Aging and Memory in Humans
AGING AND MEMORY IN HUMANS
One of the commonest complaints of older people is that their memory is not what it used to be. The validity of such subjective reports is borne out by the scientific literature: Memory performance does decline as a function of the normal aging process in healthy adults, although the decline is much more evident with some materials and tasks than it is with others. This variability has given researchers useful clues to the specific memory components or processes that are particularly vulnerable to the effects of aging. Some of this work is described below.
Nearly all the studies described use the cross-sectional method of age comparison; that is, a group of young adults (often college students in their early twenties) is compared with a group of older adults (usually community-dwelling volunteers in their sixties and seventies). Additionally, some studies incorporate middle-aged groups of people in their forties and fifties. The cross-sectional method is much more practicable than within-subject longitudinal studies, but it does leave open the possibility that the differences observed between the groups may be attributable to causes other than aging—to differences in education or motivation, for example. Obviously researchers take pains to minimize the possibility of such artifacts, and they do this by matching the groups by educational level, by vocabulary (a rough measure of verbal intelligence), and by other indicators of intelligence and socioeconomic status. Many crucial experimental results take the form of interactions between age and some experimental variable; that is, one condition of the experiment is associated with large age-related differences, whereas another condition is associated with much smaller differences, even though the same subjects are used. The finding of such differential effects makes it harder to argue that group differences are a function, say, of reduced motivation in the older sample.
Like other experimental explorations of individual differences in memory ability, the work on aging has been carried out within various theoretical frameworks and with a view to establishing some theoretical point. In this brief article it is not possible to go into details of theoretical motivation in most cases, but, in summary, four main age-related changes that arguably underlie changes in memory performance are a decline in general processing speed; a decline in processing resources or "attentional energy" (Craik and Byrd, 1982); an age-related reduction in the efficiency of inhibitory processes; and an impairment in the executive control of cognitive processing. These four theoretical viewpoints, along with supporting evidence, are discussed in Kester, Benjamin, Castel, and Craik (2002); and in Zacks, Hasher, and Li (2000). The present article focuses on empirical work since the 1980s, pointing out the implications for theory where relevant.
Indirect Memory Tests
When we think about memory, it is usually in the sense of the conscious retrieval of a past event or the retrieval of a previously learned fact. However, there are also many cases in which previous events can affect present behavior in the absence of conscious awareness of the past event in question. Such cases of "implicit memory" are revealed by indirect memory tests—examples include word-stem completion and word-fragment completion. In word-stem completion, the participant is given the first few letters of a word (e.g., MAR___ or DRA___) and asked to complete the stem with the first word that comes to mind. Word-fragment completion is similar; here the participant is given words with letters missing and is asked to complete them (e.g., M _ _ K _ T or _ R A _ _ R). The general finding in such experiments is that participants can complete the words more readily if they have studied the target words (e.g., MARKET, DRAWER) previously, even though the person may be quite unaware that the words are drawn from the studied list. This phenomenon is known as "priming," and it has typically been found that age differences in tests of priming are much smaller than those found in direct tests (Fleischman and Gabrieli, 1998). A similar discrepancy between direct and indirect tests is found in amnesic patients, who do poorly on explicit or direct tests but comparatively well on implicit or indirect tests. These results are sometimes taken as evidence that encoding processes are therefore intact in elderly individuals and amnesic patients, and that the observed decrements are failures of retrieval. But encoding processes may be somewhat impaired in older people and amnesics; they might be sufficient to support later indirect tests, but insufficient to support later direct tests.
Semantic memory refers to a person's learned knowledge of facts and concepts. Typically we are unaware of where and when we first learned that Paris is the capital of France, the meaning of rhinoceros, and that 7 × 9 = 63; the notion of semantic memory thus stands in contrast to episodic memory, where details of time and place are crucial. There are different views on whether episodic and semantic memory are different "memory systems" or whether they simply exemplify different degrees of abstraction from the original events that gave rise to the encoded knowledge (Craik, 2002). Whatever the resolution of this debate, older adults typically show fewer losses in semantic memory than in episodic memory, provided that the knowledge in question is used on a regular basis. Thus, older people show minimal losses in vocabulary and in knowledge of facts and concepts.
