حافظه اپیزودیک و ساخت رویداد در افزایش سن و فراموشی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|33670||2012||15 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Memory and Language, Volume 67, Issue 2, August 2012, Pages 270–284
Construction of imaginative or fictitious events requires the flexible recombination of stored information into novel representations. How this process is accomplished is not understood fully. To address this problem, older adults (mean age = 74.2; Experiment 1) and younger patients with MTL lesions (mean age = 54.2; Experiment 2), both of whom have deficient LTM compared to their respective controls, were given a setting (e.g. jungle) and 3–6 words (e.g. tiger, tree, snake) and asked to imagine an event in that setting by relating the words to each other. Both older adults and patients showed deficits in forming coherent mental representations relative to younger adult and healthy control groups, respectively. Moreover, the ability to form coherent events was associated with subsequent memory for the items. These findings suggest that deficits in LTM, or processes mediating it, are one factor that affects event construction, which in turn leads to poorer encoding and/or retention of the studied materials. These results have implications for theories of the cognitive processes underlying the construction of imaginative events in the laboratory and everyday life.
The functional utility of a memory system that can obligatorily store and retrieve unique events consciously must go beyond mere retrospection: as the Queen in Alice’s Adventures in Wonderland noted, “It’s a poor sort of memory that only works backwards” ( Carroll, 1886, p. 56). Confirming the Queen’s pronouncement, an emerging body of evidence suggests that LTM interacts with, and contributes to, various cognitive tasks, such as problem solving ( Chen et al., 2004 and Sheldon et al., 2011) and imagination ( Hassabis et al., 2007a and Hassabis et al., 2007b). Of particular interest is imagination, which involves construction of mental representations of novel events, whether deliberately (i.e. prospection) or inadvertently (i.e. daydreaming) ( Delaney et al., 2010, Hassabis and Maguire, 2009 and Pillemer, 2003). This ability to construct a novel mental representation has been posited as a means by which humans use memory to guide decision-making and subsequent behavior: specifically, generating possible future outcomes may allow us to pre-experience the consequences of choices before they happen, thus giving useful feedback provided such representations are accurate approximations of real-life ( Atance and O’Neilll, 2001, Benoit et al., 2011, Boyer, 2008, Buckner and Carroll, 2007, Gilbert and Wilson, 2007, Peters and Büchel, 2010 and Schacter and Addis, 2007). The processes that govern imagination are beginning to be elucidated. It is clear that imagining a novel event depends, in part, on retrieving relevant information from episodic memories of similar experiences and their concomitant neural substrates. Evidence from studies of patients with brain lesions (Addis et al., 2009, Hassabis et al., 2007a, Hassabis et al., 2007b and Rosenbaum et al., 2009), of functional neuroimaging of healthy people (Addis et al., 2009, Addis and Schacter, 2008, Addis et al., 2007, Addis et al., 2009, Hassabis et al., 2007a, Hassabis et al., 2007b, Okuda et al., 2003 and Spreng et al., 2009) and even of electrophysiological studies in rats (Johnson & Redish, 2007), have suggested that the hippocampus, a structure long-known to be necessary for formation and retrieval of episodic memories in long-term memory (LTM; Scoville & Milner, 1957), is also implicated during construction of imaginary or anticipated events. Other structures that form part of the autobiographical memory and default network, such as the medial parietal and medial prefrontal cortex, are also recruited (Buckner et al., 2008 and Spreng et al., 2009). This evidence has led investigators to propose that one function of LTM, mediated by the hippocampus and related structures, is to supply elements from long-term episodic memory that are needed for event construction (Buckner and Carroll, 2007, Hassabis and Maguire, 2007, Hassabis and Maguire, 2009, Moscovitch, 2008 and Schacter and Addis, 2007). In addition to retrieval, another key aspect to imagining a novel event is the actual construction of the mental representation itself (i.e. event construction). Because an imaginary event has not been previously experienced, it cannot be evoked in its entirety merely by retrieving items from memory. To imagine a novel, coherent event, these items must be recombined or reordered in new ways (Rosenbaum et al., 2009, Schacter and Addis, 2007 and Suddendorf and Corballis, 2007), and it is presumably the coherence of a constructed event (or lack thereof) that would dictate whether imagined items in consciousness are perceived as a unified scene/event, or merely unrelated mental images (Addis and Schacter, 2012, Blumenfeld et al., 2010, Hassabis et al., 2007a, Hassabis et al., 2007b and Hassabis and Maguire, 2009). Furthermore, given that one property of episodic memory is that information is encoded in a manner that it may be flexibly recombined, and that episodic memory retrieval is a reconstructive process, it is reasonable to expect that constructing/recombining information into a coherent mental representation is an important aspect of imagination (Bartlett, 1932, Eichenbaum and Cohen, 2001, Martin et al., 2011, Morris et al., 1977, Roediger and McDermott, 1995, Schacter and Addis, 2007 and Schacter et al., 1998). The processes that govern event construction, however, are still poorly understood. To date, most imagination studies have used an open-ended cueing paradigm, emphasizing the creation of detailed imagined scenes/events: such a task would require both retrieving episodic and semantic elements from