Critical Review

Physical Exercise and Cognitive Recovery in

Acquired Brain Injury: A Review of the Literature
Jennifer M. Devine, MD, Ross D. Zafonte, DO

Objective: Physical exercise has been shown to play an ever-broadening role in the
maintenance of overall health and has been implicated in the preservation of cognitive
function in both healthy elderly and demented populations. Animal and human studies of
acquired brain injury (ABI) from trauma or vascular causes also suggest a possible role for
physical exercise in enhancing cognitive recovery.
Data Sources: A review of the literature was conducted to explore the current under-
standing of how physical exercise impacts the molecular, functional, and neuroanatomic
status of both intact and brain-injured animals and humans.
Study Selection: Searches of the MEDLINE, CINHAL, and PsychInfo databases yielded an
extensive collection of animal studies of physical exercise in ABI. Animal studies strongly tie
physical exercise to the upregulation of multiple neural growth factor pathways in brain-injured
animals, resulting in both hippocampal neurogenesis and functional improvements in memory.
Data Extraction: A search of the same databases for publications involving physical
exercise in human subjects with ABI yielded 24 prospective and retrospective studies.
Data Synthesis: Four of these evaluated cognitive outcomes in persons with ABI who
were involved in physical exercise. Three studies cited a positive association between
exercise and improvements in cognitive function, whereas one observed no effect. Human
exercise interventions varied greatly in duration, intensity, and level of subject supervision,
and tools for assessing neurocognitive changes were inconsistent.
Conclusions: There is strong evidence in animal ABI models that physical exercise facilitates
neurocognitive recovery. Physical exercise interventions are safe in the subacute and rehabilita-
tive phases of recovery for humans with ABI. In light of strong evidence of positive effects in
animal studies, more controlled, prospective human interventions are warranted to better
explore the neurocognitive effects of physical exercise on persons with ABI.

INTRODUCTION
The role of exercise in the maintenance of cardiovascular health has been well established.
Epidemiological studies suggest that individuals who expend energy greater than 2000 kCal
per week in some form of moderate-to-intense exercise significantly lower their risk of
mortality from all causes [1] and decrease their cardiovascular morbidity when compared
with those who exercise less intensely or frequently [2]. Further benefits of exercise include J.M.D. Department of Physical Medicine &
Rehabilitation, Spaulding Rehabilitation Hos-
improvements in cardiovascular control [2], glycemic control [3], weight management [4], pital, Harvard Medical School, Boston, MA
bone density [5], depression [6], and breast and colon cancer [7]. In light of this evidence, Disclosure: nothing to disclose
the American College of Sports Medicine and the American Heart Association currently R.D.Z. Department of Physical Medicine & Re-
recommend that all adults aged 18 to 65 years participate in 30 minutes of moderate-intense habilitation, Spaulding Rehabilitation Hospital,
Harvard Medical School, 125 Nashua St, Bos-
aerobic (endurance) physical activity per day at least 5 days per week [8]. Adding to the
ton, MA 02214. Address correspondence to:
body of evidence showing its many established health benefits are findings that regular R.D.Z.; e-mail: [email protected]
exercise may also have a positive impact on memory and cognition; thus, investigational Disclosure: 2A, BHR and Allergan; 7A, Neuro-
healing - Drug Phase II Study Cosponsored by
interest into the specific effects of exercise on cognitive function is growing.
FDA, Funded by Orphan Drug; 8B, NIH, NIDRR,
Numerous population-based studies suggest a decreased incidence of cognitive decline in DOD-Supported Research
individuals who participate in physically and cognitively stimulating activities later in life [9-13]. Disclosure Key can be found on the Table of
Friedland et al [13] further demonstrated that a decreased risk of cognitive delay is seen in elders Contents and at www.pmrjournal.org
who participated in cardiovascular conditioning throughout midlife. Finally, in an effort to Submitted for publication October 1, 2008;
further explore potential dose-dependency for this benefit, Larson et al [14] showed a decreased accepted March 29, 2009.

PM&R © 2009 by the American Academy of Physical Medicine and Rehabilitation

1934-1482/09/$36.00 Vol. 1, 560-575, June 2009
560
Printed in U.S.A. DOI: 10.1016/j.pmrj.2009.03.015
PM&R Vol. 1, Iss. 6, 2009 561

relative risk of dementia in elders who exercise more than 3 days and effectively exercise at the same intensity as recom-
per week. Most recent, Burns et al [15] showed that individuals mended for the general population. Recent studies of con-
with early stage Alzheimer dementia who had better cardiovas- temporary stroke rehabilitation programs, for example, find
cular fitness experienced less cortical loss than their less-fit that traditional inpatient stroke therapy is not sufficiently
counterparts. Although many investigators have focused their intense to induce a cardiovascular training effect [31]. Add-
attention on older populations at risk for dementia, some evi- ing to the physical impediments to exercise, external barriers
dence suggests that the connection between aerobic condition- to participation and safety concerns can further decrease
ing and cognitive function may not be restricted to the elderly. A access to and compliance with cardiovascular conditioning
population-based study by Van Boxtel et al [16] showed a regimens in people with brain injuries [32].
positive association between age- and gender-adjusted maxi- A recent Cochrane review of cardiorespiratory training in
mum oxygen consumption (V̇O2max) levels and performance people with traumatic brain injury (TBI) concluded that aerobic
on tests of complex cognitive speed in a broad range of ages. conditioning is a safe intervention when applied in the rehabil-
Further evidence taken from the British 1946 birth cohort itation phase of moderate-to-severe ABI [33]. However, because
shows a positive impact of physical exercise on memory in only a subset of these investigators followed markers of exercise
mid-life [17]. Although the majority of the literature appears to efficacy, such as peak V̇O2, authors could not fully characterize
focus on endurance training, resistance exercise also may have the physiologic benefit of the exercise intervention. In addition,
the potential to improve free recall and recognition [18]. To- although several of the studies included in the analysis found a
gether, these findings have spurred questions about the exact positive effect of exercise on mood and social integration, none
role and mechanism of exercise as a prophylaxis against cogni- conducted a rigorous assessment of the possible role of aerobic
tive decline. conditioning on cognitive function. Meta-analyses of cardiovas-
There is broad speculation as to the possible mechanism by cular conditioning for stroke sufferers also support the safety of
which exercise might impact cognition. Churchill et al [19] such interventions and suggest other benefits on mood and the
presented specific evidence showing that improved aerobic fit- ability to perform activities of daily living [34]. Nonetheless,
ness increased cerebral blood flow, oxygen extraction, and glu- direct evidence of the role of controlled physical exercise specif-
cose utilization. Other investigators have proposed that the ically on cognitive function in ABI is not well represented in the
apparent cognitive benefits of physical exercise were likely the literature.
byproduct of other physiologic responses to exercise [20-23]. Along with concerns about overall musculoskeletal safety
However, early work by Cotman and Berchtold [24] also and efficacy of physical exercise come questions about the
showed that exercise improved neural plasticity, suggesting that metabolic demands it may make on damaged areas of the
cardiovascular conditioning itself may improve cognition by brain. Specifically, there is legitimate concern that the in-
way of specific changes to the neurochemical microenviron- creased overall oxidative stress of exercise might attenuate
ment that may ultimately lead to neurogenesis. recovery from the neurometabolic cascade of TBI. Current
Focal neurogenesis in the adult brain, the pursuit of understanding of the pathophysiology of brain injury sug-
modulatory factors that encourage it, and the question of gests that biomechanical trauma to the brain causes the
whether new cells confer function are all areas of active indiscriminate release of excitatory neurotransmitters, which
investigation in the neuroscientific community. Histopatho- in turn acutely increases glucose metabolism and decreases
logic evidence in animals [25,26] and humans [27] docu- cerebral blood flow, creating an energy mismatch. Although
ments neurogenesis in the adult hippocampus, a region of the it is during this period, hours to days after injury, when the
brain involved in spatial and temporal memory. In animals, brain is thought to be least able to respond to secondary
several studies have shown that proliferative activity in this insults, there is also evidence that decreased cerebral blood
area is enhanced by voluntary exercise, and that this phe- flow, decreased glycolysis, mitochondrial dysfunction, and
nomenon correlates directly with improved performances on intracellular calcium accumulation may persist for weeks,
temporospatial memory tasks [28-30].Given these findings, rendering the recovering brain persistently more sensitive to
there is understandable question and interest into whether the metabolic demands, such as those made by physical
the human brain holds similar capacity for focal neurogenesis exercise [35-37]. Clinical evidence to support this has been
and whether physical exercise may confer similar neurocog- reported in athletes with postconcussive syndromes, in
nitive benefits to humans. which the increased oxidative stress of aggressive physical
In humans with brain injury caused by trauma or cerebro- exercise in the days after injury was shown to worsen neuro-
vascular events, memory and cognitive impairments can limit cognitive outcomes [38]. Additional studies in animal mod-
function. Given the strong need for new therapies to address els of TBI suggest that hippocampal plasticity-related protein
these deficits, there is understandable interest in the possible expression is curtailed in animals exercised acutely after the
role that physical exercise may play in cognitive recovery. induction of injury, and that this change corresponds to
However, people with acquired brain injury (ABI) can also poorer performance on functional memory tasks [39,40].
show a decreased capacity for exercise because of the periph- Thus, there is legitimate reason for concern about the overall
eral and central sequelae of their injuries. Spasticity, paresis, safety of physical exercise for people in recovery from ABI.
and peripheral sensory and central balance disturbances can The goals of this review are 3-fold. First, it will introduce
raise concerns about the ability of persons with ABI to safely the reader to the current literature on how physical exercise
562 Devine and Zafonte PHYSICAL EXERCISE AND COGNITIVE RECOVERY IN ACQUIRED BRAIN INJURY

