Cognition in Context
Research on cognition in aging has traditionally sought to develop generalizations about changes associated with age that hold regardless of the context within which a cognitive process operates. However, the most impressive generic feature of the human mind may be its remarkable flexibility in adapting to diverse environments. As the cognitive sciences expand their scope, it becomes increasing clear that minds are adapted and often quite finely tuned to particular environments. To understand cognitive functioning, it is necessary to pay attention to the context of cognition (Super and Harkness, 1986; Goodnow, 1990; Shweder, 1991; Baltes and Staudinger, 1996; D'Andrade, 1981, 1995; Fiske et al., 1998). This context includes not only evolutionary and biological constraints and affordances but the cultures in which minds reside, including culturally shared ideas, expectations, habits of mind, communication patterns, and technologies. Contextual factors that enable some older people to function particularly well may be employed to improve functioning for others.
Recognition of cultural influences on cognition is particularly significant for research on the aging mind because it often seems natural to regard age-related changes in cognition as reflecting biological (e.g., neural, cellular, hormonal) change or activity, and equally natural to seek the sources of this variation in biological systems alone. Yet, if cognitive functioning in fact importantly reflects the contexts and environments within which people engage, some of the most important sources of variation in cognition will be found in the meanings, artifacts, practices, and institutions that structure
these contexts and environments (Fiske et al., 1998). Note in this context that ''old age" is relatively new as a cultural phenomenon and that as longevity and quality of life for older adults increase, the cultural and social aspects of old age are likely to continue changing.
RECENT SCIENTIFIC DEVELOPMENTS
Recent scientific developments in understanding the aging mind suggest the need to expand research to pay more careful attention to various aspects of the contexts of aging minds. Most important are four growing bodies of evidence: (1) that life experience can change the brain; (2) that individuals adapt in various ways to maintain cognitive functioning and task performance in the face of changes in the brain and in their social contexts; (3) that systematically different life experiences yield systematically different cognitive contents and processes; (4) and that advanced technology can modify the context of cognition to greatly improve functioning for older people.
Neurobiology of Life Experience
The coevolution of brain and culture has long been recognized in the study of human evolution (e.g., Durham, 1991), in which there has been some joining of developmental neuroscience and the behavioral-social developmental sciences. More recently, evidence of neural plasticity has revealed the importance of experiences within an individual's life span as a cause of change in the brain. The concept of "experience" was initially identified as a determining or modulating factor of brain development (e.g., Squire and Kandel, 1999). For example, in developmental work on perception, many studies demonstrated that certain aspects of brain development (such as the size of function-specific locations and the complexity and density of synaptic and dendritic architectures) are conditioned by the nature of early "sensory-input" experiences, such as the richness of an animal's environment (Merzenich and Sameshima, 1993; Kolb and Whishaw, 1998). In addition, experimental studies have shown that training in specific tasks also affects brain structure (see Kolb and Whishaw, 1998, for a review). A good recent example, though nonexperimental, is the finding that players of stringed instruments of the violin family have a larger cortical representation of the fingers of the left hand than of the right hand and that this is particularly true of string players who began musical practice before age 13 (Elbert et al., 1995). The presumed reason is that in playing these instruments, the fingers of the left hand are manipulated individually while those of the right hand, which holds the bow, move together. The evidence suggests that architecture in this area of the brain may be more plastic before age 13. These findings demonstrate links among enriching experiences, improved performance, and change
in neural representation. Further research is needed to clarify the relationships between particular kinds and durations of life experience and particular changes in the brain and thus to clarify which aspects of "experience" might be protective against cognitive decline.
The intellectual creations of human culture also leave imprints on the brain. For example, letters and digits are culturally created symbols that can be used to represent the same concepts (e.g., "four" and "4"). Yet letter and digit recognition depend on different neural regions in literate subjects (Polk and Farah, 1998). Handwriting can be selectively impaired by brain damage that does not affect other sensory-motor functions of the hand (Alexander et al., 1992). And bilinguals show regional segregation of their different languages (Ojemann and Whitaker, 1978), illustrating that the products of specific cultures can sometimes be recognized in the brain.
