"Reduced false recognition in amnesia could be a result of impaired"
Reduced false recognition in amnesia could be a result of impaired item-specific memory: the relationship between item-specific memory and gist memory. Jack Nissan Supervisor: Sharon Abrahams Abstract It is a common finding that amnesic patients produce fewer false recognitions than healthy controls, and this has led to assumptions that gist memory is damaged in these patients (Schacter et al., 1996, Budson et al., 2000). Two experiments used false recognition paradigms to ascertain whether this result could instead be a consequence of impaired item-specific memory. Experiment 1 aimed to reduce the item-specific memory of healthy adults to reflect that of an amnesic patient, by using an articulatory suppression task, while Experiment 2 aimed to increase the item-specific memory of amnesic patient JY to reflect that of a healthy adult, by bringing her to criterion on the relevant study-lists. Results indicated that when item-specific memory was sufficiently reduced in healthy adults, they produced a similar pattern of results to that found in amnesic patients, and when JY’s item-specific memory was increased, she produced a similar pattern of results to healthy adults. This suggests that the previous assumption that gist memory is damaged in amnesic patients might be flawed. The implications of this are discussed in terms of the relationship between item-specific memory and gist memory. 2 Introduction Amnesic disorders can cause profound disruptions to people’s lives, both in terms of one’s general wellbeing and also with regard to one’s ability to live independently. It is thus important to find effective ways to treat and/or cope with these problems, and crucial to this is an accurate understanding of the specific disorder and precisely which areas of memory it affects. There is comprehensive evidence that amnesic patients show substantial deficits in episodic memory. Episodic memory is the ability to remember certain events, episodes or information within a particular temporal or spatial context, and patients perform consistently badly on tasks which assess this, such as recalling stories, sentences and individual words, and recognising words, pictures and faces (Nebes, 1992). Two subdivisions of episodic memory have been described as item-specific memory (Balota et al., 1999; Budson et al., 2000) and source monitoring ability (for a review, see Johnson et al., 1993). Item specific memory is the ability to remember individual items in themselves, without putting them into any semantic context. It can be tested using some of the methods described above, such as recall of unrelated words, pictures and faces, and as mentioned, amnesic patients are severely impaired in their ability to perform these tasks. Source monitoring is the ability to identify and memorise the source of information or the context in which it was learnt. It is inherently related to, and in a sense included in, item-specific memory, as one needs to be able to remember the origin of specific information in order to recall it appropriately. For example, in order to accurately recall a list of words, one needs to be able to distinguish between the words that were on the particular list from any other words that might come to mind from different sources. In real-life situations, 3 source monitoring allows us to form opinions and judgements, as to do so we need to discriminate between various sources of information, and it also enables us to distinguish what is real from what is imagined (Johnson et al, 1993). Numerous studies have found source memory deficits in amnesic patients (eg. Schacter et al, 1984; Multhaup and Balota, 1997; Dalla Barba et al., 1999; Smith and Knight, 2002; Pierce et al., 2005). For example Multhaup and Balota (1997) found that Alzheimer’s disease (AD) patients struggled to distinguish between information they had read and information generated by themselves, and Dalla Barba et al. (1999) found AD patients were impaired in their ability to discriminate between objects they had seen and imagined. While it is widely accepted that amnesic syndromes are extremely damaging to the various elements in episodic memory, their effect on semantic memory is more controversial. Semantic memory refers to the general meaning of information. It is a memory for information within a semantic context as opposed to a temporal or spatial context, as is the case with episodic memory. Tulving (1972, p386) describes it as a “mental thesaurus, organised knowledge a person possesses about words and other verbal symbols, their meaning and referents, about relations among them and about rules, formulas, and algorithms for the manipulation of the symbols, concepts and relations”. So whereas episodic memory might allow a person to recall that he walked passed a cinema yesterday, or that he saw the word cinema written on a memory test, semantic memory is responsible for a general understanding of what a cinema is: that it relates to films, a big screen, going out etc. Several studies report that amnesic patients show deficits on various semantic memory tasks, such as category fluency tasks (Troster et al, 1989) and freely producing associations to words (Gollan et al., 4 2006). However, other researchers have argued that these particular tasks require large attentional demands and that the deficits observed are actually the result of damage to an attentional control system and do not necessarily reflect any damage to the patients’ semantic memory (Balota et al., 1999, Watson et al., 2001). Various studies report that AD patients are deficient in attentional and inhibitory processes (Spieler et al., 1996; Faust et al., 1997, Simone and Baylis, 1997a, 1997b), and evidence that their semantic memory is actually in tact comes from certain semantic priming tasks, whereby patients demonstrate in tact priming effects under conditions that do not reflect attentional processes (eg. Balota and Duchek, 1991), but disrupted priming effects under conditions that do. (eg. Ober and Shenaut, 1995). The literature on semantic memory in amnesia is thus divided and the key debate is about whether amnesic syndromes damage semantic memory itself or whether they damage other functions which interfere with the access to semantic memory. Another area of memory that forms a part of semantic memory is known as gist memory (Reyna & Brainerd, 1995). This is the ability to put information into a larger semantic context. It allows us to obtain an understanding of the general meaning of information and get the gist of what it is about. Gist memory has been explored extensively through studies on false memories, as the two seem to be related. This has been demonstrated in a series of experiments that asked healthy participants to study semantically related word lists, which converged on certain target words, sometimes called critical lures (these were not presented to the subject in the study list). For example, one list from an experimental paradigm called the DRM (Roediger and McDermott, 1995) contains the words: HOT, SNOW, WARM, WINTER, ICE, WET, FRIGID, CHILLY, HEAT, WEATHER, FREEZE, AIR, SHIVER, ARCTIC, 5 and FROST, and the critical lure for this list is COLD. After studying several of these lists, participants completed a recall or recognition phase, and results have shown that healthy adults consistently recalled or recognised the critical lures that were not on the original lists (Underwood, 1965; Cermak, 1973; Roediger and McDermott, 1995; Norman and Schacter, 1997). The fact that healthy participants produce these lures has been attributed to their gist memory. The theory is that through seeing the list of related words, a memory for the general gist of the list is formed, which encompasses representations of both the words on the actual lists and any others which are related to the gist, such as the critical lures. Hence when asked if a critical lure was on the list, participants falsely respond “yes”, as they are using their gist memory to identify the word. These experiments using related word lists have been modified and extended to study patterns of false recognition across different groups, and these patterns have revealed some interesting findings about memory and how different aspects of memory interact with each other. For instance, when false recognitions are compared between healthy young and old adults, it is found that old adults falsely recognise more critical lures than young adults (Schacter, Koustaal et al., 1997; Balota et al., 1999; Norman and Schacter, 1997). In one way this result might seem surprising, as it could imply that old adults have a better gist memory than young adults, but further experiments reveal this is not the case. Dehon and Bredart (2004) carried out an experiment to assess false recognitions in young and old adults but added an extra phase where participants had to report any other words that came to mind during the experiment. They found that young and old adults reported the same number of related lures but that young adults tended to think of them in the learning stage while old adults produced them in 6 the recall phase. This implies that there is no difference in the gisting processes in young and old adults, since they thought of the same number of related lures, and that the difference in false recall relates more to source monitoring ability. Other experiments have demonstrated an age difference in source monitoring (eg. Tun et al., 1998; McIntyre and Craik, 1987; Dywan and Jacoby, 1990), and this could therefore be one reason behind the age difference in performance on false recognition tasks. Another explanation focuses on the interaction between item-specific memory and gist memory (Kensinger and Schacter, 1999; Balota, 1999; Budson et al., 2000). Remembering related word lists is likely to involve both these processes, where gist memory assists in the recognition of all related words (both studied and unstudied), and item-specific memory enables us to remember the individual words in themselves, and hence exclude any unstudied related words that our