axonal properties. However, longitudinal studies are required to study differential trajectories of neurodevelopment due to age, sex, and puberty. Despite advances in microstructural modelling and analysis techniques, there is a dearth of longitudinal processing and analysis pipelines, as well as applications – particularly at the whole-brain level. This may be due to the difficulty of retaining longitudinal imaging samples, or due to the lack of ‘out-of-the-box’ image processing pipelines.

Here, we outline and demonstrate a novel longitudinal fixel-based analysis pipeline upon a developmental sample of children aged 9–13, to identify key white matter pathways of accelerated maturation. We interrogate specifically whether fibre properties differentially change as a function of age and sex. In addition, we investigate whether the rate of fibre development is driven by pubertal stage and pubertal progression, hypothesising that faster progression would result in accelerated fibre density increases in posterior commissural fibres.