Citation
Effects of Improved Physical Fitness on Cognitive/Psychological Functioning in Community-Dwelling, Sedentary Middle-Aged and Older Adults

Material Information

Title:
Effects of Improved Physical Fitness on Cognitive/Psychological Functioning in Community-Dwelling, Sedentary Middle-Aged and Older Adults
Creator:
Morgan, Adrienne
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (201 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Psychology
Clinical and Health Psychology
Committee Chair:
Marsiske, Michael
Committee Members:
McCrae, Christina S.
Bauer, Russell M.
Kneipp, Shawn M.
Price, Catherine
Graduation Date:
8/9/2008

Subjects

Subjects / Keywords:
Age groups ( jstor )
Anxiety ( jstor )
Business executives ( jstor )
Cognition ( jstor )
Control groups ( jstor )
Exercise ( jstor )
Memory ( jstor )
Older adults ( jstor )
Physical fitness ( jstor )
Symptomatology ( jstor )
Clinical and Health Psychology -- Dissertations, Academic -- UF
aerobic, aging, cognition, fitness, psychosocial
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Psychology thesis, Ph.D.

Notes

Abstract:
A growing corpus of research suggests that physical exercise can improve cognition, particularly executive functioning, in older adults. However, limitations of existing research have included (a) insufficient attention to the recruitment of sedentary older adults (who would most likely benefit from exercise interventions); (b) insufficient guidance in test selection drawing on neuropsychological theory and practice; and (c) failure to elucidate the multiple pathways or components of exercise effects on cognition. The current study sought to better clarify these routes to cognitive improvement via (a) assessment of both potential physical fitness and psychosocial mediators of exercise effects on cognition, and (b) inclusion of a control group that received a comparable psychoeducational intervention, matched in study contact hours and study-related non-exercise activities, but which did not receive a physical exercise enhancement intervention. Two randomized groups of 35 (control) and 34 (invention) adults aged 50 years and older were recruited from the Gainesville/Alachua County, Florida region. Both groups underwent pre- and post-intervention cognitive, fitness, and psychosocial/socio-emotional assessment. The exercise promotion intervention group received 16 weeks of intervention (health and fitness education, weekly peer motivational coaching and group support, etc.) in small groups with a peer mentor, while a control/comparison group received 16 weeks of 'health hygiene' instruction, consisting of 16 weeks of education about general health conditions in aging (also in small groups with a peer mentor). Repeated-measures MANOVA indicated no significant between-subjects effect of the intervention (p > .05). There were multivariate within-subjects effects for occasion; however, there were no study group-by-occasion interaction effects. Follow-up univariate analyses revealed within-subjects effects for 9 cognitive variables. There was a modest study group-by-occasion interaction on the COWA test, with intervention group participants improving significantly more across testing occasions. Next, exploratory age group analyses revealed significant multivariate between-subjects effects of age on executive measures only. Follow-up univariate analyses demonstrated age group effects for 4 cognitive variables. For each cognitive measure, younger participants performed significantly better than their older counterparts. In addition, there were study group-by-occasion interaction effects that suggested younger control participants performed better on the One-Back Mean RT SD task, while older intervention group participants performed significantly better on LM Delayed Recall. A three-way interaction suggested that younger intervention group participants improved significantly more over time than younger controls and older participants on the Trails B test. Finally, there was modest, but inconsistent, evidence for correlated change between cognitive, physical fitness/activity, and psychosocial variables. These findings lend some support to the previous literature suggesting the benefits of physical fitness/exercise improvements on cognitive function and the frontal aging hypothesis (West, 1996; Zimmerman et al., 2006). Future research should explore the benefits of physical and cognitive interventions in diverse samples of middle-aged and older individuals. Future studies should also explore the use of alternate cognitive and physical fitness assessment tools in elucidating the cognition-fitness relationship. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2008.
Local:
Adviser: Marsiske, Michael.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-08-31
Statement of Responsibility:
by Adrienne Morgan.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Morgan, Adrienne. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/31/2010
Classification:
LD1780 2008 ( lcc )

Downloads

This item has the following downloads:


Full Text






APPENDIX C
QUALITY CONTROL CHECKLISTS

EXERCISE PROMOTION INTERVENTION GROUP


Mentor/Coach


Date


Procedure
CREATING A SUPPORTIVE ENVIRONMENT
1. Reviews participants' exercise during previous
week.
2. Use of Open-Ended Questions
ex.: "In what ways has exercise been helpful to
you?"
3. Use of Affirmations
ex.: "You are a very conscientious person. That
quality will help you to begin an exercise program."
4. Use of Reflective Responses
ex.: "It sounds like you are frustrated. How do you
deal with that?"
5. Use of Summary Statements
6. Effective group management
Keeps group on topic
Manages time and pace of discussion
Maintains leadership of discussion

COMMUNICATION ROADBLOCKS
7. Avoids Lecturing
8. Avoids giving advice
9. Avoids interpreting or analyzing
10. Avoids questioning participant

CONCLUSION
11. Assures participant that all instructions are in
the Workbook & reminds to bring all logs and
complete any homework for next session.
12. Makes appropriate referrals regarding questions
participants may have about the study.
13. Makes appropriate referrals regarding mental or
physical health concerns observed during sessions.


None Part Full Score

0 5 10

0 5 10


0 5 10


0 5 10


0 5 10


0 5 10

0 5 10


Total


Comments for Mentor:









Finally, collaboration with researchers in the field with expert experience in the recruitment and

retention of diverse samples would be ideal.

Further conceptualization of the target cognitive domains

A third future direction, further conceptualization of executive cognitive outcome

measures, has theoretical implications for the continued examination of the differential effect of

improved cardiovascular fitness on executive cognitive function. The term "executive function"

is used as an umbrella term to describe an individual's ability to engage successfully in

independent, purposive, self-serving behavior and involve strategies necessary to approach, plan,

or carry out cognitive tasks, or the strategies needed in the monitoring of performance (Lezak et

al., 2004). Given such a broad definition of this construct, it is understandable that previous

literature has shown little consensus regarding appropriate measures of executive function or its

underlying factor structure (Kemper & McDowd, 2008). Throughout the literature, there have

been many measures of executive function used, with both macro and micro approaches to

assessment (Kemper & McDowd, 2008). This variation has contributed to differing factor

structures observed for executive function based on the sample studied (i.e., young adults,

normal elders, and Alzheimer's patients). One example is how Royal et al. (2003) found one

three-factor structure of executive function in older adults (abstraction, procedural control, and

attention switching), while others have found a different three-factor structure (shifting,

updating, and inhibition) in younger samples (Miyake et al, 2000; Friedman et al., 2004). In

general, convergent validity and discriminant validity among executive measures have not been

well supported (Luszcz & Lane, 2008). Salthouse (2005) has posited that executive function

tests are essentially tests of fluid ability or general intelligence ("g").

Two specific problems with executive function tests that have been cited include a "task

impurity" problem and the fact that performance on executive measures may reflect many









Furthermore, control beliefs and self-efficacy are also influenced by physical exercise

interventions. One's self-perception beliefs have a critical impact on whether or not he will

engage in an exercise or health intervention. The major social-cognitive theories of exercise

include the basic set of social-cognitive principles: self-efficacy expectations, outcome

expectations, outcome values, and intentions to change, modify, or initiate a certain behavior

(Maddux, 1993). These factors are key mediators of the relationship between exercise behaviors

and increased fitness, and these factors must be addressed and often intervened upon in order to

ensure individuals will participate and adhere to any physical health intervention (Bandura,

1997). Thus, in the present study, these factors will be directly addressed by the exercise

promotion intervention, with the aim of improving self-efficacy/control beliefs, thereby

increasing exercise participation and desired physical fitness and cognitive outcomes.

Conclusion

As the preceding review of the literature has discussed, there may be multiple routes by

which increasing exercise contributes to cognitive improvement. Exercise intervention may

effect cognitive change through improvement of physical factors (particularly cardiovascular

fitness) and/or improvement of psychosocial factors (self-perception influences, such as

depression/well-being, control beliefs, and self-efficacy). Extant research has not yet paid

adequate attention to disentangling these multiple pathways, thus the present study seeks to

disentangle such exercise-related effects on cognition in older adults, thereby building on

existing literatures, which have mostly examined these.










If eligible say:


"This study has several sections. First, based on our phone call today, I would like to schedule
an in-person visit, during which I can assess aspects of your mental and physical fitness, your
health, and your everyday functioning, as well as to begin your enrollment at our Living Well
facility. I would need to schedule this meeting within the next few weeks. Depending on
individual circumstances, this session can take anywhere from 1 to 3 hours."



16. Are you willing to schedule this in-person meeting?

Y E S .............................. ..... .............. ...... 1
NO 2 = INELIGIBLE

"In order to be eligible to participate in this study and to ensure that it is safe for you to be
physically active, we also need a signed checklist from your doctor or nurse clearing you to
exercise. We have prepared this checklist, and you should be receiving it by mail soon. It should
only take your doctor a few moments to complete it. We will ask you to get it filled out in the
next 1-2 weeks."

17. Will you be able to make sure that your doctor or health care provider completes this
checklist?

Y E S ............................... .. ......... .......1. .
NO 2 = INELIGIBLE

"After our first in-person session and clearance from your doctor, you may be eligible to
participate in our program. At the outset, however, you should know that this is an involved
study and will require a significant time commitment on your part. Over the 16-week study
period, you will be asked to devote approximately 1.5 hours each week to meeting with research
staff. Although we provide convenient parking and flexible scheduling, it is important for you to
consider whether this is reasonable for you."

18. Are you able to participate in the study for the entire 16-week study period?

Y E S .................. ................... ............. ........
NO 2 = INELIGIBLE

"In addition to the weekly time commitment, you will be asked to wear an activity monitoring
device throughout the study. These devices, either an accelerometer or pedometer, should not
restrict any of your normal daily activity. However it is important that you wear this device daily









Reitan, R. M. (1992). Trail Making Test: Manual for Administration and Scoring. Tucson:
Reitan Neuropsychology Lab.

Reitan, R. M., & Wolfson, D. (1994). A selective and critical review of neuropsychological
deficits and the frontal lobes. Neuropsychology Review, 4, 161-198.

Royall, D. R., Chiodo, L. K., & Polk, M. J. (2003). Executive dyscontrol in normal aging:
Normative data, factor structure and clinical correlates. Current Neurology and
Neuroscience Reports, 3, 487-493.

Ruitenberg, A., Skoog, I., & Otta, A. (2001). Blood pressure and risk of dementia: Results from
the Rotterdam Study and the Gothenburg H-70 Study. Dementia & Geriatric Cognitive
Disorders, 12(1), 33-39.

Ruppar, T. M., & Schneider, J. K. (2007). Self-reported exercise behavior and interpretations of
exercise in older adults. Western Journal of Nursing Research, 29, 140-157.

Sabatini, T., Frisoni, G. B., Barbisoni, P., Bellelli, G., Rozzini, R., Trabucchi, M. (2000). Atrial
fibrillation and cognitive disorders in older people. Journal of the American Geriatrics
Society, 48, 387-390.

Sallis J. F., & Owen, N. (1999). Physical activity and behavioral medicine. Thousand Oaks, CA
Sage Publications.

Salthouse, T. A. (2005). Relations between cognitive abilities and measures of executive
functioning. Neuropsychology, 19, 532-545.

Salthouse, T. A., Atkinson, T. M., & Berish, D. E. (2003). Executive functioning as a potential
mediator of age-related cognitive decline in normal adults. Journal ofExperimental
Psychology: General, 132, 566-594.

Saxby, B. K., Harrington, F., McKeith, I. G., Wesnes, K., & Ford, G. A. (2003). Effects of
hypertension on attention, memory, and executive function in older adults. Health
Psychology, 22(6), 587-591.

Schimdt, R., Fazekas, F., Offenbacher, H., et al. (1993). Neuropsychologic correlates of MRI
white matter changes. European Neurology, 39, 80-89.

Schiller, J., Adams, P., & Coriaty, N. (2005). Summary health statistics for the US population:
National health interview survey 2003. National Center for Health Statistics. Vital
Statistics, 10(224).

Seeman, T. E., & Crimmins E. (2001). Social environment effects on health and aging:
integrating epidemiologic and demographic approaches and perspectives. Population
Health andAging Annals of the New York Academy of Sciences, 954, 88-117.









TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ..............................................................................................................4

L IS T O F T A B L E S ................................................................................. 9

LIST OF FIGURES ............ ...................................... .....................11

ABSTRAC T ................................................... ............... 12

CHAPTER

1 STATEMENT OF THE PROBLEM..................................................................... 14

2 REVIEW OF THE LITERA TURE ............................................................. ....... ........16

O overview ................. .... .......................... .......... ................................ 16
C onceptu al m odel ............................................................................................................ 17
S ig n ific a n c e .............. ...................................................................................................... 1 7
Continuum of Late Life Cognition ................... .. .......... ....... ........ ............ 18
Cardiovascular and Physical Health Influences on Late Life Cognition.............................22
Relationship between hypertension and cognition .............. .. .. .............. ...................23
Underlying physiological mechanisms...................... ... ........... ...............25
Synergistic effect of hypertension and other cardiovascular conditions.........................27
Environm mental factors ...................... ........ .......................27
Self-Perception Influences on Late Life Cognition.......................... ................. 28
D expression and W ell-B eing.................................................. ............................... 28
Vascular Depression Hypothesis ............... ............... .................... 31
Control Beliefs and Self-Efficacy ............................................................................33
Plasticity of L ate L ife C ignition ......... ................. ................................... .........................34
Cognitive Interventions ................................. ..... ...... ................ ............. 35
Physical Exercise Interventions: Cardiovascular Fitness.................... ... .............36
Relationship between fitness and cognition..........................................................37
Psychological benefits of exercise interventions ............................................. 40
C o n clu sio n ................... ...................4...................1..........

3 METHODS .........................................43

O v erv iew ................... ...................4...................3..........
P articip an ts ...................... ................................................................................................. 4 4
Statistical Power Considerations ................. .................................44
Sam ple C characteristics ..............................................................45
C o g n itiv e E x clu sio n s ................................................................................................. 4 6
Physician's Permission ................................. ........................... .... ....... 46




6









Table 4-21. Correlations between physical activity, physical fitness, and cognitive change scores
VO2 LTEQ Mean MVPA
Pedometer Steps
NAART -0.09 -0.01 0.06 -0.02
Trails A Time -0.03 -0.13 -0.02 -0.18
LM Immediate Recall *0.31 0.01 0.14 0.01
LM Delayed Recall 0.14 0.03 0.09 0.06
LM Recognition 0.09 -0.01 0.06 -0.09
LM Learning Slope -0.02 -0.13 *-0.29 -0.16
COWA -0.18 0.03 -0.05 0.08
Trails B Time -0.13 0.23 -0.06 0.16
Letter Number Sequencing 0.09 -0.01 -0.15 -0.10
One-Back: Mean RT 0.21 -0.07 0.13 0.04
One-Back: Mean RT SD 0.11 *-0.25 0.04 -0.21
Two-Back: Number Correct -0.11 0.00 0.09 -0.06
Two-Back: Mean RT -0.21 -0.01 0.11 -0.03
Two-Back: Mean RT SD -0.21 -0.13 0.02 -0.08
Two-Back: Number Correct -0.20 0.09 0.02 0.10
Note: *p < .05; **p < .01; LM = Logical Memory; NAART = North American
Adult Reading Test; COWA = Controlled Oral Word Association; RT = Reaction
Time; SD = Standard Deviation; VO2 = Modified Balke Submax; LTEQ =
Leisure Time Exercise Questionnaire; MVPA = Minutes of Moderate and
Vigorous Physical Activity









Table 4-7. Univariate analyses of variance for physical activity and fitness
Partial
Effects Hypothesis df Error df F P-value 112


Between Subjects: Study Group
Mean Pedometer Steps
LTEQ
VO2
MVPA
BSE
EXSE


Within-Subjects: Occasion
Mean Pedometer Steps
LTEQ
VO2
MVPA
BSE
EXSE

Interactions: Study Group by
Occasion
Mean Pedometer Steps
LTEQ
VO2
MVPA
BSE
EXSE


0.00
0.45
0.68
0.37
3.37
7.30



3.19
17.78
7.97
17.98
1.05


0.96
0.50
0.41
0.54
0.07
**0.01



0.08
**0.00
**0.01
**0.00
0.31


0.00
0.01
0.01
0.01
0.05
0.10



0.05
0.21
0.11
0.21
0.02


5.82 *0.02 0.08


0.95
0.00
0.36
0.59
0.25
1.71


0.33
1.00
0.55
0.45
0.62


0.01
0.00
0.01
0.01
0.00


0.19 0.02


Note: *p < .05; **p < .01; LTEQ = Leisure Time Exercise Questionnaire; VO2 = Modified Balke
Submax; BSE = Barriers Self Efficacy Scale; EXSE = Exercise Self-Efficacy Scale; MVPA=
Minutes of Moderate and Vigorous Exercise










o. Do you have a pacemaker 1= INELIGIBLE 2 8
or internal defibrillator?

p. Do you use portable 1= INELIGIBLE 2 8
oxygen?
q. Do you take steroids or 1= INELIGIBLE 2 8 Meds
for
cortisone? for
Asthm
a,OK
r. Do you use a cane or 1= INELIGIBLE 2 8
walker?

s. Did you ever have medical 1= INELIGIBLE 2 8
problems as a consequence of
alcohol or drug use?
t. Did you ever have legal 1= INELIGIBLE 2 8
problems as a consequence of
alcohol or drug use?
u. Did you ever have 1= INELIGIBLE 2 8
withdrawal symptoms related
to alcohol or drug use?
v. Are you currently on any 1 = ASK NEXT QUESTION 2 8
medication?









NAART, Trails A Time, LM Immediate Recall, LM Delayed Recall, and LM Recognition,

indicating all participants significantly improved across occasions on these measures, regardless

of study group random assignment (Figure 4-5). There was a study group-by-age group

interaction effect for LM Delayed Total Recall, with the older individuals in the exercise

promotion intervention group performing significantly better than younger individuals

(regardless of study group) and older controls.

There was a study group-by-age group interaction for the One-Back Mean RT SD (Figure

4-6). Younger control group participants were significantly less inconsistent in their mean

reaction time on this measure than younger intervention participants and all older participants.

Additionally, there was an age-by-occasion interaction effect for the One-Back Mean RT task.

This interaction effect suggested that younger participants improved significantly more than

older participants in mean reaction time over time. Finally, there was a study group-by-age

group-by-occasion interaction for Trails B Time. Younger intervention group participants

improved significantly more on this task than their younger control group counterparts and older

participants of both study groups.

Executive cognitive variables with significant age group effects were the One-Back Mean

RT, Two-Back Mean RT, and Two-Back Mean RT SD. These findings indicate that the younger

age group had significantly faster mean reaction time and were significantly less inconsistent in

their mean reaction time across occasions, when compared to their older counterparts. Further,

within-subjects effects for executive cognitive variables were the COWA, One-Back Mean RT,

One-Back Mean RT SD, Two-Back Mean RT SD, and Two-Back Number Correct. All

participants improved significantly on these measures across time (i.e., there was improvements









the study's dropouts reported significantly higher depression symptoms. In addition, dropouts

had significantly poorer performance on the Two-Back condition of the N-back task.

Missing Data

To preserve the present sample size with complete data, cell-specific mean replacement

was implemented for three participants (4.3% of cases; Baltes & Mayer, 2001) on baseline and

post-test cognitive variables (N-back and Logical Memory (LM)). Each missing data point for

the cognitive variables was replaced with the mean value of scores obtained by individuals of the

same gender, age group (i.e., 50-64 years old or 65 and older), and study group (control or

intervention condition).

Each missing data point for baseline physical fitness/activity and self-efficacy variables

was replaced with the mean value of those with the same gender, age group, and BMI category

(i.e., BMI <18.5; 18.5-25; 25-30; and >30). This strategy was chosen as a better approximation

of the physical variables, and because there were no baseline differences between the study

groups on physical fitness, physical activity, or self-efficacy data (Buman, 2008). Six cases

(8.7% of the sample) had missing data for the BMI variable. These data were replaced with

mean BMI for individuals of the same gender and age group. Additionally, there was a total of 6

cases (8.7% of the sample) with missing baseline physical fitness data (pre-V02 = 1 missing

case), baseline physical activity (pre-LTEQ (leisure time exercise) = 1 missing case; pre-MVPA

(minutes of moderate and vigorous physical activity) = 2 cases; and pre-mean pedometer steps =

4 missing cases), and/or baseline self-efficacy data (pre-EXSE (exercise self-efficacy) = 1

missing case). After BMI mean replacement, one case did not share the same gender, age group,

and BMI category with anyone, thus that person's missing data (for pre-mean pedometer steps)

were replaced with mean values of individuals of the same gender and age group.









Table 4-3. Baseline correlations between physical activity, physical fitness, and cognitive
variables
VO2 LTEQ Mean Pedometer MVPA
Steps
NAART -0.107 0.100 -0.111 0.128
Trails A Time -0.111 0.150 0.010 0.140
LM Immediate Recall 0.105 -0.199 -0.122 *-0.306
LM Delayed Recall 0.168 *-0.240 -0.171 **-0.317
LM Recognition 0.085 *0.295 -0.122 **-0.365
LM Learning Slope -0.004 -0.206 -0.226 -0.105
COWA -0.090 0.036 0.130 0.093
Trails B Time -0.171 *0.269 -0.015 0.177
Letter Number Sequencing 0.072 -0.073 0.101 -0.111
One-Back: Mean RT 0.008 0.018 -0.036 -0.001
One-Back: Mean RT SD 0.020 -0.039 -0.024 -0.001
Two-Back: Number Correct 0.010 -0.013 0.018 -0.039
Two-Back: Mean RT -0.132 -0.087 -0.063 -0.160
Two-Back: Mean RT SD -0.189 -0.066 -0.062 -0.122
Two-Back: Number Correct 0.136 -0.032 0.125 0.005
Note: *p < .05; **p < .01; LM = Logical Memory; NAART = North American
Adult Reading Test; COWA = Controlled Oral Word Association; RT = Reaction
Time; SD = Standard Deviation; VO2 = Modified Balke Submax; LTEQ = Leisure
Time Exercise Questionnaire; MVPA = Minutes of Moderate and Vigorous Physical
Activity









Table 4-9. Follow-up univariate analyses of variance for depression and anxiety measures
Effects Hypothesis df Error df F P-value Partial q2
Between-Subjects: Study Group
STAI State Anxiety 1 67 0.04 0.85 0.00
STAI Trait Anxiety 1 67 1.10 0.30 0.02
GDS 1 67 1.33 0.25 0.02
BDI-2 1 67 1.46 0.23 0.02

Within-Subjects: Occasion
STAI State Anxiety 1 67 1.07 0.31 0.02
STAI Trait Anxiety 1 67 11.58 **0.00 0.15
GDS 1 67 0.01 0.91 0.00
BDI-2 1 67 22.68 **0.00 0.25

Interactions: Study Group by Occasion
STAI State Anxiety 1 67 1.53 0.22 0.02
STAI Trait Anxiety 1 67 2.14 0.15 0.03
GDS 1 67 0.30 0.58 0.00
BDI-2 1 67 0.03 0.87 0.00
Note: *p < .05; **p < .01; GDS = Geriatric Depression Scale; BDI-2 = Beck Depression Inventory- Second Edition; STAI = State
Trait Anxiety Inventory









Table 4-23. Correlations between psychosocial and cognitive change scores
GDS BDI-2 STAI State Anxiety STAI Trait Anxiety BSE EXSE
NAART -0.01 -0.14 -0.01 -0.06 0.06 -0.15
Trails A Time 0.03 0.05 -0.06 -0.11 0.18 0.12
LM Immediate Recall -0.22 -0.09 -0.05 0.15 0.11 0.08
LM Delayed Recall *-0.24 0.01 0.03 0.21 -0.01 -0.02
LM Recognition -0.02 -0.03 -0.12 0.16 0.07 -0.12
LM Learning Slope -0.03 -0.03 0.04 -0.20 0.13 0.04
COWA -0.17 -0.01 -0.10 0.00 0.22 0.02
Trails B Time -0.09 0.21 0.18 -0.02 0.04 -0.09
Letter Number Sequencing 0.08 0.19 0.08 0.15 0.08 -0.16
One-Back: Mean RT 0.04 -0.08 0.01 -0.08 -0.10 -0.13
One-Back: Mean RT SD -0.02 -0.03 0.00 0.01 -0.11 -0.14
One-Back: Number Correct 0.00 -0.20 0.03 -0.17 0.15 0.09
Two-Back: Mean RT 0.12 -0.19 0.20 0.01 0.12 -0.01
S Two-Back: Mean RT SD 0.02 *-0.26 0.21 -0.08 0.18 -0.03
Two-Back: Number Correct 0.06 -0.11 0.15 -0.01 *-0.28 -0.02
Note: *p < .05; **p < .01; LM = Logical Memory; NAART = North American Adult Reading Test; COWA = Controlled Oral Word
Association; RT = Reaction Time; SD = Standard Deviation; GDS = Geriatric Depression Scale; BDI-2 = Beck Depression Inventory-
Second Edition; STAI = State Trait Anxiety Inventory; BSE = Barriers Self Efficacy Scale; EXSE = Exercise Self-Efficacy Scale









expectations are directional), and several effect size possibilities. This is a lower-bound

estimate, since it is based on a post-test-only comparison; the actual analyses were conducted as

a mixed within-between analysis of variance. With expected high test-retest reliability in the

primary outcome measures (executive cognitive function), under the same effect sizes, power

would have exceeded that reported here. While cognition was a primary outcome in this

dissertation study, it was a secondary outcome of the larger trial in which it is embedded. The

study was therefore powered to affect primary outcomes of the larger trial (exercise participation,

exercise self efficacy, and fitness). In the event that, as a secondary outcome of the larger trial,

power is not adequate to detect significant differences on cognitive outcomes, the study will still

be useful as a preliminary study for future work (specifically, to aid in effect size determination

for future studies). In addition, the correlational aims of this study (to examine the extent to

which fitness changes and psychosocial changes serve as proximal predictors of cognitive

changes) did not require significant intervention effects to be tested and have substantially

greater power.

Sample Characteristics

Table 3-2 describes participant characteristics. Overall, the sample had an average age of

63.9 years. The vast majority of the sample identified themselves as Caucasian/White (91.3%)

and female (84.1%). There was an under-representation of male participants, as is customary in

cognitive aging research, as well as ethnic minorities (i.e., African-American/Black, Hispanic,

and Asian). On average, the sample was college-educated (16.2 years of education), and the

average estimated IQ was 113, which falls in the high average range of intelligence. The sample

reported a minimal level of depression and anxiety symptoms.









performance were expected to be dependent on the anticipated improvements in physical fitness

and exercise self-efficacy.

First, the critical group-by-occasion effect was examined in univariate analyses of variance

(ANOVA). This effect would address whether there were experimental group differences in

exercise change over time, assessing the hypothesis of disproportionate improvement in the

experimental group. The critical group-by-occasion interaction did not reach significance for

any of the larger study's primary outcomes, suggesting that the intervention was not effective in

producing disproportionate improvements in intended outcome variables. This suggested that a

core assumption in the study's conceptual model (that the intervention would affect fitness and

activity) was not supported. Despite this, the planned primary analyses were completed, to

assess whether there was, nonetheless, a direct association between exercise group assignment

and secondary cognitive outcomes.

It should be noted that there was a single between subjects effect for one measure: exercise

self-efficacy (EXSE) (F(1, 67) = 7.30; p = .01; Partial r2 = 0.10). The exercise promotion

intervention group had significantly higher exercise self-efficacy than the control group, but this

was true across both pre-test and post-test (i.e., it did not indicate the presence of group

differences in self-efficacy change, which was hypothesized).

Independent of group assignment, significant within-subjects effects (i.e., occasion effects,

or significant changes from pretest to posttest) were found for four measures. For the LTEQ

(F(1, 67) = 17.78; p < .001; Partial rl2 = 0.21), MVPA(F(1, 67) = 17.98;p < .001; Partial r2

0.21), and V02 (F(1, 67) = 7.97; p = .006; Partial 12 = 0.11), participants across both groups

improved across time on these measures. Somewhat unexpectedly, for exercise self-efficacy, the

significant occasion effect (F(1, 67) = 5.82; p = .02; Partial r2 = 0.08) indicated that, across









I especially thank my partners in crime, Matt and Joe, for challenging me to be a better

researcher. I offer my heartfelt appreciation also to my research assistants, without whom, I

would not have been able to collect these data: Katherine Blasewitz, Lauren Cohen, Robin Ezzel,

Joseph Gullet, Mario Jimenez, Phuong Nguyen, Jacqueline Memminger, Meghan Saculla, and

Howin Tsang. Thanks also to the many community partners, peer group mentors, and other

support staff that made this project possible.

Next, I thank my parents, Bennie and Gail Aiken, for their unconditional love. My parents

have been there to support me through the lowest of lows and the highest of highs throughout my

life. This journey has not always been easy, but they sure helped to carry many of my burdens. I

thank them for their encouragement to always strive to do and be my best- no matter the

challenge set before me. I thank them for always pushing me to reach for the highest and

brightest stars. I am indebted to them for teaching me, and also often reminding me, that I can

indeed do all things through Christ who strengthens me.

Last, to my dear husband and my rock, Rashad- thank you for holding me down and being

the best friend and partner I could ever imagine finding. Words cannot adequately express the

gratitude and love that I have for you. Thank you for pushing me forward all of the times I felt I

would only go backwards. Thank you for understanding me like no one else can. Thank you for

all of the little things you do to remind me how blessed I am to have you in my life. If I had one

wish, it would be to simply spend the rest of my days growing old right beside you.









Table 4-17. Multivariate analyses of variance for cognitive domains with study group and age group as between-subjects factors
Wilks' Lambda F Hypothesis df Error df P-value Partial j2
Non-Executive Measures
Between Subjects: Study Group 0.89 1.20 6 60 0.32 0.11
Age Group 0.83 2.09 6 60 0.07 0.17
Study Group by Age Group 0.90 1.17 6 60 0.33 0.10
Within Subjects: Occasion 0.55 8.11 6 60 **0.00 0.45
Interaction: Study Group by Occasion 0.92 0.83 6 60 0.55 0.08
Age Group by Occasion 0.97 0.34 6 60 0.91 0.03
Study Group by Age Group by Occasion 0.94 0.63 6 60 0.70 0.06

Executive Measures
Between Subjects: Study Group 0.96 0.27 9 57 0.98 0.04
Age Group 0.73 2.33 9 57 *0.03 0.27
Study Group by Age Group 0.77 1.91 9 57 0.07 0.23
o Within Subjects: Occasion 0.64 3.58 9 57 **0.00 0.36
Interaction: Study Group by Occasion 0.86 1.01 9 57 0.44 0.14
Age Group by Occasion 0.89 0.80 9 57 0.62 0.11
Study Group by Age Group by Occasion 0.83 1.34 9 57 0.24 0.17
Note: *p <.05; **p <.01;









Just as depression has been linked to frontal lobes and executive cognitive function, the

vascular depression hypothesis has also linked vascular-related depression to frontal/executive

cognition (Alexopoulos, 2006). Clinically, vascular depression has been described as a medial

frontal lobe syndrome (Krishnan et al., 2004), with symptoms including psychomotor

retardation, apathy, and disability. Further, cerebrovascular lesions have been associated with

persisting, unstable remission of depression and increased dementia risk (Alexopoulos et al.,

2002). Two streams of literature have proposed and used separate terms related to the vascular

depression hypothesis. These include subcortical ischemic depression and depression-executive

dysfunction syndrome, and these terms differ in their assumed etiology (Alexopoulos, 2006).

First, subcortical ischemic depression (Taylor et al., 2006) asserts that the subcortical impairment

and associated depression is due specifically to cerebrovascular disease; however, the etiology of

depression-executive dysfunction syndrome is less defined by specific factors (Alexopoulos,

2006). Instead, this syndrome may be the result of vascular disease, general age-related changes,

degenerative brain disease, or and accumulation of these and other factors (Alexopoulos, 2006).

Some in this field have called for internal consistency studies to better classify vascular

depression as a syndrome and to decide upon a uniform etiology (Alexopoulos, 2006; Sneed et

al, 2006). Nonetheless, regardless of a specific etiology, in the present study, it is expected that

exercise-related improvements in cardiovascular health will work to improve depression

symptoms, which is hypothesized to lead to cognitive improvements.

Control Beliefs and Self-Efficacy

In older adults, beliefs about control and self-efficacy are importance factors with regard to

cognition. Study has suggested that control beliefs in late life show a significant relationship

with cognitive performance (Miller et al., 1999). These findings highlight the importance of

considering the impact control beliefs and other background variables may have when















V-


Baseline


U


Interventi


Control


Post-Test


Baseline


Occasion


S tContron

Intervention


Post-Test


Occasion


Figure 4-2. Mean anxiety and depression scores by study group and occasion A) STAI-State Anxiety. B) STAI-Trait Anxiety. C)
GDS. D) BDI-2.


33
31
S29
S27












2

Goal
Setting


Initial
Readiness,
Motivation, and
Self-Efficacy
Towards
Exercise


Relatedness
S- social -Verbal Persuasion
Support Vicarious
Experience
4

Mental 7
Imagery] Vicarious
Experience
Physiological
Arousal


Increased
Exercise
Self-Efficacy

Increased
Intrinsic
Motivation


Figure 3-2. Guiding theoretical framework for exercise promotion group intervention (Buman,
2008)


Autonomy
Perceptions of Comp.
Previous
Performance
Accomplish


Increased
Exercise
Frequency
And
Intensity

Improved
Physical
Fitness









LIST OF REFERENCES


Aiken-Morgan, A. T., Marsiske, M. & Whitfield, K. E. (2008). Characterizing and explaining
differences in cognitive test performance between African American and European
American older adults. Experimental Aging Research, 34(1), 80-100.

Allaire, J.C., Tamez, E., & Whitfield, K. E. (2007). Examining the association between lung
functioning and cognitive performance in African American adults. Journal ofAging and
Health, 19, 106-122.

Alexander, G.E., DeLong, M.R., & Strick, P.L. (1986). Parallel organization of functionally
segregated circuits linking basal ganglia and cortex. Annual Review ofNeuroscience, 9,
357-381.

Alexopoulos, G. S. (2006). The vascular depression hypothesis: 10 years later. Biological
Psychiatry, 60(12), 1304-1305.

Alexopoulos, G. S., Kiosses, D. N., Choi, S. J., Murphy, C. F., & Lim, K. O. (2002). Frontal
white matter microstructure and treatment response of late-life depression: A preliminary
study. American Journal ofPsychiatry, 159, 1929-1932.

Alexopoulos, G. S., Meyers, B. S., Young, R. C., Campbell, S., Silbersweig, D., & Charlson, M.
(1997). The "vascular depression" hypothesis. Archives of General Psychiatry, 54(10),
915-922.

Angevaren. M., Aufdemkampe, G., Verhaar, H. J. J., Aleman, A., Vanhees, L. (2008). Physical
activity and enhanced fitness to improve cognitive function in older people without known
cognitive impairment (Review). Cochrane Database of Systematic Reviews, Apr 16
(2),CD005381.

Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman

Ball, K., Berch, D. B., & Helmers, K. F. (2002). Effects of cognitive training interventions with
older adults: A randomized controlled trial. Journal of the American Medical Association,
288(18), 2271-2281.

Ball, K., & Owsley, C. (2000). Increasing mobility and reducing accidents in older drivers. In
K.W. Schaie (Ed.), Societal Impacts on Mobility in the Elderly. New York: Springer.

Baltes, P. B., & Mayer, K. U. (Eds.). (2001). The Berlin Aging Study: Aging from 70 to 100
(Paperback ed.). New York: Cambridge University Press.

Baltes, P. B., Staudinger, U. M., Lindenberger, U. (1999). Lifespan psychology: Theory and
application to intellectual functioning. AnnualReview ofPsychology, 50, 471-507.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social
psychological research: Conceptual, strategic and statistical considerations. Journal of
Personality and Social Psychology, 51, 1173-1182.









peer mentors who did not adhere well to study protocol were not asked to continue for future

groups. In order to ensure quality control of the intervention delivery, all group sessions were

either video- or audio-recorded, and peer mentors were evaluated on several criteria

(supportiveness, effective communication, etc; Appendix C). Feedback was recorded and

discussed with the peer mentors prior the next group session.

Each intervention group weekly session lasted between 45 and 60 minutes and began with

an initial check-in regarding physical activity completed between sessions. The schedule of

intervention topics for the study is listed in Table 3-6. The group discussed challenges and

barriers to engaging in physical activity, as well as successful participation in exercise behaviors.

The remainder of the session was devoted to introduction and discussion of the topic for the

week. At the end of each session, homework was assigned to the group. Half of the participants

arrived 30 minutes prior to each weekly intervention group session to complete 15-30 minutes of

computerized assessments of cognition, exercise self-efficacy, and barriers to exercise self-

efficacy. The second half of the participants remained 30 minutes after each group session to

complete the computerized assessments.

Design and Rationale for Control Group

The 35 participants randomized to the "health hygiene" control group received a set of

weekly topics based on materials from the National Institutes of Health's SeniorHealth website

(NIHSeniorHealth.gov). This website featured up-to-date, easily understandable and accessible

health information for seniors and their family and friends. Control participants met with a peer-

mentored group each week to receive didactic instruction based on health and aging topics

discussed on the NIH SeniorHealth website (Alzheimer's disease, arthritis, balance problems,

COPD, diabetes, etc.). Similar to quality control practices employed for the intervention

condition, control group sessions were also recorded so that specific feedback could be given to









Cognitive Interventions

Cognitive training (e.g., Jobe et al. 2001, Ball et al., 2002) has been shown to lead to an

improvement in cognitive function. Early studies have examined the plasticity of late life

cognition in laboratory settings and found these interventions to be successful in improving

cognitive function (e.g., Rebok & Balcerak, 1989). In addition, these early investigations have

shown cognitive interventions can lead to improved performance on specific mental and

perceptual functions, as well as certain aspects of everyday cognition (Leirer, Morrow, Pariante,

& Sheikh, 1998; Ball & Owsley, 2000). More recent study has continued to explore the benefits

of cognitive training intervention with older adults. Specifically, the Advanced Cognitive

Training for Independent and Vital Elders (ACTIVE) Clinical trial, a randomized controlled trial

of three cognitive intervention approaches (speed, reasoning, and memory) for older adults,

demonstrated strong, broad, and durable cognitive ability-specific training effects, comparable to

or greater than the amount of cognitive decline observed in other longitudinal studies, suggesting

the interventions have the potential to reverse age-related decline (Ball et al., 2002). There was

minimal transfer to training effects to everyday activities (i.e., functional competence); however,

through a two-year follow-up period, there was no evidence of a significant decline in ADL and

IADL status (Ball et al., 2002). At a five-year follow-up period, there was evidence for transfer

of cognitive training gains to IADL function, with individuals receiving cognitive intervention

showing slower rates of functional decline relative to controls (Willis et al., 2006).

In general, improvements from cognitive training interventions have been found to be very

specific to the cognitive skill/domain trained, with few broader cognitive improvements (Kramer

& Willis, 2002). This is a major limitation of cognitive training protocols when it comes to

intervening upon functional aging-related declines and reducing the risk and/or progression of









Table 4-18. Follow-up univariate analyses of variance for non-executive measures: Study group
and age group as between subjects factors
Hypothesis Error Partial
Effects df df F P-value l2


NAART
Study Group
Age Group
Study Group by Age Group
Occasion
Study Group by Occasion
Age Group by Occasion
Study Group by Age Group by
Occasion
Trails A Time
Study Group
Age Group
Study Group by Age Group
Occasion
Study Group by Occasion
Age Group by Occasion
Study Group by Age Group by
Occasion
LM Immediate Total Recall
Study Group
Age Group
Study Group by Age Group
Occasion
Study Group by Occasion
Age Group by Occasion
Study Group by Age Group by
Occasion
LM Delayed Total Recall
Study Group
Age Group
Study Group by Age Group
Occasion
Study Group by Occasion
Age Group by Occasion
Study Group by Age Group by
Occasion


1.76
2.42
0.48
10.50
0.00
0.29


65 0.05


0.08
7.50
0.48
7.74
0.73
0.36


65 2.16


2.12
0.08
1.80
23.11
0.17
0.21


65 0.14


1.03
0.08
3.17
29.03
0.49
0.00


65 0.43


Note: *p < .05; **p < .01; LM = Logical Memory; NAART
Reading Test


North American Adult


0.19
0.12
0.49
**0.00
0.95
0.59

0.83


0.77
**0.01
0.49
**0.01
0.40
0.55

0.15

0.15
0.78
0.18
**0.00
0.68
0.65

0.71


0.31
0.77
0.08
**0.00
0.49
0.95

0.52


0.03
0.04
0.01
0.14
0.00
0.00

0.00


0.00
0.10
0.01
0.11
0.01
0.01

0.03

0.03
0.00
0.03
0.26
0.00
0.00

0.00


0.02
0.00
0.05
0.31
0.01
0.00

0.01









Regarding post-test data, there was one case missing physical fitness (post-VO2) data at

follow-up. However, for post-test physical activity and self-efficacy variables, there were

substantially more missing data (post-LTEQ = 14 missing cases; post-mean pedometer steps =

17 missing cases; post-MVPA = 23 missing cases; post-BSE = 21 missing cases; and post-EXSE

= 21 missing cases). Since physical activity and self-efficacy data were collected on a weekly

basis, missing post-test values were replaced by the most recent available value for preceding

weeks 13-16 (i.e., if week 16 was missing, this data point was replaced by the value for week 15,

etc.). In the event that there were no available data, these cases were replaced with age, gender,

and BMI group posttest means. There were 19 such cases without any data for weeks 13-16;

thus by-cell mean replacement was completed for these individuals (post-LTEQ = 7 missing

cases; post-mean pedometer steps = 10; post-MVPA = 2; post-BSE = 1 missing case; post-EXSE

= 1 missing case).

Distributions of Dependent Variables and Outliers

Prior to analysis of the cognitive data, values that were +3 standard deviations from the

mean of each cognitive and psychosocial variable were set to missing and replaced with the

mean value of individuals with the same gender, age group, and study group (as described

above). There were 28 specific outlier values across all baseline and post-test cognitive

variables. There were 12 total specific outlier values for baseline variables of LM Leaning

Slope, Letter-Number Sequencing, Geriatric Depression Scale (GDS), Beck Depression

Inventory-2 (BDI-II), One-Back Mean Reaction Time (RT), One-Back Mean Reaction Time

(RT) Standard Deviation (SD), Two-Back Mean RT, Two-Back Mean RT SD, and Two-Back

Number Correct. There were 16 total outliers on post-test measures of LM Leaning Slope,

COWA, Trails B Time, GDS, BDI-II, STAI Trait Anxiety, One-Back Mean RT, One-Back Mean

RT SD, One-Back Number Correct, Two-Back Mean RT, and Two-Back Mean RT SD.









To improve upon these methodological flaws, the present study will examine a sample of

community-dwelling, cognitively intact elders, without major cognitive, physical or mental

health conditions. Additionally, a stream of research suggests that complex social activity

participation may be a protective factor for cognition (e.g., Brown, 1992; McAuley, 1993).

Elders who have had more complex, "engaged" lifestyles, higher levels of education, and who

are more active in later adulthood have been shown to perform at a higher level (cognitively),

and may evince attenuated rates of cognitive decline. For many adults, exercise is a possible

source of complex activity; it may bring some adults out of house, provide them with complex

regimens to monitor, and may increase their social participation with other exercisers ("mall

walking", exercising at a fitness facility, etc.). In order to account for the effects of complex

social activity, the control condition of this exercise promotion intervention will be designed to

control for the influences of complex social engagement on improvements in both cognitive and

affective functioning.

Psychological benefits of exercise interventions

A strong body of research suggests that along with the physical health benefits of exercise

in older adults, there are also strong psychological benefits of engagement in physical activity.

Particularly, there is strong evidence that participation in regular physical activity is associated

with psychological health and well-being for older adults age 50 and older (Brown, 1992;

Elavskey et al., 2005; Elavskey & McAuley, 2007; McAuley, 1993; Motl et al., 2005; Morgan &

Bath, 1998; Pescatello & DiPietro, 1993). The engagement in structured and unstructured

physical activity regimens are related to reduced symptoms of depression and anxiety and

improvements in mood in older adults (Fukukawa et al., 2004). In sum, increasingly the field

has come to accept the link between physical activity and a variety of positive emotional,

behavioral, and physical health outcomes for older adults to be a viable one.









LIST OF TABLES


Table page

3-1 Pre-data collection power analysis results, indicating power expected at different
effect sizes with two cells of 32 participants, and alpha = .05........................................59

3-2 Participants' m ean characteristics........................................................... ............... 60

3-3 Proposed study timeline for each replicate of participants..............................................61

3-4 Number of participants by age category, replicate, and experimental group .................62

3-5 Baseline and post-intervention study protocol ......... ......... ......................................63

3-6 Schedule of intervention topics............ ............... ...................... .................. ............... 64

3-7 Health hygiene control group design ................ ............ ...........................................65

3-8 Schedule of control group topics ............................................... ............................ 66

4-1 Baseline mean comparison of study completers and dropouts................ ............ ....86

4-2 Distributions of dependent variables at baseline .................................... ............... 88

4-3 Baseline correlations between physical activity, physical fitness, and cognitive
variables .............. ..................... .......... ...... ........................ ................. 89

4-4 Baseline correlations between physical fitness, physical activity, and psychosocial
v a ria b le s ....................................................... ..................... ................ 9 0

4-5 Baseline correlations between psychosocial and cognitive variables............................91

4-6 Cognitive intercorrelations at baseline........................... ......................... ....... ........ 92

4-7 Univariate analyses of variance for physical activity and fitness................................95

4-8 Repeated-measures multivariate analysis of variance on anxiety and depression
m e a su re s ..................................................................9 6

4-9 Follow-up univariate analyses of variance for depression and anxiety measures ............97

4-10 Repeated measures multivariate analyses of variance for cognitive domains with
study group as between-subjects factor ... ......................................98

4-11 Follow-up univariate analyses of variance for non-executive measures .........................99

4-12 Follow-up univariate analyses of variance for executive measures .............................100











50.00
48.00
. 46.00
44.00
" 42.00
S40.00
S38.00
Z 36.00
34.00
32.00


Baseline


Occasion


Intervention


Control


Post-Test


108.00
98.00
88.00
78.00
68.00
58.00
48.00


Baseline


Control


Interventio


Post-Test


Occasion


Figure 4-4. Mean cognitive scores by study group and occasion for executive cognitive variables A) COWA. B) Trails B. C) Letter-
Number Sequencing. D) One-Back Mean RT. E) One-Back Mean RT SD. F) One-Back Number Correct. G) Two-Back
Mean RT. H) Two-Back Mean RT SD. I) Two-Back Number Correct.









MET score, by weighting the three intensity levels and summing: 3(mild activity) + 5(moderate

activity) + 9(strenuous activity). Previous research has supported the validity and reliability of

LTEQ score interpretations with adult (Godin et al., 1986; Jacobs et al., 1993) and older adult

populations (Karvinen et al., 2007; Ruppar & Schneider, 2007).

Minutes of Moderate and Vigorous Physical Activity (MVPA)

Minutes of moderate and vigorous physical activity (MVPA) were computed from the LTEQ

by adding the number of moderate and strenuous bouts reported and multiplying by 20. This

value was then added across the week to obtain a weekly measure of MVPA. Recently, to better

approximate the true level of activity needed for reliable health benefits (moderate to vigorous;

Pate et al., 1995), study has suggested that minutes of mild activity be excluded from

calculations of physical activity (Karvinen et al., 2007).

Pedometer

The AE 120 pedometer (Yamax SW200 engine) was used as a second measure of physical

activity. The pedometer was worn on participant's hip to measure steps taken during a given

day, and participants recorded the final number of steps for each day the next morning, in a log.

These values were averaged across each week of the intervention period to obtain mean steps

taken. This measure was included as an objective measure of physical activity. Convergent

validity of pedometers with has been shown self-report measures of physical activity has been

shown (Tudor-Locke, Williams, Reis, & Pluto, 2002).

Modified Balke Submax (VO2)

Cardiorespiratory fitness was measured using a modified Balke treadmill protocol (which

is widely used) to obtain V02 max estimates. The modified Balke is a treadmill protocol that

involves slope increases while speed is held constant. Participants' heart rates were monitored

throughout the protocol and for two minutes prior to testing. Within the first minute, treadmill









FOR ELIGIBLE PARTICIPANTS


"Thank you for answering the questions. As I already implied, we would like to continue with
you in the program and meet you in person and have you meet us. At this in-person meeting, we
will ask you some more questions regarding your mental and physical fitness, health, and
everyday functioning to determine if you are eligible for participation in the program. This
meeting will take up to three hours depending on how much information is needed from you. It
will be held at the Living Well Center at the University of Florida."

"Do you have any questions for me at this time?"

"You have already indicated that you would be willing to schedule this session. To help me with
scheduling an appointment, could you tell me what other commitments you typically have during
the week, such as work, volunteering, caring for others, or social activities?"

23. What days of the week work best for you?
24. Is morning or afternoon better for you?

"We will be getting in touch with you soon to schedule your first visit. In the meantime, you
should be receiving the form for your doctor or nurse to complete to be cleared to participate in
exercise, and we will need this clearance before we can enroll you in the study."

"Thank you very much for your time."


Could I schedule you for (date/time)?

25. Date / /

26. Time: AM / PM

27. Test Site:

28. Person: (Tester ID/Initials)

Do you know where Living Well is located? I will send you a letter and map with directions to
our center.

Thank you. If you wear glasses for distance or reading or wear a hearing aid, please bring them
with you. You should have received, or will be receiving shortly, a one-paged form that needs to
be signed by your physician. Please bring this signed physician's form and any medications you
take with you to your appointment. We look forward to seeing you on (day/date).









Liu-Ambrose et al., 2004), while older adults who engage in regular physical activity reduce

their risk for mortality associated with chronic disease states and premature mortality (Bean, et

al., 2004).

Relationship between fitness and cognition

Early studies of the relationship between cognition and fitness made group comparisons

between physically fit and unfit individuals (McAuley et al., 2004; Kramer, Erickson, &

Colcombe, 2006). Like many early psychological investigations, these studies were cross-

sectional in nature, and they failed to control for a host of factors that influence results observed

(McAuley et al., 2004; Kramer et al., 2006). Such studies came to conclude that physically fit

individuals showed better performance on cognitive measures than did physically unfit persons

(McAuley et al., 2004; Kramer et al., 2006). Today, the dangers of making such conclusions are

better understood. Factors such as age, education, socioeconomic status, race/ethnicity, may

influence performance on cognitive measures. Thus, improving on such methodological

problems of these studies, more recent studies employed longitudinal designs and compared

individuals undergoing an exercise intervention to those who were not (McAuley et al., 2004;

Kramer et al., 2006). These results indicated that while many people demonstrated cognitive

performance improvements in these studies, other participants did not (McAuley et al., 2004;

Kramer et al., 2006). These inconsistencies have left many a bit baffled and motivated to

investigate the underlying truth further (Colcombe & Kramer 2003; McAuley et al., 2004).

Cited methodological reasons for these inconsistencies include reliance on self-report activity

data, failure to distinguish between activities that are aerobic and anaerobic in nature, poor

assessment of duration, intensity, and frequency of exercise activity, non-exclusion of

participants with subclinical dementia, and low statistical power (Kramer et al., 2006).

Accordingly, Colcombe & Kramer (2003) conducted a meta-analysis of 18 studies that have









decomposing age-related variance in cognitive performance (Miller et al., 1999). Additionally,

late life cognitive performance has an impact on control beliefs, suggesting a bi-directional

relationship. While control beliefs have not gained as much attention in the fitness and cognition

literature reviewed above, consideration of these factors is important in an exercise intervention

context. Related to control beliefs is the concept of self-efficacy, which is among various

theories of health behavior change (health belief model and self-efficacy theory) that include

cognitive elements related to decision-making. These cognitive elements of decision-making

abilities are critical determinants of engagement in health and/or exercise behaviors (Bandura,

1997). Such decision-making may reflect individuals' actual barriers to exercise behavior, lack

of motivation, or ambivalence toward exercise behavior. The present study will help individuals

resolve ambivalence towards exercise through techniques (goal setting and mental imagery)

designed to improve self-efficacy/control beliefs. Thus, an important test of our conceptual

model will be to determine whether cognitive performance also benefits from such improvement

in self-efficacy and control beliefs.

Plasticity of Late Life Cognition

Despite the various, and often detrimental influences, of physical and mental health on late

life cognitive function, research examining both cognitive and physical training interventions

with older adults have encouragingly found some positive results for the improvement of

cognitive functions in aged individuals. This evidence supports the notion of the plasticity of

late life cognition, which is relevant to the current proposal, which seeks to intervene through a

physical exercise promotion intervention. Even in later life, cognitive function can show

positive change, and there are both environmental and physical routes to such improvement.