can impact the molecular, functional, and neuroanatomic accident (CVA), stroke, focal cerebral ischemia, global isch-
status of intact animals and those with ABI from a variety of emia] and [exercise, running wheel, memory, hippocampus].
causes. This will include a discussion of the molecular mech- All studies in which physical exercise interventions were
anisms by which physical exercise is thought to impact given to animals with ABI are included in this discussion. To
cellular regeneration in the adult brain and the possible provide readers the most current understanding of the mo-
pathways by which this process can affect cognitive function. lecular, anatomic, and behavioral sequelae of ABI, additional
Animal models of ABI are useful and relevant to the study of studies of brain-injured animals were also included if these
human ABI rehabilitation because they provide investigators studies were thought to provide novel or relevant informa-
the opportunity to obtain information, specimens, and data tion to this discussion. Human studies were obtained using
that are largely unobtainable from humans with brain injury. the following terms: [brain injury, brain damage, CVA,
Specifically, animal studies can be carefully controlled and stroke] and [exercise, aerobic, fitness, physical activity] and
data can be gathered as a function of time after the induction [cognition, cognitive function, memory], which were identi-
of the brain injury itself. Animal models for ABI also allow fied as the most effective search terms. Although searches
investigators to simultaneously consider the molecular, ana- were not restricted to original research, review articles were
tomic, and functional sequelae of brain injuries without other not included per se in the evaluation. Rather, studies refer-
potentially confounding comorbidities, such as concomitant enced in all review articles were checked for relevance and
disease or pharmacologic interference. Although not directly included in the discussion if appropriate. Because the data-
analogous to human disease, the findings obtained under base of human studies of physical exercise and cognition was
these carefully controlled conditions can provide valuable not extensive, all studies involving physical exercise inter-
information as well as suggest possible future paths of inquiry ventions in humans with ABI are presented in this review.
and intervention in the treatment of humans. A discussion of Those studies that measured cognitive outcomes specifically
this literature is especially vital for the clinically oriented in this population are discussed individually. Reference lists
reader, as it provides the necessary background to appreciate of all included studies were mined for previously unidenti-
the rationale for physical exercise studies in humans with fied relevant studies.
ABI. The second goal of this article is to review the available
literature showing the impact of physical exercise on persons
with ABI. Because this literature is limited, a discussion of all RESULTS
existing studies involving physical exercise interventions in
Animal Studies
people with brain injury, as well as retrospective surveys of
the exercise habits of brain-injured individuals as they relate The neuroscientific community has long pursued rodent mod-
to cognitive function, will come first. This review will further els as a means of understanding the cellular and anatomic
present the subset of these exercise studies in which mea- physiology of the brain, both in its native state and under
sures of cognitive function were made and discuss the find- changing environmental conditions. Rodent studies were
ings of these measures, with special attention to the metrics among the first to demonstrate a capacity for neurogenesis in
used to assess cognitive function. Finally, the implications of specific areas of the adult brain [41], prompting similar discov-
these data for future therapeutic interventions in people with eries in nonhuman primates [42] and humans [27]. These
ABI will be summarized and discussed. findings raised immediate questions about the factors responsi-
ble for stimulating neurogenesis and the functional role new
cells might play in the adult brain. Subsequent work with
METHODS
rodents showed that environmental enrichment—a multifacto-
For the purpose of this review, exercise is defined as physical rial stimulus that includes physical activity, a larger, more varied
movement designed to improve cardiovascular conditioning living space, and social interactions with other rodents— helped
or global musculoskeletal strength. Other forms of exercise, to stimulate neurogenesis specifically in the adult hippocampus
such as cognitive exercises, physical and occupational ther- [43]. Van Praag et al [25] further analyzed the components of
apy, skill- or task-specific exercises or therapy, meditation, environmental enrichment and found that voluntary exercise
stretching, and relaxation are not discussed, unless they were alone was more effective in stimulating hippocampal neurogen-
compared directly with interventions in which physical ex- esis than were all other factors of environmental enrichment.
ercise was used. Papers discussed in this article were obtained Further studies suggested this neurogenesis was associated with
through searches of MEDLINE, Cochrane, CINAHL, and the upregulation of brain-derived neurotrophic factor (BDNF), a
PsychINFO databases. For all searches, articles included ubiquitous central nervous system (CNS) growth factor, in the
were published between 1965 and 2008 in English. For the dentate gyrus region of the rodent hippocampus and that this
results section on animal studies, the search was limited to phenomenon was linked to improved performance on tempo-
animal studies; the results section on human studies was rospatial memory tasks [28,30]. The link between cellular re-
limited to studies of human subjects. After limiting the search generation and improved function generated significant interest
to animal studies, literature reviewed in the results section on in the signaling pathways by which exercise stimulates cell
animal studies was collected using the most effective search growth in the hippocampus and in the potential capacity for
terms, which included [brain injury, TBI, cerebrovascular new neurons to function in the setting of traumatic cell loss. This
PM&R Vol. 1, Iss. 6, 2009 563

also provided circumstantial evidence of the possible role for showed that the same biphasic signaling pathway observed in
exercise as a therapeutic intervention in neurorecovery. exercise-induced neurogenesis is also triggered by transient
Many investigators have contributed to an understanding global ischemia in the rodent hippocampal dentate gyrus.
of the molecular and cellular changes seen in the hippocam- This pathway relies on ischemia-induced activation of calcium/
pus after exercise. This work has greatly contributed to an calmodulin-dependent protein kinase IV, which phosphoryl-
understanding of how exercise might provide improved out- ates CREB, a binding protein that in turn activates BDNF
comes after traumatic or ischemic insult, as well as prophy- directly and through its own positive feedback loop. Further
laxis against future brain injury. Rodent studies performed by work by Ploughman et al [49] established an additional role
Molteni et al [44,45] have shown that voluntary exercise for insulin-like growth factor I (ILGF-1) in exercise-mediated
upregulates multiple proteins involved in signal transduction neurogenesis in the rodent cerebral ischemia model. This
(Ca2⫹/calmodulin-dependent protein kinase II, mitogen-ac- study showed that animals exercising at varying intensities
tivated extracellular signal-regulated protein kinase, protein after induced ischemia produced different profiles of neuro-
kinase C), synaptic trafficking (synapsin-1, synaptogamin, trophic factors, with moderate intensity exercise (walking)
syntaxin), transcriptional regulation (cyclic AMP response exhibiting the most potent effect on contralateral hippocam-
element binding protein), neurotrophins (BDNF), and neu- pal levels of neurotrophins BDNF, synapsin-1, and ILGF-1.
rotransmitters (upregulation of N-methyl-D-aspartate recep- However, levels of the same neurotrophins exhibited in-
tors 2A and 2B; downregulation of ␥-aminobutyric acid creases in the ipsilateral hippocampus that were dose-depen-
[GABA] receptor A and glutamate decarboxylase). Current dent with increasing intensities of voluntary running.
work suggests that the molecular mechanism by which exer- Ploughman’s group also showed increasing levels of ILGF-1
cise stimulates cellular regeneration occurs via the upregula- in the hippocampus and sensorimotor cortex with exercise
tion of BDNF through a biphasic pathway that includes a corresponded to decreased ILGF-1 levels in the periphery,
cyclic-AMP responsive element binding protein (CREB) and suggesting that exercise works to increase the permeability of
a calmodulin-dependent protein kinase. It is thought that a the blood-brain barrier to neurotrophic and growth factors in
calcium influx in the setting of metabolic demand stimulates the setting of focal ischemic insult [49]. Efforts to explore the
the calmodulin-dependent protein kinase, which in turn functional gains conferred by various intensities of exercise in
phosphorylates CREB. CREB subsequently drives the synthe- the focal ischemic model were undertaken by Luo et al [50],
sis of BDNF, providing further positive feedback to CREB who showed that animals who performed voluntary wheel
[46]. Further work also suggests that multiple proteins asso- running ad libitum performed better on water maze testing than
ciated with energy metabolism and synaptic plasticity are did those who were exposed to twice-daily, high-intensity
upregulated in the setting of physical exercise. Ding et al swimming trials. This work further showed that animals who
[47,48] identified proteins associated with glycolysis, ATP undertook voluntary wheel running showed increased survival
synthesis, ATP transduction, and glutamate turnover that of new cells in the dentate gyrus as well as upregulated phos-
were increased after a brief, voluntary exercise period. phorylation of CREB. Specifically, it was found that animals that
In addition to the extensive study of the factors responsi- performed well on water maze testing showed a higher number
ble for exercise-induced neurogenesis, still other investiga- of surviving new cells in the dentate gyrus [50]. In addition to
tors have contributed to an understanding of the temporal establishing a link among physical activity, cellular survival, and
relationship among cell signaling molecules, cellular sur- improved cognitive function, this work suggested that im-
vival, and functional gains. Epp et al [29] showed that the age proved fitness alone could not fully explain the improved per-
of new cells in the hippocampus played a role in how sensi- formance on water maze testing.
tive they were to external stimulation by spatial learning An area of particular interest in the study of exercise in
challenges. This finding suggests that cell survival in the CVA damage focuses on the blood-brain barrier and the
hippocampus is dependent on stimulation at key time points molecular mechanisms by which preischemic exercise ap-
during the window of cellular maturation. Table 1 summa- pears to reduce the harmful effects of ischemic bouts. In
rizes the available literature involving hippocampal molecu- earlier studies, Lee et al [51] showed that prehemorrhagic
lar, cellular, and functional memory changes in rodent mod- exercise decreases cerebral lesion size when experimental
els of exercise and brain injury. animals are pretrained before the induction of intracerebral
hemorrhage. Davis et al [52] subsequently showed that that
Animal Models for Exercise in Ischemic dysfunction of the blood-brain barrier in the setting of stroke
is mitigated by preconditioning with treadmill exercise by
Brain Injury
way of upregulating the expression of collagen IV. The work
Rodent models of ischemic injury have shown that regions of of Ding et al [53] also provides strong and compelling evi-
the brain demonstrating the capacity for neurogenesis also dence that preischemic exercise upregulates mRNA for vas-
appear in many cases to be resistant to ischemic damage. cular endothelial growth factor (VEGF) and members of the
Evidence gained from the rodent middle cerebral artery angiopoietin family, while also downregulating the expres-
occlusion model suggests that ischemic resistance in the sion of tumor necrosis factor-␣ (TNF-␣) receptors in dam-
hippocampus is conferred via similar molecular pathways as aged brain tissue. Although these findings do not have direct
seen in exercise-induced neurogenesis. Blanquet et al [46] bearing on the possible therapeutic effects of exercise after
564 Devine and Zafonte PHYSICAL EXERCISE AND COGNITIVE RECOVERY IN ACQUIRED BRAIN INJURY

Table 1. Molecular, cellular, and functional changes seen in exercise and animal models of acquired brain injury

