Prospective Memory and Intention Deactivation: Challenges, Mechanisms and Modulators

From the simple act of picking up a glass of water while talking to someone at a party, to remembering to swing by the bike shop to pick up an inner tube while riding through traffic on our way home from the office, intentions guide and alter our behavior—often while we are busily engaged in other ongoing tasks. Particularly, performing delayed intentions, like stopping at the bike shop on our way home, relies on a set of cognitive processes summarized as prospective memory (PM) that enable us to postpone intended actions until a later point in time (time-based PM) or until specific reminders or PM cues signal the appropriate opportunity to retrieve and perform an intended action (event-based PM). Interestingly, over the past decades a growing number of studies showed that successfully completing an event-based intention does not necessarily lead to its immediate deactivation. Instead, no-longer-relevant PM cues can incur so-called aftereffects that impair task performance and sometimes even trigger erroneous repetitions of the intended action (i.e., commission errors). Although in our everyday lifes we frequently rely on both PM and intention deactivation, still relatively little is known about how our cognitive system actually manages to deactivate completed intentions, under which conditions this may fail, and how well PM and intention deactivation function under extreme conditions, like acute stress. In order to answer these questions, I first conducted a comprehensive review of the published literature on aftereffects of completed intentions. Here, I found that although intentions can incur aftereffects in terms of commission errors and performance costs that most likely result from continued intention retrieval, they generally seem to be deactivated or even inhibited at some point. Most importantly, this deactivation process does not operate like a light switch but dynamically moves along a continuum from complete reactivation to complete deactivation of intentions, and is substantially modulated by factors that also affect retrieval of intentions prior to their completion. Specifically, intention deactivation is most likely to fail when we remain within the same context in which we originally completed the intention and encounter no-longer-relevant PM cues that are extremely salient and were strongly linked to the intended action. Subsequently, in Study 1 I directly tested a dual-mechanisms account of aftereffects of completed intentions. Building on findings of impaired intention deactivation in older adults who often show deficits in cognitive-control abilities, this account posits that aftereffects and commission errors in particular stem from a failure to exert cognitive control when no-longer-relevant PM cues trigger retrieval of an intention. Accordingly, intention deactivation should hinge on the availability of cognitive-control resources at the moment we encounter no-longer-relevant PM cues. In order to test this, I assessed aftereffects of completed intentions in younger and older adults while manipulating transient demands on information processing during encounters of no-longer-relevant PM cues on a trial-by-trial basis. In Experiment 1, nominally more older adults than younger adults made a commission error. Additionally, medium demands on cognitive control substantially reduced aftereffects compared to low and high demands (i.e., u-shaped relation). In Experiment 2, which extended this manipulation but only tested younger adults, however, this control-demand effect did not replicate. Instead, aftereffects occurred regardless of cognitive-control demands. The lack of a consistent control-demand effect on aftereffects across two experiments, suggested that cognitive control either only plays a minor role for the occurrence of aftereffects or that, more likely, intention deactivation hinges on other specific cognitive-control abilities, like response inhibition. In two subsequent studies, I extended this research and tested the effects of acute stress—a potent modulator of cognitive-control functioning—on PM and intention deactivation. Previous studies showed that, under moderate demands, acute stress had no effect on PM-cue detection, intention deactivation or performance costs that presumably arise from monitoring for PM cues. Importantly, however, based on these studies it remained unclear if acute stress affects PM and intention deactivation under high demands, as has been observed, for instance, with working-memory performance. To test such a potential demand-dependence of acute stress effects on PM, I first assessed the effects of psychosocial stress induction with the Trier Social Stress Test on PM and intention deactivation when detecting PM cues and intention deactivation were either low or high demanding (Study 2). Building on this work, I then tested the effects of combined physiological and psychosocial stress induction with the Maastricht Acute Stress Test on PM and the ability to track one’s own performance (i.e., output monitoring), when PM-cue detection was difficult and ongoing tasks additionally posed either low or high demands on working memory (Study 3). Despite successful stress induction (e.g., increased levels of salivary cortisol and impaired subjective mood), both studies showed that PM-cue detection and intention retrieval were not affected by acute stress under any of these conditions. Study 2 revealed a tendency for a higher risk of making commission errors under stress when no-longer-relevant PM cues were salient and difficult to ignore. Study 3 additionally showed that acute stress had no effect on output monitoring. Most importantly, however, across the different PM tasks and stress-induction protocols in these studies, acute stress substantially reduced performance costs from monitoring for PM cues, but did so only when PM-cue detection was difficult. This effect suggested that, depending on task demands, acute stress might shift retrieval processes in PM away from costly monitoring-based retrieval towards a more economic spontaneous retrieval of intended actions. In summary, the present thesis suggests that the processes underlying prospective remembering and intention deactivation are tightly woven together and are only selectively affected by cognitive-control availability and effects of acute stress. With this, it contributed substantially to our understanding of these essential cognitive capacities and their reliability. My research showed that PM is remarkably resilient against effects of acute stress experiences when remembering intended actions is supported by external reminders. Acute stress may actually make monitoring for such reminders more efficient when they are hard to detect. Additionally, it showed that, in most circumstances, we seem to be able to successfully and quickly deactivate intentions once they are completed. It is only under some conditions that intention deactivation may be slow, sporadic or fail, which can lead to continued retrieval of completed intentions. While this seems not to be affected by transient demands on information processing during encounters of no-longer-relevant PM cues, intention deactivation might become difficult for older adults and stressed individuals when no-longer-relevant reminders of intentions easily trigger the associated action and are hard to ignore