Eastern Michigan UniversityDigitalCommons@EMU

Master's Theses and Doctoral Dissertations Master's Theses, and Doctoral Dissertations, andGraduate Capstone Projects

2017

Contributors to processing speed deficit in old age:A focus on history of obesity and medicalconditionsMelissa E. Pulcini

Follow this and additional works at: http://commons.emich.edu/theses

Part of the Psychiatry and Psychology Commons

This Open Access Dissertation is brought to you for free and open access by the Master's Theses, and Doctoral Dissertations, and Graduate CapstoneProjects at DigitalCommons@EMU. It has been accepted for inclusion in Master's Theses and Doctoral Dissertations by an authorized administrator ofDigitalCommons@EMU. For more information, please contact lib-ir@emich.edu.

Recommended CitationPulcini, Melissa E., "Contributors to processing speed deficit in old age: A focus on history of obesity and medical conditions" (2017).Master's Theses and Doctoral Dissertations. 762.http://commons.emich.edu/theses/762

Contributors to processing speed deficit in old age:

A focus on history of obesity and medical conditions

by

Melissa E. Pulcini

Submitted to the Department of Psychology

Eastern Michigan University

in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

in

Clinical Psychology

Dissertation Committee:

Karen Saules, Ph.D., Chair

Renee Lajiness-O’Neill, Ph.D.

Jin Bo, Ph.D.

Judi Brooks, Ph.D.

August 24, 2017

Ypsilanti, Michigan

ii

Acknowledgments

This work would not have been possible without support from the graduate school at

Eastern Michigan University, its department of psychology, and the members of my

dissertation committee. I am especially grateful to my advisor, Dr. Karen Saules, for her

patience, guidance, and support in both personal and professional matters throughout the past

seven years.

I am also indebted to the clinical supervisors who nurtured and encouraged my

interests relevant to this project. They have made indirect yet significant contributions to my

work. In particular, I would like to thank Dr. James Mikesell for his incredible generosity in

creating an externship for me at The Centers at St. Camillus and sharing his time, expertise,

and support.

Finally, I would like to thank my family. My work likely would not have been

possible without the support and efforts of my parents and brother to help provide me with

opportunities to pursue my goals. I am also appreciative of the invaluable support and

friendship from members of my graduate cohort. Lastly, I would like to acknowledge my

little girl, Snow White, in the highest degree of love and gratitude. She was, to say the least,

my constant companion, greatest joy, and wisest teacher.

Abstract

Obesity has been associated with impairment in most cognitive domains, but the nature of this

relationship is unclear. Most studies have examined the relationship between cognitive

impairment and current obesity and/or history of obesity-related medical conditions. However, a

number of “lower-level” obesity factors (e.g., increased adipokines, low grade inflammation,

chronic mild hypoventilation) are present even among the otherwise healthy obese population,

and it is possible that many of these factors have differential impact on cognition according to

magnitude and duration of exposure. Knowledge of a person’s weight history, which may serve

as a proxy for a number of lower-level obesity variables that cannot otherwise be measured, may

provide additional utility in predicting cognitive functioning. Of the cognitive domains,

processing speed is particularly important to study in this context, as its decline plays a large role

in driving cognitive aging. Despite this, little is known about the impact of weight history on

processing speed in old age. The present study added to the scientific body of literature by using

data from the 1999–2000 and 2001–2002 NHANES cycles to examine the relationship between

weight history and current processing speed among young-old adults (i.e., adults aged 65–74)

using data at three timepoints: young adulthood, late mid-life, and young-old adulthood. A

statistically significant relationship between persistent obesity in young adulthood and late

middle-age and processing speed in young-old adulthood was identified, but it was found to be

fully mediated by vascular health history. Additionally, slow processing speed was associated

with the weight trajectory pattern of obesity in late middle-age coupled with the absence of

obesity in both young adulthood and young-old adulthood. In regard to study implications, it

may be possible that efforts toward prevention of new onset obesity during mid-life, if

successful, would protect against slowed processing speed in young-old adulthood.

iii

iv

Table of Contents

Acknowledgments........................................................................................................................... ii

Abstract .......................................................................................................................................... iii

List of Tables ................................................................................................................................. vi

List of Figures ............................................................................................................................... vii

Introduction ..................................................................................................................................... 1

Research Hypotheses .................................................................................................................... 70

Method .......................................................................................................................................... 80

Results ..........................................................................................................................................104

Discussion ....................................................................................................................................158

References ................................................................................................................................... 197

Appendices

A: Demographics Questionnaire .................................................................................... 247

B: Occupation Questionnaire ......................................................................................... 249

C: Weight History Questionnaire................................................................................... 250

D: Blood Pressure Questionnaire ................................................................................... 253

E: Diabetes Questionnaire.............................................................................................. 254

F: Medical Conditions Questionnaire Variables ............................................................ 255

G: Kidney Conditions Questionnaire ............................................................................. 258

H: Physical Activity Questionnaire ............................................................................... 259

I: Physical Activity Individual Activities File .............................................................. 261

J: Diet Behavior and Nutrition ...................................................................................... 264

K: Smoking Questionnaire ............................................................................................. 265

v

L: Alcohol Use History Questionnaire ......................................................................... 267

M: IRB Exempt Letter ....................................................................................................268

N: Consent Form ............................................................................................................269

vi

List of Tables

Tables

1: Hypotheses 8a–8c groups ............................................................................................75

2: Participant characteristics ............................................................................................81

3: Summary of correlations between DSST score and obesity indexes.........................122

4: Summary of correlations between obesity indexes ....................................................125

5: Summary of multiple regression analyses for obesity index variables predicting DSST

score ...........................................................................................................................127

6: Means and standard deviations of DSST score by weight trajectory group ..............135

7: Results of t-tests comparing groups on DSST score by self-reported history of medical

condition ....................................................................................................................141

8: Confirmatory factor analysis for vascular health history...........................................150

9: Factor loading summary of revised hypothesized structural model of DSST score ..153

10: Factor loading summary of final structural model of DSST score ............................157

vii

List of Figures

Figures

1: Theoretical model of relationship between persistent obesity and DSST score ..........79

2: DSST scores by folate depletion and hyperhomocysteinemia status ........................113

3: Interaction effects between gender, current vitamin D status, and obesity status on

DSST score for current obesity status and obesity status ten years ago ....................116

4a: DSST scores by current BMI category and gender ....................................................129

4b: DSST scores by late middle-age BMI category and gender ......................................131

4c: DSST scores by young adult BMI category and gender ............................................133

5: Hypothesized measurement model of vascular health history ...................................147

6: Revised hypothesized structural model of DSST score .............................................148

7: Revised hypothesized structural model of DSST score with standardized regression

weights .......................................................................................................................152

8: Final structural model of DSST score........................................................................155

9: Final structural model of DSST score with standardized regression weights ...........156

Introduction

Statement of the Problem

The prevalence of obesity, a leading cause of preventable death, has nearly tripled among

adults in the United States since the late 1970s, and it was estimated that 35.0% and 40.4% of

American men and women were obese in 2016, respectively (Sassi, Devaux, Cecchini, &

Rusticelli, 2009; Zylke & Bauchner, 2016). At the same time, the percentage of the United

States population who are elderly (i.e., 65+ years old) has grown from approximately 9.9% in

1970 to an estimated 14.9% in 2015, with projections to reach 22.1% by 2050 (He, Goodkind, &

Kowal, 2016; United States Census Bureau, 2017; United States Department of Health & Human

Services, 2014a). Cognitive decline is normative with increasing age, but there is growing

evidence that obesity (or obesity-related factors) accelerate this decline (Harada, Natelson Love,

& Triebel, 2013; Sabia, Kivimaki, Shipley, Marmot, & Singh-Manoux, 2009; Xu et al., 2011). It

follows that America will most likely see an increased number of obese elderly, elderly with a

history of obesity, and elderly with greater cognitive decline than what is typically found in

normal aging in the coming decades. This has highlighted the importance of understanding the

unique cognitive challenges elderly individuals might face as a consequence of obesity or obesity

history. It also underscores the importance of understanding the unique value of each obesity

index in predicting cognitive functioning in the elderly; this is an especially salient topic given

that the most widely used obesity index, body mass index (BMI), underestimates adiposity in the

elderly (Heo, Faith, Pietrobelli, & Heymsfield, 2012). Furthermore, it also raises larger

questions of (a) whether current and/or history of obesity may be considered independent risks

for cognitive dysfunction, and (b) what processes might underlie the relationship between

obesity and cognitive dysfunction.

1

2

To briefly address the first question, obesity is associated with impairments in attention,

processing speed, memory, and executive functioning, although many of the relationships

between obesity and cognitive domains appear to vary according to age and sex, and there is

conflicting evidence regarding whether these associations are independent or can be sufficiently

explained by mediating variables (e.g., hypertension, diabetes). Recently, there has been more

focus on lower-level obesity-related factors that may impact cognition (e.g., low grade

inflammation, increased circulation of adipocyte hormones, endothelial dysfunction). For

example, one recent population-based study of the elderly found that circulating biomarkers of

two phenomenon typically seen in obesity (namely, low-grade inflammation and endothelial

dysfunction) were associated with impairments in processing speed and executive functioning,

and controlling for vascular risk factors only slightly attenuated their relationships (Heringa et

al., 2014). It may be that these obesity-related factors that contribute to cognitive decline have

an increased effect if the individual is exposed to a greater magnitude (e.g., more hormones

released from adipose tissue) and/or over a greater time period (e.g., chronic low-grade

inflammation in the chronically obese person, chronic endothelial dysfunction, chronic mild

cerebral hypoxia due to chronic mild hypoventilation).

Therefore, rather than studying the relationship between current weight status and

cognition, it may be more informative to study long-term weight history and cognition. Indeed,

there does appear to be a relationship between weight history and cognitive functioning. For

example, obesity during midlife appears to be an independent risk factor for vascular dementia

(VaD) and Alzheimer disease (AD) onset in later-life (Xu et al., 2011). Individuals who go on to

develop dementia are more likely to have lost a significant amount of weight between midlife

and late-life compared with individuals who remain free of dementia (Stewart et al., 2005).

3

Therefore, the combination of obesity in midlife and weight loss between midlife and late-life,

the latter of which is likely to be unintentional, may indicate especially elevated odds of

developing dementia.

In addition to those with dementia, many have asked whether body weight history is also

related to cognitive functioning both among those without dementia and among the general

population. Most of this work has focused on older adults, which is likely due to (a) the idea that

older adults will have a greater time period to have experienced impact of weight-related

processes on their cognitive functions, and (b) the simple fact that older adults are at greater risk

for cognitive decline than any other age group. As expected, preliminary evidence suggests that

body weight history is also related to cognitive functioning among those without dementia and in

the general population. For example, long-term underweight or obese status during adulthood

has been associated with impaired memory and executive functions in late midlife after

controlling for sex, age, and education (Sabia et al., 2009). Furthermore, the extent of weight

loss or weight gain in middle-aged and older men and women has been associated with poorer

performance on visual memory tasks and poorer overall cognitive performance (Brubacher,

Monsch, & Stähelin, 2004; Lo, Pachana, Byrne, Sachdev, & Woodman, 2012). However, this

area of study is in its infancy, and findings have not been adequately synthesized. There are a

myriad of potential mediating and moderating factors that have not yet been accounted for in

current research, and very little is known about the potential impact of weight history on

processing speed, a domain many consider to be foundational to other cognitive processes (e.g.,

Vance, 2009).

Processing speed is an especially important cognitive domain in the elderly for many

reasons, including the following: (a) In normal cognitive aging, processing speed (along with

4

spatial ability) declines more rapidly with age than other cognitive abilities (e.g., memory, verbal

abilities; Finkel, Reynolds, McArdle, & Pedersen, 2007). Many consider processing speed to be

a foundational resource for other cognitive processes, and, according to the processing speed

theory of aging, reduced processing speed in old age restricts other cognitive processes to the

extent that these cognitive processes rely on processing speed (Finkel et al., 2007; Salthouse,

1996; Vance, 2009). Therefore, processing speed may be a main driver of cognitive aging. (b)

Of the cognitive domains, only processing speed appears to be independently associated with

longer time to complete everyday tasks in the elderly (Owsley, Sloane, McGwin, & Ball, 2002).

(c) Processing speed and spatial abilities, but not memory or other cognitive domains, appear to

predict greater life satisfaction prospectively in the very old (Enkvist, Ekström, & Elmståhl,

2013). (d) Greater processing speed has been linked to longer survival time in the elderly,

independent of other cognitive impairments and medical conditions (Iwasa et al., 2014). These

reasons highlight the need to understand contributors to processing speed in the elderly.

Study Aims

The present study added to the body of scientific literature in this area by examining the

relationship between weight history (beginning in early adulthood) and current processing speed

among men and women aged 65–74, a population typically referred to as the young-old (e.g.,

Chen, Tsai, Lee, & Lee, 2014; Zizza, Ellison, & Wernette, 2009). The period of early adulthood

to young-old adulthood greatly extends the length of weight history in any published study on

the relationship between weight history and processing speed. Furthermore, because the present

study used data from the 1999–2000 and 2001–2002 cycles of the National Health and Nutrition

Examination Survey (NHANES), it was in a unique position to examine the impact of a great

number of variables (e.g., current and past medical conditions, blood work, physical activity,

5

blood pressure, diet, nutrition) that may impact the relationship between weight history and

processing speed in a large sample of elderly individuals. While it was hypothesized that weight

trajectory would play a role in processing speed in old age, other factors were expected to be

very important, and NHANES data provided bountiful opportunity to explore them. The present

study made several hypotheses regarding the relationships among NHANES variables and

culminated with creation of a model to explain poor processing speed among those with long-

term obesity.

It should be noted here that the NHANES is an ongoing study designed to assess the

health and nutrition among a sample representative of the United States population, and it

involves a number of components, including interviews, laboratory work, and physical

examinations. NHANES and other research projects were created by the National Center for

Health Statistics (NCHS) in order to collect and provide information to the Centers for Disease

Control and Prevention (CDC) for the purposes of guiding public health action and policy

(NHANES Plan and Operations, 1999–2010). The 1999–2000 and 2001–2002 NHANES cycles

were chosen for the present study due to their inclusion of a processing speed measure, whereas

subsequently completed cycles have not.

A secondary aim of the present study was to examine the relationship between various

obesity indexes (e.g., body mass index [BMI], percent of total body mass that is comprised of fat

as measured by dual energy x-ray absorptiometry [% total fat DEXA], percent of trunk mass that

is comprised of fat as measured by DEXA [% trunk fat DEXA], waist circumference [WC],

waist-thigh ratio [WTR], waist-height ratio [WHtR]) and processing speed, and determine

whether there is an interactive effect of obesity indexes on processing speed. Most studies on the

relationship between weight and cognition have relied on BMI to define obesity; however, there

6

is reason to believe that regional body composition is also important in impacting cognition. In

fact, a community study of the elderly found BMI and waist circumference to interact such that

high BMI only predicted poor performance on a brief screener for cognitive impairment when

abdominal obesity was present (Jeong, Nam, Son, Son, & Cho, 2005). Unfortunately, little more

is known due to the paucity of such studies. The present study sought to determine the

relationship between processing speed and multiple obesity indexes, as well as determine the

best combination of indexes in predicting processing speed.

Overweight and Obesity Trends and Prevalence

The prevalence of overweight and obesity in adults aged 20 years or older in 2011–2012

in the United States was estimated with NHANES data using standard World Health

Organization (WHO) BMI cut-offs to be 68.5% and 34.9%, respectively (Ogden, Carroll, Kit, &

Flegal, 2014). This obesity prevalence represents almost a three-fold increase compared with

12.6% in 1978 (Sassi et al., 2009). A number of other countries, such as Australia, Canada, and

England, are nearly as affected by the epidemic, as they have seen their obesity rates more than

double in recent decades (Sassi et al., 2009). Although the U.S. may now finally be nearing a

stabilization of obesity rates, those countries with historically low rates of obesity (e.g., South

Korea, France) are expected to show significant increases in the coming years (Flegal, Carroll,

Ogden, & Curtin, 2010; Sassi et al., 2009).

Rates of estimated overweight and obesity vary by age, sex, and race/ethnicity. In the

United States, men have higher estimated rates of overweight, but women have higher estimated

rates of obesity (Ogden et al., 2014). Overweight rates are also elevated in adults aged 40–59

years (75.3%) and 60+ years (71.6%) compared with adults aged 20–39 years (60.3%; Ogden et

al., 2014). Furthermore, Hispanic women and non-Hispanic Black women are estimated to have

7

higher rates of overweight at each of the above age ranges compared to their non-Hispanic Asian

and non-Hispanic White counterparts (Ogden et al., 2014).

Elderly Trends and Prevalence

Similar to obesity, the prevalence of the elderly (adults aged 65+) in the United States has

climbed significantly throughout the past century. For example, approximately 8.1%, 9.9%,

12.6%, and 13.0% of the population was elderly in 1950, 1970, 1990, and 2010, respectively

(United States Department of Health & Human Services, 2014a). The latest available data

indicates that approximately 14.9% of the US population was elderly in 2015 (United States

Census Bureau, 2017). Elderly prevalence is expected to climb dramatically in the upcoming

decades, partly due to the aging of the baby boomer generation; it has been projected that 22.1%

of the U.S. population will be elderly by 2050 (He et al., 2016). Taken together with the

aforementioned obesity prevalence data, these data suggest that there will be dramatic increase in

the number of obese elderly individuals and elderly individuals with a significant history of

obesity in the coming decades in the United States.

Assessing Obesity in the Elderly

To examine the impact of obesity and history of obesity on the elderly, it is first

necessary to examine the utility of obesity assessment methods. In doing so, it is also useful to

understand how body composition and these indexes typically change with age.

Body mass index. The most commonly used method of assessing obesity among adults

is the body mass index (BMI), which is calculated by dividing weight in kilograms (kg) by

height in meters squared (m²). The WHO uses BMI as its standard unit for obesity statistics for

adults and recommends the following set of international cut-offs for researchers: underweight <

18.50 kg/m²; normal 18.50–24.99 kg/m²; overweight ≥ 25.00 kg/m²; obese ≥ 30.00 kg/m²; obese

8

class I 30.00–34.99 kg/m²; obese class II 35.00–39.99 kg/m²; and obese class III ≥ 40.00 kg/m²

(WHO, 2010).

Although BMI is the most widely used method for estimating adiposity, there are several

known problems with its use. BMI, which relies solely on height and weight, cannot distinguish

fat from bone, muscle, and other lean body mass. Given this deficit, it is not surprising that BMI

tends to overestimate body fat percentage in those with muscular builds (e.g., athletes) and to

underestimate body fat percentage in those who have lost muscle or bone density (e.g., the

elderly; NIH, 2010). Indeed, researchers routinely consider types of body fat estimation methods

that do not rely solely on anthropometric measurements (e.g., hydrostatic weighing, dual energy

X-ray absorptiometry [DEXA]) to be more accurate than BMI (e.g., Duncan, Duncan, &

Schofield, 2009; Kupusinac, Stokić, & Doroslovački, 2014). Furthermore, percentage of body

fat, when ascertained by these more accurate methods, in individuals with the same BMI value

has been shown to vary significantly across age, ethnic, and racial populations (Deurenberg,

Yap, & van Staveren, 1998). For example, the elderly have significantly more body fat than

young and middle-aged adults at a given BMI (Heo et al., 2012). Moreover, Asians have

significantly more body fat than Whites at a given BMI, and White and Hispanic women have

significantly more body fat than Black women at a given BMI (Deurenberg, Yap, & Guricci,

2002; Rahman & Berenson, 2010). These findings suggest the potential for underestimation of

body fat in Asians and the elderly as well as overestimation of body fat in Black individuals.

Another weakness of BMI is its inability to provide data on body fat distribution or body shape,

factors which carry additional information regarding various health risks (Després, 2012). As a

result, many researchers continue to advocate for the use of alternative measures of body fat in

9

research and/or use of BMI cut-offs specific to the population served (Burkhauser & Cawley,

2008; Kagawa, Uenishi, Kuroiwa, Mori, & Binns, 2006).

Central obesity indexes. Many other obesity indexes have focused on abdominal

adiposity (e.g., waist circumference [WC], waist-to-hip ratio [WHR], waist-thigh ratio [WTR],

waist-height ratio [WHtR], % trunk fat DEXA). Interest in these central obesity indexes are

related to findings that they predict increased risk for certain obesity-related diseases beyond that

conferred by BMI classification and, in fact, are even better predictors of hypertension,

dyslipidemia, metabolic syndrome, and type-2 diabetes mellitus (T2DM) than BMI (Guasch-

Ferré et al., 2012; Janssen, Katzmarzyk, & Ross, 2002, 2004; Radzevičienė & Ostrauskas, 2013;

van Dis, Kromhout, Geleijnse, Boer, & Verschuren, 2009; Yusuf et al., 2004). Reasons for

advantages of these central obesity indexes have been theorized to include differences in

metabolic activity of visceral and subcutaneous adipose tissues, shared genetic influences

between waist circumference and T2DM, and findings that larger hip and thigh circumferences

are associated with decreased risk of T2DM (Conway et al., 2011; Mamtani et al., 2014; Power

& Schulkin, 2008; Snijder et al., 2003). It should be noted that, although these are called central

obesity indexes, WHR, WTR, and WHtR each reflect information about one’s body proportion

that goes beyond waist only.

Optimal cutoffs in predicting health risks using the aforementioned central obesity

indexes have been found to vary by sex, race/ethnicity, and age (Dobbelsteyn, Joffres, MacLean,

& Flowerdew, 2001; Lear, James, Ko, & Kumanyika, 2010). The most common cutoffs for

central obesity in men include WC of ≥ 102 centimeters (cm) and WHR of ≥ 0.95, whereas the

most common cutoffs for women include WC of ≥ 88cm and WHR of ≥ 0.8 (Lear et al., 2010).

However, these cutoffs are primarily based on studies of Caucasians of European origin, and

10

there is some evidence to suggest that optimal WC cutoffs for Asians are lower than those of

European Caucasians (Lear et al., 2010; Obesity in Asia Collaboration, 2008). There is less

conclusive evidence regarding specific central obesity cutoffs for other groups, including

African-American, Hispanic, and Middle Eastern populations (Lear et al., 2010). In regard to

age, higher cutoff points for WC and WHR appear to be more appropriate for older aged

populations up to age 65, at which point the appropriate cutoffs appear to lower slightly

(Dobbelsteyn et al., 2001). One study found optimal WC cutoffs in individuals aged 60–74 to be

similar to the aforementioned commonly used cutoffs (Gouveia, Marucci, Lebrão, & Duarte,

2014). More research is needed to better understand appropriate central obesity cutoffs for

predicting health risks in all age ranges within the elderly populations, particularly in those

above 75 years of age.

There is no clear consensus regarding which of the central obesity indexes is most useful

in predicting medical outcomes, although one recent systematic review and meta-analysis

concluded that WHtR is better in predicting hypertension, T2DM, dyslipidemia, and metabolic

syndrome among adults than WC or BMI (Ashwell, Gunn, & Gibson, 2012). Among the elderly,

WHtR and WC appear to predict metabolic diseases about equally as well, both surpassing BMI

(Guasch-Ferre et al., 2012). WC, but not WHR, is associated with hypertension among the

elderly, and WC has been found to predict mortality better than WHR and BMI in this

population (de Hollander et al., 2012; Visscher et al., 2001; Woo, Ho, Yu, & Sham, 2002).

Relatively less has been published about the value of WTR or % trunk fat DEXA as central

obesity indexes, but preliminary data suggest that WTR may be a better predictor of peripheral

vascular disease than WC and a better predictor of T2DM in men than WC, WHR, and WHtR

(Li, Ford, Zhao, Kahn, & Mokdad, 2010; Lu, Zhou, Waring, Parker, & Eaton, 2010). Taken

11

together, these results suggest that WTR, WHtR, and WC have great potential as health

screening tools. More system comparisons (especially those involving WHtR and WTR) are

needed to determine the best index to use to screen for each medical condition.

Other obesity indexes. Methods typically considered to be more accurate in estimating

adiposity than BMI and the aforementioned central obesity indexes include DEXA,

computerized tomography (CT), magnetic resonance imaging (MRI), bioelectrical impedance

analysis (BIA), hydrostatic weighing, total body electrical conductivity (TOBEC), and skinfold

thickness measurements (Moyad, 2001). These methods tend to be costly, not widely available,

and/or necessitate administration by trained technicians, and thus are often not practical to

implement in large studies. The NHANES, however, is an ongoing large epidemiological study

that does obtain body composition data using one of these methods, namely DEXA. This

method uses two x-ray beams of differing wavelengths to analyze body composition.

Information on bone mineral content, bone mineral density, and fat-free mass can be yielded

with this technique, and estimates of percent body fat can be calculated indirectly from this

information (Clasey et al., 1997). Standard DEXA cut-offs for obesity are fat mass ≥ 25% for

men and ≥ 35% for women (Donini et al., 2013). Advantages of DEXA include ability to

differentiate between fat and lean mass, validity in determining percent body fat within obese

and elderly populations, ability to yield regional body composition measurements (e.g., arm,

trunk), and good reproducibility in estimates of abdominal fat and lean mass (He, Li, & Kung,

1999; Salamone et al., 2000; Tallroth, Kettunen, & Kujala, 2013). Disadvantages include

measurement error introduced by variations in hydration in fat-free mass across individuals and

lack of consistency of DEXA scan mode used in research (Clasey et al., 1997; Salamone et al.,

2000).

12

Changes in body composition and obesity indexes with age. Aging is associated with

significant changes in body composition, including those involving total adiposity, distribution

of adiposity, and fat-free mass, the latter of which includes muscle mass, bone mass, and other

nonfat parts of the body (Obisesan et al., 2005; Oldroyd, Stewart, Truscott, Westmacott, &

Smith, 1998). Total fat mass and percentage of body fat generally increase with age (Jackson,

Janssen, Sui, Church, & Blair, 2012; Kyle, Genton, Slosman, & Pichard, 2001). Some data

suggest that total fat mass and body fat percentage stabilizes after about age 80, but other

research suggests that they continue to increase well into the 90s in healthy individuals (Jackson

et al., 2012; Kyle et al., 2001). Subcutaneous fat (e.g., on arms and legs) decreases begin in late

midlife and continues until at least one’s 80s, whereas fat in the viscera and other areas (e.g.,

organs, muscles) increase during midlife (Oldroyd et al., 1998; WHO Expert Consultation,

2008). Declines in fat-free mass (FFM) and fat-free mass index (FFMI) typically begin during

one’s 50s and become significant by ages 65–69 in Black men, 70–74 in Black women and

White men, and 75–79 in White women (Obisesan et al., 2005). Bone loss begins even sooner at

approximately 30 years of age and is accelerated in the early post-menopausal period in women

(Ahlborg et al., 2001; Nuti, Martini, & Gennari, 1995). Furthermore, significantly more lean

mass is lost during weight loss in the elderly than is gained during weight gain in this population

(Newman et al., 2005). These body composition changes explain why the elderly are prone to

both sarcopenia, an age-related loss of muscle tissue, and sarcopenic obesity, the combination of

sarcopenia and high fat mass (Evans & Campbell, 1993).

Obesity indexes also tend to vary with age. The obesity indexes of BMI, WHR, and WC

increase up to age 64; in the decade thereafter, BMI declines in both sexes, WC increases slightly

for men but declines slightly among women, and WHR shows no significant change

13

(Dobbelsteyn et al., 2001). Currently, there appear to be no published studies specifically on

changes in WHtR or WTR throughout adulthood into old age. It should be cautioned that

changes in obesity indexes are not necessarily consistent with body composition changes in the

elderly. For example, BMI typically declines in old age, but total fat mass does not appear to do

so. Likewise, WC decreases slightly among women after age 64, but visceral fat mass is actually

thought to increase in this population. For these reasons, care should be taken when selecting

obesity indexes to assess adiposity or related health risks in the elderly. In regard to the present

study, it was hypothesized that age would be positively associated with percent body fat

estimated by DEXA but would show a slight negative correlation with BMI. It was also

hypothesized that WC would show little change with age, but visceral adiposity estimated by %

trunk fat DEXA would increase with age.

Obesity Comorbidity/Mortality

Obesity is associated with reduced quality of life, financial burden, and increased risk for

many medical and psychological disorders, certain types of cognitive decline, and mortality. The

present section summarizes the relationships between obesity and these comorbidities, focusing

on relationships within the elderly population when data are available.

Medical. Overweight/obesity has been named one of the five leading global risks for

mortality by the WHO and is associated with three of the other four identified risks:

hypertension, high blood glucose levels, and physical inactivity (WHO, 2009). Also a risk factor

for hypercholesterolemia and raised triglyceride levels, obesity is associated with increased

prevalence of cardiovascular diseases (CVDs), the worldwide leading cause of death (Mendis,

2010; WHO, 2003). Obesity is also a primary risk factor for T2DM, as excess adipose tissue can

have a contributory role in increased insulin resistance, resulting in increased blood glucose

14

levels (Amini et al., 1997). Other potentially life-threatening events or conditions with well-

established links to obesity include several types of cancers, chronic kidney disease, obstructive

sleep apnea, gallstones, asthma, micronutrient deficiencies, and complications following certain

types of surgical procedures (Arif, Rohrer, & Delclos, 2005; Dowsey, Liew, Stoney, & Choong,

2010; Must et al., 1999; Whaley-Connell, Pavey, Afroze, & Bakris, 2006; Xanthakos, 2009).

Moreover, obesity is associated with many non-life threatening medical conditions, including

osteoarthritis, stress incontinence, gout, gastroesophageal reflux disease, and polycystic ovary

syndrome, all of which have been shown to negatively impact health-related quality of life

(Jeong, 2008; Roddy, Zhang, & Doherty, 2007; Tincello et al., 2010; Yildiz et al., 2010).

Given this, it is not surprising that obesity has been repeatedly linked to impaired health-

related quality of life and disability (e.g., Alvarez-Blasco, Luque-Ramírez & Escobar-Morreale,

2010; Corica et al., 2006; Strain et al., 2010). Those with class III obesity appear to suffer

greater health-related limitations in performing physical activities (e.g., walking one mile,

walking up steps) than those with class I–II obesity (Castres, Folope, Dechelotte, Tourny-

Chollet, & Lemaitre, 2010). Furthermore, there is a well-documented association between

obesity and chronic pain in individuals of diverse ages, ranging from children to the elderly

(McCarthy, Bigal, Katz, Derby, & Lipton, 2009; Wilson, Samuelson, & Palermo, 2010). Obesity

has also been found to have a negative mediating effect on pain outcomes of various procedures,

including knee arthroscopy, knee arthroplasty, and hip arthroplasty (Gandhi, Razak, Davey, &

Mahomed, 2010; Harrison, Morrell, & Hopman, 2004; Núñez et al., 2009).

Obesity in midlife has been linked to shorter life span, which is consistent with its

association with a number of life-threatening conditions (Peeters et al., 2003). In contrast,

obesity among the elderly, when measured by BMI, has been associated with survival (Auyeung

15

et al., 2010; Lee et al., 2014). The origins of these discrepant findings are unclear but are

theorized to be related to health risks and conditions in normal weight elderly, including illness,

smoking, and normal weight obesity, the latter of which is a condition marked by a normal BMI

but a high level of body fat (Richman & Stampfer, 2010). More specifically, a variety of clinical

or subclinical illnesses (or related medical treatments) that are more common in the aging

population, such as dementia and cancers, often directly or indirectly lead to weight loss (Stewart

et al., 2005; Tisdale, 2003; Wannamethee, Shaper, Whincup, & Walker, 2000). The extent of

weight loss is associated with worse prognosis in many of these illnesses, and weight loss

amongst the elderly is predictive of mortality (Bamia et al., 2010; Martin et al., 2013; Thivat et

al., 2010). Therefore, an elderly population sampled at a particular time point will include many

who have recently lost weight unintentionally and may have subclinical or clinical illness that

put them at greater risk for mortality. Furthermore, elderly individuals who smoke, a behavior

which carries a well-documented mortality risk, tend to have a lower BMI and a lower risk of

obesity than the elderly who do not smoke (Sulander, Rahkonen, Nissinen, & Uutela, 2007;

United States Department of Health & Human Services, 2014d; Wilson, Habib, & Philpot,

2002). Finally, the elderly demonstrate increased cases of normal weight obesity, which is

associated with cardiovascular mortality independent of BMI and central fat distribution in the

elderly (Batsis et al., 2013). It has been hypothesized that although normal BMI predicts greater

mortality in the elderly, the aforementioned factors associated with normal BMI in the elderly,

rather than the normal BMI itself, are likely to cause mortality (Richman & Stampfer, 2010). At

present, it continues to be debated as to whether there are any true benefits of obesity in old age,

or whether this idea is simply an artifact of bias in our research due to reverse causation (Flegal,

16

Graubard, Williamson, & Cooper, 2011; Richman & Stampfer, 2010). More research is needed

to better answer these questions.

Psychopathology. Obesity is associated with psychopathology, including depressive

disorders, social phobia, panic attacks, agoraphobia, binge eating disorder, attention deficit

hyperactivity disorder, and history of alcohol use disorders (Barry & Petry, 2009; Herpertz et al.,

2006; Mather, Cox, Enns, & Sareen, 2009; Pagoto et al., 2009). These associations vary

according to a host of factors, including age and gender. Generally, women and younger adults

evidence greater associations between BMI and psychiatric disorders than men and older adults,

respectively (Mather et al., 2009; McCrea, Berger, & King, 2012). Age and gender may also

interact. For example, one study found that the probability of a young woman having a common

psychiatric disorder (e.g., depressive disorders, anxiety disorders) increases between

approximately BMI 20–30 kg/m2 with a relative stabilization thereafter (McCrea et al., 2012). In

contrast, the association between BMI and common psychiatric disorders takes a “U shaped”

pattern in young men, such that both underweight and obesity increases the probability of a

common psychiatric disorder (McCrea et al., 2012). The relationship between BMI and

common psychiatric disorders attenuates significantly by age 60 in both men and women

(McCrea et al., 2012).

Financial burden. Economic costs of obesity have been conceptualized on the

following three basic levels: individual (e.g., lost opportunities), employer (e.g., lost

productivity, higher insurance premiums), and government (e.g., Medicare and Medicaid

spending; Runge, 2007). Obesity-related medical costs, which are shared by all three of these

levels, are estimated to be approximately $147 billion per year, accounting for 9.1% of the

nation’s total medical spending (Finkelstein, Trogdon, Cohen, & Dietz, 2009). Obese

17

individuals tend to use more health care resources, which may be due to increased prevalence of

conditions for which obesity is a risk factor, such as T2DM and asthma, as well as other factors

including reduced physical activity, lower quality of life, increased depressive symptoms, and

increased risks and costs of surgery (Ahn, Smith, Matthew Lee Dickerson, & Ory, 2012; Kirk,

Kuhle, Ohinmaa, Colman, & Veugelers, 2012; Rosemann, Grol, Herman, Wensing, & Szecsenyi,

2008; Trakas, Lawrence, & Shear, 1999). Little has been published on the financial impact of

obesity on the elderly, but one study used simultaneous equation system modeling to estimate

that individuals who are overweight or obese at age 65 have 6–17% greater health care

expenditures during the rest of their lifetimes than those who are normal weight at age 65 (Yang

& Hall, 2008).

Cognition. There is growing evidence that obesity is associated with impaired cognition

and may even have a contributory role in its development in some cases. Associations have been

found between obesity and impairments in attention, processing speed, memory, and executive

functions, as detailed below.

Attention. Attention is fundamental to virtually all cognitive tasks, and research suggests

obese adults have specific deficits in both automatic and controlled processing within this

domain. More specifically, obese adults have been found to have reduced capacity in perceiving

or attending to negative facial expressions compared to healthy-weight controls (Cserjési,

Vermeulen, Lénárd, & Luminet, 2011). Additionally, Gunstad et al. (2007) found significant

negative correlations between body mass index (BMI) and performances on tests requiring

attentional processing, including those of working memory, choice reaction, and visual memory

span. Higher BMI has also been associated with poorer selective attention and greater rate of

decline in selective attention abilities among middle-aged women; however, the same

18

relationship was not found among men (Cournot et al., 2006). Most recently, Stanek et al.

(2013) reported an independent association between elevated BMI and poor performance on a

composite of tasks designed to assess attention and processing speed among both men and

women.

Processing speed. Middle-aged obese adults demonstrate impaired performance on tests

of processing speed. In addition to the aforementioned findings by Stanek et al. (2013), another

recent study reported higher BMI to be independently associated with poorer performance on the

Digit Symbol Substitution Test (DSST), a traditional measure of processing speed, among

middle-aged adults without dementia after adjusting for a number or potentially confounding

variables (Cournot et al., 2006). Furthermore, Sanz and colleagues (2013) found higher BMI to

predict worse performance on the DSST after adjusting for age, sex, education, occupational

status, income, smoking, alcohol use, physical activity, psychotropic drugs, and metabolic

syndrome variables (i.e., hypertension, dyslipidemia, vascular disease, C-reactive protein) among

middle-aged individuals. In contrast to these findings among middle-aged adults, Kuo et al.

(2006) reported healthy elderly obese individuals to perform similarly to their normal-weight

counterparts on the DSST and to have better performance on the Useful Field of View (UFOV),

which relies more heavily on visuospatial speed of processing and less heavily on motor

demands, indicating that the relationship between obesity and reduced processing speed may

shift with age. In regard to the present study, which includes both healthy and unhealthy

volunteers, DSST scores were not expected to be significantly related to current BMI. The

extent to which impairment in fine motor speed, which has been found among obese adults, may

impact these findings is unclear, although one study did not find evidence for fine motor speed to

be a significant contributor to DSST score (Joy, Fein, & Kaplan, 2003; Stanek et al., 2013).

19

Graphomotor coordination, which could potentially contribute to DSST score, has not been

found to differ between normal-weight and obese children (Mond, Stich, Hay, Kraemer, &

Baune, 2007). It is unclear if graphomotor coordination or speed differs among the elderly

population according to weight category.

Memory. Results of several studies indicate the presence of auditory and visual memory

deficits in obese adults (Gunstad et al., 2007; Gunstad, Paul, Cohen, Tate, & Gordon, 2006;

Wolf, Beiser, Elias, Vasan, & Seshadri, 2007). There is, however, a lack of clarity regarding

whether these memory deficits are independently associated with obesity. For example, Gunstad

and colleagues (2006) reported an independent association between higher BMI and a variety of

memory deficits (as evidenced by poorer performance on immediate and delayed recall and

recognition tasks) among obese adults across the adult lifespan. However, Stanek et al. (2013)

failed to find an independent association between BMI and memory functioning in a similar

study that controlled for a slightly different set of potentially confounding variables.

It is not clear which variables, if any, significantly affect the relationship between BMI

and memory. This relationship has not been found to vary with age, but results of studies

investigating whether this relationship varies by sex have been inconsistent (Gunstad et al., 2006;

Smith, Hay, Campbell, & Trollor, 2011; Stanek et al., 2013). Two studies found higher BMI to

be associated with poorer immediate and delayed recall abilities among men without dementia,

independent of educational level, blood pressure, diabetes, and other potential confounds, but

only one of these studies found such a relationship among women (Gunstad et al., 2006; Smith,

Hay, Campbell, & Trollor, 2011; Stanek et al., 2013). Another study found memory deficits in

both obese men and women and reported that the relationship between obesity and memory

function did not vary significantly by sex (Gunstad et al., 2006).

20

Executive functioning. Executive functions are complex capabilities that can be used to

initiate, manage, and coordinate the use of other mental processes, and allow for complex goal-

oriented thought and behavior (Lichtenberger & Kaufman, 2009). Deficits in executive

functioning in obese adults have frequently been reported, with most work being published on

young and middle-aged adults (Boeka & Lokken, 2008; Fergenbaum et al., 2009; Gunstad et al.,

2007; Lokken, Boeka, Yellumahanthi, Wesley, & Clements, 2010; Smith et al., 2011). Among

two comprehensive studies that controlled for a host of potentially confounding variables across

the adult lifespan, one found an independent association between higher BMI and impaired

executive functioning, but the other failed to do so (Gunstad et al., 2007; Stanek et al., 2013).

Interestingly, the study that did not find an independent association between BMI and executive

dysfunction reported a significant age by BMI interaction, suggesting that obesity-related

executive dysfunction may increase with age (Stanek et al., 2013). The finding that long-term

obesity during adulthood is associated with greater executive dysfunction in later midlife is

consistent with this idea (Sabia et al., 2009). Regarding older adults, fewer studies have

examined the relationship between obesity and executive dysfunction, but the majority of these

studies support the existence of this relationship (Benito-León, Mitchell, Hernández-Gallego, &

Bermejo-Pareja, 2013; Gunstad et al., 2007; Kuo et al., 2006; Walther, Birdsill, Glisky, & Ryan,

2010).

Weight trajectory and risk of cognitive dysfunction and dementia. In addition to current

obesity (as detailed above), dramatic weight change and long-term obese status are also linked to

cognitive dysfunction. For example, one study found long-term obesity and the extent of weight

gain between early adulthood and late middle-age to be associated with executive dysfunction,

though this study did not assess processing speed (Swabia et al., 2009). Another study of elderly

21

subjects (mean age 69.4 years) found that the extent of weight gain or weight loss over the

previous 10-year period was associated with poorer performance on a neuropsychological battery

(Brubacher et al., 2004). Regarding the present study, it was hypothesized that similar

relationships would be found with processing speed. It should be noted here that although

weight changes during specific ages may be linked with cognitive impairment and mortality risk,

weight cycling itself has not been conclusively linked to either (Stevens et al., 2012). Stevens

and colleagues (2012) point out that most studies on the impact of weight cycling do not evaluate

whether weight loss was intentional, which may differentiate between weight loss due to illness

and weight loss due to active attempts to improve one’s health or appearance. Regarding the

present study, it was hypothesized that individuals endorsing significant past-year unintentional

weight loss, which may reflect an underlying disease process, would perform worse on the

processing speed measure than those who achieved significant past-year weight loss that was

intentional and those who were weight-stable in the past year.

The relationship between dementia and body weight is complex. There is some evidence

to suggest that obesity during midlife is an independent risk factor for VaD and AD onset in

later-life (Xu et al., 2011). However, individuals who go on to develop dementia are more likely

to have lost a significant amount of weight between midlife and late-life compared with

individuals who remain free of dementia; this weight loss is likely to be unintentional (Stewart et

al., 2005). Therefore, the combination of obesity in midlife and unintentional weight loss

between midlife and late-life may indicate especially elevated odds of developing dementia.

Obesity indexes and cognition. Little is known about which obesity indexes are most

closely associated with cognition in the elderly. Most studies examining the link between

obesity and cognition use BMI as their sole obesity index, making comparisons impossible (e.g.,

22

Kuo et al., 2006; Sabina et al., 2009). It may be the case that other obesity indexes, or

combinations of indexes, predict cognition better than BMI among the elderly. For example, a

community study of the elderly found BMI and waist circumference to interact such that high

BMI only predicted poor performance on a brief screener for cognitive impairment when

abdominal obesity was present (Jeong et al., 2005). However, little more is known due to the

paucity of such studies. The present study sought to determine the relationship between

processing speed and multiple obesity indexes. DSST scores were not expected to be

significantly correlated with the overall obesity indexes (i.e., BMI, % DEXA total body fat);

however, they were expected to show significant negative correlations with the central obesity

indexes (i.e. WC, WTR, WHtR, and % DEXA trunk fat). The theoretical importance of central

obesity in predicting cognition is due to the aforementioned differences between metabolic

activity in visceral and other fat tissue, as well as the superior ability of central obesity indexes to

predict a number of health conditions that may impact cognition (e.g., metabolic syndrome,

T2DM) relative to BMI. Furthermore, among those who are centrally obese (defined here as

WC ≥ 102 cm for men and 88 cm for women), it was hypothesized that individuals meeting BMI

obesity criteria (i.e., BMI ≥ 30 kg/m2) and/or DEXA obesity criteria (fat mass ≥ 25% for men

and ≥ 35% for women) would perform worse on the DSST. This hypothesis was based on the

aforementioned study by Jeong and colleagues (2005). Finally, the best combination of obesity

indexes to predict DSST score was explored. It was expected that the best combination would

involve % DEXA trunk fat due to the aforementioned theoretical importance of central obesity,

as well as it being a more accurate indicator of visceral fat than the other central obesity indexes

in this study.

23

Hypothesized Origins of Impaired Cognitive Functioning in Obese Adults

Given the findings that obesity is associated with impairment in virtually all cognitive

domains and may exacerbate age-related cognitive decline, it may prove beneficial to identify the

processes that account for this relationship. This section reviews the relationship between

obesity and processes that are thought to contribute to cognitive decline.

Medical etiological factors in impaired cognitive functioning in obese adults.

Disrupted oxygen flow. Obesity and physical inactivity, which is more prevalent in the

obese population, adversely impact heart and lung functioning, key organs that are responsible

for ensuring an adequate supply of blood and oxygen throughout the body (Chlif, Keochkerian,

Mourlhon, Choquet, & Ahmaidi, 2005; Sharp, Henry, Sweany, Meadows, & Pietras, 1964). For

example, an increased body mass requires the heart to pump a greater volume of blood, thereby

placing additional strain on the heart. Furthermore, breathing can become difficult due to the

additional upper body compressing the lungs. If the heart and the lungs cannot fully meet the

body’s demands for blood and oxygen circulation, the brain may receive inadequate blood and

oxygen supplies, conditions respectively known as cerebral hypoperfusion and hypoxia. These

conditions may cause neuronal death and decreased brain metabolism, which may lead to

impairments in cognition; this idea is supported by animal model studies of cerebral

hypoperfusion (Kim et al., 2012). Decreased blood flow in the prefrontal cortex and temporal

pole have been found in otherwise healthy overweight and obese adults, and this decreased blood

flow appears to be associated with various cognitive deficits (Willeumier, Taylor, & Amen,

2011). Thus, it could be hypothesized that chronic mild hypoxia could have a deleterious impact

on cognition in otherwise healthy obese individuals.

24

Cerebral hypoperfusion and/or cerebral hypoxia is also thought to be a main contributor

to cognitive dysfunction in a number of pathological conditions that are associated with obesity,

including heart failure (HF), ischemic stroke, vascular dementia (VaD), and several respiratory

diseases (Azagra-Calero, Espinar-Escalona, Barrera-Mora, Llamas-Carreras, & Solano-Reina,

2012; Baena-Diez et al., 2010; Strazzullo et al., 2010). Among adults with HF, cerebral

hypoperfusion is associated with poorer performance on tests of attention, memory, and

executive functioning (Alosco, Spitznagel, et al., 2012). Interestingly, one recent study

concluded that deficits in cognitive functioning associated with cerebral hypoperfusion are

exacerbated by elevated BMI (Alosco, Spitznagel, et al., 2012).

In regard to ischemic stroke and VaD, interrupted blood supply to the brain is a hallmark

of these conditions and appears to be the result of multiple etiological factors. Among those who

experience cognitive impairment after a stroke, the most common areas affected appear to be

attention, processing speed, and executive function (Cumming, Marshall, & Lazar, 2013). The

occurrence of one or more strokes raises an individual’s risk of developing dementia

considerably; for example, one study found that 30% of individuals who did not exhibit dementia

prior to suffering a stroke developed dementia within three months (Barba et al., 2000). Changes

in white matter, including white matter hyperintensities, lesions, and diffusity in the frontal and

parietal regions, are associated with worse cognitive performance in stroke survivors and may

play a role in the progression from cognitive deficits during the acute period following stroke to

subsequent progressive cognitive decline and dementia (Dufouil et al., 2009; Rasquin, Verhey,

Lousberg, & Lodder, 2005; Williamson et al., 2010). Reductions in grey matter in the thalamus

have also been associated with cognitive impairment following ischemic stroke and may

contribute to cognitive impairment in that population (Stebbins et al., 2008).

25

Respiratory diseases associated with obesity that have the potential to adversely impact

cognition include obstructive sleep apnea syndrome (OSAS), obesity hypoventilation syndrome

(OHS), chronic obstructive pulmonary disease (COPD), and chronic asthma (Guo et al., 2013;

Murugan & Sharma, 2008). In regard to OSAS, individuals affected by this condition

experience multiple episodes of upper airway obstruction while sleeping and, therefore,

intermittent deprivation of oxygen (Carlucci, Smith, & Corbridge, 2013). Obese individuals are

more likely to experience sleep apnea, which is thought to be at least partly due to a greater

amounts of adipose tissue on the neck and soft palate (Azagra-Calero, Espinar-Escalona,

Barrera-Mora, Llamas-Carreras, & Solano-Reina, 2012). Adults diagnosed with OSAS have

demonstrated cognitive dysfunction in several areas, most notably executive functioning (Daurat,

Huet, & Tiberge, 2010; Gale & Hopkins, 2004). Reduced blood oxygen saturation resulting

from the intermittent deprivation of oxygen caused by the upper airway obstructions may

represent a significant contributor to the cognitive dysfunction in this population. Indeed, animal

models of OSAS, which are created through intermittent deprivation of oxygen, implicate the

presence of oxidative stress and neuronal cell death via increased oxygen-derived free radicals

known as reactive oxygen species (ROS) (Wang, Zhang, & Gozal, 2010). Furthermore, many

OSAS neuroimaging studies have reported reduced hippocampal volume, and there is a

significant positive correlation between hippocampal volume and blood oxygen saturation in

patients with OSAS (Gale & Hopkins, 2004; Lal, Strange, & Bachman, 2012). It is also worth

noting here that daytime sleepiness associated with OSAS can impair performance on cognitive

tests, although there appears to be a pattern of impaired executive functioning in individuals with

OSAS even after controlling for this variable (Daurat et al., 2010).

26

Compared to OSAS, much less is known about OHS, an often-misdiagnosed syndrome

defined as the combination of obesity, abnormal blood levels of carbon dioxide in the daytime,

and sleep-disordered breathing in the absence of a known cause (Basoglu & Tasbakan, 2014;

Marik & Desai, 2013). Preliminary evidence suggests that OHS, which often presents with

OSAS, high C-reactive protein (CRP) levels, and vitamin D deficiency, may result in multiple

organ system dysfunction and can be fatal (Marik & Desai, 2013). Observations of mental

deterioration have been noted in cases of OHS, but future research is needed to determine

specific cognitive changes that occur as a result of this syndrome (Troester, Palfner, Kovacs, &

Olschewski, 2013).

COPD, a progressive disease that adversely impacts airflow, typically involves both

chronic bronchitis and emphysema (Poulain et al., 2006). Smoking is the leading cause of this

disease (United States Department of Health & Human Services, 2014c). It is unclear if obesity

is a risk factor for COPD, but obesity appears to be more prevalent in those with COPD and is

thought to exacerbate COPD symptoms (Franssen, O’Donnell, Goossens, Blaak, & Schols,

2008). Although many studies have reported cognitive dysfunction in COPD, a specific

cognitive profile has not been consistently identified. Some research has indicated a diffuse

pattern of cognitive impairment in this population, whereas as other research has only found

impairments in memory (De Carolis et al., 2011; Fioravanti, Nacca, Amati, Buckley, & Bisetti,

1995; Incalzi et al., 1993). It has been hypothesized that COPD contributes to cognitive

dysfunction via chronic hypoxia (De Carolis et al., 2011). Furthermore, a recent longitudinal

study found COPD and asthma in midlife to be associated with an increased risk of mild

cognitive impairment (MCI) and dementia in late life (Rusanen et al., 2013). In regard to the

present study, current or history of respiratory disease was expected to contribute to processing

27

speed impairment. Duration of respiratory disease was also expected to be negatively correlated

with processing speed task performance.

Oxidative stress and nutrition. Oxidative stress can be described as a disturbance in the

balance between free radicals derived from oxygen known as ROS and antioxidant defenses on

the cellular level (Betteridge, 2000; Wang et al., 2010). Increased amounts of ROS are thought

to contribute to tissue damage and cognitive dysfunction (Betteridge, 2000; Wang et al., 2010).

Oxidative stress has been found to increase after consumption of high fat foods, and

chronic consumption of high fat foods has been linked to cognitive impairment (Bloomer &

Fisher-Wellman, 2009; Winocur & Greenwood, 2005). Additionally, there is an inverse

relationship between oxidative stress and fruit and vegetable consumption, which may be related

to the high vitamin and antioxidant-rich carotenoid content of these foods (Vetrani, Costabile, Di

Marino, & Rivellese, 2013). Indeed, deficiencies in B vitamins and thiamine have been linked to

cognitive impairments and Alzheimer disease (Lu’o'ng & Nguyen, 2011; Selhub, Bagley, Miller,

& Rosenberg, 2000).

A number of pathological conditions for which obesity is a risk factor (e.g., OSAS, CVD,

various types of cancers) are associated with oxidative stress (Vetrani et al., 2013). Even

otherwise healthy obese individuals may be at greater risk for oxidative stress than their normal

weight counterparts. This idea is supported by studies that have found significant associations

between obesity and the dietary characteristics of lower fruit and vegetable consumption and

higher total fat consumption, though it should be noted that these associations have not been

consistent across studies (Davis, Hodges, & Gillham, 2006; Tohill, Seymour, Serdula, Kettel-

Khan, & Rolls, 2004).

28

In addition to increasing the risk for oxidative stress, obesity and poor nutrition may also

contribute to cognitive dysfunction via an increase in free fatty acid plasma concentration (Dali-

Youcef, Mecili, Ricci, & Andres, 2013; Haltia et al., 2007). There are at least two mechanisms

by which this impairment may occur. First, free fatty acids may contribute to altered brain fat

metabolism and elevated fat accumulation in brain white matter (Haltia et al., 2007). This

hypothesis is consistent with the findings of greater white matter volume in several basal brain

areas in obese individuals relative to healthy-weight controls (Haltia et al., 2007). Second, free

fatty acids may contribute to impaired insulin signaling, which may in turn contribute to

cognitive impairment via processes discussed later in this document (Dali-Youcef et al., 2013;

Haltia et al., 2007).

Endothelial dysfunction. Oxidative stress may also contribute to endothelial

dysfunction, which refers to a pathological state of the inner lining of blood and lymphatic

vessels, known as the vascular endothelium (Münzel, Sinning, Post, Warnholtz, & Schulz, 2008).

Biological markers of endothelial dysfunction have been associated with a number of conditions

that may lead to impaired cognition, including hypertension, dyslipidemia, and insulin resistance,

and these associations may be bidirectional (Münzel et al., 2008). Interestingly, circulating

biomarkers of endothelial dysfunction and low-grade inflammation (as discussed below) were

associated with impairments in processing speed and executive functioning in a population-based

study of the elderly, and accounting for vascular risk factors only slightly attenuated their

relationship (Heringa et al., 2014).

Chronic low-grade inflammation. Obesity is characterized by increased number and

size of fat cells, known as adipocytes, which secrete proinflammatory (e.g., interleukin 6) and

anti-inflammatory (e.g., adiponectin) cell-signaling proteins called cytokines (Alexaki et al.,

29

2009; Tsatsanis et al., 2006; Visser, Bouter, McQuillan, Wener, & Harris, 1999). An increase in

adipose tissue, visceral adipose tissue in particular, has been suggested to result in a higher ratio

of proinflammatory to anti-inflammatory cytokines (Trujillo & Scherer, 2005; Tsatsanis et al.,

2006). Reduced production of adiponectin, an anti-inflammatory cytokine, in response to greater

adipose tissue, and visceral adipose tissue in particular, may have a large contributory role in this

process (Tsatsanis et al., 2006). This proinflammatory cytokine imbalance is thought to lead to

peripheral inflammation, and, indeed, waist circumference is positively correlated with

peripheral inflammation markers and negatively correlated with adiponectin (Ackermann et al.,

2011). Moreover, obese individuals demonstrate elevated levels of CRP, a general marker of

peripheral inflammation, independent of other known sources of inflammation, such as health

and smoking status (Visser et al., 1999).

Inflammation is a protective process performed by the body in order to clear damaging

stimuli and assist with tissue repair (Michaud et al., 2013). The biological benefit of obesity-

caused peripheral inflammation appears to be increased energy expenditure, with inflammation

thereby acting as an energy balance mechanism (Ye & Keller, 2010). However, there are a

number of negative consequences to chronically heightened inflammation that could contribute

to the processing speed and executive functioning impairments that are found in this population

(Heringa et al., 2014). First, results from animal studies suggest that peripheral inflammation is

able to evoke central neuroinflammatory response, and this appears to be especially true in older

age animals (Barrientos, Hein, Frank, Watkins, & Maier, 2012; Godbout et al., 2005). Second,

inflammation resistance may develop over time, causing a greater build-up of proinflammatory

cytokines in the bloodstream (Ye & Keller, 2010). These proinflammatory cytokines may

interact with free fatty acids to impair insulin signaling (Dali-Youcef et al., 2013; Haltia et al.,

30

2007). Conditions involving insulin-resistance (as described below) are strongly linked with

cognitive impairment. Finally, there is evidence to suggest that central inflammation

(specifically, in the hypothalamus) may be triggered by high fat diet in addition to being

triggered by the metabolic consequences of excess adipose tissue (Boitard et al., 2014; Velloso,

Araujo, & de Souza, 2008). Regarding the present study, higher levels of chronic peripheral

inflammation (as suggested by a history of obesity and high level of CRP) were expected to be

associated with cognitive impairment.

