Visuomotor transformations for reaching to memorized targets: a PET study.

Abstract

Positron emission tomography (PET) was used to identify cortical and subcortical regions involved in the control of reaching to visual targets. Regional cerebral blood flow (rCBF) was measured in eight healthy subjects using H2(15)O PET during the performance of three different tasks. All tasks required central fixation while a 400-ms target was flashed every 5 s at a random location around a virtual circle centered on the fixation target. Additional instructions differed according to the task: (i) visual detection of the target without overt responses; (ii) immediate pointing to the most recent target in the sequence, and (iii) pointing to the previous target in the sequence. By design, the two motor tasks differed in the cognitive processing required. In each trial of immediate pointing, the spatial location of only the most recent target needed to be processed. In each trial of pointing to the previous, instead, while the most recent target was stored in memory for the movement of the next trial, the previous target had to be retrieved from memory to direct the current movement. Limb trajectories were comparable between the two motor tasks in terms of most spatiotemporal parameters examined. Significant rCBF increases were identified using analysis of covariance and t statistics. Compared with visual detection there was activation of primary sensorimotor cortex, ventrolateral precentral gyrus, inferior frontal gyrus in the opercular region, supramarginal gyrus, and middle occipital gyrus, all these sites in the hemisphere (left) contralateral to the moving limb, and cerebellar vermis, during both immediate pointing and pointing to the previous. During immediate pointing there was additional activation of left inferior parietal lobule close to the intraparietal sulcus, and when compared with pointing to the previous, dorsolateral prefrontal cortex bilaterally. During pointing to the previous, instead, there was additional activation of supplementary motor cortex, anterior and midcingulate, and inferior occipital gyrus in the left hemisphere; superior parietal lobule, supramarginal gyrus, and posterior hippocampus in the right hemisphere; lingual gyri and cerebellar hemispheres bilaterally; anterior thalamus; and pulvinar. The activation of two partially distinct cerebral networks in these two motor tasks reflects the different nature of signal processing involved. In particular, the specific activation of intraparietal sulcus and prefrontal cortex in immediate pointing appears characteristic of a network for visuospatial working memory. By contrast, the corticolimbic network engaged in pointing to the previous could mediate spatial attention and the sequence of encoding, recording, and decoding of spatial memories required by a dual task with two competing targets.