| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
RESEARCH ARTICLE |
Department of Psychology, University of Oregon, Eugene.
Address correspondence to Kirby Deater-Deckard, Department of Psychology, 1227, University of Oregon, Eugene, OR 97403-1227. E-mail: kirbydd{at}darkwing.uoregon.edu
 |
Abstract |
|---|
 TOP Abstract Successful Aging in the...Gene-Environment Correlation and...Empirical MethodsTheoretical FrameworksConclusionReferences |
|---|
environment theory of Scarr and McCartney and the selection/optimization/compensation theory of Baltes and Baltes.
STAVING off the effects of cognitive decline is critical to successful aging. From 20 to 70 years of age, cognitive performance decreases by about 1.5 SD among healthy individuals (e.g., Salthouse, 1991
). This is at least partly responsible for declining productivity seen among individuals in intellectually demanding professions (e.g., Charness, Krampe, & Mayr, 1996; for other possible factors, see Simonton, 1997). Furthermore, more pathologic decrements in cognitive functioning, such as the dementia arising from Alzheimer's disease, carry huge costs for individuals, families, and society. In the current article, our goal is to consider some of the ways in which genetic and environmental influences work together to cause changes in individual differences in cognitive skills and in executive functions in particular.
|
SUCCESSFUL AGING IN THE COGNITIVE DOMAIN |
|---|
|
TOPAbstractSuccessful Aging in the...Gene-Environment Correlation and...Empirical MethodsTheoretical FrameworksConclusionReferences |
|---|
]). Heritability can be estimated in a narrow sense (to include only additive genetic effects) or in a broad sense (to include additive as well as nonadditive effects, i.e., dominance, epistasis) (Loehlin, 1992).
The growing quantitative genetics literature shows that genetic influences on cognitive skills increase in magnitude over childhood, level out in adulthood, and then may decrease late in life. Over the entire life span, most of the environmental effects on cognitive performance are nonshared, with shared environmental influences limited to early childhood and perhaps old age (Finkel & Pedersen, 2001; McCartney, Harris, & Bernieri, 1990). Furthermore, individual differences in the degree of change in performance appear to be driven more by nonshared environmental than genetic influences (McGue & Christensen, 2001, 2002; Reynolds, Finkel, Gatz, & Pedersen, 2002). To our knowledge, there are no data regarding genetic and nongenetic influences on executive control in aging, a noteworthy gap that we turn to in some detail later in this article.
Dimensions of Cognitive Aging
When describing changes in cognitive functioning in adulthood, there is an important distinction to be made between fluid/mechanic abilities and crystallized/pragmatic abilities (e.g, Baltes, 1997; Horn & Cattell, 1966). In theory, fluid abilities reflect the biologically constrained hardware of the cognitive system and thus are assessed by tasks that ideally are immune to effects of experience and knowledge. In contrast, crystallized abilities reflect lifelong learning and acculturation. Usually, fluid abilities show gradual decrements across the life span, starting with age 20, supposedly due to biological factors. In contrast, crystallized abilities show stability into the 70s, with some decline thereafter. At least at first glance, this distinction may suggest the obvious prediction that environmental influences should be more strongly reflected in knowledge-dependent, crystallized abilities. In fact, existing data show, if anything, the opposite, with somewhat higher estimates of genetic influences on crystallized than fluid abilities (e.g., Pedersen, Plomin, Nesselroade, & McClearn, 1992; Tambs, Sundet, & Magnus, 1984). It may be the very fact that crystallized abilities reflect the accumulated effect of lifelong learning that actually makes them particularly reliable indicators of genetic influences of learning ability. In contrast, fluid abilities may provide a better snapshot of current processing efficiency.
When focusing on aspects of intellectual functioning that do decline (i.e., mainly fluid abilities), another important finding to bear in mind is that age-related differences seem to be accounted for in terms of one, or maybe two, very powerful "aging factors." For example, most of the age differences in a broad range of cognitive abilities can be accounted for in terms of age differences in processing speed and/or working memory (e.g., Lindenberger, Mayr, & Kliegl, 1993; Verhaeghen & Salthouse, 1997). More recent evidence suggests that any measure that reflects the integrity of the central nervous system (e.g., age differences in sensory functioning or balance) can account for age differences in cognitive abilities (e.g., Lindenberger & Baltes, 1994). From a quantitative genetic perspective, it is critical to examine the degree to which the one or two aging factors (if they exist) are influenced differentially by genetic and environmental effects. So far, this has been done only rarely. Usually, researchers simply report heritability estimates separately for different tasks or psychometric factors and how they change as a function of cohort or age. This does little to answer the question of the degree to which the factors underlying cognitive change across functions is influenced by genetic or environmental factors.