Ischemic Brain Injury Traumatic Brain Injury

Primary Preischemic Postischemia Postischemia Primary Posttraumatic

Exercise Ischemia Low-Intensity Low-Intensity High-Intensity Trauma Low-Intensity
Alone Response Exercise Exercise Exercise Response Exercise
Molecular Changes
Brain-derived ⫹ ⫹ ⫹⫹ ⫹ ⫹⫹
neurotrophic factor
(BDNF)
[26,28,30,40,44,46,49,54,55]
Insulin-like growth factor-1 ⫹ ⫹ ⫹⫹
(ILGF-1) [48,49]
Fibroblast growth factor-2 ⫹
(FGF-2) [54]
Vascular endothelial ⫹
growth factor (VEGF)
[53]
Cyclic-AMP response ⫹ ⫹ ⫹ ⫹ ⫺ ⫹ ⫹⫹
element binding protein
(CREB) [39,40,44-46,50]
Ca2⫹-dependent protein ⫹ ⫹
kinase [44-46,48]
Growth-associated ⫺ ⫹
protein 43 (GAP 43) [65]
Protein kinase C [39] ⫺ ⫹
Synapsin-1 ⫹ ⫹ ⫹⫹
[39,44,45,48,49,55]
Synaptogamin [26,44,45] ⫹
Syntaxin [44,45] ⫹
Synaptophysin (SYP) [65] ⫺ ⫹
Collagen IV [52] ⫺ ⫹
Matrix metalloproteinase ⫹ ⫺
(MMP) 9 [52]
Tumor necrosis factor-␣ ⫹ ⫺
(TNF-␣) [53]
Myelin-associated ⫹ ⫹ ⫺
glycoprotein (MAG)
[57-59,65]
Nogo-A [57-59,65] ⫹ ⫹ ⫺
N-methyl-D-aspartate ⫹
receptor (NMDAR-2A,
NMDAR-2B) [44,45]
␥-Aminobutyric acid ⫺
receptor (GABA-A)
[44,45]
Glutamate ⫺
decarboxylase (GAD65)
[44,45]
Cellular Changes
Hippocampal ⫹ ⫹ ⫺ ⫹ ⫹⫹
neurogenesis
[50,54,62-64]
Cerebral ⫹
cortex—dendritic
arborization [97]
Functional Changes
Performance on Morris ⫹⫹ ⫺ ⫹ ⫺ ⫺ ⫹
Water Maze (MWM)
[25,28,30,40,43,48,49,50,54]

CVA, they do provide strong evidence for the molecular the molecular signaling cascade involved in cellular regener-
mechanism for exercise in secondary stroke prevention. ation. Some of the earliest work performed in this area by
Gomez-Pinilla, So, and Kesslak [54] showed that spatial
learning and physical activity induced cerebral fibroblast
Animal Models for Exercise in TBI growth factors in intact animals, which correlated with in-
Extensive work in rodent models for TBI has further charac- creased hippocampal astrocyte density. Subsequent work by
terized the pathways by which BDNF stimulates neuronal Griesbach et al [55] demonstrated upregulation of BDNF and
recovery and has provided valuable insight into the timing of synapsin 1 in the hippocampus of rodents after induced mild
PM&R Vol. 1, Iss. 6, 2009 565

TBI. However, there was also evidence that exercise too early tion can occur rapidly in the acute and subacute periods after
after brain injury, within the first few days of injury, dis- injury, rendering potential benefits from exercise easier to
rupted this process [39]. Building on these findings, Gries- detect in real time. Moreover, in the subacute TBI recovery
bach et al [40] sought to correlate growth factor upregulation phase, potential subjects are often involved in either inpa-
to exercise at specific times after injury, and to characterize tient or outpatient rehabilitation and may thus be easier to
the link between growth factor upregulation to functional attract and retain in an exercise intervention. Studies of the
improvements in spatial memory tasks. This study demon- effects of physical exercise on humans with ABI can be
strated upregulation of BDNF as well as CREB and synapsin 1 divided by the subjects’ mechanism of injury (vascular or
in the rodent hippocampus after fluid percussion injury traumatic), severity of injury, and time since injury. For the
(FPI). This upregulation was attenuated in animals that par- purposes of this article, the available literature has been
ticipated in voluntary exercise more than 14 days after injury. grouped by cause of injury and is presented in Tables 2
However, animals that participated in voluntary exercise through 4. A total of 1290 subjects with ABI were evaluated
early after their injury (days 0-6) experienced a decrease in in 23 prospective studies and 1 retrospective survey. Eleven
CREB and synapsin 1. This change in the molecular response studies involved interventions that focused principally on
to injury was accompanied by an observed decrease in per- improving cardiovascular endurance [66-76]. Three studies
formance on water maze memory exercises. Although the focused exclusively on muscle strengthening [77-79]. Ten
direct mechanism by which exercise triggers growth factor studies applied interventions with components of both, or
transcription is not clear, building evidence suggests it is applied “physical exercise” interventions that were oriented
through the generation of oxidative stress. Davis et al [56] toward functional strength or function-specific activities, or
showed that oxidative stress could play a role in impairing not otherwise characterized [34,80-88]. No adverse events
recovery in rodents after moderate controlled cortical impact were reported secondary to exercise intervention. Of the 24
(CCI). Specifically, they noted that animals who fasted in the studies included in this review, 4 examined the neurocogni-
first 24 hours after injury showed a decrease in biomarkers of tive effects of physical exercise on individuals with ABI.
oxidative stress that correlated with improved performance Despite a wide diversity of exercise interventions, time after
on cognitive assessments. Important work has also focused injury, and primary outcome measures, available literature
on the role of exercise in regulating inhibitors of neuronal strongly suggests that individuals with ABI can safely participate
growth in the setting of TBI. Myelin-associated glycoprotein in physical exercise. Strong evidence exists to support the safety
(MAG) and Nogo-A have been described as potent inhibitors of cardiovascular endurance exercise for people with ABI. Al-
of neuronal growth in rodent CVA models [57,58] as well as though fewer studies involving muscular strength training have
in nonhuman primate CVA models [59], and there is evi- been reported in this population, no adverse events were asso-
dence that these pathways are also activated in rodent models ciated with this type of physical exercise. A subset of the litera-
for TBI [60,61]. In both CVA and TBI models, evidence ture shows that exercise interventions can improve indices of
suggests that interference with these pathways can spur neu- cardiovascular health, such as hemodynamics and peak V̇O2, as
ronal cell growth [62-64]. Work by Chytrova et al [65] well as peak power, maximum strength, and range of motion. Of
showed that FPI elevated hippocampal levels of MAG and the 4 studies assessing the neurocognitive effects of exercise on
Nogo-A but that voluntary wheel running overcame MAG subjects with brain injury, 3 found that exercising subjects
and Nogo-A upregulation. Furthermore, blockade of BDNF scored better on measures of memory and cognition than did
restored MAG and Nogo-A levels to those seen in sedentary the nonexercising brain-injured controls. None of these studies
animal after FPI, suggesting that BDNF acts to promote the included both neurocognitive and cardiovascular outcome mea-
growth of hippocampal cells at least in part by downregulat- sures. Numerous authors raised the question of whether their
ing inhibitory pathways [65]. interventions were of sufficient intensity and duration to achieve
specific outcomes. Moreover, it was plausibly suggested that the
intensity or duration needed to achieve threshold effects in one
Human Studies
determinant were not equal when measuring another. Given the
A growing body of population-based, retrospective studies relative paucity of literature exploring the role of physical exer-
supports the role for physical exercise on human cognitive cise on cognitive recovery, there is no way of stating, based on
function. However, as compelling and numerous as these the current literature, whether and how exercise may potentiate
studies are, they show mere association, not causation. Pro- known pathways for recovery, such as plasticity, regeneration,
spective studies of memory and cognition during controlled or functional variation.
exercise interventions offer some insight into the effects of
exercise. But because cognitive decline occurs over the
Exercise in Acquired Brain Injury
course of years, applying a consistent intervention during
this time period poses a significant logistical challenge. Hu- Seven studies of physical exercise were done involving subjects
man studies of physical exercise after ABI offer a unique with brain injury as a result of trauma, hemorrhage, or CVA and
window into the possible benefits of exercise on cognition for are outlined in Table 2 [68-71,80-82]. Two hundred eleven
several important reasons. In hemorrhagic, ischemic, and subjects between 12 and greater than 52 weeks after injury
traumatic brain injury, the restoration of function and cogni- submitted to exercise interventions lasting 8 to 12 weeks. Sev-
566 Devine and Zafonte PHYSICAL EXERCISE AND COGNITIVE RECOVERY IN ACQUIRED BRAIN INJURY

Table 2. Physical exercise in humans with acquired brain injury (trauma or CVA)

Study Type/Participants Intervention

Bateman Single blind RCT 12 weeks intervention
et al, 157 consecutive inpatients with moderate to severe acquired Group 1: 3 ⫻ 30 minutes per week cycle ergometer sessions
2001 [68] brain injury due to trauma or vascular event (GCS ⱕ 13). per week. Session duration and workload intensity
Subjects averaged 25.5 weeks (control group) and 22.2 increased individually to 60-80% of age-predicted
weeks (experimental group) after injury maximum heart rate.
Group 2: 3 ⫻ 30 minutes per week relaxation therapy
sessions
Jackson Retrospective analysis of subjects participating in Bateman et 12 weeks intervention
et al, al RCT 3 ⫻ 30 minutes per week cycle ergometer sessions per
2001 [69] 90 participants in exercise intervention arm of RCT conducted week. Session duration and workload intensity increased
by Bateman et al individually to 60-80% of age-predicted maximum heart
rate.
Driver RCT 8 weeks intervention
et al, 16 outpatients age 33-45 with acquired brain injury from one Group 1: Aquatics therapy 3 ⫻ 60 min/wk with individual
2004 [70] rehabilitation facility. Ranchos Los Amigos scales ⱖ 6, ⱖ 12 instruction. Aquatic exercises fell into one of three
months since injury categories: aerobic, resistance and a combination of the
two. Different specific exercises were administered at
random for each session. Heart rate monitors were used
throughout exercise to insure intensity was kept between
50-70% of age-predicted maximum.
Group 2: Vocational rehabilitation.

Driver RCT 8 weeks intervention

et al, 18 outpatients age 25-47 with acquired brain injury from one Group 1: Aquatics therapy 3 ⫻ 60 min/wk with individual
2006 [71] rehabilitation facility. Ranchos Los Amigos scales ⱖ 6, ⱖ 12 instruction. Heart rate monitors were used throughout
months since injury. exercise to insure intensity was kept between 50-70% of
age-predicted maximum.
Group 2: Vocational rehabilitation.

