Reflection and Reaction http://neurology. thelancet. com Vol 7 May 2008 381 who is now aged 9 years, had generalised dystonia that spared only the cranial and facial muscles, his walking was severely impaired, but his cognition was normal. In addition to the frameshift mutation, we found two new silent changes in PRKRA in two other patients: c. 126C/T, which was absent in all the other patients, and c. 795C/T, which was found in four of the 127 patients with dystonia and in five of the controls. The c. 665C> T mutation in PRKRA and the GAG deletion in the DYT1 gene were excluded in all 127 patients with dystonia. Although RNA samples were unavailable to test for any possible effects of the two silent changes on splicing, the nonsense mutation is most probably pathogenic. DYT16 dystonia seems to follow a recessive inheritance pattern, which suggests that we might have missed a gene dosage mutation or a sequence change in a noncoding region of the gene with the methods we used. Alternatively, a heterozygous (truncating) mutation might have had a pathogenic effect in its own right. Our findings should prompt additional mutational analyses and association studies to elucidate further the role of mutations in PRKRA in the aetiology of different forms of dystonia.