| Working memory is a limited-resource cognitive system that plays a crucial role in goal-oriented activities.fMRI studies have shown that working memory tasks elicit widespread activation in bilateral frontoparietal regions,the activation levels of which are modulated by the content and process of experimental tasks.The effect of aging on brain activation is mainly manifested in the decreased activation of task-positive brain regions and the increased activation of task-negative brain regions.Few studies have used task difficulty above 2-back to examine the neural activity in the bilateral frontoparietal region during working memory tasks,as well as the changes in connectivity between different brain regions within the frontoparietal region across high and low working memory loads.Moreover,the connectivity between the core brain regions of the resting state brain network is significantly correlated with working memory performance,and aging also influences this relationship.However,most studies on resting state prediction tend to use seed analysis or specific functional networks.This paper attempts to use the whole brain networks to explore the interactions between broader brain networks and its relation to working memory performance.Furthermore,researchers tend to use a single behavioral indicator to explore the prediction of the resting state networks.This paper expects to use both task accuracy and reaction time as behavioral indicators to provide convergent evidence for the intricate network relationships in the aging process.This paper,using fMRI,investigated the effects of aging on the neural activity of working memory from two perspectives: the neural activity and functional connectivity of different brain regions influenced by working memory load,and the prediction ability of the resting state network functional connections on working memory performance.Study 1 aimed to investigate age differences in neural activity and functional connectivity in the brain regions affected by working memory load.We used the n-back task with four difficulty levels to conduct fMRI scanning during the working memory task in younger and elder people.During data analysis,8 brain regions of working memory load effect were obtained according to meta-analysis,and spherical ROIs of these brain regions were extracted with a radius of 8mm.First,ROI analysis was used to explore the changes of BOLD signal in the load-effect brain regions of young and old people during the n-back task under four loading conditions and the relationship between BOLD signal and task performance(i.e.,accuracy and reaction time).Next,g PPI analysis was used to explore functional connectivity differences between brain regions of load effect under high and low working memory load levels,the relationship between connectivity differences and task performance,and within-group differences in such relationship.Study 2 aimed to investigate the predictive ability and age differences of resting state network functional connectivity strength in working memory performance.We used the brain atlas to parcel the resting brain network,the strength of the resting functional network connection was used as the feature,and the performance of working memory task was used as the predictive variable.The RVR algorithm was used to establish a prediction model for the performance of working memory task of the elderly and young people respectively at four difficulty levels.At the same time,the prediction differences at different working memory load levels,the prediction differences in different behavioral indicators,and the age differences in predictability were also explored.In study 1,ROI analysis found that only left v PMC showed significant age difference,while the BOLD signals of FEF,right DLPFC and right v PMC had significant interaction between age groups and load level.Only the BOLD signals of ACC and FEF brain regions of young people showed a linear trend with the increasing task difficulty.By the g PPI method,we observed the age difference in the interaction between the brain regions of working memory load under high vs.low working memory load conditions.In young people,the functional connection strength between ACC and bilateral DLPFC,left v PMC and right FEF under low load condition was significantly higher than that under high load condition.The functional connection strength between ACC and left v PMC and left v PMC and right DLPFC in the elderly under low load condition was significantly higher than that under high load condition.Compared with the young people,ACC in the elderly lost the ability and sensitivity to adjust the coupling between frontoparietal brain region to cope with the increased difficulty.In study 2,RVR was used to establish a prediction model,and we found that the functional connection strengths of resting brain networks could predict the working memory performance.We also found that functional connection strengths could predict the task accuracy of the elder better,and the age-group advantage of this prediction ability was maintained in all four task conditions.In contrast,the age differences in the predictive ability of the models in response to tasks were more complex,with better predictive ability for young people at low load levels and better predictive ability for older people at high load levels.More importantly,prediction of task performance in older adults requires support not only from higher executive control networks such as DMN and SAN,but also from primary information systems such as VIN,cerebellum and sensorimotor networks,as well as subcortical networks.This paper investigates the aging effect on the neural activities of working memory,and supports the decline in working memory neural activities due to aging from the perspectives of task state and resting state.Study 1 investigated the effects of aging on the activation and functional connectivity of brain regions involved in working memory load.It revealed that the elderly brain exhibited reduced neural efficiency,neural dedifferentiation and neural compensation in the working memory process,which shed light on the mechanism of cognitive decline.Study 2 used the resting state whole brain functional network to model and predict working memory performance.It found that the resting state functional connectivity had a better predictive ability for working memory accuracy in the elderly,and revealed the complex interactive relationship between the higher cognitive network and the primary information processing network.It also reflected the reorganization of functional network,the reduction of functional network stability and the enhancement of functional network flexibility in the resting state of the elderly brain.This paper enriches the research content of predicting working memory performance by resting state,and provides convergent evidence for the complex relationship between the higher cognitive network and the primary information processing network in the aging process.In practical application,the use of resting state network can avoid the influence of factors such as operation complexity in task state,cooperation degree of special groups such as the elderly and task difficulty.It can also improve the accurate assessment of individual cognitive ability and risk factors,and select the optimal resting state network connectivity features for prediction according to different task performance indicators.This can improve sensitivity to individual diagnosis and intervention of working memory decline,and enhance quality of life in older adults. |
PDF Full Text Request