You are on page 1of 17

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/263935402

The Effects of Poststroke Aerobic Exercise on Neuroplasticity: A Systematic


Review of Animal and Clinical Studies

Article  in  Translational Stroke Research · July 2014


DOI: 10.1007/s12975-014-0357-7 · Source: PubMed

CITATIONS READS

50 1,339

4 authors:

Michelle Ploughman Mark W Austin


Memorial University of Newfoundland Memorial University of Newfoundland
87 PUBLICATIONS   1,482 CITATIONS    9 PUBLICATIONS   229 CITATIONS   

SEE PROFILE SEE PROFILE

Lindsay Alcock Dale Corbett


Memorial University of Newfoundland University of Ottawa
40 PUBLICATIONS   412 CITATIONS    195 PUBLICATIONS   9,697 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Using transcranial magnetic stimulation to detect cognitive, motor, and fatigue outcomes in multiple sclerosis: A systematic review. View project

Intracerebral hemorrhage View project

All content following this page was uploaded by Michelle Ploughman on 17 July 2014.

The user has requested enhancement of the downloaded file.


Transl. Stroke Res.
DOI 10.1007/s12975-014-0357-7

REVIEW ARTICLE

The Effects of Poststroke Aerobic Exercise on Neuroplasticity: A


Systematic Review of Animal and Clinical Studies
Michelle Ploughman & Mark W. Austin & Lindsay Glynn &
Dale Corbett

Received: 16 April 2014 / Revised: 2 June 2014 / Accepted: 30 June 2014


# Springer Science+Business Media New York 2014

Abstract Aerobic exercise may be a catalyst to promote Keywords Aerobic exercise . Neurotrophins . BDNF .
neuroplasticity and recovery following stroke; however, the Neuroplasticity . Rehabilitation . Stroke . Cerebrovascular .
optimal methods to measure neuroplasticity and the effects of Review
training parameters have not been fully elucidated. We con-
ducted a systematic review and synthesis of clinical trials and Abbreviations
studies in animal models to determine (1) the extent to which AE Aerobic exercise
aerobic exercise influences poststroke markers of BDNF Brain-derived neurotrophic factor
neuroplasticity, (2) the optimal parameters of exercise re- EEG Electroencephalography
quired to induce beneficial effects, and (3) consistent out- FE Forced exercise
comes in animal models that could help inform the design of fMRI Functional magnetic resonance imaging
future trials. Synthesized findings show that forced exercise at GAP43 Growth-associated protein-43
moderate to high intensity increases brain-derived neurotroph- IGF-I Insulin-like growth factor-I
ic factor (BDNF), insulin-like growth factor-I (IGF-I), nerve NGF Nerve growth factor
growth factor (NGF), and synaptogenesis in multiple brain MAP2 Microtubule-associated protein 2
regions. Dendritic branching was most responsive to moderate MEG Magnetoencephalography
rather than intense training. Disparity between clinical stroke PET Positron emission tomography
and stroke models (timing of initiation of exercise, age, gen- VE Voluntary exercise
der) and clinically viable methods to measure neuroplasticity VEGF Vascular endothelial growth factor
are some of the areas that should be addressed in future
research.

Introduction
M. Ploughman (*) : M. W. Austin
Recovery and Performance Laboratory, Rehabilitation Research With few treatment options available, stroke is the leading
Unit, Faculty of Medicine, Memorial University, Rm 400, L.A. cause of adult neurological disability, affecting one in six
Miller Centre, 100 Forest Rd., St. John’s, NL, Canada
e-mail: michelle.ploughman@med.mun.ca
people worldwide [1, 2]. Although rehabilitative interventions
improve outcome, full recovery is often incomplete [3]. In
L. Glynn animal models, neurotrophins such as brain-derived neuro-
Health Sciences Library, Memorial University of Newfoundland, St. trophic factor (BDNF), insulin-like growth factor-I (IGF-I),
John’s, NL, Canada
and nerve growth factor (NGF) are dramatically upregulated
D. Corbett during the first 4 weeks after stroke [4], creating a
Department of Cellular and Molecular Medicine, University of “neuroplastic milieu” in which cortical remodeling within
Ottawa, Ottawa, ON, Canada the intact brain is optimized [5]. BDNF belongs to a group
of secreted proteins that mediate activity-dependent processes
D. Corbett
Heart and Stroke Foundation Canadian Partnership for Stroke such as synapse formation and neuronal growth [6, 7]. In-
Recovery, Ottawa, ON, Canada creased IGF-I is associated with better neurological recovery
Transl. Stroke Res.

[8], while NGF promotes the growth and survival of sensory Methods
and sympathetic neurons: processes that are likely beneficial
to poststroke plasticity and recovery [9]. Animal models of Using PRISMA guidelines [26], a search of English language
stroke have elucidated the postischemic molecular events in articles (up to December 2013) in CINAHL, PubMed,
which growth of new synapses (synaptogenesis) and dendritic PsychInfo, and the Cochrane Library was guided by a librar-
branching modify the intact sensorimotor cortex and subcor- ian expert in systematic review (LG). Using a priori search
tical structures contributing to recovery of cognitive and mo- criteria (Table 1), records were screened at the title and ab-
tor performance [10]. Much less is known about plasticity stract stage by two authors (MP, MWA) who then examined
events in human stroke. full-text manuscripts. Experimental methods and results of
Aerobic exercise (AE) is one of several behavioral studies meeting inclusion criteria were sorted in a spreadsheet
interventions known to influence neurotrophins, and results were consolidated by the same two authors, who,
neuroplasticity, and cognition [11]. “Aerobic exercise” by consensus, categorized study outcomes by theme. In order
is defined by the Consensus Panel recommending AE to compare results across preclinical studies, data derived
for stroke, as “planned, structured and repetitive physical from text and graphs were recalculated as a percentage of
activity performed for extended periods of time and at sedentary stroke animals. We did not determine statistical
sufficient intensity to improve or maintain physical fit- significance of synthesized findings but instead reported when
ness” [12]. In healthy humans and animals, AE increases the p value was <0.05 in the original source study. To protect
a variety of plasticity-promoting factors including BDNF against inflation of findings from one or two studies, we only
[13] and IGF-I [14]. AE also increases synaptogenesis synthesized outcomes that were examined in three or more
[15] and maintains brain volume in older adults [16], and studies.
two meta-analyses [11, 17] report that AE significantly We divided exercise interventions into voluntary exercise
enhances several domains of cognitive function. (VE) or forced exercise (FE). VE used a freely rotating run-
Chronic stroke survivors demonstrate significant im- ning wheel which the rodent could propel voluntarily for a
provements in aerobic fitness following AE interventions specific duration designated by the experimenter, while FE
[18] and the use of AE to optimize neuroplasticity and utilized a motorized treadmill, wheel, or rod in which the
recovery has begun to garner considerable interest [19, speed and duration were completely controlled by the
20]. However, due to inadequately powered clinical trials experimenter.
and lack of standardized outcomes, recent systematic
reviews have reported that there is insufficient evidence
to recommend poststroke AE to augment plasticity and Results
cognition [18, 21, 22]. Almost all research examining the
effects of AE after stroke have utilized animal models so We retrieved 4,250 titles of which 115 were assessed at the
the optimal methods to measure neuroplasticity (such as manuscript level (Fig. 1). Thirty measured neuroplasticity
peripheral blood levels of neurotrophins or noninvasive outcomes: neurotrophic factors (BDNF, IGF-I, and NGF),
brain imaging) in humans are not known. Furthermore, neuronal morphology (synaptic and dendritic change), and
the effects of training parameters such as time of initia- cortical reorganization (Table 2). All studies except two (ex-
tion, frequency, and intensity have not been fully eluci- amining brain activity in clinical ischemic stroke) utilized
dated—important information in order to translate find- animal stroke models, with the majority of these employing
ings to clinical trials. ischemic damage within the middle cerebral artery territory
This review and synthesis of animal and human studies (25 of 28). Two studies used a global ischemia model [50, 51]
aimed to determine (1) the extent to which AE influences and one a hemorrhagic model of striatal stroke [52]. Although
poststroke markers of neuroplasticity (e.g., neurotrophin the postinjury events after hemorrhage differ from that of
levels, synaptogenesis), (2) the optimal parameters of exercise ischemia [57], we included the study utilizing the hemor-
required to induce beneficial effects, and (3) consistent out- rhage model since it examined synaptic and dendritic
comes in animal models that could help inform the design of change (rather than molecular events directly post-
clinical trials. Although there is a body of literature exam- injury). The data from that study is pointed out separate-
ining the neuroprotective effects of aerobic exercise, and ly in the text and in Tables 4 and 5.
hypoxia/hypercapnia instituted before stroke [23, 24], we We examined the intensity of exercise regimens across
focused on AE initiated post-stroke. This paper, which studies, and based on previous research [31] and study con-
synthesizes data from studies examining neuroplasticity, sensus, we categorized rodent FE into three levels: low inten-
is the second of three reviews examining the effects of AE sity (<10 m/min), moderate intensity (10–17 m/min), and high
on poststroke repair and recovery (inflammation, cognition, intensity (>17 m/min). Based on average and peak heart rate
skilled motor performance, etc.) [25]. values [31], VE was classified as low-/moderate-intensity
Transl. Stroke Res.