Impaired insulin production or signaling. Insulin, a hormone primarily produced in the

pancreas, is important for regulation of carbohydrate and lipid metabolism, as well as neuronal

survival and synaptic plasticity (van der Heide, Ramakers, & Smidt, 2006). Peripheral insulin is

generally thought to be able to travel through the blood-brain barrier to the brain, although there

is some evidence to suggest that central production may occur (Woods, Seeley, Baskin, &

Schwartz, 2003; van der Heide et al., 2006). Obese individuals are more likely to experience

peripheral impairments in insulin signaling, which may be due, at least in part, to increased

systemic inflammation, oxidative stress, and high fat diet in this population (Dali-Youcef et al.,

2013; Greenwood & Winocur, 2005; Henriksen, Diamond-Stanic, & Marchionne, 2011).

Peripheral insulin resistance or insufficient insulin production is a hallmark of T2DM, and post-

mortem studies have shown central insulin resistance to be greater in T2DM patients than age-

matched controls (Liu, Liu, Grundke-Iqbal, Iqbal, & Gong, 2011). The relationship between

T2DM and deficits in processing speed, memory, and executive functioning among older adults

have been well-established (Blanco et al., 2012; Ryan & Geckle, 2000; van den Berg et al., 2008,

2009; Yaffe et al., 2012; Yeung, Fischer, & Dixon, 2009). Mechanisms for cognitive

dysfunction in T2DM are not well understood, but some evidence suggests the role of

31

hippocampal neuron damage due to inadequate supply of glucose to the brain, reduced central

cholinergic activity, decreased functionality of glutamate receptors, oxidative stress, and

disruption of the hypothalamic-pituitary adrenal (HPA) axis, in addition to central insulin

resistance (Kodl & Seaquist, 2008; Raber, 1998; Shonesy et al., 2012; Vijayakumar, Sirisha,

Begam, & Dhanaraju, 2012). Decreased frontal glucose metabolism and impaired neuronal

integrity have also been related to BMI and poorer memory and executive function in adults

without T2DM but who may be somewhat insulin-resistant (Gazdzinski et al., 2010; Volkow et

al., 2009).

Insulin-resistance may also lead to cognitive dysfunction other ways. For example,

insulin-resistance, along with other cardiometabolic syndrome symptoms, has been implicated in

the pathogenesis of chronic kidney disease, which, in turn, is associated with impaired

psychomotor efficiency and processing speed (Jassal, Roscoe, LeBlanc, Devins, & Rourke,

2008; Whaley-Connell et al., 2006). Insulin-resistance may also cause cognitive impairment via

promotion of the build-up of beta amyloid (Aβ) plaques found in AD, and in fact, insulin-

resistance has been proposed as the primary precipitating event of this disease (Kodl & Seaquist,

2008; Salkovic-Petrisic & Hoyer, 2007). Indeed, autopsies performed in patients with AD show

central insulin resistance, and there is considerable similarity between pathology of experimental

inducement of central insulin resistance in animal studies and AD pathology in humans (Liu,

Liu, Grundke-Iqbal, Iqbal, & Gong, 2011; Shonesy et al., 2012). Such findings have led to the

proposal of the term type 3 diabetes to describe this mechanism of neurodegeneration (Steen et

al., 2005).

As per the current literature, the potential contributory role of peripheral insulin

resistance in impaired processing speed is unclear. Among middle-aged individuals without

32

diabetes mellitus, two biological markers of peripheral insulin-resistance (i.e., hyperinsulinemia

and homeostasis model assessment of insulin resistance [HOMA-IR]), have been associated with

poor performance on tests of processing speed; however, the only study that appears to have

controlled for BMI when testing the relationship between insulin resistance and poor processing

speed to-date reported that it became insignificant after doing so (Sanz et al., 2013; Young,

Mainous, & Carnemolla, 2006). This suggests the possibility that peripheral insulin resistance,

itself, may not contribute to processing speed decline in a middle-aged population. The

relationship between central insulin resistance and AD pathology, however, raises the question of

whether insulin resistance would be associated with processing speed decline among the elderly

independent of BMI. The present study investigated this question as part of its exploratory

analyses.

Hyperglycemia/hypoglycemia. Blood glucose level and peripheral insulin resistance are

closely related, but distinct, concepts. Insulin is needed to reduce blood glucose levels, and the

pancreas of an individual developing insulin resistance typically produces extra insulin in order

to compensate for the insulin resistance. The individual can eventually develop hyperglycemia if

this extra insulin is no longer adequate to reduce blood glucose levels to a normal range.

Therefore, peripheral insulin resistance onset precedes peripheral hyperglycemia onset (United

States Department of Health & Human Services, 2017). Once an individual’s fasting glucose

level reaches 100–125 mg/dl or HbA1c, a marker of average glucose level in the past 2–3 months,

reaches 5.7% to 6.4%, prediabetes is diagnosed (American Diabetes Association, 2014). When

fasting glucose level reaches ≥ 126 mg/dl or HbA1c reaches ≥ 6.5%, diabetes mellitus is

diagnosed (American Diabetes Association, 2014).

33

Hyperglycemia in the diabetic or prediabetic range may have an adverse impact on

processing speed. Hyperglycemia has been identified as the major contributor to the relationship

between metabolic syndrome and impairment in tasks requiring fast perceptual speed among a

sample of elderly individuals (Dik et al., 2007). Furthermore, HbA1c levels in the prediabetic

range have been linked to impaired processing speed independent of BMI in middle-aged

individuals (Sanz et al., 2013). It is unclear as to how hyperglycemia might contribute to

cognitive impairment, although specific biological pathways in the brain have been proposed

(Kodl & Seaquist, 2008). More broadly, Dik and colleagues (2007) noted that those with

diabetes mellitus also show perceptual speed deficits, and suggested that similar glucose

intolerance pathology in fronto-subcortical structures may be present in both those with diabetes

and prediabetes. There is also strong evidence that hyperglycemia increases the risk for

atherosclerosis, possibly via endothelium damage, oxidative stress, and inflammation (Aronson

& Rayfield, 2002).

In regard to the present study, it was hypothesized that participants with fasting glucose

or HbA1c levels in the prediabetic range would perform worse on the processing speed task than

would those with fasting glucose and HbA1c levels in the normal range; individuals diagnosed

with diabetes mellitus or on antidiabetic medication were excluded from the sample investigated

by this hypothesis. In addition, it was hypothesized that those with diabetes mellitus would

perform worse on the processing speed task than the prediabetic and the non-diabetic groups.

Hypoglycemia has also been hypothesized to contribute to worsening cognition in those

with diabetes mellitus (Kodl & Seaquist, 2008). Namely, repeated episodes of severe

hypoglycemia have been linked to worse cognition in a variety of domains within this

population, although findings are mixed and most work has focused on type 1 diabetes, a

34

condition for which obesity is not a risk factor (e.g., DCCT/EDIC Study Research Group, 2007;

Northam et al., 2001). Hypoglycemia history is not included in the NHANES database;

therefore, no hypotheses were made regarding history of this condition.

Adipose tissue hormones. It has been proposed that differential levels of adipose tissue-

derived hormones, including adiponectin and leptin, may have a complex role in cognitive

dysfunction pathology (Gustafson, 2010). Adiponectin (discussed above) has proinflammatory

effects that may contribute to the development of cognitive impairment. Circulating levels of

leptin, a neuroprotective hormone which is also thought to assist in the regulation of food intake

and body weight by acting on the hypothalamus to signal satiety and increase energy

expenditure, tend to be elevated in the obese, and obese individuals are also more likely to show

peripheral leptin resistance, central leptin insufficiency, and brain atrophy (Harvey, 2007; Heini

et al., 1998; Rajagopalan, Toga, Jack, Weiner, & Thompson, 2013). In addition to the

hypothalamus, leptin also has receptor sites in several other areas of the brain, including the

cerebellum and hippocampus (Funahashi, 2003; Savioz et al., 1997). It has been proposed that

leptin deficiencies and leptin resistance in lean and obese individuals, respectively, are related to

cognitive dysfunction (Labad et al., 2012). In line with this argument, rats with dysfunctional

leptin receptors were found to perform poorly on spatial memory tasks (Li et al., 2002). Another

study found higher levels of leptin to be associated with poorer mental flexibility and executive

functioning among a group of primarily obese men with T2DM (Labad et al., 2012).

Hypertension. Over 70% of obese adults suffer from hypertension, a condition that is

associated with cognitive dysfunction and is linked to deficits in attention, memory, executive

functioning, psychomotor functioning, and motor speed among those with heart failure

independent of medication status (Alosco, Brickman, et al., 2012; Bramlage et al., 2004; van den

35

Berg, Kloppenborg, Kessels, Kappelle, & Biessels, 2009). Hypertension has also found to be

significantly associated with cognitive dysfunction independent of obesity among men, although

the same relationship was not found among women (Elias et al., 2003). There is some evidence

to suggest that, among the general population, only those with long histories of hypertension and

those not on antihypertensive medications evidence these cognitive deficits (Farmer et al., 1990;

Power, Tchetgen, Sparrow, & Schwartz, 2013).

Prolonged hypertension may impair cognition through disturbed cerebral blood flow

distribution and alterations to small cerebral blood vessels, which may lead to inadequate blood

supply to brain tissue, blood-brain barrier alterations, and eventually lesions, white matter

changes, silent strokes that do not yield any outward symptoms, stroke, and VaD (Elias, Elias,

Sullivan, Wolf, & D’Agostino, 2003; Pantoni & Simoni, 2003). Hypertension has also been

associated with exacerbated morphological changes that occur during the natural aging process,

particularly in the prefrontal cortex, as well as in the temporal and occipital lobes and underlying

white matter (Raz, Rodrigue, & Acker, 2003; Strassburger et al., 1997).

Hypertension in midlife is a risk factor for VaD and AD development (Kivipelto, Laakso,

Tuomilehto, Nissinen, & Soininen, 2002; Sharp, Aarsland, Day, Sønnesyn, & Ballard, 2011).

Interestingly, among those with midlife hypertension who go on to develop dementia, blood

pressure tends to decline near the time cognitive functioning starts to decline (Hajjar, Keown, &

Frost, 2005). This may result in many individuals with dementia no longer meeting criteria for

hypertension. Nevertheless, if hypertension is present in an individual with dementia, it may

accelerate cognitive decline (Hajjar et al., 2005).

Dyslipidemia. Dyslipidemia (e.g., unhealthy levels of total cholesterol, high-density

lipoprotein [HDL] cholesterol, low-density lipoprotein [LDL] cholesterol, and/or triglycerides) is

36

a risk factor for cardiovascular diseases, which in turn are risk factors for cognitive decline, late-

life dementia, and AD onset in carriers of the ApoE4 allele (Eriksson, Bennet, Gatz, Dickman, &

Pedersen, 2010; Reiner, Muacevic-Katanec, Katanec, & Tedeschi-Reiner, 2011). For example,

60–70 and 70–80-year olds with plasma cholesterol levels of ≥ 8 mmol/L demonstrate a coronary

heart disease (CHD) risk that is 2.3 and 1.6 times greater than those with cholesterol levels of < 5

mmol/L, respectively (Iversen, Jensen, Scharling, & Schnohr, 2009). The link between plasma

lipids and CHD is thought to be primarily due to atherosclerosis, a thickening of artery walls that

limits blood flow (Elwood, Pickering, Bayer, & Gallacher, 2002; Iversen et al., 2009).

Atherosclerosis itself has also been associated with slower processing speed in the elderly, which

may be due to decreased blood flow to the brain (Vinkers et al., 2005). It has been theorized that

plasma lipids, including low HDL cholesterol and high LDL cholesterol, may also contribute to

cognitive impairment via the build-up of amyloid plaque deposits in the brain (Reed et al., 2014).

Given these theoretical etiological factors, it was hypothesized that dyslipidemia would be

related to worse processing speed in the present study.

Some studies have already linked dyslipidemia with cognitive impairment, although

findings have been mixed. For example, there is some evidence that high triglycerides are

related to executive dysfunction in elderly without dementia after controlling for age, gender,

education, LDL cholesterol, ApoE4 allele status, and a rating of clinical dementia (Parthasarathy

et al., 2014). Low HDL cholesterol may be independently related to memory dysfunction and

decline in middle-aged adults and has been associated with reductions in grey matter volume in

regions associated with neurodegenerative disease, such as the parahippocampal region (Singh-

Manoux, Gimeno, Kivimaki, Brunner, & Marmot, 2008; Ward et al., 2010). Overall, however,

studies of the relationship between dyslipidemia and cognitive functioning have yielded

37

inconsistent results, which may be related to the common pattern of favorable body weight

changes prior to dementia onset (Stewart, White, Xue, & Launer, 2007; van den Berg et al.,

2009). Unfortunately, the present study did not include a direct measure of dementia.

Other factors. Obesity appears to be a risk factor for alcoholic fatty liver disease and

non-alcoholic fatty liver disease, conditions which appear to be causally linked to cognitive

impairment (Lee et al., 2006). For example, those with end-stage liver disease can experience

hepatic encephalopathy, which is thought to be caused by a build-up of toxins in the bloodstream

and may involve inattention, confusion, and disorientation (Rahimi & Rockey, 2014). Cognition

appears to return to a normal state after hepatic encephalopathy is treated; however, recent

evidence suggests that learning deficits may continue after treatment (Umapathy, Dhiman,

Grover, Duseja, & Chawla, 2014). Less severe stages of liver disease may also be linked to

impaired cognition in certain domains. For example, one study that compared those with mild

liver disease and hepatitis C virus (HCV) to controls demonstrated that the former group had

worse concentration and memory speed after controlling for depression and fatigue, although

DSST scores were not significantly different (Forton et al., 2002). The authors of this study

suggested that an impact of chronic HCV on the central nervous system or preexisting individual

differences may account for these findings (Forton et al., 2002, 2005). Others have suggested

that substance use history accounts for some of the relationship between HCV and cognitive

impairment. For example, Huckans et al. (2009) reported impairments in processing speed, as

estimated by the DSST, in a group with HCV and SUD history but not in a group with HCV and

no SUD history. Nevertheless, deficits in verbal learning, auditory attention, reasoning, and

mental flexibility persisted even in the absence of SUD history in those with HCV. The

NHANES database used in the present study contains yes/no questions on history of “any liver

38

condition” and whether those who endorsed this item still have “any liver condition.”

Unfortunately, this did not allow for identification of specific liver conditions nor did it allow for

specific hypotheses to be made in the present study. Nonetheless, the relationship between “any

liver condition” and current processing speed was investigated as part of the present study’s

exploratory analyses. Based on the above information, history of “any liver condition” was not

expected to be significantly related to processing speed task performance, though this was not a

formal hypothesis.

It should also be noted here that hypothyroidism has been linked to subtle cognitive

deficits when left untreated, whereas less evidence is available for an adverse impact of

hyperthyroidism (Lillevang-Johansen, Petersen, Christensen, Hegedüs, & Brix, 2014; Samuels,

2008, 2014). Hypothyroidism has been associated with obesity; however, the direction of this

relationship is unclear (Verma, Jayaraman, Kumar, & Modi, 2008). The NHANES data used in

the present study contains yes/no questions on history of “thyroid disease.” Unfortunately, this

did not allow for differentiation between types of thyroid disease. In regard to the present study,

thyroid disease was not expected to be significantly linked with impaired processing speed

unless it was not treated. That being said, upon implementation of the present study, it was

discovered to be impossible to determine whether thyroid disease had been treated, as the

NHANES data does not include information on history of many of the treatments for thyroid

diseases, such as those of radioiodine and surgery (Abraham & Acharya, 2010).

Combinations of risk factors. The above subsections have included information on the

potential interactive effects of various health conditions on cognition. Some researchers have

also attempted to study the relationship between cognition and clusters of risk factors for

diseases/events, most notably stroke and cardiovascular disease. In doing so, most of these

39

studies use full or modified versions of the Framingham Stroke Risk Profile (FSRP) or the

Framingham Risk Score (FRS) for general cardiovascular disease (CVD; e.g., Bangen et al.,

2010; Elias et al., 2004; Joosten et al., 2013; Kaffashian et al., 2011; Llewellyn et al., 2008).

Both the FSRP and FRS for general CVD were developed using data from the Framingham

Study, an ongoing cohort study in Framingham, Massachusetts (Boston Medical Center, 2014).

In regard to the FSRP, the following stroke risk factors were identified based on the incidence of

stroke over a 10-year period among participants aged 55–84 without a history of stroke at

baseline: age, systolic blood pressure, antihypertensive medication, diabetes, smoking status,

atrial fibrillation, left venticular hypertrophy, and history of CVD (defined as one or more of the

following: myocardial infarction, angina pectoris, coronary insufficiency, intermittent

claudication, congestive heart failure [CHF]; D'Agostino, Wolf, Belanger, & Kannel, 1994;

Wolf, D'Agostino, Belanger, & Kannel, 1991). In regard to the FRS for general CVD, the

following CVD risk factors were identified based on the incidence of CVD (defined as one or

more of the following: heart failure, coronary insufficiency, myocardial infarction, angina

pectoris, ischemic or hemorrhagic stroke, transient ischemic attack, peripheral artery disease, or

coronary death) over a 10-year period among participants aged 30–74 without a history of CVD

at baseline: age, total cholesterol, HDL cholesterol, systolic blood pressure, antihypertensive

medication, smoking, and diabetes (D'Agostino et al., 2008). FSRP and FRS for general CVD

scores can be calculated by assigning points to each of these risk factors using sex-specific

formulas that can be found within their respective publications (D'Agostino et al., 1994, 2008).

As mentioned above, many studies have used full or modified versions of the FSRP and

FRS for general CVD to assess the relationship between these groups of risk factors and

cognition. Modified versions appear to be used due to availablity of data in a study rather than

40

theoretical differences. For example, Llewellyn and colleagues (2008) calculated FSRP scores

without left venticular hypertrophy and used a different definition of history of CVD (i.e., CHD,

CHF, and peripheral vascular disease [PVD]). Llewellyn and colleagues (2008) found that the

modified FSRP was related to worse performance in a number of cognitive domains, including

processing speed, in their United Kingdom population sample of individuals aged 50 years and

older. Research has also shown high FRS for general CVD scores to be associated with worse

cognition in a number of areas, most notably executive functioning (Joosten et al., 2013;

Kaffashian et al., 2011). The relatonship between FRS for general CVD scores and processing

speed has not been reported to date. In regard to the present study, it was hypothesized that

scores yielded by modified versions of the FSRP and FRS for general CVD would be inversely

associated with processing speed performance. The present study calculated FSRP scores

without left venticular hypertrophy and atrial fibrillation, and it used a different definition of

CVD (i.e., CHF, angina pectoris, heart attack). The present study calculated FRS for general

CVD scores in the same manner as the full version but used a different definition of CVD (i.e.,

CHF, angina pectoris, heart attack, stroke).

Psychological etiological factors in impaired cognitive functioning in obese adults.

Depression. Impaired ability to think, concentrate and make decisions are among the

symptoms of major depressive disorder (MDD; American Psychiatric Association, 2013). Not

surprisingly, research indicates that MDD is associated with impairment in a variety of cognitive

domains, including attention, processing speed, memory, and executive functioning (Luo et al.,

2013; Marazziti, Consoli, Picchetti, Carlini, & Faravelli, 2010). Obese adults are more likely to

experience depressive disorders compared to their non-obese counterparts, with stronger

associations among those with greater obesity severity (Castres, Folope, Dechelotte, Tourny-

41

Chollet, & Lemaitre, 2010; Mather et al., 2009). The connection between depression and obesity

is not fully understood, but longitudinal evidence suggests a bidirectional relationship (Luppino

et al., 2010). Body image dissatisfaction, reduced levels of social and physical activity, and use

of psychotropic medication appear to mediate this relationship (de Wit et al., 2010; Gavin,

Simon, & Ludman, 2010; Smits et al., 2010). Additional factors accounting for the relationship

between elevated body weight and depression may include HPA-axis dysregulation,

hypercortisolemia, unhealthy diet, emotional eating, carbohydrate craving, binge eating, and

social stigmatization (Carels et al., 2010; Gibson, 2012; Hawkins & Stewart, 2012; Ouwens, van

Strien, & van Leeuwe, 2009; Peterson, Latendresse, Bartholome, Warren, & Raymond, 2012;

Wurtman & Wurtman, 1995). Among the elderly, it has been proposed that physical functioning

might be a particularly important mediator between obesity and depression, whereas factors

important in explaining this relationship in young adults (e.g., body image) may become less

important in old age (McCrea et al., 2012).

Anxiety. Anxiety has been strongly linked to attentional bias toward threatening

information and increased distractibility, which in turn may contribute to short-term memory and

executive dysfunctions in this population (Lapointe & Blanchette, 2013). Elevated rates of past-

year and lifetime anxiety disorders and symptoms, particularly social phobia and panic attacks,

have been documented in the general obese population (Mather et al., 2009). The nature and

direction of the relationship between anxiety and obesity is not well understood, but it has been

hypothesized that more frequent stress exposure predisposes obese individuals to anxiety

disorders (Pickering, Grant, Chou, & Compton, 2007). Additionally, longitudinal evidence

suggests that severe anxiety may predispose one to increases in waist circumference and

decreases in HDL cholesterol, both of which are predictive of CVD risk and cognitive

42

impairment (Bombelli et al., 2013; Reiner et al., 2011; Singh-Manoux et al., 2008; van Reedt

Dortland, Giltay, van Veen, Zitman, & Penninx, 2013; Yoon et al., 2012). Other potential

factors accounting for the relationship between anxiety and higher BMI include unhealthy diet,

HPA-axis dysregulation, hypocortisolemia, and psychotropic medication (Boyer, 2000; Hawkins

& Stewart, 2012; Smits et al., 2010).

Attention-deficit/hyperactivity disorder. Persistent pattern of inattention, difficulties

with organization, and impulsivity are among the symptoms of attention-deficit/hyperactivity

disorder (ADHD), and each of the three subtypes of this disorder (i.e., predominantly inattentive,

predominantly hyperactive/impulsive, and combined) is associated with executive dysfunction

(American Psychiatric Association, 2013; Pagoto et al., 2009). Adults with ADHD also exhibit

elevated rates of obesity (Pagoto et al., 2009). The relationship between ADHD and obesity is

not well understood, but three main theories have been proposed. First, effects of obesity-related

comorbidities (e.g., decreased alertness as a consequence of obstructive sleep apnea syndrome)

may be mistaken for ADHD symptoms (Cortese & Vincenzi, 2012). Second, cognitive deficits

common to ADHD may interact with environmental factors to predispose individuals with

ADHD to overeat or make unhealthy food choices (Davis, 2010). This idea is partly supported

by findings that impulsivity and/or impaired inhibitory control are positively correlated with

external eating, taste-preference driven food choices, and quantity of ready-to-eat and away-

from-home foods consumed (Appelhans et al., 2012; Jasinska et al., 2012). Third, it has been

hypothesized that deficits in dopaminergic activity in reward centers of the brain underlie both

ADHD and obesity and drive their association (Campbell & Eisenberg, 2007; Liu, Li, Yang, &

Wang, 2008). For example, individuals with reduced dopaminergic activity in these reward

centers may attempt to compensate via behaviors that promote dopaminergic activity, including

43

overeating and consumption of unhealthy foods (Davis, 2010; Wang et al., 2001). The finding

that binge eating disorder, which may be associated with downregulation of the dopamine

receptor D2 in the striatum, partially mediates the association between ADHD and obesity is

consistent with this hypothesis (Johnson & Kenny, 2010; Pagoto et al., 2009).

Intellectual disability. Intellectual disability (ID) is a neurodevelopmental disorder that

involves impairment in both intellectual and adaptive functioning (American Psychiatric

Association, 2013). Although there is not a strict IQ criterion for diagnosis, individuals with this

disorder typically obtain IQ scores of 65–75 or less (American Psychiatric Association, 2013).

Adults with ID demonstrate increased prevalence of obesity and earlier age of obesity onset

compared with the general population (de Winter, Bastiaanse, Hilgenkamp, Evenhuis, & Echteld,

2012; Melville, Hamilton, Hankey, Miller, & Boyle, 2007). The nature of the association

between ID and obesity is uncertain. One longitudinal study found lower IQ at age 11 to predict

weight gain and obesity at age 42, a relationship which was partly mediated by education level

and diet quality (Chandola, Deary, Blane, & Batty, 2006). A number of risk factors for obesity

within the intellectually disabled population have been identified, including female gender,

Down syndrome, and ability to eat, prepare meals, and shop for groceries independently. Obese

individuals with ID do not appear to have less nutrition knowledge than lean individuals with ID

(Caton et al., 2012; Golden & Hatcher, 1997).

Future research directions in the obesity-related etiology of cognitive dysfunction.

To summarize, there is conflicting evidence regarding whether associations between obesity and

cognitive dysfunctions are independent or can be sufficiently explained by mediating variables

(e.g., hypertension, diabetes status). In addition to focusing on medical diagnoses as mediating

variables, it may be of equal or greater importance for future research to study the contribution of

44

lower-level obesity-related factors (e.g., low grade inflammation, increased circulation of

adipocyte hormones) to cognition decline. This idea is supported by a recent population-based

study of the elderly that found that circulating biomarkers of low-grade inflammation and

endothelial dysfunction were associated with impairments in processing speed and executive

functioning and that accounting for vascular risk factors only slightly attenuated their

relationship (Heringa et al., 2014). It may be further hypothesized that these lower-level obesity-

related factors that contribute to cognitive decline have an increased effect if the individual is

exposed to a greater magnitude (e.g., more hormones released from adipose tissue that contribute

to structural brain changes) and/or over a greater time period (e.g., chronic low-grade

inflammation in the chronically obese person). Therefore, it follows that knowledge of a

person’s weight history, which may serve as a proxy for a number of lower-level obesity

variables, may provide additional utility in predicting cognitive functioning than history of

medical diagnoses alone. Indeed, there is some evidence to support the notion that weight

history may impact cognition. However, this area of study is in its infancy, and its data has not

been adequately synthesized. Future research should work to clarify the impact of weight history

on cognition.

Importance of Processing Speed in the Elderly

Of the cognitive domains that weight history may impact, processing speed may have

some of the worst repercussions among the elderly. First, processing speed is a crucial driving

factor in many accounts of cognitive aging, which will be elaborated on in the following section.

This suggests that weight history-related cognitive deficits may speed the process of cognitive

aging. Second, elderly individuals with faster processing speed appear to have more satisfactory

outcomes in a variety of areas involving quality, satisfaction and duration of life, and these

45

outcomes are typically independent of other cognitive domains. For example, faster processing

speed in the elderly has been found to be associated with shorter time to complete everyday tasks

and greater survival time independent of medical conditions and performance in other cognitive

domains (Iwasa et al., 2014; Owsley et al., 2002). Furthermore, processing speed and spatial

abilities, but not memory or other cognitive domains, were found to predict greater life

satisfaction prospectively in a large population-based sample of 78–98-year-old men and women

(Enkvist et al., 2013). Processing speed training may also protect against health-related quality

of life decline; a large randomized controlled trial reported that elderly individuals who

underwent processing speed training were significantly less likely to evidence extensive health-

related quality of life decline at a two-year follow-up evaluation compared with those who

underwent memory training, reasoning training, or no training (Wolinsky et al., 2006).

Additionally, performance on the digit symbol substitution test (DSST), a commonly used

measure of processing speed, has been shown to be a more useful predictor of driving safety in

the elderly than performance on cognitive tasks measuring global functioning, visuospatial

perception, working memory, semantic fluency, and selective attention (Lafont et al., 2010).

Taken together, these findings suggest the importance of efforts to minimize processing speed

decline in the elderly. Research that will allow us to better understand the process of cognitive

aging and the contributors to processing speed decline among the elderly is also warranted.

Cognitive Aging

Humans experience declines in cognitive functioning as they age. Although normal age-

related cognitive decline has been difficult to clearly differentiate from pathological cognitive

decline due to difficulty ruling out underlying disease states and accounting for impact of

46

medical history, there is a consensus that declines in most cognitive domains occur with aging

even in the absence of known health problems (e.g., Aine et al., 2011; Harada, et al., 2013).

Determining when age-related cognitive decline begins is complex. Decline onset varies

by a variety of factors including cognitive domain and birth cohort (Schaie & Willis, 2010).

Generally, it believed that fluid abilities, such as problem solving, begin declining in late

adolescence or early adulthood, whereas crystallized abilities, such as vocabulary, are maintained

or improve into late adulthood (e.g., Kaufman & Horn, 1996). Recently, attention has been

drawn to the impact of study design on our conclusions regarding when age-related decline

begins (Rönnlund, Nyberg, Bäckman, & Nilsson, 2005; Salthouse, 2009; Schaie, 2009). Most

research has been cross-sectional rather than longitudinal for practical reasons, and cross-

sectional research appears to be more likely to show continuous age-related decline throughout

the adult lifespan than longitudinal research, which is not likely to show significant decline until

approximately age 55 or later (Rönnlund et al., 2005; Salthouse, 2009; Schaie, 2009; Schaie &

Willis, 2010). Those who contend that age-related cognitive decline begins in early life argue

that longitudinal studies do not evidence decline due to retest effects, whereas those who contend

that age-related cognitive decline begins in late life argue against the validity of conclusions that

have been drawn on the basis of cross-sectional studies due to cohort effects (Salthouse, 2009;

Schaie, 2009).

Processing speed, inhibitory control, and working memory are among the factors most

frequently hypothesized to be important in explaining, or even driving, normal age-related

cognitive decline (e.g., Hasher, Stoltzfus, Zacks, & Rypma, 1991; Rozas, Juncos-Rabadán, &

González, 2008; Vance, 2009). Among these, processing speed may be particularly important,

as the effect of age on processing speed may mediate the effect of age on working memory and

47

inhibition (Rozas et al., 2008). Furthermore, changes in the processing factor are a leading

indicator of changes in memory and spatial factors, whereas the inverse relationship does not

hold nor are changes in the processing factor a leading indicator of changes in the verbal factor

(Finkel et al., 2007). These findings are consistent with the processing speed theory of aging,

which holds that reduced processing speed in old age restricts other cognitive processes to the

extent that these cognitive processes rely on processing speed (Salthouse, 1996). In other

words, these findings suggest that processing speed, which is related to fluid ability more so than

crystallized ability, drives fluid ability decline but not crystallized ability decline (Finkel et al.,

2007).

In addition to these traditionally studied cognitive components, some researchers have

sought to understand age-related cognitive decline via the concept of cognitive control, which

Manard, Carabin, Jaspar, and Collette (2014) describe as being “required to adjust and flexibly

guide people’s behavior in changing environmental circumstances, especially in situations where

distracting information or a predominant response tendency must be ignored in order to

successfully act in a goal-oriented manner.” Therefore, cognitive control includes aspects of

executive and attentional control, goal maintenance, and response selection and inhibition. A

thorough review of cognitive control theories and models is beyond the scope of this document.

Briefly, information processing theory, among the earliest cognitive control theories proposed,

contended automaticity to be on a continuum based on processing speed and interference effects

rather than an all-or-nothing phenomenon (Cohen, Dunbar, & McClelland, 1990). Conflict

monitoring theory subsequently built on information processing theory by theorizing that

cognitive control processes are initiated when a monitoring system in the anterior cingulate

cortex detects conflicts in information processing (Botvinick, Braver, Barch, Carter, & Cohen,

48

2001). Later work also suggested the dorsolateral prefrontal cortex to have a special role in

allowing one to actively maintain task-relevant memory representations, referred to as “context

processing,” during times of great interference (Kane & Engel, 2002).

It has been theorized that difficulties with context processing play a role in age-related

cognitive decline, an idea which became known as the “context processing theory of cognitive

aging” (Braver et al., 2001; Braver, Satpute, Rush, Racine, & Barch, 2005). Results of one

study support this theory over the role of processing speed or inhibition deficit alone in

explaining age-related cognitive decline, although more research is needed to clarify the relative

utility of context processing theory of cognitive aging (Rush, Barch, & Braver, 2006).

Within the context processing theory of cognitive aging, context processing has been

divided into reactive control and proactive control (Braver et al., 2005; Braver, 2012). Reactive

control, which is sometimes referred to as interference resolution, initiates temporary reminders

of context information after a probe stimulus (Bélanger, Belleville, & Gauthier, 2010; Braver et

al., 2005). In contrast, proactive control, which is sometimes referred to as goal maintenance,

actively updates and maintains context information and communicates with other attention

systems to prepare for an upcoming probe stimulus (Bélanger et al., 2010; Braver et al., 2005).

For example, a person can use proactive control to stop at the bank on her way home from work

by keeping her goal of stopping at the bank in mind as she approaches the intersection in which

she must make a left turn to go to the bank. In contrast, a person relying on reactive control

would not be keeping the goal of stopping at the bank in her mind as she approaches the

intersection, but the sight of the intersection may trigger a reminder of her goal of stopping at the

bank.

49

Proactive control processes appear to decline with age more so than reactive control

processes (Bélanger et al., 2010; Braver et al., 2005). Preliminary evidence supports the

hypothesis that successful use of proactive control processes in the older population is partly

dependent upon amount of available cognitive resources, consisting of processing speed and

fluid intelligence (Manard et al., 2014). Indeed, older adults are more vulnerable to interference

effects on the Stroop test, which is believed to be partially attributable to slowed processing

speed, as well as inhibitory deficits (Bugg, DeLosh, Davalos, & Davis, 2007; Troyer, Leach, &

Strauss, 2007). Therefore, although the context processing theory of cognitive aging involves

many cognitive components, processing speed clearly remains a crucial concept in this account.

Hypothesized Origins of Processing Speed Decline in the Elderly

Chronic inflammation. Aging is associated with low grade chronic inflammation

(Michaud et al., 2013). Within the aging process, chronic inflammation is believed to be

primarily caused by changes in the immune system and increased secretion of adipokines

(Michaud et al., 2013). This state of low grade chronic inflammation may be detrimental for the

same reasons explained in regard to obesity-related chronic inflammation.

Neurobiological factors. The brain’s morphology, activation patterns, and metabolism

change throughout the lifespan. During the aging process, reduced total brain volume, reduced

total gray matter volume, reduced total white matter volume, and increased volume of ventricles

occur (Fjell & Walhovd, 2010). It has been theorized that reduced grey matter is primarily due

to factors such as shrinkage of neurons and synaptic spines, and factors such as resultant widened

synaptic gaps may impair neural communication and contribute to slower processing speed (Fjell

& Walhovd, 2010). In regard to white matter, it has been hypothesized that breakdown of

50

myelin and/or shortened internodes due to remyelination result in slowed neural communication

and, therefore, reduced processing speed (e.g., Lu et al., 2013; Peters, 2009).

Although changes in total brain volume are characteristic of aging, brain matter volume

does not decrease uniformly across brain regions. The volume of frontal cortex (particularly in

the lateral prefrontal and orbitofrontal cortices), temporal cortex, thalamus, putamen, caudate

nucleus, nucleus accumbens, cerebellum, and hippocampus appear to be especially affected by

age (Eckert, 2011; Fjell & Walhovd, 2010; Raz et al., 2003; Van Der Werf et al., 2001). Smaller

volume and/or magnitude of atrophy over time of many of these areas negatively correlate with

processing speed, leading to the proposal of region-specific neurobiological explanations for age-

related processing speed decline (e.g., Nadkarni et al., 2014; Van Der Werf et al., 2001). For

example, one explanation focuses on specific areas of age-related grey matter reductions that

predict poor processing speed, including the bilateral inferior frontal and superior parietal

regions and the lingual and left superior frontal gyri (Chee et al., 2009). Other accounts focus on

the role of declines in white matter integrity, and, more specifically, volume of white matter

hyperintensities and myelin breakdown (Borghesani, Madhyastha, Aylward, & Reiter, 2014; Lu

et al., 2013; Papp et al., 2014). Specific areas of these declines in white matter integrity have

been found to be particularly important. For example, it has been found that age-related myelin

breakdown is more likely to occur in areas of the brain that form myelin later in life, including

the frontal lobe (especially in the prefrontal region; Lu et al., 2013). Myelin breakdown in these

late-myelinating areas have been found to mediate the relationship between age and processing

speed slowing, whereas early-myelinating brain areas have not (Lu et al., 2013). These findings

are consistent with accounts of age-related processing speed decline that focus on the role of the

prefrontal cortex, or, more broadly, the frontal lobe (e.g., Eckert, 2011). Despite degradation in

51

the prefrontal area of the brain, research has shown increased activity in the prefrontal cortex in

the elderly during a variety of cognitive tasks, including attention/executive functioning and

motor planning, compared with younger adults (Berchicci, Lucci, Pesce, Spinelli, & Di Russo,

2012; Ohsugi, Ohgi, Shigemori, & Schneider, 2013). Increased activity in the prefrontal cortex,

however, is actually associated with worse ability in many domains, including processing speed

(e.g., Woodward, Duffy, & Karbasforoushan, 2013). It has been theorized the increased activity

in the prefrontal cortex of the elderly during these tasks may reflect the brain’s attempt to

compensate for adverse age-related brain changes or may be the result of the aging brain’s

difficulty in recruiting specialized brain regions (Cabeza & Dennis, 2013).

Age-related change in cellular functioning and brain glucose metabolism are thought to

have adverse effects on cognition. On a cellular level, oxidative stress within mitochondria,

which is believed to occur with age, may lead to the overproduction of reactive oxygen species,

which, in turn, may cause cell damage and death (Bergamini, Gambetti, Dondi, & Cervellati,

2004; Yin, Boveris, & Cadenas, 2014). In regard to brain glucose metabolism, declines with age

can be seen in many areas of the frontal cortex, as well as in the caudate nuclei and superior

lateral temporal cortex (Chételat et al., 2013). Decreased glucose utilization, which is closely

liked to regulation of neuronal death, is thought to reflect cognitive impairments (e.g., Gage,

Kelly, & Bjorklund, 1984; Mergenthaler, Lindauer, Dienel, & Meisel, 2013).

Changes in activity of a number of other brain chemicals are also seen in aging, and

many of these changes are thought to have adverse effects on cognition. For example, older

adults evidence reduced dopamine metabolism due to factors such as reduced dopamine

transporters in the striatum and reduced binding potential in the striatum and frontal cortex, and

impaired dopamine activity has been linked to declines in glucose metabolism in the frontal and

52

temporal cortices, anterior cingulated gyrus, and caudate regardless of age (Suhara et al., 1991;

Volkow et al., 1996, 2000). Based on these findings, it has been suggested that age-related

changes in dopamine metabolism may play a contributory role in decreased glucose metabolism

in these brain areas, although it could also be hypothesized that decreased glucose metabolism

may contribute to reduced dopamine metabolism, as glucose is needed for the synthesis of

neurotransmitters (Mergenthaler et al., 2013; Volkow et al., 2000). The proportion of dopamine

relative to acetylcholine has been found to be particularly important in the processing of novel

stimuli by neurons in the medial temporal lobe and prefrontal cortex (Eckart & Bunzeck, 2013).

Another endogenous chemical for which the elderly develop a lack of sensitivity toward and may

subsequently experience adverse cognition is the neuroprotective hormone, leptin (Al Hazzouri,

Stone, Haan, & Yaffe, 2013; Folch et al., 2012). Central leptin resistance may be a contributor

to cognitive decline in obesity, as stated earlier in this document, but it has also been found to be

linked to aging even after the effects of body weight are accounted for (Gabriely, Ma, Yang,

Rossetti, & Barzilai, 2002). Changes in activity of other brain chemicals seen in aging and

thought to contribute to processes that impair cognition include thyroid stimulating hormone,

estrogen, and androgen (Akishita, 2004; Moon et al., 2004; Peeters, 2008).

Medical events and conditions. Many medical events and conditions that are more

common among the obese also have increased prevalence among the elderly; some examples

include T2DM, hypertension, hypercholesterolemia, heart failure, ischemic stroke, vascular

dementia, and many cancers (Centers for Disease Control and Prevention, 2014; Copeland et al.,

1999; Moon et al., 2013; National Heart Lung and Blood Institute, 2014). Ways in which these

events and conditions may contribute to cognitive decline have already been elaborated in this

document with the exception of cancer, which will be addressed in the medical treatments

53

section below. Furthermore, it should be emphasized here that vascular risk factors (e.g.,

hypertension, hypercholesterolemia, T2DM, cerebrovascular diseases) have been implicated in

the progression of MCI to dementia, both of which are more prevelant in the elderly and the

latter for which midlife obesity has been documented as a risk factor (Copeland et al., 1999; Xu

et al., 2011). That being said, there are certainly medical events and conditions that are not

associated with obesity but which could have an adverse impact on the cognition of affected

older individuals, including those of hypotension and type 1 diabetes mellitus (Frier, 2011;

Hebert et al., 2004).

Medical treatments with potential adverse cognitive effects. Elderly individuals have

an increased risk for medication-induced confusion, which is thought to be due to

pharmacokinetics and pharmacodynamics that have been altered as a result of the normal aging

process or age-related diseases (Flaherty, 1998). Risk increases signficantly with the number of

medications taken (Flaherty, 1998). Anticholinergic medications in particular often have adverse

consequences for cognition in older individuals (Campbell et al., 2009). Use of these

medications is prevalent among the elderly, with the most common types including bladder

antimuscarinics, first-generation antihistamines, and tricyclic antidepressants (Gray et al., 2015).

Acutely, anticholinergic medication use is associated with confusion and slowed information

processing speed (Low, Anstey, & Sachdev, 2008; Nebes, Pollock, Halligan, Houck, & Saxton,

2011). Long-term anticholinergic medication use also seems to be problematic, as recent

longitudinal studies have linked persistent use with faster decline in episodic memory and higher

cumulative burden with increased risk for dementia (Gray & Hanlon, 2016; Papenberg et al.,

2017).

54

In addition, certain chemotherapy and radiation therapy treatments for cancer have been

linked to decline in performance across measures of processing speed and executive functioning

in longitudinal studies, as well as self-reported increase in severity of memory and concentration

problems (Avila et al., 2015; Kohli et al., 2007). There is, however, little evidence to suggest

that many types of cancer are associated with cognitive impairment prior to systemic

chemotherapy or radiation therapy localized to the brain (Anstey, Sargent-Cox, Cherbuin, &

Sachdev, 2015; Avila et al., 2015; Mandelblatt et al., 2014).

Medical treatments with potential cognitive benefits. Among individuals with

hypertension, current and past use of antihypertensive medications have each been associated

with better cognitive performance than among those who have never taken antihypertensive

medications or who are not currently taking antihypertensive medication (Farmer et al., 1990).

These medications may protect cognition by avoiding vascular damage to the brain secondary to

hypertension (Hajjar et al., 2005). Additionally, it was discovered that certain antihypertensive

medications protect against cognitive decline to a greater extent than control of blood pressure

alone; these include antihypertensive medications that cross the blood brain barrier and impact

the renin-angiotensin-aldosterone system (RAAS) or intracellular calcium metabolism (Hajjar et

al., 2005). In particular, angiotensin-converting-enzyme (ACE) inhibitors are believed to be the

most efficacious class of hypertensive medication in preventing cognitive decline (Gard, 2010).

Comparatively less is known about the protective effects of anti-diabetic medication use

on cognition. Anti-diabetic medication is generally believed to improve cognition or protect

against cognitive decline; however, findings are mixed, and improved glycemic control does not

consistently predict quality of cognition (Alagiakrishnan, Sankaralingam, Ghosh, Mereu, &

Senior, 2013; Gradman, Laws, Thompson, & Reaven, 1993; Hewer, Mussell, Rist, Kulzer, &

55

Bergis, 2003). Furthermore, anti-diabetic medication usage has not been reliably shown to

protect against development of MCI or AD, conditions for which diabetes is a risk factor

(Alagiakrishnan et al., 2013).

In regard to hypothyroidism, adverse cognitive effects are generally believed to be

reversible with proper medical treatment (Samuels, 2014). Moreover, individuals with long-term

use of thyroid medication do not show cogntive abilities that are significantly different than those

without a thyroid disorder (Kramer, von Mühlen, Kritz-Silverstein, & Barrett-Connor, 2009).

Psychosocial factors. Several psychosocial factors may also be important in explaining

cognitive functioning in the elderly, including gender, race, education, mood, sleep, exercise,

smoking, diet, and nutrition factors. There is some evidence to suggest that elderly women may

experience faster rate of age-related cognitive decline than elderly men; however, other studies

have found no differences between genders (Barnes et al., 2003; Mortensen & Hogh, 2001;

Zhang, 2006). Furthermore, there is preliminary evidence to suggest that elderly men may

perform worse on tests of processing speed than elderly women (MacDonald, Hultsch, Strauss,

& Dixon, 2003). In regard to race, African Americans may be more vulnerable to faster rate of

cognitive decline, although this also has not been fully established (Lee et al., 2012; Obidi et al.,

2008).

Low level of education has been shown to be associated with lower IQ and greater

cognitive impairment among the elderly; however, education does not appear to have a

significant relationship with processing speed in this population (Kirton & Dotson, 2016;

Ritchie, Bates, Der, Starr, & Deary, 2013; Strandberg, Pitkala, Eerola, Tilvis, & Tienari, 2005;

Zahodne et al., 2011). Although many have theorized that high education level may have a

protective role against rate of cognitive decline, research findings are mixed (e.g., Muniz-

56

Terrera, Matthews, Dening, Huppert, & Brayne, 2009; Small, La Rue, Komo, Kaplan, &

Mandelkern, 1995; Zahodne et al., 2011). It may be the case that high education level is only

protective in certain populations or circumstances. This idea is supported by findings that

number of stressful life events predict faster rate of cognitive decline in those with a low level of

education, but not in those with a high level of education (Tschanz et al., 2013). Additionally, a

longitudinal study of community-dwelling elderly without dementia found that, among carriers

of two ApoE4 alleles but not among carriers of only one ApoE4 allele, the magnitude of

cognitive decline increased as years of education decreased even after adjustments for age,

gender, ethnicity, depression, diabetes, history of ischemic heart disease, and history of

cerebrovascular disease were made (Shadlen et al., 2005). More research is needed to

understand the potential protective role of education against cognitive decline.

Mood and sleep factors have also been hypothesized to impact cognitive functioning in

the elderly. Regarding mood, current depressed mood and long-term history of depressive

symptoms among the elderly have been associated with worse current cognitive functioning and

greater cognitive decline, respectively, and processing speed is among the cognitive domains

found to be affected by depression (Gualtieri, Johnson, & Benedict, 2006; Santos et al., 2013;

Singh-Manoux et al., 2010). Interestingly, self-reported frequency of feeling lonely has been

found to be associated with slower processing speed independent of depression, social networks,

and premorbid IQ (O’Luanaigh et al., 2012). Regarding sleep factors, shorter duration and worse

quality of sleep have been associated with both worse performance in several cognitive domains

(e.g., working memory, attention, abstract problem solving) and increased risk for cognitive

impairment among elderly without dementia (Keage et al., 2012; Miyata et al., 2013; Nebes,

Buysse, Halligan, Houck, & Monk, 2009; Potvin et al., 2012). However, preliminary evidence

57

suggests that processing speed does not appear to be impacted by sleep quality (Nebes et al.,

2009). Interestingly, daytime napping among the elderly may be associated with less cognitive

decline in this population (Keage et al., 2012).

Lack of exercise may also play a key role in cognitive decline in many elderly

individuals. Both aerobic exercise and a combination of aerobic and strength training forms of

exercise have been shown to improve processing speed task performance (Nouchi et al. 2014;

Tam, 2013). More specifically, these effects have typically been found in regard to two or three

weekly sessions of exercise for both short-term (e.g., 4 weeks) and long-term (e.g., 3 years)

periods of time (e.g., Nouchi et al. 2014; Rikli & Edwards, 1991; van de Rest et al., 2014). The

duration of exercise sessions in studies examining their impact on cognition is typically around

30 minutes, totaling a minimum of two sessions or 60 minutes of exercise per week (e.g., Nouchi

et al., 2014; Voss et al., 2013). It is not yet well established if less frequent exercise or shorter

exercise duration may also benefit processing speed. Compared with anaerobic exercise, aerobic

exercise appears to be more beneficial for executive control processes that are frontally-mediated

(e.g., switching between tasks, ability to disregard task-irrelevant stimuli, ability to discontinue a

planned action; Kramer et al., 1999). Combination training, which includes both aerobic and

anaerobic exercises, appears to be more beneficial than either type of exercise alone regarding

overall cognition (Colcombe & Kramer, 2003).

In regard to the present study, it was then hypothesized that the group of elderly

individuals who engaged in either a minimum of 60 minutes of aerobic exercise weekly or a

minimum of two sessions of strength training sessions weekly would exhibit higher scores on the

DSST, a traditional measure of processing speed, than those who did not exercise. Further, those

who engaged in only aerobic exercise were expected to perform better on the DSST than those

58

engaged in only strength training exercise, as the DSST has been shown to activate the frontal

lobe and involve elements of executive functioning (Davis & Pierson, 2012; Nakahachi et al.,

2008). Finally, it was hypothesized that elderly individuals who engaged in combination training

would perform better on the DSST compared to the aerobic only, strength training only, and no

exercise groups.

Other lifestyle factors, including smoking, alcohol use, and diet history, may also

contribute to cognitive impairment in the elderly. Current smoking has been associated with

deficits in a number of cognitive domains, including processing speed (Lo, Woodman, Pachana,

Byrne, & Sachdev, 2014). It could be hypothesized that smoking causes a decrease in lung

functioning that may lead to mild cerebral hypoxia or that inhaled toxins adversely impact the

brain. Smoking may also indirectly lead to cognitive dysfunction such that it is a risk factor for a

number of conditions that are thought to be causally related to cognitive decline, including CHD

(e.g., Wilson et al., 1998). Alcohol use may also be a salient lifestyle factor that contributes to

cognitive impairment in this population. Approximately 8.9% of the elderly meet criteria for

alcohol abuse, which is associated with dementia in this population, as well as impairments in

recent visuospatial memory, verbal memory, and executive functioning (Hudetz & Warltier,

2007; Thomas & Rockwood, 2001). Length of alcohol abuse history is also relevant for

cognition outcomes. For example, years of alcohol abuse has been inversely associated with

psychomotor speed and short-term memory (DeFranco, Tarbox, & McLaughlin, 1985).

Although heavy alcohol use is known to impair cognition, some evidence suggests that a history

of moderate alcohol use may have a protective influence on cognition (Lo et al., 2014). Given

the available variables in the NHANES, it was hypothesized that those who endorsed a likely

problem with lifetime history of excess alcohol consumption (i.e., answered yes to “Was there

59

ever a time or times in your life when you drank 5 or more drinks of any kind of alcoholic

beverage almost every day?”) would exhibit worse processing speed task performance.

Likewise, it was hypothesized that those who endorsed a likely problem with past-year excess

alcohol consumption (i.e., endorsed consuming 5 or more drinks on half or more of the days

during the past year) would exhibit worse processing speed. Finally, it was hypothesized that

current smoking status would be associated with worse processing speed. It should be noted that

those with current or history of chronic use of other drugs, such as cocaine and heroin, have also

been found to have impaired cognition in various domains (Lundqvist, 2005). Unfortunately,

NHANES did not assess for these factors among the elderly.

Regarding diet, the most consistent findings in this area involve the association between a

Mediterranean-type diet and slower cognitive decline and reduced risk for AD (Cheung, Ho,

Chan, Sea, & Woo, 2014; Kuczmarski, Allegro, & Stave, 2014). Characteristics of this type of

diet include high consumption of vegetables and fruit and low consumption of processed foods,

sugar, and saturated fat. This pattern may have a beneficial impact on cognition for a number of

reasons. For example, it has also been hypothesized that regular consumption of certain fruits

and vegetables that contain polyphenolic compounds with antioxidant, anti-inflammatory, and

neuroprotective properties may protect against cognitive decline (Darvesh, Carroll, Bishayee,

Geldenhuys, & Van der Schyf, 2010). It has also been hypothesized that high levels of dietary

saturated fat may make clearing beta-amyloid proteins from the brain more difficult (Hanson et

al., 2013). Other ways in which diets high in fat may lead to cognitive impairment (e.g., central

inflammation, impaired insulin signaling) have already been discussed in this document.

A number of vitamins, nutrients, and related chemicals have also found to be important

for healthy cognitive functioning among the elderly, including homocysteine, folate, vitamin B12,

60

iron, and vitamins C, D, and E. Although these factors are being discussed under the

psychosocial/diet section of this document, it should be noted here that abnormal levels of these

chemicals can occur not only as a result of diet, but also as a result of certain neurologic,

psychiatric, hematologic, or genetic reasons (e.g., Hector & Burton, 1988). Homocysteine,

folate, and vitamins B12 will be addressed first due to their interdependent relationship. Namely,

vitamin B12, vitamin B6, and folate are thought to be needed to convert homocysteine to an

amino acid called methionine, which is needed to produce a variety of other chemicals (Leishear,

Ferrucci, et al., 2012). Therefore, if vitamin B12, vitamin B6, and/or folate levels are depleted,

homocysteine levels may increase (Leishear, Ferrucci, et al., 2012). Likewise, increasing folate

and B12 via supplements has been shown to decrease homocysteine levels (Graham et al., 1997).

Levels of homocysteine, an amino acid, tend to increase with age and are typically

greater in men (Feng et al., 2013). High homocysteine level has been linked to reduced

information processing speed and visuoconstructional ability independent of folate levels,

although some have argued that high homocysteine levels are only associated with worse

cognition in those with low folate levels (Feng et al., 2013; Feng, Ng, Chuah, Niti, & Kua, 2006;

Vidal, Dufouil, Ducros, & Tzourio, 2008). Hyperhomocysteinemia, the condition of having high

homocysteine levels, has also been linked to lower global cognition (Di Bonito et al., 2007). The

pathways by which high homocysteine levels may negatively impact cognition are not well-

established; however, it is thought that high levels of homocysteine damage the vascular

endothelium (i.e., interior lining of blood and lymphatic vessels) and increase risk for thrombosis

and atherosclerosis (Arnesen et al., 1995; Graham et al., 1997). Therefore,

hyperhomocysteinemia could be considered an indicator for endothelial dysfunction (Meigs et

al., 2001). These factors could, in turn, lead to cognitive impairment. It has also been proposed

61

that direct neurotoxic effects might explain the link between high homocysteine levels and

cognitive impairment, or that insulin resistance might mediate this relationship (Di Bonito et al.,

2007). Indeed, T2DM, the hallmarks of which are insulin resistance and hyperglycemia, is

among the strongest predictors of peripheral arterial disease (PAD), a vascular disease marked by

atherosclerosis (American Diabetes Association, 2003). Interestingly, lowering homocysteine

levels (e.g., via folate or B12 supplements) has not been reliably shown to slow cognitive decline

(Vogel, Dali-Youcef, Kaltenbach, & Andrès, 2009). This may suggest a number of possibilities,

including hyperhomocysteinemia being an indicator, but not necessarily a cause, of endothelial

dysfunction. It may also suggest an inability to recover from previous damage, the strong

importance of other factors (e.g., diet, exercise) that likely remain unchanged, or limitations in

research methodology.

Low folate levels have been associated with worse episodic memory and language

ability, but its association with processing speed is less clear (Feng et al., 2006; Jelicic, Jonker, &

Deeg, 2001). Findings on the ability of folic acid supplementation to improve processing speed

have been mixed (Durga et al., 2007; Malouf & Evans, 2008; Vogel et al., 2009). In addition,

some have suggested that low vitamin B12 levels impact the relationship between folate levels

and cognition, but findings on the direction of this effect have been contradictory (Doets et al.,

2014; Moore et al., 2014).

Vitamin B12 deficiency (typically defined within the range of ≤ 145–260 pmol/L) is

associated with a variety of cognitive impairments (e.g., memory problems, slowed thinking,

confusion), and it has been hypothesized that cognitive impairments may also occur among

elderly with a low-normal range of this vitamin (Hector & Burton, 1988; Leishear, Ferrucci, et

al., 2012; Smith & Refsum, 2009). One recent study also found lower vitamin B12 plasma level

62

to be associated with greater processing speed decline over a period of six years (Leishear,

2011). Furthermore, lower vitamin B12 plasma levels are related to poor sensory and motor

peripheral nerve function (Leishear, Boudreau, et al., 2012).

