801558

research-article2018
NNRXXX10.1177/1545968318801558Neurorehabilitation and Neural RepairNi et al

Review Article
Neurorehabilitation and

Exercise Guidelines for Gait Function in

Neural Repair
2018, Vol. 32(10) 872­–886
© The Author(s) 2018
Parkinson’s Disease: A Systematic Article reuse guidelines:
sagepub.com/journals-permissions

Review and Meta-analysis DOI: 10.1177/1545968318801558

https://doi.org/10.1177/1545968318801558
journals.sagepub.com/home/nnr

Meng Ni, PhD1, Joseph B. Hazzard, EdD1, Joseph F. Signorile, PhD2,3,

and Corneliu Luca, MD, PhD3

Abstract
This systematic review and meta-analysis is to provide comprehensive evidence-based exercise recommendations
targeting walking function for adults with Parkinson’s disease. Methods. Fixed- or random-effect meta-analyses estimated
standardized effect sizes (Hedge’s g), comparing treatment effects from exercise with nonexercise and another form of
exercise (non-EXE control and EXE control). Cuing and exercise duration were used as moderators for subanalyses.
Results. The 40 included randomized controlled trials comprised 1656 patients. The exercise group showed significantly
superior performance in timed up-and-go (g = −0.458; g = −0.390) compared with non-EXE control and EXE control;
significantly greater improvement in comfortable walking speed (g = 0.449), fast walking speed (g = 0.430), and stride
or step length (g = 0.379) compared with non-EXE control; and significantly greater cadence (g = 0.282) compared with
EXE controls. No significant differences between intervention and control groups were observed for double-leg support
time (DLST), dynamic gait index (DGI), 6-minute walk test, or freezing of gait questionnaire (FOG-Q). Notably, treatment
effect from the exercise of interest compared with a standard exercise was greater than for nonexercise for cadence
and FOG-Q. Moreover, EXE control was favored for DLST and DGI. Cuing had a significantly positive effect on stride
length alone. Exercise duration significantly, but negatively, influenced the treatment effect on comfortable walking speed.
Conclusion. Gait-specific training, rather than a general exercise program, should be emphasized if gait is the outcome of
interest. Further investigation is needed on exercise dosage and its selective effect on more challenging walking tasks,
endurance, and freezing of gait.

Keywords
task-specific training, walking, exercise prescription, movement disorder