There is evidence that the aging brain continues to change. For example, dendritic growth increases, possibly in compensation for cell loss (Kolb and Whishaw, 1998). New research presents evidence of neurogenesis in adulthood: new neurons continually appear in areas of the adult primate brain associated with higher cognitive functions (Gould et al., 1999b). The role of life experiences in such processes has hardly begun to be investigated.
These lines of evidence that experience changes the brain in lasting ways support the idea that many outcomes of brain development from infancy through old age are the expression of experiential-cultural factors and suggest that simple reductionistic and deterministic models in which cognitive capabilities flow exclusively from brain development are inappropriate. They suggest a program of research to examine in detail the proposition that although evolution-based brain development lays the basic foundation of brain architecture, subsequent differentiation and development of the brain is importantly influenced by how societies are organized and by how individuals live their lives. This research program would aim to clarify the relationships between particular kinds and durations of life experience and particular changes in the brain.
Adaptivity of Cognitive Functioning
Older people often continue to perform well the cognitive tasks of living despite declines in some of the underlying cognitive capabilities. For cognitive tasks that require new learning or that depend on speed of responding, performance diminishes with age (Burke and MacKay, 1997); such performance is also related to the ability to enact certain tasks of everyday life, such as such as paying bills or filling out tax forms (Diehl, 1998; Willis and Marsiske, 1991). However, performance on laboratory tasks that measure cognitive processes does not map perfectly onto performance in many important life domains, such as performance in the workplace and the exercise of
"practical intelligence" (e.g., solving social problems, planning meals) (Sternberg, 1986).
The research suggests that the more an everyday problem is complex, ambiguous, and dependent on a variety of skills, the weaker the relationship between traditional measures of intelligence and performance. For instance, older adults do well in performing tasks of wisdom (Baltes and Staudinger, 2000). Wisdom involves the coordinated use in making life choices of factual and procedural knowledge about life; knowledge about life conflicts, contexts, and priorities; and knowledge about recognizing and managing uncertainty. It is measured in this research by the number of criteria people consider when they think aloud about problems of life planning, such as how to advise a 15-year-old girl who wants to get married right away, or what to say to a close friend who is thinking about suicide. Performance on such complex and ambiguous tasks shows only a weak relationship to measures of intelligence; rather, it is primarily associated with indicators of life experience, personality, and cognitive style. Thus, age-related declines in measured cognitive processes may not imply equivalent declines in the ability to perform cognitive tasks of living. Some research actually shows positive age trends, with older people outperforming younger people despite lower performance on some measures of cognitive processing (Marsiske and Willis, 1995).
One reason practical functioning may decline more slowly than some kinds of cognitive capability may be that experience-based procedural and declarative knowledge is relatively well maintained in later life (Baltes et al., 1999; Blanchard-Fields and Hess, 1996). To the extent that people come to old age with greater stores of knowledge, adaptivity in later life benefits from those reserves. Adaptivity may also be promoted because older people consider and rely more on other people in solving problems in everyday life (Sansone and Berg, 1993; Blanchard-Fields, 1997). Memory is known to benefit from social support (Dixon and Gould, 1998), and older people's memories are significantly enhanced when they recall them in the presence of another. Margrett (1999) found that older people's performance on a range of tasks, including understanding medication labels, social dilemmas, and map reading, benefited from collaboration; moreover, those who perceived themselves as needing assistance benefited more. In other words, the people with the poorest independent performance were able to compensate by collaboration.
Researchers in the field of what is now called "everyday competence" have come to adopt a framework of "person-environment fit" (Lawton, 1982), which suggests that functioning in real-world situations is determined by the fit between the individual and the environmental and social context. Although this model is applicable to people of all ages, research suggests that "fit" is more important for people whose cognitive abilities are compromised. For example, among people with serious visual impairment, environ-
mental fit is far more relevant to everyday competence than among people with mild or no visual impairment (Wahl et al., 1999). The strong relationship of sensory functioning to cognitive performance among aging individuals (Lindenberger and Baltes, 1994) is just as relevant to everyday activity competence (Marsiske et al., 1997). It appears that older people can sometimes improve their "fit" and their everyday competence by drawing on other people in their environments.
Another aspect of adaptivity is that people's motivations change as they approach the end of life (Carstensen et al., 1999). One change is that emotional goals are more salient to older as opposed to younger people. Older people's adaptation to their life situation seems to give them a relative advantage in certain kinds of cognitive tasks. For instance, older adults perform better than younger adults when solving social dilemmas involving problems high in emotional salience. They are more likely to generate good solutions and to consider other peoples' feelings in their recommendations (Blanchard-Fields, 1997).
