Victims of minimal traumatic brain injury (mTBI) do not show clear morphological brain defects, but frequently suffer lasting cognitive deficits, emotional difficulties and behavioral disturbances. In the present study we adopted a non-invasive closed-head weight-drop mouse model to produce mTBI. We examined the effects of 20, 25, or 30 g weight drop 7, 30, 60 and 90 days following injury on mice’s ability to perform the Morris water maze. The mice suffered profound long-lasting learning and memory deficits that were force- and time-dependent. Although the injured mice could acquire the task, they could not improve their initial escape latency by more than 50%, while normal mice improved by up to 450% ( P<0.001). In order to directly compare the learning ability of individual mice following our mTBI we have devised a new measure which we term learning rate. We define learning rate as the rate the mouse improved its own performance in consecutive trials in a given experimental day. The learning rate of control mice increased linearly throughout the testing period with a slope of approximately 0.9. Injured mice that sustained 20 and 25 g weight drop could also improve their learning rate linearly but with a slope of only 0.2. Mice who sustained 30 g weight drop could not improve their learning rate linearly and reached a plateau after the third experimental learning day. These results indicate that the severity of injury may correlate with the degree of integration of the learning task. These cognitive deficits occurred without any other clear neurological damage, no evident brain edema, no notable damage to the blood–brain barrier and no early anatomical changes to the brain (observed by magnetic resonance imaging imaging). These results demonstrate that persistent deficits of cognitive learning abilities in mice, similar to those observed in human post-concussive syndrome, can follow mTBI without any anatomical damage to the brain and its surrounding tissue.