Research Report
Sensory-Specific Balance Training in
KP Westlake, PT, PhD, MSc, is
Post-Doctoral Fellow, Rehabilita-
tion Research and Development
Center, VA Palo Alto HCC, 3801
Miranda Ave, Palo Alto, CA 94304
(USA). Address all correspondence
to Dr Westlake at: westlake@rrd.
stanford.edu.
EG Culham, PT, PhD, is Professor
and Director, School of Rehabilita-
tion Therapy, Queens University,
Kingston, Ontario, Canada.
[Westlake KP, Culham EG.
Sensory-specific balance training
in older adults: effect on proprio-
ceptive reintegration and cogni-
tive demands. Phys Ther.
2007;87:1274 1283.]
2007 American Physical Therapy
Association
Post a Rapid Response or
find The Bottom Line:
www.ptjournal.org
1274 f Physical Therapy Volume 87
Older Adults: Effect on Proprioceptive
Reintegration and Cognitive Demands
Kelly P Westlake, Elsie G Culham
Background and Purpose
Age-related changes in the ability to adjust to alterations in sensory information
contribute to impaired postural stability. The purpose of this randomized controlled
trial was to investigate the effect of sensory-specific balance training on propriocep-
tive reintegration.
Subjects
The subjects of this study were 36 older participants who were healthy.
Methods
Participants were randomly assigned to a balance exercise group (n17) or a falls
prevention education group (n19). The primary outcome measure was the center-
of-pressure (COP) velocity change score. This score represented the difference
between COP velocity over 45 seconds of quiet standing and each of six 5-second
intervals following proprioceptive perturbation through vibration with or without a
secondary cognitive task. Clinical outcome measures included the Fullerton Ad-
vanced Balance (FAB) Scale and the Activities-specific Balance Confidence (ABC)
Scale. Assessments were conducted at baseline, postintervention, and at an 8-week
follow-up.
Results
Following the exercise intervention, there was less destabilization within the first 5
seconds following vibration with or without a secondary task than there was at
baseline or in the falls prevention education group. These training effects were not
maintained at the 8-week follow-up. Postintervention improvements also were seen
on the FAB Scale and were maintained at follow-up. No changes in ABC Scale scores
were identified in the balance exercise group, but ABC Scale scores indicated
reduced balance confidence in the falls prevention education group postintervention.
Discussion and Conclusion
The results of this study support short-term enhanced postural responses to propri-
oceptive reintegration following a sensory-specific balance exercise program.
Number 10 October 2007