Introduction
Parkinson’s disease (PD) affects between 0.5% and 1.0% of
adults aged 65 to 69 years, and 1% to 3% of those 80 years
of age and older.1 PD gait disturbance is produced by a
combination of multiple factors, including bradykinesia,
impaired postural reflexes, abnormal postures, rigidity, and
tremor and characterized by a paucity of movement of the
trunk, upper-body, and lower-body muscles, resulting in
slow and shuffling steps that are short and rapid. Because of
the shortened step length, the feet barely clear the ground,
and the soles of the feet shuffle and scrape the floor.2
Furthermore, freezing of gait in PD, one of the most dis-
abling symptoms, is characterized by an inability to gener-
ate or sustain an effective stepping sequence,3 which leads
to sudden blocks in walking.4 PD increases the risk of fall-
ing because of these gait disturbances.
Exercise has been widely used in allied health care for
PD to address symptoms and improve motor functions,5,6
and the exercise mode includes balance and gait training,7
resistance training,8 treadmill training,9 and complementary
therapies.10 However, the lack of comprehensive effective
1
Bloomsburg University of Pennsylvania, Bloomsburg, PA, USA
2
University of Miami, Coral Gables, FL, USA
3
University of Miami, FL, USA
Supplementary material for this article is available on the
Neurorehabilitation & Neural Repair website at http://nnr.sagepub.com/
content/by/supplemental-data.
Corresponding Author:
Meng Ni, PhD, Department of Exercise Sciences, Bloomsburg University
of Pennsylvania, 400 E 2nd St Centennial Hall Rm 132, Bloomsburg, PA
17815, USA.
Email: mni@bloomu.edu
Ni et al 873
exercise guidelines, specifically focusing on gait function,
makes it reasonable to evaluate the efficacy of a variety of
exercise modalities and programs. The purpose of this sys-
tematic review and meta-analysis is to provide exercise rec-
ommendations (frequency, intensity, type, and time) for
clinical program implementation by evaluating the training
effect of exercise on gait function in adults with PD, thereby
allowing a wider range of use and accessibility by patients
and therapist.
Methods
Data Sources and Searches
The following databases were searched: MEDLINE
(EBSCO), CINAHL Plus, PubMed, and SPORTDiscus to
identify potential studies. Studies published from inception
to June 2017 were included. The combinative keywords
used for searching were the following: PD/Parkinsonism,
exercise/physical therapy/physical activity/training, and
gait/walking.
Inclusion Criteria
Inclusion criteria were a randomized control trial that
included the following:
1. the effect of exercise interventions (include cuing
strategies);
2. outcomes including gait/walking-related perfor-
mance (gait speed, including comfortable and fast
gait speeds; stride or step length; cadence; double
support time [percentage of gait cycle]; timed up-
and-go [TUG]; 6-minute walk test [6MWT]; freez-
ing of gait questionnaire [FOG-Q]; and dynamic
gait index [DGI]); and
3. articles available in English.
Exclusion Criteria
Exclusion criteria were as follows:
1. studies where the effects of a computer-based exer-
cise intervention (such as computer games), walk-
ing-assist equipment (such as walkers or robotics),
or exercise intervention with electrical/magnetic
stimulation were evaluated and
2. studies where outcomes did not include gait
assessments.
Methodological Quality
Information extracted from randomized controlled trials
(RCTs) comprised a description of participants, details of the
exercise, and details of outcome measures. The PEDro Scale
was used to rate the methodological quality of the RCTs. Two
independent investigators (MN and JBH) screened and
scored RCTs. Interrater reliabilities for individual items of
the PEDro Scale were calculated as Cohen’s κ. Differences
of opinion on rating were resolved through discussion by the
2 investigators (MN and JBH). PEDro scores of 6 to 8 points
were classified as “good quality,” 4 to 5 points as “fair qual-
ity,” and less than 4 points as “poor quality.”11
Calculating Effect Size
Meta-analyses were conducted for 9 outcome measures.
The primary outcome was comfortable walking speed.
Other gait measurements included stride or step length,
cadence, and double-leg support time (DLST; percentage of
gait cycle) at comfortable walking speed, fast walking
speed, TUG, DGI, 6MWT, and FOG-Q. The selected out-
comes (gait speed, stride length, cadence, TUG, 6MWT,
and FOG-Q) are the most frequently used measures in clini-
cal settings and reported in published trails. The results of
this article would be more clinically applicable. Meanwhile,
double support time (percentage of gait cycle) and DGI pro-
vide more advanced gait analysis and comprehensive gait
evaluation.
The standardized mean difference (Hedge’s g) was cal-
culated for all meta-analyses. The effect size was calculated
using the group mean difference of the change score
between the pretest and posttest for each group and the
pooled SD during patients’ “on” medication state. The
group contrast was unbiased by the correction factor J = 1
− 3/(4 × [n1 + n2 − 2] − 1). Finally, the mean difference for
the first assessment period was used for crossover studies.
Moderators
A univariate moderator model using categorical variables
was conducted to assess the presence of cuing (yes/no or
unclear) and estimates of the effect of cuing on gait-related
performances. A univariate metaregression was conducted
to assess the associations between the number of hours for
the interventions in the trials. The metaregressions were
only performed for those outcome measures where the stud-
ies included more than 10 trials to ensure sufficient statisti-
cal power.
Statistical Analysis
For all the outcome measures, the analysis was separated
into 2 categories, exercise versus nonexercise control (non-
EXE control), and exercise versus another form of exercise
control (EXE control). Statistical heterogeneity was
quantified using Q statistics. The overall fixed- or random-
effect (restricted maximum likelihood) model was used
874 Neurorehabilitation and Neural Repair 32(10)
based on the Q statistics. Publication bias for each outcome
measure was assessed using Egger’s test. Post hoc analysis
was used to examine the difference in the presence of cuing
(yes/no or unclear) and the number of minutes of the inter-
ventions on the gait performance. The threshold for signifi-
cance was set at P <.05. All data were analyzed using R
Statistical Software 3.2.3.
Results
Identified Literature
Searching identified 669 RCTs, of which 47 were appropri-
ate for inclusion and 40 were included in the meta-analysis
(Supplement, PRISMA flow). A total of 1656 participants
contributed to the studies reported in this meta-analysis.
The minimal number of participants in a study was 14, and
the maximum was 199. Among the included studies, 60.7%
were male participants (1005/1656), and 2 studies did not
report the sex.12,13 PD stage was described using Hoehn and
Yahr’s disease stages (H&Y) in all trials except in 4 studies
that used the UPDRS motor score. All studies included par-
ticipants at H&Y stages from I to IV (Supplement Table 1).
Methodological Quality
PEDro scores ranged from 3 to 8 points (out of a maximum
of 10 points), indicating low to high methodological quality
(Supplement Table 2). The 2 reviewers who assessed the
RCTs had 20 disagreements among the 470 ratings, yield-
ing a Cohen’s κ of 0.914.The percentage of agreement was
91.4%, which suggested a “substantial” agreement between
the 2 investigators.14
Heterogeneity and Publication Bias
The test of heterogeneity indicated that the homogeneity of
the population studied was not met for comfortable walk-
ing speed (non-EXE control: Q = 40.0, df = 20, P = .005;
EXE control: Q = 38.0, df = 18, P = .004), stride or step
length non-EXE control (Q = 38.1, df = 15, P < .001),
and DGI non-EXE control (Q = 5.3, df = 1, P = .021).
The homogeneity of the population was met for fast walk-
ing speed (non-EXE control: Q = 1.16, df = 3, P = .76;
EXE control: Q = 3.3, df = 1, P = .07), stride or step
length cadence EXE control (Q = 22.3, df = 15, P = .10),
cadence (non-EXE control: Q = 5.9, df = 8, P = .663;
EXE control: Q = 5.0, df = 11, P = .933), DLST (non-
EXE control: Q = 1.1, df = 3, P = .770; EXE control: Q
= 4.5, df = 7, P = .726), TUG (non-EXE control: Q =
10.0, df = 9, P = .354; EXE control: Q = 14.9, df = 13, P
= .315), 6MWT (non-EXE control: Q = 2.1, df = 2, P =
.354; EXE control: Q = 10.5, df = 8, P = .229), DGI
EXE control (Q = 0.81, df = 1, P = .370), and FOG-Q
(non-EXE control: Q = 2.7, df = 2, P = .259; EXE con-
trol: Q = 1.5, df = 1, P = .216). Egger’s test indicated that
outcome measures were not significantly biased across all
publications.
Outcomes
Comfortable Walking Speed. The effect size for exercise
compared with non-EXE control was 0.449 with a standard
error (SE) of 0.11 and was statistically significant (95% CI
= 0.23 to 0.66, P < .001; Figure 1). The effect size for
exercise compared to EXE control was 0.241 with a SE of
0.13 but not statistically significant (95% CI = −0.052 to
0.50, P = .065; Figure 2).
Fast Walking Speed.  The Hedge’s g for fast walking speed
when comparing exercise with non-EXE control was 0.430,
with a SE of 0.19 and statistical significance (95% CI =
0.06 to 0.80; P = .023); the g when comparing exercise
with EXE control was 0.324, with a SE of 0.23, but not
statistically significant (95% CI = −0.12 to 0.77, P = .152;
Figure 3).
Stride or Step Length. The effect size comparing exercise
with non-EXE control was statistically significant (0.379),
with a SE of 0.14 (95% CI = 0.11 to 0.65, P = .006; Figure
4). But it was not significant when comparing exercise with
EXE control, with a g of 0.198 and a SE of 0.13 (95% CI =
−0.05 to 0.45, P = .118; Figure 5).
Cadence. The standard difference between exercise inter-
vention and non-EXE control (g = 0.223; SE = 0.12; 95%
CI = −0.01 to 0.46; P = .062) was not statistically signifi-
cant; however, for EXE control (g = 0.282; SE = 0.11;
95% CI = 0.06 to 0.51; P = .014), it was statistically sig-
nificant (Figure 6).
Double-Leg Support Time. No significant difference was
observed neither when comparing exercise with non-EXE
control, with a g of −0.135 and SE of 0.20 (95% CI = −0.52
to 0.25; P = .495), nor with EXE control, with a g of 0.095
and SE of 0.15 (95% CI = −0.19 to 0.38, P = .515; Figure 7).
Timed Up-and-Go. A significant difference was observed
when comparing exercise with non-EXE control, with a g
of −0.458 and SE of 0.13 (95% CI = −0.71 to −0.21, P <
.001), and EXE control, with a g of −0.390 and SE of 0.08
(95% CI = −0.55 to −0.23, P < .001; Figure 8).
Dynamic Gait Index.  No statistical significance was observed
when comparing exercise with non-EXE control, with a g
of 0.443 and SE of 0.24 (95% CI = −0.02 to 0.91; P =
.063), and EXE control, with a g of −0.063 and SE of 0.29
(95% CI = −0.63 to 0.50, P = .827; Figure 9).
Ni et al 875

Figure 1.  Forest plot of exercise effect on comfortable walking speed compared with nonexercise controls.
Abbreviations: EC, external cueing; PT, physical therapy.