LTM in response to a general cue (e.g. imagine a beach scene), and then constructing the imagined event from those elements. Consequently, variations in task performance may be due to differences in the ability to search memory and retrieve the requisite elements from LTM (e.g. umbrella, beach balls, people playing volleyball, etc.), and/or from differences in recombining and binding of retrieved elements into a coherent representation. Some evidence suggests that imagination performance depends partially on event construction ability. Studies of patients with hippocampal lesions have shown that in addition to being sparsely detailed, the imagined scenes produced by patients are also rated as less-coherent by the patients themselves and by raters (Hassabis et al., 2007a, Hassabis et al., 2007b and Rosenbaum et al., 2009; though see Maguire et al., 2010 and Squire et al., 2011). It is not clear, however, whether the patients’ deficit in retrieving details from memory precluded them from constructing spatially coherent scenes, or whether two separate deficits exist. Indeed, in a recent review, Addis and Schacter (2012) identify initial retrieval and elaboration as two of the many process involved in imagination, noting that less is known about the processes and neural substrates governing event constructions. Drawing on Hassabis et al.’s findings regarding the importance of coherence in scene construction, we reasoned that coherence may also be implicated in event construction. Evidence from humans with medial temporal lesions and fMRI in healthy controls suggests that the hippocampus is implicated in a variety of processes, such as transitive inference (Preston, Shrager, Dudukovic, & Gabrieli, 2004), maintaining continuity across discourse (Duff, Gupta, Hengst, Tranel, & Cohen, 2011; Duff Hengst, Tranel, & Cohen, 2009) and story-telling (Schmitter-Edgecombe & Creamer, 2010; Rosenbaum et al., 2009) all of which would suggest a role for the hippocampus in constructing coherent events. Recent evidence has indicated that aging is also associated with deficits in successfully incorporating specific event details (i.e. ‘person, place, and object’) during imagination tasks (Addis, Musicaro, Pan, & Schacter, 2010). When given three details to incorporate into an imagined event, older adults showed a deficit in integrating all the details within one time period. However, it remains unclear whether this was due to poor construction of an event per se, or due to other age-related cognitive changes, such as decreased monitoring ability or manipulation within working memory (Osaka, Logie, & D’Esposito, 2007; Petrides, 2005). A related question is whether there are consequences for how well an imaginary event is initially constructed: Specifically, how does the coherence of a constructed event relate to its subsequent memory? One might predict that more coherent imagined events would be remembered better than less coherent ones, perhaps because of more elaborative encoding and organization that boosts recall for those items (Bower, 1970, Craik and Lockhart, 1972 and Staresina et al., 2009). This interpretation is supported by a recent finding that hippocampal activation for an imagined event predicts subsequent memory for it (Martin et al., 2011) but no-one has shown more directly that the coherence of the imagined event is a contributing factor. Taking these concerns into account, we constructed a novel task similar to those of Summerfield et al., 2010 and Addis et al., 2010 that separates retrieval of the elements of the event from the construction process itself. By testing older adults with episodic memory loss presumably caused by medial temporal lobe deterioration, and amnesic people with confirmed MTL lesions, we hoped to gain insight into the contribution of episodic memory and the MTL to event construction. If deficits in event construction, and subsequent memory, in these populations are still evident when the elements comprising the event are provided, then the impairment cannot be ascribed simply to impaired retrieval of goal-relevant information, but must also include processes implicated in the construction and retention of these elements into a coherent narrative (Craik and Salthouse, 2000, Hasher et al., 1999, Mitchell et al., 2000 and Old and Naveh-Benjamin, 2008). In Experiment 1, we compared younger and older adults’ event construction and subsequent memory performance when the elements for constructing the event were provided rather than requiring the participants to retrieve the elements from LTM themselves (as in Summerfield et al., 2010). Participants were shown a “context” word such as “jungle” and 3–6 words such as “tiger, tree, snake” and were asked to imagine a novel event by relating the various objects that were named to one another within that imagined context. Providing names for the to-be-imagined events allowed participants to complete the task with minimal retrieval demands, so that performance would be based largely on event constructional ability, rather than differences in retrieving the required elements from LTM. In addition, by varying set size from three to six items, we hope to extend previous research on imagination by examining at what level of mnemonic load do event construction deficits appear, if such group differences exist. Given the recent debate as to whether the MTL are truly necessary for constructing novel events (See Addis & Schacter, 2012 for a review), in Experiment 2, we tested patients with confirmed MTL lesions and amnesia, who were otherwise cognitively intact. Doing so allowed us to determine directly whether MTL damage impairs the ability to construct coherent imagined events. Moreover, we could ascertain whether the predicted deficits in event construction in older adults are due in part to MTL-mediated episodic memory functions, one aspect of cognitive functioning that declines with age.