PURPOSE: To investigate the feasibility of functional MR imaging of motor language function and its usefulnessin the determination of hemispheric language dominance. MATERIALS AND METHODS: In order to activate the motorcenter of language, six subjects(5 right-handed, 1 left-handed; 3 males, 3 females) generated words. They wererequested to do this silently, without physical articulation, in response to English letters presented visually.Gradient-echo images (TR/TE/flip angle, 80/60/40o; 64x128 matrix; 10mm thickness) were obtained in three axialplanes including the inferior frontal gyrus. Functional maps were created by the postprocessing of gradient-echoimages, including subtraction and statistics. Areas of activation were topographically analyzed and numbers ofactivated pixels in each region were compared between right and left sides. The reproducibility of functional mapswas tested by repetition of functional imaging in the same subjects. RESULTS: Statistically significant activationsignals were demonstrated in five of six subjects, in whom the distribution of those signals was predominantly inboth frontal lobes. Hemispheric lateralization of activation, when activated pixels were compared between bothinferior frontal gyri, was in all cases on the left. In four subjects, functional maps were reproduced in asimilar fashion. CONCLUSION: Our results suggest that functional MR imaging can depict the activation of motorlanguage function in the brain and can be used as a useful non-invasive method for determining the hemisphericdominance of language.