Iron deficiency has also been associated with lower scores on a screening instrument for

cognitive impairment among a sample of elderly hospital patients, although it is not known if

iron deficiency is specifically associated with processing speed in the elderly (Yavuz et al.,

2012). Among those with iron-deficiency, anemia, a condition that is sometimes caused by

deficiencies in iron, folate, or vitamin B12, has been linked to worse global cognition (Murray-

Kolb & Beard, 2007). Furthermore, severity of anemia has been found to adversely impact

processing speed to a greater extent than other cognitive domains (Murray-Kolb & Beard, 2007).

Low levels of vitamin D are associated with poor processing speed, with this relationship

being most salient when 25-hydroxyvitamin D level is below 35 nmol/L (Lee et al., 2009). It

may also be possible that vitamin E and vitamin C also impact processing speed. For example,

vitamin E intake was shown to predict better performance on a battery of cognitive tests that

included a processing speed measure (Morris, Evans, Bienias, Tangney, & Wilson, 2002).

Results of another study suggest that vitamin C deficiency, but not necessarily vitamin C level

itself, may be associated with greater age-related cognitive decline in the general population

(Harrison, 2012).

In regard to the present study, it was hypothesized that individuals with low vitamin D

level would perform worse on the processing speed task. Low vitamin D level was operationally

defined as 25-hydroxyvitamin D level < 35 nmol/L based on the aforementioned work of Lee

and colleagues (2009). Note that the current Merck Manual, Professional Edition states that

target 25-hydroxyvitamin D levels are approximately 50 to 60 nmol/L for maximal bone health

63

but does not provide a specific definition of vitamin D deficiency (Johnson, 2016b). Concerning

homocysteine and folate, high homocysteine levels were expected to be related to worse

performance independent of folate levels. Folate levels, however, were not expected to be

significantly correlated with processing speed independent of homocysteine levels. It was

difficult to make further hypotheses with confidence given the complicated nature of

homocysteine, folate, vitamin B12, iron, and vitamins C, D, and E, but a number of exploratory

hypotheses were made in the hypothesis section of this document. For the purposes of these

exploratory analyses, folate deficiency was defined as serum folate < 1.4 ng/mL, as suggested by

the 1999–2000 CDC Laboratory Procedure Manual in regard to interpretation of results from the

Bio-Rad Laboratories’ “Quantaphase II Folate/Vitamin B12” Radioassay Kit. Iron deficiency

was operationally defined as a serum ferritin level of < 22µg/L, as this has been found to be the

optimal serum ferritin cutoff point for the diagnosis of iron deficiency within community-

dwelling older adults (Choi et al., 2005). Vitamin B12 deficiency was defined as < 145 pmol/L

and low-normal vitamin B12 level as 145–260 pmol/L, as per the current Merck Manual,

Professional Edition (Johnson, 2016a). Vitamin E deficiency was defined as alpha-tocopherol <

11.6 µmol/L or, for adults with hyperlipidemia, a ratio of serum alpha-tocopherol to total lipids

of < 0.8 mg/g (Johnson, 2016c). Alpha-tocopherol was chosen rather than gamma-tocopherol,

another form of vitamin E with values also available in the NHANES database, because research

has shown low alpha-tocopherol to be associated with cognitive impairment and dementia in

community-dwelling older adults (Cherubini et al., 2005). Also note that, although both serum

and red cell folate assays were performed as part of the NHANES, all analyses in the present

study used the serum assay results due to a number of recent findings indicating its relative

64

utility in determining folate deficiency (De Bruyn, Gulbis, & Cotton, 2014; Farrell, Kirsch, &

Herrmann, 2013; Galloway & Rushworth, 2003).

Other correlates. Usual slow gait speed and gait speed variability when fast-walking

have been linked to worse global cognition and MCI, respectively (Beauchet, Allali, Launay,

Herrmann, & Annweiler, 2013; Deshpande, Metter, Bandinelli, Guralnik, & Ferrucci, 2009).

Such gait changes typically precede onset of cognitive symptoms in those who go on to develop

MCI or dementia (Buracchio, Dodge, Howieson, Wasserman, & Kaye, 2010). These gait

patterns are not causally linked to cognitive dysfunction in these individuals, but they are

believed to be a result of the same underlying pathological process. Other factors that may slow

an individual’s gait speed include psychomotor slowing secondary to depression and relative

difficulty walking secondary to obesity, injury, or disability (Atkinson et al., 2007). In regard to

the present study, it was hypothesized that gait speed would remain positively correlated with

processing speed task performance after controlling for BMI.

Processing Speed Assessment and the Digit-Symbol Substitution Test

Processing speed can generally be defined as the rate at which information can be

processed (Lichtenberger & Kaufman, 2009). However, it is difficult to obtain a pure measure of

this construct due to a number of confounds in testing. Processing of visual information, for

example, depends on a number of other factors, including the integrity of one’s vision, ability to

perceive the visual information properly as a whole, and attention capacity. When the processing

speed measure requires a written response, motor control and speed may confound results.

Furthermore, certain processing speed tests place more attention on visual scanning and attention

to detail than others, and some researchers have distinguished processing speed tests based on

such considerations. For example, although both require visual processing in a divided-attention

65

format, Kuo et al. (2006) refers to the Useful Field of View (UFOV) as a measure of visual

processing and the DSST as a measure of motor processing speed, seemingly due to the

relatively high visual scanning demands of the former and motor demands of the latter.

However, labels applied by researchers to the same processing speed test are often different; for

example, the DSST, which is historically among the most used measures of processing speed,

has been referred to as a measure of motor processing speed, information processing speed,

visuo-attentional psychomotor speed, psychomotor functioning, and general cognitive ability by

various researchers (González-Blanch et al., 2011; Kuo et al., 2006; Lafont et al., 2010; Lee et

al., 2012; Pieters, Maas, & Hulstijn, 2004). Some researchers also decline to label the DSST,

instead providing a description of domains it has been used to assess (e.g., Nakahachi et al.,

2008). This raises the question of what the DSST actually measures, the answer to which is

especially important given the vast amount of research using this measure.

The DSST is a test in which numbers are paired with symbols in a key at the top of a

page, and the participant has 120 seconds to copy symbols to corresponding numbers as quickly

and accurately as possible. Individuals must be able to complete the sample items and have no

significant visual or motor limitations in order to be administered the test. Scores are based on

the number of symbols that were copied correctly. Research indicates that Symbol Copy, an

optional measure on the Wechsler Adult Intelligence Scale, Third Edition (WAIS-III) and the

Wechsler Adult Intelligence Scale-Revised as a Neuropsychological Instrument (WAIS-R-NI)

that simply requires subjects to copy symbols from one box to another box below it, accounts for

35–50% of the DSST score variance (Joy et al., 2003; Joy, Fein, Kaplan, & Freedman, 2000;

Kreiner & Ryan, 2001). Symbol Copy has generally been considered to be a measure of “motor

skill” or “speed” (Joy et al., 2003; Kreiner & Ryan, 2001). Joy and colleagues (2003) further

66

broke down this variable by assessing for graphomotor speed in a name-writing task. They

concluded that the Symbol Copy variable likely contains both graphomotor and perceptual speed

components, with the graphomotor component accounting for approximately half of the DSST

score variance that is explained by Symbol Copy. Visual scanning efficiency is also very

important in the DSST, as Joy and colleagues (2003) found 34% of the DSST score variance to

be explained by performance on a visual scanning measure. Memory/learning, although still

statistically significant, appear to be less important in explaining DSST scores, accounting for

only 4–5% of the variance (Joy et al., 2000, 2003; Kreiner & Ryan, 2001). One study even

failed to find a statistically significant association between DSST performance and paired

associative learning in older adults (Stephens & Kaufman, 2009). These findings imply that

subjects may not benefit from memorizing number-symbol pairs on the DSST, which Joy and

colleagues (2003) attempt to explain by hypothesizing that accessing memory to retrieve a

symbol may actually take longer than looking at the key to retrieve the symbol, especially as the

number of memorized symbols increases. Recently, it has been proposed that the DSST also

taps executive functioning on the basis that DSST scores correlate with the Letter-Number

Sequencing subtest of the Delis-Kaplan Executive Functions System (D-KEFS) Trail-Making

Test to a greater extent than the other subtests (Davis & Pierson, 2012). Finally, it has been

hypothesized that fine motor speed and manipulative dexterity may also contribute to DSST

scores (Joy et al., 2003). Their roles have not been established, but gait speed has been

positively associated with DSST scores (Killane et al., 2014). To summarize, DSST scores

appear to primarily reflect graphomotor speed, perceptual speed, visual scanning ability, and

executive functioning, although other areas such as learning and memory may play a smaller, yet

67

still significant, role. It should be noted here that there are a number of DSSTs, including the

WAIS-III Digit Symbol—Coding subtest (CD) that is used in the NHANES.

The multifaceted nature of the DSST makes interpreting its scores with regard to one

specific cognitive domain difficult; however, this characteristic may also allow for its clinical

utility in screening for cognitive dysfunction arising from a variety of sources, and especially

those arising from generalized processes (González-Blanch et al., 2011; Joy, Kaplan, & Fein,

2004). For example, the DSST is particularly sensitive to cognitive impairment in multiple

sclerosis and schizophrenia, and is a predictor of stroke and heart disease independent of known

vascular risk factors (Elkins, Knopman, Yaffe, & Johnston, 2005). The latter suggests the

possible utility of DSST scores as an early indicator of poor health prognosis as a function of a

premorbid disease state. The DSST is also highly sensitive to aging and age-related changes in

brain morphology, and is useful in predicting incident disability, decline in basic and high levels

of functional capacity in daily activities, AD, and mortality (Iwasa et al., 2008, 2014; Mosley et

al., 2005; Rapp & Reischies, 2005; Salthouse, 1992).

Justification for the Present Study and Structure of NHANES Database

Obesity has been associated with cognitive impairment in nearly all cognitive domains.

The exact etiological processes that explain this relationship have not been fully elucidated. For

example, it is unclear whether the association between obesity and cognitive impairment is

independent or can be sufficiently explained by mediating variables (e.g., T2DM, hypertension).

In addition to focusing on medical diagnoses as mediating variables (as the majority of research

in this area does), it is important to consider the contribution of lower-level obesity-related

factors (e.g., low grade inflammation, increased circulation of adipocyte hormones, endothelial

dysfunction, insulin resistance, leptin resistance, hypoventilation) that may underlie a number of

68

obesity-related disorders, as well as be present among obese individuals without obesity-related

disorders. Furthermore, it may be the case that lower-level obesity-related factors that contribute

to cognitive decline have an increased effect if the individual is exposed to a greater magnitude

(e.g., more hormones released from adipose tissue) and/or over a greater time period (e.g.,

chronic low-grade inflammation in the chronically obese person). Therefore, it follows that

knowledge of a person’s weight history, which may serve as a proxy for a number of lower-level

obesity variables, may provide additional utility in predicting cognitive functioning than history

of medical diagnoses alone. There has been some evidence to support the notion that weight

history may impact cognition; however, this area of study is in its infancy, and its data have not

been adequately synthesized.

The present study adds to the body of scientific literature in this area by examining the

relationship between weight history (beginning in early adulthood) and current processing speed

among the young-old (i.e., ages 65–74). The period of early adulthood to age 65–74 greatly

extends the length of weight history in any published study on the relationship between weight

history and processing speed. The young-old population and the domain of processing speed

were chosen in the present study for a number of reasons, including: (a) the number of elderly

with significant history of obesity is growing in the United States; (b) processing speed may be

foundational to other cognitive processes and drive cognitive aging (Finkel et al., 2007;

Salthouse, 1996; Vance, 2009); (c) processing speed in the elderly is predictive of a number of

outcomes, including time to complete everyday tasks, life satisfaction, and longevity, to an equal

or greater extent than other cognitive domains (Enkvist et al., 2013; Owsley et al., 2002).

Furthermore, it should be noted here that the young-old population was selected for the

present study due to consideration of available NHANES variables (e.g., weight “ten years

69

ago”), consideration of age ranges studied in extant literature in this area (in order to inform this

study’s hypotheses), and consideration of the utility of limiting the study to a relatively small age

range due to the sensitivity of a number of the relationships of interest to age. Furthermore, it

was noted that the large sample size of NHANES allow for the narrowing of age range without

sacrificing the necessary statistical power for the present study’s analyses.

The 1999–2000 and 2001–2002 NHANES cycles were chosen for the present study due

to their inclusion of a processing speed measure, whereas subsequently completed cycles have

not. The variables of these cycles that are relevant to the present study are included in

Appendixes A–L and in the methods section of this document. The present study makes a

number of hypotheses regarding the relationships among NHANES variables and culminates

with creation of a model to explain poor processing speed among those with long-term obesity.

70

Research Hypotheses

Basic Relationships

The following hypotheses or groups of hypotheses involve basic relationships that would

have either replicated those previously reported in the literature or have been very similar to

those already reported in the literature:

(1) Age and DSST performance:

There would be a strong negative correlation between age and processing speed task

performance (DSST score).

(2) Age and obesity indexes:

(a) Overall indexes: There would be a weak, non-significant negative correlation between

age and BMI but a significant positive correlation between age and % total body fat

measured by DEXA.

(b) Central indexes: There would be no significant correlation between age and WC but a

significant positive correlation between age and % trunk fat measured by DEXA.

(3) Substance use and DSST performance:

(a) Those with probable lifetime history of excess alcohol consumption (i.e., answered

yes to “Was there ever a time or times in your life when you drank 5 or more drinks of

any kind of alcoholic beverage almost every day?”) would exhibit worse processing

speed task performance.

71

(b) Those with probable past-year excess alcohol consumption (i.e., endorsed consuming

5 or more drinks on half or more of the days during the past year) would exhibit worse

processing speed task performance.

(c) Those who currently smoke cigarettes would perform worse on the DSST than those

who do not currently smoke cigarettes.

(4) Exercise and DSST performance:

For the purposes of the following hypotheses, the following groups are defined:

Aerobic-only group: reported at least 60 minutes of moderate or vigorous aerobic

exercise per week but no instances of past-month strength training

Strength training-only group: reported at least 8 instances of past-month strength training

but no moderate or vigorous aerobic past-month exercise

Combination training group: reported at least 60 minutes of moderate or vigorous aerobic

exercise per week and at least 8 instances of past-month strength training

No exercise group: denied past-month vigorous exercise, moderate exercise, and strength

training

(a) The no exercise group would perform worse on the DSST than the aerobic-only,

strength training-only, and combination training groups.

(b) The aerobic-only group would perform better on the DSST than the strength training-

only group.

(c) The combination training group would perform better on the DSST compared to the

aerobic-only, strength training-only, and no exercise groups.

72

(5) Diabetes status and DSST performance:

For the purposes of these hypotheses, the following groups were operationally defined:

Diabetes group: Those with diagnosed history of diabetes or with fasting glucose or

HbA1c levels in the diabetic range.

Prediabetes group: Those without diagnosed history of diabetes who are not currently

taking antidiabetic medication and have fasting glucose or HbA1c levels in the prediabetic

range.

No diabetes group: Those without diagnosed history of diabetes who are not currently

taking antidiabetic medication and have fasting glucose and HbA1c levels in the normal

range.

(a) The diabetes group would perform worse on the DSST than the prediabetes and no

diabetes groups.

(b) The prediabetes group would perform worse on the DSST than the no diabetes group.

Main Hypotheses

The following hypotheses or groups of hypotheses involve relationships that have either

not been investigated to-date or have not been reported in the literature with consistency:

(6) Nutritional biochemistry and DSST performance:

(a) Homocysteine levels would be negatively correlated with processing speed

independent of serum folate levels.

73

(b) Serum folate levels would not be significantly correlated with processing speed

independent of homocysteine levels.

(c) Those with low 25-hydroxyvitamin D levels (< 35 nmol/l) would perform worse on

the processing speed measure than those with healthy vitamin D levels.

Exploratory (less support from the literature):

(d) Those with iron deficiency would perform worse on the processing speed task than

those without iron deficiency.

(e) Vitamin E levels would be positively correlated with the processing speed measure

performance.

(f) Total number of nutrient deficiencies (i.e., vitamin B12, vitamin D, vitamin E, folate,

iron) would be negatively correlated with processing speed task performance.

(7) Obesity indexes and DSST performance:

(a) DSST scores would not be significantly correlated with the overall obesity indexes

(i.e., BMI, % DEXA total body fat).

(b) DSST scores would be significantly negatively correlated with the central obesity

indexes (i.e. WC, WTR, WHtR, and % DEXA trunk fat).

(c) Among those who are centrally obese (defined here as WC ≥ 102 cm for men and 88

cm for women):

(1) Individuals meeting BMI obesity criteria (i.e., BMI ≥ 30 kg/m2) would

perform worse on the DSST.

(2) Individuals meeting DEXA obesity criteria (fat mass ≥ 25% for men and ≥

35% for women) would perform worse on the DSST.

74

Exploratory (less support from the literature):

(d) The best combination of obesity indexes to predict DSST score is unknown; however,

it was hypothesized that the combination would involve % DEXA trunk fat.

(8) Weight trajectory and DSST performance:

The groups in Hypotheses 8a-8c are defined in Table 1. Note that individuals who were

underweight (i.e., BMI < 18.5 kg/m²) at one or more of the three timepoints were

excluded from these analyses.

(a) Groups 1 and 2 would perform the worst on the DSST.

(b) Group 8 would perform the best on the DSST.

(c) Groups 4, 6, and 7 would perform better than Groups 1 and 2.

(d) Percent of weight gained between early adulthood (age 25) and late middle-age (10

years ago, which would be ages 55–64 in this study) would be associated with worse

performance on the DSST.

(e) Absolute value of percent of weight change over the previous 10-year period would

be associated with worse performance on the DSST.

(f) Individuals endorsing significant past-year unintentional weight loss would perform

worse on the DSST than those who achieved significant past-year weight loss that

was intentional and those who were past-year weight-stable.

75

Table 1

Hypotheses 8a–8c Groups

Obese at age 25 Obese 10 years ago Obese currently

Group 1 Y Y Y

Group 2 Y Y N

Group 3 Y N Y

Group 4 Y N N

Group 5 N Y Y

Group 6 N Y N

Group 7 N N Y

Group 8 N N N

76

(9) Peripheral inflammation and DSST performance:

There would be a significant negative correlation between a marker of peripheral

inflammation (i.e., CRP level) and processing speed performance, which would remain

significant after controlling for health condition and smoking confounds.

(10) Medical history and DSST performance:

(a) Those who endorsed one or more of the following conditions would perform worse on

the DSST than those who did not endorse any: problems of blood flow to the heart

typically leading to or involving heart muscle damage (i.e., congestive heart failure,

coronary heart disease, heart attack), stroke, diabetes, weak/failing kidneys, chronic

respiratory disease (i.e., chronic bronchitis or emphysema), and thyroid disease (if

untreated).

(b) Duration of respiratory disease would be negatively correlated with processing speed

task performance.

(11) Dyslipidemia and DSST performance:

Those with dyslipidemia (i.e., unhealthy levels of total cholesterol, HDL cholesterol,

LDL cholesterol, and/or triglycerides) would demonstrate worse processing speed task

performance.

(12) Framingham risk scores and DSST performance:

77

(a) Risk of first stroke, as indicated by score on the modified version of the FSRP, would

be inversely associated with processing speed. Note that the modified version of the

FSRP, as defined previously in this document, includes age, systolic blood pressure,

antihypertensive medication, diabetes, smoking status, CHD, and congestive heart failure.

(b) Risk of onset of CVD, as indicated by FRS for general CVD scores, would be

inversely associated with processing speed task performance.

(13) Gait speed and DSST performance:

There would be a significant positive correlation between speed of usual gait and

processing speed task performance, which would remain significant after controlling for

BMI.

Final Model

Finally, the present study used structural equation modeling (SEM) to develop a model of

the relationship between weight trajectory and DSST performance. Specifications of the model

depended on observed relationships between study variables, which were hypothesized to be

consistent with theoretical relationships addressed in the literature review and tested by many of

the aforementioned hypotheses. A theoretical mapping of relationships based on the literature

review is shown in Figure 1, although this mapping was modified extensively based on the

results of tests of relevant hypotheses and practical limitations. In regard to Figure 1, it should

be noted that potential causes of processing speed decline that underlie a number of variables on

this map (e.g., oxidative stress, endothelial dysfunction, central inflammation) can be causes of a

number of factors, while also contributing to the development or persistence of these factors.

78

Due to their complicated interactions and the lack of biological markers for many of these

variables in the NHANES database, they were omitted from the theoretical mapping of

relationships in Figure 1, although they did influence reasoning behind many of the hypothesized

relationships in the theoretical mapping.

79

Figure 1. Theoretical model of relationship between persistent obesity and DSST score. Error

terms are not shown.

Persistent obesity

Insulin resistance

Hyperglycemia

Hypertension hx

Depression

Peripheral inflammation

Physical

inactivity hx

Current

physical

inactivity

Brain oxygen flow disruption/ insufficiency

indicators:

Atherosclerosis

Problems of blood flow to the heart typically

leading to or involving heart muscle damage:

CHD, heart attack, CHF

Ischemic stroke

Chronic hypoventilation during sleep or awake

periods: e.g., OSAS, OHS

Chronic respiratory disease: chronic asthma

chronic bronchitis, emphysema

DSST score

Dyslipidemia

Current diet/nutrients

Diet/nutrient hx

80

Method

Participants

The present study’s sample consists of the young-old men and women aged 65–74 (n =

1,381) who completed the National Health and Nutrition Examination Survey (NHANES) during

the years 1999–2002. NHANES participants were chosen to be representative of the United

States population; however, individuals aged 12–19 and over 70 years, as well as certain

racial/ethnic/socioeconomic groups (i.e., Non-Hispanic Black, Mexican American, and low-

income Caucasian individuals), were oversampled due to being “of particular public health

interest” (NHANES Plan and Operations, 1999–2010). NHANES participants were drawn from

urban, suburban, and rural locations across the United States, and are of diverse education

backgrounds. The aforementioned number of participants aged 65–74 who completed the

NHANES during the years 1999–2002 was found to be sufficient for the present study’s

secondary analyses based on an a priori power analysis using an expected medium effect size (f

= .25) and power = 0.8, which yielded a minimum total sample size for a one-way analysis of

variance (ANOVA) with eight groups to be 240 for a 95% confidence interval. See Table 2 for

participant characteristics.

81

Table 2

Participant Characteristics

Demographic variable All participants (n = 1,381)a

Age (years) 69.3 (2.9); 65–74

BMI (kg/m²) 28.5 (5.8); 10.8-56.3

BMI category

Underweight 1.2

Normal weight 26.1

Overweight 38.5

Obese 34.2

Gender

Male 50.5

Female 49.5

Race/Ethnicity

Non-Hispanic White 50.6

Non-Hispanic Black 19.0

Mexican American 24.0

Other Hispanic 4.5

Other Race–Including Multiracial 1.9

Education

Less Than 9th Grade 28.2

9-11th Grade (Includes 12th grade with no diploma) 17.8

High School Grad/GED or Equivalent 21.8

Some College or AA degree 18.3

College Graduate or above 13.8

Marital status

Married/Living with Partner 64.2

Divorced/Separated 12.4

Widowed 18.9

Never married 4.3

Employment status

Not employed 81.1

Employed full-time 10.2

Employed part-time 8.7

Annual household income

< $10,000 12.8

$10,000-$19,999 23.8

$20,000-$34,999 24.8

$35,000-$54,999 17.5

$55,000-$75,000 8.4

> $75,000 10.9 Note. Data are presented as percentages for all categorical variables, and as M (SD); range for Age and BMI. an = 1,381 except for BMI and BMI category (n = 1,373), education (n = 1,374), marital status (n = 1,317),

employment status (n = 1,379), and annual household income (n = 1,133).

82

Measures

It should be noted here that, due to the length and number of interviews in the NHANES,

only questions from the interviews that are theoretically relevant to the present study are

included in the appendices. These questions remain in the order in which they were asked. It

should also be noted here that the NHANES “questionnaires” in the appendices are actually

personal interviews conducted by NHANES staff or computerized questionnaires completed with

the assistance of NHANES staff. Questions were directed toward the sampled person, but a

proxy was used if the sampled person (abbreviated SP in the appendixes) needed assistance.

Socio-demographics. Socio-demographic information was collected in the Demographics

Questionnaire (Appendix A) and Occupation Questionnaire (Appendix B), which included

gender, age, race/ethnicity, education, marital status, number of people living in the household,

household income, and employment status.

Anthropometric measures. Current height and weight were measured with a stadiometer

and digital scale, respectively (NHANES Anthropometry Procedures Manual, 2000). Using the

Weight History Questionnaires, which is a personal interview conducted by NHANES staff,

participants were also asked to self-report their current weight and height, as well as recall their

weight at 1-year ago and 10-years ago, and weight and height at age 25 (Appendix C). Self-

reported highest and lowest adult weight (excluding times of pregnancy) and corresponding ages

were also collected. Height and weight allow body mass index (BMI) to be calculated at each of

these timepoints. Current waist circumference (WC) and current thigh circumference were

determined using measuring tape at a standard position, which allows current waist-thigh ratio

(WTR) to be calculated. Additionally, current WC and height allow current waist-height ratio

(WHtR) to be calculated.

83

Dual energy x-ray absorptiometry (DEXA) was conducted with a Hologic Densitometer

QDR4500A, allowing the following to be measured or estimated: total mass, fat mass, bone

mineral content (BMC), lean mass excluding BMC, lean mass including BMC, percent body fat,

and trunk fat (NHANES Body Composition Procedures Manual, 2000). DEXA has good

psychometric properties regarding measuring total fat and trunk fat. For example, total body fat

measured by DEXA shows strong correlations with total body fat measured by computerized

tomography (CT) and magnetic resonance imaging (MRI) -based T-1 mapping (Kullberg et al.,

2009). Furthermore, trunk fat measured by DEXA has been shown to strongly correlate with

trunk fat measured by CT scan among the elderly (r = 0.87 to 0.98 depending on gender and

race/ethnicity), indicating good convergent validity in this population (Snijder et al., 2002).

Wechsler Adult Intelligence Scale, Third Edition (WAIS-III) Digit Symbol—Coding

subtest. The WAIS-III Digit Symbol—Coding subtest (CD) is a digit symbol substitution test

(DSST). It is broadly intended to measure processing speed, and is one of two subtests of the

Processing Speed Index (PSI) of the WAIS-III. In the CD, numbers are paired with symbols in a

key at the top of the page, and the subject has 120 seconds to copy the symbols to corresponding

numbers below the key as quickly and accurately as possible. Individuals who cannot complete

the sample items or have significant visual or motor limitations are excluded from taking this

test. The CD score is the number of symbols that were copied correctly. There are 133 total

items, and therefore, scores range from 0–133, with higher scores indicating better performance.

Broadly, the CD is a measure of processing speed; however, more specifically, as reported in

detail previously in this document, DSST scores appear to primarily reflect graphomotor speed,

perceptual speed, and visual scanning ability, although other areas such as executive functioning

and learning/memory may play a smaller, yet still significant, role.

84

The WAIS-III Digit Symbol—Coding subtest (CD) has acceptable test-retest reliability

among 65–74-year olds, which is the age range of the population in the present study (r = .86;

The Psychological Corporation, 1997). Acceptable reliability is also indicated by a standard

error of measurement of 1.12 among 65–74-year olds (The Psychological Corporation, 1997).

The CD also shows acceptable criterion validity, as evidenced by a correlation of .77 between

WAIS-III and Wechsler Adult Intelligence Scale-Revised (WAIS-R) versions of this subtest

among a sample of 16–74-year olds, as well as a correlation of .77 between WAIS-III and

WISC-III versions of this subtest among a sample of 16-year-olds (The Psychological

Corporation, 1997). Discriminant validity of the CD is reflected by a weak correlation (r = .25)

with the Raven’s Standard Progressive Matrices (SPM), an untimed test of reasoning.

Convergent validity is indicated by a strong negative correlation (r = -.46) between scores on the

WAIS-R version of the CD (which is very similar to the WAIS-III version of the CD) and the

Useful Field of View Test (UFOV), a test of visual processing speed in which higher scores

indicate worse performance (Lunsman et al., 2008).

Memory impairment. Participants are asked if they are limited in any way due to

difficulty remembering or periods of confusion. They may answer yes, no, don’t know, or refuse

to answer.

Medical conditions. Information regarding medical conditions was collected using a

number of interviews, including the Blood Pressure Questionnaire (Appendix D), the Diabetes

Questionnaire (Appendix E), the Medical Conditions Questionnaire (Appendix F), and the

Kidney Conditions Questionnaire (Appendix G). More specifically, participants were asked

whether they had ever been diagnosed with certain medical conditions, as well as age at first

diagnosis. Diagnoses relevant to this study include asthma, cancer, congestive heart failure,

85

coronary heart disease, diabetes, hypercholesterolemia, hypertension, kidney failure, stroke,

emphysema, thyroid disease, and chronic bronchitis. If endorsed, participants were asked to

identify type(s) of cancer as well as whether they still had thyroid disease or chronic bronchitis,

whether they had an asthma attack in the past-year, and whether they’d been to the emergency

room or urgent care for asthma in the past year. If endorsed, they were also asked whether they

took insulin for diabetes (and for how long), whether they took pills for diabetes, and whether

they took prescription drugs for hypertension or hypercholesterolemia. Participants were asked

if they had been treated for anemia in the past three months, although they were not asked if they

had anemia. Finally, participants were asked whether any family members had Alzheimer

disease (AD) and, if so, to identify the family member’s relationship to them (e.g., mother,

father).

Physical activity. Information regarding physical activity was collected using the

Physical Activity Questionnaire (Appendix H) and the Physical Activity Individual Activities

File (Appendix I). Participants were asked to rate their average level of daily physical activity

out of four options. The interviewer then defined moderate and vigorous activities, gave

examples for the participant, and then asked if the participant had engaged in any moderate or

vigorous activities lasting at least ten minutes in the past month. If moderate or vigorous

activities were endorsed, the type of activity, past-month frequency, and duration was collected.

Frequency of any strength training exercises was then collected. Participants were asked to

report any strength training exercises even if they had included them in the moderate or vigorous

activity report.

There is no psychometric data available regarding this particular set of questions. For the

purposes of the present study, minutes of aerobic exercise per week was calculated using items

86

PAD200 and PAD320 of the Physical Activity Questionnaire (Appendix H) and items

PADACTIV, PADTIMES, and PADDURAT of the Physical Activity Individual Activities File

(Appendix I). Likewise, number of sessions of strength training per week was calculated using

items PAD440 and PAD460 of the Physical Activity Questionnaire (Appendix H).

Generally, there is little agreement on how to best assess physical activity in the elderly

(Kowalski, Rhodes, Naylor, Tuokko, & MacDonald, 2012). The only systematic review and

meta-analysis of validity of indirect measures of physical activity in older adults to-date found

that self-reported physical activity in the elderly showed a moderate correlation with direct

measures of physical activity (Kowalski et al., 2012). However, it should be noted that the

validity of many direct measures of physical activity that are developed primarily on middle-

aged samples are questionable among the elderly population due to metabolic or physical activity

characteristics. For example, accelerometers are generally less accurate recording lower

intensity level activities, which elderly individuals are more likely to do (Garatachea, Luque, &

Gallego, 2010; Kowalski et al., 2012). Therefore, there currently appears to be no clear “gold

standard” for assessing physical activity in the elderly.

Diet. Information regarding diet was collected using the Diet Behavior and Nutrition

Questionnaire and the Dietary Interview. In regard to the former, participants were asked to

estimate the number of fruits/fruit juices and vegetables they typically consume per day (see

Appendix J). In regard to the latter, participants were asked to recall all food and beverage

consumed during the previous day (i.e., midnight to midnight). To improve accuracy of recall,

participants were asked time, occasion, and place of consumption of each food and beverage they

recalled and were shown a list of foods that individuals tend to forget to report. Participants

were then asked about specific details of each food and beverage consumed, including amount.

87

Finally, all reported foods and beverages were repeated to the participant in chronological order

for confirmation or correction. Additional information regarding the Dietary Interview can be

found in the NHANES Dietary Interview Manual (NHANES Dietary Interview Manual, 1999–

2000). Previous day total fat and total saturated fat are estimated using the Dietary Interview

data. There are no psychometric data available on the Diet Behavior and Nutrition Questionnaire

or the Dietary Interview.

Blood pressure. NHANES blood pressure protocol was based on recommendations of

the American Heart Association. More specifically, blood pressure was taken by a physician or

health technologist using a mercury sphygmomanometer. Three to four trials were obtained. In

order to calculate a final blood pressure score, the first blood pressure reading was dropped and

the other readings were averaged. For participants who only completed one blood pressure trial,

the first blood pressure score was used for the final blood pressure score. Any diastolic readings

that were zero were dropped from computing the average unless there were only zero readings.

A zero reading likely indicates that the diastolic pressure was not measurable on that trial.

Cholesterol, triglycerides, C-reactive protein, glucose, and insulin. Total cholesterol,

high-density lipoprotein (HDL) cholesterol, triglycerides, C-reactive protein (CRP), fasting

glucose, and fasting insulin were measured using blood samples. Low-density lipoprotein (LDL)

cholesterol was calculated from total cholesterol, HDL cholesterol, and triglycerides.

Homeostasis model assessment of insulin resistance (HOMA-IR) was calculated from fasting

glucose and fasting insulin.

Nutritional biochemistries. Measurements of vitamin D, vitamin E, folate (serum and

red cell assays), vitamin B12, homocysteine, and iron were obtained via blood tests.

88

Substance use history. Information regarding current smoking and smoking history was

collected using the Smoking Questionnaire (Appendix K). Information regarding current and

past alcohol use was collected using the Alcohol Use Questionnaire (Appendix L). There are no

psychometric data available on these questionnaires.

20-foot-walk test. The 20-foot-walk test is a frequently used measure of gait speed (e.g.,

Atkinson et al., 2007; Okoro et al., 2006). In this test, participants are asked to walk down a 20-

foot unobstructed hallway or track. Instructions to the participant vary according to the study.

Typically, participants are either told to walk at their usual pace or walk at a fast pace. The

present study used the 20-foot-walk test to assess usual pace of walking. Assistive devices (e.g.,

canes, walkers) were allowed if typically used. Time was measured using a hand-held

stopwatch.

Despite the frequency of use of this test, little has been published on its psychometric

properties. It has shown to produce reliable ratings of gait speed relative to the 8-foot-walk test

(ICC 0.912; Newton, Kilma, & Cromwell, 2006). Some concerns have been raised regarding

test-retest reliability in similar tests, namely the 10-meter-walk and 20-meter-walk tests. In these

tests, the first trial appears to be significantly slower than subsequent trials (Green, Forster, &

Young, 2002; Motyl, Driban, McAdams, Price, & McAlindon, 2013). Information on the test-

retest reliability of the 20-foot-walk test has not been published to-date.

Medical treatments. Participants were asked to list all prescription medications that they

are currently taking, as well as number of times each medication was taken in the past 30 days.

They were also asked to show the prescription bottles to NHANES staff. Additionally,

participants were asked whether they had received cancer chemotherapy in the past 4 weeks or

planned to receive cancer chemotherapy in the next 4 weeks. Although this question on

89

chemotherapy was asked to determine eligibility for venipuncture, it can be used to more

inclusively describe medical treatments.

Procedures

The present study involved analysis of data from the National Health and Nutrition

Examination Survey (NHANES) collected throughout the country between 1999 and 2002. The

NHANES created by the National Center for Health Statistics (NCHS) in order to collect health

and nutrition data on individuals of all ages living in the United States to provide to the Centers

for Disease Control and Prevention (CDC) for the purposes of guiding public health action and

policy (NHANES Plan and Operations, 1999–2010). The NHANES protocol received approval

from the NCHS Research Ethical Review Board and was reviewed annually. The present study

used publically-available NHANES data. No additional permissions were required, and

exemption status from the Eastern Michigan University’s Institutional Review Board was

obtained in April 2015 (see Appendix M).

NHANES procedures are explained elsewhere (NHANES Plan and Operations, 1999–

2010). Briefly, identification of potential participants was accomplished via a multistage

probability sampling design aimed at enrolling a sample that is representative of the United

States civilian noninstitutionalized household population, although certain populations

(individuals aged 12–19, those aged over 70 years, Non-Hispanic black, Mexican American, and

low-income Caucasian individuals) were oversampled due to being “of particular public health

interest” (NHANES Plan and Operations, 1999–2010). Counties or groups of nearby small

counties were divided into neighborhoods, which were selected at random. Residences within

each selected neighborhood were then selected at random. The residence was then visited by an

NHANES researcher, who determined the eligibility of each resident using a screening

90

questionnaire that solicited basic demographic data (e.g., age, gender, race). The researcher

entered this information into an electronic device that reported whether each participant was

eligible, which was based on match with the latest census data. Those identified as eligible were

provided information about the study and asked if they would be willing to participate. For each

of the two phases of NHANES participation, the home interview and the health examination,

documented written consent among adults and parents of participating children, as well as

documented assent of participants aged 7–17, was required. The informed consent process and

personal interview were typically conducted in English or Spanish, but interpreters for those

more comfortable with other languages were used, as needed. The English version of the

consent form is included in Appendix N. Remuneration, as detailed below, was only provided

after completion of the health examination, which is the second phase of NHANES participation.

Individuals who were eligible and consented/assented then completed the personal

interview. This interview took place in the participant’s home unless the participant strongly

preferred to complete the interview elsewhere. The personal interview consisted of a series of

“questionnaires” about a variety of topics, such as health history, weight history, and alcohol use.

Most questionnaires were verbally administered by NHANES staff, whereas the others were

completed on a computer with the assistance of NHANES staff. If an individual needed

assistance in answering questions, a proxy (e.g., family member) was consulted. During the

personal interview, individuals were also administered the WAIS-III Digit Symbol—Coding

subtest. Prior to the start of the Coding subtest, participants were asked to put on any glasses that

they usually use to read. Those who could not take the Coding subtest without distraction in the

home and those who could not complete the sample items on the subtest (e.g., due to physical or

cognitive limitations) were not administered the test.

91

At the end of the personal interview, participants were asked to schedule an appointment

for their health examination at a nearby Mobile Examination Center (MEC) during a randomized

portion of the day (i.e., morning, afternoon, or evening). Participants were asked to attend

during these assigned appointment times, but were allowed to attend during alternate times, if

needed. These MEC appointments were made approximately two weeks from the time of the

personal interview. A postal letter was sent to participants one week before their scheduled

MEC appointment, which included health exam location information, approximate duration of

appointment, and remuneration amount. Duration of MEC appointment ranged from 40 minutes

to 4 hours depending on the tests administered (which was based on participant age, gender, and

health conditions). The amount of the cash remuneration varied from $30 to $100 according to a

number of characteristics, including survey year, participant age, and type of appointment

attended (i.e., assigned vs. alternate). Namely, adult participants were offered more money than

child participants. Adult participants were also offered additional money as an incentive to

attend their assigned health examination appointment. To improve attendance, a reminder phone

call was also made two days prior to this appointment.

Those who attended the MEC appointment underwent a variety of tests, including those

of vision, hearing, blood pressure, walking pace, body composition (i.e., DEXA) and blood (e.g.,

glucose, iron, CRP). At the end of the health examination, participants were provided their cash

remuneration, as well as reimbursement for any incurred travel expense. Approximately 3-4

months after the MEC appointment, participants were mailed a report of results of several of the

health tests (e.g., vision, hearing, blood). Individuals with results outside of healthy ranges were

recommended to consult a doctor.

92

For each of the 2-year NHANES cycles in the present study (i.e., 1999–2000 and 2001–

2002), approximately 30 counties were sampled and 12,000 individuals were invited to

participate. During the 1999–2000 cycle, 9,965 individuals completed the personal interview

and, of those, 9,282 completed the health examination (NHANES Plan and Operations, 1999–

2010). During the 2000–2001 cycle, 11,309 individuals completed the personal interview and,

of those, 10,477 completed the health examination (NHANES Plan and Operations, 1999–2010).

Exact response rates of these two cycles have not been reported.

Design

The present study analyzed data collected as part of NHANES during the years 1999–

2002. NHANES uses a multistage probability sampling design aimed at enrolling a sample that

is representative of the United States civilian noninstitutionalized household population

(NHANES Plan and Operations, 1999–2010). As explained above, counties or groups of small

counties were divided into neighborhoods, which were selected at random. Residences within

each selected neighborhood were then selected at random. Occupants of those residences were

then screened to determine study eligibility.

This was a cross-sectional study, as tests were conducted but no interventions were made.

Participants were assessed on two occasions: the personal interview and the health exam. The

personal interview occurred first, at which time the health exam was scheduled for a date

approximately two weeks later. Because these data were collected so close together and do not

include repeated measures, the present study used them as the same timepoint. This study also

used retrospective pretests (e.g., self-report history of weight and alcohol use).

Analysis

93

For hypotheses 1–13, data were analyzed using IBM SPSS version 21. All data were

screened using frequency distributions and descriptive statistics, and checked for outliers.

Relevant data were then assessed for the multivariate statistical testing assumptions of normality,

linearity, and homoscedasticity. More specifically, visual inspection of a histogram and normal

Q-Q plot for each variable of interest was used to assess univariate normality. Significance of

skewness and kurtosis was evaluated at an alpha level of .01. Bivariate scatterplots for all

subsets of variables were also examined; roughly elliptical patterns supported the assumption of

multivariate normality and linearity. Homoscedasticity was assessed through a visual inspection

of bivariate scatterplots and Levene’s Test for Equality of Variances.

Inferential statistical procedures also assume simple random sampling. This assumption

was technically not met, as all noninstitutionalized civilians living in the United States did not

have equal chance of being selected due to oversampling of certain groups of interest. However,

the assumption of independence in independent samples t-tests and Analysis of Variance

(ANOVA) was met because participant responses did not influence one another.

The minimum sample size needed within each group in order to run a two-tailed t-test or

chi-square test (1 df) using an alpha level of .05 was set at 26, as per results of an a priori power

analysis using an expected large effect size (d = .8) and power = 0.8. Unless otherwise specified,

all statistical analyses excluded underweight participants and used a two-tailed test of

significance at alpha level .05.

Analysis for Hypothesis 1. It was hypothesized that there would be a strong negative

correlation between age and DSST score. To test this hypothesis, a Pearson product-moment

correlation coefficient was calculated for age and DSST score, and its significance was tested

using a two-tailed test of significance at alpha level .05.

94

Analysis for Hypotheses 2a and 2b. Hypotheses 2a and 2b are concerned the

relationship between age and overall and central obesity indexes, respectively. More

specifically, it was hypothesized that there would be a weak, non-significant negative correlation

between age and BMI but a significant positive correlation between age and % total body fat

measured by DEXA. In addition, it was hypothesized that there would be no significant

correlation between age and WC but a significant positive correlation between age and % trunk

fat measured by DEXA. To test these hypotheses, four Pearson product-moment correlation

coefficients were calculated, and significance of these correlations were tested using two-tailed

tests of significance at alpha level .05.

Analysis for hypotheses 3a and 3b. It was hypothesized that participants with probable

lifetime history of excess alcohol consumption (i.e., answered yes to “Was there ever a time or

times in your life when you drank 5 or more drinks of any kind of alcoholic beverage almost

every day?”) would exhibit worse processing speed. To test this hypothesis while controlling for

gender as a potential confound, a two-way between-subjects ANOVA using a test of significance

at alpha level .05 was conducted with gender and probable lifetime history of excess alcohol

consumption as the independent variables and DSST score as the dependent variable.

Likewise, it was hypothesized that participants with probable past-year excess alcohol

consumption (i.e., endorsed consuming five or more drinks on half or more of the days during

the past year) would exhibit worse processing speed. An independent samples t-test using a two-

tailed test of significance at alpha level .05 was planned to compare those with and without

probable past-year history of excess alcohol consumption on DSST score; however, this analysis

could not be conducted as planned due to the probable past-year excess alcohol consumption

95

group having fewer participants than the minimum group size of 26 determined by the a priori

power analysis.

Analysis for Hypothesis 3c. It was hypothesized that individuals who currently smoke

would perform worse on the DSST than those who do not currently smoke. To test this

hypothesis, those who endorsed currently smoking tobacco via cigarettes, cigars, and/or pipe

were defined as current smokers, whereas those who denied currently smoking tobacco via

cigarettes, cigars, or pipe were defined as current non-smokers. An independent samples t-test

using a two-tailed test of significance at alpha level .05 was conducted to compare those who do

and do not currently smoke on DSST score.

Analysis for Hypothesis 4. Hypotheses 4a–4c stated the following: The no exercise

group would perform worse on the DSST than the aerobic-only, strength training-only, and

combination training groups. The aerobic-only group would perform better on the DSST than

the strength training-only group. The combination training group would perform better on the

DSST compared to the aerobic-only, strength training-only, and no exercise groups. To test

these hypotheses, a one-way ANOVA at alpha level .05 was run with exercise group as the

independent variable and DSST score as the dependent variable, followed by Tukey’s HSD post-

hoc test.

Analysis for Hypothesis 5. It was hypothesized that the diabetes group would perform

worse on the DSST than the prediabetes and no diabetes groups. It was also hypothesized that

the prediabetes group would perform worse on the DSST than the no diabetes group. To test

these hypotheses, a one-way ANOVA at alpha level .05 was run with diabetes group as the

independent variable and DSST score as the dependent variable, followed by Tukey’s HSD post-

hoc test.

96

Analysis for Hypothesis 6a. It was hypothesized that homocysteine levels would be

negatively correlated with processing speed independent of serum folate levels. To test this

hypothesis, a Pearson product-moment correlation coefficient was calculated for homocysteine

level and DSST score, and significance of this correlation was tested using a two-tailed test of

significance at alpha level .05. A partial correlation calculated for homocysteine level and DSST

score while controlling for folate level was then conducted using a two-tailed test of significance

at alpha level .05.

Analysis for Hypothesis 6b. It was hypothesized that folate levels would not be

significantly correlated with processing speed independent of homocysteine levels. To test this

hypothesis, a Pearson product-moment correlation coefficient was calculated for folate level and

DSST score, and significance of this correlation was tested using a two-tailed test of significance

at alpha level .05. Given the significant bivariate correlation, a partial correlation controlling for

homocysteine level was then conducted using a two-tailed test of significance at alpha level .05.

Analysis for Hypothesis 6c. It was hypothesized that those with low vitamin D levels (<

35 nmol/l) would perform worse on the processing speed measure than those with healthy

vitamin D levels. To test this hypothesis, an independent samples t-test using a two-tailed test of

significance at alpha level .05 was conducted to compare those with and without low vitamin D

levels on DSST score.

Analysis for Hypothesis 6d (exploratory). It was hypothesized that participants

suffering from an iron deficiency (serum ferritin level of < 22µg/L) would perform worse on the

processing speed task than participants without an iron deficiency. An independent samples t-

test using a two-tailed test of significance at alpha level .05 was conducted to compare

participants with and without iron deficiency on DSST score.

97

Analysis for Hypothesis 6e (exploratory). It was hypothesized that vitamin E level

(i.e., alpha-tocopherol) would be positively correlated with processing speed performance. As an

additional exploratory analysis, gamma-tocopherol, another type of Vitamin E, was also

investigated. For each of the forms of vitamin E included in the NHANES dataset (i.e., alpha-

tocopherol and gamma-tocopherol), a Pearson product-moment correlation coefficient was

calculated for vitamin E level and DSST score, and significance of this correlation was tested

using a two-tailed test of significance at alpha level .05. The Bonferroni post-hoc test was then

applied to prevent alpha inflation.

Analysis for Hypothesis 6f (exploratory). It was hypothesized that the total number of

nutrient deficiencies (i.e., vitamin B12, vitamin D, vitamin E, folate, iron) would be negatively

correlated with processing speed task performance. To test this hypothesis, a Pearson product-

moment correlation coefficient was calculated for number of nutrient deficiencies and DSST

score, and significance of this correlation was tested using a two-tailed test of significance at

alpha level .05. Prior to running this analysis, cutoffs for defining nutrient deficiency statuses

were reviewed, and relationships between single nutrient deficiencies and processing speed task

performance were assessed to determine which nutrients should be included in the analysis.

Analysis for Hypotheses 7a and 7b. It was hypothesized that DSST scores would not

be significantly correlated with the overall obesity indexes (i.e., BMI, % DEXA total body fat)

but that they would be significantly negatively correlated with the central obesity indexes (i.e.

WC, WTR, WHtR, and % DEXA trunk fat). A bivariate correlation matrix was computed, and

significance of these correlations was tested using a two-tailed test of significance at alpha level

.05.

98

Analysis for Hypothesis 7c. It was hypothesized that, among those who are centrally

obese (WC ≥ 102 cm for men and 88 cm for women), (a) individuals meeting BMI obesity

criteria (BMI ≥ 30 kg/m2) would perform worse on the DSST and (b) individuals meeting DEXA

obesity criteria (fat mass ≥ 25% for men and ≥ 35% for women) would perform worse on the

DSST. To test the first part of the hypothesis, an independent samples t-test using a two-tailed

test of significance at alpha level .05 was conducted to compare those with central obesity and

BMI obesity to those with central obesity but no BMI obesity. Likewise, to test the second part

of the hypothesis, an independent samples t-test using a two-tailed test of significance at alpha

level .05 was conducted to compare those with central obesity and DEXA obesity to those with

central obesity but no DEXA obesity.

Analysis for Hypothesis 7d (exploratory). A stepwise multiple regression using

backward elimination was run to determine which obesity indices make meaningful

contributions to the overall prediction of DSST scores.

Analyses for BMI category by gender interaction on DSST score (exploratory). A

series of 2x3 ANOVAs were conducted to investigate main and interaction effects of BMI

weight category (i.e., normal weight, overweight, obese) and gender on current DSST score for

each of three timepoints (i.e., current, 10 years ago, and age 25), followed by Fischer’s LSD

post-hoc tests.

Analysis for Hypotheses 8a–8c. Hypotheses 8a–8c concerned the relationship between

weight trajectory and DSST score. Eight groups were created based on whether participants

were obese or not obese at age 25, 10 years ago (ages 55–64), and currently (ages 65–74). It was

hypothesized that those who were obese at age 25 and 10 years ago would perform the worst on

the DSST, whereas those with no history of obesity would perform best on the DSST. It was

99

also hypothesized that those who were obese at only one timepoint would perform better than

those who were obese at age 25 and 10 years ago. To test this hypothesis, a one-way ANOVA

with eight groups at alpha level .05 was planned with weight trajectory group as the independent

variable and DSST score as the dependent variable, followed by Tukey’s HSD post-hoc test.

However, the decision was made to remove three groups from this analysis after it was

discovered that each contained a very low number of participants. After removal of these three

groups from the analysis, five groups remained. To test this hypothesis using the new groups, a

one-way ANOVA with five groups at alpha level .05 with weight trajectory group as the

independent variable and DSST score as the dependent variable was conducted, followed by

Tukey’s HSD post-hoc test.

Analysis for Hypothesis 8d. It was hypothesized that percent of weight gained between

early adulthood (age 25) and late middle-age (10 years ago, which is ages 55-64 in the present

study) would be associated with worse performance on the DSST. To test this hypothesis, a

Pearson product-moment correlation coefficient was calculated for percent of weight gained and

DSST score, and significance of this correlation was tested using a two-tailed test of significance

at alpha level .05. As planned, individuals who were underweight in early adulthood were

excluded from this analysis. Additionally, since the question addressed the percent of weight

gained between early adulthood and late middle-age, individuals who did not gain weight

between these two timepoints were excluded from this analysis.

Analysis for Hypothesis 8e. It was hypothesized that the absolute value of the percent

of weight change over the previous 10-year period would be associated with worse performance

on the DSST. To test this hypothesis, a Pearson product-moment correlation coefficient was

calculated for absolute value of percent of weight change and DSST score, and significance of

100

this correlation was tested using a two-tailed test of significance at alpha level .05. As planned,

individuals who were underweight 10 years ago (during their “late middle-age” period) were

excluded from this analysis, as well as the Hypothesis 8e exploratory analyses.

Analysis for Hypothesis 8f. It was hypothesized that individuals endorsing significant

past-year unintentional weight loss would perform worse on the DSST than those who achieved

significant past-year weight loss that was intentional and those who were past-year weight-

stable. No specific hypotheses were made regarding how participants with significant past-year

weight gain would perform on the DSST relative to the other three groups.

For the purposes of this hypothesis, “significant past-year weight loss” was defined as

having a current body weight that is less than or equal to 95% of the body weight reported one

year ago, representing a 5% or greater decrease in body weight. Likewise, “significant past-year

weight gain” was defined as having a current body weight that is greater than or equal to 105%

of the body weight reported one year ago, representing a 5% or greater increase in body weight.

Finally, “past-year weight-stable” was defined as having a weight that was greater than 95% of

the body weight reported one year ago but less than 105% of the body weight reported one year

ago, representing less than a 5% change in body weight.

To test this hypothesis, a one-way ANOVA at alpha level .05 was run with past-year

weight group as the independent variable (four levels: unintentional significant past-year weight

loss, intentional significant past-year weight loss, significant past-year weight gain, and past-year

weight stable) and DSST score as the dependent variable. The Tukey-Kramer modification of

Tukey’s HSD test, which was chosen due to the presence of unequal group sizes, was then used

to make post-hoc comparisons of means.

101

Analysis for Hypothesis 9. It was hypothesized that there would be a significant

negative correlation between a marker of peripheral inflammation (i.e., CRP level) and DSST

score, which would remain significant after controlling for health condition and smoking

confounds. To test this hypothesis, a Pearson product-moment correlation coefficient was

calculated for CRP level and DSST score, and significance of this correlation was tested using a

two-tailed test of significance at alpha level .05. Given that the bivariate correlation was not

significant, there was no need to conduct a partial correlation controlling for health conditions

and smoking, as planned.

Analysis for Hypothesis 10a. It was hypothesized that participants who endorsed one or

more of the following conditions would perform worse on the DSST than those who did not

endorse any: problems of blood flow to the heart typically leading to or involving heart muscle

damage (i.e., congestive heart failure, coronary heart disease, heart attack), stroke, diabetes,

weak/failing kidneys, chronic respiratory disease (i.e., chronic bronchitis or emphysema), and

thyroid disease (if untreated). As previously explained, upon implementation of the present

study, it was discovered not to be possible to determine whether thyroid disease had been treated;

therefore, hypothyroidism was not included in this analysis as planned. To test the hypothesis,

an independent samples t-test followed by a two-tailed test of significance at alpha level .05 was

conducted to compare DSST scores of those with none of these health conditions and those with

≥ 1 of these health conditions.

Analysis for Hypothesis 10b. It was hypothesized that duration of respiratory disease

would be negatively correlated with DSST score. To test this hypothesis, a Pearson product-

moment correlation coefficient was calculated for duration of respiratory disease and DSST

score, followed by a two-tailed test of significance at alpha level .05.

102

Analysis for Hypothesis 11. It was hypothesized that participants with dyslipidemia

(i.e., unhealthy levels of total cholesterol, HDL cholesterol, LDL cholesterol, and/or

triglycerides) would demonstrate worse processing speed task performance. To test this

hypothesis, an independent samples t-test followed by a two-tailed test of significance at alpha

level .05 was conducted to compare DSST scores of participants with and without dyslipidemia.

Analysis for Hypothesis 12a. It was hypothesized that risk of first stroke, as indicated

by score on the modified version of the FSRP, would be inversely associated with processing

speed. To test this hypothesis, a Pearson product-moment correlation coefficient was calculated

for modified FSRP score and DSST score, and significance of this correlation was tested using a

two-tailed test of significance at alpha level .05. Individuals with a history of stroke were

excluded from this analysis, as planned.

Analysis for Hypothesis 12b. It was hypothesized that risk of onset of CVD, as

indicated by modified FRS for general CVD scores, would be inversely associated with

processing speed task performance. To test this hypothesis, a Pearson product-moment

correlation coefficient was calculated for modified FRS for general CVD score and DSST score,

and significance of this correlation was tested using a two-tailed test of significance at alpha

level .05. Individuals with a history of CVD (defined as one or more of the following: CHF,

angina pectoris, heart attack, stroke) were excluded from this analysis, as planned.

Analysis for Hypothesis 13. It was hypothesized that there would be a significant

positive correlation between speed of usual gait and DSST score, which would remain significant

after controlling for BMI. To test this hypothesis, a Pearson product-moment correlation

coefficient was calculated for gait speed and DSST score, and significance of this correlation

was tested using a two-tailed test of significance at alpha level .05. Given that the bivariate

103

correlation was significant, a partial correlation controlling for BMI was then conducted using a

two-tailed test of significance at alpha level .05.

