Microstructural neuroimaging of white matter tracts in persistent post-concussion syndrome: A prospective controlled cohort study

Abstract

Introduction: Children with mild traumatic brain injury (mTBI) typically recover quickly, however approximately 15% experience persistent post-concussive symptoms (PPCS) past 3 months. The microstructural pathology associated with underlying persistent symptoms is poorly understood but is suggested to involve axonal injury to white matter tracts. Diffusion tensor imaging (DTI) can be used to visualize and characterize damage to white matter microstructure of the brain.

Objective: We aimed to investigate white matter microstructure in children with persistent concussive symptoms as compared to typically developing controls, alongside evaluating differences in white matter changes over time and how this relates to symptom recovery.

Methods: The current study is a prospective, longitudinal, controlled cohort study of children with mTBI. 104 children aged 8 to 18 years with a mTBI (72 symptomatic; 32 asymptomatic) were recruited from the Alberta Children's Hospital and compared to 20 healthy controls. Microstructural evidence of white matter injury was evaluated using DTI one month post injury and repeated 4 to 6 weeks later. Primary outcomes included fractional anisotropy and mean diffusivity of the corticospinal tracts, uncinate fasciculi, and motor fibers of the corpus callosum. Post-concussive symptoms were also measured using the Post-Concussion Symptom Inventory (PCSI) taken at both time points.

Results: Fractional anisotropy of the left uncinate fasciculi was lower in symptomatic children compared to controls (F(2,119) = 3.582, p = 0.031). No other significant differences were observed.

Conclusions: Our findings provide evidence of microstructural injury following mTBI in children with ongoing post-concussive symptoms one month post injury. The changes were persistent 4-6 weeks later. Further longitudinal studies of white matter microstructure in PPCS will be helpful to clarify whether these white matter alterations resolve over time.

Keywords: Diffusion tensor imaging; Mild traumatic brain injury; Pediatrics; Persistent post-concussive symptoms; Uncinate fasciculus.

Fig. 1

Participant recruitment diagram. Visual flow chart of participant recruitment, screening, and data collection. Note that children who had contraindications for MRI (e.g., metal braces) were included within the population unable to commit to study requirements.

Fig. 2

Tract Inclusion ROIs and Visual Depiction of Isolated Fibers. Inclusion ROIs used to delineate the motor fibers of the corpus callosum (A), left corticospinal fibers (B), and the left uncinate fasciculus fibers (C), and the resulting fiber tracts shown on the T1-weigthed anatomical image from the target representative scan. Right CST and UF fibers were isolated using similar inclusion ROIs to the above. Note that cortical inclusion ROIs used to isolate the left and right CST were used to delineate the corpus callosum fibers projecting to the cortical motor strip.

Fig. 3

Group comparisons of white matter microstructure one month post injury. (A) Group differences in FA; significantly lower FA of the left UF was observed in the symptomatic group. (B) Groups differences in MD values; no significant group differences were noted. (C) Group differences in RD values; no significant group differences were noted. D) Group differences in AD; significantly lower AD of the left UF in the symptomatic group. Note that box plots represent the median and interquartile range, with thick black lines indicating the group mean. Black circles indicate data points falling outside of the 95% confidence interval, while error bars indicate the range in diffusion values.

Fig. 4

Group comparisons of left uncinate fasciculus along fiber analysis. (A) MD of left UF for controls (blue line), asymptomatic (green line), and symptomatic (red line) groups; significantly higher MD in the mTBI groups was noted compared to controls. (B) FA of left UF for controls (blue line), asymptomatic (green line), and symptomatic (red line) groups; significantly lower FA in the mTBI groups was noted compared to controls. Note that black bars indicate significant differences between control and symptomatic groups, while grey bars indicate significant differences between control and asymptomatic groups. No significant differences were identified between symptomatic and asymptomatic groups.

Fig. 5

Observations in the canonical discriminant analysis space for each group. The probabilities of group membership before analysis is shown. Classification results for original and cross-validated groups are shown. Two discriminant factors (linear combinations of variables) were used to classify group membership.