Craig P. McFarland a & Elizabeth L. Glisky a a Department of Psychology, University of Arizona, Tucson, AZ, USA. Published online: 28 Oct 2011.

Transcription

1 This article was downloaded by: [University of Montana] On: 20 November 2013, At: 09:32 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: Registered office: Mortimer House, Mortimer Street, London W1T 3JH, UK Aging, Neuropsychology, and Cognition: A Journal on Normal and Dysfunctional Development Publication details, including instructions for authors and subscription information: Implementation intentions and prospective memory among older adults: An investigation of the role of frontal lobe function Craig P. McFarland a & Elizabeth L. Glisky a a Department of Psychology, University of Arizona, Tucson, AZ, USA Published online: 28 Oct To cite this article: Craig P. McFarland & Elizabeth L. Glisky (2011) Implementation intentions and prospective memory among older adults: An investigation of the role of frontal lobe function, Aging, Neuropsychology, and Cognition: A Journal on Normal and Dysfunctional Development, 18:6, , DOI: / To link to this article: PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content ) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &

2 Conditions of access and use can be found at

3 Aging, Neuropsychology, and Cognition, 2011, 18 (6), ISSN: print; online Implementation intentions and prospective memory among older adults: An investigation of the role of frontal lobe function Craig P. McFarland and Elizabeth L. Glisky Department of Psychology, University of Arizona, Tucson, AZ, USA ABSTRACT Prospective memory (PM) among older adults has been shown to be influenced by frontal lobe (FL) function. An implementation intention (e.g., if situation X occurs, I will do Y ) is a mnemonic strategy that may be particularly beneficial for individuals with low-fl function, as it has been suggested that implementation intentions produce heightened accessibility to environmental cues, and automatic triggering of previously formed intentions. The present study investigated the effectiveness of implementation intentions among 32 older adults characterized as possessing high- or low-fl function. Participants were placed into one of two conditions: Read-Only or Implementation Intentions, before being tested on a laboratory prospective memory task. Results indicated that older adults with high-fl composite scores demonstrated better PM than those with low-fl scores, and that those who made implementation intentions outperformed those who simply read task instructions. Of particular interest is the finding that high-fl participants benefited from implementation intentions, suggesting that implementation intentions may improve PM of all older adults regardless of FL function. Theoretical underpinnings of implementation intentions are discussed in the context of FL function. Keywords: Aging; Prospective memory; Frontal lobe function; Implementation intentions; Executive function. This research was supported by National Institute on Aging Grant AG14792 as well as by a grant from the Social and Behavioral Sciences Research Institute at the University of Arizona. We would like to thank Drs John Allen, Al Kaszniak, and Lee Ryan for their insight and assistance with this project. We would also like to thank Alex Ayala for testing participants. Address correspondence to: Craig P. McFarland, Department of Psychology, University of Arizona, 1503 East University Blvd, Tucson, AZ 85721, USA. cpm3@ .arizona.edu 2011 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business

4 634 CRAIG P. MCFARLAND AND ELIZABETH L. GLISKY The necessity of remembering to perform actions in the future is a burden known to all. From remembering to tell a friend about an upcoming social event, to remembering to take one s medication, life is filled with intentions. Despite being highly motivated to get in better shape or to achieve more in the workplace, however, many intentions go unfulfilled. In the interest of reducing such failures, Gollwitzer (1993, 1996, 1999) devised a strategy referred to as implementation intentions. Gollwitzer contends that by framing one s goals of action or behavior in the context of an if, then statement, one can reduce failures of execution of previously formed intentions. To that end, implementation intentions rely upon the development of just such a relation. Implementation intentions are essentially a means of planning for the successful execution of actions to be performed at some point in the future, and consist of two key, inter-related components. The first component specifies the situation in which an action is to be executed (e.g., if situation X occurs... ), while the second component specifies the desired action that is to be completed in the specified future situation (e.g., if situation X occurs, I will perform Y ) (Gollwitzer, 1993, 1996, 1999). The study of the execution of intentions has recently begun to receive considerable research attention, particularly with regard to the cognitive operations that may be required. Prospective memory (PM) is a term that is used to describe the memorial processes involved in remembering to perform a specific action at a future point. In investigations of PM, participants are engaged in some ongoing activity, and may have to search or monitor the environment for the presence of a cue, in the case of event-based PM, or for the arrival of a particular time, in the case of time-based PM. Upon recognizing the appropriate cue, people must then interrupt their performance of the ongoing task in order to successfully complete the intended action. The nature of PM tasks, including the need to recognize a cue, as well as the need to remember the intention associated with a particular cue, makes them uniquely suited for the use of implementation intentions. Several studies have provided support for the effectiveness of implementation intentions in improving the execution of goal behaviors, or PM. Gollwitzer and Brandstätter (1997), for example, found that compared to students who simply made an intention to write a paper during Christmas break, students who were instructed to form an implementation intention that included both the specific time and place at which they would write the paper were more likely to do so. Similarly, Kardiasmenos, Clawson, Wilken, and Wallin (2008) reported positive effects of implementation intentions among adults with multiple sclerosis. The PM task in their study involved the execution of a variety of tasks presented in the context of a board game, in which half of the cues and tasks were strongly associated (e.g., When you look for pictures of your dad, write a letter to your mom ), and the other