Final model. Mplus 6.1 was used for the final structural equation model (Muthén &

Muthén, 2008). Missing data was treated with weighted least squares with mean and variance

adjustment (WLSMV) due to the use of categorical data. A theoretical model of the relationship

between persistent obesity and DSST score, as outlined in Figure 1, was extensively revised prior

to testing due to practical limitations, results of the previously described analyses, and theoretical

considerations. Confirmatory modeling was used, with an absolute fit index cut-off of root mean

square error of approximation (RMSEA) ≤ 0.06 and an incremental index cut-off of comparative

fit index (CFI) ≥ 0.95 (Bentler, 1990; Hu & Bentler, 1999). Additionally, chi-square statistics

were examined for significance, with nonsignificant results suggestive of acceptable fit. Further,

standardized factor loadings of > .35 were required for indicators to be retained in the

measurement model (Brown, 2006).

104

Results

Prior to each parametric analysis, data were assessed with respect to all assumptions

relevant to the planned statistical tests. Namely, group data for relevant variables were checked

for normality using visual inspection of a histogram and Q-Q plot, significance of skewness and

kurtosis was evaluated at an alpha level of .01, bivariate scatterplots for all subsets of variables

were examined for roughly elliptical patterns to support the assumptions of multivariate

normality and linearity, and homoscedasticity was assessed though a visual inspection of

bivariate scatterplots and Levene’s Test for Equality of Variances. Nonparametric tests were

used to analyze data that were not normally distributed and could not achieve normality with the

appropriate transformation.

Of the total 1,381 participants in the study sample, 1,143 had a recorded DSST score, and

238 did not have a recorded DSST score. Common reasons for lack of DSST score included

decision not to attempt the DSST or its sample exercise due to distractions in the home

environment (n = 90), refusal to complete the DSST sample exercise (n = 33), and inability to

complete the DSST sample exercise due to cognitive (n = 26) or physical (n = 25) limitations.

Independent samples t-tests failed to reveal a statistically significant difference in mean age, t

(1379) = -.705, p = .481, or mean BMI, t (1371) = -.372, p = .710, between the group with

recorded DSST scores and the group without recorded DSST scores. See Table 2 for participant

characteristics. Of note, gender was identified as a potential covariate for analyses involving

DSST score. Namely, men were found to have significantly lower DSST scores (M = 40.3, SD =

18.2) compared to women (M = 44.2, SD = 18.6), t (1141) = -3.643, p < .001. Therefore, gender

was considered as a potential covariate in relevant analyses.

105

Hypothesis 1 Results

It was hypothesized that there would be a strong negative correlation between age and

processing speed task performance, as indicated by DSST score. A Pearson product-moment

correlation coefficient calculated for age and DSST score revealed a very weak, but statistically

significant, negative relationship between age and DSST score, r (1141) = -.084, p = .005.

Hypothesis 2a Results

It was hypothesized that there would be a negative correlation between age and BMI that

would not reach statistical significance. A Pearson product-moment correlation coefficient

calculated for age and BMI yielded a very weak, but statistically significant, negative correlation,

r (1371) = -.071, p = .008.

In contrast, it was hypothesized that there would be a significant positive correlation

between age and % total body fat measured by DEXA. Given that the NHANES database

contained multiply imputed data for DEXA measurements, a Pearson product-moment

correlation coefficient was calculated for each of the five versions of completed data, and these

estimates were then pooled by calculating their mean. Results were not statistically significant, r

(1239) = .003, p = .916.

Hypothesis 2b Results

It was hypothesized that there would be no significant correlation between age and WC.

Consistent with this hypothesis, the correlation between age and WC was not statistically

significant, r (1210) = -.006, p = .835.

In contrast, it was hypothesized that there would be a significant positive correlation

between age and % trunk fat measured by DEXA. A Pearson product-moment correlation

coefficient was calculated for each of the five versions of completed data. Results of these

106

estimates were then pooled, failing to reveal a significant relationship between age and % total

trunk fat measured by DEXA, r (1239) = -.009, p = .752.

Hypothesis 3a Results

It was hypothesized that those with probable lifetime history of excess alcohol

consumption (i.e., answered “yes” to “Was there ever a time or times in your life when you

drank 5 or more drinks of any kind of alcoholic beverage almost every day?”) would exhibit

worse processing speed task performance than those without probable lifetime history of excess

alcohol consumption (i.e., answered “no” to this question). In order to control for gender as a

potential confound in this analysis, a two-way between-subjects ANOVA was run with gender

and probable lifetime history of excess alcohol consumption as the independent variables and

DSST score as the dependent variable. Results revealed a main effect for probable lifetime

history of excess alcohol consumption on DSST score, F (1, 1039) = 5.742, p = .017, but failed

to reveal a significant main effect for gender on DSST score or a significant interaction effect

between the two independent variables. More specifically, participants with probable lifetime

history of excess alcohol consumption scored significantly lower on the DSST (M = 36.5, SD =

16.3) than those without probable lifetime history of excess alcohol consumption (M = 43.7, SD

= 18.7), indicating worse processing speed task performance.

Hypothesis 3b Results

Hypothesis 3b concerned the survey question “In the past 12 months, on how many days

did you have 5 or more drinks of any alcoholic beverage?” It was predicted that those with

probable past-year excess alcohol consumption (i.e., endorsed consuming five or more drinks on

half or more of the days during the past year) would exhibit slower processing speed than those

without probable past-year excess alcohol consumption (i.e., reported consuming five or more

107

drinks on fewer than half of the days during the past year). Due to the number of participants in

the probable past-year excess alcohol consumption group (n = 14) being less than the minimum

group size of 26 determined by the a priori power analysis, an independent samples t-test could

not be conducted as planned.

Hypothesis 3c Results

It was hypothesized that individuals who currently smoke would perform worse on the

DSST than those who do not currently smoke. As expected, individuals who reported that they

currently smoke cigarettes, cigars, and/or pipes scored significantly lower on the DSST (M =

37.0, SD = 17.0) than those who reported that they do not currently smoke (M = 43.4, SD =

18.1), indicating worse processing speed task performance, t (624) = -3.989, p < .001.

In exploratory analyses, the relationship between current smoking frequency and DSST

score was investigated. A one-way between subjects ANOVA was performed to compare the

effect of smoking frequency on DSST score in the “not at all,” “some days,” and “everyday”

conditions. There was a statistically significant effect of smoking frequency on DSST score for

the three groups, F (2, 623) = 8.661, p < .001. A post hoc Tukey’s HSD test revealed that the

“not at all” group (M = 43.4, SD = 18.1) scored significantly higher on the DSST than the

“everyday” group (M = 36.3, SD = 17.1), whereas the “some days” group (M = 40.3; SD = 16.4)

did not differ significantly from either group. Results indicate that nonsmokers performed

significantly better on the processing speed task than everyday smokers, whereas individuals

who smoked only “some days” of the week performed in a manner that was not statistically

different from the nonsmokers or the everyday smokers.

108

Hypothesis 4 Results

Using the operational definitions for no exercise, aerobic-only, strength training-only,

and combination training specified in the hypotheses section of this document, it was

hypothesized that (a) the no exercise group would perform worse on the DSST than the aerobic-

only, strength training-only, and combination training groups; (b) the aerobic-only group would

perform better on the DSST than the strength training-only group; and (c) the combination

training group would perform better on the DSST compared to the aerobic-only, strength

training-only, and no exercise groups. A one-way ANOVA conducted with exercise group as the

independent variable and DSST score as the dependent variable revealed a statistically

significant effect of exercise group on DSST score, F (3, 959) = 31.697, p <.001. Post hoc

Tukey’s HSD tests were then performed, revealing that the combination training group (M =

53.4, SD = 15.1) scored significantly higher on the DSST than the aerobic-only (M = 47.3, SD =

17.5; p = .017), strength training-only (M = 41.9, SD = 17.9; p = .009), and no exercise groups

(M = 37.8, SD = 18.1; p < .001). It also revealed that the aerobic-only group scored significantly

higher on the DSST than the no-exercise group, p < .001. No other group differences were

statistically significant.

Hypothesis 5 Results

Using the operational definitions for diabetes, prediabetes, and no diabetes specified in

the Hypotheses section of this document, it was predicted that the diabetes group would perform

worse on the DSST than the prediabetes and no diabetes groups. It was also expected that the

prediabetes group would perform worse on the DSST than the no diabetes group. To test these

hypotheses, a one-way ANOVA at alpha level .05 was run with diabetes status as the

independent variable (three levels) and DSST score as the dependent variable, revealing a

109

significant effect of diabetes status on DSST score [F (2, 758) = 14.04, p < .001]. Post hoc

comparisons using the Tukey HSD test indicated that the mean DSST score of the diabetes group

(M = 36.5, SD = 17.0) was significantly lower than both the prediabetes group (M = 43.9, SD =

18.3; p < .001) and the no diabetes group (M = 43.7, SD = 19.8; p < .001). The difference

between the prediabetes group and the no diabetes group was not statistically significant, p =

.992.

In exploratory analyses, the potential impact of past-month blood glucose regulator

medication on DSST was investigated among participants with diabetes, using the same

operational definition for this condition as above. An independent samples t test failed to find a

statistically significant difference in mean DSST scores between participants with diabetes who

reported past-month use of blood glucose regulator medication and participants with diabetes

who denied past-month use of blood glucose regulator medication, t (273) = .519, p =.604.

Next, the relationship between fasting blood glucose and DSST score was investigated.

Individuals were divided into the following groups based on their fasting blood glucose levels:

diabetic range (> 125 mg/dl), prediabetic range (100–125 mg/dl), normal range (70–99 mg/dl),

and hypoglycemic range (< 70 mg/dl). Only one individual was found to have tested within the

hypoglycemic range; therefore, this category was excluded from subsequent analyses. A one-

way ANOVA with fasting blood glucose group as the independent variable (three levels) and

DSST score as the dependent variable failed to find a significant effect of fasting blood glucose

on DSST score [F (2, 465) = 2.39, p = .093].

The relationship between HbA1c and DSST score was then investigated. Individuals were

divided into the following groups based on their HbA1c values: diabetic range (≥ 6.5%),

prediabetic range (5.7% to 6.4%), and normal range (< 5.7%). A one-way ANOVA with HbA1c

110

group as the independent variable (three levels) and DSST score as the dependent variable

revealed a significant effect of HbA1c group on DSST score [F (2, 1042) = 18.22, p < .001]. Post

hoc comparisons using the Tukey HSD test indicated that the mean DSST score of the diabetic

range group (M = 36.0, SD = 17.0) was significantly lower than both the prediabetic range group

(M = 41.3, SD = 18.3; p = .007) and the normal range group (M = 45.3, SD = 18.6; p < .001). In

addition, the prediabetic range group scored significantly lower than the normal range group, p =

.007.

Finally, the relationship between HOMA-IR and DSST score was investigated. HOMA-

IR was calculated by multiplying fasting insulin (U/l) by fasting blood glucose (mg/dl) and

dividing by 405, as described by Matthews and colleagues (1985). Individuals taking insulin

were excluded from this analysis. Peripheral insulin resistance was operationally defined as a

HOMA-IR value of greater than the 75th percentile; this criterion was chosen in accordance with

the practice of a number of other studies (e.g., Lee et al., 2006; Nasution, Setiati, Trisnohadi, &

Oemardi, 2006; Wilson et al., 2007). An independent samples t-test conducted to compare

DSST scores of participants with peripheral insulin resistance (M =38.7, SD = 18.3) and without

peripheral insulin resistance (M =43.6, SD = 17.6) revealed a statistically significant difference

between groups, t (389) = -2.311, p = .021. The same analysis among only participants without

diabetes was attempted, but it could not be completed due to the number of participants in the

peripheral insulin resistance group (n = 6) being lower than the minimum sample size of 26

determined by the a priori power analysis.

Hypothesis 6a Results

It was hypothesized that homocysteine levels would be negatively correlated with

processing speed task performance independent of folate levels. First, a Pearson product-

111

moment correlation between homocysteine level and DSST score was calculated, which revealed

a statistically significant weak negative relationship between these two variables, r (1039) = -

.108, p < .001. Next, a partial correlation was run to determine the relationship between these

two variables while controlling for folate level. There was a statistically significant very weak

negative partial correlation between homocysteine level and DSST score while controlling for

folate level, r (1038) = -.084, p = .007. A comparison of the zero-order correlation and partial

correlation suggests that folate had relatively little influence in controlling for the relationship

between homocysteine level and processing speed task performance. Results indicate that higher

homocysteine levels are very weakly associated with worse processing speed task performance

independent of folate levels.

In an exploratory analysis, participants with and without hyperhomocysteinemia were

compared on DSST score. Hyperhomocysteinemia was defined as ≥ 12 μmol/L based on the

work of Castañon, Lauricella, Kordich, and Quintana (2007). An independent samples t-test

revealed that participants with hyperhomocysteinemia scored significantly lower on the DSST

(M = 37.2, SD = 18.0) than participants without hyperhomocysteinemia (M = 44.0, SD = 18.5),

indicating worse processing speed task performance, t (1041) = -4.764, p < .001. Cohen’s effect

size (d = .37) suggested weak-to-moderate practical significance.

Hypothesis 6b Results

It was hypothesized that folate levels would not be significantly correlated with

processing speed independent of homocysteine levels. A Pearson product-moment correlation

between folate level and DSST score revealed a statistically significant weak positive association

between these two variables, r (1039) = .204, p < .001. In addition, there was a statistically

significant weak positive partial correlation between folate level and DSST score while

112

controlling for homocysteine level, r (1038) = .193, p < .001. A comparison of the zero-order

correlation and partial correlation suggests that homocysteine level had very little influence in

the relationship between folate level and processing speed task performance. Results indicate

that higher folate levels are weakly associated with better processing speed task performance

regardless of homocysteine levels.

No participants in the present study met folate deficiency criteria (i.e., serum folate < 1.4

ng/mL). Given the lack of consensus on the lower limit of serum folate reference values in the

literature and the results from the present study’s previous correlation analyses that showed

folate to be positively associated with processing speed independent of homocysteine levels, it

was deemed acceptable to include a folate variable in this analysis that would capture folate

depletion rather than folate deficiency. A serum folate value < 8 ng/mL was chosen for a cutoff

for “folate depletion” based on the results of De Bruyn and colleagues (2014). An independent

samples t-test conducted to compare those with and without folate depletion on DSST score

showed that participants with folate depletion scored significantly lower on the DSST (M = 35.0,

SD = 18.5) than participants without folate depletion (M = 43.6, SD = 18.3), indicating worse

processing speed task performance, t (1026) = -4.406, p < .001. Cohen’s effect size (d = .47)

suggested moderate practical significance.

Next, a two-way ANOVA conducted to investigate interaction effects between

hyperhomocysteinemia and folate depletion on DSST score failed to show a statistically

significant interaction, F (1, 1022) = .797, p = .372. See Figure 2. Note that results of main

effects for this analysis are identical to the previously reported t-test results.

113

Figure 2. DSST scores by folate depletion and hyperhomocysteinemia status. Significant main

effects for hyperhomocysteinemia, p < .001, and folate depletion, p < .001, on DSST score. Non-

significant interaction effect between hyperhomocysteinemia and folate depletion on DSST

score, p = .372.

114

Hypothesis 6c Results

It was hypothesized that participants with low 25-hydroxyvitamin D levels (< 35 nmol/L)

would perform worse on the processing speed task than those with healthy vitamin D levels.

Results of an independent samples t-test conducted to compare those with low vitamin D levels

and those without low vitamin D levels on DSST score showed that participants with low

vitamin D levels scored significantly lower on the DSST (M = 39.2, SD = 17.9) than participants

without low vitamin D levels (M = 46.0, SD = 16.8), indicating worse processing speed task

performance, t (516) = -2.855, p = .004.

Exploratory analyses were used to investigate possible interaction effects between

individual obesity variables, gender, and 25-hydroxyvitamin D level on DSST score. First, a

2x2x2 factorial ANOVA was conducted to investigate interaction effects of current obesity status

(two levels: obese, non-obese), gender (two levels: male, female), and 25-hydroxyvitamin D

status (two levels: low vitamin D, adequate vitamin D) on DSST score. The interaction effect

between vitamin D status and gender trended toward significance, F (1, 518) = 2.957, p = .086,

such that men’s DSST scores tended to be more negatively impacted by low vitamin D status

than women’s DSST scores. No other interaction effects were significant.

Second, a 2x2x2 factorial ANOVA was conducted to investigate potential interaction

effects between late middle-age obesity history, gender, and current 25-hydroxyvitamin D level

on DSST score. The interaction effect between late middle-age obesity history and vitamin D

status was statistically significant, F (1, 504) = 5.116, p = .024, such that low vitamin D status

appeared to have little impact DSST score in those with history of late middle-age obesity but

did appear to have significant impact on DSST score in those without history of late middle-age

obesity. More specifically, among those without late middle-age obesity, those with current low

115

vitamin D level had lower DSST scores (M = 37.5, SD = 18.2) than those without current low

vitamin D level (M = 48.2, SD = 16.1). See Figure 3 for more information. No other interaction

effects were significant.

116

Women’s DSST scores by obesity and vitamin D status

Low Vitamin D

Yes No Yes No Obese 10 Years Ago Currently Obese

Men’s DSST scores by obesity and vitamin D status

Low Vitamin D

Yes No Yes No Obese 10 Years Ago Currently Obese

Figure 3. Interaction effects between gender, current vitamin D status, and obesity status on

DSST score for current obesity status and obesity status ten years ago. Significant interaction

effect between obesity status ten years ago (i.e., late middle-age obesity history) and vitamin D

status on DSST score, p = .024. Interaction effect between vitamin D status and gender on DSST

score trended toward significance, p = .086, when data was analyzed by current obesity status.

All other interaction effects were non-significant.

DS

ST

Sco

re

DS

ST

Sco

re

117

Hypothesis 6d Results (exploratory)

It was hypothesized that participants with iron deficiency (serum ferritin level < 22µg/L)

would perform worse on the processing speed task than participants without iron deficiency. An

independent samples t-test was conducted to compare participants with iron deficiency to

participants without iron deficiency on DSST score. The iron deficient group scored lower on

the DSST (M = 39.3, SD = 18.0) compared to the non- iron deficient group (M = 43.0, SD =

18.5), but this difference was not statistically significant, t (1026) = -1.391, p = .165.

Hypothesis 6e Results (exploratory)

It was hypothesized that vitamin E level (i.e., alpha-tocopherol) would be positively

correlated with processing speed task performance. A Pearson product-moment correlation

coefficient calculated for alpha-tocopherol (µmol/L) and DSST score using a two-tailed test of

significance at alpha level .05 revealed a statistically significant weak-to-moderate positive

relationship between these two variables, r (490) = .258, p < .001, indicating that higher alpha-

tocopherol levels are associated with better processing speed task performance.

A set of exploratory analyses were conducted to investigate the relationship between

gamma-tocopherol and DSST score. In contrast to the findings involving alpha-tocopherol, a

Pearson product-moment correlation coefficient calculated for gamma-tocopherol and DSST

score revealed a statistically significant weak inverse relationship between these two variables, r

(417) = -.140, p = .004, indicating that higher gamma-tocopherol levels are associated with

worse processing speed task performance. Next, a correlation was calculated for alpha-

tocopherol and gamma-tocopherol, which revealed a moderate inverse relationship, r (497) = -

.395, p < .001. Finally, a partial correlation between gamma-tocopherol level and DSST score

while controlling for alpha-tocopherol level was run, which failed to indicate a significant

118

relationship, r (409) = -.012, p = .804. Results suggest that higher gamma-tocopherol levels are

not associated with worse processing speed task performance after alpha-tocopherol levels are

controlled.

Another exploratory analysis was attempted in order to assess the relationship between

vitamin E deficiency status and DSST score. It was discovered that only 2 of the 492

participants with both an alpha-tocopherol value and DSST score met vitamin E deficiency

criteria (i.e., alpha-tocopherol < 11.6 µmol/L or, for adults with hyperlipidemia, a ratio of serum

alpha-tocopherol to total lipids of < 0.8 mg/g). Therefore, the relationship between vitamin E

deficiency status and DSST score could not be assessed.

Hypothesis 6f Results (exploratory)

It was hypothesized that the total number of nutrient deficiencies (i.e., vitamin D, iron,

vitamin B12, vitamin E, folate) would show a negative correlation with processing speed task

performance. Prior to running this analysis, cutoffs for defining nutrient deficiency statuses were

reviewed, and relationships between single nutrient deficiencies and processing speed task

performance were assessed. Low vitamin D level (25-hydroxyvitamin D < 35 nmol/L), folate

depletion (serum folate < 8 ng/mL), iron deficiency (serum ferritin level < 22µg/L), and low

vitamin E level (alpha-tocopherol < 11.6 µmol/L or, for adults with hyperlipidemia, a ratio of

serum alpha-tocopherol to total lipids of < 0.8 mg/g) have already been addressed in hypotheses

6b-e. In regard to vitamin B12, a one-way ANOVA run at alpha level .05 with vitamin B12 group

as the independent variable (three levels: deficient [< 145 pmol/L], low-normal range [145–260

pmol/L], normal range [> 260 pmol/L]) and DSST score as the dependent variable failed to

reveal a significant effect of vitamin B12 group on DSST score [F (2, 1025) = .132, p = .876].

119

Based on the above analyses, the following nutrient deficiencies/depletions were chosen

to be included in the main analysis for this hypothesis: 25-hydroxyvitamin D level (< 35

nmol/L), serum ferritin level (< 22µg/L), and serum folate depletion (< 8 ng/mL). A Pearson

product-moment correlation coefficient calculated for total number of these nutrient

deficiencies/depletions and DSST score using a two-tailed test of significance at alpha level .05

revealed a statistically significant weak negative relationship between these two variables,

r (1141) = -.111, p < .001, indicating that greater number of nutrient deficiencies/depletions is

associated with worse processing speed task performance.

In an exploratory analysis, a one-way ANOVA was run with methylmalonic acid quartile

as the independent variable and DSST score as the dependent variable, revealing a significant

effect of methylmalonic acid quartile on DSST score [F (3, 1036) = 5.104, p = .002]. Post hoc

comparisons using the Tukey-HSD test indicated that the highest methylmalonic acid quartile

scored significantly lower on the DSST (M = 38.7, SD = 18.9) than the second highest quartile

(M = 44.0, SD = 17.9), third highest quartile (M = 43.9, SD = 18.4), and bottom quartile (M =

43.8, SD = 18.6). No other group differences were statistically significant.

Results of Exploratory Analyses on Self-Reported Dietary Habits

The relationship between self-reported dietary habits and processing speed task

performance was explored. Results of vegetables and fruits/fruit juices are presented first, as

these were the only self-reported eating habit variables for which exploratory analyses were

planned. Note that these analyses were only possible for participants examined during the 1999–

2000 data collection cycle due to subsequent changes in the NHANES Diet Behavior and

Nutrition questionnaire.

120

A Pearson product-moment correlation coefficient calculated for self-reported number of

helpings of vegetables consumed per day and DSST score revealed a statistically significant

weak-to-moderate positive relationship between these two variables, r (556) = .231, p < .001,

indicating that greater self-reported number of helpings of vegetables consumed per day is

associated with better processing speed task performance. A Pearson product-moment

correlation coefficient calculated for self-reported number of helpings of fruits/fruit juices

consumed per day and DSST score revealed a statistically significant weak positive relationship

between these two variables, r (555) = .137, p < .001, indicating that greater self-reported

number of helpings of fruits/fruit juices consumed per day is associated with better processing

speed task performance. Self-reported numbers of helpings of vegetables and fruits/fruit juices

consumed per day were found to have a moderate positive relationship with one another,

r (708) = .344, p < .001.

A stepwise multiple regression was then run to determine whether daily number of

fruits/fruit juices made a meaningful contribution to the overall prediction of DSST score above

and beyond that of daily number of helpings of vegetables. Note that the statistical assumptions

of stepwise multiple regression, including the lack of multicollinearity, were met prior to running

this analysis. A significant regression equation was found, F (2,554) = 16.725, p < .001 with an

R2 of .053 with the predictor daily number of helpings of vegetables, but daily number of

helpings of fruits/fruit juices did not meaningfully contribute to the model.

Relationship between DSST score and self-reported number of helpings of each of the

following types of foods were investigated in additional exploratory analyses: breads/grain

foods, milk/dairy foods, and protein foods. None of the Pearson product-moment correlation

coefficients calculated for these variables and DSST score reached statistical significance.

121

Hypotheses 7a and 7b Results

It was hypothesized that DSST scores would not be significantly correlated with the

overall obesity indexes (i.e., BMI, % DEXA total body fat) but would show significant negative

correlations with the central obesity indexes (i.e., WC, WTR, WHtR, and % DEXA trunk fat).

First, bivariate correlations were calculated to analyze the relationship between DSST

score and each of the obesity indexes that were not derived from DEXA (i.e., BMI, WC, WTR,

and WHtR). Results are displayed in Table 3. As hypothesized, these analyses revealed

statistically significant inverse relationships between DSST score and each of the central obesity

indexes (i.e., WC, WTR, WHtR) but no statistically significant relationship between DSST score

and the overall obesity index (i.e., BMI).

Next, the relationship between DSST score and each DEXA variable (i.e., % DEXA total

body fat and % DEXA trunk fat) was analyzed using multiply imputed data. For each DEXA

variable, bivariate correlations between DSST score and each of its five versions of completed

data were calculated, and results of these estimates were pooled. In contrast to what was

hypothesized, a very weak but statistically significant positive correlation between DSST score

and % total body fat measured by DEXA was found. Further, results failed to show a

statistically significant correlation between DSST score and % trunk fat measured by DEXA.

See Table 3 for more information.

122

Table 3

Summary of Correlations between DSST Score and Obesity Indexes

DSST score p

Overall obesity indexes

BMI -.046 .121

DEXA % total body fat .076a .013

Central obesity indexes

WC -.072 .019

WHtR -.135 <.001

WTR -.154 <.001

DEXA % trunk fat .050a .103

Note. BMI = body mass index; WC = waist circumference; WHtR = waist-height ratio; WTR = waist-thigh ratio;

DEXA = dual energy x-ray absorptiometry. apooled estimates using multiply imputed data

123

Hypothesis 7c Results

It was hypothesized that, among participants who met WC central obesity criteria, those

who also met BMI obesity criteria would perform worse on the DSST than those who did not.

An independent samples t-test conducted to compare centrally obese individuals with

BMI obesity (M = 40.9, SD = 18.0) and without BMI obesity (M = 44.3, SD = 18.3) on DSST

score revealed a statistically significant difference between groups, t (667) = -2.448, p = .015.

Similarly, it was hypothesized that, among participants who met WC central obesity

criteria, those who also met DEXA obesity criteria would perform worse on the DSST than those

who did not. An independent samples t-test to compare centrally obese individuals with and

without DEXA obesity on DSST score was run for each of the five versions of the completed

DEXA data. Results of these estimates were then pooled. In contrast to what was hypothesized,

results failed to show a statistically significant difference between centrally obese individuals

with and without DEXA obesity on DSST score, t (660) = .642, p = .521.

Hypothesis 7d Results (exploratory)

The question of which obesity indexes make meaningful contributions to the overall

prediction of DSST scores was explored. Note that, for Hypothesis 7d exploratory analyses,

single imputation of missing DEXA data was used rather than multiple imputation to allow for

multiple regression analysis. Also note that correlations between the dependent variable and the

predictor variables as shown in Table 3 are weaker than ideal for multiple regression analysis;

however, the decision was made to continue given the exploratory nature of this analysis.

While testing the present data against statistical assumptions of stepwise multiple

regression, the problem of multicollinearity was identified on the basis of a bivariate correlation

124

matrix of the obesity indices using a cut-off value of .7, as shown in Table 4. It was further

confirmed by examining the tolerance and variance inflation factor for each predictor variable.

125

Table 4

Summary of Correlations between Obesity Indexes

BMI WC WHtR WTR DEXA %

total body

fat

DEXA %

trunk fat

BMI 1.000

WC .859 1.000

WHtR .902 .896 1.000

WTR .126 .468 .422 1.000

DEXA % total

body fat

.595 .362 .589 -.086 1.000

DEXA % trunk fat .674 .514 .701 .100 .949 1.000 Note. BMI = body mass index; WC = waist circumference; WHtR = waist-height ratio; WTR = waist-thigh ratio;

DEXA = dual energy x-ray absorptiometry. All correlations significant at p < .01.

126

Based on this data, the following obesity indices were excluded from the multiple

regression analysis: BMI, WC, and DEXA % trunk fat. A stepwise multiple regression using

backward elimination was then run to determine which of the remaining obesity indices (i.e.,

WHtR, WTR, DEXA % total body fat) make meaningful contributions to the overall prediction

of DSST scores. A significant regression equation was found (F(2, 1025) = 29.188, p < .001)

with an R2 of .054 with the predictors WHtR and DEXA % total body fat, which indicates that

5.4% of variance in DSST score is explained by movement in WHtR and DEXA % total body

fat. WTR did not meaningfully contribute to the model. Results are further detailed in Table 5.

127

Table 5

Summary of Multiple Regression Analyses for Obesity Index Variables Predicting DSST Score

b SE b β T p Part

value

Model 1

Constant

WHtR

WTR DEXA % total body fat

63.738

-55.845

-2.625

.489

6.098

10.330

3.307

.095

-.251

-.030

.218

10.452

-5.406

-.794

5.170

< .001

< .001

.427

< .001

-.164

-.024

.157

Model 2

Constant

WHtR DEXA % total body fat

60.203

-60.654

.523

4.166

8.366

.084

-.273

.234

14.452

-7.250

6.217

< .001

< .001

< .001

-.220

.189 Note. WHtR = waist-height ratio; WTR = waist-thigh ratio; DEXA = dual energy x-ray absorptiometry.

128

Results of Exploratory Analyses on BMI Category by Gender Interaction on DSST Score

In exploratory analyses, a series of 2x3 ANOVAs were conducted to investigate main and

interaction effects of BMI weight category (i.e., normal weight, overweight, obese) and gender

on current DSST score for each of the following three timepoints: current, 10 years ago, and age

25. These are described below.

First, regarding the current timepoint, results revealed a significant main effect for

gender, F (1, 1120) = 20.880, p < .001 and a trend toward a significant main effect for current

weight category, F (2, 1120) = 2.919, p = .054. More specifically, men scored significantly

lower on the DSST (M = 40.4, SD = 18.2) than women (M = 44.4, SD = 18.6). Additionally, the

currently obese group tended to score lower on the DSST (M = 40.7, SD = 18.0) than the

currently overweight (M = 43.4, SD = 18.1) and currently normal weight (M = 43.0, SD = 19.7)

groups, though these differences are not assumed to be significant, as p values were not

calculated for these group differences. The interaction effect between gender and current weight

category was statistically significant, F (2, 1120) = 9.753, p < .001. More specifically, there was

a statistically significant simple effect of gender within the normal weight group, F (1, 1114) =

33.039, p < .001, but not within the overweight or obese groups. Fischer’s LSD post hoc tests

revealed that currently normal weight females scored significantly higher on the DSST (M =

49.3, SD = 18.2) than currently normal weight males (M = 37.1, SD = 18.8). Further, as Figure

4a reveals, the mean DSST score of the female group decreased when the weight category

increased from normal weight to overweight, whereas the mean DSST score of the male group

increased during that interval.

129

Figure 4a. DSST scores by current BMI category and gender. Significant interaction effect

between gender and current weight category on DSST score, p < .001. Additionally, there is a

significant main effect for gender, p < .001, and a trend toward significance for current weight

category, p = .054.

130

Second, regarding the late middle age timepoint (i.e., 10 years ago), results revealed

significant main effects for gender, F (1, 1098) = 13.740, p < .001 and late middle age weight

category, F (2, 1098) = 11.415, p < .001. More specifically, men scored significantly lower on

the DSST (M = 40.8, SD = 18.1) than women (M = 44.8, SD = 18.3). Additionally, Fischer’s

LSD post-hoc tests revealed that the late middle age obese group scored significantly lower on

the DSST (M = 38.2, SD = 17.7) than the late middle age overweight (M = 43.2, SD = 17.7) and

the late middle age normal weight (M = 45.5, SD = 18.3) groups. The interaction effect between

gender and late middle age weight category was statistically significant, F (2, 1098) = 6.207, p =

.002. There was a statistically significant simple effect of gender within the normal weight

group, F (1, 1092) = 24.328, p < .001, but not within the overweight or obese groups. Fischer’s

LSD post hoc tests revealed that females who were normal weight in late middle age scored

significantly higher on the DSST (M = 49.8, SD = 18.3) than males who were normal weight in

late middle age (M = 40.3, SD = 18.4). Moreover, as Figure 4b reveals, the mean DSST score of

the female group decreased when the weight category increased from normal weight to

overweight, whereas the mean DSST score of the male group increased during that interval.

131

Figure 4b. DSST scores by late middle-age BMI category and gender. Significant interaction

effect between gender and late middle-age BMI category (i.e., BMI category 10 years ago) on

current DSST score, p = .002. Additionally, there are significant main effects for gender, p <

.001, and late middle-age weight category, p < .001, on current DSST score.

132

Finally, regarding the age 25 timepoint, results revealed significant main effects for

gender, F (1, 999) = 4.255, p = .039 and age 25 weight category, F (2, 999) = 5.782, p = .003.

More specifically, men scored significantly lower on the DSST (M = 41.2, SD = 17.9) than

women (M = 45.4, SD = 17.6). Additionally, Fischer’s LSD post-hoc tests revealed that the age

25 normal weight group scored significantly higher on the DSST (M = 44.2, SD = 17.6) than the

age 25 overweight (M = 40.8, SD = 18.2) and the age 25 obese (M = 35.6, SD = 17.8) groups.

No other group differences were statistically significant. See Figure 4c.

133

Figure 4c. DSST scores by young adult BMI category and gender. Non-significant interaction

effect between gender and young adult BMI category (i.e., BMI category at age 25) on current

DSST score. Significant main effects for gender, p = .039, and young adulthood weight

category, p = .003, on current DSST score.

134

Hypothesis 8a–8c Results

Eight groups were created based on whether participants were obese at age 25, 10 years

ago (ages 55–64), and currently (ages 65–74), as depicted in Table 6. It was hypothesized that

participants who were obese at age 25 and 10 years ago would perform the worst on the DSST,

whereas those with no history of obesity would perform best on the DSST. It was also

hypothesized that those who were obese at only one timepoint would perform better than those

who were obese at age 25 and 10 years ago.

To test these hypotheses, a one-way ANOVA with eight groups at alpha level .05 with

weight trajectory group as the independent variable and DSST score as the dependent variable

was planned. Prior to this analysis, however, it was discovered that three of the eight groups

contained very few participants, as shown in Table 6. The decision was made to remove these

groups from this analysis rather than combine them with similar groups because it was unclear as

to which other group, if any, was most similar to each of the three groups in question. For

example, the mean DSST score of Group 4 was extremely low and may suggest that the

individuals in this group are relatively unique, perhaps due to illness, and could not be

appropriately combined into another group for this analysis. After removal of these groups, five

groups remained. Means and standard deviations for each of the new and original groups are

included in Table 6.

135

Table 6

Means and Standard Deviations of DSST Score by Weight Trajectory Group

Hypothesis 8 Group Obese at

age 25

Obese 10

years ago

Obese

currently n

DSST score

Original New M SD

1 A Y Y Y 29 37.2 18.0

2 __ Y Y N 7 34.1 18.6

3 __ Y N Y 2 39.5 31.8

4 __ Y N N 5 26.6 12.7

5 B N Y Y 152 41.9 16.6

6 C N Y N 54 32.8 16.7

7 D N N Y 163 43.7 17.3

8 E N N N 558 45.4 17.9

Note. Groups removed from Hypothesis 8 analyses due to low sample size are shaded in grey.

136

Given the results of preliminary analyses that showed a statistically significant

relationship between gender and DSST score, conducting an ANCOVA rather than an ANOVA

was considered. However, the relationship between the covariate (i.e., gender) and predictor

variable (i.e., weight trajectory group) would have violated the ANCOVA assumption of

independence of the covariate and treatment effect (Miller & Chapman, 2001). Therefore, the

decision was then made to conduct the one-way ANOVA as planned, with the addition of an

exploratory analysis to assess for interaction effects between gender and weight trajectory group

on DSST score.

A one-way ANOVA with five groups with weight trajectory group as the independent

variable and DSST score as the dependent variable was conducted. Results showed a significant

effect of weight trajectory group on DSST score, F (4, 951) = 7.851, p < .001. Post hoc

comparisons using the Tukey-HSD test revealed that Group C scored significantly lower on the

DSST than Groups B, D, and E. No other group differences were statistically significant, though

a slight trend toward significance for Group A scoring lower than Group E was noted, p = .102.

In an exploratory analysis, a two-way ANOVA conducted to investigate interaction effects

between gender and weight trajectory group on DSST score failed to reveal a statistically

significant interaction, F (4, 946) = .577, p = .679.

Group differences in mean processing speed between participants with history of

persistent obesity (i.e., obesity in both early adulthood and late middle-age) and participants with

history of persistent non-obesity (i.e., obesity in neither early adulthood nor late middle-age)

were also investigated in exploratory analyses. The persistent obesity group consisted of a

combination of the original Groups 1 and 2, whereas the persistent non-obesity group consisted

of a combination of the original Groups 7 and 8. Results of an independent samples t-test

137

conducted to compare mean DSST scores of those with persistent obesity (M = 36.6, SD = 17.9)

to those with persistent non-obesity (M = 45.0, SD = 17.8) revealed a statistically significant

difference between groups, t (755) = -2.770, p = .006. Other exploratory analyses showed that

persistently obese males (M = 33.1, SD = 17.5) had significantly lower DSST scores than

persistently non-obese males (M = 42.9, SD = 18.1), t (409) = -2.423, p = .006. No statistically

significant difference in DSST score between persistently obese females (M = 41.4, SD = 18.0)

and persistently non-obese females (M = 47.4, SD = 17.1) was found, t (344) = -1.326, p = .186;

however, caution was taken when interpreting results of this analysis, as the number of

participants in the persistently obese female group (n = 15) was lower than the minimum group

size of 26 needed to run a two-tailed t-test at alpha level .05 determined by an a priori power

analysis.

Hypothesis 8d Results

It was hypothesized that the percent of weight gained between early adulthood (age 25)

and late middle-age (age 55–64) would be associated with worse performance on the DSST. To

test this hypothesis, a Pearson product-moment correlation coefficient was calculated for DSST

score and percent weight gain between early adulthood and late middle-age. Results of this

analysis uncovered a statistically significant weak negative correlation between DSST score and

percent weight gain between early adulthood and late middle-age, r (805) = -.14, p < .001.

In exploratory analyses, the relationship between performance on the DSST and extent of

weight gain (i.e., absolute weight gain in pounds) between early adulthood and late middle-age

was assessed. Results of a Pearson product-moment correlation coefficient revealed a

statistically significant weak negative relationship between DSST score and extent of weight

gain between early adulthood and late middle-age, r (805) = -.14, p < .001. Next, the

138

relationship between performance on the DSST and extent of weight loss (i.e., absolute weight

loss in pounds) between early adulthood and late middle-age was assessed. The correlation

between DSST score and extent of weight loss between early adulthood and late middle-age was

not significant, r (83) = .03, p = .80.

Hypothesis 8e Results

It was hypothesized that the absolute value of the percent of weight change over the

previous 10-year period (i.e., from late middle-age to young-old) would be associated with worse

performance on the DSST. To test this hypothesis, a Pearson product-moment correlation

coefficient was calculated for DSST score and the absolute value of percent of weight change

over the previous 10-year period, revealing a statistically significant weak negative relationship,

r (1096) = -.11, p < .001.

In exploratory analyses, a Pearson product-moment correlation coefficient was calculated

for DSST score and the extent of weight change over the previous 10 year period (i.e., absolute

value of weight change in pounds), revealing a statistically significant weak negative correlation

between these two variables, r (1096) = -.13, p < .001. Exploratory analyses also showed

statistically significant weak inverse relationships between performance on the DSST and the

extent of weight loss (i.e., absolute value of weight loss in pounds) over the previous 10 year

period, r (409) = -.20, p < .001, as well as between performance on the DSST and the extent of

weight gain (i.e., absolute value of weight gain in pounds) over the previous 10 year period, r

(685) = -.13, p < .001.

Hypothesis 8f Results

It was hypothesized that individuals with unintentional significant past-year weight loss

would perform worse on the DSST than those who achieved intentional significant past-year

139

weight loss and those who were past-year weight-stable. No specific hypotheses were made

regarding how participants with significant past-year weight gain would perform on the DSST

relative to the other three groups. A one-way ANOVA was run with past-year weight group as

the independent variable (four levels) and DSST score as the dependent variable, revealing a

significant effect of past-year weight group on DSST score [F (3, 1022) = 12.53, p < .001]. Post

hoc comparisons using the Tukey-HSD test indicated that the unintentional significant past-year

weight loss group scored significantly lower on the DSST (n = 65, M = 31.0, SD = 17.1) than the

intentional significant past-year weight loss group (n = 83, M = 41.0, SD = 17.9), the significant

past-year weight gain group (n = 187, M = 42.1, SD = 18.8), and the past-year weight-stable

group (n = 691, M = 44.8, SD = 17.6). No other group differences were statistically significant.

Hypothesis 9 Results

It was hypothesized that there would be a significant negative correlation between CRP

level and DSST score and that this correlation would remain significant after controlling for

health and smoking confounds. In contrast to the predicted outcome, a Pearson product-moment

correlation coefficient calculated for CRP level and DSST score failed to reveal a statistically

significant relationship between these two variables, r (1027) = -.045, p = .153.

Hypothesis 10a Results

It was hypothesized that individuals who endorsed one or more of the following

conditions would perform worse on the DSST than those who did not: problems of blood flow to

the heart typically leading to or involving heart muscle damage (i.e., congestive heart failure

[CHF], coronary heart disease [CHD], or heart attack), stroke, diabetes, weak/failing kidneys,

chronic respiratory disease (i.e., chronic bronchitis or emphysema), and thyroid disease (if

untreated). As previously explained, upon implementation of the present study, it was

140

discovered not to be possible to determine whether thyroid disease had been treated; therefore,

thyroid disease was not included in this analysis. Results of an independent samples t-test

conducted to compare mean DSST scores of those with none of these health conditions (M =

44.8, SD = 18.9) to those with ≥ 1 of these health conditions (M = 38.4, SD = 17.3) showed a

statistically significant difference between groups, t (1113) = -5.642, p < .001.

In exploratory analyses, a series of t-tests were conducted in order to investigate the

relationship between individual self-reported medical conditions and DSST score. Results of

these analyses are listed in Table 7. In addition, a one-way ANOVA was run with cardiac

diagnostic group as the independent variable (four levels: no history of CHF, CHD, or heart

attack; history positive for CHF; history positive for only CHD; history positive for only heart

attack) and DSST score as the dependent variable, revealing a significant effect of cardiac

diagnostic group on DSST score [F (3, 1090) = 5.91, p = .001]. Post hoc comparisons using

Tukey’s HSD test indicated that the group without history of any of these cardiac diagnoses

scored significantly higher on the DSST (M = 42.9, SD = 18.7) than the group with history of

CHF (M = 35.6, SD = 17.1; p = .013) and the group with history positive for only heart attack

(M = 33.2, SD = 15.5; p = .014). No other group differences were statistically significant.

141

Table 7

Results of t-tests Comparing Groups on DSST Score by Self-reported History of Medical

Condition

M SD df T p

Stroke

Yes

No

34.5

42.7

16.7

18.5

1137

-3.555

< .001

Diabetes

Yes

No

36.0

43.6

16.5

18.8

1113

-5.308

< .001

Hypertension

Yes

No

39.6

44.5

17.8

18.8

1136

-4.452

< .001

Weak or failing kidneys

Yes

No

35.6

42.6

16.0

18.6

1136

-2.746

.006

Cardiac

Yes

No

39.3

42.9

17.0

18.7

1129

-2.349

.019*

Chronic respiratory disease

Yes

No

39.5

42.5

17.0

18.6

1138

-1.638

.102

Any liver condition

Yes

No

44.3

42.2

17.0

18.6

1136

.670

.503

Note. *Did not remain statistically significant after Bonferroni correction for alpha inflation. Cardiac = self-reported

history of congestive heart failure, coronary heart disease, and/or heart attack. Chronic respiratory disease = chronic

emphysema and/or chronic bronchitis.

142

Hypothesis 10b Results

It was hypothesized that duration of respiratory disease would be negatively correlated

with processing speed task performance. In contrast to the predicted outcome, a Pearson

product-moment correlation coefficient calculated for duration of respiratory disease and DSST

score revealed a nonsignificant positive relationship between these two variables, r (85) = .178, p

= .104.

Hypothesis 11 Results

It was hypothesized that individuals with dyslipidemia (i.e., high levels of total

cholesterol, low levels of HDL cholesterol, high levels of LDL cholesterol, and/or high levels of

triglycerides) would demonstrate worse processing speed. To test this hypothesis, an

independent samples t-test was conducted to compare DSST scores of those with dyslipidemia

(M = 42.2, SD = 18.4) and without dyslipidemia (M = 44.4, SD = 18.4), which failed to reveal a

statistically significant difference between groups, t (543) = -1.306, p = .192.

In exploratory analyses, the relationship between each of the above types of dyslipidemia

(i.e., unhealthy levels of total cholesterol, HDL cholesterol, LDL cholesterol, and triglycerides)

and DSST score was investigated. The relationship between DSST and high levels of non-HDL

cholesterol was also examined. A series of independent samples t-tests using a two-tailed test of

significance at alpha level .05 was conducted to compare participants with and without each

dyslipidemia type on DSST score. Results showed significantly lower DSST scores in those

with low HDL cholesterol (M = 39.9, SD = 18.5) compared to those without low HDL

cholesterol (M = 44.4, SD = 18.2), t (1018) = -3.699, p < .001. Significantly lower DSST scores

were also found among those with high non-HDL cholesterol (M = 40.6, SD = 18.8) compared to

those without high non-HDL cholesterol (M = 43.5, SD = 18.4), t (1018) = -2.045, p =.041.

143

Additionally, results showed a trend toward significance for lower DSST scores in those with

high LDL cholesterol (M = 39.7, SD = 19.1) compared to those without high LDL cholesterol

(M = 43.9, SD = 18.1), t (1018) = -1.859, p = .064. In contrast to what would have been

expected, results showed significantly higher DSST scores in those with high triglycerides (M =

46.8, SD = 16.8) compared to those without high triglycerides (M = 42.1, SD = 18.6), t (459) =

2.332, p = .020. Further, results failed to show a significant difference between DSST scores in

those with high total cholesterol compared to those without high total cholesterol, t (1018) = -

.585, p = .559. After the Bonferroni correction was applied to prevent alpha inflation, mean

group DSST scores only remained significantly different between participants with and without

low HDL cholesterol.

Additional analyses were later used to explore whether the above relationships differed

by sex. Results showed sex-specific relationships between DSST score and two types of

dyslipidemia: high non-HDL cholesterol and high triglycerides. Namely, men with high non-

HDL cholesterol scored significantly lower on the DSST (M = 36.8, SD = 17.2) than men

without high non-HDL cholesterol (M = 41.9, SD = 18.4), t (517) = -2.442, p = .015, but the

same relationship was not found among women, t (499) = -1.013, p = .312. Moreover, women

with high triglycerides scored significantly higher on the DSST (M = 51.0, SD = 17.2) than

women without high triglycerides (M = 43.3, SD = 18.3), t (227) = 2.725, p = .007, but the same

relationship was not found among men, t (230) = .486, p = .628.

Hypothesis 12a Results

It was hypothesized that risk of first stroke, as indicated by score on a modified version

of the FSRP, would be inversely associated with processing speed task performance. A Pearson

product-moment correlation coefficient calculated for modified FSRP score and DSST score

144

revealed a statistically significant weak negative correlation between these two variables, r (216)

= -.136, p = .045.

Hypothesis 12b Results

It was hypothesized that risk of onset of CVD, as indicated by score on a modified

version of the FRS for general CVD, would be inversely associated with processing speed task

performance. A Pearson product-moment correlation coefficient calculated for modified FRS for

general CVD score and DSST score revealed a statistically significant weak negative correlation

between these two variables, r (171) = -.205, p = .007.

Hypothesis 13 Results

It was hypothesized that there would be a significant positive correlation between speed

of usual gait and processing speed and that this correlation would remain significant after

controlling for BMI. Speed of usual gate (ft/s) was calculated using results from the 20-foot-

walk test. A Pearson product-moment correlation coefficient calculated for gait speed and DSST

score revealed a statistically significant moderate positive correlation between these two

variables, r (999) = .369, p < .001. Next, a partial correlation was run to determine the

relationship between gait speed and DSST score while controlling for BMI. A statistically

significant moderate positive correlation between gait speed and DSST score remained after

controlling for BMI, r (998) = .366, p < .001. A comparison of the zero-order correlation and

partial correlation suggests that BMI had very little influence in controlling for the relationship

between gait speed and DSST score. Results indicate that faster usual gait speed is moderately

associated with faster processing speed regardless of BMI.

145

Final Model

The originally proposed theoretical model of the relationship between persistent obesity

and DSST score, as depicted in Figure 1, required extensive revision prior to testing due to

practical limitations, results of the previously described analyses, and theoretical considerations.

Most notably, it was deemed inappropriate to hypothesize that persistent obesity (which refers to

the span of time from young adulthood to late middle-age in the present study) would have a

bidirectional relationship with current habits and characteristics (e.g., diet, physical activity,

cholesterol levels), as there was a 10-year gap between late middle-age and the time of data

collection (i.e., young adulthood); therefore, these observed variables were removed from the

original model. Additionally, a number of proposed latent constructs (i.e., depression, physical

inactivity history, diet/nutrient history) were removed from the model due to lack of available

indicators to represent them. In other words, this information was not collected as part of the

NHANES cycles under analysis. Likewise, one proposed indicator of “brain oxygen flow

disruption/ insufficiency” (i.e., chronic hypoventilation during sleep or awake periods) was

removed due to this data not having been collected. Furthermore, some observed variables (i.e.,

chronic respiratory disease, peripheral inflammation) were removed from the model due to the

lack of significant relationship with DSST score found in earlier analyses. The remaining

variables were then examined for theoretical considerations. It was noted that each of the

remaining variables were vascular risk factors that conceivably could have a) been partially

“caused” by a history of persistent obesity and b) contributed to a decline in processing speed.

Another consideration relating to the decision to proceed with a model that focused on vascular

health history was the understanding that vascular risk factors are more theoretically relevant to a

decline in processing speed than risk factors associated with only Alzheimer disease (AD), as

146

deficits of processing speed and executive functioning are typical of a vascular pattern of

cognitive impairment whereas deficits in memory and language are typical of cognitive

impairment seen in AD, particularly in this age group when AD is less likely to be advanced

stages in which additional domains tend to become impaired (American Psychiatric Association,

2013; Gasecki, Kwarciany, Nyka, & Narkiewicz, 2013; Lezak, Howieson, Bigler, & Tranel,

2012). As a result of the above considerations, a measurement model of vascular health history

was hypothesized using the remaining variables as indicators, as depicted in Figure 5.

Additionally, a structural model in which vascular health history partially mediates the

relationship between persistent obesity and DSST score was hypothesized, as shown in Figure 6.

Note that cardiac history in Figure 5 refers to a binary variable that reflects the presence or

absence of self-reported CHD, CHF, and/or heart attack, which is consistent with its use in

previous analyses in the present study.

147

Hypertension history

Stroke history

Diabetes history

Cardiac history

Vascular health

history

Figure 5. Hypothesized measurement model of vascular health history. Error terms

not shown.

148

DSST score

Vascular health

history

Persistent obesity

Figure 6. Revised hypothesized structural model of DSST score. Error terms not shown.

149

The hypothesized measurement model for vascular health history was tested using

confirmatory factor analysis (CFA) prior to assessment of the hypothesized structural model.

The hypothesized measurement model included one latent variable (i.e., vascular health history)

and four observed indicators (i.e., self reported history of hypertension, stroke, diabetes, and

cardiac event/ condition). Results of the CFA for vascular health history revealed an acceptable

fit between the measurement model and the present study’s data [χ2 (2, 1380) = 1.404, p = .496;

CFI = 1.000; RMSEA = .000 (90% CI = .000–.048)]. Further, the standardized factor loadings

for each of the four indicators exceeded the cut-off criterion of > .35 and, therefore, were

retained in the measurement model. See Table 8 for complete information on factor loadings.

150

Table 8

Confirmatory Factor Analysis for Vascular Health History

Estimate SE p

Stroke 0.57 0.08 0.000

Hypertension 0.66 0.07 0.000

Diabetes 0.45 0.06 0.000

Cardiac 0.44 0.06 0.000

Note. Standardized regression weights and two-tailed p values are reported.

151

Next, the revised hypothesized structural model of DSST score, as previously depicted in

Figure 6, was tested, revealing inadequate fit with the present study’s data [χ2 (8, 874) = 15.409,

p = .052; CFI = 0.949; RMSEA = 0.033 (90% CI = .000–.057)]. See Figure 7 and Table 9 for

results regarding individual factor loadings of the revised hypothesized structural model.

152

Figure 7. Revised hypothesized structural model of DSST score with standardized regression

weights. Error terms not shown. **p < .001.

DSST score

Vascular health

history

Persistent obesity

.35**

.00

.29**

153

Table 9

Factor Loading Summary of Revised Hypothesized Structural Model of DSST Score

Estimate SE Estimate/SE p

DSST score on:

Vascular health history 10.27 (0.35) 2.50 4.11 0.000

Persistent obesity 0.05 (0.00) 3.91 0.01 0.990

Vascular health history on:

Persistent obesity 0.82 (0.29) 0.17 4.81 0.000

Note. Standardized estimates are in parentheses. Two-tailed p values are reported.

154

Modification indices and factor loadings were then examined to determine if an

adjustment to the structural model could be made to improve model fit in the event that the

adjustment made theoretical sense. In doing so, the direct path between persistent obesity and

DSST score was removed. The resulting structural model, as depicted in Figure 8, was then

tested, revealing adequate fit with the present study’s data [χ2 (9, 874) = 14.661, p = .101; CFI =

0.961; RMSEA = 0.027 (90% CI = .000–.051)]. See Figure 9 and Table 10 for results regarding

individual factor loadings of the final model.

155

Figure 8. Final structural model of DSST score. Error terms not shown.

DSST score Persistent obesity Vascular

health history

156

Figure 9. Final structural model of DSST score with standardized regression weights. Error

terms not shown. **p < .001.

DSST score Persistent obesity Vascular

health history

.37** .29**

157

Table 10

Factor Loading Summary of Final Structural Model of DSST Score

Estimate SE Estimate/SE p

DSST score on:

Vascular health history 10.93 (0.37) 2.34 4.68 0.000

Vascular health history on:

Persistent obesity 0.81 (0.29) 0.16 5.03 0.000

Note. Standardized estimates are in parentheses. Two-tailed p values are reported.

158

Discussion

Hypothesis 1 Discussion

It was hypothesized that there would be a strong negative correlation between age and

processing speed task performance, as indicated by Digit Symbol Substitution Test (DSST)

score. This hypothesis was based on the consensus that declines in most cognitive domains

occur with aging even in the absence of known health problems as well as research that supports

the notion that processing speed is among the most important factors in explaining normal age-

related cognitive decline (e.g., Aine et al., 2011; Harada et al., 2013; Hasher et al., 1991; Rozas

et al., 2008; Vance, 2009).

This hypothesis was partly supported by the data. Age was found to have a significant

inverse relationship with DSST score, which suggests that processing speed task performance

declines as age increases within the young-old age category. This finding is consistent with

previous research on other tests of processing speed within this age category, including the

Wechsler Adult Intelligence Scale, Fourth Edition (WAIS-IV) Processing Speed Index (PSI)

subtests Coding and Symbol Search (Wisdom, Mignogna, & Collins, 2012).

The strength of the relationship between age and DSST score in the present study was

weak, whereas it was predicted to be strong. This finding may be related to the cross-sectional

nature of the present study, as longitudinal effects have been shown to be larger than cross-

sectional effects for the Wechsler Adult Intelligence Scale-Revised (WAIS-R) DSST and a

variety of other cognitive measures among a sample of older adults aged 66–92 at baseline who

were followed for up to five 18-month intervals (Sliwinski & Buschke, 1999). In other words,

cross-sectional age differences in processing speed may be less obvious than longitudinally-

measured age-related changes in this population (Sliwinski & Buschke, 1999).

159

Additionally, evidence of increased intragroup variability on tests of processing speed as

age increases, as captured by Wisdom, Mignogna, and Collins (2012) with the use of the

coefficient of variation (CV), should be considered when interpreting these results. The CV,

which is calculated by dividing the standard deviation by the mean and multiplying by 100, has

utility as a measure of score dispersion above and beyond that of standard deviation, the latter of

which has been shown to vary relatively little on the WAIS-IV PSI subtests for participants

within this age category (Wisdom et al., 2012). In other words, in regard to a variable whose

mean decreases with age but whose standard deviation remains stable with age (similar to the

DSST variable in the present study), the variable will show increased score dispersion with age,

which would then yield a weaker correlation between that variable and age.

