U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Research Council (US) Committee on Aging Frontiers in Social Psychology, Personality, and Adult Developmental Psychology; Carstensen LL, Hartel CR, editors. When I'm 64. Washington (DC): National Academies Press (US); 2006.

Cover of When I'm 64

When I'm 64.

Show details

5Social Engagement and Cognition

The ability to maintain mental vitality into late adulthood and function independently are major goals of most older Americans. There is great popular interest in knowing what can be done to preserve good cognitive function and to prevent Alzheimer's disease and other dementias. Although there have been tremendous gains in understanding changes in neurocognitive function with age, there is a surprising dearth of information on social factors that lead to healthy minds in late adulthood. Findings from a few studies suggest that social engagement and absorbing leisure activities may play a role in maintaining cognitive function in late adulthood and even in delaying or preventing dementia. These studies have largely been correlational in nature, so that it is impossible to know whether people who age well have selected engaged life-styles or whether those life-style behaviors themselves promote healthy aging. Caution is also required in assuming that activities that improve cognitive functioning in older adults who do not have dementia will contribute to preventing Alzheimer's disease. However, adopting a more engaged lifestyle or maintaining participation in the work force might be easier behaviors to adopt than some others that lead to healthy aging, such as changing diet and exercise habits, but are very difficult for most people to do.

Our specific focus is the association between social engagement and cognition. There is a recent report on contextual factors that affect cognitive function in late adulthood (National Research Council, 2000), as well as a burst of research activity that is examining mentally engaging leisure activities as a possible protective factor against dementia (e.g., Verghese et al., 2003). However, the construct of social engagement as a behavioral variable supportive of cognitive health has not been systematically considered in either the cognitive or social psychological literature. Although there is a long history of research on the relationship between social ties and psychological and physical health, less is known about the relationship of social engagement to cognitive health. Thus, because it is a promising yet neglected topic at the intersection of social psychology and adult development and aging, more fully investigating the causal relationships between social engagement and cognition is scientifically—and socially—essential.


It is widely accepted that social engagement is related to many positive outcomes in older adults. People with more social ties have been found to live longer (see reviews by Antonucci, 2001; Bowling and Grundy, 1998), to have better health (see Berkman, 1985; Vaillant, Meyer, Mukamal, and Soldz, 1998), and to be less depressed (Antonucci, Fuhrer, and Dartiques, 1997).

A number of factors occur with older age that may increase the likelihood of decreased participation in meaningful social and intellectual activities in older adulthood. Older adults are retired, not raising children, and have fewer positions of authority than middle-aged and younger adults, so that their lives are not necessarily structured to maintain engagement. In a recent study, older adults (up to age 85) reported being less busy than younger adults (age 35 and older), and older adults described themselves as leading more routine and predictable lives (Martin and Park, 2003). While it would be inaccurate to characterize most older adults as plagued by loneliness, irreplaceable relationships are inevitably lost as people age. The differences in the types of social networks that older adults and younger adults have are well-established: as discussed in Chapter 2, older people have a smaller circle of close relatives and friends, particularly those who are important to them and from whom they derive the most pleasure (Carstensen, Isaacowitz, and Charles, 1999).

Social activities do not inevitably lead to meaningful engagement with others. Moreover, social ties are not always positive, and it is important to recognize that increased engagement may not reliably enhance life satisfaction, as engagement with other people has the potential to be stressful as well as supportive (see Rook, 1992). Thus, it is essential that research on social engagement and cognition simultaneously focus on characterizing the relevant features of social relations that facilitate cognition while testing the effects of social engagement on cognition.

Neurocognitive Function and Cognition

As people age, changes in the brain affect cognitive function. There are significant decreases in many behavioral domains of cognitive function. Although knowledge and verbal skills remain strong, there are measurable decreases in the speed of processing information (Salthouse, 1995), working memory function, and long-term memory processes between people in their 20s and people in their 90s (Park et al., 1996, 2002).