These environmental and physical interventions will be briefly reviewed here.









CHAPTER 2
REVIEW OF THE LITERATURE

Overview

The present investigation examined the effects of improved physical fitness on cognitive

functioning in a group of adults, ages 50 and older. Thus, the review of the literature will first

consider the evidence for normative and non-normative cognitive declines with aging,

highlighting the cognitive continuum in later life. Evidence to date suggests aging

disproportionately affects executive functions and processes, and the present literature review

will discuss these findings. Next, the literature review will consider evidence for the plasticity of

cognition in late life. Specifically, the review will consider research that has shown cognition

may be improved through cognitive training interventions, as well as behavioral interventions.

Such behavioral interventions include physical exercise interventions aimed at improving

cognitive function through improved physical fitness.

One route by which physical exercise interventions may improve cognition is by

improving cardiovascular fitness, and this literature will be discussed. Next, additional routes by

which physical exercise interventions may improve cognition (i.e., via improving mood/affect,

self-efficacy, and control beliefs) will also be explored. This review is intended to set the stage

for the current study, which examined pre-post exercise intervention cognitive performance

changes on selected neuropsychological measures. Additionally, this study empirically

investigated whether measures of executive control processes were particularly sensitive to

physical exercise intervention (as a recent meta-analysis has suggested), relative to other

cognitive domains. This study also explored whether physical (fitness) and psychosocial

variables (mood, affect, and self-efficacy) changed as a result of an exercise promotion

intervention, and whether these helped explain any pre-post changes in cognition









LIST OF FIGURES


Figure pe

2-1 Conceptual model of the effects of exercise intervention on cognitive function ..............42

3-1 Flow chart of participants at each study phase. ............. ................... ................. 67

3-2 Guiding theoretical framework for exercise promotion group intervention (Buman,
2 0 0 8 ) ........................................................................................ . 6 8

4-1 Mean scores by study group and occasion for physical fitness, activity, and self-
efficacy variables ............... .... ... .............. ................ 117

4-2 Mean anxiety and depression scores by study group and occasion..............................120

4-3 Mean cognitive scores by study group and occasion for non-executive cognitive
v a ria b le s ................... ........................................................................ 1 2 2

4-4 Mean cognitive scores by study group and occasion for executive cognitive variables .125

4-5 Mean cognitive scores by study group, age group, and occasion for non-executive
cognitive variables .............................................................. .... .......... 130

4-6 Mean cognitive scores by study group, age group, and occasion for executive
cognitive variables .............................................................. .... .......... .. 133









al., 2004). There tends to be a relative sparing on memory and language tests (Libon et al.,

2004).

These findings guide the current proposed study with regard to the hypothesized selective

benefit of exercise on executive control processes. Furthermore, there is evidence to suggest that

related to the selective decline in executive brain function in older individuals, physical exercise

interventions work to improve these vulnerable frontal and white matter brain regions.

Specifically, brain mechanisms for improvement with physical exercise include increased

cerebrovascular perfusion, cerebrovascular sufficiency, and improved cortical plasticity

(McAuley et al. 2004); these changes occur through increased proliferation of neurotrophin

factors, dendritic branching, neurogenesis, and formation of new synaptic connections (McAuley

et al. 2004). This is a fundamental premise which drives the current proposal, as the hope is to

demonstrate that by improving physical, particularly cardiovascular, fitness levels through

physical exercise intervention, brain cerebrovascular perfusion and sufficiency and cortical

plasticity may be enhanced in a group of vulnerable individuals (due to older age and sedentary

lifestyle), thereby enhancing the neurocognitive (executive control) functions dependent on the

integrity of brain white matter structure and function.

Cardiovascular and Physical Health Influences on Late Life Cognition

There are certain factors that may exacerbate cognitive aging and the aging process in

general. These include behavioral factors, such as diet, physical activity, and other health

promoting behaviors. In addition, environmental factors, such as socioeconomic status, lifetime

educational and occupational opportunity, stress, and a host of other environmental influences,

interfere with "optimal" aging (Burke et al., 2001). Furthermore, the presence of cardiovascular

risk factors can influence cognitive function with aging. It has been documented hypertension is

a major risk for many other cardiovascular diseases (stroke, diabetes, congestive heart failure,









efficacy of various types of interventions, as well as exploration of the optimal combination of

cognitive and physical training that would result in reliable cognitive outcomes for aging

populations.

Conclusion

As the aging population in the US and world continues to rapidly grow, the call for

lifestyle interventions to prevent the onset of both physical disease and cognitive impairment and

dementia will also grow stronger. While the present study found weak effects of an exercise

promotion program, as well as only modest support for association between physical

fitness/activity and psychosocial improvements and cognitive performance improvements over

time, it is critical that future intervention work in the field of cognitive aging build upon the

present study. This work has significant public health implications, as the societal burden of

caring for older adults with disabling medical diseases, including Alzheimer's disease and other

dementias, is great and threatens to become even greater in the future without effective,

preventive lifestyle intervention.












APPENDIX B
NURSE/PHYSICIAN CHECKLIST AND PERMISSION


Project AAMP (Active Adult Mentoring Project)
College ofPuble Health and Health Professions
Adrienne Aiken Morgan


PO Box 100165
Gainesville, FL 32610-0165
Phone: (352) 273-5098


Please type or print clearly


Physician's Name

Patient's Name


Phone #


(Project AAMP participant)


Program exclusion checklist (please check any that apply to this patient):
0 Terminal illness with life expectancy of< 12 months E History of cardiac arrest
E Myocardial infarction in the last 6 months D Uncontrolled angina
D Chronic heart failure (New York Classification III to IV) D Stroke or TIA
D Aortic stenosis D Peripheral vascular disease
D Cardiac arrhythmia D Pulmonary disease requiring oxygen or steroid treatment
D Cardiac stent D Receiving chemotherapy or radiation for cancers
D Cardiac arrest D Ambulation with assistive devices
D Implanted cardiac defibrillator D Poorly controlled diabetes
D Pacemaker D Smoked regularly (>4 cigarettes per day) in past 3 years
D Coronary artery bypass graft D Any of the following calcium channel or beta blockerS


Beta Blockers
Acebutolol (Monitan, Sectral)
Atenolol (Apo-Atenolol, Novo-Atenol, Tenormin)
Betaxolol (Kerlone)
Bisoprolol (Zebeta)
Carteolol (Cartrol)
Labetalol (Normodyne, Trandate)
Oxprenolol (Trasicor, Slow-Trasicor)


Calcium Channel Blockers
Bepridil (Vascor)
Diltiazem (Cardizem, Cardizem CD, Cardizem LA, Cardizem SR,
Dilacor-XR)
Betaxolo (Kerlone)


Note: Trade Names in parenthesis


I hereby give my patient permission to:
1. Participate in an exercise program

2. Complete a health and fitness assessments*


YES

YES


*The fitness assessment includes resting heart rate and blood pressure measurements and an 85% sub-maximal cardiovascular test (heart rate
only, no EKG)

Special instructions or indicated activities:

Contraindications to any activities:


Nurse/Physicians' Signature [Required]









Krishnan, K. R. R., Taylor, W. D., McQuaid, D. R. E., MacFall, J. R., Payne, M. E., Provenzale,
J. M., & Steffens, D. C. (2004). Clinical characteristics of magnetic resonance imaging-
defined subcortical ischemic depression. BiologicalPsychiatry, 55, 390-397.

Krishnan, M., Mast, B. T., Flicker, L. J., Lawhorne, L., & Lichtenberg, P. A. (2005). The effects
of preexisting depression on cerebrovascular health outcomes in geriatric continuing care.
Journal of Gerontology: Medical Sciences, 60(7), 915-919.

Krishnan, M. S., O'Brien, J. T., Firban, M. J., Pantoni, L., Carlucci, G., Erkinjuntti, T., Wallin,
A., Wahlund, L. O., Scheltens, P., van Straaten, E. C. W., & Inzitari, D. (2006).
Relationship between periventricular and deep white matter lesions and depressive
symptoms in older people. The LADIS Study. International Journal of Geriatric
Psychiatry, 21, 983-989.

La Rue, A. (1992). Depression, Aging and neuropsychological assessment (pp. 259-289). New
York: Plenum Press.

Launer, L. J., Ross, G. W., & Petrovitch, H. (2000). Midlife blood pressure and dementia: The
Honolulu-Asia Aging Study. Neurobiology of Aging. 21(1), 49-55.

Leirer, V. O., Morrow, D. G., Pariante, G. M., & Sheikh, J. I. (1988). Elders' nonadherence, its
assessment, and computer assisted instruction for medication recall training. Journal of the
American Geriatrics Society, 36, 877-884.

Lezak, M., Howieson, D., & Loring, D. (2004). Neuropsychological assessment, (4 Edition).
New York: Oxford University Press.

Libon, D., Price, C., Garrett, K., & Giovannetti, T. (2004). From Binswanger's disease to
Leuokoaraiosis: What we have learned about subcortical vascular dementia. The Clinical
Neuropsychologist, 18(1), 83-100.

Liu-Ambrose, T., et al. (2004). Resistance and agility training reduce fall risk in women aged 75-
85 with low bone mass: A 6-month randomized, controlled trial. Journal ofAmerican
Geriatrics Society, 52(5), 657-665.

Lopez, OL; Jagust, WJ; DeKosky, ST; Becker, JT; Fitzpatrick, A; Dulberg, C; Breitner, J;
Lyketsos, C; Jones, B; Kawas, C; Carlson, M; Kuller, LH. (2003). Prevalence and
classification of mild cognitive impairment in the Cardiovascular Health Study Cognition
Study: part 1. Archives ofNeurology, 10(60), 1385-1389.

Luszcz, M. A. & Lane, A. P. (2008). Executive function in cognitive, neuropsychological, and
clinical aging. In S. M. Hofer & D. F. Alwin (Eds.), Handbook of cognitive aging:
Interdisciplinary perspectives. (pps. 193-206). Los Angeles: Sage Publications.

Madden, D. J., Blumenthal, J. A., Allen, P. A., & Emery, C. F. (1989). Improving aerobic
capacity in healthy older adults does not necessarily lead to improved cognitive
performance. Psychology and A'gini,. 4(3), 307-320.









their overall performance on non-executive measures from baseline to post-test. The direction of

this finding will be explored in follow-up univariate analyses presented below.

Next, for the executive cognitive measures, a second repeated-measures MANOVA again

showed no significant occasion-by-group interaction, suggesting that the exercise promotion

intervention group did not improve significantly more or less than the health hygiene control

group over time (Wilks' A= 0.86; F(9, 59) = 0.96, p = .48; Partial r2 = 0.14). There was no

significant between subjects effect (Wilks' A= 0.97; F(9, 59) = 0.31, p = .97; Partial rl2 = 0.03),

but there was a significant within-subjects effect of occasion. (Wilks' A= 0.65; F(9, 59) = 4.01, p

< .001; Partial r2 = 0.35). Follow-up univariate analyses were conducted to ascertain the

direction of the effect and are explored below.

The presence of significant overall occasion effects required follow-up univariate analyses

to ascertain their directions in each measure (Tables 4-11 and 4-12). Beyond occasion analyses,

these follow-up analyses also examined the between-group effect and interactions; while these

effect examinations were not "protected" by the omnibus test, they were still useful as follow-up,

exploratory analyses. The analyses that follow are reported with their original, uncorrected

probabilities, since there is little consensus on alpha adjustment for these follow-up tests

(Tabachnick & Fidell, 2007). However, with six non-executive measures, a simple Bonferroni

correction would require probabilities ofp = 0.05/6 = 0.0083 to be labeled as significant.

Similarly, with nine executive measures, a simple Bonferroni correction would require

probabilities ofp = 0.05/9 = 0.0056 to be labeled as significant. The implications of these

adjustments will be considered in the discussion chapter that follows.

While there were no significant between-subjects effects, there was an occasion-by-group

interaction for one cognitive measure (COWA), suggesting intervention group participants
































Figure 4-1 Continued.


Control
~000


S208

S158

: 108

S58

" 8


Baseline


Post-Test


Occasion


Baseline


Post-Test


Occasion









Table 4-10. Repeated measures multivariate analyses of variance for cognitive domains with study group as between-subjects factor
Effects Wilks' Lambda F Hypothesis df Error df P-value Partial j2
Non-Executive Measures
Between Subjects: Study Group 0.90 1.17 6 62 0.33 0.10
Within Subjects: Occasion 0.55 8.35 6 62 **0.00 0.45
Interaction: Study Group by Occasion 0.92 0.89 6 62 0.51 0.08

Executive Measuresa
Between Subjects: Study Group 0.97 0.23 9.00 59.00 0.99 0.03
Within Subjects: Occasion 0.65 3.59 9.00 59.00 **0.00 0.35
Interaction: Study Group by Occasion 0.86 1.09 9.00 59.00 0.38 0.14
Note: *p< .05; **p < .01; aThere were unequal variance-covariance matrices for these measures.










"I am now going to ask you a few questions about your recent exercise habits. To do this, I will
need to read to you a definition of what we mean by "regular exercise" so that we understand
each other. Are you ready to hear the definition?" [WAIT UNTIL PARTICIPANT SEEMS
ATTENTIVE AND READY TO LISTEN]

"Regular exercise is any planned voluntary physical activity (such as brisk walking, aerobics,
jogging, bicycling, swimming, basketball, etc.) performed to increase physical fitness. Such
activity should be performed 3 to 5 times per week for a minimum of 20 minutes per session.
Exercise does not have to be painful to be effective, but should be done at a level that increases
your breathing rate and causes you to break a sweat. Is this definition clear to you?" [IF YES,
CONTINUE. IF NO, CLARIFY ANY CONFUSIONS, PROBE FOR EXAMPLES OF
ACTIVITIES THEY SUGGEST]

(Record Persons report of Routine)







15a. Based on this definition, do you currently exercise regularly?

YES.......................................... ............. 1 = INELIG IBLE

N O ........................................ .....................2 G O TO Q U E STION #15b

[QUESTIONS #2 AND #3 ARE CODED AS ONE ITEM]

15b. Do you intend to begin exercising regularly?

YES........................................................1 GO TO QUESTION #15c

N O........................................... ...............2 SKIP QUESTION #15c

15c. Do you intend to begin exercising regularly in the next 30 days or the next 6 months?

Next 30 days........1 Next 6 months.......2


Determine eligibility here, before continuing. If ineligible, skip to last page and read
closeout script.







APPENDIX A
TELEPHONE SCREENING PROCEDURE
ELIGIBLE:
SCHEDULED:
INELIGIBLE:


THE AAMP


STUDY


TELEPHONE SCREENING PROCEDURE


This telephone screener includes (a) telephone consent, (b) exclusionary criteria,
(c) basic demographic questions, (d) the Telephone Interview of Cognitive Status
(TICS), and (e) the Exercise Staging Algorithm (ESA)



Participant ID Number:

Date of Screening:

Name of AAMP Researcher conducting the Screening:









moderate and vigorous physical activity, suggesting the sample improved over time;

nevertheless, there were no group-by-time interactions to suggest the intervention group

improved significantly more than controls. Cohen's d effect sizes for the significant effects of

time were small to medium (LTEQ, d= 0.51; MVPA, d= 0.51; mean pedometer steps, d= 0.22;

V02, d= 0.34). In this study, there was an average 7% increase in V02, and a 67% increase in

self-reported metabolic equivalents of activity (LTEQ). As a meta-analysis (reviewed below)

suggests, these V02 improvements did not reach the threshold of clinical significance. The more

improved self-reported measure has not been studied with regard to cognitive outcomes, thus the

clinical significance of the improvement on this measure is harder to judge.

Previous studies have shown that improvement in aerobic capacity by 11% was unrelated

to cognitive improvement above a practice effect (Madden et al., 1989); nevertheless, a 27%

increase in aerobic capacity was associated with significant cognitive improvement in a prior

study (Dustman et al., 1984). A more recent meta-analysis studying the aerobic capacity/fitness-

cognition link reviewed eleven randomized, controlled trials and found that aerobic fitness

interventions resulted in an approximate 14% increase in cardiorespiratory fitness (Angevaren et

al., 2008). This 14% increase was associated with significant improvement in cognitive

function: The largest effects on cognition were found for motor function (1.17), auditory

attention (0.52), and delayed memory functions (0.50). When compared to effect sizes of the

current cognitive variables, these published effects are generally much larger than those here.

Pre-post changes in self-efficacy, anxiety, and depression

With regard to psychosocial outcomes, for one self-efficacy measure (EXSE) there was a

between-subjects effect, suggesting the intervention group had significantly higher mean

exercise self-efficacy scores at both baseline and post-test; nevertheless, a within-subjects effect

for occasion on this measure indicated there was reduced self-efficacy over time for both groups.










Table 3-6. Schedule of intervention topics
Intervention Topic Homework Facilitator
week
1 Introduction to program and orientation to fitness Read Health Fitness


facility. Computation of target heart rate.

Getting to Know You; Goal Setting Part I

Goal Setting Part II

Defining exercise behavior: What is exercise? What is
physical activity? How much? How hard?


Types of exercise / exercise options
5 Barriers to Exercise



6 All About Mental Imagery
Define imagery
Practice general example of imagery
Practice exercise specific imagery
7 Discuss Exercise Imagery Questionnaire and progress
toward stated goals
8 Getting Good at Exercise accessing available
supports to help you improve in your exercises of
choice
9 Revisiting barriers and goals: discuss "pesky" barriers
and do status update on progress toward stated goals



10 Group discussion of exercise behavior and
accomplishments
11 Sustainability: Continuing progress and maintaining
success, accessing family/social support
12 Discuss maintenance goals, long-term versus short-
term goals. Review basics of goal-setting
(Exercise facility membership ends)
13 No meeting
14 Share future goals; PARTY
15 No meeting
16 No meeting
Note: (Giacobbi et al. 2008)


Benefits of
Physical Activity
What is exercise?
Activity
Write out your
goal
Read Mental
Imagery Article

Media Activity
3 ways to
overcome
barriers to
exercise
Fill out the
Exercise Imagery
Inventory


None


Identifying
Improvement
Resources
Apply one
suggestion for
overcoming
barriers) to
exercise
None


None


Write out your
goals for 3, 6 and
12 months


None


center
staff
Peer mentor

Peer mentor

Peer mentor


Peer mentor



Peer mentor


Peer mentor

Peer mentor


Peer mentor




Peer mentor

Peer mentor

Peer mentor


None
Peer mentor
None
None














Control



Intervention


Baseline


Intervention'


Post-Test


Occasion


C
Figure 4-2 Continued.


Control


Baseline


Post-Test


Occasion









As discussed later in the study limitations of this chapter, it may be the case that participants

became more realistic regarding their personal goals for physical activity change and felt less

confident in their ability to make such changes. Cohen's d effect sizes were small for these

significant occasion effects (Cohen, 1992; BSE, d= -0.12; EXSE, d= -0.29). There were no

group-by-occasion interactions. Examination of changes in anxiety and depression symptoms

indicated that there was only a multivariate effect of occasion and univariate within-subjects

effects for the STAI-Trait Anxiety and BDI-II Depression measures. Overall, the sample

improved in trait anxiety and depression symptoms over time; however these were small to

medium effects (Cohen, 1992; GDS, d= 0.01; BDI-II, d= 0.58; STAI State Anxiety, d= 0.13;

and STAI-Trait Anxiety, d= 0.41).

Given the results of these preliminary analyses, it is important to note that the core

assumption of this study's conceptual model was that improvements in physical activity, fitness,

and self-efficacy and affect would mediate or moderate any changes in cognitive, particularly

executive, function. Thus, the likelihood that the present study would observe an effect of the

intervention on cognitive improvements over time was greatly reduced. Despite this, the planned

analyses were completed to assess for any direct associations between the exercise promotion

intervention and cognitive variables.

Primary Analyses

Aim One

The goal of the first specific aim of the study was to examine the effect of the exercise

promotion intervention on cognitive performance. It was hypothesized that participants

receiving the exercise promotion intervention would show improved performance on cognitive

measures, particularly those assessing executive control processes, relative to control

participants. However, contrary to this hypothesis, the present findings indicated no significant









Table 4-5. Baseline correlations between psychosocial and cognitive variables
GDS BDI-2 STAI State Anxiety STAI Trait Anxiety BSE EXSE
NAART 0.185 0.086 -0.005 0.124 -0.130 -0.197
Trails A Time 0.146 *0.249 0.166 0.059 0.129 -0.001
LM Immediate Recall -0.017 -0.003 -0.139 0.001 *-0.307 *-0.248
LM Delayed Recall -0.033 -0.028 -0.144 -0.010 **-0.320 *-0.288
LM Recognition *-0.239 -0.187 *-0.253 *-0.251 -0.074 -0.055
LM Learning Slope -0.009 -0.117 0.050 -0.040 0.008 -0.048
COWA 0.080 -0.012 -0.002 0.125 -0.018 0.005
Trails B Time **0.341 **0.314 0.142 0.169 0.112 -0.065
Letter Number Sequencing 0.095 0.131 -0.031 0.066 0.023 -0.065
One-Back: Mean RT 0.008 0.017 0.022 -0.101 -0.063 -0.138
One-Back: Mean RT SD 0.023 0.014 -0.010 -0.031 0.074 -0.019
One-Back: Number Correct -0.036 -0.057 -0.136 -0.039 0.210 0.099
Two-Back: Mean RT -0.069 -0.005 0.099 -0.133 0.137 0.034
Two-Back: Mean RT SD 0.057 0.029 0.068 -0.101 0.174 0.052
Two-Back: Number Correct 0.024 -0.123 -0.071 -0.008 0.002 0.054
Note: *p < .05; **p < .01; LM = Logical Memory; NAART = North American Adult Reading Test; COWA = Controlled Oral Word
Association; RT = Reaction Time; SD = Standard Deviation; GDS = Geriatric Depression Scale; BDI-2 = Beck Depression
Inventory- Second Edition; STAI = State Trait Anxiety Inventory; BSE = Barriers Self Efficacy Scale; EXSE = Exercise Self-
Efficacy Scale









Table 4-22. Correlations between physical fitness, physical activity, and psychosocial change
scores
V02 LTEQ Mean Pedometer Steps MVPA
GDS *-0.25 0.05 -0.14 0.08
BDI-2 0.02 0.02 -0.06 -0.04
STAI State Anxiety -0.04 0.18 0.08 0.20
STAI Trait Anxiety -0.05 -0.11 0.14 -0.11
BSE 0.04 0.12 -0.06 0.08
EXSE -0.15 0.11 0.05 0.09
Note: *p < .05; **p < .01; GDS = Geriatric Depression Scale; BDI-2 = Beck
Depression Inventory- Second Edition; STAI = State Trait Anxiety
Inventory; BSE = Barriers Self Efficacy Scale; EXSE = Exercise Self-
Efficacy Scale; V02 = Modified Balke Submax; LTEQ = Leisure Time
Exercise Questionnaire









Primary Outcomes

North American Adult Reading Test (NAART)

The NAART (Blair & Spreen, 1989) was used to measure pre-morbid intelligence. It was

developed specifically for use with adults suspected to have compromised cognition. The

NAART consisted of 61 irregular, rare words that participants were asked to pronounce. The

NAART has been found to correlate between 0.40 and 0.80 with other measures of intelligence.

Test-retest reliability of the NAART has been established at 0.92 within one year. The score

used in was derived from using a prediction equation that took into account the number of words

incorrectly pronounced. The NAART was selected as a non-executive cognitive measure in the

present test battery because it is thought to be a measure of pre-morbid intellectual function (i.e.,

"crystallized" abilities), which would be expected to be relatively stable and less sensitive to

improvements in physical fitness over time.

Logical Memory subtest of the Wechsler Memory Scale, 3r Edition (WMS-III)

The Logical Memory subtest of the Wechsler Memory Scale-Third Edition (WMS-III)

(Wechsler, 1997) measured the ability to learn and retain verbal memories for brief stories.

Individuals heard a brief story of 25 propositions and were asked to recall as many story

propositions as possible in an immediate recall of the story. Participants received one point for

each proposition correctly recalled. Next, a second story of 25 propositions was read, followed

by an immediate story recall trial. This second story was read twice to assess verbal learning

slope. Delayed recall and recognition trials for each story were completed following a 25-35

minute delay interval. The LM subtest was selected as a measure of non-executive cognitive

function because it measures episodic memory abilities in a structured way. Relative to word list-

learning memory tasks, LM is thought to rely less on executive functioning, since individuals do









and avoidance of individual exercise barriers. A control/comparison group received 16 weeks of

"health hygiene" instruction, consisting of education about general health issues relevant to older

populations (i.e., osteoporosis, Alzheimer's disease, nutrition). There was one session devoted to

non-specific discussion of exercise/physical activity. Both groups received unlimited use/access

to a fitness facility; amount of exercise was monitored in both groups, such that the incremental

benefits of the treatment could be carefully assessed.

Participants

The sample of participants consisted of 69 middle-aged and older adults who were

sedentary and community-dwelling. Participants were recruited in the local Gainesville/Alachua

County, Florida area, primarily through the use of local newspaper advertisement in the

Gainesville Sun, Senior Times, and the Gainesville Regional Utilities (GRU) newsletter.

Additionally, flyers were posted at locations around the Gainesville and University of Florida

communities, and flyers were mailed to individuals on the University of Florida Older Adult

participant registry. To attempt to maximize recruitment of diverse populations, an ad was also

placed in the Gainesville Guardian, a publication targeted to the African-American community.

Finally, recruitment extended to community organizations, including a seniors' group a at local

community college, a local church, a retirement community affiliated with the university, a

grandparents support group, local health fairs for seniors, and a direct mailing list. Figure 3-1

outlines the number of participants that were initially contacted and/or expressed interest in the

study, screened by telephone, randomized to the study conditions, and included in the present

analyses.

Statistical Power Considerations

Table 3-1, displays the results of a power analysis conducted to examine the study

design's ability to detect effects with 32 participants per group, alpha = .05 (one-tailed;









patients. Their findings suggested that performance on neuropsychological measures was

improved with exercise; nonetheless, there was no control group comparison to determine

whether these patients' improvements were above mere practice effects. Also with such a small,

clinical sample, these findings are difficult to generalize to other groups. Another study included

in the Colcombe & Kramer (2003) meta-analysis with methodological concerns compared the

effects of exercise on cognitive performance in a group of depressed older adults (Khatri 2001).

This study compared depressed elders who performed 16 weeks of aerobic exercise to those

prescribed anti-depressant medication and found while both groups improved in depression

symptoms, only the exercise group showed improved performance on the Stroop Interference

task and Visual Reproductions of the Wechsler Memory Scale (WMS) (Khatri 2001). As with

the previous study, there were was no control for factors that would confound the results,

particularly the mere personal contact involved in being in a study, or practice/re-test effects.

Furthermore, in a study of patients with chronic obstructive lung disease (COPD), they examined

whether aerobic exercise would result in improved performance on cognitive and psychological

outcomes (Emery et al., 1998). These results showed that while patients did not improve lung

functioning with exercise, there were improvements in depression symptoms, though not more

than controls (Emery et al., 1998). Also, the only cognitive improvements above re-test effects

were found for verbal fluency performance, which is thought to have a strong frontal-executive

component (Emery et al., 1998). These three studies all contained clinical samples and were

include in the Colcombe & Kramer (2003) meta-analysis, along with 15 other studies of normal,

healthy older adults, which may obscure the clean examination of the true exercise-cognition

link.









4-13 Raw cognitive estimated marginal means by study group..............................................101

4-14 Calculation of reliable change index scores adjusted for control group practice effect..102

4-15 Participants who reliably changed on non-executive cognitive measures.......................103

4-16 Participants who reliably changed on executive cognitive measures ............................ 104

4-17 Multivariate analyses of variance for cognitive domains with study group and age
group as betw een-subjects factors .......................................... ............................... 106

4-18 Follow-up univariate analyses of variance for non-executive measures: Study group
and age group as between subjects factors ........................................... ............... 107

4-19 Follow-up univariate analyses of variance for executive measures: Study group and
age group as between subjects factors .......................... ........................ ............ ... 109

4-20 Raw cognitive estimated marginal means by age and study group ........... ....................112

4-21 Correlations between physical activity, physical fitness, and cognitive change scores..114

4-22 Correlations between physical fitness, physical activity, and psychosocial change
sc o re s................. .. .......... ................. ............................................... 1 1 5

4-23 Correlations between psychosocial and cognitive change scores.................................116









indicate inverse relationships between these sets of variables, such that improved leisure time

exercise was associated with decreased inconsistency in mean reaction time on the One-Back

task, and improved mean pedometer steps was related to reduced learning slope across testing

occasions. It is important to note that these significant relationships represent the minority of

those tested. For the 60 correlations tested between physical fitness/activity and cognition, only

3 (5%) reached significance.

Next, in correlating changes in physical fitness/activity and psychosocial variables, the

only significant association was between the GDS and VO2 (r = -0.25,p < .05). This inverse

relationship indicated that a decrease in GDS depression symptoms was associated with an

increase in cardiorespiratory fitness. Of the 24 correlations between physical fitness/activity and

self-perception, only 1 (4.2%) was significant. Finally, correlations between psychosocial

changes and cognitive score changes over time revealed three significant negative correlations

between changes in GDS depression and LM Delayed Recall (r = -0.24, p < .05), BDI-II

depression and Two-Back Mean RT SD (r = -0.26, p < .05), and BSE and Two-Back Number

Correct (r = -0.28, p < .05). These negative correlations indicated that reduced GDS depression

symptoms was associated with improved delayed recall of the LM stories over time, reduced

BDI-II depression symptoms was related to increased reaction time inconsistency on the Two-

Back task over time, and higher self-efficacy on the BSE was related to an improvement in the

number correct on the Two-Back task over time. As with fitness variables, only the minority of

associations tested reached the p < .05 criterion of significance. Of the 90 correlations tested

between self-perception and cognition, only 3 (3.3%) reached significance. It should also be

noted that, given the total number of correlations examined, if thep-values for these correlations

were Bonferroni-corrected, the critical value of alpha would bep = .05/176 = 0.00029.









years or 65 and older) (Table 3-4). For replicates 6-8, the groups were age-mixed in order to

meet the larger study's recruitment goals more quickly. Due to attrition of the sample, 35

control group participants and 34 intervention group participants were included in the present

analyses. A detailed description of the intervention is found in a later section. Finally, following

this period, participants underwent post-intervention assessment.

Table 3-5 outlines the pre-post assessment protocol (organized in a thematic grouping, not

test order). Executive cognitive measures, which are thought to be more sensitive to physical

fitness improvements (Colcombe & Kramer, 2003), are shown in the first row.

Rationale for Measures

Neuropsychological measures were selected to assess aspects of cognitive function

expected to be affected by physical exercise intervention, as well as those expected to remain

stable with exercise. Specifically, following the reviewed literature, executive control processes

are hypothesized to be most positively influenced by improved aerobic fitness (Colcombe &

Kramer, 2003); however, other cognitive functions, such as estimated intellectual ability and

memory for structured information (story memory), are hypothesized to be remain relatively

unchanged by improved aerobic fitness; these are included to help evaluate the putative

specificity of exercise effects. Psychosocial, exercise, and aerobic fitness measures were selected

to measure the physical and psychosocial components of exercise participation, as outlined by

the conceptual model above, that are potential mediating variables in the relationship between

exercise and cognitive function. The specific rationale for selecting each cognitive measure will

be described in the section that follows.

Detailed descriptions of each measure are as follows:










Schedule of control group topics


Control Topic Facilitator
week
1 Introduction to program and orientation to fitness facility; Getting to Fitness center
know you staff
2 Exercise Peer mentor
Why is exercise important?
How much exercise is too much?
What types of exercise are recommended?
3 Osteoporosis Peer mentor
What is osteoporosis?
What are the warning signs?
How do I prevent it?
4 Alzheimer's Disease Peer mentor
What do I look for?
What affects the development of Alzheimer's?
5 Cancer Screening Peer mentor
What types of cancer should I be concerned about?
How often should I be checked?
6 Hearing Loss Peer mentor
How common is hearing loss in older adults?
How do I know if I have hearing loss?
7 Arthritis Peer mentor
What are the different types of arthritis?
What can I do to treat arthritis?
8 Vision Loss Peer mentor
Is vision loss just part of getting older?
Is all vision loss the same?
9 Sleep Peer mentor
Do older adults need as much sleep as young adults?
How can I get a good night's sleep?
10 Balance Problems Peer mentor
What causes me to lose my balance?
Is there anything I can do to prevent it?
11 COPD* Peer mentor
What is COPD?
What type of exercise can I do if I have COPD?
12 Heart Failure* Peer mentor
How can I prevent heart failure?
Who is most at risk for heart failure?
(Exercise facility membership ends)
13 No meeting None
14 Share future goals; PARTY Peer mentor
15 No meeting None
16 No meeting None
Note: (Giacobbi et al. 2008). For Replicate 8, weeks 11 and 12 included discussion of diabetes
and high blood pressure health topics.


Table 3-8.









groups, participants experienced self-efficacy reductions over time. Cohen's d estimates of the

time effects were as follows: LTEQ, d= 0.51; MVPA, d= 0.51; mean pedometer steps, d= 0.22;

V02, d= 0.34; BSE, d= -0.12; EXSE, d= -0.29, with the latter two coefficients reflecting

poorer mean self-efficacy at post-test. It should be noted that all of these effects were small to

medium (Cohen, 1992). Examining the means in Figure 4-1, with a baseline V02 of 28, and an

average 2 point increase, the mean level fitness increase in this study represented only 7%. The

self-reported gain in metabolic equivalents, based on self-reported activity, was an increase of 4

points over a baseline value of 6, which represented a more sizeable 67% increase.

Anxiety and Depression: Pre-Post Changes

Since changes in psychosocial variables of anxiety and depression were hypothesized to

be one route by which changes in executive cognitive function would occur, these variables were

included in a repeated-measures MANOVA to determine intervention effects on these variables

(Table 4-8 and Figure 4-2). Again, the critical occasion-by-group interaction did not reach

significance (Wilks' A= 0.94; F(4, 64) = 1.00, p < 0.41; Partial r2 = 0.06). Turning to the main

effects, while there was no overall significant between subjects effect (Wilks' A = 0.97; F(4, 64)

= 0.41, p = .80; Partial r12 = 0.03), there was a significant overall within subjects effect of

occasion (Wilks' A= 0.71; F(4, 64) = 6.68, p < .001; Partial r12 = 0.29). Post-hoc univariate

ANOVAs (Table 4-9) on each anxiety and depression measure revealed significant reductions

over time for the STAI Trait Anxiety (F(1, 67) = 11.58; p < .001; Partial r2 = 0.15) and BDI-II

measures (F(1, 67) = 22.68; p < .001; Partial r2 = 0.25), suggesting that there were

improvements in anxiety and depression symptoms for the combined groups. Cohen's d

estimates of effect sizes for anxiety and depression within-subjects effects were small to medium









not have to impose a semantic organization on the information to maximize recall. It has been

found to be a reliable memory measure for use in adult and older adult groups (Wechsler, 1997).

Control Oral Word Association (COWA)

To assess spontaneous word production, the COWA (Benton & Hamsher, 1989) was

administered. Participants were read a letter ("F", "A", and "S") and asked to generate as many

non-proper nouns beginning with that letter as possible within sixty seconds. Participants were

not allowed to repeat previously stated words during each trial. The rationale for selecting

COWA as an executive cognitive measure was due to its sensitivity to frontal lobe/executive

function (Salthouse et al., 2003). Previous literature has identified the COWA as a "gold

standard" of executive function (Crawford et al, 2000 & 2005).

Trail Making Test A and B (Trails A, Trails B)

The Trail Making Test (Trails A and Trails B) (Reitan, 1992) assessed attention, working

memory, psychomotor speed, visual scanning and sequencing, and cognitive flexibility by

requiring individuals to connect circles containing numbers (Trails A) and numbers and letters

(Trails B). Trails B was conceptualized as a more demanding task, due to the increased

cognitive flexibility and executive skills required for successful task completion. The rationale

for inclusion of the Trail Making Test was due to both being one of the most widely used

measures in clinical neuropsychological practice (Rabin et al., 2005) and due to its documented

high sensitivity to brain function (Reitan & Wolfson, 1994). Trails A was categorized as a non-

executive measure because it is conceptualized as a measure of attention and processing speed

that is less reliant on working memory, as is Trails B. Trails B was selected as an executive

cognitive measure because of the added mental flexibility and working memory required to

complete this task (Lezak et al., 2004).









Table 4-19 Continued.
Hypothesis Error Partial
df df F P-value r2
Two-Back: Number Correct
Study Group 1 65 0.07 0.79 0.00
Age Group 1 65 2.12 0.15 0.03
Study Group by Age Group 1 65 0.48 0.49 0.01
Occasion 1 65 7.43 **0.01 0.10
Study Group by Occasion 1 65 0.83 0.37 0.01
Age Group by Occasion 1 65 0.20 0.65 0.00
Study Group by Age Group by Occasion 1 65 0.03 0.86 0.00
Note: *p < .05; **p < .01; COWA = Controlled Oral Word Association; RT = Reaction Time;
SD = Standard Deviation

































2008 Adrienne T. Aiken Morgan















-U Control Middle
Aged
- Control Older
Adult
Intervention
Middle Aged
g & Intervention Older
Adult


7.50
Baseline Post-Test

Age by Study Group
C

Figure 4-6 Continued.


1070

970

870

770

670

570


- Control Middle
Aged
Control Older
Adult
Intervention
Middle Aged
Intervention Older
Adult


Baseline Post-Test

Age by Study Group


12.50

11.50

10.50

9.50

8.50









examined this exercise-cognition link in older adults to understand these inconsistencies better.

The goal of the meta-analysis was to test several hypotheses have been proposed in the literature

regarding the specific effects of exercise: atheoretical- no specific hypotheses; speed hypothesis-

measures of speed of processing show the greatest improve with exercise; visuospatial

hypothesis- these measures show the greatest improvements with exercise; controlled processing

hypothesis- these sustained attentional tasks show the greatest improvements; and executive

control hypothesis- working memory, problem solving, and inhibition type tasks show the most

exercise-related gains (Colcombe & Kramer 2003). Results from the meta-analysis showed that

indeed, exercise was related to the improvements in cognitive performance, with executive

control processes showing the greatest exercise related improvement (Colcombe & Kramer

2003). Women and older participants showed the most cognitive gains, as combined aerobic and

strength training exercise programs showed the most benefits to participants (Colcombe &

Kramer 2003).

Nonetheless, there are several methodological problems with this meta-analytic study.

First, the meta-analysis included 18 studies that had well over 50 different

neuropsychological/cognitive measures. The authors allowed for multiple cognitive domain

categorizations along the four major hypothesized cognitive domains (i.e., one measure could be

considered a speed, visuospatial, and executive tasks all at once) (Colcombe & Kramer 2003).

While they control for the influence of executive control processing on performance on other

domains, they fail to control for the influence of the other three domains on executive control

processing. This meta-analysis also suffered from including some studies that were flawed.

Palleschi et al (1996) examined the effect of 12 weeks of exercise on cognitive performance

(MMSE, attentional matrix, verbal span, supraverbal span tests) in a group of 15 Alzheimer's









education, and many involved samples of people with more severe cardiovascular disease, and

not just hypertension (Waldstein et al., 1991). More recent studies with fewer methodological

flaws have come to show relationships between hypertension and cognitive decline, but these

findings have often been inconsistent (Waldstein et al., 1991). While many studies have found

high blood pressure and cognition to be related, others have found low blood pressure to be more

related to cognition, or no relationship between blood pressure and cognition at all (Waldstein et

al., 1991). Additionally, there have been inconsistencies with regard to the various

neuropsychological domains most affected by high blood pressure; for instance, some studies

find relationships between memory or attention and blood pressure, while others may not

(Verghaegen et al., 2002; Madden et al., 2003). Furthermore, studies have found various

patterns of neuropsychological and cognitive deficits, such that relative effects of blood pressure

on neuropsychological domains are difficult to discern (Waldstein et al., 1991). Overall,

relationships between blood pressure and cognition have been found cross-sectionally and

longitudinally, while chronicity of hypertension seems to be an important predictor

longitudinally (Waldstein et al., 1991). Specifically, much research suggests high blood pressure

negatively affects cognitive function (Deary et al., 1998; Elias et al., 2003; Elias et al., 2004;

Starr & Whalley, 2005; Waldstein et al., 1991; Waldstein & Katzel, 2001; Waldstein et al., 2005;

Nilsson et al., 2004), with the effects of hypertension exacerbated by other factors, including

APOE 4 allele status (Carmelli et. al., 1998) and obesity and hyperglycemia (Elias et al., 2003).

Across various studies, a common finding is the U-shaped effect of blood pressure, with

regard to actual blood pressure levels and age (Carmelli, 1998; Launer, 2000; den Heijer, 2002;

Madden, 2003). High blood pressure is associated with worse performance across various

cognitive domains (i.e., attention, executive functions, memory and learning, processing speed,









Table 4-1 Continued.
Overall Completers Dropouts Df t/X2 p-value
n= 90 n = 69 n= 21
One-Back 361.1 (313.1) 330.3 (224.9) 462.4 (500.4) 88 1.712 0.090
Mean RT
SD
One-Back 96.1 (12.4) 95.5 (14.1) 98.0 (3.3) 88 0.799 0.427
Number
Correct
aTwo-Back 1521.2 (796.7) 1756.0 (748.2) 749.9 (325.7) 77.2 -8.769 <.001
Mean RT
aTwo-Back 1058.9 (713.3) 912.8 (538.1) 1538.9 (981.7) 23.8 2.797 0.010
Mean RT
SD
aTwo-Back 71.8 (34.8) 89.6 (14.3) 13.3 (3.3) 85 -40.843 <.001
Number
Correct
Note: Mean (Standard Deviation). aThere were unequal variances for these measures. LM =
Logical Memory; NAART = North American Adult Reading Test; COWA = Controlled Oral
Word Association; RT = Reaction Time; SD = Standard Deviation









executive processes (Luszcz & Lane, 2008). Nonetheless, verbal fluency (COWA) has been

identified by previous studies as a "gold standard" measure of executive function, due to its

sensitivity to frontal lobe function (Crawford et al., 2000 & 2005; Salthouse et al., 2003). The

present study's findings that COWA was most sensitive to the intervention effect (albeit very

modestly) are consistent with this previous work.

This lack of consensus in conceptualization and assessment of executive cognitive function

likely contributes to the inconsistencies in the extant literature regarding the relationship between

physical/aerobic fitness and cognitive function. To date, the studies that have concluded that

executive cognitive function is disproportionately affected by improvements in

cardiovascular/physical fitness have examined predominately one type of executive function:

executive control processes or working memory (Colcombe & Kramer, 2003). However,

abstract reasoning, conceptualization, and problem-solving are also considered executive

functions that are sub-served by frontal lobe-subcortical connections in the brain via white matter

tracts, which are disproportionately affected by both normal brain aging (e.g., Sullivan &

Pfefferbaum, 2006) and are particularly sensitive to cardiovascular and cerebrovascular health

declines (e.g., van Boxtel et al., 2006). Nevertheless, little attention in the literature has been

given to these other types of executive cognitive functions, which tend to depend on

conceptualization and abstraction and less on one's ability to mentally hold and manipulate

information or rapidly shift/alternate between competing mental sets, as do working memory

tasks. This distinction is important because working memory tasks rely more on intact

attentional abilities, and these tasks often have a speed component to them (such as the Trail

Making Test and N-back task). Executive measures of conceptualization and abstract

reasoning/problem-solving abilities do not require the same level of attentional control and future









and silent lacunar infarcts often occur in the deep and periventricular white matter brain regions,

leading to observable executive control deficits (Libon et al., 2004).

Raz and colleagues (2003) and den Heijer et al. (2003) have demonstrated the relationship

between high blood pressure and pathological brain changes; in their work, hypertension was

related to increased brain atrophy. Even further, Raz et al. (2003) found atrophy of the prefrontal

brain regions to be most related to high blood pressure, while den Heijer et al. (2003) showed

there was a U-shaped relationship between hypertension and brain atrophy. Thus, both high and

low blood pressure levels were both found to be associated with disproportionate amounts of

brain atrophy (den Heijer, 2003). Pantoni (1999) has argued evidence demonstrates vascular

cognitive impairment tends to be more frequent in individuals with white matter lesions, as well

as in those with cerebral circulatory dysfunction. Similarly, Waldstein and Katzel (2001)

reported evidence suggesting the effects of decreased cerebral blood flow, related to increased

peripheral vascular resistance in blood vessels in the body, may also be an underlying

mechanism in this relationship. It has also been reported that increased activity of both the

sympathetic nervous system and hypothalamic-pituitary-adrenocortical axis is involved in stress

responses to environmental challenges, and this appears to play a role in the hypertension-

cognition relationship (Waldstein & Katzel, 2001).

As was explored further in the present study, these physiological brain changes have direct

implications for cognitive function and declines mediated by these brain areas, which explain the

pattern of neuropsychological deficits often observed (i.e., deficits in executive function,

processing speed and attention, more so than deficits in memory and language) (Pantoni, 1999;

Deary et al., 2003; Waldstein & Katzel, 2001; Raz, 2003; den Heijer, 2003; Kramer, 2001).















U Control Middle
Aged
d Control Older
Adult
- intervention
Middle Aged
Oq intervention
Older Adult


125.00

105.00

85.00

65.00

45.00

25.00


Baseline Post-Test

Age by Study Group


- Control Middle
Aged
- Control Older
Adult
- Intervention
Middle Aged
- intervention Older
Adult


Baseline Post-Test

Age by Study Group


A
Figure 4-6. Mean cognitive scores by study group, age group, and occasion for executive cognitive variables A) COWA. B) Trails B.
C) Letter-Number Sequencing. D) One-Back Mean RT. E) One-Back Mean RT SD. F) One-Back Number Correct. G)
Two-Back Mean RT. H) Two-Back Mean RT SD. I) Two-Back Number Correct.


52.00


S47.00

42.00

" 37.00

32.00


27.00









neuropsychological, psychological, and exercise/fitness testing. After randomization to either

the control (health hygiene) or experimental (exercise promotion) condition, weeks three through

eighteen (sixteen weeks total) of the study involved the intervention period. However, it is

important to note that several factors determined the actual study timeline for each participant.

The actual study timeline for each participant was influenced by his or her availability to devote

sixteen weeks to participation in a group (i.e., schedule conflicts, vacation, etc), the status of

participant recruitment (all the available slots of each group needed to be filled to maximize use

of resources), the status of required completion and receipt of physician's permission form, and

the ability of study staff to screen participants in a timely manner. Due to these factors, there

was a lag of several months between initial participant contact, completion of telephone

screening, and/or baseline testing and the start of the study group period for most participants.

Specifically, there was an average of 5.6 weeks between baseline testing and the start of the

small group sessions (range 0 to 26.6 weeks). After the 16 week intervention period, most (66

out of 69) participants were post-tested within one or two weeks following the end of the group

sessions. The remaining 3 participants were tested longer than two weeks post-intervention due

to scheduling conflicts (range 3 to 6.4 weeks).

During the intervention period, all participants received a free membership to either a

University of Florida campus fitness facility or a community-based, church fitness center. There

were a total of eight replicates of peer/support groups, with one control and intervention group

each per replicate for a total of sixteen small peer/support groups. In total, 47 participants were

randomly assigned to the control condition, while 44 participants were assigned to the

intervention condition. In addition, study groups for replicates 1-5 were composed of people

randomly assigned to peer groups including other individuals in their same age range (50-64









* To investigate whether, relative to matched non-exercising control participants, sedentary
adults receiving a physical exercise promotion intervention experience improvements in
the primary outcome of cognitive function (particularly executive control processes).

Hypothesis: As shown in previous studies, participants receiving the exercise intervention

will show improved performance on cognitive measures, particularly those assessing executive

control processes, relative to control participants.

* To determine the separate and joint roles of improvement in proximal outcomes (fitness,
activity, and affect) in mediating exercise intervention effects on cognition.

Hypothesis: It is further expected that changes in measures of physical fitness, physical

activity engagement, and emotional well-being mediates some or all of the intervention effect on

cognitive change.









Cognitive Exclusions

All potential participants were screened by telephone to exclude individuals based on the

following cognitive criteria: dementing illness, lifetime history of significant head injury

requiring hospitalization, other neurological or major medical illnesses (Parkinson's disease,

epilepsy, stroke, cancer (and/or chemotherapy and radiation above the chest)), severe

uncorrected vision or hearing impairments, inpatient psychiatric treatment, extensive drug or

alcohol abuse, any use of an anticholinesterase inhibitor (such as Aricept), or unavailability at

future follow-up time points. Telephone screening included the 11-item Telephone Interview for

Cognitive Status (TICS; Brandt, Spencer, & Folstein, 1988) for a standardized assessment of

cognitive status. The TICS has a sensitively of 94% and a specificity of 100%. The cut-off

score of 27 points was used to exclude demented individuals from the study (Brandt, Spencer &

Folstein). The TICS, embedded in the general telephone screening protocol, is included in

Appendix A.

Physician's Permission

To assure individuals were properly excluded due to cardiovascular/physical

conditions/diseases and medications presented below, participants were required to submit a

completed and signed nurse/physician's checklist and permission form prior to enrolling in the

study. This checklist is included in Appendix B.

Physical/Cardiovascular Exclusions

Exclusion criteria for diseases or conditions likely to adversely affect the safety of elders

in the exercise promotion intervention are as follows: terminal illness with life expectancy less

than 12 months, cardiovascular disease (myocardial infarction in last 6 months, chronic heart

failure, aortic stenosis, history of cardiac arrest, implanted cardiac defibrillator, or uncontrolled

angina), pulmonary disease requiring oxygen or steroid treatment, and ambulation with assistive















U Control Middle
Aged
I Control Older
Adult
- Intervention
Middle Aged
Intervention
Older Adult


1505

1305

1105

905

705

505

305


-0 Control Middle
Aged
- Control Older
Adult
a* intervention
Middle Aged
- Intervention
Older Adult


Baseline Post-Test

Age by Study Group


Figure 4-6 Continued.


2340
2140
1940
1740
1540
1340
1140
940
740


Baseline Post-Test

Age by Study Group









"predispose, precipitate, and perpetuate" depression in the elderly and that neuropathology found

in the white matter regions of the brain can be an etiological factor in the expression of

depression (Alexopoulos et al., 1997; Alexopoulos, 2006). One investigation showed, through

path modeling, a stronger relationship between cerebrovascular risk factors (i.e., hypertension,

diabetes, and heart disease) and depression symptoms in a group of oldest-old (85 and older),

after controlling for co-morbid health/medical conditions and limitations (Mast et al., 2005). Co-

morbid health/medical conditions and limitations mediated the relationship between

cerebrovascular risk factors and depression in individuals age 50-84 in this sample (Mast et al.,

2005). Additionally, a second study examining the role of cardiovascular risk factors found that

among the oldest-old living in a retirement community, both depression and number of

cardiovascular risk factors at baseline predicted stroke (Krishnan et al. 2005). Depression

accounted for twelve percent of the variance in stroke incidence and partially moderated the

effect of cardiovascular risk factors (Krishnan et al., 2005). Furthermore, studies have shown

that white matter lesion load significantly predicts depression in older samples (e.g., Godin et al.

2008; Krishnan et al. 2004). In one cross-sectional study, white matter lesion volume was

significantly associated with lifetime of major depression after controlling for covariates (sex,

age, hypertension, cardiovascular disease, and alcohol and tobacco consumption) (Godin et al.

2008). Furthermore, in another cross-sectional study investigating the strength of association

between various vascular-related neuropathology linking to depression (i.e., cerebral white

matter hyperintensities signaling periventricular and deep white matter lesions), the findings

demonstrated that deep white mater lesions was more strongly correlated with depression

symptoms as measured by the Geriatric Depression Scale (15-item Short-Form) (Krishnan et al.,

2006).









by age (Wilks' A= 0.94; F(4, 60) = 0.63, p = .70; Partial r2 = 0.06). There was also no study

group-by-age interaction (Wilks' A= 0.90; F(6, 60) =1.17, p = .33; Partial rl2 = 0.10). In

addition, overall study group (Wilks' A= 0.89; F(6, 60) = 1.20, p = .32; Partial rl2 = 0.11) and

overall age group effects (Wilks' A= 0.83; F(6, 60) = 2.09, p = 0.07; Partial rl2 = 0.17) did not

reach significance. Although, it is worth nothing that the age group effect approached our

criterion of significance (p = .05). As with the initial multivariate analyses for study aim one,

there was a significant within-subjects effect (Wilks' A= 0.55; F(6, 60) = 8.11,p <.001; Partial

2 = 0.45).

