Abstract

Working Memory refers to cognitive functions that support the ready availability of a small amount of information temporarily, while we undertake ongoing actions and mental activities, and is viewed as a core mechanism underpinning higher-order cognitive abilities. Moreover, the functioning of WM abilities is important for autonomy and wellbeing in older adults. As assessed, WM suffers pronounced, linear decline during adult ageing. However, merely establishing that younger adults outperform older adults on a given cognitive task is of limited value given that it is uninformative regarding how and why WM declines with age. This thesis was inspired by research that has suggested that some aspects of WM decline faster than others. Indeed, verbal WM appears least susceptible, and visuospatial WM most susceptible to age-related decline. In six experiments, I moved beyond the ‘Dull Hypothesis’ and tested whether older adults approached WM tasks differently from younger adults, perhaps relying on relatively intact verbal abilities while performing visual memory tasks. Crucially, visual material – in everyday life as well as in memory experiments – may be remembered via verbal codes or visual traces, or both. In some WM theories, visual and verbal material is seen as maintained by separate mechanisms. The Multiple-Component model of WM is based on the postulate that visuospatial and verbal information is stored separately in dedicated storage buffers, which may also rely on separate rehearsal mechanisms. If one mechanism declines more with age, perhaps older adults strategically recruit a different mechanism. This led to our central research question: Do we approach cognitive tasks differently as we age? I investigated this in several WM paradigms, as outlined below. The first series of experiments addressed the debate about whether older adults have a specific deficit in the ability to bind and remember conjunctions of features, by investigating the consequences of allowing verbal rehearsal in visual feature-binding tasks. In experiments 1 to 4, I studied the role of verbal labels in two different feature-binding paradigms to test whether discrepancies in the literature can be explained by verbal rehearsal of visual features, which might vary by age group. I found that overall, visual memory for difficult-to-label, noncategorical, visual information appeared especially limited for older adults, likely because it impedes engagement of other systems, such as verbal WM or longterm memory. Results regarding the potential implications for discrepancies in the feature-binding literature were mixed. Next, I looked at the effect of instructing or preventing verbal labelling in a continuous colour memory paradigm, using a mixture model which allowed comparison of continuous and categorical memory representations. Labelling improved memory performance in both age groups, but older adults appeared to spontaneously rehearse verbal labels sub-vocally more than younger adults when simply instructed to perform the task in silence. Finally, I investigated the role of strategic approaches in an N-back WM training paradigm. In this study, I instructed participants to use a visualisation strategy previously found to improve N-back performance in younger adults. I found that both younger and older adults benefitted from the instructed strategy and performed better than uninstructed controls, but some evidence suggested that the strategy was more difficult to implement for older adults. I also found significant associations between N-back performance and the type, and level of detail, of self-generated strategies in the uninstructed participants. Combined, the results suggested that measures of performance and capacity partly reflect the extents to which participants apply appropriate strategies. Strategic mediation should be considered in research aiming to understand memory for visual features, continuous colour memory, and the mechanisms of WM training. Our results highlighted that strategic differences between younger and older adults may be crucial to interpret the age-related decline of memory, as measured in these paradigms, thus illustrating the importance of controlling differences in age-related strategic preferences in visual memory tasks. These differences may be informative for our understanding of age-related cognitive decline, suggesting that older adults may compensate for decline of some functions by approaching tasks differently.