gist memory might endorse. As young adults have a better item-specific memory than old adults, they are more able to use this information to regulate their gist memory, and hence they produce less false recognitions than old adults. Support for this theory comes from an experiment by Kensinger and Schacter (1999) in which the entire study-test procedure was repeated across five trials. Results indicate that young adults are able to use this repetition to reduce the number of false recognitions, whereas older adults cannot, and it is suggested that this is due to an increasing item-specific memory in the younger adults. When false recognitions are studied in people suffering from amnesic disorders, the results are quite surprising. Given that both their source memory and item-specific memory is so poor, we might expect these patients to show an even greater increase in the number of lures they falsely recognise. However, the results indicate the complete 7 opposite, finding that they falsely recognise even less related lures than healthy young adults (eg. Budson et al., 2000; Budson, Desikan et al., 2001). A large body of literature explains this result by inferring that amnesic patients have an impaired gist memory (eg. Budson et al, 2000; Budson, Desikan et al., 2001; Budson et al., 2006; Gallo, 2006). This certainly seems like a rational explanation, but there may be another way of looking at the results that would lead to a different conclusion. It could be that in order to gist, we need a certain amount of information to gist from. Whilst it may not have been tested, it seems common-sense that we would get a clearer gist from a list of twenty related items than we would from a list of two. Hence an increased item-specific memory, whilst being able to oppose gist memory in one sense, might at the same time enhance it, since the more items one has available to them, the better their gisting resources. Further to this, there might be a sort of threshold, whereby one cannot gist effectively if there are not enough items available to them. This would not imply that their ability to gist is damaged, just that there is simply not enough information for it to act on, and it is possible that this is the case with amnesic patients. In a similar repeated trials experiment to the one described above (Kensinger and Schacter, 1999), Budson et al (2000) found that false recognition in AD patients actually increased over trials, in sharp contrast with young adults whose false recognition decreased (due to an improved memory for specific items) and old adults whose false recognition remained fairly stable. While the authors explained this result in terms of an impaired gist memory in AD patients that improves with repetition, it could instead be the AD patients’ impaired item-specific memory that improves, which in turn enables their intact gist memory to come into play. However, unlike healthy adults, they may still be unable to remember enough specific items to suppress the gist representations now available to them. 8 Some evidence that gist memory might be in tact in amnesic patients comes from studies conducted by Balota and colleagues (Balota et al, 1999; Watson et al 2001), which analyse the number of false recognitions of related lures with respect to the number of true recognitions of studied items. They matched AD patients who performed well on true recall with healthy older adults who performed poorly on this measure, and found that under these conditions AD patients actually showed a higher rate of false recognition than the older adults. This seems to imply that the reduced number of false recognitions found in AD patients is more a result of poor item- specific memory than of a damaged gist memory. Another observation that raises questions about the use of false recognition results to infer a damaged gist memory in amnesic patients, is the fact that this reduction in false recognitions has been found in patients suffering a variety of amnesias and amnesic syndromes, such as Korsakoff’s syndrome (Schacter et al., 1996; Schacter, Verfaille et al.,1997; Koustaal et al., 2001), semantic dementure (Simons et al., 2005), Alzheimer’s disease (Budson et al, 2000; 2006), frontal lobe lesions (Verfaille et al., 2004) medial temporal lobe lesions (Verfaille et al., 2004), and from mixed etiologies including anoxia and encephalitis (Schacter et al., 1996; Koustaal et al., 2001, Verfaille et al., 2002). These disorders all affect different areas of the brain, so it seems unlikely that they all damage the same function. However, it is more feasible that gist memory, as a particular memory function, could be accessed in various ways, and thus it seems more likely that each disorder prevents access to this function in some way as opposed to damaging it directly. 9 Hence, in a similar way that the results of certain semantic memory tasks might imply either damage to semantic memory itself or an inability to access it, the false recognition results of amnesic patients could imply either damage to their gist memory itself or an inability to access it, perhaps due to the poor item-specific memory caused by the amnesia. The current experiment tests for the possibility that the patterns of false recognition observed amnesic patients are more a product of a weak item-specific memory than of a damaged gist memory. Experiment 1 requires healthy young and old adults to perform two false recognition paradigms, one of which involves carrying out an articulatory suppression task during the presentation of the word lists. The purpose of the articulatory suppression task is to prevent participants from rehearsing the words and hence to reduce their item-specific memory so that it reflects that of an amnesiac. Since the participants are healthy and their gist memory is in tact, if they produce less false recognitions under the articulatory suppression condition, this would mimic the pattern observed in amnesic patients and thus suggest that this pattern does not necessarily reflect a damaged gist memory, but could instead be the result of an impaired item-specific memory. Experiment 2 is a case study on amnesic patient JY. Again there are two false recognition paradigms to be completed, and in one of these the intention is to increase the item-specific memory of JY to mimic that of a healthy adult, by bringing her to criterion on the various lists. If, after being brought to criterion, JY produces more false recognitions than in the control condition, this would again suggest that it is a damaged item-specific memory as opposed to a damaged gist memory that causes the observed reduction in false recognitions in amnesic patients. 10 Experiment 1 Previous literature has shown that patients with various amnesic syndromes produce less false recognitions than healthy young and old adults (eg. Schacter et al., 1996; Budson, 2000), and these results have been used to conclude that this is due to a damaged gist memory in the amnesic groups (eg. Schacter et al., 1996; Budson et al., 2000, Gallo et al., 2006). However, when false recognition results are analysed with respect to veridical memory the opposite effect is found, and patients show an increase in false recognitions (Balota et al., 1999; Watson et al., 2001). This fits with literature on semantic memory in AD that suggests it is access to semantic memory, rather than semantic memory itself, that is damaged (Balota and Duchek, 1991; Balota et al., 1999; Watson et al., 2001). Hence, the theory posited by this paper is that the reduced number of false recognitions observed in these patient populations does not reflect a damaged gist memory but rather an inability to access it. Access to gist memory might require a certain level of item-specific memory and since these patients’ item-specific memory is so poor, it may be below the required level for them to gist effectively. This experiment compares performance of healthy young and old adults on two false recognition paradigms. One of these is a control condition, and the other involves participants performing an articulatory suppression task whilst studying the word lists. The articulatory suppression task aims to reduce the participants’ item-specific memory to reflect that of an amnesiac. Since it is hypothesised that this level of item- specific memory is too low to gist effectively, it is predicted that subjects will produce less false recognitions under the articulatory suppression condition than under the control condition. However, in the control condition, the level of item-specific 11 memory in both young and old adults is high enough to gist effectively, and it is thus predicted that young adults will produce less false recognitions than old adults, since they should be able to use their better item-specific memory to regulate their gist memory more so than old adults. Method: Participants: Four subjects were excluded from the analysis as they did not complete the experiment appropriately. Of the participants included in the analysis, younger adults were 32 students (16 male, 16 female) from Edinburgh University, all aged between 18 and 25 years old (mean=21.59, SD=1.48). Older adults were 33 residents (11 male, 22 female) from Edinburgh or nearby. They were all aged between 60 and 75 (mean=68.30, SD=4.00) and were recruited partly through a departmental volunteer list and partly through advertising around the city in garden bowls clubs, charity shops and churches. Materials and tasks: False recognition task: There were two false recognition paradigms (A and B) in total and both were presented on a computer. Each paradigm consisted of twelve related word lists, and each list contained fifteen related words, which were presented in the study phase, and an associated critical lure which was presented in the recognition phase. The lists were taken from an experiment by Bellamy and Shillcock (2007), and were mostly the same as those used in a well recognised false recognition paradigm called the DRM (Roediger and McDermott, 1995), though some words were adapted to better suit the native language of British English as opposed to 12 American