By the year 2075, more than 50% of the European population will be aged 60 or more, and there will be a three-fold increase in individuals aged 80 or more, relative to the present day. It is important that our society prepares for this demographic change and endeavours to enable older adults to optimise their quality of life and autonomy for as long as possible. To the extent that age-related cognitive decline is one of the biggest threats to independent living and well-being for this cohort, the field of cognitive neuroscience is arguably the discipline with the most potential to help in this regard. Non-pathological ageing is accompanied by several cognitive and brain changes that are a product of the natural ageing process, one’s environment, and one’s ability to compensate for them. These changes become evident in multiple cognitive dimensions. On the one hand, older adults have improved regulation of emotion, better vocabulary, better culture-related knowledge, and have better life satisfaction, compared to younger adults. On the other, they have reduced acuity of the senses, they require more time to both process, and respond to, sensory information, and invariably, they undergo declines in a number of other important physical and cognitive capacities. As the deterioration of older adults’ cognitive capacities begins to occur, the ability to monitor and evaluate the success of their cognitive processes is of paramount importance for detecting errors, and calibrating their daily activities to suit their strengths and weaknesses. Yet, the extent to which these metacognitive monitoring processes are affected by the natural ageing process has rarely been considered in the literature. A wealth of evidence from research on clinical populations indicates that metacognitive capacities are highly susceptible to disruption in several diverse neurological conditions, particularly those with damage to right frontal regions. Considering there is much evidence to suggest that the frontal lobe is one of the brain regions that undergoes the most extensive age-related changes (Dempster, 1992; Mosocovitch & Wincour, 1992; Raz, Gunning, Head et al., 1997; West, 1996), the question follows whether metacognitive capacities are also vulnerable to disruption due to the natural ageing process.
The present chapter is organised in six main sections. In the proceeding section, an overview of different perspectives on ageing at the neuropsychological and neurobiological level is provided. The third section introduces the topic of metacognition and draws on the clinical literature surrounding anosognosia to highlight the importance of metacognitive abilities, how they are measured, and what is known about their neuropsychological and neuroanatomical bases. The fourth and fifth sections provide more focussed reviews of the cognitive neuroscience literature on performance monitoring and conscious error awareness, respectively, which in the context of this thesis, are hypothesised to be critical to the accuracy of many metacognitive abilities. The sixth and final section provides an overall summary and an outline of the objectives of this thesis.
Age-Related Cognitive and Cerebral Decline
A robust, and positive, finding to emerge from cognitive ageing research is that age-related losses are not necessarily seen across all cognitive functions. Patterns of relative preservation versus decline are usually particularly apparent for what are known as crystallized versus fluid intelligence domains (Horn & Cattell, 1967). These two clusters of intellectual abilities have also been discussed in terms of the pragmatics and mechanics of cognition (Baltes, Lindenberger, & Staudinger, 1998). The former constitutes a culture-related knowledge base that is accumulated through experience, whereas the latter constitutes content-free information processing that relies on fundamental biological processes, and therefore mainly on the integrity of the central nervous system. Both cross-sectional (Lindenberger & Baltes, 1995) and longitudinal studies (Schaie 1996; 2005) have indicated that processes in the fluid intelligence domain begin to decline from middle adulthood on, whereas capacities in the crystallized domain improve from childhood right through to adulthood and then either remain stable or continue to improve until very late in life. As such many authors have described cognitive ageing as a multi-dimensional and multi-directional process.
Perspectives on cognitive ageing at the behavioural level
Many theories have been proposed to explain age-related declines and individual differences in cognitive functioning. Given that it is not within the objectives of this thesis to test any particular theory of cognitive ageing, this section provides a brief overview, as opposed to a comprehensive review, of the main perspectives on cognitive ageing at the behavioural level. This section is then followed by an overview of literature on cognitive ageing at the neurobiological level.
