NeuroRehabilitation 20 (2005) 183–190 183

IOS Press

Gait and step training to reduce falls in

Parkinson’s disease
Elizabeth J. Protasa,∗, Katy Mitchelld , Amanda Williamsb , Huma Qureshyb, Kavitha Carolineb and
Eugene C. Laib,c
a
Department of Physical Therapy, University of Texas Medical Branch, Galveston, TX, USA
b
Parkinson’s Disease Research, Education and Clinical Center, Michael E. DeBakey Veterans Affairs Medical
Center, Houston, TX, USA
c
Department of Neurology, Baylor College of Medicine, Houston, TX, USA
d
School of Physical Therapy, Texas Woman’s University, Houston, TX, USA

Abstract. Introduction: Frequent falls and risk of injury are evident in individuals with Parkinson’s disease (PD) as the disease
progresses. There have been no reports of any interventions that reduce the incidence of falls in idiopathic PD.
Purpose: Assess the benefit of gait and step perturbation training in individuals with PD.
Design: Randomized, controlled trial.
Setting: Outpatient research, education and clinical center in a tertiary care Veterans Affairs Medical Center.
Outcome measures: Gait parameters, 5-step test, report of falls
Subjects: Eighteen men with idiopathic PD in stage 2 or 3 of the Hoehn and Yahr staging
Methods: Subjects were randomly assigned to a trained or control group. They were asked about any falls 2 weeks prior to and
after an 8 week period. Gait speed, cadence, and step length were tested on an instrumented walkway. Subjects were timed while
stepping onto and back down from an 8.8 cm step for 5 consecutive steps. Gait training consisted of walking on a treadmill at
a speed greater than over ground walking speed while walking in 4 directions and while supported in a harness for safety. Step
training consisted of suddenly turning the treadmill on and off while the subject stood in the safety harness facing either forwards,
backwards, or sideways. Training occurred 1 hour per day, three times per week for 8 weeks. A two-factor (time and group)
analysis of variance with repeated measures was used to compare the groups.
Results: Substantial reduction occurred in falls in the trained group, but not in the control group. Gait speed increased in the
trained group from 1.28 ± 0.33 meters/sec to 1.45 ± 0.37 meters/sec, but not in the control group (from 1.26 to 1.27 m/s). The
cadence increased for both groups: from 112.8 to 120.3 steps/min for the trained group and 117.7 to 124.3 steps/min for the
control group. Stride lengths increased for the trained group, but not the control group. The 5-step test speed increased in the
trained group from 0.40 ± 0.08 steps/sec to 0.51 ± 0.12 steps/sec, and in the control group (0.36 ± 0.11 steps/sec to 0.42 ± 0.11
steps/sec).
Conclusion: Gait and step perturbation training resulted in a reduction in falls and improvements in gait and dynamic balance.
This is a promising approach to reduce falls for patients with PD.

Keywords: Parkinson’s disease, falls, gait training, rehabilitation

1. Introduction

Gait abnormalities are one of the most common dis-

∗ Address
abling conditions in Parkinson’s disease (PD) [31]. In-
for correspondence: Elizabeth J. Protas, P.T., Ph.D.,
dividuals with PD have a gait pattern characterized by
FACSM, Department of Physical Therapy, University of Texas Med-
ical Branch, Galveston, TX 77555-1144, USA. Tel.: +1 409 772 hesitant, shuffling steps that are short and quick. Diffi-
3068; Fax: +1 409 747 1613; E-mail: [email protected]. culties in gait initiation and changes in postural control

ISSN 1053-8135/05/$17.00 © 2005 – IOS Press and the authors. All rights reserved
184 E.J. Protas et al. / Gait and step training to reduce falls in Parkinson’s disease