Table 1 Search inclusion/exclu-


sion criteria Clinical studies Animal models of stroke

1. Adults over the age of 18 1. Utilization of a known model of ischemic


2. Acute, subacute, or chronic stroke (focal or global) or hemorrhagic stroke
3. Any poststroke AE-based intervention, alone or in 2. Poststroke intervention involving either
combination with another intervention, including voluntary (free access to a running wheel)
but not limited to overground locomotion, arm or forced (treadmill running, swimming,
ergometry, swimming, and treadmill training etc.) aerobic exercise, alone or in combination
with another intervention
4. Exercise lasting more than 2 min to meet aerobic
metabolism threshold [27] 3. At least one indicator of neuroplasticity measured
5. At least one component of aerobic fitness
measured (heart rate, perceived exertion, etc.)
6. At least one indicator of neuroplasticity measured

exercise. All animal studies, with the exception of one [33], 22 studies, all employing animal models of stroke. Neuro-
which compared old and young rodents, examined the effects trophic factors examined included BDNF (n=15), IGF-I (n=
of exercise in young adult or juvenile animals. 4), and NGF (n=4). There were too few studies examining
Twelve of the 28 animal studies employed prestroke train- neurotrophin-4 (NT-4; [59]), vascular endothelial growth fac-
ing habituation which may help to reduce stress during initi- tor (VEGF; [40]), midkine [47], and phosphorylated CREB
ation of poststroke AE, especially in FE protocols. In some [29, 31] to synthesize findings so they were excluded from
cases, this technique may have affected findings which we further analysis. Other than two studies that examined IGF-I
point out in the relevant sections. levels in peripheral blood in animal models [29, 43], all
studies examined neurotrophin levels using histology, West-
The Effects of Exercise on Neurotrophins ern blot, or other assays of brain tissue.

Levels of neurotrophic factors, proteins that support neuronal


development, function, and survival [58], were measured in BDNF

The 15 studies investigating the effects of AE on BDNF and its


receptor trkB all employed focal ischemic stroke in rodents.
Outside of the two studies investigating BDNF messenger
RNA (mRNA) [28, 32], all studies measured protein levels of
BDNF in the brain by immunoassay or immunohistochemistry.
All study protocols instituted either VE or FE, 1 to 5 days post-
stroke, with the exception of one study [37], which delayed
exercise for 1 week after stroke. Because there were 15 studies
examining BDNF, we report the findings below by brain region
studied (hippocampus, subcortical structures, and cortex).

Hippocampal Levels of BDNF Levels of exercise-induced


changes in BDNF in contralesional and ipsilesional hippo-
campus, important for learning and memory, were examined
in 10 studies. Regardless of timing of exercise onset post-
stroke (1–7 days), mode of exercise, or type and size of stroke,
there are immediate increases in hippocampal levels of BDNF.
A single bout of VE or moderate-intensity FE produces a short
burst of hippocampal BDNF [29, 31]. Short duration training
for 1 week results in similar effects [36].

Optimal Parameters Longer term exercise durations for two


or more weeks results in changes in hippocampal BDNF
expression that appear to be at least partly intensity dependent.
Fig. 1 PRISMA flow chart of search results For example, low- to moderate-intensity FE for 2 weeks [34]
Transl. Stroke Res.

Table 2 Search findings by


theme Theme Subtheme (number of articles) References

Neurotrophins BDNF, n=15 [28–42]


IGF-I, n=4 [29, 31, 43, 44]
NGF, n=4 [45–48]
Neuronal modification Synaptic change, n=15 [29, 31–35, 37, 40–42, 45, 49–52]
Dendritic branching, n=8 [32, 33, 40, 41, 48, 52–54]
Brain activity n=3 [49, 55, 56]

or 5 weeks [32] and VE for 4 [28] or 6 weeks [33] do another increased [39], and a third showed no effect
not affect hippocampal BDNF. However, three studies [37]. Careful examination of methodological differences
examining high-intensity FE (18–20 m/min) report con- suggests that longer duration FE (3 weeks) instituted
flicting findings: one showed decreased BDNF [41], 24 h post-stroke may increase hippocampal BDNF [39]

Table 3 BDNF level in cortical regions following exercise (as a percentage of stoke sedentary animals)

Articles Duration of VE VE VE VE+reach FE FE FE FE+reach FE+reach


exercise IL CL PL PL IL CL PL IL CL

Risedal et al. [31]: anterior brain 30 100 90 104


Risedal et al. [31]: posterior brain 30 89 104 71
Ploughman et al. [29]: 30 min 11 m/min 1 115 143*
or VE
Ploughman et al. [29]: 30 min 14 m/min 1 120
Ploughman et al. [29]: 60 min 11 m/min 1 86
Kim et al. [30]: moderate deficit group 12 90 150
Ploughman et al. [31] 1 100 121*
Ploughman et al. [32] 35 104 102 142 132
Maldonado et al. [33]: young group 42 100 100
Maldonado et al. [33]: old group 42 100
Seo et al. [34]: treadmill 14 133 115
Seo et al. [34]: rotorod 14 267 130
Ke et al. [36]: VE or FE 7 155 95
Ke et al. [36]: FE with hindlimb 7 181
stimulation
Sakakima et al. [38] 14 200*
Quirie et al. [37]: mature BDNF lesion rim 14 70
Quirie et al. [37]: mature BDNF lesion 14 206*
border
Quirie et al. [37]: mature BDNF 14 433*
perilesional
Quirie et al. [37]: proBDNF lesion rim 14 70
Quirie et al. [37]: proBDNF lesion border 14 75
Quirie et al. [37]: proBDNF perilesional 14 52*
Quirie et al. [37]: total BDNF 14 100
Zhang et al. [39] 28 200*
Lan et al. [42] 14 230*
Mean 96.3 102.3 110.0 100.0 148.5 120.9 159.3 142.0 132.0
SD 6.4 10.3 42.3 0.0 81.0 20.8 107.9 NA NA

Values are derived and recalculated from the included article text and graphs
VE voluntary exercise, FE forced exercise, IL ipsilesional cortex, CL contralesional cortex, PL perilesional cortex, +Reach combined with skilled reach
training, NA not available
*p<0.05, significantly different from the stroke sedentary group
Table 4 Synthesis of findings from studies reporting the effects of exercise on indicators of synaptic modification (as a percentage of stroke sedentary animals)

Articles Duration VE IL VE VE VE VE VE+ FE IL FE FE PL FE FE FE+ FE+ FE IL FE FE FE FE IL FE CL


of CTX CL PL IL CL reach CTX CL CTX IL CL reach reach STR CL IL CL THAL THAL
exercise CTX CTX HIP HIP IL CTX HIP HIP IL CL STR DG DG
Transl. Stroke Res.

Synapsin-I
Ploughman et al. [29]: 1 84 110 96 106 140 112
30 min 11 m/min or VE
Ploughman et al. [29]: 1 105 97 87
30 min 14 m/min
Ploughman et al. [29, 31]: 1 124 130 131
60 min 11 m/min
Ploughman et al. [32] 35 100 100 100 100
Chang et al. [35] 14 165*
Zhang et al. [40] 24 220* 150*
Shih et al. [41]: 14 112 135*
low intensity
Shih et al. [41]: 14 100 125
high intensity
Lan et al. [42] 14 194*
Synaptophysin
Marin et al. [49] 14 100
Seo et al. [34]: treadmill 14 80 97 183* 140 183* 150 150* 150
Seo et al. [34]: rotorod 14 84 95 206* 135 194* 130 175* 140
Quirie et al. [37] 14 400*
Mean 84 100 110 96 121 105 297 138 124 100 100 158 189 140 163 145
SD 67 10 146 42 19 11 8 14 18 7
Synapse to neuron ratio
Briones et al. [50]: anterior 14 95
Briones et al. [50]: medial 14 96
Briones et al. [51]: 14 96
inner layer
Briones et al. [51]: 14 95
outer layer
Receptor number
Dahlqvist et al. [45]: 30 84
serotonin dentate gyrus
Dahlqvist et al. [45]: 30 85*
serotonin (average dorsal
subregions; sig CA4)
Table 4 (continued)

Articles Duration VE IL VE VE VE VE VE+ FE IL FE FE PL FE FE FE+ FE+ FE IL FE FE FE FE IL FE CL


of CTX CL PL IL CL reach CTX CL CTX IL CL reach reach STR CL IL CL THAL THAL
exercise CTX CTX HIP HIP IL CTX HIP HIP IL CL STR DG DG

