Frontoamygdala hyperconnectivity predicts autonomic dysregulation and persisting symptoms in sports-related concussion

Abstract

OBJECTIVEWe investigated longitudinal trajectories of resting-state fMRI (rsfMRI), autonomic function, and symptoms after sport-related concussion (SRC). BACKGROUNDLimbic circuitry may be particularly vulnerable to traumatic brain injury (TBI), which could explain the affective and autonomic dysfunction some patients develop. Relatively few studies have performed longitudinal rsfMRI analyses in concussion and fewer have combined imaging with autonomic and symptom data. We leveraged published limbic rsfMRI networks centered on the amygdala that include central autonomic structures in frontal and temporal lobes and their visceromotor targets. We hypothesized that frontoamygdala connectivity would differentiate athletes with SRC from matched, in-sport controls (ISC), predict autonomic function, and predict symptom recovery. DESIGN/METHODSUsing independent-samples t-tests, we compared rsfMRI connectivity strength in amygdala networks in college athletes with SRC (SRC: n=31, female=14) at three time points after concussion (T1[≤]4 days, T2=10-14 days, T3=60-90 days) and healthy, matched controls without a concussion in the same sport (ISC: n=36, female=17). RESULTSSRC athletes showed significantly greater frontoamygdala connectivity compared to ISCs at the acute and subacute post-injury time points (T1 p=0.003, T2 p=0.014) that normalized to control-level connectivity by the chronic time point (T3 p=0.182). When testing whether autonomic function interacts with network connectivity trajectory, we found that opposing trajectories of frontoamygdala connectivity between SRC athletes with higher versus lower acute heart rate variability (HRV), as measured by pNN50 (percentage of intervals between successive normal sinus beats greater than 50ms). SRC athletes with higher HRV acutely post-injury had significantly greater acute frontoamygdala connectivity compared to ISCs at T1; connectivity in these high-HRV SRC athletes normalized to control level over time (T1 p=0.001, T2 p=0.055, T3 p=0.576). SRC athletes with lower HRV acutely post-injury had control-level frontoamygdala connectivity at T1; connectivity in these low-HRV SRC athletes significantly exceeded control-level connectivity at T3 (T1 p=0.429, T2 p=0.050, T3 p=0.002). Furthermore, those SRC athletes with the greatest frontoamygdala connectivity at T3, had the most persistent symptoms on the graded symptom checklist at T3 (r=0.635, p=0.001). Differences in diffusion-tensor-imaging-based measures of structural connectivity in the uncinate fasciculus, the fiber bundle most critical to our frontoamygdala network, could not account for these relationships. CONCLUSIONSThese results suggest that increased connectivity in amygdala circuitry acutely after a concussion and its normalization over time may be protective or compensatory; acute measures of amygdala network connectivity and measures of HRV may be valuable biomarkers for predicting symptom persistence.