are also problematic [2]. Turning is difficult because individuals with neurologic deficits [16,23]. Several
it requires a series of gait initiations [3]. Freezing and recent reports suggest that ambulation training using
motor blocks, balance deficits, and frequent falls oc- body weight support for individuals with Parkinson’s
cur during latter stages of PD. Although antiparkiso- disease (PD) results in improvements in gait [12,19,20].
nian medications improve gait, their effectiveness de- These reports were either case reports or of small sam-
creases as the disease progresses. Other forms of inter- ple sizes. The studies used varying training intervals,
ventions have limited impact on gait and balance over or used no balance or fall measures.
time. Therefore, there is a need to explore alternative, We previously demonstrated with a case report that
rehabilitation interventions to improve gait and balance a gait and step training strategy improved gait and bal-
impairments. ance, and reduced falls in an individual with a parkin-
Few studies have tried to isolate different compo- sonian syndrome [40]. The purpose of this article is to
nents of rehabilitation interventions. Several studies report the results of a pilot controlled study of gait and
examined the efficacy of using motor control strate- step training in individuals with idiopathic PD who had
gies in patients with PD. Stefaniwsky and Bilowit [38] reduced balance, and/or recent problems with falls.
comparing 10 patients with PD and 5 healthy subjects
found that movement initiation was significantly slower
in the patients with PD compared to healthy subjects.
After daily in-home exercises using sensory stimuli to 2. Methodology
facilitate movement initiation for a 3 week period, the
patients demonstrated comparable movement initiation 2.1. Sample
speeds to the healthy individuals. In another study,
standing weight shifting was compared in 34 people Eighteen men with idiopathic PD diagnosed at the
with stage I or II PD and 34 neurologically intact sub- Houston VA Parkinson’s Disease Research, Education
jects [15]. Reduced ability to shift weight from one
and Clinical Center (PADRECC) were recruited for this
lower extremity to another was observed in both dis-
study. The demographic and clinical characteristics of
ability stages of PD when compared to the controls.
the subjects are shown in Table 1. Inclusion criteria
All groups improved weight shifting responses using
were: 1) postural instability-gait difficulty predomi-
continuous video feedback when compared to end of
nant PD, 2) experiences with freezing episodes, and/or
trial feedback. Although this study only used a single
a history of falls, 3) stable regimen of antiparkinso-
session, and did not use a training strategy over sev-
eral weeks, it supports the notion that sensory stimuli nian medications, 4) ability to stand and walk with or
can influence movement control in individuals with PD without assistance, 5) stage 2 or 3 of the Hoehn and
when performing complex motor tasks. In a similar Yahr staging [11], and 6) scores of moderate or higher
fashion, visual and auditory sensory cues as well as on all scales of the Neurobehavioral Cognitive Status
the anti-parkinsonian medication levodopa can modify Examination (Cognistat) [24]. The subjects provided
gait movements and muscle activation in some patients informed consent as approved by the Institutional Re-
with PD [21,28,41]. Schenkman et al. [34] reported view Boards of Baylor College of Medicine and Texas
that improved spinal flexibility improved measures of Woman’s University, and were randomly assigned to
balance in individuals with PD. Similarly, posteroven- either the training or control groups.
tral pallidotomy has been reported to influence postural
control in PD [30].
2.2. Measurement
Studies have been published demonstrating that hu-
mans with neurological conditions can improve gait by
either training ambulation on a treadmill with no body Subjects were tested in the morning, and were asked
weight support or with body weight support [6,7,10,17, to take their morning dose of medications about 1 hour
27,43,44,46–48]. The degree of locomotor recovery prior to the test in order to assure that subjects were
has been shown to be significantly related to the training at their best ‘on’ state. All testing except for the fall
used in patients with neurological conditions [10,29, record was conducted by a physical therapist and a
43]. Other studies comparing task-specific gait training technician who were blinded to the subject’s group as-
with body-weight supported training report that simi- signment. A physical therapist who was not blinded to
lar gait outcomes occur with either training strategy in group assignment obtained fall records.
 E.J. Protas et al. / Gait and step training to reduce falls in Parkinson’s disease 185

Table 1
Subject characteristics by group
Variable Group
Trained (n = 9) Control (n = 9)
Mean SD Mean SD
Age 71.3 7.4 73.7 8.5
Height (m) 1.81 0.07 1.78 0.07
Weight (kg) 83.5 12.3 83.4 15.0
Years since Diagnosis 7.1 5.1 8.1 4.4
Hoehn & Yahr Stage 2.8 0.35 2.9 0.17
UPDRS Motor Scorea 28.3 13.6 30.4 8.0
Activities of Daily Living 82.2 8.7 80.6 7.3
a Unified Parkinson’s Disease Rating Scale.

