Content uploaded by Willem Bossers
Author content
All content in this area was uploaded by Willem Bossers on Aug 10, 2018
Content may be subject to copyright.
A 9-Week Aerobic and Strength Training
Program Improves Cognitive and Motor
Function in Patients with Dementia:
A Randomized, Controlled Trial
Willem J.R. Bossers, Ph.D., Lucas H.V. van der Woude, Ph.D.,
Froukje Boersma, M.D., Ph.D., Tibor Hortobágyi, Ph.D., Erik J.A. Scherder, Ph.D.,
Marieke J.G. van Heuvelen, Ph.D.
Objective: To compare training and follow-up effects of combined aerobic and strength
training versus aerobic-only training on cognitive and motor function in institutional-
ized patients with dementia and to explore whether improved motor function mediates
improved cognitive function. Methods: Using a 9-week, parallel, three-group, single-
blind, randomized, controlled trial with a follow-up assessment at week 18, we assessed
109 patients with dementia (age 85.5 5.1 years) in a psycho-geriatric nursing home.
Each 9-week intervention consisted of 36, 30-minute sessions. A combined group (N ¼
37) received and completed two strength and two walking sessions per week, an aerobic
group (N ¼36) completed four walking sessions, and a social group (N ¼36)
completed four social visits per week. Cognitive and motor functions were assessed at
baseline, after the 9-week intervention, and after a consecutive 9 weeks of usual care.
Results: Baseline corrected post-test scores in the combined versus the social group were
higher for global cognition, visual memory, verbal memory, executive function, walking
endurance, leg muscle strength, and balance. Aerobic versus social group scores were
higher for executive function. Follow-up effects reversed toward baseline values. Motor
improvement did not significantly mediate cognitive improvement. Conclusion:
Compared with a nonexercise control group, a combination of aerobic and strength
training is more effective than aerobic-only training in slowing cognitive and motor
decline in patients with dementia. No mediating effects between improvements in cogni-
tive function via improved motor function were found. Future research into the under-
lying mechanistic associations is needed. (Am J Geriatr Psychiatry 2015; 23:1106e1116)
Key Words: Dementia, aerobic exercise, resistance training, cognition, physical fitness,
mediation
Received May 8, 2014; revised December 17, 2014; accepted December 25, 2014. From the Center for Human Movement Sciences (WJRB,
LHVvdW, TH, EJAS, MJGvH), Center for Rehabilitation (LHVvdW), and Department of General Practice, Elderly Care Medicine (FB),
University of Groningen, University Medical Center Groningen, Groningen, The Netherlands; and the Department of Clinical Neuropsy-
chology (EJAS), VU University, Amsterdam, The Netherlands. Send correspondence and reprint requests to Willem J.R. Bossers, Ph.D.,
UMCG, P.O. Box 196, 9700 AD, Groningen, The Netherlands. e-mail: w.j.r.bossers@umcg.nl
Supplemental digital content is available for this article in the HTML and PDF versions of this article on the journal’s Web site (www.
ajgponline.org).
Ó2015 American Association for Geriatric Psychiatry
http://dx.doi.org/10.1016/j.jagp.2014.12.191
1106 Am J Geriatr Psychiatry 23:11, November 2015
REGULAR RESEARCH ARTICLES
INTRODUCTION
Dementia is associated with a decline in cognitive
and motor function that in turn necessitates
increasing daily assistance and care. As a care-
intensive condition, dementia is expected to exact a
severe social and financial impact, making it a public
health priority.
1
To date there is no cure for dementia. Pharmaco-
logic treatments attempt to slow the decline in key
elements of cognitive functions, including memory
and executive function. However, the clinical and
cost-effectiveness of these pharmacologic treatments
are controversial, and the medications may cause
side effects.
2
Moreover, current medications fail to
counteract the motor decline associated with de-
mentia, as illustrated by the continued loss of
endurance, muscle strength, mobility, and balance.
3
Therefore, there is an urgent need for affordable
alternative treatments to combat both the cognitive
and motor decline of patients with dementia, pref-
erably with fewer or no side effects.
In light of the recent recognition by the American
College of Sports Medicine that “exercise is medi-
cine,”exercise is suggested to be a potential treat-
ment for slowing a dementia patient’s decline in
cognitive and motor function.
4,5
However, potential
cognitive benefits of exercise treatment should be
interpreted with caution because the results have
been inconsistent.
4,6
Small to moderate improve-
ments in mixed cognitive domains (e.g., communi-
cation, global cognition, executive function, memory)
were found after mainly aerobic exercise (i.e.,
walking, cycling),
7e10
whereas similar aerobic exer-
cise studies in patients with dementia found no ef-
fects.
11,12
In studies consisting of healthy older
people, a combination of aerobic and strength
training showed the strongest effects in improving
cognitive function compared with single-component
aerobic or strength interventions.
13
Therefore, we
suggest that after combined training these beneficial
effects might also apply to patients with dementia.
However, the effects of combined exercise training on
cognition have not yet been studied in patients with
dementia.
Based on previous evidence in healthy older peo-
ple,
13
we hypothesized that combining aerobic and
strength training in patients with dementia will result
in stronger effects on cognitive function than aerobic-
only training. Leg muscle strength training (e.g., m.
quadriceps femoris, m. gastrocnemius, m. biceps
femoris) has the potential to improve motor function,
such as balance and mobility.
14,15
These motor im-
provements may promote higher aerobic training
ability, thereby eliciting higher metabolic and car-
diovascular responses. With these higher exercise
responses, a mediating cascade of neuromotor events
may be initiated.
15e19
A pilot study we previously conducted showed
that alternating aerobic and strength training sessions
is safe and feasible for patients with dementia.
