We assessed the utility of a brief MRI protocol, appropriate for the acute setting, to detect acute traumatic brain injury (TBI) in patients with suspected mild TBI (mTBI) and distinguish traumatic from nontraumatic brain injury by comparing trauma with nontrauma patients. Twenty-two patients with suspected mTBI were included in this exploratory study over a period of 9 months. Median time from injury to MR scanning was 5.4 h (interquartile range 3.6-15.3). To determine the specificity of certain findings for TBI, 61 patients presenting with suspected minor acute stroke were included as a comparative group using the same MRI methods. A selected series of MRI sequences (diffusion-weighted imaging, fluid attenuated inversion recovery [FLAIR], and T2* weighted) were independently evaluated by two neuroradiologists blinded to clinical diagnosis, for presence of specific findings. In a separate session, all cases in which at least one MRI sequence above was positive were classified as TBI, stroke, or indeterminate. Intracranial MRI abnormalities were observed in 47 (57%) of the 83 studied patients. Based on findings on MRI, 12 (55%) of 22 suspected mTBI patients were classified as having traumatic injury. Nine (47%) of the 19 suspected mTBI patients with a negative CT had findings on MRI. Abnormalities on MRI consistent with trauma were observed most frequently on postcontrast FLAIR (83%) and T2*-weighted (58%) sequences. We demonstrated the ability of a fast MRI protocol to identify trauma-related abnormalities not seen on CT, and differentiate acute trauma from nonspecific chronic disease in a blinded cohort of mTBI patients.