On the other hand, it is possible that the relationship between age and processing speed

would be stronger in a population older than the present study’s sample of 65–74-year olds.

Salthouse (2012) presented data suggestive of stability in rate of decline in processing speed in

adults up to age 75, but future reseach is needed to understand whether this rate of decline

remains stable, or perhaps accelerates, beyond age 75. Finally, the finding that age was only

weakly correlated with DSST score in the present study highlights the importance of other

contributors to processing speed task performance in this age group.

Hypothesis 2 Discussion

It was hypothesized that age would show a significant positive association with % body

fat as measured by dual energy x-ray absorptiometry (DEXA) but would show a slight, negative,

statistically insignificant correlation with body mass index (BMI). Likewise, it was hypothesized

that visceral adiposity estimated by % trunk fat DEXA would increase significantly with age but

waist circumference (WC) would not be significantly correlated with age. These hypotheses

160

were based on a number of interrelated findings that highlight the changing body composition

with age as well as the discrepancy between this group’s body composition markers and obesity

indexes. The hypotheses involving anthropometrically derived obesity indexes (i.e., BMI, WC)

were largely supported by the data, whereas those involving DEXA-derived obesity indexes (i.e.,

% body fat DEXA, % trunk fat DEXA) were not.

The hypothesis that % total fat DEXA would increase with age had been informed by a

study that showed increase in bioelectrical impedance analysis (BIA)-derived percentage of total

body fat with age across the adult lifespan (Kyle, et al., 2001). The present study, however, did

not find a statistically significant relationship between age and % total fat DEXA among its

young-old participants. The absence of this finding may be related to the specific age decade

under investigation (ages 65–74) and the timing of the onset of significant age-related decline in

fat-free mass. Namely, Obisesan and colleagues (2005) found that declines in fat-free mass

typically began during one’s 50s but did not become significant until ages 65–69 in Black men,

70–74 in Black women and White men, and 75–79 in White women. It may be the case that

many participants in the present study had not yet experienced significant decline in fat-free

mass, a change in body composition that would have contributed to increased % total fat DEXA

values. Another possible explanation for the apparent discrepancy between the results of Kyle

and colleagues (2001) and the present study pertains to the health of their respective participants,

as the former attempted to exclude individuals with acute or chronic diseases (a number of which

are known risk factors for dementia). Because dementia risk increases with age and individuals

who go on to develop dementia are more likely to have lost a significant amount of weight

between midlife and late-life compared to individuals who remain free of dementia, it may be the

161

case that the present study contained more individuals at relatively higher ages with recently

reduced body weights (Stewart et al., 2005).

The hypothesis that % trunk fat DEXA would increase with age in the young-old

population had been informed by results of a cross-sectional study reported by Oldroyd and

colleagues (1998), which showed increased % trunk fat DEXA in males over the age of 50 (M =

59.2 years) and postmenopausal women (M = 61.6 years) relative to males below the age of 50

(M = 37.5 years) and premenopausal women (M = 37.2 years), respectively. Results of the

present study suggest that the increase in abdominal fat that appears to occur between

approximately midlife and late midlife may not continue to occur during the young-old portion

of the lifespan (i.e., 65–74 years of age). This idea is supported by findings from a recent cross-

sectional study of women using BIA, which found statistically significant increases in visceral

fat area during the third, fourth, and fifth decades of life but statistically insignificant increases in

visceral fat area in the sixth decade of life and beyond (Gába & Přidalová, 2014).

Hypothesis 3 Discussion

The hypothesis that participants with probable lifetime history of excess alcohol

consumption would exhibit worse processing speed task performance than participants without

this history was supported by the data. This finding is not surprising given that individuals with

a history of chronic alcohol use are at risk for alcohol-induced neurocognitive disorders. In

addition, there is evidence of impaired cognition during periods of abstinence subsequent to

chronic alcohol use, mostly with regard to tasks dependent on the hippocampus and/or prefrontal

cortex (e.g., working memory and executive functioning; Staples & Mandyam, 2016). Further,

history of chronic alcohol use is associated with other lifestyle variables that may be

independently linked to worsened processing speed task performance, such as those of smoking.

162

The present study adds information to the current body of literature on associations between

history of chronic excess alcohol use and processing speed task performance in a nationally

representative sample of young-old adults.

The hypothesis that participants who currently smoke would perform worse on the DSST

than those who do not currently smoke was supported by the data. Moreover, exploratory

analyses showed that everyday smokers performed worse on the DSST than nonsmokers,

whereas occasional smokers did not differ significantly from either group. Results indicate that

smoking everyday is associated with worsened processing speed in adults aged 65–74 compared

to not smoking. This finding was expected, as smoking has been hypothesized to lead to

impaired cognition through the initiation of atherogenesis and subsequent development of

atherosclerosis, which can lead to cardiovascular and/or cerebrovascular diseases that negatively

impact cognition (Messner & Bernhard, 2014). It is also possible that daily smoking may

decrease lung functioning, which in turn may discourage individuals from participation in

regular aerobic exercise, which itself is linked with improved cognitive outcomes and decreased

risk of cardiovascular and cerebrovascular diseases. Previous research has found that, among

women aged 40–79, current smokers and previous smokers scored lower on the WAIS-III PSI

than never-smokers (Lo et al., 2014). The present study’s findings add information to the current

body of literature on associations between frequency of current smoking behavior and processing

speed task performance in a nationally representative sample of young-old adults.

Hypothesis 4 Discussion

Participants were divided into the following groups for this set of hypotheses: participants

who did not engage in regular exercise (no exercise group); participants who engaged in both

regular aerobic exercise and regular strength training exercise (combination training group);

163

participants who engaged in regular aerobic exercise only (aerobic-only group); and participants

who engaged in regular strength training exercise only (strength training-only group). Among

these groups, it was hypothesized that the no exercise group would perform worst on the DSST,

whereas the combination training group would perform the best and the aerobic-only group

would outperform the strength training-only group.

These hypotheses were based on results of experimental studies showing the benefit of

immediate and regular aerobic exercise on processing speed and other aspects of cognition (e.g.,

Nouchi et al. 2014; Tam, 2013). Additionally, they were informed by an experimental study that

found evidence for a program of aerobic exercise to selectively improve performance of

executive control processes supported by the frontal and prefrontal regions of the brain (e.g.,

switching between tasks, ability to disregard task-irrelevant stimuli, ability to discontinue a

planned action), which is relevant to the present study given that the DSST involves elements of

executive functioning and has been shown to activate the frontal lobe (Davis & Pierson, 2012;

Nakahachi et al., 2008). Further, these hypotheses were informed by results of a meta-analysis

on the effects of exercise on the cognitive functioning of older adults that yielded the finding that

those in combination training programs demonstrated improvements in cognitive outcomes to a

greater degree than those in aerobic training groups (Colcombe & Kramer, 2003).

This group of hypotheses was partially supported by the results of the present study. As

predicted, the present study found that, in regard to the DSST, the combination training group

outperformed each of the other three exercise groups, and the aerobic-only group outperformed

the no-exercise group. However, in contrast to what was hypothesized, the strength training-only

group did not perform in a significantly different manner on the DSST than the aerobic-only

group or the no exercise group. Results suggest that, in this nationally representative sample of

164

adults aged 65–74, regular strength training may have an additive effect on regular aerobic

exercise with regard to its association with improved processing speed task performance but that

regular strength training alone does not have a significant association with processing speed task

performance relative to that of no exercise. Interestingly, results from two recent 24-week,

randomized, double-blind, placebo-controlled trials that tested the effect of strength training (vs.

no exercise training) with or without protein supplementation in older adults found that strength

training with protein supplementation improved information processing speed relative to no

exercise with protein supplementation, but the same benefits of strength training were not found

among participants who did not receive protein supplementation (van de Rest et al., 2014). This

suggests the possibility of diet having a moderating role with regard to the impact of strength

training exercises on information processing speed outcomes. Indeed, many tasks created to

measure processing speed have a strong motor component, and it would seem that increased

muscle strength (which requires adequate dietary protein to achieve) could have a beneficial

impact on motor speed; however, efforts taken by van de Rest and colleagues (2014) to control

for motor speed on performance outcomes suggest that the benefits of strength training on

information processing speed in those who received protein supplementation is not simply due to

increased motor speed secondary to strengthened muscles. Future research is needed to clarify

the role of diet and nutrition in regard to cognitive outcomes of exercise programs for older

adults.

Hypothesis 5 Discussion

Using the operational definitions for diabetes, prediabetes, and no diabetes specified in

the Hypotheses section, participants with diabetes were expected to perform worse on the DSST

than participants with prediabetes and participants without diabetes. In addition, participants

165

with prediabetes were expected to perform worse on the DSST than those without diabetes.

These hypotheses were informed by studies that have shown processing speed impairment in

those with type-2 diabetes mellitus (T2DM), as well as findings of similar impairment in those

with metabolic syndrome, suggesting that cognitive decline may develop in the early stages of

T2DM (e.g., van den Berg et al., 2008). The present study’s hypotheses were partially supported

by the results. Namely, the mean DSST score of the diabetes group was significantly lower than

both the prediabetes group and the no diabetes group; however, the mean DSST score of the

prediabetes group did not differ significantly from the no diabetes group.

Exploratory analyses were used to investigate the relationship between individual

markers of diabetes status and processing speed task performance. Concerning fasting blood

glucose group (i.e., diabetic range, prediabetic range, normal range), group membership did not

have a statistically significant effect on DSST score. Regarding HbA1c group (i.e., diabetic

range, prediabetic range, normal range), those in the diabetic range performed worst on the

DSST, those in the normal range performed best, and those in the prediabetic range performed

better than those in the diabetic range but worse than those in the normal range; each of these

group differences was statistically significant.

These results showed that individuals with diabetes demonstrated worse processing speed

task performance relative to individuals free of diabetes when HbA1c level was used to define

diabetes status but not when fasting blood glucose level was used to define diabetes status. It

may be the case that chronic hyperglycemia, as represented by HbA1c diabetic range, is involved

in the etiology of impaired processing speed in young-old adults or otherwise associated with

this impairment, whereas current hyperglycemia status itself is less relevant. Alternatively, it is

possible that measurements of HbA1c, a non-fasting test, were more accurate than measurements

166

of fasting blood glucose given that the latter would have been vulnerable to human error in recall

(e.g., inaccurate report of hours of fasting status).

Further, the finding that participants in the HbA1c prediabetic range performed better than

those in the HbA1c diabetic range but worse than those in the HbA1c normal range supports the

theory that diabetes-associated cognitive decline may begin developing in the early stages of the

disease (van den Berg et al., 2008). The present study’s finding that young-old adults with

prediabetic HbA1c levels demonstrate impaired processing speed relative to same aged peers with

normal HbA1c are consistent with previous research that has found this relationship among

middle-aged adults and adds to the current literature (Sanz et al., 2013; Young, Mainous, &

Carnemolla, 2006).

Exploratory analyses were also used to investigate the relationship between peripheral

insulin resistance and processing speed task performance. Participants with peripheral insulin

resistance, operationally defined as a homeostasis model assessment of insulin resistance

(HOMA-IR) score above the 75th percentile, performed significantly worse on the DSST than

participants without peripheral insulin resistance. This could be taken to support the theory that

cognitive decline develops in the early stages of T2DM, as peripheral insulin resistance proceeds

the onset of hyperglycemia in the development of T2DM; however, in the present study, the

majority of individuals with peripheral insulin resistance also had diabetes mellitus. This is not

surprising because average onset of T2DM during the years 1999 to 2002 ranged from 52 to 53

years of age, whereas the participants in the present study were aged 65–74 (Centers for Disease

Control and Prevention, 2017). The low number of participants who had peripheral insulin

resistance but were free of diabetes prevented investigation of the potential impact of peripheral

insulin resistance on processing speed task performance among non-diabetic young-old adults.

167

Hypothesis 6 Discussion

Homocysteine. Folate and vitamin B12 are needed to convert homocysteine to

methionine; therefore, homocysteine levels may increase when levels of folate or vitamin B12 are

depleted (Leishear, Ferrucci, et al., 2012). Vitamin B12 level and DSST score were not

significantly associated in the present study. It was, therefore, unnecessary to control for vitamin

B12 level in either the homocysteine or folate analyses. The discussion that follows focuses on

the relationship between homocysteine and processing speed as it relates to the potential impact

of folate.

Some research has found high homocysteine levels to be associated with reduced

information processing speed independent of folate levels, but results of other studies suggest

that high homocysteine levels are only associated with worse cognition in those with low folate

levels (Feng et al., 2013; Feng et al., 2006; Vidal et al., 2008). Results from the present study

indicate that, among a nationally representative sample of adults aged 65–74, homocysteine level

has a weak inverse relationship with processing speed and a very weak inverse relationship with

processing speed after controlling for folate level.

Although the above results were statistically significant, their practical value is low. In

contrast, hyperhomocysteinemia, defined as homocysteine level ≥ 12 μmol/L, showed a

moderate relationship to DSST score and may be of greater practical significance than

homocysteine level. This relationship was independent of folate status, as no statistically

significant interaction between hyperhomocysteinemia and folate depletion was found.

Results support the hypothesis that higher homocysteine level, and

hyperhomocysteinemia in particular, is associated with worse processing speed independent of

folate and vitamin B12 levels in young-old adults. In other words, reduced processing speed in

168

young-old adults with hyperhomocysteinemia is not simply a consequence of low folate or

vitamin B12 levels.

Results from the present study are consistent with findings from a recent study showing

that, among adults aged 65–80 with normal serum levels of folate and vitamin B12, DSST scores

of those with hyperhomocysteinemia were significantly lower than those without

hyperhomocysteinemia; the same relationship was not found among adults aged 50–64 in that

study, indicating that this relationship may be sensitive to age (Hsu et al., 2016). Moreover,

hyperhomocysteinemia was not significantly associated with performance in any other

neuropsychological domain, which adds to the body of literature that supports the idea that

reduced processing speed is the type of cognitive impairment most associated with

hyperhomocysteinemia (Hsu et al., 2016; Feng et al., 2006; Feng et al., 2013).

Whether hyperhomocysteinemia causes cognitive dysfunction or is a marker of this

dysfunction is uncertain. Hyperhomocysteinemia has been theorized to be a common risk factor

for many dementia subtypes (including Alzheimer disease [AD] and vascular dementia [VaD])

via disrupted cerebrovascular homeostasis (Kamut, Vacek, Kalani, & Tyagi, 2015). Specific

mechanisms of action suggested by research include, but are not limited to, facilitation of

oxidative distress, direct neurotoxic effects, increased risk for atherosclerosis and thrombosis via

damage to the vascular endothelium, and disruption of the blood-brain barrier leading to micro-

vascular degeneration and accumulation of toxic substances (Arnesen et al., 1995; Di Bonito et

al., 2007; Graham et al., 1997; Kamut et al., 2015).

Folate. There is a lack of consensus on the lower limit of serum folate reference values

in the literature. The 1999–2000 CDC Laboratory Procedure Manual defined folate deficiency

as serum folate < 1.4 ng/mL with regard to interpretation of results from the particular radioassay

169

kit that was used (CDC Laboratory Procedure Manual). In contrast, the current Merck Manual

states that folate deficiency is likely if serum folate is < 3 ng/mL (Johnson, 2016d). Further,

according to the World Health Organization (WHO), serum folate level that is < 3 ng/mL

indicates folate deficiency, whereas the 3 ng/mL to 5.9 ng/mL range indicates “possible folate

deficiency” (WHO, 2015). These criteria are generally consistent with results of an early

experimental study of the development of folate deficiency through a low folate diet that was

conducted by a physician on himself, which concluded that a persistent serum folate level of less

than 3 ng/mL led to folate deficiency, as evidenced by megaloblastic changes in his blood cells

(Herbert, 1962). More recently, the use of a higher cut-off value has also been argued, as some

metabolic changes associated with folate deficiency can be observed prior to changes in blood

cells (Green et al., 2008; De Bruyn, Gulbis, & Cotton, 2014).

The present study had selected an operational definition of folate deficiency to be serum

folate < 1.4 ng/mL in order to be consistent with the 1999–2000 CDC Laboratory Procedure

Manual. However, none of the participants in the present study met that criterion. Given that

folate and DSST score were found to have a statistically significant positive relationship

independent of homocysteine level in a previous analysis in the present study, the utility of an

alternate folate cut-off value was explored. Namely, a cutoff value of < 8 ng/mL was selected

for the present study’s operational definition of folate depletion on the basis of recent research,

with 9.5% of the present study’s sample meeting this criterion (De Bruyn, Gulbis, & Cotton,

2014). Those with folate depletion performed significantly worse on the processing speed task

than those without folate depletion independent of hyperhomocysteinemia status. Therefore,

based on results of the present study, folate depletion is of greater practical significance than

folate level in regard to its relationship to processing speed among young-old adults.

170

Results of the present study are not consistent with results of an earlier study of adults

aged 55–85 living in the Netherlands that failed to find a statistically significant difference in

processing speed task performance between participants with and without low folate level

(Jelicic et al., 2001). Relative to the present study, Jelicic and colleagues (2001) used a similar

folate cutoff (i.e., 7 ng/mL) and processing speed task (i.e., a letter substitution task). This

discrepancy highlights the need for additional research to clarify this relationship, as it is not yet

known whether low folate level has a causal or contributory relationship to slowed processing

speed in young-old adults and/or whether low folate level is a marker of another condition that

explains this association.

Folate is believed to be important to cognitive and emotional functioning. Symptoms of

folate deficiency can include depression and confusion (Johnson, 2016d; Reynolds, 2002).

Furthermore, some have speculated that low folate level contributes to an increased risk for

dementia (e.g., Reynolds, 2002). There is some evidence to support this idea. For example, one

large prospective study of older adults over a 2.4-year period found that baseline folate level, but

not baseline B12 or homocysteine levels, predicted incident dementia (Kim et al., 2008). During

the follow-up period of that study, those with incident dementia showed an accelerated decline in

folate and a steeper increase in homocysteine compared to those who did not develop dementia.

Moreover, one recent study using amyloid positron emission tomography (PET) scans found that

those whose imaging was positive for amyloid plaques had lower folate levels than those who

tested negative (Yoshinaga, Nishimata, Kajiya, & Yokoyama, 2017). Of course, these studies

are not experimental in nature and cannot show evidence for a causal role of folate in the

development of dementia.

171

Folate is also important to physical functioning, and a chronic low level of this vitamin

can cause medical problems that lead to cognitive and/or emotional dysfunction (Reynolds,

2002). For example, chronic alcohol use can impair intestinal absorption of folate, leading to

folate deficiency and subsequent development of alcoholic liver disease and its associated

cognitive impairments (Medici & Halsted, 2013). Additionally, folate assists in proper

deoxyribonucleic acid (DNA) methylation, without which cancer and other diseases are more

likely to develop (Crider, Yang, Berry, & Bailey, 2012; Duthie, 2011). Though most types of

cancers themselves are not believed to impair cognition, cancer treatments of systemic

chemotherapy and radiation localized to the brain may negatively impact processing speed and

executive functioning (Anstey et al., 2015; Avila et al., 2015; Mandelblatt et al., 2014).

Vitamin D. The hypothesis that participants with low 25-hydroxyvitamin D levels

would perform worse on the processing speed task than those with healthy vitamin D levels was

supported by the data. This finding is consistent with previous research among adults of various

ages (Karakis et al., 2016; Lee et al., 2009). The present study contributes to the literature by

reporting this relationship specifically within a nationally representative sample of young-old

adults. Findings that suggest that, among young-old adults, men’s processing speed tends to be

more negatively impacted by low vitamin D status than women’s processing speed are also

noteworthy.

It is unclear if, and to what extent, low 25-hydroxyvitamin D level contributes to slowed

processing speed or whether it is primarily a biomarker for one or more other conditions. This

question is complicated by the dual functioning of vitamin D through both endocrine (e.g.,

regulation of calcium and phosphate absorption) and intracellular (e.g., regulation of target gene

expression) signaling mechanisms (Vanherwegen, Gysemans, & Mathieu, 2017). Moreover, the

172

vitamin D receptor (VDR) and the enzyme needed for conversion of vitamin D to its active form

are both present in the human brain, as well as multiple other tissues throughout the human body

(Eyles, Smith, Kinobe, Hewison, & McGrath, 2005; Feldman & Malloy, 2014). Given this,

vitamin D functioning is likely to have a variety of bodily effects. Consistent with this idea is

empirical evidence that implicates vitamin D in skeletal, immune functioning, cardiovascular

health, and cancer outcomes (Wilson, Tripkovic, Hart, & Lanham-New, 2017).

There are, therefore, several possible explanations for the present study’s finding that

young-old adults with low vitamin D levels evidenced slower processing speed than young-old

adults with healthy vitamin D levels. Select possibilities relevant to the present study are briefly

discussed here. First, it may be the case that a certain level of 25-hydroxyvitamin D is needed in

order to support good vascular health, as low 25-hydroxyvitamin D has been associated with

endothelial dysfunction and arterial stiffness (Mheid et al., 2011). There is also evidence to

suggest that 25-hydroxyvitamin D may contribute to impaired glucose metabolism (Binkley,

Ramamurthy, & Krueger, 2010). These findings are consistent with the associations between

low 25-hydroxyvitamin D level and the two cognitive domains most likely to be impacted by

cerebrovascular damage: processing speed and executive functioning (Karakis et al., 2016).

Second, there may be a more direct impact of 25-hydroxyvitamin D on the brain. Empirical

studies have suggested a role for 25-hydroxyvitamin D in protecting neuronal integrity via

detoxification pathways and synthesis of proteins necessary for neuronal survival (Buell &

Dawson-Hughes, 2008). Third, research suggests that 25-hydroxyvitamin D may be protective

against depression, a disorder that often includes slowed cognitive and psychomotor speed

(Marazziti et al., 2010). Mechanisms by which 25-hydroxyvitamin D may be protective against

depression are not well-established, but one hypothesis is that 25-hydroxyvitamin D reduces the

173

increased neuronal levels of calcium ions that drive depression (Berridge, 2017). Fourth, lower

25-hydroxyvitamin D has been associated with the presence of obesity, a condition that has also

been associated with slowed processing speed (Mai, Chen, Camargo, & Langhammer, 2012).

Per the current literature, it is not clear whether obesity promotes low serum 25-hydroxyvitamin

D levels, low serum 25-hydroxyvitamin D promotes obesity, they are related through other

common variables (e.g., diet), or some combination thereof (Mai, et al., 2012). In any case,

results of the present study indicate that history of persistent obesity in young adulthood and late

middle-age does not account for the relationship between low 25-hydroxyvitamin D level and

slow processing speed in young-old adults once vascular health history is controlled.

The present study’s finding that current low vitamin D level significantly and negatively

impacts DSST score in the absence of late middle-age obesity history (but not in the presence of

late middle-age obesity history) in young-old adults is interesting, though its meaning is unclear.

Perhaps reasons for low vitamin D tend to differ between young-old adults with and without

history of late middle-age obesity. It could be hypothesized that reasons for low vitamin D in

young-old adults without history of late middle-age obesity tend to contribute to slowed

processing speed whereas reasons for low vitamin D in young-old adults with history of late

middle-age obesity tend not to contribute as much to slowed processing speed.

Iron deficiency. Research on the relationship between iron deficiency and processing

speed has primarily focused on infants and children (e.g., Hamid, Amal, Rohani, & Norimah,

2010). Little has been reported on this relationship among adults, and what has been reported is

mostly limited to women of childbearing age (e.g., Murray-Kolb & Beard, 2007; Scott &

Murray-Kolb, 2016). Findings from that body of research suggest that severity of anemia may

adversely impact processing speed to a greater extent than other cognitive domains, whereas

174

severity of iron deficiency may be associated with a broader pattern of cognitive dysfunction

(Murray-Kolb & Beard, 2007). Regarding older adults, one study reported an association

between iron deficiency and lower scores on a cognitive impairment screening measure (Yavuz

et al., 2012). The present study adds to the literature by reporting that young-old adults with iron

deficiency do not have significantly slower processing speed than young-old adults without iron

deficiency, though it should be emphasized that this finding does not imply that these two groups

have equivalent processing speed abilities. More research is needed to better understand the

relationship between iron deficiency and processing speed in young-old adults.

Vitamin E. There are multiple forms of vitamin E, including tocopherols, tocotrienols,

and stereoisomers (Johnson, 2016c). Of these, alpha-tocopheral is the most biologically active,

whereas gamma-tocopherol is the most abundant in the typical American diet (Brigelius-Flohe &

Traber, 1999; Camara et al., 2014; Johnson, 2016c). Vitamin E is believed to be important in

preventing diseases caused by oxidative stress through its antioxidant properties, though

additional mechanisms for health promotion are also likely (Brigelius-Flohe & Traber, 1999;

Niki, 2015).

Results from the present study indicate that higher alpha-tocopherol levels are associated

with better processing speed task performance in young-old adults. This finding is consistent

with a study of community-dwelling older adults that reported that those with alpha-tocopherol

levels in the highest tertile were significantly less likely to be suffering from cognitive

impairment or dementia than those in the lowest tertile (Cherubini et al., 2005). These results,

though they do not imply a causal role of vitamin E on health outcomes, are consistent with the

idea that vitamin E may aid in the prevention and/or treatment of diseases driven by oxidative

stress. Clinical trials testing this theory have been inconsistent and somewhat disappointing thus

175

far, though there is growing evidence that d-l-alpha-tocopherol supplementation can slow

functional decline in AD (Dysken et al., 2014; Niki, 2015).

Only 2 of the 492 participants with both an alpha-tocopherol value and DSST score in the

present study met vitamin E deficiency criteria; therefore, the relationship between vitamin E

deficiency and processing speed could not be assessed as planned. Results suggest that vitamin

E deficiency is rare in young-old adults in the United States and that a study of greater sample

size would be needed to properly assess this relationship. Of note, the low percentage of

participants meeting criteria for vitamin E deficiency in the present study is similar to results

from a study of European adults aged 55–87 that used a similar vitamin E deficiency criterion

(Polito et al., 2005). These findings are consistent with the assertion that vitamin E deficiency is

uncommon in developed countries (Johnson, 2016c).

Surprisingly, exploratory analyses revealed higher levels of gamma-tocopherol to be

associated with worse cognitive functioning in young-old adults, though this relationship nearly

vanished once the effect of alpha-tocopherol level was controlled. There is a paucity of

published information on the relationship between gamma-tocopherol and cognition in the

literature to date. When literature in areas outside of cognition is consulted for context, a few

important differences between alpha-tocopherol and gamma-tocopherol emerge. First, recent

research has shown that tocopherols regulate signaling pathways in the inflammatory process;

namely, during allergic inflammation, alpha-tocopherol is anti-inflammatory, whereas gamma-

tocopherol is pro-inflammatory (Abdala-Valencia, Berdnikovs, & Cook-Mills, 2013; Berdnikovs

et al., 2009). Second, there is some difference in associations between health outcomes and

serum levels of these two tocopherols. A recent meta-analysis, for example, concluded that

serum alpha-tocopherol level, but not serum gamma-tocopherol level, was inversely associated

176

with the risk of prostate cancer (Cui, Liu, & Xu, 2014). Results of the present study, as well as

the studies described above, suggest that the role of gamma-tocopherol should be more closely

studied. If better physical and cognitive outcomes are seen for alpha-tocopherol than gamma-

tocopherol, it may be advisable to consider efforts to increase alpha-tocopherol in the typical

American diet.

Vitamin B12. Vitamin B12 status, which was defined based on serum vitamin B12 level,

was not found to have a significant relationship with DSST score in the present study. Results

are in contrast to a population-based study of older adults in the Netherlands that found that those

with low serum vitamin B12 levels performed significantly worse on a letter substitution task

designed to measure processing speed (Jelicic et al., 2001). Reasons for this discrepancy are

unclear, and there does not appear to be any other published studies that have assessed this

particular relationship.

One limitation in research on the relationship between vitamin B12 and cognition has

been the considerable difficulty in accurately assessing the former. Vitamin B12 is often assessed

using serum B12 levels, as was done in the present study. Some, however, have expressed

concern that this method may not accurately reflect intracellular vitamin B12 levels (e.g., Hin et

al., 2006). Instead, other potential indicators of low vitamin B12 status (i.e., high methylmalonic

acid and high total homocysteine) have been suggested for use either alone or in combination

with serum B12 level (Clarke et al., 2003; Hin et al., 2006; Smith & Refsum, 2009). Of those two

potential indicators, high methylmalonic acid has been described as a more specific marker of

vitamin B12 deficiency (Tangney, Tang, Evans, & Morris, 2009).

Interestingly, exploratory analyses in the present study indicated that individuals in the

highest quartile for methylmalonic acid had significantly slower processing speed than

177

individuals in each of the other quartiles. These results are consistent with a community-based

study of older adults in England that found that participants in the top two quartiles of

methylmalonic acid level had a 3.7-fold risk for cognitive impairment, as operationally defined

as a Mini-Mental State Examination (MMSE) score of < 22/30, compared to those in the bottom

two quartiles (Hin et al., 2006). It is difficult to conclude that low vitamin B12 is associated with

slower processing speed in young-old adults on the basis of the present study, as there is still

some uncertainty regarding whether methylmalonic acid accurately reflects intracellular vitamin

B12 (Miller, 2006). In any case, high methylmalonic acid is associated with slowed processing

speed in young-old adults.

Hypothesis 7 Discussion

Individual obesity indexes. The present study sought to determine the relationship

between processing speed and various obesity indexes in young-old adults. There was no

statistically significant relationship between BMI and DSST score. This was expected because

individuals with midlife obesity have an increased risk for developing dementia, and individuals

who go on to develop dementia are more likely to have lost a significant amount of weight

between midlife and late-life (Xu et al., 2011; Stewart et al., 2005). Furthermore, those who are

obese during midlife may develop dementia earlier; a recent study found overweight and obese

BMI at age 50 to be associated with earlier onset of AD compared to normal BMI, with each unit

increase in BMI at age 50 predicting earlier onset by 6.7 months (Chuang et al., 2016).

The present study found statistically significant inverse relationships between DSST and

each of the following central obesity indexes: WC, waist-height ratio (WHtR), and waist-thigh

ratio (WTR). This was expected due to the greater utility of central obesity indexes in predicting

a number of health conditions that have associated cognitive decline relative to BMI, which is in

178

part believed to be due to differences between metabolic activity in visceral and other fat tissue.

Of these relationships, results of the present study suggest that those involving WHtR and WTR

have greater practical value than WC in regard to their relationship with processing speed in

young-old adults. It is possible that these findings reflect these indexes’ abilities to predict the

presence of relevant medical conditions (i.e., those associated with cognitive decline) within the

young-old adult population. For example, WHtR and WC were found to predict diabetes

mellitus, hyperglycemia, atherogenic dyslipidemia, and metabolic syndrome better than BMI in a

large cohort of Mediterranean adults aged 55–80 at high cardiovascular risk (Guasch-Ferre et al.,

2012).

The present study found a very weak, statistically significant positive correlation between

DSST score and % total body fat DEXA. These results were not as predicted and suggest that

greater adiposity is related to better processing speed in young-old adults. It should, however, be

noted that the magnitude of this finding suggests low practical value. Nevertheless, it may be the

case that young-old adults with greater adiposity performed better on the cognitive task since

they may have been less likely to be in the early stages of development of AD or other type of

dementia process (Xu et al., 2001). Results of the present study are generally consistent with a

recent community-based study of adults aged 74–94 that found that individuals in the lowest

tertile of DEXA body fat evidenced poorer executive functioning compared to those in the

middle tertile for men and the middle and highest tertiles for women (Smith et al., 2014).

The present study failed to find a statistically significant correlation between DSST score

and % trunk fat measured by DEXA. This finding is in contrast to the statistically significant

inverse relationships between DSST and the other central obesity indexes included in this study

(i.e., WC, WHtR, and WTR). It may be the case that central adiposity itself (i.e., % trunk fat

179

measured by DEXA) is not predictive of medical conditions associated with cognitive decline

among young-old adults as was previously hypothesized. Rather, it may be the case that an

increase in waist circumference (resulting in increased WC, WHtR, and WTR) occurs as a result

of complications from medical conditions associated with cognitive decline. For example,

gastroparesis, a disorder in which gastric emptying is delayed in the absence of physical

obstruction, is a common complication of uncontrolled diabetes and often includes the symptom

of bloating (Dickman et al., 2013; Koch & Calles-Escandón, 2015). Results of the present study

highlight the importance of factors other than trunk adiposity in the relationship between

anthropometrically derived central obesity indexes and processing speed in young-old adults.

Combinations of obesity indexes. The hypothesis that, among participants who met WC

central obesity criteria, those who also met BMI obesity criteria would perform worse on the

DSST was supported by the data. Results suggest that BMI obesity adds some value in

predicting poor processing speed task performance beyond WC alone among a nationally

representative sample of young-old U.S. adults. The reason this may be the case, however, is

unclear. This hypothesis was based on findings from a community study of elderly individuals

in Korea that found obese BMI (defined as ≥ 25 kg/m2 per modified WHO guidelines) predicted

a positive result on a brief screener for dementia using an empirically supported cut-point only

when abdominal obesity was also present (Jeong et al., 2004; Jeong et al., 2005).

The related hypothesis that, among participants who met WC central obesity criteria,

those who also met DEXA obesity criteria would perform worse on the DSST was not supported

by the data. Results suggest that BMI adds greater value than a more accurate estimate of

current adiposity in this population. It is unclear as to why this may be the case, but, as

180

hypothesized above, BMI may somehow better take into account other factors important in this

population. Future research is needed to better understand this relationship.

Finally, an attempt was made to find which combination of obesity indexes made the

most meaningful contribution to the prediction of DSST score among young-old adults. Given

the weak to very weak relationships between individual obesity indexes and DSST score

uncovered in prior analyses in the present study, it was not expected that findings of good

practical significance would be made in this planned analysis. After excluding a number of

indexes from a stepwise multiple regression based on issues of multicollinearity (i.e., BMI, WC,

DEXA % trunk fat), WHtR, DEXA % total body fat, and WTR remained. Results indicated that

WHtR and DEXA % total body fat accounted for 5.4% of variance in DSST score, and that

WTR did not meaningfully contribute to the model. Although results suggest WHtR and DEXA

% total body fat did make significant contributions to the overall prediction of DSST scores, the

low R2 value indicates that this model was not a good fit. This suggests that current obesity

indexes may not be particularly valuable in the prediction of processing speed in young-old

adults. Indeed, based on these findings as well as on the literature previously discussed, it would

appear that midlife obesity indexes are of greater predictive value on cognition in young-old

adults than are current obesity indexes.

BMI category. The present study found that young-old adults who were obese in young

adulthood have slower processing speed than young-old adults who had normal range BMI in

young adulthood. Additionally, young-old adults who were obese in late middle age were found

to have slower processing speed than young-old adults who were overweight in late middle age

as well as young-old adults who had normal range BMI in late middle age. These results are

generally consistent with the theory that history of persistent obesity, which is operationally

181

defined in this study as obesity throughout young adulthood and late middle age, is associated

with slower processing speed in young-old adults.

BMI category and gender interaction. The present study found that young-old males

had slower processing speed than young-old females. When potential interactions with BMI

were assessed, it was found that this relationship was mostly limited to individuals with normal

range BMI. More specifically, findings from the present study suggest that normal range BMI is

associated with slower processing speed in males aged 65–74 (i.e., young-old males) compared

to their female counterparts. The same relationship was found between BMI and current

processing speed when BMIs at various time points were used. These time points included

current BMI (i.e., young-old BMI), BMI 10 years ago (i.e., late middle age BMI), and BMI at

age 25 (i.e., young adult BMI). In contrast, no significant differences in processing speed

between males and females in the other BMI categories studied (i.e., obese, overweight) were

found at any of the time points, with the exception of a trend toward men who were obese in

young adulthood having slower processing speed than women who were obese at that time.

It is unclear as to why this relationship would be almost entirely limited to those with

normal range BMI. One possible explanation is that men in the normal range BMI category may

tend to have a greater number of, or more severe, medical problems that can adversely impact

processing speed over time than do women in the normal range BMI category. Another possible

explanation is that women may tend to have faster processing speed than men under “normal

conditions” (i.e., normal range BMI) but are, perhaps, more sensitive to the cognitive effects of

some of the comorbidities of overweight and obesity that can contribute to slowed processing

speed over time.

Hypothesis 8 Discussion

182

Pattern of obesity. Participants were divided into eight groups based on all possible

combinations of obesity status (yes/no) at age 25, 10 years ago, and the present, as depicted in

Table 5. It was hypothesized that participants who were obese at age 25 (i.e., young adulthood)

and 10 years ago (i.e., late middle age) would perform the worst on the DSST. In addition, it

was hypothesized that participants with no history of obesity would perform best on the DSST.

Further, it was hypothesized that those who were obese at only one timepoint would perform

better than those who were obese at age 25 and 10 years ago.

The analyses could not be run as originally planned due to three of the eight groups

containing very few participants. Each of these three groups shared the characteristic of obesity

at age 25. In retrospect, it seems likely that the small numbers of participants in these groups is

due to a lower prevalence of obesity in the past relative to the current day. Namely, in the

present study, data from participants aged 65–74 was collected during the years 1999–2002,

which means that self-reported young adult weight (i.e., age 25) referred to the participants’

weights during the years 1950–1962. The prevalence of obesity in U.S. adults aged 20–29

during the years 1960–1962, however, was estimated to be only 7.6% (Flegal, Carroll,

Kuczmarski, & Johnson, 1998). Obesity rates rose dramatically in the United States and many

other developed countries between approximately 1976 and 1994 (Flegal et al., 1998).

Consistent with this increase, the present study evidenced an adequate number of participants

with self-reported late middle age (i.e., age 55–64) weight, which referred to the participants’

weights during the years 1989–1992, in the obese range.

After removal of the three groups containing very few participants from the planned

analysis, five groups remained; they are shown in Table 6. Results revealed that Group C

(defined as those who were obese at late mid-life but not during young adulthood or the present)

183

scored significantly lower on the DSST than Groups B, D, and E. No other group differences

were statistically significant, though a slight trend toward significance for Group A (defined as

those who were obese at each of the three time points) scoring lower than Group E (defined as

those who were not obese at any of the three time points) was noted.

The finding that Group C (defined as those who were obese at late mid-life but not during

young adulthood or the present) scored significantly lower on the DSST than Groups B, D, and E

suggests that, of the young-old adults who were not obese at age 25, the pattern of the presence

of obesity in late middle age and the absence of obesity at the present is associated with

significantly slower processing speed than any other possible pattern of obesity status at those

two time points (i.e., presence of obesity at late middle age and at the present; absence of obesity

at late middle age and the present; absence of obesity at late middle age and presence of obesity

at the present).

More specifically, when comparing Groups B and C, both groups gained significant

weight by late middle age and were possibly facing similar obesity-related health risks at that

time. It may be the case that Group B (those who remained obese at the present) outperformed

Group C because Group C represents those individuals who lost between late middle age and the

present secondary to physical illness (obesity-related or otherwise) or preclinical dementia.

Further, it is possible that Group E outperformed Group C because those in Group E did not gain

as much weight by late middle age or reach the obesity category, at which point risk for many

diseases is known to increase. Finally, it is possible that Group D outperformed Group C

because becoming obese in young-old adulthood is less of a health risk than becoming obese in

middle age. Generally, results suggest that slowed processing speed in young-old adults may be

more related to the timing of their obesity status rather than number of time points for which they

184

can be categorized as obese, though some caution is taken with regard to this conclusion due to

inability to compare each of the eight groups as planned.

In any case, exploratory analyses were then conducted to better understand the

relationship between persistent obesity (i.e., obesity during young adulthood and late middle-

age) and processing speed in young-old adults. Young-old adults with a history of persistent

obesity were found to have slower processing speed than young-old adults with a history of

persistent non-obesity (i.e., obesity neither during young adulthood nor during late middle-age).

These findings are consistent with the hypothesis that long-term obesity in adulthood is

associated with slower processing speed in older adults.

Extent of weight gain between early adulthood and late middle-age. The hypothesis

that the extent of weight gain between early adulthood (i.e., age 25) and late middle-age (i.e., age

55–64) would be associated with slower processing speed was supported by the data. Results are

consistent with Sabia and colleagues (2009), who reported the extent of weight gain between

early adulthood and late middle-age among a large sample of London-based office staff to be

associated with executive dysfunction. Further, findings from an exploratory analysis in the

present study suggest that there is no significant relationship between processing speed in young-

old adults and the extent of weight loss between early adulthood and late middle-age. Taken

together, results suggest that weight gain, but not weight loss, between early adulthood and late

middle-age is associated with slower processing speed in young-old adulthood. It may be the

case that weight gain during this time is detrimental to processing speed via onset of weight-

related diseases or disorders that also impact processing speed over time. Alternatively, it may

be the case that onset of new diseases or disorders during this time led to both weight gain and

slower processing speed, or that slow processing speed in early adulthood led to weight gain over

185

time. Overall, however, results are generally consistent with the contraindication of gaining a

great amount of weight between early adulthood and late middle-age and suggest that amount of

weight lost between these two timepoints, assuming that one is not underweight at either time

point, is not detrimental to processing speed in young-old adulthood.

Extent of weight change between late middle-age and young-old adulthood. The

hypothesis that the extent of weight change over the previous 10-year period (i.e., from late

middle-age to young-old) would be associated with worse performance on the DSST was

supported by the data. Likewise, exploratory analyses uncovered inverse relationships between

processing speed and the extent of weight loss over the previous 10-year period as well as

between processing speed and the extent of weight gain over the same time period. These

findings are consistent with a previous study that have found that the extent of weight change

during a 10-year period between late middle-age (mean age 59) and early young adulthood

(mean age 69) was associated with poorer overall cognitive performance (Brubacher et al..,

2004). The present study adds to the literature by reporting this relationship to also be true of

processing speed among a nationally-representative sample of young-old adults. The findings of

the present study and that of Brubacher and colleagues (2004) may be related to the pattern of

significant weight loss between midlife and late-life prior to onset of dementia (Stewart et al.,

2005). However, because the present study did not include a baseline processing speed measure,

it is not possible to rule out the possibility that slower baseline processing speed led to, or

contributed to, significant weight change during the 10-year period in question.

Past-year weight change. The present study found that young-old adults with

unintentional significant past-year weight loss performed worse on the DSST than participants

who experienced significant intentional past-year weight loss, significant past-year weight gain,

186

or past-year weight stability. This finding is consistent with the tendency for unintentional

weight loss to be associated with an underlying illness, which may or may not also be impacting

cognition. Results may have implications for the utility of conducting brief cognitive screenings

of young-old adults who present to their physicians with past-year unintentional weight loss.

Hypothesis 9 Discussion

Peripheral inflammation is a metabolic consequence of excess adipose tissue, and obese

individuals tend to evidence higher levels of peripheral inflammation markers independent of

other known sources of inflammation (Visser et al., 1999). Obesity creates a state of chronic

low-grade peripheral inflammation, and research suggests that this state may lead to cognitive

impairment through mechanisms such as invocation of central neuroinflammatory response,

development of inflammation resistance, accumulation of proinflammatory cytokines in the

bloodstream, and interaction of these cytokines with free fatty acids to impair insulin signaling

(Dali-Youcef et al., 2013; Ye & Keller, 2010).

The present study’s hypothesis that C-reactive protein (CRP), a marker of peripheral

inflammation, would have a statistically significant inverse association with processing speed

among young-old adults was not supported by the data; their association was in the expected

direction but did not reach statistical significance. The few other extant studies that have

reported on the relationship between CRP and processing speed in older adult populations have

yielded mixed findings (Tegeler et al., 2016; Trollor et al., 2012).

There are a number of possible explanations for the lack of a clear relationship between

CRP and processing speed in older adult populations. First, CRP is only one of several markers

of peripheral inflammation, which also include interleukins-1β, -6, -8, -10, -12, plasminogen

activator inhibitor, serum amyloid A, tumour necrosis factor-α, and vascular adhesion molecule-

187

1 (Trollor et al., 2012). It may be the case that some of these other markers of peripheral

inflammation are more closely associated with processing speed in young-old adults. Of these,

interleukins -6 and -12 may be of particular importance (Palta et al., 2015; Rafnsson et al., 2007;

Trollor et al., 2012). Second, it is possible that CRP does not have a significant relationship to

processing speed but does have a significant relationship to one or more other cognitive domains

among young-old adults. There is, in fact, some evidence to suggest an association between

CRP and executive functioning in this age group (Schram et al., 2007; Sweat et al., 2008;

Tegeler et al., 2016). Third, it is possible that a history of elevated CRP level in midlife is more

relevant to processing speed performance in older adult populations than current CRP level. No

published study has reported on this to-date; however, high CRP in midlife has been found to be

associated with decline in inductive reasoning ability ten years later (Gimeno, Marmot, & Singh-

Manoux, 2008).

Hypothesis 10 Discussion

The hypothesis that individuals who endorsed history of one or more of the following

conditions would perform worse on the DSST was supported by the data: cardiac (i.e.,

congestive heart failure [CHF], coronary heart disease [CHD], or heart attack), stroke, diabetes,

weak/failing kidneys, and chronic respiratory disease (i.e., chronic bronchitis or emphysema).

Potential mechanisms for cognitive decline with regard to these diseases and conditions have

already been discussed.

Exploratory analyses were used to investigate group differences in processing speed with

respect to the presence or absence of each of the aforementioned conditions, as well as

hypertension and any liver condition. Among the conditions tested, all but chronic respiratory

disease, any liver condition, and CHD in the absence of other cardiac conditions were

188

significantly linked to slower processing speed. Furthermore, length of chronic respiratory

disease was not significantly correlated with processing speed task performance. Results suggest

that history of any liver condition, history of chronic respiratory disease, length of chronic

respiratory disease, and history of CHD in the absence of other cardiac conditions may not be

relevant factors with regard to current processing speed in young-old adults.

Hypothesis 11 Discussion

The hypothesis that participants with one or more types of dyslipidemia, as determined

by blood tests, would demonstrate worse processing speed was not supported by the data. This

finding suggests that overall lipid health status, as operationally defined as the presence or

absence of one or more types of dyslipidemia, may not be important with regard to current

processing speed in young-old adults. Results of exploratory analyses, however, indicate that

three individual types of dyslipidemia are salient to this relationship: low high-density

lipoprotein [HDL] cholesterol, high non-HDL cholesterol, and high triglycerides, the latter two

of which are sex-specific. Each of these three is discussed below.

Low HDL cholesterol. Among both men and women in the present study, young-old

adults with low HDL cholesterol had worse DSST performance than those without low HDL

cholesterol. Because young-old adults are in the age range that is particularly susceptible to a

preclinical stage of dementia, the present study’s finding is consistent with results of large

prospective studies that show HDL cholesterol to have an inverse relationship with risk of all-

cause dementia (Koch & Jensen, 2016; Reitz et al., 2010). The present study’s finding is also

consistent with research that shows low HDL cholesterol to be independently associated with

memory dysfunction and decline in middle-aged adults and linked to structural brain changes in

regions associated with AD (Singh-Manoux et al., 2008; Ward et al., 2010). How low HDL

189

cholesterol, in particular, may contribute to the development of cognitive impairment and

dementia is not well established, but indirect pathways are believed to include atherosclerosis,

which may lead to decreased cerebral blood flow, as well as increased risk for other vascular

events and diseases that in turn can cause cognitive impairment and dementia (Ancelin et al.,

2013; Vinkers et al., 2005). There also appears to be a relationship between low HDL and build-

up of amyloid plaque deposits in the brain, though the nature of this relationship is unclear (Reed

et al., 2014). In addition to low levels of HDL cholesterol, emerging evidence suggests that

other aspects of HDL cholesterol may be problematic for some individuals with cognitive

impairment or dementia. Namely, one recent study reported decreased efficiency of HDL

cholesterol among those with AD (Camponova et al., 2017).

High non-HDL cholesterol. Among men only, those with high non-HDL cholesterol

evidenced significantly slower processing speed than those without high non-HDL cholesterol.

The relationship between processing speed and low-density lipoprotein [LDL] cholesterol, on the

other hand, did not reach statistical significance for either gender.

Non-HDL cholesterol is calculated by subtracting HDL cholesterol from total cholesterol,

which yields an estimate of cholesterol concentration among all atherogenic lipoproteins: LDL

cholesterol, very-low density lipoprotein (VLDL) cholesterol, intermediate-density lipoprotein

(IDL) cholesterol, and lipoprotein (a) [Lp(a)] cholesterol (Virani, 2011). Therefore, non-HDL

cholesterol is a more comprehensive measure of atherogenic lipoproteins than LDL cholesterol.

Although LDL cholesterol is the current primary target for treatment of dyslipidemia,

many doctors have argued for the use of non-HDL cholesterol in place of, or in addition to, LDL

cholesterol in clinical practice on the basis that it is more predictive of cardiovascular risk than

LDL cholesterol or, at the least, adds additional value (Rana, Boekholdt, Kastelein, & Shah,

190

2012; van Deventer et al., 2011; Virani, 2011). Nevertheless, LDL cholesterol continues to

receive greater attention than non-HDL cholesterol in the empirical literature, and, not

surprisingly, there are numerous published studies that include information on the relationship

between LDL cholesterol and cognition but very few, if any, that address the relationship

between non-HDL cholesterol and cognition. Results from the present study provide support to

the value of studying non-HDL cholesterol.

Because LDL cholesterol comprises a large portion of non-HDL cholesterol, empirical

findings regarding the relationship between LDL cholesterol and cognition are briefly described

here. Concerning the present study, mean difference in DSST score between groups with and

without high LDL cholesterol was not statistically significant but did show a trend toward

significance, suggesting lower processing speed in those with high LDL cholesterol. This

relationship was stronger among men than women but, again, did not reach statistical

significance when analyses were limited to men. In regard to the literature to date, the

relationship between processing speed and LDL cholesterol level in older adults is not well

established, but one recent study of older adults found that greater variability in LDL cholesterol

readings across study visits was associated with worse processing speed and other aspects of

cognition, as well as lower cerebral blood flow and greater white matter hyperintensity (WMH)

load (Smit et al., 2016). The findings on the relationship between LDL cholesterol and other

cognitive domains have been mixed. For example, one recent study of older adults with

cardiovascular risk factors found higher LDL cholesterol to be related to worse performance on a

working memory task, whereas another recent study of healthy middle-aged and older aged

adults found LDL cholesterol to be associated with better episodic memory (Leritz, McGlinchey,

Salat, Milberg, 2016; Meusel et al., 2017).

191

High triglycerides. Among women only, those with high triglycerides evidenced

significantly faster processing speed than those without high triglycerides. This relationship was

surprising, as high triglyceride level is considered to be a biomarker of cardiovascular disease

(CVD) risk, though it should be noted that triglyceride itself is not directly atherogenic (Miller et

al., 2011). There have been few published studies on the relationship between current

triglyceride levels and processing speed in older adults to date. One such study failed to find a

statistically significant relationship between processing speed and triglycerides, but this study did

not report on sex-specific relationships (van den Kommer, Dik, Comijs, Jonker, & Deeg, 2012).

More research is needed in this area to determine whether this is a stable finding and, if so, to

explore the nature of this relationship. Reasons for such a relationship are unclear, though the

authors of a prospective study that found a relationship between high triglyceride levels and

decreased risk of AD in women hypothesized that this finding may reflect a survival bias

(Ancelin et al., 2013). Findings on the relationship between triglycerides and other cognitive

domains among older adults have been mixed (Parthasarathy et al., 2014; van den Kommer et al.,

2012).

Hypothesis 12 Discussion

The hypotheses that risk of first stroke, as indicated by score on a modified version of the

Framingham Stroke Risk Profile (FSRP), and risk of CVD onset, as indicated by score on a

modified version of the Framingham Risk Score (FRS) for general CVD, would each be

inversely associated with DSST score was supported by the data. Scores reflecting risk of first

stroke and risk of CVD onset each showed a statistically significant weak inverse relationship

with DSST score. Findings indicate that, among this nationally representative cross-sectional

sample of young-old adults, processing speed tends to decrease as risk of first stroke or risk of

192

CVD onset increases. Although only weak relationships were found, results suggest the

potential impact of stroke and CVD risk factors on processing speed in young-old adults without

history of stroke or CVD. They also suggest the possibility that some cognitive decline or

impairment may be present in these individuals prior to first stroke or CVD onset.

Of course, because these equations were designed to predict risk for first stroke or onset

of CVD rather than processing speed or other aspects of cognition, some of their components

may not be optimal when the goal is to predict current processing speed in young-old adults.

FRS for general CVD, for example, takes total cholesterol into account, whereas the present

study did not support the utility of total cholesterol in predicting processing speed in the young-

old adult population.

Hypothesis 13 Discussion

The hypothesis that there would be a significant positive correlation between usual gait

speed and processing speed after controlling for BMI was supported by the data. Therefore,

although walking speed has been found to be slower among obese older adults compared to their

normal weight counterparts, BMI does not explain the association between slower walking speed

and slower processing speed in young-old adults (Xu, Houston, Gropper, & Zizza, 2009).

Results from the present study are consistent with those from another recent study that found

lower processing speed to be a major contributor to slower gait speed after controlling for other

cognitive domains, age, BMI, and other potential confounds in late-middle aged and older adults

(Killane et al., 2014).

The finding that gait speed decelerates faster in the years preceding mild cognitive

impairment (MCI) diagnosis in affected individuals compared to the normal cognitive aging

population suggests the role of shared pathological processes of organic etiology in the

193

relationship between slowed gait speed and cognitive decline (Albers et al., 2015; Buracchio,

Dodge, Howieson, Wasserman, & Kaye, 2010; Kikkert, Vuillerme, van Campen, Hortobagyi, &

Lamoth, 2016). For example, cerebral white matter hyperintensities (WMHs), which are areas of

abnormal change in white matter found on head computerized tomography (CT) and brain

magnetic resonance imaging (MRI) scans that reflect changes in interstitial fluid mobility and

water content, demyelination, and/or axonal damage, are increasingly common with age and are

associated with vascular risk factors, stroke, and dementia (Wardlaw, Valdés Hernández, &

Muñoz-Maniega, 2015). WMHs are believed to be causally related to slowed gait speed and

reductions in the cognitive domains of processing speed and executive functioning in affected

individuals depending on their size, location, and total volume (Wakefield et al., 2010). It has

been hypothesized that WMHs can impact gait speed directly through disrupting mobility-related

pathways or indirectly through disrupting pathways important for executive functions

(Bolandzadeh et al., 2014). Likewise, it can be hypothesized that slower processing of

visuospatial information may contribute to slower gait speed. The finding from a recent study

that performance on the DSST significantly mediates the relationship between WMHs and gait

speed supports these hypotheses, as the DSST is a measure of processing speed that has also

been shown to reflect visual scanning ability and executive functioning (Bolandzadeh et al.,

2014; Davis & Pierson, 2012; Joy et al., 2003).

Final Model Discussion

It had been reasoned that bodily states caused by obesity that tend to be present even

among otherwise healthy obese individuals (e.g., increased adipokines, low grade inflammation)

might contribute to worsened processing speed above and beyond that of diagnosed diseases,

disorders, and medical events for which obesity has been shown to often have a contributory role

194

(e.g., T2DM, stroke). Additionally, it was thought that these obesity-related bodily states (e.g.,

increased adipokines, low grade inflammation) might have a differential impact on cognition

according to magnitude and duration of exposure. Persistent obesity was conceptualized as a

way of identifying long duration of exposure to these bodily states.

Results of the present study’s final model suggest that vascular health history mediates

the relationship between persistent obesity and processing speed among adults aged 65–74;

however, no evidence for the direct contribution of persistent obesity to current processing speed

was found. Likewise, results of Hypothesis 7 analyses did not provide evidence for persistent

obesity being related to processing speed to a greater extent than obesity in late middle-age.

Taken together, results of the present study do not provide support for the cumulative impact of

obesity during young adulthood and late middle-age on processing speed in young-old adults.

One possible explanation for this finding may be that the brain is more vulnerable to the effects

of obesity-related bodily states (e.g., increased adipokines, low grade inflammation) during late

middle-age than during young adulthood. Additionally, the recent finding that the low grade

systemic inflammation associated with obesity tends to impair the blood brain barrier only in

those with already diseased brains would also support the lack of importance of obesity in young

adulthood relative to obesity in late middle-age (Elwood, Lim, Naveed, & Galea, 2017).

Consistent with this idea, it may be the case that those who are obese when preclinical dementia

begins experience more rapid decline.

