Functional anatomy of the mental representation of upper extremity movements in healthy subjects.

1. Differences in the distribution of relative regional cerebral blood flow during motor imagery and execution of a joy-stick movement were investigated in six healthy volunteers with the use of positron emission tomography (PET). Both tasks were compared with a common baseline condition, motor preparation, and with each other. Data were analyzed for individual subjects and for the group, and areas of significant flow differences were related to anatomy by magnetic resonance imaging (MRI). 2. Imagining movements activated a number of frontal and parietal regions: medial and lateral premotor areas, anterior cingulate areas, ventral opercular premotor areas, and parts of superior and inferior parietal areas were all activated bilaterally when compared with preparation to move. 3. Execution of movements compared with imagining movements led to additional activations of the left primary sensorimotor cortex and adjacent areas: dorsal parts of the medial and lateral premotor cortex; adjacent cingulate areas; and rostral parts of the left superior parietal cortex. 4. Functionally distinct rostral and caudal parts of the posterior supplementary motor area (operationally defined as the SMA behind the coronal plane at the level of the anterior commissure) were identified. In the group, the rostral part of posterior SMA was activated by imagining movements, and a more caudoventral part was additionally activated during their execution. A similar dissociation was observed in the cingulate areas. Individual subjects showed that the precise site of these activations varied with the individual anatomy; however, a constant pattern of preferential activation within separate but adjacent gyri of the left hemisphere was preserved. 5. Functionally distinct regions were also observed in the parietal lobe: the caudal part of the superior parietal cortex [medial Brodmann area (BA) 7] was activated by imagining movements compared with preparing to execute them, whereas the more rostral parts of the superior parietal lobe (BA 5), mainly on the left, were additionally activated by execution of the movements. 6. Within the operculum, three functionally distinct areas were observed: rostrally, prefrontal areas (BA 44 and 45) were more active during imagined than executed movements; a ventral premotor area (BA 6) was activated during both imagined and executed movements; and more caudally in the parietal lobe, an area was found that was mainly activated by execution presumably SII. 7. These data suggest that imagined movements can be viewed as a special form of "motor behavior' that, when compared with preparing to move, activate areas associated heretofore with selection of actions and multisensory integration.(ABSTRACT TRUNCATED AT 400 WORDS)