The frontal cortex is a large area important for memory and strategic processing of information while the hippocampus, along with prefrontal cortex, plays a critical role in encoding and long-term retention of information. The frontal cortex shows significant shrinkage with age, as does the hippocampus, although not as much as the frontal areas (Moss and Raz, 2001). There is also evidence that older adults have fewer dopamine receptors than young adults (Volkow et al., 1998). Functional imaging studies present a consistent picture of decreased hippocampal activation and increased frontal recruitment during working memory and long-term memory tasks (Gutchess et al., 2005; Park et al., 2003). The increased frontal activation appears to represent the brain's reorganization to compensate for brain shrinkage, decreased hippocampal function, deterioration in white matter, and loss of dopamine receptors (Cabeza, 2002; Logan et al., 2002; Reuter-Lorenz, 2002). These studies have been cross-sectional and include an age range from college students to those over 85 years old. An important question is what kinds of social and intellectual experiences might support neural function and optimize compensatory activation. Also worth consideration is whether there are any differences in experiences in successive birth cohorts—for example, differences in educational experiences (Castro-Caldas, Petersson, Reis, Stone-Elander, and Ingvar, 1998), exposure to television, and the structure of society—that might contribute to the observed results in brain function.

Evidence indicates that the brain has developed specialized mechanisms for dealing with the social environment. The amygdala appears to be important for using social cues, such as judging whether someone is trustworthy (Adolphs, Tranel, and Damasio, 1998). That is, there are dissociable brain regions that are involved in social and emotional processing. It has further been suggested that aging may have a differential impact on hippocampal as opposed to amygdala function (Denburg, Buchanan, Tranel, and Adolphs, 2003). This finding would help to explain why emotionally significant stimuli—especially positive—are readily remembered by older adults (Charles, Mather, and Carstensen, 2003) and may lead individuals to adopt strategies in which they use relatively more preserved brain regions to support those that are less preserved.

In contrast to normal cognitive aging, dementia is defined as declines in multiple cognitive domains sufficient to interfere with daily functioning. Alzheimer's disease is the most common cause of dementia (Cummings and Cole, 2002). A variety of terms have been applied to intermediate conditions in which there is cognitive decline that is not sufficiently severe to be diagnosed as dementia (e.g., mild cognitive impairment). By age 85, approximately 50 percent of individuals experience significant memory loss (Graham et al., 1997), and 25 percent show dementia that ultimately limits their ability to manage independently (von Strauss et al., 1999). Determining whether heightened social engagement can defer cognitive aging and maintain more years of independent functioning is therefore a critical public health issue.

Mental Stimulation and Cognitive Aging

The relationship between mental stimulation and cognitive aging is currently an active area of investigation. This work has implications both methodological and substantive for the study of social engagement. Engagement is a behavior that involves a high level of both intellectual and social function. There is a growing literature (based on longitudinal data) suggesting that leading an intellectually stimulating life may foster cognitive vitality (Hultsch, Hertzog, Small, and Dixon, 1999; Schooler, Mulatu, and Oates, 1999). Other literature suggests that cognitive stimulation in everyday life can even be an important buffer in delaying onset of Alzheimer's disease or slowing its progression (Wilson, Bennett, Gilley, Beckett, Barnes, and Evans, 2000). It is quite well established that higher education serves a protective effect with respect to dementia. What is not known is the mechanism for this association, although it may be that those with higher education have generally more cognitively stimulating lives (Wilson and Bennett, 2003).

It is important to differentiate the construct of engagement from that of cognitive training. Cognitive training is a manipulation that is a systematic attempt to improve cognitive function through focused training activities. The training literature involves traditional laboratory-based manipulations and resides almost exclusively in the realm of cognitive psychology and is thus considered only briefly here. In contrast, the engagement literature is drawn primarily from social psychology and large epidemiological studies. Overall, the training literature suggests that specific, but not general, increases in cognitive function can be realized with cognitive training. For example, memory training with older adults (who do not have dementia) from their 60s through their 80s can be effective if—in addition to teaching mnemonic techniques—the intervention provides information about memory and aging and includes cognitive restructuring (i.e., altering pessimistic beliefs about memory abilities) and stress reduction, which, if untreated, can reduce scores on memory tests (Lachman et al., 1992; Yesavage, 1984). More recently, Ball and colleagues (2002) completed a major randomized clinical trial with older adults up to age 95 that showed modest effects of training on memory and perceptual speed. The effects of training on cognition tended to be very specific, and there was little evidence for broad improvement in everyday cognitive functioning after specific cognitive tasks were trained (Ball et al., 2002). Notably, this trial included a large African American sample as well as European Americans.