5 IMPLEMENTATION INTENTIONS 635 half non-associated (e.g., When you go to buy more paint, buy more pens ). Findings indicated that although PM performance on highly associated tasks did not differ between an implementation intention group and a rote rehearsal group, implementation intentions did provide for significant improvement on non-associated PM tasks. According to Gollwitzer (1993, 1996, 1999), implementation intentions affect the execution of previously formed intentions in two ways. First, the act of specifying a situation in which an intended action is to be executed results in heightened accessibility to, or apprehension of environmental cues. Second, the establishment of a link between a desired action and a specified situation may result in automatic triggering of the intention in the presence of the appropriate cues. More specifically, it is thought that implementation intentions reduce the need for cognitively demanding processes, such as maintaining an intention over a delay, monitoring the environment for an appropriate time to execute a task, or retrieving an intention once a cue is encountered, by establishing heightened accessibility to environmental cues, and providing for automatic, or reflexive responding to those cues, resulting in the successful execution of intentions (Cohen & Gollwitzer, 2008; Gollwitzer, 1993, 1996, 1999; McDaniel, Howard, & Butler, 2008). A second critical component of implementation intentions appears to be the linking of a cue with an intended action. (Although standard PM instructions also link a cue with an intended action, implementation intentions require that participants develop an associative link between cue and intention by focusing on the relation between the two at encoding, whereas standard instructions provide less explicit instructions for how to encode the relation between cue and intention.) If implementation intentions do in fact provide for automatic responding in the presence of cues, then the successful execution of PM tasks should exact no cost on the performance of an ongoing task, contrary to what has been seen in PM tasks performed without implementation intentions (Smith, 2003; Smith, Hunt, McVey, & McConnell, 2007). Recent research provides support for this idea. For example, in a study conducted by Cohen and Gollwitzer (2008), participants were engaged in a lexical decision task (ongoing activity), and were instructed to say the second word of a previously studied word pair if the first word of that pair appeared during the lexical decision task (PM task). The implementation intention group not only completed more PM tasks successfully, but also exhibited less cost of the PM task on the ongoing activity. That is, response times in the lexical decision task were not slowed by the concurrent PM task when implementation intentions were formed, although they were under the standard condition. These results suggest that implementation intentions may provide for automatic retrieval of an intended action in the presence of an associated cue.

6 636 CRAIG P. MCFARLAND AND ELIZABETH L. GLISKY McDaniel et al. (2008) addressed the question of automaticity in implementation intentions by manipulating the attentional demands of the ongoing activity. While engaged in a word rating task, participants were periodically required to generate random number sequences, thus providing for alternating conditions of divided and full attention within an ongoing task. The PM task involved making a key press when particular words (e.g., spaghetti or doll ) were presented for rating. With regard to the ongoing word rating task, no performance differences emerged among an implementation intention group, a PM group (without implementation intentions), and a no PM control group, indicating that neither the implementation intention group, nor the PM group diverted attention away from the ongoing task in order to complete the PM task. Of particular interest, however, was that when attention was divided, PM task performance suffered, but only for the groups that did not form implementation intentions. These results provide further support for the notion that implementation intentions foster automatic responding in the presence of relevant cues. The authors suggest that implementation intentions facilitate automatic responding as a result of the associative link between cue and intention established at encoding. In particular, implementation intentions force the encoding of a cue-intention association. In contrast, standard PM instructions are not so explicit, which may result in some participants developing a strong association, and others not. A possible inference that may be made from these findings is that individuals with reduced processing resources should benefit from the use of implementation intentions. Evidence exists that older adults have reduced processing resources and often perform similarly on cognitive tasks to young adults under divided attention (Anderson, Craik, & Naveh-Benjamin, 1998; Craik, 2005; Craik & McDowd, 1987). Craik has suggested more specifically that older adults are impaired in their use of self-initiated processing. Self-initiated processing (Craik, 1986) is any subject-directed operation, including the active monitoring of the environment, recognition of a cue, the retrieval of the intended action associated with a cue, and the execution of the intended action processes often required in tasks of PM. Evidence from work with frontal lobe patients suggests that self-initiated processing may rely upon the frontal lobes (Craik, Luo, & Sakuta, 2010; Fuster, 1997; Stuss & Benson, 1984). Such a relation has direct implications for the performance of older adults on PM tasks, as evidence suggests an age-related decline in frontal lobe function (Fuster, 1997; Raz, 2000; West, 1996), support for which has come from a variety of sources (Haug & Eggers, 1991; Raz et al., 2005; Van Petten et al., 2004). Consistent with this line of reasoning, two studies have found that older adults do in fact benefit from the use of implementation intentions. Chasteen, Park, and Schwarz (2001) reported an experiment in which participants were required to write the date on the top of several pieces of paper as they were