20
The
potentially favorable motor and cognitive responses
in the pilot study prompted us to launch a large-scale
randomized controlled trial. The objectives of the
present study compare the training and follow-up
effects of a combined aerobic and strength versus
an aerobic-only intervention on cognitive and motor
function in patients with dementia. Furthermore, we
examine whether cognitive change is mediated by a
change in motor function.
METHODS
The Medical Ethics Committee of the University
Medical Center Groningen approved the research
protocol. Patients orally agreed to participate and in
conjunction their legal representative gave written
consent.
Design Overview
We compared the effects of three 9-week in-
terventions on cognitive and motor function in a
parallel, three-group, single-blind, randomized,
controlled trial. Institutionalized patients with de-
mentia were assigned to one of the three intervention
groups: combined aerobic and strength exercise
(combined group), aerobic-only exercise (aerobic
group), and social visits (social group). Post-tests
were conducted after the 9-week intervention (time
9 weeks, T1) and follow-up tests 9 weeks thereafter
(time 18 weeks, F1).
Setting and Participants
A three-group study design, alpha ¼0.05,
power ¼0.80, and a medium to large effect size
Am J Geriatr Psychiatry 23:11, November 2015 1107
Bossers et al.
determined that approximately 35 participants per
group were required.
20
Between January 2011 and
May 2013, patients with mild to severe dementia
21
who lived in seven Dutch psychogeriatric nursing
homes were recruited. First, a geriatrician checked
the following initial eligibility criteria: age over 70,
diagnosis of dementia reported in the patient’s
medical file by Team 290 (i.e., Dutch dementia
diagnosis team) or a medical specialist, and absence
of serious health problems that could preclude safe
participation. Next, a trained Human Movement
Sciences (HMS) research assistant tested the patients
for the following additional inclusion criteria: Mini-
Mental State Exam (MMSE) score of at least 9 to no
more than 23 and the ability to perform the timed up
and go test.
22
Those who passed these five criteria
completed a neuropsychological and physical test
battery in the nursing home at baseline (time zero,
T0) administered by an experienced and trained
HMS research assistant.
Randomization and Interventions
After T0, stratification took place according to
gender (male/female), MMSE score (low subgroup,
high subgroup), and location (allocation ratio 1:1:1).
A researcher, unrelated to the study, performed the
randomization procedure.
Patients in each group participated in 36 individ-
ualized sessions. Each 30-minute session consisted of
one patient being guided by 1 of 18 assigned HMS
research assistants. The combined group participated
in two strength sessions (s) and two walking sessions
(w) per week. In accordance with the American
College of Sports Medicine guidelines, strength and
walking sessions were alternated (wesewes).
23
The
aerobic group participated in four walking sessions
(wewewew) and the social group participated in
four social visits each week. An intervention diary
was used to monitor fidelity. A trainer used a rate of
perceived exertion (RPE) score (range: 6e20) to
monitor exercise intensity after each session.
Furthermore, pre- and post-session differences in
manually measured heart rates in seated positions
were measured. Trainer-monitored RPE scores were
used because patient self-reported RPE feedback is
insufficient for monitoring exercise intensity.
24
Our
aim was to offer a moderate to high intensity exercise
program at a rate of perceived exertion 12e15
(e.g., “somewhat hard”to “hard”)
23
and at 50%e85%
of maximum heart rate.
25
Strength Exercises. Strength exercises for the
combined group focused on lower-limb strength-
ening because such large muscle group exercises are
assumed to lead to enhanced positive responses in
gait speed, balance, and mobility.
26e28
The exercises
were as follows: seated knee extension, plantar
flexion through toe raises while holding both hands
of the trainer, hip abduction by moving the straight
leg sideways while standing behind and holding
onto a chair, and hip extension by moving the
straight leg backward while standing behind and
holding onto a chair. Exercise intensity increased
gradually by increasing the number of repetitions
and by affixing weights around the ankles.
To minimize the chance for injury, overload, and
drop-out, all participants started with three sets of
eight repetitions for each leg without weights. When
a participant performed an exercise with ease and
according to protocol, the number of repetitions was
increased to 10 in the next session and to 12 in the
session thereafter. When a participant was able to
correctly perform 12 repetitions without weights, at
an RPE less than 12, a weight of 0.5 kg was attached
to the ankles. After the weight was attached, partic-
ipants performed eight repetitions and progressed as
prescribed above. The trainer increased weights from
0 kg to a maximum of 1.5 kg in 0.5-kg increments. For
plantar flexion the number of repetitions increased in
increments of 2 per session to a maximum of 30
repetitions.
Aerobic Training. The combined group and aero-
bic group performed moderate to high-intensity
walking sessions.
23
Walking sessions usually took
place indoors. However, if the weather permitted and
the nurse gave permission, walking took place out-
doors. If rest was requested, an appropriate rest
period was included in the 30-minute session. The
training intensity was adjusted by varying the dis-
tances the participants walked per session.
Social Program. To control for social and intellec-
tual engagement,
29
each participant in the social
group (N ¼36) received 30-minute one-on-one social
visits at the same frequency as the exercise groups.
The same HMS research assistants who provided the
exercise interventions assessed these sessions.
1108 Am J Geriatr Psychiatry 23:11, November 2015
Effects of Physical Exercise on Cognitive and Motor Function in Dementia Patients
Outcomes and Follow-Up
Assessment of Cognitive Function. A previously
described neuropsychological test battery with the
corresponding references
20
was used to assess
cognitive function. An HMS research assistant, who
was blinded to the treatment conditions,
administered the tests. Global cognition was
measured with the MMSE. Both short- and long-term
verbal memory were measured with the eight-words
test direct recall, eight-words test recognition, and
digit span forward test (from the Wechsler Memory
Scale Revised [WMS-R]). Short- and long-term visual
FIGURE 1. Consolidated Standards of Reporting Trials (CONSORT) flowchart of the study design and patient flow.
