| The major goal of this research was to identify the neural correlates of working memory (WM) development in the human brain by comparing activations, as measured by functional magnetic resonance imaging (fMRI), between children (ages 7--12 years) and adults (ages 20--29 years). WM refers to mental resources supporting maintenance of goal-relevant information in mind, and it is known that WM is taxed by the amount of information kept in mind (WM load). We compared children and adults performing separate tests of verbal and spatial WM, while WM load was varied. Behaviorally, children made a greater number of errors than adults and the difference between groups grew as WM load increased for both verbal and spatial tasks. Consistent with hemispheric lateralization, children and adults demonstrated greater activation in the right hemisphere for spatial WM, and greater activation in the left hemisphere for verbal WM. The major difference between groups was that for both verbal and spatial WM maintenance, children failed to exhibit the same degree of increasing brain activation across WM loads. Thus, there were interactions between age and WM load in left and right inferior parietal cortex (BA 40), right superior parietal cortex (BA 40), left and right inferior frontal gyrus (BA 44,45), left medial frontal gyrus (BA 6), and left middle frontal gyrus (BA 46). The difference between children and adults remained large even when groups were equated for accuracy at different WM loads. Individual differences in WM accuracy were correlated with magnitude of fMRI response in children. Therefore, it appears that children are unable to marshal neural resources as well as adults in order to meet the demands of maintaining a large amount of information in WM. Additional studies were conducted to address the primary challenges of fMRI research in children. Those studies showed similar results in children and adults in the distribution and strength of fMRI signal response and also in participant views on the subjective experience of being scanned. Where results of these studies differed by group, additional methods for measuring and calibrating group differences in BOLD response were tested. |