Limitations of the Present Study

The present study had several limitations. As a non-experimental, cross-sectional study,

causality of processing speed cannot be inferred. Longitudinal data would have provided many

advantages, including the opportunity for collection and analysis of repeated cognitive tests and

195

biochemistry data. The use of retrospective pretests (e.g., self-report of weight ten years ago,

self-report of weight at age 25) did allow for investigation of associations between weight

trajectory and other variables, but reliance on these retrospective pretests may have been

problematic with regard to their accuracy. For example, one recent study of older adults found

that 20 year retrospectively recalled weights tend to underestimate both 20-year prior assessed

weights and 20 year prior self-reported weights by an average of 4.2 ± 13.0 and 1.2 ± 11.5

pounds, respectively (Dahl & Reynolds, 2013). Another limitation of the present study was the

use of only one measure of processing speed. Cognitive tests are rarely able to measure one

domain in isolation. Given that the DSST primarily reflects graphomotor speed, perceptual

speed, visual scanning ability, and executive functioning, results need to be interpreted within

this specific context. Moreover, the present study was limited by the low number of participants

in some of the weight trajectory groups, which prevented some planned analyses from being run.

Finally, the present study would have benefitted from the inclusion of weight data at time points

between young adulthood and late middle-age. More specifically, the present study assumed that

the combined presence of obesity in young adulthood and late middle-age was generally

consistent with a state of “persistent obesity” throughout that time; however, this assumption

may not be fair, as weight can fluctuate considerably over time.

Conclusions and Recommendations for Future Research

The present study did not provide evidence for the cumulative impact of obesity during

young adulthood and late middle-age on processing speed in young-old adults. In contrast, it

highlighted the link between slower processing speed and the weight trajectory pattern of obesity

in late middle-age coupled with the absence of obesity in both young adulthood and young-old

adulthood. That being said, results should be interpreted with caution due to study limitations,

196

especially because the low number of participants in some of the weight trajectory groups

prevented the processing speed outcomes of those groups from being investigated. Future

studies that use similarly large sample sizes but analyze more recent data would be less likely to

encounter the same problem given the history of obesity prevalence trends in the United States.

An indirect effect of persistent obesity during young adulthood and late middle-age on

processing speed in young-old adults through vascular health history was identified by the

present study, but there was no evidence for a direct effect. It may be the case that persistent

obesity does not have a direct effect on processing speed in young adulthood; however,

limitations in the present study prevent this conclusion from being drawn. Future research

should include weight data at time points between young adulthood and late middle-age in order

to more accurately classify individuals as persistently obese throughout that time period. In

addition, future studies should strive to use measured weights rather than self-reported weights,

as this would improve the accuracy of the collected weight data and, by extension, would likely

improve the accuracy of weight category classification. Moreover, future studies should include

a measure of processing speed at baseline to investigate the relationship between weight

trajectory and change in processing speed as well as to investigate the possibility that slower

baseline processing speed predicts subsequent weight patterns.

197

References

Abdala-Valencia, H., Berdnikovs, S., & Cook-Mills, J. M. (2013). Vitamin E isoforms as

modulators of lung inflammation. Nutrients, 5(11), 4347–4363. doi:10.3390/nu5114347

Abraham, P., & Acharya, S. (2010). Current and emerging treatment options for Graves'

hyperthyroidism. Therapeutics and Clinical Risk Management, 6, 29–40.

Ackermann, D., Jones, J., Barona, J., Calle, M. C., Kim, J. E., LaPia, B.,…Fernandez, M. L.

(2011). Waist circumference is positively correlated with markers of inflammation and

negatively with adiponectin in women with metabolic syndrome. Nutrition Research, 31(3),

197–204. doi:10.1016/j.nutres.2011.02.004

Ahlborg, H. G., Johnell, O., Nilsson, B. E., Jeppsson, S., Rannevik, G., & Karlsson, M. K.

(2001). Bone loss in relation to menopause: a prospective study during 16 years. Bone,

28(3), 327–331.

Ahn, S., Smith, M.L., Dickerson, J. B., & Ory, M. G. (2012). Health and health care utilization

among obese and diabetic baby boomers and older adults. American Journal of Health

Promotion, 27(2), 123–132. doi:10.4278/ajhp.111013-QUAN-366

Aine, C. J., Sanfratello, L., Adair, J. C., Knoefel, J. E., Caprihan, A, & Stephen, J. M. (2011).

Development and decline of memory functions in normal, pathological and healthy

successful aging. Brain Topography, 24(3-4), 323–339. doi:10.1007/s10548-011-0178-x

Akishita, M. (2014). Sex hormones and cognitive function. Nihon Rinsho, 72(4), 657–661.

Al Hazzouri, A. Z., Stone, K. L., Haan, M. N., & Yaffe, K. (2013). Leptin, mild cognitive

impairment, and dementia among elderly women. The Journals of Gerontology. Series A,

Biological Sciences and Medical Sciences, 68(2), 175–180. doi:10.1093/gerona/gls155

Alagiakrishnan, K., Sankaralingam, S., Ghosh, M., Mereu, L., & Senior, P. (2013). Antidiabetic

drugs and their potential role in treating mild cognitive impairment and Alzheimer’s

disease. Discovery Medicine, 16(90), 277–286.

Albers, M. W., Gilmore, G. C., Kaye, J., Murphy, C., Wingfield, A., Bennett, D. A.,…Zhang, L.

I. (2015). At the interface of sensory and motor dysfunctions and Alzheimer’s disease.

Alzheimer’s & Dementia, 11(1), 70–98. doi:10.1016/j.jalz.2014.04.514

Alexaki, V. I., Notas, G., Pelekanou, V., Kampa, M., Valkanou, M., Theodoropoulos, P.,

…Castanas, E. (2009). Adipocytes as immune cells: differential expression of TWEAK,

BAFF, and APRIL and their receptors (Fn14, BAFF-R, TACI, and BCMA) at different

stages of normal and pathological adipose tissue development. The Journal of Immunology,

183(9), 5948–5956. doi:10.4049/jimmunol.0901186

198

Alosco, M. L., Brickman, A. M., Spitznagel, M. B., van Dulmen, M., Raz, N., Cohen, R., Sweet,

L. H.,...Gunstad, J. (2012). The independent association of hypertension with cognitive

function among older adults with heart failure. Journal of the Neurological Sciences, 323(1-

2), 216–220. doi:10.1016/j.jns.2012.09.019

Alosco, M. L., Spitznagel, M. B., Raz, N., Cohen, R., Sweet, L. H., Colbert, L. H., Josephson,

R.,…Gunstad, J. (2012). Obesity interacts with cerebral hypoperfusion to exacerbate

cognitive impairment in older adults with heart failure. Cerebrovascular Diseases Extra,

2(1), 88–98. doi:10.1159/000343222

Alvarez-Blasco, F., Luque-Ramírez, M., & Escobar-Morreale, H. F. (2010). Obesity impairs

general health-related quality of life (HR-QoL) in premenopausal women to a greater extent

than polycystic ovary syndrome (PCOS). Clinical Endocrinology, 73(5), 595–601.

doi:10.1111/j.1365-2265.2010.03842.x

American Diabetes Association (2014, August 18). Diagnosing diabetes and learning about

prediabetes. Retrieved from http://www.diabetes.org/diabetes-basics/diagnosis/

American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental

Disorders (5th ed.). Washington, DC.

Amini, M., Afshin-Nia, F., Bashardoost, N., Aminorroaya, A, Shahparian, M., & Kazemi, M.

(1997). Prevalence and risk factors of diabetes mellitus in the Isfahan city population (aged

40 or over) in 1993. Diabetes Research and Clinical Practice, 38(3), 185–90.

Anstey, KJ, Sargent-Cox, K, Cherbuin, N, & Sachdev, PS. (2015). Self-reported history of

chemotherapy and cognitive decline in adults aged 60 and older: The PATH through life

project. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences,

70(6), 729–735. doi:10.1093/gerona/glt195

Appelhans, B. M., Waring, M. E., Schneider, K. L., Pagoto, S. L., DeBiasse, M. A., Whited, M.

C., & Lynch, E. B. (2012). Delay discounting and intake of ready-to-eat and away-from-

home foods in overweight and obese women. Appetite, 59(2), 576–584.

doi:10.1016/j.appet.2012.07.009

Arif, A. A., Rohrer, J. E., & Delclos, G. L. (2005). A population-based study of asthma, quality

of life, and occupation among elderly Hispanic and non-Hispanic whites: A cross-sectional

investigation. BMC Public Health, 5, 97. doi:10.1186/1471-2458-5-97

Arnesen, E., Refsum, H., Bonaa, K., Ueland, P. M., Forde, O. H., & Nordrehaug, J. E. (1995).

Serum total homocysteine and coronary heart disease. International Journal of

Epidemiology, 24(4), 704–709.

Aronson, D., & Rayfield, E. J. (2002). How hyperglycemia promotes atherosclerosis: Molecular

mechanisms. Cardiovascular Diabetology, 1(1). doi:10.1186/1475-2840-1-1

199

Ashwell, M., Gunn, P., & Gibson, S. (2012). Waist-to-height ratio is a better screening tool than

waist circumference and BMI for adult cardiometabolic risk factors: systematic review and

meta-analysis. Obesity Reviews, 13(3), 275–286. doi:10.1111/j.1467-789X.2011.00952.x

Atkinson, H. H., Rosano, C., Simonsick, E. M., Williamson, J. D., Davis, C., Ambrosius, W. T.,

…Health ABC Study (2007). Cognitive function, gait speed decline, and comorbidities: the

health, aging and body composition study. The Journals of Gerontology, 62(8), 844–850.

Auyeung, T. W., Lee, J. S. W., Leung, J., Kwok, T., Leung, P. C., & Woo, J. (2010). Survival in

older men may benefit from being slightly overweight and centrally obese—a 5-year

follow-up study in 4,000 older adults using DXA. The Journals of Gerontology. Series A,

Biological Sciences and Medical Sciences, 65(1), 99–104. doi:10.1093/gerona/glp099

Avila, R., Chapman, C. H., Tao, Y., Schipper, M., Buchtel, H., Ten Haken, R. K.,…Kim, M. M.

(2015). Dose-Volume Effects of Partial-Brain Radiation Therapy on Late Neurocognitive

Processing Speed and Executive Function. International Journal of Radiation Oncology

Biology Physics, 93(3S), S175. doi: 10.1016/j.ijrobp.2015.07.420

Azagra-Calero, E., Espinar-Escalona, E., Barrera-Mora, J. M., Llamas-Carreras, J. M., & Solano-

Reina, E. (2012). Obstructive sleep apnea syndrome (OSAS). Review of the literature.

Medicina Oral Patología Oral y Cirugia Bucal, 17(6), e925–e929.

doi:10.4317/medoral.17706

Baena-Diez, J., Byram, A. O., Grau, M., Gomez-Fernandez, C., Vidal-Solsona, M., Ledesma-

Ulloa, G.,…Schroder, H. (2010). Obesity is an independent risk factor for heart failure:

Zona Franca Cohort study. Clinical Cardiology, 33(12), 760–764. doi:10.1002/clc.20837

Bamia, C., Halkjaer, J., Lagiou, P., Trichopoulos, D., Tjønneland, A, Berentzen, T.

L.,…Trichopoulou, A. (2010). Weight change in later life and risk of death amongst the

elderly: The European prospective investigation into cancer and nutrition-elderly network

on ageing and health study. Journal of Internal Medicine, 268(2), 133–44.

doi:10.1111/j.1365-2796.2010.02219.x

Bangen, K. J., Delano-Wood, L., Wierenga, C. E., McCauley, A., Jeste, D. V., Salmon, D. P., &

Bondi, M. W. (2010). Associations between stroke risk and cognition in normal aging and

Alzheimer’s disease with and without depression. International Journal of Geriatric

Psychiatry, 25(2), 175–82. doi:10.1002/gps.2317

Barba, R., Martinez-Espinosa, S., Rodriguez-Garcia, E., Pondal, M., Vivancos, J., & Del Ser, T.

(2000). Poststroke dementia: Clinical features and risk factors. Stroke, 31(7), 1494–1501.

doi:10.1161/01.STR.31.7.1494

Barnes, L. L., Wilson, R. S., Schneider, J. A., Bienias, J. L., Evans, D. A., & Bennett, D. A.

(2003). Gender, cognitive decline, and risk of AD in older persons. Neurology, 60(11),

1777–1781.

200

Barrientos, R. M., Hein, A. M., Frank, M. G., Watkins, L. R., & Maier, S. F. (2012).

Intracisternal interleukin-1 receptor antagonist prevents postoperative cognitive decline and

neuroinflammatory response in aged rats. The Journal of Neuroscience, 32(42), 14641–

14648. doi:10.1523/JNEUROSCI.2173-12.2012

Barry, D., & Petry, N. M. (2009). Associations between body mass index and substance use

disorders differ by gender: Results from the National Epidemiologic Survey on Alcohol and

Related Conditions. Addictive Behaviors, 34(1), 51–60. doi:10.1016/j.addbeh.2008.08.008

Basoglu, O. K., & Tasbakan, M. S. (2014). Comparison of clinical characteristics in patients with

obesity hypoventilation syndrome and obese obstructive sleep apnea syndrome: A case-

control study. The Clinical Respiratory Journal, 8(2), 167–174. doi:10.1111/crj.12054

Batsis, J. A., Sahakyan, K. R., Rodriguez-Escudero, J. P., Bartels, S. J., Somers, V. K., & Lopez-

Jimenez, F. (2013). Normal weight obesity and mortality in United States subjects ≥ 60

years of age (from the Third National Health and Nutrition Examination Survey). American

Journal of Cardiology, 112(10), 1592–1598. doi:10.1016/j.amjcard.2013.07.014

Beauchet, O., Allali, G., Launay, C., Herrmann, F. R., & Annweiler, C. (2013). Gait variability

at fast-pace walking speed: A biomarker of mild cognitive impairment? The Journal of

Nutrition, Health & Aging, 17(3), 235–239. doi:10.1007/s12603-012-0394-4

Bélanger, S., Belleville, S., & Gauthier, S. (2010). Inhibition impairments in Alzheimer’s

disease, mild cognitive impairment and healthy aging: Effect of congruency proportion in a

Stroop task. Neuropsychologia, 48(2), 581–590.

doi:10.1016/j.neuropsychologia.2009.10.021

Benito-León, J., Mitchell, A. J., Hernández-Gallego, J., & Bermejo-Pareja, F. (2013). Obesity

and impaired cognitive functioning in the elderly: a population-based cross-sectional study

(NEDICES). European Journal of Neurology, 20(6), 899–906. doi:10.1111/ene.12083

Berchicci, M., Lucci, G., Pesce, C., Spinelli, D., & Di Russo, F. (2012). Prefrontal hyperactivity

in older people during motor planning. Neuroimage, 62(3), 1750–1760.

doi:10.1016/j.neuroimage.2012.06.031

Bergamini, C. M., Gambetti, S., Dondi, A., & Cervellati, C. (2004). Oxygen, reactive oxygen

species and tissue damage. Current Pharmaceutical Design, 10(14), 1611–1626.

Berdnikovs, S., Abdala-Valencia, H., McCary, C., Somand, M., Cole, R., Garcia, A.,…Cook-

Mills, J. M. (2009). Isoforms of vitamin E have opposing immunoregulatory functions

during inflammation by regulating leukocyte recruitment. The Journal of Immunology.

182(7), 4395-4405. doi: 10.4049/jimmunol.0803659

Berridge, MJ (2017). Vitamin D and depression: Cellular and regulatory mechanisms.

Pharmacological Reviews, 69(2), 80-92. doi: 10.1124/pr.116.013227

201

Betteridge, D. J. (2000). What is oxidative stress? Metabolism, 49(2), 3–8.

Binkley, N, Ramamurthy, R, & Krueger, D (2010). Low vitamin D status: Definition,

prevalence, consequences and correction. Endocrinology and Metabolism Clinics of North

America, 39(2), 287–301. doi:10.1016/j.ecl.2010.02.008

Blanco, R., Eloisa, M., Ballesteros, A., Rosa, M., Sanchez Martin, M., Rabana Basalo, A., &

Olivas Dominguez, Antonio Garcia Lacalle, C. (2012). Prevalence of insulin resistance in a

young adult population. Relationship with weight status. Endocrinologia y Nutricion, 59(2),

98–104. doi:10.1016/j.endonu.2011.10.009

Bloomer, R. J., & Fisher-Wellman, K. H. (2009). Systemic oxidative stress is increased to a

greater degree in young, obese women following consumption of a high fat meal. Oxidative

Medicine and Cellular Longevity, 2(1), 19–25.

Boeka, A. G., & Lokken, K. L. (2008). Neuropsychological performance of a clinical sample of

extremely obese individuals. Archives of Clinical Neuropsychology, 23(4), 467–474.

doi:10.1016/j.acn.2008.03.003

Boitard, C., Cavaroc, A., Sauvant, J., Aubert, A., Castanon, N., Laye, S., & Ferreira, G. (2014).

Impairment of hippocampal-dependent memory induced by juvenile high-fat diet intake is

associated with enhanced hippocampal inflammation in rats. Brain, Behavior, and Immunity,

40, 9–17.

Bolandzadeh, N., Liu-Ambrose, T., Aizenstein, H., Harris, T., Launer, L., Yaffe, K.,…Rosano,

C. (2014). Pathways linking regional hyperintensities in the brain and slower gait.

Neuroimage, 99, 7-13. doi: 10.1016/j.neuroimage.2014.05.017

Bombelli, M., Facchetti, R., Fodri, D., Brambilla, G., Sega, R., Grassi, G., & Mancia, G. (2013).

Impact of body mass index and waist circumference on the cardiovascular risk and all-cause

death in a general population: Data from the PAMELA study. Nutrition, Metabolism, and

Cardiovascular Diseases, 23(7), 650–656. doi:10.1016/j.numecd.2012.01.004

Borghesani, P. R., Madhyastha, T. M., Aylward, E. H., & Reiter, M. A. (2014). The association

between higher order abilities, processing speed, and age are variably mediated by white

matter integrity during typical aging. Neuropsychologia, 51(8), 1–26.

doi:10.1016/j.neuropsychologia.2013.03.005

Boston Medical Center (2014, August 18). Framingham study. Retrieved from

http://www.bmc.org/stroke-cerebrovascular/research/framinghamstudy.htm

Botvinick, M.M., Braver, T.S., Barch, D.M., Carter, C.S., & Cohen, J.D. (2001). Conflict

monitoring and cognitive control. Psychological Review, 108(3), 624-652.

Boyer, P. (2000). Do anxiety and depression have a common pathophysiological mechanism?

Acta Psychiatrica Scandinavica. Supplementum, 24(6), 24–29.

202

Bramlage, P., Pittrow, D., Wittchen, H.-U., Kirch, W., Boehler, S., Lehnert, H., Hoefler, M., et

al. (2004). Hypertension in overweight and obese primary care patients is highly prevalent

and poorly controlled. American Journal of Hypertension, 17(10), 904–910.

doi:10.1016/j.amjhyper.2004.05.017

Braver, T. S. (2012). The variable nature of cognitive control: a dual mechanisms framework.

Trends in Cognitive Sciences, 16(2), 106–113. doi:10.1016/j.tics.2011.12.010

Braver, T. S., Barch, D. M., Keys, B. A., Carter, C. S., Cohen, J. D., Kaye, J. A., Janowsky, J. S.,

et al. (2001). Context processing in older adults: Evidence for a theory relating cognitive

control to neurobiology in healthy aging. Journal of Experimental Psychology: General,

130(4), 746–763. doi:10.1037//0096-3445.130.4.746

Braver, T. S., Satpute, A. B., Rush, B. K., Racine, C. A., & Barch, D. M. (2005). Context

processing and context maintenance in healthy aging and early stage dementia of the

Alzheimer’s type. Psychology and Aging, 20(1), 33–46. doi:10.1037/0882-7974.20.1.33

Brigelius-Flohe, R & Traber, MG (1999). Vitamin E: function and metabolism. The FASEB

Journal, 13, 1145–1155.

Brubacher, D., Monsch, A. U., & Stähelin, H. B. (2004). Weight change and cognitive

performance. International Journal of Obesity and Related Metabolic Disorders, 28(9),

1163–1167. doi:10.1038/sj.ijo.0802721

Buell, J. S. & Dawson-Hughes, S. (2008). Vitamin D and neurocognitive dysfunction:

Preventing ‘‘D”ecline? Molecular Aspects of Medicine, 29, 415–422.

doi:10.1016/j.mam.2008.05.001

Bugg, J. M., DeLosh, E. L., Davalos, D. B., & Davis, H. P. (2007). Age differences in Stroop

interference: Contributions of general slowing and task-specific deficits. Neuropsychology,

Development, and Cognition. Section B, Aging, Neuropsychology and Cognition, 14(2),

155–167. doi:10.1080/138255891007065

Buracchio, T, Dodge, H, Howieson, D, Wasserman, D., & Kaye, J (2010). The trajectory of gait

speed preceding MCI. Archives of Neurology, 67(8), 980–986.

doi:10.1001/archneurol.2010.159

Burkhauser, R. V., & Cawley, J. (2008). Beyond BMI: the value of more accurate measures of

fatness and obesity in social science research. Journal of Health Economics, 27(2), 519–

529. doi:10.1016/j.jhealeco.2007.05.005

Cabeza, R. & Dennis, N. A. (2013). Frontal lobes and aging: Deterioration and compensation. In

D.T. Stuss & R.T. Knight (Eds). Principles of Frontal Lobe Function, 2nd Edition. Oxford

University Press, New York.

203

Campbell, B. C., & Eisenberg, D. (2007). Obesity, attention deficit-hyperactivity disorder and

the dopaminergic reward system. Collegium Antropologicum, 31(1), 33–38.

Campbell, N., Boustani, M., Limbil, T., Ott, C., Fox, C., Maidment, I., Schubert, C.C., et al.

(2009). The cognitive impact of anticholinergics: A clinical review. Clinical Interventions

in Aging, 4, 225–233.

Camponova, P., Le Page, A., Berrougui, H., Lamoureux, J., Pawelec, G., Witkowski, M. J.,

…Khalil, A. (2017). Alteration of high-density lipoprotein functionality in Alzheimer’s

disease patients. Canadian Journal of Physiology and Pharmacology, 95(8), 894–903.

doi:10.1139/cjpp-2016-0710

Carels, R., Wott, C., Young, K., Gumble, A., Koball, A., & Oehlhof, M. (2010). Implicit,

explicit, and internalized weight bias and psychosocial maladjustment among treatment-

seeking adults. Eating Behaviors, 11(3), 180–185. doi:10.1016/j.eatbeh.2010.03.002

Carlucci, M., Smith, M., & Corbridge, S. J. (2013). Poor sleep, hazardous breathing: An

overview of obstructive sleep apnea. The Nurse Practitioner, 38(3), 20–28.

doi:10.1097/01.NPR.0000426409.62476.fa

Castañon, M. M., Lauricella, A. M., Kordich, L., & Quintana, I. (2007). Plasma homocysteine

cutoff values for venous thrombosis. Clinical Chemistry and Laboratory Medicine, 45(2),

232–236. doi:10.1515/CCLM.2007.038

Castres, L., Folope, V., Dechelotte, P., Tourny-Chollet, C., & Lemaitre, F. (2010). Quality of life

and obesity class relationships. International Journal of Sports Medicine, 31, 773–778.

doi:10.1055/s-0030-1262830

Caton, S., Chadwick, D., Chapman, M., Turnbull, S., Mitchell, D., & Stansfield, J. (2012).

Healthy lifestyles for adults with intellectual disability: knowledge, barriers, and facilitators.

Journal of Intellectual and Developmental Disability, 37(3), 248–259.

doi:10.3109/13668250.2012.703645

Cefalu, W. T. (2001). Insulin resistance: cellular and clinical concepts. Experimental Biology

and Medicine, 226(1), 13-26.

Centers for Disease Control and Prevention (2017, May 7). Diabetes public health resource.

Retrieved from https://www.cdc.gov/diabetes/statistics/age/fig2.htm

Chandola, T., Deary, I. J., Blane, D., & Batty, G. D. (2006). Childhood IQ in relation to obesity

and weight gain in adult life: The National Child Development (1958) Study. International

Journal of Obesity, 30(9), 1422–1432. doi:10.1038/sj.ijo.0803279

Chee, M. W. L., Chen, K. H. M., Zheng, H., Chan, K. P. L., Isaac, V., Sim, S. K. Y.,…Ng, T. P.

(2009). Cognitive function and brain structure correlations in healthy elderly East Asians.

Neuroimage, 46(1), 257–269. doi:10.1016/j.neuroimage.2009.01.036

204

Chen, Y. J., Tsai, Y. F., Lee, S. H., & Lee, H. L. (2014). Protective factors against suicide among

young-old Chinese outpatients. BMC Public Health, 14(1), 372. doi:10.1186/1471-2458-14-

372

Cherubini, A., Martin, A., Andres-Lacueva, C., Di Iorio, A., Lamponi, M., Mecocci,

P.,…Ferrucci, L. (2005). Vitamin E levels, cognitive impairment and dementia in older

persons: The InCHIANTI study. Neurobiology of Aging, 26, 987–994.

doi:10.1016/j.neurobiolaging.2004.09.002

Chesnokova, V., Pechnik, R. N., & Wawrowsky, K. (2016). Chronic peripheral inflammation,

hippocampal neurogenesis, and behavior. Brain, Behavior, and Immunity, 58, 1–8.

doi: 10.1016/j.bbi.2016.01.017

Chételat, G., Landeau, B., Salmon, E., Yakushev, I., Bahri, M. A., Mézenge, F.,…Fellgiebel, A.

(2013). Relationships between brain metabolism decrease in normal aging and changes in

structural and functional connectivity. Neuroimage, 76, 167–177.

doi:10.1016/j.neuroimage.2013.03.009

Cheung, B. H. K., Ho, I. C. H., Chan, R. S. M., Sea, M. M. M., & Woo, J. (2014). Current

evidence on dietary pattern and cognitive function. Advances in Food and Nutrition

Research, 71, 137–163.

Chlif, M., Keochkerian, D., Mourlhon, C., Choquet, D., & Ahmaidi, S. (2005). Noninvasive

assessment of the tension-time index of inspiratory muscles at rest in obese male subjects.

International Journal of Obesity, 29(12), 1478–1483. doi:10.1038/sj.ijo.0803030

Choi, C. W., Cho, W. R., Park, K. H., Choi, I. K., Seo, J. H., Kim, B. S.,…Lee, J. (2005). The

cutoff value of serum ferritin for the diagnosis of iron deficiency in community-residing

older persons. Annals of Hematology, 84(6), 358–361. doi:10.1007/s00277-005-1033-5

Chuang, Y. F., An, Y., Bilgel, M., Wong, D. F., Troncoso, J. C., O’Brien, R. J.,… Thambisetty,

M. (2016). Midlife adiposity predicts earlier onset of Alzheimer’s dementia, neuropathology

and presymptomatic cerebral amyloid accumulation. Molecular Psychiatry, 21(7), 910-915.

doi: 10.1038/mp.2015.129

Clarke, R., Refsum, H., Birks, J., Evans, J. G., Johnston, C., Sherliker, P.,… Scott, J. M. (2003).

Screening for vitamin B-12 and folate deficiency in older persons. American Journal of

Clinical Nutrition, 77(5), 1241–1247.

Clasey, J. L., Hartman, M. L., Kanaley, J., Wideman, L., Teates, D. C., Bochard, C., & Weltman,

A. (1997). Body composition by DEXA in older adults, accuracy and influence of scan

mode. Medicine & Science in Sports & Exercise, 29(4), 560–567.

Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control of automatic processes: A

parallel distributed processing account of the Stroop effect. Psychological Review, 97(3),

332–361.

205

Colcombe, S., & Kramer, A. F. (2003). Fitness effects on the cognitive function of older adults:

A meta-analytic study. Psychological Science, 14(2), 125–130. doi:10.1111/1467-9280.t01-

1-01430

Conway, B., Xiang, Y. B., Villegas, R., Zhang, X., Li, H., Wu, X.,…Shu, X. O. (2011). Hip

circumference and the risk of type 2 diabetes in middle-aged and elderly men and women:

the Shanghai women and Shanghai men’s health studies. Annals of Epidemiology, 21(5),

358–366. doi:10.1016/j.annepidem.2011.02.005.Hip

Copeland, J. R. M., McCracken, C. F. M., Dewey, M. E., Wilson, K. C. M., Doran, M., Gilmore,

C.,…Larkin, B. A. (1999). Undifferentiated dementia, Alzheimer’s disease, and vascular

dementia: age- and gender-related incidence in Liverpool. British Journal of Psychiatry,

175, 433–438.

Corica, F., Corsonello, A., Apolone, G., Lucchetti, M., Melchionda, N., & Marchesini, G.

(2006). Construct validity of the short form-36 health survey and its relationship with BMI

in obese outpatients. Obesity, 14(8), 1429–1437.

Cortese, S., & Vincenzi, B. (2012). Obesity and ADHD: Clinical and neurobiological

implications. Current Topics in Behavioral Neurosciences, 9, 199–218.

doi:10.1007/7854_2011_154

Cournot, M., Marquié, J. C., Ansiau, D., Martinaud, C., Fonds, H., Ferrières, J., & Ruidavets, J.

B. (2006). Relation between body mass index and cognitive function in healthy middle-aged

men and women. Neurology, 67(7), 1208–1214. doi:10.1212/01.wnl.0000238082.13860.50

Crider, K. S., Yang, T. P., Berry, R. J., & Bailey, L. B. (2012). Folate and DNA methylation: A

review of molecular mechanisms and evidence for folate's role. (2012). Advances in

Nutrition, 3(1), 21-38. doi:10.3945/an.111.000992

Cserjési, R., Vermeulen, N., Lénárd, L., & Luminet, O. (2011). Reduced capacity in automatic

processing of facial expression in restrictive anorexia nervosa and obesity. Psychiatry

Research, 188(2), 253–257. doi:10.1016/j.psychres.2010.12.008

Cui, R, Liu, Z, & Xu, Q. (2014). Blood α-tocopherol, γ-tocopherol levels and risk of prostate

cancer: A meta-analysis of prospective studies. PLoS One, 9(3), e93044.

doi:10.1371/journal.pone.0093044

Cumming, T. B., Marshall, R. S., & Lazar, R. M. (2013). Stroke, cognitive deficits, and

rehabilitation: Still an incomplete picture. International Journal of Stroke, 8(1), 38–45.

doi:10.1111/j.1747-4949.2012.00972.x

D’Agostino, R. B., Vasan, R. S., Pencina, M. J., Wolf, P. A., Cobain, M., Massaro, J. M., &

Kannel, W. B. (2008). General cardiovascular risk profile for use in primary care: The

Framingham Heart Study. Circulation, 117(6), 743–753.

doi:10.1161/CIRCULATIONAHA.107.699579

206

D’Agostino, R. B., Wolf, P. A., Belanger, A. J., & Kannel, W. B. (1994). Stroke risk profile:

adjustment for antihypertensive medication. The Framingham Study. Stroke, 25(1), 40–43.

doi:10.1161/01.STR.25.1.40

Dahl, A. K. & Reynolds, C. A. (2013). Accuracy of recalled body weight – A study with 20-

years of follow-up. Obesity, 21(6), 1293–1298. doi:10.1002/oby.20299

Dali-Youcef, N., Mecili, M., Ricci, R., & Andres, E. (2013). Metabolic inflammation:

Connecting obesity and insulin resistance. Annals of Medicine, 45, 242–253.

Darvesh, A. S., Carroll, R. T., Bishayee, A., Geldenhuys, W. J., & Van der Schyf, C. J. (2010).

Oxidative stress and Alzheimer’s disease: dietary polyphenols as potential therapeutic

agents. Expert Review of Neurotherapeutics, 10(5), 729–745. doi:10.1586/ern.10.42

Daurat, A., Huet, N., & Tiberge, M. (2010). Metamemory beliefs and episodic memory in

obstructive sleep apnea syndrome. Psychological Reports, 107(1), 289–302.

doi:10.2466/10.13.20.22.PR0.107.4.289-302

Davis, A. S., & Pierson, E. E. (2012). The relationship between the WAIS-III digit symbol

Coding and executive functioning. Applied Neuropsychology: Adult, 19(3), 192–197.

doi:10.1080/09084282.2011.643958

Davis, C. (2010). Attention-deficit/hyperactivity disorder: associations with overeating and

obesity. Current Psychiatry Reports, 12(5), 389–395. doi:10.1007/s11920-010-0133-7

Davis, J. N., Hodges, V. A., & Gillham, M. B. (2006). Normal-weight adults consume more fiber

and fruit than their age- and height-matched overweight/obese counterparts. Journal of the

American Dietetic Association, 106(6), 833–840. doi:10.1016/j.jada.2006.03.013

DCCT/EDIC Study Research Group (2007). Long-term effect of diabetes and its treatment on

cognitive function. New England Journal of Medicine, 356(18), 1842-1852.

De Bruyn, E., Gulbis, B., & Cotton, F. (2014). Serum and red blood cell folate testing for folate

deficiency: New features? European Journal of Haematology, 92(4), 354-359.

doi:10.1111/ejh.12237

De Carolis, A., Giubilei, F., Caselli, G., Casolla, B., Cavallari, M., Vanacore, N.,…Orzi, F.

(2011). Chronic obstructive pulmonary disease is associated with altered

neuropsychological performance in young adults. Dementia and Geriatric Cognitive

Disorders Extra, 1(1), 402–408. doi:10.1159/000333079

de Hollander, E. L., Bemelmans, W. J., Boshuizen, H. C., Friedrich, N., Wallaschofski, H.,

Guallar-Castillón, P.,…WC elderly collaborators (2012). The association between waist

circumference and risk of mortality considering body mass index in 65- to 74-year-olds: A

meta-analysis of 29 cohorts involving more than 58,000 elderly persons. International

Journal of Epidemiology, 41(3), 805–817. doi:10.1093/ije/dys008

207

de Winter, C. F., Bastiaanse, L. P., Hilgenkamp, T. I. M., Evenhuis, H. M., & Echteld, M. A.

(2012). Overweight and obesity in older people with intellectual disability. Research in

Developmental Disabilities, 33(2), 398–405. doi:10.1016/j.ridd.2011.09.022

de Wit, L. M., Fokkema, M., van Straten, A., Lamers, F., Cuijpers, P., & Pennin, B. W. J. H.

(2010). Depressive and anxiety disorders and the association with obesity, physical, and

social activities. Depression and Anxiety, 27(11), 1057–1065. doi:10.1002/da.20738

DeFranco, C., Tarbox, A., & McLaughlin, E. J. (1985). Cognitive deficits as a function of years

of alcohol abuse. American Journal of Drug and Alcohol Abuse, 11(3-4), 279–293.

Deshpande, N., Metter, E. J., Bandinelli, S., Guralnik, J., & Ferrucci, L. (2009). Gait speed under

varied challenges and cognitive decline in older persons: A prospective study. Age and

Ageing, 38(5), 509–514. doi:10.1093/ageing/afp093

Després, J. P. (2012). Body fat distribution and risk of cardiovascular disease: an update.

Circulation, 126(10), 1301–1313. doi:10.1161/CIRCULATIONAHA.111.067264

Deurenberg, P., Yap, M., & Guricci, S. (2002). Asians are different from Caucasians and from

each other in their body mass index/body fat per cent relationship. Obesity Reviews, 3(3),

141–146. doi:10.1046/j.1467-789X.2002.00065.x

Deurenberg, P., Yap, M., & van Staveren, W. A. (1998). Body mass index and percent body fat:

a meta analysis among different ethnic groups. International Journal of Obesity and Related

Metabolic Disorders, 22(12), 1164–1171.

Di Bonito, P., Di Fraia, L., Di Gennaro, L., Vitale, A., Lapenta, M., Scala, A.,…Capaldo, B.

(2007). Impact of impaired fasting glucose and other metabolic factors on cognitive

function in elderly people. Nutrition, Metabolism, and Cardiovascular Diseases, 17(3),

203–208. doi:10.1016/j.numecd.2005.07.011

Dickman, R., Kislov, J., Boaz, M., Ron, Y., Beniashvili, Z., Raz, I.,…Wainstein, J. (2013).

Prevalence of symptoms suggestive of gastroparesis in a cohort of patients with diabetes

mellitus. Journal of diabetes and its complications, 27(4), 376–379.

doi:10.1016/j.jdiacomp.2013.02.004

Dik, M. G., Jonker, C., Comijs, H. C., Deeg, D. J. H., Kok, A., Yaffe, K., & Penninx, B. W.

(2007). Contribution of metabolic syndrome components to cognition in older individuals.

Diabetes Care, 30(10), 2655–2660. doi:10.2337/dc06-1190

Dobbelsteyn, C. J., Joffres, M. R., MacLean, D. R., & Flowerdew, G. (2001). A comparative

evaluation of waist circumference , waist-to-hip ratio and body mass index as indicators of

cardiovascular risk factors. The Canadian Heart Health Surveys. International Journal of

Obesity, 25, 652–661.

208

Doets, E. L., Ueland, P. M., Tell, G. S., Vollset, S. E., Nygård, O. K., Van’t Veer, P.,…Eussen,

S. J. (2014). Interactions between plasma concentrations of folate and markers of vitamin

B(12) status with cognitive performance in elderly people not exposed to folic acid

fortification: the Hordaland Health Study. British Journal of Nutrition, 111(6), 1085–1095.

doi:10.1017/S000711451300336X

Donini, L. M., Poggiogalle, E., Del Balzo, V., Lubrano, C., Faliva, M., Opizzi, A.,…Rondanelli,

M. (2013). How to estimate fat mass in overweight and obese subjects. International

Journal of Endocrinology, 2013, 1–9. doi:10.1155/2013/285680

Dowsey, M., Liew, D., Stoney, J., & Choong, P. (2010). The impact of obesity on weight change

and outcomes at 12 months in patients undergoing total hip arthroplasty. Medical Journal of

Australia, 193(1), 17–21.

Dufouil, C., Godin, O., Chalmers, J., Coskun, O., MacMahon, S., Tzourio-Mazoyer,

N.,…PROGRESS MRI Substudy Investigators (2009). Severe cerebral white matter

hyperintensities predict severe cognitive decline in patients with cerebrovascular disease

history. Stroke, 40(6), 2219–2221. doi:10.1161/STROKEAHA.108.540633

Duncan, J. S., Duncan, E. K., & Schofield, G. (2009). Accuracy of body mass index (BMI)

thresholds for predicting excess body fat in girls from five ethnicities. Asia Pacific Journal

of Clinical Nutrition, 18(3), 404–411.

Durga, J., van Boxtel, M. P. J., Schouten, E. G., Kok, F. J., Jolles, J., Katan, M. B., & Verhoef,

P. (2007). Effect of 3-year folic acid supplementation on cognitive function in older adults

in the FACIT trial: A randomised, double blind, controlled trial. Lancet, 369(9557), 208–

216. doi:10.1016/S0140-6736(07)60109-3

Duthie, S. J. (2011). Folate and cancer: how DNA damage, repair and methylation impact on

colon carcinogenesis. Journal of Inherited Metabolic Disease, 34(1), 101–109.

doi:10.1007/s10545-010-9128-0

Dysken, M. W., Sano, W., Asthana, S., Vertrees, J. E., Pallaki, M., Llorente, M.,…Guarino, P.

D. (2014). Effect of vitamin E and memantine on functional decline in Alzheimer Disease:

the TEAM-AD VA cooperative randomized trial. Journal of the American Medical

Association, 311(1), 33–44. doi:10.1001/jama.2013.282834

Eckart, C., & Bunzeck, N. (2013). Dopamine modulates processing speed in the human

mesolimbic system. Neuroimage, 66, 293–300. doi:10.1016/j.neuroimage.2012.11.001

Eckert, M. A. (2011). Slowing down: age-related neurobiological predictors of processing speed.

Frontiers in Neuroscience, 5(March), 25. doi:10.3389/fnins.2011.00025

Elias, M. F., Elias, P. K., Sullivan, L. M., Wolf, P. A., & D’Agostino, R. B. (2003). Lower

cognitive function in the presence of obesity and hypertension: the Framingham heart study.

209

International Journal of Obesity and Related Metabolic Disorders, 27(2), 260–268.

doi:10.1038/sj.ijo.802225

Elias, M. F., Sullivan, L. M., D’Agostino, R. B., Elias, P. K., Beiser, A., Au, R.,…Wolf, P. A.

(2004). Framingham stroke risk profile and lowered cognitive performance. Stroke, 35(2),

404–9. doi:10.1161/01.STR.0000103141.82869.77

Elkins, J. S., Knopman, D. S., Yaffe, K., & Johnston, S. C. (2005). Cognitive function predicts

first-time stroke and heart disease. Neurology, 64, 1750–1755.

Elwood, E., Lim, Z., Naveed, H., & Galea, I. (2017). The effect of systemic inflammation on

human brain barrier function. Brain, Behavior, and Immunity, 62, 35–40.

doi: 10.1016/j.bbi.2016.10.020

Elwood, P. C., Pickering, J., Bayer, A., & Gallacher, J. E. J. (2002). Vascular disease and

cognitive function in older men in the Caerphilly cohort. Age and Ageing, 31(1), 43–48.

Enkvist, A., Ekström, H., & Elmståhl, S. (2013). Associations between cognitive abilities and

life satisfaction in the oldest-old. Results from the longitudinal population study Good

Aging in Skåne. Clinical Interventions in Aging, 8, 845–853. doi:10.2147/CIA.S45382

Eriksson, U. K., Bennet, A. M., Gatz, M., Dickman, P. W., & Pedersen, N. L. (2010). Nonstroke

cardiovascular disease and risk of Alzheimer disease and dementia. Alzheimer Disease and

Associated Disorders, 24(3), 213–219. doi:10.1097/WAD.0b013e3181d1b99b

Evans, W. J., & Campbell, W. W. (1993). Sarcopenia and Age-Related Changes in Body

Composition and Functional Capacity. American Institute of Nutrition, Feb 123(2 Suppl),

465–468.

Eyles, D. W., Smith, S., Kinobe, R., Hewison, M., & McGrath, J. J. (2005). Distribution of the

Vitamin D receptor and 1α-hydroxylase in human brain. Journal of Chemical

Neuroanatomy, 29, 21-30.

Farmer, M. E., Kittner, S. J., Abbott, R. D., Wolz, M. M., Wolf, P. A., & White, L. R. (1990).

Longitudinally measured blood pressure, antihypertensive medication use, and cognitive

performance: The Framingham Study. Journal of Clinical Epidemiology, 43(5), 475–480.

Feldman, D. & Malloy, P. J. (2014). Mutations in the vitamin D receptor and hereditary vitamin

D-resistant rickets. BoneKEy Reports, 3, 510. doi: 10.1038/bonekey.2014.5

Feng, L., Isaac, V., Sim, S., Ng, T. P., Krishnan, K. R. R., & Chee, M. W. L. (2013).

Associations between elevated homocysteine, cognitive impairment, and reduced white

matter volume in healthy old adults. American Journal of Geriatric Psychiatry, 21(2), 164–

172.

210

Feng, L., Ng, T. P., Chuah, L., Niti, M., & Kua, E. H. (2006). Homocysteine, folate, and vitamin

B-12 and cognitive performance in older Chinese adults: findings from the Singapore

Longitudinal Ageing Study. American Journal of Clinical Nutrition, 84(6), 1506–1512.

Fergenbaum, J. H., Bruce, S., Lou, W., Hanley, A. J. G., Greenwood, C., & Young, T. K. (2009).

Obesity and lowered cognitive performance in a Canadian First Nations population.

Obesity, 17(10), 1957–1963. doi:10.1038/oby.2009.161

Finkel, D., Reynolds, C. A., McArdle, J. J., & Pedersen, N. L. (2007). Age changes in processing

speed as a leading indicator of cognitive aging. Psychology and Aging, 22(3), 558–568.

doi:10.1037/0882-7974.22.3.558

Finkelstein, E. A., Trogdon, J. G., Cohen, J. W., & Dietz, W. (2009). Annual medical spending

attributable to obesity: payer-and service-specific estimates. Health Affairs (Project Hope),

28(5), w822–w831. doi:10.1377/hlthaff.28.5.w822

Fioravanti, M., Nacca, D., Buckley, A. E., & Bisetti, A. (1995). Chronic obstructive pulmonary

disease and associated patterns of memory decline. Dementia, 6, 39–48.

Fjell, A. M., & Walhovd, K. B. (2010). Structural brain changes in aging: courses, causes and

cognitive consequences. Reviews in the Neurosciences, 21(3), 187–221.

Flaherty, J. H. (1998). Psychotherapeutic agents in older adults. Commonly prescribed and over-

the-counter remedies: causes of confusion. Clinics in Geriatric Medicine, 14(1), 101–127.

Flegal, K. M., Carroll, M. D., Kuczmarski, R. J., & Johnson, C. L. (1998). Overweight and

obesity in the United States: prevalence and trends, 1960-1994. International Journal of

Obesity, 22, 39-47.

Flegal, K. M., Carroll, M. D., Ogden, C. L., & Curtin, L. R. (2010). Prevalence and trends in

obesity among US adults, 1999-2008. Journal of the American Medical Association, 303(3),

235–241. doi:10.1001/jama.2009.2014

Flegal, K. M., Graubard, B. I., Williamson, D. F., & Cooper, R. S. (2011). Reverse causation and

illness-related weight loss in observational studies of body weight and mortality. American

Journal of Epidemiology, 173(1), 1–9. doi:10.1093/aje/kwq341

Folch, J., Pedrós, I., Patraca, I., Sureda, F., Junyent, F., Beas-Zarate, C., Verdaguer, E., et al.

(2012). Neuroprotective and anti-ageing role of leptin. Journal of Molecular

Endocrinology, 49(3), R149–R156. doi:10.1530/JME-12-0151

Forton, D. M., Allsop, J. M., Cox, I. J., Hamilton, G., Wesnes, K., Thomas, H. C., & Taylor-

Robinson, S. D. (2005). A review of cognitive impairment and cerebral metabolite

abnormalities in patients with hepatitis C infection. AIDS, 19(S3), S53–S63.

211

Forton, D. M., Thomas, H. C., Murphy, C. A., Allsop, J. M., Foster, G. R., Main, J.,…Taylor-

Robinson, S. D. (2002). Hepatitis C and cognitive impairment in a cohort of patients with

mild liver disease. Hepatology, 35(2), 433–9. doi:10.1053/jhep.2002.30688

Framingham Heart Study (2014, August 18). Framingham risk functions. Retrieved from

https://www.framinghamheartstudy.org/risk-functions

Frier, B. M. (2011). Cognitive functioning in type 1 diabetes: The Diabetes Control and

Complications Trial (DCCT) revisited. Diabetologia, 54(2), 233–236. doi:10.1007/s00125-

010-1983-6

Funahashi, H. (2003). Distribution, function, and properties of leptin receptors in the brain.

International Review of Cytology, 224, 1–27. doi:10.1016/S0074-7696(05)24001-9

Gába, A., & Přidalová, M. (2014). Age-related changes in body composition in a sample of

Czech women aged 18-89 years: a cross-sectional study. European Journal of Nutrition,

53(1), 167–176. doi:10.1007/s00394-013-0514-x

Gabriely, I., Ma, X. H., Yang, X. M., Rossetti, L., & Barzilai, N. (2002). Leptin resistance during

aging is independent of fat mass. Diabetes, 51(April), 1016–1021.

Gage, F. H., Kelly, P. A. T., & Bjorklund, A. (1984). Regional changes in brain glucose

metabolism reflect cognitive impairments in aged rats. The Journal of Neuroscience, 4(11),

2856–2865.

Gale, S. D., & Hopkins, R. O. (2004). Effects of hypoxia on the brain: Neuroimaging and

neuropsychological findings following carbon monoxide poisoning and obstructive sleep

apnea. Journal of the International Neuropsychological Society, 10(1), 60–71.

doi:10.1017/S1355617704101082

Gandhi, R., Razak, F., Davey, J. R., & Mahomed, N. N. (2010). Metabolic syndrome and the

functional outcomes of hip and knee arthroplasty. The Journal of Rheumatology, 37(9),

1917–1922. doi:10.3899/jrheum.091242

Garatachea, N., Luque, G. T., & Gallego, J. G. (2010). Physical activity and energy expenditure

measurements using accelerometers in older adults. Nutricion Hospitalaria, 25(2), 224–230.

doi:10.3305/nh.2010.25.2.4439

Gard, P. R. (2010). Non-adherence to antihypertensive medication and impaired cognition:

Which comes first? The International Journal of Pharmacy Practice, 18(5), 252–259.

doi:10.1111/j.2042-7174.2010.00045.x

Gasecki, D., Kwarciany, M., Nyka, W., & Narkiewicz, K. (2013). Hypertension, Brain Damage,

and Cognitive Decline. Current Hypertension Reports, 15(6), 547-558. doi:10.1007/s11906-

013-0398-4

212

Gavin, A. R., Simon, G. E., & Ludman, E. J. (2010). The association between obesity,

depression, and educational attainment in women: the mediating role of body image

dissatisfaction. Journal of Psychosomatic Research, 69(6), 573–581.

doi:10.1016/j.jpsychores.2010.05.001

Gazdzinski, S., Millin, R., Kaiser, L. G., Durazzo, T. C., Mueller, S. G., Weiner, M. W., &

Meyerhoff, D. J. (2010). BMI and neuronal integrity in healthy, cognitively normal elderly:

a proton magnetic resonance spectroscopy study. Obesity, 18(4), 743–748.

doi:10.1038/oby.2009.325

Gibson, E. L. (2012). The psychobiology of comfort eating: implications for

neuropharmacological interventions. Behavioural Pharmacology, 23(5-6), 442–460.

doi:10.1097/FBP.0b013e328357bd4e

Gimeno, D., Marmot, M.G., & Singh-Manoux, A. (2008). Inflammatory markers and cognitive

function in middle-aged adults: the Whitehall II study. Psychoneuroendocrinology. 33(10),

1322–1334. doi: 10.1016/j.psyneuen.2008.07.006

Godbout, J. P., Chen, J., Abraham, J., Richwine, A. F., Berg, B. M., Kelley, K. W., & Johnson,

R. W. (2005). Exaggerated neuroinflammation and sickness behavior in aged mice after

activation of the peripheral innate immune system. The FASEB Journal. doi:10.1096/fj.05-

3776fje

Golden, E., & Hatcher, J. (1997). Nutrition knowledge and obesity of adults in community

residences. Mental Retardation, 35(3), 177–184.

González-Blanch, C., Pérez-Iglesias, R., Rodríguez-Sánchez, J. M., Pardo-García, G., Martínez-

García, O., Vázquez-Barquero, J. L., & Crespo-Facorro, B. (2011). A digit symbol coding

task as a screening instrument for cognitive impairment in first-episode psychosis. Archives

of Clinical Neuropsychology, 26(1), 48–58. doi:10.1093/arclin/acq086

Gouveia, L. A. G., Marucci, M. D. F. N., Lebrão, M. L., & Duarte, Y. A. O. (2014). Association

between waist circumference (WC) values and hypertension, heart disease (HD) and

diabetes, reported by the elderly—SABE Survey: Health, Wellness and Aging, 2000 and

2006. Archives of Gerontology and Geriatrics, 59(1), 62–68.

doi:10.1016/j.archger.2014.02.005

Gradman, T. J., Laws, A., Thompson, L. W., & Reaven, G. M. (1993). Verbal learning and/or

memory improves with glycemic control in older subjects with non-insulin-dependent

diabetes mellitus. Journal of the American Geriatrics Society, 41(12), 1305–1312.

Graham, I. M., Daly, L. E., Refsum, H. M., Robinson, K., Brattstrom, L. E., Ueland, P.

M.,…Andria, G. (1997). Plasma homocysteine as a risk factor for vascular disease. Journal

of the American Medical Association, 277(22), 1775–1781.

213

Gray, S. L., Anderson, M. L., Dublin, S., Hanlon, J. T., Hubbard, R., Walker, R.,…Larson, E. B.

(2015). Cumulative use of strong anticholinergics and incident dementia: a prospective

cohort study. JAMA Internal Medicine, 175(3), 401–417.

doi:10.1001/jamainternmed.2014.7663

Gray, S. L. & Hanlon, J. T. (2016). Anticholinergic medication use and dementia: latest evidence

and clinical implications. Therapeutic Advances in Drug Safety, 7(5), 217–224.

doi:10.1177/2042098616658399

Green, J., Forster, A., & Young, J. (2002). Reliability of gait speed measured by a timed walking

test in patients one year after stroke. Clinical Rehabilitation, 16(3), 306–314.

doi:10.1191/0269215502cr495oa

Greenwood, C. E., & Winocur, G. (2005). High-fat diets, insulin resistance and declining

cognitive function. Neurobiology of Aging, 26(Suppl 1), 42–45.

doi:10.1016/j.neurobiolaging.2005.08.017

Gualtieri, C. T., Johnson, L. G., & Benedict, K. B. (2006). Neurocognition in depression :

Patients on and off medication versus healthy comparison subjects. The Journal of

Neuropsychiatry & Clinical Neurosciences, 18(2), 217–225.

Guasch-Ferré, M., Bulló, M., Martínez-González, M. Á., Corella, D., Estruch, R., Covas, M.

I.,…Salas-Salvadó J. (2012). Waist-to-height ratio and cardiovascular risk factors in elderly

individuals at high cardiovascular risk. PLoS One, 7(8), e43275.

doi:10.1371/journal.pone.0043275

Gunstad, J., Paul, R., Cohen, R., Tate, D., & Gordon, E. (2006). Obesity is associated with

memory deficits in young and middle-aged adults. Eating and Weight Disorders, 11, e15–

e19.

Gunstad, J., Paul, R. H., Cohen, R. A, Tate, D. F., Spitznagel, M. B., & Gordon, E. (2007).

Elevated body mass index is associated with executive dysfunction in otherwise healthy

adults. Comprehensive Psychiatry, 48(1), 57–61. doi:10.1016/j.comppsych.2006.05.001

Guo, R. B., Sun, P. L., Zhao, A. P., Gu, J., Ding, X., Qi, J.,…Hu, G. (2013). Chronic asthma

results in cognitive dysfunction in immature mice. Experimental Neurology, 247, 209–217.

doi:10.1016/j.expneurol.2013.04.008

Gustafson, D. R. (2010). Adiposity hormones and dementia. Journal of the Neurological

Sciences, 299(1-2), 30–34. doi:10.1016/j.jns.2010.08.036

Hajjar, I., Keown, M., & Frost, B. (2005). Antihypertensive agents for aging patients who are at

risk for cognitive dysfunction. Current Hypertension Reports, 7(6), 466–473.

Haltia, L. T., Viljanen, A., Parkkola, R., Kemppainen, N., Rinne, J. O., Nuutila, P., & Kaasinen,

V. (2007). Brain white matter expansion in human obesity and the recovering effect of

214

dieting. Journal of Clinical Endocrinology & Metabolism, 92(8), 3278–3284.

doi:10.1210/jc.2006-2495

Hanson, A. J., Bayer-Carter, J. L., Green, P. S., Montine, T. J., Wilkinson, C. W., Baker, L.

D.,…Craft, S. (2013). Effect of apolipoprotein E genotype and diet on apolipoprotein E

lipidation and amyloid peptides: randomized clinical trial. JAMA Neurology, 70(8), 972–

980. doi:10.1001/jamaneurol.2013.396

Harada, C. N., Natelson Love, M. C., & Triebel, K. L. (2013). Normal cognitive aging. Clinics in

Geriatric Medicine, 29, 737–752.

Hamid, J. J. M., Amal, K. M., Rohani, A., & Norimah, A. K. (2010). Association of iron

deficiency with or without anemia and cognitive functions among primary school children

in Malaysia. Malaysian Journal of Nutrition, 16(2), 261-270.

Harrison, F. E. (2012). A critical review of vitamin C for the prevention of age-related cognitive

decline and Alzheimer’s disease. Journal of Alzheimer’s Disease, 29(4), 711–726.

doi:10.3233/JAD-2012-111853

Harrison, M., Morrell, J., & Hopman, W. (2004). Influence of obesity on outcome after knee

arthroscopy. Arthroscopy, 20(7), 691–695. doi:10.1016/j.arthro.2004.06.004

Hasher, L., Stoltzfus, E. R., Zacks, R. T., & Rypma, B. (1991). Age and inhibition. Journal of

Experimental Psychology: Learning, Memory, & Cognition, 17(1), 163–169.

Harvey, J. (2007). Leptin: a diverse regulator of neuronal function. Journal of Neurochemistry,

100(2), 307–313. doi:10.1111/j.1471-4159.2006.04205.x

Hawkins, M. A. W., & Stewart, J. C. (2012). Do negative emotional factors have independent

associations with excess adiposity? Journal of Psychosomatic Research, 73(4), 243–250.

doi:10.1016/j.jpsychores.2012.07.009

He, M., Li, E.T.S, & Kung, A.W.C. (1999). Dual-energy X-ray absorptiometry for body

composition estimation in Chinese women. European Journal of Clinical Nutrition, 53,

933–937.