One important point to note is that cross-sectional designs are not sufficient for examining the dimensionality of change from a quantitative genetic perspective. Cross-sectional designs do not provide estimates of change in cognitive functioning that can be submitted to a genetic analysis. In addition, in any type of study, cross-sectional data confound age-related and interindividual differences (Lindenberger & Poetter, 1998) and therefore may lead to incorrect estimates of the amount of variance that is common to both measures of functioning and age. The ultimate answer to the question of whether cognitive aging is a general factor or a multifaceted phenomenon will come from careful longitudinal data that allow for testing whether changes in one ability explain changes in other abilities (e.g., Ghisletta & Lindenberger, 2003; Hertzog, Dixon, Hultsch, & MacDonald, 2003; Sliwinski & Buschke, 1999).
In contrast to their cross-sectional counterparts, the few available longitudinal studies suggest more modest estimates of a general change factor (e.g., Sliwinski & Buschke, 1999). At the same time, it is difficult to know the degree to which these longitudinal studies have allowed reliable measurement of change, due to assessment over brief time spans and the possible contamination of data through learning effects. It is interesting that more recent structural equation models have shown a trend toward somewhat higher estimates of shared change scores across domains of functioning (e.g., Ghisletta & Lindenberger, 2003; Hertzog et al., 1999). It will be critical to combine longitudinal studies that allow uncovering the dimensionality of change across a wide range of cognitive abilities, on the one hand, with quantitative genetic analyses, on the other. Such work will be essential to answering questions regarding to what extent—and, ultimately, in what ways—genetic and environmental influences cause change in the one or more cognitive aging factors.
Whereas it is certainly important to carefully relate the dimensions underlying age-related change to genetic and environmental factors, we believe it will be also very useful to take a functional perspective toward the gene–environment question in the context of age-related changes in cognition. In this regard, it is important to ask which abilities (a) are likely to be affected through experience and interventions and (b) at the same time are likely to exert broad, domain-general effects. Little attention has been paid to this perspective, and our proposal in this regard is speculative. However, we believe that there are initial reasons for focusing on a cluster of abilities often referred to as executive control or executive attention. Executive control functions are responsible for organizing basic processes and resources in a goal-directed manner (e.g., Posner & DiGirolamo, 1998; for an overview with a focus on aging, see Mayr, Spieler, & Kliegl, 2001). These functions become particularly critical when dominant behavior patterns need to be changed. Given their role in realigning information processing to changing demands, executive functions should be of primary importance whenever active compensation (e.g., age-related losses in basic functions) becomes necessary. Furthermore, by most accounts, executive functions subserve efficiency of information processing in a domain-general manner. Thus, effective interventions that target executive functions would have the potential for generalizable effects on cognitive efficiency. This is a central point that we emphasize in some detail below in the section on "candidate" environmental factors.
|
GENE–ENVIRONMENT CORRELATION AND NONSHARED ENVIRONMENTAL MECHANISMS |
|---|
|
TOPAbstractSuccessful Aging in the...Gene-Environment Correlation and...Empirical MethodsTheoretical FrameworksConclusionReferences |
|---|
provided an integrative perspective on the ways in which genetic and environmental influences play out in development over the life span. Genetic influences operate, in part, through their effects on environments (via gene–environment correlation), and these correlated environmental factors can powerfully moderate genetic influences on phenotypes (gene–environment interaction): heritable attributes are modifiable; environments are controllable (in general, particularly when compared to genes). It is time for developmental science to move beyond descriptive variance estimation of heritability, shared environment, and nonshared environment and move toward precise specification of these gene–environment mechanisms. The presence of robust, and possibly increasingly influential, nonshared environmental factors is an obvious indicator of the presence of powerful and potentially identifiable environmental lifestyle- and health-related factors that causally influence cognitive change. Furthermore, the presence of moderate to substantial heritable influences does not rule out environmental influences, because many of these genetic and environmental factors are correlated and operate in conjunction.