In short, aging people are often successful in using other minds and environmental supports to buffer the effects of cognitive decline. However, the adaptive transactions that occur between persons and environments remain poorly understood. Very little research to date has focused on the ways in which older people actually use environmental resources, including social or cultural resources, to maintain functioning. Relatedly, there is growing recognition of the need to consider the broader sociocultural contexts in which people age (Baltes and Carstensen, 1999). In order to understand everyday functioning, it may be necessary to shift from a conception of cognition as something that occurs entirely within the individual to one that takes into consideration distributed, interactive processes that shape and are shaped by the social world and the technological environment.
Cognitive Effects of Life Experiences
If the brain is shaped by experience, then it should not be surprising that individuals who differ systematically in the kinds of experience that shape the brain differ in their cognitive functioning (for a thorough review of social influences on cognition, see Levine et al., 1993). A reasonable hypothesis is that if experiential differences occur at formative periods or persist throughout the life span, then the cognitive effects will be persistent. Another reasonable hypothesis is that so-called crystallized or pragmatic aspects of cognition are likely to be more strongly influenced by life experience than so-called fluid or mechanical aspects.
Considerable evidence is accumulating of cognitive differences between aging members of social groups that differ systematically in experience through the life cycle: groups defined by occupation, socioeconomic status,
ethnicity, race, and culture (e.g., Avolio and Waldman, 1994; Burton and Bengston, 1982; Jackson, 1985). For instance, better cognitive performance in aging individuals is generally associated with higher levels of education (e.g., Birren and Morrison, 1961; Blum and Jarvik, 1974; Denny, 1979; Denny and Palmer, 1981; Green, 1969; Kesler et al., 1976; Ripple and Jaquish, 1981; Schaie and Strother, 1968; Selzer and Denny, 1980), higher-status emploment and higher income levels (e.g., Arbuckle et al., 1986; Schaie, 1983; Gribbon et al., 1980; Owens, 1966), and being white rather than black in the United States (e.g., Fillenbaum et al., 1988). The most precipitous decline in adaptive functioning, which occurs during the eighth decade of life, is far less pronounced in people with greater as opposed to lesser education (Schaie, 1996). The association of education with cognitive functioning in old age varies, however, by country. In Germany, for instance, this association is much weaker than in the United States (Baltes and Mayer, 1999). In addition, there is a growing body of evidence that adult members of sharply different cultures differ systematically in the ways they habitually attend, process, and interpret information as well as in the approaches they take toward the aging process (for a review, see Kitayama, Appendix F).
The true meaning of such associations is not yet understood. However, several interesting hypotheses are available to account for them and to suggest causal mechanisms leading from particular life experiences to their presumed cognitive effects. One set of hypotheses centers on cognitive practice or training. The central idea is that formal education, occupational experience, and the like provide cognitive practice that shapes cognitive abilities and maintains particular ones—perhaps a broad range of abilities for general education and a narrower range for certain occupational experiences. Similarly, cultural differences may provide members of certain cultures with training in ways of thinking that are not practiced in other cultures, resulting in lifelong differences in cognitive skills (Gauvain, 1995). An example from a recent study is the finding that adult native speakers of Italian and English show distinct patterns of brain activation in the temporal regions involved in reading and naming tasks during language processing (Paulesu et al., 2000). This finding is thought to be due to the sharp difference between English and Italian in how closely phonemes (sounds) and graphernes (letter combinations) map onto one another. Importantly, this difference is observed for both words and nonsense words, suggesting that the acquisition of a particular language leads to persistent differences in the processing of new language input.
Another example is the report that formally schooled children are more likely to use the learning strategy called clustering, in which they group like items together explicitly, making recall much easier (Brislin, 1993). This sort of training may help explain differences in memory associated with educational attainment later in the life span. Formal education and experience in wage-labor occupations have been found to be associated with increased con-
cern for time, punctuality, and planning activities in advance (Inkeles and Smith, 1974). The complexity of work roles has also been associated with general characteristics of cognitive functioning in adults (Kohn et al., 1978, 1982). Technology also shapes cognition through training. Abacus users have been found to make different kinds of errors in solving mathematical problems from people who use Arabic numerals (Stigler, 1984; Stigler et al., 1986). It has been argued that in medieval times, when few people could use writing as a memory aid, memory was much more detailed and rote than it is in advanced societies today (see Yates, 1966; Carruthers, 1990; Olson, 1994). And it has been suggested that television shapes a mind adept in rapid processing of images and comfortable with attending for repeated short periods rather than extended ones (Greenfield, 1984).
