PURPOSE: Conventional imaging study, such as CT and T1 or T2 weighted MRI, usually give valuable information, but cannot always differentiate early parenchymal lesion in the acute stage of hypoxic-ischemic brain injury from unmyelinated white matter in the neonate. Diffusion-weighted image already has been proven highly sensitive and specific for the detection of acute stroke in adult population. This study was performed to evaluate the usefulness of this technique in the diagnosis of neonatal seizure. METHODS: 11 neonates, aged from 1 to 6 days admitted at Ajou University Hospital for neonatal seizure, were studied with diffusion-weighted imaging and conventional images at within 72 hours of seizure onset. Follow-up MRIs were obtained from 4 patients at about 3 months of age. Comparisons were made among conventional T1, T2 weighted images, diffusion-weighted images and images from follow-up MRIs. RESULTS: 9 of 11 patients showed abnormal findings on diffusion-weighted images, 2 patients showed focal cerebral infarction and 7 patients showed diffuse injury consistent with HIE. Diffusion-weighted image showed more accurate images of involved lesion than T1 or T2 weighted images in some cases. But most of all, diffusion-weighted image could detect lesions of white matter injury which was not visualized on conventional MRI. The white matter injury later result in cortical atrophy and leukomalatic changes by follow-up MRI. CONCLUSION: Our study provides the usefulness of diffusion-weighted imaging in the evaluation of neonatal seizure along with conventional images. For the confirmation, a larger number of infants and clinical evaluation for development will be needed.