Accordingly, one of the tasks is to identify the specific environments and their roles in the functioning of correlated genotypes. Furthermore, because the majority of environmental influences appear to be nonshared in their effects (i.e., leading to differentiation between family members), we also must assess these environmental influences in a way that is sensitive to within-family differentiation (Plomin, 1994). We now turn to a more detailed discussion of these gene–environment correlations and nonshared environment mechanisms.
Gene–Environment Correlation
Genetic and environmental influences on a phenotype can be correlated with each other. Much of this effect is subsumed within the heritability estimate in a traditional quantitative genetic analysis, unless the gene–environment correlation (rg-e) effect is measured and specified. Nevertheless, quantitative genetic techniques can be extremely useful for specifying rg-e influences. Testing for rg-e often points to some of the ways in which the genotype can influence the content of individuals' experiences (Plomin, 1994).
Two broad types of rg-e include passive and nonpassive effects. Passive rg-e is defined as exposure to an environmental factor that is being provided by another individual who is also genetically related to the individual. For example, consider a parent who is above average in cognitive performance ability. This parent is more likely to provide higher amounts of cognitively stimulating materials and interactions for her or his child. On average, the child also will show above average cognitive performance. However, these cognitive skills are, in part, genetically influenced, and the parent also is transmitting genetic influences to the child. Because the parent and child are genetically related, the cognitively enriched environment and the genetic factors that are associated with better cognitive performance are confounded.
Nonpassive rg-e includes active and evocative/reactive person–environment transactions. Active rg-eis synonymous with the self-selection of environmental inputs, whereby the individual experiences certain things as a result of her or his genetically influenced skills or attributes. For example, an adult with genes that positively influence cognitive skill and capacity may be more likely to find problem solving inherently rewarding and is therefore more likely to seek out cognitive challenge. This is exemplified in data showing a connection between improvements in cognitive functioning later in life and the complexity of paid and unpaid work (defined as activity that requires "thought and independent judgment" [Schooler, Mulatu, & Oates, 1999]) and leisure activity (Schooler & Mulatu, 2001). There is evidence of bidirectional influences between cognitive functioning and activity complexity, with each influencing the other. To the extent that genetic influences on cognitive functioning operate, in part, through the individual's selection of work and leisure activities that are cognitively demanding and stimulating, this would represent a nonpassive rg-e mechanism.
Evocative or reactive rg-e is synonymous with elicitation of environmental inputs from the social milieu, whereby individuals experience certain things because their genetically influenced attributes consistently evoke similar reactions from other people. For example, individuals who are aggressive and hostile in their interactions with others are exposed to many more instances of rejection and hostility from others, which in turn serve to reinforce these negative social behaviors (Anderson, Lytton, & Romney, 1986). These noxious behaviors are heritable, and part of this genetic influence operates through the elicited negative reactions of others (O'Connor, Deater-Deckard, Fulker, Rutter, & Plomin, 1998).
Nonshared Environment
All, or nearly all, of the nongenetic variance in cognitive functioning is attributable to nonshared environmental mechanisms. These are the nongenetic influences that cause family member differentiation, rather than similarity. Even when parent–child and sibling pairs are living together in the early stages of the family life cycle, each individual has many experiences that are unique. These can be seen as family member differences in relationships within the same family (e.g., mother–daughter) and outside of the family (e.g., peers and romantic partners). In addition, the very same event or experience—for example, a parental divorce—can be perceived in very different ways and have very different effects on the outcomes of those who are affected. These within-family differences can be quite large, and it is possible to reliably assess them and test whether these differences have predictable effects, while controlling for potential genetic influences on family member differentiation (Plomin, 1994).
Adult siblings' own "created" home environments are surprisingly different from each other, even though they come from the same family of origin. Schaie and Zuo (2001) found that adult siblings' reports of their current family environments pointed to substantial differences in levels of family cohesion and conflict. At the same time, the siblings' retrospective reports of their home environments from childhood (as well as parent–adult offspring retrospective reports) pointed to a substantial consensus about the nature of the environment in the family of origin. Outside of the home environment, there may be sibling differences in the cognitive complexity of occupations and leisure activities that manifest themselves as sources of nonshared environmental influences on cognitive changes throughout adulthood (Schooler & Mulatu, 2001; Schooler et al., 1999). These are just a few examples—in fact, there are many potential sources of nonshared environmental influences.
