Is retaining the youthful functional anatomy underlying speed of information processing a signature of successful cognitive ageing? An event-related fMRI study of inspection time performance

Abstract

It has been hypothesized that individual differences in cognitive ageing might in part be based on the relative preservation of speed of information processing. However, the biological foundations of processing speed are not understood. Here we compared two groups of non-demented older people who had relatively similar IQs at age 11 but differed markedly in non-verbal reasoning ability at age 70: ‘cognitive sustainers’ (n = 25), and ‘cognitive decliners’ (n = 15). Using an event-related fMRI design, we studied the BOLD response while they performed an inspection time task. Inspection time is a two-alternative forced choice, backward masking test of the speed of the early stages of visual information processing. Inspection time has a well-established, significant association with higher cognitive abilities. The group of cognitive sustainers showed a pattern of BOLD activation–deactivation in response to inspection time stimulus duration differences that was similar to a healthy young sample [Deary, I.J., Simonotto, E., Meyer, M., Marshall, A., Marshall, I., Goddard, N., Watdlaw, J.M., 2004a. The functional anatomy of inspection time: an event-related fMRI study. NeuroImage 22, 1466–1479]. The group of cognitive decliners lacked these clear neural networks. The relative preservation of complex reasoning skills in old age may be associated with the preservation of the neural networks that underpin fundamental information processing in youth.

Introduction

There is a growing proportion of older people in society. Therefore, the problems of the older person have become a higher priority for researchers (House of Lords, 2005). Age-related cognitive decline is among the most feared and most burdensome aspects of growing old (Martin, 2004, Stern and Carstensen, 2000). Cognitive ageing is the harbinger of pathological states of cognitive decline, such as the dementias (Tierney et al., 2005). In order to provide rational bases for treatments to counter or ameliorate cognitive ageing, it is necessary to understand its psychological and biological foundations.

There is much descriptive data on the cognitive ability changes that occur with ageing (Hedden and Gabrieli, 2004, Salthouse and Ferrer-Caja, 2003, Schaie, 2005). Some cognitive capabilities are well retained in older people. Examples are vocabulary, some number skills, and general knowledge. Other cognitive functions show decline, similar in trajectory to physical functions, such as grip strength (Frederiksen et al., 2006). Examples are reasoning, some aspects of memory, spatial ability, executive functions, and speed of information processing. There are marked individual differences in human cognitive ageing (Schaie, 2005, Wilson et al., 2002). A large proportion of this variation appears to be caused by general cognitive decline, such that, when one domain of cognitive ability starts to decline, others also tend to deteriorate, though there is also additional age-related decline in specific cognitive abilities (Salthouse, 1996, Salthouse and Czaja, 2000, Wilson et al., 2002).

There are attempts within cognitive ageing to explain the behavioral phenomena using a smaller number of more “fundamental cognitive mechanisms or ‘primitives’” (Braver and West, 2008, p.311). In this mode, there are accounts which place working memory and, especially, executive control in explanatory positions regarding the ageing of disparate cognitive functions. Another influential hypothesis in cognitive ageing is that speed of information processing has a special place in understanding what happens in the older person's brain (Bugg et al., 2006, Charlton et al., in press, Finkel et al., 2005, Hertzog et al., 2003, Salthouse, 1996, Verhaeghen and Salthouse, 1997, Zimprich and Martin, 2002). In this view, processing speed might not merely be one domain of cognitive function that declines with age. It might be a partial foundation of other cognitive functions that decline; age-related changes in processing speed might account for some of the age-related variance in other mental functions. That is, if speed of information processing slows down, then perhaps other cognitive functions deteriorate too, by their being partly dependent on intact speed of processing for their efficiency. Speed of information processing can be assessed in humans by psychometric tests, such as the Wechsler battery's digit symbol and similar tests (Hoyer et al., 2004, Salthouse, 1996). However, these are complex, involving, probably, other high-level functions when people perform them (Deary, 2000, Chapter 8). Better evidence for the ageing of processing speed comes from conceptually simpler, lower-level tasks. These include reaction times that come from cognitive-experimental psychology, and the inspection time task that comes from psychophysics. Reaction times speed up during childhood, and are fastest in late adolescence/early adulthood. Thereafter, there is, on average, a marked non-linear slowing in reaction time with age, starting from early adulthood, which parallels age-related changes in physical functions (Der and Deary, 2006). There is statistical evidence that reaction times can account for a substantial proportion of the age-related variation in other cognitive domains, including memory (Salthouse, 1996). Slowing of reaction time is associated with an increased risk of mortality, as well as cognitive decline (Deary and Der, 2005, Shipley et al., 2006, Shipley et al., 2007). It is possible that the speeding up of reaction times during childhood and the slowing during middle and later adulthood are caused in part by the maturation and deterioration, respectively, of white matter tracts in the brain (Deary et al., 2006a). Despite this evidence for their importance in old age, a limitation of reaction times is that they involve speeded physical reactions, which could be slowed for other, non-cognitive reasons in older people, such as arthritis.

