The acknowledgement of risks for traumatic brain injury in American football players has prompted studies for sideline concussion diagnosis and testing for neurological deficits. While concussions are recognized etiological factors for a spectrum of neurological sequelae, the consequences of sub-concussive events are unclear. We tested the hypothesis that blood-brain barrier disruption (BBBD) and the accompanying surge of the astrocytic protein S100B in blood may cause an immune response associated with production of autoantibodies. We also wished to determine whether these events result in disrupted white matter on diffusion tensor imaging (DT) scans. Players from three college football teams were enrolled (total of 67 volunteers). None of the players experienced a concussion. Blood samples were collected before and after games (n = 57); the number of head hits in all players was monitored by movie review and post-game interviews. S100B serum levels and auto-antibodies against S100B were measured and correlated by direct and reverse immunoassays (n = 15 players; 5 games). A subset of players underwent DTI scans pre- and post-season and after a 6-month interval (n = 10). Cognitive and functional assessments were also performed. After a game, transient BBB damage measured by serum S100B was detected only in players experiencing the greatest number of sub-concussive head hits. Elevated levels of auto-antibodies against S100B were elevated only after repeated sub-concussive events characterized by BBBD. Serum levels of S100B auto-antibodies also predicted persistence of MRI-DTI abnormalities which in turn correlated with cognitive changes. Even in the absence of concussion, football players may experience repeated BBBD and serum surges of the potential auto-antigen S100B. The correlation of serum S100B, auto-antibodies and DTI changes support a link between repeated BBBD and future risk for cognitive changes.