Objectives - The aim of this study was to re-evaluate the clinical features of stroke in children, their outcome and the place of the different mechanisms, in the light of CT-scan and magnetic resonance imaging. Methods - A 10-year review of the Dijon Childhood Neurology Clinic experience (1985–1995) identified 54 patients with arterial stroke. Diagnosis was established by CT-scan and angiography and by magnetic resonance imaging from 1987. When an hemorrhagic stroke was identified, a cerebral arteriogram and an investigation of the coagulation factors were performed. When an ischemic stroke was identified, the following tests were performed: an ultrasound examination of the cervical arteries, a cerebral arteriogram, a lumbar puncture, an investigation of the coagulation factors and lipid status, a measurement of homocystine in the plasma and the urine, an electrocardiogram, a Holter procedure and a cardiac echography. Results - During the 11 full calendar years of this study, we observed 54 stroke patients. There were 31 cases of ischemic stroke, representing some 57% of the total, as well as 23 cases of hemorrhagic stroke, representing 43% of the total. Of the 31 cases with ischemic stroke, 4 had no known predisposing condition, 6 occurred in children with pre-existing heart disease, 2 had moya-moya disease, 2 had leukemia, 2 had a carotid dissection, 1 had an hemolytic-uremic syndrome, and 14 were observed in patients with other associated conditions, such as infections (7 cases) or slight cranial trauma (7 cases). On CT-scan, a basal ganglia infarction was identified in 14 cases. Among the 23 hemorrhagic strokes, 9 were due to arteriovenous malformations, 2 to arterial aneurysm and 5 to cavernomas. Two cases occurred in hemophilia, 2 in idiopathic thrombocytopenic purpura, and 2 after throat infections. One case had no etiologic factor. Clinical course was marked by a low mortality rate even in hemorrhagic stroke, long-lasting seizures, and hemidystonia. Conclusion - This neuro-imaging survey focused on the real place of hemorrhage and basal ganglia infarctions in children under 16 years of age, compared to adulthood. Follow-up demonstrated good or complete resolution of neurological deficits despite the frequent late hemidystonia and late epilepsy.