Am J Geriatr Psychiatry 23:11, November 2015 1109
Bossers et al.
memory were measured with the visual memory
span forward test (WMS-R), face recognition test
(Rivermead Behavioral Memory Test), and picture
recognition test (Rivermead Behavioral Memory
Test). Executive function was measured using the
visual memory span backward test (WMS-R), digit
span backward test (WMS-R), the Stroop test, verbal
fluency test (animals and professions), picture
completion test (Groningen Intelligence Test), and
trail making test-A.
30
Assessment of Motor Function. Apreviously
described physical test battery with corresponding
references
20
was used to assess motor function. An
HMS research assistant, who was blinded to the
treatment conditions, administered the tests.
Walking endurance was measured with the
TABLE 1. Baseline Characteristics of the Combined Group (N [37), Aerobic Group (N [36), and Social Group (N [36)
Characteristic Combined Group Aerobic Group Social Group F or
c
2
Test Value, (df), p Value
Mean age, (SD) 85.7 (5.1) 85.4 (5.4) 85.4 (5.0) F ¼0.03, (2,106), p ¼0.973
a
Males 21.6% 22.2% 30.6%
c
2
¼0.97, (2), p ¼0.616
b
Dutch education level
c
2
¼5.87, (10), p ¼0.826
b
Finished primary school or lower 16.7% 24.3% 18.5%
Lower than finished higher education 48.6% 45.9% 56.8%
Finished higher education 34.7% 29.8% 24.7%
Use of walking aid 43.2% 63.9% 61.1%
c
2
¼3.73, (2), p ¼0.155
b
Mean MMSE (SD), range
c
15.8 (4.3), 9e23 15.2 (4.8), 9e23 15.9 (4.2), 9e23 F ¼0.24, (2, 106), p ¼0.784
a
Mild dementia (score 20e23) 27.0% 22.2% 22.2%
Moderate dementia (score 11e20) 59.5% 44.5% 61.1%
Severe dementia (score 9e11) 13.5% 33.3% 16.7%
Mean number of medications used (SD) 6.0 (2.4) 6.9 (2.6) 7.4 (2.9) F ¼2.45, (2, 106), p ¼0.091
a
Mean functional Comorbidity Index (SD)
d
3.6 (1.8) 3.2 (1.8) 3.6 (1.8) F ¼2.23, (2, 106), p ¼0.112
a
Mean Katz index (SD)
e
10.5 (1.8) 11.0 (2.6) 10.9 (2.8) F ¼0.55, (2, 106), p ¼0.580
a
Cause of dementia
c
2
¼0.97, (6), p ¼0.360
b
Alzheimer disease 54.1% 58.3% 47.2%
Vascular dementia 13.5% 22.2% 13.9%
Alzheimer disease/vascular dementia 13.5% 16.7% 19.4%
Type not reported 18.9% 2.8% 19.5%
a
Differences between the groups were tested with one-way analysis of variance.
b
Differences between the groups were tested with
c
2
test.
c
Theoretical range 0e30 and a higher score indicates better performance.
d
Theoretical range 0e18 and a higher score indicates more comorbidities.
e
Theoretical range 6e18 and a higher score indicates higher activities of daily living dependency.
TABLE 2. Training Characteristics of the 9-Week Interventions for the Combined Group (N [37), Aerobic Group (N [36), and
Social Group (N [36)
Training Characteristic Combined Group Aerobic Group Social Group F or t Test Value, (df), p Value
Mean % adherence rate 89.2 (9.8) 89.1 (10.6) 93.2 (6.7) F ¼2.34, (2,106), p ¼0.101
a
Mean duration per session, min 30.9 (1.9) 30.9 (2.9) 29.9 (2.0) F ¼2.31, (2,106), p ¼0.104
a
Mean resting heart rate, beats/min
1
73.1 (7.9) 74.3 (7.2) 73.4 (8.0) F ¼2.26, (2,106), p ¼0.775
a
Mean heart rate after exercise, beats/min
1
83.4 (10.4) 86.7 (10.3) 74.2 (8.3) F ¼16.22, (2,106), p ¼<0.001
a
t¼4.22, (71), p <0.001
b
t¼5.73, (70), p <0.001
c
Mean heart rate difference, beats/min
1
10.3 (4.7) 12.3 (6.5) 0.8 (2.0) F ¼59.98, (2,106), p <0.001
a
t¼11.15, (71), p <0.001
b
t¼10.19, (70), p <0.001
c
Mean rate of perceived exertion after session
d
13.2 (2.9) 13.8 (1.7) Not collected t ¼1.11 (71), p ¼0.272
Notes: Values are means with standard deviation in parentheses.
a
Differences between groups were tested with one-way ANCOVA.
b
For post-hoc comparison between the combined and social groups, independent-sample t test was done.
c
For post-hoc comparison between the aerobic and social groups, independent-samples t test was done.
d
Theoretical range 6e20, where a score of 6 indicates the lowest intensity level and a score of 20 indicates the highest intensity level.
1110 Am J Geriatr Psychiatry 23:11, November 2015
Effects of Physical Exercise on Cognitive and Motor Function in Dementia Patients
6-minute walk test. Leg strength was measured with
the 30-second sit-to-stand test, and the maximal
knee extension strength was measured with a
dynamometer. Mobility was measured with the
6-meter walk test and the timed up and go test.
Balance was measured with the Frailty and Injuries
Cooperative Studies of Intervention Techniques-
Subtest 4, the figure of eight test, and the Gronin-
gen Meander Walking Test.
20,31
Statistical Analysis
SPSS Statistics for Windows (version 20.0, IBM,
Armonk, NY) was used for data management with
two-tailed significance set at p <0.05. We corrected
the p values for multiple hypotheses testing by
multiplying the obtained p value by four (i.e., four
cognitive and motor domains, respectively). Group
characteristics, compliance with the exercise pro-
gram, and the intensity measures of the programs
were compared between the three groups with one-
way analyses of variance, independent sample t
tests, and
c
2
tests.