2.2.1. Gait parameters This was repeated for the two-week period after the
Gait was assessed by gait speed, cadence, and stride completion of the training.
length. The patient was asked to walk as fast as possible
with any assistive device necessary on an instrumented, 2.2.4. Balance parameters
3-meter walkway (GAITRite, CIR Systems Inc., PO Dynamic balance was assessed by timing 5 consec-
Box 4402, Clifton, NJ) while the subject was guarded utive steps up and back down a 8.8 cm step (step test).
by a physical therapist to prevent falls. The patient The patient was asked to perform the steps as quickly
completed 2 trials on the walkway, and the average of and safely as possible while being guarded for safety.
the results from these 2 trials was used as data. Gait This test has been reported to be reliable in elderly
speed was calculated from the time to walk 3 meters as subjects [22].
meters/second. Cadence was the number of steps/sec.,
and the stride length was the length (in centimeters) of 2.2.5. Patient characteristics
two consecutive footfalls of the same extremity. Spatial The investigators asked each patient for a history
and temporal parameters measured with the GAITRite of the PD disease and medications. Cognitive status
have been reported to be reliable (ICC > 0.93) and valid was determined by the Cognistat, a 10-item, reliable
(ICC > 0.93) [8,18]. Gait speed has also been reported measure [24]. The Unified Parkinson’s Disease Rating
to be a reliable measure for individuals with PD with Scale (UPDRS), a standardized measure of PD impair-
intra-class correlation coefficients of 0.87 [33]. ment, was administered at baseline [9].

2.2.2. Freezing of gait 2.3. Procedures

The provocative test for freezing and motor blocks
was used to assess freezing [42]. The subject was After the subject gave informed consent, the subject
asked to stand, walk 5 meters between two chairs that was asked to provide demographic information, a his-
were placed 1 meter apart in the path, turn 180 ◦ , and tory of his PD, and current medications. An investi-
walk back through the chairs, and sit back down. The gator completed the Cognistat and the UPDRS. If the
rater scored start hesitation, sudden transient blocks that subject met the inclusion/exclusion criteria, he was ran-
interrupt gait, motor blocks on turning, motor blocks domly assigned to either the gait and step training inter-
on reaching a target (chairs), and motor blocks when vention group or a control group who only received the
walking through the chairs (narrow space). The tasks pre- and post-testing. An investigator called the subject
were rated as not observed (0 = no) or observed (1 = daily for 2 weeks to establish a falls history prior to
yes). The maximum score for freezing was 5. completing the pre-testing and after the 8-week period
for both groups.
2.2.3. Fall frequency
Each subject was contacted daily by telephone for a 2.3.1. Gait and step training
period of 2 weeks prior to starting the 8 week training Subjects assigned to the training group received
or control sessions. The patient was asked if he fell training 3 times per week for 8 weeks. The subject was
that day, under what circumstances, and whether or not fitted with a harness that was attached snugly around the
the fall resulted in any injuries. The number of falls trunk (Quinton Pneu-Weight harness,Seattle, WA). The
and fall history for this two-week period was recorded. harness was attached to a pneumatic support system po-
186 E.J. Protas et al. / Gait and step training to reduce falls in Parkinson’s disease

Table 2
Gait and step test results by group
Variable Group
Trained (n = 9) Control (n = 9)
Pre Post Pre Post
Gait speed (m/s) 1.28 1.45 1.26 1.27
(0.33) (0.37) (0.19) (0.25)
Cadence (steps/min) 112.8 120.3 117.7 124.3
(7.2) (8.2) (13.0) (15.1)
Stride length right (cm) 66.5 71.1 60.2 60.4
(13.7) (14.4) (13.3) (10.0)
Stride length left (cm) 68.7 72.9 61.0 60.8
(14.9) (17.0) (15.4) (10.9)
Step test (steps/s) 0.40 0.51 0.36 0.42
(0.08) (0.12) (0.11) (0.11)