Figure 2.  Forest plot of exercise effect on comfortable walking speed comparing to exercise controls.a
Abbreviations: EC, external cueing; IC, internal cueing; OG, overground; PT, physical therapy; TT, treadmill; UB exe, upper-body exercise.
a
Exercises in parentheses are the exercise control groups.
876 Neurorehabilitation and Neural Repair 32(10)

Figure 3.  Forest plot of exercise effect on fast walking speed. A. Comparison of exercise interventions with nonexercise controls. B.
Comparison of exercise interventions with exercise controls.a
Abbreviations: OG, overground; PT, physical therapy.
a
Exercises in parentheses are the exercise control groups.

Figure 4.  Forest plot of exercise effect on stride or step length compared with nonexercise controls.
Abbreviation: PT, physical therapy.

6-Minute Walk Test. No statistical significance was seen
comparing exercise with non-EXE controls (g = 0.341; SE
= 0.28; 95% CI = −0.21 to 0.90; P = .23) or EXE controls
(g = 0.190, SE = 0.15, 95% CI = −0.10 to 0.48, P = .20;
Figure 10).
Freezing of Gait Questionnaire. The random-effect model
was not statistically significant comparing exercise with
non-EXE controls (g = −0.057; SE = 0.24; 95% CI =
−0.52 to 0.41; P = .808) or EXE controls (g = −0.236, SE
= 0.33, 95% CI = −0.89 to 0.42, P = .480; Figure 11).
Ni et al 877

Figure 5.  Forest plot of exercise effect on or step length compared with exercise controls.a
Abbreviations: OG, overground; PT, physical therapy; TT, treadmill.
a
Exercises in parentheses are the exercise control groups.

Figure 6.  Forest plot of exercise effect on cadence. A. Comparison of exercise interventions with nonexercise controls. B.
Comparison of exercise interventions with exercise controls.a
Abbreviations: OG, overground; PT, physical therapy; TT, treadmill.
a
Exercises in parentheses are the exercise control groups.
878 Neurorehabilitation and Neural Repair 32(10)
Figure 7.  Forest plot of exercise effect on double-leg support time. A. Comparison of exercise interventions with nonexercise
controls. B. Comparison of exercise interventions with exercise controls.a
Abbreviations: TT, treadmill; PT, physical therapy.
a
Exercises in parentheses are the exercise control groups.
Effect of Moderators
Cuing. A significant effect of cuing (yes/no or unclear) was
only observed for the stride or step length during exercise com-
paring with EXE controls, where a larger effect for exercise
with cuing (g = 0.543; SE = 0.24; 95% CI = 0.06 to 1.0; P =
.026) than exercise without cuing or with unclear cuing (g =
0.357; SE = 0.21; 95% CI = −0.59 to 0.21) was observed.
Exercise Duration.  When the total length of intervention (6
to 45 hours or 360 to 4320 minutes) in the trials was used as
a moderator for gait performance, a significant effect of
exercise duration was only observed for the comfortable
walking speed during exercise comparing with EXE con-
trols. The estimated intercept and slope were 0.66 (SE =
0.22; 95% CI = 0.22 to 1.09; P = .003) and −0.023 (SE =
0.01; 95% CI = −0.03 to −0.003; P = .024). The estimated
mean effect size was 0.66 when the intervention of interest
was 6 hours; for each additional increase in hour of inter-
vention, the effect size decreased by 0.023, comparing with
another form of exercise.
Discussion
Exercise recommendation for improving gait functions are
shown in Table 1.
Exercise Mode
Our results indicate that variations in the magnitude of gait
improvement may depend on the types of training. Walking
function appears to be improved by (1) LSVT BIG; (2) multi-
dimensional physical training, including balance and gait train-
ing (excluding treadmill); and aquatic exercise; (3) treadmill
and cycling training; (4) resistance training focusing on lower-
body musculatures; and (5) Tai Chi, yoga, dance, and boxing.
LSVT BIG Program.  The LSVT BIG group15 showed a sig-
nificant improvement in TUG compared with nonsuper-
vised home exercise, which also included high-amplitude
movements. This suggests that patient-therapist face-to-
face training is recommended for achieving the greatest
improvement in motor function.
Multidimensional Physical Training. Researchers reported
results of function-based physical therapy with external
cues (music,16 auditory, and visual12,17) and physical ther-
apy combined with medication.18 Sensory-enhanced physi-
cal therapy produces significantly greater improvement in
gait parameters, including comfortable walking speed,17,18
stride length,16,17,19 cadence,16,17,19 and DLST.19 The
improvements in gait parameters may be attributable to
enhanced motor learning and retention.20
Ni et al 879

Figure 8.  Forest plot of exercise effect on timed up-and-go (TUG) test. A. Comparison of exercise interventions with nonexercise
controls. B. Comparison of exercise interventions with exercise controls.a
Abbreviations: NW, Nordic walk; OG, overground; PT, physical therapy; RT, resistance training; TT, treadmill.
a
Exercises in parentheses are the exercise control groups.

Figure 9.  Forest plot of exercise effect on Dynamic Gait Index (DGI). A. Comparison of exercise interventions with nonexercise
controls. B. Comparison of exercise interventions with exercise controls.a
Abbreviations: EC, external cueing; IC, internal cueing.
a
Exercises in parentheses are the exercise control groups.
880 Neurorehabilitation and Neural Repair 32(10)

Figure 10.  Forest plot of exercise effect on 6-minute walk test (6MWT). A. Comparison of exercise interventions with nonexercise
controls. B. Comparison of exercise interventions with exercise controls.a
Abbreviations: OG, overground; PT, physical therapy; TT, treadmill.
a
Exercises in parentheses are the exercise control groups.

Figure 11.  Forest plot of exercise effect on freezing of gait questionnaire (FOG-Q), compared with nonexercise or exercise
controls.a
Abbreviations: PT, physical therapy; TT, treadmill.
a
Exercises in parentheses are the exercise control groups.

Multidimensional physical intervention, emphasizing balance, step training with trunk rotation, functional train-
balance and gait components, have been gaining promi- ing, and postural re-education in one training session.
nence, where training includes flexibility and strengthening, Training adopting highly challenging tasks, such as obstacle
Ni et al 881

Table 1.  Evidence-Based Exercise Recommendations for Improving Gait Function for Parkinson’s Disease (H&Y I-III).