On the other hand, some aspects of semantic memory do appear to decline with age—the ability to learn completely new facts, for example. Even well-learned material can be more difficult to retrieve for older adults, and difficulty in remembering names is possibly the most frequent age-related memory complaint. It is unclear, however, whether names show a disproportional impairment with age (Maylor, 1997); a more general statement might be that older adults have sporadic difficulty in retrieving any information that they use infrequently. Having a word or name "on the tip of the tongue" is an experience that occurs more frequently as people grow older, and older adults report less partial information about the target word. In addition, speed of retrieval slows with advancing age, in line with the slowing of many other cognitive processes. In summary, then, whereas the representation of learned knowledge remains reasonably intact into old age, older adults experience occasional difficulty in assessing that knowledge.
Episodic memory refers to the ability to recollect specific events, and this form of memory has been extensively studied in the laboratory using tests of recall and recognition. Many such studies have used rather artificial materials—lists of unrelated words, for example. The benefit has been greater experimental control over encoding and retrieval processes, but a possible drawback is that the principles emerging from such studies might not apply to real-life remembering. Work carried out beginning in the 1980s on autobiographical memory has allayed these fears to a large extent, however. In these studies, memory for personally experienced events in the subject's life have shown that real-life memories are affected by the same factors and subject to the same laws as are materials learned in the laboratory. Age-related differences in autobiographical memory have been studied by presenting people of different ages with cue words such a flag or school and asking participants to generate a personal memory that each word evokes. For adults of all ages, recent memories were generated most often, and the incidence of recollected events declined from the present to the past. One interesting exception to this general trend is that all adults showed a disproportionate "bump" of generated memories from late teenage and early adult years (Rubin, Wetzler, and Nebes, 1986). Presumably this bump reflects the many important career-related and emotionally significant events that occur during this period of our lives.
Episodic memory for events occurring in the last few minutes, hours, or even weeks typically shows large age decrements, and much of the effort of cognitive aging researchers has been directed to understanding the factors underlying this problem. One set of findings points to an age-related decline in the efficiency of retrieval processes. Several investigators have shown larger age differences in tests of free recall (in which participants must recall a list of words or a paragraph of text with no cues or reminders) than in tests of recognition memory (in which participants must pick out the originally studied items from a mixed set of targets and distractors). Details of these experiments are summarized by Kester, Benjamin, Castel, and Craik (2002). The fact that older adults have difficulty recalling studied items but can recognize them suggests that the difficulty lies at the retrieval stage, although recognition is usually the easier test. It seems likely that there are also problems at encoding, however. One piece of evidence supporting the encoding hypothesis is that, when encoding processes are "guided" appropriately by means of questions that emphasize the semantic aspects of the item, age-related differences are often reduced (Craik and Jennings, 1992). One similarity between encoding and retrieval is thus that, when appropriate processes are supported or guided by additional information (e.g., semantic orienting tasks at encoding, retrieval cues or a recognition test at retrieval), age differences are typically reduced. This pattern of findings led Craik (1983) to suggest that, whereas "unsupported" encoding and retrieval processes are often inefficient in older people, possibly due to a reduction in available processing resources, these inefficiencies can be overcome through the "environmental support" provided by the experimental situation or by the context of a person's familiar surroundings.
The conclusion that age-related difficulties in episodic memory are consequences of impaired processing at both encoding and retrieval is supported by findings from studies of functional neuroimaging. Studies using PET (positron emission tomography) and fMRI (functional magnetic resonance imagery) have shown that older adults exhibit less activation of the ventral left prefrontal cortex during memory encoding than do their younger counterparts. Many studies have demonstrated that this left prefrontal region is associated with the processing of semantic information and with good later memory of processed items. Other studies have shown that, whereas young adults activate regions that are predominantly lateralized in the right prefrontal cortex during retrieval, older adults show activations in both right and left prefrontal regions when retrieving information. The additional left-sided activations in older people may reflect the brain's attempt to compensate for the declining efficiency of regions that subserve retrieval in young adulthood. An account of this exciting line of research is provided by Prull, Gabrieli, and Bunge (2000).