English (see appendix A for a table of all the lists used and the respective critical lures). The words from each list were displayed one at a time at the centre of a computer screen for a duration of two seconds per word. There was a one second pause between each word during which the participants focused on a fixation cross in the centre of the screen. After each list, written instructions appeared informing participants that they had to complete some simple maths problems. The maths problems were then displayed and participants had to type in the answer and hit “enter”, after which another problem would be presented. After thirty seconds, a message appeared on the screen saying “well done” and the next list of words was displayed. After all twelve lists had been displayed, there was a recognition phase. Instructions appeared on the screen informing participants that random words would appear and that they had to decide whether or not these words were on any of the lists they had just seen. If they thought the word was on any of the lists, they had to hit the “yes” key and if they thought the word was not on any the lists they had to hit the “no” key (the actual keys were on a standard computer keyboard but were covered with “yes” and “no” stickers). In order to avoid any right- or left-handed bias, they were told to use only the index finger of their preferred hand to press both “yes” and “no”. There were 48 words in total in the recognition phase and these consisted of the 12 critical lures, 24 studied words (two from each list) and 12 completely unrelated words that were neither on or related to any of the lists. The computer programme recorded the results of the recognition phase. Free recall task: Two lists were used for each participant and both consisted of unrelated words (see appendix B for a table of these lists). The function of these lists was to gain an independent measure of item-specific memory. Each word was printed 13 on an A5 sized flashcard. All words were taken from “Birkbeck Word Association Norms” (Moss and Older, 1996) and this book was used to ensure that the words were as semantically unrelated as possible, and also that neither of the lists would be subject to semantic interference with each other or any of the lists used in the false recognition paradigm. In each list, the most common 10 associates of every word were cross-referenced with each other and all the words in the false recognition paradigms to ensure they were completely unrelated. If any word or associated word was on any of the lists in the false recognition paradigms or on any of the free recall lists then that word was discarded. The lists were also equally split between concrete and abstract words to try and replicate the difficulty of the lists in the false recognition paradigm. For each of the recall lists, participants were instructed to try to remember as many words as possible. The list was then presented, one word at a time and a rate of approximately 2 per second. After the last word was presented, participants were asked to count down in intervals of three from a number above 200, and after thirty seconds of doing so, they were told to recall as many of the words as possible in any order. The number of words correctly recalled was recorded on a score sheet. Articulatory suppression task: Throughout one of the unrelated free recall tests and one of the false recognition paradigms, participants had to perform an articulatory suppression task while they studied the words. The task was the same as that used by Beaman and Jones (1997), and involved saying the letters A-G repeatedly at a rate of approximately 2 letters per second. Participants were instructed to keep the flow of letters as constant and regular as possible and to avoid leaving a pause between the G and following A. The articulatory suppression task was only performed whilst 14 studying the word lists and not throughout the recall or recognition phases or during the maths problems. National Adult Reading Test - Revised (NART) (Nelson and Willison, 1991): The NART was used to match the young and old groups by a general measure of intelligence, so as to ensure that any effect of age would not be due to a difference in IQ. The test itself is a list of words which participants have to read out loud. The marking is based on the pronunciation of the words and participants are thus instructed to attempt to pronounce all of the words, even if they do not recognise them. Procedure: Participants were told that most of the experiment would take place on the computer and that in addition there would be a few tests carried out in person. The general procedure for the false recognition tests, free-recall tests and articulatory suppression task was described to participants and they were asked to sign a consent form. Participants first carried out one of the free recall tasks, and this was followed by one of the false recognition tests. They then had a five minute break, after which they completed the other free recall task, followed by the remaining false recognition test. Throughout one of the unrelated free recall tasks and one of the false recognition tests, participants had to carry out the articulatory suppression task whilst studying the words. They were instructed to perform the articulatory suppression only throughout the presentation of the words and not during the maths problems or throughout the recall or recognition phases. The order that the false recognition, free recall, and articulatory suppression tasks were performed was varied equally between 15 participants to control for any effect of test order. After the second false recongition task had been completed, participants carried out the NART. Results: First, a word must be said about correction procedures that are frequently used in analysing false recognition paradigms such as this one. These are carried out on critical lures and studied words to account for biases in the results. There are two principal methods that have been employed in previous studies. The most common method involves subtracting the proportion of “false alarms” (falsely recognised unrelated words) from the proportion of “hits” for critical lures (falsely recognised critical lures) and studied words (correctly recognised studied words). So for each participant, the corrected score for critical lures would be the proportion of falsely recognised unrelated words subtracted from the proportion of falsely recognised critical lures, and the corrected score for studied words would be the proportion of falsely recognised unrelated words subtracted from the proportion of correctly recognised studied words. The other method is signal detection analysis (see Banks, 1970 for a review), in which the corrected scores for critical lures and studied words are obtained by entering the proportions of false alarms and hits into certain equations. Given that these experiments aim to assess one’s ability to gist by looking at false recognitions of critical lures, there is a potential problem with the first correction procedure in its assessment of critical lures. While it accounts for recognition errors of unrelated words, it does not seem to account for errors made in the recognition of studied words. Recognition of studied words involves a combination of item-specific and gist memory (Budson, 2000), since the studied words all come from semantically related lists. Hence failing to recognise these 16 words should be a sign of an impaired gist memory, in the same way that failing to recognise critical lures is taken as a sign of an impaired gist memory. This should therefore be taken into account when analysing critical lures. Signal detection analysis uses a more independent methodology to correct for biases and is a widely accepted and acclaimed procedure. As a result, a signal detection analysis of the data from the current experiment is presented in this paper. However, the data was also analysed using the standard correction procedure and the main differences are mentioned at the end of the results section. The type of signal detection analysis used in the current study was a simplified version of d’, the equation for which is A’ = 0.5 + [(H - FA)(1 + H - FA)] / [4H(1 - FA)] (Radvansky, 2006). This method will be referred to from here on as the A’ correction procedure, and any critical lure scores or studied word scores corrected by this method will be referred to as A’ critical lures or A’ studied words respectively. Table 1 shows the means and standard deviations for the percentages of A’ critical lures, A’ studied words and unrelated words endorsed in the young and old adult groups and within each experimental condition. Control Articulatory suppression (AS) A' A' A' A' Critical Studied Unrelated Critical Studied Unrelated lures words words lures words words mean 85.58 89.03 6.78 86.63 85.72 10.42 Young SD 9.88 7.28 9.57 8.77 8.20 11.20 mean 89.11 89.72 9.09 86.17 86.23 15.15 Old SD 6.03 3.89 9.85 8.47 6.08 11.87 mean 87.37 89.38 7.95 86.39 85.98 12.82 Total SD 8.28 5.77 9.71 8.55 7.15 11.70 Table 1: means and standard deviations (SD) of A’ critical lures, A’ studied words and unrelated words endorsed by young and old adults within the control and articulatory suppression (AS) experimental conditions. 17 Unrelated words: A mixed ANOVA, with one within subject factor of “experimental condition” (2 levels: control and AS) and one between subjects factor of “age” (2 levels: young and old), revealed that there was no main effect of age on unrelated words, but that there was a significant main effect of experimental condition on unrelated words (F=16.84, df=1, p<0.01). Participants falsely recognised more unrelated words in the AS condition (mean = 12.82) than in the control condition (mean =7.95), and post hoc analysis using repeated measures ANOVAs found that this effect was present in both old adults (F=11.506, df=1, p<0.01) and young adults (F=5.61, df=1, p<0.05). This pattern of results is depicted in Figure 1. 