Next, for the executive cognitive measures there were no significant multivariate occasion-

by-study group (Wilks' A= 0.86; F(9, 57) = 1.01,p = 0.44; Partial r2 = 0.14), occasion-by-age

group (Wilks' A= 0.89; F(9, 57) = 0.80, p = 0.62; Partial r2 = 0.11), or occasion-by-study group-

by-age group (Wilks' A= 0.83; F(9, 57) = 1.34,p = 0.24; Partial r2 = 0.17) interactions. A study

group-by-age group interaction effect approached significance (Wilks' A= 0.77; F(9, 57) = 1.91,

p = 0.07; Partial r2 = 0.23). While there was no effect of study group (Wilks' A= 0.96; F(9, 57)

= 0.27, p = .98; Partial r2 = 0.04), there was a significant age group effect (Wilks' A= 0.73; F(9,

57) = 2.33,p = 0.03; Partial r2 = 0.27). Additionally, there was a significant within-subjects

effect for occasion (Wilks' A= 0.64; F(9, 57) = 3.58, p < .001; Partial r12 = 0.36), suggesting that

overall, there was an effect of time on performance.

Follow-up univariate ANOVAs are presented in Tables 4-16 and 4-17, and estimated

marginal means are displayed in Table 4-18. The only non-executive cognitive variable showing

an age group effect was Trails A Time, indicating that the younger age group (50-64 year olds)

performed this task significantly faster than the 65 and older group (score was number of

seconds). The non-executive variables showing significant within-subjects effects included the










CONTROL GROUP


Mentor/Coach


Procedure
GROUP MANAGEMENT
1. Begin sessions with discussion about last week's
topic, including quiz.
2. Keeps group on topic
3. Manages time and pace of discussion
4. Maintains leadership of discussion

PRESENTATIONAL SKILLS
5. Use of Open-ended questions
6. Clearly presents topic to be discussed
7. Promotes discussion by asking questions
8. Actively makes efforts to include all members of
the group into discussion.

COMMUNICATION ROADBLOCKS
9. Avoids Lecturing
10. Avoids giving advice
11. Avoids interpreting or analyzing
12. Avoids questioning participant

CONCLUSION
13. Assures participant that all instructions are in the
Workbook & reminds to bring all logs and complete
any homework for next session.
14. Makes appropriate referrals regarding questions
participants may have about the study.
15. Makes appropriate referrals regarding mental or
physical health concerns observed during sessions.


None Part Full Score

0 5 10


0 5 10


0 5 10

0 5 10


Comments for Mentor:


Date


Total









Table 4-15. Participants who reliably changed on non-executive cognitive measures
Control (n= 35) Intervention (n = 34) Total (n = 69)


NAART
Reliable Decline
No Reliable Change
Reliable Improvement

LM Immediate Recall
No Reliable Change
Reliable Improvement

LM Delayed Recall
Reliable Decline
No Reliable Change
Reliable Improvement

LM Learning Slope
Reliable Decline
No Reliable Change

LM Recognition
Reliable Decline
No Reliable Change
Reliable Improvement

Trails A Time
Reliable Decline
No Reliable Change
Reliable Improvement


2 (5.7%)
31 (88.6%)
2 (5.7%)


25(71.4%)
10 (28.6%)


3 (8.6%)
31 (88.6%)
1 (2.9%)


13 (37.1%)
22 (62.9%)


3 (8.6%)
29 (82.9%)
3 (8.6%)


3 (8.6%)
31 (88.6%)
1 (2.9%)


1 (2.9%)
32(94.1%)
1 (2.9%)


27 (79.4%)
7 (20.6%)


0 (0%)
33 (97.1%)
1 (2.9%)


14(41.2%)
20 (58.8%)


2 (5.9%)
31 (91.2%)
1 (2.9%)


6 (17.6%)
27 (79.4%)
1 (2.9%)


Note: NAART = North American Adult Reading Test; LM


3 (4.3%)
63 (91.3%)
3 (4.3%)


52 (75.4%)
17 (24.6%)


3 (4.3%)
64 (92.8%)
2 (2.9%)


27(39.1%)
42 (60.9%)


5 (7.2%)
60 (87.0%)
4 (5.8%)


9 (13%)
58(84.1%)
2 (2.9%)
Logical Memory









Table 4-19. Follow-up univariate analyses of variance for executive measures: Study group and
age group as between subjects factors
Hypothesis Error P- Partial
Effects df df F value ?2
COWA
Study Group 1 65 0.04 0.84 0.00
Age Group 1 65 0.15 0.70 0.00
Study Group by Age Group 1 65 0.50 0.48 0.01
Occasion 1 65 7.00 *0.01 0.10
Study Group by Occasion 1 65 4.03 *0.05 0.06
Age Group by Occasion 1 65 0.51 0.48 0.01
Study Group by Age Group by Occasion 1 65 0.23 0.64 0.00
Trails B Time
Study Group 1 65 0.57 0.45 0.01
Age Group 1 65 3.46 0.07 0.05
Study Group by Age Group 1 65 1.93 0.17 0.03
Occasion 1 65 0.92 0.34 0.01
Study Group by Occasion 1 65 0.20 0.65 0.00
Age Group by Occasion 1 65 0.29 0.59 0.00
Study Group by Age Group by Occasion 1 65 5.93 *0.02 0.08
Letter-Number Sequencing
Study Group 1 65 0.16 0.69 0.00
Age Group 1 65 0.86 0.36 0.01
Study Group by Age Group 1 65 0.19 0.67 0.00
Occasion 1 65 0.10 0.75 0.00
Study Group by Occasion 1 65 1.39 0.24 0.02
Age Group by Occasion 1 65 0.33 0.57 0.01
Study Group by Age Group by Occasion 1 65 0.01 0.94 0.00
One-Back: Mean RT
Study Group 1 65 1.17 0.28 0.02
Age Group 1 65 6.54 *0.01 0.09
Study Group by Age Group 1 65 1.19 0.28 0.02
Occasion 1 65 11.21 **0.00 0.15
Study Group by Occasion 1 65 0.57 0.45 0.01
Age Group by Occasion 1 65 5.45 *0.02 0.08
Study Group by Age Group by Occasion 1 65 2.26 0.14 0.03
Note: *p < .05; **p < .01; COWA = Controlled Oral Word Association; RT = Reaction Time;
SD = Standard Deviation









Conceptual model

The current study sought to confirm the effects of exercise on cognition using a more

individualized approach to exercise intervention than many previous studies (this is described in

greater detail below). In addition to examining exercise effects on cognition, the study also

explored several possible mediators of exercise effects, including physical factors (changes in

physical fitness) and psychosocial factors (changes in self-perception, depression, well-being,

and self-efficacy). Thus, the conceptual model below guides the current study (Figure 2-1). As

noted above, exercising may also be construed as a form of complex activity, which may

increase social engagement. Many previous studies have failed to control for these social

participation components of exercise, using wait-list or no-contact control groups. The present

study included a "health hygiene" control group, matched with experimental participants in study

contact hours and interaction with study staff and other participants and in exposure to cognitive

testing. These factors were designed to reduce the confounding role of social/complex activity

participation as a plausible alternative explanation of exercise effects.

Significance

The proposed study sought to further the understanding of the relationship between

cognition and exercise in a middle-aged and older adult sample. It further attempted to explore

the unique and combined contributions of fitness and self-perception changes as mediators of

exercise effects on cognition. This is unique in that, to our knowledge, few studies have

attempted to examine both process and outcome exercise-related variables as routes to cognitive

improvement in an older adult population. Furthermore, this study was thought to be the first in

a line of future research aimed at exploring the effect of one potentially remediable antecedent of

late life cognitive decline, physical fitness.










39.00
119.00 37.00
117.00 -Control T 35.00
Control
S 115.0 0 -. 33. 00 -
.~ 113.00 .-
Itvt 29.00
111.00 Interventio 29.00 Intervention
1110 27.00
109.00 25.00
107.00 23.00
Baseline Post-Test Baseline Post-Test
Occasion Occasion
A B

Figure 4-3. Mean cognitive scores by study group and occasion for non-executive cognitive variables A) NAART. B) Trails A. C) LM
Immediate Recall. D) LM Delayed Recall. E) LM Recognition. F) LM Learning Slope.









final state. In other words, helping to more actively shape the goals of participants seems a

likely ingredient for higher success. Consequently, an intervention that is more behavioral in

nature, with individualized targets more clearly defined a priori, may be more suitable for older

groups.

Selection of the correct follow-up interval

Next, as mentioned above, the present exercise promotion was designed with long-term

behavioral lifestyle outcomes in mind. As such, it may be the case that our lack of significant

findings at post-test may not tell the whole story. Longer-term (e.g., 6-, 12-, or 18-month)

longitudinal follow up may be necessary to detect a delayed effect of the intervention. A true

test of the effectiveness of the intervention may be its effect at long-term follow up, rather than at

an immediate post-test interval.

One hypothesis to consider is that intervention effects might cumulate. Specifically, the

present study tried to introduce health habits and personal goal setting that would lead to an

altered fitness lifestyle. If successful, the intervention would produce small incremental gains

that would continue long after the study is completed. (This is admittedly idealistic; most

follow-ups of exercise studies find that effects dissipate after cessation of treatment).

Another hypothesis is that intervention effects are delayed because they are revealed

downstream, with a separation of the decline trajectories of those with and without the

intervention. There is a recent example of such a delayed effect, albeit in the cognitive domain.

In the ACTIVE clinical trial of cognitive interventions for older adults (Willis et al., 2006),

ACTIVE study investigators did not find a transfer effect of cognitive (i.e., memory, reasoning,

and speed) training on daily function (i.e., IADLs, self-ratings, everyday problem solving) until a

5-year follow-up assessment. Two reasons cited for this delayed intervention effect were

previous work suggesting a lag between the onset of cognitive decline and the onset of functional









and Delayed recall trials (0.605 and 0.668, respectively), while the smallest were for Letter-

Number Sequencing and Trails B (0.049 and 0.087, respectively).

As a follow-up to these analyses, reliable change index scores (adjusted for the practice

effect demonstrated by the control group) were calculated to examine intraindividual trajectories

of change. The goal was to determine the proportion of the sample that experienced sizable

gains above the typical effect of practice. Results of these analyses indicated little evidence for

reliable improvement for either the control and intervention groups. The LM Immediate Recall

variable showed the largest percentage of the sample experiencing improvements across time

(24.6% of the sample); however, relative to controls, there were fewer intervention participants

who reliably improved (28.6% of controls improved versus 20.6% of intervention participants),

although this difference (nor any other) was significantly different between study groups. This

suggested little evidence for reliable cognitive improvement, not due to change alone, for our

sample.

Next, as additional follow-up analyses, the repeated-measures MANOVAs described

above were re-run with age group as a second between-subjects factor. This permitted analysis

of whether age moderated intervention effects (did one age group improve more than the other?).

Further, it allowed exploration of potential two-way (study group-by-age group) and three-way

(study group-by-age group-by-occasion) interactions. At the multivariate level, there were no

significant effects of age group for non-executive measures, though this effect did approach

significance. There were also no two-way or three-way interactions involving age group.

Univariate analyses demonstrated an age group effect for Trails A Time (the younger age group

was significantly faster on this task), as well as study group-by-age group interaction effect for

LM Delayed Recall (older individuals in the intervention group performed significantly better









efficacy). This may have been in part due to the tapering of group sessions towards the end of

the 16-week intervention period.

Interestingly, the current study drop-out rate was higher than that of various previous

physical activity/exercise training studies (reporting attrition rates), which had retention rates

ranging from 83 to 92 percent (Madden et al., 1989; Blumenthal et al. 1991; Hawkins et al.,

1992; Emery et al., 1998; Elavskey et al., 2005; Motl et al., 2005; Elavsky & McAuley, 2007).

In these studies, participants were either highly motivated to participate (Madden et al., 1989;

Blumenthal et al. 1991) or the interventions employed a structured exercise training protocol that

followed American College of Sports Medicine (ACSM) guidelines for physical activity

intensity and duration minimums to improve self-efficacy, self-esteem, and quality of life.

Employing more structure in the exercise promotion lifestyle intervention may have helped

participants to feel more confident in the physical activity changes they were making. Using

ACSM guidelines also may have ensured sufficient exercise behavior changes to effect physical

fitness at the magnitude needed to impact cognitive function significantly. Furthermore, as

mentioned previously, participants that dropped out of the present study differed significantly

form those completing the study only in self-reported depression and anxiety symptoms.

Perhaps, initially intervening upon these symptoms would have resulted in a higher rate of

retention. Finally, having more financial resources to recruit and retain participants may have

resulted in a higher final sample size. Katula and colleagues (2007) used numerous methods

(brochures, newspaper, radio, television, etc.) to recruit participants during the course of a four-

site randomized, controlled pilot study of the benefits of a physical activity intervention for

immobility prevention in older adults. In the end, the study successfully randomized 424 older

adults (a 13.5% recruitment yield) to two intervention conditions after spending approximately









TAKE A PAUSE WHILE YOU BRIEFLY ASSESS THE FOLLOWING TWO ITEMS.


INTERVIEWER ASSESSMENT OF PARTICIPANT COMMUNICATION

USING THE SCORING CRITERIA ON THE NEXT PAGE, CODE YOUR ASSESSMENT OF
PARTICIPANT'S ABILITY TO MAKE SELF UNDERSTOOD AND TO UNDERSTAND
OTHERS. THESE JUDGMENTS CAN BE MADE BOTH ON THE BASIS OF COGNITIVE
UNDERSTANDING, AND ALSO OF ENGLISH-AS-SECOND-LANGUAGE ISSUES.

21. MAKING SELF UNDERSTOOD

U N D E R ST O O D ........................................ .....................0

USUALLY UNDERSTOOD ...............................................1
(DIFFICULTY FINDING WORDS OR FINISHING
THOUGHTS.)

SOMETIMES UNDERSTOOD 2= INELIGIBLE
(ABILITY IS LIMITED TO MAKING CONCRETE
REQUESTS I

RARELY NEVER UNDERSTOOD 3 = INELIGIBLE






22. ABILITY TO UNDERSTAND OTHERS

U N D E R ST A N D S ........................................ .....................0

USUALLY UNDERSTANDS..........................................1...
(MAY MISS SOME PART/INTENT OF MESSAGE.)

SOMETIMES UNDERSTANDS 2 = INELIGIBLE
(RESPONDS ADEQUATELY ONLY TO SIMPLE.
DIRECT COMNIIUNICATION)

RARELY NEVER UNDERSTANDS 3 = INELIGIBLE


MAKING SELF UNDERSTOOD SCORING


0 = Understood: The participant expresses ideas clearly.
1 = Usually Understood: The participant has difficulty finding the right words or finishing









age groups had similar mean years of education and percent gender, and when follow up

exploratory analyses controlled for age group assignment, the overall pattern of results remained

unchanged. In fact, the inclusion of the age group variable revealed significant relationships that

would have otherwise gone unobserved. Specifically, the three-way interaction showing greater

improvements in performance over time on the Trails B task for younger intervention group

participants was an interesting finding and one of few significant study findings that was in the

expected direction. This finding was considered a small effect (Partial q2= 0.08) and was no

longer significant with Bonferroni correction; however, this trend is consistent with study

hypotheses.

Retention issues

A second study limitation involves retention of the sample. The larger, 16-week

intervention study, in which the present study is embedded, required a considerable time

commitment from participants, which made recruitment and retention of participants

challenging. Of the 433 participants identified through various recruitment methods, 90 were

randomized to a study group and pre-tested. This amounted to a 20.9% recruitment yield. While

only twelve participants dropped out of the study during the intervention period, there were nine

participants who underwent baseline neuropsychological testing and did not attend any study

group sessions. Had the study retained these 21 participants that were randomized and pre-

tested, but not post-tested, this would have equated to roughly a 30% increase in the present

sample's numbers. Other retention issues involve the collection of complete data for each

participant. As noted in the Methods chapter, there was a substantial amount of missing physical

activity and self-efficacy data towards the end of the study, due to participant non-adherence to

study protocols (i.e., completing daily and weekly questionnaires on physical activity and self













99.00
t
Control
S94.00


89.00 Intervention


84.00
Baseline Post-Test

Occasion

Figure 4-4. Continued
Figure 4-4. Continued









improved significantly more on this measure than did controls over time (F(1, 67) = 4.15; p <

.046; Partial r2 = 0.06); Table 4-12 and Figure 4-4). To examine this further, a pre-post COWA

mean change score was computed for each participant and used as the dependent variable in a

follow-up, one-way analysis of variance to examine whether there was a significant group

difference in the magnitude of change on this measure. The results indicated a significant

difference in the magnitude of mean change in COWA scores over time, with the intervention

group demonstrating greater improvement in COWA scores over time than controls (mean (SD)

change for intervention group: 4.44 (6.46) ; for control group: 0.58 (9.02); F(1, 67) = 4.15;p =

.046, Partial r2 =0.06).

Next, there were within-subjects effects of occasion for several cognitive variables. Tables

4-10 and 4-11 show F- statistics, degrees of freedom, and significance values for the within-

subjects effects of each repeated-measures ANOVA. These cognitive variables include NAART,

Trails A, and LM Immediate and Delayed recall (non-executive measures) and COWA, One-

Back Mean RT, Two-Back Mean RT, and Two-Back Mean RT SD (executive measures).

Estimated marginal means for each group are presented in Table 4-13, and Figures 4-3 and 4-4

display line graphs of estimated marginal means by group. Overall, there were significant

improvements on these cognitive variables for the entire sample, regardless of study group, over

time. These practice/occasion effects were of small to medium size for most of the cognitive

variables (Cohen, 1992). Cohen's d estimates of effect were as follows: NAART, d= 0.40;

Trails A, d= 0.32; LM Immediate Recall, d= 0.61; LM Delayed Recall, d= 0.67; LM

Recognition, d= 0.27; LM Learning Slope, d= 0.13; COWA, d= 0.31; Trails B, d=0.09; Letter

Number Sequencing, d= 0.05; One-Back Mean RT, d= 0.41; One-Back Mean RT SD, d= 0.49;









etc.) in small groups with a peer mentor, while a control/comparison group received 16 weeks of

"health hygiene" instruction, consisting of 16 weeks of education about general health conditions

in aging (also in small groups with a peer mentor). Repeated-measures MANOVA indicated no

significant between-subjects effect of the intervention (p >.05). There were multivariate within-

subjects effects for occasion; however, there were no study group-by-occasion interaction

effects. Follow-up univariate analyses revealed within-subjects effects for 9 cognitive variables.

There was a modest study group-by-occasion interaction on the COWA test, with intervention

group participants improving significantly more across testing occasions. Next, exploratory age

group analyses revealed significant multivariate between-subjects effects of age on executive

measures only. Follow-up univariate analyses demonstrated age group effects for 4 cognitive

variables. For each cognitive measure, younger participants performed significantly better than

their older counterparts. In addition, there were study group-by-occasion interaction effects that

suggested younger control participants performed better on the One-Back Mean RT SD task,

while older intervention group participants performed significantly better on LM Delayed Recall.

A three-way interaction suggested that younger intervention group participants improved

significantly more over time than younger controls and older participants on the Trails B test.

Finally, there was modest, but inconsistent, evidence for correlated change between cognitive,

physical fitness/activity, and psychosocial variables. These findings lend some support to the

previous literature suggesting the benefits of physical fitness/exercise improvements on cognitive

function and the frontal aging hypothesis (West, 1996; Zimmerman et al., 2006). Future research

should explore the benefits of physical and cognitive interventions in diverse samples of middle-

aged and older individuals. Future studies should also explore the use of alternate cognitive and

physical fitness assessment tools in elucidating the cognition-fitness relationship.









BIOGRAPHICAL SKETCH

My academic and research interest in cognitive aging began when I was an undergraduate

student at Florida A&M University in Tallahassee, Florida. During this time, I was a

Distinguished Scholar, with a full academic scholarship. After graduating summa cum laude

with a Bachelor of Arts in psychology in April 2002, I attended the University of Florida on a

University of Florida Alumni Fellowship. I received a Master of Science in clinical psychology

in May 2004. Throughout my graduate career, I was the recipient of various awards and grants,

including a National Institute on Aging (NIA) Aging Research Dissertation Award to Increase

Diversity (Grant #1R36AG029664-01) to fund the present dissertation research. In June 2008, I

completed a pre-doctoral clinical internship in clinical neuropsychology at the University of

Chicago. In addition, I will earn a Doctor of Philosophy degree in clinical psychology, with a

specialty in neuropsychology and a Graduate Certificate in gerontology in August 2008 from the

University of Florida. I will begin a post-doctoral fellowship in geropsychology and geriatric

rehabilitation at Rush University Medical Center in July 2008, and I look forward to a career

studying the influence of heath and disease on cognitive aging in racial/ethnic minority elders.









INELIGIBLE SCRIPT

"This concludes your participation in this study. Thank you for answering these questions. This
has been very helpful. Based on our interview today, you are not eligible to participate in the
study at this time. This is typically because individuals have health conditions that have been
identified as exclusion characteristics for this study. We appreciate the time you have spent
answering these questions. Although you are not eligible for this study, we may want to call you
in the future about your interest in some other study."

29. May we have permission to share your interest in research with our aging research
colleagues here at the University of Florida, so that other researchers can call you to invite you to
consider participating in future research studies?

Y E S ............................... .. ...............1. .
N O ...................................... ............ 2









with LM Recognition at baseline (p < .05), indicated that, as would be expected, lower state and

trait anxiety was associated with higher performance on the LM Recognition task.

Lastly, with regard to exercise-related self-beliefs, both BSE and EXSE scores were

negatively associated with performance on LM Immediate Recall and LM Delayed Recall (p <

.05). This suggested that, contrary to expectation, higher perceived self-efficacy was related to

lower performance on both LM recall trials at baseline.

Next, baseline intercorrelations on all cognitive variables were computed to determine the

inter-relatedness of all executive and non-executive cognitive variables (Table 4-6). Executive

cognitive variables were expected to be more correlated with each other and less correlated with

non-executive cognitive variables. Nevertheless, all executive cognitive measures except for

One-Back Mean RT, Two-Back Mean RT, One-Back Number Correct, and Two-Back Mean RT

SD were significantly correlated with two or more non-executive cognitive measures. Thus,

these findings indicated that the executive cognitive variables were often more correlated with

non-executive cognitive variables than with each other in this sample, suggesting the theoretical

groupings of the cognitive measures was not supported empirically. It further suggested that the

hypothesized separation of executive and non-executive measures with regard to treatment

effects might not be supportable, given the lack of distinctiveness between the two domains.

Physical Fitness and Self-Efficacy Measures: Pre-Post Changes

Prior to examining the effect of the intervention on cognitive outcomes, the influence of

the intervention on intended primary outcomes of the larger study (physical fitness/activity and

exercise self-efficacy) was explored (Table 4-7 and Figure 4-1). While the present document

considers the cognitive measures as primary outcomesfor this study, the larger project in which

it was embedded considered physical fitness and self-efficacy as primary outcomes and cognitive

function as secondary outcomes. As such, the hypothesized improvements on cognitive









CHAPTER 1
STATEMENT OF THE PROBLEM

The present study sought to extend the growing research literature that has shown

physical exercise can improve cognition, particularly executive functioning, in older adults.

Increased physical activity has been shown to be related to many positive outcomes, including

lower mortality and morbidity rates, lower cardiac risks, improved psychological well-being

(lower rates of depression), improved aerobic capacity, and improved functional capacity (e.g.,

McAuley et al., 2004). There is also evidence that cognitive function may be improved with

improved physical fitness through increased exercise behaviors (e.g., Colcombe & Kramer

2003). Thus, there are several premises guiding the current investigation:

* Cognition in later life has plasticity, and can be improved (e.g., Jobe et al., Ball et al.)

* In addition to conventional behavioral interventions, physical exercise interventions have
also been shown to be effective in improving cognition (e.g. Colcombe & Kramer, 2003)

* One route by which physical exercise interventions may improve cognition is by
improving cardiovascular fitness (e.g., Haskell et al., 1992)

* Physical exercise interventions improve cognitive performance (McAuley et al., 2004).

* Physical exercise interventions may improve cognition by increasing complex
activity/social participation, and by improving affective functioning (e.g., Brown, 1992;
McAuley, 1993).

* Depression is another correlate of late-life cognition. Both aging and depression are
independently associated with difficulties in executive functioning and high-level cognitive
control processes that mediate other aspects of cognition (e.g. Hartlage, et al., 1993).

* Exercise may disproportionately affect executive control processes (Colcombe & Kramer,
2003).

A major goal of this study was to elucidate several potential pathways or components of

exercise effects on cognition, separating psychosocial and activity-engagement from physical

fitness changes as possible mediators of these effects. The present study sought to address this

goal by examining the following specific aims:









Table 4-11. Follow-up univariate analyses of variance for non-executive measures


Hypothesis df Error df F P-value Partial q2


1.31
0.02
2.52
1.49
0.55
1.14


Effects
Between-Subjects: Study Group
NAART
Trails A Time
LM Immediate Total Recall
LM Delayed Total Recall
LM Recognition
LM Learning Slope

Within-Subjects: Occasion
NAART
Trails A Time
LM Immediate Total Recall
LM Delayed Total Recall
LM Recognition
LM Learning Slope


0.26
0.88
0.12
0.23
0.46
0.29


**0.00
*0.01
**0.00
**0.00
*0.03
0.29


0.02
0.00
0.04
0.02
0.01
0.02


0.14
0.09
0.27
0.31
0.07
0.02


Interactions: Study Group by
Occasion
NAART 1 67
Trails A Time 1 67
LM Immediate Total Recall 1 67
LM Delayed Total Recall 1 67
LM Recognition 1 67
LM Learning Slope 1 67
Note. *p < .05; **p < .01; LM = Logical Memory; NAART = North
Test


0.01
0.44
0.12
0.61
1.78
0.88
American


0.93 0.00
0.51 0.01
0.73 0.00
0.44 0.01
0.19 0.03
0.35 0.01
Adult Reading


10.68
6.89
24.85
30.69
4.82
1.15









wonders if the best "first-line" approach for improving cognition in the future might not be direct

cognitive intervention. Of course, there are many questions to be answered. For example, the

literature suggests that cognitive training effects may be highly specific to the domains studied

(e.g., Willis et al., 2006), whereas physical training effects may be more broadly global (e.g.,

Kramer et al., 2006). Thus, one possibility is that exercise improves general brain health, and

with it, promotes broad, low-magnitude cognitive gain. However, to improve specific functions

to a higher magnitude, those specific functions need to be practiced and trained.

Dealing with mood and affect

This study experienced almost 30% attrition. In examining the predictors of attrition, the

variables that emerged as most uniquely predictive were pre-test depression symptoms.

Although the exercise promotion intervention in this study was self-efficacy focused, it did not

specifically target mood issues. One wonders if better sample retention, and larger intervention

effects, might have been obtained if a mood intervention component had been included in this

study.

One could imagine an altered multi-step intervention model in future work. If individuals

with higher levels of depression and anxiety symptoms had poor coping skills, and found it

especially challenging to engage in a lifestyle physical activity program (the extreme form of

which was dropout), then perhaps these depression and anxiety symptoms should have been

addressed first.

It is important to note that participants were not, for the most part, at clinical levels of

depression and anxiety. Rather, the argument is that subsyndromal depression and anxiety may

interfere with full participation in the intervention. If this is true, then future research might first

employ cognitive-behavioral approaches (either in the full group, or individually in one-on-one

sessions with persons experiencing elevated levels) for dealing with mood issues. The idea is










Table 4-20. Raw cognitive estimated marginal means by age and study group
Younger Group


Control Group


Exercise Promotion Group


Non-Executive


NAART
Trails A Time
LM Immediate Recall
LM Delayed Recall
LM Recognition
LM Learning Slope


Mean
113.02
28.38
41.0(
26.0(
25.8
4.7(


Executive


COWA
^ Trails B Time
Letter Number Sequencing
One-Back: Mean RT
One-Back: Mean RT SD
One-Back: Number Correct
Two-Back: Mean RT
Two-Back: Mean RT SD
Two-Back: Number Correct


Mean
40.8
68.C
10.8
760.9
268.2
92.8
1518.7
773.4
91.7


Baseline
Std. Error
7 1.56
8 2.66
0 2.06
3 1.83
8 0.57
3 0.65

Baseline
Std. Error
8 2.68
)1 9.29
8 0.65
>3 39.14
:9 43.28
8 3.46
'8 128.37
r9 117.76
'8 2.33


Post-Test
Mean Std. Error
114.63 1.50
27.16 1.75
46.06 2.22
30.06 1.73
26.88 0.42
3.12 0.54

Post-Test
Mean Std. Error


42.65
70.40
10.65
755.63
200.10
98.65
1341.44
623.45
94.71


2.40
8.97
0.55
36.15
24.83
0.68
143.18
100.35
1.44


Baseline
Mean Std. Error
110.19 1.71
31.87 2.93
42.00 2.27
24.79 2.02
26.31 0.63
3.71 0.72

Baseline
Mean Std. Error
38.21 2.95
82.68 10.24
10.57 0.71
776.21 43.13
333.60 47.69
95.79 3.82
1566.19 141.46
720.47 129.76
93.21 2.57


Post-Test
Mean Std. Error
111.53 1.65
25.25 1.93
45.50 2.44
28.86 1.91
26.93 0.47
4.00 0.60

Post-Test
Mean Std. Error
42.93 2.65
65.52 9.88
10.93 0.60
741.20 39.84
241.58 27.36
98.07 0.74
1470.00 157.78
623.33 110.58
94.71 1.58


Note: LM = Logical Memory; NAART = North
Reaction Time; SD = Standard Deviation


American Adult Reading Test; COWA = Controlled Oral Word Association; RT












19

-^-
8 95.00
U

1 .


85.00


O I Control Middle
Aged
Control Older
Adult
- Intervention
Middle Aged
- Intervention
Older Adult


Baseline Post-Test
Age by Study Group
I

Figure 4-6 Continued.


a









Finally, one limitation inherent in the larger study's design was permitting the health

hygiene control group access to an exercise facility in exchange for study participation. This

decision was made to avoid the potential confound of the anticipated intervention effects and

access to a fitness facility and also mounted a very stringent test of the added value of this

motivational intervention. Nonetheless, this decision precluded complete study manipulation of

exercise/physical activity, as there was no way to limit the amount or intensity of exercise in the

control group. Also, the social contact and use of a peer-mentorship model may have increased

motivation in the control group to make changes in physical activity. It may be the case that the

combination of these factors made for an uncharacteristic control group, instead of what would

be expected from a more traditional, no-contact control condition.

Conceptual Issues

A number of broader conceptual issues were introduced in the Introduction Chapter, and

several others emerged as a result of the analyses. This section considers several of the key

issues that emerged.

Selection of the correct intervention

A first issue to consider is whether a 16-week intervention period, with only 12 weeks of

actual instruction, is sufficient enough to reasonably expect adequate physiological change.

Unfortunately, inconsistencies in the existing literature provide less concrete guidance in the

types and lengths of exercise intervention protocols to ensure consistent effects on both physical

and cognitive outcomes (Kramer et al., 2006). Review of various exercise intervention trials,

with cognitive outcomes, shows that the length of training programs may range from as few as

10 weeks (Hawkins et al., 1992; Emery et al., 1998) to as many as 14 months (Blumenthal et al.,

1991) and provide mixed evidence for the relationship between physical fitness and cognition.

Methodological reasons for such inconsistent findings that have been discussed include









research should examine these other types of executive functions to gain a more complete

understanding of the effects of cardiovascular/physical fitness.

It may be the case that the fitness-cognition relationship is mediated primarily by the

attentional system, such that this is the mechanism for improved executive control processing. In

fact, in Colcombe & Kramer's (2003) meta-analysis, there was no control for the overlap

between executive measures that could also be categorized as measuring a related cognitive

function (i.e., speed, visuospatial ability, or sustained attention). Their "executive" category

included the largest number of studies; thus, this category may have had the most variance, and

that could be the explanation for why executive measures appeared to be most-improved by

exercise interventions. Very concretely, almost every measure in their meta-analysis was

classified as executive and something else. Thus, Colcombe & Kramer's conclusion that

executive control processes show the most exercise-related improvements may require further

examination.

Future research should examine more than executive tasks that are highly dependent on the

attentional system in the quest to better understand the link between fitness and executive

cognitive function. Including measures of higher-order conceptual abilities would provide a

deeper level of analysis of all of the executive functions that are supported by frontal lobe and

subcortical brain function.

Future work should also pay attention to sensitivity and specificity of measures in the

selection of cognitive batteries. More sensitive measures might be selected in a highly-

advantaged sample, such as the present one, to maximize the likelihood of detecting individual

differences in baseline performance, which could then be more sensitive to a physical









CHAPTER 4
RESULTS

This study examining the effects of an exercise intervention on executive cognitive function

involved a between (group assignment) and within (measure/instrument, occasion of

measurement) subjects design. The primary outcome was cognitive performance, while

proximal or mediating outcomes measured included physical fitness and self-perception

(depression, well-being, and self-efficacy). The following results are presented with respect to

two specific aims:

* To investigate whether, relative to matched non-exercising control participants, sedentary
adults receiving a physical exercise promotion intervention experienced improvements in
the primary outcome of cognitive function (particularly executive control processes).

* To determine the separate and joint roles of improvement in proximal outcomes (fitness,
activity, and affect) in mediating exercise intervention effects on cognition.

Preliminary Analyses

Attrition of the sample

A total of 22 participants were considered "dropouts" in this study, resulting in the final

sample of 69 participants. Of the 91 individuals that were randomized to either the control or

intervention conditions following baseline assessment, 10 dropped out of the study prior to the

16-week study group period (Figure 3-1). One participant was randomized to a group with the

expectation that pre-testing would be completed just prior to the first session; however, the

participant did not undergo baseline assessment or remain in the study for the small group

sessions. Post-randomization, an additional twelve participants completed part or all of the small

group sessions and did not complete post-testing. As shown in Table 4-1, in comparing

demographic characteristics of individuals who completed post-testing (n = 69) to those

individuals who only completed pre-testing (n = 21), independent samples t-tests indicated that









Table 4-8. Repeated-measures multivariate analysis of variance on anxiety and depression measuresa
Effects Wilks' Lambda F Hypothesis df Error df P-value Partial qj2
Between Subjects: Study Group 0.97 0.41 4 64 0.80 0.03
Within Subjects: Occasion 0.71 6.68 4 64 **0.00 0.29
Interaction: Study Group by Occasion 0.94 1.00 4 64 0.41 0.06
Note. *p < .05; **p < .01; aThere were unequal variance-covariance matrices for these measures.




Full Text

PAGE 1

1 EFFECTS OF IMPROVED PHYSICAL FI TNESS ON COGNITIVE/PSYCHOLOGICAL FUNCTIONING IN COMM UNITY-DWELLING, SEDE NTARY MIDDLE-AGED AND OLDER ADULTS By ADRIENNE T. AIKEN MORGAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Adrienne T. Aiken Morgan

PAGE 3

3 To my grandparents and great-grandparents, all of whom have inspired me to pursue a career in minority cognitive aging research.

PAGE 4

4 ACKNOWLEDGMENTS This dissertation project was supported primar ily by a National Institute on Aging (NIA) Aging Research Dissertation Award to In crease Diversity (1R 36AG029664-01 PI: Aiken Morgan). In addition, this work was supported by other sources that funded the larger study in which it was em bedded: a Research Opportunity Fund in the College of Health and Human Performance at the University of Florida (P I: Giacobbi), an Age Network Multidisciplinary Research Enhancement grant at the Univers ity of Florida (PI: McCrae), a Mentorship Opportunity Grant from the Graduate Student Counc il at the University of Florida (PI: Buman), a Graduate Student Research Grant, University of Florida, College of Public Health and Health Professionals (PI: Dzierzewski), and a Divisi on 20, American Psychol ogical Association and Retirement Research Foundation Thesis Pr oposal Award (PI: Dzierzewski). There are many people without whom this diss ertation would not have been successfully accomplished, and I would be remiss if I did not properly acknowledge them here. First, I thank my mentor, Dr. Michael Marsiske, for his gui dance and unwavering s upport throughout my graduate school career. I trul y appreciate his dedication to providing outstanding mentorship, and I owe the better part of my writing skills, ability to think critically about research, and knowledge of statistics and methodology to him. I also owe him my sincerest gratitude for the countless hours of work that he willfully co mpleted both with and for me to ensure my matriculation through graduate school. I tha nk my doctoral committee for their patience and support throughout this process: Dr. Russell Bauer, Dr. Shawn Kneipp, Dr. Michael Marsiske, Dr. Christina McCrae, and Dr. Catherine Price. I also thank the Project AAMP research team: Peter Giacobbi, Jr., Michael Marsiske, Christin a McCrae, Beverly Roberts, Matthew Buman, and Joe Dzierzewski. This project would not have been completed without the full strength and support of this collaborative group, and I look forward to our conti nued successful collaboration.

PAGE 5

5 I especially thank my partners in crime, Matt and Joe, for challenging me to be a better researcher. I offer my heartfelt appreci ation also to my research assistants, without whom, I would not have been able to collect these data : Katherine Blasewitz, Lauren Cohen, Robin Ezzel, Joseph Gullet, Mario Jimenez, Phuong Nguyen, Jacqueline Memminger, Meghan Saculla, and Howin Tsang. Thanks also to the many commun ity partners, peer group mentors, and other support staff that made this project possible. Next, I thank my parents, Bennie and Gail Aike n, for their unconditional love. My parents have been there to support me through the lowest of lows and the highest of highs throughout my life. This journey has not always been easy, but they sure helped to carry many of my burdens. I thank them for their encouragement to always strive to do and be my bestno matter the challenge set before me. I thank them for al ways pushing me to reach for the highest and brightest stars. I am indebted to them for t eaching me, and also often reminding me, that I can indeed do all things through Christ who strengthens me. Last, to my dear husband and my rock, Ra shadthank you for holding me down and being the best friend and partner I could ever imagin e finding. Words cannot adequately express the gratitude and love that I have for you. Thank you for pushing me forward all of the times I felt I would only go backwards. Thank you for understanding me like no one else can. Thank you for all of the little things you do to remind me how blessed I am to have you in my life. If I had one wish, it would be to simply spend the rest of my days growin g old right beside you.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES.......................................................................................................................11 ABSTRACT...................................................................................................................................12 CHAPTER 1 STATEMENT OF THE PROBLEM......................................................................................14 2 REVIEW OF THE LITERATURE........................................................................................ 16 Overview....................................................................................................................... ..........16 Conceptual model............................................................................................................ 17 Significance................................................................................................................... ..17 Continuum of Late Life Cognition......................................................................................... 18 Cardiovascular and Physical Health Influences on Late Life Cognition ................................22 Relationship between hypertension and cognition..........................................................23 Underlying physiological mechanisms............................................................................ 25 Synergistic effect of hypertension and other cardiovascular conditions......................... 27 Environmental factors..................................................................................................... 27 Self-Perception Influences on Late Life Cognition................................................................ 28 Depression and Well-Being............................................................................................. 28 Vascular Depression Hypothesis..................................................................................... 31 Control Beliefs and Self-Efficacy...................................................................................33 Plasticity of Late Life Cognition............................................................................................ 34 Cognitive Interventions................................................................................................... 35 Physical Exercise Interventi ons: Cardiovascular Fitness................................................ 36 Relationship between fitness and cognition.............................................................37 Psychological benefits of exercise interventions..................................................... 40 Conclusion..............................................................................................................................41 3 METHODS.............................................................................................................................43 Overview....................................................................................................................... ..........43 Participants.............................................................................................................................44 Statistical Power Considerations..................................................................................... 44 Sample Characteristics....................................................................................................45 Cognitive Exclusions....................................................................................................... 46 Physicians Permission.................................................................................................... 46

PAGE 7

7 Physical/Cardiovascular Exclusions................................................................................ 46 Medication Exclusions.................................................................................................... 47 Inclusion: Staging Algorithm..........................................................................................47 Procedure................................................................................................................................47 Overview.........................................................................................................................47 Rationale for Measures.................................................................................................... 49 Primary Outcomes........................................................................................................... 50 North American Adult Reading Test (NAART)...................................................... 50 Logical Memory subtest of the Wechsler Memory Scale, 3rd Edition (WMSIII).........................................................................................................................50 Control Oral Word Association (COWA)................................................................ 51 Trail Making Test A and B (Trails A, Trails B)....................................................... 51 Letter-Number Sequencing subtest of the Wechsler Mem ory Scale, 3rd Edition (WMS-III)............................................................................................................. 52 N-Back task..............................................................................................................52 Proximal Outcomes......................................................................................................... 53 Geriatric Depression Scale (GDS)...........................................................................53 Beck Depression Inventory2nd Edition (BDI-II).................................................. 53 State-Trait Anxiety Inventory (STAI) ...................................................................... 54 Exercise self-efficacy...............................................................................................54 Leisure Time Exercise Questionnaire (LTEQ)........................................................54 Minutes of Moderate and Vigorous Physical Activity (MVPA).............................. 55 Pedometer................................................................................................................. 55 Modified Balke Submax (VO2)................................................................................55 Design and Rationale fo r Experime ntal Group............................................................... 56 Design and Rationale for Control Group......................................................................... 57 4 RESULTS...............................................................................................................................69 Preliminary Analyses..............................................................................................................69 Attrition of the sample..................................................................................................... 69 Missing Data....................................................................................................................70 Distributions of Dependent Variables and Outliers.........................................................71 Baseline comparisons and correlations...........................................................................72 Physical Fitness and Self-Efficacy Measures: Pr e-Post Changes................................... 74 Anxiety and Depression: Pre-Post Changes.................................................................... 76 Aim 1: Exercise Promotion Interventi on-Related Improve ments in Cognition..................... 77 Study Group Comparisons.............................................................................................. 77 Reliable Change...............................................................................................................80 Exploratory Analyses: Age Group Com parisons............................................................ 81 Aim 2: Cognitive changes correlated with activity, fitness and psychosocial ch anges.......... 84 5 DISCUSSION.......................................................................................................................138 Review of Study Findings....................................................................................................138 Preliminary Analyses..................................................................................................... 139

PAGE 8

8 Baseline correlations among measures.................................................................. 139 Pre-post changes in fitness and activity................................................................. 140 Pre-post changes in self-efficacy, anxiety, and depression.................................... 141 Primary Analyses........................................................................................................... 142 Aim One................................................................................................................. 142 Aim Two................................................................................................................ 146 Study Limitations.............................................................................................................. ....147 Sampling issues.............................................................................................................147 Retention issues............................................................................................................. 149 Measurement issues....................................................................................................... 151 Intervention issues.........................................................................................................152 Conceptual Issues.............................................................................................................. ...154 Selection of the correct intervention.............................................................................. 154 Selection of the correct follow-up interval.................................................................... 156 On the relative importance of physical vs. mental exercise.......................................... 157 Dealing with mood and affect....................................................................................... 159 Future Directions..................................................................................................................160 Longitudinal follow-up.................................................................................................. 160 Enhanced sampling........................................................................................................160 Further conceptualization of the target cognitive doma ins........................................... 162 Further conceptualization of f itness and activity measure ment.................................... 165 Relative effect size of exercise and combinatorial interventions .................................. 166 Conclusion............................................................................................................................167 APPENDIX A TELEPHONE SCREENING PROCEDURE....................................................................... 168 B NURSE/PHYSICIAN CHECKLIST AND PERMISSION................................................. 186 C QUALITY CONTROL CHECKLISTS............................................................................... 187 LIST OF REFERENCES.............................................................................................................190 BIOGRAPHICAL SKETCH.......................................................................................................201

PAGE 9

9 LIST OF TABLES Table page 3-1 Pre-data collection power analysis resu lts, indicating power e xpected at different effect sizes with two cells of 32 participants, and alpha = .05...........................................59 3-2 Participants mean characteristics...................................................................................... 60 3-3 Proposed study timeline for each replicate of participants............................................... 61 3-4 Number of participants by age categ ory, replicate, and experime ntal group....................62 3-5 Baseline and post-inte rvention study protocol................................................................... 63 3-6 Schedule of intervention topics.......................................................................................... 64 3-7 Health hygiene control group design................................................................................. 65 3-8 Schedule of control group topics....................................................................................... 66 4-1 Baseline mean comparison of study completers and dropouts.......................................... 86 4-2 Distributions of depende nt variables at baseline............................................................... 88 4-3 Baseline correlations between physical activity, physical fitness, and cogn itive variables.............................................................................................................................89 4-4 Baseline correlations between physical f itness, physical activity, and psychosocial variables...................................................................................................................... .......90 4-5 Baseline correlations between psychosocial and cognitive variables................................ 91 4-6 Cognitive intercorre lations at baselin e............................................................................... 92 4-7 Univariate analyses of variance for physical activ ity and fitness...................................... 95 4-8 Repeated-measures multivariate analys is of variance on anxiety and depression measuresa...........................................................................................................................96 4-9 Follow-up univariate analyses of vari ance for depression and anxiety me asures............. 97 4-10 Repeated measures multivariate analys es of varia nce for cognitive domains with study group as between -subjects factor.............................................................................98 4-11 Follow-up univariate analyses of variance for non-executive measures ........................... 99 4-12 Follow-up univariate analyses of variance for executive me asures................................. 100

PAGE 10

10 4-13 Raw cognitive estimated ma rginal means by study group............................................... 101 4-14 Calculation of reliable cha nge index scores adjusted fo r control group practice effect.. 102 4-15 Participants who reliably change d on non-executive cognitive measures ....................... 103 4-16 Participants who reliably changed on executive cognitive measures.............................. 104 4-17 Multivariate analyses of variance for cognitive dom ains with study group and age group as between-subjects factors................................................................................... 106 4-18 Follow-up univariate analyses of va riance for non-executive measures: Study group and age group as between subjects factors ...................................................................... 107 4-19 Follow-up univariate analyses of vari ance for executive me asures: Study group and age group as between subjects factors.............................................................................109 4-20 Raw cognitive estimated marginal means by age and study group................................. 112 4-21 Correlations between physical activity, physical fitness, and cognitive change scores .. 114 4-22 Correlations between phys ical fitness, physical activ ity, and psychosocial change scores................................................................................................................................115 4-23 Correlations between psychosocia l and cognitive change scores.................................... 116

PAGE 11

11 LIST OF FIGURES Figure page 2-1 Conceptual model of th e effects of exercise intervention on cognitive function.............. 42 3-1 Flow chart of participants at each study phase.................................................................. 67 3-2 Guiding theoretical framework for exerci se promotion group intervention (Buman, 2008) ..................................................................................................................................68 4-1 Mean scores by study group and occasion for physical fitness, activity, and selfefficacy variables............................................................................................................. 117 4-2 Mean anxiety and depression sc ores by study group and occasion................................. 120 4-3 Mean cognitive scores by study group and occasion for non-executive cognitive variables...................................................................................................................... .....122 4-4 Mean cognitive scores by study group and occasion for executive cogn itive variables. 125 4-5 Mean cognitive scores by study group, age group, and occasion for non-executive cognitive variables........................................................................................................... 130 4-6 Mean cognitive scores by study group, age group, and occasion for executive cognitive variables........................................................................................................... 133

PAGE 12

12 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EFFECTS OF IMPROVED PHYSICAL FI TNESS ON COGNITIVE/PSYCHOLOGICAL FUNCTIONING IN COMM UNITY-DWELLING, SEDE NTARY MIDDLE-AGED AND OLDER ADULTS By Adrienne T. Aiken Morgan August 2008 Chair: Michael Marsiske Major: Psychology A growing corpus of research suggests th at physical exercise can improve cognition, particularly executive functioning, in older adults. However, limitations of existing research have included (a) insufficient atte ntion to the recruitment of sedentary older adults (who would most likely benefit from exercise interventions ); (b) insufficient guidance in test selection drawing on neuropsychological theory and practi ce; and (c) failure to elucidate the multiple pathways or components of exercise effects on cognition. The current study sought to better clarify these routes to cognitive improvement via (a) assessment of both potential physical fitness and psychosocial mediators of exercise e ffects on cognition, and (b) inclusion of a control group that received a comparable psychoeduca tional intervention, matched in study contact hours and study-related non-exercise activities, but which did not receive a physical exercise enhancement intervention. Two randomized groups of 35 (control) and 34 (invention) adults aged 50 years and older were recruited from the Gainesville/Alachua County, Florida region. Both groups underwent preand post-intervention cognitive, fitness, and psychosocial/socioemotional assessment. The exercise promotion intervention group received 16 weeks of intervention (health and fitness education, weekly peer motivati onal coaching and group support,

PAGE 13

13 etc.) in small groups with a peer mentor, while a control/comparison group received 16 weeks of health hygiene instruction, consisting of 16 weeks of education about general health conditions in aging (also in small groups with a peer mentor). Repeated-measures MANOVA indicated no significant between-subjects effect of the intervention ( p >.05). There were multivariate withinsubjects effects for occasion; however, th ere were no study group-by-occasion interaction effects. Follow-up univariate analyses revealed within-subjects effects for 9 cognitive variables. There was a modest study group-by-occasion inte raction on the COWA test, with intervention group participants improving significantly more acr oss testing occasions. Next, exploratory age group analyses revealed significant multivariate between-subjects eff ects of age on executive measures only. Follow-up univariate analyses demonstrated age group effects for 4 cognitive variables. For each cognitive measure, younger pa rticipants performed significantly better than their older counterparts. In addition, there were study groupby-occasion interaction effects that suggested younger control participants performed better on the One-Back Mean RT SD task, while older intervention group part icipants performed significantly better on LM Delayed Recall. A three-way interaction sugge sted that younger intervention group participants improved significantly more over time than younger controls and older partic ipants on the Trails B test. Finally, there was modest, but inconsistent, evidence for correlated change between cognitive, physical fitness/activity, and psychosocial variables. Thes e findings lend some support to the previous literature suggesting th e benefits of physical fitness/ex ercise improvements on cognitive function and the frontal aging hypothesis (West, 1996; Zimmerman et al., 2006). Future research should explore the benefits of physical and cogn itive interventions in diverse samples of middleaged and older individuals. Future studies should also explore the use of alternate cognitive and physical fitness assessment to ols in elucidating the cognition-fitness relationship.

PAGE 14

14 CHAPTER 1 STATEMENT OF THE PROBLEM The present study sought to ex tend the growing research li terature that has shown physical exercise can improve cognition, particularly executive functioning, in older adults. Incre ased physical activity has been shown to be related to many positive outcomes, including lower mortality and morbidity rates, lower ca rdiac risks, improved psychological well-being (lower rates of depression), improved aerobic capacity, and improved fu nctional capacity (e.g., McAuley et al., 2004). There is also evidence that cognitive function may be improved with improved physical fitness through increased ex ercise behaviors (e.g., Colcombe & Kramer 2003). Thus, there are several premises guiding the curren t investigation: Cognition in later life has plas ticity, and can be improved (e. g., Jobe et al., Ball et al.) In addition to conventional behavioral interv entions, physical exercise interventions have also been shown to be effective in im proving cognition (e.g. Colcombe & Kramer, 2003) One route by which physical exercise in terventions may improve cognition is by improving cardiovascular fitness (e.g., Haskell et al., 1992) Physical exercise interventions improve cognitive performanc e (McAuley et al., 2004). Physical exercise interventions may improve cognition by increasing complex activity/social participati on, and by improving affective functioning (e.g., Brown, 1992; McAuley, 1993). Depression is another correlate of late-l ife cognition. Both aging and depression are independently associated with difficulties in executive functioning and high-level cognitive control processes that mediat e other aspects of cognition (e .g. Hartlage, et al., 1993). Exercise may disproportionately affect execu tive control processes (Colcombe & Kramer, 2003). A major goal of this study was to elucidate several potential pathwa ys or components of exercise effects on cognition, separating psychos ocial and activity-engage ment from physical fitness changes as possible mediators of these e ffects. The present study sought to address this goal by examining the following specific aims:

PAGE 15

15 To investigate whether, relative to matched non-exercising control pa rticipants, sedentary adults receiving a physical ex ercise promotion intervention experience improvements in the primary outcome of cognitive function (par ticularly executive co ntrol processes). Hypothesis : As shown in previous studies, particip ants receiving the exercise intervention will show improved performance on cognitive measur es, particularly those assessing executive control processes, relative to control participants. To determine the separate and joint roles of improvement in proximal outcomes (fitness, activity, and affect) in mediating exerci se intervention eff ects on cognition. Hypothesis : It is further expected that changes in measures of physical fitness, physical activity engagement, and emotional well-being mediat es some or all of the intervention effect on cognitive change.