sitioned over the treadmill (Quinton Pneu-Weight Sup- of 15–20 perturbations in the forward and backward
port System, Seattle, WA). The harness and support direction and 10 to 15 for both right and left sideways
system was used for safety in case the subject fell. No directions. If the subject fatigued during step training,
body-weight support was used. The subject was also the subject was allowed to take a short rest. If a sub-
guarded by a physical therapist during the training. Ini- ject missed more than 3 training sessions in a row for
tially, the treadmill speed was set at the fastest over- medical reasons or inability to participate, the subject
ground speed noted on the pre-test while the subject could be discontinued from the study. If an occasional
walked in a forward direction on the treadmill. The session was missed, the subject was allowed to make
patient walked forward in the support system for 5 to up the session until a total of 24 training sessions had
7 minutes each session. The subject was then asked to occurred before post-testing.
walk at his fastest, self-selected speed that allowed a
full step while walking backwards on the treadmill for 2.4. Data analysis
5 to 7 minutes. The subject then walked at his fastest,
self-selected speed that allowed a full step sideways The data were analyzed descriptively in order to de-
both right and left. The subject walked sideways for termine group means and standard deviations for the
2 to 3 minutes in each direction. The therapist cued measures. Group means were compared with a multi-
the patient verbally, and assisted the patient stepping variate analysis of variance for repeated measures with
if necessary. The treadmill speed was reassessed at two factors (time and group). Post-hoc analysis was
the end of every training week, with the goal of gradu- used for significant differences. The alpha level was
ally increasing the treadmill speed and time as training set at < 0.05.
progressed. The subject was allowed to rest if fatigue
occurred during gait training.
After a 5 minute rest, the patient underwent step 3. Results
training while standing on the treadmill in the support
system when the treadmill was suddenly turned on at a 3.1. Pre- and post-gait outcomes
sufficiently fast speed to perturb the subject’s standing
balance, but not fall. The subject was asked to take There were no significant differences on the subject
several steps to recover balance in response to this sud- characteristics between the trained and control groups
den perturbation. The treadmill was then turned off, (Table 1). Overall, the subjects were moderately im-
and the subject had to recover balance again. The step paired from the PD. Descriptive statistics for gait speed,
training occurred while the patient stood in four direc- cadence, right and left stride length, and the step test
tions: forward, backwards, and sideways right and left. are shown in Table 2. Significant differences occurred
Initially, the subject was allowed to hold the handrail, post-test compared to pre-test for the trained group for
but, as training progressed, the subject stood without gait speed, cadence and the step test (Table 3). Sig-
holding the handrail during perturbations and the tread- nificant differences occurred for the control group on
mill speed was gradually increased. The number of post-testing compared to pre-testing on cadence and
trials in each position varied, but generally consisted the step test. No significant differences occurred pre-
 E.J. Protas et al. / Gait and step training to reduce falls in Parkinson’s disease 187

Table 3
Results of the analyses of the effects of time and group by time for each of the measures
Variable Sum of squares df Mean square F Sig. Power
Time
Gait speed 7.86 1 7.86 8.81 0.009 0.80
Cadence 440.3 1 440.3 9.26 0.008 1.0
Stride Length Right 50.43 1 50.43 1.19 0.29 0.18
Stride Length Left 33.25 1 33.25 0.94 0.35 0.15
Step Test 6.82 1 6.82 20.2 0.000 0.99
Group by Time
Gait Speed 4.95 1 4.95 5.55 0.032 0.60
Cadence 1.91 1 1.91 0.04 0.84 0.05
Stride Length Right 43.6 1 43.6 1.03 0.33 0.16
Stride Length Left 43.6 1 43.6 1.23 0.28 0.18
Step Test 5.21 1 5.21 1.54 0.23 0.22
Error
Gait Speed 0.143 16 8.92
Cadence 760.9 16 47.6
Stride Length Right 676.8 16 42.3
Stride Length Left 565.9 16 35.4
Step Test 5.42 16 3.39

4. Discussion

This is the first report of reduced falls in individuals

with PD as a result of a physical training intervention.
This is an important finding since falling is a serious
problem as the disease progresses that can lead to injury
and possibly death. This observation was accompanied
by a statistically significant increase in gait speed, ca-
dence, and the step test in the trained group. Although
not significant, the stride length also increased in the
trained group. We conclude that reduced falls were as-
Fig. 1. Reported falls prior to and following an 8 week period for the
sociated with improvements in gait speed and dynamic
trained and control groups. balance.
Ashburn and her group compared individuals with
PD who were fallers and non-fallers and reported that
to post-test for stride length, although the right and left
40% of the subjects had fallen within the last 12
stride length increased for the intervention group but
months [1]. The fallers reported a median of 3 falls
not for the control. The only significant gait measure
in 12 months. Furthermore, the fallers were more im-
between the groups was for post-test gait speed. None paired and had poorer measures of mobility and bal-
of the subjects demonstrated freezing during our motor ance than the non-fallers. Seventy-five percent of the
provocation testing. fallers were either in Stage 2 or 3 of the Hoehn & Yahr
staging, and, in contrast to our cohort, had lower UP-
3.2. Falls outcomes DRS motor scores (22 for the fallers compared to 28.3
and 30.4 for our study). We only recorded falls for
two 2 week periods before and after the 8 weeks of
Eleven subjects out of the 18 reported falling during training; therefore, only 11 of our subjects experienced
the 2 week period prior to the start of the study. Those a fall during these times (5 in the trained group and 6
in the trained group experienced a significant decrease in the control group). The trained group had half the
in reported falls after the intervention compared to be- number of falls in the follow-up period compared to the
fore. However, the difference between the groups is control group. A longer period of time to observe falls
not statistically significant although the trained group is warranted in a follow-up study.
experienced half the falls during follow-up compared The gait speeds of our subjects were within normal
to the control group (Fig. 1). limits for their age; however, we asked them to walk
188 E.J. Protas et al. / Gait and step training to reduce falls in Parkinson’s disease