Training designs have also been applied to older adults, ages 70 to over 95, with dementia. For example, Camp and colleagues (Bourgeois et al., 2003; Camp et al., 1997) showed that cognition can be improved in older adults with dementia through training that uses the types of learning that are least affected by aging in the brain and through training in the use of appropriate aids to memory. These interventions sometimes include a social component, such as mentoring of children by adults with dementia, conducting training in group settings where participants are asked to engage in conversational exchanges, and so forth.


Studies that suggest that social engagement and socially absorbing leisure activities play an important role in maintaining cognitive function in late adulthood are not as well developed as those connecting mental activation to cognitive outcomes. The data on engagement and cognitive function are relatively limited and restricted to epidemiologic or longitudinal studies in which engagement is self-selected, and there are few, if any, using clinical trials or random assignment. Greater social engagement—measured in terms of contact with family and friends and participation in social activities—has been shown to reduce risk of cognitive decline over 3, 6, and 12 years (Bassuk, Glass, and Berkman, 1999). Another study suggests that individuals (average age = 65) who engage in many social interactions or who have large social networks show better cognitive function than those who do not (Arbuckle, Gold, Andres, Schwartzman, and Chaikelson, 1992).

Better social connectedness has also been shown to be related to reduced risk of subsequently developing dementia (Fratiglioni, Wang, Ericsson, Maytan, and Winblad, 2000). And individuals who are married and presumably engage in more social interactions than single people are also less susceptible to dementia (Helmer et al., 1999). A number of studies show that more engagement in leisure activities is related to reduced risk of dementia or Alzheimer's disease (Crowe, Andel, Pedersen, Johansson, and Gatz, 2003; Fabrigoule, Letenneur, Dartigues, Zarrouk, Commenges, and Barberger-Gateau, 1995; Scarmeas, Levy, Tang, Manly, and Stern, 2001; Wang, Karp, Winblad, and Fratiglioni, 2002). These studies include international samples, and U.S. samples with African American and Latino participants. In most of these studies, leisure encompasses a range of activities that may have social, intellectual, or physical components, so it is not entirely clear which component is responsible for the protective effect. However, Wang et al. (2002) did demonstrate that engagement in social types of leisure activities (theatre, cards, traveling) was related to reduced risk of dementia. A summary of the literature on social engagement and dementia is available in Fratiglioni, Paillard-Borg, and Winblad (2004).

Other research has examined social and intellectual engagement in occupations rather than leisure. Schooler et al. (1999) reported that individuals who engaged in complex jobs that required decisions and engagement showed better intellectual functioning than adults who had less demanding jobs. In a similar vein, data from the Maastricht Longitudinal Study indicated that individuals with cognitively demanding jobs showed less cognitive decline over a 3-year period than those with less demanding jobs (Bosma et al., 2003). Stern et al. (1995) found lower risk of dementia among those whose occupations had high interpersonal and physical demands. Although randomized clinical trials to test the effect of social engagement on cognition and dementia are not available (see Fratiglioni et al., 2004), interventions whose purpose is to strengthen social ties and improve social support are common in long-term care institutions and senior citizen community settings. Randomization to conditions is unusual. Cognitive tests may be included in the outcomes, and follow-up tests of participants in some of these programs have shown improved memory: for example, Newman, Karip, and Faux (1995) found improved memory in older adults in their 60s and 70s who participated in a volunteer program.

