In reply
Quinn pointed out that the result of our study1 could be influenced by selection bias of both control subjects and patients. In particular, he suggested that the increased frequency and extent of periventricular hyperintensities in patients with PD with accelerated disease progression could be due to performing magnetic resonance imaging more often in patients with atypical features. He also claimed that we do not specify how the "healthy" subjects have been selected for magnetic resonance imaging.The control group included subjects referred to the Institute of Neurology of the Univeristy of Pisa, Pisa, Italy, over a period of 3 years. The subjects complained of neurologic symptoms not known to alter magnetic resonance imaging white matter characteristics. None of the subjects had a history of psychiatric or neurologic diseases, and none of them had cerebrovascular risk factors or cognitive deficits as scored by the Mini-Mental State Examination. Examples