Dahlqvist et al. [45]: serotonin 30 89


(average ventral subregions)
Maldonado et al. [33]: NMDA 42 100 100
young group
Synaptic spines/boutons
Briones et al. [50]: no. of 14 89
multiple boutons anterior
Briones et al. [50]: no. of 14 97
multiple boutons medial
Briones et al. [50]: no. of 14 105
perforated synpases anterior
Briones et al. [50]: no. of 14 104
perforated synapses medial
Maldonado et al. [33]: 42 176 160
(spinophilin) young group
Takamatsu et al. [52]: PSD-95 11 64 89
synaptic spinesa
Takamatsu et al. [52]: synaptic 11 107 100
spine densitya
Shih et al. [41]: dendritic spine 14 110 126* 135* 118*
density, low intensity
(CA1 apical)
Shih et al. [41]: dendritic spine 14 97 114 110 111
density, high intensity
(CA1 apical)
Shih et al. [41]: dendritic spine 14 120* 114*
density, low intensity
(CA1 basal)
Shih et al. [41]: dendritic spine 14 93 103
density, high intensity
(CA1 basal)
Shih et al. [41]: dendritic spine 14 108 108
density, low intensity
(CA3 apical)
Shih et al. [41]: dendritic spine 14 101 105
density, high intensity
(CA3 apical)
Shih et al. [41]: dendritic spine 14 123* 121*
density, high intensity
Transl. Stroke Res.
Transl. Stroke Res.

FE CL

VE voluntary exercise, FE forced exercise, IL ipsilesional cortex, CL contralesional cortex, PL perilesional cortex, +Reach combined with skilled reach training, CTX cortex, HIP hippocampus, STR
THAL
more than delayed (2 or 7 days) shorter duration
(2 weeks) FE.
THAL
FE IL

Subcortical Levels of BDNF In subcortical regions, the effects


of AE on BDNF levels were variable and were only examined
DG

in five studies: two examining the thalamus [28, 34] and three
CL
FE

the striatum [35–37]. In the ipsilesional or contralesional


thalamus, there was no significant effect of exercise whether
DG
FE
IL

VE for 30 days [28] or low intensity FE for 14 days [34].


STR

High-intensity protocols have not been examined. In the


CL
FE

ipsilesional striatum, 7 days VE [36] beginning 24 h post-


FE IL

stroke did not significantly elevate BDNF while a 2-week


STR

burst of high-intensity FE (without prestroke habituation) or


4 weeks training at moderate intensity commencing 24 h post-
reach
FE+

CL

stroke increased striatal BDNF [35, 40]. When exercise was


accompanied by prestroke habituation [36, 37] or delayed by
reach
FE+

IL

7 days [37], there was no effect on striatal BDNF.


HIP

114
CL
FE

These findings suggest that the thalamus and striatum, even


on the ipsilesional side, are relatively insensitive to exercise-
HIP

110
FE

induced changes in BDNF compared to cortical and hippo-


IL

campal regions, and protocols must be high intensity or mod-


FE PL
CTX

erate intensity/long duration to affect striatal levels of BDNF.


CTX

Cortical Levels of BDNF Of the 12 studies examining cortical


CL
FE

levels of BDNF, five reported increases due to exercise [29,


31, 38, 40, 42], six reported no significant effects [28, 30,
FE IL
CTX

32–34, 36], and one reported mixed findings depending on the


cortical region and form of BDNF examined [37]. In order to
reach
VE+

synthesize study findings, cortical BDNF was recalculated in


IL

all studies as a percentage of BDNF in the corresponding


HIP
VE
CL

region of sedentary, socially housed stroke animals. Our syn-


thesized findings (Table 3) show that VE (1, 30, or 42 days)
HIP
VE
IL

does not alter cortical BDNF levels in the contralesional


Values are derived and recalculated from included article text and graphs

cortex (102.3±10.3 %), the ipsilesional cortex distant from


CTX
VE
PL

the lesion (96.3± 6.4 %), and perilesional cortex (110 ±


*p<0.05, significantly different from the stroke sedentary group

42.3 %) even when combined with skilled reach training [33].


CTX
VE
CL

Optimal Parameters Analysis of the effects of FE regimens


VE IL
CTX

on cortical BDNF levels (Table 3) suggested an effect of


region (proximity to lesion), intensity, and form of BDNF
Duration

exercise

striatum, DG dentate gyrus, THAL thalamus

measured (matureBDNF, proBDNF, or total BDNF). In the


contralesional cortex, exercise-induced changes in BDNF are
14

Utilized hemorrhagic model of stroke


of

only detected after one episode of moderate (30 or 60 min)


wheel running [29] [31]. Exercise for longer periods whether
using wheel, treadmill or rotating rod, or adding skilled reach
Shih et al. [41]: dendritic

intensity (CA3 basal)

training does not significantly affect BDNF levels in the


spine density, high

contralesional cortex (120.9±20.8 %; Table 3). There were


Table 4 (continued)

also no significant effects of FE on the ipsilesional cortex


(CA3 basal)

(148.5±81.0 %; Table 3) even when combined with reach


training [32]. In the perilesional cortex however, FE signifi-
Articles

cantly increased BDNF (159.3±107.9 %) whether measured


by immunostaining (200 %, [38]) or Western blot analysis
a
Table 5 Synthesis of findings from studies reporting effects of exercise on dendritic modification (as a percentage of stroke sedentary animals)

Articles Duration of VE IL VE CL VE PL VE+reach FE IL FE CL FE PL FE IL FE CL FE+reach FE+reach FE IL FE CL FE IL FE CL


exercise CTX CTX CTX PL CTX CTX CTX CTX HIP HIP IL CTX CL CTX STR STR DG DG

Microtubule-associated protein (MAP-2)


Ploughman et al. [32]: 35 84 95 95 109
MAP2 stain (cingulate)
Ploughman et al. [32]: 35 83 88
MAP2 stain (sensory)
Ploughman et al. [32]: 35 101 124
MAP2 stain (motor)
Ploughman et al. [32]: 35 100 100 100 100
MAP2 mRNA
Maldonado et al. [33]: 42 100 100
young group MAP2 stain
Shimada et al. [54]: MAP2 28 131 157* 145* 101
stain, low intensity (CA3)
Shimada et al. [54]: MAP2 28 80 84 90 81
stain, high intensity (CA3)
Shimada et al. [54]: MAP2 28 149 125
stain, low intensity (CA1)
Shimada et al. [54]: MAP2 28 85 63
stain, high intensity (CA1)
Dendritic morphology
Takamatsu et al. [52]: total 11 100 137
dendritic lengtha
Takamatsu et al. [52]: peak 11 100 129
no. intersectionsa
Shih et al. [41]: peak no. 14 114 133* 131* 142*
intersections, low intensity
(CA3)
Shih et al. [41]: peak no. 14 100 107 100 100
intersections, high intensity
(CA3)
Shih et al. [41]: peak no. 14 100 125*
intersections, low intensity
(CA1)
Shih et al. [41]: peak no. of 14 100 105
intersections, high intensity
(CA1)
Mean 100 100 92 95 107 112 98 105 100 133 117 106
SD 11 8 23 29 4 15 0 6 26 26
Dendritic growth indicators
Transl. Stroke Res.
Table 5 (continued)

Articles Duration of VE IL VE CL VE PL VE+reach FE IL FE CL FE PL FE IL FE CL FE+reach FE+reach FE IL FE CL FE IL FE CL


exercise CTX CTX CTX PL CTX CTX CTX CTX HIP HIP IL CTX CL CTX STR STR DG DG

Mizutani et al. [48]: GAP43 14 162*


Transl. Stroke Res.

Mizutani et al. [48]: pSer41- 14 175*


GAP43
Liu et al. [53]: netrin protein 28 113
day 4
Liu et al. [53]: netrin protein 28 140*
day 7
Liu et al. [53]: netrin protein 28 141*
day 14
Liu et al. [53]: netrin protein 28 170*
day 28
Liu et al. [53]: netrin receptor 28 117
DCC day 4
Liu et al. [53]: netrin receptor 28 181*
DCC day 7
Liu et al. [53]: netrin receptor 28 171*
DCC day 14
Liu et al. [53]: netrin receptor 28 188*
DCC day 28
Liu et al. [53]: netrin receptor 28 109
Unc5B day 4
Liu et al. [53]: netrin receptor 28 109
Unc5B day 7
Liu et al. [53]: netrin receptor 28 51*
Unc5B day 14
Liu et al. [53]: netrin receptor 28 47*
Unc5B day 28
Zhang et al. [40]: GAP43 23 87 90
Zhang et al. [40]: NogoA 23 22* 48
(inhibitor)

Values are derived and recalculated from the included article text and graphs
VE voluntary exercise, FE forced exercise, IL ipsilesional cortex, CL contralesional cortex, PL perilesional cortex, +Reach combined with skilled reach training, CTX cortex, HIP hippocampus, STR
striatum, DG dentate gyrus, THAL thalamus
*p<0.05, significantly different from the stroke sedentary group
a
Utilized hemorrhagic model of stroke
Transl. Stroke Res.