Multiple imputation was used to account for
missing values (8.9% of the values missing; 5.2% at
T0, 6.4% at T1, and 14.6% at F1). Characteristic var-
iables of the sample and cognitive and physical test
scores set at T0, T1, and F1 were included in the
imputation model. The following imputation settings
were used: automatic model setting, 100 iterations,
and 70 imputations to preserve 99.5% of the relative
estimation of the imputed dataset.
Mean z scores of composite factors for cognitive
domains (i.e., global cognitive function, verbal
memory, visual memory, and executive function) and
motor domains (i.e., walking endurance, leg strength,
balance, and mobility) were calculated. Analyses of
covariance (ANCOVAs) were performed to deter-
mine group effects in each domain after the 9-week
intervention and after the 18-week total observation
period. Baseline domain scores were used as cova-
riates. To specify significant group effects, ANCOVA
Bonferroni corrected post-hoc tests were done.
For each cognitive and motor domain, mean
Cohen’s d effect sizes were calculated by using a
reference group (i.e., combined versus social; aerobic
versus social; combined versus aerobic). The
following formula was used: d ¼[(post
exp.
pre
exp.
)(post
ref.
pre
ref.
)]/O[([s
2
pre
exp.
(n
exp.
)þ
s
2
pre
ref.
(n
ref.
)]/[n
exp.
þn
ref.
]) þ([s
2
post
exp.
(n
exp.
)þ
s
2
post
ref.
(n
ref.
)]/[n
exp.
þn
ref.
])/2],
32
where post is the
mean post-test, pre is the mean pretest, s is the
standard deviation, exp. is the experimental group,
and ref. is the reference group. Cohen’s benchmarks
were used to indicate small (d ¼0.20), moderate (d ¼
0.50), and large (d ¼0.80) effect sizes.
32
To explore mediation of motor effects on cognitive
effects, the Hayes and Preacher mediation analysis
was completed.
33
The difference scores in global
cognition, visual memory, verbal memory, and ex-
ecutive function were the outcome variables. Differ-
ence scores in walking endurance, leg muscle
strength, balance, and mobility were the mediators in
motor domain. Dummy coding was used to test the
direct and indirect (i.e., via mediators) effects of the
combined and aerobic group using the social group
as a reference. A bootstrapping method with 5000
resamples was used.
33
RESULTS
Participant Characteristics and Compliance with
the Intervention
Figure 1 shows the study design and patient flow.
In total, 495 patients were screened for eligibility, and
132 were enrolled in the study. There were no sig-
nificant between-group differences for descriptive
characteristics, and the severity of dementia ranged
between mild to severe dementia (Table 1).
21
Table 2
presents the training characteristics for each group.
After exercise the mean heart rate increased more in
the combined and aerobic groups than in the social
group (Table 2).
Intervention Effects
Results on the individual neuropsychological tests
per group and time point are presented in Appendix
1(available online). On the cognitive domain level,
Table 3 shows a significant intervention effect for
global cognitive function, visual memory, verbal
memory, and executive function. ANCOVA Bonfer-
roni corrected post-hoc tests revealed that the com-
bined group (I) improved compared with the social
group (J) on global cognitive function (mean T1eT0
difference (I-J) ¼0.430; 95% confidence interval [CI]:
0.176e0.685; t(71) ¼4.12, p <0.001), visual memory
Am J Geriatr Psychiatry 23:11, November 2015 1111
Bossers et al.
(mean T1eT0 difference (I-J) ¼0.527; 95% CI:
0.231e0.823; t(71) ¼4.17, p <0.001), verbal memory
(mean difference (I-J) ¼0.353; 95% CI: 0.078e0.628;
t(71) ¼3.10, p ¼0.003), and executive function (mean
T1eT0 difference (I-J) ¼0.305; 95% CI: 0.129e0.482;
t(71) ¼4.10, p <0.001). Furthermore, post-hoc com-
parisons revealed that the aerobic (I) versus social
group (J) only differed in favor of the aerobic group
in executive function (mean T1eT0 difference (I-J) ¼
0.183; 95% CI: 0.005e0.360; t(70) ¼2.37, p ¼0.021).
After the 9-week delayed post-test, no significant ef-
fects of the intervention group in any of the cognitive
domains remained (Table 3).
Results on the individual motor tests per group and
time point are presented in Appendix 2 (available on-
line). On the motor domain level, Table 4 shows a
significant intervention effect for walking endurance,
leg muscle strength, and balance. ANCOVA Bonfer-
roni corrected post-hoc tests revealed that the com-
bined group (I) improved compared with the social
group (J) on walking endurance (mean T1eT0 differ-
ence (I-J) ¼0.578; 95% CI: 0.211e0.935; t(71) ¼3.01, p ¼
0.004), leg muscle strength (mean T1eT0 difference
(I-J) ¼0.358; 95% CI: 0.067e0.649; t(71) ¼4.41,
p<0.001), and balance (mean T1eT0 difference (I-J) ¼
0.329; 95% CI: 0.100e0.558; t(71) ¼3.27, p ¼0.002).
Furthermore, post-hoc comparisons revealed that the
combined group (I) scored higher than the aerobic
group (J) in walking endurance (mean T1eT0 differ-
ence (I-J) ¼0.363; 95% CI: 0.006e0.719; t(71) ¼3.00, p ¼
0.004) and leg muscle strength (mean T1eT0 difference
(I-J) ¼0.359; 95% CI: 0.068e0.650; t(71) ¼3.41, p ¼
0.001). After the 9-week delayed post-test, no signifi-
cant intervention effects in any of the motor domains
remained (Table 4).