It is possible that these nonshared environmental mechanisms are highly idiosyncratic and nonsystematic in their effects. It also is possible that the finding of a substantial and pervasive nonshared environment variance estimate merely points to substantial and pervasive measurement error. These are viable explanations (although far less interesting ones, from a scientific perspective) that must be ruled out and can be ruled out only if candidate environmental factors are assessed and examined within genetically informative designs.
|
EMPIRICAL METHODS |
|---|
|
TOPAbstractSuccessful Aging in the...Gene-Environment Correlation and...Empirical MethodsTheoretical FrameworksConclusionReferences |
|---|
).
All that is required is careful measurement of the outcomes and environmental factors of interest and a quantitative genetic design in which parent–offspring and/or sibling similarity is estimated for subgroups of pairs that differ in their genetic similarity. This includes twin, adoption, and stepfamily studies in which multiple family members are assessed. Although these studies are usually correlational, the designs also can be used for experimental studies. Very little of this kind of work has been done in aging research. Therefore, we describe several examples from research in childhood. Tests of these sort are much more common in studies of child and adolescent development (for a recent overview, see Reiss, Neiderhiser, Hetherington, & Plomin, 2000) and provide useful templates for how these empirical methods can be applied to answer questions regarding gene–environment mechanisms in successful aging.
Tests of Gene–Environment Correlation Mechanisms
The sibling quantitative genetic design (e.g., comparison of identical and fraternal twins or adoptive and nonadoptive siblings) can be used to test for the presence of gene–environment correlation by investigating whether there are signs of genetic variance in measured environmental factors. For instance, if nonpassive rg-e is present, it means that genetic influences on particular attributes are further enhanced by reinforcement from correlated environmental influences.
Consider as an example the research on warm, supportive family relationships, which have been identified as robust protective factors for a number of developmental outcomes (Conger & Conger, 2002). Nearly all of the studies on parent–child warmth have examined parenting and child outcomes for only one parent–child dyad in the family. However, when a parent's relationships with two children in the same family are examined, the warmth and acceptance in each parent–child dyad differ, often markedly (Dunn, 1993).
Because individuals behave in ways that consistently elicit certain responses from other people, some of the genetic influences on traits of interest may operate, in part, through these evoked effects. When evocative rg-e is detected using a child genetic design in which there is genetic variation between siblings (e.g., observing twins interacting with the same parent), finding genetic influences on aspects of parenting behavior provides some evidence that a child effect may be operating. There have been a number of studies that have documented evidence of evocative rg-e in parent–child relationships and interactions, although few of these have used observational measures (Reiss et al., 2002).
We have found evidence of evocative rg-e for a number of child outcomes and components of parent–child interactions and relationships, using both parents' reports and observers' ratings (Deater-Deckard & O'Connor, 2000; Deater-Deckard & Petrill, 2004). In one twin study of 3 year olds, mothers completed a questionnaire regarding their feelings about each of their twin children and also rated each child's behavioral and emotional problems. We found that overlapping genetic variance accounted for a significant proportion of the correlation between parent-rated negativity/positivity toward the child and child behavioral and emotional problems. This suggests that genetically influenced child attributes may be eliciting or evoking feelings of parental negativity and warmth.
These rg-e effects are not limited to participants' reports. We have found a similar result in the examination of very brief observations of parent–child dyadic mutuality (i.e., joint positive affect, coherent and well-organized interaction, cooperation, co-responsiveness). Children's behavioral problems are more common in mother– and father–child dyads that are low in observed mutuality. More to the point, sibling similarity in mutuality with the same parent is accounted for entirely by sibling genetic similarity, with sibling correlations ranging from.6 for identical twins to 0 for unrelated adoptive siblings. This strongly implicates an evocative rg-elinking children's aggressive behavior problems and problems in parent–child interaction quality.
Tests of Nonshared Environment Mechanisms
Nonshared environmental influences are those that account for family member differences after genetically based differences have been controlled. There are several methods for identifying specific nonshared environment mechanisms, but the examination of identical twin differences is the most straightforward. Because identical twins have the same genotypes, differences between them must have some environmental source: Identical twin differences are not due to differences in genes in the way that fraternal twin or nontwin sibling differences arise. The task is to reliably assess identical twin differences in the phenotypes of interest and then attempt to identify twin differences in environmental factors that predict phenotype differences.