Salazar Single blind RCT 8 week intervention

et al, 120 consecutive active duty military inpatients with moderate Group 1: Standardized 8-week inpatient cognitive
2000 [80] to severe brain injury (GCS ⱕ 13) or focal cerebral contusion rehabilitation program, including fitness, planning and
or hemorrhage visible on CT/MRI within 3 months of organization, cognitive skills, work skills, medication and
randomization milieu groups.
Group 2: Limited home rehabilitation program with weekly
telephone support from a psychiatric nurse.
Braverman Subgroup analysis of the treatment arm of Salazar et al RCT; 8 weeks intervention
et al, designed to better describe standardized 8-week inpatient Group 1: Standardized 8-week inpatient cognitive
1999 [81] rehabilitation used by Salazar et al. rehabilitation program, including fitness, planning and
67 Active-duty military personnel within 3 months of moderate organization, cognitive skills, work skills, medication, and
to severe TBI (GCS ⱕ 12 or posttraumatic amnesia 24 h or milieu groups.
focal cerebral contusion hemorrhage on CT/MRI), with
Rancho Los Amigos 7 within 90 days of injury.
Warden Subgroup analysis of the home-based treatment arm of 8 weeks intervention
et al, Salazar et al RCT; designed to better describe standardized Group 1: Individualized home-based multifaceted therapy
2000 [82] 8-week home-based rehabilitation program used by Salazar program monitored by weekly telephone visits with a
et al. master’s-level psychiatric nurse. Individualized therapy
53 Active-duty military personnel within 3 months of moderate included at least 30 minutes of physical exercise daily, at
to severe TBI (GCS ⱕ 12 or posttraumatic amnesia 24 h or individual’s own pace. Adherence to exercise, as with all
focal cerebral contusion hemorrhage on CT/MRI), with home-based therapy, was self-reported.
Rancho Los Amigos 7 within 90 days of injury.

eral teams performed retrospective or subgroup analyses of subjects underwent a standardized, 8-week, inpatient, cognitive
existing studies [69,81,82]; although each study is reported rehabilitation program that included fitness but did not specify
separately by author, subjects who are included in more than the intensity or duration of training. Although no direct mea-
one study were only counted once. One study had components surement of neurocognitive status was made in any of the
of resistance and cardiovascular endurance training [70]; all studies completed of this population, no differences were seen
other studies focused strictly on cardiovascular endurance. in return to work or fitness for military duty after 6, 12, or 24
Salazar et al [80], Braverman et al [81], and Warden et al [82] all months [80-82].
evaluated an active-duty military population in which their
primary outcome measure was fitness for military duty or work.
Exercise in CVA
Experimental groups underwent an 8-week, individualized,
home-based therapy program that included at least 30 minutes Thirteen studies of physical exercise in persons after CVA
of daily physical exercise at the subject’s own pace. Control were identified and described in Table 3 [34,72-79,83,86-
PM&R Vol. 1, Iss. 6, 2009 567

Outcome Measures Findings Neurocognitive Effects

Primary: peak heart rate (measured through graded Experimental group showed improvements in peak Not assessed.
submaximal cycle exercise test), Ashworth scores, heart rate.
Berg balance. No difference was seen in functional
Secondary: Barthel index (disability), FIM, independence, mobility, psychological function,
Nottingham Extended ADLs, Hospital Anxiety and either immediately after intervention or 12 weeks
Depression Scale after intervention ended.

Primary: number of sessions at which full 30 minutes 55 of 90 patients completed full intervention. 35 Not assessed.
of training was achieved. withdrew. 45 were able to perform at least 20
Secondary: average length of training session, minutes of training per session. 18 were able to
percentage of time spent ⱖ 60% age-predicted attain 60% age-predicted max.
maximum heart rate
Primary: cardiovascular endurance (submaximal Aquatic therapy intervention caused significant Not assessed.
cycle ergometry), body composition, muscular increases in range of motion in bilateral elbow
strength (grip dynamometer), flexibility (sit and flexion, bilateral hip flexion, right-sided hip
reach, goniometer). extension, bilateral knee flexion. Significant
improvements in sit-and-reach and dynamometer
were seen. Improvements in body composition,
cycle ergometry time, and total wattage were
also seen.
No significant changes in outcome measures were
reported for the vocational rehabilitation group.
Primary: Health Promoting Lifestyles Profile II HPLP-II questionnaire showed aquatics therapy Not assessed.
Questionnaire , Physical Self-Description group had significant increases in health
Questionnaire (PSDQ) administered before and responsibility, physical activity, nutrition, spiritual
after intervention. growth, and interpersonal relationships. No
significant changes were seen in vocational
rehabilitation group.
PSDQ questionnaire showed aquatics therapy
group had significant improvement in self-esteem,
coordination, body fat, strength, flexibility, and
endurance.
Primary: Return to gainful employment. Overall benefit seen with both. Not assessed.
Secondary: Quality of life, verbal/visual memory,
attention, cognitive function, psychiatric status.

Primary: Fitness for military duty or full/part-time 96% return to work; 66% return to military duty. Not assessed.
gainful employment 1 year after intervention.
Assessments performed at 6, 12, and 24 months
after intervention.

Primary: Fitness for military duty or full/part-time 94% return to work; 66% return to military duty. Not assessed.
gainful employment 1 year after intervention.
Assessments performed at 6, 12, and 24 months
after intervention.

88]. Patients were between 4 and 288 weeks after stroke at evaluated cognitive functioning before and after the interven-
the time of enrollment, with exercise interventions lasting tion using the Functional Independence Measure (FIM) and
between 8 and 19 weeks. Five studies applied interventions Stroke Impact Scale (SIS). They reported that scores specific
that focused on cardiovascular endurance training alone [72- to cognition and memory did not differ between exercise and
76]; 3 interventions involved predominantly resistance- control groups at time points immediately after the exercise
based training [77-79]. Five studies applied a mixed endur- intervention and 6 months after exercise [83].
ance and strengthening or unspecified fitness intervention
[34,83,86-88]. Only one group, Studenski et al [83], specif-
Exercise in TBI
ically assessed neurocognitive outcomes. In their 2005 sec-
ondary analysis of the 2003 12-week cardiovascular exercise The most detailed and specific assessments of the neurocog-
intervention involving 100 community-dwelling stroke sur- nitive effects of physical exercise in persons with ABI have
vivors reported by Duncan et al [86], Studenski et al [83] been done in studies looking exclusively at those who are
Table 3. Physical exercise in humans with acquired brain injury due to cerebrovascular accident (CVA)
Study Type/Participants Intervention

Katz-Leurer RCT 8 weeks intervention

et al, 92 inpatients admitted to Group 1: 5 ⫻ weekly sessions of a progressive leg cycle ergometry, building intensity and duration of exercise
2003 [72] rehabilitation unit within 30 days of in first 2 weeks to reach 3 ⫻ 30 minutes at 60% of peak heart rate, performed in addition to regular inpatient
1st CVA with unilateral symptoms. stroke therapy.
Group 2: 5 ⫻ weekly sessions of regular inpatient stroke therapy, including group activities, physical,
occupational, and speech therapy as needed.
Katz-Leurer 6-month follow-up of RCT performed 8 weeks intervention
et al, by Katz-Leurer et al, 2003. Group 1: 5 ⫻ weekly sessions of a progressive leg cycle ergometry, building intensity and duration of exercise
2003 [73] 92 inpatients admitted to in first 2 weeks to reach 3 ⫻ 30 minutes at 60% of peak heart rate, performed in addition to regular inpatient
rehabilitation unit within 30 days of stroke therapy.
1st CVA with unilateral symptoms. Group 2: 5 ⫻ weekly sessions of regular inpatient stroke therapy, including group activities, physical,
occupational, and speech therapy as needed.
Moreland Single blind RCT Study intervention ⫽ length of stay (median ⫽ 62 days for group 1, 56 days for group 2)
et al, 133 consecutive admissions to 5 stroke Group 1: 3 ⫻ weekly sessions of 9 lower extremity exercises performed with progressive resistance, in addition
2003 [77] rehabilitation units within 6 months of to conventional inpatient physical therapy.
CVA who had lower extremity motor Group 2: 3 ⫻ weekly sessions of 9 lower extremity exercises performed without resistance, in addition to
function sufficient to permit conventional inpatient physical therapy.
participation.

Potempa Randomized prospective study 10 weeks intervention

et al, 42 subjects at least 6 months after Group 1: 3 ⫻ 30 min/wk cycling increasing from 30% to 50% of preintervention peak V̇O2.
1995 [74] stroke and residual hemiparesis. Group 2: 3 ⫻ 30 min/wk passive range of motion sessions.
Duncan Single blind RCT 8 weeks intervention (4 weeks supervised, 4 weeks independent)
et al, 20 subjects 30-90 days after stroke Group 1: 3 ⫻ 60 min/wk progressing to 20 min sustained cycling in each session, proprioceptive, balance,
1998 [75] and upper extremity functional activities.
Group 2: 2 ⫻/week assessment of physical activities
Duncan Single blind RCT 12 weeks intervention
et al, 100 community-dwelling subjects age Group 1: 36 exercise sessions (90 min duration) for 12-14 weeks. Exercise program consisted of progressive
2003 [86] ⱖ 50 within 30 and 150 days after strengthening, balance, upper extremity balance, range of motion, and endurance activities. Endurance
stroke, taken from the Kansas City exercise increased to up to 30 minutes, then interval training with progressive intensity was implemented. No
Stroke Registry other therapy, aside from speech, was provided.
Group 2: Usual care as prescribed by subject’s own MD. Usual care subjects received home visits by
research staff every 2 weeks for health education, vital signs, and assessment of oxygen saturation
Studenski Secondary analysis of Duncan et al 12 weeks intervention
et al, 2003 single blind RCT Group 1: 36 exercise sessions (90 min duration) for 12-14 weeks. Exercise program consisted of progressive
2005 [83] 100 community-dwelling subjects age strengthening, balance, upper extremity balance, range of motion, and endurance activities. Endurance
ⱖ 50 within 30 and 150 days after exercise increased to up to 30 minutes, then interval training with progressive intensity was implemented. No
stroke, taken from the Kansas City other therapy, aside from speech, was provided.
Stroke Registry Group 2: Usual care as prescribed by subject’s own MD. Usual care subjects received home visits by
research staff every 2 weeks for health education, vital signs, and assessment of oxygen saturation

Lai et al, Secondary analysis of Duncan et al 12 weeks intervention

2006 [34] 2003 single blind RCT Group 1: 36 exercise sessions (90 min duration) for 12-14 weeks. Exercise program consisted of progressive
100 community-dwelling subjects age strengthening, balance, upper extremity balance, range of motion, and endurance activities. Endurance
ⱖ 50 within 30 and 150 days after exercise increased to up to 30 minutes, then interval training with progressive intensity was implemented. No
stroke, taken from the Kansas City other therapy, aside from speech, was provided.
Stroke Registry Group 2: Usual care as prescribed by subject’s own MD. Usual care subjects received home visits by
research staff every 2 weeks for health education, vital signs, and assessment of oxygen saturation
Teixeira- Unblinded RCT 10 weeks intervention
Salmela 13 community-dwelling subjects with Group 1: 3 ⫻ 60-90 min/wk individualized sessions, including increasing intensity of aerobic activity to goal of
et al, unilateral stroke, ⱖ 9 months after 20 minutes @ 50-70% of predicted max HR.
1999 [76] CVA Group 2: No intervention.