Mediating Effect of Motor Domain Difference
Scores on Cognitive Domain Difference Scores
A mediation effect equals the multiplication of the
direct effect of each intervention group on motor
function (i.e., “path a”) by the direct effect of each
motor function on cognitive function (i.e., “path b”)
represented as “ab”.
33
In line with results presented
in Table 4, the aerobic group showed no significant
direct effects on path a. However, in the combined
group, significant path a effects on walking endur-
ance (B ¼0.69, t(36) ¼2.13, p ¼0.040), leg strength
(B ¼0.50, t(36) ¼2.18, p ¼0.036), and balance
(B ¼0.43, t(36) ¼2.40, p ¼0.021) were found. Despite
this result, no significant path b effects were found.
Subsequently, mediating effects path ab in both
intervention groups were nonsignificant.
DISCUSSION
This is the first study that provides evidence for the
effectiveness of a combined aerobic and strength
training program to improve cognitive and motor
function in older patients with dementia. Cognitive
and motor function improved after 9 weeks of
training. Moreover, an alternating form of aerobic
and strength training sessions were more effective
than aerobic-only training.
Patients with dementia suffer from a continuous
loss in global cognition.
34
Previous studies show an
average 12-month global cognitive decline of 2.3
MMSE points.
34
However, our study results demon-
strate that the combined and aerobic group improved
by 2.1 and 1.0 MMSE points after the intervention,
respectively. These values were computed by
adjusting for the natural decline of 0.72 MMSE points
in the social group. These short-term improvements
suggest that the combined and aerobic group
compensated, respectively, 11 (1.35 to 0.72 ¼2.1
MMSE points) and 5 months (0.28 to 0.72 ¼1.0
MMSE points) of the average 12-month 2.3 MMSE
deterioration,
34
which is of high clinical relevance.
Thus, the current data suggest that a combined aer-
obic and strength training slows the natural global
cognitive decline in patients with dementia and is
more effective than aerobic-only training.
In addition to global cognitive benefits, patients
who received the combined exercise program also
improved their verbal memory, visual memory, and
executive function. This is in line with previous
findings in healthy elderly people.
5,13
In our previous
6-week pilot study, we showed only a moderate
nonsignificant effect size in visual memory,
20
whereas the current longer training period resulted
in larger significant cognitive effect size in all cogni-
tive domains. These larger significant effects can be
explained by the longer training duration and larger
sample size, which result in higher statistical power.
Improvements in verbal memory, visual memory,
and executive function have been previously linked
to brain circuitry in the posterior and anterior
1112 Am J Geriatr Psychiatry 23:11, November 2015
Effects of Physical Exercise on Cognitive and Motor Function in Dementia Patients
hippocampi and the frontal lobes.
35
These circuits are
vulnerable in normal aging and critically involved in
dementia progression.
5,36
In the current study, the 9-
week combined aerobic and strength training strat-
egy may have induced changes in different brain
regions (i.e., frontal, hippocampal) that led to cogni-
tive improvement; however, this cannot be ascer-
tained because brain imaging measures were not
collected. A recent study in older women with
probable mild cognitive impairment showed
increased hippocampal volume after a 6-month aer-
obic training program.
37
Based on these findings,
future studies in patients with dementia should
collect brain imaging data to visualize whether
cognitive changes are elicited by changes in brain
activity, connectivity, and/or composition.
Our aerobic-only training improved cognitive func-
tion, because there were significant increases in execu-
tive function. However, aerobic-only training did not
significantly improve motor function. Surprisingly, the
improved executive function after aerobic-only training
contradicts the results presented in a review by
Scherder et al.,
5
which showed that aerobic-only
training did not improve executive function in pa-
tients with dementia and had only an effect in cogni-
tively nonimpaired older people. A possible
explanation could be the high level of participation,
combined with the one-on-one training program in the
current study.
38
Although aerobic-only training
improved executive function, it did not improve
memory function. Improvements in both executive and
memory function were reached after combining aerobic
with strength training, as was done in the combined
group. These findings are in line with a recent 16-week
nonrandomized controlled multimodal exercise study
in patients with Alzheimer disease.
39
Furthermore,
compared with the nonexercise social group, the com-
bined group scored higher than the aerobic group on
walking endurance, leg muscle strength, and balance
after the 9-week program. These findings agree with
our pilot study, which showed medium to large effect
sizes on motor function after a 6-week combined aer-
obic and strength training program.
20
Taken together,
aerobic-only training may not be sufficient to elicit both
improvements in cognitive (i.e., executive and memory
function) and motor function. Therefore, we suggest
combining aerobic and strength training.
In a recent strength exercise study in healthy older
people, a potential mediating effect between improved
TABLE 3. Cognitive Data Representing Group Effects and Effect Size for the 9-Week Intervention Period (T0eT1) and for the Total 18-Week Period (T0eF1), Including
Post-Hoc Group Comparisons
Cognitive Domains
a
F Value,
(df), p
b
Mean Effect Size:
Combined
vs. Social (range)
Mean Effect Size:
Aerobic
vs. Social (range)
Mean Effect Size:
Combined
vs. Aerobic (range)
F Value,
(df), p
c
Mean Effect Size:
Combined
vs. Social (range)
Mean Effect Size:
Aerobic
vs. Social (range)
Mean Effect Size:
Combined
vs. Aerobic (range)
Global cognition F ¼8.46, (2,105),
p<0.001
d
0.48 (-) 0.21 (-) 0.23 (-) F ¼3.96, (2,105),
p¼0.088
0.45 (-) 0.28 (-) 0.15 (-)
Visual memory F ¼6.72, (2,105),
p¼0.008
d
0.46 (0.35e0.61) 0.26 (0.05e0.62) 0.20 (0.40e0.32) F ¼1.83, (2,105),
p¼0.660
0.26 (0.17e0.32) 0.14 (0.06e0.27) 0.12 (0.05e0.39)
Verbal memory F ¼5.10, (2,105),
p¼0.032
d
0.37 (0.32e0.43) 0.29 (0.15e0.39) 0.06 (0.00e0.15) F ¼3.23, (2,105),
p¼0.176
0.34 (0.17e0.56) 0.28 (0.04e0.52) 0.28 (0.05e0.22)
Executive function F ¼9.26, (2,105),
p<0.001
d,e
0.37 (0.13e0.51) 0.17 (0.04e0.33) 0.18 (0.05e0.30) F ¼3.31, (2,105),
p¼0.160
0.23 (0.12e0.31) 0.04 (0.08e0.11) 0.17 (0.06e0.38)
a
Composite factors for cognitive function domains global cognition, visual memory, verbal memory, and executive function were calculated by averaging z values of test scores
that measured these domains (see Appendix 1 for separate test values; available online).