The cognitive control hypothesis
Common to many of the processes that witness age-related declines is a reliance on cognitive control. Cognitive control is critical to a range of higher order processes that allow for the regulation of sensory information and behaviour in accordance with one’s goals. These processes include monitoring, sequencing, initiation of action, inhibiting pre-potent responses, formulating goals, focusing attention and generating response alternatives (Fuster, 2000; Miller, 2000; Miller & Cohen, 2000). These higher order control processes are also frequently referred to as executive functions (Baddeley, 1986; Norman & Shallice, 1986; Shallice, 1998), and are predominantly mediated by the frontal lobes. Age-related differences are consistently observed on tasks that place high demands on cognitive control, including working memory (Borella, Ghisletta, & de Ribaupierre, 2011; Hasher & Zacks, 1988; Salthouse, 1994), attention (McAvinue; McDowd, 1986; Milham, Erickson, Banich et al., 2002; Hawkins, Kramer, & Capaldi, 1992; West, 2004), multi-tasking (Clapp, Rubens, Sabharwal & Gazzaley, 2011; Jimura & Braver, 2010), as well as episodic and source memory (Craik, Morris, Morris, & Loewen, 1990). In contrast, older adults’ performance on measures of non-declarative or implicit memory, which are believed to rely on more automatic and less control demanding processes, has been found to be largely age invariant (Bergerbest, Gabrieli, Whitfield-Gabrieli et al., 2009; Fleischman & Gabrieli, 1998; Light & Singh, 1987; La Voie & Light, 1994). Such observations have prompted many authors to propose that age-related cognitive decline may arise from impaired or inefficient deployment of cognitive control processes due to age-related degeneration of frontal lobe structures (Braver & Barch, 2002; Crawford, Bryan, Luscez, Obonsawin, & Stewart, 2000; Glisky, 2007; Greenwood, 2000; West, 2000; Rodriguez-Aranda & Sundet, 2006). This general idea has been variously termed the “cognitive control hypothesis” (West, 1996; 2000; Gallo, Bell, Beier, & Schacter, 2006; Koutstaal, 2006) “frontal lobe hypothesis” (West, 2000), “frontal ageing hypothesis” (Greenwood, 2000), “executive decline hypothesis” (Crawford et al., 2000), and “frontal hypothesis” (Rodriguez-Aranda & Sundet, 2006). In support of this idea executive functions have been found to mediate the relationship between age and general cognitive capacities (Salthouse, Atkinson, & Berish, 2003) and have explained age-related differences in learning and memory (Brooks, Kempe, & Sionova, 2006; Crawford et al., 2000). Furthermore, when young and older adults’ performance on putative tests of frontal, temporal, and parietal functions were compared, the strongest correlation to emerge was between age and frontal measures, with advancing age being predictive of decreasing performance on frontal lobe measures (Mittenberg, Seidenberg, O’Leary, & Digioulo, 1989).
The processing-speed hypothesis
Salthouse (1996) has argued that age-related deficits in controlled processing are secondary to a generalised reduction in the processing speed of underlying cognitive operations. Behavioural slowing has long been considered a primary concomitant of the ageing process. Christensen & Kumar (2003) have suggested that processing speed peaks in the early 20s and then declines by approximately 20% by the age of 40, and by up to 40-60% by the age of 80. Age-related declines in processing speed have been attributed to a general slowing of information processing (Birren & Fisher, 1995) or increased neural noise (Welford, 1965) within the central nervous system with advancing age. In support of the processing-speed theory it has been observed that age differences on several capacities in the fluid domain, such as abstract reasoning, working memory, and problem solving were attenuated after statistically controlling for processing speed (Bors & Farrin, 1995; Salthouse, 1996; Salthouse & Babcock, 1991; Zimprich & Martin, 2002). Speed of processing was also found to be the main predictor of age-related changes in memory and spatial ability (Finkel & McGue, 1993).
The inhibitory deficit hypothesis
Hasher and Zacks (1988) advanced that a selective deficit in inhibitory control processes may constitute a global cognitive ageing phenomenon. More specifically, this theory assumes that in order for goals to be fulfilled effectively, automated responses to non-goal relevant information need to be suppressed. However, age-related reductions in inhibitory control enable non-goal relevant information to vie for attentional resources, which results in greater distractibility, slowed and error-prone behaviour, and greater forgetting rates (Lustig, Hasher & Zacks, 2007; Hasher & Zacks, 1998). Age-related declines in inhibitory control and increased susceptibility to distractors have been found to explain a considerable proportion of age-related variance in working memory capacity (Hasher, Zacks, & May, 1999). In a more recent study, both processing speed and inhibition were identified as independent mediators of age differences in working memory capacity (Borella, Ghisletta, & de Ribaupierre, 2011).
