Neurocognitive Aging and the Compensation Hypothesis

The most unexpected and intriguing result from functional brain imaging studies of cognitive aging is evidence for age-related overactivation: greater activation in older adults than in younger adults, even when performance is age-equivalent. Here we examine the hypothesis that age-related overactivation is compensatory and discuss the compensation-related utilization of neural circuits hypothesis (CRUNCH). We review evidence that favors a compensatory account, discuss questions about strategy differences, and consider the functions that may be served by overactive brain areas. Future research directed at neurocognitively informed training interventions may augment the potential for plasticity that persists into the later years of the human lifespan. KEYWORDS—plasticity; dedifferentiation; brain imaging; working memory Brain imaging has become a method of great importance for studying cognitive aging, which makes sense because the latter presumably results from neurobiological aging. Therefore, brain-based measurements that can be linked to cognitive processes expand the range of questions that can be addressed about the aging mind. The emerging answers have prompted new ways to think about the normal aging process and about functional brain organization across the lifespan. Before the advent of brain imaging, the behavioral methods and interpretive logic of clinical neuropsychology guided brain-based theories of cognitive aging. This approach assumes that minimal age differences in performance imply minimal alterations in underlying cognitive mechanisms and, by extension, age-invariance in the neural substrates that mediate them. In our assessment, one of the most far-reaching discoveries to have thus far emerged from brain imaging studies of aging is that this assumption is erroneous. The initial neuroimaging studies of cognitive aging, which measured brain activation via the distribution of a radioactive isotope (i.e., positron emission tomography, PET; Grady et al., 1994), noted that older adults display activation in regions that are not activated by younger adults performing the same tasks. In some studies, sites of overactivation co-occur with regions that are underactive relative to younger adults. In other studies, regions of overactivation are the only indication that older brains function differently than younger brains (for reviews, see Grady & Craik, 2000; Reuter-Lorenz, 2002). The terms overactivation and underactivation are purely relative, referring to sites that senior adults activate more or less, respectively, than their younger counterparts (Fig. 1). Overactivation is frequently observed in prefrontal sites (Cabeza et al., 2004; Reuter-Lorenz et al., 2000). Overactivation in seniors is often found in regions that approximately mirror active sites in younger adults but in the opposite hemisphere (e.g., Cabeza, 2002; Reuter-Lorenz et al., 2000; see the lower left panel of Fig. 1). This pattern of reduced asymmetry in older adults has been referred to as hemispheric asymmetry reduction in older age, or HAROLD for short (Cabeza, 2002). INTERPRETING OVERACTIVATION Age-related underactivation is typically interpreted as a sign of impairment due to poor or underutilized strategies or due to structural changes such as atrophy. However, the cognitive and neural mechanisms associated with age-specific regions of overactivation are more ambiguous. Determining whether overactivations are neural correlates of processes that are beneficial, detrimental, or inconsequential to cognitive function is the crux of many research efforts in the cognitive neuroscience of aging (Reuter-Lorenz & Lustig, 2005). Because overactivation has been found for a broad range of tasks, across a variety of brain regions, with or without age differences in performance, and with or without concurrent underactivation, it is highly unlikely that all instances stem from a single cause. Unsurprisingly, when overactivation is found in association with poor performance, it is interpreted as impairAddress correspondence to Patricia A. Reuter-Lorenz, Department of Psychology, University of Michigan, 530 Church Street, Ann Arbor, MI 48109-1043; e-mail: [email protected]. CURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE Volume 17—Number 3 177 Copyright r 2008 Association for Psychological Science ment and is typically attributed to any of several potentially related mechanisms: the use of multiple and/or inefficient cognitive strategies; disinhibition because communication between the left and right hemispheres declines; or dedifferentiation, whereby the specificity and selectivity of neural processors break down. In many studies, however, overactivation is accompanied by age-equivalent performance, raising the possibility that the additional activity serves a beneficial, compensatory function without which performance decrements would result (see Fig. 1). Regardless of whether performance matching is achieved by selecting younger and older subgroups that perform at equivalent levels, providing different amounts of training, adopting age-tailored stimulus parameters, or otherwise altering task demands for each age group, overactivation has been found consistently across perceptual, motoric, mnemonic, verbal, and spatial domains. The compensation hypothesis predicts that, even while performance is matched at the group level, overactivation across individuals should be correlated with higher performance in the older group. Although significant correlations may sometimes be lacking due to insufficient variability or a lack of statistical power, positive activation–performance correlations have been reported, lending support to the compensatory account of age-specific overactivations (Fig. 1; Cabeza et al., 2004; Reuter-Lorenz & Lustig, 2005). Establishing that overactive sites in older adults contribute to and are necessary for successful performance would provide especially strong support for the compensation hypothesis. Transcranial magnetic stimulation (TMS) is a technique that applies a series of focally directed magnetic pulses to the scalp to stimulate the underlying neural tissue. TMS can be applied in either a deactivating or an activating mode. In the deactivating mode, TMS temporarily disrupts the underlying neural signals, producing a virtual, transient lesion. Using this mode, Rossi et al. (2005) showed that overactive sites in seniors contributed to performance success: Older adults, who typically show bilateral prefrontal activation during recognition memory, were impaired by TMS to either hemisphere, suggesting that recognition relies on both sides. Younger adults, who activate unilaterally during recognition memory, were impaired by TMS to only one side. When used in the activating mode, TMS increases the contribution of the underlying tissue. Another study found that, when TMS was applied prefrontally in the activating mode, a group of low-performing elderly showed improvement; furthermore, functional magnetic resonance imaging (fMRI) showed their brain activation to be unilateral before TMS and bilateral after TMS, in association with their improved performance (Sole-Padulles et al., 2006). COMPENSATION FOR WHAT? The compensation hypothesis assumes that overactive sites in older adult brains are ‘‘working harder’’ than the corresponding regions in their younger counterparts. In the aging brain, a network may work harder, and thus overactivate, to make up either for its own declining efficiency or for processing deficiencies elsewhere in the brain. Although definitive support for the first possibility is currently lacking, such support could come from work using multiple measures to assess structural and functional integrity within the same subjects. For example, volumetric measures could reveal age-related atrophy in a region that also displays overactivation. When also coupled with preserved performance, such a pattern would suggest that increased recruitment compensates for decline (cf., Persson et al., 2006). Alternatively, a network may need to work harder and thus becomes overactive because the input it receives is degraded or compromised. By this account, overactivation is compensating for functional declines elsewhere. We see three types of evidence as being consistent with this possibility. First, several studies Fig. 1. Results typically referred to as ‘‘underactivation’’ (top) and ‘‘overactivation’’ (bottom). When older adults activate a brain region at lower levels or show a smaller extent of activation compared to younger adults, as illustrated in the top pair of images, the results are often interpreted to indicate that the older group is functionally deficient in the processing operations mediated by this region. The overactivation pattern in the bottom pair of images illustrates the hemispheric asymmetry reduction in older age (or HAROLD) effect: Younger adults show activation that is lateralized to the left hemisphere, whereas the older adults are activating homologous brain regions in the opposite hemisphere also. See Reuter-Lorenz and Lustig (2005) for examples of studies reporting these age-specific activation patterns. 178 Volume 17—Number 3 Neurocognitive Aging and the Compensation Hypothesis