In an analysis of parents' reports, interviewers' ratings, and observers' ratings, we were able to derive and estimate correlations between two sets of difference scores for 3-year-old identical twins. We estimated identical twin differences in outcomes (e.g., emotional reactivity, behavioral problems, noncompliance, prosocial behaviors) and twin differences in parenting environments (e.g., harshness of discipline, maternal warmth/negativity). Within each family, identical twin differences in these outcomes were correlated with differences in their interactions and relationships with the same parent, in anticipated ways. The twin who was exposed to higher levels of maternal negative emotions and harsh discipline and lower levels of maternal warmth and supportive control strategies (e.g., praise) showed more behavioral and emotional problems, more negative affect and less positive affect, more noncompliance, and less "on-task" behavior. It is important to emphasize that these results inform us about possible nonshared environment processes, but not about direction of effect. The maternal differences in parenting behaviors could be leading to the identical twin differences in behaviors, or the twin differences in behaviors could arise from other nonshared environmental sources (e.g., prenatal environment) and be driving the observed differences in maternal behaviors (Deater-Deckard, Pike, Petrill, Cutting, Hughes, & O'Connor, 2001).
Although these quantitative genetic techniques have been available for some time, there are surprisingly few analyses of this kind that have been done in adulthood, and these have tended to rely exclusively on participant self-reports. Nevertheless, there is no reason why the very same study designs and quantitative genetic analyses that are used in studies in childhood and adolescence cannot be used in studies of aging and cognitive change (for a review and examples, see Finkel & Pedersen, 2001; Reynolds et al., 2002). However, such designs are of no use unless viable candidate environmental factors are identified and incorporated into these genetic analyses. We turn now to a consideration of some of these environmental variables for tests of rg-e and nonshared environment mechanisms in cognitive aging.
Candidate "Environmental" Variables
Because of our focus on successful aging, we concentrate here on factors that might enhance intellectual functioning. Unfortunately, relevant evidence is scarce and, if it exists, is at least, at first glance, mostly discouraging with regard to the question of general enhancement of performance as a result of experience or interventions. There is, of course, substantial evidence of specific effects—that is, functions that are practiced throughout the life span are protected from the negative effects of aging. For example, Krampe and Ericsson (2000) showed that old professional pianists who have maintained a high level of deliberate practice played equally well as young professionals. However, lifetime practice with piano playing had no discernable effects on other simple cognitive tasks. (As an aside: In hindsight, it would have been very interesting to test for transfer on executive control tasks in that study. The continuous engagement in deliberate practice activities that require a high degree of self-monitoring and planning might be expected to have generalizable executive control effects.)
Salthouse and Mitchell (1990) looked at the effects of professional experience on psychometric abilities. For example, they examined the degree to which lifelong experience with a "visual–spatial" profession (architect) protects visual–spatial abilities from age-related decline. The sobering result was overall better performance for architects accompanied by rates of decline that were parallel for both architects and nonarchitects. In another study, Hambrick, Salthouse, and Meinz (1999) refuted the often-made claim that crossword puzzling may have general positive effects on cognitive functioning.
In sum, these studies provide little in terms of positive evidence in favor of effective experiential factors. However, from the viewpoint that executive control may be a particularly important cluster of functions to emphasize in the hunt for specific gene–environment correlation and nonshared environmental effects, the existing studies are not optimal. For example, this research has examined naturally occurring variation in the context of hobbies or professions—but a hallmark of skill acquisition in any one domain is the rise of automatic processing and the accompanying reduction in need for resource-demanding executive control. If it is actually the frequent use of executive regulation that produces enhancing effects on cognitive functioning, many natural skills may not be appropriate model cases to study the effects of experience on general abilities. Instead, we should be looking for situations with continuing demands on resource-sensitive processing, with little opportunity for routinization.
An interesting study in this regard (Shimamura, Berry, Mangels, Rusting, & Jurica, 1996) showed a lower effect of aging on cognitive function in a group of university professors. A faculty position at a prestigious academic institution may represent a case in which demands for flexible top-down control persist throughout work life. More generally, complexity of both paid work (defined as activity that requires "thought and independent judgment" [Schooler et al., 1999]) and leisure activity (Schooler & Mulatu, 2001) predicts later intellectual functioning, independent of prior levels of intellectual functioning. The cognitive complexity of daily activities, performed for decades throughout adulthood and into old age, includes a set of highly probable candidates for testing rg-e and nonshared environment mechanisms.