PAGE 16

16 CHAPTER 2 REVIEW OF THE LITERATURE Overview The present investigat ion examined the effects of i m proved physical fitness on cognitive functioning in a group of adults, ages 50 and older. Thus, the review of th e literature will first consider the evidence for normative and non-normative cognitive declines with aging, highlighting the cognitive continuum in late r life. Evidence to date suggests aging disproportionately affects executive functions and processes, and the present literature review will discuss these findings. Next, the literature re view will consider evidence for the plasticity of cognition in late life. Specifica lly, the review will consider re search that has shown cognition may be improved through cognitive training interventions, as well as beha vioral interventions. Such behavioral interventions include physical exercise in terventions aimed at improving cognitive function through im proved physical fitness. One route by which physical exercise in terventions may improve cognition is by improving cardiovascular fitness, and this literature will be discussed. Next, additional routes by which physical exercise interventions may impr ove cognition (i.e., via improving mood/affect, self-efficacy, and control beliefs) will also be expl ored. This review is intended to set the stage for the current study, which examined pre-post exercise intervention cognitive performance changes on selected neuropsychological measures. Additionally, this study empirically investigated whether measures of executive cont rol processes were particularly sensitive to physical exercise intervention (as a recent meta -analysis has suggested), relative to other cognitive domains. This study also explored whether physical (fitn ess) and psychosocial variables (mood, affect, and self-efficacy) change d as a result of an exercise promotion intervention, and whether these helped explain any pre-post changes in cognition

PAGE 17

17 Conceptual model The current study sought to confirm the eff ects of exercise on cognition using a m ore individualized approach to exercise intervention than many previous studies (this is described in greater detail below). In addition to examin ing exercise effects on cognition, the study also explored several possible mediators of exercise effects, including physical factors (changes in physical fitness) and psychosocial factors (cha nges in self-perception, depression, well-being, and self-efficacy). Thus, the conceptual model below guides the current study (Figure 2-1). As noted above, exercising may also be construe d as a form of complex activity, which may increase social engagement. Many previous studies have failed to control for these social participation components of exercise, using wait -list or no-contact control groups. The present study included a health hygiene co ntrol group, matched with experi mental participants in study contact hours and interaction with study staff and other participants and in exposure to cognitive testing. These factors were designed to reduce the confounding role of so cial/complex activity participation as a plausible alternativ e explanation of exercise effects. Significance The proposed study sought to further th e understanding of the relationship between cognition and exercise in a m iddle-aged and older adult sample. It further attempted to explore the unique and combined contributions of fitne ss and self-perception changes as mediators of exercise effects on cognition. This is unique in that, to ou r knowledge, few studies have attempted to examine both process and outcome exercise-related vari ables as routes to cognitive improvement in an older adult popu lation. Furthermore, this study was thought to be the first in a line of future research aimed at exploring the ef fect of one potentially remediable antecedent of late life cognitive decline, physical fitness.

PAGE 18

18 Continuum of Late Life Cognition Late life cognitive function appears to repr es ent a continuum, ranging from normative agerelated declines in abilities to pathological cognitive deficits. Clinic al and neuroanatomical evidence support such a continuum notion of normal and pathological cognitive aging; nevertheless, there are certain factors which exacerbate (and op timize) this cognitive aging process, such that a smooth, linear transition al ong all points of this continuum may not be observed for all individuals. Thus, some individuals may be at a le sser or greater risk for optimal cognitive outcomes in late life. Normal cognitive aging has been characteri zed in a multidimensional and multidirectional way; with aging, there is evidence for gains an d loses in cognitive functi on (Baltes et al., 1999). Evidence suggests a gradual decline in some c ognitive areas, as well as stability in other cognitive functions. Specifically, th ere appears to be linear declines in fluid abilities as well as crystallized abilities. Fluid abilities refer to those that are nece ssary for processing and learning novel information and coping with new situations in the environment (i.e., abstract reasoning, processing speed, episodic memory processing). In contrast, crystallized abilities represent the knowledge garnered from a lifetime of experien ce (i.e., general in telligence, verbal skills, and semantic knowledge). Interestingly, it appears fluid abilities show a more steady and accelerated decline with aging, wh ile there is some preservation and even perhaps gain in crystallized knowledge until late life, at which poin t there is accelerated dec line in these abilities (around 75-80 years). Drawing from the neuropsychological literatu re, there are several normal aging related declines observed. For instance, while memory declines are commonly cited, executive functions also have been found to be closely related to the aging process and sustained by specific brain regions (e.g., We st, 1996; Greenwood, 2000; Charot & Feyereisen, 2005; Gunstad,

PAGE 19

19 et al., 2006; Zimmerman et al., 2006). Executive functions, while not con ceptualized the same, are similar to fluid abilities, in that such skills are important for the successful processing and learning of novel information and coping with new situations. Sp ecifically, executive functions, governed by the frontal brain regions, generally consist of the capacities to engage successfully in independent, purposive, self-serving behavior (Lezak et al., 2004). These behaviors include formulating goals with regard for long-term consequences, generating multiple response alternatives, choosing an d initiating goal-directed behaviors, self-monitoring the adequacy and correctness of the behavior, correcting and modi fying behaviors when conditions changes, and persisting in the fact of distraction (Synder & Nussbaum, 1998). With specific regard to cognitive behaviors, executive functions refer to the strategies invol ved in approaching, planning, or carrying out cognitive tasks, or in monitoring of performance (Lezak et al., 2004). These cognitive components, often termed executiv e control functions, of executive functions are sub-served by the dorsolateral prefrontal cortical brain re gions (Synder & Nussbaum, 1998). Examples of executive control functions in clude working memory, mental flexibility, maintenance of task set, and task-switching. Next, evidence has suggested frontal lobe structure and (execu tive) function undergo disproportionate declines with aging (e.g., We st, 1996; Greenwood, 2000; Charot & Feyereisen, 2005; Gunstad, et al., 2006; Zimmerman et al., 2006); this consistent findi ng has been termed the frontal aging hypothesis. The frontal aging hypothesis, which is supported by many in the literature, states frontal-executive abilities seem to be disproportionately negatively affected by aging relative to non-frontal, more posterior brain mediated abilitie s (West, 1996; Greenwood, 2000). Nonetheless, while with normal agi ng, execution of executive functions becomes inefficient (e.g., Pfefferbaum, Ad alstein, & Sullivan, 2005; Brickman et al., 2006; Nordahl et al.,

PAGE 20

20 2006; Sullivan & Pfefferbaum, 2006), it is not im possible. Instead, it becomes necessary for recruitment of wider-spread brain networks to perform at the same level on executive tasks (i.e., HAROLD model, Cabeza, 2002; Cherry et al., 2005). Study in this area has shifted to not only exam ine cortical, or gray matter, influences on executive function, but also supporting white matte r regions (Pfefferbaum et al., 2005; Brickman et al., 2006; Nordahl et al., 2006; Sullivan & Pfefferbaum, 2006). In one study (Pfefferbaum et al., 2005) using diffusion tensor imag ing (DTI), the integr ity of white matter fiber microstructure in a group of healthy, highly educated young and older adults was examined. DTI results indicated that older adults s howed greater frontal white matte r loss (anisotropy) relative to younger subjects. There were no age-related differences found for posterior white matter regions. In general, normal age-related white ma tter loss features degradation of myelin and microtubules and number and length of myelinated fibers, particularly in the precentral gyrus and corpus callosum. In addition, thin fibers, which are greatest in numb er in the frontal lobes, seem to have a predilection for loss. Axon deletion se ems to also be related to the white matter normal aging process. Further study has confirmed this age-associated d ecline in frontal white matter, and such decline has been shown to mediate declines in neuropsychological functioning in a more general fashion (Brickman et al., 2006; Nordahl et al., 2006). Such cortical and subcortical declines with aging have implications for the corticalsubcortical circuitry that underl ies executive control pr ocessing. Specifically, there are at least five parallel circuits that conn ect the frontal cortex to subcorti cal structures. The input segment of the basal ganglia (BG) is th e striatum, which consists of the caudate, putamen, and nucleus accumbens. The striatum receives input from mu ltiple cortical areas and projects through the globus pallidus and substantia nigra to the thal amus, ultimately closing the circuit back to the

PAGE 21

21 area of the cortex from which it received (A lexander, DeLong, & Strick, 1986). One circuit relevant to the present study (the associative ci rcuit, implicated in working memory function) originates in the dorsolateral pr efrontal and lateral orbitofrontal cortex and includes the caudate nucleus, part of the putamen, the anterodorsal globus pallidus internus, the substantia nigra pars reticulata, and the ventroanterior and mediodorsal thalamic nuclei. The integrity of both cortical and subcortical structures is important to the proper function of these connections, and normal aging-related brain changes may se rve to jeopardize this circuitry. These same vulnerable white matter regions th at support executive control functions are those that are particularly se nsitive to cardiovascular and cerebrovascular health declines (Pantoni et al., 1999; Desmond, 2002; van Boxtel et al., 2006). White matter lesions frequently occur in adults aged 60 and olde r, and among these older individua ls, those with vascular risk factors and cerebrovascular disease are more likel y to have white matter lesions (Pantoni et al., 1999; Desmond, 2002; van Boxtel et al., 2006). In normal elders without dementia, white matter lesions are associated with subtle cognitive defi cits on tests of attention, speed of processing, planning, and other executive functions (Schmidt et al, 1993). Nevertheless, such white matter abnormalities may lead to the pathological aging e nd of the late life cognition continuum and results in more severe cognitive deficits. This notion is supported by evidence suggesting there are forms of vascular cognitive impairment in which the main form of neuroanatomical pathology is found in the white matter (e.g., Pantoni et al., 1999). This pathology results in the disruption of cortical-subcorti cal circuitry described previous ly (Cummings et al., 1993). Particularly, in subcortical vasc ular dementia, the presence of deep and periventricular white matter alterations and lacunar infa rcts tends to result in difficu lties on tasks involving mental manipulation, and more notably, difficulties in es tablishing and maintaini ng mental set (Libon et

PAGE 22

22 al., 2004). There tends to be a relative sparing on memory a nd language tests (Libon et al., 2004). These findings guide the current proposed study with regard to the hypothesized selective benefit of exercise on executive cont rol processes. Furthermore, th ere is evidence to suggest that related to the selec tive decline in executive br ain function in older indivi duals, physical exercise interventions work to improve these vulnerable frontal and white matter brain regions. Specifically, brain mechanisms for improvement with physical exercise include increased cerebrovascular perfus ion, cerebrovascular sufficiency, a nd improved cortical plasticity (McAuley et al. 2004); these ch anges occur through increased proliferation of neurotrophin factors, dendritic branching, neurogenesis, and fo rmation of new synaptic connections (McAuley et al. 2004). This is a fundamental premise whic h drives the current prop osal, as the hope is to demonstrate that by improving physical, particul arly cardiovascular, fitness levels through physical exercise intervention, br ain cerebrovascular perfusion and sufficiency and cortical plasticity may be enhanced in a group of vulnera ble individuals (due to older age and sedentary lifestyle), thereby enhancing th e neurocognitive (executive contro l) functions dependent on the integrity of brain white ma tter structure and function. Cardiovascular and Physical Health In fluences on Late Life Cognition There are certa in factors that may exacerbate cognitive aging and the aging process in general. These include behavioral factors, such as diet, physi cal activity, and other health promoting behaviors. In additi on, environmental factors, such as socioeconomic status, lifetime educational and occupational opportu nity, stress, and a host of other environmental influences, interfere with optimal aging (Burke et al., 2001) Furthermore, the presence of cardiovascular risk factors can influence cogniti ve function with aging. It has been documented hypertension is a major risk for many other cardiovascular diseases (stroke, diabetes, congestive heart failure,

PAGE 23

23 etc.), and its pathogenesis is linked to cognitive function (Wal dstein & Katzel, 2001; Waldstein et al., 2005; Waldstein & Katzel 2005). Hypertension is a co mmon health problem; often, the presence of hypertension precedes and is comorbid with more serious cardiovascular diseases. Hypertension is thought to be related to incr eased cardiac output, increas ed peripheral vascular resistance, increased sympathetic nervous syst em activation, and compromised renal function (Waldstein & Katzel, 2001). It also has been sh own to exacerbate age-related arterial stiffness (Benetos, 2002). Furthermore, hypertension has b een linked to cognitive function in many ways. Mechanisms include decreased cerebral blood flow and metabolism, metabolic syndrome, atherosclerosis, white matter disease, brain at rophy, and activation of stress response systems (i.e., the hypothalamic-pituitaryadrenocortical axis) (Waldstein & Katzel, 2001; Gold et al., 2005). These mechanisms, as mentioned above, have detrimental implications for the integrity of gray (Gianaros et al., 2006) and white matte r structure and executive control function. Relationship between hypertension and cognition In reviewing resea rch exploring the relationship between cardiovascular illness and cognition, several common themes emerged. First, with specific regard to blood pressure, there appears to be a U-shaped association with cogni tion. Next, cardiovascular risk factors seem to work synergistically. Th ird, cardiovascular illness seems to have direct effects on the integrity and function of the brain, which in turn affects cognition and neuropsychological function. Lastly, cardiovascular illness appears highly related to genetic and environmental factors. Historically, early studies (before the 1970s) examining the relationship between hypertension and cognitive decline have come to find strong, dramatic results (Waldstein et al., 1991). These early studies concluded hype rtension was related to very severe neuropsychological and cognitive deficits; how ever, many of these studies were poorly controlled (Waldstein et al., 1991). They faile d to account for variables such as age and

PAGE 24

24 education, and many involved samples of people with more severe cardiovascular disease, and not just hypertension (Waldstein et al., 1991). More recent studies with fewer methodological flaws have come to show relationships between hypertension and cogniti ve decline, but these findings have often been inconsistent (Waldstein et al., 1991). While many studies have found high blood pressure and cognition to be related, others have found low blood pressure to be more related to cognition, or no relati onship between blood pressure and cognition at all (Waldstein et al., 1991). Additionally, there have been inc onsistencies with regard to the various neuropsychological domains most affected by hi gh blood pressure; for instance, some studies find relationships between memory or attent ion and blood pressure, while others may not (Verghaegen et al., 2002; Madden et al., 2003). Furthermore, studies have found various patterns of neuropsychological and c ognitive deficits, such that relative effects of blood pressure on neuropsychological domains are difficult to di scern (Waldstein et al., 1991). Overall, relationships between blood pr essure and cognition have b een found cross-sectionally and longitudinally, while chronicity of hyperten sion seems to be an important predictor longitudinally (Waldstein et al., 1991). Specifically, much research suggests high blood pressure negatively affects cognitive func tion (Deary et al., 1998; Elias et al., 2003; Elias et al., 2004; Starr & Whalley, 2005; Waldstein et al., 1991; Waldstein & Katzel 2001; Waldstein et al., 2005; Nilsson et al., 2004), with the effects of hypert ension exacerbated by ot her factors, including APOE 4 allele status (Carmelli et. al., 1998) and ob esity and hyperglycemia (Elias et al., 2003). Across various studies, a common finding is th e U-shaped effect of blood pressure, with regard to actual blood pressure levels and age (Carmelli, 1998; Launer, 2000; den Heijer, 2002; Madden, 2003). High blood pressure is associated with worse performance across various cognitive domains (i.e., attention, executive functions, memory and learning, processing speed,

PAGE 25

25 visuospatial and perceptu al functions, etc.) and higher dementia incidence, and this relationship is strongest in middle-aged adults both cross-s ectionally (Madden et al., 2003) and longitudinally (Carmelli et al., 1998; Launer et al ., 2000; Lopez et al., 2003). Nonetheless, low blood pressure levels are also associated with poorer cognitiv e performance and higher dementia risk; some work has shown that while high blood pressu re at baseline may predict poor cognitive performance at baseline, low blood pressure at follow up may be more predictive of cognitive decline over time (Ruitenberg et al., 2002; Zuccala et al., 2002; Launer et al., 2000). Finally, low blood pressure seems to be more influential on cognitive decline in older adults than in younger age groups, and some suggest this is du e to hypoperfusion in the brain (Waldstein & Katzel, 2001). Underlying physiological mechanisms A common finding in this literatur e is the direct effect cardi ovascular disease has on brain integrity and function, which in turn negativel y impacts neuropsychological function (Pantoni et al., 1999; D eary et al., 2003; Waldstein & Kat zel, 2001; Raz, 2003; den Heijer, 2003; Kramer, 2001). Cardiovascular conditions, particularly hypertension, cause changes in overall weight, plus changes in brain vasculature (Raz, 2003; den Heijer, 2003; Pant oni, 1999; Kramer, 2001). There is evidence for hypertension-related atrophy in prefrontal brain re gions (Raz, 2003). This finding is significant since, as aforementioned, prefrontal brain regions are responsible for executive control functions, which are vulnera ble to both normal aging and poor cardiovascular health. Furthermore, there is evidence hypertension and other cardiovascular illnesses exacerbate atherogenesis, arterial stiffness, white matter changes, and there is evidence linking them to large and small vessel infarcts, silent la cunar infarcts, and ischem ic attacks (Benetos et al., 2002, Patoni, 1999; Kramer, 2001; Waldstein & Katzel, 2001). These small vessel infarcts

PAGE 26

26 and silent lacunar infarcts often occur in the d eep and periventricular white matter brain regions, leading to observable executive c ontrol deficits (Libon et al., 2004). Raz and colleagues (2003) and den Heijer et al (2003) have demonstrated the relationship between high blood pressure and pathological brain changes; in their work, hypertension was related to increased brai n atrophy. Even further, Raz et al. (2003) found atrophy of the prefrontal brain regions to be most related to high blood pressure, while den Heije r et al. (2003) showed there was a U-shaped relationship between hypert ension and brain atrophy. Thus, both high and low blood pressure levels were both found to be associated with disproportionate amounts of brain atrophy (den Heijer, 2003). Pantoni (1999 ) has argued evidence demonstrates vascular cognitive impairment tends to be more frequent in individuals with white matter lesions, as well as in those with cereb ral circulatory dysfunction. Sim ilarly, Waldstein and Katzel (2001) reported evidence suggesting the effects of decrea sed cerebral blood flow, related to increased peripheral vascular resistance in blood vesse ls in the body, may also be an underlying mechanism in this relationship. It has also b een reported that increa sed activity of both the sympathetic nervous system and hypothalamic-pitu itary-adrenocortical axis is involved in stress responses to environmental challenges, and this appears to play a ro le in the hypertensioncognition relationship (Waldstein & Katzel, 2001). As was explored further in the present study, these physiological brain changes have direct implications for cognitive function and declines mediated by these brain areas, which explain the pattern of neuropsychological deficits often observed (i.e., deficits in executive function, processing speed and attention, more so than de ficits in memory and language) (Pantoni, 1999; Deary et al., 2003; Waldstei n & Katzel, 2001; Raz, 2003; de n Heijer, 2003; Kramer, 2001).

PAGE 27

27 Synergistic effect of hypertension and other cardiovascular conditions Another common finding across studi es in this area is that cardiovascular risks seem to have synergistic effects on cogni tion (Elias et al., 2003; Martins et al., 2006; Waldstein & Katzel, 2006). One study found the effect of hypertension and obesity in some individuals, as well as the effect of hypertension and hyperglycemia in others, more negatively affected cognitive function than those with either hypertension, obesity, or hyperglycemia onl y (Elias et al., 2003). Furthermore, another study showed the effect of hypertension was more severe for individuals with one or more APOE 4 alleles than for thos e with hypertension or AP OE 4 status alone, or with neither condition (Carmelli et al., 1998). C onversely, individuals with APOE 4 and treated with antihypertensive medication showed better cognitive performance those with APOE 4 and untreated by antihypertensives in another study (Hestad & Engedal, 200 6). In general, it appears to be the case many cardiovascular illnesses and diseases are comorbid, and this comorbidity likely results in worse cognitive and medical outco mes. The present study explored whether an exercise promotion intervention improved these comorbid conditions eith er together, or in isolation, and how potential health imp rovements affected cognitive function. Environmental factors Finally, a summation resulting fr om various findings in th is literature regarding the relationsh ip between cardiovascular illness and c ognition is the strong evidence that genetic and environmental factors seem to greatly influen ce cardiovascular risks (e .g., Waldstein & Katzel, 2001). Certain genetic factors include APOE 4 status, family history of cholesterol, blood pressure, and glucose/insulin (met abolic disease) levels (Walds tein & Katzel, 2001; Martins et al., 2006). These all influence the expression of cardiovascular dis ease and may act as predisposing factors, if concep tualizing genetic influences from a diathesis-stress model framework. In addition, the environment plays a role in cardiovascular risks. Dietary habits and

PAGE 28

28 exercise and physical activity are all obvious influences (Waldstein & Katzel, 2001); nevertheless, social environment also plays a role (Seeman & Crimmins, 2001). Individual socioeconomic status is a major factor in mo rtality and morbidity rates among most diseases (Dowd & Goldman, 2006; Seeman & Crimmins 2001; Valanis, 1999). Lastly, social relationships are important to consider in the expression of cardiovasc ular and other illness (Seeman & Crimmins, 2001). Studies show social relationships to be stable across the lifespan, and the more social integration one has, the lo wer the risk for mortality (Seeman & Crimmins, 2001). Nevertheless, to the extent social relationships cause st ress in a person and cause the development of poor health habits (diet, exerci se, etc), social interac tion can have a negative health impact (Dowd & Goldman, 2006; Seeman & Crimmins, 2001). The present study addressed many of these enviro nmental factors that play a role in not only cardiovascular, but cogniti ve risks, and disentangle th e often confounded relationship between physical fitness and ps ychosocial exercise outcomes. Self-Perception Influences on Late Life Cognition Depression and Well-Being Depression is not a norma l part of aging, but it is considered one of the more common mental health concerns in the elderly. In ol der adults, depression is the second most commonly diagnosed mental disorder (LaRue, 1992). Accord ing to the National Institu tes of Health (NIH), major depression affects about 2 million of the 35 million Americans 65 and older, and another 5 million suffer from depression symptoms; however, only about 10% receive treatment (Anthony & Arboraya, 1992; Gatz et al. 1996; NIMH, 2003). While clinica lly diagnosed depression has lower prevalence rates in older ad ults, some studies have observed a curvilinear relationship with age (Haynie et al., 2001). High depression scores in young adulthood, lower in middle age, then higher in old age. Certain factors noted to acc ount for age differences in depressive symptoms

PAGE 29

29 include physical disability, chronic illness, and loss of close re latives (Haynie et al., 2001). However, chronological age is not necessarily the critical variable for depressive symptoms in late life, and instead, it has been found that negative changes in health and psychological functioning are associated with depression (Haynie et al., 2001). There is also much suggestion in the literature that depression may differ in ol der adults, such that diagnostic criteria may underestimate the prevalence of the disorder in the elderly. Thus, a minor depression category may be more appropriate to characterize depr ession in older adults (Knight, 2004). Minor depressive disorder is listed in DSM-IV-TR (American Psychological Association, 2000) as needing further study and is defined as fewer sy mptoms than major depression and lacking the 2 year timeframe for dysthymic disorder (Knight, 2004; Blazer et al., 1991; American Psychological Association, 2000). Furthermore, some have argued that a depletion syndrome of depression, marked by loss of interest and fa tigue, rather than intensely depressed mood and guilt, is a better way of conceptualizing and ch aracterizing late life depression (Newmann et al., 1991; Blazer, 2003). Complaints that physical pr oblems have gotten worse (i.e., arthritis or headaches) is often a predominant symptom, and symptoms of anxiety or irritability may be present (Newmann et al., 1991; Blazer, 2003). A nother term in the lit erature is depression without sadness, such that instead of feeling sa d, a depressed older adul t may complain of low motivation, lack of energy, or physical problems (Newmann et al., 1991; Blazer, 2003). This may be characterized by withdrawal, apathy, and a lack of vigor. In general, older depressed adults are less likely to admit dysphoria, guilt, or suicidal ideation compared to young depres sed adults (LaRue, 1992). They instead present with somatic symptoms and a lack of positive feelings (LaR ue, 1992). In addition, depression symptoms in the elderly are difficult to inte rpret because they may mimic or be a precursor to medical

PAGE 30

30 conditions (Gatz, 2000). Late life depression is often comorbid with chronic medical illnesses or disabilities and can precede or be a risk for the development of illness (Herbst et al., 2007). Study has suggested that lifetime major depression may be associ ated with increased risk of coronary heart disease in older adults (Herbst et al., 2007). Depression ha s also been shown to increase risk of death while decreasing the ability to benef it from rehabilitation. This suggests the prevalence of depression in the elderly may be higher than current estimates. Furthermore, older adults experienci ng depression symptoms may experience similar effects of Major Depressive Disorder (the clinical diagnosis for depres sion), including health problems, malnutrition, disabilit y, functional and cognitive impairm ent, and increased mortality rates (Gatz, 2000). Similar to the effect of normal aging on cognitive function, depression tends to be associated with difficulties perf orming cognitive tasks. There have been a number of terms in the literature that refer to the potentially revers ible nature of cognitive impairments associated with depression in older adults. Such terms include pseudodementia (K iloh, 1961) and dementia syndrome of depression (Folstei n & McHugh, 1978). However, thes e terms may not be the most appropriate, because the cognitive effects of depression can range in scope and severity and often not meet the criteria of dementia (Houston & Bo ndi, 2006). Instead, the term depression-related cognitive dysfunction (Stoudemire et al., 1989) has b een proposed as a better alternativ e, since it refers to depression and is general enough to encompass a range of cognitive deficit severity. Regarding specific cognitive deficits common wi th depression, depressed individuals show deficits on controlled, effortful tasks (such as those highly dependent on executive functions), while verbal skills and other automatic processes are relatively spared (e.g. Hartlage, Alloy, Vazquez, & Dykman, 1993). Overall, there is consistent evidence for neuropsychological

PAGE 31

31 deficits in depression for older adult samples on tasks of execu tive function, visuospatial skills, and psychomotor speed (Houston & Bondi, 2006). Moreover, depression is associated with impaired performance on a number of traditional neuropsychological tests purported to measure executive functioning (i.e., Wisconsin Card Sorti ng Test, Stroop Color-Word Test), and various working memory measures (Moritz et al., 2002). These executive functioning deficits predict difficulty performing instrumental activities of daily living (Cahn-Weiner, Boyle, & Malloy, 2002), which is often related to functional de cline in the elderly. One study sought to characterize the neuropsychological presentation of geriatric de pression and to determine if depression-related executive dysfunction was mo re pronounced in old age (Lockwood et. al, 2002). Results indicated for measures of atte ntion, there was no signi ficant age-depression interaction; however, for measures of inhibitory control and focused effort there was an agedepression interaction, with depressed older adults performi ng worst (Lockwood et. al, 2002). They discussed two models to explain their fi ndings: in one model, depression is viewed as unmasking executive dysfunction in patients with compromised frontal -subcortical pathways. In a second model, dysfunction of frontal-subcortica l pathways is thought to be predisposing to depression and executive dysfunction. Neverthe less, not all older depressed patients have significant cognitive dysfunction. These individu als may show only mild deficits, but often continue to have subjective memo ry complaints (Houston & Bondi, 2006). Vascular Depression Hypothesis One possible explanation for such depression-re lated executive control deficits ma y be due to the fact both depression and cognitive d ecline share a common car diovascular etiology (Alexopoulos et al., 1997). As such, the vascular depression hypothesis has been proposed as a way to explain the common link between depre ssion and cardiovascular health, as well as cognitive impairment. The hypothesis originally purported that cerebr ovascular disease can

PAGE 32

32 predispose, precipitate, and pe rpetuate depression in the elde rly and that neuropathology found in the white matter regions of the brain can be an etiological factor in the expression of depression (Alexopoulos et al., 1997; Alexopoulos, 2006). One investigation showed, through path modeling, a stronger relationship between cerebrovascular risk factors (i.e., hypertension, diabetes, and heart disease) and depression symp toms in a group of oldest-old (85 and older), after controlling for co-morbid health/medical c onditions and limitations (M ast et al., 2005). Comorbid health/medical conditions and limita tions mediated the relationship between cerebrovascular risk factors and depression in individuals age 50-84 in this sample (Mast et al., 2005). Additionally, a second study examining the ro le of cardiovascular risk factors found that among the oldest-old living in a retirement community, both depression and number of cardiovascular risk factors at baseline predicte d stroke (Krishnan et al. 2005). Depression accounted for twelve percent of the variance in stroke inciden ce and partially moderated the effect of cardiovascular risk f actors (Krishnan et al., 2005). Fu rthermore, studies have shown that white matter lesion load significantly predicts depression in older sa mples (e.g., Godin et al. 2008; Krishnan et al. 2004). In one cross-sectional study, white matter lesion volume was significantly associated with lifetime of major de pression after controlling for covariates (sex, age, hypertension, cardiovascular disease, and alc ohol and tobacco consumption) (Godin et al. 2008). Furthermore, in another cross-sectional study investigati ng the strength of association between various vascular-related neuropathology linking to depression (i.e., cerebral white matter hyperintensities signaling periventricular and deep white matter lesions), the findings demonstrated that deep white mater lesions was more strongly corre lated with depression symptoms as measured by the Geriatric Depressi on Scale (15-item Short-Form) (Krishnan et al., 2006).

PAGE 33

33 Just as depression has been linked to frontal lobes and ex ecutive cognitive function, the vascular depression hypothesis ha s also linked vascular-related depression to frontal/executive cognition (Alexopoulos, 2006). Clinically, vascular depression has been described as a medial frontal lobe syndrome (Kris hnan et al., 2004), with sympto ms including psychomotor retardation, apathy, and disability. Further, cerebrovascular lesions have been associated with persisting, unstable remission of depression and increased dementia risk (Alexopoulos et al., 2002). Two streams of literature have proposed and used separate terms related to the vascular depression hypothesis. These include subcortical ischemic depr ession and depression-executive dysfunction syndrome, and these terms differ in their assumed etiology (Alexopoulos, 2006). First, subcortical ischemic depression (Taylor et al., 2006) assert s that the subcortical impairment and associated depression is due specifically to cerebrovascular disease; however, the etiology of depression-executive dysfunction syndrome is le ss defined by specific factors (Alexopoulos, 2006). Instead, this syndrome may be the result of vascular disease, gene ral age-related changes, degenerative brain disease, or and accumulati on of these and other factors (Alexopoulos, 2006). Some in this field have called for internal c onsistency studies to be tter classify vascular depression as a syndrome and to decide upon a uniform etiology (Alexopoulos, 2006; Sneed et al, 2006). Nonetheless, regardless of a specific et iology, in the present stud y, it is expected that exercise-related improvements in cardiovascul ar health will work to improve depression symptoms, which is hypothesized to lead to cognitive improvements. Control Beliefs and Self-Efficacy In older adults, beliefs about c ontrol and self-efficacy are im port ance factors with regard to cognition. Study has suggested that control beliefs in late life show a significant relationship with cognitive performance (Mil ler et al., 1999). These findings highlight the importance of considering the impact cont rol beliefs and other backgr ound variables may have when

PAGE 34

34 decomposing age-related variance in cognitive performance (Miller et al., 1999). Additionally, late life cognitive performance has an impact on control beliefs, suggesting a bi-directional relationship. While cont rol beliefs have not gained as much attention in the fitness and cognition literature reviewed above, consider ation of these factors is importa nt in an exercise intervention context. Related to control beliefs is the concept of self-efficacy, which is among various theories of health behavior change (health be lief model and self-efficacy theory) that include cognitive elements related to decision-making. These cognitive elements of decision-making abilities are critical de terminants of engagement in health and/or exercise behaviors (Bandura, 1997). Such decision-making may reflect individuals actual barriers to exercise behavior, lack of motivation, or ambivalence toward exercise be havior. The present study will help individuals resolve ambivalence towards exercise through techniques (goal setting and mental imagery) designed to improve self-efficacy/control beliefs. Thus, an important test of our conceptual model will be to determine whether cognitive perfo rmance also benefits from such improvement in self-efficacy and control beliefs. Plasticity of Late Life Cognition Despite the various, and often detrimental influe nces, of physical and mental health on late life cognitive function, research exam ining both cognitive and physical training interventions with older adults have encouragingly found so me positive results fo r the improvement of cognitive functions in aged indi viduals. This evidence supports the notion of th e plasticity of late life cognition, which is relevant to the cu rrent proposal, which seeks to intervene through a physical exercise promotion intervention. Even in later life, cognitive function can show positive change, and there are both environmental and physical routes to such improvement. These environmental and physical interven tions will be briefly reviewed here.

PAGE 35

35 Cognitive Interventions Cognitive training (e.g., Jobe et al. 2001, Ball et al., 2002) has been shown to lead to an im provement in cognitive function. Early studies have examined the plasticity of late life cognition in laboratory settings and found these interventions to be successful in improving cognitive function (e.g., Rebok & Bal cerak, 1989). In addition, thes e early investigations have shown cognitive interventions can lead to improved performance on specific mental and perceptual functions, as well as certain aspects of everyday cogn ition (Leirer, Morrow, Pariante, & Sheikh, 1998; Ball & Owsley, 2000). More recent study has continue d to explore the benefits of cognitive training intervention with older adults. Specifically, the Advanced Cognitive Training for Independent and Vital Elders (ACTIVE) Clinical trial, a randomized controlled trial of three cognitive intervention approaches (s peed, reasoning, and memory) for older adults, demonstrated strong, broad, and durable cognitive abil ity-specific training effects, comparable to or greater than the amount of c ognitive decline observed in othe r longitudinal studies, suggesting the interventions have the potenti al to reverse age-rela ted decline (Ball et al., 2002). There was minimal transfer to training effects to everyday activities (i.e., functional competence); however, through a two-year follow-up peri od, there was no evidence of a significant decline in ADL and IADL status (Ball et al., 2002). At a five-year follow-up period, th ere was evidence for transfer of cognitive training gains to IADL function, w ith individuals receiving cognitive intervention showing slower rates of functional declin e relative to controls (Willis et al., 2006). In general, improvements from cognitive traini ng interventions have been found to be very specific to the cognitive skill/domain trained, with few broader cognitive improvements (Kramer & Willis, 2002). This is a major limitation of cognitive training protocols when it comes to intervening upon functional aging-related declines and reduci ng the risk and/or progression of

PAGE 36

36 cognitive impairment disorders and dementias. Thus, the consideration of interventions with potentially broader cognitiv e effects is warranted. Physical Exercise Interventions: Cardiovascular Fitness It has well been established the effects of phys ical activity on health and disease (increased longevity, decreased risk of coronary heart disease, blood pressure, various cancers, type-2 diabetes, etc.; Sallis & Ow en, 1999) across the li fespan. For older adults, the engagement in physical activity can improve cardio vascular health, increase metabolism, and slow declines in bone mineral density (Singh, 2002). Since there is a higher rate of chronic disease in older populations, the health benefits of physical activity and exercise may be even greater in older adults (Pescatello, & DiPietro, 1993 ). Specifically, older adults ca n benefit from participating in regular, moderate-intensity exer cise (i.e., 30 minutes of brisk walking, five or more times a week) (Schiller et al., 2005). Even when physical activity is initiated late in life, health and mortality rates are still substant ially affected, even after accoun ting for factors such as smoking, family history, weight gain, and hypertension (Blair et al., 1995). Additionally, there is evidence that regular physical activity bene fits older adults who have chronic conditions as varied as cancer, cardiovascular disease, re spiratory disease, and dementia (e.g., Morgan &, Bath, 1998). Over the past several decades, research ha s examined the benefits of physical fitness training on cognitive function, partic ularly in older individuals (Colcombe & Kramer, 2003). As mentioned above, one route by which physical exer cise interventions may improve cognition is by improving cardiovascular fitness. Regular ph ysical activity can reduce adults risk of coronary heart disease (Haskell et al., 1992) cancer (Blair et al., 1989) o ffer protection against non-insulin dependent diabetes (Helmrick, et al. 1991), and reduce hyperc holesterolemia (Harris et al., 1991). Studies have also linked physical activity with dimi nished declines of functional capacity due to age and reduced risk factors asso ciated with falls in th e elderly (DiPietro, 2001;

PAGE 37

37 Liu-Ambrose et al., 2004), while older adults w ho engage in regular physical activity reduce their risk for mortality associated with chronic disease states and premature mortality (Bean, et al., 2004). Relationship between fitness and cognition Early studies of the relationship between cognition and fitness ma de group comparisons between physically fit and unf it individuals (McAuley et al., 2004; Kramer, Erickson, & Colcombe, 2006). Like many ear ly psychological investigations these studies were crosssectional in nature, and they failed to control for a host of factor s that influence results observed (McAuley et al., 2004; Kramer et al., 2006). Such studies came to conclu de that physically fit individuals showed better performance on cognitive measures than did physically unfit persons (McAuley et al., 2004; Kramer et al., 2006). Toda y, the dangers of making such conclusions are better understood. Factors such as age, educat ion, socioeconomic status race/ethnicity, may influence performance on cognitive measur es. Thus, improving on such methodological problems of these studies, more recent studies employed longitudinal designs and compared individuals undergoing an exercise intervention to those who were not (McAuley et al., 2004; Kramer et al., 2006). These results indicated that while many people demonstrated cognitive performance improvements in these studies, other participants did not (McAuley et al., 2004; Kramer et al., 2006). These inconsistencies ha ve left many a bit baffled and motivated to investigate the underlying truth further (Col combe & Kramer 2003; McAuley et al., 2004). Cited methodological reasons for these inconsiste ncies include reliance on self-report activity data, failure to distinguish be tween activities that are aerobi c and anaerobic in nature, poor assessment of duration, intensity, and freque ncy of exercise activ ity, non-exclusion of participants with subclinical dementia, and low statistical power (Kramer et al., 2006). Accordingly, Colcombe & Kramer (2003) conduct ed a meta-analysis of 18 studies that have

PAGE 38

38 examined this exercise-cognition lin k in older adults to understand these inconsistencies better. The goal of the meta-analysis was to test severa l hypotheses have been propos ed in the literature regarding the specific effects of exercise: atheoreticalno spec ific hypotheses; speed hypothesismeasures of speed of processing show the gr eatest improve with exercise; visuospatial hypothesisthese measures show the greatest impr ovements with exercise; controlled processing hypothesisthese sustaine d attentional tasks show the grea test improvements; and executive control hypothesisworki ng memory, problem solving, and inhibition type tasks show the most exercise-related gains (Colcombe & Kramer 2003). Results from the meta-analysis showed that indeed, exercise was related to the improveme nts in cognitive performa nce, with executive control processes showing the greatest exerci se related improvement (Colcombe & Kramer 2003). Women and older participants showed the most cognitive gains, as combined aerobic and strength training exercise programs showed the most benefits to par ticipants (Colcombe & Kramer 2003). Nonetheless, there are several methodological problems with this meta-analytic study. First, the meta-analysis included 18 st udies that had well over 50 different neuropsychological/cognitive measures. The authors allowed for multiple cognitive domain categorizations along the four ma jor hypothesized cognitive domain s (i.e., one measure could be considered a speed, visuospatial, and executive tasks all at once) (Col combe & Kramer 2003). While they control for the influence of execu tive control processing on performance on other domains, they fail to control for the influence of the other three domains on executive control processing. This meta-analysis also suffered from including some studies that were flawed. Palleschi et al (1996) examined the effect of 12 weeks of exercise on cognitive performance (MMSE, attentional matrix, verbal span, supraver bal span tests) in a group of 15 Alzheimers

PAGE 39

39 patients. Their findings suggested that performance on ne uropsychological measures was improved with exercise; nonethel ess, there was no control group comparison to determine whether these patients improvements were above mere practice effects. Also with such a small, clinical sample, these findings are difficult to ge neralize to other groups. Another study included in the Colcombe & Kramer (2003) meta-analysi s with methodological concerns compared the effects of exercise on cognitive performance in a group of depressed older adults (Khatri 2001). This study compared depressed elders who perf ormed 16 weeks of aerobic exercise to those prescribed anti-depressant medication and found while both groups improved in depression symptoms, only the exercise group showed improved performan ce on the Stroop Interference task and Visual Reproductions of the Wechsler Memory Scale (WMS) (Khatri 2001). As with the previous study, there were was no control for factors that woul d confound the results, particularly the mere personal contact involved in being in a study, or prac tice/re-test effects. Furthermore, in a study of patients with chroni c obstructive lung disease (COPD), they examined whether aerobic exercise woul d result in improved performance on cognitive and psychological outcomes (Emery et al., 1998). These results sh owed that while patients did not improve lung functioning with exercise, there were improve ments in depression symptoms, though not more than controls (Emery et al., 1998). Also, the only cognitive improvements above re-test effects were found for verbal fluency performance, which is thought to have a strong frontal-executive component (Emery et al., 1998). These three st udies all contained clin ical samples and were include in the Colcombe & Kramer (2003) meta-analysis, along with 15 other studies of normal, healthy older adults, which may obscure the clean examination of the true exercise-cognition link.

PAGE 40

40 To improve upon these methodological flaws, the present study will examine a sample of community-dwelling, cognitively intact elders, w ithout major cognitive, physical or mental health conditions. Additionally, a stream of re search suggests that complex social activity participation may be a prot ective factor for cognition (e.g., Brown, 1992; McAuley, 1993). Elders who have had more complex, engaged li festyles, higher levels of education, and who are more active in later adult hood have been shown to perform at a higher level (cognitively), and may evince attenuated rates of cognitive decline. For many adults, exercise is a possible source of complex activity; it may bring some a dults out of house, provide them with complex regimens to monitor, and may increase their soci al participation with other exercisers (mall walking, exercising at a fitness facility, etc.). In order to account for the effects of complex social activity, the control condition of this exer cise promotion intervention will be designed to control for the influences of complex social engagement on improvements in both cognitive and affective functioning. Psychological benefits of exercise interventions A strong body of research suggests that along with the physical health be nefits of exercise in older adults, there are also strong psychological benefits of engageme nt in physical activity. Particularly, there is strong evidence that participation in regular physical ac tivity is associated with psychological health and well-being fo r older adults age 50 and older (Brown, 1992; Elavskey et al., 2005; Elavskey & McAuley, 2007; McAuley, 1993; Motl et al., 2005; Morgan & Bath, 1998; Pescatello & DiPietro, 1993). The engagement in structured and unstructured physical activity regimens are related to reduc ed symptoms of depression and anxiety and improvements in mood in older adults (Fukukawa et al., 2004). In sum, increasingly the field has come to accept the link between physical ac tivity and a variety of positive emotional, behavioral, and physical hea lth outcomes for older adul ts to be a viable one.

PAGE 41

41 Furthermore, control beliefs and self-efficacy are also influenced by physical exercise interventions. Ones self-perception beliefs ha ve a critical impact on whether or not he will engage in an exercise or hea lth intervention. The major social -cognitive theories of exercise include the basic set of soci al-cognitive principles: self-efficacy expectations, outcome expectations, outcome values, and intentions to change, modify, or initiate a certain behavior (Maddux, 1993). These factors are key mediators of the relationshi p between exercise behaviors and increased fitness, and these factors must be addressed and often intervened upon in order to ensure individuals will participate and adhere to any physical health intervention (Bandura, 1997). Thus, in the present study, these factors will be directly addressed by the exercise promotion intervention, with the aim of im proving self-efficacy/control beliefs, thereby increasing exercise participation and desire d physical fitness and cognitive outcomes. Conclusion As the preceding rev iew of the literature has discussed, there may be multiple routes by which increasing exercise contributes to cogn itive improvement. Exercise intervention may effect cognitive change through improvement of physical factors (particularly cardiovascular fitness) and/or improvement of psychosocial factors (self-perception influences, such as depression/well-being, control beliefs, and self-efficacy). Extant research has not yet paid adequate attention to disentangling these multip le pathways, thus the present study seeks to disentangle such exercise-related effects on cognition in older adul ts, thereby building on existing literatures, which have mostly examined these.

PAGE 42

42 Figure 2-1. Conceptual model of the effects of exercise intervention on cognitive function Exercise Promotion Intervention Increased Exercise Participation Changes in self perception (e.g., reduced depression and higher self-efficacy) Improved Cognitive (Executive) Function Changes in physical fitness (improved performance on fitness tasks )

PAGE 43

43 CHAPTER 3 METHODS Overview To achieve the specific aims outlined above, the present study was part of a larger, multidisciplinary, pilot study examining the effects of an exercise pr omotion intervention on exercise behavior/beliefs and fitness in sedentary, community-dwelling adults aged 50 and older. To keep the scope of the current dissertation ma nageable, the present study focuses on immediate pre-test-post-test changes in c ognitive functioning in adults w ho did and did not receive this intervention. The larger project assessed dyna mic (intraindividual) cognitive and psychosocial changes throughout the period of intervention, co nsidered other outcomes (cardiorespiratory fitness, exercise behaviors, etc.), and also has a longer-term followup (12-months) of the cognitive and non-cognitive outcomes in this study planned. Two randomized groups of (35 in the control group, and 34 in the intervention group ) adults were recruited from the Gainesville/Alachua County, Florida region. Both the control and intervention groups underwent preand post-intervention cognitive, fi tness, self-efficacy, sleep, and emotional wellbeing assessment. Additionally, the intervention was such that the control and intervention participants were divided into smaller replicates of five to twelve pe ople, with trained peer mentors. There was quality cont rol monitoring of the interventi on and control groups to insure fidelity of protocol adherence. The exercise promotion (experimental) group received 16 weeks of intervention (motivational interviewing, h ealth and fitness education, w eekly peer mentoring and group support session, adaptive goal se tting for individual performances improvements, mental imagery, daily home-based self-monitoring, etc.). The intervention was informed by current state-of-the art interventions fo r sedentary adults and was custom ized for individual fitness goals

PAGE 44

44 and avoidance of individual exer cise barriers. A control/compari son group received 16 weeks of health hygiene instruction, consisting of education about general heal th issues relevant to older populations (i.e., osteoporosis, Alzh eimers disease, nutrition). There was one session devoted to non-specific discussion of exercise /physical activity. Both groups received unlimited use/access to a fitness facility ; amount of exercise was monitored in both groups, such that the incremental benefits of the treatment could be carefully assessed. Participants The samp le of participants consisted of 69 middle-aged and older adults who were sedentary and community-dwelling. Participants were recruited in the local Gainesville/Alachua County, Florida area, primarily through the us e of local newspaper advertisement in the Gainesville Sun, Senior Times, and the Gainesville Regional Utilities (GRU) newsletter. Additionally, flyers were posted at locations around the Gainesville and University of Florida communities, and flyers were mailed to individua ls on the University of Florida Older Adult participant registry. To attempt to maximize r ecruitment of diverse popula tions, an ad was also placed in the Gainesville Guardian, a publication targeted to th e African-American community. Finally, recruitment extended to community organizations, including a seniors group a at local community college, a local church, a retirement community affiliated with the university, a grandparents support group, local health fairs for se niors, and a direct mailing list. Figure 3-1 outlines the number of participants that were init ially contacted and/or expressed interest in the study, screened by telephone, randomized to the study conditions, and included in the present analyses. Statistical Power Considerations Table 3-1, displays the results of a po wer analysis conducted to exam ine the study designs ability to detect eff ects with 32 participants per gr oup, alpha = .05 (one-tailed;

PAGE 45

45 expectations are direc tional), and several effect size possi bilities. This is a lower-bound estimate, since it is based on a post-test-only co mparison; the actual anal yses were conducted as a mixed within-between analysis of variance. With expected hi gh test-retest reliability in the primary outcome measures (executive cognitive f unction), under the same effect sizes, power would have exceeded that reported here. While cognition was a primary outcome in this dissertation study, it was a secondary outcome of the larger trial in which it is embedded. The study was therefore powered to affect primary outco mes of the larger trial (exercise participation, exercise self efficacy, and fitness) In the event that, as a seconda ry outcome of the larger trial, power is not adequate to detect significant differences on cognitive outcomes, the study will still be useful as a preliminary study for future work (specifically, to aid in effect size determination for future studies). In additi on, the correlational aims of this study (to examine the extent to which fitness changes and psychosocial changes serve as proximal predictors of cognitive changes) did not require signifi cant intervention effects to be tested and have substantially greater power. Sample Characteristics Table 3-2 describes participant characteristics. Overall, the sample had an average age of 63.9 years. The vast majority of the sample id entified themselves as Caucasian/White (91.3%) and female (84.1%). There was an under-representat ion of male participants, as is customary in cognitive aging research, as we ll as ethnic minorities (i.e., Af rican-American/Black, Hispanic, and Asian). On average, the sample was colle ge-educated (16.2 years of education), and the average estimated IQ was 113, which falls in the hi gh average range of inte lligence. The sample reported a minimal level of depr ession and anxiety symptoms.

PAGE 46

46 Cognitive Exclusions All potentia l participants were screened by te lephone to exclude individuals based on the following cognitive criteria: dementing illness, lifetime history of significant head injury requiring hospitalization, other neurological or major medical illnesses (Parkinsons disease, epilepsy, stroke, cancer (and/or chemotherapy and radiation above th e chest)), severe uncorrected vision or hearing im pairments, inpatient psychiatri c treatment, extensive drug or alcohol abuse, any use of an anticholinesterase inhibitor (such as Aricep t), or unavailability at future follow-up time points. Telephone screen ing included the 11-item Telephone Interview for Cognitive Status (TICS; Brandt, Spencer, & Fols tein, 1988) for a standardized assessment of cognitive status. The TICS has a sensitively of 94% and a specificity of 100%. The cut-off score of 27 points was used to exclude demented individuals from the study (Brandt, Spencer & Folstein). The TICS, embedded in the general telephone screening protocol, is included in Appendix A. Physicians Permission To assure individuals were properly excluded due to cardi ovascular/physical conditions/diseases and me dications presented below, participan ts were required to submit a completed and signed nurse/physicians checklist and permission form prior to enrolling in the study. This checklist is included in Appendix B. Physical/Cardiovascular Exclusions Exclusion criteria for dis eases or conditions likely to adversely affect the safety of elders in the exercise promotion intervention are as follows: terminal illness with life expectancy less than 12 months, cardiovascular disease (myocardial infarction in last 6 months, chronic heart failure, aortic stenosis, history of cardiac arrest, implanted cardi ac defibrillator, or uncontrolled angina), pulmonary disease requ iring oxygen or steroid treatment, and ambulation with assistive

PAGE 47

47 devices (such as cane, walker, or wheelchair). Exclusion criteria for fact ors adversely affecting compliance with study protocols and intervention s include: inability to read and understand English, not willing to consent to random a ssignment to one of the two study conditions, hospitalization for current psychiat ric illness, daily alcohol consum ption of more than two (for women) or three (for men) drinks, and h earing or speech impairments making verbal communication difficult. Medication Exclusions This study excluded individuals taking calcium channel blockers or beta-blockers. The mechanisms of these drugs are likely to adversely affect the safety of older adults participation in an exercise intervention, since they interfere with heart rate reactivity during exercise. Inclusion: Staging Algorithm The Stages of Exercise Change Questionnair e (S ECQ; Reed, Velicer, & Prochaska, 1997) was administered during the initi al phone contact to assess part icipants stages of change regarding the initiation of a physical exercise program. The SECQ contains five orderedcategory items, which assess change readin ess along a continuum: pre-contemplation, contemplation, preparation, action, and maintenance. Potential pa rticipants in the action and maintenance stages of physical activity were ex cluded, while persons in the earliest stage (i.e., pre-contemplation) were unlikel y to volunteer/be compliant w ith an exercise intervention, by definition. Procedure Overview The planned study timeline for a single particip ant was divided into approxim ately twenty weeks (Table 3-3). In the first two weeks, participants were scr eened, and eligible persons were enrolled into the study to undergo baseline assessment. This assessment included cognitive,

PAGE 48

48 neuropsychological, psychological, and exercise/fitness testing. After randomization to either the control (health hygiene) or experimental (exercise promotion) condition, weeks three through eighteen (sixteen weeks total) of the study involved the intervention period. However, it is important to note that several factors determined the actual study timeline for each participant. The actual study timeline for each par ticipant was influenced by his or her availability to devote sixteen weeks to participation in a group (i.e., schedule conflicts, vacation, etc), the status of participant recruitment (all the av ailable slots of each group needed to be filled to maximize use of resources), the status of required completion and receipt of physicians permission form, and the ability of study staff to screen participants in a timely manner. Due to these factors, there was a lag of several months between initial participant contact, completion of telephone screening, and/or baseline testing and the start of the study group period for most participants. Specifically, there was an average of 5.6 weeks between baseline testing and the start of the small group sessions (range 0 to 26.6 weeks). Af ter the 16 week intervention period, most (66 out of 69) participants were pos t-tested within one or two weeks following the end of the group sessions. The remaining 3 participants were te sted longer than two w eeks post-intervention due to scheduling conflicts (r ange 3 to 6.4 weeks). During the intervention period, all participants received a free membership to either a University of Florida campus fitness facility or a community-based, church fitness center. There were a total of eight replicat es of peer/support groups, with on e control and intervention group each per replicate for a total of sixteen small p eer/support groups. In total, 47 participants were randomly assigned to the control condition, while 44 participants were assigned to the intervention condition. In addi tion, study groups for replicates 1-5 were composed of people randomly assigned to peer groups including other individuals in their same age range (50-64

PAGE 49

49 years or 65 and older) (Table 34). For replicates 6-8, the groups were age-mixed in order to meet the larger studys recruitment goals more quickly. Due to attrition of the sample, 35 control group participants and 34 intervention group participants were included in the present analyses. A detailed description of the intervention is found in a later section. Finally, following this period, participants underwen t post-intervention assessment. Table 3-5 outlines the pre-post assessment protocol (organized in a thematic grouping, not test order). Executive cognitive measures, which are thought to be more sensitive to physical fitness improvements (Colcombe & Kramer, 2003), are shown in the first row. Rationale for Measures Neuropsycho logical measures were selected to assess aspects of cognitive function expected to be affected by physic al exercise intervention, as well as those expected to remain stable with exercise. Specifically, following th e reviewed literature, ex ecutive control processes are hypothesized to be most positively influe nced by improved aerobic fitness (Colcombe & Kramer, 2003); however, other cognitive functions such as estimated intellectual ability and memory for structured information (story memory), are hypothesized to be remain relatively unchanged by improved aerobic fitness; these ar e included to help evaluate the putative specificity of exercise effects. Psychosocial, exerci se, and aerobic fitness me asures were selected to measure the physical and psychosocial componen ts of exercise participation, as outlined by the conceptual model above, that are potential mediating variables in the relationship between exercise and cognitive function. The specific rationale for sel ecting each cognitive measure will be described in the section that follows. Detailed descriptions of each measure are as follows:

PAGE 50

50 Primary Outcomes North American Adult Re ading Te st (NAART) The NAART (Blair & Spreen, 1989) was used to measure pre-morbid intelligence. It was developed specifically for use with adults su spected to have compromised cognition. The NAART consisted of 61 irregular, rare words that participants were asked to pronounce. The NAART has been found to correlate between 0.40 a nd 0.80 with other measures of intelligence. Test-retest reliability of the NAART has been established at 0.92 within one year. The score used in was derived from using a prediction equa tion that took into acco unt the number of words incorrectly pronounced. The NAART was selected as a non-executive cognitive measure in the present test battery because it is thought to be a measure of pre-mo rbid intellectual function (i.e., crystallized abilities), which w ould be expected to be relativel y stable and less sensitive to improvements in physical fitness over time. Logical Memory subtest of the Wechsler Memory Scale, 3rd Edition (WMS-III) The Logical Memory subtest of the Wech sler Memory Scale-Third Edition (WMS-III) (Wechsler, 1997) measured the ability to learn and retain verbal memories for brief stories. Individuals heard a brief story of 25 propositions and were asked to recall as many story propositions as possible in an im mediate recall of the story. Part icipants received one point for each proposition correctly recalled. Next, a seco nd story of 25 propositions was read, followed by an immediate story recal l trial. This second story was r ead twice to assess verbal learning slope. Delayed recall and rec ognition trials for each story were completed following a 25-35 minute delay interval. The LM subtest was se lected as a measure of non-executive cognitive function because it measures episodic memory abili ties in a structured way. Relative to word listlearning memory tasks, LM is thought to rely less on executive functioning, since individuals do

PAGE 51

51 not have to impose a semantic organization on the information to maximize recall. It has been found to be a reliable memory measure for use in adult and older adu lt groups (Wechsler, 1997). Control Oral Word Association (COWA) To assess spontaneous word production, the COWA (Benton & Ham sher, 1989) was administered. Participants were read a letter (F, A, and S) and asked to generate as many non-proper nouns beginning with that letter as possible within sixt y seconds. Participants were not allowed to repeat previously stated words during each trial. The rationale for selecting COWA as an executive cognitive measure was du e to its sensitivity to frontal lobe/executive function (Salthouse et al., 2003). Previous lite rature has identified the COWA as a gold standard of executive functi on (Crawford et al, 2000 & 2005). Trail Making Test A and B (Trails A, Trails B) The Trail Making Test (Trails A and Trails B) (Reitan, 1992 ) assessed attention, working m emory, psychomotor speed, visual scanning and sequencing, and cognitive flexibility by requiring individuals to connect circles containing numbers (Trails A) and numbers and letters (Trails B). Trails B was conceptualized as a more demanding task, due to the increased cognitive flexibility and executive skills required for successful task completion. The rationale for inclusion of the Trail Making Test was due to both being one of the most widely used measures in clinical neuropsychological practice (Rabin et al., 2005) and due to its documented high sensitivity to brain function (Reitan & Wo lfson, 1994). Trails A was categorized as a nonexecutive measure because it is conceptualized as a measure of attention and processing speed that is less reliant on working memory, as is Trai ls B. Trails B was selected as an executive cognitive measure because of th e added mental flexibility and working memory required to complete this task (Lezak et al., 2004).