7 IMPLEMENTATION INTENTIONS 637 presented during the course of an experiment. Older adults who formed an implementation intention successfully performed the PM task more often than did older adults who did not form an implementation intention. However, implementation intentions did not improve performance on a second PM task, in which the PM cue was integrated with the ongoing task, arguably because there was no requirement to search the environment for the relevant cue. The authors interpreted this pattern of results as evidence that implementation intentions enhance PM performance, but only under conditions that require substantial self-initiation. Furthermore, these findings were presented as supportive of the idea that implementation intentions facilitate automatic responding to the cue, thereby circumventing any task requirements for effortful processing, which has been shown to be more age-sensitive than automatic processing (Park, 1999). Liu and Park (2004) came to a similar conclusion when they found that compared to older adults who simply rehearsed instructions to test their blood glucose levels over a period of 3 weeks, older adults who were instructed to form implementation intentions for testing their blood remembered to do so more often. As did Chasteen et al. (2001), Liu and Park (2004) interpreted their results as evidence that implementation intentions impart an automatic quality to the processing of, and responding to, environmental cues, thereby supporting successful PM completion among a group of older adults. Although these studies demonstrate that older adults can benefit from implementation intentions, the mechanism by which implementation intentions have their effects in this population is still unclear. It is possible that by creating a strong link between environmental cues and intended actions, implementation intentions provide for automatic responding in the presence of such cues (Gollwitzer, 1993, 1996, 1999). If implementation intentions have their effect by increasing the automaticity with which environmental cues are responded to, thereby decreasing the need for effortful, self-initiated processing, then older individuals with reduced frontal lobe function should benefit from their employment. However, any such argument with regard to these studies is entirely speculative, as few neuropsychological tests, particularly indices thought to tap frontal lobe function, have been included in existing research. Therefore, what underlies the benefit seen among older adults remains an empirical question. One method of exploring the nature of the effect of implementation intentions in older adults would be to include a measure by which older adults could be classified as possessing either high- or low-frontal lobe (FL) function. If the use of implementation intentions results in the automatic processing and recognition of environmental stimuli, as well as the automatic retrieval of previously formed intentions upon the recognition of associated environmental cues, then one might expect to see enhanced PM performance, especially among older adults possessing low-fl function. Previous studies

8 638 CRAIG P. MCFARLAND AND ELIZABETH L. GLISKY with older adults have suggested that the frontal lobes are involved in PM tasks (McDaniel, Glisky, Rubin, Guynn, & Routhieaux, 1999; McFarland & Glisky, 2009) in a variety of component processes, including maintaining an intention over a delay, dividing attention, or monitoring the environment for a cue (Fuster, 1997; Stuss & Benson, 1984). Thus, any increase in the automaticity of these processes should benefit older adults with low-fl function. Older adults having high-fl function could presumably rely upon more efficient self-initiated processing, which may enable them to create a stronger link between cue and intended action, resulting in more successful PM performance. However, findings of improved PM among young adults (McDaniel et al., 2008) suggest that all older adults should benefit from the use of implementation intentions, regardless of frontal function. If however, implementation intentions selectively benefit older adults with low-fl function, such a finding might speak more directly to the question of automaticity than did either of the previous studies that investigated implementation intentions among older adults. So, although two studies have demonstrated positive effects of implementation intentions among older adults, the nature of the effect of implementation intentions and how it relates to FL function in this population remains uncertain. The current study was designed to investigate the relation between FL function, as characterized by performance on neuropsychological tests, and implementation intentions in a group of older adults, in a task of event-based PM. It was hypothesized that overall PM performance would be greater for the high-fl group than for the low-fl group, replicating previous findings (McDaniel et al., 1999; McFarland & Glisky, 2009). With regard to the effect of implementation intentions on PM performance, it was hypothesized that both low- and high-fl groups would benefit from their use. Despite the fact that high-fl participants may be more likely to spontaneously create a strong link between a cue and its related intention under standard instructions, they too could be expected to benefit from the more explicit direction to link cue and intention provided in implementation intention instructions. The magnitude of improved performance from standard to implementation intention instructions may, however, be greater for the low-fl group, as they are less likely to generate a link between cue and intention under standard instruction than is the high-fl group. To test these hypotheses, a measure of frontal function developed by Glisky, Polster, and Routhieaux (1995; for recent update see Glisky & Kong, 2008) was employed, in which older adults were categorized as possessing high- or low-frontal function based on their performance on a battery of neuropsychological tests. This measure has been shown to predict performance on tasks of source memory (Glisky et al., 1995; Glisky, Rubin, & Davidson, 2001; Glisky & Kong, 2008) and PM (McDaniel et al., 1999; McFarland & Glisky, 2009).

