Age-related memory decline is the consequence of multiple biological factors that lead to brain structural and functional change, including gray matter atrophy, white matter injury, and loss of functional coordination between regions. However, the independent roles that each of these brain changes play in mediating memory decline is not clear. Therefore, we used magnetic resonance imaging (MRI) to measure gray matter (GM) volume, white matter hyperintensity (WMH) volumes, and blood oxygen level-dependent (BOLD) functional magnetic resonance imaging-based functional connectivity among default mode network nodes in 76 cognitive normal older adults. We found that GM, WMH, and connectivity between left inferior parietal and medial prefrontal cortex (MPF_LIP) were independently associated with episodic memory performance. Within the group with GM volumes below the median, greater MPF_LIP connectivity was associated with better memory performance, whereas this association was not present for individuals with GM volume above the median. These findings confirm the heterogeneous nature of brain-behavior relationships in cognitive aging. In addition, the relationship between resting state functional connectivity and memory performance, particularly amongst those individuals with more brain atrophy, strongly suggests compensation against the effects of neuronal injury.
Decline in memory performance is a hallmark of the aging process (Grady, 2000 and Grady and Craik, 2000). Successful performance on memory tasks, including encoding, consolidation, and retrieval, requires recruitment of several brain regions within medial temporal (Scoville and Milner, 1957, Squire, 2004 and Squire et al., 2004), prefrontal, and parietal cortical areas (Frings et al., 2010, Jenkins and Ranganath, 2010 and Staresina and Davachi, 2010). While activity in these regions tends to increase during memory task performance, activity in an additional set of brain regions in the prefrontal, parietal, precuneus, posterior cingulate, and medial temporal cortices, termed the default network (DMN), correspondingly decreases (Raichle et al., 2001). Because blood oxygen level-dependent (BOLD) functional magnetic resonance imaging (fMRI)-based measures of connectivity among nodes in the DMN are known to be associated with memory performance in cognitively healthy elderly individuals (Damoiseaux and Greicius, 2009 and Dickerson and Sperling, 2009), DMN connectivity has emerged as a potentially novel and sensitive measure of subtle brain injury associated with diminished cognition among otherwise healthy elders.
However, the independent value of DMN connectivity as a marker of age-related memory decline is unclear because its reductions occur against a backdrop of aging-associated neuronal injury (Greicius et al., 2004 and Seshadri et al., 2007). Advancing age is associated with reduced integrity of white matter (WM) in frontal regions that participate in memory function (O'Sullivan et al., 2001 and Pfefferbaum et al., 2000), as well as posterior regions (Lee et al., 2009 and Lee et al., 2010), white matter hyperintensities (WMH) are common (Ylikoski et al., 1995) and can reflect damage to WM pathways required for memory performance (Pantoni and Garcia, 1997). Gray matter (GM) volume loss across the brain in older adults is also prominent, especially in frontal regions required for memory function (Raz et al., 1998). Growing evidence suggests that both GM and WM injury may independently contribute to age-related memory declines (Buckner, 2004 and Hedden and Gabrieli, 2004).
Thus, in the context of brain aging, it is unclear whether DMN connectivity reduction is simply a downstream result of injury to the underlying neuronal architecture, a functional consequence of neuropathology on cognition independent of structural injury to neurons (e.g., functional toxicity; Shanker et al., 2008), a background factor that helps to buffer the brain against the effects of neuronal injury, or some combination of all these factors. Addressing the independent role that DMN integrity has on cognitive aging in the absence of clinical impairment, therefore, may prove important to understanding some of the most basic mechanisms by which older individuals either maintain memory ability or develop impairment in memory.
In this study, we examined the independent effects of GM volume, WMH volume, and DMN integrity on memory performance in a group of cognitively normal older adults. We hypothesized that because aging-associated brain changes exacerbate neuronal dysfunction, even among neurons that are not yet structurally injured, retained coordination among DMN nodes would be associated with better memory performance even after accounting for the effects of GM and WM injury on memory performance..