Underlying Mechanisms

The causal mechanisms underlying the cognitive benefits that accrue from social and even intellectual stimulation are unknown. There are many possibilities. One is that social and intellectual stimulation alters neural tissue and pathways. Some evidence for this comes from animal studies in which rats raised in social environments showed greater brain volume than animals raised alone (Bennett, Rosenzweig, and Diamond, 1969). More recently, studies have indicated that older animals in environmentally complex, stimulating environments showed both enhanced learning and increased neural tissue and dendrite growth in old age (Greenough, McDonald, Parnisari, and Camel, 1986; Greenough, West, and DeVoogd, 1978). Other work has suggested that neurogenesis (birth of new neurons) occurs in hippocampal structures and dentate gyrus in old mice that are exposed to an enriched environment (Kempermann, Kuhn, and Gage, 1997, 1998). Finally, there is recent research suggesting that physical exercise in older humans enhances both cognitive function and prevents age-related brain atrophy (Colcombe and Kramer, 2003; Colcombe et al., 2004). These data suggest that neurocognitive function in late adulthood is malleable, and they are consistent with the popular “use it or lose it” argument in American culture (a hypothesis described as the “disuse hypothesis” by Orrel and Sahakian, 1995). In other words, by stimulating neural pathways, the health of the brain is maintained. It is plausible that social engagement could provide sustained neurocognitive stimulation, possibly due to the high degree of comprehension, memory, and problem solving required to manage and sustain social relationships (Byrne and Whiten, 1988). It is also plausible that racial, cultural, and ethnic groups that place a high value on social engagement might provide a cognitive advantage to individuals in those groups over others in more isolated or individualistic groups.

Another plausible interpretation of engagement effects is that highly social individuals in late adulthood are healthier than less engaged individuals. It may be that engagement effects in old age, although real, do not result from a causal relationship, but that individuals with greater physical health also have better cognition and engage in more activities than their frailer counterparts. Because the data collected thus far are largely from studies in which individuals have self-selected whether to be socially active, this interpretation cannot be ruled out until experimental studies of engagement are conducted in which volunteers are randomly assigned to participate in engaging activities (which could include the individual's choosing which social activities to engage in) or in an appropriate control group.

Another related explanation for the observed effect of social engagement on cognition may be long-standing differences in individuals that are responsible both for their greater social engagement and for their being relatively protected from cognitive decline and dementia. Results from the nun study (Danner, Snowdon, and Friesen, 2001; Snowdon, Greiner, and Markesbery, 2000) and from Scottish studies of birth cohorts who received a national survey of mental ability (Whalley, Starr, Athawes, Hunter, Pattie, and Deary, 2000) demonstrated that there are notable differences early in life—in language, emotions, and intelligence—that predict cognitive outcomes in late life. These differences imply that both social engagement and better cognitive function in late life stem from these earlier influences.

Another possible explanation is that social engagement prevents the onset of depression, although some studies have controlled for depressive symptoms and still found protective effects for leisure activities (e.g., Bassuk et al., 1999). An extension of this argument is that social engagement leads to reduced stress. Depression and stress can lead to negative changes in cognitive and neural function, including loss of neurons in the hippocampus through glucocorticoid dysregulation (Sapolsky, Krey, and McEwen, 1986). Thus, the observed effect of social engagement in maintaining cognitive function could be due to the positive effects of social engagement on mental health.

Social engagement effects, particularly those associated with evidence for later onset of Alzheimer's disease, may also be understood in terms of how engagement might foster more assistance with cognitive maintenance. That is, individuals who are engaged may engender a larger network of collaboration and support for tasks for which individual cognitive resources may be insufficient. In line with this thinking, Dixon and colleagues (Dixon, Hopp, Cohen, de Frias, and Bäckman, 2003) have suggested that compensatory mechanisms for declining cognition can operate at the social level. Recruitment or engagement of social partners by older adults to consult or to work together in managing problems of everyday living may compensate for declining cognition and help maintain independence. Dixon and Gould (1998) observed that older couples in their late 60s and 70s, with a history of social interaction, were particularly adept at developing specific compensatory strategies to manage problems of everyday living, while dyads who did not previously know one another were only able to construct mechanisms of general social and emotional support. Here, again, cultural differences in endorsement of communal versus individual values may affect social engagement and, hence, cognitive maintenance.

Finally, the manner in which social engagement may foster cognition could be viewed through the work of social psychologists on the role of social context on performance. Zajonc's (1965) theory of social facilitation and subsequent work identifying conditions under which social facilitation occurs (Bargh and Apsley, 2001) provides a framework for understanding when social engagement might provide cognitive benefits. While learning new skills can be more difficult with an audience, well-practiced skills are performed better in the presence of others. The importance of social context may also be understood in terms of its potential negative effects on cognitive function through the operation of age-related stereotypes (see Chapter 6). Thus, social contexts that emphasize age-related declines while demanding acquisition of new skills could diminish rather than enhance cognitive function.