Dedifferentiation and cognitive permeation
Many studies have reported that the statistical correspondence between sensory and sensorimotor abilities such as vision, hearing, balance, and gait, and intellectual abilities in both fluid and the crystallized domain is significantly greater in older adults than in young adults (e.g. Baltes & Mayer, 1999). Moreover, it has been found that for older adults sensory functioning is a stronger predictor of capacities in the fluid domain than a comprehensive set of sociobiographic factors (Baltes & Lindenberger, 1997). This apparent loss of domain specificity with increasing age has been termed “dedifferentiation.”
A number of authors have proposed that this apparent dedifferentiation of functions may be attributable to sensory and sensorimotor functions placing greater demands on attentional control resources. This has become known as the cognitive permeation hypothesis (e.g. Lindenberger, Marsiske, & Baltes, 2000; Schafer, Huxhold, Lindenberger, 2006). According to this hypothesis, resource overlap and competition amongst domains increases with advancing age, and compensation in the form of resource allocation trade-offs become more frequent (Li & Lindernberger, 2002; Schafer et al., 2006). In accord with this, Li et al. (Li, Lindenberger, Freund & Baltes, 2001) have shown that balance during walking was preserved at the expense of performance of a simultaneously executed cognitive task. Such findings suggest that age-related declines in cognitive domains could be attributable to increased allocation of attentional resources to processes that were previously automated.
Another important conceptual framework labelled ‘cognitive reserve,’ concerns how older adults may be able to draw on a pool of accumulated resources to maintain cognitive function. The notion of cognitive reserve emerged from recurrent observations that levels of cognitive impairment did not always manifest to the extent that would be expected from a given brain pathology (Stern, 2002). For instance, Katzman et al. (Katzman, Terry, DeTeresa et al., 1998) have reported that older adults can be cognitively intact up until they die, but exhibit advanced AD-related cerebral pathology at post-mortem. Such discrepancies have also been observed in a range of other conditions including stroke (Ojala-Oksala, Jokinen, Kopsi et al., 2012) and traumatic brain injury (TBI; Kesler, Adams, Blasey, & Bigler, 2003).
This apparent elevation of threshold for cognitive impairment appears to be promoted by factors such as high levels of education, occupational complexity, and participation in cognitively stimulating leisure activities (Mortimer, 1997). It has been proposed that cognitive reserve may mediate individual differences in non-pathological cognitive ageing by fostering more efficient utilisation of brain networks or an enhanced ability to recruit alternate networks (Stern, 2002).
The natural ageing process is associated with myriad cognitive changes. Some of the most pronounced and consistently reported are on tasks that challenge cognitive control processes and working memory, or that require long term working memory (Hedden & Gabrieli, 2004; Piguet & Corkin, 2007). Several hypotheses about cognitive ageing at the behavioural level have been advanced, and each hypothesis described above continues to feature prominently in recent literature. However, it is difficult to arbitrate between these theories in the absence of neural evidence. The next sub-section will outline how the increasing availability of neuroimaging technologies has provided important new insights into the relationship between age-related changes in brain structure and function, and concomitant changes in cognitive abilities.
Perspectives on cognitive ageing at the neurobiological level
In the same way that ageing does not have an equal impact on all cognitive domains, ageing does not result in a general deterioration of the brain. Rather, the ageing brain is characterised by a ‘patchwork pattern of differential declines and relative preservation,’ not only at the structural level, but also at the functional level (Raz, 2000).
Grey matter integrity
Magnetic resonance imaging (MRI) based studies consistently show a global age-related reduction in grey matter volumes, but considerable regional differences exist in terms of the magnitude and relative rate of change. In a longitudinal study, which spanned five years, Raz et al. (Raz, Lindenberger, Rodrigue et al., 2005) found a significant negative association between age and volume in the lateral prefrontal cortex, the orbitofrontal cortex, the cerebellum, the caudate and the hippocampus. These associations were found to be stronger after five years for the prefrontal regions, the cerebellum, the caudate and the hippocampus, indicating age-related accelerations in the shrinkage of these regions. Conversely, volumes in areas such as the primary visual cortex, the fusiform cortex and the inferior parietal lobes were not significantly associated with age, and there was no change in these associations over the course of five years. Several other studies using a variety of methods have reported similar findings, and in particular, an ever-growing literature documents the most dramatic age-related grey matter