OBJECTIVE. Our goal was to compare the patterns of brain injury detected by computed tomography, conventional MRI (T1- and T2-weighted sequences), and diffusion-weighted MRI in a cohort of term newborns with neonatal encephalopathy studied uniformly with all 3 modalities on the third day of life. METHODS. Term newborns (>= 36 weeks' gestation) admitted to our center with neonatal encephalopathy were scanned with computed tomography, MRI, and diffusion-weighted MRI at 72 (+/- 12) hours of life (n = 48). Each modality was scored independently of the other with previously validated scoring systems. The predominant pattern of brain injury was classified as: normal, watershed, basal nuclei, total (maximal basal nuclei and watershed), and focal-multifocal (presence of strokes and/or white matter injury alone). RESULTS. The agreement for the predominant pattern of injury was excellent between MRI and diffusion-weighted MRI (77% agreement). The agreement for the pattern of injury was also good for computed tomography and diffusion-weighted MRI (67% agreement). The extent of cortical injury and focal-multifocal lesions, such as strokes and white matter injury, were less apparent on computed tomography than diffusion-weighted MRI. In 19 newborns with a repeat MRI in the second week of life, the predominant pattern seen on the day 3 diffusion-weighted MRI was confirmed. CONCLUSIONS. Diffusion-weighted MRI is the most sensitive technique with which to assess brain injury on day 3 of life in term newborns with neonatal encephalopathy, particularly for cortical injury and focal-multifocal lesions such as stroke and white matter injury. All 3 modalities identify the most serious patterns of brain injury similarly. Pediatrics 2009;123:319-326