Chu et al, Single blind RCT 8 weeks intervention

2004 [87] 12 community-dwelling subjects (ages Group 1: 3 ⫻ 60 min/wk water-based exercise in chest-level water. Intervention consisted of 10 minutes land-
51-70, 1-7 years after stroke) with mild based stretching, 5 minutes of light aerobic warm-up in water, and 30 minutes of moderate to high aerobic
to moderate residual motor deficits. activity at progressively increasing target heart rates, ranging from 50-80% of initially measured maximum.
Group 2: 3 ⫻ 60 min/wk land-based upper extremity strengthening and fine motor skills. Subjects performed 5
minutes warm-up of upper extremities, followed by 7-minute rotations through a 6-station circuit that focused on
gross and fine upper extremity motion and strengthening of the shoulder, elbow, wrist, and fingers.
Pang et al, Single blind RCT 19 weeks intervention
2004 [88] 63 community-dwelling subjects (age Group 1: 3 ⫻ 60 min/wk group exercise sessions. Program consisted of 3 stations, including a progressively
50, time since stroke ⱖ 1 year) increasing cardiovascular training station (leading up to 30 minutes of continuous exercise at intensity 40-
80% predicted max heart rate), mobility and balance, and leg muscle strength.
Group 2: 3 ⫻ 60 min/wk group exercise sessions. Program consisted of 3 stations, including progressively
increasing shoulder muscle strengthening, progressively increasing elbow/wrist strengthening, and range of
motion and hand activities, including putty/grip strengthening, fine motor activities, and electrical
stimulation to wrist extensors.
Ouellette RCT 12 weeks intervention
et al, 42 community-dwelling subjects (age Group 1: 3 ⫻ weekly supervised sessions of progressive resistance training exercises of the lower extremities.
2004 [78] ⱖ 50 years), 6-72 months after single Subjects performed 3 sets of 8-10 repetitions at 70% of 1-rep maximum (1RM). 1RM was reassessed biweekly.
unilateral mild-moderate stroke, with Group 2: 3 ⫻ weekly supervised sessions of range of motion and upper body flexibility exercises.
residual lower extremity hemiparesis.

Eng et al, Single group repeated measure 8 weeks intervention

2003 [79] design Group 1: 3 ⫻ weekly supervised sessions of functional strengthening, balance, and mobility exercises.
25 community-dwelling subjects
(mean age 63 years, 12 months after
stroke, with residual hemiparesis.
PM&R Vol. 1, Iss. 6, 2009 569
Outcome measures Findings Neurocognitive effects

Primary: Maximum workload, exercise time, resting and Resting heart rate, maximum workload, and work time improved significantly in the Not assessed.
submaximal hemodynamics, functional abilities intervention group. An interaction was seen between age and aerobic training
(walking distance, speed, stair climbing, and FIM on walking distance.
score).

Primary: Frenchay Activities Index (FAI) to measure FAI mean scores did not differ between intervention and control groups 6 months Not assessed.
independence in daily and social activities at after intervention.
enrollment and 6 months after enrollment in Subgroup assessment showed an interaction effect between event severity,
intervention. intervention, and FAI total score such that mean FAI scores declined significantly
less in less severely impaired intervention group subjects than in control subjects
with similar degrees of impairment.
Primary: Rate of change to Chedoke-McMaster Stroke No functional advantage was conferred by progressive resistance strengthening Not assessed.
Assessment (CMSA) Disability Inventory and rate of exercises.
change to 2-minute walk test score at time of No difference was seen in living arrangements between the 2 groups at 6 months.
discharge. No cognitive assessments were made.
Secondary: Change to CMSA disability inventory, 2-
minute walk test, and muscle tone after 4 weeks;
change to above, plus living arrangements and
adverse effects 6 months after discharge.
Primary: Fugl-Meyer (functional independence), peak Improved peak V̇O2 and improved Fugl-Meyer scores in intervention group. Not assessed.
V̇O2, hemodynamics No significant differences in resting hemodynamics.

Primary: Fugl-Meyer (functional independence), 6- Improved Fugl-Meyer scores, 6-minute walk scores, SF-36 scores, Berg balance Not assessed.
minute walk, 10-meter walk, Lawton Scale (extended scores in intervention group.
ADLs), SF-36, Berg balance No change to Lawson Scale, Barthel index.

Primary: Orpington Prognostic Scale (stroke severity), Intervention group showed significantly greater gains than the usual care group in Not assessed.
Fugl-Meyer (functional independence), Wolf Motor peak V̇O2, exercise duration, Berg balance scores, 6-minute walk distance, gait
Function Test (upper extremity), grip strength, velocity, Wolf Motor Function scores.
isometric ankle and knee strength, 10-meter walk, 6- The usual care group showed improvement from baseline in strength, balance,
minute walk, Berg balance scale, functional reach, upper and lower extremity motor control, upper extremity function, and gait
peak V̇O2, maximum HR, maximum exercise duration. velocity, but did not show gains in endurance.
All outcomes measured at 3 months.
Primary: Baseline, posttreatment, and 6-month Intervention group showed greater improvements than usual care group FIM scores specific to
posttreatment daily functioning and quality of life immediately after intervention in SF-36 scores relating to social function and SIS cognition and SIS scores
assessed via Barthel Index, Functional Independence scores relating to strength, emotion, social participation, and physical function. specific to memory and
Measure (FIM), instrumental activities of daily living, Intervention group showed marginally improved Barthel scores and SF-36 scores thinking did not differ
Medical Outcomes Study short form 36-item (SF-36), relating to physical function, physical role function, and SIS scores relating to significantly between
and Stroke Impact Scale (SIS) upper extremity function. the 2 groups either
All treatment effects were diluted at the 6-month follow-up assessment. immediately after
intervention or 6 months
after intervention.
Primary: Depressive symptoms measured via Geriatric Intervention group showed significantly fewer subjects with depressive symptoms at Not assessed.
Depression Scale (GDS), quality of life measured via both 3 months and 6 months after intervention than the usual care group.
SF-36, and Stroke Impact Scale (SIS), Stroke severity Intervention subjects, both with and without depressive symptoms, had equivalent
via Orpington Prognostic Scale, perceived social treatment-related gains in impairments and functional limitations, but only
support as measured through the Pearlin Expressive intervention subjects with depressive symptoms reported improved quality of life
Social Support Scale after the intervention.
Depressive symptoms did not limit gains in physical function as a result of exercise.
Primary: Peak isokinetic torque in lower extremity Improved walking speed, HAP, NHP scores, increased muscle strength. Not assessed.
muscles, walking speed, stair climbing speed, Human No change to lower extremity spasticity.
Activity Profile (general level of physical activity).
Secondary: Nottingham Health Profile (quality of life),
lower extremity spasticity.
Primary: V̇O2 max on stationary bicycle. Intervention group experienced significant gains when compared with control Not assessed.
Secondary: Maximal workload, muscle strength, gait group subjects in measures of postexercise V̇O2 max, (22% higher than control)
speed, Berg Balance Scale (BBS) score. maximal workload, gait speed (19% higher than control), and strength on their
more affected side.
Control group showed greater improvement in BBS than intervention group, but
showed no improvement in V̇O2 max, workload, gait speed, or muscle strength.

Primary: V̇O2max on cycle ergometer, 6-minute walk Intervention group had significantly more improvements than control group in V̇O2 Not assessed.
test, maximum strength in knee extension, Berg max, 6-minute walk test, paretic leg muscle strength.
Balance Scale (BBS), Physical Activity Scale for Intervention group maintained baseline bone mineral density (BMD) in paretic hips,
Individuals with Physical Disabilities (PASIPD) scale, where controls experienced 2.5% decline in BMD.
Secondary: Bilateral hip DEXA scan. Both groups improved PASIPD and BBS scores

Primary: Lower extremity muscle strength and peak Intervention group showed significant gains as compared with control group in Not assessed.
power, performance-based functional measures lower extremity muscle strength and peak power in all activities except for
(including 6-minute walk, stair climb, chair rise, and nonparetic ankle dorsiflexion.
habitual and maximal gait velocities), Late-Life Advanced lower extremity function, limitation dimension, and instrumental role in
Function and Disability Instrument (LLFDI) disability aspects of the LLFDI improved significantly in the intervention group as
Secondary: Geriatric Depression Scale (GDI), Sickness compared with the control group.
Impact Profile (SIP), and Exart’s Self Efficacy Scale. Both groups improved in 6-minute walk and maximal gait velocity, GDS and SIP
scores after 12 weeks; no significant difference was seen between groups.
No difference was seen in Self-Efficacy Scale or PF-10 scores in either group.
Primary: Berg Balance Test, 12-minute Walk Test, self- Improvements were seen in all physical outcome measures, both immediately Not assessed.
paced gait speed, fast-paced gait speed, self- after intervention and after 4 weeks.
paced stair speed, fast-paced stair speed, Subjects improved COPM scores both immediately after and 4 weeks after
Reintegration to Normal Living Index (RNL), Canadian intervention. Overall, subjects did not improve RNL scores as a result of the
Occupational Performance Measure (COPM) intervention.
assessments performed at baseline, after intervention,
and 4 weeks after intervention.
570 Devine and Zafonte PHYSICAL EXERCISE AND COGNITIVE RECOVERY IN ACQUIRED BRAIN INJURY

Table 4. Physical exercise in humans with acquired brain injury due to trauma (TBI)

Study Type/Participants Intervention

McMillan et al, 2002 [84] Single blind RCT 4 weeks intervention
145 TBI patients age 16-65 y who showed problems Group 1: 5 ⫻ 45 minutes of supervised attentional
with attention on neuropsychological testing, or control therapy (audio tape), plus daily self-
who reported attentional problems in everyday administered sessions with the same tape.
life. Subjects were between 3 and 12 months Group 2: 5 ⫻ 45 minutes of supervised physical
after injury and living at home. No restrictions on exercise (unspecified nature/duration/intensity)
severity of injury were made. plus audio tape– based training in physical fitness
training.
Group 3: No therapy contact.

Gordon et al, 1998 [66] Retrospective survey Subjects were grouped based on self-reported
240 community-dwelling individuals with TBI and exercise habits. Those who exercised (jogging,
139 individuals without disability selected from the biking, or swimming) ⱖ 30 minutes 3 ⫻/wk for at
Quality of Life and Health Interviews questionnaire least 6 months and an unspecified intensity were
participant database, developed by the deemed exercisers. Those who exercised ⬍ 1 ⫻/
Research and Training Center (RTC) on wk were deemed nonexercisers.
Community Integration of Individuals with TBI in Group 1: 64 TBI subject exercisers.
New York City. Subjects with TBI were ⱖ 1 year Group 2: 176 TBI subject nonexercisers.
after injury. Group 3: 66 nondisabled subject exercisers.
Group 4: 73 nondisabled subject nonexercisers.

Grealy, Johnson and Rushton, RCT 4 weeks intervention

1999 [67] 26 consecutive moderate TBI patients, at least 6 Group 1: 3 ⫻ 25 min/wk @ Borg 10-12.
weeks after injury Group 2: Single-session intervention was one
25 control TBI patients matched for age/severity/ workout like this tests:
time since injury.

