Explaining complex age-effects on prospective memory

The general aim of this thesis was to investigate complex age-related differences in prospective memory (PM). An extensive review of PM and cognitive ageing (an accepted peer reviewed article for Oxford Research Encyclopedias) was the starting point for identifying key areas of further research in the age-PM field. Specifically, two areas of further research were identified and then investigated empirically in the present research project. First, the need to use conceptually parallel PM task types across settings to illuminate mechanisms of the age-PM paradox was identified. The paradox refers to the general finding that young outperform older adults in the laboratory, but vice versa in naturalistic-settings; and young-old outperform old-old in laboratory but show equivalent levels of performance in naturalistic-setting studies. A second area identified as requiring further empirical investigation was the role of executive functions in mediating age-effects across a wide range of PM tasks. The current research project made a significant contribution to the age-PM literature by undertaking a series of empirical investigations into both of these areas.

The first key empirical investigation is related to illuminating the mechanisms of the age-PM paradox. A paper reporting two empirical experiments is presented (which have been submitted to the Journal of Experimental Psychology: General). In Experiment 1, young (19–30 years; n = 40) and older adults (65–86 years; n = 53) were tested on conceptually parallel PM tasks in the laboratory (using the Virtual Week paradigm that simulates naturalistic PM tasks and routine daily activities) and in a naturalistic-setting using a recently developed novel smartphone paradigm, the MEMO. PM tasks were conceptually parallel in terms of the type of cue and inherent level of environmental support afforded by three PM task types: event-based, time-of-day, and time-interval. The latter two time-based tasks were hypothesised to largely account for the age-PM paradox, in particular by not being sufficiently distinguished and investigated separately in previous studies of the age-PM association in performance across settings. In Experiment 1, participants completed two simulated days of Virtual Week with the three PM tasks types embedded, and two separate blocks of three days for the event-based PM (a photo task when particular events were encountered) and two time-based PM tasks (scheduled and “pop up” quizzes, e.g., come back in 10 minutes to open app on phone to complete a quiz). It was found that young adults outperformed older adults in the laboratory, however, in the naturalistic-setting older adults only outperformed younger adults on time-of-day tasks (i.e., appointment like tasks; in this case completing a scheduled quiz), while similar levels of performance were revealed on event-based tasks (relatively high performance) and time-interval tasks (relatively low performance). In Experiment 2 young-old (60–74 years; n = 64) and old-old (75–87 years; n = 40) adults were compared on the same two PM task types. However, this time the naturalistic-setting paradigm, MEMO, was made more challenging, and conceptually closer to Virtual Week, by combining all PM task types over the same three day period. Participants were also permitted to use external aids. The results showed that young-old outperformed old-old in the laboratory, with both age groups showing better performance on event-based than the time-interval tasks (involving monitoring a stop clock in Virtual Week). Together, these experiments show that the age-PM paradox is only apparent when comparing different types of time-based tasks across settings, and that older adults are vulnerable to forgetting delayed intentions over short intervals with relatively few time cues in a naturalistic-setting. However, when permitted to use external aids older adults can compensate for this cognitive vulnerability, and show similar high levels of performance to their own performance on event-based and time-of-day tasks.

The second key empirical investigation is related to the generally hypothesised mediating role of facets of executive functioning on age-effects on PM, particularly for putatively high demand tasks. In a very novel study in the age-PM field, for the first time four different PM measures were combined in a single study to investigate individual differences in facets of executive functioning mediating age-effects on PM. The study used a large sample of older adults (n = 104; range 60–87 years) who performed the laboratory paradigm Virtual Week, two clinical measures, the Cambridge Prospective Memory Test (CAMPROMT), and the Memory for Intentions Test (MIST), and the MEMO. The results showed that there were age-effects for the PM measures that presumably had the highest cognitive demands in so far as they do not permit external aids, that is Virtual Week and the MIST (external aids are allowed in the CAMPROMT and in the present study’s use of the MEMO paradigm). Contrary to previous studies using abstract PM paradigms in the laboratory, older adults’ performance on these relatively naturalistic (even familiar) PM task types showed some relationship to retrospective memory processes, processing speed, and age, but virtually no relationship to separable facets of executive functioning. The models tested included both parallel mediation models, in which both executive and non-executive cognitive processes were measured, and a serial mediation model with processing speed hypothesised to impact executive processes which in turn influence PM task performance. The lack of evidence for both these models suggest that PM is a relatively independent functional construct that is affected by age in some circumstances, regardless of individual differences in executive functioning, and is also sometimes independent of both age and executive functioning processes. The present thesis thus indicates potential boundary conditions for both the manifestation of the age-PM paradox and the mediating role of executive functions on age-related changes in PM.