xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 002020860
007 cr mnu|||uuuuu
008 090722s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002490
Hughes, Tiffany F.
The role of lifestyle factors in cognitive aging and dementia
h [electronic resource] /
by Tiffany F. Hughes.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 104 pages.
Dissertation (Ph.D.)--University of South Florida, 2008.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
ABSTRACT: It is widely accepted that cognitive abilities decline with normal aging. At the same time it is also recognized that there is variability in the magnitude and rate of decline among aging individuals. A similar phenomenon exists for dementia, where individuals with similar neuropathologic burden present with varying degrees of cognitive impairment. Of importance is determining what factors account for this variability, and whether individuals can modify these factors in order to preserve their cognitive abilities with aging or delay the onset of dementia. The purpose of this dissertation was to examine three potentially modifiable lifestyle factors' association with age-related differences/change in cognitive performance and risk for dementia by conducting three separate studies. The first study examined the association between engagement in lifestyle activities and concurrent cognitive speed performance.The second study examined whether there are differential associations between social resource factors and change in cognitive performance. The final study estimated the risk of late-life dementia in Swedish twins as a function of fruit and vegetable consumption in midlife. Taken together, the results of these studies provide evidence that individuals may be able to protect themselves against age-related cognitive decline or dementia by modifying their lifestyle. Specifically, individuals may benefit their cognitive speed performance by engaging in more cognitively demanding activities. Declines in episodic memory performance may be alleviated by being more satisfied with social support, and declines in general cognitive performance and speed and attention in young-old adults may be attenuated by having a larger social network of friends.Finally, the risk of all types of dementia and Alzheimer's disease may be reduced by consuming a moderate amount of fruits and vegetables in the diet, especially for females, those with self-reported angina, and those who consumed alcohol in midlife. These findings contribute to the literature on potential strategies to maintain cognitive health with aging and serve as groundwork for future intervention studies.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Co-advisor: Brent J. Small, Ph.D.
Co-advisor: Ross Andel, Ph.D.
t USF Electronic Theses and Dissertations.
The Role of Lifestyle Factors in Cognitive Aging and Dementia by Tiffany F. Hughes A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy School of Aging Studies College of Arts and Sciences University of South Florida Co-Major Professor: Brent J. Small, Ph.D. Co-Major Professor: Ross Andel, Ph.D. Cathy L. McEvoy, Ph.D. James A. Mortimer, Ph.D. Huntington Potter, Ph.D. Date of Approval: June 24, 2008 Keywords: aging, cognition, dementia, lifestyle, diet, so cial resources Copyright 2008, Tiffany F. Hughes
i Table of Contents List of Tables ................................................................................................................ ..... iii Abstract ...................................................................................................................... ........ iv Preface ...................................................................................................................... ......... vi Chapter One: Introduction ...................................................................................................1 Chapter Two: Literature Review .........................................................................................5 Cognitive Functioning and Aging ...........................................................................5 Normal Aging .............................................................................................5 Dementia (with focus on AlzheimerÂ’s disease) .........................................10 Theoretical Foundations: Environmental Complexity and Cognitive Reserve ....11 Environmental Complexity and Cognitive Reserve ..................................12 Factors Associated with Cognitive Functioning ....................................................13 Modifiable Factors in Midand Late-Life under Investigation .................14 Study I: Lifestyle Activities .......................................................................14 Study II: Social Resources .........................................................................18 Study III: Dietary Factors ..........................................................................21 Summary ................................................................................................................23 Chapter Three: Study I ......................................................................................................2 4 Abstract ..................................................................................................................25 Introduction ...........................................................................................................26 Methods..................................................................................................................29 Participants ................................................................................................29 Measures ....................................................................................................30 Cognitive Speed Variables .............................................................30 Simple Reaction Time ........................................................30 Choice Reaction Time ........................................................30 Lexical Decision Time ......................................................30 Semantic Decision Time ....................................................31 Activity Lifestyle ...........................................................................32 Health Composite...........................................................................32 Data Analysis .............................................................................................33 Results ....................................................................................................................34 Sample Characteristics ...............................................................................34 Correlations ................................................................................................35 Regression Analyses ..................................................................................36
ii Discussion .............................................................................................................37 Chapter Four: Study II ......................................................................................................4 8 Abstract ..................................................................................................................49 Introduction ...........................................................................................................50 Methods..................................................................................................................51 Participants ................................................................................................51 Measures ...................................................................................................52 Cognitive Measures ......................................................................52 Social Resources ...........................................................................52 Covariates ......................................................................................53 Analyses .........................................................................................53 Results ....................................................................................................................54 Descriptive Analyses ................................................................................54 Random Effects Models .............................................................................54 Discussion .............................................................................................................55 Chapter Five: Study III .....................................................................................................6 0 Abstract ..................................................................................................................61 Introduction ...........................................................................................................62 Methods..................................................................................................................63 Participants ................................................................................................63 Measures ....................................................................................................64 Dementia Diagnosis .......................................................................64 Dietary Assessment .......................................................................64 Covariates ......................................................................................65 Statistical Analyses ....................................................................................65 Results ....................................................................................................................67 Case-Control Analysis ..............................................................................67 Co-Twin Analysis ......................................................................................69 Discussion .............................................................................................................69 Chapter Six: Concluding Remarks ....................................................................................78 Limitations .............................................................................................................81 Future Directions ...................................................................................................82 References .................................................................................................................... ......85 Appendices .................................................................................................................... .....97 Appendix A: Action Letter ....................................................................................98 Appendix B: Curriculum Vitae ............................................................................99 About the Author ................................................................................................... End Page
iii List of Tables Table 1.1 Descriptive Char acteristics of the Study Variables by Age Group ...................43 Table 2.1 Correlations of Demographic Char acteristics and Hea lth with Lifestyle Activities and Reaction Time Performance .......................................................................44 Table 3.1 Correlations Between Lifest yle Activities and Reaction Time Performance ................................................................................................................... ....45 Table 4.1 Standardized Regression Estim ates for the Association between Lifestyle Activities and Cognitive Speed ..........................................................................46 Table 5.1 Standardized Regression Estimat es for the Association between Age and Select Cognitive Speed Measures by Activity Engagement .......................................47 Table 1.2 Models Predicting Cognitive Perf ormance as a Function of Social Resources ..................................................................................................................... ......59 Table 1.3 Characteristics of the Participan ts in the Case-Control Study by Disease Status ....................................................................................................................... ..........73 Table 2.3 Characteristics of Participants in the Case -Control Study by Relative Consumption of Fruits and Vegetables at Midlife .............................................................74 Table 3.3 Case-Control Analyses of the Association between Midlife Fruit and Vegetable Consumption and Dementia or Alzheimer's Disease (AD) ..............................75 Table 4.3 Stratified Analyses for Fruit and Vegetable Consumption and Risk of Alzheimer's Disease ........................................................................................................... 76 Table 5.3 Co-Twin Control Analyses of th e Association between Midlife Fruit and Vegetable Consumption and Risk of Dementia and Alzheimer's Disease ..................77
iv The Role of Lifestyle Factors in Cognitive Aging and Dementia Tiffany F. Hughes ABSTRACT It is widely accepted that cognitive abili ties decline with normal aging. At the same time it is also recognized that there is variability in the magnitude and rate of decline among aging individuals. A sim ilar phenomenon exists for dementia, where individuals with similar neuropathologic burden present with varying degrees of cognitive impairment. Of importance is de termining what factors account for this variability, and whether individuals can modify these factors in order to preserve their cognitive abilities with aging or delay the onset of dementia. The purpose of this dissertation was to examine three potentially modifiable lifestyle factorsÂ’ associat ion with age-related differences/change in cognitive performance and risk for dementia by conducti ng three separate studies. The first study examined the association between engagement in lifestyle activities and concurrent cognitive speed performance. The second st udy examined whether there are differential associations between social resource factors and change in cognitive performance. The final study estimated the risk of late-life dementia in Swedis h twins as a function of fruit and vegetable consumption in midlife. Taken together, the results of these studi es provide evidence that individuals may be able to protect themselves against ag e-related cognitive decline or dementia by modifying their lifestyle. Sp ecifically, individuals may be nefit their cognitive speed
v performance by engaging in more cognitively demanding activities. Declines in episodic memory performance may be alleviated by bein g more satisfied with social support, and declines in general cognitive performance and speed and at tention in young-old adults may be attenuated by having a larger social ne twork of friends. Fina lly, the risk of all types of dementia and AlzheimerÂ’s disease may be reduced by consuming a moderate amount of fruits and vegetables in the diet especially for females, those with selfreported angina, and those who consumed alc ohol in midlife. These findings contribute to the literature on poten tial strategies to maintain cogniti ve health with aging and serve as groundwork for future intervention studies.
vi Preface I would like to thank all that have played a role in the su ccessful completion of this dissertation work. Without your patience a nd support this work would not have been possible. I would like to fu rther extend my gratitude to the following individuals: Dr. Brent Small: I am extremely fortunate to have had you as a mentor. You were there every step of the way providing just the right amount of guidance and support that I needed to grow into an independent researcher I can only hope to be able to mentor my future students as well as you have mentored me. Dr. Ross Andel: Your patience, encouragement, and focus always kept me on track when I was discouraged or overwhelmed. I will always be grateful for the Â“pep talksÂ”. I hope that I am able to develop the same type of professional relationshi p with my students as we have shared. Dr. Cathy McEvoy: Thanks for giving me a chance and making an exception for me. I have had a lot of fun working with you a nd appreciate all of the guidance and support you provided in many aspects of my professional development. Dr. Huntington Potter: Without your faith in my abilities I would not have had the opportunity to complete the Ph.D. in Aging Studies. Thank you for supporting my decision to explore other disciplines and to di scover where my true passions for research lie. Drs. Mortimer and Borens tein: Thank you for introducing me to the world of neuroepidemiology. The training you have prov ided has opened doors that I never would have thought to look behind. Gail, Amy, and SAS staff: No one (myself included) could make it through the Ph.D. program without you. Your support is invalu able. Thank you for always coming to the rescue during emergencies and doing your best to make sure that things went as smoothly as possible. My family and friends: Thank you for your support and understanding, even when you didnÂ’t understand, during th ese last 5 years. Without you I would not be able to say, Â“I am finally not a studentÂ”.
1 Chapter One: Introduction The study of determinants of successful cognitive aging is receiving increased attention as research demonstrates that th ere is considerable heterogeneity in the magnitude of and rate of cognitive change experienced by older adults (Christensen, 2001). While some individuals remain relativel y cognitively intact in to old age, others become cognitively impaired, suggesting that cognitive decline or impairment is not an inevitable part of aging. As we approach a time where older adults will comprise a larger portion of the total populati on, a clearer understanding of what factors modify ageassociated and pathological changes in cogniti on is necessary in order for older adults to maintain cognitive health with aging. Research now supports that both genetic and environmental factors contribute to age-associated changes in cognition and to de mentia risk (Finkel & Pedersen, 2004; Gatz, Reynolds, Fratiglioni, Johansson, Mortimer, Berg, et al., 2006). Throughout the life course various factors including genetic (A shford & Mortimer, 2002; Farrer, Cupples, Haines, Hyman, Kukull, Mayeux et al., 1997; Finkel & Pedersen, 2004; Plomin, 1999), demographic (Anstey, 2000; Katzman, 1993), health (Anstey, 2000; Christensen, Jorm, Henderson, Mackinnon, Korten, & Scott, 1994; Rosnick, Small, McEvoy, Borenstein, & Mortimer 2004), and lifestyle factors (Crowe, Andel, Peders en, Johansson, & Gatz, 2003; Fratiglioni, Paillard-Borg, & Winblad, 2004; Fritsch, Smyth, Debanne, Petot, & Friedland, 2005; Hultsch, Hammer, & Sma ll, 1993; Hultsch, Hertzog, Small, & Dixon, 1999; Small, Hughes, Hultsch, & Dixon, 2007) in teract to either protect against or
2 increase susceptibility for cognitive decline or dementia. Although cognitive decline and dementia manifest in old age, the current demographic trends in the aging population demand that interventions in midand late-l ife be elucidated. Th e current dissertation focuses on how environmental factors in midand late-life are associ ated with cognitive functioning in non-demented individuals and w ith risk for dementia. Specifically, the dissertation examines three environmental factor s; lifestyle activities, social resources, and fruit and vegetable consumption and their associations with age-related differences/changes in cognitive functioning and w ith dementia risk. The ability of these factors to modify age-related differences/cha nges and dementia risk is examined within the framework of the environmental complex ity and cognitive rese rve hypotheses. Considerable interest now exists rega rding whether engagement in lifestyle activities can decrease risk for cognitive de cline or dementia. Since participation in lifestyle activities is potentially modifiable ev en in midand late-lif e, the implications of the research findings could be profound. Unfo rtunately, the existing body of literature on the association between engagement in lif estyle activities and cognition is mixed. Several issues related to the measurement of lifestyle activities, their association with different domains of cognitive functioning, a nd the directionality of the association between lifestyle activities and cognitive performance have contributed to the inconsistencies in the literature (Small et al., 2007). In addition, few studies have examined whether lifestyle activities are associated with variability in cognitive performance, which may be more indicativ e than mean level performance of how lifestyle activities a ffect the integrity of the central nervous system (Hultsch, MacDonald, Hunter, Levy-Bencheton, & Straus s, 2000; Li, & Lindenberger, 1999).
3 A substantial body of literature suggests th at social resources are associated with physical and mental health outcomes (Seem an & Crimmins, 2001). Recently, social resources have begun to be examined for their potential association with cognitive functioning. Various aspects of social reso urces, such as social network (Holtzman, Rebok, Saczynski, Kouzis, Wilcox Doyle, & Eaton, 2004) and social support (Seeman, Lusignolo, Albert, & Berkman, 200 1), have been found to be associated with cognitive functioning. However, there are inconsistencie s in the literature when different types of social resources are examined and associat ed with different cognitive outcomes. Determining whether different aspects of soci al resources, such as social network or satisfaction with social support, are more beneficial to cogni tive functioning than others is important for older adultsÂ’ maintenance of cognitive abilities. It is known that proper nutrition is an important factor in the promotion of cognitive health with aging (Fillett, Butler, OÂ’Connell, Albert, Birren, Cotman, et al., 2002), but the extent to which nutrition-relate d factors affect the risk for cognitive impairment in old age remains to be eluc idated (Gonzalez-Gross, Marcos, & Pietrzik, 2001). Although studies have shown that dietar y intake of specific macroand micronutrients are associated with risk for deme ntia (see Gillette Guyonnet, Abellan Van Kan, Andrieu, Barberger Gateau, Berr, Bonnefoy, et al., 2007 for review), there is limited evidence regarding the associations betw een fruit and vegetable consumption and dementia that Â“reflect the e ffects of dietary exposures and not changes in diet secondary to the diseaseÂ” (Luchsinger & Mayeux, 2004, p.580). In addition, to our knowledge no studies have examined the risk for dementia in twins to account for genetic and early-life factors.
4 The development of strategies that ma y reduce risk for cognitive decline or dementia hinges upon research that examines f actors that are malleable in midand latelife. The present doctoral dissertation focuse s on lifestyle activities and social resources in late-life as potential strategies to mitigate age-related differences or to slow age-related cognitive decline, and on fruit and vegeta ble consumption in midlife as a potential strategy to reduce risk for dementia. The di ssertation is organized as three individual studies with the overarching theme reflecting environmental (leisure activities, social resources, and diet) predictors of cognitiv e performance and cognitive impairment. The first study examines whether lifestyle activ ities influence older adultsÂ’ mean-level performance and variability on measures of cognitive speed. The second study assesses the association between seven aspects of soci al resources; social network of family, social network of friends, instrumental suppor t, emotional support, informational support, satisfaction with support, and negative inte ractions, and performance on a battery of cognitive tests including global ability, atte ntion and perceptual speed, and episodic memory at baseline and over five years. Finally, the third study examines fruit and vegetable consumption assessed prospectiv ely in midlife and risk for dementia controlling for potential conf ounds including early environm ental factors and genetic background by using a sub-sample composed ex clusively of complete twin pairs. The following chapter is a review of the existing body of literature on cognitive functioning with aging and dementia and fact ors that potentially influence cognitive health in old age.
5 Chapter Two: Literature Review A key determinant of successful aging is maintaining cognitive functioning (Baltes & Baltes, 1990). Maintaining cognitive abilities with aging is important for older adults to perform the activities of daily livi ng that are necessary fo r independent living; therefore understanding ways in which cogni tive functioning is maintained or cognitive impairment can be delayed or prevented is cr ucial for older adultsÂ’ quality of life. The overarching theme of the dissertation is to ex plore environmental fact ors that contribute to successful cognitive aging in older adults First, what is currently known about normative age-related changes and changes as sociated with dementia, the study outcomes of the dissertation, is reviewed. Two theoretic al models that attemp t to account for the link between environmental fact ors and cognitive functioning and dementia risk are then discussed. Finally, the factor s currently under inve stigation for their association with differences/change in cognitive functioning and with risk for dementia are described. Cognitive Functioning and Aging Normal Aging Despite disagreement about the nature and extent of change in cognitive functioning across the lifespan (Christe nsen, Mackinnon, Jorm, Henderson, Scott, & Korten, 1994), it is generally accepted that there are declines ac ross multiple cognitive abilities, regardless of testing in lab or re al world settings (Kramer, Bherer, Colcombe, Dong, & Greenough, 2004; Wilson, Beckett, Barnes, Schneider, Bach, Evans, et al., 2002). The declines are not unitary acro ss cognitive domains, however (Park, 2000).
