To assess volume of activation within posterior parietal and left prefrontal ROIs, we calculated the number voxels showing significantly greater activation during SWM relative to vigilance for each participant within each ROI. We then performed regression analyses to predict volume of activation from age, gender, and their interaction. To determine whether movement during fMRI scanning might affect results, we examined relationships between age and bulk motion in two ways. Both total number of removed repetitions and average movement in each direction throughout the task were examined in relation to age and gender using correlational analyses. The number of repetitions removed for excessive motion during the task declined with age . However, in brain regions demonstrating a relationship between SWM response and age, number of removed repetitions did not significantly relate to brain response , and the relationship between age and brain response in each cluster remained significant after controlling for number of removed repetitions. Mean rotational and translational motion were not significantly related to age. The average rotational movement throughout the task was 0.07, 0.22, and 0.09 degrees for roll, pitch, and yaw, respectively; the average translational movement was 0.16, 0.06, and 0.09 mm for superior, left, and posterior, respectively. There were no significant gender differences for number of repetitions removed for movement or on any directional movement parameter, with the exception of males demonstrating significantly greater rotational motion than females in the pitch direction = -2.08, p < .05.Age positively predicted SWM brain response in bilateral medial portions of superior frontal gyrus ; left superior and middle frontal gyri ;inferior aspects of the left precuneus and angular gyrus ; and a cluster encompassing the right inferior parietal lobule, postcentral gyrus, and insula .
A negative relationship between age and SWM response was observed in the left superior frontal gyrus ; left precuneus and superior parietal lobule ; superior portions of the right inferior parietal lobule ; and the right lingual gyrus . Exploratory follow-up analyses revealed that in the medial superior frontal cluster, teens evidenced less response during SWM than during vigilance,commercial drying racks with younger youths showing greater vigilance response than older teens. Further, in the right lingual gyrus, youths demonstrated less response during SWM than during rest , with older teens showing a greater decrease in SWM response relative to fixation than younger teens . In the left superior frontal gyrus , most participants showed no significant response to SWM relative to vigA significant age × gender interaction was observed in the right fronto polar superior frontal gyrus , in the same location as the gender difference described above . In this cluster, males showed a negative relationship between age and SWM response, but females showed a positive relationship.To understand whether age and gender related differences in BOLD response could be accounted for by task performance , we examined mediational models using a series of regressions . As vigilance reaction time was the only task performance index related to age or gender, regression analyses examined whether it mediated the relationship between age or gender and BOLD response in any of the clusters listed in Table 2. Vigilance reaction time was not significantly related to brain response in any region that was related to age or gender, and therefore did not mediate the relationship between age or gender and BOLD response. The posterior parietal ROI encompassing areas demonstrating significant activation to SWM relative to vigilance in young and old youths was 96,876 microliters, and spanned bilateral portions of the precuneus and superior and inferior parietal lobules. Within that cluster, ROI analyses demonstrated that activation for young adolescents was mostly in superior regions of parietal cortex, while response for old teens was mostly in inferior parietal areas .
Although age and the age × gender interaction did not predict volume of parietal activation within the combined parietal ROI, males demonstrated larger volumes of activations than females =5.62, p<.025. Further, males showed a significant negative relationship between age and volume of activation , while females showed no relationship between age and volume of response. Within the left prefrontal ROI , age significantly predicted the volume of activation, with larger volumes of activation demonstrated by older teens =4.95, p=.03. There was no significant effect of gender or the age × gender interaction on the volume of left prefrontal activation. This cross-sectional study examined the effects of age and gender on brain response during a SWM task among 12- to 17-year-olds. In general, we observed comparable task performance across the age range and between genders, and all teens showed typical response patterns for SWM, with activation in bilateral prefrontal and posterior parietal cortices. This pattern parallels adult activation during spatial working memory tasks and supports occipitoparietal, or “dorsal stream,” processing of spatial locations , suggesting that, in general, teens use similar working memory and spatial processing strategies as adults. However, specific localization and intensity of response varied across the adolescent age range, and males and females showed slightly different activations. These differential patterns emerged despite similar task performance across the age range and between genders, suggesting that developmental changes in SWM brain response are driven by factors other than task performance.In contrast to the literature suggesting that SWM abilities on n-back tasks improve across the adolescent age range , we did not observe age related improvements in performance on our SWM task. While this was likely due to the low difficulty level of the task used which approached ceiling effects, it is a benefit to the neuroimaging component of this study as it prevented confounding performance effects on the neural activation patterns observed across this age range.