PAGE 52

52 Letter-Number Sequencing subtest of the Wechsler Memory Scale, 3rd Edition (WMS-III) The Letter-Number Sequencing subtest of the WMS-III (Wechsler, 1997) is a verbal working memory task that requir ed individuals to repeat a gr oup of numbers and letters after mental manipulation. Individuals heard a sequence of numbers a nd letters (ranging from 2 to 8 numbers/letters) and were then required to respon d such that all numbers were stated first, in numerical order, followed by all lett ers stated in alphabetical order. This task was selected for inclusion in the cognitiv e battery as a measure of executi ve function, particularly working memory and attention. The Letter-Number Sequencing subtest has been found to be a reliable measure of executive function for use in adul t and older adult populations (Wechsler, 1997). N-Back task The N-back task was the only cognitive measur e administered by computer. In the N-Back task, participants saw a single lett er in 48-point Arial font in th e center of the computer screen and decided whether that letter matched a target letter. They were told to respond as fast and accurately as they could. The computer collected both accuracy and response time information. In the One-Back condition, participants j udged whether the current letter matched the immediately preceding letter. They were instru cted to push the Yes button indicated on the computer keyboard when that lett er appeared, and the No button when the letter did not match the previously viewed letter. In the Two-Back conditi on, participants judged whether the current letter matched the letter presented two letter s previously. Letters remained visible until a response was made. There was a 1 sec. inter-stim ulus interval (Cohen, et al., 1994). The N-back task was selected as a measure of executive functi on due to previous work that has suggested its sensitivity to brain activation in dorsolateral prefrontal regions, thought to sub-serve executive control processes, in the brain (Cohen, et al., 1997).

PAGE 53

53 Proximal Outcomes Geriatric Depression Scale (GDS) The GDS (Yesavage & Brink, 1983a) was ad mi nistered to assess depressive symptomatology. The GDS is a 30item self-report scale of yes/no quest ions about symptoms of depression (Do you feel that your life is em pty?). A point was given for each depressive symptom endorsed. This measure has been sh own to be a reliable and valid measure of depressive mood in older adults. Clinical cut-o ff scores for this measure are as follows: 0-10 normal depression symptoms, 11-14 mild depr ession symptoms, 15-21 moderate depression symptoms, and 22+ severe depression symptoms (Yesavage & Brink, 1983a). The GDS is a commonly used self-report measure of depression that has been validated for use in older adult samples (Yesavage & Brink, 1983b). Beck Depression Inventory2nd Edition (BDI-II) The BDI-II (Beck, Brown, & Steer, 1996) was al s o administered to assess depressive symptomatology. The BDI-II consists of 21 groups of statements related to cognitive and somatic depression symptoms (Sadness, Changes in Sleeping Pattern, etc.). Individuals selected from one of four statements the one best describing the severity of their symptoms over the past two weeks (0 = I do not feel sad, 1 = I feel sad mu ch of the time, 2 = I am sad all the time, 3 = I am so sad or unhappy that I cant stand it). The scale has a clinical ra nge of 0-13 for minimal depression symptoms, 14-19 points for mild depression symptoms, 20-28 for moderate depression symptoms, and 29-63 se vere depression symptoms. The BDI-II is a commonly used self-report measure of depression th at has been shown to have reli ability and validity with adults, both younger and older than 65 years (Beck et al., 1996; Brink et al., 1983).

PAGE 54

54 State-Trait Anxiety Inventory (STAI) The STAI (Spielberger, 1983) was admi nistered to assess curren t (state) and typical (trait) anxiety symptoms. The STAI consists of 40 stat ements to which participants responded based on the degree to which they felt that way. Fo r the first 20 statements (I feel calm, etc.), participants responded with respect to how they felt right now (i.e., not at all, somewhat, moderately so, or very much so), while for the second 20 items (I feel pleasant, etc.), they responded according to how they generally felt (i.e., almost never, sometimes, often, almost always). Clinical cut-offs depend on age and gender corrected normative data. The STAI has been used in numerous studies of the relationships between an xiety and various psychological constructs (Spielberger, 1983). Exercise self-efficacy Two measures were used to measure exercise self-efficacy expectati ons. Both measures were based on social cognitive theories, particularly Banduras theory (1997) of self-efficacy. The Barriers Self-Efficacy Scale (McAuley, 1992) is a 13-item measure of perceptions of confidence to maintain regular ex ercise (at least three times a week), despite common exercise barriers. The Exercise Self-Efficacy Scale (McA uley, 1993) is an 8-item measure of confidence to maintain exercise (three times a week at m oderate intensity for 40 minutes) consecutively for a period of 8 weeks. Both measures have been shown to be predictive of exercise behavior and have adequate internal consistency across many research studies, incl uding those with older adults (Blissmer & McAuley, 2002; McAuley et al., 2003; McAuley et al., 2005). Leisure Time Exercise Questionnaire (LTEQ) The LTEQ is a three-item scale that asked indi viduals to rate how often they engaged in mild, m oderate, and strenuous leisure-time exer cise (Godin & Shephard, 1985). The LTEQ is a reliable and valid measure of adult exercise beha vior, and it allows for th e calculation of a total

PAGE 55

55 MET score, by weighting the three intensity leve ls and summing: 3(mild activity) + 5(moderate activity) + 9(strenuous activity). Pr evious research has supported th e validity and reliability of LTEQ score interpretations with adult (Godin et al., 1986; Jacobs et al., 1993) and older adult populations (Karvinen et al., 2007; Ruppar & Schneider, 2007). Minutes of Moderate and Vigorou s Physical Activity (MVPA) Minutes of moderate and vigor ous physical activity (MVPA) were compu ted from the LTEQ by adding the number of moderate and strenuous bouts reported and multiplying by 20. This value was then added across the week to obtain a weekly measure of MVPA. Recently, to better approximate the true level of activity needed fo r reliable health benefits (moderate to vigorous; Pate et al., 1995), study has s uggested that minutes of mild activity be excluded from calculations of physical activ ity (Karvinen et al., 2007). Pedometer The AE 120 pedometer (Ya max SW200 engine) was used as a second measure of physical activity. The pedometer was worn on participan ts hip to measure steps taken during a given day, and participants recorded the final number of steps for each day the next morning, in a log. These values were averaged across each week of the intervention period to obtain mean steps taken. This measure was included as an objec tive measure of physical activity. Convergent validity of pedometers with has been shown self-report measures of physical activity has been shown (Tudor-Locke, Williams, Reis, & Pluto, 2002). Modified Balke Submax (VO2) Cardiorespiratory fitness was measured using a modified Balke treadmill protocol (which is widely used) to obtain VO2 max estimates. The modified Balke is a treadmill protocol that involves slope increases while speed is held constant. Participan ts heart rates were monitored throughout the protocol and for two minutes prior to testing. Within the first minute, treadmill

PAGE 56

56 grade was increased 1% and continued to increase at 1-minute intervals until the heart rate reached 85% of age approximated maximal heart rate. Physicians permission was a must for this test. It is important to note that the m odified Balke treadmill protocol results in a less accurate estimate of cardiorespi ratory fitness than a maximal graded exercise stress test (VO2max, the gold standard); however, this modifi ed protocol is considered much safer for sedentary older adults as were studied in the present project (McAuley, 1992). Design and Rationale for Experimental Group The design of the exercise intervention was not the key focus of the present study; rather, it was the focus of the larger proj ect in which this study was embe dded. The exercise intervention was designed to produce individu ally tailored, goal-reference d exercise plans that were sustainable and situ ated in the everyday life cont exts of middle-aged and older individuals (n = 34). The intervention was a hybrid design, which integrated various theoretical models and components of behavior change, including motiv ational interviewing, mentorship and social support, self-efficacy, mental imagery, and goa l setting (Figure 3-2). Accordingly, the intervention was designed to maximi ze the probability that individuals would continue to engage in exercise once their study participation wa s finished. The intervention was a 13-session, 16week psycho-educational interv ention designed to enhance the adoption and maintenance of regular physical activity. Intervention sessions were c onducted by trained peer mentor s, with frequent quality control observations by study invest igators, and it was completely manualized, with peer mentors receiving extensive instruction in the use of th e manual. Peer mentors were selected based on successful completion of the prescribed interven tion previously and willingness and ability to undergo additional training related to exercise behavior change. Peer mentors that did exceptionally well at delivering the intervention protocol were aske d to lead future groups, while

PAGE 57

57 peer mentors who did not adhere well to study protocol were not asked to continue for future groups. In order to ensure quality control of the intervention delivery, all group sessions were either videoor audio-recorded, and peer mentors were evaluated on several criteria (supportiveness, effective communication, etc; Appendix C). Feedback was recorded and discussed with the peer mentor s prior the next group session. Each intervention group weekly session lasted between 45 and 60 minutes and began with an initial check-in rega rding physical activity completed be tween sessions. The schedule of intervention topics for the study is listed in Table 3-6. Th e group discussed challenges and barriers to engaging in physical act ivity, as well as successful particip ation in exercise behaviors. The remainder of the session was devoted to in troduction and discussion of the topic for the week. At the end of each session, homework was a ssigned to the group. Half of the participants arrived 30 minutes prior to each weekly intervention group session to complete 15-30 minutes of computerized assessments of cognition, exercise self-efficacy, and barriers to exercise selfefficacy. The second half of the participants remained 30 minutes after each group session to complete the computerized assessments. Design and Rationale for Control Group The 35 participants randomized to the heal th hygiene control group received a set of weekly topics based on m aterials from the Nationa l Institutes of Healths SeniorHealth website (NIHSeniorHealth.gov). This website featured up-to-date, easily understandable and accessible health information for seniors and their family a nd friends. Control participants met with a peermentored group each week to receive didactic instruction based on health and aging topics discussed on the NIH SeniorHealth website (Alz heimer's disease, arthritis, balance problems, COPD, diabetes, etc.). Similar to quality control practices employed for the intervention condition, control group sessions were also recorded so that specifi c feedback could be given to

PAGE 58

58 peer mentors delivering control gr oup didactic material. The goal was to focus on general health and overall lifestyle promotion, w ith little emphasis on exercise or physical activity. There was no exposure to any of the motivational, goal se tting, or imagery techniques used with the experimental, exercise promotion intervention gr oup. As mentioned previously, this group also received a free membership to a University of Florida campus fitness facility or a communitybased, church fitness center. This choice was made to control for the possible confound of access to fitness facilities and to enable the isol ation of the unique peer mentoring, motivational, and coaching aspects of the exercise promo tion/experimental group. Table 3-7 shows an overview of the design of this health hygiene control group. Table 3-8 shows the schedule of control group topics. While no formal homework assignments were given, peer mentors often challenged control participants to independently research additional information regarding the health topics discussed. As with the intervention group sessions, control group sessio ns lasted between 45 and 60 minutes each week. The first 15 minutes of the sessions consisted of a review the previous weeks discussion topic and the remainder of the time (30-45 minutes) was spent on the current weeks health topic. Half of the participants arrived 30 minutes prior to the groups session to complete weekly computerized assessment (15-30 minutes for cognitive and self-efficacy measures). The second half of the participants remained 30 minutes after each group session to complete computer assessments.

PAGE 59

59 Table 3-1. Pre-data collection pow er analysis results, indicating power expected at different effect sizes with two cells of 32 participants, and alpha = .05 Effect size Weak (d = .20) Medium (d = .50) Strong (d = .80) Power (1) .31 .76 .97

PAGE 60

60 Table 3-2. Participants mean characteristics Overall Health Hygiene Control Group Exercise Promotion Intervention Group n = 69 n = 35 n = 34 Df t / 2 Pvalue Age 63.9 (8.7) 63.7 (9.2) 64.2 (8.3) 67 -0.21 0.838 84.1% F 80% F 88.2% F Gender 15.9% M 20% M 11.8% M 1 0.87 0.35 91.3% White/Caucasian 94.3% White/Caucasian 88.2% White/Caucasian Race 8.7% Other 5.7% Other 11.8% Other 3 3.46 0.326 Years of Education 16.2 (2.2) 16.5 (2.1) 15.9 (2.3) 67 1.24 0.22 GDS 4.1 (3.4) 4.4 (3.9) 3.71 (2.7) 67 0.87 0.387 BDI-2 5.4 (4.3) 5.9 (4.2) 4.8 (4.5) 67 1.06 0.293 STAIState 29.7 (7.5) 30.3 (7.4) 29.1 (7.6) 67 0.68 0.502 STAITrait 30.5 (8.1) 31.9 (8.7) 29.2 (7.3) 67 1.39 0.171 Note : Mean (Standard Deviation) GDS= Geriatric Depression Scale; BDI-2 = Beck Depression InventorySecond Edition; STAI = State Trait Anxiety Inventory

PAGE 61

61 Table 3-3. Proposed study timeline for each replicate of participants Weeks 1-2 Weeks 3-18 Weeks 19-20 Pretesting Intervention Post-testing

PAGE 62

62 Table 3-4. Number of participants by age category, replicate, a nd experimental group Replicate Number Age Category 1 2 3 4 5 6 7 8 Total Control 4 6 1 4 2 17 Intervention 2 5 2 5 0 14 Ages 5064 Total 6 11 3 9 2 31 Control 3 4 3 2 1 5 18 Intervention 2 4 4 4 1 5 20 Ages 65 and older Total 5 8 7 6 2 10 38 Control 3 4 3 4 6 3 5 7 35 Intervention 2 4 4 2 5 6 6 5 34 Total 5 8 7 6 11 9 11 11 69 Replicate Grand Totals Dates of active participation 6/061/07 3/076/07 3/076/07 4/077/07 4/078/07 6/0710/07 12/074/08 1/085/08

PAGE 63

63 Table 3-5. Baseline and postintervention study protocol Domain Measure Primary Outcomes: Executive Control Cognitive Measures N-Back Task (Cohen et al., 1994) Controlled Oral Word Association (COWA; Benton & Hamsher, 1989) Trail Making Test B (Trails B; Reitan, 1992) Letter-Number Sequencing of WMS-III (Wechsler, 1997) Non-Executive Control Cognitive Measures North American Adult Reading Test (NAART; Blair & Spreen, 1989) Trail Making Test A (Trails A) (Reitan, 1992) Logical Memory of WMS-III (Wechsler, 1997) Proximal Outcomes: Psychosocial/Socio-emotional Measures: Emotional Well-being and Exercise Self Efficacy Geriatric Depression Scale (GDS; Yesavage & Brink, 1983) Beck Depression InventorySecond Edition (BDI-II; Beck, et al., 1996) State-Trait Anxiety Inventory (STAI; Spielberger, 1983) Exercise self-efficacy (EXSE; McAuley, 1993) Exercise barriers self-efficacy (BSE; McAuley, 1992) Exercise Behavior Measures Leisure Time Exercise Questionnaire (LTEQ ; Godin & Shephard, 1985) AE 120 pedometer (Yamax SW200 engine) Aerobic Fitness Measure Modified Balke Submax VO2 (VO2)

PAGE 64

64 Table 3-6. Schedule of intervention topics Intervention week Topic Homework Facilitator 1 Introduction to program and orientation to fitness facility. Computation of target heart rate. Read Health Benefits of Physical Activity Fitness center staff 2 Getting to Know You; Goal Setting Part I What is exercise? Activity Peer mentor 3 Goal Setting Part II Write out your goal Peer mentor 4 Defining exercise behavior: What is exercise? What is physical activity? How much? How hard? Types of exercise / exercise options Read Mental Imagery Article Media Activity Peer mentor 5 Barriers to Exercise 3 ways to overcome barriers to exercise Peer mentor 6 All About Mental Imagery Define imagery Practice general example of imagery Practice exercise specific imagery Fill out the Exercise Imagery Inventory Peer mentor 7 Discuss Exercise Imagery Questionnaire and progress toward stated goals None Peer mentor 8 Getting Good at Exercise accessing available supports to help you improve in your exercises of choice Identifying Improvement Resources Peer mentor 9 Revisiting barriers and goals: discuss pesky barriers and do status update on progress toward stated goals Apply one suggestion for overcoming barrier(s) to exercise Peer mentor 10 Group discussion of exercise behavior and accomplishments None Peer mentor 11 Sustainability: Continui ng progress and maintaining success, accessing family/social support None Peer mentor 12 Discuss maintenance goals, long-term versus shortterm goals. Review b asics of goal-setting (Exercise facility membership ends) Write out your goals for 3, 6 and 12 months Peer mentor 13 No meeting None 14 Share future goals; PARTY None Peer mentor 15 No meeting None 16 No meeting None Note : (Giacobbi et al. 2008)

PAGE 65

65 Table 3-7. Health hygiene control group design Design: Controls For: o Didactic Information about health and disease topics (Alzheimer's disease, arthritis, balance problems, COPD, diabetes, etc.) o Free Fitness Facility Membership o Assigns treatment effects to the unique didactic, motivation, and coaching components of the intervention o Controls for Social contact with study staff and group peers o Controls for out of house activity o Controls for retest effects o Controls for access to fitness facilities

PAGE 66

66 Table 3-8. Schedule of control group topics Control week Topic Facilitator 1 Introduction to program and orientation to fitness facility; Getting to know you Fitness center staff 2 Exercise Why is exercise important? How much exercise is too much? What types of exercise are recommended? Peer mentor 3 Osteoporosis What is osteoporosis? What are the warning signs? How do I prevent it? Peer mentor 4 Alzheimers Disease What do I look for? What affects the development of Alzheimers? Peer mentor 5 Cancer Screening What types of cancer should I be concerned about? How often should I be checked? Peer mentor 6 Hearing Loss How common is hearing loss in older adults? How do I know if I have hearing loss? Peer mentor 7 Arthritis What are the different types of arthritis? What can I do to treat arthritis? Peer mentor 8 Vision Loss Is vision loss just part of getting older? Is all vision loss the same? Peer mentor 9 Sleep Do older adults need as much sleep as young adults? How can I get a good nights sleep? Peer mentor 10 Balance Problems What causes me to lose my balance? Is there anything I can do to prevent it? Peer mentor 11 COPD* What is COPD? What type of exercise can I do if I have COPD? Peer mentor 12 Heart Failure* How can I prevent heart failure? Who is most at risk for heart failure? (Exercise facility membership ends) Peer mentor 13 No meeting None 14 Share future goals; PARTY Peer mentor 15 No meeting None 16 No meeting None Note : (Giacobbi et al. 2008). For Replicate 8, weeks 11 and 12 included discussion of diabetes and high blood pressu re health topics.

PAGE 67

67 Figure 3-1. Flow chart of partic ipants at each study phase. Note : *Three participants met more than one exclusion criteria. **O ne participant was randomized and did not complete baseline assessment. 433 Participants Identified and Contacted 306 Interested and Screened 160 Ineligible*: 10 Sensory 17 Cognitive/TICS 52 Physical/Medical 36 Medication 44 Exercise 4 Not Willing 127 Not Interested and Not Screened 146 Eligible 91 Randomized after Baseline Assessment** 55 Dropped-Out Prior to Randomization 47 Assigned to Control Group 44 Assigned to Intervention Group 12 Dropped-Out Prior to, During, or After Group 10 Dropped-Out Prior to, During, or After Group 35 Included in Current Analyses 34 Included in Current Analyses

PAGE 68

68 Initial Readiness, Motivation, and Self-Efficacy Towards Exercise Goal Setting Social Support Mental Imagery Autonomy Perceptions of Comp. Previous Performance Accomplish Relatedness Verbal Persuasion Vicarious Experience Vicarious Experience Physiological Arousal Increased Exercise Self-Efficacy Increased Intrinsic Motivation Increased Exercise Frequency And Intensity Improved Physical Fitness 1 2 3 4 5 6 7 89 Figure 3-2. Guiding theoretical fr amework for exercise promoti on group intervention (Buman, 2008)

PAGE 69

69 CHAPTER 4 RESULTS This study examining the effects of an exerci se intervention on executi ve cognitive function involved a between (group assignm ent) and within (measure/instrument, occasion of measurement) subjects design. The primary outcome was cognitive performance, while proximal or mediating outcomes measured in cluded physical fitness and self-perception (depression, well-being, and self-efficacy). The fo llowing results are presented with respect to two specific aims: To investigate whether, relative to matched non-exercising control pa rticipants, sedentary adults receiving a physical ex ercise promotion intervention experienced improvements in the primary outcome of cognitive function (par ticularly executive cont rol processes). To determine the separate and joint roles of improvement in proximal outcomes (fitness, activity, and affect) in mediating exerci se intervention eff ects on cognition. Preliminary Analyses Attrition of the sample A total of 22 participants were considered dropouts in th is study, resulting in the final sample of 69 participants. Of the 91 individuals that were rando m ized to either the control or intervention conditions following baseline assessm ent, 10 dropped out of the study prior to the 16-week study group period (Figure 3-1). One pa rticipant was randomized to a group with the expectation that pre-testing would be complete d just prior to the first session; however, the participant did not undergo baseline assessment or remain in the study for the small group sessions. Post-randomization, an additional twelve participants completed part or all of the small group sessions and did not complete post-testi ng. As shown in Table 4-1, in comparing demographic characteristics of individuals w ho completed post-testi ng (n = 69) to those individuals who only completed pre-testing (n = 21), independent samples t-tests indicated that

PAGE 70

70 the studys dropouts reported significantly hi gher depression symptoms In addition, dropouts had significantly poorer performance on the Two-Back condition of the N-back task. Missing Data To preserv e the present sample size with co mplete data, cell-specific mean replacement was implemented for three participants (4.3% of cases; Baltes & Maye r, 2001) on baseline and post-test cognitive variables (N-back and Logical Memory (LM)). Each missing data point for the cognitive variables was replaced with the mean value of scores obtained by individuals of the same gender, age group (i.e., 50-64 years old or 65 and older), and study group (control or intervention condition). Each missing data point for baseline physical fitness/activity and self-efficacy variables was replaced with the mean value of those w ith the same gender, age group, and BMI category (i.e., BMI <18.5; 18.5-25; 25-30; and >30). This strategy was chosen as a better approximation of the physical variables, and because there were no baseline differences between the study groups on physical fitness, physical activity, or self-efficacy data (Buman, 2008). Six cases (8.7% of the sample) had missing data for the BMI variable. These data were replaced with mean BMI for individuals of the same gender an d age group. Additionally, there was a total of 6 cases (8.7% of the sample) w ith missing baseline physical fitn ess data (pre-VO2 = 1 missing case), baseline physical activity (p re-LTEQ (leisure time exercise) = 1 missing case; pre-MVPA (minutes of moderate and vigorous physical act ivity) = 2 cases; and pre-mean pedometer steps = 4 missing cases), and/or baseline self-efficacy data (pre-EXSE (exercise self-efficacy) = 1 missing case). After BMI mean replacement, one case did not share the same gender, age group, and BMI category with anyone, thus that persons missing data (for pre-mean pedometer steps) were replaced with mean values of indivi duals of the same gender and age group.

PAGE 71

71 Regarding post-test data, there was one case missing physical fitness (post-VO2) data at follow-up. However, for post-test physical act ivity and self-efficacy variables, there were substantially more missing data (post-LTEQ = 14 missing cases; post-mean pedometer steps = 17 missing cases; post-MVPA = 23 missing cases; post-BSE = 21 missing cases; and post-EXSE = 21 missing cases). Since physical activity and self-efficacy data were collected on a weekly basis, missing post-test values were replaced by the most recent available value for preceding weeks 13-16 (i.e., if week 16 was missing, this data point was repl aced by the value for week 15, etc.). In the event that there were no available data, these cases were replaced with age, gender, and BMI group posttest means. There were 19 su ch cases without any data for weeks 13-16; thus by-cell mean replacement was completed for these individuals (post-LTEQ = 7 missing cases; post-mean pedometer steps = 10; post-MV PA = 2; post-BSE = 1 missing case; post-EXSE = 1 missing case). Distributions of Dependent Variables and Outliers Prior to analysis of the cognitive data, values that were standard deviations from the mean of each cognitive and psychosocial variable were set to missing and replaced with the mean value of individuals with the same gender, age group, and study group (as described above). There were 28 specific outlier values across all baseline and post-test cognitive variables. There were 12 total specific outli er values for baseline va riables of LM Leaning Slope, Letter-Number Sequencing, Geriatric Depression Scale (GDS), Beck Depression Inventory-2 (BDI-II), One-Back Mean Reaction Time (RT), One-Back Mean Reaction Time (RT) Standard Deviation (SD), Two-Back Mean RT, Two-Back Mean RT SD, and Two-Back Number Correct. There were 16 total outliers on post-test m easures of LM Leaning Slope, COWA, Trails B Time, GDS, BDI-II, STAI Trait Anxiety, One-Back Mean RT, One-Back Mean RT SD, One-Back Number Correct, Two-Back Mean RT, and Two-Back Mean RT SD.

PAGE 72

72 Prior to outlier replacement, the distributions of all cognitive dependent variables were explored for skewness and kurtosis. Many execu tive cognitive were significantly skewed at baseline. Variables that were significantly positively skewed (absolute skewness values > 2) were as follows: Trails B Time (Skewness = 2.23), One-Back Mean RT SD (Skewness = 2.57), and Two-Back Mean RT (Skewness = 2.13). Va riables that were significantly negatively skewed were One-Back Number Correct (Skewness = -5.48) a nd Two-Back Number Correct (Skewness = -3.88). Regarding kurtosis, variab les that were significan tly kurtotic (absolute valuates > 2) were as follows: Trails A Time (Kurtosis = 2.65), Trails B Time (Kurtosis = 6.01), One-Back Mean RT (Kurtosis = 3.35), One-Ba ck Mean RT SD (Kurtosis = 8.25), One-Back Number Correct (Kurtosis = 33.40) Two-Back Mean RT (Kurtosi s = 6.29), Two-Back Number Correct (Kurtosis = 20.67), GDS (Kurtosi s = 4.77), and BDIII (Kurtosis = 5.36). After outlier mean repl acement, the following baseline variables were significantly skewed positive: Trails B Time (Skewness = 2.23) and One-Back Mean RT SD (Skewness = 2.09). The only variable negatively skewed was One-Back Number Correct (Skewnes s = -5.48). Significant kurtosis values were observed for Trails A Time (Kurtosis = 2.65), Trai ls B Time (Kurtosis = 6.01), One-Back Mean RT SD (Kurtosis = 5.69), One-Back Number Correct (Kurtosis = 33.40), and Two-Back Number Correct (Kurtosis = 4.39). All skewne ss and kurtosis values (after outlier replacement) are presented in Table 4-2. Baseline comparisons and correlations Baseline independent s amples t-tests for all cognitive, psychosocial (affect and selfefficacy), and physical fitness/activity data in dicated no significant study group differences on study variables, except for EXSE ( t (67) = -3.289; p = .002), with the exercise promotion intervention group reporting a higher mean level of exercise self-e fficacy at baseline. However, with Bonferroni correction (.05/27 = 0.00185), this finding was no longer significant. Next,

PAGE 73

73 correlations among all cognitive, psychosocial, and physical fitness/activity data were examined to determine their associations at baseline (Tables 4-3, 4-4, a nd 4-5). Overall, of the 153 correlations assessed, only 16 (10.5%) reached the p < .05 level of significance. With regard to physical variables, bivariat e correlations indicated modest significant positive relationships between the LTEQ and LM Recognition and Trails B Time, and a significant negative association between the LTEQ and LM Delayed Recall (p < .05). Thus, greater levels of leisure time exercise was a ssociated with better LM Recognition scores, but worse Trails B Time and LM Delayed Recall pe rformance, which was in the opposite direction of what would be expected. Furthermore, MVPA was significantly negatively associated with LM Immediate (p < .05) and Delayed Recall and LM Recognition (p < .01), also in the opposite direction. Next, regarding psychosocial variables, th e GDS, STAI-State Anxi ety, and STAI-Trait Anxiety had significant positive correlations with the LTEQ ( p < .05), suggesting that also counterintuitive to what would be expected, hi gher levels of leisure time exercise were associated with higher levels of GDS depressi on and state and trait anxi ety symptoms. The GDS and STAI-Trait Anxiety measures were also significantly positively correlated with MVPA (p < .01), with higher minutes of modera te-vigorous physical activity asso ciated with higher levels of GDS depression and trait anxiety. Furthermore, GDS depression symptoms were negatively correlated with LM Recognition scores ( p < .05), and GDS depression symptoms were positively correlated with Trails B Time (p < .01). These correlations were in the directions that would be anticipated. Similarly as would be hypothesi zed, BDI-II depression symptoms were positively correlated with both Trails A ( p < .05) and Trails B Time ( p < .01). Regarding associations with anxiety scores, both STAI State and STAI Trait Anxiety variable s showed negative associations

PAGE 74

74 with LM Recognition at baseline ( p < .05), indicated that, as woul d be expected, lower state and trait anxiety was associated with higher performance on the LM Recognition task. Lastly, with regard to exercise-related se lf-beliefs, both BSE and EXSE scores were negatively associated with performance on LM Immediate Recall and LM Delayed Recall (p < .05). This suggested that, contra ry to expectation, higher percei ved self-efficacy was related to lower performance on both LM recall trials at baseline. Next, baseline intercorrelations on all cognitive variables were computed to determine the inter-relatedness of all executive and non-executive cognitive variables (Table 4-6). Executive cognitive variables were expected to be more correlated with e ach other and less correlated with non-executive cognitive variables. Nevertheless, all executive cognitive measures except for One-Back Mean RT, Two-Back Mean RT, One-B ack Number Correct, and Two-Back Mean RT SD were significantly correlated with two or more non-executive cognitive measures. Thus, these findings indicated that the executive cogni tive variables were often more correlated with non-executive cognitive variables than with each ot her in this sample, suggesting the theoretical groupings of the cognitive measures was not supporte d empirically. It furt her suggested that the hypothesized separation of execu tive and non-executive measures with regard to treatment effects might not be supportable, given the lack of distinctiveness between the two domains. Physical Fitness and Self-Efficac y Measures: Pre-Post Changes Prior to examining the effect of the interv ention on cognitive outcomes, the influence of the intervention on intended primary outcomes of the larger study (physical fitness/activity and exercise self-efficacy) was explored (Table 4-7 and Figure 4-1). While the present document considers the cognitive measures as primary outcomes for this study the larger project in which it was embedded considered physical fitness and self-efficacy as primary outcomes and cognitive function as secondary outcomes. As suc h, the hypothesized improvements on cognitive

PAGE 75

75 performance were expected to be dependent on the anticipated improveme nts in physical fitness and exercise self-efficacy. First, the critical group-by-occas ion effect was examined in univariate analyses of variance (ANOVA). This effect would address whether there were experimental group differences in exercise change over time, assessing the hypothe sis of disproportionate improvement in the experimental group. The critical group-by-occasion interaction did not reach significance for any of the larger studys primary outcomes, sugges ting that the interventi on was not effective in producing disproportionate improveme nts in intended outcome variab les. This suggested that a core assumption in the studys conceptual model (that the intervention would affect fitness and activity) was not supported. Despite this, the planned primary analyses were completed, to assess whether there was, nonetheless, a direct association between ex ercise group assignment and secondary cognitive outcomes. It should be noted that there was a single betwee n subjects effect for one measure: exercise self-efficacy (EXSE) (F (1, 67) = 7.30; p = .01; Partial 2 = 0.10). The exercise promotion intervention group had significantly higher exercise self-efficacy th an the control group, but this was true across both pre-test a nd post-test (i.e., it did not in dicate the presence of group differences in self-efficacy ch ange, which was hypothesized). Independent of group assignment, significant w ithin-subjects effects (i.e., occasion effects, or significant changes from pretest to posttest) were found for four measures. For the LTEQ ( F (1, 67) = 17.78; p < .001; Partial 2 = 0.21), MVPA( F (1, 67) = 17.98; p < .001; Partial 2 = 0.21), and VO2 ( F (1, 67) = 7.97; p = .006; Partial 2 = 0.11), participants across both groups improved across time on these measures. Somewh at unexpectedly, for exer cise self-efficacy, the significant occasion effect ( F (1, 67) = 5.82; p = .02; Partial 2 = 0.08) indicated that, across

PAGE 76

76 groups, participants experi enced self-efficacy reductions over time. Cohens d estimates of the time effects were as follows: LTEQ, d = 0.51; MVPA, d = 0.51; mean pedometer steps, d = 0.22; VO2, d = 0.34; BSE, d = -0.12; EXSE, d = -0.29, with the latter tw o coefficients reflecting poorer mean self-efficacy at post-test. It should be noted that all of these effects were small to medium (Cohen, 1992). Examining the means in Fi gure 4-1, with a baseline V02 of 28, and an average 2 point increase, the mean level fitness increase in this study represented only 7%. The self-reported gain in metabolic equivalents, based on self-reported activity, was an increase of 4 points over a baseline value of 6, which represented a more sizeable 67% increase. Anxiety and Depression: Pre-Post Changes Since changes in psychosocial variables of anxiety and depression were hypothesized to be one route by which changes in executive cognitive function would occur, these variables were included in a repeated-measures MANOVA to determine intervention effects on these variables (Table 4-8 and Figure 4-2). Again, the critical occasion-by-group interaction did not reach significance (Wilks = 0.94; F(4, 64) = 1.00, p < 0.41; Partial 2 = 0.06). Turning to the main effects, while there was no overall signi ficant between subjects effect (Wilks = 0.97; F(4, 64) = 0.41, p = .80; Partial 2 = 0.03), there was a significant over all within subjec ts effect of occasion (Wilks = 0.71; F(4, 64) = 6.68, p < .001; Partial 2 = 0.29). Post-hoc univariate ANOVAs (Table 4-9) on each anxiety and depr ession measure revealed significant reductions over time for the STAI Trait Anxiety ( F (1, 67) = 11.58; p < .001; Partial 2 = 0.15) and BDI-II measures ( F (1, 67) = 22.68; p < .001; Partial 2 = 0.25), suggesting that there were improvements in anxiety and depression symptoms for the combined groups. Cohens d estimates of effect sizes for anxiety and depressi on within-subjects effects were small to medium

PAGE 77

77 and are as follows: GDS, d = 0.01; BDI-II, d = 0.58; STAI State Anxiety, d = 0.13; and STAITrait Anxiety, d = 0.41. Aim 1: Exercise Promotion Intervention-Relat ed Improvements in Cognition Study Group Comparisons To address the first aim, repeated-measures multivariate analyses of variance (MANOVA) was employed to determine whether there were cognitive changes with time, across instruments, both within and between groups. Of critical interest was the occasion-by-group interaction, which addressed whether subjects receiving the exercise promo tion intervention improved more than health hygiene controls. As shown in previous stud ies, it was hypothesized that participants receiving the exercise intervention would show improved performance on cognitive measures, particularly those assessing executi ve control processes, relative to control participants. Two repeated-measures MANOVAs were conduct ed to test this hypothesis, one for nonexecutive cognitive measures and a second for executive cognitive tests (Table 4-10). First, for the non-executive cognitive measures, there was no significant occasion-by-group interaction as hypothesized, suggesting that th e exercise promotion interv ention group did not improve significantly more than the health hy giene control group across time (Wilks = 0.92; F(6, 62) = 0.51, p = .80; Partial 2 = 0.08). Next, there was no signifi cant between subjects (intervention group) effect (Wilks = 0.90; F(6, 62) = 1.12; p = 0.36; Partial 2 = 0.10). This suggests there were no overall mean group differences in cognitive performance on non -executive measures. There was a significant within-subje cts effect of occasion (Wilks = 0.55; F(6, 62) = 8.48, p < .001; Partial 2 = 0.45), suggesting that overa ll, there was an effect of time on performance. Specifically, the entire sample, regardless of ra ndom group assignment, cha nged significantly in

PAGE 78

78 their overall performance on non-executive measures from baseline to post-test. The direction of this finding will be explored in follow-up univariate analyses presented below. Next, for the executive cognitive measures a second repeated-measures MANOVA again showed no significant occasion-by -group interaction, suggesting th at the exercise promotion intervention group did not improve significantly more or less than the health hygiene control group over time (Wilks = 0.86; F(9, 59) = 0.96, p = .48; Partial 2 = 0.14). There was no significant between subjects effect (Wilks = 0.97; F(9, 59) = 0.31, p = .97; Partial 2 = 0.03), but there was a significant within-s ubjects effect of occasion. (Wilks = 0.65; F(9, 59) = 4.01, p < .001; Partial 2 = 0.35). Follow-up univariate analys es were conducted to ascertain the direction of the effect and are explored below. The presence of significant overall occasion e ffects required follow-up univariate analyses to ascertain their directions in each measure (Tables 4-11 and 4-12). Beyond occasion analyses, these follow-up analyses also examined the betw een-group effect and interactions; while these effect examinations were not protected by the om nibus test, they were still useful as follow-up, exploratory analyses. The analyses that follo w are reported with thei r original, uncorrected probabilities, since there is li ttle consensus on alpha adjustment for these follow-up tests (Tabachnick & Fidell, 2007). However, with si x non-executive measures, a simple Bonferroni correction would requ ire probabilities of p = 0.05/6 = 0.0083 to be labeled as significant. Similarly, with nine executive measures, a simple Bonferroni correction would require probabilities of p = 0.05/9 = 0.0056 to be labeled as signifi cant. The implications of these adjustments will be considered in the discussion chapter that follows. While there were no significan t between-subjects effects, th ere was an occasion-by-group interaction for one cognitive measure (COWA) suggesting interventi on group participants

PAGE 79

79 improved significantly more on this meas ure than did controls over time ( F (1, 67) = 4.15; p < .046; Partial 2 = 0.06); Table 4-12 and Figure 4-4). To examine this further, a pre-post COWA mean change score was computed for each participant and used as the dependent variable in a follow-up, one-way analysis of variance to examine whether there was a significant group difference in the magnitude of change on this measure. The results indicated a significant difference in the magnitude of mean change in COWA scores over time, with the intervention group demonstrating greater improvement in COWA scores over time than controls (mean (SD) change for intervention group: 4.44 (6.46) ; for control group: 0.58 (9.02); F(1, 67) = 4.15; p = .046, Partial 2 =0.06). Next, there were within-subjects effects of occasion for severa l cognitive variables. Tables 4-10 and 4-11 show F statistics, degrees of freedom, a nd significance values for the withinsubjects effects of each repeated-measures ANOVA. These cognitive variables include NAART, Trails A, and LM Immediate and Delayed r ecall (non-executive measures) and COWA, OneBack Mean RT, Two-Back Mean RT, and TwoBack Mean RT SD (executive measures). Estimated marginal means for each group are pres ented in Table 4-13, and Figures 4-3 and 4-4 display line graphs of estimated marginal m eans by group. Overall, there were significant improvements on these cognitive variables for the entire sample, regardless of study group, over time. These practice/occasion effects were of small to medium size for most of the cognitive variables (Cohen, 1992). Cohens d estimates of effect were as follows: NAART, d = 0.40; Trails A, d = 0.32; LM Immediate Recall, d = 0.61; LM Delayed Recall, d = 0.67; LM Recognition, d = 0.27; LM Learning Slope, d = 0.13; COWA, d = 0.31; Trails B, d =0.09; Letter Number Sequencing, d = 0.05; One-Back Mean RT, d = 0.41; One-Back Mean RT SD, d = 0.49;

PAGE 80

80 One-Back Number Correct, d = 0.20; Two-Back Mean RT, d =0.17; Two-Back Mean RT SD, d = 0.35; and Two-Back Number Correct, d = 0.34. Reliable Change As follow up analyses, reliabl e change index scores were calculated to examine intraindividual trajectories of change. The goa l was to determine the proportion of the sample that experienced sizable gains above the usual effect of practice that would be observed in an untreated control group. This analysis allowed the subset of particip ants who experienced especially large improvement to be identified to determine whether the proportion of such participants differed by study group. Traditional reliable change index sc ores (RCI) use published normative data to determine a confidence interval (90% or 95% ar e commonly used) that indicates the expected, normal variation of change scores. In this ca lculation, the average post-test score for each measure was subtracted from the average baseline scores and divided by the standard error of the difference between scores ([(SD2x + SD2y)(1-rxy)]1/2). The baseline change, around which the RCI confidence interval is com puted, is traditionally zero. A modified reliable change sc ore, employed in this study, assumes that the mean level of practice-related gain in the cont rol group serves as the baseline around which the reliable change confidence interval should be established. This is done because even untreated participants typically experience some gain or change. When RCI scores are adjusted for practice (RCI-P), performance of a control condition is taken into account by adju sting the RCI for the average change in scores experienced by the control grou p. The RCI-P scores were calculated for the present sample by adding and subtracting the aver age change in scores to each RCI-P. Because of the conservative nature of the RCI-P scores, an al pha criterion of p = .10 was used. Table 414 displays, for each measure, the pre-test and po st-test standard deviatio ns, correlations between

PAGE 81

81 pre-test and post-test sc ores, the standard error of the differe nce, and the standa rd error of the difference times 1.64 (the crit erion of significance at p = .10). Also, the control groups mean change scores and the thresholds for reliable d ecline and reliable improvement, adjusted for the control groups mean change, are presented. Upon examination of the present samples scores to determine how many individuals reliably changed from preto post-test, there was little evidence for reliable improvement across cognitive measures (Tables 415 and 4-16). Logical Memory Immediate Recall was most sensitive to reliable improvement (24.6% of the sample); however, relative to controls, fewer intervention participants experienced reliable improvement (28.6% of controls improved versus 20.6% of intervention participants ), although this difference was not significant (nor was any other tested) using chi-square analysis. Overall, most participants showed no reliable change between baseline and post-test. In addition, a small number of par ticipants demonstrated reliable decline across measures. Exploratory Analyses: Age Group Comparisons As additional exploratory, follow-up an alyses, repeated-m easures MANOVAs were conducted, using age group as a second between subjects factor, to determine whether intervention effects might be moderated by age (i.e., Did younger participants experience more improvement than older particip ants? Did the occasion-by-group interaction exist for one age group, but not for the other?). This time, tw o three-way repeated-measures MANOVAs were conducted (Table 4-15). The repeated-m easures MANOVA for non-executive cognitive measures demonstrated, as with the previous an alyses, no significant in teractions for occasionby-study group (Wilks = 0.92; F(4, 60) = 0.83, p = .55; Partial 2 = 0.08), occasion-by-age group, (Wilks = 0.97; F(4, 60) = 0.34, p = 0.91; Partial 2 = 0.03), or occasion-by-study group-

PAGE 82

82 by age (Wilks = 0.94; F(4, 60) = 0.63, p = .70; Partial 2 = 0.06). There was also no study group-by-age interaction (Wilks = 0.90; F(6, 60) = 1.17, p = .33; Partial 2 = 0.10). In addition, overall study group (Wilks = 0.89; F(6, 60) = 1.20, p = .32; Partial 2 = 0.11) and overall age group effects (Wilks = 0.83; F(6, 60) = 2.09, p = 0.07; Partial 2 = 0.17) did not reach significance. Although, it is worth nothing that the age group effect approached our criterion of significance ( p = .05). As with the initial multiv ariate analyses for study aim one, there was a significant within -subjects effect (Wilks = 0.55; F(6, 60) = 8.11, p <.001; Partial 2 = 0.45). Next, for the executive cognitive measures there were no significant multivariate occasionby-study group (Wilks = 0.86; F(9, 57) = 1.01, p = 0.44; Partial 2 = 0.14), occasion-by-age group (Wilks = 0.89; F(9, 57) = 0.80, p = 0.62; Partial 2 = 0.11), or occasion-by-study groupby-age group (Wilks = 0.83; F(9, 57) = 1.34, p = 0.24; Partial 2 = 0.17) interactions. A study group-by-age group interac tion effect approached significance (Wilks = 0.77; F(9, 57) = 1.91, p = 0.07; Partial 2 = 0.23). While there was no effect of study group (Wilks = 0.96; F(9, 57) = 0.27, p = .98; Partial 2 = 0.04), there was a significant age group effect (Wilks = 0.73; F(9, 57) = 2.33, p = 0.03; Partial 2 = 0.27). Additionally, there was a significant within-subjects effect for occasion (Wilks = 0.64; F(9, 57) = 3.58, p < .001; Partial 2 = 0.36), suggesting that overall, there was an effect of time on performance. Follow-up univariate ANOVAs are presente d in Tables 4-16 and 4-17, and estimated marginal means are displayed in Table 4-18. The only non-executive cognitive variable showing an age group effect was Trails A Time, indicating that the younger age group (50-64 year olds) performed this task significan tly faster than the 65 and ol der group (score was number of seconds). The non-executive variables showing significant within-subjects effects included the

PAGE 83

83 NAART, Trails A Time, LM Immediate Recall, LM Delayed Recall, and LM Recognition, indicating all participants significantly improve d across occasions on these measures, regardless of study group random assignment (Figure 4-5) There was a study group-by-age group interaction effect for LM Delayed Total Recall, with the older individuals in the exercise promotion intervention group performing signi ficantly better than younger individuals (regardless of study group) and older controls. There was a study group-by-age gr oup interaction for the OneBack Mean RT SD (Figure 4-6). Younger control group participants were si gnificantly less inconsistent in their mean reaction time on this measure than younger intervention participants and all older participants. Additionally, there was an age-byoccasion interaction effect for the One-Back Mean RT task. This interaction effect suggest ed that younger participants impr oved significantly more than older participants in mean reaction time over time. Finally, there was a study group-by-age group-by-occasion interaction fo r Trails B Time. Younger intervention group participants improved significantly more on this task than their younger contro l group counterparts and older participants of both study groups. Executive cognitive variables wi th significant age group effects were the One-Back Mean RT, Two-Back Mean RT, and Two-Back Mean RT SD. These findings indicate that the younger age group had significantly faster mean reaction time and were sign ificantly less inconsistent in their mean reaction time across occa sions, when compared to their older counterparts. Further, within-subjects effects for executive cognitive variables were the COWA, One-Back Mean RT, One-Back Mean RT SD, Two-Back Mean RT SD, and Two-Back Number Correct. All participants improved significantly on these measures across time (i.e., there was improvements

PAGE 84

84 in mean reaction time on the One-Back Mean RT and less inconsistency on the One-Back and Two-Back RT SD variables). Aim 2: Cognitive changes correlated with ac tivity, fitness and psychosocial changes In the aforementioned specific aim s, examinati on of whether intervention-affected changes in fitness, activity, and self-perception variables might mediate expected intervention-related changes in cognitive variables (particularly executive outcomes) was originally planned. However, the general absence of intervention effects on cognitive outcomes renders this mediator analysis unwarrant ed (per Baron & Kenney, 1986). Nonetheless, since the study results indicated at least practice-re lated improvement on most cognitive variables, and since participants in both exercise groups did have access to facilities and support to increase their activity, the a ssociations between observed changes in cognition and changes in activity, fi tness, or self-belief variables were nonetheless investigated. This was done to determine whether there was an y support that improvements in exercise-related outcomes would be associated with cognitive improvements, regardless of group membership (since there was evidence that both groups experienced VO2 and leisure time activity improvements). Consequently, to address the second aim, pre-post change scores across cognitive, selfperception, and fitness variables were computed and correlated (Tables 4-21, 4-22, and 4-23). Bivariate correlations between phys ical fitness and activ ity and cognitive performance indicated a significant positive correlation between change in LM Immediate Recall score and change in VO2 (r = 0.31, p < .05). This suggests that improvements in immediate recall of the LM stories were positively associated with improvements in cardiorespiratory fitness over time. There were significant negative correlations between the LTEQ and One-Back Mean RT SD (r = -0.25, p < .05) and mean pedometer steps and LM Learning Slope (r = -0.29, p < .05). These findings

PAGE 85

85 indicate inverse relationships between these sets of variables, such that improved leisure time exercise was associated with decreased incons istency in mean reaction time on the One-Back task, and improved mean pedometer steps was rela ted to reduced learning slope across testing occasions. It is important to note that these significant relationships represent the minority of those tested. For the 60 correlations tested be tween physical fitness/activity and cognition, only 3 (5%) reached significance. Next, in correlating changes in physical fitn ess/activity and psychos ocial variables, the only significant association was between the GDS and VO2 (r = -0.25, p < .05). This inverse relationship indicated that a de crease in GDS depression symptoms was associated with an increase in cardiorespiratory fitness. Of the 24 correlations between physi cal fitness/activity and self-perception, only 1 (4.2%) was significant. Finally, correlations between psychosocial changes and cognitive score changes over time revealed three significant negative correlations between changes in GDS depressi on and LM Delayed Recall (r = -0.24, p < .05), BDI-II depression and Two-Back Mean RT SD (r = -0.26, p < .05), and BSE and Two-Back Number Correct (r = -0.28, p < .05). These negative correlations indicated that reduced GDS depression symptoms was associated with improved delaye d recall of the LM stories over time, reduced BDI-II depression symptoms was related to incr eased reaction time inconsistency on the TwoBack task over time, and higher self-efficacy on the BSE was related to an improvement in the number correct on the Two-Back task over time. As with fitness variables, only the minority of associations tested reached the p < .05 criterion of significance. Of the 90 correlations tested between self-perception and cognition, only 3 (3.3%) reached signi ficance. It should also be noted that, given the total number of correlations examined, if the p-values for these correlations were Bonferroni-corrected, the cr itical value of alpha would be p = .05/176 = 0.00029.