6 During the course of adult development two patterns of change in cognitive functioning are evident. Some abilities have been shown to remain relatively stable into old age, while others follow a trajectory of decline (Baltes, Staudinger, & Lindenberg, 1999; Horn & Cattell, 1966; Horn & Hofer, 1992; Salthouse, 1999). The abilities that show stability are known as crystallized abilities. Verbal abilities, including vocabulary, information, and comprehension are considered to be cr ystallized. Abilities such as speed of processing, memory, spatial ability, and reason ing are those that te nd to decline with aging and are known as fluid abilities. Diffe rent mechanisms have been proposed to explain age-related declines in cognitive performance. Th ese mechanisms include, but may not be limited to, the processing speed theory, working memory, inhibition and sensory function (Park, 2000). These proce sses, especially processing speed, often account for a large proportion of age-related variance across a wide assortment of tasks and environments, but are rarely able to en tirely eliminate age-related differences in cognitive functioning. A great deal of research on cognitive aging has focused on age differences in cognition. This is primarily the result of th e use of cross-sectiona l study designs which are faster and less expensive to conduct. Mo st cross-sectional studies of cognitive aging have reported that older adults perfor m more poorly than you nger adults on most cognitive tasks, with the exception of voca bulary performance (Park, 2000). This was supported by Park, Smith, Lautenschlager, Ea rles, Frieske, Zwahr, et al. (1996) who found that cognitive performance on tasks of processing speed, working memory, and cued and free recall all declined in a generally lin ear pattern across the lifespan from age 20 to age 90 In contrast, vocabulary perf ormance was found to remain relatively stable
7 across the lifespan. Cross-sectional studies of older adults also find that young-old adults perform better than old-old adults on cognitive measures. For example, Christensen and colleagues (1994b) found age differences be tween the 70-74, 75-79, and 80 and over age groups in crystallized intel ligence, fluid intelligence, and memory where there was a decrease in performance across the age groups. In their study, fluid abilities were found to decline more than crystal lized abilities as a function of age, and all individual tests were found to show decline with the excepti on of vocabulary. The results of these studies demonstrate that age differences in cognition exist even into very old age. However, an accurate description of cognitive aging requires that change in cognition be examined through longitudinal methods, which are able to estimate individual rates of decline or risk factors for decline, and can examine the association between changes in cognitive functioning and changes in ot her cognitive and non-cognitive domains (Christensen, 2001). In order to better understand what normal c ognitive changes can be expected with aging and to discover the mechanisms that explain adult development, researchers are now recognizing the importance of longitudina l studies of aging (Schaie & Hofer, 2001). Wilson and colleagues (2002) examined cognitive changes on 21 cognitive tasks consisting of the domains of episodic me mory, semantic memory, working memory, perceptual speed, and visuospatial ability over six years in a cognitive ly intact sample of older adults. They found that performan ce on each of the cognitive domains declined over the 6-year period and that the rate of d ecline in each domain was related to age at baseline such that those who were older e xperienced greater dec lines. Moreover, the authors also found that there was substant ial heterogeneity in each of the cognitive
8 domains, indicating that while on average there are declines, some older adults remain stable or improve slightly over time as well. It has not been until recently that longitudinal studies of cognitive aging have exam ined change into very old age. A study conducted by Singer, Verhaeghen, Ghisletta, Lindenberger, and Baltes (2003) examined change in four intellectual abilities (i .e. perceptual speed, memory, fluency, and knowledge) over six years in a sample of adu lts ranging between 70 and 100 years of age from the Berlin Aging Study. Overall, the resu lts of their analyses showed declines with age in perceptual speed, memory, and fl uency. Performance on the knowledge tasks (Vocabulary and Spot-the-Word), however, re mained stable until age 90, but declined thereafter. Despite the advantages provide d by longitudinal studies of cognitive aging compared to cross-sectional stud ies, it is important to consid er issues such as attrition, survival effects, missing data, and practice effects when interpreting the results of longitudinal studies (Schaie & Hofer, 2001). Cognitive differences/changes associated with normal aging can vary both between and within older adults with re gard to the magnitude (Bckman, Small, & Wahlin, 2001) and rate (Wilson et al., 2002) of decline. There is evidence that both betweenand within-person va riability increases with normal aging. Between-person variability refers to the fact that individuals begin and end at differe nt places on tests of cognitive performance. By contrast, within-p erson variability refers to the fact that persons may perform differently from day to day, as well as from trial to trial. For example, Hultsch, MacDonald, & Dixon, ( 2002) found that cognitive performance on reaction time tasks became increasingly more diverse, dispersed, and inconsistent with age, and that these measures were positively associated with each other and negatively
9 associated with performance on other cogniti ve composites. Although the majority of research on variability in cognitive func tioning in older age has focused on betweenperson variability, or diversity, there is increasing interest in the study of within-person variability in the study of cognitive performance. Two types of within-person change that have been described are intraindividual change and intraindividual variabil ity (MacDonald, Hultsch, & Dixon, 2003). Intraindividual change refers to change that is relatively slow and enduring (e.g. learning, development, and changes in traits), while intr aindividual variability refers to change that is relatively rapid and transient (e.g. m ood states, emotions, and fluctuations in performance). Variability associated with a personÂ’s performance on multiple cognitive tasks is known as dispersion, whereas variabil ity associated with a personÂ’s performance on a single task measured on multiple occasions is referred to as inconsistency (Hultsch et al., 2002). Traditionally, performance in consistency was regard ed as random error (Williams, Hultsch, Strauss, Hunter, & Ta nnock, 2005); however recent studies have shown that inconsistency is not a statistica l artifact and can be measured reliably apart from other within-person vari ability (e.g. practice effects, learning to learn, materials effects) (as cited by Hultsch et al., 2000). Si nce inconsistency has been shown to predict cognitive performance independently of mean level performance, and also to predict mean level change, performance variability may be a useful behavioral indictor in addition to mean level performance (Hu ltsch et al., 2000; Hultsch et al., 2002; MacDonald, Hultsch, & Dixon, 2003; Williams et al., 2005).
10 Dementia (with focus on AlzheimerÂ’s disease) AlzheimerÂ’s disease is characterized by a progressive loss of intellectual and cognitive functioning that ultimately impair s everyday functional abilities (Sloane, Zimmerman, Suchindran, Reed, Wa ng, Boustani, et al., 2002). In the mild stages of the disease, patients generally suffer from memo ry loss and may have some difficulty with executive functions, language, and attention (Slo ane et al., 2002). This is consistent with the diagnostic criteria of impairment in me mory plus an additional cognitive domain (4th ed.; DSM-IV; American Psychiatric Association, 1994). As the disease progresses, memory impairment worsens and assistance is needed with instrumental activities of daily living, dressing and with handling the de tails of toileting. During this stage AD patients may also experience significant pers onality changes and behavioral symptoms and may wander or get lost. Finally, in the mo st severe stages of AD, the individual is no longer able to speak or comprehend langua ge and requires cons tant care, including supervision and assistance in the basic ev eryday functional abilitie s of eating, toileting, bathing, and transferring (Sloane et al., 2002) Although the rate of progression through these stages varies greatly in AD cases, the clin ical course of the di sease is generally well established (McDowell, 2001). Currently, AlzheimerÂ’s disease affects approximately 5.1 million Americans, 4.9 million of whom are aged 65 years and over. Future projections of the prevalence of AD estimate that as many as 11-16 million older adults could have the disease by 2050 if no cure is found or if no strategy is devel oped to slow its progression (AlzheimerÂ’s Association, 2007). Sloane and colleagues (2002) examined how therapeutic advances may alter disease prevalence through 2050 and re ported three different scenarios. First,
11 there could be a delay in the onset of AD. Second, there could be a reduced rate of progression. And finally, th ere could be both a delayed onset and reduced rate of progression. Based on their analyses, if no therapeutic advances occur, 10.2 million persons will have the disease in 2050, of which 3.8 million will have mild disease and 6.5 million will have moderate/severe disease. In the delayed onset model, it was projected that there would be 35.6% fewer cases of AD by 2050 than baseline projection, but moderate/severe disease would constitute the majority of cases. If the disease progression were slowed, they pr ojected a slight increase (1.2%) in the number of cases, but with a higher proportion of patients in th e mild stage versus moderate to severe stages. Finally, a combined model projecte d a reduction in the to tal number of cases (37.4%) where mild cases would predomin ate (56%; Sloane et al., 2002). Taken together, these studies demonstrate th at there is substantial heterogeneity in cognitive functioning during the course of agi ng. In general, some individuals maintain their cognitive functioning, some decline gradua lly, and some decline sharply to a clinical diagnosis of dementia. The focus of the curr ent project is to examine factors that may play a role in determining where older adults fall on this cognitive continuum with aging. Theoretical Foundations: Environmental Complexity and Cognitive Reserve It has been suggested that genetics play a large role in develo pment of age-related brain changes or disease-related pathology, and that environmental factors play a larger role in the expression of cognitive impai rments (Mortimer, Borenstein, Gosche, & Snowdon, 2005). Two theoretical perspectives environmental complexity and cognitive reserve, may provide explanations as to how environmental factors affect the expression of cognitive impairment.
12 Environmental Complexity and Cognitive Reserve The concepts of environmental comple xity and cognitive reserve have great salience for understanding heterogeneity in individual performance and how the following factors contribute to this heterogene ity (Stern, 2002). Alt hough they originated from different empirical root s, the sociology of work a nd brain injury, respectively, together they may help to identify envi ronmental factors that reduce age-related differences and change and e xplain how these same envir onmental factors act at the neurophysiologic level (Small et al., 2007). Simply stated, the environmental comple xity hypothesis suggests that complex environments have a positive effect on cogni tive functioning, and simple environments have a negative effect on cognitive functioni ng. More specifically, the complexity of an environment is a function of th e diversity the stimuli, the nu mber of decisions required, the number of considerations to be taken into account in making decisions, and the illdefined and apparently contra dictory contingencies result ing from these decisions. Accordingly, complex environments are expe cted to reward cognitive effort, where individuals should be motivated to develop thei r intellectual capacitie s and to generalize their use to other situations. Continued exposure to relatively simple environments may have the opposite effect sin ce the low level of environmen tal demand does not foster the development of or maintenance of intellect ual functioning (Schooler, 1984; Small et al., 2007). The concept of cognitive reserve has been proposed to explain the heterogeneity in clinical outcomes between individuals who have the similar neural deficits related to disease pathology or normal age-related brain changes. Two types of cognitive reserve
13 have been proposed by Stern (2002) to describe this variation, passi ve and active. The passive model of reserve suggests that neur on and synapse number or brain size provide the basis for reserve, and consequently is determined primarily by genetics but may be influenced to some degree by environmental in fluences. The active model of reserve is concerned more with neural processing and synaptic organization than neuroanatomical differences. Neural proce ssing and synaptic organization are more sensitive to environmental influences; therefore it is these changes that provide the greatest potential for increasing reserve (Stern, 2002). It is likely a combination of active and passive reserve that provides the most comprehens ive explanation of th e cognitive variation between individuals at th e neurophysiologic level. Factors Associated with Cognitive Functioning Age-related and pathological changes in cognitive functioning are influenced by a combination of genetic and environmental fa ctors (Finkel & Pedersen, 2004). It is generally thought that genetic and environmental factors operate in conc ert to affect risk for cognitive decline and dementia, with genetic factors affecting risk for age-related or pathological changes in the ne ural substrate, and environmen tal factors affecting risk for clinical expression of these changes (M ortimer, Snowdon, & Markesbery, 2003; Stern, 2002). Although the focus of this dissertation is to examine the influence of midand late-life environmental fact ors on cognitive functioning in older age, one should be mindful of the fact that gene tic factors and earl y-life factors, especially educational attainment (see Andel, Hughes, & Crowe, 2005 for review), are interacting with these factors to produce the cognitive outcomes observed in older age. Furthermore, the selected midand late-life factors under inve stigation are likely influenced by other late-
14 life factors, such as medical comorbidies, that are also being studied for their role in cognitive vitality with aging, and that adju stment for these factors is made whenever possible. Modifiable Factors in Midand Late-Life under Investigation Evidence suggests that environmenta l factors become increasingly more important for individual differen ces in cognitive ability in late-life (Finke l & Pedersen, 2004), and that environmental factors can con tinue to affect olde r adultsÂ’ cognitive functioning (Schooler, 1984). Three factors th at can be modified in middle and older age, but may be just as importa nt in early-life, are lifestyle activities, social resources and diet. The following sections provide a review of the literature describing how each of these factors has been linked to age-rela ted cognitive differences/changes and risk for dementia and serve as the background for the three studies of the dissertation. Study I: Lifestyle Activities Age-related differences/changes in late-life leisure activity participation and cognitive performance have been studied by a number of researchers; however, the findings are inconsistent. A positive relationshi p between activity participation level and cognitive functioning has been found in seve ral studies (e.g. Bielak, Hughes, Small, Dixon, 2007; Newson & Kemps, 2005; Wilson, Barnes, & Bennett, 2003). For example, Newson and Kemps (2005) found that particip ation in four categories of everyday activities (household maintenance, domestic chores, social activities, and service to others) was a significant predictor of baseli ne level of speed, picture naming, incidental recall, and verbal fluency; and of cognitive change over six years in speed, picture naming, and incidental recall after controlling for sensor y functioning. A life course
15 approach was taken by Wilson and colleagues (2003) to examine how participation in cognitively stimulating activities measured at ages 6, 12, 18, 40 and current age was related to function in different cognitive sy stems. They found that lifetime cognitive activity was related to better cognitive performance, but onl y for certain tasks. Finally, using a lifestyle activity questionnaire to m easure frequency of engagement in cognitive, social, and physical activities Bielak and colleagues (2007) found that higher frequency of engagement in cognitively complex activit ies was related to faster latency and less inconsistency on speeded tasks concurrently, but that change activity was not associated with change in speed performance. These resu lts suggest that current activity level may influence cognitive functioning to a greater exte nt than change in activity participation with aging. In contrast to these studies, a number of studies have failed to observe that engagement in leisure activities that were socially, experienti ally, developmentally (Aartsen, Smits, van Tilburg, Knipscheer, & Deeg, 2002), or cognitively (Salthouse, Berish, & Miles, 2002) stimulating provided a ny protection against c ognitive decline. It should be noted that the lack of support for an association between lif estyle activities and cognitive functioning in these studies is c onceivably due to met hodological/statistical issues. For example, younger adults were included in the study by Salthouse and colleagues (2002) who likely exhibited less variability in cognitive performance and engagement in activities than the older a dults, which may have reduced the authorsÂ’ ability to detect age-relate d differences in cognitive performance as a function of cognitively stimulating activity pa rticipation (Small et al., 2007).
16 Age-related changes in lifestyle activities and changes in cognitive performance have been studied in order to provide cl ues about the temporal relationship between cognitive functioning and engagement in life style activities. Hu ltsch and colleagues (1999) used a measure of active lifestyle that included 70 activities classified into six groups: physical activities, self-maintenance activities, social activities, hobbies and home maintenance activities passive information processing, and novel information processing. Longitudinal examin ation of the influence of engagement in these activities on maintenance of cognitive f unctioning over time revealed that higher engagement in novel activities at baseline and over time buf fered against decline in working memory, however, the possibility that cognitive decline preceded decline in engagement was suggested. Another longitudinal study examined the association betw een participation in activities and cognitive performance on tests of memory, cognitive speed, and crystallized intelligence and f ound that declines in activity participation and cognition occurred simultaneously, and that declines in cognition occurred even in a sub-sample of individuals who maintained thei r level of activity participa tion over the course of seven years (Mackinnon, Christensen, Hofer, Korten, & Jorm, 2003). Although the authors concluded that participation in lifestyle activities may not offer protection against cognitive decline because those who maintained their level of activity participation still declined, the fact that those who maintained their activity levels were those who had lower levels of activity at baseline does not preclude the possibility that older adults who are able maintain a higher level of activity participation may be able to stave off decline. A more recent examination of whether ch anges in lifestyle activities lead to a change in cognitive functioning or if changes in cognitive functioning lead to a change in
17 lifestyle activities was c onducted by Ghisletta, Bickel, and Lvdn (2006) using a bivariate dual change score model appro ach. This method allows for the direct assessment of the temporal order among the variables tested, and thus can determine whether changes in lifestyle activities preced e or follow changes in cognitive functioning or whether changes in both variables influen ce changes in the other. The results of the study revealed that increased frequency of pa rticipation in media(e.g. listen to radio, watch television) and leisure(e.g. play games, crossword puzzles) type activities was associated with less decline in perceptual speed. In general, there appears to be some support for a positive association between engagement in leisure activities in late-lif e and cognitive functioning that warrants further research. One area that has received only mi nimal attention in the literature is the association between lifestyle activity pa rticipation and variability in cognitive performance. In normal aging, within-person variability, or inconsistency, refers to variability associated with a personÂ’s performance on a single task measured on multiple occasions (Hultsch et al., 2002). Evidence fr om studies such as Hultsch and colleagues (2002) suggest that increased inconsistency with aging coul d be predictive of impending cognitive decline or an indicator of neurol ogical dysfunction, and may a more sensitive measure than mean level performance. Ther efore, the purpose of the first study of the dissertation is to examine the concurrent as sociation between partic ipation in lifestyle activities and both mean level performance and inconsistency on four reaction time tasks. A lifestyle activities questionna ire that assessed frequency of participation in activities that are cognitively, socially, and physically stimulating and range in cognitive demands is used to measure engagement in lifestyle activities. How engagement in these lifestyle
18 activities is related to mean level performa nce and inconsistency in cognitive speed; as measured by performance on simple, choice, lexi cal, and semantic reaction time tasks, is tested controlling for age, gender, educatio nal attainment, and physical health. The relationships found further our knowledge regard ing (1) the types of lifestyle activities that are most strongly related to neuroc ognitive speed, and (2) whether mean-level performance or inconsistency is more str ongly related to engagement in lifestyle activities. Study II: Social Resources There is now a large body of evidence s upporting the association between social resources and health. Both the quantity and qua lity of social resources have been shown by numerous studies to be associated w ith mortality and morbidity (see Seeman & Crimmins, 2001; Uchino, Cacioppo, & Kiecolt-Gla ser, 1996 for reviews). Diseases such as heart disease (Krumholz, Butler, Miller, Vaccarino, W illiams, Mendes de Leon, et al., 1998; Mookadam & Arthur, 2004) and depr essive symptoms (Jang, Borenstein, Chiriboga, & Mortimer, 2005; Yaffe, Lui, Grady, Stone & Morin, 2002) have been shown to be reduced in individuals with greater social resources. The strong link between social resources and health outcome s has led to investigations of a possible relationship between social resources a nd cognitive functioning in older adults. Studies of the association between so cial resources and cognitive functioning have found different aspects of social resour ces to be related to cognitive functioning. For example, Holtzman and colleagues (2004) examined the association between social network characteristics and global cognitive functioning in a sample of high-functioning older adults. They found that a larger social network at baseline was associated with
19 better maintenance of cognitive ability and with a reduced risk for general cognitive decline over 12 years of follow-up. In contrast, Seeman and colleagues (2001) examined the association between social ties and s upport and patterns of c ognitive aging and found that cognitive performance was better at ba seline and at the follo w-up seven and a half years later for those who received more emo tional support, while they did not find that the size of social network was associated w ith cognitive functioning. Interestingly, being unmarried and reporting greater conflict with members of the social network was also related to better cognitive pe rformance in their study. A more powerful study of the relationship between social ties, social integratio n and social engage ment and cognitive functioning using random effects models al so found differential importance for these social relation measures in terms of cognitive f unctioning (Beland, Zunzunegui, Alvarado, Otero, & Del Ser, 2005). Specifically, social engagement and social integration were found to be more important for general cognitive functioning over seven years than social ties after controlling for gender, edu cation, depressive symptoms, functional limitations, and chr onic conditions. Finally, strong er perceived social support has also been found to be related to bette r cognitive functioning by Yeh and Lui (2003), where having a good friend to talk to was a ssociated with better scores on the short portable mental status questionnaire. Several issues have created challeng es for the study of social resources and cognitive performance and may account for the mi xed results. First, social resources are defined differently across studies. For exam ple, social network was defined by Holtzman and colleagues (2004) as the number of re latives and family members outside the household and the number of friends and neighb ors with whom one kept in touch with by
20 phone or visits. In comparison, Seeman and co lleagues (2001) define d social network as marital status, number of close ties with child ren, family, and friends, and participation in religious or other groups. To measure soci al support, some studies have examined emotional support available (Bassuk, Glass, & Berkman, 1999; Seeman et al., 2001), while others have also used instrumental support (Seeman et al ., 2001) and indices of living arrangement and loneliness (Yeh & Li u, 2003). Further, most studies have not included measures of perceived support or ne gative interactions, which have been shown to be associated with cognitive functioning (Seeman et al., 2001; Yeh & Liu, 2003). The ability to find an association be tween social resources and cognitive performance is likely related to the types of cognitive outcomes measured. For example, few studies have examined measures of cognitive functioning that may be most susceptible to decline with age, such as attention and perceptual speed, and domains that may be more strongly associated with envir onmental factors (e.g. social resources), such as episodic memory (Kramer et al., 2004). The majority of studies examining the association between social resources a nd cognitive functioning have only included measures of global cognitive ability (Bassuk et al., 1999; Holtzman et al., 2004; Yeh & Liu, 2003). Seeman and colleagues (2001) di d assess the higher order functions of language, verbal and nonverbal memory, abstra ct reasoning, and spatia l ability in their study; however, they did not assess more basi c cognitive abilities a nd neither did any of the other studies. In summary, evidence suggests that so cial resources may be important for cognitive functioning similar to health outcomes; however, further resear ch is needed that addresses the described limitations before c onclusions can be drawn. The purpose of the
21 second study of the dissertati on is to examine the associa tion between social resources and cognitive performance while addressing these limitations. This is accomplished by including multiple measures of social resources including social network of family, social network of friends, instrumental support, emotional support, informational support, satisfaction with support, and negative intera ctions and by using a more comprehensive cognitive battery that has a test of genera l cognitive ability, and tests of basic (i.e. attention and perceptual sp eed) and higher order (i.e. memory) cognitive abilities. Increasing our knowledge about th e association between social resources in this way is important for tailoring interventions that may a lter or modify social resources in order to improve cognitive health in older age. Study III: Dietary factors Dietary factors are important for cognitive functioning in late-life since they function directly by maintaining oneÂ’s health and indirectly by preventing other diseases that are related to cognitive impairment, such as vascular disease. Several studies have now reported that certain micro-nutrients including Vitamins B6, B12, folate, antioxidants (Vitamins C, E and carotenes), and polyphenols are related to the risk of dementia (Commenges, Scotet, Renaud, Jacqmi n-Gadda, Barberger-Gateau, & Dartigues, 2000; Dai, Borenstein, Wu, Jackson, & Lars on, 2006; Engelhart, Geerlings, Ruitenber, Van Swieten, Hofman, Witteman, et al., 2002; Morris, Evans, Bienias, Tangney, Bennett, Aggarwal, et al., 2002). Low serum levels of each of these are char acteristic of people with cognitive impairment (Solfrizzi, Panza, & Capurso, 2003); therefor e it is of interest whether maintaining adequate levels of thes e vitamins either through supplementation or through diet can reduce risk for impairment. These micro-nutrients are thought to be
22 important for cognitive health because they pl ay a role in processes such as oxidative stress and inflammation, which are known to play a role in both age-related and dementia-related brain changes. Care should be taken when studying the e ffect of diet on cognitive impairment in late life since there is the potential for prec linical disease processes to affect dietary patterns. Few studies have prospectively ex amined the relationship between midlife diet and risk of late-life dementia. Laurin, Masaki, Foley, White, & Launer (2004) examined the association between midlife dietary inta ke of antioxidants estimated based on a food questionnaire and the incidence of late-life dementia. Over all, the authors did not find support for a protective effect of midlife antioxidant intake on de mentia risk in later life. Despite this negative finding, this study high lights the need for additional studies of dietary factors in midlife and risk of latelife dementia and its subtypes, including a more basic examination of intake of whole foods. The role of diet in cognitive impairme nt with aging has not been extensively studied. The fact that diet is an importan t part of a healthy lifestyle and has a large impact on disease profiles suggests that dietar y factors deserve more attention in future research. The focus of the final study of th e dissertation is to e xplore the association between midlife fruit and vegetable consump tion and the risk for dementia using two study designs: case-control and co-twin control. These designs test whether genetic or early environmental familial factors account for any infl uence of midlife fruit and vegetable intake on the risk of dementia. The risk of late-life dementia was prospectively assessed using data collected in midlife, which reduces the possibility of measuring changes in diet secondary to dementia since dementia has a long latency period.