Although task performance was not related to age, teens performed more accurately on vigilance than SWM, yet reaction times were faster on SWM. Our previous studies using this task demonstrated similar findings, with slightly faster performance on SWM than vigilance . While the reason for this difference is unclear, it could be that the small visual discrimination necessary for dot detection during vigilance blocks is more time consuming than the broader location detection required during SWM blocks. Future studies should attempt to eliminate this difference in reaction times between experimental and control conditions, perhaps by designing a task with easier visual discrimination .As hypothesized, age positively predicted SWM activation within the prefrontal cortex. Specifically,rolling grow table age was positively associated with both the intensity and extent of brain response in the left middle and superior frontal gyri . This cluster spanned frontopolar cortex but also encompassed parts of dorsolateral prefrontal cortex. Frontopolar prefrontal cortex activation has been associated with subgoal processing and evaluation of internally generated information . Thus, older teens may invoke more self-generated strategies, including rule induction or more efficient retrieval processes. This cluster also included portions of dorsolateral prefrontal cortex, which has been consistently implicated in working memory tasks. Adult studies have suggested that prefrontal activation is often left-lateralized during verbal working memory tasks ; thus the greater left prefrontal response among older teens may suggest that older teens employ more verbal rehearsal strategies during the task than younger adolescents. While the task was designed to minimize verbal encoding, older teens may have imagined the eight possible stimulus locations as positions on a clock face, facilitating verbal labeling and resulting in greater left prefrontal activity. Also consistent with our hypotheses, we observed a positive relationship between age and SWM response in posterior parietal regions. However, while we detected a positive relationship between age and SWM activation in bilateral inferior parietal regions, including inferior aspects of the right precuneus and left inferior parietal lobule, our data also revealed a negative relationship between age and brain response in bilateral superior parietal cortex, comprising superior portions of the right inferior parietal lobule, left precuneus and left superior parietal lobule. Exploratory ROI analyses confirmed these findings, demonstrating that while both young and old teens evidenced overlapping posterior parietal activation, younger youths showed activation mostly in superior regions, but older teens showed activation mostly in inferior regions. Together, these results indicate a shift from superior to inferior parietal areas utilized during SWM across adolescence. Previous fMRI studies of SWM have suggested that parietal activation intensity increases across adolescence , yet small sample sizes and different task designs may have prevented the observation of additional negative relationships identified in the current study. While functional parcellation of parietal involvement in sub-components of spatial working memory is largely unknown, some researchers of adult populations have suggested that superior parietal regions may be important for spatial rehearsal during working memory , while inferior parietal regions may be implicated in short-term storage during working memory .
Therefore, the superior to inferior shift in parietal activation across adolescence could represent a change in spatial working memory strategies. Younger adolescents may rely more on spatial rehearsal, which could become more automated throughout adolescence, requiring less superiorparietal activation. Along those lines, older adolescents may be better able to engage inferior parietal regions involved with spatial storage, and rely less on spatial rehearsal. Moreover, if older adolescents are employing greater verbal rehearsal strategies as discussed above, then spatial rehearsal may be less efficient, and therefore utilized to a lesser degree. In addition, stage of pubertal development was negatively associated with response in the superior right inferior parietal lobule cluster, above and beyond the effects of chronological age. Previous literature has demonstrated the impact of sex hormones on the development of cerebral lateralization , and pubertal timing has been related to functional asymmetry . Similarly, in this study, right parietal maturation appears linked to pubertal stage while left parietal development is not, suggesting asymmetrical cortical development that may be hormonally influenced. This finding points to the importance of individual variation in biological maturation that may not be accounted for by chronological age, and suggests that indices of pubertal development may further characterize neural maturation and help explain changes in SWM brain response patterns and cognitive strategies across adolescence. As well as showing changing fMRI response patterns to SWM tasks across adolescence , previous adolescent research has also demonstrated age-related increases in the spatial extent of frontal and parietal SWM activation . We found a greater number of significantly activated voxels in left prefrontal cortex with increasing adolescent age, suggesting that in some regions, both the magnitude of response and the volume of significant activation increase across adolescence. However, the results of our spatial extent analysis in posterior parietal cortex showed no significant relationship between age and volume of significant SWM response. Taken together with our results demonstrating age-related regional changes in the intensity of activation, these findings suggest that in late developing frontal brain regions, intense and more widespread activation emerges, while in slightly earlier developing posterior parietal networks, there is a focal shift in localization of activity. When examined in light of the adult working memory literature, adolescent age-related changes in frontal and parietal networks involved in SWM support the evolution of more efficient cognitive strategies. In the lingual gyrus, we observed deactivation that increased with age. Increasing occipital deactivation across adolescence could represent enhanced priming, and therefore greater recognition and reprocessing of repeated spatial locations among older youths. Teens in this study also demonstrated less response during SWM than during the vigilance condition in medial superior frontal cortex, yet this discrepancy dissipated across adolescence, such that this area was no longer “under active” in older teens. Medial frontal cortex is highly active at rest, during which it is involved in attentional monitoring of various internal and external stimuli . Medial frontal cortex underactivation during a cognitive task may represent reallocation of limited attentional resources to areas directly involved in task performance . Thus, SWM task demands may be more difficult for younger youths, who require greater attentional allocation to maintain performance, and therefore greater under-activation of medial frontal cortex. fMRI Response and Gender This is the first known fMRI study to attempt to examine the role of gender in relation to the neural substrates involved in SWM across adolescent development. While our findings do not entirely support the hypothesis that females would evidence more mature SWM response patterns than males, several interesting gender specific findings suggest that males and females utilize slightly different brain regions to perform well on a SWM task. Specifically, females demonstrated more right anterior cingulate response during the vigilance condition than did males.