(206 and 433 % [37], 230 % [42], and 200 % [40]). There was studies confirmed increased NGF-positive cells in the
an effect of the form of BDNF with mature BDNF signifi- ipsilesional cortex after 14 to 28 days of exercise training:
cantly increased by FE, while proBDNF was decreased in the either moderate- to high-intensity FE [46, 47] or 14 days VE
same area [37]. The apparently contradictory effect may be (12 h/day) [48]. Both Chung et al. and Matsuda et al. showed
explained by the different roles of each form; conversion of an effect of exercise duration with increasing NGF-positive
proBDNF to matureBDNF by activity assists with synapse cells over time, peaking at days 16 [46] and 28 [47], but not
stabilization [60]. Furthermore, Quirie et al. demonstrated that earlier (days 1, 3, 5, 7, and 14 [47]). The exceptions to these
although FE increased matureBDNF in the perilesional cor- findings were those of Dahlqvist and colleagues [45]. They
tex, there was no effect at the rim of the lesion (Table 3; [37]). observed that 1 month of daily VE resulted in lower levels of
Synthesized findings showed that increased BDNF in the mRNA expression of NGF-induced gene A (ipsilesional and
perilesional cortex was elicited using similar FE parameters: contralesional cortex and hippocampus) than animals exposed
12 to 18 m/min (moderate), 30 min/day for 14 or more days, to social housing or environmental enrichment. In contrast to
3–7 days post-stroke [37, 40, 42], but not high-intensity the previously mentioned studies, animals in the VE group
protocols (immediate training at 20 m/min for 7 days [36]). were provided access to running wheels 24 h/day and were
The exception was increased BDNF following rotating rod socially deprived which could have produced unintended
exercise reported by Sakakima et al. (low intensity 4.8 m/min stressful effects.
[38]), suggesting a training mode-specific effect. Our findings
also suggest that BDNF changes in the cortex are not uni- The Effects of Exercise on Neuronal Modification
formly distributed so studies in which the entire hemisphere
was homogenized or tissues pooled for analysis [30, 32, 34] After stroke, synaptogenesis (the development of new synap-
may have been unable to detect subtle differences in BDNF ses) and dendritic branching represent examples of
within ipsilesional areas. neuroplastic change implicated in stroke recovery [61]. Fif-
teen studies investigated the effects of AE on synaptic modi-
IGF-I fication after stroke (Table 1) and eight studies investigated
the effects of AE on dendritic change (Table 1).
Four studies examined the effects of AE on IGF-I levels
following stroke [29, 31, 43, 44]. Two studies, which mea- Synaptic Change
sured both peripheral blood and brain levels [29, 43], con-
firmed the apparent paradoxical relationship between levels of Of the 15 studies examining exercise-induced changes in syn-
IGF-I in brain and blood. There was a strong negative corre- apses after stroke, eight reported significant increases in synap-
lation between plasma levels of IGF-I [43] or serum [29] and tic plasticity, one reported a decrease [45], while the remaining
FE exercise duration [43] or VE distance [29], suggesting that six reported no effect of exercise compared to the control. The
exercise training depletes circulating IGF-I and enhances its synthesized data from these studies are shown in Table 4.
uptake into the injured brain. Notably, the four studies examining the effects of VE [29, 33,
45, 49] showed no benefit of VE on synaptic modification.
Optimal Parameters In terms of brain levels of IGF-I, a single
bout of moderately high-intensity FE (14 m/min), but not Optimal Parameters Synthesized findings (Table 4)
lower intensity FE (11 m/min) or VE [29, 31], increased showed that moderate- to high-intensity (8–20 m/min)
IGF-I levels in contralesional hippocampus (but not the FE regimens for 14 or more days increased synapsin-I
contralesional cortex or ipsilesional hippocampus). Addition- or synaptophysin in the perilesional cortex (297±146 %
ally, 2 weeks of daily high-intensity FE increased IGF-I con- [37, 42]), ipsilesional cortex distant from the lesion
centration in the ipsilesional motor cortex (about 309 % sed- (121 ± 67 % [40]), ipsilesional hippocampus (124 ±
entary) and striatum (about 182 % sedentary [43]). The find- 19 % [34]), ipsilesional striatum (189 ± 8 % [34]),
ings were also confirmed by Zhang et al. showing that ipsilesional dentate gyrus (189 ± 8 % [34]), and
moderate-intensity FE increased the number of IGF-I- ipsilesional thalamus (163±18 % [34]). Only one study
positive cells within the peri-infarct cortex after 7 days showed an increase in synapsin-I in the contralateral
(163 % sedentary), 14 days (220 % sedentary), or 21 days hemisphere, but the increase was short-lived, dropping
(192 % sedentary) of exercise post-stroke [44]. below sedentary levels 30 min after cessation of exer-
cise [29]. Other indicators of synaptic change such as
NGF synapse to spine ratio, number of receptors, or synaptic
spine density were examined in five studies [33, 45,
Of the four studies investigating NGF or its receptor trkA, all 50–52], and none of these reported any beneficial ef-
examined brain tissue levels in animal models of stroke. Three fects of poststroke exercise (Table 4).
Transl. Stroke Res.

Dendritic Growth following VE in an animal model of stroke [49] and two using
functional magnetic resonance imaging (fMRI) [55, 56] to
Eight studies examined the effects of AE on dendritic modi- map subcortical activation associated with movement of the
fication (Tables 1 and 5), microtubule-associated protein affected lower extremity (Table 2). Both clinical studies
(MAP2) [32, 33, 54], growth-associated protein-43 (GAP43) employed moderate-intensity treadmill training in patients
[40, 48], netrin [53], and dendritic length and complexity [41, about 3 years post-stroke. Longer duration (6 month) training
52]. Two studies employing VE showed no benefit of VE or resulted in increased subcortical activation during movement
VE combined with skilled reaching on MAP2 levels in the of the paretic limb [56] while 4 weeks training did not [55]. In
ipsilesional cortex but evidence of increases in GAP43 and an animal model, 14 days VE did not alter EEG power in the
pSer41-GAP43 in the area directly adjacent to the focal lesion ipsilesional parietal cortex [49]. There were no studies that
(Table 5; VE PL CTX [33, 48]). The remaining six studies examined the effects of AE on reorganization in brain areas
used FE paradigms from moderate to high intensity for 11 to responsible for cognition, language, or upper extremity motor
35 days, with differential effects depending on the intensity control.
and brain region examined.

Optimal Parameters In terms of the effect of FE in cortical


regions, FE increased GAP43, pSer41-GAP43 [48], netrin, Discussion
and netrin receptors [53] in the cortical areas adjacent to the
lesion (Table 5; FE PL CTX) but not MAP2 in more distant Reduced mortality and better poststroke care results in more
cortical regions [32]. Furthermore, both Shimada et al. and stroke survivors with serious poststroke disability [1]. This,
Shih et al. showed that low-intensity (8 m/min), but not high- along with anticipated increases in the rate of stroke due to an
intensity (20–22 m/min) AE resulted in greater MAP2 stain- aging population and rising rates of obesity coupled with a
ing and dendritic arborization within the CA1, CA3, and sedentary lifestyle, highlights the importance of carefully
dentate gyrus of the contralesional (but not ipsilesional) hip- designed clinical trials to improve recovery during the reha-
pocampus [41, 54]. Only two studies examined the dendritic bilitative phase of stroke. As recommended by stroke research
change in the perilesional cortex following forced exercise. experts, our aims were to identify consistencies and determine
Liu et al. demonstrated an increase in the expression of netrin the optimal methods in animal models and preliminary phase I
and netrin receptor (DCC) following 16 days of moderate FE and II clinical trials in order to advance the field and inform
[53], while Zhang and group [40] showed decreases in out- the design of future phase III and IV clinical trials [63].
growth inhibitor NogoA following 4 weeks of a similar exer-
cise paradigm. Subject Characteristics in Preclinical Trials Compared
to Stroke Patients
Exercise-induced dendritic change within the striatum has
been examined in only two studies [40, 52]. They showed that All preclinical studies, with the exception of one, which
although there was a decrease in NogoA in the ipsilesional compared young and old female animals [33], utilized juve-
side, dendritic length and complexity was not significantly nile or young adult male rats. The two clinical studies on the
increased in either side with FE employed after a striatal other hand, examined men and women who were on average
hemorrhagic stroke [52]. However, Takamatsu et al. used only 63.2 [56] and 59.8 years of age [55]. Over two thirds of
11 days of training, one of the shortest interventions reported, strokes in Canada [64] and the USA [65] occur over the age
and the increase in dendritic length and intersections in the of 65, evenly balanced between men and women. The consis-
contralesional striatum, although not significant, was 137 and tent use of young male rodents in stroke models likely limits
129 % of sedentary [52]. In summary, evidence from the the ability to detect age and sex effects that have been dem-
synthesized findings suggests that after stroke, low- to onstrated in other models of neurological injury [66–70].
moderate-intensity FE, but not VE, enhances markers associ- Future research in animal models should focus on understand-
ated with dendritic branching in brain regions adjacent to and ing poststroke rehabilitation and plasticity in the older brains
distant from the lesion. of males and females.
In humans, variants of genetic polymorphisms of BDNF,
The Effects of Exercise on Brain Activity apolipoprotein E (APOE), and other genes may alter response
to interventions that would be expected to enhance plasticity
Imaging techniques that map brain activation patterns have the and recovery [71]. About a third of the population has a
potential to detect neuroplasticity after stroke [62]. Only three variation of the amino acid chain in the proBDNF gene
studies that met inclusion criteria examined the effects of AE (val66met) which results in decreased activity-induced secre-
on brain activation, one using electroencephalography (EEG) tion of BDNF. The impact of BDNF gene polymorphism has
Transl. Stroke Res.