23 Determining whether dietary factors in midlife can modify the risk for dementia in latelife has important implications for the types of dietary recommendations made by governmental agencies and physicians to pr omote the cognitive heal th of the population. Summary The current dissertation examines thr ee aspects of environment that may be predictors of cognitive performance or cogni tive impairment; lifestyle activities, social resources, and fruit and vegetable inta ke. This is accomplished through three independent studies which attempt to build on the previous litera ture and aid in the understanding of whether lifestyle factors are a feasible strategy to slow down or delay age-related cognitive decline or reduce the risk for dementia. The first study of the dissertation examines the infl uence of lifestyle activity pa rticipation on mean-level and inconsistency in reaction time performance. The analyses determines whether activities of varying cognitive demand are differentially related to reaction time performance, and whether inconsistency in reaction time perf ormance is a more sensitive marker of cognitive functioning than reaction time perfor mance at the mean-level. In the second study, the relation between seven aspects of social resources and multiple domains of cognitive functioning over five years is exam ined. Lastly, the third study examines the role of midlife fruit and vegetable consumption in the risk for dementia in late-life. Each of these studies is carried out using the concepts of environmental complexity and cognitive reserve as the theo retical underpinnings.
24 Chapter Three: Study I Does Engagement in Lifestyle Activities Affect Inconsistency in Cognitive Speed Performance in Older Adults? Tiffany F. Hughes, Allison A. M. Bielak, Brent J. Small & Roger Dixon
25 ABSTRACT The Â“use it or lose itÂ” h ypothesis of cognitive aging c ontends that engagement in stimulating activities moderates age-related differences in cognitive performance. Using data from the Victoria Longitudinal Study (n = 511), we examined whether frequency of engagement in lifestyle activities is asso ciated with concurrent mean-level and intraindividual variability in cognitive speed performance. Multiple regression analyses revealed that higher frequency of engage ment in integrative and novel information processing activity was associated with fa ster and less variab le cognitive speed performance. Age-differences in lexical a nd semantic decision time performance were moderated by engagement in self-mai ntenance and novel information processing activities. Overall, the findi ngs support the notion that a hi gher level of engagement in cognitively demanding activities is associated with better cognitive speed performance, and that there are greater differences in reaction time performance across age as a function of engagement in self-maint enance and novel lifestyle activities.
26 INTRODUCTION The potential for older adults to assume an active role in their cognitive health has recently been the focus of public health in itiatives (Centers fo r Disease Control and Prevention and the AlzheimerÂ’s Association, 200 7). Several actions have been proposed as potential strategies to maintain or improve the cogniti ve performance of older adults including engagement in ac tivities that are physically, socially, and cognitively stimulating. The Â“use it or lose itÂ” para digm suggests that cognitive abilities are maintained through stimulation of the cogniti ve system, while age-related declines are attributable to Â“disuseÂ” of cognitive abili ties (see Small, Hughes, Hultsch, & Dixon, 2007). However, the current literature desc ribing the association between engagement in activities and cognitive functioning is inconc lusive, where some studies show a positive association and others do not support an association. The concepts of environmental comple xity and cognitive reserve propose psychological and physiological mechanisms to explain how engagement in lifestyle factors may affect the level of cognitive f unctioning in older adults and the rate of cognitive change with aging. The envir onmental complexity hypothesis suggests that exposure to complex environments rewards cogn itive effort such that individuals will be motivated to develop their intellectual capac ities and to generaliz e their use to other situations, whereas continued exposure to rela tively simple environments will not foster the development of or maintenance of in tellectual functioning (Schooler, 1984). This cognitive effort is then tran slated at the neurophysiologic le vel to more efficient neural processing and synaptic orga nization as described by the active model of cognitive reserve (Stern, 2002). Cognitive reserve is believed to be amenable to change
27 throughout the life course and to modify the tr ajectory of decline associated with aging such that those individuals with higher re serve will experience a slower rate of agerelated cognitive change compared to those with lower reserve. Several observational studies have been conducted in the field of cognitive aging that examine the relation between lifestyle activities and cognitive performance. For example, Newson and Kemps (2005) found that participation in household maintenance, domestic chores, social activities, and serv ice to others was a si gnificant predictor of current cognitive performance in speed of pr ocessing, picture naming, incidental recall, and verbal fluency after controlling for se nsory functioning. Similarly, Hultsch, Hertzog, Small, and Dixon (1999) also found that highe r engagement in in tellectual activities buffered against decline in working memor y, however, the possibil ity that cognitive decline preceded decline in engagement was su ggested. To clarify this issue, Ghisletta, Bickel, and Lvdn (2006) used the bivariat e dual change score model approach and found that increased frequency of participation in media(e.g. listen to radio, watch television) and leisure(e.g. play games, crossword puzzles) type activities was associated with less decline in perceptual speed, whereas the reverse association was not found. In contrast to these studies, others have failed to observe that engagement in leisure activities buffers against cognitive dysfunction with aging. Aartsen Smits, van Tilburg, Knipscheer, and Deeg (2002) f ound that socially, experientially, and developmentally stimulating ac tivities did not provide any pr otection against decline in general cognitive ability, immediate recall a nd learning, and fluid intelligence. Similarly, Mackinnon, Christensen, Hofer, Korten, and Jo rm (2003) found that d eclines in activity
28 participation and cognition o ccurred simultaneously, and that declines in cognition occurred even in a sub-sample of individua ls who maintained their level of activity participation over the course of seven years. Finally, Salthou se, Berish, and Miles (2002) found that engagement in 22 cognitively dema nding activities did not buffer against agerelated differences on four cognitive compos ite measures reflecting fluid intelligence, episodic memory, and crysta llized intelligence. The mixed findings in the li terature are likely due to different methodological and analytical considerations (see Ghisletta et al., 2006; Salthouse, 2006; Small, et al., 2007 for reviews). One consideration is that det ecting an association between engagement in lifestyle activities and cognitive performance requires a measure sensitive to the integrity of the central nervous system (e.g. synaptic connections). Intraindi vidual variability, or inconsistency, in cognitive performance refers to variability associat ed with a personÂ’s performance on a single task measured on multiple occasions (Hultsch & MacDonald, 2004). Increased inconsistency has been show n to be associated with older adults compared to younger adults (e.g., Hultsc h, MacDonald, & Dixon, 2002; Nesselroade & Salthouse, 2004; Williams, Hultsch, Strau ss, Hunter, & Tannock, 2005), poorer overall cognitive performance (Hultsch & MacDona ld, 2004; Li, Aggen, Nesselroade, & Baltes, 2001), and conditions known to affect neurological functioning (e.g. dementia, ParkinsonÂ’s disease, MCI, traumatic brai n injury; Burton, Strauss, Hultsch, Moll, & Hunter, 2006; Christensen, Dear, Anste y, Parslow, Sachdev, & Jorm, 2005; Dixon, Garrett, Lentz, MacDonald, Strauss, & Hultsc h, 2007; Strauss, Bielak, Bunce, Hunter, & Hultsch, 2006; Stuss, Pogue, Buckle, & Bondar, 1994; Walker, Ayre, Perry, Wesnes, McKeith, Tovee, et al., 2000). Therefore, examining inconsistency in cognitive speed,
29 in addition to mean-level performance, may in crease the ability to detect an association between engagement in activities and cogniti on since slower and mo re variable reaction time performance with increased age could be attributed to lower frequency of engagement in stimulating activities. The purpose of the current study is to examine whether age-related slowing, measured by mean level and inconsistency in reaction time tasks, is related to engagement in lifestyle activiti es using data from the Vict oria Longitudinal Study (VLS). We hypothesized that 1) increasing age would be associated with d ecreased participation in lifestyle activities and slower mean-level and greater inconsistency in reaction time performance, 2) higher frequency of participatio n in activities, particularly those that are more cognitively engaging, will be associated with faster mean-level cognitive speed and with less inconsistency in cognitive speed perf ormance, and 3) age-related differences in reaction time performance between individuals will vary as a function of the level of engagement in lifestyle activities. METHODS Participants Participants were drawn from the Victor ia Longitudinal Study (VLS). The VLS is a longitudinal-sequential research project in which participants are retested every three years on an extensive battery of cognitive, physical, sens ory, health, and psychological tests with new samples added every six years. A more detailed description of the VLS design, procedures, and measures can be found elsewhere (Dixon & de Frias, 2004; Hultsch, Hertzog, Dixon, & Small, 1998). The present study is based on cross-sectional
30 data from Wave 1 of Sample 3 that cons isted of 577 community-dwelling older adults aged 55-90 years. Measures Cognitive Speed Variables The measurement of cognitive speed is based on reaction time (RT) latencies in milliseconds from four multi-trial computer-bas ed RT tasks. These four tasks require rapid responses to simple nonverbal stimu li and complex language-based stimuli, and require participants to make a decision about the stimulus by pressing a key on the response console. Simple Reaction Time. In the simple reaction time task, the participants were presented with a warning stimulus (***) follo wed by a signal stimulus (+) in the middle of the screen. They were inst ructed to press a key as quick ly as possible in response to the signal stimulus. The latency scores of fi fty trials after completion of 10 practice trials served as the outcome measure. Choice Reaction Time. In the choice reaction time task, a warning stimulus (++) was presented on the left and right of the sc reen followed by one changing to a square. The participants were instru cted to the press key corresponding to square location as quickly as possible. Following 10 practice trials a total of 50 trials were administered and the latency was used as the outcome measure. Lexical Decision. The objective of the lexical decision task was for the participants to judge as rapi dly as possible whether a set of 5-7 letters presented was a real English word (e.g. island vs. nabion ). The time to respond by pressing one of two keys across 60 trials (30 words and 30 nonw ords) served as the outcome measure.
31 Semantic Decision. The semantic decision task required the participants to judge as rapidly as possible whether a se ntence presented was plausible (e.g. The tree fell to the ground with a loud crash vs. The pig gave birth to a litte r of kittens this morning.; Palmer, MacLeod, Hunt, and Davidson (1985)) Time to press one of two keys was recorded across 50 trials (25 plausible and 25 implausible sentences) and used as the outcome measure. Following the recommendations of previous research (Hultsch et al., 2002); we defined the lower bound for legitimate respons es as 150 ms for simple reaction time, 150 ms for choice reaction time 400 ms for lexical decisi on, and 1,000 ms for semantic decision. The upper bound limit was determined by calculating the mean and standard deviation for each task and o ccasion of measurement and removing any trials that exceed the mean by three or more standard deviati ons. Any latency scores falling above or below these bounds were dropped from the an alyses since extremely fast or slow responses most likely represent various s ources of measurement error (e.g., accidental key press, distraction of participant). We imputed missing value estimates usi ng a regression substitution procedure that assesses the correlations among response time across all trials (Hultsch, MacDonald, Hunter, Levy-Bencheton, &Strauss, 2000). Th is method of eliminating outlying trials and imputing estimates for the missing values decreases within-subj ect variation such that the data represent a conservative approach to examining intraindi vidual variability in response time performance. Following data preparation proce dures, the level of cognitive speed was computed in the trad itional manner as the mean RT of each individualÂ’s latency for each task. Our meas ure of inconsistency was computed as the
32 across-trial within-person individual standa rd deviation (ISD) about each individualÂ’s mean RT. Activity Lifestyle The VLS Activity Lifestyle Questionna ire (VLS-ALQ) is a validated, 70-item self-report questionnaire desi gned to measure frequency of participation in activities during the past two years. Frequency of participation is rate d on a 9-point scale ( never, less than once a year, about once a year, 2 or 3 times a year, about once a month, 2 or 3 times a month, about once a week, 2 or 3 times a week, daily ), and scaled such that higher scores are associated with greater freque ncy of activity. For the present study, we classified 66 items into a 7-category clas sification based upon th e previous validation work of Hultsch, Hammer, and Small (1993) and Hultsch and colleagues (1999). The seven subscales were defined as: (1) physical activity, such as gard ening or jogging (n = 4); (2) self-maintenance, such as preparing a m eal or shopping (n = 6); (3) social, such as attending church or eating out (n = 7); (4) travel, such as trav eling within Canada (n = 3); (5) passive information processing, such as wa tching a sporting event or listening to the radio (n = 8); (6) integrative information pr ocessing, such as drivi ng a car or playing a musical instrument (n = 13) ; and (7) novel information pr ocessing, such as learning a new language or preparing income tax forms (n = 25). Items within each of the seven categories were summed to form composite activity measures. Health Composite At each wave of measurement, the VL S Personal Data Sheet was administered and included questions pertaini ng to health beliefs, health conditions, and health risk factors. These questions were designed to re present four of five measures of physical
33 health identified by Liang (1986): chroni c illness, number of illness episodes, instrumental health, and subj ective health. A measure of self-reported physical health was created by summing the average scores of 34 items assessing the presence and severity of chronic illnesses in the past tw o years, six items assessing the number of illness episodes in the last four weeks and three items assessing the number of illness episodes in the past year, eight items assessi ng the extent to which health affected their daily activity patterns (i.e. instrumental health), and two items assessing how the participants rated their health in comparison to perfect health and to others their age (see Hultsch et al., 1993 and Hultsch et al., 1999 for a detailed description of the health measures). Higher scores on the composite heal th measures were representative of worse health. Data Analysis Descriptive characteristics of the sa mple according to age group (young-old = 5364 years; old-old = 65-74 years; oldest-o ld = 75-90 years) were calculated and comparisons were made using analysis of variance between age groups. Correlation analyses were performed to examine the uni variate associations among the demographic characteristics, health, frequency of activity engagement and reaction time performance. Regression analyses were conducted for each reaction time task to examine the main effects of each lifestyle activities controlli ng for the effects of age, gender, education, health, and the other activity measures. Intera ction effects were then tested to examine whether the association between age and reaction time performance depended upon the level of engagement in lifestyle activities Each interaction term was computed by centering the scores of the main effects to avoid multicollinearity (Aiken & West, 1991),
34 and was added separately to the main effect model in order to maintain a sufficient ratio between the number of subjects and predictor variable s entered in the regression model. Significant interaction effects were interprete d by stratifying the sele ct activity domains into tertiles (low, medium, and high) and ex amining the whether the point estimates of each activity tertile fell with in the 95% confidence interval of the other two groups. RESULTS Sample Characteristics Following list-wise deletion procedures, a to tal of 511 participants were eligible for the present analyses. Characteristics of the sample by age group are presented in Table 1.1. Compared to the young-old, the old-ol d and oldest-old were less likely to be women, and the oldest-old were less likel y to have completed as many years of education. For engagement in lif estyle activities, the old-old were more likely to engage in passive activities and less likely to engage in physical activities compared to the young-old participants. The oldest-old were le ss likely to engage in travel, physical and integrative information processing activities compared to the young-old, and they were also less likely to engage in travel and integrative information processing activities in comparison to the old-old. There were also differences between each of the age groups in the frequency of self-maint enance and novel information processing activities where engagement in these activities decreased with increasing age group membership. In terms of cognitive speed performance, the oldest-old were slower and more variable on all reaction time tasks compared to both the young-old and old-ol d. The old-old were also slower and more variable on the c hoice IM and ISD and the lexical IM tasks compared to the young-old participants.
35 Correlations Prior to regression analyses, we examin ed univariate correlations between the demographic characteristics and health a nd the lifestyle activity and reaction time measures (Table 2.1). Increasing age was co rrelated with less frequency of engagement in travel, self-maintenance, physical, in tegrative information processing, and novel information processing activities, and with sl ower reaction times and greater variability on all measures. Men tended to engage in fe wer social and self-m aintenance activities, and to engage more frequently in travel, integrative and novel activ ities. They also performed faster on choice reaction time. Reporting a lower level of educational attainment was related to less frequent enga gement in travel, physical, integrative and novel activities as well as with slower mean level reaction time on all tasks, and more variability in lexical and semantic decision tim e. Finally, those in poorer health were less likely to engage in passive, travel, physical integrative, or novel activities, and were slower and more variable on the semantic decision time task. Table 3.1 shows the correlations between the activity measures and reaction time performance. More frequent engagement in travel activities was associated with faster performance on all tasks and with less vari ability on choice and lexical decision time. A higher level of self-maintenance activities was associated with faster and less variable choice reaction time performan ce. Higher level of physical activity engagement was related to faster and less va riable simple, choice and lexi cal decision time performance as well as faster semantic decision time. Mo re frequent engagement in integrative information processing activities was associated with faster and less variable performance
36 on the simple and choice reaction time meas ures, while engagement in novel activities was associated with faster and less variable performa nce on all tasks. Regression Analyses The results of the multiple regression an alyses are summarized in Table 4.1. After controlling for age, gender, e ducation and health, social, self-maintenance, integrative and novel activities were associ ated with reaction time perfor mance. Specifically, more frequent engagement in social activities was related to sl ower choice reaction time, and more frequent engagement in self-mainten ance activities was associated with more variable lexical decision time performance. Conversely, more frequent engagement in integrative information processing activities wa s associated with faster simple and choice reaction time and with slower lexical and mo re variable lexical and semantic decision time performance. More frequent engagement in novel activities was associated with faster and less variable performance on lexi cal and semantic decision time tasks. Significant interactions between self-mai ntenance and novel activity engagement and age were found for performance on the lexi cal and semantic decision tasks (Table 5.1). Slowing of performance on each of th e lexical and semantic time tasks with increasing age varied as a function of the level of engagement in self-maintenance activities such that more fr equent engagement was genera lly associated with greater slowing as age increased. The opposite was found for novel activity participation such that more frequent engagement was genera lly associated with a less slowing on the lexical and semantic time tasks as age increased.