Bhambhani et al, 2005 [85] Prospective study 12 weeks intervention; 18-20 weeks total follow-up.
26 inpatient with severe TBI (average GCS 4.6), Group 1: 3 ⫻ ⱖ60 min/wk training consisting of 10
who had completed acute rehabilitation and min warm-up, 45 min circuit training, 15-25
were avg 17.4 months since injury. minutes of aerobic exercise @ 60% predicted
max.

post trauma. Four studies examined the role of cardiovascu- same self-reported exercise habits. Gordon et al [66] used
lar endurance training on individuals with brain injury The Institute for Rehabilitation Research Symptom Checklist
caused specifically by trauma and are described in Table 4 (TIRR) to evaluate a broad range of self-reported symptoms
[66,67,84,85]. Two studies involved cardiovascular endur- in all 4 groups studied. Additional surveys were administered
ance-based interventions [66,67] and 2 examined outcomes to the group, including the Beck Depression Inventory (BDI),
after mixed endurance and strengthening or unspecified fit- SF-36, Community Integration Questionnaire (CIQ), and
ness interventions [84,85]. The 1998 retrospective survey by Craig Handicap Assessment and Reporting Technique
Gordon et al [66] of 240 community-dwelling TBI survivors (CHART). Gordon et al [66] found that TBI exercisers re-
evaluated subjects with a history of TBI recruited from the ported fewer cognitive symptoms on TIRR assessment. Addi-
Research and Training Center on Community Integration in tionally, TBI exercisers scored better on BDI scales, were
New York. Subjects were stratified as exercisers or nonexer- more productive in their communities as reflected in CIQ,
cisers on the basis of self-reported habits. Criteria for inclu- and performed better on CHART assessment than did TBI
sion in the exercise group included jogging, biking, or swim- nonexercisers. Although nondisabled control subjects per-
ming for 30 minutes or greater 3 or more times weekly for the formed better on all assessments than did TBI subjects,
preceding 6 months; criteria for inclusion in the nonexercise controls did not differ significantly in their scores based on
group included exercise less than once per week for the exercise habits. Gordon’s team suggested the need for pro-
preceding 6 months. Nondisabled control subjects were re- spective, controlled studies of the impact of exercise on
cruited from the community and stratified on the basis of the cognitive function in the TBI population. Although their
PM&R Vol. 1, Iss. 6, 2009 571

Outcome Measures Findings Neurocognitive Effects

Primary: Test of Everyday Attention, Adult Intervention groups showed higher scores at baseline in self-reported No significant differences
Memory and Information Processing cognitive failures than control group; a significant reduction in self- in cognitive function
Battery, Paced Auditory Serial Addition reported cognitive failures was then seen in both intervention were seen in either
Test, Trail Making Test, Sunderland groups at 12 months. intervention group
Memory Questionnaire, Cognitive No other differences were observed between groups. immediately after or at
Failures Questionnaire, Hospital Anxiety 12 months after
and Depression Questionnaire, General intervention.
Health Questionnaire, Rivermead Post-
Concussional Symptoms Questionnaire
Assessments made before, immediately
after, and 12 months after intervention.
Subjects were also given a 6-month
follow-up via postal questionnaire
survey.

Primary: The Institute for Rehabilitation TIRR: Nondisabled subjects reported lower TIRR scores than TBI TBI subjects from the
Research Symptom Checklist (TIRR), subjects. Subjects with TBI who exercised showed significantly fewer exercise group reported
Beck Depression Inventory (BDI), SF-36 symptoms than nonexercisers. Of the 23 symptoms seen more significantly fewer
Health Survey, Community Integration frequently in TBI nonexercisers, 17 symptoms were cognitive. cognitive symptoms
Questionnaire (CIQ), Craig Handicap Exercise did not impact the number of symptoms seen in than TBI nonexercisers.
Assessment Capacity Technique nondisabled subjects.
(CHART) BDI: Nondisabled subjects reported lower scores (less depression) on
BDI than TBI subjects. TBI subjects who exercised scored lower than
TBI nonexercisers; exercise did not impact BDI scores in nondisabled
group.
SF-36: Nondisabled subjects reported better self-perceived health
status than TBI subjects. TBI subjects who exercised reported better
health status than nonexercisers in all areas except bodily pain.
CIQ: Nondisabled subjects reported higher levels of community
integration than TBI subjects. Exercising TBI subjects were more
productive than nonexercising TBI counterparts.
CHART: Nondisabled performed better than TBI subjects; TBI
exercisers were more mobile than TBI nonexercisers.

Attention, info processing, learning, 4-week exercise group did better than single-session group on Improved verbal, visual
memory, reaction and movement times verbal, visual learning tasks. Reaction times and movement times learning seen
were improved with single bout. immediately after 4-
week cardiovascular
exercise intervention.

Primary: Body composition, peak Improved peak power output, oxygen uptake, and ventilation with Not assessed.
cardiorespiratory responses. intervention.
No changes is body composition, peak heart rate, or blood lactate.

findings suggested a link between physical exercise and im- trol subjects (age, severity of injury, and postinjury time-
proved cognition after brain injury, association only could be matched TBI patients at the same inpatient rehabilitation
inferred from this study. Moreover, as a retrospective survey, facility) underwent routine inpatient TBI rehabilitation. Out-
the study by Gordon et al depended completely on subjective come measures were entirely cognitive: Digit Span, Digit
reports for the initial stratification of subjects into exercise Symbol, Trails A and B, Auditory Verbal Learning, Visual
groups and for their subjective reports of memory and recall. Learning, Logistical Memory, and Complex Figure (Rey)
In conclusion, Gordon et al [66] cited the need for prospec- tests. After intervention, the exercise group showed improve-
tive, controlled studies of exercise and cognition in this ments relative to controls in digital symbol testing as well as
population with objective assessments of neurocognitive out- auditory and visual learning; no improvements were seen in
come measures. Grealy et al [67] followed in 1999 with a complex figure or logical memory tasks. The authors thought
novel prospective pilot study examining the impact of a this showed that the intervention was successful in enhanc-
combined cardiovascular exercise and virtual-reality inter- ing working memory but did not impact impairments in
vention on cognition TBI patients. Thirteen inpatient TBI attention, as evidenced by the lack of improvement on Digit
subjects at least 6 weeks after injury underwent 4 weeks of Span and Trails B testing [67]. The question of how physical
3-times-weekly, 25-minute sessions of nonimmersive virtual- exercise might impact attention was again raised in the 2002
reality exercise at an intensity of 10 to 12 on the Borg Scale. study by McMillan et al [84] comparing multiple cognitive
Subjects used a recumbent cycle ergometer that provided measures in outpatient TBI subjects receiving a 4-week,
visual, tactile, vestibular, and auditory stimuli. Thirteen con- audio-taped, attentional therapy intervention; a supervised,
Critical Review

Physical Exercise and Cognitive Recovery in

Acquired Brain Injury: A Review of the Literature
Jennifer M. Devine, MD, Ross D. Zafonte, DO

Objective: Physical exercise has been shown to play an ever-broadening role in the
maintenance of overall health and has been implicated in the preservation of cognitive
function in both healthy elderly and demented populations. Animal and human studies of
acquired brain injury (ABI) from trauma or vascular causes also suggest a possible role for
physical exercise in enhancing cognitive recovery.
Data Sources: A review of the literature was conducted to explore the current under-
standing of how physical exercise impacts the molecular, functional, and neuroanatomic
status of both intact and brain-injured animals and humans.
Study Selection: Searches of the MEDLINE, CINHAL, and PsychInfo databases yielded an
extensive collection of animal studies of physical exercise in ABI. Animal studies strongly tie
physical exercise to the upregulation of multiple neural growth factor pathways in brain-injured
animals, resulting in both hippocampal neurogenesis and functional improvements in memory.
Data Extraction: A search of the same databases for publications involving physical
exercise in human subjects with ABI yielded 24 prospective and retrospective studies.
Data Synthesis: Four of these evaluated cognitive outcomes in persons with ABI who
were involved in physical exercise. Three studies cited a positive association between
exercise and improvements in cognitive function, whereas one observed no effect. Human
exercise interventions varied greatly in duration, intensity, and level of subject supervision,
and tools for assessing neurocognitive changes were inconsistent.
Conclusions: There is strong evidence in animal ABI models that physical exercise facilitates
neurocognitive recovery. Physical exercise interventions are safe in the subacute and rehabilita-
tive phases of recovery for humans with ABI. In light of strong evidence of positive effects in
animal studies, more controlled, prospective human interventions are warranted to better
explore the neurocognitive effects of physical exercise on persons with ABI.

INTRODUCTION
The role of exercise in the maintenance of cardiovascular health has been well established.
Epidemiological studies suggest that individuals who expend energy greater than 2000 kCal
per week in some form of moderate-to-intense exercise significantly lower their risk of
mortality from all causes [1] and decrease their cardiovascular morbidity when compared
with those who exercise less intensely or frequently [2]. Further benefits of exercise include J.M.D. Department of Physical Medicine &
Rehabilitation, Spaulding Rehabilitation Hos-
improvements in cardiovascular control [2], glycemic control [3], weight management [4], pital, Harvard Medical School, Boston, MA
bone density [5], depression [6], and breast and colon cancer [7]. In light of this evidence, Disclosure: nothing to disclose
the American College of Sports Medicine and the American Heart Association currently R.D.Z. Department of Physical Medicine & Re-
recommend that all adults aged 18 to 65 years participate in 30 minutes of moderate-intense habilitation, Spaulding Rehabilitation Hospital,
Harvard Medical School, 125 Nashua St, Bos-
aerobic (endurance) physical activity per day at least 5 days per week [8]. Adding to the
ton, MA 02214. Address correspondence to:
body of evidence showing its many established health benefits are findings that regular R.D.Z.; e-mail: [email protected]
exercise may also have a positive impact on memory and cognition; thus, investigational Disclosure: 2A, BHR and Allergan; 7A, Neuro-
healing - Drug Phase II Study Cosponsored by
interest into the specific effects of exercise on cognitive function is growing.
FDA, Funded by Orphan Drug; 8B, NIH, NIDRR,
Numerous population-based studies suggest a decreased incidence of cognitive decline in DOD-Supported Research
individuals who participate in physically and cognitively stimulating activities later in life [9-13]. Disclosure Key can be found on the Table of
Friedland et al [13] further demonstrated that a decreased risk of cognitive delay is seen in elders Contents and at www.pmrjournal.org
who participated in cardiovascular conditioning throughout midlife. Finally, in an effort to Submitted for publication October 1, 2008;
further explore potential dose-dependency for this benefit, Larson et al [14] showed a decreased accepted March 29, 2009.