b
Model used for data analysis: ANCOVA with T0 as covariate, T1 as dependent variable, and group as a fixed factor.
c
Model used for data analysis: ANCOVA with T0 as covariate, T2 as dependent variable, and group as a fixed factor.
d
Significant ANCOVA Bonferroni corrected post-hoc test between the combined and social group (see text for statistics).
e
Significant ANCOVA Bonferroni corrected post-hoc test between the aerobic and social group (see text for statistics).
Am J Geriatr Psychiatry 23:11, November 2015 1113
Bossers et al.
muscle strength on inhibitory processes was found.
40
However, in the current study in patients with de-
mentia, we did not find such a mediating effect be-
tween improved motor and cognitive function. There
may be other possible underlying mechanisms
involved that explain how physical exercise is related
to cognition in patients with dementia. Previous
research suggests that complementary neurobiologic
and physiologic mechanisms may arise from alter-
nating aerobic and strength training sessions,
compared with those arising from single training
programs. Such a combined exercise program could
conceptually be more effective for slowing the
dementia-related cognitive deterioration. Both aerobic
and strength training can favorably influence brain-
derived neurotropic factors and levels of insulin-like
growth factor 1, which mediate cell growth, prolifer-
ation, survival, and differentiation.
16
Strength training
may lower levels of neurotoxic homocysteine, which is
related to improved cognition.
17,19
The improvements
in motor function after a combination of aerobic and
strength training, as seen in the present study, have
the potential to increase angiogenesis via vascular
endothelial growth factor, thereby increasing cerebral
blood flow, which is a key factor related to cognitive
function.
15,19
The current study lacks imaging, blood,
and more comprehensive cardiovascular data that
could support such a mechanistic understanding as
described above. Future studies should include these
outcome measures.
In addition to the mechanistic exerciseecognition
discussion above, another possible explanation why
the combined intervention produced higher cognitive
and motor scores could be related to a stronger
neuromotor stimulus after the strength exercises.
These exercises required the patient to perform
complex cognitive tasks, such as comprehending the
instructions, moving the limbs in the correct order,
and mimicking the trainer’s movements. In addition,
strength exercises incorporate motor coordination
and balance tasks, known to activate specific
cerebellarecortical connections, which can act as a
stimulus for concurrent improvements in cognitive
function and balance.
41
Such responses were
demonstrated after Tai-Chi training in older persons
with dementia.
42,43
In the current study no data were
collected to confirm the cerebellarecortical hypothe-
sis, and future studies in patients with dementia are
needed.
TABLE 4. Motor Data Representing Group Effects and Effect Size for the 9-Week Intervention Period (T0eT1) and for the Total 18-Week period (T0eF1), Including Post-
Hoc Group Comparisons
Motor Domain
a
F Value,
(df), p
b
Mean Effect Size:
Combined
vs. Social (range)
Mean Effect Size:
Aerobic
vs. Social (range)
Mean Effect Size:
Combined
vs. Aerobic (range)
F Value,
(df), p
c
Mean Effect Size:
Combined
vs. Social (range)
Mean Effect Size:
Aerobic
vs. Social (range)
Mean Effect Size:
Combined
vs. Aerobic (range)
Walking endurance F ¼4.53, (2,105),
p<0.049
d,e,f
0.47 (-) 0.08 (-) 0.38 (-) F ¼1.23, (2,105),
p¼0.296
0.19 (-) 0.06 (-) 0.27 (-)
Leg muscle strength F ¼7.07, (2,105),
p¼0.004
d,e,f
0.38 (0.26e0.49) 0.04 (0.00e0.07) 0.36 (0.15e0.56) F ¼2.86, (2,105),
p¼0.247
0.27 (0.20e0.33) 0.25 (0.14e0.36) 0.01 (0.15e0.17)
Mobility F ¼1.28, (2,105),
p¼0.282
0.28 (0.18e0.39) 0.06 (0.03e0.12) 0.26 (0.17e0.33) F ¼2.36, (2,105),
p¼0.398
0.29 (0.12e0.41) 0.00 (0.03e0.04) 0.34 (0.20e0.44)
Balance F ¼5.36, (2,105),
p¼0.024
d,e,f
0.30 (0.14e0.65) 0.08 (0.06e0.16) 0.33 (0.10e0.72) F ¼2.86, (2,105),
p¼0.247
0.24 (0.02e0.54) 0.19 (0.43e0.00) 0.48 (0.23e0.67)
a
Composite factors for motor function domains walking endurance, leg muscle strength, mobility, and balance were calculated by averaging z values of test scores that measured
these domains (see Appendix 2 for separate test values; available online).
b
Model used for data analysis: ANCOVA with T0 as covariate, T1 as dependent variable, and group as a fixed factor.
c
Model used for data analysis: ANCOVA with T0 as covariate, T2 as dependent variable, and group as a fixed factor.
d
p<0.05 (Bonferroni corrected for multiple hypothesis testing).
e
Significant ANCOVA Bonferroni corrected post-hoc test between the combined and social groups (see text for statistics).
f
Significant ANCOVA Bonferroni corrected post-hoc test between the combined and aerobic groups (see text for statistics).
