Mapping Developmental Trajectories of Brain White Matter from Birth to Six Months Using Diffusion Tensor Imaging: A Preliminary Study

Background:  Research on brain development may provide groundbreaking insights into the identification of brain systems that underlie the development of social cognitive functions(Johnson,2005) and developmental disabilities, such Autism Spectrum Disorder (ASD)(Wolff, 2013). Diffusion tensor imaging (DTI) quantifies brain white matter microstructures and has become a prominent tool for indexing white matter development(Basser,1995).

Objectives:  To reveal critical transitions in regional brain development in typically developing infants from birth to 6 months as a benchmark against which to compare atypical transitions of brain development that may be a sign or symptom of ASD.

Methods:  24 typical developing infants (mean age: 115 days (27-218 days); 6 females) were enrolled. DTI data were collected from each infant at three pseudorandom time points between birth and 6 months, yielding a total of 53 scans. Data were collected on a 3T Siemens Trio scanner with 32-channel head coil and multiband sequence(Moeller, 2010). DTI parameters are: TR/TE of 6200/74ms, a multiband factor of 2 combined with a GRAPPA of 2, FOV of 184×184, spatial resolution of 2mm isotropic, b=0/700 s/mm2, 61 diffusion directions, extra 6 of b0s in both phase encoding directions. Distortion corrected(Andersson, 2003) brain tensor maps from each participant were aligned to a sample-specific common space using tensor-based registration(Zhang, 2006). The John Hopkin’s Neonate Atlas(Oishi, 2011) was aligned to our common space, grouped into nine brain areas and multiplied with a white matter mask (mean fractional anisotropy, FA>0.25) (Fig.1A). Principal Component Analysis through Conditional Expectation (PACE)(Yao, 2005)—a method designed to overcome missing values, a common problem in longitudinal infant research—was used to fit changes in FA and its derivatives over time in each brain area.

Results:  FA increases in all brain areas from birth to 6 months (Fig.1B), with subcortical, cerebellum and occipital lobes having the highest FA increase during infants’ first 6 months. FA change rates decrease over time (Fig.1C), with orbitofrontal areas showing the highest FA change rate (although its FA value is lower relative to other brain regions at birth). By month 6, rate of change in FA decreases by approximately 8.6 fold. All brain regions show negative FA accelerations at birth, especially in cortical areas, followed by a rapid decrease in acceleration toward zero. Interestingly, FA accelerations reach a plateau around 5 months of age in most cortical regions.

Conclusions:  We identified a potential critical time period around 5 months of age, in which FA accelerations reach a plateau for the majority of brain regions. Although the exact causes of this plateau are still under investigation, possible explanations include pruning processes(LaMantia, 1990), divergent trends of FA changes in white and gray matter, and/or the development of association pathways(Dubois, 2014). This work highlights the importance of using non-parametric curve fitting techniques, such as PACE, for modeling data and their derivatives, as approaches that specify the shape of brain developmental profiles a priori may miss critical dynamic information in brain development, especially in early infancy.