at their fastest, but safest speed during the tests. Our voluntary step initiation time [31]. Improved move-
control group had similar speeds during the pre- and ment initiation could also be related to the outcomes
post-tests, but did not show an improvement. Sev- we observed in our study.
eral authors reported increased gait speeds after tread- Falls can be reduced in older adults with gait and
mill training [19,26] or repetitive training of compen- balance training programs, and fall reduction is associ-
satory stepping in individuals with PD [13]. Pohl ated with improvements in gait and balance abilities in
and his group [26] reported increased gait speeds and elder fallers [35]. Gait disturbances and instability are
stride length immediately after a single session of either often linked to falls in individuals with PD. Our data
speed-dependent treadmill training or limited progres- suggests that improvements in fastest walking speed
sive treadmill training compared to conventional gait and dynamic balance may contribute to a reduction in
training or a no-intervention waiting period. Miyai et falls in PD.
al. [19] reported an increase in gait speed from 0.93 m/s This study has a number of limitations. Our sam-
to 1.18 m/s following gait training, using body-weight ple size was relatively small, especially given the vari-
supported treadmill training 3 days/week for 1 month. ability in individuals with PD. We choose to randomly
In another study, patients with PD underwent train- assign our subjects to the intervention as well as a no-
ing consisting of pull perturbations for 14 days and treatment group in order to control for this variabil-
demonstrated an increase of gait speed from 0.64 m/s ity. As a result, we saw no difference between groups
to 0.77 m/s [13]. This increase was accompanied by an in some of our measures before and after intervention
increase in cadence (0.80 steps/s to 0.87 steps/s) and (cadence and stride length) even though the measure
step length (0.80 m to 0.87 m), but there was no control changed for the trained group but not for the control
group comparison. group (stride length) or had a larger improvement for
We located only one study that used a dynamic bal- the trained group (step test). Interestingly, both groups
ance test somewhat similar to the one used in this study increased cadence, while only the trained group had
before and after a physical therapy intervention [37]. an increase in stride length. Increased stride length
Forty individuals with PD underwent a program con- is frequently a desired outcome for physical therapy
sisting of exercise, cued walking, stepping, and motor gait training with this population [21]. The power of
function strategies for 30 days. The program resulted our pre- and post-training measures were high for gait
in improvements in static balance (timed tandem and speed (0.80), cadence (1.0), and the step test (0.99).
The power for our group comparisons were modest for
single limb stance), as well as the number of single
gait speed (0.60), and low for all other measures. De-
limb steps in 15 seconds for both groups of patients
spite this, these results are promising enough to expand
who fell and did not fall. The initial performance of
this small study to a larger trial. Our fall record was
our subjects on the 5-step test (0.40 steps/s for the inter-
a self-report from either the patient or a caregiver, and
vention group and 0.36 steps/s for the control) is sim-
was collected by an investigator who was not blinded
ilar to values we have previously reported for elderly
to group assignment, both of which could lead to some
subjects 0.30 steps/s) [22]. Although the values do
bias in the number of reported falls. One potential out-
not suggest a deficiency in dynamic balance, the speed
come of our training could be a reduced fear of falling;
of performance of this activity was improved for the
however, we did not have a measure for falls efficacy.
trained group.
Our training was task-specific to some of the chal-
Our study and others suggest that gait and balance lenges to balance experienced by these patients, such
can be improved for people with PD using a variety as difficulty stepping or walking backwards or side-
of motor learning strategies. It is difficult to deter- ways. Other approaches may produce different out-
mine the underlying mechanisms for these improve- comes. We were not able to address the intensity, fre-
ments. We used a multidirectional gait and step training quency, and duration of the training intervention in this
strategy for training based on the outcomes of a single study. Therefore, we do not know if the intervention
case study [40]. Morris suggests that visual and audi- we tested is most optimal.
tory cueing as well as attentional strategies can impact
gait and balance in PD [21]. The treadmill could pro-
vide some visual cueing during walking, and auditory 5. Conclusion
cues could have occurred during step training when
the treadmill was turned on and off. Step perturbation Task-specific gait and step training resulted in a re-
training in community-dwelling older adults improved duction in falls and improvement in gait speed and dy-
 E.J. Protas et al. / Gait and step training to reduce falls in Parkinson’s disease 189