PM&R © 2009 by the American Academy of Physical Medicine and Rehabilitation

1934-1482/09/$36.00 Vol. 1, 560-575, June 2009
560
Printed in U.S.A. DOI: 10.1016/j.pmrj.2009.03.015
PM&R Vol. 1, Iss. 6, 2009 573

concomitant mood disorders increases its attractiveness as a 16. Van Boxtel MP, Paas FG, Houx PJ, Adam JJ, Teeken JC, Jolles J. Aerobic
nonpharmacologic intervention for people with ABI. capacity and cognitive performance in a cross-sectional aging study.
Med Sci Sports Exerc 1997;29:1357-1365.
Finally, further investigation into possible interactions 17. Richards M, Hardy R, Wadsworth ME. Does active leisure protect
between exercise and other pharmacologic treatments is war- cognition? Evidence from a national birth cohort. Soc Sci Med 2003;
ranted if cardiovascular conditioning is to be used to improve 56:785-792.
cognitive function. Although evidence suggests that exercise 18. Perrig-Chiello P, Perrig WJ, Ehrsam R, Staehelin HB, Krings F. The
and selective serotonin reuptake inhibitors (SSRIs) can ele- effects of resistance training on well-being and memory in elderly
volunteers. Age Ageing 1998;27:469-475.
vate hippocampal BDNF, a recent rodent study of these 19. Churchill JD, Galvez R, Colcombe S, Swain RA, Kramer AF, Greenough
interventions undertaken simultaneously did not show syn- WT. Exercise, experience and the aging brain. Neurobiol Aging 2002;
ergism between the two [95]. Preliminary evidence in rodent 23:941-955.
brain injury models also suggests a lack of synergism between 20. Aleman A, De Haan EHF, Verhaar HJJ, Samson MM, de Vries WR,
neurostimulants and voluntary exercise [96]. These findings Koppeschaar HP. Relation between cognitive and physical function in
healthy older men; a role for aerobic power? J Am Geriatr Soc 2000;48:
further accentuate our limited understanding of the path- 104-105.
ways through which neurorecovery takes place and point to 21. McAuley E, Kramer AF, Colcombe SJ. Cardiovascular function in older
a need for further research in this area. adults: A brief review. Brain Behav Immun 2004;18:214-220.
22. Prins ND, Den Heijer T, Hofman A, et al. Homocysteine and cognitive
function in the elderly: The Rotterdam Scan Study. Neurology 2002;
59:1375-1380.
REFERENCES 23. Rogers RL, Meyer JS, Mortel KF. After reaching retirement age physical
activity sustains cerebral perfusion and cognition. J Am Geriatr Soc
1. Lee IM, Skerrett PJ. Physical activity and all-cause mortality: What is the 1990;38:123-128.
dose-response relation? Med Sci Sports Exerc 2001;33(Suppl):S459- 24. Cotman CW, Berchtold NC. Exercise: A behavioral intervention to
S471. enhance brain health and plasticity. Trends Neurosci 2002;25:295-
2. Thompson PD, Buchner D, Pina IL, et al. Exercise and physical activity 301.
in the prevention and treatment of atherosclerotic cardiovascular dis- 25. Van Praag H, Christie BR, Sejnowski TJ, Gage FH. Running enhances
ease: A statement from the Council on Clinical Cardiology (Subcom- neurogenesis, learning, and long-term potentiation in mice. Proc Natl
mittee on Exercise, Rehabilitation, and Prevention) and the Council on Acad Sci U S A 1999;96:13427-13431.
Nutrition, Physical Activity, and Metabolism (Subcommittee on Physi- 26. Cotman CW, Engesser-Cesar C. Exercise enhances and protects brain
cal Activity). Circulation 2003;107:3109-3116. function. Exerc Sport Sci Rev 2002;30:75-79.
3. Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in the 27. Eriksson PS, Perfilieva E, Bjork-Eriksson T, et al. Neurogenesis in the
incidence of type 2 diabetes with lifestyle intervention or metformin. adult human hippocampus. Nat Med 1998;4:1313-1317.
N Engl J Med 2002;346:393-403. 28. Vaynman S, Ying Z, Gomez-Pinilla F. Hippocampal BDNF mediates the
4. Wing RR, Hill JO. Successful weight loss maintenance. Ann Rev Nutr efficacy of exercise on synaptic plasticity and cognition. Eur J Neurosci
2001;21:323-341. 2004;20:2580-2590.
5. Vuori IM. Health benefits of physical activity with special reference to 29. Epp JR, Spritzer MD, Galea LA. Hippocampus-dependent learning
interaction with diet. Public Health Nutr 2001;4:517-528. promotes survival of new neurons in the dentate gyrus at a specific time
6. Brosse AL, Sheets ES, Lett HS, Blumenthal JA. Exercise and the treat- during cell maturation. Neuroscience 2007;149:273-285.
ment of clinical depression in adults: Recent findings and future direc- 30. Farmer J, Zhao X, Van Praag H, Wodtke K, Gage FH, Christie BR.
tions. Sports Med 2002;32:741-760. Effects of voluntary exercise on synaptic plasticity and gene expression
7. Thune I, Furberg AS. Physical activity and cancer risk: Dose-response in the dentate gyrus of adult male Sprague-Dawley rats in vivo. Neuro-
and cancer, all sites and site-specific. Med Sci Sport Exerc 2001; science 2004;124:71-79.
33(Suppl):S530-S550. 31. MacKay-Lyons MJ, Makrides L. Cardiovascular stress during a contem-
8. Haskell WL, Lee IM, Pate RR, et al. Physical activity and public health: porary stroke rehabilitation program: Is the intensity adequate to
Updated recommendation for adults from the American College of induce a training effect? Arch Phys Med Rehabil 2002;83:1378-1383.
Sports Medicine and the American Heart Association. Med Sci Sports 32. Gordon NF, Gulanick M, Costa F, et al. Physical activity and exercise
Exerc 2007;39:1423-1434. recommendations for stroke survivors: An American Heart Association
9. Weuve J, Kang JH, Manson JE, Ware JH, Grodstein F. Physical activity, scientific statement from the Council on Clinical Cardiology, Subcom-
including walking, and cognitive function in older women. JAMA mittee on Exercise, Cardiac Rehabilitation, and Prevention; the Council
2004;292:1454-1461. on Cardiovascular Nursing; the Council on Nutrition, Physical Activity,
10. Yaffe K, Barnes D, Nevitt M, Lui LY, Covinsky K. A prospective study of and Metabolism; and the Stroke Council. Stroke 2004;35:1230-1240.
physical activity and cognitive decline in elderly women: Women who 33. Hassett LM, Moseley AM, Tate R, Harmer AR. Fitness training for
walk. Arch Intern Med 2001;161:1703-1708. cardiorespiratory conditioning after traumatic brain injury. Cochrane
11. Verghese J, Lipton RB, Katz MJ, et al. Leisure activities and the risk of Database of Systematic Reviews 2008 (2), CD006123. Available at
dementia in the elderly. N Engl J Med 2003;348:2508-2516. http://www.URL. Accessed August 25, 2008.
12. Wilson RS, Mendes De Leon CF, Barnes LL, et al. Participation in 34. Lai SM, Studenski S, Richards L, et al. Therapeutic exercise and
cognitively stimulating activities and risk of incident Alzheimer disease. depressive symptoms after stroke. J Am Geriatr Soc 2006;54:240-247.
JAMA 2002;287:742-748. 35. Giza CC, Hovda DA. The neurometabolic cascade of concussion. J Athl
13. Friedland RP, Fritch T, Smyth KA, et al. Patients with Alzheimer’s Train 2001;36:228-235.
disease have reduced activities in midlife compared with healthy con- 36. Vagnozzi R, Tavazzi B, Signoretti S, et al. Temporal window of meta-
trol-group members. Proc Natl Acad Sci U S A 2001;98:3440-3445. bolic brain vulnerability to concussions: Mitochondrial-related impair-
14. Larson EB, Wang L, Bowen JD, et al. Exercise is associated with reduced ment—part I. Neurosurgery 2007;61:379-388; discussion 388-389.
risk for incident dementia among persons 65 years of age and older. 37. Tavazzi B, Vagnozzi R, Signoretti S, et al. Temporal window of meta-
Ann Intern Med 2006;144:73-81. bolic brain vulnerability to concussions: Oxidative and nitrosative
15. Burns JM, Cronk BB, Anderson HS, et al. Cardiorespiratory fitness and stresses—part II. Neurosurgery 2007;61:390-395; discussion
brain atrophy in early Alzheimer disease. Neurology 2008;71:210-216. 395-396.
574 Devine and Zafonte PHYSICAL EXERCISE AND COGNITIVE RECOVERY IN ACQUIRED BRAIN INJURY