P
ostural control depends on the
ability to extract peripheral sen-
sory inputs, integrate this infor-
mation within the central nervous
system (CNS), and coordinate and
execute an appropriate motor re-
sponse. Proprioception is an essen-
tial component of this sequence of
events, providing orientation infor-
mation about passive and active
movements and positions of the
joints as well as the force resulting
from muscular contractions. Age-
related changes in the ability to as-
sess the contribution of propriocep-
tive inputs relative to those of other
sensory inputs become evident un-
der conditions in which the propri-
oceptive inputs are distorted or dis-
torted and then suddenly restored.1
Whereas young adults who are
healthy are able to restore balance
quickly by taking advantage of sen-
sory redundancy and centrally re-
weighing available information,2 4
older adults do not as readily inter-
pret misleading cues or recognize
and reintegrate accurate propriocep-
tive information and therefore can
experience postural instability.3,5
These effects are particularly evident
when attentional resources are
divided.3,6 8
Given that the sensory inputs related
to various environmental conditions
are constantly changing,9 the ability
to adjust instantly to a change in sen-
sory information is central to the re-
duction of fall risk in older adults.10
To date, there have been no reports
of training interventions designed to
enhance the ability of older adults to
use proprioceptive information in
balance control. The successful iden-
tification of training effects necessar-
ily involves a randomized controlled
trial with an exercise intervention
designed to induce specific changes
in the recognition and effective use
of sensory information. Findings that
sensory-specific balance exercises,
such as training on unsteady support
surfaces with transitions between
October 2007
sensory environments, result in in-
creased postural stability compared
with the effects of nonspecific activ-
ity interventions, such as running or
strength training, lend support to
this theory.1113
One method used to evaluate the
contribution of proprioceptive in-
puts to postural control and the in-
tegrity of the integrative mechanisms
within the CNS is to measure
changes in postural sway during or
following vibration applied over the
muscle belly or tendon.14 This tech-
nique directly targets the primary
muscle afferents contributing to pro-
prioception and may effectively re-
flect a perturbation of this system.
Muscle vibration evokes a sensation
of movement in a direction that nor-
mally would cause elongation of the
vibrated muscle. Accordingly, vibra-
tion of antagonistic muscle groups
results in immediate disruption of
the proprioceptive system.3 Postural
responses to such a perturbation
then can be assessed in a quiet stand-
ing position by recording center-of-
pressure (COP) outcomes on a force
platform.3,15 This position effectively
reduces confounding variables, such
as muscle activation, torque genera-
tion, and biomechanical changes,
that are present during more dy-
namic tasks. Such variables inevita-
bly would become a source of accu-
rate proprioceptive information that
could override the controlled effect
of proprioceptive perturbation at the
ankle joints. Because the present
study represents one of the earliest
reports on the ability to train sensory
integration immediately following vi-
bration perturbation, the quiet stand-
ing protocol was considered advan-
tageous as a means to isolate changes
in this ability.
Thus, the primary hypothesis of the
present study was that older adults,
having completed a sensory-specific
exercise program, would demon-
Volume 87
Balance Training in Older Adults
strate reduced postural destabiliza-
tion and earlier restabilization imme-
diately following the termination of
proprioceptive perturbation through
vibration in comparison with pre-
exercise outcomes or with the out-
comes in a falls prevention educa-
tion group. These effects were
postulated to improve during a con-
comitant cognitive task. The second-
ary hypothesis was that the en-
hanced postural stability would be
reflected in superior scores on a bal-
ance performance scale and a
balance confidence questionnaire.
Method
Study Participants and
Study Design
Participants were volunteers over 65
years of age who were healthy and
recruited through advertisements
and flyers in the community. Exclu-
sion criteria were pre-existing major
lower-extremity pathology (eg,
chronic ankle instability or severe
osteoarthritis), neurological disor-
ders or balance difficulties (eg, ver-
tigo, poor vision, dizziness, stroke,
or epilepsy) that would prevent
standing for the duration of the test-
ing procedures without the aid of an
assistive device, and health condi-
tions (eg, heart disease, uncontrolled
hypertension, chronic obstructive
pulmonary disease, or osteoporosis)
that would preclude participation in
a balance exercise program.
A brief clinical examination was used
to screen for symptoms of peripheral
neuropathy, which are considered a
risk factor for falls.16 This examina-
tion identified the presence, diminu-
tion, or absence of sensation to light
touch on the dorsal and plantar as-
pects of the foot, the Achilles tendon
reflex, and position sensation of the
big toe. Subjects demonstrating the
absence or diminution of one or
more of these characteristics were
excluded from participation. Physi-
cian approval was required before
the subjects were allowed to partic-
Number 10 Physical Therapy f 1275
Balance Training in Older Adults
Figure 1.
Experimental setup of the vibrators at the Achilles and tibialis anterior tendons. Par-
ticipants stood as steadily as possible with arms alongside the body, heels positioned
according to height, and forefeet splayed to a comfortable stance.
ipate in the exercise program. All
subjects gave written informed con-
sent prior to data collection.
In this single-blind, randomized con-
trolled trial, participants were as-
signed to an exercise group or a falls
prevention education group. Both
groups were assessed at baseline
and within 1 week postintervention.
Follow-up testing was done for the
exercise group only at 8 weeks
postintervention.
Of the 64 older adults who re-
sponded to study advertisements, 44
met study criteria and were random-
ized into the exercise or education
group. Eight participants dropped
out of the study for reasons such as
time commitment issues, lack of
transportation, language barrier, and
disinterest. By the end of the 8-week
interventions, 17 and 19 participants
remained in the exercise and educa-
tion programs, respectively. The
mean numbers of all visits attended
1276 f Physical Therapy Volume 87
by participants in the exercise group
and the education group were 21.5
(89.9%) and 5.4 (66.3%), respec-
tively. The 36 participants who com-
pleted the exercise (n17) and
education (n19) interventions re-
turned for postintervention testing.
By the 8-week follow-up, conducted
only with the exercise group, 15 par-
ticipants returned for testing.
Sensory-specific balance classes
were held 3 times per week, for 1
hour each session, over an 8-week
period. The exercise protocol fol-
lowed the FallProof Program,17
which emphasizes static and dy-
namic balance exercises with transi-
tions between different sensory con-
ditions. Activities were designed to
optimize and force use of the so-
matosensory system. Tasks included
standing or walking on various sup-
port surfaces, such as a rocker board,
foam, or narrow beam, and standing
in a tandem position, a semitandem
position, on one leg, or in a feet-
Number 10
together position. Progressions to
these tasks included simultaneous
alterations of visual and vestibular
inputs. To alter visual cues, partici-
pants were instructed to close their
eyes, to engage vision with a reading
or tracking secondary task, or to per-
form balance tasks with a distracting
background, such as a checked pat-
tern or moving people. To modify
vestibular cues, participants were in-
structed to tilt their head backward
or to quickly move their head side to
side and up and down.
Measurement of Central
Integration and Attentional
Capacity
The mean COP velocity for the total
COP path length was measured on a
force platform* as an estimate of the
frequency of postural corrections.
Of the COP stability parameters,
COP velocity generally is considered
to be most useful in identifying age-
related changes and fall risk.18 20
The mean velocity also demon-
strated the highest sensitivity to the
effects of vibration on posturo-
graphic measurements21 and had the
smallest reproducibility error (intra-
individual standardized coefficient of
variation of 14) over a 1-week
period.20
Data were sampled at 200 Hz and
smoothed with a fourth-order
double-pass Butterworth filter with a
cutoff frequency of 10 Hz. Proprio-
ceptive input was perturbed by use
of 4 vibrators oscillating at 80 Hz, 1
mm in amplitude,2,22 and secured at
both ankles with 3-cm-wide elastic
bands (Fig. 1).
The COP velocity was measured for
each of the 45-second postural con-
ditions shown in Table 1, completed
as one set in a random order, fol-
lowed by 3 minutes of seated rest.
* Biodex Medical Systems Inc, 20 Ramsay Rd,
Shirley, NY 11967.
October 2007
 Balance Training in Older Adults