Frequency Intensity Time Duration Outcomes

LSVT BIG 4 d/wk “80% Of maximal energy” 60 min/d >4 Weeks GS, TUG
Multidimensional physical 3 d/wk — 30-60 min/d >4 Weeks Comfortable GS, stride
therapy, including length,a cadence, TUG,
balance and gait training 6MWT
(excluding treadmill)
Treadmill (0%-10% 2-3 d/wk Self-selected speed, increase 30-60 min/d >4 Weeks GS, stride length,a cadence,
loaded, 20% BWS, speed by 0.2 km/h, incline by double-leg support time,
downhill) 1%-2%, to 40%-50% HRR TUG, FOG-Q, 6MWT, DGI
Cycling 5 d/wk Progressively increased 60 min/d >3 Weeks GS, stride length, cadence,
workload with RPE of 11-14; double-leg support time,
60%-75% heart rate maximum TUG, 6MWT
Aquatic exercise Need further investigation
Resistance training 2-3 d/wk, 48 •• Weight machine: progressive, — >6 Weeks GS, stride length,a cadence,
Hours in high speed, 40%-60% 1RM, TUG, FOG-Q, 6MWT
between 10-12 repetitions, 3 sets/d,
major muscle groups
•• Weighted vest: lower-body
exercises with 1% initial body
weight, progress to 5% body
weight, RPE 15 (out of 20)
•• Ankle weights and elastic
band for multidimensional
ankle movement
Complementary therapy 2-3 d/wk — 60 min/d >12 Weeks GS, stride length,a TUG,
(Tai Chi/Yoga/Tango) FOG-Q

Abbreviations: 1RM, 1 repetition maximum; 6MWT, 6-minute walk test; BWS, body weight support; DGI, dynamic gait index; FOG, freezing of gait;
GS, gait speed; HRR, heart rate reserve; H&Y, Hoehn and Yahr; RPE, rate of perceived effort; TUG, timed up-and-go.
a
Cueing improves the performance.