namic balance in individuals with postural instability [11] M. Hoehn and M.D. Yahr, Parkinsonisam: onset, progression,
gait difficulty pre-dominant PD and moderate disease and mortality, Neurology 17 (1967), 427–442.
[12] A.J. Jackson, J.W. Porter, K.A. Merrell and B.T. Burt, The
symptoms. This is a promising approach that warrants effects of harness supported treadmill ambulation training on
further research. the gait characteristics of a person with Parkinson’s disease,
Medicine & Science of Sports & Exercise 32 (2000), S236.
[13] M. Jöbges, G. Heuschkel, C. Pretzel, C. Illhardt, C. Renner and
H. Hummelsheim, Repetitive training of compensatory steps:
Acknowledgments A therapeutic approach for postural instability in Parkinson’s
disease, J Neurol Neurosurg Psychiatry 75 (2004), 1682–1687.
This research was made possible by a grant from the [14] W.C. Koller, S. Glatt, B. Vitare-Overfield et al., Falls and
Parkinson’s disease, Clinical Neuropharmacology 12 (1989),
Parkinson’s Disease Research, Education, and Clincal 98–105.
Center, Michael E. Debakey Veterans Affairs Medical [15] G. Krasilovsky and J. Gianutsos, Effect of video feedback
Center, Houston, TX. (Department of Veteran’s Affairs on the performance of a weight shifting controlled tracking
# B2728-R). We gratefully recognize Barry McKay and task in subjects with parkinsonism and neurologically intact
individuals, Experimental Neurology 113 (1991), 192–201.
Teresa Joe for their assistance in subject testing and [16] Y. Laufer, R. Dickstein, Y. Chefez and E. Marcovitz, The effect
Mary Green for project coordination. of treadmill training on the ambulation of stroke survivors in
the early stages of rehabilitation: A randomized study, Journal
of Rehabilitation Research & Development 38 (2001), 69–78.
[17] R.F. Macko, C.A. DeSouza, L.D. Tretter, K.H. Silver, G.V.
References Smith, P.A. Anderson, N. Tomoyasu, P. Gorman and D.R.
Dengel, Treadmill aerobic exercise training reduces the energy
[1] A. Ashburn, E. Stack, R.M. Pickering and C.D. Ward, A expenditure and cardiovascular demands of hemiparetic gait
community-dwelling sample of people with Parkinson’s dis- in chronic stroke patients, Stroke 28 (1997), 326–330.
ease: Characteristics of fallers and non-fallers, Age and Aging [18] A.L. McDonough, M. Batavia, F.C. Chen et al., The valid-
30 (2001), 47–52. ity and reliability of the GAITRite systems’ measurements:
[2] S. Bagley, B. Kelly, N. Tunnicliffe, G.I. Turnbill and J.M. A preliminary evaluation, Archives of Physical Medicine &
Walker, The effect of visual cues on the gait of independently Rehabilitation 82 (2001), 419–425.
mobile Parkinson’s disease patients, Physiotherapy 77 (1991), [19] I. Miyai, Y. Fujimoto, H. Yamamoto et al., Long-term effect
415–420. of body-weight supported treadmill training in Parkinson’s
[3] F.I. Caird, Rehabilitation in Parkinson’s Disease, New York: disease: A randomized, controlled trial, Archives of Physical
Chapman and Hall, 1991. Medicine & Rehabilitation 83 (2002), 1370–1373.
[4] D.B. Calne, B.J. Snow and C. Lee, Criteria for diagnosing [20] I. Miyai, Y. Fujimoto, Y. Ueda, H. Yamamoto, S. Nozaki,