not been fully elucidated with some studies showing an im- plasticity, their dynamics in response to exercise must be more
pairment in memory and cognition and others showing no fully understood.
effect [20]. Future studies examining the influence of AE post- Our review revealed four studies investigating the effects
stroke will have to consider patient populations with BDNF of AE on brain tissue levels (but not serum or plasma levels) of
and other genetic polymorphisms. NGF, with three of the four confirming that moderate- to high-
intensity exercise for 2 to 4 weeks increases NGF-positive
Timing of Interventions cells in the ipsilesional cortex [46–48], suggesting that NGF
may be a potentially important neurobiological marker. How-
In studies involving animals, exercise training began 1 to ever, the sensitivity of peripheral blood NGF as a marker of
7 days post-stroke compared to 3 years post-stroke in the activity-induced plasticity is under debate with some studies
two clinical studies included in this review. Treadmill training showing increases and others showing no effect with exercise
during the subacute rehabilitation phase of stroke improves [81, 82]. Future research should investigate peripheral NGF as
endurance and fitness [72, 73]; however, how soon to imple- a viable and accessible measure of poststroke plasticity.
ment aerobic exercise and the optimal intensity to influence Brain imaging likely provides the most accessible and
plasticity after stroke is not known. In an animal model, sensitive indicator of training-induced plasticity in clinical
delayed initiation of rehabilitation (30 days post-stroke) did stroke trials. Using transcranial magnetic stimulation,
not produce beneficial effects compared to training onset at 5 rehabilitation-induced changes in motor threshold and map
and 14 days, suggesting that sensitivity to training declines size of target muscles can be detected [56, 83]. Functional
with time post-stroke [74]. It is likely that the window of time MRI [84], EEG, magnetoencephalography (MEG) [85], and
in humans when the brain is most responsive is within the first positron emission tomography (PET) also hold promise in
few months post-stroke; therefore, future clinical trials should identifying brain activity in response to rehabilitative treat-
focus efforts within this period when the brain is most ame- ments [86, 87]. As PET probes continue to advance [88], there
nable to change. will be opportunities to map the properties of neurobiological
markers such as VEGF [89].
Identifying Optimal Biological Indicators of Plasticity
for Clinical Trials
Regions of Interest in Poststroke Recovery
All studies examining the effects of AE on BDNF included in
this review measured protein levels of BDNF in the brain, not Early after ischemic stroke, fMRI and neuronal tracing
in peripheral blood, which would be more practical in clinical studies show that motor-related neural activity is en-
studies. Although lower BDNF blood levels have been impli- hanced in both hemispheres, but that over subsequent
cated in neurological disorders [75, 76] and higher peripheral months, good motor performance and recovery of the
BDNF is associated with exercise [13], there is no correlation affected side is accompanied by more focused activity
between brain and blood levels of BDNF after stroke [77], in the injured hemisphere particularly the ipsilesional
making it difficult to identify clinically appropriate biological dorsal premotor cortex and striatum [10]. It was surpris-
markers of plasticity. Furthermore, there is debate as to which ing that only 3 of 30 studies in this review examined
component of peripheral blood is most sensitive to behavioral plastic changes in the striatum as a result of AE. Clearly,
interventions: serum, plasma, or platelets [78, 79]. Future more research needs to be undertaken to determine the
animal studies must identify and optimize the most sensitive role of distal brain regions in poststroke recovery.
methods to detect peripheral BDNF that can be translated to In this review, we found a preponderance of studies
clinical populations. examining the effects of AE on hippocampal plasticity
Voss and group [80] in a review of exercise studies in post-stroke. Although the role of the hippocampus in
elderly people note that in addition to BDNF, other peripheral motor relearning and motor recovery after stroke has
markers, such as IGF-I and VEGF, may be useful indicators of not been examined, recent imaging studies using longi-
neuroplasticity. Our review did not find any studies examining tudinal high-resolution MRI suggests that gray matter
peripheral levels of VEGF; however, IGF-I was measured in volume of the hippocampus and precuneus is positively
either brain or blood in four studies (Table 1). Interestingly, correlated with motor recovery [90]. Attention, motiva-
both Ploughman et al. [29] and Chang et al. [43] showed that tion, and short- and long-term memory, which are regu-
lower peripheral IGF-I and higher brain levels of IGF-I were lated in part by the hippocampus, are likely important
associated with exercise duration, suggesting that blood levels contributors to practice and motor relearning post-stroke
were not positively correlated with brain levels and the re- [91]. Whether or not the hippocampus plays a role in
sponse of IGF-I is dynamic and time dependent. Although poststroke recovery and its potential as a plasticity target
IGF-I and VEGF have potential as peripheral biomarkers of are important to determine in future research.
Transl. Stroke Res.

Optimal Aerobic Exercise Parameters to Affect Plasticity & Future clinical trials should move the AE intervention to
the subacute phase of stroke when the brain is most
The majority of studies examined in this systematic review amenable to change.
employed FE paradigms which are more likely to mimic & Create a suite of peripheral (blood, saliva, etc.)
poststroke rehabilitation protocols in which the intensity and neurotrophin markers or noninvasive imaging techniques
duration are modified by the therapist. Overall, VE paradigms to detect exercise-induced repair and recovery.
did not significantly alter neurotrophins or synaptic and den- & Focus on FE protocols in which the timing and dosage can
dritic structure compared to FE protocols. We found that be titrated in order to translate to post-stroke rehabilitation
intensity but not duration of FE was an important parameter trials, aiming to reach moderate- to high-intensity exercise
that increased the expression of neurotrophins and modified three to five times a week.
synapses and dendrites. The critical threshold of intensity & Explore the combination of AE with skilled task and
depended on the outcome and the brain region examined. cognitive and language training.
Increased BDNF was consistently observed in the perilesional & In preclinical and imaging studies, examine the effects of
cortex with moderate-intensity FE, 30 min/day for 14 or more AE on brain networks that may underlie recovery such as
days beginning 3 to 7 days post-stroke. Our synthesized those in the supplementary motor cortex and subcortical
findings suggest that brain levels of IGF-I and NGF as well structures. There is a preponderance of studies of the
as markers of synaptic modification were increased consis- hippocampus even though its role in post-stroke recovery
tently across studies using moderate- to high-intensity FE is not known.
protocols. Post-AE dendritic change was particularly sensitive & AE effects may differ among certain populations with
to intensity, with two studies demonstrating that dendritic BDNF and other genetic polymorphisms. Understanding
modification is less robust and even downregulated with this group will help in the application of AE and other
immediate high-intensity FE (20–22 m/min), suggesting that interventions, aimed at modifying plasticity to a broader
moderate-intensity training may be optimal. group of patients.
Intensity of inpatient stroke rehabilitation varies substan-
tially across hospitals, health authorities, and countries. In the
Netherlands, it has been reported that patients attending stroke
rehabilitation spend 114 min/day in their heart rate training Conclusion
zone [92], whereas in North America, patients spent less than
4 min per day at an intensity that would lead to a cardiovas- Our synthesized findings show that forced exercise at moder-
cular training effect [93]. According to best practices for ate to high intensity increases BDNF, IGF-I, NGF, and
aerobic training after stroke [12], treadmill or bicycle synaptogenesis in multiple brain regions at least in animal
ergometry, 20 min, three or more times per week for at least models of stroke. Clinical stroke and animal models need to
8 weeks with graduated intensity (depending on fitness base- become more similar in terms of choice of research subject
line) is fundamental to obtain cardiovascular fitness in stroke (age, sex, time post-stroke) and selection of plasticity indicators
patients. However, these guidelines were derived from studies (peripheral vs. brain levels).
examining cardiovascular fitness, not brain plasticity, so al-
though they may be a reasonable starting point, the optimal
Acknowledgments We acknowledge the work of Chelsea Harris, Ste-
parameters to induce neuroplastic change are not known. phen Hogan, and Michael Monks in obtaining and cataloguing articles.
Furthermore, from our understanding of task specificity train-
ing [94], aerobic exercise on its own is unlikely to improve Conflict of Interest All authors declare that they have no conflicts of
unrelated skills such as fine motor performance or executive interest.
function. Combined therapy in which aerobic exercise is
Compliance with Ethical Requirements This article is a synthesis of
paired with other interventions to “prime” the recovering brain
human and animal study findings. Studies included in the review identi-
holds promise for future trials [63]. fied that they were in accordance with institutional and national guide-
lines for clinical and animal research.