1114 Am J Geriatr Psychiatry 23:11, November 2015
Effects of Physical Exercise on Cognitive and Motor Function in Dementia Patients
Nine weeks after the intervention was ended,
follow-up effects reversed toward baseline values.
Thus, the current data stress that a structural exercise
program is essential to maintain the demonstrated
positive effects of regular exercise. Long-term main-
tenance of physical exercise and activities could
potentially slow disease progression, which is of high
clinical relevance (i.e., “exercise is medicine”).
Limitations
We were limited in collecting and using data of
patients who dropped out before the start of the
intervention. Also, only a specific segment of insti-
tutionalized patients with dementia participated
(e.g., those who were mobile and motivated enough).
When generalizing the results, both factors should be
taken into account.
This is the first study that used trainer-monitored
RPE to determine exercise intensity in patients with
dementia. Indeed, we were confronted with the fact
that no clinimetric data on external RPE in patients
with dementia is available. However, a significant
correlation between RPE and heart rate data in the
exercise groups was found (Spearman’sr¼0.314,
N¼73, p ¼0.007), indicating sessions that led to a
larger difference in heart rate were scored with a
higher RPE. More research is needed in measuring
exercise intensity in patients with dementia to eval-
uate the doseeresponse relationship between exer-
cise and cognition.
CONCLUSION
We conclude that compared with a nonexercise
control group, a combination of aerobic and
strength training elicits stronger effects than the
aerobic-only training in slowing the cognitive and
motor function decline in institutionalized patients
with dementia. However, we did not confirm that
improved motor function mediates improved
cognitive function. Future research is needed to
study the underlying mechanisms that are
involved in the link between exercise and the
cognitive and motor deterioration processes in
dementia.
Healthcare institutions ZINN, Zorggroep Gronin-
gen, Dignis-Lentis, participants and their families, and
research assistants are acknowledged. The authors thank
professors Eddy van der Zee, Paul Luiten, and Wiebo
Brouwerfortheiradvice.TheTechnicalDepartmentof
the Center for Human Movement Sciences, University
Medical Center Groningen is acknowledged. Finally,
Garett Vanderveen is acknowledged for language
revision.
Fonds NutsOhra provided funding for the execution
of this study (project number 1003-76) and had no role in
study design, data collection, analysis, decision to publish,
or manuscript preparation. The authors declare no con-
flicts of interest.
References
1. World Health Organization: Dementia: A Public Health Priority.
Geneva, Switzerland, WHO Press, 2012
2. Bond M, Rogers G, Peters J, et al: The effectiveness and cost-
effectiveness of donepezil, galantamine, rivastigmine and
memantine for the treatment of Alzheimer’s disease (review of
technology appraisal no. 111): a systematic review and eco-
nomic model. Health Technol Assess 2012; 16:1e470
3. Landi F, Russo A, Liperoti R, et al: Anticholinergic drugs and
physical function among frail elderly population. Clin Pharmacol
Ther 2007; 81:235e241
4. Forbes D, Thiessen EJ, Blake CM, et al: Exercise programs for
people with dementia. Cochrane Database Syst Rev 2013; 12:
CD006489
5. Scherder E, Scherder R, Verburgh L, et al: Executive functions of
sedentary elderly may benefit from walking: a systematic review
and meta-analysis. Am J Geriatr Psychiatry 2013; 22:782e791
6. Scarmeas N, Luchsinger JA, Brickman AM, et al: Physical activity
and Alzheimer disease course. Am J Geriatr Psychiatry 2011; 19:
471e481
7. Friedman R, Tappen RM: The effect of planned walking on
communication in Alzheimer’s disease. J Am Geriatr Soc 1991;
39:650e654
8. Palleschi L, Vetta F, De Gennaro E, et al: Effect of aerobic training
on the cognitive performance of elderly patients with senile de-
mentia of Alzheimer type. Arch Gerontol Geriatr 1996; 22(suppl
1):47e50
9. Venturelli M, Scarsini R, Schena F: Six-month walking program
changes cognitive and ADL performance in patients with Alz-
heimer. Am J Alzheim Dis Other Demen 2011; 26:381e388
10. Kwak YS, Um SY, Son TG, et al: Effect of regular exercise on
senile dementia patients. Int J Sports Med 2008; 29:471e474
11. Cott CA, Dawson P, Sidani S, et al: The effects of a walking/
talking program on communication, ambulation, and functional
status in residents with Alzheimer disease. Alzheimer Dis Assoc
Disord 2002; 16:81e87
12. Eggermont L, Swaab D, Hol E, et al: Walking the line: a randomised
trial on the effects of a short term walking programme on cogni-
tion in dementia. J Neurol Neurosurg Psychiatry 2009; 80:802e804
Am J Geriatr Psychiatry 23:11, November 2015 1115
Bossers et al.
13. Colcombe S, Kramer AF: Fitness effects on the cognitive function
of older adults: a meta-analytic study. Psychol Sci 2003; 14:
125e130
14. LaRoche DP, Millett ED, Kralian RJ: Low strength is related to
diminished ground reaction forces and walking performance in
older women. Gait Posture 2011; 33:668e672
15. Wiesmann M, Kiliaan AJ, Claassen JA: Vascular aspects of cogni-
tive impairment and dementia. J Cereb Blood Flow Metab 2013;
33:1696e1706
16. Huang T, Larsen KT, Ried-Larsen M, et al: The effects of physical
activity and exercise on brain-derived neurotrophic factor in
healthy humans: a review. Scand J Med Sci Sports 2014; 24:1e10
17. Kim G, Kim H, Kim KN, et al: Relationship of cognitive function
with B vitamin status, homocysteine, and tissue factor pathway
inhibitor in cognitively impaired elderly: a cross-sectional survey.