20.00 % Unrelated Words 15.00 Young 10.00 Old 5.00 0.00 Control AS Experimental Condition Figure 1: graph to illustrate the mean numbers of unrelated words falsely recognised by young and old adults across the different experimental conditions. A’ Studied words: A mixed ANOVA, with one within subject factor of “experimental condition” (2 levels: control and AS) and one between subjects factor of “age” (2 levels: young and old), revealed that there was no main effect of age on A’ studied words but that there was a significant main effect of experimental condition (F=22.40, df=1, p<0.01). More A’ studied words were correctly recognised in the control condition (mean = 18 89.38) than in the AS condition (mean = 85.98) and post hoc analysis using repeated measures ANOVAs found a similar effect in both the young adult group (F=9.02, df=1, p<0.01) and the old adult group (F=14.17, df=1, p<0.01). This pattern of results is depicted in Figure 2. 95.00 % Studied Words 90.00 Young Old 85.00 80.00 Control AS Experimental Condition Figure 2: graph to illustrate the mean numbers of studied words correctly recognised by young and old adults across the different experimental conditions. A’ Critical lures A mixed ANOVA, with one within subject factor of “experimental condition” (2 levels: control and AS) and one between subjects factor of “age” (2 levels: young and old), revealed that there was no main effect of age or experimental condition but that there was a significant interaction between experimental condition and age (F=4.77, df=1, p<0.05). Post hoc analysis using repeated measures ANOVAs revealed that the experimental condition had no effect in the young adult group (F=0.52, df=1, p=0.48) but that it was significant in the old adult group (F=6.95, df=1, p<0.05), where old adults falsely recognised more A’ critical lures in the control condition (mean = 89.11) than in the AS condition (mean = 86.17). This pattern of results is depicted in Figure 3. 19 90.00 % Critical Lures Young 85.00 Old 80.00 Control AS Experimental Condition Figure 3: graph to illustrate the mean numbers of critical lures falsely recognised by young and old adults across the different experimental conditions. An analysis of the relationship between critical lures and studied words was also carried out within both experimental conditions. A’ studied words were positively correlated with A’ critical lures in both the control condition (r=0.55, p<0.01) and the AS condition (r=0.79, p<0.01). These correlations are illustrated in figure 4. AS Control 100 100 95 95 % Studied Words 90 % Studied Words 85 90 80 85 75 70 80 65 75 60 55 70 55 60 65 70 75 80 85 90 95 100 55 60 65 70 75 80 85 90 95 100 % Critical Lures % Critical Lures Figure 4: scatter plots to illustrate the correlations between studied words and critical lures in the control and AS conditions. 20 Free recall analysis: A mixed ANOVA revealed that there were significant main effects of experimental condition (F=38.624, df=1, p<0.01) and age (F=26.926, df=1, p<0.01) on the number of words recalled, but there was no interaction between them. Participants recalled more words in the control condition (mean = 43.79) than in the AS condition (mean = 30.36), and young adults recalled more words than old adults in both the control (t=3.714, df=63, p<0.01) and AS (t=3.853, df=63, p<0.01) free recall tasks. Differences between the signal detection analysis (A’) and the standard correction (SC) procedures for the current results: The analysis using the SC procedure produced similar patterns and results to those of the A’ correction procedure. The only difference was that, in the SC procedure, no significant interaction was found between experimental effect and age in the number of critical lures produced. However, there was a trend toward significance (F=3.246, df=1, p=0.076), and when young and old adults were analysed separately, the results were the same as the A’ analysis, in that there was a significant effect of experimental condition on old adults (F=4.423, df=1, p<0.05) but not on young adults (F=0.515, df=1, p=0.478). A possible reason for the lack of interaction in the SC analysis might be due to the large amount of variance in SC critical lures in the young adult group (SD=27.17). 21 Discussion Previous studies using false recognition paradigms such as this one have found that amnesic patients recognise fewer false recognitions than healthy adults (eg. Schacter et al., 1996; Budson et al., 2000; Gallo et al., 2006), and it has been concluded that this is due to a damaged gist memory in the patient group (eg. Budson et al., 2000; Gallo et al., 2006). The main objective of Experiment 1 was to see if a similar pattern of results could be produced in a healthy sample (ie. participants with an in tact gist memory) whose item-specific memory was reduced so that it reflected that of an amnesic patient. The analysis of the studied and unrelated words suggests that this may indeed be possible. There was a clear effect of experiment on both studied and unrelated words, where young and old adults both recognised significantly more studied words and significantly less unrelated words in the AS condition than in the control condition. This mimics the pattern of results of previous studies which find that amnesic patients produce less studied words and more unrelated words than healthy controls (eg. Schacter et al., 1996; Budson et al., 2001). However, the analysis of critical lures in this experiment does not bear such a clear resemblance to these studies. They find that amnesic patients produce less critical lures than healthy adults (eg. Schacter et al., 1996; Budson et al., 2000; Gallo et al., 2006), while the results of this study find that there was no main effect of experiment with respect to critical lures. However, there was a significant interaction between experiment and age, which revealed that, while there was no effect of experiment in the young adults, old adults produced significantly less critical lures in the AS condition than in the control condition. It would appear then, that the results from the old adults do reflect those of previous 22 studies that compare amnesic patients to healthy controls, while the results from young adults do not. Since critical lures are taken as the key measure of one’s gisting ability, this age difference raises some issues about the theories on gist and its relationship to item- specific memory put forward in this paper. Indeed, if these theories are to carry any weight at all, they need to be able to account for this interaction. One possible explanation posits that the AS task did not reduce the item-specific memory of young and old adults to the same level. The theory proposed by this paper is that item- specific memory needs to be above a certain threshold in order for one to gist effectively, and hence if one’s item-specific memory is below this level one should produce very few critical lures. However, beyond this threshold, an increased item- specific memory should result in a reduction of critical lures, since one should be able to use this memory to suppress the false gist representations now available to them. Hence, if one’s item-specific memory were to be reduced to a level that is still above the threshold, they would not produce fewer critical lures than when their item- specific memory was unaltered, and should in fact produce more. This might be the case in the current experiment, wherein the AS task reduced the item-specific memory of old adults to a level below the threshold, but failed to do so with the young adults. Evidence that the AS task reduced the item-specific memory of old adults more than it did young adults can be found in the analysis of the AS free-recall tests, in which young adults recalled significantly more words than old adults. Hence, it is argued that the reason the young adults did not produce less critical lures in the AS condition is because the AS task did not reduce their item-specific memory enough to impair their access to gist memory. In the old adult group, however, item specific memory 23 was reduced to a level that prevented them from gisting and they thus produced less critical lures than in the control condition. Explained as such, the results of this experiment are in keeping with the general theory espoused by this paper. There are a couple of potential problems with this argument however. The main issue is that dual-task performance might be the reason behind why the AS task reduces item-specific memory of old adults more than that of young adults. Previous studies have demonstrated that old adults are worse at dual-task performance than young adults (Hartley, 1992; McDowd and Shaw, 2002), and if this were the case in the present experiment, it could be that the cognitive load required by the old adults in performing the AS task might have interfered with their ability to gist. Hence, the fact that they were less able to gist in the AS condition than in the control condition may not be due their reduced item specific memory, but rather it could be an effect of their poor performance on dual-tasks. Another hole in the argument concerns the assumption that, above a certain threshold, an increase in item-specific memory should reduce the number of false recognitions of critical lures. If the AS task failed to reduce the item-specific memory of young adults to a level below this threshold, but did reduce it with respect to the control condition, we might expect young adults to produce more critical lures in the AS condition than in the control condition, and no such result was found. One possibility is that the assumption itself may be inaccurate. The theory that an increased item- specific memory reduces the number of critical lures produced was inferred from previous experiments which found that old adults produced more critical lures than young adults (Schacter, Koustaal et al., 1997; Norman and Schacter, 1997). Since 24 young adults have a better item-specific memory than old adults, it was assumed that they could use it to override false gist representations more so than old adults. However, other experiments failed to find this effect of age on critical lures (Kensinger and Schacter, 1999; Budson, 2000), and this experiment