One further topic pertaining to episodic memory is the ability to remember the source of encoded information, or details of the context in which an event occurred. Everyone has experienced the situation in which a person's face seems very familiar, and yet we cannot say how we know the person. Usually it turns out that we have encountered the familiar person in a very different context from his or her habitual location, leading George Mandler to refer to the experience as "the butcher on the bus" phenomenon. A similar failure of memory for context occurs when we know a fact but cannot recollect whether someone told us the information or whether we read it in the newspaper or heard it on the radio. Older adults are particularly prone to such failures to bind contextual information to the core aspects of the event, or to the item of information (see Kester, Benjamin, Castel, and Craik, 2002, for details). The effects are also seen during the output of information; for example, an older person may be cued by a conversational companion's background to exclaim, "Ah—you will be interested to hear this!" and then proceed to retell the same story that the hapless listener has endured many times previously.
What lies behind the phenomenon of source forgetting? Some researchers have linked the failure to bind contextual details to core features to an inefficiency of frontal lobe function. Other cognitive neuroscientists have suggested that the hippocampus is centrally concerned with such binding functions (Prull, Gabrieli, and Bunge, 2000). At the behavioral level, source forgetting may be viewed as one instance of a general age-related difficulty of association or integration. Naveh-Benjamin (2000) lays out convincing evidence that older adults have particular difficulty remembering the associative links between items of information, although the items themselves may be remembered quite well.
Short-Term and Working Memory
The phrase short-term memory has unfortunately been used in a number of slightly different ways, and this can give rise to confusions about findings. Clinicians typically use the phrase to mean recent memory—events that have happened in the last few hours or days—whereas experimental psychologists have used the phrase to refer to information still held in mind, as when we look up a telephone number and rehearse the information until we have dialed it. This latter type of memory (sometimes also referred to as primary memory) shows very little decline with age. Similarly, memory span—the longest list of digits or words that a person can repeat back accurately—declines only slightly from the twenties to the eighties. If short-term memory is used in the first sense, however, it falls into the general category of episodic memory, and substantial age-related decrements are found, as discussed in the previous section.
The term working memory has been adopted widely to refer to information held and manipulated in mind. Thus, solving a verbal problem or performing mental arithmetic is considered to involve working memory (WM). In this sense WM incorporates executive processes as well as relatively automatic auxiliary systems such as the articulatory loop and visuo-spatial sketchpad (Baddeley, 1986). With regard to aging, performance relying largely on auxiliary systems holds up well; memory span falls into this category. But when good performance requires executive processes or complex manipulations of information held in mind, then older people typically do less well than their younger counterparts (Craik and Jennings, 1992; Zacks, Hasher, and Li, 2000). It seems possible that this age-related decline reflects the reduced efficiency of frontal lobe processes in older adults (Glisky, Polster, and Routhieaux, 1995).
Older adults often have difficulty dealing with dual-task situations in which they must divide their attention between two simultaneous activities. In one such demonstration Anderson, Craik, and Naveh-Benjamin (1998) found that, when younger and older adults divided their attention between a memory task and an ongoing reaction-time task, memory performance dropped equally for the younger and older groups (relative to performing the memory task on its own) but that performance on the reaction-time task dropped much more for older than for younger adults, especially during the retrieval phase of the memory task. In a similar demonstration, Lindenberger, Marsiske, and Baltes (2000) measured walking accuracy and speed in adults of different ages while they learned a list of words. They found greater dual-task costs in the older group, partly reflecting the increased need for executive processes in word learning but also partly reflecting the older adults' greater need to deploy executive processes to control accurate walking.
Prospective memory refers to the situation in which a person intends to carry out some action at a future time and then either performs the action successfully or forgets to perform it. Such situations are common in everyday life, as are failures of prospective memory—forgetting to make a phone call, mail a letter, or to pass on a message, for example. Researchers have made the useful distinction between event-based and time-based prospective memory, the first referring to situations in which the intended action should be cued by an event, such as seeing the letter to be mailed or meeting the colleague for whom the message was intended. On the other hand, time-based prospective actions are cued by times: "I should call home at 3:30 P.M.," for example.
