Background

While traumatic brain injury (TBI) is associated with subsequent depressive symptoms and sleep dysfunction, it is uncertain whether these symptoms are mediated by early neurodegeneration in specific brain regions or by psychosocial stressors.

Methods

We used cross-sectional data from 57 middle-aged, cognitively healthy, former professional rugby players and 114 propensity score matched controls. TBI burden (number of episodes of loss of consciousness), depressive symptoms and sleep dysfunction were assessed using self-reported questionnaires (Brain Injury Screening Questionnaire, Center for Epidemiological Studies Depression scale, Pittsburgh Sleep Quality Index). Grey matter volumes (GMVs), which were normalised to total intracranial volume, and diffusion tensor imaging (DTI) measures (fractional anisotropy (FA), mean (MD), axial (AxD) and radial diffusivity (RD)) were obtained using FreeSurfer (v7.1.0) and FSL (v6.0.5). Spearman correlations, uncorrected for multiple comparisons, were used to identify imaging markers which were significantly associated with either TBI burden or depressive symptoms. Partial least square structural equational modelling was used to test whether TBI burden was associated with depressive symptoms and sleep dysfunction, and whether this was mediated through differences in the significant imaging markers.

Results

Former rugby players reported a higher TBI burden (1.9vs.0.7,p<0.001), more frequent moderate (4.6vs.4.2,p=0.002) and vigorous (4.2vs.3.8,p=0.002) exercise, but were less likely to report previous sleep disorders (2%vs.15%,p=0.02), anxiety disorders (7%vs.22%,p=0.03), or depression (11%vs.23%,p=0.08) than controls. In former rugby players, there was a strong direct association between TBI burden and depressive symptoms (B:0.52[95%CI:0.21-0.74]), and a moderate indirect association between TBI and sleep dysfunction, mediated through depressive symptoms (B:0.18[95%CI:0.03-0.39]). Higher TBI burden was associated with lower AxD of the inferior cerebellar peduncle (r=-0.28,p=0.03), while more depressive symptoms were associated with higher FA of the hippocampal cingulum (r=0.27,p=0.04) and lower MD of the fornix (r=-0.28,p=0.03). Neither were associated with RD or any GMVs. Furthermore, none of the significant imaging markers mediated the association between TBI burden and depressive symptoms.

Conclusion

TBI burden was strongly associated with depressive symptoms in former rugby players, and both factors correlated with DTI measures. However, imaging markers did not explain the relationship between TBI and depression, suggesting that psychosocial stressors may be more important.