Introduction: Repetitive head impacts (RHI) sustained in contact sports are thought to be necessary for the long-term development of chronic traumatic encephalopathy (CTE). Our objectives were to: 1) characterize the magnitude and persistence of RHI-induced white matter (WM) changes; 2) determine their relationship to kinematic measures of RHI; and 3) explore their clinical relevance. Methods: Prospective, observational study of 10 Division III college football players and 5 non-athlete controls during the 2011-12 season. All subjects underwent diffusion tensor imaging (DTI), physiologic, cognitive, and balance testing at preseason (Time 1), post-season (Time 2), and after 6-months of no-contact rest (Time 3). Head impact measures were recorded using helmet-mounted accelerometers. The percentage of whole-brain WM voxels with significant changes in fractional anisotropy (FA) and mean diffusivity (MD) from Time 1 to 2, and Time 1 to 3 was determined for each subject and correlated to head impacts and clinical measures. Results: Total head impacts for the season ranged from 431-1,850. No athlete suffered a clinically evident concussion. Compared to controls, athletes experienced greater changes in FA and MD from Time 1 to 2 as well as Time 1 to 3; most differences at Time 2 persisted to Time 3. Among athletes, the percentage of voxels with decreased FA from Time 1 to 2 was positively correlated with several helmet impact measures. The persistence of WM changes from Time 1 to 3 was also associated with changes in serum ApoA1 and S100B autoantibodies. WM changes were not consistently associated with cognition or balance. Conclusions: A single football season of RHIs without clinically-evident concussion resulted in WM changes that correlated with multiple helmet impact measures and persisted following 6 months of no-contact rest. This lack of WM recovery could potentially contribute to cumulative WM changes with subsequent RHI exposures.