Three sets were performed, and the Table 1.

average of 3 trials was recorded for Postural Conditions Used to Evaluate Proprioceptive Reintegration With or Without a
each condition. Conditions 3 and 4 Secondary Task
included 5 seconds of stable stand- Testing Description
ing followed by 10 seconds of vibra- Conditiona
tion activation3,21 and then 30 sec- 1 Quiet standing, eyes closed
onds of deactivation. The initial 5
2 Quiet standing, eyes closed, secondary task
seconds was used so that partici-
pants could become acclimated to 3 Quiet standing, eyes closed, vibration
standing on the force platform prior 4 Quiet standing, eyes closed, vibration, secondary task
to the onset of vibration. The third, a
The duration of each condition was 45 seconds.
30-second interval was used because
previous work had indicated that 10
seconds was insufficient for com-
responsiveness of the FAB Scale fol- ence of strength (force-generating
plete recovery of postural stability in
lowing an exercise intervention had capacity) between and within
the absence of vision in either a
not yet been established. groups. Following one practice trial,
young or an older population.3 The
an average of the best 3 of 5 peak
COP velocity was analyzed over
The Activities-specific Balance Confi- torque values normalized to body
5-second intervals from the moment
dence (ABC) Scale was used to assess weight was recorded.
the vibration was turned off, at 15.05
participants level of balance confi-
to 20 seconds (time 1), 20.05 to 25
dence in performing particular Data Analysis
seconds (time 2), 25.05 to 30 sec-
tasks.24 Confidence in performing The effects of the interventions on
onds (time 3), 30.05 to 35 seconds
each task was rated on a scale of 0 the ability of older adults to regain
(time 4), 35.05 to 40 seconds (time
(no confidence) to 10. The ABC postural stability with or without a
5), and 40.05 to 45 seconds (time 6).
Scale showed excellent internal con- secondary task were assessed by use
For the secondary task in conditions
sistency (Cronbach alpha.96), test- of a group time interval visit
2 and 4, participants counted back-
retest reliability (r.92), and validity (2 6 2) analysis of variance
ward by 3 from a random 3-digit
for community-dwelling older (ANOVA) for repeated measures on
number as quickly and accurately as
people.24,25 the last 2 factors. The dependent
possible throughout the 45-second
variable was COP change scores, ob-
trials.7
The Physical Activity Scale for the tained by subtracting COP velocity
Elderly (PASE) was used to deter- averaged over 3 trials for each time
Clinical Measures
mine group equivalences in activity interval of conditions 3 and 4 from
The Fullerton Advanced Balance
levels outside of the treatment inter- the average COP velocity in the three
(FAB) Scale was used to measure
vention. Scores ranged from 0 to 45-second trials in conditions 1 and
functional limitations associated
greater than 400, depending on sub- 2, respectively. The equivalences of
with 10 high-level balance tasks in-
jects reported activity intensities COP velocity across the six 5-second
corporating tests that challenge sen-
and frequencies over 7 days. The intervals and over the entire 45-
sory integration.23 The FAB Scale
PASE showed good test-retest reli- second time interval during condi-
included modified clinical tests for
ability (r.75) and validity for older tion 1 were verified by use of data
sensory integration in balance: walk-
subjects who were healthy.26,27 from 10 randomly selected partici-
ing with head turns, Functional
pants (F0.37; df6,54; P.76).
Reach Test, 360-degree turn, tandem
The concentric isokinetic strength of Changes in strength and clinical mea-
walking, one-leg stance, 2-foot jump,
the hip, knee, and ankle flexor and sures were determined by use of a
step up and over, and reactive pos-
extensor muscles of the dominant group visit (2 2) repeated-
tural control. Scores ranged from 0
leg was assessed by use of an isoki- measures ANOVA for continuous
(unable) to 4, with a maximum score
netic dynamometer (AMTI Multiaxis variables or a Friedman test for cate-
of 40. Psychometric testing of this
Force Platforms, model OR 6-7) set gorical variables.
scale indicated good convergent va-
at a velocity of 60/s. These measure-
lidity compared with the Berg Bal-
ments were taken in consideration of The outcomes for the exercise group
ance Scale (.75) and demon-
the potentially confounding influ- at the 8-week follow-up were com-
strated high test-retest (.96),
pared with the outcomes at baseline
intrarater (.971.00), and inter-
Advanced Medical Technologies Inc, 176 and postintervention by use of a
rater (.94 .97) reliability.23 The Waltham St, Watertown, MA 02172.