9 IMPLEMENTATION INTENTIONS 639 METHOD Participants A total of 32 older adults (age 65 or older) were selected from the subject pool of the Aging and Cognition Unit at the University of Arizona. All older participants were healthy, community-dwelling adults without dementia, and were depression-free and had no history of drug or alcohol abuse. Participants were paid $8 per hour to take part in this study. All older adults had been previously categorized along two dimensions based on their performance on a battery of neuropsychological exams conducted within the past year. Initial exploratory factor analyses of neuropsychological test performance of older adults (Glisky et al., 1995, 2001) and recent confirmatory factor analysis with a new sample of 227 older adults (see Glisky & Kong, 2008) had revealed two separate neuropsychological components. The first factor is hypothesized to tap executive control processes that have been shown to be dependent on the frontal lobes; it will be referred to as the FL factor. Tests loading on the FL factor include the number of categories achieved on the modified Wisconsin Card Sorting Test (Hart, Kwentus, Wade, & Taylor, 1988), Mental Arithmetic from the Wechsler Adult Intelligence Scale Revised (WAIS-R; Wechsler, 1981), Mental Control from the Wechsler Memory Scale III (WMS-III; Wechsler, 1997), Backward Digit Span from the WMS-III, and the total number of words generated on the Controlled Oral Word Association Test (Benton & Hamsher, 1976). Neuropsychological tests loading on the second factor include Logical Memory I, Verbal Paired Associates I, Faces I (all from the WMS III), Visual Paired Associates II (from the WMS-R; Wechsler, 1987), and Long- Delay Cued Recall from the California Verbal Learning Test (Delis, Kramer, Kaplan, & Ober, 1987). The second factor taps basic memory processes that have been shown to be dependent on the medial temporal lobes; it will be referred to as the MTL factor. The composite scores for both factors represent average z-scores relative to the normative population of 227 older adults. Variability attributable to age has been removed from these scores. Half of the older adults had scores on the FL factor that were above the mean of the normative group (high-fl) and half had scores that were below the mean of the normative group (low-fl). Among both high- and low-fl groups, 8 participants received standard read-only instructions, and 8 received implementation intention instructions. MTL factor scores were equivalent across high- and low-fl groups, serving to equate retrospective memory between FL groups. Characteristics of each group are presented in Table 1. There were no differences in age, education, or score on the Mini-Mental Status Examination (MMSE; Folstein, Folstein, & McHugh, 1975) as a function of neuropsychological status (all Fs < 1), or instructional condition (all Fs < 1), as revealed by separate one-way analyses of variance (ANOVA).

10 640 CRAIG P. MCFARLAND AND ELIZABETH L. GLISKY TABLE 1. Characteristics of older adults as a function of neuropsychological group FL function High-FL Low-FL Variable M SD M SD Age (years) Education (years) MMSE FL score a MTL score a Note: FL, frontal lobe; MMSE, Mini-Mental Status Examination; MTL, medial temporal lobe. a z-score (see text). Materials The ongoing task in this experiment consisted of a multiple choice test of general knowledge and trivia. One hundred ninety six questions were selected from the McDaniel et al. (1999) study and were presented on a Gateway E-6610D PC, using DMDX (Forster & Forster, 2003). Questions appeared in the center of the monitor. Beneath each question were four response options corresponding to the multiple choice answers. Participants indicated their answer choice by pressing one of four keys on the computer keyboard (labeled A, B, C, or D), and were allowed a maximum of 12 seconds to respond to each question. Successive questions appeared at 12-second intervals regardless of speed of responding. Questions were presented in a different random order for each participant. Included among the multiple choice questions were eight target questions about states in which the word state appeared. Those questions appeared at approximately 5-minute intervals throughout the course of the experiment. This task is identical to one in which we previously found impaired performance in low-frontal older adults under standard conditions (McDaniel et al., 1999). Procedure The experiment lasted approximately one hour. After providing informed consent, participants were given a sheet of paper detailing task instructions and were asked to read it carefully. The instructions indicated that as a participant in the study, they would be asked to complete a multiplechoice test of general knowledge and trivia, and that they would have a total of 12 seconds to respond to each question. They were also told that they would

11 IMPLEMENTATION INTENTIONS 641 need to complete a secondary task, and at that time were instructed to direct their attention to the computer monitor, on which additional instructions were presented. At that point the instructions that followed differed, depending on whether a participant was in the Read-Only or Implementation Intention condition. In the Read-Only condition, participants read silently from the computer screen PM instructions consisting of If you see questions pertaining to states, you should press the 6 key. They were then asked to write that exact instruction one time on a sheet of paper provided and to repeat it once aloud to the experimenter. In the implementation intentions condition, participants silently read the read-only instructions from the computer screen, followed by a more specific implementation intention ( When I see the word state, I will press the 6 key ) which they also read silently from the computer. They were then asked to write the exact implementation intention on a sheet of paper and to repeat it once aloud to the experimenter. Following the presentation of instructions, participants completed a series of five practice questions so that they could become familiar with the type of questions they would encounter during the experiment. Included among the practice questions was one question containing the word state. In instances in which a participant failed to press the 6 key in response to that practice question, the participant was informed of their omission, and the practice questions were presented again, until the participant correctly pressed the 6 key when presented with a question containing the word state. After completion of the practice trials, no further mention of the PM task was made. The multiple choice trivia questions, with the embedded PM task, then proceeded for 42 minutes, after which participants were asked to describe the tasks involved (i.e., answering multiple choice questions; pressing button 6 when seeing the word state ). At that time, participants completed a brief recognition memory test as an additional measure of retrospective memory. The recognition memory test took approximately 12 minutes to complete, and consisted of a study phase in which 56 words were presented visually on the computer at a rate of approximately one word every 2 seconds. Immediately following completion of the study phase, participants completed the test phase, which consisted of visual presentation of 112 words, half previously studied, half new. For each word, participants were asked to press a key labeled yes if they believed they had seen the word in the previous study phase, and to press a key labeled no if they believed they had not seen it in the previous study phase. Following completion of the recognition test, participants were debriefed regarding the purpose of the study, and were provided time to ask any questions they might have had.