negotiation, utilizing unstable surfaces, and dual task seems
effective to improve general motor function21-23 as well as
cognitive status.21,22
Another type of physical training is water-based exercise.
Three trials24-26 included in the current study (1 study27
included in the meta-analysis) compared the effects of
aquatic physiotherapy with conventional land therapy.
Neither group nor time differences were seen in overground
gait-related performance after these interventions. Therefore,
the capacity of aquatic training effects to transfer into over-
ground walking function needs further investigation.
Treadmill and Cycling Training.  These training methods have
been presented as a potentially effective approach to
reduce freezing of gait and promote a comfortable step-
ping pattern.28 Treadmill endurance training could be clas-
sified into 4 types: regular treadmill walking (unloaded),
body weight support (BWS), loaded, and downhill walk-
ing. Treadmill training combined with auditory and visual
cues yielded better performances in gait speed and stride
cycle than conventional physical therapy with external
cues.29 Three studies30-32 compared treadmill training with
overground gait training and reported that both training
modalities improve gait performances, including speed,30,32
stride or step length,32 cadence,30 and TUG.31 However,
the persistence of treatment effect between treadmill and
overground training seems to be different. In the work of
Bello et al,33 treadmill training produced significant
increases in comfortable walking speed, cadence, stride
length, and fast walking stride length after a 5-week train-
ing, and these improvements were maintained for 1 month.
In contrast, overground training only improved preferred
walking speed and cadence, and these improvements were
not maintained across the 1-month follow-up period. For
the treadmill with BWS or external load, Miyai et al34
showed that a 20% BWS was the most comfortable per-
centage of weight reduction compared with 0%, 10%, and
30% (least comfortable) BWS, whereas Toole et al35
reported that the 5% loaded and 25% BWS treadmill train-
ing produced similar performances in stride length and
cadence; but Trigueiro et al13 suggested that 10% loaded,
5% loaded, and unloaded treadmill training yielded simi-
lar improvements in gait speed, stride length, and DLST.
This might suggest that regular treadmill training would
be sufficient to induce improvement in gait function and
that loads from 20% BWS to 10% loaded treadmill train-
ing could also be utilized as alternatives. Downhill tread-
mill walking has also been shown to be effective for
882 Neurorehabilitation and Neural Repair 32(10)
generating higher walking speed and stride length, which
may be attributed to the imposed trunk extension in an
effort to pull the center of mass backward and decrease the
forward momentum produced by gravity during downhill
walking, which results in reduced trunk flexion and
improved posture.36 Also, it may be associated with the
increased knee extensor strength resulting from eccentric
contraction but lower oxygen demand during downhill
walk.37,38 Finally, it seems that the training session should
be supervised in order to induce significant improvements
in walking parameters because it was shown that a semisu-
pervised home-based treadmill training program was not
sufficient to generate improvement.39
Two trials evaluated the effect of cycling on gait per-
formance and compared the treatment effect with no
intervention19 and treadmill training.40 Sage and Almeida19
only observed improved step length, but Arcolin et al40
showed improved gait speed, step length, cadence, sin-
gle-leg support time and DLST, TUG, 6MWT, Mini-
BESTest, and UPDRS motor score, even though in this
analysis, the treatment effect of cycling is not as high as
that of treadmill (see Figures 1-6, 8). The difference in the
findings might result from the exercise dosage (Sage and
Almeida: 30 min/d, 3 d/wk; Arcolin et al: 2 sessions of 30
min/d, 5 d/wk). However, the smaller treatment effect
from cycling compared with treadmill may be a result of
biomechanical specificity because treadmill training is
similar to the task of walking, with sensory input pro-
vided by the treadmill.30 Moreover, it should be noted that
in the study by Arcolin et al,40 cycling and full body exer-
cises (strengthening, balance, and stretching) were com-
bined in the training program; this suggests that
intervention programs should be designed to be multidi-
mensional while maintaining a level of specificity in
order to maximize the treatment effect.
Resistance Training. Progressive resistance training, using
resistance machines41-43 or lower-body strengthening exer-
cise with a weighted vest25,44 or ankle weights44 or elastic
bands,41 has been shown to significantly increase gait
speed,41,43,44 stride length,41,44 cadence,41,43,44 and FOG-Q.25
Power-based resistance training,45,46 using high-speed and
moderate intensity, was shown to reduce bradykinesia,
increase movement speed,46 and improve balance,46 gait
performance,46 6-MWT time,45 and leg press strength45,46;
however, whether this intervention is superior to traditional
resistance training in PD is unknown. Two trials25,47 com-
bined resistance training with a balance component, but
only the work of Silva-Batista et al47 showed improved bal-
ance and strength after the resistance and stability training
compared with resistance training alone. The nature of
resistance may lead to improved postural control and walk-
ing ability through enhanced ankle and hip strategies48 and
neuromuscular relearning.44
Tai Chi, Yoga, Dance, and Boxing.  Three RCTs examined the
impact of Tai Chi on gait.44,49,50 They used Yang Short or
8-form styles, and 2 of them produced significant improve-
ment in the gait speed and stride length compared with no
intervention or stretching.44,50 Li et al44 showed a signifi-
cantly greater increase in stride length through Tai Chi com-
pared with resistance training. The improvement in gait
seen may be attributable to the characteristics of Tai Chi,
which uses large and alternating steps to effectively reduce
rigidity and enhance the ability to control the center of grav-
ity, and sustains knee bending during squatting and step-
ping. It is highly likely that Tai Chi could improve muscle
strength in the lower extremity and subsequently enhance
the gait performance, as evidenced by increased muscle
strength of bilateral knee extensors and flexors after Tai Chi
training.44 Another complementary therapy that continues
to gain attention is yoga. A specially designed power yoga
program using Vinyasa yoga style, provided by Ni et al,46
significantly improved mobility, including comfortable and
fast walking speed, balance, and agility (TUG). The fast
transition from one posture to another may have contributed
to improved postural control and increased movement
speed, whereas the repetitive and multidirectional step
training may have facilitated the motor learning process.
Finally, Argentinian tango, a type of dance therapy, has
been increasingly utilized because of its potential to address
freezing of gait51 and enhance internal rhythm by imposing
external rhythm from the music.52 Three RCTs evaluated
the effect of tango, but the findings varied. Romenets et al53
reported improvement in TUG performance but no signifi-
cant change in FOG-Q, whereas Duncan and Earhart54
reported improved FOG-Q, and McKee et al55 detected no
significant improvement in TUG or FOG-Q.
The combination of resistance, gait training, and boxing
specific activities was effective in producing increases in
walking speed and aerobic capacity compared with tradi-
tional physical therapy.56 In the study, participants were
trained as intensely as was tolerated, and repetitions were
increased during each 3-minute training interval. This high-
intensity interval training may be an alternative approach to
a task-specific training with a focus on gait.
Importance of Task-Specific Training. Current evidence sug-