Following on reported differences in information processing styles between people from East Asian and European cultural groups (Hsu, 1983; Liu, 1974; Lloyd and Moodley, 1990; Nagashima, 1973), Park and colleagues (1999) found that cues in a memory task had different effects on people from the two groups. Compared with people from European cultures, supportive cues helped the East Asians more, and distracting cues harmed their performance more. The explanation offered was in terms of the previously published claim that memory among East Asians is more sensitive to contextual cues, whereas Europeans focus more narrowly on the object at hand. The same authors offered a contextual-support explanation for their finding that Americans perform better than Chinese on a free recall task involving six words from five natural categories.
A second set of causal hypotheses involves health as an intervening variable between social context and cognitive aging. A central idea is that the shared life experiences of certain social groups may lead them to suffer more from diseases like hypertension, cardiovascular disease, and diabetes that directly affect cognitive functioning (see Waldstein, Appendix E, for a discussion of these health effects). For example, people from lower-status social groups, including low-status ethnic minority groups, have poorer health histories, including a higher incidence of chronic diseases that have cognitive effects (Williams, 2000). Similarly, people lacking in social support—resources available through social ties to other individuals and groups—may suffer more from the effects of stress (Caplan, 1974; Cassel, 1976; Cobbs, 1976; Payne and Jones, 1987; Seeman et al., 1996) and experience negative effects on blood pressure and immune function (Uchino et al., 1996). Differences in social support may help account for poorer health outcomes among black Americans, although there are compensatory effects of cultural factors, such as religion (Jackson et al., 1995; Ortega et al., 1983), and the effects may be moderated by demographic variables such as socioeconomic status, marital status, age, and gender.
Another health-related explanation of group differences in cognitive func-
tioning concerns personal control of life outcomes. Lack of personal control in the form of unemployment and working in high-demand, low-control jobs creates stress that sets predisease mechanisms in motion (Karasek and Theorell, 1990; Rushing et al., 1992; Schnall et al., 1994). In addition, beliefs about personal control, which are associated with educational level (Lachman and Weaver, 1998; Marmot et al., 1998; Markus et al., in press), are also associated with a variety of health-promoting behaviors (see Taylor, 1999, for a review).
Stressful experiences may also help explain individual and group differences in cognitive functioning by way of health effects (for reviews summarizing mechanisms linking stressful experience to impairments of health, see Manuck et al., 1990; Seeman and Robbins, 1994; Seeman et al., 1997; McEwen, 1998). For example, several authors have proposed that the experience of racial discrimination leads to chronic stress-induced sympathetic activation that leads in turn to hypertension (see Clark et al., 1999, for a review). Clark and colleagues (1999) have proposed a biopsychosocial model to account for such effects of the experience of discrimination.
Health-related and practice-related variables may also combine to produce intergroup differences in cognitive outcomes. For example, in a recent study of "successful aging," being white predicted greater maintenance of cognitive performance over time (Albert et al., 1995). Education and strenuous activity were intervening factors, suggesting that both of these factors affect cognitive outcomes regardless of race, and that race-related differences in these mediators may help explain between-group differences in cognitive outcomes. The situation is probably even more complex than this. As already noted, the association of educational level with cognitive functioning in old age varies by country. Moreover, even within a country, educational attainment is not the same variable for all groups because they experience systematic differences in the quality of education (e.g., Anyon, 1980; Giroux, 1981; Sieber, 1982; Oakes, 1985; Willis, 1982).
Yet a third set of hypotheses about intergroup differences in cognitive aging is that they are shaped by different cultural meanings of aging. It has often been claimed, for instance, that the dominant American cultural image of aging is one of people who are losing their faculties and are less than fully competent. In contrast, some other societies are said to value aging individuals as having special qualities that entitle them to be treated with added respect compared with their juniors (see, for example, Kitayama, Appendix F). Such social attitudes may affect aging individuals by leading them to think and function in the ways their cultures expect. This social expectation hypothesis is sufficiently plausible to warrant further study.