12 642 CRAIG P. MCFARLAND AND ELIZABETH L. GLISKY RESULTS For PM performance, we recorded the number of times participants remembered to press the 6 key. A PM score was computed, given a total of eight possible responses. Correct responses were defined as an appropriate button press (i.e., press of the 6 key) made during or following the presentation of the target question, and prior to the presentation of the first question to succeed each target question. A value of 1 was given to 6 button presses made prior to the presentation of succeeding questions. A value of 0 was given for omissions and to button presses made following the presentation of subsequent questions. The mean proportion of correct responses is shown in Table 2 as a function of neuropsychological group and instructional condition. A 2 (high- vs. low-fl group) 2 (Read-Only vs. Implementation Intention) ANOVA revealed a main effect of frontal status, with the high- FL group completing significantly more PM tasks than the low-fl group, F(1,28) = 8.23, MSE = 0.06, p <.01, n p2 =.24. There was also a main effect of instructional condition, with participants in the implementation intention condition completing significantly more PM tasks than participants in the read-only condition, F(1,28) = 11.65, MSE = 0.06, p <.005, n p2 =.29. There was no group by condition interaction, F < 1, as both groups benefitted from implementation intention instructions. There were no main effects of FL group or instructional condition on response times to the PM cue-word state, Fs < 1. Although high- and low-fl groups were equivalent with regard to MTL status, it was possible that MTL status influenced PM task performance. However, correlational analysis revealed that there was no relation between MTL status and PM task performance, r = 0.08, ns. Ongoing Task Performance The number of questions answered on the secondary task was assessed. There was no main effect of FL function on the number of trivia questions answered correctly, nor did it interact with instructional condition, Fs < 1. TABLE 2. Mean (SD) proportion of PM tasks executed as a function of neuropsychological group and instructional condition FL function Instructional condition Low FL High FL Mean Read-Only.0.30 (.0.32).0.55 (.0.29).0.42 (.0.32) Implementation Intention.0.60 (.0.21).0.85 (.0.11).0.72 (.0.21).0.45 (.0.30).0.70 (.0.26) Note: FL, frontal lobe.

13 IMPLEMENTATION INTENTIONS 643 There was also no main effect of instructional condition on the number of trivia questions answered correctly, F < 1. The time taken to respond to trivia questions was also recorded and assessed. There was no main effect of FL function on response time in answering trivia questions, nor did it interact with instructional condition, Fs < 1. There was no main effect of instructional condition on response time, F < 1. Retrospective Memory Three separate measures were used to assess retrospective memory. As noted above, participants were categorized as possessing high- or low- MTL function, which served as the first measure of retrospective memory. Participants in read-only and implementation intention conditions did not differ with regard to MTL status, t(30) = 0.76, p >.05. A brief recognition memory test, presented upon completion of the trivia questions, provided a second measure of retrospective memory. Again, participants in the readonly and implementation intention conditions did not differ, t(30) = 1.55, p >.05. There was a significant relation between MTL function and recognition memory performance, r =.37, p <.05, such that the higher one s MTL status, the more items correctly recognized. Finally, after completing the recognition memory test, participants were asked to describe the tasks involved in the experiment. All participants were able to correctly recall the requirement to answer multiple choice trivia questions and to press the 6 key whenever a question included the word state. As anticipated, we did not find a significant difference in retrospective memory between participants in the two instructional conditions. Also, as expected, performance on the recognition memory test was related to one s MTL factor score. DISCUSSION The current study demonstrated a relation between neuropsychological function and performance of an event-based PM task, thereby replicating a previous finding (McDaniel et al., 1999). Older adults characterized as possessing high-fl function completed significantly more PM tasks than did participants characterized as possessing low-fl function. This finding fits nicely with the results of patient studies (Burgess et al., 2008; Kliegel, Jäger, Altgassen, & Shum, 2008a) and also with the results of imaging work (Okuda et al., 2007; Poppenk, Moscovitch, McIntosh, Ozcelik, & Craik, 2010; Reynolds, West, & Braver, 2009), and was anticipated in light of theoretical views of frontal function. The FL factor used in this study, though it provides only an indirect measure of frontal function, has previously predicted performance on tasks of source memory (Glisky et al., 1995, 2001;