gests that for training designed to enhance motor skill learn-
ing, both repetition and task-specific practice is required to
maximize training effect,57,58 and repeated practice of chal-
lenging movement tasks results in larger brain representa-
tions of the practiced movement.59,60 Based on the calculated
effect sizes, the treatment effect from a task-specific exer-
cise compared with a standard exercise control, such as
physical therapy or stretching, is even greater than that
compared with a nonexercise control. It is, therefore, logi-
cal that exercise emphasizing task specificity would con-
tribute to a higher quality of gait, especially during an
Ni et al 883
activity that requires greater physical and cognitive
demands. This is reflected by the successful outcomes in
variables included in our analysis, such as cadence and
FOG-Q. Finally, in addition to functional improvements,
task-specific practice has been associated with neuroplastic
changes within the cortical and subcortical areas of the
brain.61,62 These results provide strong support for the use of
gait-specific training when improvements in Parkinsonian
gait are the major goal.
Exercise Frequency/Intensity/Duration
Overall, it seems that 12 to 14 training sessions spreading
out over 2 to 4 weeks is the minimum exercise dosage for a
short-term improvement. Even though the metaregression
yielded a significant but negative relationship between total
exercise duration and exercise interventions on the comfort-
able walking speed (the treatment effect would decrease
with an increase in the duration of exercise), it should be
noted that the variations in the training effect resulting from
differences in population sample, exercise modality, and
training volume may have contributed to this finding or the
lack of significance.
LSVT BIG and Multidimensional Physical Training.  For a general
physical therapy program, the required intensity has not
been determined because of the variations in therapeutic
goal. A notable exception was the LSVT BIG program,
where intensity was specified as “80% of subject’s maximal
energy” for each repetition,15 with H&Y stage I to III. This
provides a guideline for implementing the LSVT BIG pro-
gram for PD patients with mild to moderate mobility limita-
tions; however, the LSVT program only yielded a trend
toward increased walking speed. This may be a result of the
low dosage of training, where the total training time was 960
minutes with a relatively short training duration (4 weeks)
and insufficient training frequency (60 min/d, 4 d/wk). In
contrast, a strategy-based physical therapy program used
with inpatients (45 min/d, 7 d/wk, 2 weeks) induced signifi-
cant increases in walking speed and 2-minute walking
endurance at discharge.12 However, it should be noted that
only the improvement in 2-minute walking endurance was
maintained during the 3-month follow-up. Consequently,
long-term effects need to be verified in future investigations
using higher exercise doses. For other programs,16,19,21-23,63,64
the exercise frequency varied among studies (2-7 times/wk),
but the majority of the trials used a frequency of 3 times/wk,
30 to 60 minutes per session, for 4 to 12 weeks, which
induced significantly improved gait function.
Treadmill and Cycling Training.  For treadmill training, training
with low intensity and longer duration produced the most
consistent improvements in gait speed. It appears unneces-
sary to utilize high-intensity walking to achieve these
benefits. This is fortunate because low-intensity exercise is
more accessible and safer for a larger group of patients.
Because the treadmill imposes external pace and enhances
the focus on gait, starting training at a comfortable walking
speed and gradually increasing to 40% to 50% heart rate
reserve is sufficient to make meaningful improvements.65
Other trials used the approach of starting with patients’ self-
selected comfortable pace and increasing the speed by 0.2
km/h,66 increasing the incline by 1% to 2%,35,36,65 or a com-
bination of both,65,66 thereby reaching 70% to 80% heart
rate maximum.13,67 This was accomplished without any step
difficulty or gait deviation. Training sessions ranging from
20 to 60 minutes, 2 to 3 times/wk appear sufficient to
increase gait function.13,31,32,34-36,65-72
For cycling training, moderate-intensity training (60%-
75% heart rate maximum,19 rate of perceived effort of 11-14
out of 2040), 30 to 60 minutes per day, can induce significant
improvement in stride length. However, other components,
including strengthening and balance, should be included in
the cycling regimen to maximize the treatment effect on
gait outcomes.
Resistance Training. Resistance training, 2 to 3 times per
week, is also recommended for significantly improving gait
function, lower-body strength, and power, which are critical
factors for predicting walking and balance functions.73 For
progressive resistance training,41-43 starting with 30% to
40% of 1 repetition maximum (1RM) for upper-body exer-
cises and 50% to 60% of 1RM for lower-body exercises and
increasing with 5% loads, 12 to 20 repetitions to volitional
fatigue, 2 to 3 sets per day, is recommended. For 2 strength
training programs using weighted vests, divergent results
were seen.25,44 Allen et al25 reported no significant increase
in gait function using gradually increasing weight to a max-
imum rate of perceived exertion of 15 out of 20. In contrast,
Li et al44 reported a significant increase in gait velocity
using an initial load of 1% of body weight that gradually
increased by 1% to 2% every 5 weeks until reaching the
maximum weight of 5% body weight. The difference in the
training responses between these 2 studies was likely a
result of the variation in loading and training duration
(Allen et al: 6 weeks; Li et al: 24 weeks). Furthermore, the
intensity of 1 power-based resistance training study used
each individual’s load that generated maximal power across
7 intensities, and these loads were adjusted with each power
plateau with 8 to 12 repetitions for each exercise and 3 sets
per training session over a 12-week course of training.46
However, the training devices, pneumatic resistance
machines, have been used nearly exclusively in the research
setting and have limited access in treatment facilities.
Another power-based resistance training program adopted 5
to 8 repetitions to volitional fatigue, 2 to 3 sets per training
session, and progressed with adding 5% to 10% additional
weights over 8 weeks.45 This program only showed
884 Neurorehabilitation and Neural Repair 32(10)
significant improvement in walking endurance (6MWT)
but not in mobility (TUG).
Tai Chi, Yoga, Dance, and Boxing.  For the 3 studies using Tai
Chi,44,49,50 the total training duration was 1200 to 2880 min-
utes with 1 hour training per day, 1 to 3 times per week. For
yoga46 and tango,53-55 it is likely that training 1 hour per
session, 2 times per week, for 12 weeks would be sufficient
to induce changes in gait function.
Effect of Cuing
External cuing, including visual and auditory rhythmic
cuing, has been widely used for gait training because it
helps normalize temporal and spatial gait parameters.74
Previous systematic reviews75,76 and a meta-analysis77 have
suggested that auditory cuing could significantly improve
walking speed,75 and visual cuing significantly improved
stride length.77 The effect sizes resulting from auditory
cuing reported in that meta-analysis were moderate for
walking speed (g = 0.544), stride length (g = 0.497), and
cadence (g = 0.556).77 In our analysis, the effect size of
cuing was significant for the stride length (g = 0.543), but
not significant for speed (g = 0.501) or cadence (0.079).
This indicates that using auditory or visual feedback is an
effective strategy if stride length is the target gait parameter.
The difference in the effect size could be a result of the dif-
ference in the analysis strategy, where the control groups
were separated into non-EXE controls or EXE controls, and
the types of cuing were not specified in our analysis.