J Alzheim Dis 2013; 33:853e862
18. Liu-Ambrose T, Donaldson MG: Exercise and cognition in older
adults: is there a role for resistance training programmes? Br J
Sports Med 2009; 43:25e27
19. Nation DA, Hong S, Jak AJ, et al: Stress, exercise, and Alzheimer’s
disease: a neurovascular pathway. Med Hypotheses 2011; 76:
847e854
20. Bossers WJ, Scherder EJ, Boersma F, et al: Feasibility of a com-
bined aerobic and strength training program and its effects on
cognitive and physical function in institutionalized dementia pa-
tients. A pilot study. PLoS One 2014; 9:e97577
21. Perneczky R, Wagenpfeil S, Komossa K, et al: Mapping scores
onto stages: Mini-mental state examination and clinical dementia
rating. Am J Geriatr Psychiatry 2006; 14:139e144
22. Podsiadlo D, Richardson S: The timed “up & go”: a test of basic
functional mobility for frail elderly persons. J Am Geriatr Soc
1991; 39:142e148
23. Nelson ME, Rejeski WJ, Blair SN, et al: Physical activity and public
health in older adults: recommendation from the American Col-
lege of Sports Medicine and the American Heart Association. Med
Sci Sports Exerc 2007; 39:1435e1445
24. Yu F, Bil K: Correlating heart rate and perceived exertion during
aerobic exercise in Alzheimer’s disease. Nurs Health Sci 2010; 12:
375e380
25. American Heart Association. Target Heart Rates (online). Avail-
able at: http://www.heart.org/HEARTORG/GettingHealthy/
PhysicalActivity/FitnessBasics/Target-Heart-Rates_UCM_434341_
Article.jsp. Accessed January 22, 2015
26. Krist L, Dimeo F, Keil T: Can progressive resistance training twice
a week improve mobility, muscle strength, and quality of life in
very elderly nursing-home residents with impaired mobility? A
pilot study. Clin Interv Aging 2013; 8:443e448
27. Beijersbergen CM, Granacher U, Vandervoort AA, et al: The
biomechanical mechanism of how strength and power training
improves walking speed in old adults remains unknown. Ageing
Res Rev 2013; 12:618e627
28. Chandler JM, Duncan PW, Kochersberger G, et al: Is lower ex-
tremity strength gain associated with improvement in physical
performance and disability in frail, community-dwelling elders?
Arch Phys Med Rehabil 1998; 79:24e30
29. Bassuk SS, Glass TA, Berkman LF: Social disengagement and
incident cognitive decline in community-dwelling elderly per-
sons. Ann Intern Med 1999; 131:165e173
30. Tombaugh TN: Trail making test A and B: normative data strati-
fied by age and education. Acrh Clin Neuropsychol 2004; 19:
203e214
31. Bossers WJ, van der Woude LH, Boersma F, et al: The Groningen
Meander Walking Test: a dynamic walking test for older adults
with dementia. Phys Ther 2014; 94:262e272
32. Cohen J: Statistical Power Analysis for the Behavioural Sciences.
2nd Ed. Hillsdale, NJ, Lawrence Erlbaum, 1988
33. Hayes AF, Preacher KJ: Statistical mediation analysis with a mul-
ticategorical independent variable. Br J Math Stat Psychol 2013;
67:451e470
34. Rascovsky K, Salmon DP, Lipton AM, et al: Rate of progression
differs in frontotemporal dementia and Alzheimer disease.
Neurology 2005; 65:397e403
35. Erickson KI, Voss MW, Prakash RS, et al: Exercise training in-
creases size of hippocampus and improves memory. Proc Natl
Acad Sci USA 2011; 108:3017e3022
36. Fjell AM, McEvoy L, Holland D, et al: What is normal in normal
aging? Effects of aging, amyloid and Alzheimer’s disease on the
cerebral cortex and the hippocampus. Prog Neurobiol 2014; 117:
20e40
37. Ten Brinke LF, Bolandzadeh N, Nagamatsu LS, et al: Aerobic ex-
ercise increases hippocampal volume in older women with
probable mild cognitive impairment: a 6-month randomised
controlled trial [Epub ahead of print April 7, 2014]. Br J Sports
Med
38. Brodaty H, Burns K: Nonpharmacological management of apathy
in dementia: a systematic review. Am J Geriatr Psychiatry 2012;
20:549e564
39. de Andrade LP, Gobbi LT, Coelho FG, et al: Benefits of multi-
modal exercise intervention for postural control and frontal
cognitive functions in individuals with Alzheimer’s disease: a
controlled trial. J Am Geriatr Soc 2013; 61:1919e1926
40. Forte R, Boreham CA, Leite JC, et al: Enhancing cognitive func-
tioning in the elderly: multicomponent vs resistance training. Clin
Interv Aging 2013; 8:19e27
41. D’Angelo E: Neural circuits of the cerebellum: hypothesis for
function. J Integr Neurosci 2011; 10:317e352
42. Cheng ST, Chow PK, Song YQ, et al: Mental and physical activ-
ities delay cognitive decline in older persons with dementia. Am J
Geriatr Psychiatry 2014; 22:63e74
43. Wayne PM, Walsh JN, Taylor-Piliae RE, et al: Effect of tai chi on
cognitive performance in older adults: systematic review and
meta-analysis. J Am Geriatr Soc 2014; 62:25e39
1116 Am J Geriatr Psychiatry 23:11, November 2015
Effects of Physical Exercise on Cognitive and Motor Function in Dementia Patients