37 DISCUSSION In this study we were interested in wh ether engagement in lifestyle activities is associated with mean-level and variability in cognitive speed performance. In addition, we sought to determine if age-related differe nces in cognitive speed performance were moderated by the level of engagement in th ese activities. This st udy contributes to the findings of previous research by including a measure of inconsistency in performance, which may account for some of the mixed findings in the literature testing the association between lifestyle activities and cognitive performance. The findings of the study support the idea that a higher level of engagement in integrative and novel information processing acti vities is associated with faster and less variable performance on select reaction time measures. In ad dition, greater differences in reaction time performance were found with increasing age depending on the level of engagement in self-maintenance and novel in formation processing activity participation, suggesting that age differences in cognitiv e speed performance are influenced by the level of engagement in lifestyle activities. Our pattern of results su ggests that engagement in novel information processing activities is most strongly and consistently associated with cognitive speed performance, especially for the language-based reacti on time tasks. The novel activity category included activities such as reading, doing crossword pu zzles, and learning a new language. In comparison to the other activity categories, these types of activities can be considered more cognitively demanding as th ey primarily stimulate cognitive domains. Thus, novel activities likely prov ided the most direct benefits to the neurological system, possibly by strengthening or creating new syna ptic connections (Stern, 2002). Additional
38 associations were found between more freque nt engagement in integrative information processing and faster simple and choice r eaction time performance. These types of activities also place demands on the cognitive system, although potentially less so than novel activities, which may be why they were associated with more basic reaction time measures. We suspect that the finding that more frequent engagement in integrative information processing activities was associated with slower lexica l and more variable lexical and semantic decision time performan ce is attributed to collinearity between activity measures since the univariate correl ations were in the expected direction. Finally, the fact that higher engagement in social and self-maintenance activities was found to be associated with slower choi ce reaction time and more variable lexical decision time, respectively, may be that these ty pes of activities provide relatively little cognitive stimulation or that higher engagement in these activities may limit engagement in more cognitively stimulating activities. The presence of interactions between age and engagement in self-maintenance and novel lifestyle activities s uggests that there are differenc es in cognitive speed with increasing age in relation to activity particip ation. According to Salthouse (2006), age by activity interactions must be found to validate the differential-preservation hypothesis whereby Â“the degree to which [cognitive] perf ormance is preserved across increasing age is postulated to differ according to the leve l of mental activityÂ” (Salthouse, 2006, pg. 70). Our findings support the differential-preser vation hypothesis for self-maintenance and novel activity engagement. The extent to which age-related slow ing occurred for the lexical and semantic decision time tasks di ffered between the low, medium and high activity groups where high self-maintenan ce and low novel activity engagement were
39 associated with greater slowing and variabil ity with increasing age. Explanations for these effects are similar to the previous disc ussion of the main effect findings for selfmaintenance and novel activity engagement. The fact that we did not observe additi onal significant interactions between age and activities may be related to the match betw een the type of activities the participants engaged in and the cognitive domain te sted. Evidence from cognitive training intervention studies, including the Seattle Longitudinal Aging Study and the Advance Cognitive Training for the Independent and Vital Elderly (ACTIVE), suggest that cognitive training improves cognitive performan ce on the specific abiliti es trained (Ball, Berch, Helmers, Jobe, Leveck, Marsiske, et al., 2002; Schaie, 2005). For example, it would be expected that higher engagement in act ivities that require verbal skills, such as crossword puzzles, would be associated w ith better performance on a cognitive task primarily assessing verbal abilit y and not a task assessing me mory. Furthermore, recent evidence also suggests that crystallized abi lities may be better maintained throughout the life course for those whose are more engaged in stimulating leisure activities compared to those who are not (see Kramer, Bherer, Colcombe, Dong, and Greenough, 2004 for review). Since few of the activ ities assessed in the current st udy were closely associated with processing speed abilities, and only tw o of the reaction time tasks required some verbal skill, stronger affects between lifestyle ac tivities and inconsistency in cognitive speed performance may have been found if hi gher order cognitive tasks, such as verbal ability or semantic memory, had been tested. We expected that engagement in activ ities would be more strongly associated with inconsistency in reaction time performan ce than with mean-level performance given
40 the sensitivity of inconsis tency to the integrity of the nervous system. Although engagement in novel activities was associated with less variability in lexical and semantic decision time, overall we found more significa nt relations between activity and reaction time performance at the mean-level compared to inconsistency. This finding suggests that engagement in lifestyle activities affects the neurological system at the macro level, which is better captured by m easuring mean-level cognitive pe rformance. Alternatively, observing few associations be tween activity engagement and inconsistency may be related to the notion that only those individuals below a certain cognitive level, or whose neurological integrity has been compromise d below a certain level, would demonstrate more variable performance in relation to lower frequency of engagement in lifestyle activities (Salthouse, 2006). Limitations of the study should be addre ssed. This study is a cross-sectional examination of the association between act ivities and cognitive performance. A fundamental issue plaguing the study of this association is whet her activity decline precedes cognitive decline or the reverse (Schooler & Malutu, 2001). This question can only be addressed with longitudinal data a nd sophisticated statistical techniques that permit lead-lag relationships to be estimat ed. Furthermore, determining whether the differential-preservation hypothesis or the preserved-differentiation hypothesis, which proposes that those who are more engaged in older adulthood are likely to have been more engaged throughout life a nd that prior engagement esse ntially determines the level of cognitive performance rather than rate of change in cognitive performance, better describes the relationship between engagement in lifestyle activities and cognitive performance requires longitudinal data over the entire life course (Salthouse, 2006).
41 Although these types of analyses were beyond the scope of the present study, we feel that examining the concurrent relationship betw een frequency of activity engagement and inconsistency in cognitive speed provides new insight into how measures of processing speed ranging in cognitive effo rt are affected by engagement in lifestyle activities and whether inconsistency is a more sensitive marker of the benefits of engaging in lifestyle activities than is mean-level performance. Another limitation is the representati veness of the sample. Members of VLS have a relatively higher level education, are in relatively better h ealth, and are higher functioning than the general population (Di xon & de Frias, 2004; Hultsch et al., 1993, 1999). Since the benefits of activities on c ognition may be greatest for those whose educational attainment or c ognitive abilities are lower than average (e.g. Arbuckle, Maag, Pushkar, & Chaikelson, 1998; Christensen & Mackinnon, 1993; Gold, Andres, Etezadi, Arbuckle, Schwartzman, & Chaikelson, 1995), our results may have been attenuated because of the intact nature of our sample. However, previous comparisons of the crosssectional association between self-reported he alth, activity, and cogni tive measures in a representative sample of 1,278 community-dwe lling older adults aged 65-100 (Ball, 1998) versus a select VLS sample revealed si milar results, suggesting that our findings may not be biased due to the intact nature of the sample (Hultsch et al., 1999). Nevertheless, our results should be inte rpreted with this possibility in mind. In conclusion, the results of this study suggest that e ngaging in activities that are more cognitively complex may benefit concu rrent mean-level performance on reaction time tasks, especially those that are more cognitively demanding. We also found that engagement in self-maintenance and novel activ ities moderated age-related differences in
42 cognitive speed performance. Future studies should be conducted to examine longitudinal relationships between lifestyle activities and inconsis tency in higher order cognitive abilities that appear to be more receptive to environmental stimulation. Until research confirms or refutes the validity of the Â“use it or lose itÂ” hypothesis with respect to cognitive aging, adults of a ll ages should continue to e ngage in activities that are mentally, socially, and physically stimulating if for no other reason than to improve their quality of life.
43VariableMeanSDMeanSDMeanSD *p -value Age59.632.9569.372.9679.903.57 Gender (% Female)77.5659.8861.94<.001 Education15.732.7815.112.9214.632.99.003 Health6.975.046.593.407.113.990.54 Social activity3.200.953.150.923.101.010.64 Passive activity3.950.964.190.914.131.110.04 Travel activity2.180.802.190.761.900.900.003 Self-maintenance5.170.934.920.984.621.03<.001 Physical activity18.104.22.1681.383.631.21<.001 Integrative information1.5 40.721.510.701.290.630.003 Novel information3.230.643.010.622.800.69<.001 Simple RT IM312.5271.03316.7169.28361.0978.05<.001 Simple RT ISD78.0256.4585.8748.68102.9280.520.001 Choice RT IM772.20122.17844.88131.94951.19154.40<.001 Choice RT ISD164.4549.41180.1753.14200.5181.25<.001 Lexical RT IM992.09316.881092.49462.701240.71421.59<.001 Lexical RT ISD294.70197.603 38.56231.89408.50276.13<.001 Semanitc RT IM3268.491085.563513.761176.194050.501462.40<.001 Semantic RT ISD1017.46673.311039.45471.911240.41653.290.002 *ANOVA Table 1.1 Descriptive Characteristics of the Study Variables by Age Group Age Group 53-64 years (n = 205)65-74 years (n = 172)75-90 years (n = 134)
44Table 2.1 Correlations of Demogra phic Characteristics and Health with Lifest y le Activities and Reaction Time Performance VariableAgeGenderEducationHealth Social activity-.0.06-0.17***0.08-0.02* Passive activity0.07 -0.010.04-0.14** Travel activity-0.13**0.10*0.18***-0.12** Self-maintenance-0.22***-0.28***-0.04***0.004 Physical activity-0.22***0.050.09*-0.19*** Integrative information-0.13**0.10*0.20***-0.12** Novel information-0.28***0.11*0.39***-0.15*** Simple RT IM0.24***-0.01-0.12**0.04 Simple RT ISD0.14**-0.07-0.070.04 Choice RT IM0.51***-0.09*-0.13**0.06 Choice RT ISD0.27***-0.03-0.070.06 Lexical RT IM0.28***-0.02-0.18***0.07 Lexical RT ISD0.21***-0.06-0.23***0.08 Semanitc RT IM0.27***-0.07-0.19***0.11** Semantic RT ISD0.17***-0.02-0.15***0.09* p < .05. ** p < .01. ***p < .001.
45Table 3.1 Correlations Between Lifest y le Activities and Reaction Time Performance VariableIM ISDIMISDIMISDIMISD Social activity0.001-0.010.010.01-0.01-0.03-0.03-0.004 Passive activity-0.03-0.03-0. 004-0.02-0.04-0.02-0.03-0.02 Travel activity-0.09*-0.08-0.11**-0.10*-0.12**-0.13**-0.10*-0.07 Self-maintenance-0.06-0.03-0.11*-0.12**-0.02-0.08-0.02-0.01 Physical activity-0.14**-0.11*-0.20** *-0.17***-0.12**-0.10*-0.09*-0.06 Integrative informatio n-0.17***-0.12**-0.21***-0.11*-0.030.01-0.020.02 Novel information-0.17***-0.10**-0.26** *-0.13**-0.29***-0.30***-0.26***-0.19*** Note: IM = Intraindividual Mean; ISD = Intraindividual Standard Deviation. p < .05. ** p < .01. *** p < .001. Simple RTChoice RTLexical RTSemantic RT
46 Table 4.1 Regression Estimates for the Association between Lifestyle Activities and Cognitive Speed VariableIMISDIMISDIMISDIMISD Main Effects Age1.66***0.85*8.43***1.68***9.81***3.84**29.53***9.25** Gender2.06-9.33-41.21**-6.763.29-0.53292.53*2.08 Education-1.27-0.151.33-0.06-9.13-9.23*-42.81*-18.17 Health-0.190.080.370.281.921.5122.839.12 Social activity0.980.262.07*0.722.214.171.1242.53 Passive activity-0.29-0.12-0.31-0.09-0.331.031.270.67 Travel activity-0.81-0.92-0.39-1.12-3.95-3.10-9.26-2.05 Self-maintenance-0.01-0.05-0.43-0.751.914.45*11.381.03 Physical activity-0.72-0.55-1.38-1.02-3.12-1.91-3.69-0.83 Integrative information-1.00*-0.51-1.85**-0.294.52*4.28***12.78.44* Novel information-0.25-0.05-0.83-0.07-6.10***-3.97***-16.65***-5.95** Model R20.07***0.02*0.30***0.08***0.13***0.14***0.13***0.05*** Interaction Effects Age*Social -0.020.003-0.04-0.01-0.40-0.29-0.93-0.81 Age*Passive 0.020.020.020.040.230.100.45-0.09 Age*Travel -0.21-0.05-0.26-0.14-0.86-0.73-4.21-1.69 Age*Self-Main. 0.030.030.04-0.020.91**0.60***3.03**1.12* Age*Physical -0.040.07-0.18-0.06-0.46-0.37-1.69-0.88 Age*Integrative -0.05-0.02-0.06-0.06-0.17-0.050.01-0.22 Age*Novel -0.05*-0.03-0.07-0.01-0.30**-0.17**-0.93**-0.38* Note: IM = Intraindividual Mean; ISD = Intraindividual Standard Deviation. p < 0.05; ** p < 0.01; *** p < .001 Simple RTChoice RTLexical RTSemantic RT
47Table 5.1 Regression Estimates and 95% Confidence Intervals for the Association between Age and Select Cognitive Speed Measures by Activity Engagement VariablesLowMediumHighLowMediumHigh Age with Lexical RT IM5.27 (-3.12,13.66)7.77 (2.13,13.42)17.36 (9.62,25.10)14.63 (4.86,24.40)4.51 (-1.58,10.61)7.20 (1.97,-12.43) Lexical RT ISD0.77 (-3.21,4.76)2.83 (-0.66,6.33)8.73 (3.48,13.98)5.77 (0.50,11.05)2.37 (-1.36,6.10)1.82 (-1.22,4.87) Semantic RT IM15.46 (-8.07,39.00)18.39 (-0.56,37.33)55.39 (28.61,82.16)43.09 (16.02,70.14)12.35 (-6.94,31.65)24.69 (4.21 ,45.16) SemanticRT ISD4.91 (-5.33,15.15)5.39 (-5.82,16.59)16.82 (2.48,31.16)13.23 (0.45-26.01)2.65 (-6.40,11.70)7.05 (-5.57,19. 66) Note: IM = Intraindividual Mean; ISD = Intraindividual Standard Deviation. Self-MaintenanceNovel Information Processing
48 Chapter Four: Study II The Association between Social Resources and Cognitive Change in Older Adults: Evidence from the Char lotte County Healthy Aging Study Tiffany F. Hughes, Ross Andel & Brent J. Small School of Aging Studies, Univ ersity of South Florida Amy R. Borenstein & James A. Mortimer Department of Epidemiology and Biostati stics, University of South Florida
49 ABSTRACT We examined associations between multiple aspects of social resources and 5-year change in performance on different domains of cognitive function. Results indicated that lower satisfaction with support was associ ated with decline in episodic memory performance over 5 years. We also found signi ficant interactions be tween age and social networks of family and friends and satisfac tion with support for the separate cognitive domains. The results suggest that social re sources may be differentially important for cognitive change but that different cognitiv e domains respond in a similar pattern to social resources.
50 INTRODUCTION Evidence that an active a nd socially integrated li festyle may slow cognitive decline in old age (Bassuk, Glass, & Berkman, 1999) or reduce risk for dementia (Fratiglioni, Paillard-Borg, & Winblad, 2004) su ggests that the social environment could confer protection against cognitive declin e (Schooler & Malatu, 2001). However, few studies have simultaneously examined the effect s of different aspects of social resources (e.g., size of social network, receipt of support, or satisfaction with social support) on the cognitive health of older adults or whether separate cognitive domains respond differently to social resources. There is considerable variability in th e manner in which social resources have been operationalized and they domains of cogn itive functioning that have been measured. For example, size of social network has been th e most frequently stud ied aspect of social resources in relation to cognitive perfo rmance (Bassuk et al., 1999; Holtzman, Rebok, Saczynski, Kouzis, Wilcox, & Eaton, 2004; Se eman, Lusignolo, Al bert, & Berkman, 2001). However, other aspects of the social environment, such as emotional support, negative interactions, and satisfa ction with support, have rece ived less empirical attention despite suggestions that they may be importa nt predictors of cognitive function (Bassuk et al., 1999; Holtzman et al., 2004; Seem an et al., 2001; Yeh & Liu, 2003). Cognitive outcomes examined in previous studies have also been varied. The majority of studies have either used a meas ure of global cognitive function such as the Mini-Mental State Examination (Bassuk et al ., 1999; Holtzman et al ., 2004) or created a summary score of overall cognitive ability (Barnes, Mendes de Leon, Wilson, Bienias, & Evans, 2004; Seeman et al., 2001). However, different domains of cognitive functioning
51 have different developmental tr ajectories across the life span and also respond differently to environmental factors (Kramer, Bherer, Colcombe, Dong, & Greenough, 2004). Therefore, examination of how social resources influence multiple cognitive domains requires further research. The purpose of the current study was to examine the associations between different aspects of social resources and change in multiple domains of cognition. We examined whether changes in general cognitive ability, speed and attention, and episodic memory are differentially influenced by social network of family and friends; emotional, informational, and instrumental received s upport; satisfaction with support received; and negative social interactions. We also examin ed whether the strength of the associations between social resources and ch ange in cognitive performance va ry as a function of age. METHODS Participants Participants were members of the Ch arlotte County Healthy Aging Study, a community-based, longitudinal study of agi ng in which participants were initially enrolled in 1997/1998. Interested readers can find a detailed description of the sampling procedure and response rates elsewhere (B orenstein, Mortimer, Wu, Jureidini-Webb, Fallin, & Small, 2006; Small, Graves, McE voy, Crawford, Mullan, & Mortimer, 2000). After we excluded 38 individuals who scored less than 82 on the Modified Mini-Mental State Examination (Teng & C hui, 1987), the baseline sample comprised 417 individuals. At the follow-up approxim ately 5 years later ( M = 4.94 years, range = 4.6Â–5.3 years), 43 participants (10.3%) had died, an additional 36 (8.7%) had withdrawn after the baseline phase, 39 (9.4%) refused to participate, and 60 (14.4%) were unable to be contacted. This
52 resulted in 239 participants. Because of missi ng data on some predictors at baseline, the longitudinal sample consisted of 217 persons. Measures Cognitive Measures The neuropsychological test battery included the Modified Mini-Mental State Examination (Teng & C hui, 1987) to measure general cognitive ability, the Stroop Test (Stroop, 1935) to meas ure attention, the Trai lmaking Test (Parts A and B) to measure perceptual speed (Re itan & Wolfson, 1985), a nd the Hopkins Verbal Learning Tests (Brandt, 1991; Benedict, Schretlen, Goninger, & Brandt, 1998) comprising tests of delayed free recall, cued recall, and recognition to measure episodic memory. Based upon significant intercorrela tions, we created two composite scores: speed and attention (Trailmaking, Parts A a nd B; and Stroop Test) and episodic memory (delayed, cued, and recognition). Social Resources. The participants were asked 26 que stions with regard to social resources at baseline. Thes e items were derived from L ubbenÂ’s (1988) Social Network Scale and social support measures from the wo rk of Krause and Borawski-Clark (1995). Principal components factor analysis with varimax rota tion was used to reduce the number of variables in the models. The prin cipal components factor analysis revealed seven factors: social network of family (number of contacts with family per month, frequency of contact per month with closes t relative, and number of close relatives), social network of friends (number of close friends, number of friends in contact at least once per month, and frequency of contact with closest friend), emotional support (frequency in the past month of others providing support in difficult times, providing comfort, listening or talking about private feelings, and s howing interest or concern),
53 instrumental support (how often in the past month others provided transportation, help with housework, chores, or yard work, and help with shopping), informational support (frequency in the past month of others sugges ting some action to take to solve a problem, making a difficult situation easier to understand, helping und erstand why something was not done well, and sharing what they di d in a similar stressful situation), satisfaction with support (frequency of having someone to talk with about an important decision, and frequency of satisfaction with emotional suppor t, instrumental support, and informational support), and negative social interactions (how often in the past month others placed demands, were critical, pried into affairs, and took advantage). The factors accounted for 63% of the variance. Covariates. Demographic variables included ag e (in years), gender (men = 0, women = 1), education (in years), marital st atus (not married = 0, married = 1), and residency in Charlotte County, Florida (i n years). We include d a between-subjects variable to account for follow-up attrition status (yes = 0, no = 1). Personality was assessed with the NEO Five-Factor Inventor y (Costa & McCrae, 1992), which measures the domains of neuroticism, extraversion, openness, agreeableness, and conscientiousness. Analyses We used Proc MIXED in SAS Version 9 (Littell, Milliken, Stroup, & Wolfinger, 1996) to examine whether each social resour ce factor contributed unique variance in cognitive performance above and beyond the other factors. For each cognitive outcome, we simultaneously entered all of the baseline social resource factors and interactions between each social resource factor and conti nuous age while adjusting for baseline age,
54 gender, education, marital status, residency st atus, and personality in the model. For the analysis of cognitive change, we also modeled the influence of years of follow-up time. In order to interpret signifi cant interactions, we stratified age into young-old versus oldold based on a median split (age 73/74), rece ntered age, and examined whether the point estimates of each group fell within the 95% confidence interval (CI) of the opposite group. RESULTS Descriptive Analyses At baseline, the study particip antsÂ’ mean age was 72.4 years ( SD = 6.2); they had an average of 14.0 years ( SD = 2.7) of education; 51.8% were women; 77.5% were married; and they had lived an average of 12.6 years ( SD = 8.5) in Charlotte County, Florida. The independent samples t -tests comparing the follow-up sample ( n = 217) to those who were not followed ( n = 200) revealed that t hose followed were younger and more likely to be married, had lower ne uroticism scores, had higher scores on extraversion and conscientiousness, had liv ed fewer years in Charlotte County, and received less instrumental support at ba seline. They also performed better on the measures of speed and attention and episodic memory at baseline compared to those who were not followed ( p < .05 for all analyses). Random Effects Models As shown in Table 1.2, examination of th e fixed effects of social resources on baseline levels of cognitive performance rev ealed that more negativ e social interactions and greater satisfaction with support were asso ciated with better gene ral cognitive ability. Better performance on speed and attention was also associated with greater satisfaction
55 with support. Over the 5 years, less satisfac tion with support was marginally associated with decline in episodic memory performance. Tests of modification by age yielded significant findings for both baseline and change in cognitive performance (see Table 1.2). Stratifying the sample into young-old versus old-old based on a median split (age 73/74) revealed differences between the young-old (estimate = 0.20, 95% CI = 0.05, 0.34) and old-old (estimate = Â–0.35, 95% CI = Â–0.60, Â–0.11) for change in general cognitive abili ty as a function of social network of friends and between the young-old (estimat e = 3.55, 95% CI = 1.40, 5.69) and old-old (estimate = Â–3.61, 95% CI = Â–7.92, 0.70) for change in speed and attention as a function of social network of friends. No other intera ctions satisfied the criteria for statistical significance. DISCUSSION In this study, we examined whether soci al resource measures were differentially important for cognitive change with aging. Th e pattern of associations found between the different social resource factors and cognitive change across multiple domains adds support for the notion that social resource f actors may be important to the cognitive health of older adults. However, the results also indicate that separate cognitive domains respond similarly to social reso urces, as we saw parallel pa tterns of associations across cognitive domains. Of the social resource factors, baseli ne satisfaction with support and social network were most consistently related to cognitive performance. Consistent with Yeh and Liu (2003), who found stronger perceived po sitive support from friends to be related to better cognitive function, we found that greater satisfaction with support was
56 associated with better general cognitive perf ormance and speed and attention, as well as less decline in episodic memory performance, and that these relations were modified by age. The relation between having a smaller social network and cognitive decline as a function of age is consistent with a number of previous studie s (Barnes et al., 2004; Bassuk et al., 1999; Beland, Zunzunegui, Alvara do, Otero, & Del Ser, 2005; Holtzman et al., 2004). These findings are also in line with the conceptual model proposed by Berkman, Glass, Brissette, and Seeman (2000) whereby social networks provide the opportunity for social support and engagement and therefore have a broader influence on the social environment than othe r aspects of social resources. Previous studies have reported that r eceipt of emotional social support is associated with cognitive performance (Bassuk et al., 1999; Seeman et al., 2001). At the main effect level, we did not find that receiving less emotional, instrumental, or informational support was related to poorer c ognitive performance. However, we did find that age modified the rela tion between baseline episodic memory performance and emotional support. Because our sample was a relatively healthy sample of older adults, we did not expect that receipt of more inst rumental or informati onal support would be associated with less cognitive decline over ti me, as this would likely have been an indication of poorer health st atus, including cognitive healt h. Similar to Seeman et al. (2001), we also found that reporting more nega tive social interactions was associated with better general cognitive ab ility. This finding may be th e result of negative social interactions providing a greater level of stimulation, which benefits cognitive functioning.
