آگاهی حافظه اپیزودیک کاهش یافته در سالمندان: شواهد از احساس شناخت و تجدید خاطره

The ability to reflect on and monitor memory processes is one of the most investigated metamemory functions, and one of the important ways consciousnesses interacts with memory. The feeling-of-knowing (FOK) is one task used to evaluate individual’s capacity to monitor their memory. We examined this reflective function of metacognition in older adults. We explored the contribution of metacognition to episodic memory impairment, in relation to the idea that older adults show a reduction in memory awareness characteristic of episodic memory. A first experiment showed that age affects the accuracy of FOK when predictions are made on an episodic memory task but not on a semantic memory task, suggesting a particular role for episodic memory awareness in metacognitive evaluations. A second experiment showed that the age-difference in episodic FOK accuracy was removed if one took into account subjective reports of memory awareness, or recollection. We argue that the FOK deficit specific to episodic memory is based on a lack of memory awareness manifest as a recollection deficit.

Aging leads to a somewhat inevitable deterioration in episodic memory. One prevalent theory is that this memory dysfunction is connected to difficulties with the strategic regulation of memory, or metamemory (e.g. Shimamura, 1994). One such strategic activity is the awareness of memory function, or memory monitoring. The ability to monitor memory performance has considerable importance in everyday life and has been widely investigated previously (Berry et al., 1989, Dixon and Hertzog, 1988 and Perlmutter et al., 1987), especially with reference to age-related changes in episodic memory function (e.g. Connor et al., 1997 and Souchay et al., 2000). The rationale for such studies of metamemory is that episodic memory dysfunction is possibly caused, or at least contributed to, by a deficit in metamemory. A failure to monitor memory would mean that older adults were unable to compensate for their memory difficulties, or find it difficult to allocate cognitive resources efficiently to ensure adequate memory function. In studies of aging, one of the often overlooked features of episodic memory is recollective experience, the subjective state or memory awareness that separates episodic ‘remembering’ from semantic ‘knowing’ (Tulving, 1985). Comparisons of remembering and knowing, two forms of memory awareness have been studied neuropsychologically as well as experimentally (Tulving, 1985; see Gardiner & Richardson-Klavehn, 2000, for a review) and are proposed to map on to episodic and semantic memory respectively. Recollective experience, or remembering occurs when a rememberer has a sense or feeling of the self in the past (according to Tulving’s (1985) view, ‘autonoetic consciousness’). Images (often visual), feelings, thoughts and verbal statements directly related to the recalled episode also often come to mind during recollective remembering. A number of different models have been proposed which assume that retrieval in memory can be based on these two distinct forms of memory (see Yonelinas, 2002). (For an account of how these two states could be formed by differences in confidence, and thus rely on a single underlying memory process see Dunn (2004).) In the aging literature, findings from different paradigms provide converging evidence that aging disrupts recollection to a greater extent than familiarity (see, Yonelinas, 2002). For example, estimates from the process-dissociation procedure show that age leads to a decrease in recollection, but does not affect familiarity (Light et al., 2000 and Yonelinas, 2002). Recent evidence from signal detection approaches, and in particular, receiver operating characteristic (ROC) curves also confirms that age affects the recollection to a greater extent than familiarity (Healy, Light, & Chung, 2005). The same pattern of results has been observed using the Remember/Know procedure (e.g. Tulving, 1995) where participants report their subjective experience as either ‘remembering’ or ‘knowing,’ with the finding that older adults report less ‘remembering’ for items that they correctly recognise (e.g. Bastin and Van der Linden, 2003, Clarys et al., 2002, Comblain et al., 2004, Parkin and Walter, 1992, Perfect and Dasgupta, 1997 and Perfect et al., 1995). Our understanding of theory is that one can either be in a conscious state of remembering, or a state of knowing. Thus, our interpretation of this data is that when retrieving items from memory, older adults are less likely to be in the state of remembering. When we argue that memory awareness is diminished in aging, we therefore posit that older adults have fewer instances of recollection. Different types of information support these different remembering and knowing states. For example, ‘remember’ answers are associated with accurate retrieval of contextual information regarding the learning of the items, while this is not the case for knowing judgments (Perfect, Mayes, Downes, & Van Eijk, 1996). Thus, one might argue that older adults with a deficit in