failed to do so as well. In addition, in the current experiment, the results from the free-recall tests in the control condition imply that the item-specific memory of the young adults was better than that of the old adults. Thus if the theory is correct, it is strange that no effect of age was observed with respect to critical lures in this experiment. The analysis of the relationship between critical lures and studied words challenges this theory as well, since there was a positive correlation between these two. If the theory were correct, one might expect the correlation to be in the opposite direction, since studied words represent a rough measure of item-specific memory and an increase in item-specific memory should, according to the theory, result in a decrease in the number of critical lures endorsed. However, whilst these results do clash with previous findings, and the initial predictions for this experiment, that old adults produce more critical lures than young adults, they can still be explained by the underlying theory advocated in this paper. That is that an increased item-specific memory, whilst being able to oppose gist memory in one sense, might at the same time enhance it, since the more items one has available to them, the better equipped they are to gist. Therefore, while someone with a better item-specific memory might override more false gist representations than someone with a weaker item-specific memory, their gist memory would also be enhanced, and so they might at the same time be producing more and stronger gist representations, and it is possible that the outcome of this would be an unobservable 25 difference between the two persons. In other words, young adults might suppress their false gist representations more than old adults, but still produce the same number of critical lures, since their gist memory would also be stronger. In addition, whilst it is hypothesised that gist memory is considerably influenced by item-specific memory, it is not entirely dependent on item-specific memory, and it exists as a type of memory in its own right. Hence there is room for variation in how much people gist off the same information. This implies a complicated and unpredictable interaction between gist and item-specific memory, but it does seem to account for the inconsistent results on the age difference in critical lures. Whether or not it is accurate, the mixed findings on this matter must at the very least suggest that the relationship between item-specific memory, gist memory and number of critical lures produced is not as clear cut as it has previously been made out to be. 26 Experiment 2: Case study of patient JY JY is 30 years old and suffers from a developmental amnesia, possibly as a result of birth asphyxia. She displays markedly poor immediate and delayed verbal recall but relatively in tact verbal recognition. Her visual memory is impaired for both recall and recognition. Her executive functions seem mostly unaffected, and she demonstrates only a slight slowing in information processing. She has been diagnosed as presenting “a clear amnesic syndrome, with preserved executive functions”. The current experiment compares JY’s performance on two tests of false recognition, both in relation to each other and to healthy young and old adults as well. One of these tests is a control condition, and the other requires that JY is brought to criterion on all of the word lists used in the test before undertaking it. The purpose of bringing JY to criterion is to improve her item-specific memory for these lists to a level approaching that of a healthy adult. This paper hypothesises that amnesia does not damage gist memory directly, and that the reason previous studies have found a reduced number of false recognitions in these patients (eg. Schacter et al., 1996; Budson et al., 2000; Gallo et al., 2006) is because their impaired item-specific memory prevents them from accessing their gist memory, and not because their gist memory itself is damaged. It is thus predicted that, in the control condition, JY will produce less false recognitions than healthy young and old adults, since her item- specific memory is too low for her to gist effectively. However, after being brought to criterion, it is predicted that she will produce more false recognitions than in the control condition, since her item-specific memory will have been improved to a level that her gist memory is able to act on. In addition, it is predicted that after being 27 brought to criterion, JY will produce more false recognitions than young and old adults, since her source monitoring ability is likely to be worse (for a review on source memory in amnesia, see Johnson et al, 1993) and it is also unlikely that her item-specific memory will be able to be improved to the same level. Method: Materials and tasks: False recognition and free recall tests: There were two false recognition tests (A and B) and two free recall tasks and these were exactly the same as those describe in Experiment 1. Criterion lists: JY was brought to criterion on all of the word lists used in false recognition test B. The criterion lists were thus the same lists as those used in that paradigm (see appendix A, list B for a table of the lists used). There were twelve lists and each list consisted of fifteen words. Each word was printed on an A5 sized flashcard. Procedure: The full experiment was divided into two sections as it was very time consuming. JY was told that most of the experiment would take place on the computer and that in addition there would be a few tests carried out in person, and the general procedure for the false recognition tests, free-recall tests and the criterion phase was described to her. In the first section, she carried out false recognition test A. She then had a half hour break before starting the second section. JY began the second section by being brought to criterion on all twelve of the criterion lists. The procedure for this was as 28 follows: For each of the lists, JY was instructed to try to remember as many words as possible. The list was then presented, one word at a time and for approximately 2 seconds each word. After the last word was presented, she was told to recall as many of the words as possible in any order. The number of words correctly recalled was recorded on a score sheet. The same list was then presented again in exactly the same manner, and again the number of words correctly recalled was recorded. This process was repeated with the same list until either JY recalled 7 or more items correctly, or she failed to increase her score from the previous trial. This whole procedure was repeated for each of the twelve lists. Once JY had been brought to criterion on all of the lists, she completed false recognition test B. Results: The same correction procedures used in experiment 1 were applied to JY’s results and for the same reasons as before, it is the A’ analysis which is presented here. However, as in the first experiment, any differences in results between the A’ and SC procedures are mentioned at the end of the results section. Table 4 shows the mean percentages of A’ critical lures, A’ studied words and unrelated words endorsed by JY in each experimental condition, and this pattern of results is illustrated in Figure 6. After being brought to criterion, she produced more A’ critical lures (mean = 92.95) than in the control condition (mean = 64.29), more A’ studied words (mean = 96.69) than in the control condition (mean = 70.09), and less unrelated words (mean = 8.33) than in the control condition (mean = 41.67). 29 Control Criterion A' Critical A' Studied Unrelated A' Critical A' Studied Unrelated lures words words lures words words mean 64.29 70.09 41.67 92.95 96.69 8.33 Table 4: mean percentages of A’ critical lures, A’ studied words and unrelated words endorsed by patient JY in the control condition and after being brought to criterion. 100.00 90.00 80.00 % Words Endorsed 70.00 60.00 A' Critical Lures 50.00 A' Studied Words 40.00 Unrelated Words 30.00 20.00 10.00 0.00 Control Criterion Experimental Condition Figure 6: graph to illustrate the pattern of means of A’ critical lures, A’ studied words and unrelated words endorsed by JY in the control condition and after being brought to criterion. JY’s results were compared to those of healthy adults using a significance test devised by Crawford and Howell (1998). First, JY’s results from the control condition were compared to those of healthy young and old adults in the control condition. It was found that JY produced less critical lures than both young and old adults (young: t=- 2.122, df=31 p<0.05; old: t=-4.057, df=32, p<0.01), more unrelated lures than both young and old adults (young: t=3.591, df=31, p<0.01; old: t=3.258, df=32, p<0.01), and with regard to A’ studied words, while she did not produce significantly less than either young or old adults there was a trend towards significance in both cases (young: t=-1.857, df=31, p=0.073; old: t=-1.869, df=32, p=0.071). After being brought to 30 criterion, JY’s “criterion” results were compared to healthy adults in the control condition, and there was no significant difference in A’ critical lures, A’ studied words or unrelated words between JY and healthy adults. Differences between the signal detection analysis (A’) and the standard correction (SC) procedures for the current results. When JY’s results from the control condition were compared to those of healthy adults in the control condition, the SC analysis failed to produce the significant difference that the A’ analysis did between JY and young adults with respect to critical lures. With regard to the studied words, the only difference was that the SC analysis produced a significant difference between JY and old adults (t=-3.197, df=32, p<0.01), while the A’ analysis only found a trend towards a significant result. When JY’s results in the control condition were compared to healthy young and old adults in the AS condition, the SC analysis failed to produce a significant difference between JY and these groups with respect to critical lures and studied words, unlike