Although enzyme replacement therapy (ERT) is an effective treatment for mucopolysaccharidosis (MPS) types I, II, IVA and VI, its effectiveness in children with central nervous system (CNS) disorders is said to be poor because the blood-brain barrier cannot be penetrated by ERT drugs. To assess CNS involvement in mucopolysaccharidosis at the start of enzyme replacement therapy and to investigate the time course of ERT in the central nervous system. We performed brain MRI in 17 children and young adults who underwent ERT. The clinical severity was classified as attenuated or severe by a specialist pediatrician, based on the clinical symptoms and genotypes. At the start of ERT, we scored nine parameters using two- or three-point scales based on the severity of the disease revealed on MRI scans. After the start of ERT, we compared the initial and follow-up MRI scans, and classified the findings as no change, improved or worse. We then compared the results with the changes in clinical findings. At the start of ERT, comparison of the clinical symptoms and image scores revealed differences between severe and attenuated mucopolysaccharidosis. The scores in patients with severe MPS ranged from 9 to 16 (mean 12.2); for patients with attenuated MPS, they ranged from 2 to 11 (mean 6.4). Images of the four patients with severe MPS showed ventricular dilation and brain atrophy. Such findings were made in only 2 of 13 patients with attenuated MPS. The results after the start of ERT showed that 11/17 (65%) patients manifested improvement or no change. All five patients with MPS I experienced improvement in some regions. There were no new lesions. One patient with MPS II experienced worsening of his CNS symptoms, and his MRI findings revealed more severe ventricular dilation, brain atrophy and white matter lesions. Ventricular dilation and brain atrophy on imaging studies might represent useful markers in predicting the severity of mucopolysaccharidosis and worsening of CNS symptoms. Enzyme replacement therapy improves CNS images in MPS I and has an inhibitory effect on the occurrence of new lesions in MPS II.