Limitations
Given that PD is a progressive disease, long-term preserva-
tion of physical function is of marked importance to delay
disease-related and aging-related declines. Most published
studies confirm that exercise could have benefits on the
completion of an intervention program; however, the pro-
longed persistence of the treatment effect from exercise is
less known and should be investigated. In addition, most
existing studies evaluated the benefits of exercise for those
with mild to moderate PD (H&Y stage I-III) who were
capable of ambulation. Future research is needed to estab-
lish the feasibility and dose–response effect of exercise in
patients with a severe loss of gait function.
Conclusion
Exercise can improve gait performance, including walking
speed, stride length, cadence, and dynamic balance (double-
leg support, TUG, and DGI) in adults with PD. However, the
effect on walking endurance and freezing of gait needs fur-
ther investigation. The development of effective and sus-
tainable task-specific programs, which yield a short-term
positive effect as well as a persistent long-term positive
effect, is of great importance. This study offers evidence-
based exercise recommendations to help ensure optimal
exercise dosage to enhance gait function in patients with PD.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, author-
ship, and/or publication of this article.
References
1. Nussbaum RL, Ellis CE. Alzheimer’s disease and Parkinson’s
disease. N Engl J Med. 2003;348:1356-1364.
2. Giladi N, Balash Y, Rȗžička E, Jankovic J. Disorders of
gait. In: Jankovic J, Tolosa E, eds. Parkinson’s Disease &
Movement Disorders. 5th ed. Philadelphia, PA: Lippincott
Williams & Wilkins; 2007:436-458.
3. Morris ME. Movement disorders in people with Parkinson dis-
ease: a model for physical therapy. Phys Ther. 2000;80:578-
597.
4. Giladi N, McMahon D, Przedborski S, et al. Motor blocks in
Parkinson’s disease. Neurology. 1992;42:333-339.
5. Keus SHJ, Bloem BR, Hendriks EJ, Bredero-Cohen AB,
Munneke M; Practice Recommendations Development
Group. Evidence-based analysis of physical therapy in
Parkinson’s disease with recommendations for practice and
research. Mov Disord. 2007;22:451-460.
6. Goodwin VA, Richards SH, Taylor RS, Taylor AH, Campbell
JL. The effectiveness of exercise interventions for people with
Parkinson’s disease: a systematic review and meta-analysis.
Mov Disord. 2008;23:631-640.
7. Allen NE, Sherrington C, Paul SS, Canning CG. Balance
and falls in Parkinson’s disease: a meta-analysis of the effect
of exercise and motor training. Mov Disord. 2011;26:1605-
1615.
8. Lima LO, Scianni A, Rodrigues-de-Paula F. Progressive
resistance exercise improves strength and physical perfor-
mance in people with mild to moderate Parkinson’s disease: a
systematic review. J Physiother. 2013;59:7-13.
9. Mehrholz J, Kugler J, Storch A, Pohl M, Hirsch K, Elsner
B. Treadmill training for patients with Parkinson’s disease.
Cochrane Database Syst Rev. 2015;(9):CD007830.
10. Da Rocha PA, McClelland J, Morris ME. Complementary
physical therapies for movement disorders in Parkinson’s
disease: a systematic review. Eur J Phys Rehabil Med.
2015;51:693-704.
11. Cotoi A, Teasell R. Introduction and methods. http://www.
ebrsr.com/evidence-review/1-introduction-and-methods.
Accessed September 6, 2018.
12. Morris ME, Iansek R, Kirkwood B. A randomized controlled
trial of movement strategies compared with exercise for peo-
ple with Parkinson’s disease. Mov Disord. 2009;24:64-71.
13. Trigueiro LC, Gama GL, Simão CR, Sousa AV, Júnior Cde
OG, Lindquist AR. Effects of treadmill training with load on
Ni et al 885
gait in Parkinson disease: a randomized controlled clinical
trial. Am J Phys Med Rehabil. 2015;94(10, suppl 1):830-837.
14. Mi MY, Collins JE, Lerner V, Losina E, Katz JN. Reliability
of medical record abstraction by non-physicians for orthope-
dic research. BMC Musculoskelet Disord. 2013;14:181.
15. Ebersbach G, Ebersbach A, Edler D, et al. Comparing exer-
cise in Parkinson’s disease—the Berlin LSVT®BIG study.
Mov Disord. 2010;25:1902-1908.
16. de Bruin N, Doan JB, Turnbull G, et al. Walking with music is
a safe and viable tool for gait training in parkinson’s disease:
the effect of a 13-week feasibility study on single and dual
task walking. Parkinsons Dis. 2010;2010:483530.
17. Thaut MH, McIntosh GC, Rice RR, Miller RA, Rathbun J,
Brault JM. Rhythmic auditory stimulation in gait training for
Parkinson’s disease patients. Mov Disord. 1996;11:193-200.
18. Ellis T, de Goede CJ, Feldman RG, Wolters EC, Kwakkel
G, Wagenaar RC. Efficacy of a physical therapy program in
patients with Parkinson’s disease: a randomized controlled
trial. Arch Phys Med Rehabil. 2005;86:626-632.
19. Sage MD, Almeida QJ. Symptom and gait changes after
sensory attention focused exercise vs aerobic training in
Parkinson’s disease. Mov Disord. 2009;24:1132-1138.
20. Dam M, Tonin P, Casson S, et al. Effects of conventional and
sensory-enhanced physiotherapy on disability of Parkinson’s
disease patients. Adv Neurol. 1996;69:551-555.
21. Conradsson D, Löfgren N, Nero H, et al. The effects of highly
challenging balance training in elderly with Parkinson’s dis-
ease: a randomized controlled trial. Neurorehabil Neural
Repair. 2015;29:827-836.
22. Landers MR, Hatlevig RM, Davis AD, Richards AR, Rosenlof
LE. Does attentional focus during balance training in people
with Parkinson’s disease affect outcome? A randomised con-
trolled clinical trial. Clin Rehabil. 2016;30:53-63.
23. Wong-Yu IS, Mak MK. Multi-dimensional balance training
programme improves balance and gait performance in people
with Parkinson’s disease: a pragmatic randomized controlled
trial with 12-month follow-up. Parkinsonism Relat Disord.
2015;21:615-621.
24. Vivas J, Arias P, Cudeiro J. Aquatic therapy versus conven-
tional land-based therapy for Parkinson’s disease: an open-
label pilot study. Arch Phys Med Rehabil. 2011;92:1202-1210.
25. Allen NE, Canning CG, Sherrington C, et al. The effects
of an exercise program on fall risk factors in people with
Parkinson’s disease: a randomized controlled trial. Mov
Disord. 2010;25:1217-1225.
26. Carroll LM, Volpe D, Morris ME, Saunders J, Clifford AM.
Aquatic exercise therapy for people with Parkinson dis-
ease: a randomized controlled trial. Arch Phys Med Rehabil.
2017;98:631-638.
27. Ayan C, Cancela J, Gutiérrez-Santiago A, Prieto I. Effects of
two different exercise programs on gait parameters in indi-
viduals with Parkinson’s disease: a pilot study. Gait Posture.
2014;39:648-651.
28. Bello O, Marquez G, Camblor M, Fernandez-Del-Olmo M.
Mechanisms involved in treadmill walking improvements in
Parkinson’s disease. Gait Posture. 2010;32:118-123.
29. Frazzitta G, Maestri R, Uccellini D, Bertotti G, Abelli P.
Rehabilitation treatment of gait in patients with Parkinson’s
disease with freezing: a comparison between two physical
therapy protocols using visual and auditory cues with or with-
out treadmill training. Mov Disord. 2009;24:1139-1143.
30. Bello O, Fernandez-Del-Olmo M. How does the treadmill affect
gait in Parkinson’s disease? Curr Aging Sci. 2012;5:28-34.
31. Chaiwanichsiri D, Wangno W, Kitisomprayoonkul W,
Bhidayasiri R. Treadmill training with music cueing: a new
approach for Parkinson’s gait facilitation. Asian Biomed.
2011;5:649-654.
32. Almeida S, Riddell MC, Cafarelli E. Slower conduction
velocity and motor unit discharge frequency are associated
with muscle fatigue during isometric exercise in type 1 diabe-
tes mellitus. Muscle Nerve. 2008;37:231-240.
33. Bello O, Sanchez JA, Lopez-Alonso V, et al. The effects of
treadmill or overground walking training program on gait in
Parkinson’s disease. Gait Posture. 2013;38:590-595.
34. Miyai I, Fujimoto Y, Yamamoto H, et al. Long-term effect of
body weight-supported treadmill training in Parkinson’s dis-
ease: a randomized controlled trial. Arch Phys Med Rehabil.
2002;83:1370-1373.
35. Toole T, Maitland CG, Warren E, Hubmann MF, Panton L.
The effects of loading and unloading treadmill walking on