Another caveat to consider, with respect to the role of executive functions, pertains to the cognitive abilities that are targeted for either interventions or the measurement of intervention effects. As we noted above, executive control may be a particularly promising target domain because of its potential for generalizable effects. Currently, we know little about the degree to which executive control is affected by naturally occurring variations in intellectual experience (although some of the variables used in the above-mentioned study by Schooler and colleagues [1999] may tap variations in executive functions). However, there is some evidence that executive control can be enhanced in a somewhat general manner through training. For example, Kramer, Larish, and Strayer (1995) showed that training to flexibly shift cognitive resources between different tasks leads to benefits that transfer to new combinations of tasks in both old and young adults. Similarly, intensive exposure to attentionally demanding video games results in generalized benefits in attentional tasks among young adults (Green & Bavelier, 2003).
Further evidence, albeit less fully examined, comes from research on bilingualism. The demand that arises from coordinating two competing languages has to be resolved via top-down control and suppression of the currently irrelevant language. Thus, control functions may receive constant training in bilingual individuals. In fact, work with children has documented enhanced executive control in bilingual children (Bialystok, 1999). More recently, Bialystok and colleagues (Bialystok, Craik, Klein, & Viswanathan, 2004) have extended this work across the life span. They found that, at least for one particular task indicating suppression of dominant response tendencies, age-related control deficits were substantially reduced in elderly bilingual individuals compared with their monolingual counterparts.
A final important area of research is in regard to the effects of physical fitness on cognitive functioning in old age. Colcombe and Kramer (2003) have revisited the thus far inconclusive evidence for an association between physical fitness and cognitive functioning. They have found that only frontal lobe–based executive functions are sensitive to aerobic fitness effects. In additional, carefully controlled experiments, Colcombe and colleagues (in press) have provided very strong support for the claim that executive functions can be enhanced in old adults through aerobic exercise and that the results of these interventions are manifested as changes in brain activation during task performance.
Quantitative Genetic Studies in Aging
The evidence from the phenotypic studies described above is that complexity of work and leisure activities, bilingualism, and aerobic fitness are strong candidate domains of environmental factors for consideration. However, few of these have been examined thus far within the context of genetically informative designs.
Bergeman (1997) summarized the few genetic studies that have been done that point to a clear and consistent finding that is much like the pattern found in studies of childhood and adolescence. Most of the measures of "environmental factors" are, in fact, also influenced by individuals' genotypes (through gene–environment correlation mechanisms) and include moderate to substantial amounts of nonshared environmental variance. The candidate variables include social support (i.e., size of social network, membership in clubs and organizations, instrumental and emotional support from family and friends), perceptions of family environment (e.g., cohesion and conflict, achievement orientation), stressful life events (e.g., losses, illnesses, injuries, traumas), physical activity, socioeconomic status, and intellectual activity. Very few tests of rg-e and nonshared environment mechanisms have been conducted. There is evidence of both types of mechanisms in the links between social support, depression, and life satisfaction and evidence of rg-e in the links between physical activity, intellectual activity, socioeconomic status, and memory performance (see also Finkel & McGue, 1993).
Executive functions may be particularly important as outcomes for assessment in future work. Executive control plays a role in almost all intellectual activities, and there is mounting evidence that this can be enhanced via specific experiences. Of course, it is specifically this last point that needs further work, particularly within the context of genetically informative studies. Therefore, we expect progress from quantitative genetic studies that use measures of executive control along with other measures of cognition that also assess environmental variables relevant for executive functions.
|
THEORETICAL FRAMEWORKS |
|---|
|
TOPAbstractSuccessful Aging in the...Gene-Environment Correlation and...Empirical MethodsTheoretical FrameworksConclusionReferences |
|---|
Selective Optimization With Compensation: Baltes and Baltes
One particularly well-developed model is the Selective Optimization with Compensation model by Baltes and Baltes (1990; Freund & Baltes, 2002; for a model with similar scope, see Schulz & Heckhausen, 1996). The basic tenet of this model is that individuals' paths through life are negotiated through three different, consciously engaged "life-management" strategies: selection, optimization, and compensation. Selection is about setting goals that adequately reflect the current system of resources and the current time horizon. For example, in response to first signs of aging-related drops in mental ability, one may commit to the goal of leading a healthy life that is beneficial to high levels of mental and physical functioning. Optimization is about the acquisition and investment of goal-adequate means. For example, to work toward the aforementioned goal, an individual may begin to engage in physical exercise routines or shift to cognitively demanding leisure activities (e.g., reading instead of television). Finally, compensation is about using new means when the old means to achieve an important goal are no longer available. Psychometric work with self-report measures has documented the convergent and discriminant validity of these constructs. It has also demonstrated their developmental relevance by showing that use of each of these strategies increases across middle adulthood and that they are positively correlated with global well-being measures (Freund & Baltes, 2002).
