Region-Specific Brain Imaging Correlates of Repetitive Head Impact Exposure in Nonconcussed High School Football Players.

Repetitive, nonconcussive head impacts in collision sports are associated with lasting neurophysiological effects. This study investigates region-specific changes in neocortical mean diffusivity (MD) and resting-state functional magnetic resonance imaging (fMRI) power spectral density (fMRI-PSD) that correlate with cumulative nonconcussive linear risk-weighted exposure (RWE-Lin) sustained over one high school football season. Pre- to postseason percent changes in MD (%ΔMD) and fMRI-PSD (%ΔfMRI-PSD) were calculated for football players without diagnosed concussions during the season of observation (n = 87 player-seasons). Voxelwise linear regression analyses examined relationships between RWE-Lin with neocortical %ΔMD and %ΔfMRI-PSD. False discovery rate-corrected alpha thresholds of 0.0125 accounted for multiple comparisons and directionality within each test. While controlling for age, body mass index, days between scans, position group (skill vs. line), level (junior varsity vs. varsity), and concussion history, RWE-Lin was significantly positively correlated with %ΔMD (p = 1.12 × 10-5) and inversely correlated with %ΔfMRI-PSD in the frontal cortex (p = 7.36 × 10-4). Football players were stratified into high RWE-Lin (n = 13) and low RWE-Lin (n = 37) groups using the Jenks natural breaks algorithm. Noncollision athletes (NCAs, n = 11) served as external controls. When comparing imaging measures across the three groups, the high RWE-Lin football group showed significant differences from both NCA and low RWE-Lin football groups in frontal %ΔMD and %ΔfMRI-PSD (p < 0.05). No significant imaging differences were observed between NCA and low RWE-Lin football groups. Biomechanical modeling localized high-risk impact regions that have high spatial correspondence to the observed imaging abnormalities. Exploratory post hoc voxelwise analyses of magnetoencephalography spectral power detected no significant voxelwise associations with RWE-Lin. In the absence of diagnosed concussion, high-impact football players exhibit abnormal structural and neurovascular changes in the frontal cortex. These regions overlap with those identified to be at greatest risk from biomechanical modeling and those implicated in the pathogenesis of chronic traumatic encephalopathy. Importantly, these imaging abnormalities were not generally observed in football players with lower RWE-Lin. These findings suggest that limiting the number and severity of head impacts may be an effective strategy to mitigate subclinical brain injury in American football.