He, W., Goodkind, D., & Kowal, P. (2016). U.S. Census Bureau, International Population

Reports, P95/16-1, An Aging World: 2015, U.S. Government Publishing Office,

Washington, DC, 2016.

Hebert, L. E., Scherr, P. A., Bennett, D. A., Bienias, J. L., Wilson, R. S., Morris, M. C., & Evans,

D. A. (2004). Blood pressure and late-life cognitive function change: A biracial longitudinal

population study. Neurology, 62(11), 2021–2024.

Hector, M., & Burton, J. R. (1988). What are the psychiatric manifestations of vitamin b12

deficiency? Journal of the American Geriatrics Society, 36(12), 1105–1012.

215

Heini, A. F., Lara-Castro, C., Kirk, K. A., Considine, R. V., Caro, J. F., & Weinsier, R. L.

(1998). Association of leptin and hunger-satiety ratings in obese women. International

Journal of Obesity and Related Metabolic Disorders, 22(11), 1084–1087.

Henriksen, E. J., Diamond-Stanic, M. K., & Marchionne, E. M. (2011). Oxidative stress and the

etiology of insulin resistance and type 2 diabetes. Free Radical Biology and Medicine,

51(5), 993–999. doi:10.1016/j.freeradbiomed.2010.12.005

Heo, M., Faith, M. S., Pietrobelli, A., & Heymsfield, S. B. (2012). Percentage of body fat cutoffs

by sex, age, and race-ethnicity in the US adult population from NHANES 1999 – 2004.

American Journal of Clinical Nutrition, 95, 594–602. doi:10.3945/ajcn.111.025171.1

Herbert V. (1962). Experimental nutritional folate deficiency in man. Transactions of the

Association of American Physicians, 75, 307–320.

Heringa, S. M., van den Berg, E., Reijmer, Y. D., Nijpels, G., Stehouwer, C. D. A., Schalkwijk,

C. G., Teerlink, T., et al. (2014). Markers of low-grade inflammation and endothelial

dysfunction are related to reduced information processing speed and executive functioning

in an older population - the Hoorn Study. Psychoneuroendocrinology, 40, 108–118.

doi:10.1016/j.psyneuen.2013.11.011

Herpertz, S., Burgmer, R., Stang, A., de Zwaan, M., Wolf, A. M., Chen-Stute, A.,…Senf, W.

(2006). Prevalence of mental disorders in normal-weight and obese individuals with and

without weight loss treatment in a German urban population. Journal of Psychosomatic

Research, 61(1), 95–103. doi:10.1016/j.jpsychores.2005.10.003

Hewer, W., Mussell, M., Rist, F., Kulzer, B., & Bergis, K. (2003). Short-term effects of

improved glycemic control on cognitive function in patients with type 2 diabetes.

Gerontology, 49(2), 86–92. doi:10.1159/000067947

Hin, H., Clarke, R., Sherliker, P., Atoyebi, W., Emmens, K., Birks, J.,…Evans, J. G. (2006).

Clinical relevance of low serum vitamin B12 concentrations in older people: the Banbury

B12 study. Age and Ageing, 35, 416–22. doi:10.1093/ageing/afl033

Hoth, K. F., Poppas, A., Moser, D. J., Paul, R. H., & Cohen, R. A. (2008). Cardiac dysfunction

and cognition in older adults with heart failure. Cognitive and Behavioral Neurology, 21(2),

65–72. doi:10.1097/WNN.0b013e3181799dc8

Hsu, W. C., Chu, Y. C., Fung, H. C., Wai, Y. Y., Wang, J. J., Lee, J. D., & Chen, Y. C. (2016).

Digit symbol substitution test score and hyperhomocysteinemia in older adults.

Medicine (Baltimore), 95(35), e4535. doi:10.1097/MD.0000000000004535

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:

Conventional criteria versus new alternatives. Structural Equation Modeling: A

Multidisciplinary Journal, 6(1), 1–55. doi:10.1080/10705519909540118

216

Huckans, M., Seelye, A., Parcel, T., Mull, L., Woodhouse, J., Bjornson, D.,…Hauser, P. (2009).

The cognitive effects of hepatitis C in the presence and absence of a history of substance

use disorder. Journal of the International Neuropsychological Society, 15(1), 69–82.

doi:10.1017/S1355617708090085

Hudetz, J. A., & Warltier, D. C. (2007). Cognitive function in older diabetic subjects with a

history of alcohol abuse. Psychological Reports, 1001(16), 1125–1132.

Incalzi, R. A., Gemma, A., Marra, C., Muzzolon, R., Capparella, O., & Carbonin, P. (1993).

Chronic obstructive pulmonary disease: An original model of cognitive decline. The

American Review of Respiratory Disease, 148(2), 418–424.

Iversen, A., Jensen, J. S., Scharling, H., & Schnohr, P. (2009). Hypercholesterolaemia and risk of

coronary heart disease in the elderly: Impact of age: The Copenhagen City Heart Study.

European Journal of Internal Medicine, 20(2), 139–144. doi:10.1016/j.ejim.2008.06.003

Iwasa, H., Gondo, Y., Yoshida, Y., Kwon, J., Inagaki, H., Kawaai, C.,…Suzuki, T. (2008).

Cognitive performance as a predictor of functional decline among the non-disabled elderly

dwelling in a Japanese community: A 4-year population-based prospective cohort study.

Archives of Gerontology and Geriatrics, 47(1), 139–149. doi:10.1016/j.archger.2007.07.008

Iwasa, H., Kai, I., Yoshida, Y., Suzuki, T., Kim, H., & Yoshida, H. (2014). Information

processing speed and 8-year mortality among community-dwelling elderly Japanese.

Journal of Epidemiology, 24(1), 52–59. doi:10.2188/jea.JE20120210

Jackson, A. S., Janssen, I., Sui, X., Church, T. S., & Blair, S. N. (2012). Longitudinal changes in

body composition associated with healthy ageing: Men, aged 20–96 years. British Journal

of Nutrition, 107(7), 1085–1091. doi:10.1017/S0007114511003886

Janssen, I., Katzmarzyk, P. T., & Ross, R. (2002). Body mass index, waist circumference, and

health risk. Archives of Internal Medicine, 162, 2074–2079.

Janssen, I., Katzmarzyk, P. T., & Ross, R. (2004). Waist circumference and not body mass index

explains obesity-related health risk. American Journal of Clinical Nutrition, 79(3), 379–

384.

Jasinska, A. J., Yasuda, M., Burant, C. F., Gregor, N., Khatri, S., Sweet, M., & Falk, E. B.

(2012). Impulsivity and inhibitory control deficits are associated with unhealthy eating in

young adults. Appetite, 59(3), 738–747. doi:10.1016/j.appet.2012.08.001

Jassal, S. V., Roscoe, J., LeBlanc, D., Devins, G. M., & Rourke, S. (2008). Differential

impairment of psychomotor efficiency and processing speed in patients with chronic kidney

disease. International Urology and Nephrology, 40(3), 849–854. doi:10.1007/s11255-008-

9375-2

217

Jelicic, M., Jonker, C., & Deeg, D. J. H. (2001). Effect of low levels of serum vitamin B12 and

folic acid on cognitive performance in old age: A population-based study. Developmental

Neuropsychology, 20(3), 565–571. doi:10.1207/S15326942DN2003_1

Jeong, J. J. (2008). Chronic gastrointestinal symptoms and quality of life in the Korean

population. World Journal of Gastroenterology, 14(41), 6388. doi:10.3748/wjg.14.6388

Jeong, S. K., Cho, J. H., & Kim, J. M. (2004). The usefulness of the Korean version of the

modified MMSE (K-mMMSE) for dementia screening in community dwelling older people.

BMC Public Health, 4, 31. doi:10.1186/1471-2458-4-31

Jeong, S. K., Nam, H. S., Son, M. H., Son, E. J., & Cho, K. H. (2005). Interactive effect of

obesity indexes on cognition. Dementia and Geriatric Cognitive Disorders, 19, 91–96.

doi:10.1159/000082659

Johnson, L. E. (2016a, September). Vitamin B 12 - Nutritional Disorders. Retrieved March 25,

2017, from http://www.merckmanuals.com/professional/nutritional-disorders/vitamin-

deficiency,-dependency,-and-toxicity/vitamin-b-12#v885010

Johnson, L. E. (2016b, September). Vitamin D - Nutritional Disorders. Retrieved August 13,

2017, from https://www.merckmanuals.com/professional/nutritional-disorders/vitamin-

deficiency,-dependency,-and-toxicity/vitamin-d#v885227

Johnson, L. E. (2016c, September). Vitamin E - Nutritional Disorders. Retrieved March 25,

2017, from http://www.merckmanuals.com/professional/nutritional-disorders/vitamin-

deficiency,-dependency,-and-toxicity/vitamin-e

Johnson, L. E. (2016d, September). Folate - Nutritional Disorders. Retrieved May 14, 2017, from

http://www.merckmanuals.com/professional/nutritional-disorders/vitamin-deficiency,-

dependency,-and-toxicity/folate

Johnson, P. M., & Kenny, P. J. (2010). Dopamine D2 receptors in addiction-like reward

dysfunction and compulsive eating in obese rats. Nature Neuroscience, 13(5), 635–643.

Joosten, H., van Eersel, M. E. A., Gansevoort, R. T., Bilo, H. J. G., Slaets, J. P. J., & Izaks, G. J.

(2013). Cardiovascular risk profile and cognitive function in young, middle-aged, and

elderly subjects. Stroke, 44(6), 1543–1549. doi:10.1161/STROKEAHA.111.000496

Joy, S, Fein, D., & Kaplan, E. (2003). Decoding digit symbol: Speed, memory, and visual

scanning. Assessment, 10(1), 56–65. doi:10.1177/0095399702250335

Joy, S, Fein, D., Kaplan, E., & Freedman, M. (2000). Speed and memory in WAIS-R-NI Digit

Symbol performance among healthy older adults. Journal of the International

Neuropsychological Society, 6(7), 770–780.

218

Joy, S., Kaplan, E., & Fein, D. (2004). Speed and memory in the WAIS-III Digit Symbol—

Coding subtest across the adult lifespan. Archives of Clinical Neuropsychology, 19(6), 759–

767. doi:10.1016/j.acn.2003.09.009

Kaffashian, S., Dugravot, A., Nabi, H., Batty, G. D., Brunner, E., Kivimäki, M., & Singh-

Manoux, A. (2011). Predictive utility of the Framingham general cardiovascular disease risk

profile for cognitive function: evidence from the Whitehall II study. European Heart

Journal, 32(18), 2326–2332. doi:10.1093/eurheartj/ehr133

Kagawa, M., Uenishi, K., Kuroiwa, C., Mori, M., & Binns, C. (2006). Is the BMI cut-off level

for Japanese females for obesity set too high? Asia Pacific Journal of Clinical Nutrition,

15(4), 502–507.

Karakis, I., Pase, M. P., Beiser, A., Booth, S. L., Jacques, P. F., Rogers, G.,…Seshadri, S.

(2016). Association of serum vitamin D with the risk of incident dementia and subclinical

indices of brain aging: The Framingham heart study. Journal of Alzheimer’s Disease, 51(2),

451–461. doi:10.3233/JAD-150991

Kane, M. J., & Engle, R. W. (2002). The role of prefrontal cortex in working-memory capacity,

executive attention, and general fluid intelligence: an individual-differences perspective.

Psychonomic Bulletin & Review, 9(4), 637–671.

Kaufman, A. S., & Horn, J. L. (1996). Age changes on tests of fluid and crystallized ability for

women and men on the Kaufman Adolescent and Adult Intelligence Test (KAIT) at ages

17-94 years. Archives of Clinical Neuropsychology, 11(2), 97–121.

Keage, H. A. D., Banks, S., Yang, K. L., Morgan, K., Brayne, C., & Matthews, F. E. (2012).

What sleep characteristics predict cognitive decline in the elderly? Sleep Medicine, 13(7),

886–892. doi:10.1016/j.sleep.2012.02.003

Kikkert, L. H. J., Vuillerme, N., van Campen, J. P., Hortobagyi, T., & Lamoth, C. J. (2016).

Walking ability to predict future cognitive decline in old adults: A scoping review. Ageing

Research Reviews, 27, 1–14. doi:10.1016/j.arr.2016.02.001

Killane, I., Donoghue, O. A., Savva, G. M., Cronin, H., Kenny, R. A., & Reilly, R. B. (2014).

Relative Association of Processing Speed, Short-Term Memory and Sustained Attention

with Task on Gait Speed. The Journals of Gerontology. Series A, Biological Sciences and

Medical Sciences. 69(11), 1407–1414. doi:10.1093/gerona/glu140

Kim, H. A., Miller, A. A, Drummond, G. R., Thrift, A. G., Arumugam, T. V., Phan, T.

G.,…Sobey, C. G.(2012). Vascular cognitive impairment and Alzheimer’s disease: role of

cerebral hypoperfusion and oxidative stress. Naunyn-Schmiedeberg’s Archives of

Pharmacology, 385(10), 953–959. doi:10.1007/s00210-012-0790-7

Kim, J. M., Stewart, R., Kim, S. W., Shin, I. S., Yang, S. J., Shin, H. Y., & Yoon, J. S. (2008).

Changes in folate, vitamin B12 and homocysteine associated with incident dementia.

219

Journal of Neurology, Neurosurgery, and Psychiatry, 79(8), 864–868.

doi:10.1136/jnnp.2007.131482

Kirk, S. F. L., Kuhle, S., Ohinmaa, A., Colman, I., & Veugelers, P. J. (2012). Health care

utilization from prevalent medical conditions in normal-weight, overweight, and obese

children. Journal of Pediatrics, 160(2), 216–221. doi:10.1016/j.jpeds.2011.08.015

Kirton, J. W. & Dotson, V. M. (2016). The interactive effects of age, education, and BMI on

cognitive functioning. Neuropsychology, Development, and Cognition. Section B, Aging,

Neuropsychology, and Cognition, 23(2), 253–262. doi:10.1080/13825585.2015.1082531

Kivipelto, M., Laakso, M. P., Tuomilehto, J., Nissinen, A., & Soininen, H. (2002). Hypertension

and hypercholesterolaemia as risk factors for Alzheimer’s disease: Potential for

pharmacological intervention. CNS Drugs, 16(7), 435–444.

Koch, K. L. & Calles-Escandón, J. (2015). Diabetic gastroparesis. Gastroenterology Clinics of

North America. 44(1), 39-57. doi:10.1016/j.gtc.2014.11.005

Koch, M. & Jensen, M. K. (2016). HDL-cholesterol and apolipoproteins in relation to dementia.

Current Opinion in Lipidology, 27(1), 76–87. doi:10.1097/MOL.0000000000000257

Kodl, C., & Seaquist, E. (2008). Cognitive dysfunction and diabetes mellitus. Endocrine

Reviews, 29(4), 494–511.

Kohli, B. S., Griggs, J. J., Roscoe, J. A., Jean-Pierre, P., Bole, C., Mustian, K. M.,…Morrow, G.

R. (2007). Self-reported cognitive impairment in patients with cancer. Journal of Oncology

Practice, 3(2), 54–59.

Kowalski, K., Rhodes, R., Naylor, P. J., Tuokko, H., & MacDonald, S. (2012). Direct and

indirect measurement of physical activity in older adults: A systematic review of the

literature. International Journal of Behavioral Nutrition and Physical Activity, 9(1), 148.

doi:10.1186/1479-5868-9-148

Kramer, A. F., Hahn, S., Cohen, N. J., Banich, M. T., McAuley, E., Harrison, C. R.,…Colcombe,

A. (1999). Ageing, fitness and neurocognitive function. Nature, 400, 418–419.

Kramer, C. K., von Mühlen, D., Kritz-Silverstein, D., & Barrett-Connor, E. (2009). Treated

hypothyroidism, cognitive function, and depressed mood in old age: The Rancho Bernardo

Study. European Journal of Endocrinology, 161(6), 917–921. doi:10.1530/EJE-09-0606

Kreiner, D. S., & Ryan, J. J. (2001). Memory and motor skill components of the WAIS-III Digit

Symbol-Coding subtest. The Clinical Neuropsychologist, 15(1), 109–113.

doi:10.1076/clin.15.1.109.1906

220

Kuczmarski, M. F., Allegro, D., & Stave, E. (2014). The association of healthful diets and

cognitive function: A review. Journal of Nutrition in Gerontology and Geriatrics, 33(2),

69–90. doi:10.1080/21551197.2014.907101

Kullberg, J., Brandberg, J., Angelhed, J.-E., Frimmel, H., Bergelin, E., Strid, L., Ahlström, H., et

al. (2009). Whole-body adipose tissue analysis: Comparison of MRI, CT and dual energy

X-ray absorptiometry. British Journal of Radiology, 82(974), 123–130.

doi:10.1259/bjr/80083156

Kuo, H. K., Jones, R. N., Milberg, W. P., Tennstedt, S., Talbot, L., Morris, J. N., & Lipsitz, L. A.

(2006). Cognitive function in normal-weight, overweight, and obese older adults: An

analysis of the Advanced Cognitive Training for Independent and Vital Elderly cohort.

Journal of the American Geriatrics Society, 54(1), 97–103. doi:10.1111/j.1532-

5415.2005.00522.x

Kupusinac, A., Stokić, E., & Doroslovački, R. (2014). Predicting body fat percentage based on

gender, age and BMI by using artificial neural networks. Computer Methods and Programs

in Biomedicine, 113(2), 610–619. doi:10.1016/j.cmpb.2013.10.013

Kyle, U. G., Genton, L., Slosman, D. O., & Pichard, C. (2001). Fat-free and fat mass percentiles

in 5225 healthy subjects aged 15 to 98 years. Nutrition, 17(7-8), 534–541.

Labad, J., Price, J. F., Strachan, M. W. J., Deary, I. J., Seckl, J. R., Sattar, N., & Reynolds, R. M.

(2012). Serum leptin and cognitive function in people with type 2 diabetes. Neurobiology of

Aging, 33(12), 2938–2941. doi:10.1016/j.neurobiolaging.2012.02.026

Lafont, S., Marin-Lamellet, C., Paire-Ficout, L., Thomas-Anterion, C., Laurent, B., &

Fabrigoule, C. (2010). The Wechsler Digit Symbol Substitution Test as the best indicator of

the risk of impaired driving in Alzheimer disease and normal aging. Dementia and Geriatric

Cognitive Disorders, 29(2), 154–163. doi:10.1159/000264631

Lal, C., Strange, C., & Bachman, D. (2012). Neurocognitive impairment in obstructive sleep

apnea. Chest, 141(6), 1601–1610. doi:10.1378/chest.11-2214

Lapointe, M. L. B., Blanchette, I., Duclos, M., Langlois, F., Provencher, M. D., & Tremblay, S.

(2013). Attentional bias, distractibility and short-term memory in anxiety. Anxiety, Stress, &

Coping, 26(3), 293–313.

Lear, S. A., James, P. T., Ko, G. T., & Kumanyika, S. (2010). Appropriateness of waist

circumference and waist-to-hip ratio cutoffs for different ethnic groups. European Journal

of Clinical Nutrition, 64(1), 42–61. doi:10.1038/ejcn.2009.70

Lee, D. M., Tajar, A., Ulubaev, A., Pendleton, N., O’Neill, T. W., O’Connor, D. B.,…EMAS

study group (2009). Association between 25-hydroxyvitamin D levels and cognitive

performance in middle-aged and older European men. Journal of Neurology, Neurosurgery,

and Psychiatry, 80, 722–729. doi:10.1136/jnnp.2008.165720

221

Lee, H. B., Richardson, A. K., Black, B. S., Shore, A. D., Kasper, J. D., & Rabins, P. V. (2012).

Race and cognitive decline among community-dwelling elders with mild cognitive

impairment: findings from the Memory and Medical Care Study. Aging & Mental Health,

16(3), 372–377. doi:10.1080/13607863.2011.609533

Lee, J. S. W., Auyeung, T. W., Chau, P. P. H., Hui, E., Chan, F., Chi, I., & Woo, J. (2014).

Obesity can benefit survival-a 9-year prospective study in 1614 Chinese nursing home

residents. Journal of the American Medical Directors Association, 15(5), 342–348.

doi:10.1016/j.jamda.2013.12.081

Lee, S., Choi, S., Kim, H.J., Chung, Y. S., Lee, K. W., Lee, H. C.,…Kim, D. J. (2006). Cutoff

values of surrogate measures of insulin resistance for metabolic syndrome in Korean non-

diabetic adults. Journal of Korean Medical Science, 21(4), 695–700.

doi:10.3346/jkms.2006.21.4.695

Leishear, K. (2011) Changes in vitamin B12, homocysteine, and neurological function in older

adults. Doctoral Dissertation, University of Pittsburgh.

Leishear, K., Boudreau, R. M., Studenski, S. A., Ferrucci, L., Rosano, C., de Rekeneire,

N.,…Health, Aging, and Body Composition Study (2012). The relationship of vitamin B12

and sensory and motor peripheral nerve function in older adults. Journal of the American

Geriatrics Society, 60(6), 1057–1063.

Leishear, K., Ferrucci, L., Lauretani, F., Boudreau, R. M., Studenski, S. A., Rosano,

C.,…Strotmeyer, E. S. (2012). Vitamin B12 and homocysteine levels and 6-year change in

peripheral nerve function and neurological signs. The Journals of Gerontology, 67(5), 537–

543. doi:10.1093/gerona/glr202

Leritz, E. C., McGlinchey, R. E., Salat, D. H., & Milberg, W. P. (2016). Elevated levels of serum

cholesterol are associated with better performance on tasks of episodic memory. Metabolic

Brain Disease, 31(2), 465–473. doi:10.1007/s11011-016-9797-y

Lezak, M. D., Howieson, D. B., Bigler, E. D., & Tranel, D. (2012). Neuropsychological

Assessment, 5th edition. New York: Oxford University Press.

Li, C., Ford, E. S., Zhao, G., Kahn, H. S., & Mokdad, A. H. (2010). Waist-to-thigh ratio and

diabetes among US adults: The Third National Health and Nutrition Examination Survey.

Diabetes Research and Clinical Practice, 89(1), 79–87. doi:10.1016/j.diabres.2010.02.014

Li, X. L., Aou, S., Oomura, Y., Hori, N., Fukunaga, K., & Hori, T. (2002). Impairment of long-

term potentiation and spatial memory in leptin receptor-deficient rodents. Neuroscience,

113(3), 607–615.

Lichtenberger, E. O., & Kaufman, A. S. (2009). Essentials of WAIS-IV Assessment. (A. S.

Kaufman & N. L. Kaufman, Eds.) (p. 242). Hoboken, NJ: John Wiley & Sons, Inc.

222

Lillevang-Johansen, M., Petersen, I., Christensen, K., Hegedüs, L., & Brix, T. H. (2014). Is

previous hyperthyroidism associated with long-term cognitive dysfunction? A twin study.

Clinical Endocrinology, 80, 290–295. doi:10.1111/cen.12255

Liu, L. L., Li, B. M., Yang, J., & Wang, Y. W. (2008). Does dopaminergic reward system

contribute to explaining comorbidity obesity and ADHD? Medical Hypotheses, 70(6),

1118–1120. doi:10.1016/j.mehy.2007.10.012

Liu, Y., Liu, F., Grundke-Iqbal, I., Iqbal, K., & Gong, C. X. (2011). Deficient brain insulin

signalling pathway in Alzheimer’s disease and diabetes. The Journal of Pathology, 225(1),

54–62. doi:10.1002/path.2912

Llewellyn, D. J., Lang, I. A., Xie, J., Huppert, F. A., Melzer, D., & Langa, K. M. (2008).

Framingham Stroke Risk Profile and poor cognitive function: a population-based study.

BMC Neurology, 8, 12. doi:10.1186/1471-2377-8-12

Lo, A. H. Y., Pachana, N. A., Byrne, G. J., Sachdev, P. S., & Woodman, R. J. (2012).

Relationship between changes in body weight and cognitive function in middle-aged and

older women. International Journal of Geriatric Psychiatry, 27(8), 863–872.

doi:10.1002/gps.2797

Lo, A. H. Y., Woodman, R. J., Pachana, N. A., Byrne, G. J., & Sachdev, P. S. (2014).

Associations between lifestyle and cognitive function over time in women aged 40-79

years. Journal of Alzheimer’s Disease, 39(2), 371–383. doi:10.3233/JAD-130971

Lokken, K. L., Boeka, A. G., Yellumahanthi, K., Wesley, M., & Clements, R. H. (2010).

Cognitive performance of morbidly obese patients seeking bariatric surgery. The American

Surgeon, 76(1), 55–59.

Low, L. F., Anstey, K. J., & Sachdev, P. (2009). Use of medications with anticholinergic

properties and cognitive function in a young-old community sample. International Journal

of Geriatric Psychiatry, 24(6), 578–584. doi: 10.1002/gps.2157

Lu, B., Zhou, J., Waring, M. E., Parker, D. R., & Eaton, C. B. (2010). Abdominal obesity and

peripheral vascular disease in men and women: a comparison of waist-to-thigh ratio and

waist circumference as measures of abdominal obesity. Atherosclerosis, 208(1), 253–257.

doi:10.1016/j.atherosclerosis.2009.06.027

Lu, P. H., Lee, G. J., Tishler, T. A., Meghpara, M., Thompson, P. M., & Bartzokis, G. (2013).

Myelin breakdown mediates age-related slowing in cognitive processing speed in healthy

elderly men. Brain and Cognition, 81(1), 131–138. doi:10.1016/j.bandc.2012.09.006

Lu, Y., An, Y., Yu, H., Che, F., Zhang, X., Rong, H.,…Xiao, R. (2017). Sex-specific nonlinear

associations between serum lipids and different domains of cognitive function in middle to

older age individuals. Metabolic Brain Disease, 32(4), 1089–1097. doi:10.1007/s11011-

017-9999-y

223

Lundqvist, T. (2005). Cognitive consequences of cannabis use: comparison with abuse of

stimulants and heroin with regard to attention, memory and executive functions.

Pharmacology, Biochemistry, and Behavior, 81(2), 319–330.

doi:10.1016/j.pbb.2005.02.017

Lunsman, M., Edwards, J. D., Andel, R., Small, B. J., Ball, K. K., & Roenker, D. L. (2008).

What predicts changes in useful field of view test performance? Psychology and Aging,

23(4), 917–927. doi:10.1037/a0013466

Luo, L., Chen, X., Chai, Y., Li, J., Zhang, M., & Zhang, J. (2013). A distinct pattern of memory

and attention deficiency in patients with depression. Chinese Medical Journal, 126(6),

1144–1149. doi:10.3760/cma.j.issn.0366-6999.20122636

Luppino, F. S., de Wit, L. M., Bouvy, P. F., Stijnen, T., Cuijpers, P., Pennix, B. W., & Zitman, F.

G. (2010). Overweight, obesity, and depression: a systematic review and meta-analysis of

longitudinal studies. Archives of General Psychiatry, 67(3), 220–229.

doi:10.1001/archgenpsychiatry.2010.2

Lu’o'ng, K. V. Q., & Nguyen, L. T. H. (2011). Role of thiamine in Alzheimer’s disease.

American Journal of Alzheimer's Disease & Other Dementias, 26(8), 588–598.

doi:10.1177/1533317511432736

MacDonald, S. W. S., Hultsch, D. F., Strauss, E., & Dixon, R. A. (2003). Age-related slowing of

digit symbol substitution revisited: what do longitudinal age changes reflect? The Journals

of Gerontology, 58(3), P187–P194.

Mai, X. M., Chen, Y., Camargo, C. A., & Langhammer, A. (2012). Cross-sectional and

prospective cohort study of serum 25-hydroxyvitamin D level and obesity in adults: the

HUNT study. American Journal of Epidemiology, 175(10), 1029–1036.

doi:10.1093/aje/kwr456

Malouf, R., & Evans, J. G. (2008). Folic acid with or without vitamin B12 for the prevention and

treatment of healthy elderly and demented people (Review). Cochrane Database of

Systematic Reviews, 4:CD004514. doi:10.1002/14651858.CD004514.pub2

Mamtani, M., Kulkarni, H., Dyer, T. D., Almasy, L., Mahaney, M. C., Duggirala, R.,…Curran, J.

E. (2014). Waist circumference is genetically correlated with incident Type 2 diabetes in

Mexican-American families. Diabetic Medicine, 31(1), 31–35. doi:10.1111/dme.12266

Manard, M., Carabin, D., Jaspar, M., & Collette, F. (2014). Age-related decline in cognitive

control: the role of fluid intelligence and processing speed. BMC Neuroscience, 15(7).

doi:10.1186/1471-2202-15-7

Mandelblatt, J. S., Stern, R. A., Luta, G., McGuckin, M., Clapp, J. D., Hurria, A.,…Ahles, T.

(2014). Cognitive impairment in older patients with breast cancer before systemic therapy:

224

is there an interaction between cancer and comorbidity? Journal of Clinical Oncology,

32(18), 1909–1918. doi:10.1200/JCO.2013.54.2050

Marazziti, D., Consoli, G., Picchetti, M., Carlini, M., & Faravelli, L. (2010). Cognitive

impairment in major depression. European Journal of Pharmacology, 626(1), 83–86.

doi:10.1016/j.ejphar.2009.08.046

Marik, P. E., & Desai, H. (2013). Characteristics of patients with the “malignant obesity

hypoventilation syndrome” admitted to an ICU. Journal of Intensive Care Medicine, 28(2),

124–130. doi:10.1177/0885066612444261

Martin, L., Birdsell, L., Macdonald, N., Reiman, T., Clandinin, M. T., McCargar, L. J.,

…Baracos, V. E.(2013). Cancer cachexia in the age of obesity: skeletal muscle depletion is

a powerful prognostic factor, independent of body mass index. Journal of Clinical

Oncology, 31(12), 1539–1547. doi:10.1200/JCO.2012.45.2722

Mather, A. A., Cox, B. J., Enns, M. W., & Sareen, J. (2009). Associations of obesity with

psychiatric disorders and suicidal behaviors in a nationally representative sample. Journal

of Psychosomatic Research, 66(4), 277–285. doi:10.1016/j.jpsychores.2008.09.008

Matthews, D. R., Hosker, J. P., Rudenski, A. S., Naylor, B. A., Treacher, D. F., & Turner, R. C.

(1985). Homeostasis model assessment: insulin resistance and beta-cell function from

fasting plasma glucose and insulin concentrations in man. Diabetologia. 28(7), 412–419.

doi:10.1007/BF00280883

McCarthy, L. H., Bigal, M. E., Katz, M., Derby, C., & Lipton, R. B. (2009). Chronic pain and

obesity in elderly people: results from the Einstein aging study. Journal of the American

Geriatrics Society, 57(1), 115–119. doi:10.1111/j.1532-5415.2008.02089.x

McCrea, R. L., Berger, Y. G., & King, M. B. (2012). Body mass index and common mental

disorders: exploring the shape of the association and its moderation by age, gender and

education. International Journal of Obesity, 36(3), 414–421. doi:10.1038/ijo.2011.65

Medici, V., & Halsted, C. H. (2013). Folate, alcohol, and liver disease. Molecular Nutrition &

Food Research, 57(4), 596–606. doi:10.1002/mnfr.201200077

Meigs, J. B., Jacques, P. F., Selhub, J., Singer, D. E., Nathan, D. M., Rifai, N.,…Framingham

Offspring Study. (2001). Fasting plasma homocysteine levels in the insulin resistance

syndrome: The Framingham Offspring Study. Diabetes Care, 24(8), 1403–1410.

Melville, C. A., Hamilton, S., Hankey, C. R., Miller, S., & Boyle, S. (2007). The prevalence and

determinants of obesity in adults with intellectual disabilities. Obesity Reviews, 8(3), 223–

230. doi:10.1111/j.1467-789X.2006.00296.x

225

Mendis, S. (2010). The contribution of the Framingham Heart Study to the prevention of

cardiovascular disease: A global perspective. Progress in Cardiovascular Diseases, 53(1),

10–14. doi:10.1016/j.pcad.2010.01.001

Mergenthaler, P., Lindauer, U., Dienel, G. A., & Meisel, A. (2013). Sugar for the brain: The role

of glucose in physiological and pathological brain function. Trends in Neurosciences,

36(10), 587–597. doi:10.1016/j.tins.2013.07.001

Meusel, L. C., Anderson, N. D., Parrott, M. D., Yuen, W., Tchistiakova, E., MacIntosh, B.

J.,…Greenwood, C. E. (2017). Brain function is linked to LDL cholesterol in older adults

with cardiovascular risk. Journal of the American Geriatrics Society. 65(2). e51–e55.

doi:10.1111/jgs.14663

Mheid, I. A., Patel, R., Murrow, J., Morris, A., Rahman, A., Fike, L.,…Quyyumi, A. A. (2011).

Vitamin D status is associated with arterial stiffness and vascular dysfunction in healthy

humans. Journal of the American College of Cardiology, 58(2), 186–192.

doi:10.1016/j.jacc.2011.02.051

Michaud, M., Balardy, L., Moulis, G., Gaudin, C., Peyrot, C., Vellas, B.,…Nourhashemi, F.

(2013). Proinflammatory cytokines, aging, and age-related diseases. Journal of the

American Medical Directors Association. 14(12), 877–882.

doi:10.1016/j.jamda.2013.05.009

Miller, G. A. & Chapman, J. P. (2001). Misunderstanding analysis of covariance. Journal of

Abnormal Psychology, 110(1), 40–48. doi:10.1037/0021-843X.110.1.40

Miller, J. W. (2006). Assessing the association between vitamin B-12 status and cognitive

function in older adults. American Journal of Clinical Nutrition, 84, 1259–1260.

Miller M., Stone N. J., Ballantyne C., Bittner V., Criqui M. H., Ginsberg, H. N.,…Council on the

Kidney in Cardiovascular Disease. (2011). Triglycerides and cardiovascular disease: a

scientific statement from the American Heart Association. Circulation, 123, 2292–2333.

Miyata, S., Noda, A., Iwamoto, K., Kawano, N., Okuda, M., & Ozaki, N. (2013). Poor sleep

quality impairs cognitive performance in older adults. Journal of Sleep Research, 22(5),

535–541. doi:10.1111/jsr.12054

Mond, J. M., Stich, H., Hay, P. J., Kraemer, A., & Baune, B. T. (2007). Associations between

obesity and developmental functioning in pre-school children: A population-based study.

International Journal of Obesity, 31, 1068–1073. doi:10.1038/sj.ijo.0803644

Moon, J. Y., Park, K. J., Hwangbo, Y., Lee, M. R., Yoo, B. I., Won, J. H., & Park, Y. H. (2013).

A trend analysis of the prevalence, awareness, treatment, and control of hypertension by age

group. Journal of Preventive Medicine & Public Health, 46(6), 353–359.

doi:10.3961/jpmph.2013.46.6.353

226

Mooradian, A.D. (1987). Blood-brain barrier choline transport is reduced in diabetic rats.

Diabetes, 36(10), 1094–1097.

Moore, E. M., Ames, D., Mander, A. G., Carne, R. P., Brodaty, H., Woodward, M. C.,…Watters,

D. A. (2014). Among vitamin B12 deficient older people, high folate levels are associated

with worse cognitive function: combined data from three cohorts. Journal of Alzheimer’s

Disease, 39(3), 661–668. doi:10.3233/JAD-131265

Morris, M. C., Evans, D. A., Bienias, J. L., Tangney, C. C., & Wilson, R. S. (2002). Vitamin E

and cognitive decline in older persons. Archives of Neurology, 59, 1125–1132.

Mortensen, E. L., & Hogh, P. (2001). A gender difference in the association between APOE

genotype and age-related cognitive decline. Neurology, 57(1), 89–95.

Mosley, T. H., Knopman, D. S., Catellier, D. J., Bryan, N., Hutchinson, R. G., Grothues, C.

A.,…Szklo, M. (2005). Cerebral MRI findings and cognitive functioning: The

atherosclerosis risk in communities study. Neurology, 64(12), 2056–2062.

Motyl, J. M., Driban, J. B., McAdams, E., Price, L. L., & McAlindon, T. E. (2013). Test-retest

reliability and sensitivity of the 20-meter walk test among patients with knee osteoarthritis.

BMC Musculoskeletal Disorders, 14(1), 166. doi:10.1186/1471-2474-14-166

Moyad, M. A. (2001). Current methods used for defining, measuring, and treating obesity.

Seminars in Urologic Oncology, 19(4), 247–256.

Muniz-Terrera, G., Matthews, F., Dening, T., Huppert, F. A., & Brayne, C. (2009). Education

and trajectories of cognitive decline over 9 years in very old people: methods and risk

analysis. Age and Ageing, 38(3), 277–282. doi:10.1093/ageing/afp004

Murray-Kolb, L. E., & Beard, J. L. (2007). Iron treatment normalizes cognitive functioning in

young women. American Journal of Clinical Nutrition, 85(3), 778–787.

Murugan, A. T., & Sharma, G. (2008). Obesity and respiratory diseases. Chronic Respiratory

Disease, 5(4), 233–242. doi:10.1177/1479972308096978

Must, A., Spadano, J., Coakley, E. H., Field, A. E., Colditz, G., & Dietz, W. H. (1999). The

disease burden associated with overweight and obesity. Journal of the American Medical

Association, 282(16), 1523–1529.

Münzel, T., Sinning, C., Post, F., Warnholtz, A., & Schulz, E. (2008). Pathophysiology,

diagnosis and prognostic implications of endothelial dysfunction. Annals of Medicine,

40(3), 180–196. doi:10.1080/07853890701854702

Nadkarni, N. K., Nunley, K. A., Aizenstein, H., Harris, T. B., Yaffe, K., Satterfield, S.,…Health

ABC Study (2014). Association between cerebellar gray matter volumes, gait speed, and

information-processing ability in older adults enrolled in the Health ABC Study. The

227

Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 69(8), 996–

1003. doi:10.1093/gerona/glt151

Nakahachi, T., Ishii, R., Iwase, M., Canuet, L., Takahashi, H., Kurimoto, R.,…Takeda, M.

(2008). Frontal activity during the digit symbol substitution test determined by multichannel

near-infrared spectroscopy. Neuropsychobiology, 57(4), 151–158. doi:10.1159/000147467

Nasution, I. R., Setiati, S., Trisnohadi, H. B., & Oemardi, M. (2006). Insulin resistance and

metabolic syndrome in elderly women living in nursing homes. Acta Medica Indonesiana,

38(1), 17–22.

National Heart Lung and Blood Institute (2014, August 18). Who is at risk for heart disease?

Retrieved from http://www.nhlbi.nih.gov/health/health-topics/topics/hdw/atrisk.html

Nebes, R. D., Buysse, D. J., Halligan, E. M., Houck, P. R., & Monk, T. H. (2009). Self-reported

sleep quality predicts poor cognitive performance in healthy older adults. The Journals of

Gerontology, 64(2), 180–187. doi:10.1093/geronb/gbn037

Newman, A. B., Lee, J. S., Visser, M., Goodpaster, B. H., Kritchevsky, S. B., Tylavsky, F.

A.,…Harris, T. B. (2005). Weight change and the conservation of lean mass in old age: the

Health, Aging and Body Composition Study. American Journal of Clinical Nutrition, 82(4),

872–878.

Newton, R. A., Klima, D., & Cromwell, R. L. (2006). Gait examination in older adults: The 8-

foot versus 20-foot walk tests. Journal of Geriatric Physical Therapy, 29(3:06), 124–125.

NHANES Anthropometry Procedures Manual (2000). Retrieved from

http://www.cdc.gov/nchs/nhanes/nhanes1999–2000/manuals99_00.htm

NHANES Body Composition Procedures Manual (2000). Retrieved from

http://www.cdc.gov/nchs/nhanes/nhanes1999–2000/manuals99_00.htm

NHANES Dietary Interview Manual,1999–2000. Retrieved from

http://www.cdc.gov/nchs/nhanes/nhanes1999–2000/manuals99_00.htm

NHANES Plan and Operations, 1999-2010. Retrieved from

http://www.cdc.gov/nchs/data/series/sr_01/sr01_056.pdf

Niki, E. (2015). Evidence for beneficial effects of vitamin E. Korean Journal of Internal

Medicine, 30(5), 571–579. doi:10.3904/kjim.2015.30.5.571

Northam, E. A., Anderson, P. J., Jacobs, R., Hughes, M., Warne, G. L., & Werther, G. A. (2001).

Neuropsychological profiles of children with type 1 diabetes 6 years after disease onset.

Diabetes Care, 24(9), 1541–1546.

228

Nouchi, R., Taki, Y., Takeuchi, H., Sekiguchi, A., Hashizume, H., Nozawa, T., Nouchi, H., et al.

(2014). Four weeks of combination exercise training improved executive functions, episodic

memory, and processing speed in healthy elderly people: Evidence from a randomized

controlled trial. Age, 36(2), 787–799. doi:10.1007/s11357-013-9588-x

Núñez, M., Lozano, L., Núñez, E., Segur, J. M., Sastre, S., Maculé, F.,…Suso, S. (2009). Total

knee replacement and health-related quality of life: Factors influencing long-term outcomes.

Arthritis and Rheumatism, 61(8), 1062–1069. doi:10.1002/art.24644

Nuti, R., Martini, G., & Gennari, C. (1995). Age-related changes of whole skeleton and body

composition in healthy men. Calcified Tissue International, 57, 336–339.

Obesity in Asia Collaboration. (2008). Is central obesity a better discriminator of the risk of

hypertension than body mass index in ethnically diverse populations? Journal of

Hypertension, 26(2), 169–177. doi:10.1097/HJH.0b013e3282f16ad3./

Obidi, C. S., Pugeda, J. P., Fan, X., Dimaculangan, C. M., Singh, S. P., Chalisa, N., & Perlmuter,

L. C. (2008). Race moderates age-related cognitive decline in type 2 diabetes. Experimental

Aging Research, 34(2), 114–125. doi:10.1080/03610730701876938

Obisesan, T. O., Aliyu, M. H., Bond, V., Adams, R. G., Akomolafe, A., & Rotimi, C. N. (2005).

Ethnic and age-related fat free mass loss in older Americans: The Third National Health and

Nutrition Examination Survey (NHANES III). BMC Public Health, 5(41).

doi:10.1186/1471-2458-5-41

Ogden, C. L., Carroll, M. D., Kit, B. K., & Flegal, K. M. (2014). Prevalence of childhood and

adult obesity in the United States, 2011–2012. Journal of the American Medical

Association, 311(8), 806–814. doi:10.1001/jama.2014.732

Ohsugi, H., Ohgi, S., Shigemori, K., & Schneider, E. B. (2013). Differences in dual-task

performance and prefrontal cortex activation between younger and older adults. BMC

Neuroscience. 14(1), 10. doi:10.1186/1471-2202-14-10

Okoro, C. A., Zhong, Y., Ford, E. S., Balluz, L. S., Strine, T. W., & Mokdad, A. H. (2006).

Association between the metabolic syndrome and its components and gait speed among

U.S. adults aged 50 years and older: A cross-sectional analysis. BMC Public Health, 6, 282.

doi:10.1186/1471-2458-6-282

O’Luanaigh, C., O’Connell, H., Chin, A. V., Hamilton, F., Coen, R., Walsh, C.,…Lawlor, B. A.

(2012). Loneliness and cognition in older people: The Dublin Healthy Ageing study. Aging

& Mental Health, 16(3), 347–352. doi:10.1080/13607863.2011.62897

Oldroyd, B., Stewart, S. P., Truscott, J. G., Westmacott, C. F., & Smith, M. A. (1998). Age

Related Changes in Body Composition. Applied Radiation and Isotopes, 49(5), 589–590.

229

Ouwens, M. A., van Strien, T., & van Leeuwe, J. F. J. (2009). Possible pathways between

depression, emotional and external eating. A structural equation model. Appetite, 53(2),

245–248. doi:10.1016/j.appet.2009.06.001

Owsley, C., Sloane, M., McGwin, G., & Ball, K. (2002). Timed instrumental activities of daily

living tasks: Relationship to cognitive function and everyday performance assessments in

older adults. Gerontology, 48, 254–265.

Pagoto, S. L., Curtin, C., Lemon, S. C., Bandini, L. G., Schneider, K. L., Bodenlos, J. S., & Ma,

Y. (2009). Association between adult attention deficit/hyperactivity disorder and obesity in

the US population. Obesity, 17(3), 539–544. doi:10.1038/oby.2008.587

Palta, P., Xue, Q., Deal, J. A., Fried, L. P., Walston, J. D., & Carlson, M. C. (2015). Interleukin-6

and C-Reactive Protein Levels and 9-Year Cognitive Decline in Community-Dwelling

Older Women: The Women’s Health and Aging Study II. The Journals of Gerontology.

Series A, Biological Sciences and Medical Sciences. 70(7), 873–878. doi:

10.1093/gerona/glu132

Pantoni, L., & Simoni, M. (2003). Pathophysiology of cerebral small vessels in vascular

cognitive impairment. International Psychogeriatrics, 15(Suppl 1), 59–65.

doi:10.1017/S1041610203008974

Papenberg, G., Backman, L., Fratiglioni, L., Laukka, E. J., Fastbom, J., & Johnell, K. (2017).

Anticholinergic drug use is associated with episodic memory decline in older adults without

dementia. Neurobiology of Aging, 55, 27–32. doi:10.1016/j.neurobiolaging.2017.03.009

Papp, K. V., Kaplan, R. F., Springate, B., Moscufo, N., Wakefield, D. B., Guttmann, C. R. G., &

Wolfson, L. (2014). Processing speed in normal aging: effects of white matter

hyperintensities and hippocampal volume loss. Neuropsychology, Development, and

Cognition. 21(2), 197–213. doi:10.1080/13825585.2013.795513

Pardridge, W.M., Triguero, D., & Farrell, C.R. (1990). Downregulation of blood-brain barrier

glucose transporter in experimental diabetes. Diabetes, 39(9), 1040–1044.

Parthasarathy, V., Frazier, D., Jastrzab, L., Chao, L., Chui, H., Bettcher, B., & Kramer, J. (2014).

The effect of triglycerides on cognitive function in aging adults: A cross-sectional study.

Neurology, 82(10), Supplement S58.004.

Peeters, A., Barendregt, J. J., Willekens, F., Mackenbach, J. P., Mamun, A. A., & Bonneux, L.

(2003). Obesity in adulthood and its consequences for life expectancy: A life-table analysis.

Annals of Internal Medicine, 138, 24–32.

Peeters, R. P. (2008). Thyroid hormones and aging. Hormones, 7(1), 28–35.

Peters, A. (2009). The effects of normal aging on myelinated nerve fibers in monkey central

nervous system. Frontiers in Neuroanatomy, 3(11). doi:10.3389/neuro.05.011.2009

230

Peterson, R. E., Latendresse, S. J., Bartholome, L. T., Warren, C. S., & Raymond, N. C. (2012).

Binge eating disorder mediates links between symptoms of depression, anxiety, and caloric

intake in overweight and obese women. Journal of Obesity, 2012, 407103.

doi:10.1155/2012/407103

Petterson, J. A. (2017). Does high dose vitamin D supplementation enhance cognition? A

randomized trial in healthy adults. Experimental Gerontology. 90, 90–97. doi:

10.1016/j.exger.2017.01.019

Pickering, R. P., Grant, B. F., Chou, S. P., & Compton, W. (2007). Are overweight, obesity, and

extreme obesity associated with psychopathology? Results from the National Epidemiologic

Survey on Alcohol and Related Conditions. Journal of Clinical Psychiatry, 68(7), 998–

1009.

Pieters, G., Maas, Y., & Hulstijn, W. (2004). Differentiation of cognitive and motor aspects in a

digit symbol substitution test in anorexia nervosa patients before and after weight

restoration. Psychopathology, 37(5), 227–232.

Polito, A., Intorre, F., Andriollo-Sanchez, M.., Azzini, E., Raguzzini, A., Meunier, N.,…Maiani,

G. (2005). Estimation of intake and status of vitamin A, vitamin E, and folate in older

European adults, the ZENITH. European Journal of Clinical Nutrition, 59(Suppl. 2), S42–

S47.

Potvin, O., Lorrain, D., Forget, H., Dubé, M., Grenier, S., Préville, M., & Hudon, C. (2012).

Sleep quality and 1-year incident cognitive impairment in community-dwelling older adults.

Sleep, 35(4), 491–499. doi:10.5665/sleep.1732

Poulain, M., Doucet, M., Major, G. C., Drapeau, V., Sériès, F., Boulet, L. P.,…Maltais, F.

(2006). The effect of obesity on chronic respiratory diseases: pathophysiology and

therapeutic strategies. Canadian Medical Association Journal, 174(9), 1293–1299.

doi:10.1503/cmaj.051299

Power, M. C., Tchetgen, J. T., Sparrow, D., & Schwartz, J. (2013). Blood pressure and cognition

factors that may account for their inconsistent association. Epidemiology, 24(6), 886–893.

doi:10.1097/eDe.0b013e3182a7121c

Power, M. L., & Schulkin, J. (2008). Sex differences in fat storage, fat metabolism, and the

health risks from obesity: possible evolutionary origins. British Journal of Nutrition, 99(5),

931–940. doi:10.1017/S0007114507853347

Raber, J. (1998). Detrimental effects of chronic hypothalamic-pituitary-adrenal axis activation:

from obesity to memory deficits. Molecular Neurobiology, 18(1), 1–22.

Radzevičienė, L., & Ostrauskas, R. (2013). Body mass index, waist circumference, waist-hip

ratio, waist-height ratio and risk for type 2 diabetes in women: a case-control study. Public

Health, 127(3), 241–246. doi:10.1016/j.puhe.2012.12.001

231

Rafnsson, S. B., Deary, I. J., Smith, F. B., Whiteman, M. C., Rumley, A., Lowe, G. D. O., &

Fowkes, G. R. (2007). Cognitive decline and markers of inflammation and hemostasis: The

Edinburgh artery study. Journal of the American Geriatrics Society. 55, 700–707.

doi:10.1111/j.1532-5415.2007.01158.x

Rahimi, R. S., & Rockey, D. C. (2014). Hepatic encephalopathy: how to test and treat. Current

Opinion in Gastroenterology, 30(3), 265–271. doi:10.1097/MOG.0000000000000066

Rahman, M., & Berenson, A. B. (2010). Accuracy of current body mass index obesity

classification for white, black, and Hispanic reproductive-age women. Obstetrics &

Gynecology, 115(5), 982–988.

Rajagopalan, P., Toga, A. W., Jack, C. R., Weiner, M. W., & Thompson, P. M. (2013). Fat-

mass-related hormone, plasma leptin, predicts brain volumes in the elderly. Neuroreport,

24(2), 58–62. doi:10.1097/WNR.0b013e32835c5254

Raman, G., Tatsioni, A., Chung, M., Rosenberg, I. H., Lau, J., Lichtenstein, A. H., & Balk, E. M.

(2007). Heterogeneity and lack of good quality studies limit association between folate,

vitamins B-6 and B-12, and cognitive function. Journal of Nutrition, 137, 1789–1794.

Rana, J. S., Boekholdt, S. M., Kastelein, J. J. P. & Shah, P. K. (2012). The role of non-HDL

cholesterol in risk stratification for coronary artery disease. Current Atherosclerosis

Reports, 14, 130. doi:10.1007/s11883-011-0224-x

Rapp, M. A., & Reischies, F. M. (2005). Attention and executive control predict Alzheimer

disease in late life: Results from the Berlin Aging Study (BASE). American Journal of

Geriatric Psychiatry, 13(2), 134–141.

Rasquin, S. M. C., Verhey, F. R. J., Lousberg, R., & Lodder, J. (2005). Cognitive performance

after first ever stroke related to progression of vascular brain damage: a 2 year follow up CT

scan study. Journal of Neurology, Neurosurgery, and Psychiatry, 76(8), 1075–1079.

doi:10.1136/jnnp.2004.055541

Raz, N., Rodrigue, K. M., & Acker, J. D. (2003). Hypertension and the brain: vulnerability of the

prefrontal regions and executive functions. Behavioral Neuroscience, 117(6), 1169–1180.

doi:10.1037/0735-7044.117.6.1169

Reed, B., Villeneuve, S., Mack, W., DeCarli, C., Chui, H. C., & Jagust, W. (2014). Associations

between serum cholesterol levels and cerebral amyloidosis. JAMA Neurology, 71(2), 195–

200. doi:10.1001/jamaneurol.2013.5390

Reiner, Z., Muacevic-Katanec, D., Katanec, D., & Tedeschi-Reiner, E. (2011). Low HDL-

cholesterol―An important risk factor for cardiovascular diseases. Liječnički Vjesnik,

133(3–4), 111–116.

232

Reitz, C., Tang, M., Schupf, N., Manly, J. J., Mayeux, R., & Luchsinger, J. A. (2010).

Association of higher levels of high-density lipoprotein cholesterol in elderly individuals

and lower risk of late-onset Alzheimer Disease. Archives of Neurology, 67(12), 1491–1497.

doi:10.1001/archneurol.2010.297

Reynolds, C. A., Gatz, M., Prince, J. A., Berg, S., & Pedersen, N. L. (2010). Serum lipid levels

and cognitive change in late-life. Journal of the American Geriatrics Society, 58(3), 501–

509. doi:10.1111/j.1532-5415.2010.02739.x

Richman, E. L., & Stampfer, M. J. (2010). Weight loss and mortality in the elderly: separating

cause and effect. Journal of Internal Medicine, 268(2), 103–105. doi:10.1111/j.1365-

2796.2010.02227.x

Rikli, R. E., & Edwards, D. J. (1991). Effects of a three-year exercise program on motor function

and cognitive processing speed in older women. Research Quarterly for Exercise and Sport,

62(1), 61–67.

Ritchie, S. J., Bates, T. C., Der, G., Starr, J. M., & Deary, I. J. (2013). Education is associated

with higher later life IQ scores, but not with faster cognitive processing speed. Psychology

and Aging, 28(2), 515–521. doi:10.1037/a0030820

Roddy, E., Zhang, W., & Doherty, M. (2007). Is gout associated with reduced quality of life? A

case-control study. Rheumatology (Oxford, England), 46(9), 1441–1444.

doi:10.1093/rheumatology/kem150

Rönnlund, M., Nyberg, L., Bäckman, L., & Nilsson, L. G. (2005). Stability, growth, and decline

in adult life span development of declarative memory: cross-sectional and longitudinal data

from a population-based study. Psychology and Aging, 20(1), 3–18. doi:10.1037/0882-

7974.20.1.3

Rosano, C., Chang, Y. F., Kuller, L. H., Guralink, J. M., Studenski, S. A., Aizenstein,

H.,…Newman, A. B.(2013). Long-term survival in adults 65 years and older with white

matter hyperintensity: Association with performance on the Digit Symbol Substitution Test.

Psychosomatic Medicine, 75, 624–631.

Rozas, A. X. P., Juncos-Rabadán, O., & González, M. S. R. (2008). Processing Speed, Inhibitory

Control, and Working Memory: Three Important Factors to Account for Age-Related

Cognitive Decline. International Journal of Aging & Human Development, 66(2), 115–130.

doi:10.2190/AG.66.2.b

Rosemann, T., Grol, R., Herman, K., Wensing, M., & Szecsenyi, J. (2008). Association between

obesity, quality of life, physical activity and health service utilization in primary care

patients with osteoarthritis. International Journal of Behavioral Nutrition & Physical

Activity, 5(4). doi:10.1186/1479-5868-5-4

233

Runge, C. F. (2007). Economic consequences of the obese. Diabetes, 56(11), 2668–2672.

doi:10.2337/db07-0633

Rusanen, M., Ngandu, T., Laatikainen, T., Tuomilehto, J., Soininen, H., & Kivipelto, M. (2013).

Chronic obstructive pulmonary disease and asthma and the risk of mild cognitive

impairment and dementia: a population based CAIDE study. Current Alzheimer Research,

10(5), 549–555.

Rush, B. K., Barch, D. M., & Braver, T. S. (2006). Accounting for cognitive aging: Context

processing, inhibition or processing speed? Neuropsychology, Development, and Cognition.

Section B, Aging, Neuropsychology and Cognition, 13, 588–610.

doi:10.1080/13825580600680703

Ryan, C. M., & Geckle, M. (2000). Why is learning and memory dysfunction in Type 2 diabetes

limited to older adults? Diabetes/Metabolism Research and Reviews, 16(5), 308–315.

Sabia, S., Kivimaki, M., Shipley, M. J., Marmot, M. G., & Singh-Manoux, A. (2009). Body mass

index over the adult life course and cognition in late midlife: the Whitehall II Cohort Study.

American Journal of Clinical Nutrition, 89, 601–607. doi:10.3945/ajcn.2008.26482.1

Salamone, L. M., Fuerst, T., Visser, M., Kern, M., Lang, T., Dockrell, M.,…Lohman, T. G.

(2000). Measurement of fat mass using DEXA: a validation study in elderly adults. Journal

of Applied Physiology, 89, 345–352.