14 644 CRAIG P. MCFARLAND AND ELIZABETH L. GLISKY Glisky & Kong, 2008), and event- and time-based PM (McDaniel et al., 1999; McFarland & Glisky, 2009, respectively), each of which has been more directly linked with frontal function (Bisiacchi, 1996; Glisky, 1996; Janowsky, Shimamura, & Squire, 1998; Senkfor & Van Petten, 1998; Swick, Senkfor, & Van Petten, 2006; Okuda et al., 2007). Thus, these results add to a growing literature that suggests that FL function plays a key role in tasks of event-based PM. The nature of frontal function in event-based PM tasks is uncertain, but could involve a variety of processes. One such process is monitoring. Successful completion of PM tasks may require one to monitor the environment for cues that indicate the appropriateness of performing a previously formed intention (Smith, 2003; but, cf. McDaniel & Einstein, 2000). If the frontal lobes are involved in such a monitoring process, one might expect to see a performance disparity between high- and low-fl participants, such that high-fl participants successfully complete significantly more PM tasks than low-fl participants. Although such a discrepancy was present in the current study, the nature of the cues used in the current study does not allow for a discrimination to be made between tasks that required more or less monitoring (Einstein & McDaniel, 2005; Scullin, McDaniel, & Einstein, 2010). However, in a previous study conducted in this lab (McDaniel, 1999), cue salience was manipulated so that half of the target cues were salient (i.e., in bold font) and half were non-salient (i.e., standard font). Though there was a performance decrement for participants characterized as possessing low- FL function relative to high-fl participants, the performance of the low-fl group was equally impaired on both salient and non-salient cues, a finding contrary to what would be anticipated if a monitoring deficit were the root of the impaired PM performance of the low-fl group. Those results suggest that the frontal lobes may not be involved in the type of monitoring demanded by event-based PM tasks. It is possible, however, that the monitoring processes that may be required in event-based PM tasks are not tapped by the FL measure used in these studies. Alternatively, it is possible that the event-based PM task employed in these studies does not require monitoring processes. According to McDaniel and Einstein s (2000) multi-process framework, some event-based PM tasks, particularly those involving focal cues, can be successfully executed without demand on frontally-mediated monitoring processes. McDaniel and Einstein (2000) contend that focal cues, such as those in the present study, which are embedded and processed during the execution of an ongoing task (Einstein & McDaniel, 2005) can prompt automatic retrieval of the intention. A second process through which frontal lobe functioning may impact PM performance is strategic retrieval. Determining the significance of a cue and searching for the appropriate response to a given cue are processes that have been demonstrated to rely upon FL function (Moscovitch, 1994). It is

15 IMPLEMENTATION INTENTIONS 645 possible that the reduced PM performance of participants with low-fl function was the result of impaired search processes, particularly as they relate to retrieval of cue and intention information. However, given the 1:1 nature of cue-intention relation in this study, this possibility is unlikely. In other words, this task should have been relatively simple in light of the fact that if the cue is apprehended, the retrieval of the associated intention should take place with little effort. The development of plans is another process that is likely associated with frontal function and has been shown to have an impact on PM performance (Kliegel et al., 2000; Kliegel, Mackinlay, & Jäger, 2008b; Mäntylä, 1996; McFarland & Glisky, 2009). It is plausible that the high-fl group developed more well-defined plans for executing the PM task than did the low-fl group, and that such plans led to an increased awareness of PM cues. Additionally, the development of better plans for completing the PM task could have resulted in greater binding of cue and intention, in this case state and the pressing of the 6 key. This possibility is in line with previous findings in which the FL composite score used to characterize frontal function in this and other PM studies (McDaniel et al., 1999; McFarland & Glisky, 2009) has also been predictive of source memory among older adults (Glisky et al., 1995, 2001; Glisky & Kong, 2008). The relation between FL function and source memory performance in those studies was interpreted in light of the role of frontal function in the integration of content and context during encoding (Mitchell, Johnson, Raye, & D Esposito, 2000; Prabhakaran, Narayanan, Zhao, & Gabrielli, 2000). A similar process may have been engaged in the current study, by which a frontally-mediated integration of cue and intention was produced, resulting in better PM performance among participants of the high-fl group. Such an integration of cue and intention is one theory regarding the effectiveness of implementation intentions in improving PM performance (Cohen & Gollwitzer, 2008; Gollwitzer, 1993, 1996, 1999; McDaniel et al., 2008). The results of the current study also extend our understanding of the effectiveness of implementation intentions. Older adults in the implementation intention condition successfully completed significantly more PM tasks than did participants in the read-only condition. This finding is consistent with that of previous studies in which older adults who received implementation intention instruction outperformed those who received standard read-only instruction, both in laboratory and naturalistic settings (Chasteen et al., 2001; Liu & Park, 2004, respectively). The current study, however, was the first to investigate the effectiveness of implementation intentions among older adults characterized as possessing high- or low-fl function. Although implementation intentions have been shown to improve PM performance among a variety of populations, the mechanism by which they exert their influence remains uncertain. One theoretical notion is that a strong