PAGE 86

86 Table 4-1. Baseline mean comparis on of study completers and dropouts Overall n = 90 Completers n = 69 Dropouts n = 21 Df t /X2 p -value Age 63.6 (8.5) 63.9 (8.7) 62.3 (7.7) 88 -0.755 0.452 82.2% F 84.1% F 76.2% F Gender 17.8% M 15.9% M 23.8% M 1 0.682 0.409 90.0% White/Caucasian 91.3% White/Caucasian 85.7% White/Caucasian Race 10.0% Other 8.7% Ot her 14.3% Other 4 4.752 0.314 Years of Education 16.1 (2.2) 16.2 (2.2) 15.9 (2.4) 88 -0.504 0.615 VO2 28.1 (7.9) 27.9 (7.5) 29.0 (9.5) 88 0.544 0.588 GDS 4.8 (4.0) 4.1 (3.4) 7.1 (5.1) 88 3.228 <.001 BDI-II 6.3 (4.8) 5.4 (4.3) 9.3 (5.2) 88 3.469 <.001 STAI-State 30.3 (7.9) 29.7 (7.5) 32.2 (9.1) 88 1.282 0.200 STAI-Trait 31.5 (8.8) 30.5 (8.1) 34.9 (10.2) 88 2.010 0.470 NAART 113.2 (6.5) 113.1 (6.5) 113.5 (6.6) 88 0.293 0.770 LM Immediate Recall 41.0 (8.8) 41.0 (8.6) 41.0 (9.9) 88 -0.026 0.979 LM Delayed Recall 25.9 (7.5) 25.8 (7.5) 26.4 (7.7) 88 0.317 0.752 LM Learning Slope 4.01 (2.5) 4.1 (2.7) 4.2 (1.8) 88 0.194 0.847 LM Recognition 26.4 (2.4) 26.4 (2.3) 26.7 (2.5) 88 0.531 0.597 Trails A Time 31.6 (10.7) 32.5 (11.1) 28.4 (8.6) 88 -1.583 0.117 COWA 39.5 (10.4) 39.5 (10.8) 39.5 (9.2) 88 0.017 0.986 Trails B Time 78.1 (38.1) 81.0 (39.9) 68.5 (30.1) 88 -1.322 0.190 Letter-Num. Sequencing 10.7 (2.6) 10.5 (2.6) 10.9 (2.5) 88 0.582 0.562 One-Back Mean RT 866.4 (203.5) 847.4 (199.8) 928.6 (207.8) 88 1.616 0.110 Note : Mean (Standard Deviation). aThere were unequal variances for these measures. LM = Logical Memory; NAART = North American A dult Reading Test; COWA = Controlled Oral Word Association; RT = Reaction Time; SD = Standard Deviation

PAGE 87

87 Table 4-1 Continued. Overall n = 90 Completers n = 69 Dropouts n = 21 Df t /X2 p -value One-Back Mean RT SD 361.1 (313.1) 330.3 (224.9) 462.4 (500.4) 88 1.712 0.090 One-Back Number Correct 96.1 (12.4) 95.5 (14.1) 98.0 (3.3) 88 0.799 0.427 aTwo-Back Mean RT 1521.2 (796.7) 1756.0 (748.2) 749.9 (325.7) 77.2 -8.769 <.001 aTwo-Back Mean RT SD 1058.9 (713.3) 912.8 (538.1) 1538.9 (981.7) 23.8 2.797 0.010 aTwo-Back Number Correct 71.8 (34.8) 89.6 (14.3) 13.3 (3.3) 85 -40.843 <.001 Note : Mean (Standard Deviation). aThere were unequal variances for these measures. LM = Logical Memory; NAART = North American A dult Reading Test; COWA = Controlled Oral Word Association; RT = Reaction Time; SD = Standard Deviation

PAGE 88

88 Table 4-2. Distributions of depe ndent variables at baseline n= 69 Mean Std. DeviationSkewnessStd. Error KurtosisStd. Error NAART 113.066.46 -0.33 0.29 -0.36 0.57 LM Immediate Recall 41.018.57 -0.08 0.29 -0.19 0.57 LM Delayed Recall 25.787.52 -0.08 0.29 -0.10 0.57 LM Learning Slope 4.072.67 0.18 0.29 0.62 0.57 LM Recognition 26.352.34 -0.46 0.29 -0.37 0.57 COWA 39.4810.84 0.32 0.29 0.22 0.57 Trails A time 32.5511.10 *1.39 0.29 *2.65 0.57 Trails B time 81.0039.90 *2.23 0.29 *6.01 0.57 Letter-Number Sequencing 10.532.62 *0.87 0.29 0.98 0.57 One-Back Mean RT 837.28175.35 *0.76 0.29 *1.32 0.57 One-Back Mean RT SD 315.41184.63 *2.09 0.29 *5.69 0.57 One-Back Number Correct 95.5214.05 *-5.48 0.29 *33.40 0.57 Two-Back Mean RT 1667.01530.85 *0.83 0.29 0.18 0.57 Two-Back Mean RT SD 888.84492.85 *1.33 0.29 *1.66 0.57 Two-Back Number Correct 90.879.51 *-1.82 0.29 *4.39 0.57 GDS 4.063.36 *1.28 0.29 *1.51 0.57 BDI-II 5.354.33 *1.07 0.29 1.06 0.57 STAI State 29.717.51 0.51 0.29 -0.50 0.57 STAI Trait 30.548.09 *1.07 0.29 0.78 0.57 VO2 27.897.45 -0.28 0.29 -0.66 0.57 LTEQ 6.085.86 *2.56 0.29 *9.82 0.57 MVPA 76.23126.59 *4.22 0.29 *23.86 0.57 Pedometer Steps 6145.412860.41 *0.83 0.29 0.88 0.57 BSE 67.5723.48 *-0.65 0.29 -0.50 0.57 EXSE 74.1427.03 *-0.96 0.29 -0.03 0.57 Note : p < .05; LM = Logical Memory; NAART = No rth American Adult Reading Test; COWA = Controlled Oral Word Association; RT = Reaction Time; SD = Standard Deviation; GDS = Geriatric Depression Scale; BDI-2 = Beck Depression InventorySecond Edition; STAI = State Trait Anxiety Inventory; VO2 = Modified Balke Submax; BSE = Barriers Self Efficacy Scale; EXSE = Exercise Self-Efficacy Scale

PAGE 89

89 Table 4-3. Baseline correlations between physic al activity, physical fitness, and cognitive variables VO2 LTEQ Mean Pedometer Steps MVPA NAART -0.107 0.100 -0.111 0.128 Trails A Time -0.111 0.150 0.010 0.140 LM Immediate Recall 0.105 -0.199 -0.122 *-0.306 LM Delayed Recall 0.168 *-0.240 -0.171 **-0.317 LM Recognition 0.085 *0.295 -0.122 **-0.365 LM Learning Slope -0.004 -0.206 -0.226 -0.105 COWA -0.090 0.036 0.130 0.093 Trails B Time -0.171 *0.269 -0.015 0.177 Letter Number Sequencing 0.072 -0.073 0.101 -0.111 One-Back: Mean RT 0.008 0.018 -0.036 -0.001 One-Back: Mean RT SD 0.020 -0.039 -0.024 -0.001 Two-Back: Number Correct 0.010 -0.013 0.018 -0.039 Two-Back: Mean RT -0.132 -0.087 -0.063 -0.160 Two-Back: Mean RT SD -0.189 -0.066 -0.062 -0.122 Two-Back: Number Correct 0.136 -0.032 0.125 0.005 Note : p < .05; ** p < .01; LM = Logical Memo ry; NAART = North American Adult Reading Test; COWA = Controlled Or al Word Association; RT = Reaction Time; SD = Standard Deviation; VO2 = Modified Balke Submax; LTEQ = Leisure Time Exercise Questionnaire; MVPA = Minut es of Moderate and Vigorous Physical Activity

PAGE 90

90 Table 4-4. Baseline correlations between physical fitness, physical activity, and psychosocial variables VO2 LTEQ Mean Pedometer Steps MVPA GDS -0.128 *0.286 0.014**0.397 BDI-2 0.137 0.154 0.0860.120 STAI State Anxiety -0.136 *0.257 0.0920.215 STAI Trait Anxiety 0.017 *0.291 0.064**0.372 BSE -0.004 0.005 0.071-0.012 EXSE 0.059 0.144 0.1460.111 Note: p < .05; ** p < .01; GDS = Geriat ric Depression Scale; BDI-2 = Beck Depression InventorySecond Edition; STAI = State Trait Anxiety Inventory; BS E = Barriers Self Efficacy Scale; EXSE = Exercise Self-Efficacy Scale; VO2 = Modified Balke Submax; LTEQ = Leisure Time Exercise Questionnaire; MVPA = Minutes of Moderate and Vi gorous Physical Activity

PAGE 91

91 Table 4-5. Baseline correlations between psychosocial and cognitive variables GDS BDI-2 STAI State Anxiety STAI Trait Anxiety BSE EXSE NAART 0.185 0.086 -0.005 0.124 -0.130 -0.197 Trails A Time 0.146 *0.249 0.166 0.059 0.129 -0.001 LM Immediate Recall -0.017 -0.003 -0.139 0.001 *-0.307 *-0.248 LM Delayed Recall -0.033 -0.028 -0.144 -0.010 **-0.320 *-0.288 LM Recognition *-0.239 -0.187 *-0.253 *-0.251 -0.074 -0.055 LM Learning Slope -0.009 -0.117 0.050 -0.040 0.008 -0.048 COWA 0.080 -0.012 -0.002 0.125 -0.018 0.005 Trails B Time **0.341 **0.314 0.142 0.169 0.112 -0.065 Letter Number Sequencing 0.095 0.131 -0.031 0.066 0.023 -0.065 One-Back: Mean RT 0.008 0.017 0.022 -0.101 -0.063 -0.138 One-Back: Mean RT SD 0.023 0.014 -0.010 -0.031 0.074 -0.019 One-Back: Number Correct -0.036 -0.057 -0.136 -0.039 0.210 0.099 Two-Back: Mean RT -0.069 -0.005 0.099 -0.133 0.137 0.034 Two-Back: Mean RT SD 0.057 0.029 0.068 -0.101 0.174 0.052 Two-Back: Number Correct 0.024 -0.123 -0.071 -0.008 0.002 0.054 Note: p < .05; ** p < .01; LM = Logical Memory; NAART = North American Adult Reading Tes t; COWA = Controlled Oral Word Association; RT = Reaction Time; SD = Standard Deviation; GDS = Geriatric Depression Scal e; BDI-2 = Beck Depression InventorySecond Edition; STAI = State Tr ait Anxiety Inventory; BSE = Barriers Self Efficacy Scale; EXSE = Exercise SelfEfficacy Scale

PAGE 92

92 Table 4-6. Cognitive intercorrelations at baseline COWA Trails B TimeLetter-Number Sequenci ng One-Back: Mean RTOne-Back: Mean RT SD COWA 1 Trails B Time -0.1151 Letter Number Sequencing **0.488*-0.252 1 One-Back: Mean RT -0.1810.155 -0.182 1 One-Back: Mean RT SD -0.1380.236 -0.210 **0.727 1 One-Back: Number Correct -0.062-0.136 -0.047 0.034 -0.050 Two-Back: Mean RT -0.0930.095 -0.105 0.204 0.157 Two-Back: Mean RT SD -0.1560.104 -0.114 *0.241 0.219 Two-Back: Number Correct *0.252*-0.289 0.205 *-0.270 *-0.237 NAART **0.394-0.031 0.178 -0.023 -0.234 Trails A Time *-0.282**0.519 *-0.381 0.128 0.045 LM Immediate Recall 0.210-0.225 *0.238 -0.177 **-0.381 LM Delayed Recall *0.280**-0.359 *0.243 -0.141 **-0.367 LM Recognition 0.214*-0.293 0.077 -0.119 -0.205 LM Learning Slope -0.053-0.138 0.076 0.062 0.117 Note : p < .05; ** p < .01; LM = Logical Memory; NAART = North American Adult Reading Tes t; COWA = Controlled Oral Word Association; RT = Reaction Time; SD = Standard Deviation

PAGE 93

93 Table 4-6 Continued. One-Back: Number Correct Two-Back: Mean RT Two-Back: Mean RT SD Two-Back: Number Correct NAART One-Back: Number Correct 1 Two-Back: Mean RT 0.173 1 Two-Back: Mean RT SD 0.142 *0.841 1 Two-Back: Number Correct *0.277 0.057 0.001 1 NAART -0.033 -0.109 -0.227 0.195 1 Trails A Time -0.005 0.048 0.002 *-0.300 0.055 LM Immediate Recall -0.083 -0.142 -0.182 0.182 **0.335 LM Delayed Recall 0.011 -0.197 -0.190 *0.273 **0.337 LM Recognition 0.106 -0.038 -0.063 *0.295 0.180 LM Learning Slope 0.164 0.008 0.015 -0.054 -0.006 Note : p < .05; ** p < .01; LM = Logical Memory; NAART = North American Adult Reading Tes t; COWA = Controlled Oral Word Association; RT = Reaction Time; SD = Standard Deviation

PAGE 94

94 Table 4-6 Continued Trails A TimeLM Immediate Recall LM Delayed Recall LM RecognitionLM Learning Slope Trails A Time 1 LM Immediate Recall -0.068 1 LM Delayed Recall -0.204 **0.840 1 LM Recognition -0.118 **0.578 **0.654 1 LM Learning Slope -0.191 -0.145 -0.115 -0.031 1 Note : p < .05; ** p < .01; LM = Logical Memory; NAART = North American Adult Reading Tes t; COWA = Controlled Oral Word Association; RT = Reaction Time; SD = Standard Deviation

PAGE 95

95 Table 4-7. Univariate analyses of vari ance for physical activity and fitness Effects Hypothesis dfError df F P -value Partial 2 Between Subjects: Study Group Mean Pedometer Steps 1 67 0.00 0.96 0.00 LTEQ 1 67 0.45 0.50 0.01 VO2 1 67 0.68 0.41 0.01 MVPA 1 67 0.37 0.54 0.01 BSE 1 67 3.37 0.07 0.05 EXSE 1 67 7.30 **0.01 0.10 Within-Subjects: Occasion Mean Pedometer Steps 1 67 3.19 0.08 0.05 LTEQ 1 67 17.78 **0.00 0.21 VO2 1 67 7.97 **0.01 0.11 MVPA 1 67 17.98 **0.00 0.21 BSE 1 67 1.05 0.31 0.02 EXSE 1 67 5.82 *0.02 0.08 Interactions: Study Group by Occasion Mean Pedometer Steps 1 67 0.95 0.33 0.01 LTEQ 1 67 0.00 1.00 0.00 VO2 1 67 0.36 0.55 0.01 MVPA 1 67 0.59 0.45 0.01 BSE 1 67 0.25 0.62 0.00 EXSE 1 67 1.71 0.19 0.02 Note : p < .05; ** p < .01; LTEQ = Leisure Time Exercise Questionnaire; VO2 = Modified Balke Submax; BSE = Barriers Self Efficacy Scale; EX SE = Exercise Self-Efficacy Scale; MVPA = Minutes of Moderate and Vigorous Exercise

PAGE 96

96 Table 4-8. Repeated-measures multivariate analysis of variance on anxiety and depression measuresa Effects Wilks' Lambda F Hypothesis dfError df P -value Partial 2 Between Subjects: Study Group 0.97 0.414 64 0.800.03 Within Subjects: Occasion 0.71 6.684 64 **0.000.29 Interaction: Study Group by Occasion 0.94 1.004 64 0.410.06 Note. *p < .05; **p < .01; aThere were unequal variance-covariance matrices for these measures.

PAGE 97

97 Table 4-9. Follow-up univariate analyses of variance for depression and anxiety measures Effects Hypothesis df Error df F P -value Partial 2 Between-Subjects: Study Group STAI State Anxiety 1 67 0.04 0.85 0.00 STAI Trait Anxiety 1 67 1.10 0.30 0.02 GDS 1 67 1.33 0.25 0.02 BDI-2 1 67 1.46 0.23 0.02 Within-Subjects: Occasion STAI State Anxiety 1 67 1.07 0.31 0.02 STAI Trait Anxiety 1 67 11.58 **0.00 0.15 GDS 1 67 0.01 0.91 0.00 BDI-2 1 67 22.68 **0.00 0.25 Interactions: Study Group by Occasion STAI State Anxiety 1 67 1.53 0.22 0.02 STAI Trait Anxiety 1 67 2.14 0.15 0.03 GDS 1 67 0.30 0.58 0.00 BDI-2 1 67 0.03 0.87 0.00 Note : p < .05; ** p < .01; GDS = Geriatric Depression Scale; BDI-2 = B eck Depression InventorySecond Edition; STAI = State Trait Anxiety Inventory

PAGE 98

98 Table 4-10. Repeated measures multivariate analyses of variance for cognitive domai ns with study group as between-subjects fact or Effects Wilks' Lambda F Hypothesis dfError df P -value Partial 2 Non-Executive Measures Between Subjects: Study Group 0.90 1.17 6 62 0.330.10 Within Subjects: Occasion 0.55 8.35 6 62 **0.000.45 Interaction: Study Group by Occasion 0.92 0.89 6 62 0.510.08 Executive Measuresa Between Subjects: Study Group 0.97 0.23 9.00 59.00 0.990.03 Within Subjects: Occasion 0.65 3.59 9.00 59.00 **0.000.35 Interaction: Study Group by Occasion 0.86 1.09 9.00 59.00 0.380.14 Note : p< .05; ** p < .01; aThere were unequal variance-covariance matrices for these measures.

PAGE 99

99 Table 4-11. Follow-up univariate analyses of variance for non-executive measures Effects Hypothesis dfError df F P -value Partial 2 Between-Subjects: Study Group NAART 1 67 1.31 0.26 0.02 Trails A Time 1 67 0.02 0.88 0.00 LM Immediate Total Recall 1 67 2.52 0.12 0.04 LM Delayed Total Recall 1 67 1.49 0.23 0.02 LM Recognition 1 67 0.55 0.46 0.01 LM Learning Slope 1 67 1.14 0.29 0.02 Within-Subjects: Occasion NAART 1 67 10.68 **0.00 0.14 Trails A Time 1 67 6.89 *0.01 0.09 LM Immediate Total Recall 1 67 24.85 **0.00 0.27 LM Delayed Total Recall 1 67 30.69 **0.00 0.31 LM Recognition 1 67 4.82 *0.03 0.07 LM Learning Slope 1 67 1.15 0.29 0.02 Interactions: Study Group by Occasion NAART 1 67 0.01 0.93 0.00 Trails A Time 1 67 0.44 0.51 0.01 LM Immediate Total Recall 1 67 0.12 0.73 0.00 LM Delayed Total Recall 1 67 0.61 0.44 0.01 LM Recognition 1 67 1.78 0.19 0.03 LM Learning Slope 1 67 0.88 0.35 0.01 Note *p < .05; ** p < .01; LM = Logical Memory; NAART = North American Adult Reading Test

PAGE 100

100 Table 4-12. Follow-up univariate analyses of variance for executive measures Effects Hypothesis dfError df F P -value Partial 2 Between Subjects: Study Group COWA 1 67 0.06 0.80 0.00 Trails B 1 67 0.56 0.46 0.01 Letter-Number Sequencing 1 67 0.15 0.70 0.00 One-Back: Mean RT 1 67 0.94 0.34 0.01 One-Back: Mean RT SD 1 67 0.32 0.57 0.00 One-Back: Number Correct 1 67 0.01 0.92 0.00 Two-Back: Mean RT 1 67 0.04 0.84 0.00 Two-Back: Mean RT SD 1 67 0.01 0.91 0.00 Two-Back: Number Correct 1 67 0.20 0.66 0.00 Within-Subjects: Occasion COWA 1 67 7.04 **0.01 0.10 Trails B 1 67 0.50 0.48 0.01 Letter-Number Sequencing 1 67 0.15 0.70 0.00 One-Back: Mean RT 1 67 11.28 **0.00 0.14 One-Back: Mean RT SD 1 67 16.80 **0.00 0.20 One-Back: Number Correct 1 67 2.61 0.11 0.04 Two-Back: Mean RT 1 67 1.87 0.18 0.03 Two-Back: Mean RT SD 1 67 8.28 **0.01 0.11 Two-Back: Number Correct 1 67 7.95 **0.01 0.11 Interactions: Study Group by Occasion COWA 1 67 4.15 *0.05 0.06 Trails B 1 67 0.52 0.47 0.01 Letter-Number Sequencing 1 67 1.33 0.25 0.02 One-Back: Mean RT 1 67 0.51 0.48 0.01 One-Back: Mean RT SD 1 67 2.16 0.15 0.03 One-Back: Number Correct 1 67 0.18 0.67 0.00 Two-Back: Mean RT 1 67 0.15 0.70 0.00 Two-Back: Mean RT SD 1 67 0.68 0.41 0.01 Two-Back: Number Correct 1 67 0.84 0.36 0.01 Note : p < .05; ** p < .01; COWA = Controlled Oral Wo rd Association; RT = Reaction Time; SD = Standard Deviation

PAGE 101

101Table 4-13. Raw cognitive estimated marginal means by study group Measure Control Group Exercise Promotion Group Non-Executive Baseline Post-Test Baseline Post-Test Mean Std. Error Mean Std. Error Mean Std. Error Mean Std. Error NAART 113.88 1.09 115.14 1.05 112.22 1.11 113.42 1.06 Trails A Time 32.32 1.89 30.06 1.30 32.78 1.30 28.98 1.32 LM Immediate Recall 39.37 1.43 44.51 1.54 42.71 1.45 47.18 1.57 LM Delayed Recall 25.09 1.27 28.54 1.22 26.50 1.29 31.09 1.24 LM Recognition 26.03 0.40 26.95 0.29 26.69 0.40 26.91 0.30 LM Learning Slope 4.48 0.45 3.60 0.38 3.65 0.46 3.59 0.38 Executive Baseline Post-Test Baseline Post-Test Mean Std. Error Mean Std. Error Mean Std. Error Mean Std. Error COWA 40.14 1.84 40.73 1.67 38.79 1.87 43.24 1.69 Trails B Time 85.68 6.75 78.48 6.33 76.19 6.84 76.27 6.42 Letter Number Sequencing 10.56 0.45 10.20 0.38 10.50 0.45 10.68 0.39 One-Back: Mean RT 860.36 29.59 776.51 25.01 813.52 30.02 759.09 25.37 One-Back: Mean RT SD 339.04 31.17 218.51 17.20 291.09 31.63 234.15 17.45 One-Back: Number Correct 95.09 2.39 98.44 0.46 95.97 2.43 97.93 0.47 Two-Back: Mean RT 1663.5090.39 1588.65102.79 1670.6291.71 1535.99104.29 Two-Back: Mean RT SD 916.13 83.79 715.60 70.90 860.75 85.02 749.45 71.93 Two-Back: Number Correct 90.81 1.62 94.36 1.01 90.94 1.64 92.76 1.02 Note : LM = Logical Memory; NAART = North American Adult R eading Test; COWA = Controlle d Oral Word Association; RT = Reaction Time; SD = Standard Deviation

PAGE 102

102Table 4-14. Calculation of reliable change index scores adjusted for control group practice effect Pre SD Post SD r.t1.t2 SE of Difference Reliable Change Control Group Mean Change Reliable Decline Reliable Improvement NAART 6.46 6.21 0.85 3.47 5.70 1.26 -4.44 6.95 LM Immediate Recall 8.57 9.16 0.58 8.16 13.38 -2.27 -15.64 11.11 LM Learning Slope 2.67 2.22 -0.25 3.88 6.36 5.14 -1.22 11.50 LM Delayed Recall 7.52 7.29 0.59 6.68 10.95 3.46 -7.50 14.41 LM Recognition 2.34 1.71 0.51 2.04 3.34 0.92 -2.42 4.27 COWA 10.84 9.86 0.64 8.76 14.36 -0.89 -15.25 13.47 Trails A Time 11.10 7.68 0.70 7.45 12.22 0.58 -11.64 12.80 Trails B time 39.90 37.16 0.48 39.21 64.30 -7.19 -71.49 57.10 Letter-Number Sequencing 2.62 2.25 0.71 1.86 3.04 -0.36 -3.40 2.69 One-Back Mean RT 175.35 147.10 0.48 165.68 271.72 -83.86 -355.58 187.87 One-Back Mean RT SD 184.63 101.31 0.45 156.47 256.61 -120.52 -377.13 136.08 One-Back Number Correct 14.05 2.74 0.01 14.28 23.42 3.35 -20.07 26.77 Two-Back Mean RT 530.85 604.22 0.31 669.54 1098.05 -74.84 -1172.89 1023.21 Two-Back Mean RT SD 492.85 416.68 0.52 449.00 736.35 -200.53 -936.88 535.83 Two-Back Number Correct 9.51 5.97 0.57 7.39 12.11 3.56 -8.56 15.67 Note : NAART = North American Adult Reading Test; COWA = Cont rolled Oral Word Association; RT = Reaction Time; SD = Standard Deviation; SE = Standa rd Error; Reliable Change = Standard Error of Difference 1.64

PAGE 103

103 Table 4-15. Participants who reliably changed on non-executi ve cognitive measures Control (n= 35) Intervention (n = 34) Total (n = 69) NAART Reliable Decline 2 (5.7%) 1 (2.9%) 3 (4.3%) No Reliable Change 31 (88.6%) 32 (94.1%) 63 (91.3%) Reliable Improvement 2 (5.7%) 1 (2.9%) 3 (4.3%) LM Immediate Recall No Reliable Change 25 (71.4%) 27 (79.4%) 52 (75.4%) Reliable Improvement 10 (28.6%) 7 (20.6%) 17 (24.6%) LM Delayed Recall Reliable Decline 3 (8.6%) 0 (0%) 3 (4.3%) No Reliable Change 31 (88.6%) 33 (97.1%) 64 (92.8%) Reliable Improvement 1 (2.9%) 1 (2.9%) 2 (2.9%) LM Learning Slope Reliable Decline 13 (37.1%) 14 (41.2%) 27 (39.1%) No Reliable Change 22 (62.9%) 20 (58.8%) 42 (60.9%) LM Recognition Reliable Decline 3 (8.6%) 2 (5.9%) 5 (7.2%) No Reliable Change 29 (82.9%) 31 (91.2%) 60 (87.0%) Reliable Improvement 3 (8.6%) 1 (2.9%) 4 (5.8%) Trails A Time Reliable Decline 3 (8.6%) 6 (17.6%) 9 (13%) No Reliable Change 31 (88.6%) 27 (79.4%) 58 (84.1%) Reliable Improvement 1 (2.9%) 1 (2.9%) 2 (2.9%) Note : NAART = North American Adult R eading Test; LM = Logical Memory

PAGE 104

104 Table 4-16. Participants who reliably ch anged on executive cognitive measures Control (n= 35) Intervention (n = 34) Total (n = 69) COWA Reliable Decline 1 (2.9%) 0 (0%) 1 (1.4%) No Reliable Change 32 (91.4%) 31 (91.2%) 63 (91.3%) Reliable Improvement 2 (5.7%) 3 (8.8%) 5 (7.2%) Trails B Time Reliable Decline 2 (5.7%) 1 (2.9%) 3 (4.3%) No Reliable Change 30 (85.7%) 32 (94.1%) 62 (89.9%) Reliable Improvement 3 (8.6%) 1 (2.9%) 4 (5.8%) Letter-Number Sequencing Reliable Decline 3 (8.6%) 1 (2.9%) 4 (5.8%) No Reliable Change 32 (91.4%) 29 (85.3%) 61 (88.4%) Reliable Improvement 0 (0%) 4 (11.8%) 4 (5.8%) One-Back Mean RT Reliable Decline 2 (5.7%) 1 (2.9%) 3 (4.3%) No Reliable Change 31 (88.6%) 30 (88.2%) 30 (88.2%) Reliable Improvement 2 (5.7%) 3 (8.8%) 3 (8.8%) One-Back Mean RT SD Reliable Decline 4 (11.4%) 2 (5.9%) 6 (8.7%) No Reliable Change 31 (88.6%) 30 (88.2%) 61 (88.4%) Reliable Improvement 0 (0%) 2 (2.9%) 2 (2.9%) One-Back Number Correct No Reliable Change 34 (97.1%) 32 (94.1%) 66 (95.7%) Reliable Improvement 1 (2.9%) 2 (5.9%) 3 (4.3%) Note : COWA= Controlled Auditory Word Asso ciation; RT=Reaction Time; SD=Standard Deviation

PAGE 105

105 Table 4-16. Continued Control (n= 35) Intervention (n = 34) Total (n = 69) Two-Back Mean RT Reliable Decline 2 (5.7%) 2 (5.9%) 4 (5.8%) No Reliable Change 31 (88.6%) 30 (88.2%) 61 (88.4%) Reliable Improvement 2 (5.7%) 2 (5.9%) 4 (5.8%) Two-Back Mean RT SD Reliable Decline 3 (8.6%) 1 (2.9%) 4 (5.8%) No Reliable Change 31 (88.6%) 30 (88.2%) 61 (88.4%) Reliable Improvement 1 (2.9%) 3 (8.8%) 4 (5.8%) Two-Back Number Correct Reliable Decline 1 (2.9%) 2 (5.9%) 3 (4.3%) No Reliable Change 32 (91.4%) 30 (88.2%) 62 (89.9%) Reliable Improvement 2 (5.7%) 2 (5.9%) 4 (5.8%) Note : COWA= Controlled Auditory Word Asso ciation; RT=Reaction Time; SD=Standard Deviation

PAGE 106

106 Table 4-17. Multivariate analyses of varian ce for cognitive domains with study group and age group as between-subjects factors Wilks' Lambda F Hypothesis dfError df P -value Partial 2 Non-Executive Measures Between Subjects: Study Group 0.89 1.20 6 60 0.32 0.11 Age Group 0.83 2.09 6 60 0.07 0.17 Study Group by Age Group 0.90 1.17 6 60 0.33 0.10 Within Subjects: Occasion 0.55 8.11 6 60 **0.00 0.45 Interaction: Study Group by Occasion 0.92 0.83 6 60 0.55 0.08 Age Group by Occasion 0.97 0.34 6 60 0.91 0.03 Study Group by Age Group by Occasion 0.94 0.63 6 60 0.70 0.06 Executive Measures Between Subjects: Study Group 0.96 0.27 9 57 0.98 0.04 Age Group 0.73 2.33 9 57 *0.03 0.27 Study Group by Age Group 0.77 1.91 9 57 0.07 0.23 Within Subjects: Occasion 0.64 3.58 9 57 **0.00 0.36 Interaction: Study Group by Occasion 0.86 1.01 9 57 0.44 0.14 Age Group by Occasion 0.89 0.80 9 57 0.62 0.11 Study Group by Age Group by Occasion0.83 1.34 9 57 0.24 0.17 Note : p < .05; ** p < .01;

PAGE 107

107 Table 4-18. Follow-up univariate analyses of variance for non-executive measures: Study group and age group as between subjects factors Effects Hypothesis df Error df F P -value Partial 2 NAART Study Group 1 65 1.760.19 0.03 Age Group 1 65 2.420.12 0.04 Study Group by Age Group 1 65 0.480.49 0.01 Occasion 1 65 10.50**0.00 0.14 Study Group by Occasion 1 65 0.000.95 0.00 Age Group by Occasion 1 65 0.290.59 0.00 Study Group by Age Group by Occasion 1 65 0.050.83 0.00 Trails A Time Study Group 1 65 0.080.77 0.00 Age Group 1 65 7.50**0.01 0.10 Study Group by Age Group 1 65 0.480.49 0.01 Occasion 1 65 7.74**0.01 0.11 Study Group by Occasion 1 65 0.730.40 0.01 Age Group by Occasion 1 65 0.360.55 0.01 Study Group by Age Group by Occasion 1 65 2.160.15 0.03 LM Immediate Total Recall Study Group 1 65 2.120.15 0.03 Age Group 1 65 0.080.78 0.00 Study Group by Age Group 1 65 1.800.18 0.03 Occasion 1 65 23.11**0.00 0.26 Study Group by Occasion 1 65 0.170.68 0.00 Age Group by Occasion 1 65 0.210.65 0.00 Study Group by Age Group by Occasion 1 65 0.140.71 0.00 LM Delayed Total Recall Study Group 1 65 1.030.31 0.02 Age Group 1 65 0.080.77 0.00 Study Group by Age Group 1 65 3.170.08 0.05 Occasion 1 65 29.03**0.00 0.31 Study Group by Occasion 1 65 0.490.49 0.01 Age Group by Occasion 1 65 0.000.95 0.00 Study Group by Age Group by Occasion 1 65 0.430.52 0.01 Note : p < .05; ** p < .01; LM = Logical Memory; NAART = North American Adult Reading Test

PAGE 108

108 Table 4-18. Continued Effects Hypothesis df Error df F P -value Partial 2 LM Recognition Study Group 1 65 0.450.51 0.01 Age Group 1 65 0.360.55 0.01 Study Group by Age Group 1 65 0.010.91 0.00 Occasion 1 65 5.160.03 0.07 Study Group by Occasion 1 65 1.450.23 0.02 Age Group by Occasion 1 65 0.600.44 0.01 Study Group by Age Group by Occasion 1 65 0.230.63 0.00 LM Learning Slope Study Group 1 65 0.900.35 0.01 Age Group 1 65 0.070.79 0.00 Study Group by Age Group 1 65 0.550.46 0.01 Occasion 1 65 1.050.31 0.02 Study Group by Occasion 1 65 1.020.32 0.02 Age Group by Occasion 1 65 0.250.62 0.00 Study Group by Age Group by Occasion 1 65 1.350.25 0.02 Note : p < .05; ** p < .01; LM = Logical Memory; NAART = North American Adult Reading Test

PAGE 109

109 Table 4-19. Follow-up univariate analyses of variance for executive measures: Study group and age group as between subjects factors Effects Hypothesis df Error df F P value Partial 2 COWA Study Group 1 65 0.04 0.84 0.00 Age Group 1 65 0.15 0.70 0.00 Study Group by Age Group 1 65 0.50 0.48 0.01 Occasion 1 65 7.00 *0.01 0.10 Study Group by Occasion 1 65 4.03 *0.05 0.06 Age Group by Occasion 1 65 0.51 0.48 0.01 Study Group by Age Group by Occasion1 65 0.23 0.64 0.00 Trails B Time Study Group 1 65 0.57 0.45 0.01 Age Group 1 65 3.46 0.07 0.05 Study Group by Age Group 1 65 1.93 0.17 0.03 Occasion 1 65 0.92 0.34 0.01 Study Group by Occasion 1 65 0.20 0.65 0.00 Age Group by Occasion 1 65 0.29 0.59 0.00 Study Group by Age Group by Occasion1 65 5.93 *0.02 0.08 Letter-Number Sequencing Study Group 1 65 0.16 0.69 0.00 Age Group 1 65 0.86 0.36 0.01 Study Group by Age Group 1 65 0.19 0.67 0.00 Occasion 1 65 0.10 0.75 0.00 Study Group by Occasion 1 65 1.39 0.24 0.02 Age Group by Occasion 1 65 0.33 0.57 0.01 Study Group by Age Group by Occasion1 65 0.01 0.94 0.00 One-Back: Mean RT Study Group 1 65 1.17 0.28 0.02 Age Group 1 65 6.54 *0.01 0.09 Study Group by Age Group 1 65 1.19 0.28 0.02 Occasion 1 65 11.21 **0.00 0.15 Study Group by Occasion 1 65 0.57 0.45 0.01 Age Group by Occasion 1 65 5.45 *0.02 0.08 Study Group by Age Group by Occasion1 65 2.26 0.14 0.03 Note : p < .05; ** p < .01; COWA = Controlled Oral Wo rd Association; RT = Reaction Time; SD = Standard Deviation

PAGE 110

110 Table 4-19 Continued. Effects Hypothesis df Error df F P -value Partial 2 One-Back: Mean RT SD Study Group 1 65 0.16 0.69 0.00 Age Group 1 65 0.62 0.43 0.01 Study Group by Age Group 1 65 5.32 *0.02 0.08 Occasion 1 65 17.71 **0.00 0.21 Study Group by Occasion 1 65 1.74 0.19 0.03 Age Group by Occasion 1 65 0.24 0.63 0.00 Study Group by Age Group by Occasion1 65 3.51 0.07 0.05 One-Back Number Correct Study Group 1 65 0.01 0.91 0.00 Age Group 1 65 0.28 0.60 0.00 Study Group by Age Group 1 65 0.26 0.61 0.00 Occasion 1 65 2.64 0.11 0.04 Study Group by Occasion 1 65 0.18 0.67 0.00 Age Group by Occasion 1 65 0.62 0.44 0.01 Study Group by Age Group by Occasion1 65 0.38 0.54 0.01 Two-Back: Mean RT Study Group 1 65 0.07 0.79 0.00 Age Group 1 65 5.55 *0.02 0.08 Study Group by Age Group 1 65 1.12 0.29 0.02 Occasion 1 65 1.76 0.19 0.03 Study Group by Occasion 1 65 0.11 0.74 0.00 Age Group by Occasion 1 65 0.18 0.67 0.00 Study Group by Age Group by Occasion1 65 0.72 0.40 0.01 Two-Back: Mean RT SD Study Group 1 65 0.08 0.77 0.00 Age Group 1 65 5.78 *0.02 0.08 Study Group by Age Group 1 65 0.00 0.99 0.00 Occasion 1 65 7.78 **0.01 0.11 Study Group by Occasion 1 65 0.66 0.42 0.01 Age Group by Occasion 1 65 0.31 0.58 0.00 Study Group by Age Group by Occasion1 65 0.11 0.74 0.00 Note : p < .05; ** p < .01; COWA = Controlled Oral Wo rd Association; RT = Reaction Time; SD = Standard Deviation

PAGE 111

111 Table 4-19 Continued. Hypothesis df Error df F P -value Partial 2 Two-Back: Number Correct Study Group 1 65 0.07 0.79 0.00 Age Group 1 65 2.12 0.15 0.03 Study Group by Age Group 1 65 0.48 0.49 0.01 Occasion 1 65 7.43 **0.01 0.10 Study Group by Occasion 1 65 0.83 0.37 0.01 Age Group by Occasion 1 65 0.20 0.65 0.00 Study Group by Age Group by Occasion1 65 0.03 0.86 0.00 Note : p < .05; ** p < .01; COWA = Controlled Oral Wo rd Association; RT = Reaction Time; SD = Standard Deviation

PAGE 112

112 Table 4-20. Raw cognitive estimated marginal means by age and study group Young er Group Control Group Exercise Promotion Group Non-Executive Baseline Post-Test Baseline Post-Test Mean Std. Error Mean Std. Error Mean Std. Error Mean Std. Error NAART 113.07 1.56 114.63 1.50 110.19 1.71 111.53 1.65 Trails A Time 28.38 2.66 27.16 1.75 31.87 2.93 25.25 1.93 LM Immediate Recall 41.00 2.06 46.06 2.22 42.00 2.27 45.50 2.44 LM Delayed Recall 26.06 1.83 30.06 1.73 24.79 2.02 28.86 1.91 LM Recognition 25.88 0.57 26.88 0.42 26.31 0.63 26.93 0.47 LM Learning Slope 4.76 0.65 3.12 0.54 3.71 0.72 4.00 0.60 Executive Baseline Post-Test Baseline Post-Test Mean Std. Error Mean Std. Error Mean Std. Error Mean Std. Error COWA 40.88 2.68 42.65 2.40 38.21 2.95 42.93 2.65 Trails B Time 68.01 9.29 70.40 8.97 82.68 10.24 65.52 9.88 Letter Number Sequencing 10.88 0.65 10.65 0.55 10.57 0.71 10.93 0.60 One-Back: Mean RT 760.93 39.14 755.63 36.15 776.21 43.13 741.20 39.84 One-Back: Mean RT SD 268.29 43.28 200.10 24.83 333.60 47.69 241.58 27.36 One-Back: Number Correct 92.88 3.46 98.65 0.68 95.79 3.82 98.07 0.74 Two-Back: Mean RT 1518.78 128.37 1341.44 143.18 1566.19 141.46 1470.00 157.78 Two-Back: Mean RT SD 773.49 117.76 623.45 100.35 720.47 129.76 623.33 110.58 Two-Back: Number Correct 91.78 2.33 94.71 1.44 93.21 2.57 94.71 1.58 Note : LM = Logical Memory; NAART = North American Adult Readin g Test; COWA = Controlled Oral Word Association; RT = Reaction Time; SD = Standard Deviation

PAGE 113

113 Table 4-20 Continued. Older Group Control Group Exercise Promotion Group Non-Executive Baseline Post-Test Baseline Post-Test Mean Std. Error Mean Std. Error Mean Std. Error Mean Std. Error NAART 114.64 1.51 115.62 1.46 113.64 1.43 114.74 1.38 Trails A Time 36.04 2.59 32.79 1.70 33.42 2.45 31.60 1.61 LM Immediate Recall 37.83 2.00 43.06 2.16 43.20 1.90 48.35 2.04 LM Delayed Recall 24.17 1.78 27.11 1.68 27.70 1.69 32.65 1.59 LM Recognition 26.17 0.56 27.01 0.41 26.95 0.53 26.90 0.39 LM Learning Slope 4.21 0.63 4.05 0.53 3.60 0.60 3.30 0.50 Executive Baseline Post-Test Baseline Post-Test Mean Std. Error Mean Std. Error Mean Std. Error Mean Std. Error COWA 39.44 2.60 38.91 2.34 39.20 2.47 43.45 2.22 Trails B Time 102.36 9.03 86.12 8.72 71.65 8.57 83.79 8.27 Letter Number Sequencing 10.25 0.63 9.78 0.53 10.45 0.60 10.50 0.51 One-Back: Mean RT 954.28 38.04 796.23 35.13 839.64 36.09 771.61 33.33 One-Back: Mean RT SD 405.85 42.06 235.91 24.13 261.34 39.90 228.94 22.89 One-Back: Number Correct 97.17 3.37 98.24 0.66 96.10 3.19 97.83 0.62 Two-Back: Mean RT 1800.17 124.75 1822.14 139.15 1743.72 118.35 1582.19 132.01 Two-Back: Mean RT SD 1050.84 114.44 802.63 97.52 958.95 108.57 837.74 92.52 Two-Back: Number Correct 89.89 2.26 94.04 1.40 89.35 2.15 91.39 1.33 Note : LM = Logical Memory; NAART = North American Adult Readin g Test; COWA = Controlled Oral Word Association; RT = Reaction Time; SD = Standard Deviation

PAGE 114

114 Table 4-21. Correlations between physical activity, physical fitness, and cognitive change scores VO2 LTEQ Mean Pedometer Steps MVPA NAART -0.09 -0.01 0.06 -0.02 Trails A Time -0.03 -0.13 -0.02 -0.18 LM Immediate Recall *0.31 0.01 0.14 0.01 LM Delayed Recall 0.14 0.03 0.09 0.06 LM Recognition 0.09 -0.01 0.06 -0.09 LM Learning Slope -0.02 -0.13 *-0.29 -0.16 COWA -0.18 0.03 -0.05 0.08 Trails B Time -0.13 0.23 -0.06 0.16 Letter Number Sequencing 0.09 -0.01 -0.15 -0.10 One-Back: Mean RT 0.21 -0.07 0.13 0.04 One-Back: Mean RT SD 0.11 *-0.25 0.04 -0.21 Two-Back: Number Correct -0.11 0.00 0.09 -0.06 Two-Back: Mean RT -0.21 -0.01 0.11 -0.03 Two-Back: Mean RT SD -0.21 -0.13 0.02 -0.08 Two-Back: Number Correct -0.20 0.09 0.02 0.10 Note: *p < .05; ** p < .01; LM = Logical Memo ry; NAART = North American Adult Reading Test; COWA = Controlled Or al Word Association; RT = Reaction Time; SD = Standard Deviation; VO2 = Modified Balke Submax; LTEQ = Leisure Time Exercise Questionnaire ; MVPA = Minutes of Moderate and Vigorous Physical Activity

PAGE 115

115 Table 4-22. Correlations between physical fitness, physical ac tivity, and psychosocial change scores VO2 LTEQ Mean Pedometer Steps MVPA GDS *-0.25 0.05 -0.14 0.08 BDI-2 0.02 0.02 -0.06 -0.04 STAI State Anxiety -0.04 0.18 0.08 0.20 STAI Trait Anxiety -0.05 -0.11 0.14 -0.11 BSE 0.04 0.12 -0.06 0.08 EXSE -0.15 0.11 0.05 0.09 Note: p < .05; ** p < .01; GDS = Geriatric Depression Scale; BDI-2 = Beck Depression InventorySecond Editio n; STAI = State Trait Anxiety Inventory; BSE = Barriers Self Effi cacy Scale; EXSE = Exercise SelfEfficacy Scale; VO2 = Modified Balke Submax; LTEQ = Leisure Time Exercise Questionnaire

PAGE 116

116 Table 4-23. Correlations between psychosocial and cognitive change scores GDSBDI-2STAI State Anxiety STAI Trait AnxietyBSEEXSE NAART -0.01 -0.14 -0.01 -0.06 0.06 -0.15 Trails A Time 0.03 0.05 -0.06 -0.11 0.18 0.12 LM Immediate Recall -0.22 -0.09 -0.05 0.15 0.11 0.08 LM Delayed Recall *-0.24 0.01 0.03 0.21 -0.01 -0.02 LM Recognition -0.02 -0.03 -0.12 0.16 0.07 -0.12 LM Learning Slope -0.03 -0.03 0.04 -0.20 0.13 0.04 COWA -0.17 -0.01 -0.10 0.00 0.22 0.02 Trails B Time -0.09 0.21 0.18 -0.02 0.04 -0.09 Letter Number Sequencing 0.08 0.19 0.08 0.15 0.08 -0.16 One-Back: Mean RT 0.04 -0.08 0.01 -0.08 -0.10 -0.13 One-Back: Mean RT SD -0.02 -0.03 0.00 0.01 -0.11 -0.14 One-Back: Number Correct 0.00 -0.20 0.03 -0.17 0.15 0.09 Two-Back: Mean RT 0.12 -0.19 0.20 0.01 0.12 -0.01 Two-Back: Mean RT SD 0.02 *-0.26 0.21 -0.08 0.18 -0.03 Two-Back: Number Correct 0.06 -0.11 0.15 -0.01 *-0.28 -0.02 Note: p < .05; ** p < .01; LM = Logical Memory; NAART = North American Adult Reading Te st; COWA = Controlled Oral Word Association; RT = Reaction Time; SD = Standard Deviation; GDS = Geriatric Depression Scale; BDI-2 = Beck Depression InventorySecond Edition; STAI = State Tr ait Anxiety Inventory; BSE = Barriers Self Effi cacy Scale; EXSE = Exercise Self-Efficacy Scale

PAGE 117

117 A 3990 4990 5990 6990 7990 Baseline Post-Test OccasionMean Number of Steps Per WeekContro l Intervention 21 23 25 27 29 31 33 35 37 BaselinePost-Test OccasionLevel of Cardiorespiratory FitnessContro l InterventionB Figure 4-1. Mean scores by study group and oc casion for physical fitness, activity, a nd self-efficacy variables A) Mean pedomet er steps. B) Cardiorespiratory fitness (VO2 ). C) Leisure Time Exercise (LTEQ). D) Mean minutes of moderate or vigorous physical activity. E) Barriers of Self Efficacy (BSE). F) Exercise Self-Efficacy.

PAGE 118

118 C 0 2 4 6 8 10 12 14 16 BaselinePost-Test OccasionMETsControl Interventio -42 8 58 108 158 208 258 Baseline Post-Test OccasionMinutes of Moderate and Vigorous Physical Activity Per WeekContro l Interventio n D Figure 4-1 Continued.

PAGE 119

119 E 44 49 54 59 64 69 74 79 84 Baseline Post-Test OccasionMean BSE Score 40 50 60 70 80 90 100 Baseline Post-Test OccasionMean EXSE ScoreF Figure 4-1 Continued.

PAGE 120

120 A 23 25 27 29 31 33 35 37 Baseline Post-Test OccasionState ScoreControl Interventio 22 24 26 28 30 32 34 36 38 Baseline Post-Test OccasionTrait ScoreControl InterventionB Figure 4-2. Mean anxiety and depression scores by study group and occasion A) STAI-State Anxiety. B) ST AI-Trait Anxiety. C) GDS. D) BDI-2.

PAGE 121

121 0.5 1.5 2.5 3.5 4.5 5.5 6.5 7.5 Baseline Post-Test OccasionGDS ScoreControl Interventio 0 2 4 6 8 10 BaselinePost-Test OccasionBDI-2 ScoreControl InterventionD C Figure 4-2 Continued.

PAGE 122

122 A 107.00 109.00 111.00 113.00 115.00 117.00 119.00 BaselinePost-Test OccasionPredicted IQInterventio Control 23.00 25.00 27.00 29.00 31.00 33.00 35.00 37.00 39.00 BaselinePost-Test OccasionTime in SecondsControl InterventionB Figure 4-3. Mean cognitive scores by study group and occasion for non-execu tive cognitive variables A) NAART. B) Trails A. C) L M Immediate Recall. D) LM Delayed Recall. E) LM Recognition. F) LM Learning Slope.

PAGE 123

123 C 33.00 35.00 37.00 39.00 41.00 43.00 45.00 47.00 49.00 51.00 53.00 Baseline Post-Test OccasionNumber of PropositionsControl Intervention 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00 Baseline Post-Test OccasionNumber of PropositionsIntervention ControlD Figure 4-3 Continued.

PAGE 124

124 E 24.00 25.00 26.00 27.00 28.00 29.00 BaselinePost-Test OccasionNumber CorrectControl Intervention 2.50 3.50 4.50 5.50 6.50 Baseline Post-Test OccasionDifference Between Story B Trial 2 & Story B Trial 1Interventio ControlF Figure 4-3 Continued.

PAGE 125

125 48.00 58.00 68.00 78.00 88.00 98.00 108.00 Baseline Post-Test OccasionTime in SecondsInterventio Control A B Figure 4-4. Mean cognitive scores by study gr oup and occasion for executive cognitive vari ables A) COWA. B) Trails B. C) LetterNumber Sequencing. D) One-Back Mean RT. E) One-Back Mean RT SD. F) On e-Back Number Correct. G) Two-Back Mean RT. H) Two-Back Mean RT SD. I) Two-Back Number Correct. 32.00 34.00 36.00 38.00 40.00 42.00 44.00 46.00 48.00 50.00 Baseline Post-Test OccasionNumber of Word s Intervention Control

PAGE 126

126 C 6.00 8.00 10.00 12.00 14.00 Baseline Post-Test OccasionNumber CorrectInterventio Control 640.00 740.00 840.00 940.00 Baseline Post-Test OccasionTime in MillisecondsControl InterventioD Figure 4-4 Continued.

PAGE 127

127 E 4.00 104.00 204.00 304.00 404.00 504.00 Baseline Post-Test OccasionTime in MillisecondsControl Interventio 85.00 90.00 95.00 100.00 105.00 BaselinePost-Test OccasionNumber CorrectControl InterventioF Figure 4-4 Continued.

PAGE 128

128 G 1120.00 1320.00 1520.00 1720.00 1920.00 BaselinePost-Test OccasionTime in MillisecondsControl Interventio 375.00 575.00 775.00 975.00 1175.00 BaselinePost-Test OccasionTime in MillisecondsControl InterventioH Figure 4-4. Continued

PAGE 129

129 I 84.00 89.00 94.00 99.00 Baseline Post-Test OccasionNumber CorrectControl Intervention Figure 4-4. Continued

PAGE 130

130 A 104.00 106.00 108.00 110.00 112.00 114.00 116.00 118.00 120.00 122.00 BaselinePost-Test Age by Study GroupPredicted IQ Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older Adult 15.00 20.00 25.00 30.00 35.00 40.00 45.00BaselinePost-TestAge by Study GroupTime in Seconds Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older AdultB Figure 4-5. Mean cognitive scores by st udy group, age group, and occasion for non-ex ecutive cognitive variables A) NAART. B) Trails A. C) LM Immediate Recall. D) LM Delaye d Recall. E) LM Recognition. F) LM Learning Slope

PAGE 131

131 C 30.00 35.00 40.00 45.00 50.00 55.00 BaselinePost-Test Age by Study GroupNumber of Propositions Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older Adult 16.00 21.00 26.00 31.00 36.00 41.00 BaselinePost-Test Age by Study GroupNumber of Propositions Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older AdultD Figure 4-5 Continued.