Technology Training as Engagement

Technology plays an increasingly central role in contemporary lives, and it can be an important tool in maintaining independence in the face of physical frailty, as shopping, banking, and many everyday tasks can be managed with computer technology (Liu and Park, 2003). One provocative and important possibility is that technology training of older adults could have multiple benefits. One obvious benefit is in providing older adults with more tools to help themselves in terms of health information, medication refills, financial management, and shopping (Park, 1997). A second benefit is the potential that technology offers for people to interact easily with friends, family, and special interest groups, providing a particular type of social engagement about which little is known. Older adults report that one of their primary uses of the Internet is maintaining contact with family (Loges and Jung, 2001). One obvious question is whether such remote interactions confer the same benefits that other social interactions offer in terms of both socioemotional and cognitive well-being. A third potential benefit is the possibility that the act of acquiring technology expertise in older adults is socially and cognitively stimulating and enhances cognitive vitality, independent of the effect of using the technology. Although some research has been done on the human factors (Czaja, 2001) and cognitive aspects (Rogers and Fisk, 2001) of technology acquisition (see also National Research Council, 2003), the possibility that technology training results in cognitive gains beyond the acquisition of specific skills because of increasing engagement is worthy of further exploration. Given the multiplicative potential for benefits from learning to use may types of technology, the social and cognitive benefits of technology training and use in older adults is an area in which investigation should be encouraged. In this regard, the gap in access to technology across different communities is a topic that warrants attention.


There is a need to sort through the various mechanisms that have been postulated about the effects of engagement. First, it is important to know if social engagement does have a measurable effect on cognition and maintenance of independence in late adulthood, both independently of cognitive stimulation and in concert with it. There have not been any controlled experimental studies examining the effects of sustained social interactions on immediate and longer term changes in cognitive function, although there are sufficient data from epidemiologic studies to suggest strongly that high social function is protective of neurocognitive processes. It is important to assess not only the effects of social interactions on cognitive function, but also the magnitude and duration of such effects and whether these effects have practical value. Controlled investigations on this topic could include development of methodologies that permit characterization and manipulation of social engagement that will lead to concrete recommendations for the types of leisure activities and social interactions that promote healthy intellectual function in late adulthood. Of particular importance is determining the dose-response relationship of sustained social activity for healthy aging, as well as whether there are critical periods across the life span at which social engagement has the greatest effect. The topic of engagement raises several important questions, which are discussed in the rest of this section.

Does engagement augment cognitive function or delay disease onset? Whether any protective effect of social engagement applies chiefly to improved cognition within normal aging or whether it extends to dementia in general or Alzheimer's disease in particular, where it has been argued that social engagement might prevent or delay the onset or affect the course of the disease, must be considered. The question of whether the onset of dementia can be delayed is particularly difficult to investigate, and we do not want to underestimate the profound difficulties in establishing causal direction from engagement to cognition. Because dementia has a long and insidious onset, cognitive changes may begin to occur well before there is clinical recognition of any dementia (Elias et al., 2000). Thus, it is particularly important to be sure that reduced social interaction is not a consequence of the beginning of cognitive decline. Some analytic innovations may be helpful with epidemiological datasets in which risk factors can be both predictors and outcomes of subsequent exposure (Robins, Hernan, and Brumback, 2000; Tilling, Sterne, and Szklo, 2002). Randomized trials will be important in establishing connections between engagement and cognition without the selection effects that plague observational studies. Various strategies can help to optimize these studies, including extended follow up or sampling those most at risk of cognitive impairment.

What types of engagement are most effective? A second major research domain of importance is determining what aspects of social engagement may lead to cognitive health. Social engagement can take many forms, ranging from close familial relationships and interactions, including family caregiving, or other culturally sanctioned social activities, to participation in novel activities that result in acquisition of new behavioral repertoires and ideas. Research that specifically characterizes the types of social engagement that are facilitative of cognitive function is needed. It is also important to understand the role of social networks across the life span in maintaining cognitive function, for example, whether the size of social networks is important, or whether it is the extent of social engagement, regardless of network size, that matters. Due to decreased cognitive function associated with normal aging (Park et al., 2001), older adults may prefer more limited social networks. Alternatively, decreased social networks could hasten normal age-related cognitive decline, in which case efforts to increase social engagement might include increasing social networks.