In theory, the manner in which individuals use such strategies should be of critical importance for the way in which genetic and nongenetic effects become manifest. The higher an individual scores in these strategies, the more he or she should be in a position to cope most effectively with the "hard constraint" of cognitive decline. To the degree to which these constraints are genetically determined, the ability to use developmental strategies could be seen as a catalyst for nongenetic influences to take their effect. For example, only in someone who becomes aware of decreases in mental functioning, who selects the goal to do something about it, and who adequately engages in appropriate optimization or compensation activities can changes in environment be expected to become manifest. Of course, this may be a simplified picture because the ability to engage in these strategies itself probably is influenced by genetic factors to some degree. Nevertheless, we believe that there is a good case for incorporating assessments of life-management strategies within quantitative genetic studies with the hypothesis that they may serve as moderators of the degree to which genetic versus nongenetic influences dominate developmental outcomes and serve as mediators of gene–environment correlation and nonshared environment effects. Executive control abilities are likely to be involved in the real-time execution of these developmentally relevant strategies.
Life-Span Change in Gene–Environment Correlation: Scarr and McCartney
Another important theory is that of Scarr and McCartney (1983), who proposed that passive gene–environment correlation influences dissipate from early to late childhood, as parents have less control over the environments to which their children are exposed. At the same time, nonpassive (active and evocative) forms of gene–environment correlation increase in importance. There is some evidence for this theory in data on cognitive performance (McCartney et al., 1990), but no other phenotypes have been tested as thoroughly, therefore making it impossible to draw conclusions.
A life-span extension of Scarr and McCartney's theory is useful with respect to predictions regarding potential developmental shifts in rg-e (Figure 1). Specifically, passive rg-e effects may become important again later in life, to the extent that an individual's daily experiences become increasingly influenced by romantic partners and biological offspring. Though not usually described in this way, assortative mating creates new opportunities for passive gene–environment correlation effects to arise. Assortative mating is substantial for chronologic age and is moderate for cognitive performance and various measures of attitudes about social behavior, politics, and religion. By comparison, assortative mating effects appear to be less robust for aspects of personality (Buss, 1999). Thus, members of couples tend to resemble each other in their cognitive abilities. For those who are more cognitively adroit, part of the "environment" for each person also includes her or his partner's interest in and provision of cognitive stimulation in their interactions. This may become even more important in relationships where the partner maintains primary responsibility for care, in terms of the types of environmental inputs to which that person is exposed. As aging progresses, many become increasingly dependent on their mates, whereby the partner comes to exercise more control over the daily experiences of the spouse. Because of assortative mating, this environmental stimulation is confounded with overlapping genes between partners.
|
). It also is likely that parents who themselves are cognitively more adept also expect such performance in their children and are more likely to provide environments that are conducive to cognitive growth and skill acquisition—this is typically how passive rg-e is described. However, it is also likely that adult offspring who are cognitively adept and who were provided with cognitively stimulating and enriched environments by their parents will expect such performance in their aging parents and will provide environments that are conducive to the maintenance of cognitive abilities in their parents. To our knowledge, there is no proposed theory of species-typical change in nonshared environmental influences. Thus, we end with two hypotheses to guide future work. First, there are identifiable nonshared environmental influences that are stable over time. It is essential to test for this stability by using precise and highly reliable measures of environmental factors to rule out the alternative that nonshared environmental variance in cognitive aging simply represents measurement error. Second, there are likely to be predictable points in development when nonshared environmental influences are most influential, such as during highly salient life transitions in which the uncertainty of environments and experiences increases dramatically and constraints on time use and activities decrease. This is true for the transition to adulthood and may also be true for transition to retirement or major transitions into different types of lifestyle (e.g., shifts into and out of institutional care).
|
CONCLUSION |
|---|
|
TOPAbstractSuccessful Aging in the...Gene-Environment Correlation and...Empirical MethodsTheoretical FrameworksConclusionReferences |
|---|
|
Acknowledgments |
|---|
|
References |
|---|
|
TOPAbstractSuccessful Aging in the...Gene-Environment Correlation and...Empirical MethodsTheoretical FrameworksConclusionReferences |
|---|
environment effects. Child Development, 54,424-435.[Medline]
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
