Age Differences in Encoding versus Retrieval
For many years, researchers examining age differences in memory performance have tried to implicate either encoding or retrieval as the source of memory processing. However, such attempts have met with mixed results (Zacks et al., 2000). As we will see, however, more recent work based on brain metabolism, automaticity in retrieval, and the social context of memory raises some interesting questions about the source of memory changes. But before moving to those subjects, we will examine the evidence for the relative contribution of encoding and retrieval to memory performance in older adults.
Encoding. Results from years of research suggest an age-related decrement in encoding processes (Craik & Byrd, 1982; Koutstaal, 2003; Smith, 1996; Zacks et al., 2000). An example of an encoding
process that affects memory is elaborative rehearsal. Elaborative rehearsal involves making connections between incoming information and information already known. For example, a person presented with the word emu and told that this is a bird that does not fly may try to think of other flightless birds. With some thought, ostrich may come to mind. Linking emu and ostrich would be an example of this type of rehearsal.
Research in support of an age-related decrement in elaborative rehearsal finds that older adults have more difficulty making such connections than younger adults do (Light, 1996; Smith, 1996). Interestingly, however, once these connections have been made, older and younger adults maintain them equivalently (which supports the conclusion that whereas encoding differences exist, storage differences do not). In other words, older adults would be slower than younger adults at making the emu – ostrich connection, but once it was made, both groups would remember it just as well.
This example of the encoding process involves intentional learning of information. However, what happens when information is learned and there are no expectations of testing this memory later on (e. g., when a person witnesses a crime)? This latter type of memory processing is sometimes called incidental learning. For example, a study compared incidental and intentional learning of names in younger and older adults (Troyer, Hafliger, Cadieux, & Craik,
2006) . Their instructions for encoding the names varied from stating the first letter of the name to defining the name to simply trying to remember the name for a later test. Younger adults tended to outperform older adults for intentionally learned names, but older adults did as well as younger adults whenever they encoded the information incidentally (e. g., remembering the first letter or defining the name).
Why do these age differences in intentional versus incidental encoding occur? Anderson and colleagues (1998) conclude that attentional resources are consumed to a greater extent during intentional encoding. If you were asked to study some materials that you would be tested on while at the same time you had to tune out a conversation going on in the same room, you would place a lot of effort into using
strategies to effectively focus on the information to be studied. However, while you were studying the materials, if you were told that the conversation in the room was about you, you would probably extend more effort to listening to the conversation. You would perform better on the test in the first case than in the second case. Attentional control would be necessary while you encoded the information to succeed on the test. Thus, the distracting conversation would not hurt you when you were able to focus your attention more exclusively on the test.
One important aspect of attentional control and expending attentional resources in the example above is that when we are confronted with large amounts of information that we need to remember, we tend to use various techniques, called strategies that make the task easier and increase the efficiency of storage. We will examine strategy differences in the elderly next.
Use of Strategies during Encoding. A critical issue is whether older adults behave strategically when studying information to be remembered. Two effective strategies for learning new information are to organize it and to establish links to help you remember the information. For example, consider a student’s efforts to learn the information in college courses. Learning the necessary facts of chemistry, psychology, literature, and so forth is much easier if separate notebooks are kept for each class. Imagine the potential for confusion if all of the class notes were simply mixed together. Keeping them separate is an example of an organizational strategy.
It may be that older adults do not spontaneously use such effective encoding strategies in comparison to younger adults (Hertzog, Dunlosky, & Robinson,
2007) . For example, older adults are less likely to take advantage of similarities in meaning among words (such as the link between river and lake) presented randomly in a list as a way to organize the items. Because the number of items remembered from such a list is highly related to the use of organization, younger adults outperform older adults on such tasks. However, the available evidence suggests that the spontaneous production of effective strategies is not a sufficient account of age differences in
memory performance (Hertzog & Hultsch, 2000; Hertzog et al., 2007; Light, 1996).
Hertzog, McGuire, and Lineweaver (1998) found that only 35% of older adults compared to 49% of younger adults used optimal strategies for encoding (e. g., placing related items into meaningful categories); however, as Figure 6.2 shows, there were large age differences in recall irrespective of strategy methods used. They also found that strategy production had a very modest effect on age differences in recall. In support of this finding, Dunlosky and Hertzog
(1998) found that large age differences in recall remained even when older adults were instructed to use effective strategies. However, as Hertzog and Hultsch (2000) have pointed out, producing a strategy does not necessarily imply that the strategy was used effectively. We need more research to differentiate the quality of implementation of strategies as opposed to their production. Current thinking suggests that this is an important avenue for future research in order to better understand adaptive behavior in learning new information (Hertzog & Hultsch, 2000). We will explore this more when we discuss remediation of memory deficits later.
