Prospective Memory

In 1990, Einstein and McDaniel introduced a laboratory paradigm for investigating prospective memory (remembering to perform an action in the future), in which participants are engaged in an ongoing activity, such as a short-term memory task. The prospective memory task is then embedded in the ongoing task: Einstein and McDaniel requested that participants press a particular key on the computer keyboard whenever the target word “rake” appeared during the ongoing task. The Einstein and McDaniel paper foreshadowed several issues that continue to intrigue researchers in this area. For instance, Einstein and McDaniel suggested that prospective memory tasks vary in the degree of self-initiation required, which echoes their influential multiprocess view (McDaniel & Einstein, 2000) and its proposal that some tasks rely on automatic spontaneous retrieval of the delayed intention.

While Einstein and McDaniel (1990) employed an ongoing task in order to mimic aspects of nonlaboratory tasks, performance on the ongoing task eventually became the focus of research aimed at investigating the resource demands of the prospective memory task (e.g., Smith, 2003). Specifically, ongoing task performance was compared for participants who had the prospective memory intention and participants who perform the ongoing task without the prospective memory intention. A number of studies found that performance on the ongoing task was impaired in the prospective memory condition, sometimes called the cost to the ongoing task (for review, see Smith, Hunt, McVay, & McConnell, 2007). Einstein et al. (2005) proposed that under certain circumstances, there should be no cost to the ongoing task. Smith et al. conducted experiments in which these restrictions were met and continued to find a cost. In a commentary on Smith et al.’s article, Einstein and McDaniel (2010) suggested that spontaneous retrieval is not automatic. As noted in Smith’s (2010) reply, if spontaneous retrieval is not equated with automatic retrieval, then this makes a simple demonstration of cost or no cost far less interesting. Furthermore, a cost is often found and Smith (2010) suggested that the focus should be shifted to understanding at a deeper level the processes that contribute to the cost and how these processes will be affected by different factors and in different groups of participants. Doing so moves us beyond the cost debate.

As illustrated by the articles presented in this special issue, new paradigms and richer analytic techniques are pressing the field forward. Instead of taking an either/or approach to investigating cost, researchers are exploring ways to learn more from their analysis of the ongoing task. For instance, an alternative approach to analyzing ongoing task response time data is suggested by Brewer (2011). Specifically, Brewer illustrates the utility of the Ex-Gaussian approach to describing the distribution of response times as this provides a richer and more complete picture of how the prospective memory task affects the ongoing task. Similarly, Voigt, Aberle, Schönfeld, and Kliegel (2011) show how taking a more fine-grained approach to analyzing monitoring in a time-based prospective memory task provides information about the nature of development of monitoring in a time-based prospective memory task that would be missed when examining monitoring in less detail. Smith, Persyn, and Butler (2011) use a formal model that simultaneously considers accuracy on both the ongoing task and the prospective memory task within a single analysis. In doing so, the multinomial process tree modeling approach provides information concerning the effects of personality and working memory that was not available through a separate analysis of the ongoing task.

A number of articles in this special issue present new approaches for addressing another point raised by Einstein and McDaniel (1990): The need for ecologically valid tasks that still allow for the control afforded by a laboratory setting. This is perhaps best illustrated in Loft, Pearcy, and Remington’s (2011) investigation of prospective memory using an air traffic control (ATC) simulation. The ATC task presents a complex environment in which participants must process multiple stimuli simultaneously and make a variety of ongoing decisions. While Loft’s ATC simulation is designed to address a particular real-world task environment, the complexity of the task mimics many of our daily environmental demands (e.g., driving in traffic) and indicates that more complex laboratory tasks can provide a rich source of data for understanding prospective memory in applied settings.

Similarly, the relatively complex virtual week task was used by Margrett, Reese-Melancon, and Rendell (2011). The virtual week task is an engaging task for the participants and although the task does not afford the level of detailed ongoing task performance measures seen in the ATC task (e.g., response time and accuracy on a variety of tasks), it does facilitate the investigation of an important and largely neglected aspect of prospective memory: the social aspect. Moreover, the particular approach adopted by Margrett and her colleagues provides new insights into how couples influence one another’s prospective memory performance and points the way for future research.

As noted in the articles in this issue, prospective memory is important for functioning successfully in our everyday lives. The inability to carry out delayed intentions can have a negative impact on health and safety, success in school and work settings, and on our social relationships. Given this, it is perhaps not surprising that Einstein and McDaniel (2010) argue for spontaneous retrieval on the basis that it is more adaptive. While rats and pigeons have shown evidence for prospective coding, which could play a role in successful prospective memory, Klein, Evans, and Beran (2011) failed to find clear evidence that the rhesus and capuchin monkeys in their study rely on prospective coding. As noted by Klein et al. the animal tasks for prospective coding do not directly map onto human prospective memory performance; however, the Klein et al. study demonstrates that the investigation of prospective memory from an evolutionary perspective is likely to produce a complicated story.

In summary, some of the same issues raised by Einstein and McDaniel (1990) continue to intrigue prospective memory researchers two decades later. At the same time, the articles collected for this special issue illustrate how far the area has expanded in both the breadth and depth of the questions addressed and in the increasingly complex methodologies used by researchers in this area.