balance, gait, fall risk, and daily function in Parkinsonism.
NeuroRehabilitation. 2005;20:307-322.
36. Yang YR, Lee YY, Cheng SJ, Wang RY. Downhill walking
training in individuals with Parkinson’s disease: a randomized
controlled trial. Am J Phys Med Rehabil. 2010;89:706-714.
37. Kuster M, Sakurai S, Wood G. Kinematic and kinetic com-
parison of downhill and level walking. Clin Biomech (Bristol,
Avon). 1995;10:79-84.
38. Gault ML, Willems ME. Isometric strength and steadi-
ness adaptations of the knee extensor muscles to level and
downhill treadmill walking in older adults. Biogerontology.
2013;14:197-208.
39. Canning CG, Allen NE, Dean CM, Goh L, Fung VS. Home-
based treadmill training for individuals with Parkinson’s
disease: a randomized controlled pilot trial. Clin Rehabil.
2012;26:817-826.
40. Arcolin I, Pisano F, Delconte C, et al. Intensive cycle ergom-
eter training improves gait speed and endurance in patients
with Parkinson’s disease: a comparison with treadmill train-
ing. Restorative Neurol Neurosci. 2015;34:125-138.
41. Hass CJ, Buckley TA, Pitsikoulis C, Barthelemy EJ.
Progressive resistance training improves gait initiation in indi-
viduals with Parkinson’s disease. Gait Posture. 2012;35:669-
673.
42. Schlenstedt C, Paschen S, Kruse A, Raethjen J, Weisser B,
Deuschl G. Resistance versus balance training to improve
postural control in Parkinson’s disease: a randomized rater
blinded controlled study. PLoS One. 2015;10:e0140584.
43. Rafferty MR, Prodoehl J, Robichaud JA, et al. Effects of 2
years of exercise on gait impairment in people with Parkinson
disease: the PRET-PD randomized trial. J Neurol Phys Ther.
2017;41:21-30.
44. Li F, Harmer P, Fitzgerald K, et al. Tai chi and postural sta-
bility in patients with Parkinson’s disease. N Engl J Med.
2012;366:511-519.
45. Schilling BK, Pfeiffer RF, Ledoux MS, Karlage RE, Bloomer
RJ, Falvo MJ. Effects of moderate-volume, high-load lower-
body resistance training on strength and function in persons
886 Neurorehabilitation and Neural Repair 32(10)
with Parkinson’s disease: a pilot study. Parkinsons Dis.
2010;2010:824734.
46. Ni M, Signorile JF, Mooney K, et al. Comparative effect
of power training and high-speed yoga on motor function
in older patients with Parkinson’s disease. Arch Phys Med
Rehabil. 2016;97:345-354.e15.
47. Silva-Batista C, Corcos DM, Roschel H, et al. Resistance
training with instability for patients with Parkinson’s disease.
Med Sci Sports Exerc. 2016;48:1678-1687.
48. Judge JO, Davis RB III, Ounpuu S. Step length reductions in
advanced age: the role of ankle and hip kinetics. J Gerontol A
Biol Sci Med Sci. 1996;51:M303-M312.
49. Amano S, Nocera JR, Vallabhajosula S, et al. The effect of
Tai Chi exercise on gait initiation and gait performance in
persons with Parkinson’s disease. Parkinsonism Relat Disord.
2013;19:955-960.
50. Hackney ME, Earhart GM. Tai Chi improves balance and
mobility in people with Parkinson disease. Gait Posture.
2008;28:456-460.
51. Hackney ME, Kantorovich S, Levin R, Earhart GM. Effects
of tango on functional mobility in Parkinson’s disease: a pre-
liminary study. J Neurol Phys Ther. 2007;31:173-179.
52. Pacchetti C, Mancini F, Aglieri R, Fundarò C, Martignoni
E, Nappi G. Active music therapy in Parkinson’s disease: an
integrative method for motor and emotional rehabilitation.
Psychosom Med. 2000;62:386-393.
53. Romenets SR, Anang J, Fereshtehnejad S, Pelletier A,
Postuma R. Tango for treatment of motor and non-motor
manifestations in Parkinson’s disease: a randomized control
study. Complement Ther Med. 2015;23:175-184.
54. Duncan RP, Earhart GM. Randomized controlled trial of
community-based dancing to modify disease progression in
Parkinson disease. Neurorehabil Neural Repair. 2012;26:132-
143.
55. McKee KE, Hackney ME. The effects of adapted tango on
spatial cognition and disease severity in Parkinson’s disease.
J Mot Behav. 2013;45:519-529.
56. Combs SA, Diehl MD, Chrzastowski C, et al. Community-
based group exercise for persons with Parkinson dis-
ease: a randomized controlled trial. NeuroRehabilitation.
2013;32:117-124.
57. Hubbard IJ, Parsons MW, Neilson C, Carey LM. Task-
specific training: evidence for and translation to clinical prac-
tice. Occup Ther Int. 2009;16:175-189.
58. Bayona NA, Bitensky J, Salter K, Teasell R. The role of
task-specific training in rehabilitation therapies. Top Stroke
Rehabil. 2005;12:58-65.
59. Nudo RJ, Milliken GW, Jenkins WM, Merzenich MM. Use-
dependent alterations of movement representations in pri-
mary motor cortex of adult squirrel monkeys. J Neurosci.
1996;16:785-807.
60. Karni A, Meyer G, Jezzard P, Adams MM, Turner R,
Ungerleider LG. Functional MRI evidence for adult motor
cortex plasticity during motor skill learning. Nature.
1995;377:155-158.
61. Adkins DL, Boychuk J, Remple MS, Kleim JA. Motor
training induces experience-specific patterns of plasticity
across motor cortex and spinal cord. J Appl Physiol (1985).
2006;101:1776-1782.
62. Page SJ, Szaflarski JP, Eliassen JC, Pan H, Cramer SC.
Cortical plasticity following motor skill learning dur-
ing mental practice in stroke. Neurorehabil Neural Repair.
2009;23:382-388.
63. Kadivar Z, Corcos DM, Foto J, Hondzinski JM. Effect of step
training and rhythmic auditory stimulation on functional per-
formance in Parkinson patients. Neurorehabil Neural Repair.
2011;25:626-635.
64. Stożek J, Rudzińska M, Pustułka-Piwnik U, Szczudlik A. The
effect of the rehabilitation program on balance, gait, physical
performance and trunk rotation in Parkinson’s disease. Aging
Clin Exp Res. 2016;28:1169-1177.
65. Shulman LM, Katzel LI, Ivey FM, et al. Randomized clinical
trial of 3 types of physical exercise for patients with Parkinson
disease. JAMA Neurol. 2013;70:183-190.
66. Nadeau A, Pourcher E, Corbeil P. Effects of 24 wk of tread-
mill training on gait performance in Parkinson disease. Med
Sci Sports Exerc. 2014;46:645-655.
67. Kurtais Y, Kutlay S, Tur BS, Gok H, Akbostanci C. Does
treadmill training improve lower-extremity tasks in Parkinson
disease? A randomized controlled trial. Clin J Sport Med.
2008;18:289-291.
68. Protas EJ, Mitchell K, Williams A, Qureshy H, Caroline K,
Lai EC. Gait and step training to reduce falls in Parkinson’s
disease. NeuroRehabilitation. 2005;20:183-190.
69. Cakit BD, Saracoglu M, Genc H, Erdem HR, Inan L. The
effects of incremental speed-dependent treadmill training on
postural instability and fear of falling in Parkinson’s disease.
Clin Rehabil. 2007;21:698-705.
70. Schlick C, Ernst A, Bötzel K, Plate A, Pelykh O, Ilmberger
J. Visual cues combined with treadmill training to improve
gait performance in Parkinson’s disease: a pilot randomized
controlled trial. Clin Rehabil. 2016;30:463-471.
71. Fisher BE, Wu AD, Salem GJ, et al. The effect of exercise
training in improving motor performance and corticomotor
excitability in people with early Parkinson’s disease. Arch
Phys Med Rehabil. 2008;89:1221-1229.
72. Ganesan M, Sathyaprabha TN, Pal PK, Gupta A. Partial body
weight-supported treadmill training in patients with Parkinson
disease: impact on gait and clinical manifestation. Arch Phys
Med Rehabil. 2015;96:1557-1565.
73. Li F, Fisher KJ, Harmer P, McAuley E. Delineating the
impact of Tai Chi training on physical function among the
elderly. Am J Prev Med. 2002;23(2, suppl):92-97.
74. Behrman AL, Teitelbaum P, Cauraugh JH. Verbal instruc-
tional sets to normalise the temporal and spatial gait vari-
ables in Parkinson’s disease. J Neurol Neurosurg Psychiatry.
1998;65:580-582.
75. Lim I, van Wegen E, de Goede C, et al. Effects of external
rhythmical cueing on gait in patients with Parkinson’s dis-
ease: a systematic review. Clin Rehabil. 2005;19:695-713.
76. Rubinstein TC, Giladi N, Hausdorff JM. The power of cueing
to circumvent dopamine deficits: a review of physical therapy
treatment of gait disturbances in Parkinson’s disease. Mov
Disord. 2002;17:1148-1160.
77. Spaulding SJ, Barber B, Colby M, Cormack B, Mick T,
Jenkins ME. Cueing and gait improvement among people
with Parkinson’s disease: a meta-analysis. Arch Phys Med
Rehabil. 2013;94:562-570.