References
Summary of Recommendations

1. Towfighi A, Saver JL. Stroke declines from third to fourth leading


& Preclinical studies should focus on understanding post-
cause of death in the United States: historical perspective and chal-
stroke rehabilitation and plasticity in the older brains of lenges ahead. Stroke. 2011;42(8):2351–5. doi:10.1161/
male and female animals. STROKEAHA.111.621904.
Transl. Stroke Res.

2. Hennerici MG, Kern R, Szabo K. Non-pharmacological strategies for effects of aerobic exercise and genetic variation on brain-derived
the treatment of acute ischaemic stroke. Lancet Neurol. 2013;12(6): neurotrophic factor. Phys Ther. 2013;93(12):1707–16. doi:10.2522/
572–84. doi:10.1016/S1474-4422(13)70091-7. ptj.20130053.
3. Langhorne P, Bernhardt J, Kwakkel G. Stroke rehabilitation. Lancet. 21. McDonnell MN, Smith AE, Mackintosh SF. Aerobic exercise to
2011;377(9778):1693–702. doi:10.1016/S0140-6736(11)60325-5. improve cognitive function in adults with neurological disorders: a
4. Murphy TH, Corbett D. Plasticity during stroke recovery: from systematic review. Arch Phys Med Rehabil. 2011;92(7):1044–52.
synapse to behaviour. Nat Rev Neurosci. 2009;10(12):861–72. doi: doi:10.1016/j.apmr.2011.01.021.
10.1038/nrn2735. 22. Cumming TB, Tyedin K, Churilov L, Morris ME, Morris ME. The
5. Carmichael ST. Cellular and molecular mechanisms of neural repair effect of physical activity on cognitive function after stroke: a sys-
after stroke: making waves. Ann Neurol. 2006;59(5):735–42. doi:10. tematic review. Int Psychogeriatr. 2012;24(4):557–67. doi:10.1017/
1002/ana.20845. S1041610211001980.
6. Ploughman M, Windle V, MacLellan CL, White N, Dore JJ, Corbett 23. Romero JR, Babikian VL, Katz DI, Finklestein SP. Neuroprotection
D. Brain-derived neurotrophic factor contributes to recovery of and stroke rehabilitation: modulation and enhancement of recovery.
skilled reaching after focal ischemia in rats. Stroke. 2009;40(4): Behav Neurol. 2006;17(1):17–24.
1490–5. doi:10.1161/STROKEAHA.108.531806. 24. Zhang F, Wu Y, Jia J. Exercise preconditioning and brain ischemic
7. MacLellan CL, Keough MB, Granter-Button S, Chernenko GA, Butt tolerance. Neuroscience. 2011;177:170–6. doi:10.1016/j.
S, Corbett D. A critical threshold of rehabilitation involving brain- neuroscience.2011.01.018.
derived neurotrophic factor is required for poststroke recovery. 25. Austin MW, Ploughman M, Glynn L, Corbett D. Aerobic exercise
Neurorehabil Neural Repair. 2011;25(8):740–8. doi:10.1177/ effects on neuroprotection and brain repair following stroke: a sys-
1545968311407517. tematic review and perspective. Neurosci Res. 2014. doi:10.1016/j.
8. De Smedt A, Brouns R, Uyttenboogaart M, De Raedt S, Moens M, neures.2014.06.007.
Wilczak N, et al. Insulin-like growth factor I serum levels influence 26. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC,
ischemic stroke outcome. Stroke. 2011;42(8):2180–5. doi:10.1161/ Ioannidis JP, et al. The PRISMA statement for reporting systematic
STROKEAHA.110.600783. reviews and meta-analyses of studies that evaluate health care inter-
9. Park H, Poo MM. Neurotrophin regulation of neural circuit develop- ventions: explanation and elaboration. PLoS Med. 2009;6(7):
ment and function. Nat Rev Neurosci. 2013;14(1):7–23. doi:10.1038/ e1000100. doi:10.1371/journal.pmed.1000100.
nrn3379. 27. Medbo JI, Mohn AC, Tabata I, Bahr R, Vaage O, Sejersted OM.
10. Nudo RJ. Recovery after brain injury: mechanisms and principles. Anaerobic capacity determined by maximal accumulated O2 deficit.
Front Hum Neurosci. 2013;7:887. doi:10.3389/fnhum.2013.00887. J Appl Physiol (1985). 1988;64(1):50–60.
11. Angevaren M, Aufdemkampe G, Verhaar HJ, Aleman A, Vanhees L. 28. Risedal A, Mattsson B, Dahlqvist P, Nordborg C, Olsson T,
Physical activity and enhanced fitness to improve cognitive function Johansson BB. Environmental influences on functional outcome
in older people without known cognitive impairment. Cochrane after a cortical infarct in the rat. Brain Res Bull. 2002;58(3):315–21.
Database Syst Rev. 2008;3, CD005381. doi:10.1002/14651858. 29. Ploughman M, Granter-Button S, Chernenko G, Tucker BA, Mearow
CD005381.pub3. KM, Corbett D. Endurance exercise regimens induce differential
12. Mackay-Lyons M, Macko R, Eng J, Billinger S, Hafer-Macko C, effects on brain-derived neurotrophic factor, synapsin-I and insulin-
Suskin N, Tang A, Giacomantonio N, Prior P, Che M, Dromerick A, like growth factor I after focal ischemia. Neuroscience. 2005;136(4):
Thornton M, Reid R, Unsworth K, LaGrand C, Firth W. Aerobic 991–1001. doi:10.1016/j.neuroscience.2005.08.037.
exercise recommendations to optimize best practices in care after 30. Kim MW, Bang MS, Han TR, Ko YJ, Yoon BW, Kim JH, et al.
stroke. Halifax; 2012. http://strokebestpractices.ca/wp-content/ Exercise increased BDNF and trkB in the contralateral hemisphere of
uploads/2013/07/AEROBICS-FINAL-July-2013.pdf. the ischemic rat brain. Brain Res. 2005;1052(1):16–21. doi:10.1016/
13. Coelho FG, Gobbi S, Andreatto CA, Corazza DI, Pedroso RV, j.brainres.2005.05.070.
Santos-Galduroz RF. Physical exercise modulates peripheral levels 31. Ploughman M, Granter-Button S, Chernenko G, Attwood Z, Tucker
of brain-derived neurotrophic factor (BDNF): a systematic review of BA, Mearow KM, et al. Exercise intensity influences the temporal
experimental studies in the elderly. Arch Gerontol Geriatr. profile of growth factors involved in neuronal plasticity following
2013;56(1):10–5. doi:10.1016/j.archger.2012.06.003. focal ischemia. Brain Res. 2007;1150:207–16. doi:10.1016/j.
14. Gatti R, De Palo EF, Antonelli G, Spinella P. IGF-I/IGFBP system: brainres.2007.02.065.
metabolism outline and physical exercise. J Endocrinol Invest. 32. Ploughman M, Attwood Z, White N, Dore JJ, Corbett D. Endurance
2012;35(7):699–707. doi:10.3275/8456. exercise facilitates relearning of forelimb motor skill after focal
15. Hotting K, Roder B. Beneficial effects of physical exercise on ischemia. Eur J Neurosci. 2007;25(11):3453–60. doi:10.1111/j.
neuroplasticity and cognition. Neurosci Biobehav Rev. 2013;37(9 1460-9568.2007.05591.x.
Pt B):2243–57. doi:10.1016/j.neubiorev.2013.04.005. 33. Maldonado MA, Allred RP, Felthauser EL, Jones TA. Motor skill
16. Colcombe SJ, Erickson KI, Scalf PE, Kim JS, Prakash R, McAuley training, but not voluntary exercise, improves skilled reaching after
E, et al. Aerobic exercise training increases brain volume in aging unilateral ischemic lesions of the sensorimotor cortex in rats.
humans. J Gerontol A Biol Sci Med Sci. 2006;61(11):1166–70. Neurorehabil Neural Repair. 2008;22(3):250–61. doi:10.1177/
17. Smith PJ, Blumenthal JA, Hoffman BM, Cooper H, Strauman TA, Welsh- 1545968307308551.
Bohmer K, et al. Aerobic exercise and neurocognitive performance: a 34. Seo HG, Kim DY, Park HW, Lee SU, Park SH. Early motor balance
meta-analytic review of randomized controlled trials. Psychosom Med. and coordination training increased synaptophysin in subcortical
2010;72(3):239–52. doi:10.1097/PSY.0b013e3181d14633. regions of the ischemic rat brain. J Korean Med Sci. 2010;25(11):
18. Pang MY, Charlesworth SA, Lau RW, Chung RC. Using aerobic 1638–45. doi:10.3346/jkms.2010.25.11.1638.
exercise to improve health outcomes and quality of life in stroke: 35. Chang HC, Yang YR, Wang SG, Wang RY. Effects of treadmill
evidence-based exercise prescription recommendations. Cerebrovasc training on motor performance and extracellular glutamate level in
Dis. 2013;35(1):7–22. doi:10.1159/000346075. striatum in rats with or without transient middle cerebral artery
19. Dobkin BH. Training and exercise to drive poststroke recovery. Nat occlusion. Behav Brain Res. 2009;205(2):450–5. doi:10.1016/j.bbr.
Clin Pract Neurol. 2008;4(2):76–85. doi:10.1038/ncpneuro0709. 2009.07.033.
20. Mang CS, Campbell KL, Ross CJ, Boyd LA. Promoting 36. Ke Z, Yip SP, Li L, Zheng XX, Tong KY. The effects of voluntary,
neuroplasticity for motor rehabilitation after stroke: considering the involuntary, and forced exercises on brain-derived neurotrophic
Transl. Stroke Res.

