Carme Junqué

Healthy aging (HA) is associated with certain declines in cognitive functions, even in individuals that are free of any process of degenerative illness. Functional magnetic resonance imaging (fMRI) has been widely used in order to link this age-related cognitive decline with patterns of altered brain function. A consistent finding in the fMRI literature is that healthy old adults present higher activity levels in some brain regions during the performance of cognitive tasks. This finding is usually interpreted as a compensatory mechanism. More recent approaches have focused on the study of functional connectivity, mainly derived from resting state fMRI, and have concluded that the higher levels of activity coexist with disrupted connectivity. In this review, we aim to provide a state-of-the-art description of the usefulness and the interpretations of functional brain connectivity in the context of HA. We first give a background that includes some basic aspects and methodological issues regarding functional connectivity. We summarize the main findings and the cognitive models that have been derived from task-activity studies, and we then review the findings provided by resting-state functional connectivity in HA. Finally, we suggest some future directions in this field of research. A common finding of the studies included is that older subjects present reduced functional connectivity compared to young adults. This reduced connectivity affects the main brain networks and explains age-related cognitive alterations. Remarkably, the default mode network appears as a highly compromised system in HA. Overall, the scenario given by both activity and connectivity studies also suggests that the trajectory of changes during task may differ from those observed during resting-state. We propose that the use of complex modeling approaches studying effective connectivity may help to understand context-dependent functional reorganizations in the aging process.

Neural correlates of working memory (WM) in healthy subjects have been extensively investigated using functional MRI (fMRI). However it still remains unclear how cortical areas forming part of functional WM networks are also connected by white matter fiber bundles, and whether DTI measures, used as indices of microstructural properties and directionality of these connections, can predict individual differences in task performance. fMRI data were obtained from 23 healthy young subjects while performing one visuospatial (square location) and one visuoperceptual (face identification) 2-back task. Diffusion tensor imaging (DTI) data were also acquired. We used independent component analysis (ICA) of fMRI data to identify the main functional networks involved in WM tasks. Voxel-wise DTI analyses were performed to find correlations between structural white matter and task performance measures, and probabilistic tracking of DTI data was used to identify the white matter bundles connecting the nodes of the functional networks. We found that functional recruitment of the fusiform and the inferior frontal cortex was specific for the visuoperceptual working memory task, while there was a high overlap in brain activity maps in parietal and middle frontal areas for both tasks. Axial diffusivity and fractional anisotropy, of the tracts connecting the fusiform with the inferior frontal areas correlated with processing speed in the visuoperceptual working memory task. Although our findings need to be considered as exploratory, we conclude that both tasks share a highly-overlapping pattern of activity in areas of frontal and parietal lobes with the only differences in activation between tasks located in the fusiform and inferior frontal regions for the visuoperceptual task. Moreover, we have found that the DTI measures are predictive of the processing speed.