Figure 6.2 Mean recall for young, middle-aged, and older adults as a function of strategy groups.
Source: “Aging, Attributions, perceived control, and Strategy Use in a Free Recall Task”, by C. Hertzog, C. L. McGuire, & T. T. Lineweaver, 1998. In Aging, Neuropsychology, and Cognition, Vol. 5, pp.95, 85-106.
Retrieval. Age-related differences on explicit memory tasks are affected by the degree to which a task employs retrieval cues. It has been long known that age differences are more pronounced on free-recall tests which provide no cues to individuals (Old & Naveh – Benjamin, 2008; Spencer & Raz, 1995). However, for recognition tests, the original target to be learned is presented as a cue. Accordingly, fewer age differences are found (Spencer & Raz, 1995). In addition, older adults show greater frequency of tip-of-the-tongue states and feeling-of-knowing (feeling you know something yet you are not sure what it is) after failure to retrieve information (Burke et al., 1988; MacKay & Abrams, 1996; Maylor, 1990). This is probably due to a temporary inaccessibility. In fact, older adults are as likely as younger adults to recognize the material they cannot recall when in the TOT state.
Craik (1986) suggests that age differences in memory where there are limited cues (e. g., free recall) may be a function of limited self-initiated operations in older adults. This idea suggests that older adults have difficulties in performing tasks that do not have a high level of environmental support. For example, when you are recognizing a previous piece of information, the fact that the target is in front of you creates a high level of environmental support. However, a free-recall task demands that you access that information with limited help and cues from the environment.
Based on the evidence so far, we find that memory functioning in the aging adult is influenced by deficits in both encoding and retrieval (Zacks et al.,
2000) . However, there are other factors that may qualify this conclusion, suggesting that the picture is a more complicated. In order to explore this, we need to consider whether the cognitive process is automatic or deliberate as well as implicit or explicit. But first let’s revisit cognitive neuroscience evidence for encoding and retrieval processing.
Cognitive Neuroscience Revisited. Cognitive neuroscience (discussed in Chapter 2) also presents evidence suggesting age differences in encoding and retrieval. Using PET scans, Grady and her research team (1995) found that both younger and older age groups showed similar patterns of increased blood flow in specific areas of the brain during recognition, but differed
significantly during encoding. Compared with older adults, younger adults showed significantly greater increases in blood flow to the left prefrontal and temporal areas of the brain. (Recall that this is called lateralization to a specific hemisphere or asymmetry.) Grady and colleagues interpreted these differences as evidence that age-related memory differences may be due to older adults’ failure to encode information adequately. Inadequate encoding could result in information not getting into memory at all, or not being as elaborately encoded, making retrieval more difficult at best. Furthermore, fMRI studies have found that age-related reductions in frontal-lobe activation are associated with reduced ability in encoding words in memory (Stebbins et al., 2002). In addition, the hippocampal region has been related to current memory functioning as well as predicting the rate of further memory decline over time (Prull et al., 2000).
Cabeza and colleagues (Cabeza, 2002; Cabeza et al., 2003), however, found that prefrontal activity during memory performance is less lateralized (activity found more equally in both brain hemispheres instead of localized in one hemisphere) in older adults than in younger adults. Lateralization (found in young adults) leads to more effective processing of information. It may be, though, that those older adults who show strong bilateralization (i. e., activity in both hemispheres) when processing information also show enhanced cognitive performance relative to older adults who show less bilateralization and less lateralization. The additional activity across many brain regions that occurs when older adults process information may enhance performance on the specific task investigated.
In sum, the research on encoding and retrieval processes is important for three reasons. First, it emphasizes that age-related decrements in memory are complex; they are not due to changes in a single process. Second, intervention or training programs must consider both encoding and retrieval. Training people to use encoding strategies without also training them how to use retrieval strategies will not work. To the extent that only partial information is encoded, retrieval strategies need to focus on helping people find whatever aspects are available. Third, theories of how memory changes with age must take individual differences into account, especially differential rates of change in component processes and lateralization. Overall, finding the sources of memory decline demonstrates the importance of considering multiple levels of influence.
Theories of memory development must consider those components of processes that change and those that do not. However, not all aspects of retrieval are effortful and do not necessarily demand resources that overburden older adults’ processing. We next examine a relatively recent area of research that centers on automatic retrieval processes that tend to be equivalent for older and younger adults.