Prospective memory failure generally increases with age. As one example, Mäntylä and Nilsson (1997) reported a study in which participants were asked to remind the experimenter at the end of a testing session that they should sign a form. Successful performance dropped from 61 percent of participants aged thirty-five to forty-five to only 25 percent of participants aged seventy to eighty. Other research demonstrates that older adults do worse on time-based than on event-based, prospective memory tasks, arguably because the former type of task is less well supported by environmental cues (Craik, 1983). Finally, older people do better on real-life prospective memory tasks than on laboratory-based tasks (Kester, Benjamin, Castel, and Craik, 2002). This finding may reflect greater motivation on the part of older adults, or it may reflect their greater use of daily structures and routines.
Memory performance does decline with age, but the decline is greater in some tasks than in others. Performance is often poor on episodic memory tasks, especially if the person must recall the information without cues. Performance is also comparatively poor on source memory, on prospective memory, and on working memory tasks. On the other hand, memory for general knowledge and for routine activities holds up well with age, as does primary memory for information held briefly in mind. Finally, environmental support from familiar surroundings can be particularly helpful as an aid to remembering in older adults.
Anderson, N. D., Craik, F. I. M., and Naveh-Benjamin, M. (1998). The attentional demands of encoding and retrieval in younger and older adults: 1. Evidence from divided attention costs. Psychology and Aging 13, 405-423.
Baddeley, A. D. (1986). Working memory. London: Oxford University Press.
Craik, F. I. M. (1983). On the transfer of information from temporary to permanent memory. Philosophical Transactions of the Royal Society of London, ser. B, 302, 341-359.
—— (2002). Human memory and aging. In L. Bäckman and C. von Hofsten, eds., Psychology at the turn of the millennium, Vol. 1: Cognitive, biological, and health perspectives, pp. 261-280. Hove, Sussex, UK: Psychology Press.
Craik, F. I. M., and Byrd, M. (1982). Aging and cognitive deficits: The role of attentional resources. In F. I. M. Craik and S. Trehub, eds., Aging and cognitive processes, pp. 191-211. New York: Plenum.
Craik, F. I. M., and Jennings, J. M. (1992). Human memory. In F. I. M. Craik, and T. A. Salthouse, eds., The handbook of aging and cognition, pp. 51-110. Hillsdale, NJ: Erlbaum.
Fleischman, D. A., and Gabrieli, J. D. E. (1998). Repetition priming in normal aging and Alzheimer's disease: A review of findings and theories. Psychology and Aging 13, 88-119.
Glisky, E. L., Polster, M. R., and Routhieaux, B. C. (1995). Double dissociation between item and source memory. Neuropsychology 9, 229-235.
Kester, J. D., Benjamin, A. S., Castel, A. D., and Craik, F. I. M. (2002). Memory in the elderly. In A. Baddeley, B. Wilson, and M. Kopelman, eds., Handbook of memory disorders, 2nd edition. London: Wiley.
Lindenberger, U., Marsiske, M., and Baltes, P. B. (2000). Memorizing while walking: Increase in dual-task costs from young adulthood to old age. Psychology and Aging 15, 417-436.
Mäntylä, T., and Nilsson, L. G. (1997). Remembering to remember in adulthood: A population-based study on aging and prospective memory. Aging, Neuropsychology, and Cognition 4, 81-92.
Maylor, E. A. (1997). Proper name retrieval in old age: Converging evidence against disproportionate impairment. Aging, Neuropsychology, and Cognition 4, 211-266.
Naveh-Benjamin, M. (2000). Adult age differences in memory performance: Tests of an associative deficit hypothesis. Journal of Experimental Psychology: Learning, Memory, and Cognition 26, 1,170-1,187.
Prull, M. W., Gabrieli, J. D. E., and Bunge, S. A. (2000). Age-related changes in memory: A cognitive neuroscience perspective. In F. I. M. Craik and T. A. Salthouse, eds., The handbook of aging and cognition, 2nd edition, pp. 91-153. Mahwah, NJ: Erlbaum.
Rubin, D. C., Wetzler, S. E., and Nebes, R. D. (1986). Autobiographical memory across the adult life-span. In D. C. Rubin, ed., Autobiographical memory, pp. 202-221. Cambridge, UK: Cambridge University Press.
Zacks, R. T., Hasher, L., and Li, K. Z. H. (2000). Human memory. In F. I. M. Craik and T. A. Salthouse eds., The handbook of aging and cognition, pp. 293-357. Mahwah, NJ: Erlbaum.
Fergus I. M.Craik