Parkinson’s disease, Annals of Neurology 32 (1992), S125– T. Saito and J. Kang, Treadmill training with body weight
S127. support: Its effect on Parkinson’s disease, Archives of Physical
[5] C.L. Commella, G.T. Stabbins, N. Brown-Toms et al., Physical Medicine & Rehabilitation 81 (2000), 849–852.
therapy and Parkinson’s disease, Neurology 44 (1994), 376– [21] M.E. Morris, Movement disorders in people with Parkinson
378. disease: a model for physical therapy, Physical Therapy 80
[6] I. Cunha, A.C.P. Lim, H. Qureshy, H. Henson, T.N. Monga and (2000), 578–597.
E.J. Protas, A comparison of regular rehabilitation and regular [22] M.A. Murphy, S.L. Olson, E.J. Protas and T. Overby, Screen-
rehabilitation with supported treadmill ambulation training for ing for falls in community-dwelling elderly, Journal of Aging
acute stroke patients, Journal of Rehabilitation Research & & Physical Activity 11 (2003), 66–80.
Development 38 (2001), 245–255. [23] L. Nilsson, J. Carlsson, A. Danielsson, A. Fugl-Meyer, K.
[7] I. Cunha, A.C.P. Lim, H. Qureshy, H. Henson, T.N. Monga Hellstrom, L. Kristensen, B. Sjolund, K.S. Sunnerhagen and
and E.J. Protas, Gait outcomes after acute stroke rehabilita- G. Grimby, Walking training of patients with hemiparesis at
tion with supported treadmill ambulation training: A random- an early stage after stroke: A comparison of walking training
ized, controlled pilot study, Archives of Physical Medicine & on a treadmill with body weight support and walking training
Rehabilitation 83 (2002), 1258–1265. on the ground, Clinical Rehabililitation 15 (2001), 515–527.
[8] R.G. Cutlip, C. Mancinelli, F. Huber and J. Di Pasquale, Eval- [24] M.E. Oehlert, S.D. Hass, M.R. Freeman, M.D. Williams, J.J.
uation of an instrumented walkway for measurement of the Ryan and S.W. Sumerall, The Neurobehavioral cognitive sta-
kinematic parameters of gait, Gait & Posture 12 (2000), 134– tus examination: Accuracy of the “Screen-Metric” approach
138. in a clinical sample, Journal of Clinical Psychology 53 (1997),
[9] S. Fahn and R.I. Elton, Members of the UPDRS Develop- 733–737.
ment Committee. Unified Parkinson’s Disease Rating Scale, [25] T. Overby, D. Roberts-Warrior and E. Lai, Changes in postural
in: Recent Developments in Parkinson’s Disease, (Vol. 2), control after pallidotomy, in: Pallidal Surgery for the Treat-
S. Fahn, C.D. Marsden, D.B. Calne and M. Goldstein, eds, ment of Parkinson’s Disease and Movement Disorders, J.K.
Florham Park, NJ: Macmillan Health Care Information, 1987, Krauss, R.G. Grossman and J. Jankovic, eds, Philadelphia,
pp. 153–164. PA: Lippencott-Raven, 1998, pp. 191–221.
[10] S. Hesse, C. Bertelt, A. Schaffrin et al., Restoration of gait [26] M. Pohl, G. Rockstroh, S. Ruckriem, G. Mrass and J.
in nonambulatory hemiparetic patients by treadmill with par- Mehrholz, Immediate effects of speed-dependent treadmill
tial body-weight support, Archives of Physical Medicine & training on gait parameters in early Parkinson’s disease,
Rehabilitation 75 (1994), 1087–1093.
190 E.J. Protas et al. / Gait and step training to reduce falls in Parkinson’s disease