Here, to investigate further the biological foundations of human cognitive ageing, we employ a psychophysical task of speed of information processing that assesses the efficiency of iconic memory in the early stages of visual information processing. It is a visual backward masking task called inspection time (Deary, 2001). The task requires only a very simple visual discrimination, one that is almost error-free at longer stimulus exposure durations, such as the 150 ms and 200 ms durations included in the present study. The inspection time task, in its most usual format, is a two-alternative, forced choice procedure. Subjects are first cued that a stimulus is about to appear. In a typical version of the task, subjects are shown two parallel, vertical lines, joined at the top with a horizontal line (Fig. 1). There is a marked difference in length between the two lines. Subjects indicate, at leisure and with no requirement for a speeded response, whether the longer line was on the right or the left. The stimulus is presented at a number of durations that vary from a few hundred milliseconds (easy trials), to just a few milliseconds (difficult trials). The stimulus is backward-masked with a visual pattern mask (Fig. 1) after stimulus offset, to prevent further processing. The function of the stimulus duration versus the proportion of correct responses is described by a cumulative normal ogive (Deary et al., 1993).

There are individual differences in the efficiency of information processing assessed by the inspection time procedure. These are significantly correlated with individual differences in higher level mental abilities, including psychometric tests of intelligence (Grudnik and Kranzler, 2001). Subjects who make a higher proportion of correct discriminations in the inspection time task tend to score better on cognitive ability tests. This is the case for children (Edmonds et al., in press), healthy adults (Crawford et al., 1998), healthy older adults (Nettelbeck and Rabbitt, 1992), and cognitively impaired older adults (Bonney et al., 2006). There is a stronger association between inspection time and those tests of mental ability that decline with age than with those that don't, especially tests of processing speed (Burns and Nettelbeck, 2003, Crawford et al., 1998, Luciano et al., 2004). Inspection time mediates the association between age and the decline in other mental functions (Deary, 2000, p. 245; Nettelbeck and Rabbitt, 1992). Inspection time performance, and performance on other visual backward masking tasks, is markedly poorer in older adults with pathological cognitive decline, including mild cognitive impairment (Bonney et al., 2006, Lu et al., 2005) and dementia (Deary et al., 1991), and it is suggested that tests of this type of processing efficiency might be useful early detectors—‘biomarkers’—of age-related cognitive decline (Nettelbeck and Wilson, 2004, Nettelbeck and Wilson, 2005).