PAGE 132

132 E 24.00 25.00 26.00 27.00 28.00 29.00 BaselinePost-Test Age by Study GroupNumber Correct Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older Adult 0.50 1.50 2.50 3.50 4.50 5.50 6.50 7.50 BaselinePost-Test Age by Study GroupDifference Between Story B Trial 2 & Story B Trial 1 Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older AdultF Figure 4-5 Continued.

PAGE 133

133 A 27.00 32.00 37.00 42.00 47.00 52.00BaselinePost-Test Age by Study GroupNumber of Words Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older Adult 25.00 45.00 65.00 85.00 105.00 125.00BaselinePost-Test Age by Study GroupTime in Seconds Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older AdultB Figure 4-6. Mean cognitive scores by study group, age group, and o ccasion for executive cognitive va riables A) COWA. B) Trails B. C) Letter-Number Sequencing. D) One-Back Mean RT. E) On e-Back Mean RT SD. F) On e-Back Number Correct. G) Two-Back Mean RT. H) Two-Back Mean RT SD. I) Two-Back Number Correct.

PAGE 134

134 C 7.50 8.50 9.50 10.50 11.50 12.50 BaselinePost-Test Age by Study GroupNumber Correct Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older Adult 570 670 770 870 970 1070 BaselinePost-Test Age by Study GroupTime in Milliseconds Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older AdultD Figure 4-6 Continued.

PAGE 135

135 E 10 110 210 310 410 510 BaselinePost-Test Age by Study GroupTime in Milliseconds Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older Adult 90.00 100.00BaselinePost-Test Age by Study GroupNumber Correct Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older AdultF Figure 4-6 Continued.

PAGE 136

136 G 740 940 1140 1340 1540 1740 1940 2140 2340 BaselinePost-Test Age by Study GroupTime in Milliseconds Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older Adult 105 305 505 705 905 1105 1305 1505BaselinePost-Test Age by Study GroupTime in Milliseconds Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older AdultH Figure 4-6 Continued.

PAGE 137

137 I 85.00 95.00 BaselinePost-Test Age by Study GroupNumber Correct Control Middle Aged Control Older Adult Intervention Middle Aged Intervention Older Adult Figure 4-6 Continued.

PAGE 138

138 CHAPTER 5 DISCUSSION The purpose of this study was to examine the ef fects of an exercise prom otion intervention on executive cognitive function in middle-aged and older adults. The present chapter will review the major study findings and their implications discuss the studys methodological limitations, discuss key conceptual issues that emerged from these results, and explore future directions regarding the improvement of cognitive func tion through physical exer cise intervention. The current investigation followed several major premises in the existing literature regarding the plasticity of cognitive function in la te life (e.g., Jobe et al ., Ball et al.) and the utility of physical exercise interventions in improving cognition and cognitive performance (e.g. Colcombe & Kramer, 2003; McAuley et al., 2004). Specifically, prior research has shown that cognitive function in older adults can be impr oved through both cognitive and physical training. In addition, the literatu re inconsistently suggests that executive control processes may be disproportionately improved when cardiovascular fitness levels increase (e.g. Colcombe & Kramer, 2003; McAuley et al., 2004). Where th is research has fallen short is in drawing connections between the multiple routes by whic h cognitive function may be improved in late life. Most of this work has concerned physica l/cardiovascular fitness improvements; however, a second route to affecting late life cognitive function is through improving complex activity/social participation and mood (e.g., Brown, 1992; Mc Auley, 1993). Specifically, the vascular depression hypothesis accounts for the common link between depression and cardiovascular health, as well as cognitive impairment (Alexopoulos, 2006). Review of Study Findings The present study sought to examine two routes a cardiovascular/physic al fitness pathway and a psychosocial pathway, by which an exerci se prom otion intervention could result in

PAGE 139

139 improvements in executive cognitive function. This study examined cognition, physical fitness and activity, and mood and affective functioning prio r to and after a 16-week exercise promotion intervention and was organized with re spect to the following specific aims: To investigate whether, relative to matched non-exercising control pa rticipants, sedentary adults receiving a physical ex ercise promotion intervention experience improvements in the primary outcome of cognitive function (p articularly executive control processes). To determine the separate and joint roles of improvement in proximal outcomes (fitness, activity, and affect) in mediating exerci se intervention eff ects on cognition. The 69 participants included in the present analyses had an averag e age of 63.9 years (SD = 8.7). The majority of the sample self-identified as Caucasian/White (91.3%) and female (84.1%). In addition, participants were on averag e, college educated (16.2 years of education) with high average estimated IQ (NAART Predicted IQ = 113). The sample reported a minimal level of depression and anxiety symptoms. Preliminary Analyses Baseline correlations among measures Prior to addressing the specifi c aims of the study, prelim inar y analyses were conducted to determine whether there was an intervention eff ect on the larger studys primary outcomes of physical activity and fitness. First, at baseline there were no significant differences between the exercise promotion and control groups on cogni tive, physical fitness/act ivity, and psychosocial variables. In addition, only ba seline self-reported leisure time exercise (LTEQ) was significantly correlated with 3 of the 15 cognitive variable s (two non-executive and one executive). This pattern of correlational analyses was unexpecte d, given the extant lit erature regarding the association between physical fitness/activity and cognitive performance. One explanation for this may be the significant skewness and kurtosi s of many of the cogniti ve variables included. Even with by-cell mean replacement of ou tliers, non-normality among these variables likely

PAGE 140

140 reduced the likelihood that significant corr elations between variables would be found. Further, though power analyses indicated the present sample size provided sufficient power for the larger study, it may be the case that it wa s not large enough to detect the small effects of the cognitive variables, which were conceptualized as secondary outcomes in th e larger study. Thus, a larger sample size would increase the power of the pres ent study to detect these small cognitive effects. Also of importance to note was the lack of substantial baseline association of age and education with cognition. Age showed significant positive correlations with only two (out of 15; 13.3%) baseline cognitiv e variables: One-Back Mean RT (r = 0.36, p = .002) and Two-Back Mean RT SD (r = 0.28, p = .022). Education was significantly correlated with only four (out of 15; 26.7%) cognitive baseline variab les: NAART Estimated IQ (r = 0.30, p = 0.013), LM Immediate Recall (r = 0.37, p = .002), LM Delayed Recall (r = 0.31, p = .009), and COWA (r = 0.24, p = 0.046). This lack of expected baseline association of age and education with cognitive variables was likely due to the smaller age range than the typical extreme group design that compares young and old. The middle-aged (50-64 years) and older ( 65 and older) age groups were likely much closer in cogni tive performance than would be needed to detect significant effects of normal cognitive aging. Further, as wi ll be explored greater in the limitations section of this chapter, this sample was highly-educated and advantaged, with re stricted variability in educational attainment. Again, a larger sample size may have improved the power to detect significant baseline relationshi ps between these variables. Pre-post changes in fitness and activity Preliminary pre-post analyses of the effect s of the studys intervention on the parent studys primary outcom es revealed that there were no significant interven tion effects on physical fitness or physical activity outcomes. Both th e intervention and the control groups demonstrated significant improvement in leisure time exercise cardiorespiratory fitness, and minutes of

PAGE 141

141 moderate and vigorous physic al activity, suggesting the sample improved over time; nevertheless, there were no group-by-time inte ractions to suggest the intervention group improved significantly more than controls. Cohens d effect sizes for the significant effects of time were small to medium (LTEQ, d = 0.51; MVPA, d = 0.51; mean pedometer steps, d = 0.22; VO2, d = 0.34). In this study, there was an averag e 7% increase in VO2, and a 67% increase in self-reported metabolic equivalents of activity (LTEQ). As a meta-analysis (reviewed below) suggests, these VO2 improvements did not reach the threshold of clinical significance. The more improved self-reported measure has not been studied with regard to cognitive outcomes, thus the clinical significance of the improvement on this measure is harder to judge. Previous studies have shown that improvement in aerobic capacity by 11% was unrelated to cognitive improvement above a practice eff ect (Madden et al., 1989) ; nevertheless, a 27% increase in aerobic capacity was associated w ith significant cognitive improvement in a prior study (Dustman et al., 1984). A more recent me ta-analysis studying the aerobic capacity/fitnesscognition link reviewed eleven randomized, cont rolled trials and found that aerobic fitness interventions resulted in an a pproximate 14% increase in cardiores piratory fitness (Angevaren et al., 2008). This 14% increase was associated with significant improvement in cognitive function: The largest effect s on cognition were found for mo tor function (1.17), auditory attention (0.52), and delayed memory functions ( 0.50). When compared to effect sizes of the current cognitive variables, these published effects are generally much larger than those here. Pre-post changes in self-efficacy, anxiety, and depression With regard to psychosocial outcomes, for one s elf-efficacy measure (EXSE) there was a between-subjects effect, suggesting the inte rvention group had significantly higher mean exercise self-efficacy scores at both baseline an d post-test; nevertheless, a within-subjects effect for occasion on this measure indicated there was re duced self-efficacy over time for both groups.

PAGE 142

142 As discussed later in the study limitations of this chapter, it may be the case that participants became more realistic regarding their personal goals for physical activity change and felt less confident in their ability to make such changes. Cohens d effect sizes were small for these significant occasion effects (Cohen, 1992; BSE, d = -0.12; EXSE, d = -0.29). There were no group-by-occasion interactions. Examination of changes in a nxiety and depression symptoms indicated that there was only a multivariate eff ect of occasion and univariate within-subjects effects for the STAI-Trait Anxiety and BDI-II Depression measures. Overall, the sample improved in trait anxiety and depression sympto ms over time; however these were small to medium effects (Cohen, 1992; GDS, d = 0.01; BDI-II, d = 0.58; STAI State Anxiety, d = 0.13; and STAI-Trait Anxiety, d = 0.41). Given the results of these preliminary analys es, it is important to note that the core assumption of this studys conceptual model was that improvements in physical activity, fitness, and self-efficacy and affect would mediate or mo derate any changes in cognitive, particularly executive, function. Thus, the li kelihood that the presen t study would observe an effect of the intervention on cognitive improvements over time wa s greatly reduced. Desp ite this, the planned analyses were completed to assess for any direct associations between the exercise promotion intervention and cognitive variables. Primary Analyses Aim One The goal of the first specific aim of t he study was to examine the effect of the exercise promotion intervention on cognitive performance. It was hypothesized that participants receiving the exercise promotion intervention would show improved performance on cognitive measures, particularly those assessing executi ve control processes, relative to control participants. However, contrary to this hypothe sis, the present findings indicated no significant

PAGE 143

143 multivariate effect of the in tervention on cognitive performa nce for either executive or nonexecutive measures. There was a multivariate eff ect of occasion, suggesting there was change in overall performance on non-executive and executiv e cognitive measures across time. However, there was no evidence of the critical group-by-occas ion interactions at th e multivariate level. This indicated that the intervention was not effective in improving cognitive outcomes. Reasons for the failure to find intervention effects will be discussed in the limitations section that follows. Despite the lack of multivariat e between-subjects or interaction effects, exploratory followup univariate analyses were conducted to determine the directions of significant multivariate within-subjects effects observ ed for both non-executive and execu tive measures. Univariate results revealed no between-subjects effects; however, there was a modest group-by-occasion interaction for the COWA test indicating intervention group participants improved significantly more than controls over time. This inte raction represented a sm all effect (Partial 2= 0.06) and was no longer significant w ith Bonferroni correction. Additionally, univariate within-s ubjects effects were found fo r several executive and nonexecutive measures: NAART, Trails A Time, and LM Immediate and Delayed recall (nonexecutive measures), and COWA, One-Back Me an RT, Two-Back Mean RT, and Two-Back Mean RT SD (executive measures). The findi ngs suggested improvements on these measures for the entire sample across time, which was likely a practice effect. With Bonferroni correction, these occasion effects only reached significance for the NAART and LM Immediate and Delayed recall. There was little evidence that above and beyond this ex pected practice effect, the intervention group improved to a greater de gree than controls. As mentioned above, examination of effect sizes for each cognitive cha nge score revealed small to medium effects of time (Cohens d = 0.049 to 0.668; Cohen, 1992). The largest effects were for the LM Immediate

PAGE 144

144 and Delayed recall trials (0.605 and 0.668, respectively), while the smallest were for LetterNumber Sequencing and Trails B (0.049 and 0.087, respectively). As a follow-up to these analyses, reliable chan ge index scores (adj usted for the practice effect demonstrated by the control group) were ca lculated to examine intraindividual trajectories of change. The goal was to determine the propo rtion of the sample that experienced sizable gains above the typical effect of practice. Results of these analys es indicated little evidence for reliable improvement for either the control a nd intervention groups. Th e LM Immediate Recall variable showed the largest pe rcentage of the sample experiencing improvements across time (24.6% of the sample); however, relative to controls, there were fewer intervention participants who reliably improved (28.6% of controls improved versus 20.6% of interv ention participants), although this difference (nor any other) was significantly differen t between study groups. This suggested little evidence for re liable cognitive improvement, not due to change alone, for our sample. Next, as additional follow-up analyses, the repeated-measures MANOVAs described above were re-run with age group as a second betw een-subjects factor. This permitted analysis of whether age moderated intervention effects (did one age group improve more than the other?). Further, it allowed exploration of potential two-way (study group-by-age group) and three-way (study group-by-age group-by-occasion) interactions At the multivariate level, there were no significant effects of age group for non-executive measures, though this effect did approach significance. There were also no two-way or three-way interactions involving age group. Univariate analyses demonstrated an age group effect for Trails A Time (the younger age group was significantly faster on this task), as well as study group-by-age group interaction effect for LM Delayed Recall (older individuals in the intervention group perfor med significantly better

PAGE 145

145 than younger individuals, regardle ss of study group, and older contro ls). This finding for LM Delayed Recall was contrary to what would be expected given the consistent findings in the literature that older adults tend to perform more poorly on measures of episodic memory. It may be the case that there was a failure of randomi zation. The older intervention participants were likely a selective group of individuals that did not represent the general population. Next, for executive cognitive measures, there was a significant multivariate effect of age group, but no multivariate two-way or three-way interactions. Univariate analyses revealed significant age group effects for One-Back Mean RT, Two-Back Mean RT, and Two-Back Mean RT SD. These findings suggested the younger gro up (middle-aged) had significantly faster mean reaction time and were significantly less inconsistent in mean r eaction time. Further, an age group-by-study group interaction for the One-Back Mean RT SD variable indicated younger control group participants were significantly less inconsistent in mean reaction time than younger intervention group and older participants. This latter finding suggested that contrary to our hypothesis, the intervention was not effectiv e in improving cognitive function on executive measures. An age group-by-occasion interaction effect revealed younger participants improved significantly more in their mean reaction time on the One-Back Mean RT variable than their older counterparts. Furthermore, there was a three-way interaction for Trails B Time, which demonstrated that younger intervention group par ticipants improved signif icantly more on this task than younger controls and a ll older participants. This thr ee-way interaction provides modest support for the effect of the intervention on executive cognitive function. While, with Bonferroni correction, this eff ect no longer reached the criteri on for significance, this finding offered trend support for the studys hypothesis of differential effects of exercise promotion on cognitive function.

PAGE 146

146 Interestingly, only executive measures were se nsitive to the inclusi on of age as a second between-subjects variable in multivariate models These findings indicated that there was a stronger effect of age than for the intervention in the sample. While this is, in part, contrary to the study hypotheses, these age effects were consis tent with the broader cognitive aging literature indicating that older adults show age related declines in processing speed and fluid cognitive abilities. In addition, these findings suppor ted the frontal agin g hypothesis (West 1996; Greenwood, 2000; Zimmerman et al., 2006), which suggests that frontal-executive abilities undergo disproportionate declines with aging as compared to non-fr ontal, more posteriormediated brain abilities (West, 1996; Greenwood, 2000; Zimmerman et al., 2006). Aim Two For the second specific aim of the study, it was proposed that physical fitness/activ ity, selfefficacy, and affect variables would be used as covariates in fo llow-up ANCOVAs; however, given the lack of significant in tervention e ffects or group-by-occasion interactions, these analyses were not appropriate. Thus, instead, ch ange scores were computed for each cognitive, physical fitness/activ ity, and psychosocial variable and correl ated in order to determine whether changes in the cognitive variables were associated with change s in the proposed covariates. Overall, there was modest support for correla ted change. Improved physical fitness was associated with improved LM Immediate Reca ll performance and reduced depression on the GDS, while improved leisure time exercise (LTEQ) was associated with less inconsistency on the One-Back Mean RT SD. However, contra ry to expectations, improvement in mean pedometer steps was correlated with reduced LM Learning Slope across testing occasions. Furthermore, reduced GDS depression was associ ated with improved LM Delayed Recall, and improved self-efficacy on the BSE was correlated with improved mean performance on the TwoBack Number Correct variable. Contrary to wh at would be expected, reduced BDI-II depression

PAGE 147

147 was related to increased reaction time inconsiste ncy across occasions on the Two-Back Mean RT SD. It is important to note that with Bonfe rroni correction, these corr elations between change scores were no longer significant. Study Limitations This study has ma ny limitations that ar e important to briefly consider. Sampling issues First, the present study utilized a relatively small, convenience sam ple of predominately Caucasian/White, female, healthy, highly educat ed adults who were recruited from the Gainesville/Alachua County Florida area. While the current sample size proved adequate power to detect certain within-subjects effects, a larger sample size would have increased the studys power to detect between-subjects an d/or interaction effects that were trends in the present data and that apparently obtained effect sizes lower than those expected a priori. Diversity of the sample was an additional concern. Recruitment strategies (i.e., use of an older adult registry, newspaper advertisement, co mmunity flyers, and presen tations at health and research fairs, etc.) were not specifically focused to perm it oversampling of underrepresented racial/ethnic minorities, who trad itionally suffer from health di sparities and presumably would benefit from inclusion in a h ealth/exercise promotion, lifestyle program. The present studys timeline and resources did not allow for recruitm ent strategies more appropriate for collecting racial/ethnic minority samples, such as identifying and collabor ating with community gatekeepers (i.e., church pastors, community orga nizers), to be properly instituted. Furthermore, when attempts were made to form community pa rtnerships with African -American churches and other community liaisons, competing research st udies (with similar health promotion messages) that were in the field at the same time challenge d the successful recruitment of participants.

PAGE 148

148 Thus, as will be discussed in the future directi ons section of this chap ter, modification of the recruitment strategies used in this study would be necessary to recruit a more diverse sample. Exclusion and inclusion criteria may further ha ve constrained sample heterogeneity. While it was important for the present study to have various cognitive, physical, health/medical exclusionary criteria to prevent various conf ounding variables from being introduced into the study and to ensure the safety of participants under going cardiorespiratory fitness assessment, the implementation of such exclusions may have excluded racial/ethnic minorities, particularly African-Americans, who disproportionately suff er from cardiovascular and other health conditions. The sample also suffered from a low number of male participants. While men are generally less represented in cognitive aging rese arch due to higher mortality rates, the current studys findings regarding physical exercise inte rvention-related cognitive improvements are less generalizable to non-female populations. In genera l, this studys strict exclusionary criteria likely reduced normal variation among participants and decreased the studys power to detect major intervention effects on physical and cognitive outcomes. Future studies might consider fewer exclusionary criteria. In particular, u tilizing alternative measur es of physical fitness (versus VO2) would change the need for exclus ion of individuals ta king beta-blocker and calcium-channel blocker medications, as was done in the present study. The use of these medications is very common among this study po pulation, but the current sample was a highly select group of people who were in good cardiov ascular health and not in need of such medications. One potential limitation concerns the inclusi on of middle-aged adu lts (along with older adults) in the present sample, wh ich may have introduced more noise in the present analyses due to the increased between person heterogeneity. However, this is unlikely given that the two

PAGE 149

149 age groups had similar mean years of educat ion and percent gender, and when follow up exploratory analyses controlled for age group assignment, the overall pattern of results remained unchanged. In fact, the inclusion of the age group variable revealed significant relationships that would have otherwise gone unobserved. Specifica lly, the three-way inter action showing greater improvements in performance over time on the Trails B task for younger intervention group participants was an interesting finding and one of few significant study findings that was in the expected direction. This finding was considered a small effect (Partial 2= 0.08) and was no longer significant with Bonferroni correction; ho wever, this trend is consistent with study hypotheses. Retention issues A second study limitation involves retention of the sample. The larger, 16-week intervention study, in which the present study is embedded, required a considerable time commitment from participants, which made recruitment and retention of participants challenging. Of the 433 participants identifi ed through various recruitment methods, 90 were randomized to a study group and pre-tested. This amounted to a 20.9% recruitment yield. While only twelve participants dropped out of the study during the intervention peri od, there were nine participants who underwent base line neuropsychological testi ng and did not attend any study group sessions. Had the study re tained these 21 participants that were randomized and pretested, but not post-tested, this would have e quated to roughly a 30% increase in the present samples numbers. Other retention issues invo lve the collection of complete data for each participant. As noted in the Methods chapter, there was a subs tantial amount of missing physical activity and self-efficacy data towards the end of the study, due to participant non-adherence to study protocols (i.e., completing daily and weekly questionnaires on physi cal activity and self

PAGE 150

150 efficacy). This may have been in part due to the tapering of group sessions towards the end of the 16-week intervention period. Interestingly, the current study drop-out rate was higher than that of various previous physical activity/exercise training studies (reporting attrition rate s), which had retention rates ranging from 83 to 92 percent (Madden et al., 1 989; Blumenthal et al. 1991; Hawkins et al., 1992; Emery et al., 1998; Elavskey et al., 2005 ; Motl et al., 2005; Elavsky & McAuley, 2007). In these studies, participants were either highly motivated to participate (Madden et al., 1989; Blumenthal et al. 1991) or the in terventions employed a structured exercise training protocol that followed American College of Sports Medici ne (ACSM) guidelines for physical activity intensity and duration minimums to improve se lf-efficacy, self-esteem, and quality of life. Employing more structure in the exercise prom otion lifestyle intervention may have helped participants to feel more conf ident in the physical activity ch anges they were making. Using ACSM guidelines also may have ensured sufficient exercise behavior cha nges to effect physical fitness at the magnitude needed to impact cogni tive function significantly. Furthermore, as mentioned previously, participants that dropped out of the present study differed significantly form those completing the study only in self-reported depression and anxiety symptoms. Perhaps, initially intervening upon these symptoms would have resulted in a higher rate of retention. Finally, having more financial resource s to recruit and retain participants may have resulted in a higher final sample size. Katu la and colleagues (2007) used numerous methods (brochures, newspaper, radio, tele vision, etc.) to recruit participants during the course of a foursite randomized, controlled pilot study of the benefits of a phys ical activity intervention for immobility prevention in older adults. In the end, the study successfully randomized 424 older adults (a 13.5% recruitment yi eld) to two intervention conditions after spending approximately

PAGE 151

151 $439 in direct recruiting costs pe r randomized participant. Wh ile the numbers involved in Katula et al.s study were beyond the scope of the present study, their findi ngs give some insight into the feasibility of recrui ting geographically and ethnically diverse populations of older community-dwelling adults for a physical activity intervention trial. Measurement issues A third limitation involves the selection ne uropsychological battery of tests employed. Since the current study was designed to be embedded within a larger research protocol, it was necessary to limit the breath of neuropsychological assessment to be undertaken at baseline and post-intervention. Tradi tional neuropsychological evaluation is comprehensive in nature, and it may be the case that with a greater breadth of cognitive measures, intervention effects on cognitive performance would have been better de tected. Nonetheless, th e present selection of cognitive measures was guided by prior research that has distinguished the effects of improved cardiovascular fitness on executive cognitive function; particularly executive control processes (Colcombe & Kramer, 2003). While the present i nvestigation relied primarily on clinical and experimental measures of one aspect of ex ecutive cognitive function, working memory, it may be the case that inclusion of other measures of working memory and/or other aspects of executive function (such as response inhibitio n) would have increa sed the likelihood of significant cognitive effects. Noteworthy is the fa ct that there was a near-ceiling effect on the Nback working memory task accuracy. This was one example of the need for more extensive neuropsychological assessme nt in the present study. The present study used a modified treadmill protocol that estimated VO2 after 85% of maximal heart-rate was reached, which, while sa fe for an older sample, was a less sensitive estimation of cardiorespiratory fitness. It may be the case that the use of a VO2 max stress test instead would have resulted in stronger relationships between cardiorespiratory fitness and

PAGE 152

152 cognition; nevertheless, VO2 ma x may not be a direct indica tor of cerebral profusion. Neuroimaging should be employed in the future to better assess th e direct effect of cardiorespiratory fitness on the brain. As mentioned previously, th e timing of the post-test in terval was variable across participants. Most participants were tested at varying intervals following the intervention period. While efforts were made to ensure each participant was post-tested within two weeks of study completion, scheduling conflicts precluded this from being accomplished for some participants (3 out of 69; range 3 to 6.4 week s post-intervention). This variability in post-test intervals may have introduced error into our analyses by not capt uring each participant at the same point in the study timeline. Since it might be expected that the strongest physical ac tivity/fitness intervention effects would be found just afte r the completion of th e intervention program, all participants ideally would have been tested at the same point in th eir individual timelines. Intervention issues Fourth, the exercise promotion interventi on was a novel approach to exercise intervention that combined various theoretical a pproaches. The larger study utilized a 16-week intervention period (with 13 weeks of sm all peer group sessions), which may not have been extensive enough impact exercise behaviors, and in turn, fitness and cognitive improvements. Additionally, the program was a lifestyle intervention that was not designed to prescribe specific time-periods or intensity of exerci se behavior, with the rationale that this type of intervention would result in increased main tenance of lifestyle changes and more favorable long-term outcomes. The lack of specific behavioral target s for intervention participants, while consistent with the interventions theoreti cal underpinnings, likely resulted in this group failing to meet minimum physical activity levels that would e ffect physical fitness improvements sufficient enough to mediate cognitive change (Buman, 2008). It may be the case that without the

PAGE 153

153 concrete, structured program of other exercise studies (Madden et al., 198 9; Blumenthal et al. 1991; Hawkins et al., 1992; Emery et al., 1998; Elavskey et al., 2005; Motl et al., 2005; Elavsky & McAuley, 2007), participants simply did not ex ercise enough or at a level of intensity for physical fitness and cognitive improvements to be detected. Effect sizes for these physical fitness/activity variables were small for bot h between-subjects effects and group by occasion interaction effects (Partial Eta2 ranging from 0.0 to 0.1), and this suggests that there was little association between group membersh ip and fitness/activity levels. Another consideration is the tapering of group sessions in the final weeks of the intervention period. While this decision was made to slowly wean participants from the in fluence of social contact and peer-mentorship afforded by study participation, the strongest phy sical fitness/activity effects may have been found at week 12, which was last week of real group instruction. Also, the decline (instead of improvement) in self-efficacy over time may suggest that participants became more realistic regarding the difficulties associated with implementing behavioral changes in physical activity and thus, felt less confident in thei r ability to make such changes after gaining a better unde rstanding the actual work invol ved. This sample may have been so highly-advantaged and accustomed to being very goaland self-directed and highly motivated that they initially unde restimated the difficulty of ma king physical activity changes at this stage in their lives. Th eir decreased self-efficacy may be driven by some comparison between their current ability to make significan t activity changes and the ease with which they may have made them in the past. Furthermore, the use of multiple peer coaches may have affected treatment delivery. While quality c ontrol procedures were implemented, individual differences between the peer coac hes (i.e., interactions with pa rticipants) likely differentially impacted the success of the small peer groups.

PAGE 154

154 Finally, one limitation inherent in the larg er studys design was permitting the health hygiene control group access to an exercise facili ty in exchange for study participation. This decision was made to avoid the potential conf ound of the anticipated intervention effects and access to a fitness facility and also mounted a ve ry stringent test of the added value of this motivational intervention. Nonetheless, this decision precluded complete study manipulation of exercise/physical activity, as there was no way to li mit the amount or intensity of exercise in the control group. Also, the social contact and use of a peer-mentorship model may have increased motivation in the control group to make changes in physical activity. It may be the case that the combination of these factors made for an uncha racteristic control group, instead of what would be expected from a more traditio nal, no-contact control condition. Conceptual Issues A num ber of broader conceptual issues were introduced in the Intr oduction Chapter, and several others emerged as a result of the analys es. This section consid ers several of the key issues that emerged. Selection of the correct intervention A first issue to consider is whether a 16-w eek intervention period, with only 12 weeks of actual instruction, is sufficient enough to reasonably expect adequate physiological change. Unfo rtunately, inconsistencies in the existing l iterature provide less concrete guidance in the types and lengths of exercise in tervention protocols to ensure c onsistent effects on both physical and cognitive outcomes (Kramer et al., 2006). Revi ew of various exercise intervention trials, with cognitive outcomes, shows th at the length of training progra ms may range from as few as 10 weeks (Hawkins et al., 1992; Emery et al., 1998) to as many as 14 months (Blumenthal et al., 1991) and provide mixed evidence for the relationship between physical fitness and cognition. Methodological reasons for such inconsistent findings that have been discussed include

PAGE 155

155 differences in the nature, intensity, and durati on of aerobic fitness inte rventions, as well as, differences in sample characteristics, measuremen t of fitness, and the nature of control groups (Colcombe & Kramer, 2003). There have been se veral studies of self-efficacy and affective function (without inclusion of cognitive outcomes) that have employed a 6 month exercise training program (Elavskey et al., 2005; Motl et al., 2005; Elavsky & McAuley, 2007) and found positive results. Additionally, a more recent randomized, controlled fitness intervention (walking) study found that after 6 months, interv ention group participants performed better on a focused attention task and had increased frontal and parietal brain activation than participants in a stretching/toning control group co ndition (Colcombe et al., 2004). What is important to note is that for these studies, there was some level of structure regarding the types, duration, fre quency, and often, intensity that wa s expected of participants. While for theoretical reasons, participants in th e present study were not given specific duration or intensity prescriptions or a guideline for minimum exercise goals, it may be the case that sedentary older adults require gr eater level of instruction that includes specific procedural and declarative knowledge when it comes to implemen ting physical activity changes. Drawing from the literature on cognitive strategy training and use among older adults, studies suggest that older adults suffer from a metacognitive deficit, such that they do not spontaneously use cognitive compensation strategies (e.g., Dixon & Hultsch, 1983), even following instruction in such strategies. The present study re lied more heavily on cognitive-beha vioral strategies and likely held to traditional motivational interviewing tech niques too closely (where in goals are defined by the individual). A modification might well have b een to establish an objective, normative target for intervention intensity, duration and freque ncy (e.g., ACSM guidelines), and then use the motivational interviewing strategies to help move participants from their baseline state to this

PAGE 156

156 final state. In other words, helping to more actively shape the goals of participants seems a likely ingredient for higher succe ss. Consequently, an interventi on that is more behavioral in nature, with individualized target s more clearly defined a priori, may be more suitable for older groups. Selection of the correct follow-up interval Next, as me ntioned above, the present exer cise promotion was designed with long-term behavioral lifestyle outcomes in mind. As suc h, it may be the case that our lack of significant findings at post-test may not tell the whole story. Longer-term (e.g., 6-, 12-, or 18-month) longitudinal follow up may be necessary to detect a delayed effect of th e intervention. A true test of the effectiveness of the intervention may be its effect at long-term follow up, rather than at an immediate post-test interval. One hypothesis to consider is that interventi on effects might cumulate. Specifically, the present study tried to introduce he alth habits and personal goal se tting that would lead to an altered fitness lifestyle. If successful, the in tervention would produce sm all incremental gains that would continue long after the study is completed. (This is admittedly idealistic; most follow-ups of exercise studies find that effects dissipate after cessation of treatment). Another hypothesis is that in tervention effects are delayed because they are revealed downstream, with a separation of the decline trajectories of those with and without the intervention. There is a recent ex ample of such a delayed effect, albeit in the cognitive domain. In the ACTIVE clinical trial of cognitive interventions for older adults (Willis et al., 2006), ACTIVE study investigators did not find a transfer effect of cognitive (i.e., memory, reasoning, and speed) training on daily functi on (i.e., IADLs, self-ratings, ev eryday problem solving) until a 5-year follow-up assessment. Two reasons cited for this delayed intervention effect were previous work suggesting a lag between the onset of cognitive decline and the onset of functional

PAGE 157

157 declines, as well as the advantaged nature of the ACTIVE sample. Specifically, since cognitively and functionally impa ired individuals were initially excluded from the study, only after normative cognitive and func tional declines occurred in the control group could the protective effect of cognitive tr aining be detected (Willis et al., 2006). Thus, in the present sample, in addition to any cumulative and contin uous gain that the inte rvention might promote, the protective effect of partic ipation in a lifestyle physical activity intervention may not be clearly manifested until the onset of normative ag ing-related decline, at which point, different rates of physical declin e between the study groups may be observable. On the relative importance of physical vs. mental exercise This study was designed to evaluate the sec ondary effect of an exercise promotion intervention on cognitive function. In contrast to many other studies with longer and more structured interventions, the study fa iled to find an exercise effect. Despite this, the majority of participants (regardless of group assignment) experienced cognitive improvements on most measures from pre-test to post-test. Interestingly, despite the l ack of an intervention effect as hypothesized, the most consistent finding in th e present study was the si gnificant effect of occasion, with improvements noted across tim e on cognitive, physical, and psychosocial measures. It appears these within-subjects effect s constituted the classic practice effect that has been widely shown, due to retest, across the li fespan. In fact, RCI scores (adjusted for the average change in cognitive scores for the cont rol group) indicated few cognitive gains, above and beyond the average practice effect, for the ma jority of both control and intervention group participants. This is not surprising, given the l ack of significant differe nce between the groups. Of course, using the control group as the baseline for defining practice may be problematic if our health-hygiene + gym membership control group constituted a real intervention. In future work with this data set, it may be important to use published test-retest and pub lished practice effects

PAGE 158

158 for our dependent variables to better gauge, for each group, whether the improvements observed exceeded what one might normally expect due to retest alone. However, it is not clear whether these o ccasion effects represent pure practice or something more. Since both groups experienced fitness and activity ga ins, there is some possibility that these gains might have been ac tivity-related for both groups. Given the weak association between improvements in physical f itness, declines in self-efficacy, and improved cognitive performance, it may be the case that, un like in the conceptual model presented (Figure 2-1), the physical fitness and psychosocial path ways were not the important routes for cognitive improvements in this study, as hypothesized. However, another inadvertent interven tion was occurring throughout the study. Coinvestigators of the larger study included weekly assessments of cognition (not the measures in this study, but Letter Series, Number Copy, Reac tion Time, and Symbol Digit) in the larger studys protocol. When consid ering pre-test and post-test assessments, this means most participants had up to 18 weekly practice sessions with cognitive measures. Thus, this repeated assessment constituted a kind of intensive practice-based cognitive intervention and likely contributed to some the time effect observe d on cognitive measures (Because of the known specificity of practice effects, and the lack of overlap between weekly and pre-post cognitive measures, the magnitude of transfer from the weekly practice may have been fairly small). Furthermore, there was no age group difference in this practice effect, such that both age groups benefited equally from practice. These findings raise questions about the best approach for obtaining cognitive improvements in future work. It appears that relatively simple cogni tive practice was a far simpler route to cognitive improvement than phy sical exercise promotion. Given this, one

PAGE 159

159 wonders if the best first-line approach for improving cognition in the future might not be direct cognitive intervention. Of course there are many questions to be answered. For example, the literature suggests that cognitive training effects ma y be highly specific to the domains studied (e.g., Willis et al., 2006), whereas physical training effects may be more broadly global (e.g., Kramer et al., 2006). Thus, one possibility is th at exercise improves general brain health, and with it, promotes broad, low-magnitude cognitive gain. However, to improve specific functions to a higher magnitude, those specific functi ons need to be practiced and trained. Dealing with mood and affect This study experienced almost 30% attrition. In exam ining the predictors of attrition, the variables that emerged as most uniquely pred ictive were pre-test depression symptoms. Although the exercise promotion intervention in th is study was self-efficacy focused, it did not specifically target mood issues. One wonders if better sample retention, and larger intervention effects, might have been obtained if a mood in tervention component had been included in this study. One could imagine an altered multi-step intervention model in future work. If individuals with higher levels of depression and anxiety symptoms had poor coping skills, and found it especially challenging to engage in a lifestyle physical activity program (the extreme form of which was dropout), then perhaps these depre ssion and anxiety symptoms should have been addressed first. It is important to note that participants were not, for the most part, at clinical levels of depression and anxiety. Rather, the argument is that subsyndromal depression and anxiety may interfere with full participation in the intervention. If this is true, then future research might first employ cognitive-behavioral approaches (either in the full group, or individually in one-on-one sessions with persons experiencing elevated levels ) for dealing with mood issues. The idea is

PAGE 160

160 that mood interventions become the pre-interve ntion. One rationale of such affective preintervention would be to prepare the individual to subsequently maximize his or her ability to receive and benefit from subs equent physical and cognitive lifestyle intervention. Future Directions Longitudinal follow-up There are several di rections that the present line of research ma y take in the future. First, longitudinal follow up of current sample is planne d in the larger study to determine whether any lifestyle changes in physi cal activity and exercise are maintain ed over time. Despite the lack of significant between-group findings for physical fitn ess and activity, there were within-subjects effects for VO2, leisure time exercise, and minut es of moderate and vigorous physical activity. Thus, it would be informative to assess longit udinally whether any of these effects remain, and/or whether there is group separation in main tenance or subsequent improvements. From a cognitive standpoint, longitudinal assessment also would be interesting to determine whether any long-term gains in cognitive performance. Furt hermore, the larger study as a broader goal of using the present physical, psychosocial, and cognitiv e data as pilot data for a future, large-scale study examining these relationships. It will be im portant for future study in this area to improve upon the present studys methodological limitations in order to increase the likelihood that an intervention effect would be detected. Enhanced sampling A second future direction that is of part icular significance is improving the cultural diversity of physical exercise promotion studies in the future. Not only would im proving cultural diversity increase the heterogeneity of future samples, such improvements would also allow for more interesting questions regard ing the generalizablity of physical exercise interventions and study findings to various populations of middle-aged and older adults. This is

PAGE 161

161 of importance from a public heath perspect ive, since cardiovascular disease, obesity, hypertension, and other related chronic health conditions are among the leading causes of morbidity and/or mortality in the US (Mokdad et al., 2004) and are related to poorer performance on neuropsychological and cognitive measures (e.g., Waldstein, 1995; Breteler, van Swieten, Bots, & Grobbee, 1994). Racial/ethnic minority gr oups, particularly African Americans, tend to suffer from vast disparities in h ealth status relative to the majo rity population, and in turn, have higher rates of these cardiovascular conditions and poorer cognitive performance (National Center for Health Statistics (NCHS), 1990). T hus, the implementation of a physical exercise promotion intervention in culturally diverse samples could begin to remediate prevalent cardiovascular health conditions in underserved groups. In addition, examining physical fitnessrelated improvements in cognitive performance in a diverse sample could further augment the literature that has attempted to both characteri ze and explain racial/et hnic group differences in late life cognitive function (e.g., Aiken-Morgan et al. 2008; Manly, Jacobs, Touradji, Small, & Stern, 2002). This work might focus on whethe r physical fitness improvements resulted in similar cognitive gains for multiple racial/ethnic groups. To accomplish this goal of examining a cultu rally diverse sample in future work, recruitment methods must be modified. Specifica lly, partnerships with community gatekeepers, (such as church pastors, heads of community or ganizations, etc.) would need to be formed and maintained in a manner that would be equally beneficial to both the researcher and the community being targeted. Retention also would need to be a key focus of future work, and study operations would likely need to take place in the community at a convenient location, or even the participants homes, to ensure continued partic ipation throughout the study timeline.

PAGE 162

162 Finally, collaboration with resear chers in the field with expert experience in the recruitment and retention of diverse samp les would be ideal. Further conceptualization of the target cognitive domains A third future direction, f urther conceptu alization of executive cognitive outcome measures, has theoretical implications for the con tinued examination of the differential effect of improved cardiovascular fitness on executive cogn itive function. The term executive function is used as an umbrella term to describe an individuals ability to engage successfully in independent, purposive, self-serving behavior and involve strategies necessary to approach, plan, or carry out cognitive tasks, or the strategies needed in the monitoring of performance (Lezak et al., 2004). Given such a broad definition of this construct, it is understandable that previous literature has shown little consensu s regarding appropriate measures of executive function or its underlying factor structure (Kemper & McDowd, 2008). Throughout the literature, there have been many measures of executive function use d, with both macro and micro approaches to assessment (Kemper & McDowd, 2008). This variation has contributed to differing factor structures observed for executive function based on the sample studied (i.e., young adults, normal elders, and Alzheimers patients). One example is how Royal et al. (2003) found one three-factor structure of executi ve function in older adults (abs traction, procedural control, and attention switching), while others have found a different three-factor structure (shifting, updating, and inhibition) in younger samples (Miyake et al, 2000; Friedman et al., 2004). In general, convergent validity and discriminant validity among executive measures have not been well supported (Luszcz & Lane, 2008). Salthouse (2005) has posited that executive function tests are essentially tests of fluid abil ity or general intelligence (g). Two specific problems with executive function tests that have been cited include a task impurity problem and the fact that performa nce on executive measures may reflect many

PAGE 163

163 executive processes (Luszcz & Lane, 2008). None theless, verbal fluency (COWA) has been identified by previous studies as a gold st andard measure of execu tive function, due to its sensitivity to frontal lobe function (Crawford et al., 2000 & 2005; Salthouse et al., 2003). The present studys findings that COWA was most sens itive to the intervention effect (albeit very modestly) are consistent with this previous work. This lack of consensus in conceptualization and assessment of executive cognitive function likely contributes to the inconsiste ncies in the extant literature regarding the relationship between physical/aerobic fitness and cogniti ve function. To date, the studi es that have concluded that executive cognitive function is disproportionately affected by improvements in cardiovascular/physical fitness have examined pr edominately one type of executive function: executive control processes or working memo ry (Colcombe & Kramer, 2003). However, abstract reasoning, conceptuali zation, and problem-solving are also considered executive functions that are sub-served by frontal lobe-subcortical connectio ns in the brain via white matter tracts, which are disproportionately affect ed by both normal brain aging (e.g., Sullivan & Pfefferbaum, 2006) and are particularly sensitiv e to cardiovascular and cerebrovascular health declines (e.g., van Boxtel et al., 2006). Nevertheless, little atte ntion in the literature has been given to these other types of executive cognitive functi ons, which tend to depend on conceptualization and abstracti on and less on ones ability to mentally hold and manipulate information or rapidly shift/alternate between competing mental sets, as do working memory tasks. This distinction is important because working memory tasks rely more on intact attentional abilities, and these tasks often have a speed component to them (such as the Trail Making Test and N-back task). Executive m easures of conceptualization and abstract reasoning/problem-solving abilities do not require the same level of attentional control and future

PAGE 164

164 research should examine these other types of executive functions to gain a more complete understanding of the effects of car diovascular/physical fitness. It may be the case that the fitness-cogniti on relationship is mediated primarily by the attentional system, such that this is the mechan ism for improved executive control processing. In fact, in Colcombe & Kramers (2003) meta-ana lysis, there was no c ontrol for the overlap between executive measures that could also be categorized as measuring a related cognitive function (i.e., speed, visuospatial ability, or sustained attentio n). Their executive category included the largest number of st udies; thus, this category may ha ve had the most variance, and that could be the explanation for why executive measures appeared to be most-improved by exercise interventions. Very concretely, almost every measur e in their meta-analysis was classified as executive and something else. Thus, Colcombe & Kramers conclusion that executive control processes show the most exercise-related improvements may require further examination. Future research should examine more than executive tasks that are hi ghly dependent on the attentional system in the quest to better understand the link between fitness and executive cognitive function. Including measures of highe r-order conceptual abil ities would provide a deeper level of analysis of a ll of the executive functions that are supported by frontal lobe and subcortical brain function. Future work should also pay attention to sens itivity and specificity of measures in the selection of cognitive batteries More sensitive measures mi ght be selected in a highlyadvantaged sample, such as the present one, to maximize the likelihood of detecting individual differences in baseline performance, which could then be more sensitive to a physical

PAGE 165

165 fitness/exercise effect. In other words, there mu st be sufficient variability in our instruments so that they can be sensitive to improvements. However, if a sample included a less healt hy sample of individuals (e.g., those with cardiovascular disease) specificity of measur es might be more important to consider. Instruments that are sensitive specifically to card iovascular impairment, and that therefore might respond to improvements in cardiova scular condition would be desire d. In order to identify such cardio-sensitive measures, preliminary measuremen t work might need to be conducted in an agehomogenous sample (to eliminate age-related variab ility) or in a sample more representative of the lifespan (so that age can be covaried out before identifying m easures that have high degrees of unique cardiovascular variance). Further conceptualization of fitness and activity measurement In addition to exploring other measures of executive cognitive function, future studies in this regard would benefit from the use of altern ative measures of cardiov ascular health/fitness. The present study used a VO2 estimate, which is a measure of cardiorespiratory fitness or the circulatory and respiratory systems ability to efficiently supply oxygen to the body during sustained physical activity. However, in order to include this measure, participants taking betablockers and/or calcium-channel blockers had to be excluded due to the negative health risk of pushing these individuals to 85% of their maxi mal heart rate. This exclusionary criterion excluded a larger number of potential participants. One approach in the future may be to include a physician and/or nurse (i.e., trained clinicians) on the study team to perform this test in a safe manner. Nevertheless, despite VO2 being considered a gold-standard in the fi eld (the present study used a modified treadmill protocol that estimated VO2 after 85% of maximal heart-rate was reached), future research might consider othe r approximations of physical fitness that would

PAGE 166

166 exclude fewer older participants from this t ype of study. Colcombe et al (2004) used the Rockport 1-mile walk test as a more appropriate measure of aerobic fitness for use with older adults and reported a high correlation between the Rockport test and treadmill VO2 (r = .88, p < .01) (Colcombe et al., 2004). Furthermore, meas ures of general lung function might approximate an individuals fitness, specifically their ab ility to efficiently use oxygen necessary for meaningful physical activity. In particular, m easures that might be considered are forced expiratory volume (FEV; a measure of amount of air that can be forcibly blown one seconds) and forced vital capacity (FVC; the total amount of air that can fo rcibly be blown out after full inspiration). These spirometry measures are co nsidered primary indicators of lung function, and they may be useful predictors of cognitive function in an aging sample. For instance, one literature has investigated the relationship between lung function and cognitive performance and found positive relationships across various cognitiv e measures (e.g., Allaire, Tamez, Whitfield, 2007). Less work has been done to determine whether executive cognitive function might show disproportionate associations with spirometry measures. Relative effect size of exercise and combinatorial interventions Finally, another future directi on of th e present research is implementing interventions in older populations that combine physical and c ognitive training protocols. The literature suggesting the benefits of both physical fitn ess intervention (Colco mbe & Kramer, 2003) and cognitive training (Jobe et al., 2001) on cognitive performance in olde r adults helped to build the premises for this study that cognition in late life is plastic and may be improved through intervention. Beyond the scope of this study wa s the implementation of other experimental conditions (i.e., cognitive and combined physical a nd cognitive) that could be manipulated to see if cognition, particularly executive function might be improved. While there is little previous work in this regard, a novel intervention approach would a llow for the comparison of the

PAGE 167

167 efficacy of various types of interventions, as well as exploration of the optimal combination of cognitive and physical training that would result in reliable cognitive outcomes for aging populations. Conclusion As the aging population in the US and worl d continues to rapidl y grow, the call for lifestyle in terventions to preven t the onset of both physical dis ease and cognitive impairment and dementia will also grow stronger. While the present study found weak effects of an exercise promotion program, as well as only modest support for association between physical fitness/activity and psychosocial improvement s and cognitive performance improvements over time, it is critical that future intervention wo rk in the field of cognitive aging build upon the present study. This work has si gnificant public health implications as the societal burden of caring for older adults with disabling medical di seases, including Alzhei mers disease and other dementias, is great and threatens to become ev en greater in the future without effective, preventive lifestyle intervention.

PAGE 168

168 APPENDIX A TELEPHONE SCREENING PROCEDURE ELIGIBLE: _________________________ SCHEDULED: ______________________ INELIGIBLE: _______________________ THE AAMP STUDY TELEPHONE SCREENING PROCEDURE This telephone screener includes (a) telephone consent, (b) exclusionary criteria, (c) basic demographic questions, (d) the Telephone Interview of Cognitive Status (TICS), and (e) the Exercise Staging Algorithm (ESA) Participant ID Number: _________________ Date of Screening: _________________ Name of AAMP Researcher conducting the Screening: _________________________________________

PAGE 169

169 TELEPHONE INTRODUCTORY SCRIPT: Hello, may I speak with (NAME)?--------> WAIT TO SPEAK WITH PARTICIPANT. My name is (TESTER) and I'm calling for the AAMP Study at the University of Florida. You recently indicated your interest in participating in this peer-mentored exercise promotion research study for older adults. Do you r ecall your previous interest in this study? IF PARTICIPANT RECALLS PRIOR CONTACT: Do you recall how you heard of the AAMP Study? ________________ Do you have any questions? IF PARTICIPANT DOES NOT RECALL PRIOR CONTACT: The purpose of this program is to promote health and exercise behaviors in older adults through the use of a peer-mentoring program. Do you have any questions? IF PARTICIPANT HAS QUESTIONS. ANSWER APPROPRIATELY AND STATE: I am calling today to see if you are eligible for this program. I would like to ask you a few questions which will take about 20 minutes. Is this a good time to talk with me? IF YES: BEGIN TELEPHONE SCREENER. IF NO: GET CALL BACK TIME AND STATE: Thank you. A member of our staff will call you then. Call back date and time: __________________________________________ [If you will be unable to make this call-back appointment yourself, please make arrangements with another researcher to do so.]

PAGE 170

170 Informed Consent to Participate in Research Telephone Screening 1. Name of Participant ("Study Subject") _____________________________________________________________________ 2. Title of Research Study Trial of a multi-component exercise pr omotion intervention for older adults 3. Principal Investigator and Telephone Number(s) Peter A. Giacobbi, JR., Ph.D. THE FOLLOWING CONSENT/CONFIDENTIAL ITY STATEMENT MUST BE READ TO ALL POTENTIAL PARTICIPANTS AND INITIALED. "Before we begin, I'd like to mention a few th ings. All of your responses are completely confidential, and will only be seen or heard by people directly associated with the study. No information about any specific individual will ever be reported. Your name will never appear in any report about this study. You may refuse to answer any questions at any time. Do you have any questions before we begin?" I HAVE READ THIS STATEMENT TO THE PARTICIPANT. Signature/Initials of Te lephone Interviewer: __________________________________________ Print Name: _____________________________________ Date: _______________________

PAGE 171

171 Page 1 of 1 First, I would like to confir m some information about you. 1. Can you please give me your full name, address, and phone number? Name: __________________________________________________________________ Address: ________________________________________________________________ ________________________________________________________________________ Phone: __________________________________________________________________ E-mail (if applicable):_____________________________________________________ ASK GENDER (2) ONLY IF NOT KNOWN OR UNABLE TO DETERMINE. OTHERWISE, CODE QUEST. 2 AND NO W ASK PREFERRED TITLE QUESTION. 2. Are you male or female? MALE........................... 1 [Is that Mr., Dr., Rev., or other?] RECORD PREFERRED TITLE ON CONTACT RECORD. FEMALE.......................2 [Is that Mrs., Miss, Ms., Dr., Rev., or other?] RECORD PREFERRED TITLE ON CONTACT RECORD. 3. What is your date of birth? .___/____/._______ IS PARTICIPANT'S AGE WI THIN 6 WEEKS OF 50th BIRTHDAY OR OLDER TODAY? YES ..............................................................1 NO ...............................................................2 INELIGIBLE: READ SCRIPT BELOW AND END INTERVIEW AGE INELIGIBILITY CLOSE-OUT SCRIPT: These are the only questions I need to ask. This research study is designed for people who are age 50 or older. I would like to thank you for th e time you have taken to speak with me. We will not need to contact you again for this study, but could we contact you in the future for other studies? Thank you. 3a. Is English your fi rst language? YES NO .2 4. What is your marital status? [READ RESPONSE CATEGORIES IF UNABLE TO ANSWER] MARRIED, .........................................................................1 LIVING AS MARRIED,.....................................................2 SEPARATED,.....................................................................3

PAGE 172

172 DIVORCED,........................................................................4 WIDOWED, ........................................................................5 SINGLE, OR NEVER MARRIED?....................................6 5. Does anyone live in the home with you? YES..........1 NO..........2 6. What is the highest grade of school or level of education that you completed? [CODE ONLY ONE RESPONSE] DID NOT GO TO SCHOOL 00 GRADE 1 01 GRADE 2 02 GRADE 3 03 GRADE 4 04 GRADE 5 05 GRADE 6 06 GRADE 7 07 GRADE 8 08 GRADE 9 09 GRADE 10 10 GRADE 11 11 GRADE 12/GED 12 VOCATIONAL/TRAINING/ SOME COLLEGE AFTER HS GRAD 13 ASSOCIATE DEGREE 14 COLLEGE GRAD/BA-BS 16 SOME PROFESSIONAL SCHOOL AFTER COLLEGE GRAD 17 MASTER'S DEGREE 18 DOCTORAL DEGREE (PhD, MD, DVM, DDS, JD, etc.) 20 7. What race do you consider yourself? (PRO BE: White, Black/Afri can American, Asian, Native Hawaiian/Pacific Islander, American I ndian/Alaskan Native, or another race?) WHITE/CAUCASIAN .........................................................1 BLACK/AFRICAN AMERICAN ........................................2 ASIAN...................................................................................3 NATIVE HAWAIIAN/PACIFIC ISLANDER......................4 AMERICAN INDIAN/ALASKAN NATIVE.......................5 BIRACIAL.............................................................................6 SPECIFY:_______________________________________ OTHER .................................................................................7

PAGE 173

173 SPECIFY: ______________________________________ DON'T KNOW......................................................................8 IF PARTICIPANT IS UNABLE TO ANSWER 7, PROBE: Which race do you most identify with or consider yourself to be? 8. Are you Hispanic or Latino? YES..........1 NO..........2 9. The next questions are about your vision. Do you wear glasses or contact lenses to read? YES..........1 NO..........2 BEGINNING WITH ITEM 10a, AND FOR ALL OTHERS DO NOT TERMINATE INTERVIEW IF INELIGIBLE COLLECT ALL DATA, TH EN READ INELIGIBILITY SCRIPT BEFORE ITEM 15. 10a. How much difficulty do you have reading ordinary print in the newspaper, [wearing glasses or contact le nses]? Would you say... no difficulty................................................................1 (11) a little or some difficulty............................................2 (11) extreme difficulty ......................................................3 = INELIGIBLE (11) you stopped reading because of your eyesight..........4 = INELIGIBLE (11) you stopped reading for other reasons or you are not interested in reading...................................................5 (11) 10b. How much difficulty do you have hearing c onversation partners, when in small groups and there is background noise? (m ultiple conversations, music, white noise) What about with your hearing aid(s) on? Would you say... no difficulty................................................................1 (11) a little or some difficulty............................................2 (11) extreme difficulty ......................................................3 = INELIGIBLE (11) you stopped participating in small group/ social settings because of your hearing......................4 = INELIGIBLE (11) you stopped participating in small group/social settings for other reasons or you are not interested in reading...................................................5 (11) The next few questions are about medical conditions you might have.