16 646 CRAIG P. MCFARLAND AND ELIZABETH L. GLISKY social commitment is instantiated when participants write or state When I see X, I will do Y (Ellis & Freeman, 2008). The development of a strong commitment to completing the PM task would presumably result in greater importance being placed on its successful execution, something that has been shown to improve PM performance (Kliegel, Martin, McDaniel, & Einstein, 2001, 2004). However, an increase in perceived importance of the PM task typically comes at the expense of ongoing activity (Kliegel et al., 2001, 2004), as the importance of the PM task mandates a shift in attentional resources away from the ongoing task and in the direction of the PM task. In the current study we would have anticipated a decrement in performance of the ongoing trivia questions for participants in the implementation intention condition, if in fact a strong social commitment was made. Specifically, heightened attention to the PM task as a result of its greater importance would have been expected to result in slower response times to the ongoing trivia questions, fewer correct responses, or both. However, none of these patterns of performance emerged. Instead, participants in the read-only and implementation intention conditions responded to questions, on average, at the same speed, and both groups answered an equivalent number of questions correctly. These findings suggest that although a strong social commitment or greater perceived task importance may influence PM performance under some conditions, it is unlikely to have been the driving force behind the success of participants in the implementation intention condition in the current study. However, it may be possible, as some have suggested (Ellis & Freeman, 2008; Meeks & Marsh, 2010) that an increase in perceived importance has the effect of increasing attention to the entire task set (i.e., both the ongoing and PM tasks), rather than biasing attention towards the PM task at the expense of the ongoing task. Another means by which implementation intentions may influence PM is through the establishment of an increased level of activation of, or heightened accessibility to environmental cues. By specifying a future situation in which a cue may be encountered, implementation intentions are thought to facilitate detection of those cues, and in the process, relieve the burden of effortful search processes (Cohen & Gollwitzer, 2008; Gollwitzer, 1993, 1996, 1999). This increased accessibility to cues may be the product of a state of perceptual readiness. In a study of lexical decision making, Aarts, Dijksterhuis, and Midden (1999) reported faster response latencies to words that were associated with goal attainment than to words unrelated to a previously established goal. The authors interpreted those findings as evidence that implementation intentions produce increased accessibility to situational cues, thereby resulting in greater likelihood of achieving one s goal. In the current study, implementation intentions may have affected PM performance by producing heightened accessibility to situational cues, in this case, the word state, which in turn increased the likelihood of successful PM task

17 IMPLEMENTATION INTENTIONS 647 performance. This may have been true for both low- and high-fl groups. Although there is reason to believe that the high-fl participants in the readonly condition would have spontaneously developed more specific plans (McFarland & Glisky, 2009), and created a stronger link between cue and intention (Glisky et al., 1995, 2001; Glisky & Kong, 2008) than did low-fl participants in the read-only condition, it is certainly plausible that those plans did not involve the degree of specificity inherent in the implementation intention instructions. Those less specific plans, although providing for greater sensitivity to the presence of environmental cues than the plans developed by the low-fl group may have contributed to a level of PM task performance that left room for improvement for participants in the implementation intention condition. Therefore, perceptual readiness or heightened accessibility to situational cues resulting from the specificity of implementation intentions may have contributed to the greater PM performance of both low- and high-fl groups under implementation intention instruction compared to those under read-only instruction. A third mechanism through which implementation intentions may have their effect relates to the development of a strong associative link between cue and intention. Gollwitzer (1993, 1996, 1999) and others (Cohen & Gollwitzer, 2008; Liu & Park, 2004; McDaniel et al., 2008; Wilson & Park, 2008) contend that by specifying a future context in which a cue may be encountered, and defining a particular action to be performed in response to that cue, a strong associative link is produced. This associative link is believed to then result in reflexive responding whereby the presence and apprehension of a cue results in the effortless production of the pre-specified action. For the purposes of the current study, the implementation intention of when I see the word state, I will press the 6 key should result in relatively reflexive responding of pressing the 6 key in the presence of the word state. The argument that Gollwitzer (1993, 1996, 1999) makes regarding the idea of reflexive responding rests on the notion that by increasing the strength of the associative link between a future context, cue, and intention, any controlled, effortful search processes become unnecessary because the intended action is spontaneously triggered by the presence of the cue. If implementation intentions do in fact produce a strong associative link between cue and intention, then the low-fl group in the current study would be expected to benefit from their employment, as frontal function has been implicated in the integration of content and context (Mitchell et al., 2000; Prabhakaran et al., 2000). Furthermore, previous studies using the FL measure suggest that high-fl participants may spontaneously produce a strong link between content and context, while low-fl participants do not (Glisky et al., 1995, 2001; Glisky & Kong, 2008). Consistent with this expectation, low-fl participants in the implementation intention condition completed