the A’ analysis. However, there were trends towards significance in the old adults with respect to both critical lures (t=-1.755, df=33, p=0.089) and studied words (t=- 1.862, df=32, p=0.072). All other results were the same, and as in experiment 1, the differences could be due to the large amount of variance found in SC data, as compared to that found in the A’ data. 31 Discussion: The main objective of experiment 2 was to see if improving the item-specific memory of patient JY would increase her ability to gist, and the results suggest that it did. After being brought to criterion, JY showed clear signs of increased gisting as she produced more critical lures, more studied words and less unrelated words than in the control condition. The results also suggest that, in line with predictions, JY was less able to gist than young and old adults when they were all in the control conditions, since she produced less critical lures and more unrelated words, and also showed trends towards producing less studied words. This is in keeping with the findings of previous studies which show that amnesic patients produce less false recognitions than young and old adults (Schacter, Verfaellie et al., 1997; Norman and Schacter 1997). After being brought to criterion, it was predicted that JY would produce more false recognitions than young and old adults since, though she should be gisting at roughly the same level, her source monitoring ability should still be considerably worse (for a review, see Johnson et al., 1993). However, the results show that there were no differences between JY and young and old adults with respect to critical lures, studied words or unrelated words, and thus imply that JY was able to gist at the same level as both young and old adults. A possible reason for this is that, in bringing her to criterion, seeing the lists several times might have strengthened her memory for the source of the words, and this could have allowed her to exclude some of the false gist representations in the same way that healthy adults do. Another interpretation refers to the theory expressed in the discussion in Experiment 1 about the interaction between item-specific memory and gist memory. It could be that JY’s item-specific memory was not increased to as high a level as that of the young and old adults, and therefore although they might have been able to override more false gist 32 representations than her, their gist memory would also have been stronger, and hence the gist representations they experienced might have been greater in number and strength. Hence despite any differences that might have existed between JY, young adults and old adults with respect to gist memory, item-specific memory and their interaction, the observable outcome (ie. number of falsely recognised critical lures) could still be the same. These results demonstrate that JY showed signs of healthy gisting when her item- specific memory was increased, but signs of impaired gisting in the control condition when her item-specific memory was very low. This implies that her gist memory was in tact (since it worked fine when her item-specific memory was increased), despite the fact that she produced less false recognitions than young and old adults in the control condition. This stands in stark contrast to the conclusions drawn by previous studies (eg. Budson et al., 2000, Gallo et al., 2006) which claim that a damaged gist memory is precisely the reason behind why amnesic patients produce less false recognitions than healthy controls. Although this is only a case study, and one on a patient whose amnesic syndrome is different from those studied previously, it still points towards a possible error in the previous studies’ assumptions that gist memory is damaged in amnesic patients. This is because the pattern of results found when JY was in the control condition was exactly the same as that found in various experiments on different amnesic patients (eg. Korsakoff’s patients: Schacter, Verfaellie et al.,1997; AD patients: Budson et al, 2000; patients with mixed etiologies: Schacter et al., 1996), and since it is these results that have led to assumptions of impaired gist memory, it could be that that the assumptions have not adequately accounted for the effect of item-specific memory in the results. On a more 33 theoretical level, these results support the theory put forward by this paper that there is an interaction between gist memory and item-specific memory, since when JY’s item-specific memory was increased she showed signs of improved gisting. More specifically, the results are also in line with the idea of a threshold whereby one cannot gist effectively if their item-specific memory is below a certain level, since in the control condition when JY’s item-specific memory was very weak, she produced significantly less critical lures than the older adults. There is a potential problem, however, with the suggestion that the increased number of critical lures in the criterion condition is due to JY’s increased item-specific memory. JY was brought to criterion on the various study-lists in the false recognition paradigm by presenting them to her several times. While it is true that this might have increased her item-specific memory for the individual words on the lists, it might also have increased her gist memory for the lists in general. A repeated trials experiment by Budson et al. (2000) found similar results, in that the number of critical lures and studied words recognised by AD patients increased over trials, but these researchers argued that this increase was due to the AD patients’ damaged gist memory improving with the repetition. Since the method for bringing JY to criterion did not isolate her item-specific memory from her gist memory, this argument could theoretically apply to her results as well. However, the experiment by Budson et al. (2000) did not isolate AD patient’s item-specific memory from their gist memory either, and thus the explanation given for JY’s results (that the increase in critical lures is due to an increase in item-specific memory) could just as easily apply to the results of Budson et al’s study. Indeed, the fact that the number of studied words increased over trials in Budson et al.’s study suggests that AD patients’ item specific 34 memory was increasing, even if their gist memory was as well, so the possibility that this caused the increase in critical lures should at least be acknowledged. In sum, it seems likely that when study-lists are presented to amnesic patients several times, it results in an increase in both item-specific and gist memory, but a question remains over whether it is the increase in item-specific memory that causes the increase in gist memory. Neither the present experiment or previous experiments (Budson et al, 2000) manage to answer this question effectively as they both fail to isolate the effects of item-specific memory and gist memory on false recognitions. However, it is important to recognise the fact that an increasing item-specific memory could account for the results of previous experiments (eg. Schacter et al., 1996; Budson et al., 2000; Gallo et al., 2006) and thus that their assumptions that gist memory is damaged in amnesic patients are not conclusive. 35 General Discussion: Previous studies using false recognition paradigms such as this one have found that amnesic patients falsely recognise fewer critical lures than healthy adults (Schacter et al., 1996; Budson et al., 2000), and it has been concluded that this is due to a damaged gist memory in the patient group (eg. Budson et al., 2000; Gallo et al., 2006). However, it may instead be that this pattern of results is due to the severely impaired item-specific memory of the amnesic patients, which may be too low for their gist memory to act on. In other words, it might be that their ability to gist is in tact and it is rather their access to it which is impaired. This was the initial theory proposed by the current paper and it is supported, to an extent, by the results from Experiments 1 and 2. Experiment 1 supports the theory by demonstrating that reducing the item- specific memory of healthy old adults produced a similar pattern of results to that shown by amnesic patients, as healthy old adults produced less critical lures in the AS condition than in the control condition. Since gist memory in these adults was in tact, the results suggest that reduced recognition of critical lures does not necessarily reflect a damaged gist memory, but could instead be a consequence of a weak item- specific memory. Experiment 2 supports the theory by demonstrating that when the item-specific memory of amnesic patient JY was increased, she was able to gist as well as healthy adults, which implies that her gist memory was in tact. However, she produced less critical lures than healthy adults when in the control condition (where her item-specific memory was very weak), and this again suggests that a reduction in critical lures does not necessarily reflect a damaged gist memory but might rather be due to a weak item-specific memory. 