The above examples indicate that there are numerous plausible hypotheses about how life experiences might yield the observed intergroup differences in cognitive aging; few of these hypotheses, however, have received
detailed examination, and most of the research has focused on children or young adults rather than older people. Yet there is reason to believe that these contextual factors may become even more important with age, affecting ideas and expectations of how to age, norms of when to seek support or help, and decisions about whether to comply with advice, as well as the interpersonal and institutional supports guiding such decisions. Understanding the cognitive effects of life experience may have great practical importance because life experiences can be modified. Improved understanding may therefore lead to promising interventions to improve cognitive outcomes.
Technological Support for the Performance of Cognitive Tasks
Technology has long been used to change the context of behavior to help people adapt to declines in their capability to perform daily life tasks. Eyeglasses, hearing aids, and wheelchairs are among the most obvious examples. Recent developments in information and sensing technology promise to yield revolutionary new technological supports that can help aging individuals adapt to declining capabilities.
As computers become smaller, more powerful, and more easily embedded in other objects and processes (e.g., Norman, 1998), they provide the opportunity to devise new technologies to augment the adaptivity and functionality of the human user. An example is computerized eyeglasses that can enhance the peripheral field of vision (Jebara et al., 1998; see Fisher, Appendix D, for a discussion of additional possibilities). New technologies exist for sensing environmental variables, integrating environmental information (see also Abidi and Gonzalez, 1992), and planning possible actions or facilitating possible decisions in order to make a person's behavior more broadly context sensitive and thereby more adaptive. These technologies hold promise for maintaining the ability of older people to manage such tasks as driving motor vehicles (e.g., Hancock and Parasuraman, 1992) and operating automated teller machines (ATMs) and other technologies in spite of declines in sensory-motor and cognitive capabilities.
New technologies are also becoming available to assist with the information processing aspects of judgment and choice. For example, with the explosion of available information over the Internet, technologies to aid decision-making by reducing information overload are sure to proliferate. Some of these will address important life decisions facing older people, such as choosing health care providers and estate planning. However, these information-reducing technologies must be designed to fit well with users' capabilities and needs.
To make such technologies useful in practice, it is necessary to build understanding of the sensory-motor and cognitive processes the technologies are intended to assist and to address issues of information overload, distribu-
tion of control between the device and the user, and user acceptance. It is also important to take seriously the idea that technology, as part of social context, shapes society and cognition (e.g., Mead, 1953; McLuhan, 1964; Sclove, 1995). The technologies must also be compatible with the users' values and motives. For example, if the values of older adults shift from emphasizing task efficacy to emphasizing emotional connectedness (e.g., Carstensen et al., 1999), that shift may affect their willingness to adopt some of the new technologies intended to increase their adaptivity. It is important to consider such reactions, both to predict which technological innovations may be successfully introduced to the population of older adults and to assess the possible sociocultural consequences of adopting them. It is also important to evaluate the possibility that particular technological supports might undermine cognitive functioning by supplanting the use of mental abilities.
RESEARCH INITIATIVE ON COGNITION IN CONTEXT
The NIA should undertake a major research initiative to understand the effects of behavioral, social, cultural, and technological context on the cognitive functioning and life performance of aging individuals and to build the knowledge needed to intervene effectively in these contexts to assist individuals' functioning and performance.
An appropriate fit between person and context is necessary for effective cognitive functioning and performance at any age. If anything, the need for adaptation is greater in old age than in middle adulthood because of major changes in context that tend to accumulate in late life, such as losses of family and friends, chronic illness and physical decline, migration to new living conditions, and the contemplation of death. The recommended research initiative would promote adaptivity by pursuing three major goals: understanding adaptive processes that affect cognitive functioning during aging; understanding how differences in sociocultural context bring about systematic variation in cognitive functioning and performance; and developing the knowledge needed to design effective technologies to support adaptivity in older adults.