PAGE 174

174 11. Has a doctor or a nurse ev er told you that you have: YES NO DK N/A a. Alzheimer's Disease, or dementia 1 = INELIGIBLE 2 8 b. Huntingtons disease, with dementia symptoms 1 = INELIGIBLE 2 8 c. Parkinsons disease with dementia symptoms 1 = INELIGIBLE 2 8 d. Recurring epilepsy? 1 = INELIGIBLE 2 8 e. Stroke? 1 = ASK NEXT QUESTION 2 8 f. [Was it in the past year?] 1 = INELIGIBLE 2 8 7 g. [Do you have limb weakness or paralysis as a result?] 1= INELIGIBLE 2 8 -7 h. heart attack or myocardial infarction? 1 = ASK NEXT QUESTION 2 8 i. [Was it in the past year?] 1 = INELIGIBLE 2 8 7 j. A head injury requiring hospitalization any time in your lifetime? 1= INELIGIBLE 2 8 k. been hospitalized for psychiatric illness at any point in your lifetime, or do you currently have a psychiatric illness? 1= INELIGIBLE 2 8 l. cancer, other than skin cancer, within the past 5 years? 1= ASK NEXT QUESTION 2 (14) 8 (14) m. [Are you currently receiving chemotherapy or radiation treatment for this cancer?] 1= INELIGIBLE 2 8 -7 n. Did you ever receive radiation treatment for a cancer above the chest? 1= INELIGIBLE 2 8 -7

PAGE 175

175 o. Do you have a pacemaker or internal defibrillator? 1= INELIGIBLE 2 8 p. Do you use portable oxygen? 1= INELIGIBLE 2 8 q. Do you take steroids or cortisone? 1= INELIGIBLE 2 8 Meds for Asthm a,OK r. Do you use a cane or walker? 1= INELIGIBLE 2 8 s. Did you ever have medical problems as a consequence of alcohol or drug use? 1= INELIGIBLE 2 8 t. Did you ever have legal problems as a consequence of alcohol or drug use? 1= INELIGIBLE 2 8 u. Did you ever have withdrawal symptoms related to alcohol or drug use? 1= INELIGIBLE 2 8 v. Are you currently on any medication? 1 = ASK NEXT QUESTION 2 8

PAGE 176

176 w. [If known, are you currently on any calcium channel blockers or beta blockers?] List Meds: (Name and purpose) __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ __________________________ 1= INELIGIBLE Acebutolol (Monitan, Secral) Atenolol (ApoAtenolol, NovoAtenol, Ternormin) Betaxolol (Kerlone) Bisprolol (Zebeta) Carteolol (Cartrol) Labetalol (Normodyne) Oxprenolol (Trasicor, SlowTrasicor) Bepridi (Vascor) Diltazem (Cardizem, Cardizem CD, Cardizem LA, Cardizem SR, Dilacor-XR) Betaxolo (Kerlone) Toprol Norvasc If still unsure, look in book or Rxlist.com (clinical pharmacology) 2 8 TELEPHONE INTERVIEW FOR COGNITIVE STATUS (TICS) For the next few questions, it is very important that you turn off your television or radio, so that you can concentrate and hear me clearly. Please look around you, and move all papers, ne wspapers, pens and pencils away from where you are. I will be asking you some questions that require you to use your memory and its important that you dont write anything down for this part. [ NOTE TO SCREENERS : SINGLE REPETITIONS AR E PERMITTED ON ALL ITEMS EXCEPT FOR T5 AND T8. T1. Please tell me your full name. ______________________________ 1 point for first name, 1 point for last name ____ / 2 T2. What is todays date? ______________________________ Prompt for missing parts (month, date, year, day of week, season) 1 point for month 1 point for date 1 point for year 1 point for day of week 1 point for season ____ / 5

PAGE 177

177 T3. Please tell me your age and phone number. _______________________________ 1 point for age 1 point for phone number ____ / 2 T4. Count backwards from 20 to 1. ______________________________ 2 points if completely correct on first trial; 1 point if completely correct on second trial; 0 points for anything else ____ / 2 T5. Im going to read you a list of ten words. Please listen carefully and try to remember them. When I am done, tell me as many words as you can, in any order. Ready? The words are: cabin, pipe, elephant, chest, silk, theatre, watch, whip, pillow, giant. Now tell me all the words you can remember. _________________________________ _________________________________ _________________________________ _________________________________ _________________________________ 1 point for each correct response. No penalty for repetitions or intrusions. ( cabin, pipe, elephant, chest, silk, theatre, watch, whip, pillow, giant ) ____ / 10 T6. One hundred minus 7 equals what? _______________ And 7 from that? ____________ Keep going ____________ Keep going ____________ Keep going ____________ Stop at 5 serial subtractions. 1 point for each correct subtraction. ( 93-86-79-72-65 ) Do not inform the participant of incorrect responses, but allow subtractions to be made from his/her last response (e.g., -85-7871-65 would get 3 points). ____ / 5 T7. What do people usually use to cut paper? __________________________ How many things are in a dozen ____ What do you call the prickly green plant that lives in the desert? ____________ What animal does wool come from? ______________________________ 1 point for scissor or shears 1 point for 12 1 point for cactus 1 point sheep or lamb ____ / 4

PAGE 178

178 T8. Say this: No ifs, ands, or buts. Say this: Methodist Episcopal. 1 point for complete repetition on the first trial. Repeat only if poorly presented. ____ / 2 T9. What is the full name of the President of the United States right now? ______________________________ What is the full name of the Vice President? ______________________________ 1 point for correct first and last name. ( George W. Bush in 2002/2003 ) 1 point for correct first and last name. ( Richard Cheney in 2002/2003 ) ____ / 4 T10. With your finger, tap 5 times on the part of the phone you speak into. 2 points if 5 taps are heard 1 point if participant taps more or less than 5 times ____ / 2 T11. Im going to give you a word and I want you to give me its opposite. For example, the opposite of hot is cold. What is the opposite of west; ______________________________ What is the opposite of generous? _______________________________ 1 point for east 1 point for selfish, greedy, stingy, tight, cheap, mean, meager, skimpy, or other good antonym. ____ / 2 T12. Please tell me all the words you remember from the list I gave you before. _________________________________ _________________________________ _________________________________ _________________________________ 1 point for each correct response. No penalty for repetitions or intrusions. ( cabin, pipe, elephant, chest, silk, theatre, watch, whip, pillow, giant ) ____ / 10 TOTAL TICS SCORE If score is 27 or below, complete rest of screener, then confer with another AAMP staff member to determine eligibility. _____ / 50 15. EXERCISE STAGING ALGORITHM (ESA)

PAGE 179

179 I am now going to ask you a few questions about your recent exercise ha bits. To do this, I will need to read to you a definition of what we m ean by regular exercis e so that we understand each other. Are you ready to hear the de finition? [WAIT UNTIL PARTICIPANT SEEMS ATTENTIVE AND READY TO LISTEN] Regular exercise is any pla nned voluntary physical ac tivity (such as brisk walking, aerobics, jogging, bicycling, swimming, basketball, etc.) perfor med to increase physical fitness. Such activity should be performed 3 to 5 times per week for a minimum of 20 minutes per session Exercise does not have to be pain ful to be effective, but should be done at a level that increases your breathing rate and causes you to break a sweat Is this definition clear to you? [IF YES, CONTINUE. IF NO, CLARIFY ANY CONF USIONS, PROBE FOR EXAMPLES OF ACTIVITIES THEY SUGGEST] (Record Persons report of Routine) _________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ 15a. Based on this definition, do you currently exercise regularly? YES.................................................................1 = INELIGIBLE NO...................................................................2 GO TO QUESTION #15b [QUESTIONS #2 AND #3 ARE CODED AS ONE ITEM] 15b. Do you intend to begin exercising regularly? YES.................................................................1 GO TO QUESTION #15c NO...................................................................2 SKIP QUESTION #15c 15c. Do you intend to begin exercising regularly in the next 30 days or the next 6 months? Next 30 days..1 Next 6 months.2 Determine eligibility here, before continuing. If ineligible, skip to last page and read closeout script.

PAGE 180

180 If eligible say: This study has several sections. First, base d on our phone call today, I would like to schedule an in-person visit, during which I can assess as pects of your mental a nd physical fitness, your health, and your everyday functioning, as well as to begin your enrollment at our Living Well facility. I would need to schedule this mee ting within the next few weeks. Depending on individual circumstances, th is session can take anywhe re from 1 to 3 hours. 16. Are you willing to schedule this in-person meeting? YES................................................................1 NO..................................................................2 = INELIGIBLE In order to be eligible to partic ipate in this study and to ensure that it is safe for you to be physically active, we also need a signed check list from your doctor or nurse clearing you to exercise. We have prepared this checklist, and you should be receiving it by mail soon. It should only take your doctor a few moments to complete it. We will ask you to get it filled out in the next 1-2 weeks. 17. Will you be able to make sure that your doctor or health care provider completes this checklist? YES................................................................1 NO..................................................................2 = INELIGIBLE After our first in-person se ssion and clearance from your docto r, you may be eligible to participate in our program. At the outset, how ever, you should know that this is an involved study and will require a significant time comm itment on your part. Over the 16-week study period, you will be asked to devote approximately 1.5 hours each week to meeting with research staff. Although we provide convenient parking and flexible scheduling, it is important for you to consider whether this is reasonable for you. 18. Are you able to participate in the st udy for the entire 16-week study period? YES................................................................1 NO..................................................................2 = INELIGIBLE In addition to the weekly time commitment, you will be asked to wear an activity monitoring device throughout the study. These devices, either an acceleromet er or pedometer, should not restrict any of your normal daily activity. However it is important that you wear this device daily

PAGE 181

181 throughout the entire 16-week study period. It is important for you to consider whether this is reasonable for you. 19. Do you think you would be able to wear th e accelerometer or pedometer daily throughout the study period? YES................................................................1 NO..................................................................2 = INELIGIBLE You will also be asked to complete daily questi onnaires that ask questions about your sleep and exercise behaviors during the previous day. Although these questionnaires should require no more than 5 minutes per day, its im portant that they are complete daily and that few days are missed. It is important for you to consider whether this is reasonable for you. 20. Would you be willing to complete daily que stionnaires consistent ly throughout the study period? YES................................................................1 NO..................................................................2 = INELIGIBLE

PAGE 182

182 TAKE A PAUSE WHILE YOU BRIEFLY ASSESS THE FOLLOWI NG TWO ITEMS. INTERVIEWER ASSESSMENT OF PARTICIPANT COMMUNICATION USING THE SCORING CRITERIA ON THE NEXT PAGE, CODE YOUR ASSESSMENT OF PARTICIPANT'S ABILITY TO MAKE SELF UNDERSTOOD AND TO UNDERSTAND OTHERS. THESE JUDGMENTS CAN BE MADE BOTH ON THE BASI S OF COGNITIVE UNDERSTANDING, AND ALSO OF ENGLIS H-AS-SECOND-LANGUAGE ISSUES. 21. MAKING SELF UNDERSTOOD UNDERSTOOD....................................................................0 USUALLY UNDE RSTOOD.................................................1 (DIFFICULTY FINDING WO RDS OR FINISHING THOUGHTS.) SOMETIMES UNDERSTOOD............................................2 = INELIGIBLE (ABILITY IS LIMITED TO MAKING CONCRETE REQUESTS.) RARELY/NEVER UNDERSTOOD.....................................3 = INELIGIBLE 22. ABILITY TO UNDERSTAND OTHERS UNDERSTANDS..................................................................0 USUALLY UNDERSTANDS...............................................1 (MAY MISS SOME PART/INTENT OF MESSAGE.) SOMETIMES UNDERSTANDS..........................................2 = INELIGIBLE (RESPONDS ADEQUATELY ONLY TO SIMPLE, DIRECT COMMUNICATION) RARELY/NEVER UNDERSTANDS...................................3 = INELIGIBLE MAKING SELF UNDERSTOOD SCORING 0 = Understood: The particip ant expresses ideas clearly. 1 = Usually Understood: The participant has difficulty finding the ri ght words or finishing

PAGE 183

183 thoughts, resulting in delayed respon ses; or requires some prompting to make self understood. 2 = Sometimes Understood: The participant has limited ability, but is able to express concrete request regarding at least basic n eeds (e.g., food, drink, sleep, toilet). 3 = Rarely/Never Understood: At best, understanding is limited to interpretation of highly individual, person-specific sounds (e. g. indicated presence of pain or need to toilet). MAKING SELF UNDERSTOOD SCORING 0 = Understands: Clearly comprehends th e interviewer's message(s) and demonstrates comprehension by words or questions. 1 = Usually Understands: May miss some part of intent of the message but comprehends most of it. The participant may have periodic difficulties integrating information but generally demonstrates comprehension by responding to words or questions. 2 = Sometimes Understands: Demonstrates frequent difficultie s integrating information, and responds adequately only to simple and direct questions or directions. When the message is rephrased or simplified, the participant's comprehension is enhanced. 3 = Rarely/Never Understands: Demonstrates very limited ability to understand communication. Or, interviewer has difficulty determining whether the participant comprehends messages, based on verbal responses. Or, the participant can hear sounds but does not understand messages. DETERMINE PARTICIPANT ELIGIBILIT Y AND GO TO APPROPRIATE SCRIPT ON NEXT PAGE.

PAGE 184

184 FOR ELIGIBLE PARTICPANTS Thank you for answering the questions. As I already implied, we w ould like to continue with you in the program and meet you in person and have you meet us. At this in-person meeting, we will ask you some more questions regarding you r mental and physical fitness, health, and everyday functioning to determine if you are elig ible for participation in the program. This meeting will take up to three hours depending on how much information is needed from you. It will be held at the Living Well Center at the University of Florida. Do you have any questions for me at this time? You have already indicated that you would be willing to schedule this session. To help me with scheduling an appointment, could you tell me wh at other commitments you typically have during the week, such as work, vol unteering, caring for others or social activities? 23. What days of the week work best for you? ____________________ 24. Is morning or afternoon better for you? ______________________ We will be getting in touch with you soon to schedule your first visit. In the meantime, you should be receiving the form for your doctor or nurse to complete to be cleared to participate in exercise, and we will need this clearan ce before we can enroll you in the study. Thank you very much for your time. Could I schedule you for (date/time)? 25. Date / / ______ 26. Time: : AM / PM 27. Test Site: 28. Person: ___ (Tester ID/Initials) Do you know where Living Well is located? I will send you a letter and map with directions to our center. Thank you. If you wear glasses for distance or read ing or wear a hearing aid, please bring them with you. You should have received, or will be receiving shortly, a one-paged form that needs to be signed by your physician. Please bring this signed physicians form and any medications you take with you to your appointment. We look forward to seeing you on (day/date).

PAGE 185

185 INELIGIBLE SCRIPT This concludes your participation in this st udy. Thank you for answering these questions. This has been very helpful. Based on our interview today, you are not eligible to participate in the study at this time. This is typically because i ndividuals have health conditions that have been identified as exclusion characteristics for th is study. We appreciate the time you have spent answering these questions. Althoug h you are not eligible for this study, we may want to call you in the future about your interest in some other study. 29. May we have permission to share your interest in research with our aging research colleagues here at the University of Florida, so that other researchers ca n call you to invite you to consider participating in future research studies? YES................................................................1 NO..................................................................2

PAGE 186

186 APPENDIX B NURSE/PHYSICIAN CHECKLIST AND PERMISSION Project AAMP (Active Adult Mentoring Project) College of Public Health and Health Professions Adrienne Aiken Morgan PO Box 100165 Gainesville, FL 32610-0165 Phone: (352) 273-5098 Please type or print clearly Physicians Name ___________________________ Phone # _______________________ Patients Name ______________________________ (Project AAMP participant) Program exclusion checklis t (please check any that apply to this patient): Terminal illness with lif e expectancy of < 12 months Myocardial infarction in the last 6 months Chronic heart failure (New York Classification III to IV) Aortic stenosis Cardiac arrhythmia Cardiac stent Cardiac arrest Implanted cardiac defibrillator Pacemaker Coronary artery bypass graft History of cardiac arrest Uncontrolled angina Stroke or TIA Peripheral vascular disease Pulmonary disease requiring oxygen or steroid treatment Receiving chemotherapy or radiation for cancers Ambulation with assistive devices Poorly controlled diabetes Smoked regularly (>4 cigarettes per day) in past 3 years Any of the following calcium channel or beta blockerS Beta Blockers Acebutolol (Monitan, Sectral) Atenolol (Apo-Atenolol, Novo-Atenol, Tenormin) Betaxolol (Kerlone) Bisoprolol (Zebeta) Carteolol (Cartrol) Labetalol (Normodyne, Trandate) Oxprenolol (Trasicor, Slow-Trasicor) Calcium Channel Blockers Bepridil (Vascor) Diltiazem (Cardizem, Cardizem CD, Cardizem LA, Cardizem SR, Dilacor-XR) Betaxolo (Kerlone) Note: Trade Names in parenthesis I hereby give my patient permission to: 1. Participate in an exercise program YES NO 2. Complete a health and fitness assessments* YES NO *The fitness assessment includes resting heart rate and blood pressure measurements and an 85% sub-maximal cardiovascular test (heart rate only, no EKG) Special instructions or indicated activities: ___________________________________________________ Contraindications to any activities: _________________________________________________________ ____________________________ ____________ Nurse/Physicians Signature [Required] Date

PAGE 187

187 APPENDIX C QUALITY CONTROL CHECKLISTS EXERCISE PROMOTION INTERVENTION GROUP Mentor/Coach ______________________ Date ____________ Procedure None Part Full Score CREATING A SUPPORTIVE ENVIRONMENT 1. Reviews participants exercise during previous week. 0 5 10 _____ 2. Use of Open-Ended Questions ex.: In what ways has exercise been helpful to you? 0 5 10 _____ 3. Use of Affirmations ex.: You are a very co nscientious person. That quality will help you to begin an exercise program. 0 5 10 _____ 4. Use of Reflective Responses ex.: It sounds like you are frustrated. How do you deal with that? 0 5 10 _____ 5. Use of Summary Statements 0 5 10 _____ 6. Effective group management Keeps group on topic Manages time and pace of discussion Maintains leadership of discussion 0 5 10 _____ COMMUNICATION ROADBLOCKS 7. Avoids Lecturing 0 10 20 _____ 8. Avoids giving advice 0 10 20 _____ 9. Avoids interpreting or analyzing 0 10 20 _____ 10. Avoids questioning participant 0 10 20 _____ CONCLUSION 11. Assures participant that all instructions are in the Workbook & reminds to bring all logs and complete any homework for next session. 0 5 10 _____ 12. Makes appropriate referrals regarding questions participants may have about the study. 0 5 10 _____ 13. Makes appropriate referrals regarding mental or physical health concerns obse rved during sessions. 0 5 10 _____ Comments for Mentor: Total __________

PAGE 188

188 Session Specific Evaluations Procedure None Part Full Score GOAL SETTING SESSION 7. Use of OARS during discussions about goals 0 5 10 _____ 8. Distinguishes between longand short-term goals 0 5 10 _____ 9. Clearly discusses SMART goals 0 5 10 _____ 10. Encourages participants to set SMART goals in a non-judgmental manner. 0 5 10 _____ 11. Gives examples of SMART goals 0 5 10 _____ Comments for Mentor: Total __________ Mental imagery SESSION 12. Integrates mental imagery into discussions using OARS. 0 5 10 _____ 13. Makes connections between fitness/health goals and mental imagery with op en-ended questions. 0 5 10 _____ 14. Gives examples of vivid images that evoke all five senses and feelings/emotions. 0 5 10 _____ Comments for Mentor: Total __________

PAGE 189

189 CONTROL GROUP Mentor/Coach ______________________ Date ____________ Procedure None Part Full Score GROUP MANAGEMENT 1. Begin sessions with discussion about last weeks topic, including quiz. 0 5 10 _____ 2. Keeps group on topic 0 5 10 _____ 3. Manages time and pace of discussion 0 5 10 _____ 4. Maintains leadership of discussion 0 5 10 _____ PRESENTATIONAL SKILLS 5. Use of Open-ended questions 0 5 10 _____ 6. Clearly presents topic to be discussed 0 5 10 _____ 7. Promotes discussion by asking questions 0 5 10 _____ 8. Actively makes efforts to include all members of the group into discussion. 0 5 10 _____ COMMUNICATION ROADBLOCKS 9. Avoids Lecturing 0 10 20 _____ 10. Avoids giving advice 0 10 20 _____ 11. Avoids interpreting or analyzing 0 10 20 _____ 12. Avoids questioning participant 0 10 20 _____ CONCLUSION 13. Assures participant that all instructions are in the Workbook & reminds to bring all logs and complete any homework for next session. 0 5 10 _____ 14. Makes appropriate referrals regarding questions participants may have about the study. 0 5 10 _____ 15. Makes appropriate referrals regarding mental or physical health concerns observ ed during sessions. 0 5 10 _____ Comments for Mentor: Total __________

PAGE 190

190 LIST OF REFERENCES Aiken-Morgan, A. T., Marsiske, M. & Whitfiel d, K. E. (2008). Characterizing and explaining differences in cognitive test perform an ce between African American and European American older adults. Experimental Aging Research, 34(1), 80-100. Allaire, J.C., Tamez, E., & Whitfield, K. E. (2007). Examining the association between lung functioning and cognitive performance in African American adults. Journal of Aging and Health, 19, 106-122. Alexander, G.E., DeLong, M.R., & Strick, P.L. (1986). Parallel organi zation of functionally segregated circuits linking basal ganglia and cortex. Annual Review of Neuroscience, 9, 357-381. Alexopoulos, G. S. (2006). The vascular depression hypothesis: 10 years later. Biological Psychiatry, 60(12) 1304-1305. Alexopoulos, G. S., Kiosses, D. N., Choi, S. J ., Murphy, C. F., & Lim, K. O. (2002). Frontal white matter microstructure and treatment res ponse of late-life depression: A preliminary study. American Journal of Psychiatry, 159 1929-1932. Alexopoulos, G. S., Meyers, B. S., Young, R. C., Campbell, S., Silbersweig, D., & Charlson, M. (1997). The "vascular de pression" hypothesis. Archives of General Psychiatry, 54(10), 915-922. Angevaren. M., Aufdemkampe, G., Verhaar, H. J. J., Aleman, A., Vanhees, L. (2008). Physical activity and enhanced fitness to improve cogni tive function in older people without known cognitive impairment (Review). Cochrane Database of Systematic Reviews, Apr 16 (2), CD005381. Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman Ball, K., Berch, D. B., & Helmers, K. F. (2002). Effects of cognitive trai ning interventions with older adults: A randomized controlled trial. Journal of the American Medical Association, 288(18) 2271-2281. Ball, K., & Owsley, C. (2000). Increasing mobility and reducing accidents in older drivers. In K.W. Schaie (Ed.), Societal Impacts on Mobi lity in the Elderly New York: Springer. Baltes, P. B., & Mayer, K. U. (Eds.). (2001). The Berlin Aging Study: Aging from 70 to 100 (Paperback ed.). New York: Cambridge University Press. Baltes, P. B., Staudinger, U. M., Lindenberger U. (1999). Lifespan psychology: Theory and application to inte llectual functioning. Annual Review of Psychology, 50, 471-507. Baron, R. M., & Kenny, D. A. (1986). The modera tor-mediator variable di stinction in social psychological research: Conceptual, stra tegic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182.

PAGE 191

191 Bean, J., Vora, A., & Frontera, W. (2004). Benefits of exercise for community-dwelling older adults. Archives of Physical Medicine and Rehabilitation, 85(3) s31-s42. Beck, A. T., Brown, G., & Steer, R. A. ( 1996). Beck Depression Inventory II manual. San Antonio, TX: The Psychological Corporation. Benetos, A., Waeber, J. I., Mitchell, G., Resnick, L., Asmar, R., & Safar, M. (2002). Influence of age, risk factors, and cardi ovascular and renal disease on ar terial stiffness: Clinical applications. American Journal of Hypertension, 15, 1101-1108. Benton, A. & Hamsher, K. (1989). Multilingual Aphasia Examination Iowa City: AJA Associates. Blair, J. R. & Spreen, O. (1989). Predicting pr e-morbid IQ: A revision of the National Adult Reading Test. The Clinical Neuropsychologist, 3, 129-136. Blissmer, B. & McAuley, E. (2002). Testing th e requirements of stages of physical activity among adults: the comparative effectiveness of stage-matched, mismatched, standard care, and control interventions. Annals of Behaviora l Medicine, 24(3) 181-189. Brandt, J., Spencer, M., & Folstein, M. (1988). The Telephone Interview for Cognitive Status. Neuropsychiatry, Neuropsychology, and Behavioral Neurology, 1, 111-117. Breteler, M. M. B., van Swieten, J. C., Bots, M. L., & Grobbee, D. E. (1994). Cerebral white matter lesions, vascular risk factors, and cognitive functi on in a population-based study: The Rotterdam study. Neurology, 44(7), 1246-1252. Brown, D. R. (1992). Physical activity, ageing, an d psychological well-being: An overview of research. Canadian Journal of Sports Sciences, 17, 185-193. Bryan, J., & Luszcz, M. A. (2000). Measurement of executive function: considerations for detecting adult age differences. Journal of Clinical & Ex perimental Neuropsychology, 22(1) 40-55. Buman, M. P. (2008). Evaluation of a peer-assiste d social cognitive physical activity intervention for older adults. Unpublished doctoral di ssertation, University of Florida. Burke, G. L., Arnold, A. M., & Bild, D. E. (2001). Factors associated w ith healthy aging: The Cardiovascular Health Study. Journal of the American Geriatrics Society, 49(3), 254-262. Cabeza, R. (2002). Hemispheric asymmetry redu ction in older adults: The HAROLD model. Psychology of Aging, 17(1) 85-100. Cahn-Weiner, D. A., Boyle, P. A., & Malloy, P. F. (2002).Tests of executive function predict instrumental activities of daily living in community-dwelling older individuals. Applied Neuropsychology, 9(3), 187-191.

PAGE 192

192 Carmelli, D., Swan, G. E., Reed, T., Miller, B., Wo lf, P. A., Jarvik, G. P., & Schellenberg, G. D. (1998). Midlife cardiovascular risk factors, ApoE, and cognitive dec line in elderly male twins. American Academy of Neurology,50(6) 1580-1585. Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159. Cohen, J.D., Forman, S.D., Braver, T.S., Casey, B. J., Servan-Schreiber, D., & Noll, D.C. (1994). Activation of prefrontal cortex in a non-spat ial working memory task with functional MRI. Human Brain Map, 1, 293. Cohen, J.D., Perlstein, W.M., Braver, T.S., Nystrom, L.E., Noll, D. C., Jonides, J., & Smith, E.E. (1997). Temporal dynamics of brain activati on during a working memory task. Nature, 386, 604-608. Colcombe, S. & Kramer, A. (2003). Fitness eff ects on the cognitive function of older adults: A meta-analytic study. Psychological Science, 14(2) 125-130. Crawford, J. R., Bryan, J., & Luszcz, M. A. (2000). Executive decline hypothesis of cognitive aging: Do executive deficits qualify as differential deficits and do they mediate age-related memory decline? Aging, Neuropsychology, and Cognition, 7 9-31. Crawford, J. & Henry, J. (2005). Assessment of executive dysfunction. In P. W. Halligan & D. T. Wade (Eds.), Effectiveness of rehabili tation for cognitive deficits (pp. 233-245). Oxford, UK: Oxford University Press. Cummings, J. (1993). Frontal-subcortic al circuits in human behavior. Archives of Neurology, 50 873-880. Deary, I. J., Starr, J. M., & MacLennan, W. J. (1998). Fluid intellig ence, memory, and blood pressure in cognitive aging. Personality and Individual Differences, 25, 605-619. den Heijer, T., Skoog, I., Oudkerk, M., de Leeuw, F-E., de Groot, J. C., Hofman, A., & Breteler, M. M. B. (2003). Association between blood pressu re levels over time and brain atrophy in the elderly. Neurobiology of Aging, 24 307-313. DiPietro, L. (2001). Physi cal activity in aging: Changes in patterns and their relationship to health and function. Journals of Gerontology, 56a, 13-22. Dixon, R.A., & Hultsch, D.F. (1983). Structure a nd development of metamemory in adulthood. Journal of Gerontology, 38 682-688. Dustman, R. E., Ruhling, R. O., Russell, E. M., Sh earer, D. E., Bonekat, H. W., Shigeoka, J. W., Wood, J. S., & Bradford, D. C. (1984). Aerobic exercise training and improved neuropsychological function of older individuals. Neurobiology of Aging, 5 35-42. Elavsky, S. & McAuley, E. (2007). Exercise and self-esteem in menopausal women: A randomized controlled trial involving walking and yoga. American Journal of Health Promotion, 22(2) 83-92.

PAGE 193

193 Elavsky, S., McAuley, E., Motl, R. W., Konopack, J. F., Marquez, D. X., Hu, L., Jerome, G. J., & Diener, E. (2005). Physical activity enhances long-term quality of life in older adults: Efficacy, esteem, and affective influences. Annals of Behavioral Medicine, 30(2) 138-145. Elias, M. F., Elias, P. K., & Sullivan, L. M. (2003). Lower cognitive function in the presence of obesity and hypertension: The Framingham heart study. International Journal of Obesity & Related Metabolic Disorders, 27(2) 260-268. Emery, C. F., Schein, R. L., Hauck, E. R., & MacIntyre, N. R. (1998). Psychological and cognitive outcomes of a randomized trial of exercise among patients with chronic obstructive pulmonary disease. Health Psychology, 17(3) 232-240. Esiri, M. M., Hyman, B. T., Beyreuther, K., & Ma sters, C. L. (1997). Ageing and dementia. In D. I. Graham & P. L. Lantos (Eds.), Greenfields neuropathology, Vol 2 Arnold, London; pp 153-233. Folstein, M. F., Folstein, S. E. & McHugh, P. R. (1975). Mini Mental State: A practical method for grading the cognitive states of patients for the clinician. Journal of Psychiatric Research, 12 189-198. Friedman, N. P., & Miyake, A. (2004). The rela tions among inhibition and interference control functions: A latent-variable analysis. Journal of Experimental Psychology: General, 133 101-135. Fukukawa, Y. et al. (2004). Age differences in the effect of physical activity on depressive symptoms. Psychology and Aging, 19(2) 346-351. Gatz, M. (2000). Variations on depression in late r life. In S. H. Qualls & N. Abeles (Eds.), Psychology and the aging revolution: How we adapt to longer life. (pp. 239-254). Washington, DC: American Psychological Association. Giacobbi, Jr., P. R., Buman, M. P., & Harris, K. G. (2006). Active ad ult mentoring program: Project AAMP participant manual. Unpublis hed manual, University of Florida. Graves, A. B., Bowen, J. D., Rajaram, L., McCo rmick, W. C., McCurry, S. M., Schellenberg, G. D., & Larson, E. B. (1999). Impaired olf action as a marker for cognitive decline: Interaction with apolipoprotein E 4 status. Neurology, 53(7) 1480-1487. Greenwood, P. M. (2000). The front al aging hypothesis evaluated. Journal of the International Neuropsychological Society, 6(6) 705-726. Godin, O., Dufouil, C., Maillard, P., Delcroix, N ., Mazoyer, B., Crivello, F., Alperovitch, A., & Tzourio, C. (2008). White matter lesions as a pr edictor of depression in the elderly: The 3C-Dijon Study. Biological Psychiatry, 63 663-669. Godin, G., Jobin, J., & Bouilon, J. (1986). Assessment of leisure time exercise behavior by selfreport: A concurrent validity study. Canadian Journal of Public Health, 77, 359-361.

PAGE 194

194 Godin, G., & Shephard, R. J. (1985). A simple method to assess exercise behavior in the community. Canadian Journal of Applied Sport Science, 10 141-146. Hartlage, S., Alloy, L. B., Vazquez, C., & Dykman, B. (1993). Automatic and effortful processing in depression. Psychological Bulletin, 113(2) 247-278. Haskell, W. et al. (1992). Card iovascular benefits and assessment of physical activity and physical fitness in adults. Medicine and Science in Sports and Exercise, 201-220. Helmrick, S. P., Ragland, J. M., Leung, R. W ., & Paffenbarger, R. S., Jr. (1991). Physical activity and reduced occurrence of non-insulin dependent diabetes mellitus. New England Journal of Medicine, 325 147-152. Horn, J.L., & Cattell, R.B. (1966). Refinement and test of the theory of fluid and crystallized intelligence. Journal of Educational Psychology, 57, 253. Jacobs, D. R., Jr., Ainsworth, B. E., Hartma n, T. J., Leon, A. S. (1993). A simultaneous evaluation of 10 commonly used physical activity questionnaires. Medicine and Science in Sports and Exercise, 25(1) 81-91. Jobe, J., Smith, D., Ball, K., Tennstedt, S., Marsiske, M., Willis, S., Rebok, G., Morris, J., Helmers, K., Leveck, M., & Kleinman, K. (2001). ACTIVE: A Cognitive Intervention Trial to Promote Independence in Older Adults. Controlled Clinical Trials, 22, 453-479. Karvinin, K. H., Courneya, K. S., Campbell, K. L., Pearcy, R. G, Dundas, G., Capstick, V., et al. (2007). Correlates of exercise motivation and behavior in a populat ion-based sample of endometrial cancer survivors: an applica tion of the theory of planned behavior. Journal of Behavioral Nutrition and Physical Activity, 4, 21-31. Kemper, S. & McDowd, J. M. (2008). Dimensions of cognitive aging: Executive function and verbal fluency. In S. M. Hofer & D. F. Alwin (Eds.), Handbook of cognitive aging: Interdisciplinary perspectives. (pps. 181-192). Los Angele s: Sage Publications. Khatri, P., Blumenthal, J. A., Babyak, M. A., Craighead, W. E., Herman, S., Baldewicz, T., Madden, D. J., Doraiswamy, M., Waugh, R., & Krishman, K. R. (2001). Effects of exercise training on cognitive functioning among depressed older men and women. Journal of Aging and Physical Activity, 9 43-57. Kramer, A. F., Erickson, K. I., & Colcombe, S. J. (2006). Exercise, cognition, and the aging brain. Journal of Applied Physiology, 101 1237-1242. Kramer, J. H., Kemenoff, L. A., & Chui, H. C. (2001). The neuropsychology of subcortical ischemic vascular dementia. In S. R. Waldstein & M. F. Elias (Eds.), Neuropsychology of cardiovascular disease (pps. 83-104). Mahwah, NJ: Eribaum. Kramer, A. F., & Willis, S. L. (2002). Enhanc ing the cognitive vitali ty of older adults. Current Directions in Psychol ogical Science, 11(5) 173-177.

PAGE 195

195 Krishnan, K. R. R., Taylor, W. D ., McQuaid, D. R. E., MacFall, J. R., Payne, M. E., Provenzale, J. M., & Steffens, D. C. ( 2004). Clinical characteristics of magnetic resonance imagingdefined subcortical ischemic depression. Biological Psychiatry, 55 390-397. Krishnan, M., Mast, B. T., Flicker, L. J., Lawhorne, L., & Lichtenberg, P. A. (2005). The effects of preexisting depression on cerebrovascular he alth outcomes in geriatric continuing care. Journal of Gerontology: Medical Sciences, 60(7) 915-919. Krishnan, M. S., OBrien, J. T., Firban, M. J., Pantoni, L., Carlucci, G ., Erkinjuntti, T., Wallin, A., Wahlund, L. O., Scheltens, P., van Straaten, E. C. W., & Inzitari, D. (2006). Relationship between periventricular and d eep white matter lesions and depressive symptoms in older people. The LADIS Study. International Journal of Geriatric Psychiatry, 21 983-989. La Rue, A. (1992). Depression, Aging and neuropsychological assessment (pp. 259-289). New York: Plenum Press. Launer, L. J., Ross, G. W., & Petrovitch, H. (2000). Midlife bl ood pressure and dementia: The Honolulu-Asia Aging Study. Neurobiology of Aging, 21(1) 49-55. Leirer, V. O., Morrow, D. G., Pariante, G. M ., & Sheikh, J. I. (1988). Elders nonadherence, its assessment, and computer assisted in struction for medication recall training. Journal of the American Geriatrics Society, 36, 877-884. Lezak, M., Howieson, D., & Loring, D. (2004). Neuropsychological assessment, (4th Edition). New York: Oxford University Press. Libon, D., Price, C., Garrett, K., & Giovannetti, T. (2004). From Binswangers disease to Leuokoaraiosis: What we have learned about subcortical vascular dementia. The Clinical Neuropsychologist, 18(1) 83-100. Liu-Ambrose, T., et al. (2004). Re sistance and agility training redu ce fall risk in women aged 7585 with low bone mass: A 6-month randomized, controlled trial. Journal of American Geriatrics Society, 52(5) 657-665. Lopez, OL; Jagust, WJ; DeKosky, ST; Becker, JT ; Fitzpatrick, A; Dulb erg, C; Breitner, J; Lyketsos, C; Jones, B; Kawas, C; Carlson, M; Kuller, LH. (2003). Prevalence and classification of mild cognitive impairment in the Cardiovascular Health Study Cognition Study: part 1. Archives of Neurology, 10(60) 1385-1389. Luszcz, M. A. & Lane, A. P. (2008). Executive function in cognitive, neuropsychological, and clinical aging. In S. M. Hofer & D. F. Alwin (Eds.), Handbook of cognitive aging: Interdisciplinary perspectives. (pps. 193-206). Los Angele s: Sage Publications. Madden, D. J., Blumenthal, J. A., Allen, P. A., & Emery, C. F. (1989). Improving aerobic capacity in healthy older adults does not necessarily lead to improved cognitive performance. Psychology and Aging, 4(3) 307-320.

PAGE 196

196 Madden, D. J., Langley, L. K., Thurston, R. C ., Whiting, W. L., & Blumenthal, J. A. (2003). Interaction of blood pressure and adult ag e in memory search and visual search performance. Aging, Neuropsychology, and Cognition, 10(4) 241-254. Maddux, J. (1993). Social cognitive models of health and exercise behavior : An introduction and review of conceptual issues. Journal of Applied Sport Psychology, 5 116-140. Manly, J. J., Jacobs, D. M., Touradji, P., Small, S. A., & Stern, Y. (2002). Reading level attenuates differences in ne uropsychological test performa nce between African American and European American elders. Journal of the International Neuropsychological Society, 8 (3) 341-348. Mast, B. T., Azar, A. R., & Murrell, S. A. (2005). The vascular depression hypothesis: The influence of age on the relationship between ce rebrovascular risk factors and depressive symptoms in community dwelling elders. Aging & Mental Health, 9(2) 146-152 McAuley, E. (1992). Self-efficacy relationships with affective and exertion responses to exercise. Journal of Applied Social Psychology, 22 (4) 312-326. McAuley, E. (1993). Self-efficacy and the maintenan ce of exercise participation in older adults. Journal of Behavioral Medicine, 16 103-113. McAuley, E., Elavsky, S., Jerome, A., Konopack, J. F., & Marquez, D. X. (2005). Physical activity-related well-being in older adults: Social cognitive influences. Psychology and Aging, 20 295-302. McAuley, E., Jerome, G. J., Elavsky, S., Marquez, D. X., & Ramsey, S. N. (2003). Predicting long-term maintenance of physi cal activity in older adults. Preventive Medicine, 37 110118. McAuley, E., Kramer, A. F., & Colcombe, S. J. (2004). Cardiovascular fitness and neurocognitive function in older adults: A brief review. Brain, Behavior & Immunity, 18(3) 214-220. McAuley, E., Marquez, D. X., Jerome, G. J. U., Blissmer, B. U., & Katula, J. (2002). Physical activity and physique anxiety in older adults: Fitness and efficacy influences. Aging and Mental Health, 6 220-230. Miyake, A., Friedman, N. P., Emerson, M. J., & W itzki, A. H. (2000). The unity and diversity of executive functions and their contributions to complex frontal lobe tasks: A latent variable analysis. Cognitive Psychology, 41 49-100. Miller, L., & Lachman, M. (1999). The sense of c ontrol and cognitive agi ng: Toward a model of mediational processes. In T. Hess & F. Blanchard-Fields (Eds.), Social cognition and aging. San Diego, CA, US: Academic Press, Inc., pp. 17-41.

PAGE 197

197 Mokdad, A. H., Marks, J. S., Stroup, D. F., & Gerb erding, J. L. (2004). Actual causes of death in the United States, 2000. Journal of the American Me dical Association, 291(10) 12381245. Moritz, S., Birkner, C., Kloss, M., Jahn, H ., Hand, I., Haasen, C., & Krausz, M. (2002). Executive functioning in obsessive-compulsi ve disorder, unipolar depression, and schizophrenia. Archives of Clinical Neuropsychology, 17(5), 477-483. Morgan, K., & Bath, P. (1998). Customary physi cal activity and psychological well-being: A longitudinal study. Age & Aging, S27 35-40. Motl, R. W., Konopack, J. F., McAuley, E., Elavs ky, S., Jerome, G. J., & Marquez, D. X. (2005). Depressive symptoms among older adults: Long-term reduction after a physical activity intervention. Journal of Behaviora l Medicine, 28(4) 385-394.National Center for Health Statistics (1990). Current Estimates from the Nati onal Health Interview Survey, 1989 Hyattsville, MD: Public Health Service. Pantoni, L., Leys, D., Fazekas, F., Longstreth, W., Inzitari, D., Wallin, A., Filippi, M., Scheltens, P., Erkinjuntti, T., & Hachinski, V. (1999). Role of white matter lesions in cognitive impairment of vascular origin. Alzheimers Disease and Associated Disorders, 13(3), S49S54. Pate, R. R., Pratt, M., Blair, S. N., Haskell, W. L., Macera, C. A., & Brouchard, C. (1995). Physical activity and public health: A recommendation from the Centers for Disease Control and Prevention and the Ameri can College of Sports Medicine. Journal of the American Medical Association, 273, 402-407. Pescatello, L., & DiPietro, L. (1993). Physical activity, aging, and psychological well-being. Journal of Aging and Physical Activity, 3 67-96. Pfefferbaum, A., Adalstein, E., & Sullivan, E. V. (2005). Frontal circui try degradation marks healthy adult aging: Evidence from diffusion tensor imaging. Neuroimage, 26(3) 891-899. Rabin, L. A., Barr, W. B., & Burton, L. A. (2005). Assessment practices of clinical neuropsychologists in the United States a nd Canada: A survey of INS, NAN, and APA Division 40 members. Archives of Clinical Neuropsychology, 20 33. Raz, N. (2000). Aging of the brain and its imp act on cognitive performa nce: integration of structural and functional findings. In: Craik, F.M., Salthouse, T.A. (Eds.), The Handbook of Aging and Cognition, Volume 2. Lawrence Erlbaum Associates, Mahweh, NJ, pps. 1-90. 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) 11691180. Rebok, G. W., & Balcerak, L. J. (1989). Memory self-efficacy and performance differences in young and old adults the e ffect of mnemonic training. Developmental Psychology, 25 (5), 714-721.

PAGE 198

198 Reitan, R. M. (1992). Trail Making Test: Manual for Administration and Scoring Tucson: Reitan Neuropsychology Lab. Reitan, R. M., & Wolfson, D. (1994). A selectiv e and critical review of neuropsychological deficits and the frontal lobes. Neuropsychology Review,4 161. Royall, D. R., Chiodo, L. K., & Polk, M. J. (2003). Executive dyscontrol in normal aging: Normative data, factor struct ure and clinical correlates. Current Neurology and Neuroscience Reports, 3, 487-493. Ruitenberg, A., Skoog, I., & Otta, A. (2001). Blood pressure and risk of dementia: Results from the Rotterdam Study and the Gothenburg H-70 Study. Dementia & Geriatric Cognitive Disorders, 12(1), 33-39. Ruppar, T. M., & Schneider, J. K. (2007). Self-repor ted exercise behavior and interpretations of exercise in older adults. Western Journal of Nursing Research, 29, 140-157. Sabatini, T., Frisoni, G. B., Barbisoni, P., Bellelli G., Rozzini, R., Trabucchi, M. (2000). Atrial fibrillation and cognitive di sorders in older people. Journal of the American Geriatrics Society, 48, 387-390. Sallis J. F., & Owen, N. (1999). Physical activity and behavioral medicine. Thousand Oaks, CA Sage Publications. Salthouse, T. A. (2005). Relations between c ognitive abilities and measures of executive functioning. Neuropsychology, 19 532-545. Salthouse, T. A., Atkinson, T. M., & Berish, D. E. (2003). Executive functioning as a potential mediator of age-related cogniti ve decline in normal adults. Journal of Experimental Psychology: General, 132 566-594. Saxby, B. K., Harrington, F., McKeith, I. G., Wesnes, K., & Ford, G. A. (2003). Effects of hypertension on attention, memory, and executive function in older adults. Health Psychology, 22(6) 587-591. Schimdt, R., Fazekas, F., Offenbacher, H., et al. (1993). Neuropsychologi c correlates of MRI white matter changes. European Neurology, 39, 80-89. Schiller, J., Adams, P., & Coriaty, N. (2005). Su mmary health statistics for the US population: National health interview survey 2003. National Center for Health Statistics. Vital Statistics, 10(224). Seeman, T. E., & Crimmins E. (2001). Social environment effects on health and aging: integrating epidemiologic and demogr aphic approaches and perspectives Population Health and Aging Annals of the New York Academy of Sciences, 954, 88-117.

PAGE 199

199 Simic, G., Kostovic, I., Winblad, B., & Bogdanovic, N. (1997). Volume and number of neurons of the human hippocampal formation in normal aging and Alzheimers disease. Journal of Comparative Neurology, 379 482-494. Singh, M. A. F. (2002). Exercise comes of age: Rationale and recommendations for a geriatric exercise prescription. Journal of Gerontology: A Biol Sci Med Sci, 57(5) M262-282. Sneed, J. R., Roose, S. P., & Sackeim, H. A. (2006). Vascular depression: A distinct diagnostic subtype? Biological Psychiatry, 60(12) 1295-1298. Snow, W. G., Tierney, M. C., Zorzitto, M. L., Fisher, R. H., & Reid, D. W. (1988). One-year test-retest reliability of selected tests in olde r adults. Paper presented at the meeting of the International Neuropsychologi cal Society, New Orleans. Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (STAI) Palo Alto, CA: Consulting Psychologists Press. Synder, P., & Nussbaum, P. (1998). Clinical Neuropsychology: A Pocket Handbook for Assessment. American Psychological Association. Tabachnick B. G., Fidell, L. S. (2007). Using multivariate statistics (5th ed.) Boston, MA: Allyn & Bacon/Pearson Education. Taylor, W. D., Steffens, D. C., & Krishnan, K. R. (2006). Psychiatric disease in the twenty-first century: The case for subcortical ischemic depression. Biological Psychiatry, 60(12) 1299-1303. Tudor-Locke, C., Burkett, L., Reis, J. P., Ainsworth, B. E., Macera, C. A., & Wilson, D. K. (2005). How many days of pedometer monitori ng predict weekly physical activity in adults? Preventive Medicine, 40(3), 293-298. Valanis, B. (1999). Epidemiological methods. Epidemiology in health care, 3rd Ed. Stamford, CT: Appleton & Lange. Van Petten, C., Plante, E., Davidson, P. S. R., Ku o, T. Y., Bajuscak, L., & Glisky, E. L. (2004). Memory and executive function in older adults: relationships with tem poral and prefrontal gray matter volumes and white matter hyperintensities. Neuropsychologia, 42(10), 13131335. Verhaeghen, P., Borchelt, M., & Smith, J. (2003). Relation between cardiovascular and metabolic disease and cognition in very old age: Cross-sectio nal and longitudinal findings from the Berlin aging study. Health Psychology, 22(6), 559-569. Waldstein, S. R. (1995). Hypertension and neurop sychological function: A lifespan perspective. Experimental Aging Research, 21(4) 321-352.

PAGE 200

200 Waldstein, S. R., Manuck, S. B., Ryan, C. M., & Muldoon, M. F. (1991). Neuropsychological correlates of hypertension: Review and methodologic considerations. Psychological Bulletin, 110(3) 451-468. Waldstein, S. R., & Katzel, L. I. (2001). Hypertension and cognitive function. In S. R. Waldstein & M. F. Elias (Eds.), Neuropsychology of cardiovascular disease (pps. 83-104). Mahwah, NJ: Eribaum. Watson, D., Clark, L.A., & Tellegen, A. (1988). De velopment and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063-1070. Wechsler, D. (1997). Wechsler Memory Scale Third Edition: Manual San Antonio: The Psychological Corporation. West, R. (1996). An application of prefrontal cortex function theory to cognitive aging. Psychology Bulletin, 20 272-292. West, M. J., Kawa, C. H., Marti n, L. J., & Troncoso, J. C. (2000). The CA1 region of the human hippocampus is a hot spot in Alzheimers disease. Ann NY Academy Science, 908 255259. Yesavage, J.A. & Brink, T. L. (1983). Development and validation of a geriatric depression screening scale: A preliminary report. Journal of Psychia tric Research, 17 37-49. Yesavage, J. A., Brink, T. L., Rose, T. L ., Lum, O., Huang, V., & Adey M. B. (1983). Development and validation of a geriatric depr ession screening scale: a preliminary report. Journal of Psychiatric Research, 17, 37-49. Zuccala, G., Onder, G., Pedone, C., Carosella, L ., Pahor, M., Bernabei, R., & Cocchi, A. (2001). Hypotension and cognitive impairment; selective association in patient s with heart failure. Neurology, 57 1986-1992.

PAGE 201

201 BIOGRAPHICAL SKETCH My academic and research interest in cogniti ve aging began when I was an undergraduate student at Florida A&M University in Tallaha ssee, Florida. During this time, I was a Distinguished Scholar, with a full academic sc holarship. After gra duating summa cum laude with a Bachelor of Arts in ps ychology in April 2002, I attended the University of Florida on a University of Florida Alumni Fellowship. I rece ived a Master of Science in clinical psychology in May 2004. Throughout my graduate career, I was the recipient of vari ous awards and grants, including a National Institute on Aging (NIA) Ag ing Research Dissertation Award to Increase Diversity (Grant #1R36AG029664-01) to fund the present dissertation res earch. In June 2008, I completed a pre-doctoral clinical internship in clinical neurop sychology at the University of Chicago. In addition, I will earn a Doctor of Philosophy degree in clinical psychology, with a specialty in neuropsychology and a Graduate Certificate in gerontology in August 2008 from the University of Florida. I will begin a post-doc toral fellowship in ger opsychology and geriatric rehabilitation at Rush University Medical Cent er in July 2008, and I look forward to a career studying the influence of heath and disease on co gnitive aging in racial/ethnic minority elders.