factor and motor function recovery: a rat brain ischemia model. PLoS intracerebral hemorrhage in rats. Brain Res. 2010;1355:165–73.
One. 2011;6(2):e16643. doi:10.1371/journal.pone.0016643. doi:10.1016/j.brainres.2010.07.070.
37. Quirie A, Hervieu M, Garnier P, Demougeot C, Mossiat C, Bertrand 53. Liu N, Huang H, Lin F, Chen A, Zhang Y, Chen R, et al. Effects of
N, et al. Comparative effect of treadmill exercise on mature BDNF treadmill exercise on the expression of netrin-1 and its receptors in rat
production in control versus stroke rats. PLoS One. 2012;7(9): brain after cerebral ischemia. Neuroscience. 2011;194:349–58. doi:
e44218. doi:10.1371/journal.pone.0044218. 10.1016/j.neuroscience.2011.07.037.
38. Sakakima H, Khan M, Dhammu TS, Shunmugavel A, Yoshida Y, 54. Shimada H, Hamakawa M, Ishida A, Tamakoshi K, Nakashima H,
Singh I, et al. Stimulation of functional recovery via the mechanisms Ishida K. Low-speed treadmill running exercise improves memory
of neurorepair by S-nitrosoglutathione and motor exercise in a rat function after transient middle cerebral artery occlusion in rats. Behav
model of transient cerebral ischemia and reperfusion. Restor Neurol Brain Res. 2013;243:21–7. doi:10.1016/j.bbr.2012.12.018.
Neurosci. 2012;30(5):383–96. doi:10.3233/RNN-2012-110209. 55. Enzinger C, Dawes H, Johansen-Berg H, Wade D, Bogdanovic M,
39. Kim G, Kim E. The effects of antecedent exercise on motor function Collett J, et al. Brain activity changes associated with treadmill
recovery and brain-derived neurotrophic factor expression after focal training after stroke. Stroke. 2009;40(7):2460–7. doi:10.1161/
cerebral ischemia in rats. J Phys Ther Sci. 2013;25(5):553–6. doi:10. STROKEAHA.109.550053.
1589/jpts.25.553. 56. Luft AR, Macko RF, Forrester LW, Villagra F, Ivey F, Sorkin JD,
40. Zhang QW, Deng XX, Sun X, Xu JX, Sun FY. Exercise promotes et al. Treadmill exercise activates subcortical neural networks and
axon regeneration of newborn striatonigral and corticonigral projec- improves walking after stroke: a randomized controlled trial. Stroke.
tion neurons in rats after ischemic stroke. PLoS One. 2013;8(11): 2008;39(12):3341–50. doi:10.1161/STROKEAHA.108.527531.
e80139. doi:10.1371/journal.pone.0080139. 57. Belur PK, Chang JJ, He S, Emanuel BA, Mack WJ. Emerging
41. Shih PC, Yang YR, Wang RY. Effects of exercise intensity on spatial experimental therapies for intracerebral hemorrhage: targeting mech-
memory performance and hippocampal synaptic plasticity in tran- anisms of secondary brain injury. Neurosurg Focus. 2013;34(5):E9.
sient brain ischemic rats. PLoS One. 2013;8(10):e78163. doi:10. doi:10.3171/2013.2.FOCUS1317.
1371/journal.pone.0078163. 58. Cotman CW, Berchtold NC, Christie LA. Exercise builds brain health:
42. Lan X, Zhang M, Yang W, Zheng Z, Wu Y, Zeng Q, et al. Effect of key roles of growth factor cascades and inflammation. Trends
treadmill exercise on 5-HT, 5-HT1A receptor and brain derived Neurosci. 2007;30(9):464–72. doi:10.1016/j.tins.2007.06.011.
neurotrophic factor in rats after permanent middle cerebral artery 59. Chung JY, Kim MW, Bang MS, Kim M. Increased expression of
occlusion. Neurol Sci. 2013. doi:10.1007/s10072-013-1599-y. neurotrophin 4 following focal cerebral ischemia in adult rat brain
43. Chang HC, Yang YR, Wang PS, Kuo CH, Wang RY. Insulin-like with treadmill exercise. PLoS One. 2013;8(3):e52461. doi:10.1371/
growth factor I signaling for brain recovery and exercise ability in journal.pone.0052461.
brain ischemic rats. Med Sci Sports Exerc. 2011;43(12):2274–80. 60. Je HS, Yang F, Ji Y, Potluri S, Fu XQ, Luo ZG, et al. ProBDNF and
doi:10.1249/MSS.0b013e318223b5d9. mature BDNF as punishment and reward signals for synapse elimi-
44. Zhang L, Hu X, Luo J, Li L, Chen X, Huang R, et al. Physical nation at mouse neuromuscular junctions. J Neurosci. 2013;33(24):
exercise improves functional recovery through mitigation of autoph- 9957–62. doi:10.1523/JNEUROSCI.0163-13.2013.
agy, attenuation of apoptosis and enhancement of neurogenesis after 61. Rossini PM, Calautti C, Pauri F, Baron JC. Post-stroke plastic
MCAO in rats. BMC Neurosci. 2013;14:46. doi:10.1186/1471-2202- reorganisation in the adult brain. Lancet Neurol. 2003;2(8):493–502.
14-46. 62. Lazaridou A, Astrakas L, Mintzopoulos D, Khanchiceh A, Singhal
45. Dahlqvist P, Ronnback A, Risedal A, Nergardh R, Johansson IM, A, Moskowitz M, et al. fMRI as a molecular imaging procedure for
Seckl JR, et al. Effects of postischemic environment on transcription the functional reorganization of motor systems in chronic stroke. Mol
factor and serotonin receptor expression after permanent focal corti- Med Rep. 2013;8(3):775–9. doi:10.3892/mmr.2013.1603.
cal ischemia in rats. Neuroscience. 2003;119(3):643–52. 63. Krakauer JW, Carmichael ST, Corbett D, Wittenberg GF. Getting
46. Chung JY, Kim MW, Bang MS, Kim M. The effect of exercise on neurorehabilitation right: what can be learned from animal models?
trkA in the contralateral hemisphere of the ischemic rat brain. Brain Neurorehabil Neural Repair. 2012;26(8):923–31. doi:10.1177/
Res. 2010;1353:187–93. doi:10.1016/j.brainres.2010.06.057. 1545968312440745.
47. Matsuda F, Sakakima H, Yoshida Y. The effects of early exercise on 64. Canadian Stroke Network. The quality of stroke care in Canada.
brain damage and recovery after focal cerebral infarction in rats. Acta Ottawa, Ontario, Canada;2011.
Physiol (Oxf). 2011;201(2):275–87. doi:10.1111/j.1748-1708.2010. 65. Centre for Disease Control: Prevention. Stroke facts. National Center
02174.x. for Chronic Disease Prevention and Health Promotion, Atlanta, GA,
48. Mizutani K, Sonoda S, Yamada K, Beppu H, Shimpo K. Alteration of USA. 2013. http://www.cdc.gov/stroke/facts.htm.
protein expression profile following voluntary exercise in the 66. Chavez-Valdez R, Martin LJ, Razdan S, Gauda EB, Northington FJ.
perilesional cortex of rats with focal cerebral infarction. Brain Res. Sexual dimorphism in BDNF signaling after neonatal hypoxia-
2011;1416:61–8. doi:10.1016/j.brainres.2011.08.012. ischemia and treatment with necrostatin-1. Neuroscience. 2014;260:
49. Marin R, Williams A, Hale S, Burge B, Mense M, Bauman R, et al. 106–19. doi:10.1016/j.neuroscience.2013.12.023.
The effect of voluntary exercise exposure on histological and neuro- 67. Smith AL, Alexander M, Rosenkrantz TS, Sadek ML, Fitch RH. Sex
behavioral outcomes after ischemic brain injury in the rat. Physiol differences in behavioral outcome following neonatal hypoxia ische-
Behav. 2003;80(2–3):167–75. mia: insights from a clinical meta-analysis and a rodent model of
50. Briones TL, Suh E, Jozsa L, Hattar H, Chai J, Wadowska M. induced hypoxic ischemic brain injury. Exp Neurol. 2014;254C:54–
Behaviorally-induced ultrastructural plasticity in the hippocampal 67. doi:10.1016/j.expneurol.2014.01.003.
region after cerebral ischemia. Brain Res. 2004;997(2):137–46. 68. Timaru-Kast R, Luh C, Gotthardt P, Huang C, Schafer MK,
51. Briones TL, Suh E, Jozsa L, Rogozinska M, Woods J, Wadowska M. Engelhard K, et al. Influence of age on brain edema formation,
Changes in number of synapses and mitochondria in presynaptic secondary brain damage and inflammatory response after brain trau-
terminals in the dentate gyrus following cerebral ischemia and reha- ma in mice. PLoS One. 2012;7(8):e43829. doi:10.1371/journal.pone.
bilitation training. Brain Res. 2005;1033(1):51–7. doi:10.1016/j. 0043829.
brainres.2004.11.017. 69. Jaerve A, Schiwy N, Schmitz C, Mueller HW. Differential effect of
52. Takamatsu Y, Ishida A, Hamakawa M, Tamakoshi K, Jung CG, aging on axon sprouting and regenerative growth in spinal cord
Ishida K. Treadmill running improves motor function and alters injury. Exp Neurol. 2011;231(2):284–94. doi:10.1016/j.expneurol.
dendritic morphology in the striatum after collagenase-induced 2011.07.002.
Transl. Stroke Res.