1. Understanding adaptive processes that affect cognitive functioning and performance during aging.
Studies are now needed to clarify the ways aging individuals deploy their cognitive faculties to maintain a high level of performance of life tasks despite decline in some abilities and to improve performance on the basis of accumulated knowledge and wisdom. These studies should focus particularly on: the ways older people rely on other people, technological aids, emotion-regula-
tion skills, and task-structuring strategies; the interdependencies among adaptive strategies, cultural contexts, and biological (including health) conditions; and the possibilities for organizing environments to facilitate aging individuals' use of adaptive strategies. Such research would include the development of highly specified models of cognitive tasks in order to identify the component(s) of a task that are affected by particular contextual factors. Knowledge of contextual factors that contribute to good functioning can provide the basis for remedial measures.
One area in particular need of serious investigation concerns the ways in which older people make decisions in everyday life, an area in which scientific research is presently scanty. A few empirical findings about general cognition have been established. Older adults are less flexible in learning and revising judgment and decision strategies than middle-aged adults, they prefer less cognitively demanding strategies, and they are slower and perhaps more cautious than younger adults (see Sanfey and Hastie, 2000, and Appendix C for more thorough review). In addition, recent research on social cognition and aging has revealed considerable evidence that older people solve problems differently than younger people (Hess, 1994). To the extent that older people appraise, process, and recall information about social matters differently than younger people, such differences involve socially embedded and culture-bound knowledge (Blanchard-Fields and Hess, 1999). Interestingly, the relative old age "profile" includes superior performance in some areas, such as the ability to solve emotionally charged social problems (Blanchard-Fields, 1997; Blanchard-Fields et al., 1997, 1995), and poorer performance in other areas, such as memory about the source of information (Hashtroudi et al., 1990, 1994). Relative to younger people, older people are less likely to revise existing knowledge structures or schemas when presented with new contradictory information (Hess and Pullen, 1994; Hess and Tate, 1991), rendering them more influenced by accumulated knowledge than younger people. There is also growing consensus that some age differences in cognitive performance reflect age differences in goals (Hasher and Zacks, 1988; Hess and Pullen, 1994; Isaacowitz et al., 2000). Older people appear to be particularly sensitive to emotional aspects of situations, including interpersonal ramifications of problems (Carstensen and Turk-Charles, 1994; Kramer, 1990). Thus, processing goals may direct older people to different aspects of problems.
Thus, performance on everyday decisions that rely heavily on past experience or emotional sensitivity may be relatively well maintained in older people, but performance on decisions that require the interpretation of new information may present problems. Important life decisions, such as about health care, estate planning, and personal abilities (whether or not to continue driving, for example) require the integration of rapidly changing information, sometimes in bewildering quantity—a task that tends to be increasingly difficult as people age and as information proliferates about the decisions. Any
serious age-related decrement in the ability to make any of these kinds of decisions wisely will affect not only the decision makers but also their families and perhaps, at least in the case of financial decisions, the larger economy. Despite great interdisciplinary progress in understanding human judgment and decision making, there has been little application to the study of aging (see Peters et al., Appendix C). Studies are needed to illuminate age-related changes in decision-making processes, strategies, and skills and to clarify how these changes are related to cultural expectations, values, and beliefs, as well as to neural health and basic cognitive capabilites. Such studies can help identify the need for decision aids and inform their development. These studies should compare representative samples of populations of different ages to avoid the selection biases that have been present in some studies.
2. Understanding how differences in sociocultural context bring about systematic variation in cognitive functioning and performance.
Empirical research to date shows clear associations between cognitive functioning in old age and social class, education, occupation, language, expertise, and ethnicity. Several kinds of studies are needed to clarify these relationships and address the underlying processes responsible for such differences among individuals and groups.
Unpacking the associations between group membership and cognition . One focus of this research should be on looking more closely at the relationships between rough sociocultural classifications and specific cognitive outcome variables as a way to develop hypotheses about causal mechanisms that may explain the associations (e.g., effects of training and practice, mediation by health effects, stress-related responses). For example, researchers interested in gender differences in cognitive aging, noting the evidence that higher levels of estrogen are related to superior cognitive functioning in aging (C.A. Smith et al., 1999), have considered the potential impact of estrogen and hormone replacement therapy on memory and other cognitive functioning. Although the specific connections between estrogen and cognitive processes and/or other brain functions are still unclear, there is evidence suggesting links of estrogen to neuronal plasticity and to the modulation of neurotransmitter pathways (Costa et al., 1997; Sherwin, 1994). These results suggest that hormone replacement therapy might yield considerable benefit to cognitive functioning in postmenopausal women, but the limited research so far has generated conflicting conclusions (Erkkola, 1996).