36 However, there were several inconsistencies in the results of both experiments and these raised some interesting questions about the relationship between item-specific memory and gist memory. These questions were addressed in the discussion sections of each experiment and a theory about this relationship was suggested in order to account for the results. This theory is similar to that put forward in the introduction, in that it maintains that an increased item-specific memory might both enhance one’s gist memory and suppress it at the same time. However, before it was assumed that an increase in item-specific memory would suppress gist memory more than it would enhance it, as this explained why old adults produced more false recognitions than young adults, but the evidence from Experiment 1 and others (Kensinger and Schacter, 1999; Budson et al., 2000) that old adults do not produce more false recognitions than young adults, suggest that this is not always the case. Gist memory exists in its own right, and hence while it is heavily influenced by item-specific memory, in being strengthened or suppressed, this is not the only factor that determines how much or how well someone gists. Hence, individual differences in gist memory could mean that two people with the same item-specific memory might produce different numbers of critical lures. On top of this, it is still assumed that a certain level of item-specific memory is required in order to for the gist memory to act on, and without this, one cannot gist effectively. So in summary, the theory states that in the first place, item specific memory acts a gateway to gist memory, whereby a certain level of item-specific memory is needed to allow gist memory to become active. After this point, however, the situation becomes very complex, since an increase in item-specific memory both strengthens gist memory and suppresses it, and this can happen to different extents depending on the gist memory of the individual and any other factors which might influence it. 37 This theory accounts for the findings that amnesic patients recognise less critical lures than healthy adults (eg. Schacter et al., 1996; Budson et al.), since the item-specific memory of these patients is severely impaired and could therefore be too low to gist effectively. It therefore challenges the assumption that amnesic patients have an impaired gist memory, since the results that led to this assumption can be explained differently. Other studies use “semantic categorisation” to test for gist memory in a different way (Koustaal et al., 1999; 2001; 2003; Budson, 2006). These experiments ask participants whether they recognise items as being part of a particular category, and false recognitions in these paradigms are seen as a better measure of gist as the possibility of source monitoring errors is greatly reduced. The results of amnesic patients in these studies show the same pattern as before, whereby they falsely categorise more items than healthy adults, and again this result has been used to infer a degraded gist memory in amnesia (Koustaal et al. 2001; Budson, 2006). However, the theory proposed by this paper can account for these results in the same way that it did for the false recognition results of amnesic patients. It could be that the amnesic patients’ impaired item-specific memory for the pictures prevents them from gisting effectively, and hence they produce fewer false categorisations. Therefore the results of semantic categorisation experiments do not necessarily point to a damaged gist memory either. Other experiments have explored ways to reduce false recognitions, such as with the use of a distinctiveness heuristic (eg. Israel and Schacter, 1997), or through repetition of study-test trials (Budson et al., 2000). The distinctiveness heuristic is a method of making items more distinctive, and experiments find that this reduced the number of 38 falsely recognised critical lures in young and old adults (Schacter et al., 1999; Dodson and Schacter, 2002), but not in AD patients, who even showed a slight increase in false recognitions (Budson et al., 2002). Again this result has been used to infer that the patients have a damaged gist memory, but again the theory proposed by this paper can account for these results in a different way. It seems that increasing the distinctiveness of items improved one’s item-specific memory for them, which enabled healthy adults to suppress more false gist representations and hence recognise fewer critical lures. While the interaction between item-specific and gist memory advocated by this paper is thought to be complex and perhaps unpredictable, this possibility is still accounted for. The theory can also account for the results of the AD patients, by positing that while the distinctiveness heuristic may have improved their item-specific memory to a degree, this may still not have been strong enough to suppress their false gist memories. This would even explain the slight increase in false recognitions found in Budson et al.’s experiment (2002). A similar explanation can be used to account for the results of the repeated trials experiment (Budson et al. 2000). These showed that that repetition of the study-test procedure reduced the number of false recognitions in healthy young adults, had no effect on the number of false recognitions in healthy old adults, and increased the number of false recognitions in AD patients. It was inferred that the repetition improved the item- specific memory of healthy young adults, which enabled them to suppress their false gist memories, but did not affect the item-specific memory of the old adults. The increase in false recognitions observed in the AD patients was explained in terms of an impaired gist memory that improved with the repetition of trials. The theory posited by this paper can explain these results differently, and without the assumption of a damaged gist memory in the AD patients. It would suggest that, in all the groups, 39 item-specific memory and gist memory was improved with the repetition of trials, but to different levels. The item-specific memory of the young adults was improved the most, and to a level that could substantially suppress even their improved gist memory. The old adults’ item specific memory increased to a level that suppressed their improved gist memory enough to prevent an increase in false recognitions but not enough to cause a reduction. Finally, the AD patients’ item-specific memory was increased to a level that allowed access their gist memory but it was not high enough to suppress the false gist representations now available to them. It would appear, therefore, that this theory can account for many of the results of different false recognition experiments. The current experiment supports this theory and hence provides evidence that previous studies might be mistaken in their assumptions that gist memory is damaged in amnesic patients, since this theory can explain their results without reference to a damaged gist memory. However, while this experiment suggests that gist memory might be in tact in amnesic patients, it does not offer any evidence that gist memory actually is in tact in these patients. Experiment 2 attempts to demonstrate this, but ultimately falls into the same trap as other false recognition experiments, as it fails to isolate gist memory from item- specific memory. However, along with Experiment 1, it does demonstrate the possibility that item-specific memory plays an influential role in one’s ability to gist. Indeed if the theory suggested in this paper is correct, it would seem that there is an extremely complex and interactive relationship between the two types of memory. Hence, if false recognition tests are to be a useful means of assessing one’s ability to gist, they need to find a way to account for this intricate relationship and somehow isolate the effects of item-specific memory from those of gist memory. However, this 40 would seem to be impossible, since these tests of false recognition all require participants to explicitly study related lists, and in doing so their item-specific and gist memories become immediately intertwined. Hence, the results of these tests can only truly reflect the relationship between item-specific memory and gist memory, and cannot accurately asses the effects of either on its own. Since this relationship appears to be highly complicated, variable and also unpredictable at times, it is suggested that false recognitions are not particularly beneficial as a measure of gist memory, and thus that there is a need to find new methods of doing so. One possible approach would be to bypass item-specific memory entirely and attempt to access gist memory via implicit memory. This suggestion is influenced by Balota et al’s work on semantic memory in amnesic patients (Balota et al., 1999, Watson et al., 2001.), and following their lead, using semantic-priming exercises to assess gist memory could be a possible avenue for a new approach. 41 References Balota, D.A., Cortese, M.J., Duchek, J.M., Adams, D., Roediger, H.L., McDermott, K.B., & Yerys, B.E. (1999). Veridical and false memories in healthy older adults and in dementia of the Alzheimer’s type. Cognitive Neuropsychology, 16, 361–384. Balota, D.A., & Duchek, J.M. (1991). Semantic priming effects, lexical repetition effects, and contextual disambiguation effects in healthy aged individuals and individuals with senile dementia of the Alzheimer’s type. Brain and Language, 40, 181–201. Banks, W.P. (1970). Signal detection theory and human memory. Psychological Bulletin, 74, 81-99. Beaman, C.P. & Jones, D.M. (1997). Role of Serial Order in the Irrelevant Speech Effect: Tests of the Changing-State Hypothesis. Journal of Experimental Psychology: Learning, Memory, and Cognition, 23, 459-471. Bellamy, K.J. & Shillcock, R. (2007). A right hemisphere bias towards false memory. Laterality, 2007, 154-166. Budson, A.E., Daffner, K.R., Desikan, R., & Schacter, D.L. (2000). When false recognition is unopposed by true recognition: Gist-based memory distortion in Alzheimer’s disease. Neuropsychology, 14, 277–287. 42 Budson, A.E., Desikan, R., Daffner, K.R., & Schacter, D.L. (2001). Perceptual false recognition in Alzheimer’s disease. Neuropsychology, 15, 230–243. Budson, A.E., Daffner, K.R., Sitarski, J., & Schacter, D.L (2002). False Recognition of Pictures Versus Words in Alzheimer’s disease: the Distinctiveness Heuristic. Neuropsychology, 16, 163-173. Budson, A.E., Todman, R.W., & Schacter, D.L. (2006). Gist memory in Alzheimer’s disease: evidence from categorized pictures. Neuropsychology, 20, 113-122. Cermak, L.S., Butters, N., & Gerrein, J. (1973). The extent of the verbal encoding ability of Korsakoff patients. Neuropsychologia, 11, 85–94. Crawford, J.R. & Howell, D.C. (1998). Comparing an individual's test score against norms derived from small samples. The Clinical Neuropsychologist, 12, 482-486. Dalla Barba, G., Nedjam, Z., & Dubois, B. (1999). Confabulation, executive functions, and source memory in Alzheimer’s disease. Cognitive Neuropsychology, 16, 385–398. Dehon, H. and Bredart, S. (2004). False memories: young and older adults think of semantic associates at the same rate but young adults are more successful at source monitoring. Psychology and aging, 19, 191-197. 