A principal reason that inspection time—and its correlations with higher cognitive ability tests—has attracted much attention is the simplicity of the task (Deary, 2000). As Nettelbeck (2001, p. 460) stated, “the measurement of IT is straightforward. The procedure requires a very simple judgment about which of two vertical lines, joined horizontally across the top, is longer (or shorter). If viewing time is not restricted, then no one makes any errors.” Thus, inspection time is simple in the senses that: the very easy discrimination is made almost always without error when the stimulus duration is sufficiently long; the stimulus–response characteristics are fixed; and the task can be done reliably even by children and people with dementia. The psychophysical task of inspection time is far simpler in conception and execution than the complex mental tests with which it correlates. Indeed, the task was designed by its originator Vickers, 1979, Vickers et al., 1972 to find the exposure duration that individual subjects required in order to resolve a simple visual judgment to a given level of accuracy. Of course, being cognitively simple would not rule out higher level brain areas being activated during the performance of the task.

A major missing link in the understanding of cognitive ageing is the lack of a biological foundation for the efficiency of information processing (Deary, 2000, Chapter 8; Finkel et al., 2005, Salthouse, 2000). Tasks such as inspection time offer candidate tools to close this explanatory gap, by acting as endophenotypes between higher cognitive functions and brain biology. Thus, impaired iconic memory has been suggested as a biomarker of accelerated ageing (Lu et al., 2005). Supporting this suggestion, inspection time is partly heritable, and has a sizeable genetic correlation with higher cognitive abilities in children and in adults (Edmonds et al., in press, Luciano et al., 2001, Luciano et al., 2004). The functional anatomy of inspection time in young adults has been described using block- and event-related fMRI study designs (Deary et al., 2001, Deary et al., 2004a). In the event-related fMRI study, subjects performed an inspection time task while fMRI took place. The BOLD activation–deactivation pattern was mapped in relation to a linear function of the inspection time stimulus duration. Task difficulty was associated with bilateral activation in the inferior fronto-opercular cortex, superior/medial frontal gyrus, and anterior cingulate gyrus, and bilateral deactivation in the posterior cingulate gyrus and precuneus. Therefore, in young adults, functional imaging has already provided information about brain areas which might be sufficient for task performance (Deary et al., 2004a).

In summary, there is good evidence that retaining efficient information processing is a key aspect of successful cognitive ageing. The inspection time technique has many findings that make it an attractive procedure—biomarker or endophenotype—for further exploring the neural foundations of successful cognitive ageing. The functional anatomy of efficient information processing with the inspection time task has been studied, to date, solely in younger people. Here, for the first time, we examine whether successful lifetime cognitive ageing, based on mental test scores taken over 50 years apart, is associated with the retention or loss of those functional brain networks that support efficient information processing in younger people (cf. Deary et al., 2004a). There are four theoretical models that attempt to account for how the average older brain operates when it tackles the same mental material as the average younger brain, viz. compensatory reorganization, dedifferentiation, computational capacity limitation, and neural inefficiency (Zarahn et al., 2007). The first two models make opposite predictions about the brain activation patterns of higher and lower mentally functioning older people when compared with younger people. Compensatory reorganization hypothesizes that the lower functioning older people will more closely resemble younger people, and dedifferentiation hypothesizes that it is the higher functioning older people who would have brain activation patterns more like the young (Cabeza et al., 2002, Cabeza et al., 2004, Zarahn et al., 2007). The present study provides evidence relevant to deciding between these hypotheses in the context of inspection time performance.

We examined individuals within a uniquely valuable cohort whose cognitive ability was assessed at age 11 and then again in their middle-to-late 60s: the Aberdeen Birth Cohort 1936 (e.g. Deary et al., 2004b, Whalley et al., 2005). Briefly, the design of the experiment was as follows. We compared two groups of non-demented 70-year-olds who, at age 11, had relatively similar general cognitive ability but who, in old age, had diverged, with one group demonstrating relatively successful, and the other one unsuccessful, cognitive ageing. We examined whether, in older people with relatively successful cognitive ageing, their BOLD activation–deactivation pattern while they performed an inspection time task was more or less similar to those of healthy younger individuals than older people with relatively unsuccessful cognitive ageing.