What is the role of individual differences? A third important question involves the mediating role that individual differences may play in the relationship between cognitive function and social engagement. Individual differences include physical health, which may limit some kinds of social activities but need not preclude meaningful social engagement. Personality differences may also be relevant: for instance, social engagement may have different meanings for those who are more introverted. The effects of social engagement may be most pronounced for people with limited opportunities for social engagement because of the unusual stimulation such engagement would provide. Alternatively, the effects of social interactions may be most salient for those who have a long-standing interest in social engagement and even be unhelpful for others. Demographic variables such as socioeconomic status, education, gender, and marital status might reveal important individual differences, suggesting mechanisms for facilitative effects. These individual differences may point to ways of tailoring activities to achieve social engagement and providing opportunities for engagement in a context in which older people can engage in choice and self-regulation with respect to those activities and relationships.

What is the role of cultural and ethnic factors? Cultural context and ethnicity may play roles in minimizing or magnifying the effects of social engagement on cognitive health. Characterizing cultural differences in social engagement with respect to age will help understanding how different forms and styles of social connectedness interact with cognitive health.

It is a relatively consistent finding that there are differences among cultural, racial, and ethnic groups in cognitive performance in later life. In the United States, cross-racial studies of performance on standardized tests of neuropsychological functioning indicate that European Americans, on average, have a higher level of performance than other groups (typically African Americans) (Escobar et al., 1986; Fillenbaum et al., 1988). Recent research suggests that these mean differences in cognitive performance can be entirely or at least partly accounted for by social factors, such as education, literacy, and health status (e.g., Albert et al., 1995; Manly, Jacobs, Touradji, Small, and Stern, 2002; Whitfield et al., 2000). There may also be differences in familiarity with the types of information presented in the tests, which represent a source of difference in mean performance between groups (Whitfield, 1996; Whitfield et al., 2000; Whitfield and Willis, 1998). These data do not suggest that cognitive functioning differs fundamentally as a function of race or ethnicity; rather, since the cultural contexts in which cognition operates are relatively stable over dozens of years, the hypothesis that culture may affect basic cognitive processes is not implausible. The role of social engagement in various cultural contexts is a question that needs to be answered in order to establish the relationships among cognition, culture, and social engagement in later life.

What types of stimulation alone, or in combination, are effective? Another important question concerns the interrelationship among different types of stimulation—intellectual, social, and physical. There is increasing evidence that physical exercise may improve cognitive function (Colcombe and Kramer, 2003), and it is possible that physical and intellectual stimulation interact with social engagement, conferring maximal neurocognitive protection. Alternatively, perhaps it is the quantity rather than the type of stimulation that makes the difference. This is an entirely unexplored question that is worthy of systematic evaluation.


The relationship between social engagement and cognitive health deserves serious research attention because of its potential implications for successful aging in America. There is a broad array of techniques for measuring the interactions between social engagement and cognitive health. Specifically, from the neuroimaging literature there is increasing evidence that older adults may use different neural routes to perform a cognitive task than young adults do (Cabeza, 2002; Park et al., 2001; Reuter-Lorenz, 2002) and that there is more plasticity in neural pathways than was previously understood. The use of neuroimaging tools to understand the effect of engagement manipulations on cognition is important and may reveal other effects as well. Research in both the social and the cognitive psychology of aging has provided rich theoretical models that can be extended to understand the role of social engagement in productive aging. Even more broadly, a full complement of sociological, behavioral, physiological, and neurobiological measures are available to measure both neural and behavioral outcomes of associations between social engagement and cognitive function. Furthermore, the outcome of research on motivation and behavior change and on socioemotional influences on decision making will directly affect the design of any interventions needed to increase social engagement.

Understanding the mechanisms underlying the causes and effects of social engagement, and discovering the types of activities that might maintain and even improve cognitive function is a major issue for public health policy and for the lives of the older population. Identifying culturally appropriate programmatic interventions that would make a contribution to delaying the onset of dementia would have major effects on the number of cases and on the demands on the health care system and on families.

Copyright © 2006, National Academy of Sciences.
Bookshelf ID: NBK83766


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (979K)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...