57 The ability to find associations between social resources and cognitive performance is likely related to the types of cognitive outcomes measured. One strength of the current study was the measurement of multiple aspects of cognitive functionÂ— including general cognitive ability, speed and attention (basic cognitive abilities), and episodic memory (complex cognition)Â—in orde r to explore the associations between various types of social resources and cha nge in separate cognitive domains. We found that the cognitive domains responded similarly to social resources. In contrast to Gerstorf, Herlitz, and Smith (2006), we found that the main effect of gender was associated with change in episodic memory (estimate = 0.82, p = .0001) after we controlled for education and attrition. Although the results of the current study ar e informative, we should acknowledge several limitations. First, part icipants from the Charlotte County Healthy Aging Study are Caucasian, well educated, and in relatively good health, which may limit the generalizability of the results; in addition, there was a cons iderable amount of attrition in the sample. Second, there was relatively high st ability in cognitive function over time, which may have reduced our ability to detect associations. Third, we did not correct for multiple comparisons. However, given the paucity of previous research examining multiple aspects of social resources, we view these results as a starting point for future research. Finally, we were unable to fully e xplore the directionality of the association (i.e., whether a lower level of social re sources leads to poor cognitive function, or whether poor cognitive function leads to a lower level of social resources). In conclusion, our results suggest that the social environment may be important for the cognitive health of older adults. Thes e findings are especially important in that
58 social resources are amenable to change. Future studies need to (a) examine the determinants of social resources in older adults (e.g., socioeconomic status, geographic proximity to family, or number of children) in order to identify persons at risk for cognitive decline, (b) examine the influence of health and level of social engagement in these associations, and (c) include three or mo re waves of data to assess the reciprocal relations between social resources and cognitive change with aging.
59Table 1.2 Models predicting cognitive performance as a function of social resourcesaEstimate p Estimate p Estimate p Fixed Effects Intercept NegInt 0.42 0.03 -2.090.48-0.180.48 Satisfaction 0.45 0.02 7.43 0.01 0.290.22 Emotional -0.080.641.070.690.050.82 Informational 0.060.76-4.270.130.040.88 Instrumental 0.010.94-4.710.090.180.45 Family 0.100.574.830.080.400.09 Friends 0.130.493.740.17-0.090.70 NegInt x Age 0.01-0.610.470.28-0.010.71 Satisfaction x Age 0.040.121.04 0.01 0.12 0.001 Emotional x Age 0.030.300.560.210.09 0.02 Informational x Age -0.010.73-0.570.180.010.76 Instrumental x Age -0.020.47-0.490.290.0010.98 Family x Age -0.06 0.05 0.100.830.060.11 Friends x Age -0.06 0.05 -0.360.40-0.040.31 Fixed Effects Time NegInt 0.090.250.110.93-0.050.63 Satisfaction 0.090.221.240.300.180.06 Emotional -0.050.450.070.95-0.020.83 Informational 0.0040.950.870.450.010.93 Instrumental -0.010.88-0.0040.990.080.36 Family 0.090.170.530.620.010.92 Friends -0.050.500.540.640.060.51 NegInt x Age 0.0030.74-0.040.840.010.34 Satisfaction x Age 0.020.070.240.180.03 0.04 Emotional x Age 0.010.370.120.510.020.18 Informational x Age 0.010.320.210.230.0040.78 Instrumental x Age -0.010.430.140.470.010.44 Family x Age -0.03 0.03 -0.120.510.0040.80 Friends x Age -0.04 0.002 -0.38 0.04 -0.010.35aAll models were adjusted for age (centered), education, gender, marital status, residency, attrition status, and personality characteristics (neuroticism, extrav ersion, openess, agreeableness, conscientiousness). Note: NegInt = Negative social interactions Global CognitionSpeed and AttentionMemory
60 Chapter 5: Study III Midlife Fruit and Vegetable Consumption and Ri sk of Dementia in Later Life in Swedish Twins T. F. Hughes, BS, R. Andel, PhD, B. J. Small, PhD, A. R. Borenstein, PhD, J. A. Mortimer, PhD, A. Wolk, DMSc, B. Johanss on, PhD, L. Fratiglioni, MD, PhD, N. L. Pedersen, PhD, and M. Gatz, PhD
61 ABSTRACT Objective: To examine the association between fruit and vegetable consumption in midlife and the risk for all types of dementia and AlzheimerÂ’s disease (AD). Methods: Participants were 3,306 members of the population-based Swedish Twin Registry who completed a diet questionnair e approximately 30 years prior to cognitive screening and full clinical evaluation for dementia as part of the HARMONY study. A total of 300 twins were diagnosed with de mentia. Eighty-one complete twin pairs discordant for dementia (50 discordant for AD) were identified from among the participants. Data were analyzed with case-control and co-twin control designs. Results: In the case-control analyses medium or great consumption of fruits and vegetables, compared to no or small, was asso ciated with decreased risk of all types of dementia (OR 0.73, 95% CI 0.53-1.00) a nd AD (OR 0.59, 95% CI 0.42-0.85) after adjustment for demographic characteristics a nd lifestyle variables. Effect-modification was observed, with a stronge r inverse association be tween fruit and vegetable consumption and AD risk for women (vs. men) those who consumed one or more drinks per week (vs. none), and those who reported an gina in midlife (vs. those who did not). Results from the co-twin analyses were uni nformative because twin pairs were rarely dissimilar in their extent of fr uit and vegetable consumption. Conclusion: A diet with medium or great consumpti on of fruits and vegetables in midlife may protect against dementia and AD later in life, especially when some other risk factors are present.
62 INTRODUCTION Nongenetic risk factors are thought to play an important role in the etiology and expression of dementia and may be the focu s of interventions to reduce disease risk (Andel, Hughes, & Crowe, 2005). Several epid emiologic studies sugge st that fruit and vegetable intake may be related to demen tia (Commenges et al., 2000; Dai et al., 2006; Engelhart et al., 2002; Morris et al., 2002). Given the long pr eclinical phase of dementia (Bckman, Jones, Berger, Laukka, & Small, 2005; Elias, Beiser, Wolf, Au, White, & DÂ’Agostino, 2000) and evidence that difficultie s performing instrumental activities of daily living (Barberger-Gateau, Fabri goule, Helmer, Roach, & Dartigues, 1999) and weight loss (Stewart, Masaki, Xue, Peila, Petrovitch, White, et al., 2005) occur during this phase, the relation between diet and demen tia can be misinterpreted in the absence of a sufficient time lag between dietary asse ssment and dementia onset. Further, the observed associations found between diet a nd dementia may also be accounted for by genetic or early life influences since demen tia is highly heritable (Gatz et al., 2006) and dietary habits in adulthood may be influenced by genetics (Heitmann, Harris, Lissner, & Pedersen, 1999; Rankinen & Bouchard, 2006) or by childhood and adolescent behaviors (Mikkila, Rasanen, Raitakari, Pietinen, & Viikari, 2005). Ho wever, these relationships remain largely unexplored. We examined the relation between the rela tive intake of fruits and vegetables at midlife and the risk of dementia approximately three decades later in members of a large population-based twin study. Using case-contro l analyses, we tested whether a medium or great proportion of fruits and vegetables in the diet offered prot ection against dementia and AD compared to no or small proportion. To test whether unm easured genetic or
63 familial factors could account for observed relations using a case-control design, we analyzed the same data usi ng a co-twin control design. METHODS Participants The Swedish Twin Registry (STR) is the largest twin registry in the world and consists of three population-ba sed cohorts of like-sexed twin pairs. It was initially designed to study the relative importance of genes and environment on several diseases (Lichtenstein, De Faire, Floderus, Svarteng ren, Svedberg, & Pedersen, 2002). This study draws from the members of the cohort born between 1886 and 1925 who responded to a questionnaire mailed in 1967. In 1998, the HARMONY study was initiated to examine the relative influences of environmental and genetic fact ors on the etiology of AD and ot her dementias, to identify genes that increase the risk for dementia, and to identify specific environmental risk or protective factors for dementia and AD. Members of HARMONY consisted of all twins who were at least 65 years of age at time of assessment in HARMONY (Gatz, Fratiglioni, Johansson, Berg, Mortimer, Reynolds, et al., 2005). The HARMONY study was reviewed and approved by the institutional revi ew board of the University of Southern California and the regional ethics board at the Karolinska Institute. In all, 5,692 twins eligible for the HARMONY study had data for fruit and vegetable intake from 1967. Dementia stat us was available for 3,778 of these twins, based on two-stage case ascertainment consisting of cognitive screening followed by clinical diagnostic evaluation. Among the others, 1,020 refused to participate, 255 were not reachable, 283 were unable to be intervie wed and an informant was not available, 104
64 died before they could be screened, and 252 were screened but not seen for a clinical evaluation either by design (b ecause their co-twin was long deceased and the pair would be uninformative) or due to refusal. Afte r excluding those with incomplete covariate data, 3006 control participants and 300 participants diagnosed with dementia (199 with AD) were used for the case-control analyses ( 58.1% of those eligible ). For the co-twin analyses, we identified 81 twin pairs discordant for dementia (18 of whom were discordant for proportion of fruits and vegetabl es in diet), and 50 pairs discordant for AD (12 of whom were discordant for proportion of fruits and vegetables in diet). Measures Dementia Diagnosis. Individuals were screened and clinically evaluated for dementia as part of the HARMONY study. A random sample of HARMONY members was selected each month for the primary te lephone screening phase to identify twins positive for cognitive dysfunction and possible dementia (Gatz, Reynolds, John, Johansson, Mortimer, Pedersen, 2002). T hose who screened positive for cognitive dysfunction and their co-twin, even if they screened negative, were contacted for a follow-up clinical evaluation of dementia st atus according to the DSM-IV diagnostic criteria (details of the study design, includi ng dropout analyses can be found elsewhere). Dietary Assessment. Included in the STR 1967 ques tionnaire regarding lifestyle factors were 23 items pertaining to dietary ha bits. Fruit and vegetable consumption was assessed by a single item on a 4point relative scale with resp onse choices being Â“great partÂ”, Â“medium partÂ”, Â“small partÂ”, or Â“no partÂ”. For the current study, the Â“noÂ” and Â“smallÂ” part categories and the Â“mediumÂ” and Â“greatÂ” part categories were collapsed due
65 to the small number of participants reporting that fruits and vegetables made up Â“no partÂ” or Â“great partÂ” of their diet. Covariates The demographic characteristics of age at follow-up (continuous), sex (male/female), and education were include d as covariates. Education was measured as basic versus more than basic. In the first part of the 20th century in Sweden, a basic education consisting of 6 years or, later, 7 years of school was manda tory. Participants described their education as the highest level attained, that is, basic education, gymnasium, vocational hi gh school, or university. In addition, baseline (1967) values re ported for current smoki ng status (yes/no), alcohol consumption (no drinks per week/one or more drinks per week), exercise (hardly any or light/regular or hard), body mass i ndex (BMI=weight[kg] over height squared [m2]; measured as BMI <25 and BMI 25), self-reported angina pectoris (yes/no), and total food intake in comparison to others (l ess or as much/more or much more) were regarded as potentially confounding variables. Unmeasured genetic and familial factors were controlled by design in the co-twin control design. Statistical Analyses The data were analyzed using two study designs: case-control and co-twin control. The case-control analyses compare cas es to unrelated contro ls to evaluate the risk for disease given a partic ular exposure. The co-twin de sign uses disease-discordant twin pairs by comparing the diagnosed twin to the non-demented co-twin. The characteristics of the participants by diagnostic status and by level of fruit and vegetable consumption were compared using independent samples t -tests for continuous variables and 2 test for categorical variables. In the case-control analysis, we
66 examined the risk for dementia in relati on to the relative c onsumption of fruit and vegetables using unconditional logistic regres sion to calculate crude and adjusted odds ratios (ORs) and 95% confidence intervals (CIs ). To further reduce the possibility that participants had signs of precl inical dementia at the time of the dietary assessment, in a secondary analysis we restricted the sample to those who were 60 years of age or younger in 1967. We also examined effect-modificat ion by age group in 1967 (<51 years, 51-61 years, >61 years), sex, educa tion, smoking, alcohol, BMI, se lf-reported angina, total food consumption compared to others, and exercise by estimating multiplicative interaction terms in fully adjusted logistic regression m odels for dementia and AD. All confidence intervals were adjusted to account for data de pendence due to the inclusion of complete twin pairs (Moradi, Hans-Olov, Ekrom, Wendren, Floderus, & Lichtenstein, 2002). Similar analyses were conducted for th e co-twin control study using conditional logistic regression to estimate ORs and 95% CIs based on within-pair comparisons. Within each twin pair, the twin who reported greater consumption of fruits and vegetables was assigned a value of 1 and th e co-twin was assigned a value of 0. When twins reported the same intake level, both tw ins were assigned a value of 0 and did not contribute to the calculation of the point estimate. Crude and adjusted ORs were calculated. For adjusted models, mean valu es were imputed for smoking status, alcohol consumption, exercise, angina pe ctoris, and total food intake in comparison to others in order to retain the maximum nu mber of twin pairs in the analyses. Less than 10% of the data were missing on each imputed covariate. All analyses were conducted using SAS version 9 (SAS Institute, 2003) with pvalues less than 0.05 (two -tailed) interpreted as being significant.
67 RESULTS Case-Control Analysis Comparisons between the 3,306 participan ts and the 1914 drop-outs showed that the drop-outs were on av erage 0.92 years older ( t = 8.27, p < 0.001), more likely to be female (70% of the drop-outs ve rsus 62% of the participants, 2 = 32.56, p < 0.001), less likely to attain more than basic level of e ducation (25% of the drop-outs versus 36% of the participants, 2 = 60.48, p < 0.001), and less likely to consume one or more drinks of alcohol weekly (61% of the drop-outs versus 69% of the participants, 2 = 39.21, p < 0.001). The drop-outs did not differ from the participants in term s of current smoking status, angina pectoris, BMI, exercise, tota l food intake in comparison to others, or consumption of fruits and vegetables in 1967 ( p > 0.05). The characteristics of the cases and controls are show n in Table 1.3 for dementia and AD status. Compared to controls, cases were older at baseline and at cognitive screening, were more likely to be female and to have completed less than basic education, were less likely at midlife to report being a current smoker, to drink alcohol, or to consume some or much more food compared to others. AD cases also were also more likely to report that fruits and vegetables made up no or small part of their diet. Table 2.3 presents the char acteristics of the sample by relative consumption of fruits and vegetables at midlife. A larger num ber of participants repo rted that fruits and vegetables made up a medium or great proporti on of their diet compared to no or small part. Women reported a greater consumption of fruits and vegetabl es compared with men, as did those with more than a basic education, non-smokers, those without selfreported angina and those who engaged in ha rdly any or light phys ical exercise.
68 The association between fruit and vegeta ble consumption at midlife and dementia or AD is shown in Table 3.3. Compared w ith those reporting no or small proportion of fruits and vegetables, those whose diet consiste d of a medium or grea t part of fruits and vegetables had a reduced odds of demen tia and AD after adjusting for age, sex, education, current smoking status, alcohol c onsumption, exercise, BMI, self-reported angina, and total food consumption in comparis on to others. The risk of dementia was reduced by 27% (95% CI 0.53-1.00), and the AD risk was reduced by 41% (95% CI 0.420.85) in the fully adjusted models. Results of the analyses restricted to those less than 60 years of age (n = 2929 contro ls; n = 269 dementia cases; n = 175 AD cases) in 1967 were not substantially different. A medium or great part compared to no or small part still appeared to offer protection against de mentia (OR 0.70, 95% CI 0.51-0.96) and AD (OR 0.57, 95% CI 0.39-0.83) in the fully adjusted models. We examined whether the association between fruit and vegetable consumption and risk of AD was modified by demogra phic and lifestyle variables by estimating multiplicative interaction terms. The analyses revealed that sex, alcohol drinking and the presence of self-reported angi na in midlife modified the a ssociation between fruits and vegetables and risk of AD (Table 4.3). The inverse association between fruit and vegetable consumption and AD risk was st ronger for women, those who consumed at least one or more drinks pe r week, and those who reported having angina in midlife. Including all cases of dementia did not substantially change the results.