episodic memory show two simultaneous or accompanying deficits: a lack of the feeling of remembering, and a lack of contextual information produced at retrieval. When examining episodic memory deficits in older adults, most studies overlook this central role of subjective experience when defining episodic function, focussing more on the more objective criteria of recently learned (in an earlier experimenter-initiated study phase), versus previously learned (e.g. tests of general knowledge) to differentiate episodic and semantic memory, and measuring memory performance quantitatively, not with subjective reports of the quality of memory. However, when considering metacognition, it is very common to ask participants for reports of subjective experience. Given that episodic memory and metacognition are so often studied in aging, it seems a logical step to examine states of awareness and metacognition in parallel. Thus, in this paper, two experiments examine the relationship between metacognitive monitoring (i.e., the capacity to predict one’s own memory performance) and memory with and without recollective experience. In the first experiment, we compared monitoring in semantic memory (where recollection or ‘remembering’ is not required) and episodic memory (where it is required). In the second experiment, we took a more direct approach, considering metacognitive monitoring and remembering in the same experiment. The focus of this paper is Feeling-of-Knowing (FOK), in which predictions are made about the likelihood of subsequent recognition of currently non-recalled information (Hart, 1965 and Nelson and Narens, 1990). In this procedure, participants are asked to estimate the likelihood that they will recognize a piece of information they have failed to recall earlier, either from long-term knowledge or semantic memory (Hart, 1965 and Nelson and Narens, 1990), or from recently learned episodic memory information (Schacter, 1983 and Souchay et al., 2000). The implicit logic is to assume that the better one’s knowledge is about one’s memory, the more accurate the predictions are likely to be. There are several studies of FOK and aging, a typical finding being that younger and older adults do not differ in their ability to predict which items they will be able to recognise at least for semantic memory information (Allen-Burge and Storandt, 2000, Bäckman and Karlsson, 1985, Butterfield et al., 1988, Lachman et al., 1979 and Marquie and Huet, 2000). However, using an Episodic FOK metamemory task, Souchay et al. (2000) found a significant age effect on FOK accuracy for recently learned episodic information. Thus, as a whole, the data on FOK and aging suggest that an age-related decline in FOK accuracy may occur when judgments are made about episodic memory information but not when they are made about semantic memory information. The question of interest concerns the age effect observed specifically for episodic FOK accuracy. This suggests that episodic and semantic FOK judgements may be based on different aspects of subjective experience. According to Koriat (1993), FOK judgments are based on an inferential process that uses a variety of mnemonic cues to determine the likelihood that the target is retained in memory and will be recognized in the future. Thus, when participants fail to recall an answer, their FOK judgments are based on partial information accessed during the search for the target. Furthermore, FOK accuracy is based on the quality of the partial information retrieved (Koriat, 1993). Our novel hypothesis is that in an episodic memory task, this ‘partial information’ is akin to the contextual information, feelings, and self awareness captured in Tulving’s concept of autonoetic consciousness or the state of ‘remembering,’ and that a disruption of this state in older adults contributes to this particular deficit in monitoring episodic memory. This idea will be developed below, but we suggest that FOK judgments in older adults are particularly impaired in episodic memory tasks, because in these types of task, retrieval processes, guided by a state of ‘remembering’ are critical (see Conway, 2005). In a semantic memory task, one can assume that target item retrieval depends mainly on cues present in the person’s general knowledge, and can be made fairly accurately on the basis of reading the cue alone. Because it would be theoretically interesting if FOK was impaired for one form of memory but not the other, it seemed important to replicate the age deficit in episodic FOK, but more importantly, to demonstrate within the same group of participants that one form of monitoring is impaired, but another is intact. The purpose of the first experiment in this research was to investigate this hypothesis directly by testing the same group of participants with both an episodic and a semantic FOK task. In regards to the literature, we expected to observe an age effect only on the episodic FOK, but not semantic FOK. As will be discussed below, this would suggest that the two forms of FOK draw upon different cues, and the ability to monitor these cues is affected differentially by the aging process.