43 Dodson, C.S, & Schacter, D.L. (2002). Aging and strategic retrieval processes: reducing false memories with a distinctiveness heuristic. Psychology and aging, 17, 405-415. Dywan, J., & Jacoby, L. (1990). Effects of aging on source monitoring: Differences in susceptibility to false fame. Psychology and Aging, 5, 379-387. Faust, M.E., Balota, D.A., Duchek, J.M., Gernsbacher, M.A., & Smith, S. (1997). Inhibitory control during sentence comprehension in individuals with dementia of the Alzheimer type. Brain and Language, 57, 225-253. Gallo, D.A., Shahid, K.R., Olson, M.A., Solomon, T.M., Schacter, D.L., & Budson, A.E. (2006). Overdependence on degraded gist memory in Alzheimer’s disease. Neuropsychology, 20, 625–632. Gollan, T.H., Salmon, D.P., & Paxton, J.L. (2006). Word association in early Alzheimer’s disease. Brain and Language, 99, 289–303. Hartley, A. A. (1992). Attention. In F.I.M. Craik & T.A. Salthouse (Eds.), The handbook of aging and cognition (pp. 3– 49). Hillsdale, NJ: Erlbaum. Israel, L., & Schacter, D.L. (1997). Pictorial encoding reduces false recognition of semantic associates. Psychonomic Bulletin & Review, 4, 577–581. Johnson, M.K., Hashtroudi, S., & Lindsay D.S. (1993). Source Monitoring, Psychological Bulletin, 114, 3-28. 44 Kensinger, E.A., & Schacter, D.L. (1999). When true memories suppress false memories: effects of aging. Cognitive Neuropsychology, 16, 399–415. Koutstaal, W., Schacter, D.L., Galluccio, L., & Stofer, K.A. (1999). Reducing Gist- Based False Recognition in Older Adults: Encoding and Retrieval Manipulations. Psychology and Aging, 14, 220-237. Koustaal, W., Schacter, D.L., & Verfaille, M. (2001). Recognizing identical versus similar categorically related common objects: further evidence for degraded gist representations in amnesia. Neuropsychology, 15, 268-289. Koustaal, W., Reddy, C., Jackson, E.M., Prince, S., Cendan, D.L., & Schacter, D.L. (2003). False recognition of abstract versus common objects in older and younger adults: testing the semantic categorization account. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 499-510. McDowd, J. M., & Shaw, R. J. (2000). Attention and aging: A functional perspective. In F.I.M. Craik & T.A. Salthouse (Eds.), The handbook of aging and cognition (pp. 221–292). Hillsdale, NJ: Erlbaum. Mclntyre, J.S., & Craik, F.I.M. (1987). Age differences in memory for item and source information. Canadian Journal of Psychology, 41, 175-192. Moss, H. & Older, L. (1996) Birkbeck word association norms. Hove: Psychology Press. 45 Multhaup, K.S., & Balota, D.A. (1997). Generation effects and source memory in healthy older adults and in adults with dementia of the Alzheimer type. Neuropsychology, 11, 382–391. Nebes, R. D. (1992). Cognitive dysfunction in Alzheimer's disease. In F.I.M. Craik & T.A. Salthouse (Eds.), The handbook of aging and cognition (pp. 373-446). Hillsdale, NJ: Erlbaum. Nelson, H.E., & Willison, J. (1991). National Adult Reading Test (NART): Test Manual. Second Edition. Windsor, UK: NFER Nelson. Norman, K., & Schacter, D.L. (1997). False recognition in younger and older adults: Exploring the characteristics of illusory memories. Memory & Cognition, 25, 838– 848. Ober, B. A., & Shenaut, G. K. (1995). Semantic priming in Alzheimer's disease: Meta-analysis and theoretical evaluation. In P. A. Allen & T. R. Bashore (Eds.), Age differences in word and language processing, (pp. 247-271). Amsterdam: Elsevier. Pierce, B.H., Sullivan, A.L., Schacter, D.L., & Budson, A.E. (2005). Comparing source based and gist based false recognition in aging and Alzheimer’s disease. Neuropsychology, 19, 411-419. Radvansky, G. (2006). Human memory, pp377-378. Boston: Allyn and Bacon. 46 Reyna, V.F., & Brainerd, C.J. (1995). Fuzzy-trace theory: An interim synthesis. Learning and Individual Differences, 7, 1–75. Roediger, H.L., & McDermott, K.B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 803–814. Schacter, D.L., Harbluk, J.L., & McLachlan, D.R. (1984). Retrieval without recollection: An experimental analysis of source amnesia. Journal of Verbal Learning and Verbal Behavior, 23, 593-611. Schacter, D.L., Verfaille, M., & Pradere, D. (1996). The neuropsychology of memory illusions: false recall and recognition in amnesic patients. Journal of Memory and Language, 35, 319–334. Schacter, D.L., Koutstaal, W., & Norman, K.A. (1997). False memories and aging. Trends in Cognitive Sciences, 1, 229–236. Schacter, D.L., Verfaellie, M., & Anes, M.D. (1997). Illusory memories in amnesic patients: conceptual and perceptual false recognition. Neuropsychology, 11, 331–342. Schacter, D.L., Israel, L., & Racine, C. (1999). Suppressing false recognition in younger and older adults: The distinctiveness heuristic. Journal of Memory and Language, 40, 1–24. 47 Simone, P.M., & Baylis, G.C. (1997a). The role of attention in a spatial memory task in Alzheimer disease patients. Alzheimer Disease and Associated Disorders, 11, 140- 152. Simone, P. M., & Baylis, G. C. (1997b). Sustained attention in a reaching task: Effects of normal aging and Alzheimer's disease. Journal of Experimental Psychology: Human Perception and Performance, 23, 595-608. Simons, J.S., Verfaellie, M., Hodges, J.R., Lee, A.C.H., Graham, K.S., Koutstaal, W., Schacter, D.L. (2005). Failing to get the gist: reduced false recognition of semantic associates in semantic dementia. Neuropsychology, 19, 353–361. Smith, J.A., & Knight, R.G. (2002). Memory processing in Alzheimer’s disease. Neuropsychologia, 40, 666–682. Spieler, D.H., Balota, D.A., & Faust, M.E. (1996). Stroop performance in younger adults, healthy older adults, and individuals with senile dementia of the Alzheimer's type. Journal of Experimental Psychology: Human Perception and Performance, 22, 461-479. Troster, A.I., Salmon, D.P., McCullough, D., & Butters, N. (1989). A comparison of category fluency deficits associated with Alzheimer's and Huntington's disease. Brain and Language, 37, 500-513. 48 Tulving, E. (1972). Episodic and semantic memory. In E Tulving & W Donaldson (Eds), Organization of memory, 381-403, New York Academic Press. Tun, P.A., Wingfield, A., Rosen, M.J., & Blanchard, L. (1998). Response Latencies for False Memories: Gist-Based Processes in Normal Aging, Psychology and Aging, 13, 230-241. Underwood, B.J. (1965). False recognition produced by implicit verbal responses. Journal of Experimental Psychology, 70, 122–129. Verfaellie, M., Schacter, D. L., & Cook, S. P. (2002). The effect of retrieval instructions on false recognition: Exploring the nature of the gist memory impairment in amnesia. Neuropsychologia, 40, 2360–2368. Verfaellie, M., Rapcsak, S.Z., Keane, M.M., Alexander, M.P. (2004). Elevated false recognition in patients with frontal lobe damage is neither a general nor a unitary phenomenon. Neuropsychology, 18, 94–103. Watson, M.J, Balota, D.A., & Sergent-Marshall, S.D. (2001). Semantic, phonological, and hybrid veridical and false Memories in healthy older adults and in individuals with dementia of the Alzheimer type. Neuropsychology, 15, 254-267. 49 APPENDIX A Paradigm A Paradigm B Anger Fruit Bread Cold Doctor King Black Chair Foot Galaxy High Man rage apple butter hot nurse queen white table shoe stars low woman fear basket jam ice hospital royal dark legs ball universe sky boy hate tree board winter ill ruler cat stool mouth planet tall uncle fury juice sandwich wet injection prince blue seat toe bar tower person red pear flour freeze health crown funeral back ankle space airplane wife temper ripe milk snow stethoscope England colour desk sock cosmos altitude male violence salad yeast frozen patient palace grief wood sole infinite flying father wrath banana dough chilly prescription throne green sofa walk Milky-Way kite strong fight strawberry crust heat pills chess death cushion smell black hole rise friend chaos orange roll weather treatment sovereign ink sitting boot nebula far beard hatred dessert slice fridge office subjects bottom swivel run constellation vertigo being mean vegetables wine air medical monarch coal furniture sore satellite hopes handsome emotion bowl loaf shiver surgeon castle brown arm step moon giant muscle shouting cocktail toast Arctic clinic leader raven rocking odour sun lofty suit enrage berry bap frost cure reign grey bench hand asteroid mighty old Mountain Needle Rough Slow Spider Sweet River Sleep Soft Thief Window Music hill thread smooth fast web sour water dreams hard steal pane note steep pin ready down insect sugar stream bed warm robber glass sound climb eye ground quick fly tooth lake night comfort crook ledge pop summit sewing tough snail arachnid chocolate wide pillow feathers burglar sill score top sharp sandpaper stop crawl good boat awake cosy money curtain sheet molehill point stubble coach tarantula taste tide peace cuddly police frame stave peak prick surface delay poison sticky swim rest gentle bad house song plain thimble coarse traffic bite nice flow slumber touch law open book glacier haystack uneven tortoise creepy honey runs doze fluffy jail broken stereo goat thorn justice hesitant animal syrup barge tired furry criminal closed singing bike hurt rugged speed ugly toffee creek snore downy villain view guitar climber sting cut bus feelers heart brook nap kitten crime breeze record range stitch bark sluggish small cake fish nightmares skin bank sash piano valley cloth rocky wait nasty wrapper bridge yawn tender dishonest soul tune ski knitting gravel idle eerie pie winding drowsy snug pillage shutter orchestra Appendix A: table showing all the words and lists used in the false recognition paradigms A and B. The critical lure for the list is written in bold above the actual list that was presented. APPENDIX B LIST A LIST B Labour Candle Gravy Famish Hamster Blot Astonish Scent Daffodil Halt Refuge Tool Pamphlet Zoo Swamp Novice Toothbrush Chance Vacant Grunt Boast Wallpaper Connect Bacteria Fashion Camera Abrupt Unique Urban Major Appendix B: table showing the words used in the two free recall lists 50