As another example, educational attainment, measured as years of schooling, is likely to be a poor proxy for what actually happens in school to shape the mind and brain. Closer examination of the kinds of thinking that are practiced in school may help illuminate intergroup differences in cognition
that persist through the life cycle. For example, children from lower-status social groups tend to be tracked into less demanding academic classes and to be taught to follow directions unquestioningly rather than to exercise critical thinking faculties (e.g., Anyon, 1980; Giroux, 1981; Sieber, 1982; Oakes, 1985). If children actually learn these lessons, there may be specific long-term effects on their ways of thinking. Similarly, there is evidence to suggest that race differences in biological markers are mediated by life experiences that cause lasting physiological stress responses (Clark et al., 1999). Such differences need to be traced in time from the younger ages, where they have usually been studied, into older age.
It is important to avoid the temptation to interpret group differences in cognitive functioning, particularly between social groups of unequal status, as reflecting underlying biological capabilities rather than responses to unequal opportunity or different life conditions (see Cauce et al., 1998). Considerable analysis is usually necessary to arrive at an appropriate explanation. One problem is that some of the usual statistical assumptions underlying between-group comparisons are not met for certain intergroup comparisons (e.g., there is more variance in the cognitive scores of blacks than whites; House et al., 1990, 1994). Another is that particular cognitive measures may not be equally reliable or valid in different cultural/ethnic groups: if an item is not equally familiar across groups, responding may require different amounts of mental effort (e.g., Cauce et al., 1998). Intergroup differences in measures of crystallized abilities may not reflect underlying biological differences because these abilities may be strongly influenced by culture (Cattell, 1963). And, as already noted, intervening variables such as education may not be the same for all groups being compared. Thus, some investigators have argued that an accurate understanding of the meaning of intergroup differences cannot be attained without careful analysis of within-group variation, especially among minority groups (e.g, Markides et al., 1990; Whitfield and Baker-Thomas, 1999). Group labels (race, ethnicity, etc.) are no more than proxies for underlying psychological variables—a useful place to begin research, but not very meaningful until the mechanisms are traced that link observed cognitive differences to causative factors that connect them to group membership.
Understanding the cognitive effects of culturally shared values, beliefs, and practices. This research should describe the relevant values and beliefs (e.g., those concerning age, self, health, relationships, spirituality, technology) and the relevant cultural practices (e.g., communication, health, and everyday activities). It should identify the ways in which these values, beliefs, and practices affect cognitive functioning (including attention, reasoning, memory, and language) and the ways in which cognitive functioning shapes these contexts.
Understanding the neurobiology of life experience. Research should examine and specify the ways that life experiences, operating intensively or over long segments of the life cycle, produce lasting changes in the nervous system. An illustration of this research approach is the recent development of a model for research on the biological effects of the experience of racism (Clark et al., 1999). The model offers plausible links from experienced racism to psychological stress to biological manifestations in the endocrine and cardiovascular systems that may in turn be associated with cognitive functioning. Developments more broadly in psychoneuroimmunology (Ader and Cohen, 1993; Maier et al., 1994) and related fields that link life experiences to health outcomes (e.g., Karasek and Theorell, 1990; Rushing et al., 1992; Schnall et al., 1994; Seeman, 1996; Wang and Mason, 1999) are improving the capability to conduct research on the ways in which behavioral variables affect the complex biological systems that support cognition. Research on the neural effects of training and practice may also suggest causal mechanisms that link life experiences to specific neural changes. Such progress in linking experiential to physiological variables is bringing the field of neurobiology of life experience to the edge of development.
Research in this emerging field should focus on identifying specific kinds and durations of experience that alter the brain in ways that affect the course of cognitive aging and on identifying the mechanisms by which these effects occur. Research under this initiative could include studies of the neural consequences of professional expertise; cognitive training and practice; emotional and motivational activity; education and sociocultural involvement; retirement; changes in family structure; social interactions and social support; experiences associated with social class, race, and ethnic group membership; spirituality; and other experiential factors that may affect cognitive functioning through effects on neural processes. It could also include studies that examine the neural effects of adaptations to the above kinds of experience. Additional studies could test experiential interventions (e.g., types of training) that might affect the brain in ways that help protect against cognitive decline. The studies would include both attempts to establish causal relationships and to develop process models that further clarify chains of causation. They would also aim to specify life experiences that alter the brain, particularly those that protect against cognitive decline and that could be used to prolong the ability of older people to perform cognitive tasks.