70. Bramlett HM. Special issue of Translational Stroke: importance of 82. Coelho FM, Pereira DS, Lustosa LP, Silva JP, Dias JMD, Dias RCD,
sex in the pathophysiology and treatment of acute CNS repair. Transl et al. Physical therapy intervention (PTI) increases plasma brain-
Stroke Res. 2013;4(4):379–80. doi:10.1007/s12975-013-0264-3. derived neurotrophic factor (BDNF) levels in non-frail and pre-frail
71. Persson N, Lavebratt C, Wahlin A. Synergy effects of HbA1c and elderly women. Arch Gerontol Geriatr. 2012;54(3):415–20. doi:10.
variants of APOE and BDNFVal66Met explains individual differ- 1016/j.archger.2011.05.014.
ences in memory performance. Neurobiol Learn Mem. 2013;106: 83. Liepert J. Motor cortex excitability in stroke before and after
274–82. doi:10.1016/j.nlm.2013.08.017. constraint-induced movement therapy. Cogn Behav Neurol.
72. Mackay-Lyons M, McDonald A, Matheson J, Eskes G, Klus MA. 2006;19(1):41–7. doi:10.1097/00146965-200603000-00005.
Dual effects of body-weight supported treadmill training on cardio- 84. Yin DZ, Song F, Xu DR, Peterson BS, Sun LM, Men WW, et al.
vascular fitness and walking ability early after stroke: a randomized Patterns in cortical connectivity for determining outcomes in hand
controlled trial. Neurorehabil Neural Repair. 2013;27(7):644–53. doi: function after subcortical stroke. PloS One. 2012;7(12):e52727. doi:
10.1177/1545968313484809. 10.1371/journal.pone.0052727.
73. Stoller O, de Bruin ED, Knols RH, Hunt KJ. Effects of cardiovascular 85. Rossini PM, Dal Forno G. Integrated technology for evaluation of
exercise early after stroke: systematic review and meta-analysis. brain function and neural plasticity. Phys Med Rehabil Clin N Am.
BMC Neurol. 2012;12:45. doi:10.1186/1471-2377-12-45. 2004;15(1):263–306.
74. Biernaskie J, Chernenko G, Corbett D. Efficacy of rehabilitative expe- 86. Thiel A, Anglade C, Hartmann A, Heiss W. Imaging correlates of
rience declines with time after focal ischemic brain injury. J Neurosci. rTMS treatment for post-stroke aphasia: a randomized controlled
2004;24(5):1245–54. doi:10.1523/JNEUROSCI.3834-03.2004. study. Stroke. 2013;44(12):E178–E.
75. Eisenberg DP, Ianni AM, Wei SM, Kohn PD, Kolachana B, Apud J, 87. De Boissezon X, Marie N, Castel-Lacanal E, Marque P, Bezy C, Gros
et al. Hippocampal dysfunction in schizophrenia: association with H, et al. Good recovery from aphasia is also supported by right basal
brain-derived neurotrophic factor genotype. Mol Psychiatry. ganglia: a longitudinal controlled PET study. EJPRM-ESPRM 2008
2013;18(6):631. doi:10.1038/mp.2013.53. award winner. Eur J Phys Rehab Med. 2009;45(4):547–58.
76. Corominas-Roso M, Ramos-Quiroga JA, Ribases M, Sanchez-Mora 88. Phelps ME. Positron emission tomography provides molecular im-
C, Palomar G, Valero S, et al. Decreased serum levels of brain- aging of biological processes. Proc Natl Acad Sci U S A.
derived neurotrophic factor in adults with attention-deficit hyperac- 2000;97(16):9226–33. doi:10.1073/pnas.97.16.9226.
tivity disorder. Int J Neuropsychopharmacol. 2013;30:1–9. doi:10. 89. Backer MV, Levashova Z, Patel V, Jehning BT, Claffey K,
1017/S1461145712001629. Blankenberg FG, et al. Molecular imaging of VEGF receptors in
77. Bejot Y, Mossiat C, Giroud M, Prigent-Tessier A, Marie C. angiogenic vasculature with single-chain VEGF-based probes. Nat
Circulating and brain BDNF levels in stroke rats. Relevance to Med. 2007;13(4):504–9. doi:10.1038/Nm1522.
clinical studies. PLoS One. 2011;6(12):e29405. 90. Fan F, Zhu C, Chen H, Qin W, Ji X, Wang L, et al. Dynamic brain
78. Cho HC, Kim J, Kim S, Son YH, Lee N, Jung SH. The concentra- structural changes after left hemisphere subcortical stroke. Hum
tions of serum, plasma and platelet BDNF are all increased by Brain Mapp. 2013;34(8):1872–81. doi:10.1002/hbm.22034.
treadmill VO(2)max performance in healthy college men. Neurosci 91. Kantak SS, Winstein CJ. Learning-performance distinction and
Lett. 2012;519(1):78–83. doi:10.1016/j.neulet.2012.05.025. memory processes for motor skills: a focused review and perspective.
79. Araya AV, Orellana X, Godoy D, Soto L, Fiedler J. Effect of exercise Behav Brain Res. 2012;228(1):219–31. doi:10.1016/j.bbr.2011.11.
on circulating levels of brain-derived neurotrophic factor (BDNF) in 028.
overweight and obese subjects. Horm Metab Res. 2013;45(7):541–4. 92. Koopman AD, Eken MM, van Bezeij T, Valent LJ, Houdijk H. Does
doi:10.1055/s-0032-1333237. clinical rehabilitation impose sufficient cardiorespiratory strain to
80. Voss MW, Erickson KI, Prakash RS, Chaddock L, Kim JS, Alves H, improve aerobic fitness? J Rehabil Med. 2013;45(1):92–8. doi:10.
et al. Neurobiological markers of exercise-related brain plasticity in 2340/16501977-1072.
older adults. Brain Behav Immun. 2013;28:90–9. doi:10.1016/j.bbi. 93. MacKay-Lyons MJ, Makrides L. Cardiovascular stress during a
2012.10.021. contemporary stroke rehabilitation program: is the intensity adequate
81. Bansi J, Bloch W, Gamper U, Kesselring J. Training in MS: influence to induce a training effect? Arch Phys Med Rehabil. 2002;83(10):
of two different endurance training protocols (aquatic versus over- 1378–83.
land) on cytokine and neurotrophin concentrations during three week 94. Williams G, Kahn M, Randall A. Strength training for walking in
randomized controlled trial. Mult Scler. 2013;19(5):613–21. doi:10. neurologic rehabilitation is not task specific: a focused review. Am J
1177/1352458512458605. Phys Med Rehabil. 2014. doi:10.1097/PHM.0000000000000058.

View publication stats

You might also like