38. Majerske CW, Mihalik JP, Ren D, et al. Concussion in sports: Postcon- 58. Zhou C, Li Y, Nanda A, Zhang JH. HBO suppresses nogo-A, ng-R, or
cussive activity levels, symptoms and neurocognitive performance. J RhoA expression in the cerebral cortex after global ischemia. Biochem
Athl Train 2008;43:265-274. Biophys Res Commun 2003;309:368-376.
39. Griesbach GS, Gomez-Pinilla F, Hovda DA. The upregulation of plas- 59. Eslamboli A, Grundy RI, Irving EA. Time-dependent increase in
ticity-related proteins following TBI is disrupted with acute voluntary nogo-A expression after focal cerebral ischemia in marmoset monkeys.
exercise. Brain Res 2004;1016:154-162. Neurosci Lett 2006;408:89-93.
40. Griesbach GS, Hovda DA, Molteni R, Wu A, Gomez-Pinilla F. Volun- 60. Marklund N, Deschamps K, McIntosh TK, McKerracher L. Activation
tary exercise following traumatic brain injury: Brain-derived neurotro- of Rho after traumatic brain injury and seizure in rats. Exp Neurol
phic factor upregulation and recovery of function. Neuroscience 2004; 2006;198:361-369.
125:129-139. 61. Marklund N, Salci K, Ronquist G, Hillered L. Energy metabolic changes
41. Altman J, Das GD. Autoradiographic and histological evidence of in the early post-injury period following traumatic brain injury in rats.
postnatal hippocampal neurogenesis in rats. J Comp Neurol 1965;124: Neurochem Res 2006;31:1085-1093.
319-335. 62. Irving EA, Vinson M, Rosin C, et al. Identification of neuroprotective
42. Gould E, Reeves AJ, Fallah M, Tanapat P, Gross CG, Fuchs E. Hip- properties of anti-MAG antibody: A novel approach for the treatment of
pocampal neurogenesis in adult Old World primates. Proc Natl Acad stroke? J Cereb Blood Flow Metab 2005;25:98-107.
Sci U S A 1999;96:5263-5267. 63. Weinmann O, Schnell L, Ghosh A, et al. Intrathecally infused antibod-
43. Kempermann G, Kuhn HG, Gage FH. More hippocampal neurons in ies against nogo-A penetrate the CNS and downregulate the endoge-
adult mice living in an enriched environment. Nature 1997;386:493- nous neurite growth inhibitor nogo-A. Mol Cell Neurosci 2006;32:
495. 161-173.
44. Molteni R, Ying Z, Gomez-Pinilla F. Differential effects of acute and 64. Thompson HJ, Marklund N, LeBold DG, et al. Tissue sparing and
chronic exercise on plasticity-related genes in the rat hippocampus functional recovery following experimental traumatic brain injury is
revealed by microarray. Eur J Neurosci 2002;16:1107-1116. provided by treatment with an anti-myelin-associated glycoprotein
45. Jovanovic JN, Czernik AJ, Fienberg AA, Greengard P, Sihra TS. Syn- antibody. Eur J Neurosci 2006;24:3063-3072.
apsins as mediators of BDNF-enhanced neurotransmitter release. Nat 65. Chytrova G, Ying Z, Gomez-Pinilla F. Exercise normalizes levels of
Neurosci 2000;3:323-329. MAG and nogo-A growth inhibitors after brain trauma. Eur J Neurosci
46. Blanquet PR, Mariani J, Fournier B. Identification of a biphasic signal- 2008;27:1-11.
ing pathway involved in ischemic resistance of the hippocampal den- 66. Gordon WA, Sliwinski M, Echo J, McLoughlin M, Sheerer MS, Meili
tate gyrus. Exp Neurol 2006;202:357-372.
TE. The benefits of exercise in individuals with traumatic brain injury:
47. Ding Q, Vaynman S, Akhavan M, Ying Z, Gomez-Pinilla F. Insulin-like
A retrospective study. J Head Trauma Rehabil 1998;13:58-67.
growth factor I interfaces with brain-derived neurotrophic factor-me-
67. Grealy MA, Johnson DA, Rushton SK. Improving cognitive function
diated synaptic plasticity to modulate aspects of exercise-induced cog-
after brain injury: The use of exercise and virtual reality. Arch Phys Med
nitive function. Neuroscience 2006;140:823-833.
Rehabil 1999;80:661-667.
48. Ding Q, Vaynman S, Souda P, Whitelegge JP, Gomez-Pinilla F. Exercise
68. Bateman A, Culpan FJ, Pickering AD, Powell JH, Scott OM, Greenwood
affects energy metabolism and neural plasticity-related proteins in the
RJ. The effect of aerobic training on rehabilitation outcomes after recent
hippocampus as revealed by proteomic analysis. Eur J Neurosci 2006;
severe brain injury: A randomized controlled evaluation. Arch Phys
24:1265-1276.
Med Rehabil 2001;82:174-182.
49. Ploughman M, Granter-Button S, Chernenko G, Tucker BA, Mearow
69. Jackson D, Turner-Stokes L, Culpan J, et al. Can brain-injured patients
KM, Corbett D. Endurance exercise regimens induce differential effects
participate in an aerobic exercise programme during early inpatient
on brain-derived neurotrophic factor, synapsin-I and insulin-like
rehabilitation? Clin Rehabil 2001;15:535-544.
growth factor I after focal ischemia. Neuroscience 2005;136:991-1001.
70. Driver S, O’Connor J, Lox C, Rees K. Evaluation of an aquatics pro-
50. Luo CX, Jiang J, Zhou QG, et al. Voluntary exercise-induced neurogen-
esis in the postischemic dentate gyrus is associated with spatial memory gramme on fitness parameters of individuals with a brain injury. Brain
recovery from stroke. J Neurosci Res 2007;85:1637-1646. Inj 2004;18:847-859.
51. Lee HH, Shin MS, Kim YS, et al. Early treadmill exercise decreases 71. Driver S, Rees K, O’Connor J, Lox C. Aquatics, health-promoting
intrastriatal hemorrhage-induced neuronal cell death and increases cell self-care behaviours and adults with brain injuries. Brain Inj 2006;20:
proliferation in the dentate gyrus of streptozotocin-induced hypergly- 133-141.
cemic rats. J Diabetes Complications 2005;19:339-346. 72. Katz-Leurer M, Carmeli E, Shochina M. The effect of early aerobic
52. Davis W, Mahale S, Carranza A, et al. Exercise pre-conditioning ame- training on independence six months post stroke. Clin Rehabil 2003;
liorates blood-brain barrier dysfunction in stroke by enhancing basal 17:735-741.
lamina. Neurol Res 2007;29:382-387. 73. Katz-Leurer M, Shochina M, Carmeli E, Friedlander Y. The influence of
53. Ding YH, Luan XD, Li J, et al. Exercise-induced overexpression of early aerobic training on the functional capacity in patients with cere-
angiogenic factors and reduction of ischemia/reperfusion injury in brovascular accident at the subacute stage. Arch Phys Med Rehabil
stroke. Curr Neurovasc Res 2004;1:411-420. 2003;84:1609-1614.
54. Gomez-Pinilla F, So V, Kesslak JP. Spatial learning induces neurotro- 74. Potempa K, Lopez M, Braun LT, Szidon JP, Fogg L, Tincknell T.
phin receptor and synapsin I in the hippocampus. Brain Res 2001;904: Physiological outcomes of aerobic exercise training in hemiparetic
13-19. stroke patients. Stroke 1995;26:101-105.
55. Griesbach GS, Hovda DA, Molteni R, Gomez-Pinilla F. Alterations in 75. Duncan P, Richards L, Wallace D, et al. A randomized, controlled pilot
BDNF and synapsin I within the occipital cortex and hippocampus study of a home-based exercise program for individuals with mild and
after mild traumatic brain injury in the developing rat: Reflections of moderate stroke. Stroke 1998;29:2055-2060.
injury-induced neuroplasticity. J Neurotrauma 2002;19:803-814. 76. Teixeira-Salmela LF, Olney SJ, Nadeau S, Brouwer B. Muscle strength-
56. Davis LM, Pauly JR, Readnower RD, Rho JM, Sullivan PG. Fasting is ening and physical conditioning to reduce impairment and disability in
neuroprotective following traumatic brain injury. J Neurosci Res 2008; chronic stroke survivors. Arch Phys Med Rehabil 1999;80:1211-1218.
86:1812-1822. 77. Moreland JD, Goldsmith CH, Huijbregts MP, et al. Progressive resis-
57. Erb M, Flueck B, Kern F, Erne B, Steck AJ, Schaeren-Wiemers N. tance strengthening exercises after stroke: A single-blind randomized
Unraveling the differential expression of the two isoforms of myelin- controlled trial. Arch Phys Med Rehabil 2003;84:1433-1440.
associated glycoprotein in a mouse expressing GFP-tagged S-MAG 78. Ouellette MM, LeBrasseur NK, Bean JF, et al. High-intensity resistance
specifically regulated and targeted into the different myelin compart- training improves muscle strength, self-reported function, and disabil-
ments. Mol Cell Neurosci 2006;31:613-627. ity in long-term stroke survivors. Stroke 2004;35:1404-1409.
PM&R Vol. 1, Iss. 6, 2009 575

79. Eng JJ, Chu KS, Kim CM, Dawson AS, Carswell A, Hepburn KE. A chronic stroke: A randomized, controlled trial. J Am Geriatr Soc
community-based group exercise program for persons with chronic 2005;53:1667-1674.
stroke. Med Sci Sports Exerc 2003;35:1271-1278. 89. Béquet F, Gomez-Merino D, Berthelot M, Guezennec CY. Exercise-
80. Salazar AM, Warden DL, Schwab K, et al. Cognitive rehabilitation for induced changes in brain glucose and serotonin revealed by microdi-
traumatic brain injury: A randomized trial. Defense and Veterans Head alysis in rat hippocampus: Effect of glucose supplementation. Acta
Injury Program (DVHIP) Study Group. JAMA 2000;283:3075-3081. Physiol Scand 2001;173:223-230.
81. Braverman SE, Spector J, Warden DL, et al. A multidisciplinary TBI 90. Chaouloff F. Effects of acute physical exercise on central serotonergic
inpatient rehabilitation programme for active duty service members as systems. Med Sci Sports Exerc 1997;29:58-62.
part of a randomized clinical trial. Brain Inj 1999;13:405-415. 91. Chennaoui M, Drogou C, Gomez-Merino D, Grimaldi B, Fillion G,
82. Warden DL, Salazar AM, Martin EM, Schwab KA, Coyle M, Walter J. A Guezennec CY. Endurance training effects on 5-HT (1B) receptors
home program of rehabilitation for moderately severe traumatic brain mRNA expression in cerebellum, striatum, frontal cortex and hip-
injury patients. The DVHIP Study Group. J Head Trauma Rehabil pocampus of rats. Neurosci Lett 2001;307:33-36.
2000;15:1092-1102. 92. Fordyce DE, Farrar RP. Enhancement of spatial learning in F344 rats by
physical activity and related learning-associated alterations in hip-
83. Studenski S, Duncan PW, Perera S, Reker D, Lai SM, Richards L. Daily
pocampal and cortical cholinergic functioning. Behav Brain Res 1991;
functioning and quality of life in a randomized controlled trial of
46:123-133.
therapeutic exercise for subacute stroke survivors. Stroke 2005;36:
93. Fordyce DE, Starnes JW, Farrar RP. Compensation of the age-related
1764-1770.
decline in hippocampal muscarinic receptor density through daily
84. McMillan T, Robertson IH, Brock D, Chorlton L. Brief mindfulness
exercise or underfeeding. J Gerontol 1991;46:B245-B248.
training for attentional problems after traumatic brain injury: A ran-
94. Gomez-Merino D, Béquet F, Berthelot M, Chennaoui M, Guezennec
domised control treatment trial. Neuropsychol Rehabil 2002;12:117- CY. Site-dependent effects of an acute intensive exercise on extracellu-
125. lar 5-HT and 5-HIAA levels in rat brain. Neurosci Lett 2001;301:143-
85. Bhambhani Y, Rowland G, Farag M. Effects of circuit training on body 146.
composition and peak cardiorespiratory responses in patients with 95. Engesser-Cesar C, Anderson AJ, Cotman CW. Wheel running and
moderate to severe traumatic brain injury. Arch Phys Med Rehabil fluoxetine antidepressant treatment have differential effects in the
2005;86:268-276. hippocampus and the spinal cord. Neuroscience 2007;144:1033-
86. Duncan P, Studenski S, Richards L, et al. Randomized clinical trial of 1044.
therapeutic exercise in subacute stroke. Stroke 2003;34:2173-2180. 96. Griesbach GS, Hovda DA, Gomez-Pinilla F, Sutton RL. Voluntary
87. Chu KS, Eng JJ, Dawson AS, Harris JE, Ozkaplan A, Gylfadottir S. exercise or amphetamine therapy, but not the combination, increases
Water-based exercise for cardiovascular fitness in people with chronic hippocampal BDNF and Synapsin 1 following cortical contusion injury
stroke: A randomized controlled trial. Arch Phys Med Rehabil 2004; in rats. Neuroscience 2008;154:530-540.
85:870-874. 97. Stranahan AM, Khalil D, Gould E. Running induces widespread struc-
88. Pang MY, Eng JJ, Dawson AS, McKay HA, Harris JE. A community- tural alterations in the hippocampus and entorhinal cortex. Hippocam-
based fitness and mobility exercise program for older adults with pus 2007;17:1017-1022.