It is worth emphasizing that research on the neurobiology of experience turns on its head the usual understanding of how biology relates to behavior. Rather than reducing social categories to neural phenomena, it would attempt to understand how individual and social experiences shape the brain. Progress in this research direction requires new collaborations between social scientists and neuroscientists that focus on a neglected aspect of biology-culture interactions. We believe social scientists will be attracted to studying
these questions in cognitive science because the research will allow them to test and refine hypotheses about the ways sociocultural factors shape human behavior and development.
3. Developing the knowledge needed to design effective technologies to support adaptivity? in older adults.
As already noted, new sensing and information technology holds promise for revolutionary advances in adapting environments to suit the cognitive needs of aging individuals. To achieve this promise, it is necessary to develop a sufficient understanding of sensory-motor and higher-level cognitive functioning in aging individuals to make it possible to design devices and decision aids to work well with individuals whose level of functioning without assistance has declined. It is also important to assess the society-and culture-shaping potential of new adaptive technologies from the perspective of older adults to guard against undesirable secondary effects of the new technology.
To illustrate the need for basic research, consider a computer-controlled device that can provide information to assist an older person in driving a car. To function well, such a device should be capable of identifying what the person is trying to do—for instance, it should be able to discriminate between a pattern of collision avoidance and one of loss of attention or consciousness. Many new control technologies are being developed that can monitor speech, gaze, head movement, gesture, biopotentials, and the like as inputs (NATO Research and Technology Organization, 1998). These technologies would need to be supplemented with analytic techniques, such as hidden Markov models that yield inferences about the person's strategies and goals (see Fisher, Appendix D). Thus, the new devices would need to combine monitoring and control technology, behavioral understanding of the relevant sensory-motor and cognitive processes, and the appropriate quantitative techniques of data processing to provide the right information to the control mechanism. It would also be necessary to address issues of information display, information overload, distribution of control between the device and the user, and user acceptance.
Designing technologies that interact appropriately with the behavioral needs and capabilities of older adults thus presents a significant research challenge. The challenge includes learning how to design technologies to foster and not supplant mental abilities. Meeting this challenge would bring obvious practical benefits and would also advance science by contributing to basic understanding of how older people search, plan, locomote, navigate, and solve problems in technology-aided contexts. More detail on the nature of the research challenges and opportunities can be found in the discussion of adaptive interfaces in Appendix D.
This research also presents a special challenge of implementation because
it requires integrating behavioral science and engineering in a context of product design and development. It will be important to establish good communication between the relevant engineering and behavioral science communities so that technological applications can be designed in tandem with improved understanding of cognitive processes. It will also be important to make prototypes of proposed devices available to researchers espousing various theoretical outlooks to provide good tests both of the devices and of the behavioral theories they apply. Moreover, a successful research program will require viable working relationships between the private-sector organizations that may produce the new technologies and cognitive and behavioral researchers, many of them in universities.
We believe that innovative funding mechanisms will be necessary to encourage basic research in support of technology to extend the adaptivity of older adults. Existing programs have not yet integrated all the necessary elements. For instance, the NIA's Roybal center grants have not thoroughly integrated engineering or focused on building theory for broad application. The National Science Foundation's initiative on Knowledge and Distributed Intelligence has these capabilities but does not focus on aging. We recommend that the NIA consider joint funding of research on adaptive technology with other agencies, such as the NSF, that regularly draw on expertise in technology and engineering.
We also recommend that the NIA explore possibilities to support research by matching industrial support in an appropriate proportion with governmental funding. With this mechanism, industry could benefit from behavioral research that it would not usually conduct, and older adults might see useful adaptive technologies sooner. Also, university researchers would be able to perform basic research in the context of more realistic technological environments than they can usually afford. We recommend that the NIA hold an open meeting with appropriate members of the research and business communities to arrive at a joint plan to fund the needed research on technology for adaptivity and to address related issues of patenting, licensing, protection of proprietary information, and access to scientific results.