Archives of Physical Medicine & Rehabilitation 84 (2003), [37] I. Stankovic, The effect of physical therapy on balance of
1760–1766. patients with Parkinson’s disease, International Journal of
[27] E.J. Protas, S.A. Holmes, A. Johnson, H. Qureshy, G. Ro- Rehabilitation Research 27 (2004), 53–57.
driquez, M. Moussavi and A.M. Sherwood. Supported tread- [38] L. Stefaniwsky and D.S. Bilowit, Parkinsonism: Facilitation of
mill ambulation training after spinal cord injury, Archives of motion by sensory stimulation, Archives of Physical Medicine
Physical Medicine & Rehabilitation 82 (2001), 825–831. Rehabilitation 54 (1973), 75–77.
[28] C.L. Richards, F. Malouin, P.J. Bedard and M. Cioni, Changes [39] M. Suteerawattananon and E.J. Protas, Reliability of perfor-
induced by L-Dopa and sensory cues on the gait of parkin- mance outcome measures in individuals with Parkinson’s dis-
sonian patients, in: Posture and Gait: Control Mechanism, ease, Physiotherapy Theory & Practice 16 (2000), 211–218.
(Vol. II), M. Wollacott and F. Horak, eds, Eugene, OR: Uni- [40] M. Suteerawattananon, E. MacNeill and E.J. Protas, Sup-
versity of Oregon Books, 1992, pp. 126–129. ported treadmill training for gait and balance: A case report in
[29] C.L. Richards, F. Malouin, S. Wood-Dauphinee, J.I. Williams, progressive supranuclear palsy, Physical Therapy 82 (2002),
J.P. Bouchard and D. Brunet, Task-specific physical therapy 485–495.
for optimization of gait recovery in acute stroke patients, [41] M. Suteerawattananon, J. Jankovic, S. Morris, B. Etnrye and
Archives of Physical Medicine Rehabilitation 74 (1993), 612– E.J. Protas, Effects of cuing on gait performance of individuals
620. with Parkinson’s disease, Journal of Neurological Science 219
[30] D. Roberts-Warrior, A. Overby, J. Jankovic, S. Olson, E.C. Lai, (2004), 63–69.
J.K. Krauss and R. Grossman, Postural control in Parkinson’s [42] M. Thomas, J. Jankovic, M. Suteerawattananon, S. Wankadia,
disease after unilateral posteroventral pallidotomy, Brain 123 K. Caroline, K.D. Vuong and E. Protas, Clinical gait and bal-
(2000), 2141–2149. ance scale (GABS), Journal Neurological Science 217 (2004),
[31] M.W. Rogers, Disorders of posture, balance, and gait in 89–99.
Parkinson’s disease, Clinical Geriatric Medicine 12 (1996), [43] M. Visintin, H. Barbeau, N.K. Bitensky and N.E. Mayo, Using
825–845. a new approach to retrain gait in stroke patients through body
[32] M.W. Rogers, M.E. Johnson, K.M. Martinez, M.-L. Mille and support and treadmill stimulation, Stroke 29 (1998), 1122–
L.D. Hedman, Step training improves the speed of voluntary 1128.
step initiation in aging, Journal of Gerontology: Biology & [44] J. Waagfjord, P.K. Levanle and C.M.E. Certo, Effects of tread-
Medical Science 58 (2003), M46–M51. mill training on gait in a hemiparetic patient, Physical Therapy
[33] M. Schenkman, T.M. Cutson, M. Kuchibhatla, J. Chandler and 70 (1990), 549–560.
C. Pieper, Reliability of impairment and physical performance [45] M. Weinrich, K. Koch, R. Garcia and R.W. Angel, Axial versus
measures for persons with Parkinson’s disease, Physical Ther- distal motor impairment in Parkinson’s disease, Neurology 38
apy 77 (1997), 19–27. (1988), 540–545.
[34] M. Schenkman, T.M. Cutson, M. Kuchibjhatla, J. Chandler, [46] A. Wernig, A. Nanassy and S. Müller, Maintenance of lo-
C.E. Pieper, L. Ray and K.C. Laub. Exercise to improve spinal comotor abilities following Laufband (treadmill) therapy in
flexibility and function for people with Parkinson’s disease: para- and tetraplegic persons: follow-up studies, Spinal Cord
A randomized controlled trial, Journal American Geriatrics 36 (1998), paper #149/8/97 International Medical Society of
Society 46 (1998), 1207–1216. Paraplegia.
[35] A. Shumway-Cook, W. Gruber, M. Baldwin and S. Liao, The [47] A. Wernig, S. Müller, A. Nanassy and E. Cagol, Laufband
effect of multidimensional exercises on balance, mobility, and therapy based on ‘Rules of Spinal Locomotion’ is effective in
fall risk in community dwelling older adults, Physical Therapy spinal cord injured persons, European Journal of Neuroscience
77 (1997), 46–57. 7 (1995), 823–829.
[36] J.W. Staeheli, F.J. Frassica and F.H. Sim, Prosthetic replace- [48] A. Wernig and S. Müller, Laufband locomotion with body
ment of the femoral neck in patients who have Parkinson’s dis- weight support improved persons with severe spinal cord in-
ease, Journal of Bone and Joint Surgery 70 (1988), 565–568. juries, Paraplegia 30 (1992), 229–238.