69 Co-Twin Analysis Results of the co-twi n control analyses are displayed in Table 5.3. None of the odds ratios approached statistic al significance. Notably, only 22% of the pairs who were discordant for dementia were also discor dant for fruit and vegetable consumption. DISCUSSION Using dietary assessments collected a bout three decades prior to dementia evaluation as part of a large population-base d study of twins, we found in case-control analyses that the consumption of fruits a nd vegetables in midlife may contribute to protection against dementia and AD in later life. Fruits and vegetables contain a number of compounds with antioxidant properties (e.g. vitamins C and E, carotenoids, and polyphe nols). Previous epidemiologic studies have reported that higher intake of antioxi dant compounds from diet ary sources such as fruits and vegetables are associated with a reduced risk of dementia (Commenges et al., 2000; Dai et al., 2006; Engelhart et al., 2002; Mo rris et al., 2002). It has been proposed that these compounds offer neuroprotection by scavenging free radicals whose accumulation in the brain may contribute to th e pathogenesis of dementia (Behl, 1999). Fruits and vegetables also contain Vitamin B, which may be important for AD prevention since Vitamin B12 deficiency also has been reported to increase the risk of AD (Wang, Wahlin, Basun, Fastborn, Winbl ad, & Fratiglioni, 2001). The protective effect of higher frui t and vegetable intake in our study was modified by sex, drinking alc ohol and having self-reported angi na pectoris in midlife. These findings suggest that those who are at higher risk for AD, specifically women (Andersen, Launer, Dewey, Letenneur, Ott, Copeland, et al., 1999) and those with
70 vascular risk factors (Luchsinger, Reitz, Honig, Tang, Shea, & Mayeux, 2005; Rosendorff, Beeri, & Silverman, 2007) may also benefit more from fruits and vegetables in their diet in midlife. Certain types of alcohol, mainly wine, contain flavonoids that offer protection against dementia (Truelse n, Thudium, & Grenbaek, 2002). Despite a link between excess alcohol consumption and vitamin B1 (thiamine) deficiency and WernickeÂ–Korsakoff Syndrome (Thomson & Marshall, 2006), moderate alcohol consumption of any type has been shown to reduce the risk of dementia (Mukamal, Kuller, Fitzpatrick, Longstreth, Mittleman, & Siscovick, 2003; Ruitenberg, van Swieten, Witteman, Mehta, van Diujn, Hofman, et al., 2002), AD (Luchsinger, Tang, Siddiqui, Shea, & Mayeux, 2004), and the risk of vascular conditions that may also affect dementia risk (Mukamal, Chung, Jenny, Kuller, Longstret h, Mittleman, et al., 2006). Therefore, the consumption of a medium or great part of fruits and vegetables in the diet and moderate alcohol consumption may have a synergistic influence on dementia and AD risk. This possibility should be explored. A majority of studies on the relation between fruit and ve getable intake and incident dementia have had relatively shor t intervals between dietary assessment and dementia diagnosis, ranging between 4 and 6 years (Commenges et al., 2000; Dai et al., 2006; Engelhart et al., 2002; Morris et al., 2002). With such a short interval, it is possible that those who were to develop dementia alre ady had made changes to their diets at the baseline assessment. Our findings suggest th at fruit and vegetable intake measured in midlife, long before signs of dementia are like ly to appear, has an effect on the risk of dementia in later life. Restricting our samp le to include those 60 years or younger at the
71 time of dietary assessment did not alter our findings, further supporting the inverse association between midlife fruit and vegetable intake and subsequent risk of dementia. Our case-control findings differ from those of the Honolulu-Asia Aging Study (HAAS) where midlife intake of antioxidant s including beta-carotene, vitamin C, and flavonoids was not significantl y associated with incident dementia and its subtypes (Laurin et al., 2004). This may be related to th e different methods of data collection. We used a self-report of dietary habits whereas the HAAS study examined beta-carotene, vitamin C, vitamin E, and total energy intake from food diaries/recal ls or food frequency questionnaires. There may also be differen ces in dietary preferences between the study populations. In contrast to the case-cont rol results, the co-twin control results were statistically non-significant. When co-twin control results do not replicate casecontrol results, this pattern typically indicates that genetic and unmeasured early life influences may explain the case-control findings. Previous work from the STR and HARMONY demonstrates that dementia is strongly heri table (Gatz et al., 2006) and th at food preferences reflect the influence of both genes and shared envir onment (Heitmann et al., 1999). Early life environment influences also affect the ri sk of dementia (Borenstein, Copenhaver, & Mortimer, 2006) and dietary preferences into adult hood (Mikkila et al., 2005). In the sample of twin pairs here, only 22% of the pa irs who were discordant for dementia were also discordant for fruit and vegetable consum ption. Taken together, the results of this study suggest that the benefici al effects of fruit and ve getable consumption may be greatest if implemented in early life when lif e-long dietary habits are being formed.
72 Limitations of this study should also be mentioned. First, variables assessed at baseline, including the intake of fruits a nd vegetables, were self-reported and hence susceptible to social desirability bias. Howe ver, such misreporting is unlikely to vary across cases and controls and th erefore would be unlikely to bi as the results. Second, diet was measured only at one time point. This does not allow the stability of the diet to be assessed. However, only 13% of controls and 12% of dementia cases (11% of AD cases) reported a past change in their diet, and adju stment for this variable did not alter our findings. Third, we were unable to adjust for total energy intake. We did adjust for selfreported total food consumption compared to others and BMI, which did not confound our findings. Finally, using prevalent cases of de mentia can lead to a differential survival bias. However, if lower fruit and vegetable intake increases the risk of mortality (Genkinger, Platz, Hoffman, Comstock, & Helzlsouer, 2004), then our results would likely underestimate the association with de mentia. Future studies might examine patterns of food intake at midlife with incident dementia cases.
73Controls Dementia casesp -valueAlzheimer's Disease casesp -value*n3,006300199 Age at baseline, Mean (SD) 47.67 (4.74)52.50 (5.68)<.00152.63 (5.64)<.001 Age at cognitive screening Mean (SD) 79.16 (4.70)83.93 (5.57)<.00184.09 (5.57)<.001 Follow-up years, Mean (SD)31.48 (0 .91)31.43 (0.96)0.3331.46 (0.95)0.69 Sex, % Female59.4573.67 <.00177.39<.001 Education % more than basic38.7222.33<.00123.62<.001 Smoke, % Yes30.2721.33<.00120.600.04 Drink alcohol, % Yes71.4956.33<.00156.78<.001 Body mass index % 258.089.670.347.540.78 Angina pectoris, % Yes25.3524.000.6123.120.48 Total food intake compared to others, % some or much more9.615.330.014.020.01 Exercise, % hardly any or light81.8786.000.0786.430.10 Fruits and Vegetable Consumption0.140.03 % No or Small Part20.3624.0026.63 % Medium or Great Part79.6476.0073.37 *Independent samples t-test or Chi-square statistic comparing controls and cases. Table 1.3 Characteristics of th e Participants in the Case-Control Study by Disease Status
74Table 2.3 Characteristics of Particip ants in the Case -Control Study by Relative Consumption of Fruits and Vegetables at Midlife No or Small PartMedium or Great Part p -value* n6842622 Age at baseline, Mean (SD) 48.19 (4.97)48.09 (5.04)0.62 Age at cognitive screening, Mean (SD)79.65 (4.91)79.57 (5.00)0.70 Follow-up years, Mean (SD)31.46 (0.92)31.48 (0.91)0.55 Sex, % Female41.6765.71<.001 Education % more than basic28.9539.40<.001 Smoke, % Yes34.9428.03<.001 Drink alcohol, % Yes68.7170.480.37 Body mass index % 258.928.050.46 Angina pectoris, % Yes29.0924.220.01 Total food intake compared to others, % some or much more10.238.960.31 Exercise, % hardly any or light78.8083.140.01* Reflects Mentel-Haentzsel chi-square test for trend across levels of fruits and vegetables or test for trend within a linear univariate regression model. Relative Consumption of Fruits and Vegetables
75Table 3.3 Case-Control Analyses of the Association between Midlife Fruit and Vegetable Consumption and Dementia or Alzheimer's Disease (AD) No or Small PartMe dium or Great Part p -value Dementia Cases/Controls72/612228/2394. AD Cases/Controls53/612146/2394. Crude OR (95% CI) Dementia1.00 (ref.) 0.81 (0.61-1.07)0.14 AD1.00 (ref.) 0.70 (0.51-0.97)0.04 Adjusted OR* (95% CI) Dementia1.00 (ref.) 0.72 (0.53-0.99)0.03 AD1.00 (ref.) 0.59 (0.41-0.85)<0.01 Adjusted ORÂ† (95% CI) Dementia1.00 (ref.) 0.73 (0.53-1.00)0.04 AD1.00 (ref.) 0.59 (0.42-0.85)<0.01 Note. OR = odds ratio; 95% CI = 95% confidence interva l; ref. = reference group. models adjusted for age at cognitive screening (continuous), sex (men/women), and education (basic/more than basic). Â† models adjusted for variables mentioned above and sm oking (yes/no), alcohol drinking (yes/no), body mass index (<25/ 25), total food compared to others (less or more/some more or much more), angina pectoris (yes/no), and exercise (hardly any or light/r egular or hard). Relative Consumption of Fruits and Vegetables
76No or Small PartMedium or Great Part OROR (95% CI) Stratum Sex Male (Cases/Controls)14/37531/844 1.00 (Ref)1.08 (0.55-2.09) Female (Cases/C ontrols)39/237115/1550 1.00 (Ref)0.46 (0.31-0.70) p for interaction <.001 Alcohol Yes (Cases/Controls)32/42481/1725 1.00 (Ref)0.53 (0.33-0.85) No (Cases/Controls)21/18865/669 1.00 (Ref)0.75 (0.43-1.31) p for interaction <.001 Self-reported angina pectoris Yes (Cases/Controls)18/17528/587 1.00 (Ref)0.32 (0.16-0.63) No (Cases/Con trols)35/437118/1807 1.00 (Ref)0.74 (0.49-1.39) p for interaction 0.001 *Models adjusted for age at cognitive screen ing (continuous), sex (men/women), education (basic/more than basic), smoking (yes/no), and al cohol drinking (yes/no) body mass index (<25/ 25), total food consumption compared to ot hers (less or more/some more or much more), angina pectoris (yes/no), and exercise (hardly any or light/regular or hard). Table 4.3 Stratified Analyses for Fruit and Vege table Consumption and Risk of Alzheimer's Disease Relative Consumption of Fruits and Vegetables
77Table 5.3 Co-Twin Control Analyses of the Association between Midlife Fruit and Vegetable Consumption and Risk of Dementia and Alzheimer's Disease Total pairs Cases higher/Cotwin higher OR (95% CI) pvalue Dementia 8110/8 AD 507/5 Crude OR (95% CI) Dementia1. 25 (0.49-3.17) 0.64 AD1.40 (0.44-4.41) 0.57 Adjusted OR* (95% CI) Dementia1. 25 (0.49-3.20) 0.64 AD1.40 (0.44-4.43) 0.56 Adjusted OR Â† (95% CI) Dementia1. 32 (0.48-3.62) 0.59 AD1.53 (0.40-5.94) 0.54 Note. OR = odds ratio; 95% CI = 95% confid ence interval; ref. = reference group. models adjusted for educa tion (basic/more than basic). Â† models additionally adjusted for smoking (yes/no), alcohol drinking (yes/no), body mass index (<25/ 25), total food compared to others (less or more/some more or much more), angina pectoris (yes/no), and exercise (hardly any or light/regular or hard).
78 Chapter Six: Concluding Remarks With the growing number of older adults the identification and implementation of strategies to maintain cognitive health with aging has profound implications for future nursing home usage, healthcare costs, caregiver burden, personal and societal resources, as well as quality of life in general (Andel et al., 2005). Although the idea that older adults can play a role in th eir cognitive health is intuitively appealing, evidence from previous studies is equivocal. In an attempt to contribute to the state of knowledge of how environmental factors influence cognitive functioning with aging, the purpose of this dissertation was to conduct three studies to determine whether lifestyle activities, social resources, and fruit and vegetable consumpti on may potentially be useful strategies to maintain cognitive functioning or reduce the risk of dementia with aging. The findings of the three studies will be summarized in the following section. The first study examined the influence of engagement in lifestyle activities on age-related differences in cognitive speed perf ormance in a sample of participants from the Victoria Longitudinal Study. The fi ndings supported the notion that higher engagement in integrative and novel informati on processing activities is related to faster and less variable cognitive speed performance on select reaction time tasks. The results also suggested that the level of engage ment in activities moderated age-related differences in cognitive speed performance. These results of this study are consistent with previous studies where more cognitively demanding activities are related to cognitive performance, and that age-relate d differences in cognitive performance are
79modified by engagement in certa in types of lifestyle activitie s in old age. Overwhelming support for the idea that engagement in activi ties would be more closely associated with inconsistency in cognitive pe rformance compared to mean-level was not found, but may be attributed to the fact that the VLS sample is a select sample whose cognitive level, or neurological integrity, was not compromised enough to detect strong er associations. Data from the Charlotte County Healt hy Aging Study were used to examine the association between social res ources and cognitive change over five years in Study II. In general, the results suggest th at lower satisfaction with soci al support that is received from others is associated with decline in episodic memory performance over five years. Significant interactions between age and soci al networks of family and friends and satisfaction with support were also found for the separate cognitive domains such that these resources become more important with increasing age to main tain general cognitive ability, as well as speed and at tention abilities. The results suggest that social resources may be differentially important for cognitive change, but that different cognitive domains respond in a similar pattern to social resources. The final study of the dissertation ex amined whether midlife consumption of fruits and vegetables was associated with the risk of dementia and AlzheimerÂ’s disease in members of the Swedish Twin Registry a nd HARMONY studies. A reduced risk of all types of dementia and AD was found for those who consumed a medium or great proportion of fruits and vegetables in their di et compared to no or small proportion in the entire sample (case-control analysis). This reduced risk was greater for females compared to males, those who consumed al cohol at least once a week compared to abstainers, and for those who self -reported angina in midlife. No significant associations
80were found when the sample was reduced to only complete twin pairs in the co-twin control analysis. However, because few twin pairs were found to be discordant for both dementia and the amount of fruits and vegetabl es in their diet in midlife, these findings may be considered uninformative. The protec tive effect of fruits and vegetables in the case-control study is consistent with the majo rity of previous studi es (Commenges et al., 2000; Dai et al., 2006; Engelhart et al., 2002; Morris et al., 2002) The possibility remains that genetic and early life influences account for these findings. Collectively, these studies suggest th at environmental factors can potentially influence cognitive functioning in later life. The first study added new insight into how engagement in lifestyle ac tivities is associated with inconsistency in cognitive performance in addition to mean-level perf ormance. This is important given that inconsistency has been shown to be a sensitive marker of the health of the central nervous system (Hultsch et al., 2002). The second st udy shed light on whether different types of social resources are differentially associated w ith cognitive performance. The majority of previous studies had focused on social networ ks, but research also suggested that other types of social resources may be importan t for cognitive functioning. Understanding the relative importance of different types of social resources can help target future research. The final study drew upon existing data on fr uit and vegetable consumption collected approximately 30 years before dementia assessm ent. The findings of this study are less likely to be biased by the pr eclinical dementia phase comp ared to previous studies because of the extended time period between di et and dementia assessments, and thus the interpretation that greater consumption of fr uits and vegetables is protective against dementia is likely to be more valid.
81LIMITATIONS Although each of the studies provided new insight into the relation between environmental factors and cognitive h ealth, their limitations should also be acknowledged. First, Study I was a cross-s ectional examination of the association between lifestyle activities and cognitive speed pe rformance. This type of analysis limits our ability to conclude whether engagement in activities is asso ciated with cognitive change and to further determine the directionali ty of the association (i.e. does decline in activities proceed decline in cognition or vice ve rsa). Longitudinal data with two or more follow-up waves and the use of latent differe nce score or dual change score modeling should be carried out in future studies. Sec ond, the VLS sample can also be considered a select sample of older adults who are re latively highly educated, in good health, and whose activity level and cogniti ve performance are above average. Finally, our findings may have been attenuated since the activities assessed did not train speed of processing abilities and prior research has shown there to be little tran sfer of benefits outside the cognitive domain that is trained. Future studies should examine th e association between these activities and inconsistency in higher order cognitive tasks (e.g. memory, verbal abilities). Limitations of Study II include the fact that the CCHAS sample can also be considered a select sample of older adults, which limits the external validity of the findings as well as limited the amount of variab ility in cognitive change. In addition, a number of participants we re lost during the study fo llow-up period which may have biased the findings. Lastly, although this study examined cognitive change over time, we were not able to assess the temporal order of the associations between social resources
82and cognitive functioning. Fo llow-up studies are needed us ing three or more waves of data collection in order to assess whether a lower level of social resources leads to declines in cognitive function, or whether poo r cognitive function leads to fewer social resources. The final study was limited by the measurement of fruit and vegetable consumption being based on one question that asked participan ts to report the proportion of fruits and vegetables in their diet on a f our-point scale ranging fr om Â“no partÂ” to Â“great partÂ”. This did not permit an estimate of the vitamins cons umed or total energy intake, and also increased the chan ces of misclassification of exposure. Although the measurement of diet can be considered pros pective, the cases are considered prevalent with respect to dementia diagnosis, which ma y have led to confounding by a survival bias or underestimation of the actual effect. Fu ture studies should be conducted looking at dietary patterns since interactions among foods likely affect dementia risk. In addition, incident dementia cases should be studied in order to determine whether dietary factors affect the age of onset in tw ins concordant for dementia. A general limitation across the studies that is also noteworthy is the fact that three different samples were used for each study, wh ich does not allow for comparisons to be made with respect to the relative importance of each of the envi ronmental factors in relation to cognitive health. FUTURE DIRECTIONS This dissertation research was driven by the desire to further our understanding of how modifiable risk factors are related to nor mal cognitive aging and the risk of dementia in late life. Ultimately, these, and other similar studies, will guide future intervention
83studies designed to help older adults improve maintain, or delay the onset of cognitive impairments associated with normal aging a nd dementing processes. Key questions for future studies in this arena are with respect to timing. For exam ple, is it ever too late to change behavior, or can adults of any age benefit their cognitive health by modifying their behavior? Will the greatest benefits be seen if changes are implemented earlier rather than later in life, and how long do individuals need to engage in such behaviors to benefit their cognitive health ? It is imperative that fu ture studies investigating environmental risk factors for cognitive decl ine and dementia are conducted from a life course perspective in order to disentangle the temporal or der of the relations between these factors and cognitive function. Another key issue when examining the a ssociation between lifestyle variables and cognitive health with aging is the underlyi ng mechanism of action. The current studies were limited by the data available, but future studies should examine how each of these factors is affecting cognitive re serve at the neurophysiologic le vel. Studies will need to include measures of brain imaging and biomarke rs in order to investigate the structural and functional underpinnings of the current fa ctors under investigation as well as other lifestyle characteristics. Findings from these studies will then increase our ability to develop more effective and specific strategi es for those most vulnerable to cognitive deficits with aging. Future studies should also ex amine if there are intervening variables at play, such as vascular h ealth conditions or other comorbidites, which may affect the risk of cognitive declin e or dementia. The potential to improve or maintain cognitive abilities by controlling risk f actors in the pathway leading to cognitive detriments could be a more feasible option since many treatmen ts and strategies are already available.