Salkovic-Petrisic, M., & Hoyer, S. (2007). Central insulin resistance as a trigger for sporadic

Alzheimer-like pathology: an experimental approach. Journal of Neural Transmission.

Supplementum, (72), 217–233.

Salthouse, T. A. (1992). What Do Adult Age Differences in the Digit Symbol Substitution Test

Reflect? The Journals of Gerontology, 47(3), 121–128.

Salthouse, T. A. (1996). The processing-speed theory of adult age differences in cognition.

Psychological Review, 103(3), 403–428.

Salthouse, T. A. (2004). What and When of Cognitive Aging. Current Directions in

Psychological Science, 13(4), 140–144. doi:10.1111/j.0963-7214.2004.00293.x

Salthouse, T. A. (2009). When does age-related cognitive decline begin? Neurobiology of Aging,

30(4), 507–514. doi:10.1016/j.neurobiolaging.2008.09.023

Salthouse, T. A. (2012). Are individual differences in rates of aging greater at older ages?

Neurobiology of Aging, 33(10), 2373–2381. doi: 10.1016/j.neurobiolaging.2011.10.018

Samuels, M. H. (2008). Cognitive function in untreated hypothyroidism and hyperthyroidism.

Current Opinion in Endocrinology, Diabetes and Obesity, 15(5), 429–433.

doi:10.1097/MED.0b013e32830eb84c

234

Samuels, M. H. (2014). Thyroid Disease and Cognition. Endocrinology Metabolism Clinics of

North America, 43, 529–543.

Santos, N. C., Costa, P. S., Cunha, P., Cotter, J., Sampaio, A., Zihl, J.,…Sousa, N. (2013). Mood

is a key determinant of cognitive performance in community-dwelling older adults: a cross-

sectional analysis. Age, 35(5), 1983–1993. doi:10.1007/s11357-012-9482-y

Sanz, C. M., Ruidavets, J. B., Bongard, V., Marquié, J. C., Hanaire, H., Ferrières, J., & Andrieu,

S. (2013). Relationship between markers of insulin resistance, markers of adiposity, HbA1c,

and cognitive functions in a middle-aged population-based sample: the MONA LISA study.

Diabetes Care, 36(6), 1512–1521. doi:10.2337/dc12-1017

Sassi, F., Devaux, M., Cecchini, M., & Rusticelli, E. (2009). OECD Health Working Papers No.

45 The Obesity Epidemic: Analysis of Past and Projected Future Trends in Selected OECD

Countries. Labour.

Savioz, A., Charnay, Y., Huguenin, C., Graviou, C., Greggio, B., & Bouras, C. (1997).

Expression of leptin receptor mRNA (long form splice variant) in the human cerebellum.

Neuroreport, 8(14), 3123–3126. doi:10.1097/00001756-199709290-00023

Schaie, K. W. (2009). “When does age-related cognitive decline begin?” Salthouse again reifies

the “cross-sectional fallacy”. Neurobiology of Aging, 30(4), 528–533.

doi:10.1016/j.neurobiolaging.2008.12.012

Schaie, K. W., & Willis, S. L. (2010). The Seattle Longitudinal Study of Adult Cognitive

Development. ISSBD Bulletin, 57(1), 24–29.

Schram, M. T., Euser, S. M., de Craen, A. J. M., Witteman, J. C., Frolich, M., Hofman,

A.,…Westendorp, R. G. (2007). Systemic markers of inflammation and cognitive decline in

old age. Journal of the American Geriatrics Society, 55(5), 708–716. doi:10.1111/j.1532-

5415.2007.01159.x

Scott, S. P. & Murray-Kolb, L. E. (2016). Iron Status Is Associated with Performance on

Executive Functioning Tasks in Nonanemic Young Women. Journal of Nutrition. 146(1),

30–37. doi:10.3945/jn.115.223586

Selhub, J., Bagley, L. C., Miller, J., & Rosenberg, I. H. (2000). B vitamins, homocysteine, and

neurocognitive function in the elderly. American Journal of Clinical Nutrition, 71(2), 614S–

620S.

Sellinger, J. J., Clark, E. A., Shulman, M., Rosenberger, P. H., Heapy, A. A., & Kerns, R. D.

(2010). The moderating effect of obesity on cognitive-behavioral pain treatment outcomes.

Pain Medicine (Malden, Mass.), 11(9), 1381–1390. doi:10.1111/j.1526-4637.2010.00935.x

235

Shadlen, M. F., Larson, E. B., Wang, L., Phelan, E. A., McCormick, W. C., Jolley, L.,…van

Belle, G. (2005). Education modifies the effect of apolipoprotein epsilon 4 on cognitive

decline. Neurobiology of Aging, 26(1), 17–24. doi:10.1016/j.neurobiolaging.2004.03.005

Sharp, J. T., Henry, J. P., Sweany, S. K., Meadows, W. R., & Pietras, R. J. (1964). Noninvasive

assessment of the tension-time index of inspiratory muscles at rest in obese male subjects.

Journal of Clinical Investigation, 43(4), 728–739.

Sharp, S. I., Aarsland, D., Day, S., Sønnesyn, H., & Ballard, C. (2011). Hypertension is a

potential risk factor for vascular dementia: systematic review. International Journal of

Geriatric Psychiatry, 26(7), 661–669. doi:10.1002/gps.2572

Shen, X., Liu, H., Hu, Z., Hu, H., & Shi, P. (2012). The relationship between cerebral glucose

metabolism and age: report of a large brain PET data set. PLoS One, 7(12), e51514.

doi:10.1371/journal.pone.0051517

Shonesy, B. C., Thiruchelvam, K., Parameshwaran, K., Rahman, E. A., Karuppagounder, S. S.,

Huggins, K. W.,….Suppiramaniam, V. (2012). Central insulin resistance and synaptic

dysfunction in intracerebroventricular-streptozotocin injected rodents. Neurobiology of

Aging, 33(2), 430.e5–18. doi:10.1016/j.neurobiolaging.2010.12.002

Silber, J. H., Rosenbaum, P. R., Kelz, R. R., Reinke, C. E., Neuman, M. D., Ross, R.

N.,…Fleisher, L. A. (2012). Medical and financial risks associated with surgery in the

elderly obese. Annals of Surgery, 256(1), 79–86. doi:10.1097/SLA.0b013e31825375ef

Singh-Manoux, A., Akbaraly, T. N., Marmot, M., Melchior, M., Ankri, J., & Sabia, S. (2010).

Persistent depressive symptoms and cognitive function in late midlife: the Whitehall II

study. The Journal of Clinical Psychiatry, 71(10), 1379–1385.

doi:10.4088/JCP.09m05349gry

Singh-Manoux, A., Gimeno, D., Kivimaki, M., Brunner, E., & Marmot, M. G. (2008). Low HDL

cholesterol is a risk factor for deficit and decline in memory in midlife: the Whitehall II

study. Arteriosclerosis, Thrombosis, and Vascular Biology, 28(8), 1556–1562.

doi:10.1161/ATVBAHA.108.163998

Sliwinski, M. & Buschke, H. (1999). Cross-sectional and longitudinal relationships among age,

cognition, and processing speed. Psychology and Aging, 14(1). 18–33.

Small, G. W., La Rue, A., Komo, S., Kaplan, A., & Mandelkern, M. A. (1995). Predictors of

cognitive change in middle-aged and older adults with memory loss. American Journal of

Psychiatry, 152(12), 1757–1764.

Smit, R. A. J., Trompet, S., Sabayan, B., le Cessie, S., van der Grond, J., van Buchem, M.

A.,...Jukema, J. W. (2016). Higher visit-to-visit low-density lipoprotein cholesterol

variability is associated with lower cognitive performance. Circulation. 134(3), 212–221.

doi:10.1161/CIRCULATIONAHA.115.020627

236

Smith, A. D., & Refsum, H. (2009). Vitamin B-12 and cognition in the elderly. American

Journal of Clinical Nutrition, 89, 707–711. doi:10.3945/ajcn.2008.26947D.Am

Smith, E., Bailey, P. E., Crawford, J., Samaras, K., Baune, B. T., Campbell, L.,…Trollor, J. N.

(2014). Adiposity estimated using dual energy x-ray absorptiometry and BMI and its

association with cognition in elderly adults. Journal of the American Geriatrics Society,

62(12), 2311–2318. doi:10.1111/jgs.13157

Smith, E., Hay, P., Campbell, L., & Trollor, J. N. (2011). A review of the association between

obesity and cognitive function across the lifespan : implications. Obesity Reviews, 12, 740–

755. doi:10.1111/j.1467-789X.2011.00920.x

Smits, J. A. J., Rosenfield, D., Mather, A. A., Tart, C. D., Henriksen, C., & Sareen, J. (2010).

Psychotropic medication use mediates the relationship between mood and anxiety disorders

and obesity: findings from a nationally representative sample. Journal of Psychiatric

Research, 44(15), 1010–1016. doi:10.1016/j.jpsychires.2010.04.007

Snijder, M. B., Visser, M., Dekker, J. M., Seidell, J. C., Fuerst, T., Tylavsky, F.,...Harris, T. B.

(2002). The prediction of visceral fat by dual-energy X-ray absorptiometry in the elderly: a

comparison with computed tomography and anthropometry. International Journal of

Obesity and Related Metabolic Disorders, 26(7), 984–993. doi:10.1038/sj.ijo.0801968

Snijder, M. B., Dekker, J. M., Visser, M., Bouter, L. M., Stehouwer, C. D. A., Kostense, P.

J.,…Seidell, J. C. (2003). Associations of hip and thigh circumferences independent of

waist circumference with the incidence of type 2 diabetes: the Hoorn Study. American

Journal of Clinical Nutrition, 77(5), 1192–1197.

Stanek, K. M., Strain, G., Devlin, M., Cohen, R., Paul, R., Crosby, R. D.,…Gunstad, J. (2013).

Body mass index and neurocognitive functioning across the adult lifespan.

Neuropsychology, 27(2), 141–151. doi:10.1037/a0031988

Stebbins, G. T., Nyenhuis, D. L., Wang, C., Cox, J. L., Freels, S., Bangen, K.,…Gorelick, P. B.

(2008). Gray matter atrophy in patients with ischemic stroke with cognitive impairment.

Stroke, 39(3), 785–793. doi:10.1161/STROKEAHA.107.507392

Steen, E., Terry, B. M., Rivera, E. J., Cannon, J. L., Neely, T. R., Tavares, R.,…de la Monte, S.

M. (2005). Impaired insulin and insulin-like growth factor expression and signaling

mechanisms in Alzheimer’s disease--is this type 3 diabetes? Journal of Alzheimer’s

Disease, 7(1), 63–80.

Stephens, R., & Kaufman, A. (2009). The role of long-term memory in digit-symbol test

performance in young and older adults. Neuropsychology, Development, and Cognition.,

16(2), 219–240. doi:10.1080/13825580802573060

237

Stevens, V. L., Jacobs, E. J., Sun, J., Patel, A. V., McCullough, M. L., Teras, L. R., & Gapstur,

S. M. (2012). Weight cycling and mortality in a large prospective US study. American

Journal of Epidemiology, 175(8), 785–792. doi:10.1093/aje/kwr378

Stewart, R., Masaki, K., Xue, Q. L., Peila, R., Petrovitch, H., White, L. R., & Launer, L. J.

(2005). A 32-Year Prospective Study of Change in Body Weight and Incident Dementia.

Archives of Neurology, 62, 55–60.

Stewart, R., White, L. R., Xue, Q. L., & Launer, L. J. (2007). Twenty-six-year change in total

cholesterol levels and incident dementia: the Honolulu-Asia Aging Study. Archives of

Neurology, 64(1), 103–107. doi:10.1001/archneur.64.1.103

Strain, G. W., Faulconbridge, L., Crosby, R. D., Kolotkin, R. L., Heacock, L., Gagner,

M.,…Pomp, A. (2010). Health-related quality of life does not vary among patients seeking

different surgical procedures to assist with weight loss. Surgery for Obesity and Related

Diseases, 6, 521–525.

Strandberg, T. E., Pitkala, K., Eerola, J., Tilvis, R., & Tienari, P. J. (2005). Interaction of

herpesviridae, APOE gene, and education in cognitive impairment. Neurobiology of Aging,

26(7), 1001–1004. doi:10.1016/j.neurobiolaging.2004.09.008

Strassburger, T., Lee, H., Daly, E., Szczepanik, J., Krasuski, J., Mentis, M.,…Alexander, G. E.

(1997). Interactive effects of age and hypertension on volumes of brain structures. Stroke,

28(7), 1410–1417.

Strazzullo, P., D’Elia, L., Cairella, G., Garbagnati, F., Cappuccio, F. P., & Scalfi, L. (2010).

Excess body weight and incidence of stroke: meta-analysis of prospective studies with 2

million participants. Stroke, 41(5), e418–e426. doi:10.1161/STROKEAHA.109.576967

Suhara, T., Fukuda, H., Inoue, O., Itoh, T., Suzuki, K., Yamasaki, T., & Tateno, Y. (1991). Age-

related changes in human D1 dopamine receptors measured by positron emission

tomography. Psychopharmacology, 103, 41–45.

Sulander, T., Rahkonen, O., Nissinen, A., & Uutela, A. (2007). Association of smoking status

with obesity and diabetes among elderly people. Archives of Gerontology and Geriatrics,

45(2), 159–167. doi:10.1016/j.archger.2006.10.007

Sweat, V., Starr, V., Bruehl, H., Arentoft, A., Tirsi, A., Javier, E., & Convit, A. (2008). C-

reactive protein is linked to lower cognitive performance in overweight and obese women.

Inflammation. 31(3), 198–207. doi:10.1007/s10753-008-9065-3

Tallroth, K., Kettunen, J. A., & Kujala, U. M. (2013). Reproducibility of regional DEXA

examinations of abdominal fat and lean tissue. Obesity Facts, 6(2), 203–210.

doi:10.1159/000348238

238

Tam, N. D. (2013). Improvement of Processing Speed in Executive Function Immediately

following an Increase in Cardiovascular Activity. Cardiovascular Psychiatry and

Neurology, 2013, 212767. doi:10.1155/2013/212767

Tangney, C. C., Tang, Y., Evans, D. A., & Morris, M. C. (2009). Biochemical indicators of

vitamin B12 and folate insufficiency and cognitive decline. Neurology. 72(4), 361–367.

doi:10.1212/01.wnl.0000341272.48617.b0

Tegeler, C., O’Sullivan, J. L., Bucholtz, N., Goldeck, D., Pawelec, G., Steinhagen-Thiessen, E.,

& Demuth, I. (2015). The inflammatory markers CRP, IL-6. and IL-10 are associated with

cognitive function—data from the Berlin Aging Study II. Neurobiology of Aging. 38, 112–

117. doi:10.1016/j.neurobiolaging.2015.10.039

The Psychological Corporation. (1997). WAIS-III/WMS-III Technical Manual. San Antonio, TX:

Harcourt Brace & Company.

Thivat, E., Thérondel, S., Lapirot, O., Abrial, C., Gimbergues, P., Gadéa, E.,…Durando, X.

(2010). Weight change during chemotherapy changes the prognosis in non metastatic breast

cancer for the worse. BMC Cancer, 10(1), 648. doi:10.1186/1471-2407-10-648

Thomas, V. S., & Rockwood, K. J. (2001). Alcohol abuse, cognitive impairment, and mortality

among older people. Journal of the American Geriatrics Society, 49(4), 415–420.

Tincello, D., Sculpher, M., Tunn, R., Quail, D., van der Vaart, H., Falconer, C.,...Timlin, L.

(2010). Patient characteristics impacting health state index scores, measured by the EQ-5D

of females with stress urinary incontinence symptoms. Value in health: The Journal of the

International Society for Pharmacoeconomics and Outcomes Research, 13(1), 112–118.

doi:10.1111/j.1524-4733.2009.00599.x

Tisdale, M. J. (2003). Pathogenesis of cancer cachexia. The Journal of Supportive Oncology,

1(3), 159–168.

Tohill, B. C., Seymour, J., Serdula, M., Kettel-Khan, L., & Rolls, B. J. (2004). What

epidemiologic studies tell us about the relationship between fruit and vegetable

consumption and body weight. Nutrition Reviews, 62(10), 365–374.

doi:10.1301/nr.2004.oct.365

Trakas, K., Lawrence, K., & Shear, N. H. (1999). Utilization of health care resources by obese

Canadians. Canadian Medical Association Journal, 160(10), 1457–1462.

Troester, N., Palfner, M., Kovacs, G., & Olschewski, H. (2013). A 57-year-old woman with

obesity, respiratory insuffiency, and slowed mental state. CHEST, 144(1), 347–348.

Trollor, J. N., Smith, E., Agars, E., Kuan, S. A., Baune, B. T., Campbell, L.,…Sachdev, P.

(2012). The association between systemic inflammation and cognitive performance in the

239

elderly: the Sydney Memory and Ageing Study. Age (Dordrecht, Netherlands). 34(5),

1295–1308. doi:10.1007/s11357-011-9301-x

Troyer, A. K., Leach, L., & Strauss, E. (2006). Aging and response inhibition: Normative data

for the Victoria Stroop Test. Neuropsychology, Development, and Cognition. Section B,

Aging, Neuropsychology and Cognition, 13(1), 20–35. doi:10.1080/138255890968187

Trujillo, M. E., & Scherer, P. E. (2005). Adiponectin—journey from an adipocyte secretory

protein to biomarker of the metabolic syndrome. Journal of Internal Medicine, 257(2), 167–

175. doi:10.1111/j.1365-2796.2004.01426.x

Tsatsanis, C., Zacharioudaki, V., Androulidaki, A., Dermitzaki, E., Charalampopoulos, I., Minas,

V.,…Margioris, A. N. (2006). Peripheral factors in the metabolic syndrome: the pivotal role

of adiponectin. Annals of the New York Academy of Sciences, 1083, 185–195.

doi:10.1196/annals.1367.013

Tschanz, J. T., Pfister, R., Wanzek, J., Corcoran, C., Smith, K., Tschanz, B. T.,…Norton, M. C.

(2013). Stressful life events and cognitive decline in late life: moderation by education and

age. The Cache County Study. International Journal of Geriatric Psychiatry, 28(8), 821–

830. doi:10.1002/gps.3888

Umapathy, S., Dhiman, R. K., Grover, S., Duseja, A., & Chawla, Y. K. (2014). Persistance of

cognitive impairment after resolution of overt hepatic encephalopathy. American Journal of

Gastroenterology, 109, 1101–1119. doi:10.1038/ajg.2014.107

United States Census Bureau (2014, August 18). The Older Population in the United States:

2012. Retrieved from http://www.census.gov/population/age/data/2012.html

United States Census Bureau (2017, May 10). Facts for Features: Older Americans Month: May

2017. Retrieved from https://www.census.gov/newsroom/facts-for-features/2017/cb17-

ff08.html

United States Department of Health & Human Services (2014a, August 18). Projected Future

Growth of the Older Population. Retrieved from

http://www.aoa.gov/Aging_Statistics/future_growth/future_growth.aspx

United States Department of Health & Human Services (2014b, August 18). Projected Future

Growth of the Older Population. Retrieved from

http://www.aoa.gov/Aging_Statistics/future_growth/future_growth.aspx

United States Department of Health & Human Services (2014c, August 18). Insulin resistance

and prediabetes. Retrieved from

http://diabetes.niddk.nih.gov/dm/pubs/insulinresistance/#causes

240

United States Department of Health & Human Services (2017, May 7). Prediabetes and insulin

resistance. Retrieved from https://www.niddk.nih.gov/health-

information/diabetes/overview/what-is-diabetes/prediabetes-insulin-resistance#causes

van de Rest, O., van der Zwaluw, N. L., Tieland, M., Adam, J. J., Hiddink, G. J., van Loon, L. J.

C., & de Groot, L. C. P. G. M. (2014). Effect of resistance-type exercise training with or

without protein supplementation on cognitive functioning in frail and pre-frail elderly:

secondary analysis of a randomized, double-blind, placebo-controlled trial. Mechanisms of

Ageing and Development, 136-7, 85–93. doi:10.1016/j.mad.2013.12.005

van den Berg, E., Dekker, J. M., Nijpels, G., Kessels, R. P., Kappelle, L. J., de Haan, E.

H.,...Biessels, G. J. (2008). Cognitive functioning in elderly persons with type 2 diabetes

and metabolic syndrome: the Hoorn study. Dementia and Geriatric Cognitive Disorders,

26(3), 261–269. doi:10.1159/000160959

van den Berg, E., Kloppenborg, R. P., Kessels, R. P. C., Kappelle, L. J., & Biessels, G. J. (2009).

Type 2 diabetes mellitus, hypertension, dyslipidemia and obesity: A systematic comparison

of their impact on cognition. Biochimica et Biophysica Acta, 1792(5), 470–481.

doi:10.1016/j.bbadis.2008.09.004

van den Kommer, T. N., Dik, M. G., Comijs, H. C., Jonker, C., & Deeg, D. J. (2012). Role of

lipoproteins and inflammation in cognitive decline: do they interact? Neurobiology of

Aging, 33(1), 196.e1–12. doi:10.1016/j.neurobiolaging.2010.05.024

van der Heide, L. P., Ramakers, G. M. J., & Smidt, M. P. (2006). Insulin signaling in the central

nervous system: learning to survive. Progress in Neurobiology, 79(4), 205–221.

doi:10.1016/j.pneurobio.2006.06.003

Van Der Werf, Y. D., Tisserand, D. J., Visser, P. J., Hofman, P. A., Vuurman, E., Uylings, H. B.,

& Jolles, J. (2001). Thalamic volume predicts performance on tests of cognitive speed and

decreases in healthy aging. A magnetic resonance imaging-based volumetric analysis. Brain

Research. Cognitive Brain Research, 11(3), 377–385.

van Deventer, H. E., Miller, W. G., Myers, G. L., Sakurabayashi, I., Bachmann, L. M., Caudill,

S. P.,…Remaley, A. T. (2011). Non–HDL cholesterol shows improved accuracy for

cardiovascular risk score classification compared to direct or calculated LDL cholesterol in

a dyslipidemic population. Clinical Chemistry, 57(3), 490–501.

doi:10.1373/clinchem.2010.154773

van Dis, I., Kromhout, D., Geleijnse, J. M., Boer, J. M. A, & Verschuren, W. M. M. (2009).

Body mass index and waist circumference predict both 10-year nonfatal and fatal

cardiovascular disease risk: study conducted in 20,000 Dutch men and women aged 20–65

years. European Journal of Cardiovascular Prevention and Rehabilitation, 16(6), 729–734.

doi:10.1097/HJR.0b013e328331dfc0

241

van Reedt Dortland, A. K. B., Giltay, E. J., van Veen, T., Zitman, F. G., & Penninx, B. W. J. H.

(2013). Longitudinal relationship of depressive and anxiety symptoms with dyslipidemia

and abdominal obesity. Psychosomatic Medicine, 75(1), 83–89.

doi:10.1097/PSY.0b013e318274d30f

Vance, D. E. (2009). Speed of processing in older adults: a cognitive overview for nursing. The

Journal of Neuroscience Nurse, 41(6), 290–297.

Vanherwegen, A., Gysemans, C., & Mathieu, C. (2017). Vitamin D endocrinology on the cross-

road between immunity and metabolism. Molecular and Cellular Endocrinology. doi:

10.1016/j.mce.2017.04.018 [Epub ahead of print].

Velloso, L. A., Araujo, E. P., & de Souza, C. T. (2008). Diet-induced inflammation of the

hypothalamus in obesity. Neuroimmunomodulation, 15(3), 189–193.

doi:10.1159/000153423

Verma, A., Jayaraman, M., Kumar, H. K. V. S., & Modi, K. D. (2008). Hypothyroidism and

obesity. Saudi Medical Journal, 29(8), 1135–1138.

Vetrani, C., Costabile, G., Di Marino, L., & Rivellese, A. A. (2013). Nutrition and oxidative

stress: a systematic review of human studies. International Journal of Food Sciences and

Nutrition, 64(3), 312–326. doi:10.3109/09637486.2012.738651

Vidal, J. S., Dufouil, C., Ducros, V., & Tzourio, C. (2008). Homocysteine, folate and cognition

in a large community-based sample of elderly people—the 3C Dijon Study.

Neuroepidemiology, 30(4), 207–214. doi:10.1159/000126914

Vijayakumar, T. M., Sirisha, G. B. N., Begam, F., & Dhanaraju, M. D. (2012). Mechanism

Linking Cognitive Impairment and Diabetes mellitus. European Journal of Applied

Sciences, 4(1), 1–5.

Vinkers, D., Stek, M., van der Mast, R., de Craen, A., Le Cessie, S., Jolles, J.,...Gussekloo, J.

(2005). Generalized atherosclerosis, cognitive decline, and depressive symptoms in old age.

Neurology, 65(1), 107–112.

Virani, S. S. (2011). Non-HDL cholesterol as a metric of good quality of care: opportunities and

challenges. Texas Heart Institute Journal, 38(2), 160–162.

Visscher, T. L. S., Seidell, J. C., Molarius, A., Kuip, D. V. D., Hofman, A., & Witteman, J. C.

M. (2001). A comparison of body mass index , waist – hip ratio and waist circumference as

predictors of all-cause mortality among the elderly: the Rotterdam study. International

Journal of Obesity, 25, 1730–1735.

Visser, M., Bouter, L. M., McQuillan, G. M., Wener, M. H., & Harris, T. B. (1999). Elevated C-

Reactive Protein Levels in Overweight and Obese Adults. Journal of the American Medical

Association, 282, 2131–2135.

242

Vogel, T., Dali-Youcef, N., Kaltenbach, G., & Andrès, E. (2009). Homocysteine, vitamin B12,

folate and cognitive functions: a systematic and critical review of the literature.

International Journal of Clinical Practice, 63(7), 1061–1067. doi:10.1111/j.1742-

1241.2009.02026.x

Volkow, N. D., Ding, Y. S., Fowler, J. S., Wang, G .J., Logan, J., Gatley, S. J.,…Gur, R. (1996).

Dopamine transporters decrease with age. Journal of Nuclear Medicine, 37(4), 554–559.

Volkow, N. D., Logan, J., Fowler, J. S., Wang, G. J., Gur, R. C., Wong, C.,…Pappas, N. (2000).

Association between age-related decline in brain dopamine activity and impairment in

frontal and cingulate metabolism. American Journal of Psychiatry, 157, 75–80.

Volkow, N. D., Wang, G. J., Elang, F., Fowler, J. S., Goldstein, R. Z., Alia-Klein, N.,…Pradhan,

K. (2009). Inverse association between BMI and prefrontal metabolic activity in healthy

adults. Obesity, 17(1), 60–65. doi:10.1038/oby.2008.469

Voss, M. W., Heo, S., Prakash, R. S., Erickson, K. I., Alves, H., Chaddock, L.,… (2013). The

influence of aerobic fitness on cerebral white matter integrity and cognitive function in

older adults: results of a one-year exercise intervention. Human Brain Mapping, 34(11),

2972–2985. doi:10.1002/hbm.22119

Wakefield DB, Moscufo N, Guttmann CR, Kuchel, GA, Kaplan, RF, Pearlson, G, & Wolfson, L

(2010). White matter hyperintensities predict functional decline in voiding, mobility, and

cognition in older adults. Journal of the American Geriatrics Society. 58(2), 275–281.

doi:10.1111/j.1532-5415.2009.02699.x

Walther, K., Birdsill, A. C., Glisky, E. L., & Ryan, L. (2010). Structural brain differences and

cognitive functioning related to body mass index in older females. Human Brain Mapping,

31(7), 1052–1064. doi:10.1002/hbm.20916

Wang, G. J., Volkow, N. D., Logan, J., Pappas, N. R., Wong, C. T., Zhu, W.,...Fowler, J. S.

(2001). Brain dopamine and obesity. Lancet, 357(9253), 354–357.

Wang, Y., Zhang, S. X. L., & Gozal, D. (2010). Reactive oxygen species and the brain in sleep

apnea. Respiratory Physiology & Neurobiology, 174(3), 307–316.

doi:10.1016/j.resp.2010.09.001

Wannamethee, S. G., Shaper, A. G., Whincup, P. H., & Walker, M. (2000). Characteristics of

older men who lose weight intentionally or unintentionally. American Journal of

Epidemiology, 151(7), 667–675.

Ward, M. A., Bendlin, B. B., McLaren, D. G., Hess, T. M., Gallagher, C. L., Kastman, E.

K.,…Johnson, S. C. (2010). Low HDL cholesterol is associated with lower gray matter

volume in cognitively healthy adults. Frontiers in Aging Neuroscience, 2(July), 1–8.

doi:10.3389/fnagi.2010.00029

243

Wardlaw, J. M., Valdés Hernández, M. C., & Muñoz-Maniega, S. (2015). What are White

Matter Hyperintensities Made of? Relevance to Vascular Cognitive Impairment. Journal of

the American Heart Association, 4(6), e001140. doi:10.1161/JAHA.115.002006

Whaley-Connell, A., Pavey, B. S., Afroze, A., & Bakris, G. L. (2006). Obesity and insulin

resistance as risk factors for chronic kidney disease. Journal of the Cardiometabolic

Syndrome, 1(3), 209–214.

Willeumier, K. C., Taylor, D. V., & Amen, D. G. (2011). Elevated BMI is associated with

decreased blood flow in the prefrontal cortex using SPECT imaging in healthy adults.

Obesity, 19(5), 1095–1097. doi:10.1038/oby.2011.16

Williamson, J., Nyenhuis, D., Stebbins, G. T., Lamb, D., Simkus, V., Sripathirathan,

K.,…Gorelick, P. (2010). Regional differences in relationships between apparent white

matter integrity, cognition and mood in patients with ischemic stroke. Journal of Clinical

and Experimental Neuropsychology, 32(7), 673–681. doi:10.1080/13803390903427406

Wilson, A. C., Samuelson, B., & Palermo, T. M. (2010). Obesity in children and adolescents

with chronic pain: Associations with pain and activity limitations. Clinical Journal of Pain,

26(8), 705–711.

Wilson, L. R., Tripkovic, L., Hart, K. H., & Lanham-New, S. A. (2017). Vitamin D deficiency as

a public health issue: using vitamin D2 or vitamin D3 in future fortification strategies.

Proceedings of the Nutrition Society, 1–8. doi: 10.1017/S0029665117000349 [Epub ahead

of print].

Wilson, M. M. G., Habib, S., & Philpot, C. (2002). Cigarette smoking and weight loss in nursing

home residents. Medicine Online : The International Journal of Medicine. Retrieved from

http://www.priory.com/med/cigsmoking.htm

Wilson, P. W., D’Agostino, R. B., Levy, D., Belanger, A. M., Silbershatz, H., & Kannel, W. B.

(1998). Prediction of coronary heart disease using risk factor categories. Circulation,

97(18), 1837–1847.

Wilson, P. W. F., Meigs, J. B., Sullivan, L., Fox, C. S., Nathan, D. M., & D’Agostino, R. B.

(2007). Prediction of incident diabetes mellitus in middle-aged adults: The Framingham

offspring study. Archives of Internal Medicine, 167(10), 1068–1074.

doi:10.1001/archinte.167.10.1068

Winocur, G., & Greenwood, C. E. (2005). Studies of the effects of high fat diets on cognitive

function in a rat model. Neurobiology of Aging, 26(Suppl 1), 46–49.

doi:10.1016/j.neurobiolaging.2005.09.003

Wolf, P. A., Beiser, M. F., Elias, R. A., Vasan, R. S., & Seshadri, S. (2007). Relation of obesity

to cognitive function: importance of central obesity and synergistic influence of

concomitant hypertension. Current Alzheimer Research, 4, 111–116.

244

Wolf, P. A., D’Agostino, R. B., Belanger, A. J., & Kannel, W. B. (1991). Probability of stroke: a

risk profile from the Framingham Study. Stroke, 22(3), 312–318.

doi:10.1161/01.STR.22.3.312

Wolinsky, F. D., Unverzagt, F. W., Smith, D. M., Jones, R., Wright, E., & Tennstedt, S. L.

(2006). The effects of the ACTIVE cognitive training trial on clinically relevant declines in

health-related quality of life. The Journals of Gerontology. Series B, Psychological Sciences

and Social Sciences, 61B(5), S281–S287.

Woo, J., Ho, S. C., Yu, A. L. M., & Sham, A. (2002). Is waist circumference a useful measure in

predicting health outcomes in the elderly? International Journal of Obesity and Related

Metabolic Disorders. 26(10), 1349–1355. doi:10.1038/sj.ijo.0802080

Woods, S. C., Seeley, R. J., Baskin, D. G., & Schwartz, M. W. (2003). Insulin and the blood-

brain barrier. Current Pharmaceutical Design, 9(10), 795–800.

doi:10.2174/1381612033455323

Woodward, N. D., Duffy, B., & Karbasforoushan, H. (2013). Prefrontal cortex activity during

response selection predicts processing speed impairment in schizophrenia. Journal of the

International Neuropsychological Society, 19(7), 782–791.

doi:10.1017/S1355617713000532

World Health Organization (2003). The world health report 2003: Shaping the future. Retrieved

June 8, 2017 from http://www.who.int/whr/2003/en/whr03_en.pdf

World Health Organization (2015). Serum and red blood cell folate concentrations for assessing

folate status in populations. Retrieved June 8, 2017 from

http://apps.who.int/iris/handle/10665/162114

Wurtman, R. J., & Wurtman, J. J. (1995). Brain serotonin, carbohydrate-craving, obesity and

depression. Obesity Research, 3(Suppl 4), 477s–80s.

Xanthakos, S. A. (2009). Nutritional deficiencies in obesity and after bariatric surgery. Pediatric

Clinics of North America, 56(5), 1105–1121. doi:10.1016/j.pcl.2009.07.002

Xu, B., Houston, D. K., Gropper, S. S., & Zizza, C. A. (2009). Race/Ethnicity differences in the

relationship between obesity and gait speed among older Americans. Journal of Nutrition

for the Elderly, 28(4), 372–385. doi:10.1080/01639360903393515

Xu, W. L., Atti, A. R., Gatz, M., Pedersen, N. L., Johansson, B., & Fratiglioni, L. (2011). Midlife

overweight and obesity increase late-life dementia risk: a population-based twin study.

Neurology, 76(18), 1568–1574. doi:10.1212/WNL.0b013e3182190d09

Yaffe, K., Lindquist, K., Penninx, B. W., Simonsick, E. M., Pahor, M., Kritchevsky, S.,…Harris,

T. (2003). Inflammatory markers and cognition in well-functioning African-American and

white elders. Neurology, 61(1), 76–80.

245

Yang, Z., & Hall, A. G. (2008). The financial burden of overweight and obesity among elderly

Americans: the dynamics of weight, longevity, and health care cost. Health Services

Research, 43(3), 849–868. doi:10.1111/j.1475-6773.2007.00801.x

Yavuz, B. B., Cankurtaran, M., Haznedaroglu, I. C., Halil, M., Ulger, Z., Altun, B., & Ariogul,

S. (2012). Iron deficiency can cause cognitive impairment in geriatric patients. Journal of

Nutrition, Health and Aging, 16(3), 220–224.

Ye, J., & Keller, J. N. (2010). Regulation of energy metabolism by inflammation: A feedback

response in obesity and calorie restriction. Aging, 2(6), 361–368.

Yeung, S. E., Fischer, A. L., & Dixon, R. A. (2009). Exploring effects of type 2 diabetes on

cognitive functioning in older adults. Neuropsychology, 23(1), 1–9. doi:10.1037/a0013849

Yildiz, N., Topuz, O., Gungen, G. O., Deniz, S., Alkan, H., & Ardic, F. (2010). Health-related

quality of life (Nottingham Health Profile) in knee osteoarthritis: correlation with clinical

variables and self-reported disability. Rheumatology International, 30(12), 1595–1600.

doi:10.1007/s00296-009-1195-x

Yin, F., Boveris, A., & Cadenas, E. (2014). Mitochondrial energy metabolism and redox

signaling in brain aging and neurodegeneration. Antioxidants & Redox Signaling, 20(2),

353–371. doi:10.1089/ars.2012.4774

Yoon, D. H., Choi, S. H., Yu, J. H., Ha, J. H., Ryu, S. H., & Park, D. H. (2012). The relationship

between visceral adiposity and cognitive performance in older adults. Age and Ageing,

41(4), 456–461. doi:10.1093/ageing/afs018

Yoshinaga, T, Nishimata, H, Kajiya, Y, & Yokoyama, S. (2017). Combined assessment of serum

folate and hemoglobin as biomarkers of brain amyloid β accumulation. PLoS One, 12(4),

e0175854. doi:10.1371/journal.pone.0175854

Young, S. E., Mainous, A. G., & Carnemolla, M. (2006). Hyperinsulinemia and cognitive

decline in a middle-aged cohort. Diabetes Care, 29(12), 2688–2693. doi:10.2337/dc06-0915

Yusuf, S., Hawken, S., Ounpuu, S., Dans, T., Avezum, A., Lanas, F.,…INTERHEART Study

Investigators. (2004). Effect of potentially modifiable risk factors associated with

myocardial infarction in 52 countries (the INTERHEART study): case-control study.

Lancet, 364, 937–952. doi:10.1016/S0140-6736(04)17018-9

Zahodne, L. B., Glymour, M. M., Sparks, C., Bontempo, D., Dixon, R. A., MacDonald, S. W. S.,

& Manly, J. J. (2011). Education does not slow cognitive decline with aging: 12-year

evidence from the victoria longitudinal study. Journal of the International

Neuropsychological Society, 17(6), 1039–1046. doi:10.1017/S1355617711001044

Zhang, Z. (2006). Gender differentials in cognitive impairment and decline of the oldest old in

China. The Journals of Gerontology, 61(2), S107–S1015.

246

Zizza, C. A., Ellison, K. J., & Wernette, C. M. (2009). Total water intakes of community-living

middle-old and oldest-old adults. The Journals of Gerontology, 64A(4), 481–486.

doi:10.1093/gerona/gln045

Zylke, J. W. & Bauchner, H. (2016). The unrelenting challenge of obesity. Journal of the

American Medical Association, 315(21), 2277–2278. doi:10.1001/jama.2016.6190

247

Appendix A: Demographics Questionnaire

RIAGENDR – Gender

RIDAGEYR - Age at Screening

RIDRETH1 - Race/Ethnicity – Recode

Mexican American

Other Hispanic

Non-Hispanic White

Non-Hispanic Black

Other Race - Including Multi-Racial

DMDEDUC2 - Education Level - Adults 20+

Less Than 9th Grade

9-11th Grade (Includes 12th grade with no diploma)

High School Grad/GED or Equivalent

Some College or AA degree

College Graduate or above

DMDMARTL - Marital Status

Married

Widowed

Divorced

Separated

Never married

Living with partner

DMDHHSIZ - Total number of people in the Household

INDHHINC - Annual Household Income

$ 0 to $ 4,999

$ 5,000 to $ 9,999

$10,000 to $14,999

248

$15,000 to $19,999

$20,000 to $24,999

$25,000 to $34,999

$35,000 to $44,999

$45,000 to $54,999

$55,000 to $64,999

$65,000 to $74,999

$75,000 and Over

Over $20,000

Under $20,000

249

Appendix B: Occupation Questionnaire

OCQ150 - Type of work done last week

English Text:

(SP Interview Version) In this part of the survey I will ask you questions about

{your/SP's} work experience. Which of the following {were you/was SP} doing last

week . . .

Working at a job or business,

With a job or business but not at work,

Looking for work, or

Not working at a job or business?

OCQ180 - Hours worked last week at all jobs

English Text:

How many hours did {you/SP} work last week at all jobs or businesses?

OCQ210 - Usually work 35 or more hours per week

English Text:

{Do you/Does SP} usually work 35 hours or more per week in total at all jobs or

businesses?

250

Appendix C: Weight History Questionnaire

WHD010 - Current self-reported height (inches)

English Text:

These next questions ask about {your/SP's} height and weight at different times in

{your/his/her} life. How tall {are you/is SP} without shoes?

WHD020 - Current self-reported weight (pounds)

English Text:

How much {do you/does SP} weigh without clothes or shoes?

WHQ030 - How do you consider your weight

English Text:

{Do you/Does SP} consider {your/his/her}self now to be . . .

Overweight,

Underweight, or

About the right weight?

Refused

Don't know

WHD040 - Like to weigh more, less or same

English Text:

Would {you/SP} like to weigh . . .

More,

Less, or

Stay about the same?

Refused

Don't know

WHD050 - Self-reported weight-1 yr ago (pounds)

English Text:

How much did {you/SP} weigh a year ago?

WHD060 - Weight change intentional

English Text:

251

Was the change between {your/SP's} current weight and {your/his/her} weight a year

ago intentional?

WHQ070 - Tried to lose weight in past year

English Text:

During the past 12 months, {have you/has SP} tried to lose weight?

WHQ090 - Tried not to gain weight in past year

English Text:

During the past 12 months, {have you/has SP} done anything to keep from gaining

weight?

WHD110 - Self-reported weight-10 yrs ago (pounds)

English Text:

How much did {you/SP} weigh 10 years ago? [If you don't know {your/his/her} exact

weight, please make your best guess.]

English Instructions:

ENTER WEIGHT IN POUNDS OR KILOGRAMS. IF PREGNANT, ASK FOR

WEIGHT BEFORE PREGNANCY.

WHD120 - Self-reported weight - age 25 (pounds)

English Text:

How much did {you/SP} weigh at age 25? [If you don't know {your/his/her} exact

weight, please make your best guess.]

English Instructions:

IF PREGNANT, ASK FOR WEIGHT BEFORE PREGNANCY.

WHD130 - Self-reported height - age 25 (inches)

English Text:

How tall {were you/was SP} at age 25? [If you don't know {your/his/her} exact height,

please make your best guess.]

WHD140 - Self-reported greatest weight (pounds)

English Text:

Up to the present time, what is the most {you have/SP has} ever weighed?

WHD150 - Age when heaviest weight

English Text:

252

How old {were you/was SP} then? [If you don't know {your/his/her} exact age, please

make your best guess.]

WHD160 - Least self-reported weight since 18(lbs)

English Text:

What is the least {you/SP} ever weighed since {you were/s/he was} 18?

English Instructions:

DO NOT INCLUDE WEIGHT DURING PREGNANCY.

WHD170 - Age when lightest weight

English Text:

How old {were you/was SP} then? [If you don't know {your/his/her} exact age, please

make your best guess.]

253

Appendix D: Blood Pressure Questionnaire

BPQ020 - Ever told you had high blood pressure

English Text:

{Have you/Has SP} ever been told by a doctor or other health professional that

{you/s/he} had hypertension, also called high blood pressure?

BPQ030 - Told had high blood pressure - 2+ times

English Text:

{Were you/Was SP} told on 2 or more different visits that {you/s/he} had hypertension,

also called high blood pressure?

BPQ050A - Now taking prescribed medicine (for hypertension)

English Text:

HELP AVAILABLE (Are you/Is SP) now taking prescribed medicine

BPQ080 - Doctor told you - high cholesterol level

English Text:

{Have you/Has SP} ever been told by a doctor or other health professional that

{your/his/her} blood cholesterol level was high?

BPQ100D - Now taking prescribed medicine (for cholesterol)

English Text:

(Are you/Is SP) now following this advice to take prescribed medicine?

254

Appendix E: Diabetes Questionnaire

DIQ010 - Doctor told you have diabetes

English Text:

The next questions are about specific medical conditions. {Other than during pregnancy,

{have you/has SP}/{Have you/Has SP}} ever been told by a doctor or health professional

that {you have/{he/she/SP} has} diabetes or sugar diabetes?

DIQ040Q - Number of years of age

English Text:

How old {was SP/were you} when a doctor or other health professional first told

{you/him/her} that {you/he/she} had diabetes or sugar diabetes?

DIQ050 - Taking insulin now

English Text:

{Is SP/Are you} now taking insulin

DIQ060Q - Number of mos/yrs taking insulin

English Text:

For how long {have you/has SP} been taking insulin?

DIQ060U - Unit of measure (month/year)

English Instructions:

ENTER UNIT (months or years)

DIQ070 - Take diabetic pills to lower blood sugar

English Text:

{Is SP/Are you} now taking diabetic pills to lower {{his/her}/your} blood sugar? These are

sometimes called oral agents or oral hypoglycemic agents.

255

Appendix F: Medical Conditions Questionnaire Variables

MCQ010 - Ever been told you have asthma

English Text:

Has a doctor or other health professional ever told {you/SP} that {you have/s/he/SP has}

asthma?

Code or Value Value Description

1 Yes

2 No

7 Refused

9 Don't know

MCQ040 - Had asthma attack in past year

English Text:

During the past 12 months, {have you/has SP} had an episode of asthma or an asthma

attack?

MCQ050 - Emergency care visit for asthma/past yr

English Text:

[During the past 12 months], {have you/has SP} had to visit an emergency room or

urgent care center because of asthma?

MCQ053 - Taking treatment for anemia/past 3 mos

English Text:

During the past 3 months, {have you/has SP} been on treatment for anemia, sometimes

called "tired blood" or "low blood"? [Include diet, iron pills, iron shots, transfusions as

treatment.]

MCQ160B – MCQ160L

English Text:

Has a doctor or other health professional ever told {you/SP} that {you/s/he} . . .had

{condition}?

MCQ160B - Ever told had congestive heart failure

MCQ160C - Ever told you had coronary heart disease

MCQ160D - Ever told you had angina/angina pectoris

256

MCQ160E - Ever told you had heart attack

MCQ160F - Ever told you had a stroke

MCQ160G - Ever told you had emphysema

MCQ160I - Ever told you had thyroid disease

MCQ160K - Ever told you had chronic bronchitis

MCQ160L - Ever told you had any liver condition

MCQ170I - Do you still have thyroid disease

English Text:

{Do you/Does SP} still . . . have another thyroid disease?

Code or Value Value Description

1 Yes

2 No

7 Refused

9 Don't know

MCQ170K - Do you still have chronic bronchitis

English Text:

{Do you/Does SP} still . . . have chronic bronchitis?

MCQ170L - Do you still have a liver condition

English Text:

{Do you/Does SP} still . . . have any kind of liver condition?

MCQ180B- MCQ180L – Age when first told had {condition}

English Text:

How old {were you/was SP} when {you were/s/he was} first told {you/s/he} . . .had

{condition}?

MCQ180B - Age when told you had heart failure

MCQ180C - Age when told had coronary heart disease

MCQ180D - Age when told you had angina pectoris

MCQ180E - Age when told you had heart attack

MCQ180F - Age when told you had a stroke

MCQ180G - Age when told you had emphysema

MCQ180I - Age when told you had thyroid disease

MCQ180K - Age when told you had chronic bronchitis

MCQ180L - Age when told you had a liver condition

257

MCQ220 - Ever told you had cancer or malignancy

English Text:

{Have you/Has SP} ever been told by a doctor or other health professional that

{you/s/he} had cancer or a malignancy of any kind?

MCQ230A- MCQ230C - What kind of cancer

English Text:

What kind of cancer was it?

English Instructions:

ENTER UP TO 3 KINDS.

MCQ240A –MCQ240DD - Age cancer diagnosed

English Text:

How old {were you/was SP} when {TYPE OF CANCER/cancer} was first diagnosed?

MCQ250B - Blood relatives have Alzheimer's

English Text:

Including living and deceased, were any of {SP's/ your} biological that is, blood relatives

including grandparents, parents, brothers, sisters ever told by a health professional that

they had . . .Alzheimer's disease?

MCQ260BA- MC260BF – Which Blood relatives have Alzheimer's

English Text:

Which biological [blood] family member?

English Instructions:

CODE ALL THAT APPLY

MCQ260BA - Blood relative-Alzheimer's-mother*

MCQ260BB - Blood relative-Alzheimer's-father*

MCQ260BC - Blood relative-Alzheimer's-mom's mother*

MCQ260BD - Blood relative-Alzheimer's-mom's father*

MCQ260BE - Blood relative-Alzheimer's-dad's mother*

MCQ260BF - Blood relative-Alzheimer's-dad's father*

*Categories are either endorsed or missing

258

Appendix G: Kidney Conditions Questionnaire

KIQ020 - Ever told you had weak/failing kidneys

English Text:

{Have you/Has SP} ever been told by a doctor or other health professional that

{you/s/he} had weak or failing kidneys? Do not include kidney stones, bladder

infections, or incontinence.

259

Appendix H: Physical Activity Questionnaire

PAQ180 - Avg level of physical activity each day

English Text:

Please tell me which of these four sentences best describes {your/SP's} usual daily

activities? [Daily activities may include {your/his/her} work, housework if {you are/s/he

is} a homemaker, going to and attending classes if {you are/s/he is} a student, and what

{you/s/he} normally {do/does} throughout a typical day if {you are/he/she is} a retiree or

unemployed.] . . .

{you sit/he/she sits} during the day and {do/does} not walk about very much.

{you stand or walk/he/she stands or walks} about a lot during the day, but {do/does}not have to

carry or lift things very often

{you/he/she} lift(s) light load or {have/has} to climb stairs or hills often.

{you/he/she} {do/does} heavy work or {carry/carries} heavy loads.

PAD200 - Vigorous activity over past 30 days

The next questions are about physical activities including exercise, sports, and physically

active hobbies that {you/SP} may have done in {your/his/her} leisure time or at school

over the past 30 days. First I will ask you about vigorous activities that cause heavy

sweating or large increases in breathing or heart rate. Then I will ask you about moderate

activities that cause only light sweating or a slight to moderate increase in breathing or

heart rate. Over the past 30 days, did {you/SP} do any vigorous activities for at least 10

minutes that caused heavy sweating, or large increases in breathing or heart rate? Some

examples are running, lap swimming, aerobics classes or fast bicycling.

Yes

No

Unable to do activity

PAD320 - Moderate activity over past 30 days

[Over the past 30 days], did {you/SP} do moderate activities for at least 10 minutes that

cause only light sweating or a slight to moderate increase in breathing or heart rate?

Some examples are brisk walking, bicycling for pleasure, golf, and dancing.

Yes

No

Unable to do activity

260

PAD440 - Muscle strengthening activities

Over the past 30 days, did {you/SP} do any physical activities specifically designed to

strengthen {your/his/her} muscles such as lifting weights, push-ups or sit-ups? Include all

such activities even if you have mentioned them before.

PAD460 - Number of times past 30 days

[Over the past 30 days], how often did {you/SP} do these physical activities? [Activities

designed to strengthen {your/his/her} muscles such as lifting weights, push-ups or sit-

ups.]

PAQ500 - Activity comparison last mo - last yr

How does the amount of activity that you reported {for SP} for the past 30 days compare

with {your/his/her} physical activity for the past 12 months? Over the past 30 days,

{were you/was he/she} . . .

more active

less active, or

About the same?

PAQ540 - Compare activity with 10 years ago

Compared with {yourself/himself/herself} 10 years ago, would you say that {you are/SP

is} . . .

More active now,

Less active now, or

About the same?

261

Appendix I: Physical Activity Individual Activities File

PADACTIV - Leisure time activity

English Text:

[Over the past 30 days], what {vigorous/moderate} activities did {you/SP} do?

English Instructions:

CODE ALL THAT APPLY.

Code or Value Value Description

10 AEROBICS

11 BASEBALL

12 BASKETBALL

13 BICYCLING

14 BOWLING

15 DANCE

16 FISHING

17 FOOTBALL

18 GARDENING

19 GOLF

20 HIKING

21 HOCKEY

22 HUNTING

23 JOGGING

24 KAYAKING

25 PUSH-UPS

26 RACQUETBALL

27 ROLLERBLADING

28 ROWING

29 RUNNING

30 SIT-UPS

31 SKATING

32 SKIING - CROSS COUNTRY

33 SKIING - DOWNHILL

34 SOCCER

35 SOFTBALL

36 STAIR CLIMBING

37 STRETCHING

38 SWIMMING

262

Code or Value Value Description

39 TENNIS

40 TREADMILL

41 VOLLEYBALL

42 WALKING

43 WEIGHT LIFTING

44 YARD WORK

50 BOXING

51 FRISBEE

52 HORSEBACK RIDING

53 MARTIAL ARTS

54 WRESTLING

55 YOGA

56 CHEERLEADING AND GYMNASTICS

57 CHILDREN'S GAMES (E.G. DODGEBALL, KICKBALL)

58 ROPE JUMPING

59 SKATEBOARDING

60 SURFING

61 TRAMPOLINE JUMPING

71 OTHER

For each of the above that are coded, the following questions are asked:

PADLEVEL - Activity level

English Text:

Reported intensity level of activity

Code or Value Value Description

1 MODERATE

2 VIGOROUS

PADTIMES - # of times did activity in past 30 days

English Text:

[Over the past 30 day], how often did {you/SP} do {activity}?

PADDURAT - Average duration of activity (minutes)

English Text:

263

[Over the past 30 days], on average about how long did {you/SP} do {activity} each

time?

PADMETS - MET score for activity (calculated from above information)

English Text:

Metabolic equivalent(MET) intensity level for activity.

264

Appendix J: Diet Behavior and Nutrition

DBD270c - Helping of fruit/fruit juices eaten/day

English Text:

On an average day, how many helpings of the following kinds of foods {do you/does

SP}eat? Fruits or fruit juices.

English Instructions:

RESPONDENT SHOULD DEFINE 'HELPING'.

DBD270d - Helping of vegetables eaten/day

English Text:

On an average day, how many helpings of the following kinds of foods {do you/does

SP}eat? Vegetables, including vegetable salads.

English Instructions:

RESPONDENT SHOULD DEFINE 'HELPING’

265

Appendix K: Smoking Questionnaire

SMQ020 - Smoked at least 100 cigarettes in life

English Text:

These next questions are about cigarette smoking and other tobacco use. {Have you/Has

SP} smoked at least 100 cigarettes in {your/his/her} entire life?

SMD030 - Age started smoking cigarettes regularly

English Text:

How old {were you/was SP} when {you/s/he} first started to smoke cigarettes fairly

regularly?

SMQ040 - Do you now smoke cigarettes

English Text:

{Do you/Does SP} now smoke cigarettes . .

Every day,

Some days, or

Not at all?

SMQ050Q - How long since quit smoking cigarettes (if do not smoke currently)

English Text:

How long has it been since {you/SP} quit smoking cigarettes?

SMQ050U - Unit of measure (day/week/month/year)

Weeks

Months

Years

SMD055 - Age last smoked cigarettes regularly

English Text:

How old {were you/was SP} when {you/s/he} last smoked cigarettes {fairly regularly}?

SMD057 - # cigarettes smoked per day when quit

English Text:

At that time, about how many cigarettes did {you/SP} usually smoke per day?

266

SMD070 - # cigarettes smoked per day now

English Text:

On average, how many cigarettes {do you/does SP} now smoke per day?

SMD075 - How many years smoked this amount

English Text:

For about how many years {have you/has SP} smoked this amount?

SMD080 - # days smoked cigs during past 30 days

English Text:

On how many of the past 30 days did {you/SP} smoke a cigarette?

SMD090 - Avg # cigarettes/day during past 30 days

English Text:

During the past 30 days, on the days that {you/SP} smoked, about how many cigarettes

did {you/s/he} smoke per day?

267

Appendix L: Alcohol Use History Questionnaire

ALQ100 - Had at least 12 alcohol drinks/1 yr?

English Text:

The next questions are about drinking alcoholic beverages. Included are liquor (such as

whiskey or gin), beer, wine, wine coolers, and any other type of alcoholic beverage. In

any one year, {have you/has SP} had at least 12 drinks of any type of alcoholic beverage?

By a drink, I mean a 12 oz. beer, a 4 oz. glass of wine, or an ounce of liquor.

ALQ110 - Had at least 12 alcohol drinks/lifetime?

English Text:

In {your/SP's} entire life, {have you/has he/ has she} had at least 12 drinks of any type of

alcoholic beverage?

ALQ120Q - How often drink alcohol over past 12 mos

English Text:

In the past 12 months, how often did {you/SP} drink any type of alcoholic beverage?

PROBE: How many days per week, per month, or per year did {you/SP} drink?

ALQ130 - Avg # alcoholic drinks/day -past 12 mos

English Text:

In the past 12 months, on those days that {you/SP} drank alcoholic beverages, on the

average, how many drinks did {you/he/she} have?

ALQ140Q - #days have 5 or more drinks/past 12 mos

English Text:

In the past 12 months, on how many days did {you/SP} have 5 or more drinks of any

alcoholic beverage? PROBE: How many days per week, per month, or per year did

{you/SP} have 5 or more drinks in a single day?

ALQ150 - Ever have 5 or more drinks every day?

English Text:

Was there ever a time or times in {your/SP's} life when {you/he/she} drank 5 or more

drinks of any kind of alcoholic beverage almost every day?

268

Appendix M: IRB Exempt Letter

269

Appendix N: Consent Form