18 648 CRAIG P. MCFARLAND AND ELIZABETH L. GLISKY significantly more PM tasks than low-fl participants in the read-only condition. However, high-fl participants also benefitted from implementation intentions, as high-fl participants in the implementation intention condition successfully completed significantly more PM tasks than high-fl participants in the read-only condition. Although both the low- and high-fl groups benefitted from the use of implementation intentions, these results do not rule out the possibility that the development of a strong associative link is the critical factor underlying the effectiveness of implementation intentions. It is possible for instance that among the high-fl participants in the read-only condition, only some participants spontaneously created a strong associative link between cue and intention. The use of implementation intentions, however, could have encouraged most, if not all, high-fl participants to establish a strong associative link, thereby fostering improved PM performance. Recent findings of improved PM performance following implementation intention instruction among undergraduate students (Cohen & Gollwitzer, 2008; Meeks & Marsh, 2010; McDaniel et al., 2008; McFarland & Glisky, 2011) lend additional credence to this argument, as younger adults have been found to possess greater processing resources than older adults (Anderson et al., 1998; Craik & McDowd, 1987). The finding that younger adults can benefit from implementation intentions suggests that although younger adults are capable of developing good plans and making strong associative links, they too may fail to do so at times, and in those cases implementation intentions might encourage engagement in those processes (McDaniel et al., 2008). An important caveat to the notion that implementation intentions produce automatic, reflexive responding when a cue is encountered is highlighted by recent findings. McDaniel and Scullin (2010) reported that implementation intentions did not protect against PM decline under conditions of high demand, a finding that they interpreted as contrary to what would be anticipated if implementation intentions invoke automatic responding. It is possible that implementation intentions provide for reflexive retrieval of an intention upon encountering a cue through the establishment of a strong link between cue and intention. However, the implication that implementation intentions yield automatic responding seems less probable, given that successful PM is believed to involve additional cognitive processes, including deciding to respond and interrupting performance of an ongoing activity in order to do so. This may be particularly relevant in the current study, in which participants may have reflexively or automatically retrieved the intended action upon encountering the cue, but delayed the execution of that action until after responding to the trivia question in which the cue was embedded. In summary, the current findings contribute to both our understanding of the role of frontal function in event-based PM tasks, as well as the influence

19 IMPLEMENTATION INTENTIONS 649 of implementation intentions on PM task completion among older adults. Several candidate processes likely played some role in successful PM execution in the current study. As anticipated, the present results indicated that frontal function is critically involved in event-based PM, and are consistent with previous research (McDaniel et al., 1999), and with theoretical notions of frontal function. Additionally, the current study was the first to investigate the effect of implementation intentions on PM among older adults characterized as possessing low- or high-fl function. Furthermore, the current results suggest a role for both heightened accessibility to environmental cues and a strengthened association between cue and intention in the improved PM that follows implementation intention instruction. REFERENCES Original manuscript received 5 December 2010 Revised manuscript accepted 18 July 2011 First published online 27 October 2011 Aarts, H., Dijksterhuis, A., & Midden, C. (1999). To plan or not to plan? Goal achievement or interrupting the performance of mundane behaviors. European Journal of Social Psychology, 29, Anderson, N. D., Craik, F. I. M., & Naveh-Benjamin, M. (1998). The attentional demands of encoding and retrieval in younger and older adults: I. Evidence from divided attention costs. Psychology and Aging, 13, Bisiacchi, P. S. (1996). The neuropsychological approach in the study of prospective memory. In M. Brandimonte, G. O. Einstein & M. A. McDaniel (Eds.), Prospective memory: Theory and applications (pp ). Mahwah, NJ: Erlbaum. Burgess, P. W., Dumontheil, I., Gilbert, S. J., Okuda, J., Schölvinck, & Simons, J. S. (2008). On the role of rostral prefrontal cortex (area 10) in prospective memory. In M. Kliegel, M. A. McDaniel & G. O. Einstein (Eds.), Prospective memory (pp ). New York, NY: Erlbaum. Chasteen, A. L., Park, D. C., & Schwarz, N. (2001). Implementation intentions and facilitation of prospective memory. Psychological Science, 12, Cohen, A.-L., & Gollwitzer, P. M. (2008). The cost of remembering to remember: Cognitive load and implementation intentions influence ongoing task performance. In M. Kliegel, M. A. McDaniel & G. O. Einstein (Eds.), Prospective memory: Cognitive, neuroscience, developmental, and applied perspectives (pp ). Mahwah, NJ: Erlbaum. Craik, F. I. M. (1986). A functional account of age differences in memory. In F. Klix & H. Hagendorf (Eds.), Human memory and cognitive capabilities: Mechanisms and performances (pp ). Amsterdam, The Netherlands: Elsevier Science Publishers, B.V. Craik, F. I. M. (2005). Remembering items and their contexts: Effects of ageing and divided attention. In H. D. Zimmer, A. Mecklinger & U. Lindenberger (Eds.), Handbook of binding and memory: Perspectives from cognitive neuroscience (pp ). New York, NY: Oxford University Press. Craik, F. I. M., & McDowd, J. M. (1987). Age differences in recall and recognition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 13,