84 One final direction for future studies is examining ways to identify as early as possible those most susceptible to poor cogni tive health in late lif e who would stand to benefit most from the strategies investigat ed in the current disse rtation. Although these factors may in and of themselves help to identify these individua ls, other markers of impending cognitive decline associated with aging and dementia should also be investigated in parallel, including unmodifiab le risk factors. This will allow for the implementation of strategies at the most opportune time in order to maximize their effectiveness and ultimately reduce the inci dence of cognitive impairments and dementia in the future. In conclusion, the current dissertation addressed an im portant public health issue that is expected to affect millions of Am ericans in the near future as the baby boom generation comes of age: cogniti ve health in late life. For most, the fear of losing cognitive abilities with aging, especially memo ry, is greater than physical disability. Although the results found in the three studies de scribed here are not definitive, they lend support for the belief that individuals can pl ay an active role in their cognitive health across the lifespan. Most importantly, th e modifications suggested by the current findings pose no health risks, so adults of all ages should cons ider their everyday behaviors and environment as a potential way to successfully age in terms of cognition.
85 References Aartsen, M.J., Smits, C.H.M., van Tillburg, T., Knipscheer, K.C.P.M, & Deeg, D. J. H. (2002). Activity in older adults: cause or consequence of cognitive functioning? (2002). A longitudinal st udy on everyday activities and cognitive performance in older adults. Journal of Gerontology: Psychological Sciences, 57B P153-P162. Aiken, L.S. & West, S.G. (1991). Multiple regression: testing and interpreting interactions Thousand Oaks, CA: Sage Publications, Inc. AlzheimerÂ’s Association (2007). AlzheimerÂ’s disease facts and figures Chicago, IL: AlzheimerÂ’s Association. American Psychiatri c Association (1994). Diagnostic and statistical manual of mental disorders (DSM-IV) Washington, DC: American Psychiatric Association. Andel, R., Hughes, T.F., & Crowe, M. (2005). Strategies to reduce the risk for cognitive decline and dementia. Aging Health, 1, 107-116. Andersen, K., Launer, L.J., Dewey, M.E., Lete nneur, L., Ott, A., Copeland, J.R., et al. (1999). Gender differences in the inci dence of AD and vascular dementia: The EURODEM Studies. EURODEM Incidence Research Group. Neurology, 53, 1992-1997. Anstey, K. & Christensen, H. (2000). Educat ion, activity, health, blood pressure, and apolipoprotein E as predictors of cognitive change in old age: a review. Gerontology, 46 163-177. Arbuckle, T.Y., Maag, U., Pushkar, D., & Ch aikelson, J.S. (1998). Individual differences in trajectory of intellectual de velopment over 45 y ears of adulthood. Psychology and Aging, 13, 663-675. Ashford, J.W., & Mortimer, J.A. (2002). Non-fa milial Alzheimer's disease is mainly due to genetic factors. Journal of Alzheimer's Disease, 4 169-177. Bckman, L., Jones, S., Berger, A-K., Laukka, E.J., & Small, B.J. (2005). Cognitive impairment in preclinical Alzheime rÂ’s disease: a meta-analysis. Neuropsychology, 19, 520-531.
86Bckman, L., Small, B.J., & Wahlin, A. (2001). Aging and Memory: Cognitive and Biological Perspectives. In J.E. Birren, & Schaie, K.W. (Eds.), Handbook of the Psychology of Aging (5th ed., pp. 349-377). San Diego, CA: Academic Press. Ball, D. (1998, April). The relationship be tween activity lifestyle self-reported health, and cognition in a representative sample of older adults. Paper presented at Cognitive Aging Conference, Atlanta, GA. Ball, K., Berch D.B., Helmers, K.F., Jobe, J.B., Leveck, M.D., Marsiske, M., Morris, J.N., Rebok, G.W., Smith, D.M., Tennstedt S.L., Unverzagt, F.W., & Willis, S.W. (2002). Effects of c ognitive training interventi ons with older adults: A randomized controlled trial. Journal of the American Medical Association, 288, 2271-2281. Baltes, P.B. & Baltes, M.M. (Eds.). (1990). Successful Aging: Perspectives from the Behavioral Sciences. New York: Cambridge University Press. Baltes, P.B., Staudinger, U.M, & Lindenberg er, U. (1999). Lifespan psychology: theory and application to intellectual function. Annual Review of Psychology, 50, 471507. Barberger-Gateau, P., Fabrigoule, C., Helmer C., Roach, I., & Dartigues, J.F. (1999). Functional impairment in instrumental activ ities of daily living: an early clinical sign of dementia? Journal of the American Geriatric Society, 47, 456-462. Barnes, L.L., Mendes de Leon, C.F., Wilson, R. S., Bienias, J.L., & Evans, D.A. (2004). Social resources and cognitive decline in a population of older African Americans and whites. Neurology, 63, 2322-2326. Bassuk, S.S., Glass, T.A., & Berkman, L. F. (1999). Social disengagement and incident cognitive decline in community-dwelling elderly persons. Annals of Internal Medicine, 131, 165-173. Behl, C. (1999). AlzheimerÂ’s disease and oxidative stress: implications for novel therapeutic approaches. Progress in Neurobiology, 57, 301-323. Beland, F., Zunzunegui, M.V., Alvarado, B., Otero, A., & Del Ser, T. (2005). Trajectories of cognitive decline and social relations. Journal of Gerontology: Psychological Sciences, 60B P320-P330. Benedict, R.H., Schretlen, D., Goninger, L ., & Brandt, J. (1998). Hopkins Verbal Learning Test-Revised: normative data and an alysis of inter-form and test-retest reliability. Clinical Neuropsychology, 12, 43-55.
87Berkman, L.F., Glass, T., Brissette, I., & Seem an, T.E. (2000). From social integration to health: Durkheim in the new millennium. Social Science & Medicine, 51, 843857. Bielak, A.A.M., Hughes, T.F., Small, B.J., & Dixon, R.A. (2007). ItÂ’s never too late to engage in lifestyle activities: significant concurrent but not change relationships between lifestyle activities and cognitive speed. Journal of Gerontology: Psychological Sciences, 62B P331-P339. Borenstein, A.R., Copenhaver, C.I., Mortimer J.A. (2006). Early-life risk factors for AlzheimerÂ’s disease. Alzheimer Disease and Associated Disorders, 20, 63-72. Borenstein, A.R., Mortimer, J.A., Wu, Y., Jureidini-Webb, F.M., Fallin, M.D., Small, B.J., Mullan, M., & Crawford, F.C. (2006). Apolipoprotein E and cognition in population-based samples of African Americans and Whites. Ethnicity and Disease, 16, 9-15. Brandt, J. (1991). The Hopkins Verbal Learni ng Test: development of a new memory test with six equivalent forms. Clinical Neuropsychologist, 5, 125-142. Burton, C.L., Strauss, E., Hultsch, D.F., Moll, A., & Hunter, M.A. (2006). Intraindividual variability as a marker of neurological dysfunction: a comparison of AlzheimerÂ’s disease and ParkinsonÂ’s disease. Journal of Clinical and Experimental Neuropsychology, 28, 67-83. Centers for Disease Control and Preventi on and AlzheimerÂ’s Association. (2007). The Healthy Brain Initiative: A National P ublic Health Road Map to Maintaining Cognitive Health Chicago, IL: AlzheimerÂ’s Association. Christensen, H. (2001). What cognitive cha nges can be expected w ith normal ageing? Australian and New Zealand J ournal of Psychiatry, 35, 768-775. Christensen, H., Dear, K.B.G., Anstey, K.J ., Parslow, R.A., Sachdev, P., & Jorm, A.F. (2005). Within-occasion intraindividual variability and preclinical diagnosis status: is intraindividual variability an indicator of mild cognitive impairment? Neuropsychology, 19 309-317. Christensen, H., Jorm, A.F., Henderson, A.S., Mackinnon, A.J., Korten, A.E., & Scott, L. R. (1994a). The relationship between health and cognitive functioning in a sample of elderly people in the community. Age and Ageing, 23 204-212. Christensen, H., & Mackinnon A. (1993). Th e association between mental, social, and physical activity and cognitive performance in young and old subjects. Age and Ageing, 22, 175-182.
88Christensen, H., Mackinnon, A., Jorm, A.F., Henderson, A.S., Scott, L.R., & Korten, A. E. (1994b). Age differences and interi ndividual variation in cognition in community-dwelling elderly. Psychology and Aging, 9 381-390. Commenges, D., Scotet, V., Renaud, S., J acqmin-Gadda, H., Barberger-Gateau, P., & Dartigues, J.-F. (2000). Intake of flavonoids and risk of dementia. European Journal of Epidemiology, 16, 357-363. Costa, P.T. & McCrae, R.R. (1992). Revise d NEO Personality Inventory (NEO-PI-R) and NEO-Five-Factor Inventory (NEO-FFI); Professional Manual. Psychological Assessment Resources: Odessa, FL. Crowe, M., Andel, R., Pedersen, N.L., Johansson, B., & Gatz, M. (2003). Does participation in leisure activ ities lead to reduced risk of Alzheimer's disease? A prospective study of Swedish twins. Journal of Gerontology: Psychological Sciences 58B, P249-P255. Dai, Q., Borenstein, A.R., Wu, Y., Jackson, J.C., & Larson, E.B. (2006). Fruit and vegetable juices and AlzheimerÂ’s disease: The Kame Project. American Journal of Medicine, 119, 751-759. Dixon, R.A., & de Frias, C.M. (2004). Th e Victoria Longitudinal Study: from characterizing cognitive aging to illust rating changes in memory compensation. Aging, Neuropsychology, and Cognition, 11 346-376. Dixon, R.A., Garrett, D.D., Lentz, T.L., MacDona ld, S.W.S., Strauss, E., & Hultsch, D.F. (2007). Neurocognitive markers of mild cogn itive impairment: exploring the roles of speed and inconsistency. Neuropsychology, 21, 381-399. Elias, M.F., Beiser, A., Wolf P.A., Au, R., White, R.F., & DÂ’Agostino, R.B. (2000). The preclinical phase of AlzheimerÂ’s dise ase: a 22-year prospective study of the Framingham Cohort. Archives of Neurology 57, 808-813. Engelhart, M.J., Geerlings, M.I., Ruitenbe r, A., van Swieten, J.C., Hofman, A., Witteman, J.C.M., et al. (2002). Dietary intake of antioxidants and risk of Alzheimer disease. Journal of the Americ an Medical Association, 287, 32233229. Farrer, L.A., Cupples, L.A., Haines, J.L., Hy man, B.T., Kukull, W.A., Mayeux, R. et al. (1997). Effects of age, sex, and ethnicity between apolipoprote in E genotype and AlzheimerÂ’s disease: a meta-analysis. Journal of the American Medical Association, 278, 1349-1356. Fillett, H.M., Butler, R.N., OÂ’Connell, A.W., Albert, M.S., Birren, J.E., Cotman, C.W., et al., (2002). Achieving and maintain ing cognitive vitality with aging. Mayo Clinic Proceedings, 77, 681-696.
89 Finkel, D., & Pedersen, N.L. (2004). Processi ng speed and longitudina l trajectories of change for cognitive abilities: The Sw edish Adoption/Twin Study of Aging. Aging, Neuropsychology, and Cognition, 11 325-345. Fratiglioni, L., Paillard-Borg, S., & Winblad, B. (2004). An active and socially integrated lifestyle in late life migh t protect against dementia. Lancet Neurology, 3 343-353. Fritsch, T., Smyth, K.A., Debanne, S.M., Petot, G.J., & Friedland, R.P. (2005). Participation in novelty-seeking leisur e activities and Alzheimer's disease. Journal of Geriatric Ps ychiatry and Neurology, 18 134-141. Gatz, M., Fratiglioni, L., Johansson, B., Ber g, S., Mortimer, J.A., Reynolds, C.A., et al. (2005). Complete ascertainment of demen tia in the Swedish Twin Registry: The HARMONY study. Neurobiology of Aging, 26, 439-447. Gatz, M., Reynolds, C.A., Fratiglioni, L., Joha nsson, B., Mortimer, J.A., Berg, S., et al. (2006). Role of genes and environments for explaining AlzheimerÂ’s disease. Archives of General Psychiatry, 63, 168-174. Gatz, M., Reynolds, C.A., John, R., Johansson, B., Mortimer, J.A., & Pedersen, N.L. (2002). Telephone screening to identify pot ential dementia cases in a populationbased sample of older adults. International Ps ychogeriatrics, 14, 273-289. Genkinger, J.M., Platz, E.A., Hoffman, S. C., Comstock, G.W., & Helzlsouer, K.J. (2004). Fruit, vegetable, and antioxidant intake and all-cause, cancer, and cardiovascular disease mortality in a community-dwelling population in Washington County, Maryland. American Journal of Epidemiology, 160, 12231233. Gerstorf, D., Herlitz, A., & Smith, J. (2006). Stability of sex differences in cognition in advanced old age: the role of education and attrition. Journal of Gerontology: Psychological Sciences, 61B P245-P249. Ghisletta, P., Bickel, J.-F., Lvdn, M. (2006) Does activity engage ment protect against cognitive decline in old age? Methodol ogical and analytical considerations. Journal of Gerontology: Psychological Sciences, 61B, P253-P261. Gold, D.P., Andres, D., Etezadi, J., Arbuckl e, T., Schwartzman, A., & Chaikelson, J. (1995). Structural equation model of in tellectual change a nd continuity and predictors of intelligence in older men. Psychology and Aging, 10, 294-303. Gonzalez-Gross, M., Marcos, A., & Pietrz ik, K. (2001). Nutrition and cognitive impairment in the elderly. British Journal of Nutrition, 86, 313-321.
90Gillette Guyonnet, S., Abellan Van Kan,G., An drieu, S., Barberger Gateau, P., Berr, C., Bonnefoy, M., et al., (2007). IANA task force on nutrition and cognitive decline with aging. The Journal of Nutrition, Health & Aging, 11, 132-152. Heitmann, B.L., Harris, J.R., Lissner, L., & Pedersen, N.L. (1999). Genetic effects on weight change and food inta ke in Swedish adult twins. American Journal of Clinical Nutrition 69, 597-602. Holtzman, R.E., Rebok, G.W., Saczynski, J.S., Kouzis, A.C., Wilcox Doyle, K., & Eaton, W.W. (2004). Social network characteri stics and cognition in middle-aged and older adults. Journal of Gerontology: Ps ychological Sciences, 59B P278-P284. Horn, J.L., & Cattell, R.B. (1966). Refineme nt and test of the theory of fluid and crystallized general intelligence. Journal of Educational Psychology, 57, 253-270. Horn, J.L., & Hofer, S.M. (1992). Major abili ties and development in the adult period. In R.J. Sternber & C.A. Berg (Eds.), Intellectual Development (pp. 44-99). New York: Cambridge University Press. Hultsch, D.F., Hammer, M., & Small, B.J. (1993). Age differences in cognitive performance in later life: relationships to self-reported health and activity life style. Journal of Gerontology: Psychological Sciences, 48B, P1-P11. Hultsch, D.F., Hertzog, C., Dixon, R.A., & Small, B.J. (1998). Memory Change in the Aged. New York: Cambridge University Press. Hultsch, D.F., Hertzog, C., Small, B.J., & Di xon, R.A. (1999). Use it or lose it: engaged lifestyle as a buffer of cognitive decline in aging? Psychology and Aging, 14 245263. Hultsch, D.F., & MacDonald, S.W.S. (2004). In traindividual variabil ity in performance as a theoretical window onto cognitive ag ing. In R. A. Dixon, L. Bckman, & L.G. Nilsson (Eds.), New Frontiers in Cognitive Aging (pp. 65-88). Oxford: Oxford University Press. Hultsch, D.F., MacDonald, S.W.S., & Dixon, R. A. (2002). Variability in reaction time performance of younger and older adults. Journal of Gerontology: Psychological Sciences, 57B P101-P115. Hultsch, D.F., MacDonald, S.W.S., Hunter, M.A., Levy-Bencheton, J., & Strauss, E. (2000). Intraindividual vari ability in cognitive perf ormance in older adults: comparison of adults with mild dementia, ad ults with arthritis, and healthy adults. Neuropsychology, 14, 588-598.
91Jang, Y., Borenstein, A.R., Chiriboga, D. A., & Mortimer, J.A. (2005). Depressive symptoms among African Ameri can and White older adults. Journal of Gerontology: Psychosocial Sciences, 60B P313-P319. Katzman, R. (1993). Education a nd the prevalence of dementia and Alzheimer's disease. Neurology, 43 13-20. Kramer, A.F., Bherer, L., Colcombe, S.J., Dong, W., & Greenough, W.T. (2004). Environmental influences on cognitive and brain plasticity during aging. Journal of Gerontology: Medical Sciences, 59A M940-M957. Krause, N. & Borawski-Clark, E. (1995). So cial class differences in social support among older adults. The Gerontologist, 35, 498-508. Krumholz, H.M., Butler, J., Miller, J., V accarino, V., Williams, C.S., Mendes de Leon, C.F., et al. (1998). Prognostic importance of emotional support for elderly patients hospitalized with heart failure. Circulation, 97 958-964. Laurin, D., Masaki, K.H., Foley, D.J., White, L. R., & Launer, L.J. (2004). Midlife dietary intake of antioxidants and risk of latelife incident dementia: the Honolulu-Asia Aging Study. American Journal of Epidemiology, 159, 959-967. Li, S.-C., Aggen, S.H., Nesselroade, J.R., & Balt es, P.B. (2001). Shor t-term fluctuations in elderly peopleÂ’s sensorimotor functi oning predict text and spatial memory performance: The MacArthur Successful Aging Studies. Gerontology, 47 100116. Liang, J. (1986). Self-reported phys ical health among aged adults. Journal of Gerontology, 41, 248-260. Lichtenstein, P., De Faire, U., Floderus, B ., Svartengren, M., Svedberg, P., Pedersen, N. L. (2002). The Swedish Twin Regi stry: a unique source for clinical, epidemiological, and genetic studies. Journal of Inter nal Medicine, 252, 184-205. Littell, R.C., Milliken, G.A., Str oup, W.W., & Wolfinger, R.D. (1996). SAS systems for mixed models Cary, NC: SAS Institute. Lubben, J.E. (1988). Assessing social networks among elderly populations. Family and Community Health, 11, 42-52. Luchsinger, J.A., & Mayeux, R. (2004). Diet ary factors and Alzheimer's disease. Lancet Neurology, 3 579-587. Luchsinger, J.A., Reitz, C., Honig, L.S ., Tang, M.-X., Shea, S., Mayeux, R. (2005). Aggregation of vascular ri sk factors and risk of inci dent AlzheimerÂ’s disease. Neurology 65, 545-551.
92 Luchsinger, J.A., Tang, M.-X., Siddiqui, M., Sh ea, S., Mayeux, R. (2004). Alcohol intake and risk of dementia. Journal of the American Geriatric Society, 52, 540-546. MacDonald, S.W.S., Hultsch, D.F., & Dixon, R. A. (2003). Performance variability is related to change in cognition: evidence from The Victoria Longitudinal Study. Psychology and Aging, 18, 510-523. Mackinnon, A., Christensen, H., Hofer, S.M., Korten, A.E., and Jorm, A.F. (2003). Use it and still lose it? The as sociation between activity and cognitive performance established using latent growth techniques in a community sample. Aging, Neuropsychology, and Cognition, 10, 215-229. McDowell, I. (2001). Alzheimer's dis ease: insights from epidemiology. Aging (Milano), 13 143-162. Mikkila, V., Rasanen, L., Raitakari, O.T., Pie tinen, P., & Viikari, J. (2005). Consistent dietary patterns identified from childhood to adulthood: The Cardiovascular Risk in Young Finns Study. British Journal of Nutrition, 93, 923-931. Mookadam, F., & Arthur, H.M. (2004). Social support and its relati onship to morbidity and mortality after acute myocardial infarction. Archives of Internal Medicine, 164 1514-1518. Moradi, T., Hans-Olov, A., Ekro m, A., Wendren, T.P., Floderu s, B., & Lichtenstein, P. (2002). Physical activity and risk fo r breast cancer a prospective cohort study among Swedish twins. International Journal of Cancer, 100, 76-81. Morris, M.C., Evans, D.A., Bienias, J.L., Ta ngney, C.C., Bennett, D.A., Aggarwal, N., et al. (2002). Dietary intake of antioxidant nutrients and the risk of incident Alzheimer disease in a biracial community study. Journal of the American Medical Association, 287, 3230-3237. Mortimer, J.A., Borenstein, A.R., Gosche K.M., & Snowdon, D.A. (2005). Very early detection of Alzheimer neuropathology and the role of brain reserve in modifying its clinical expression. Journal of Geriatric Psychiatry and Neurology, 18 218223. Mortimer, J.A., Snowdon, D.A., & Markesbe ry, W.R. (2003). Head circumference, education and risk of dementia: findings from the Nun Study. Journal of Clinical and Experimental Neuropsychology, 25 671-679. Mukamal, K.J., Chung, H., Jenny, N.S., Kuller, L.H., Longstreth, W.T., Mittleman, M.A., et al. (2006). Alcohol consumption and ri sk of coronary hear t disease in older adults: The Cardiovascular Health Study. Journal of the American Geriatric Society, 54, 30Â–37.
93 Mukamal, K.J., Kuller, L.H., Fitzpatrick, A.L., Longstreth, W.T., Mittleman, M.A., & Siscovick, D.S. (2003). Prospective study of alcohol consumption and risk of dementia in older adults. Journal of the American Medical Association 289, 1405-1413. Nesselroade, J.R., & Salthouse, T.A. (2004). Me thodological and theoreti cal implications of intraindividual variability in perceptual-motor performance. Journal of Gerontology: Psychological Sciences, 59B P49-P55. Newson, R.S., & Kemps, E.B. (2005). General life style activities as a predictor of current cognition and cognitive change in older a dults: a cross-sectio nal and longitudinal examination. Journal of Gerontology: Ps ychological Sciences, 60B P113-P120. Palmer, J., MacLeod, C.M., Hunt, E., & Davi dson, J.E. (1985). Information processing correlates of reading. Journal of Memory and Language, 24, 59-88. Park, D.C. (2000). The Basic Mechanisms accounting for age-related decline in cognitive functioning. In D. C. Park and Schwarz, N. (Eds.), Cognitive Aging: A Primer (pp. 3-21). Philadelphia, PA: Psychology Press. Park, D.C., Smith, A.D., Lautenschlager, G., Earles, J.L., Frieske, D., Zwahr, M. et al. (1996). Mediators of long-term memory performance across the life span. Psychology and Aging, 11, 621-637. Plomin, R. (1999). Genetics a nd general cognitive ability. Nature, 402 C25-C29. Rankinen, T. & Bouchard, C. (2006). Genetic s of food intake and eating behavior phenotype in humans. Annual Review of Nutrition, 26, 413-434. Reitan, R.M. & Wolfson, D. (1985). The Halstead-Reitan neuropsychological test battery. Tucson, AZ: Neuropsychology Press. Rosendorff, C., Beeri, M.S., & Silverman, J.M. (2007). Cardiovascular risk factors for AlzheimerÂ’s disease. American Journal of Geriatric Cardiology, 16, 143-149. Rosnick, C.B., Small, B.J., Borenstein Gr aves, A., & Mortimer, J.A. (2004). The association between health and cognitiv e performance in a population-based study of older adults: The Charlotte C ounty Healthy Aging Study (CCHAS). Aging, Neuropsychology, and Cognition, 11 89-99. Ruitenberg, A., van Swieten, J.C., Witteman, J.C.M., Mehta, K.M., van Duijn C.M., Hofman, A., et al. (2002). Alcohol co nsumption and risk of dementia: The Rotterdam Study. The Lancet, 359, 281-286.
94Salthouse, T.A. (2006). Mental exercise and me ntal aging: evaluati ng the validity of the Â“use it or lose itÂ” hypothesis. Perspectives on Psychol ogical Science, 1 68-87. Salthouse, T.A., Berish, D.E., Miles, J.D. ( 2002). The role of cogni tive stimulation on the relations between age and cognitive functioning. Psychology and Aging, 17 548557. SAS Institute. (2003). SAS System for Mi crosoft Windows (Version 9) [Computer software]. Cary, NC: SAS Institute Inc. Schaie, K.W. (2005). Developmental Influences on Adult Intelligence: The Seattle Longitudinal Study. New York: Oxford University Press. Schaie, K.W. & Hofer, S.M. (2001). Longitudi nal studies in aging research. In J.E. Birren & K.W. Schaie (Eds.), Handbook of the Psychology of Aging (pp. 53-77). San Diego, CA: Academic Press. Schooler, C. (1984). Psychological effects of complex environments during the life span: a review and theory. Intelligence, 8 259-281. Schooler C., & Malutu, M.S. (2001) The reciprocal effects of leisure time activities and intellectual functioning in older people: a longitu dinal analysis. Psychology and Aging, 16, 466-482. Seeman, T.E., & Crimmins, E. (2001). Social environment effects on health and aging: integrating epidemiologic and demogr aphic approaches and perspectives. Annals of the New York Academy of Sciences 954, 88-117. Seeman, T.E., Lusignolo, T.M., Albert, M., & Berkman, L. (2001). Social relationships, social support, and patterns of cognitive aging in healthy, high-functioning older adults: MacArthur studies of successful aging. Health Psychology, 20 243-255. Singer, T., Verhaeghen, P., Ghisletta, P., Li ndenberger, U., & Baltes, P.B. (2003). The fate of cognition in very old age: sixyear longitudinal findings in the Berlin Aging Study (BASE). Psychology and Aging, 18, 318-331. Sloane, P.D., Zimmerman, S., Suchindran, C., Reed, P., Wang, L., Boustani, M., et al. (2002). The public health impact of Al zheimer's disease, 2000-2050: potential implication of treatment advances. Annual Review of Public Health, 23 213-231. Small, B.J., Graves, A.B., McEvoy, C.L., Crawford, F.C., Mullan, M., & Mortimer, J.A. (2000). Is APOE-e4 a risk factor for c ognitive impairment in normal aging? Neurology, 54, 2082-2088.
95Small, B.J., Hughes, T.F., Hultsch, D.F., & Dixon, R.A. (2007). Lifestyle activities and late-life changes in cognitive performance. In Y. Stern (Ed.), Cognitive Reserve New York: Psychology Press. Solfrizzi, V., Panza, F., & Capurso, A. (2003). The role of diet in cognitive decline. Journal of Neural Transmission, 110 95-110. Stern, Y. (2002). What is cognitive reserve? Theory and research application of the reserve concept. Journal of the International Neuropsychological Society, 8 448460. Stewart, R., Masaki, K., Xue, Q.L., Peila, R., Petrovitch, H., White, R., et al. (2005). A 32-year prospective study of change in body weight and incident dementia: The Honolulu-Asia Aging Study. Archives of Neurology, 62, 55-60. Strauss, E., Bielak, A.A.M., Bunce, D., H unter, M.A., & Hultsch, D.F. (2006). Withinperson variability in response speed as an indicator of mild cognitive impairment. Aging, Neuropsychology, and Cognition, 14, 608-630. Stroop, J.R. (1935). Studies of interf erence in serial verbal reaction. Journal of Experimental Psychology, 18, 643-662. Stuss, D.T., Pogue, J., Buckle, L., & Bondar, J. (1994). Characteriza tion of stability of performance in patients with traumatic br ain injury: variability and consistency on reaction time tests. Neuropsychology, 8 316-324. Teng, E.L. & Chui, H.C. (1987). The modi fied Mini-Mental State Examination. The Journal of Clinical Psychiatry, 48, 314-318. Thomson, A.D. & Marshall, E.J. (2006). The natural history and pathophysiology of WernickeÂ’s encephalopathy and KorsakoffÂ’s psychosis. Alcohol and Alcoholism, 41, 151-158. Truelsen, T., Thudium, D., & Grenbaek, M. ( 2002). Amount and type of alcohol and risk of dementia: The Copenhagen City Heart Study. Neurology, 59, 1313-1319. Uchino, B.N., Cacioppo, J.T., & Kiecolt-Glaser J.K. (1996). The relationship between social support and physiological pr ocesses: a review with emphasis on underlying mechanisms and implications for health. Psychological Bulletin, 119, 488-531. Walker, M.P., Ayre, G.A., Perry, E.K., Wesn es, K., McKeith, I.G., Tovee, M., et al. (2000). Quantification and characterization of fluctuating cognition in dementia with Lewy bodies and AlzheimerÂ’s disease. Dementia and Geriatric Cognitive Disorders, 11 327-335.
96Wang, H.-X., Wahlin, A., Basun, H., Fastbom, J., Winblad, B., Fratiglioni, L. (2001). Vitamin B12 and folate in relation to the de velopment of AlzheimerÂ’s disease. Neurology 56, 1188-1194. Williams, B.R., Hultsch, D.F., Strauss, E.H., Hunter, M.A, & Tannock, R. (2005). Inconsistency in reaction time across the life span. Neuropsychology, 19 88-96. Wilson, R.S., Barnes, L.L., & Bennett, D.A. (2003). Assessment of lifetime participation in cognitively stimul ating activities. Journal of Clinical and Experimental Neuropsychology, 25 634-642. Wilson, R.S., Beckett, L.A., Barnes, L.L., Schneider, J.A., Bach, J., Evans, D.A., et al. (2002). Individual differences in rates of change in cognitive abilities of older persons. Psychology and Aging, 17 179-193. Yaffe, K., Lui, L.Y., Grady, D., Stone, K. & Morin, P. (2002). Estrogen receptor 1 polymorphisms and risk of cognitive impairment in older women. Biological Psychiatry, 51, 67-82. Yeh, S.-C. J., & Liu, Y.-Y. (2003). Influence of social support on cognitive function in the elderly. BMC Health Services Research, 3, 9.
98 Appendix A: Action Letter Dear Ms. Hughes: Thank you for submitting a revised version of your manuscript, "The Association between Social Resources and Cognitive Cha nge in Older Adults: Evidence from the Charlotte County Healthy Aging Study," to the Journal of Ger ontology: Psychological Sciences (JG: PS). Your lette r clearly describes the changes you have made in response to the Reviewers' and my comments, and I be lieve you have satisfact orily addressed most of our concerns. After carefully reading the revised manuscript and your detailed letter, I have decided that the manuscript is now ready for publication in JG: PS. I will forward a copy of the manuscript to our Director of Publications, Patricia Walker, for copyediting. You will be sent a copy of the editorial changes for your approval. In addition, you will receive page pr oofs after the manuscript is typeset, which you should check very carefully. You will be give n an opportunity to purchase reprints at that time. Also, I have included a copy of the certi fication form that all authors of the manuscript must sign. Please make sure the cont act information at the top of the form is correct and follow the directions on the form for returning it to us as soon as possible. I look forward to the publication of this manuscript. I believe it will make an important contribution to the literature. C ongratulations on comple ting this excellent research. Thank you for submitting your interesting work to JG: PS. Sincerely yours, Thomas M. Hess, Ph.D. Editor Journal of Gerontology: Psychological Sciences Department of Psychology North Carolina State University Box 7801 Raleigh, NC 27695-7801 (919) 515-1729 (phone) (919) 515-1716 (fax) firstname.lastname@example.org
99 Appendix B: Curriculum Vitae Tiffany F. Hughes University of South Florida School of Aging Studies 4202 E. Fowler Avenue MHC 1352 Tampa, FL 33620-8100 Telephone: (813) 974-3237 Fax: (813) 974-9754 Email: email@example.com Education University of South Florida, Tampa, FL Ph.D. in Aging Studies, 2008 Dissertation: The role of lifestyle f actors in cognitive aging and dementia. Advisors: Dr. Brent J. Small, Dr. Ross Andel University of South Florida M.P.H. Epidemiology, 2008 Advisor: Dr. James A. Mortimer Allegheny College, Meadville, PA B.S. Neuroscience and Psychology with honors, 2001 Professional Experience Graduate Research Assistant University of South Florida, Tampa, FL, 2005-2008 Supervisor: Dr. Brent J. Small Johnnie B. Byrd, Sr. Alzheimer's Cent er & Research Inst itute, Tampa, FL, 2005-2008 Supervisors: Dr. Huntington Potter, Dr. Brent J. Small Research Assistant University of South Florida, Depart ment of Biochemistry and Molecular Biology, Tampa, FL, 2002-2005 Supervisor: Dr. Huntington Potter
100 University of Pittsburgh, Department of Neurology, Pittsburgh, PA, 2001-2002 Supervisor: Dr. Paula Clemens Internship The Memory Disorder Clinic of Saraso ta Geriatrics at Sarasota Memorial Hospital, Sarasota, FL, Summer 2005 Supervisors: Kathleen Housewart, Dr. Bruce Robinson Mayo AlzheimerÂ’s Disease Research Ce nter, Mayo Clinic, Jacksonville, FL, Summer 2000 Supervisors: Dr. Michelle Nicole, Dr. Michael McKinney Teaching Experience Instructor University of South Florida, School of Aging Studies, Fall 2007 and Spring 2008 Course Title: Physic al Changes and Aging University of South Florida, School of Aging Studies and Department of Psychology, Summer 2006 and Summer 2007 Course Title: Psychology of Aging Teaching Assistant University of South Florida, School of Aging Studies, Summer 2008 Course Title: Health Promotion and Aging University of South Florida, School of Aging Studies, Spring 2007 Course Title: Intr oduction to Gerontology University of South Florida, School of Aging Studies and Department of Psychology, Fall 2006 Course Title: Psychology of Aging Allegheny College, Fall 2000 Course Title: Hea lth and Psychophysiology Allegheny College, Fall 1999 Course Title: Physiological Psychology Publications Hughes, T.F ., Andel, R., Small, B.J., Borenstein, A. R., & Mortimer, J.A. (in press). The association between social resources and cognitive change in older adults: Evidence from the Charlotte County Healthy Aging Study. Journal of Gerontology: Psychological Sciences.
101Bielak, A.A.M., Hughes, T.F ., Small, B.J., & Dixon, R.A. (2007) It's never too late to engage in lifestyle activities: Significan t concurrent but not change relationships between lifestyle activities and both mean level and intraindividual variability in cognitive speed. Journal of Gerontology: Psychological Sciences, 62, P331P339. Small, B.J., Hughes, T.F. Hultsch, D.F., & Dixon, R.A. (2007). Lifestyle activities and late-life changes in cognitive performance. In Y. Stern (ed.), Cognitive Reserve: Theory and Applications (pp.173-186). New York: Taylor & Francis. Costa, D.A., Cracchiolo, J.R., Bachstetter, A.D., Hughes, T.F. Bales, K.R., Paul, S.M., et al. (2007). Enrichment im proves cognition in AD mice by amyloidrelated and unrelated mechanisms. Neurobiology of Aging, 28(6), 831-844. Andel, R., Hughes, T.F. & Crowe, M.G. (2005). Stra tegies to reduce the risk of cognitive decline and dementia. Aging Health, 1(1), 107-116 Bilboa, R., Srinivasan, S ., Raey, D., Goldberg, L., Hughes, T ., Roelvink, P. W., et al. (2003). Binding of adenoviral knob to the coxsackievirus-adenovirus receptor is crucial for transduction of fetal muscle. Human Gene Therapy, 14(17), 645-649. Bilboa, R., Raey, D., Hughes, T ., Biermann, V., Volpers, C., Goldberg, L., et al. (2003). Fetal muscle gene tran sfer is not enhanced by an RGD capsid modification to high-capacity adenoviral vectors. Gene Therapy, 10(21), 18211829. Manuscripts in Progress Hughes, T.F. Andel, R., Small, B.J., Borenstein A.R. Mortimer, J.A., Wolk, A., Johansson, B., Fratiglioni, L., Pederse n, N.L., & Gatz, G. (submitted). Midlife Fruit and Vegetable Consumption and Risk of Dementia in Later Life in Swedish Twins. Hughes, T.F. Bielak, A.A.M., Small, B.J., & Dixon, R.A. Does Engagement in Lifestyle Activities Affect Inconsiste ncy in Cognitive Speed Performance in Older Adults? Presentations Hughes, T.F. Andel, R., Borenstein, A.R., Mor timer, J.A., Wolk, A., & Gatz, M. (November, 2007). Mid-life fruit and vegeta ble consumption and risk for dementia in Swedish twins. Poster presentation at the 60th Annual Scientific Meeting of the Gerontological Society of America, San Francisco, CA.
102Small, B. J., Jacobsen, P. B., Andrykowski, M. A., Hughes, T. F. Sharp Rawson, K., & Iser, L. (November, 2007). Genetic varia tion and cognitive performance in breast cancer survivors and non-cance r controls. Paper presente d at the National Cancer Institute Small Grants Meeting, Rockville, MD. Hughes, T.F., Small, B.J., Potter, H., Borenstein, A.R., & Mortimer, J.A. (November, 2006). Pupillary response to tropicamid e is related to cognitive performance in a non-demented sample of older a dults. Poster presentation at the 59th Annual Scientific Meeting of the Gerontolog ical Society of America, Dallas, TX. Hughes, T.F., Andel, R., Small, B.J., Borenste in, A.R., & Mortimer, J.A. (April, 2006). Social resources, health and cognitive performance in a populationbased sample of older adults: The Charlotte County Healthy Aging Study. Poster presentation at The Cognitive Aging Conference, Atlanta, GA. Hughes, T.F., Andel, R., Small, B.J., Borenste in, A.R., & Mortimer, J.A. (April, 2006). Social resources, health and cognitive performance in a populationbased sample of older adults: The Charlotte County Healthy Aging Study. Poster presentation at Epidemiology of Alzh eimerÂ’s Disease Scientific Research Conference, San Diego, CA. Hughes, T.F. Bielak, A.A.M., Small, B.J., & Di xon, R.A. (November, 2005). Effects of lifestyle activities and cognitive reserv e on level and variabil ity of cognitive speed. Poster presentation at the 58th Annual Scientific Meeting of the Gerontological Society of America, Orlando, FL. Bielak, A.A.M., Hughes, T.F ., Small, B.J., & Dixon, R.A. (November, 2005). Does an engaged lifestyle predict cognitiv e level and inconsistency 3 and 6 years later? Poster presentation at the 58th Annual Scientific Meeting of the Gerontological Society of America, Orlando, FL. Costa, D.A., Cracchiolo, J.R., Bachstetter, A.D., Hughes, T.F ., Bales, K.R., Paul, S.M., Mervis, R.F., Arendas h, G.W., Potter, H. (November, 2005). Enriched housing improves cognitio n in AD mice by amyloid-related and unrelated mechanisms. Poster presentation at the 35th Annual Meeting of the Society for Neuroscience, Washington, D.C. Hughes, T.F ., Bielak, A., Small, B.J., & Dixon R.A. (April, 2005). Lifestyle activities, cognitive reserve, and intraindividual variability in cognitive functioning. The USF College of Arts and Scienc es Session of the Fourth Annual Graduate Student Research Symposium Tampa, FL. Honors and Professional Activities Member, Gerontological Soci ety of America, 2005-present
103University of South Florida Student Representative to th e Gerontological Society of America, 2006-2007 President, Student Association for Aging Studies (SAAS), University of South Florida, 2006-2007 Admission to Doctoral Candidacy with Distinction, Spring 2006 Travel grant recipient, USF College of Arts and Sciences Social Sciences Session of the Fourth Annual Graduate Student Research Symposium, April 2005 School of Aging Studies Fellowship, Un iversity of South Florida, 2004-2005 Student Member, Psi Chi, 1999-2001, Treasurer, 2000-2001
104 ABOUT THE AUTHOR Tiffany F. Hughes received her BachelorÂ’s of Science degree in Neuroscience and Psychology from Allegheny College, Meadville PA in May of 2001. She entered the Ph.D. in Aging Studies program and the Univ ersity of South Flor ida in the Spring of 2004 with an interest in lifes tyle factors that influence cognitive aging and dementia. While in the Ph.D. program, Ms. Hughes was employed as a Graduate Research Assistant in the School of Aging Studies and at the Johnnie B. Byrd AlzheimerÂ’s Center and Research Institute where she assisted in ongoing research projec ts. In addition, she also was employed as a Graduate Teaching Assistant in the School of Aging Studies where she was primarily responsible for t eaching the undergraduate Psychology of Aging and Physical Changes with Aging courses. Ms. Hughes first authored one peer-reviewed publication and co-authored three additional publications while enrolled as a student, as well as presented her research at multiple national conferences.