October 2007 Volume 87 Number 10 Physical Therapy f 1277

Balance Training in Older Adults

A less destabilization occurring in the

COP Velocity Change (cm/s) 4.0 exercise group (change score
3.5 Exercise Group [XSD], 1.310.91 cm/s) than in
* Education Group the education group (2.001.05
3.0
cm/s) postintervention (Fig. 2A).
2.5
Separate analysis of the exercise
2.0 group revealed a visit time inter-
1.5 action (F8.62; df5,80; P.001),
1.0 indicating training effects on both
0.5 the time to restabilize and the
* amount of destabilization (Fig. 3A).
0.0
-0.5 Change scores began to stabilize dur-
-1.0 ing time interval 4 at baseline and
1 2 3 4 5 6 during earlier time interval 3 postin-
Time Interval tervention. A reduction in the extent
of destabilization was indicated by a
B significant decrease in change scores
4.0 from baseline to postintervention
COP Velocity Change (cm/s)

3.5 Exercise Group during time interval 1 (2.231.18

Education Group versus 1.310.91 cm/s) (P.002),
3.0
* with a trend toward significance dur-
2.5 ing time interval 2 (0.720.76 versus
2.0 0.430.52 cm/s) (P.081). Analysis
1.5 of the education group revealed a
1.0 nonsignificant visit time interac-
tion (F0.42; df5,90; P.70), indi-
0.5
* cating no change in the ability to
0.0 stabilize following vibration.
-0.5
-1.0 The COP change scores for the ex-
1 2 3 4 5 6 ercise group were different postin-
Time Interval tervention (1.310.91 cm/s) and at
Figure 2. the 8-week follow-up (2.111.50
Change in center-of-pressure (COP) outcomes following vibration in exercise and cm/s) during time interval 1
education groups postintervention (XSD). (A) COP sway velocity change score fol- (P.029), suggesting that postinter-
lowing vibration perturbation. (B) COP sway velocity change score following vibration vention improvements were not
perturbation and while performing a secondary task. *Significant difference between maintained. No differences were
groups (P.044).
identified between baseline (2.29
1.19 cm/s) and the 8-week follow-up
(P1.00).
repeated-measures ANOVA or a Results
Friedman test with visit as the Proprioceptive Reintegration Quiet standing with vibration
within-subject factor. Significant in- Quiet standing with vibration and secondary task (condition 4).
teraction effects (P.05) were ana- (condition 3). The first objective The second objective of this study
lyzed with Bonferroni-adjusted post of this study was to examine postural was to examine postural stability fol-
hoc tests. Statistical procedures were recovery following vibratory pertur- lowing vibratory perturbation during
performed with SPSS, version 11.5. bation without a secondary task. A secondary task performance. One
visit time group interaction was outlier in the education group was
identified (F3.82; df5,170; identified as having a mean change
P.019). Group differences were score greater than 3 standard devia-
found only during time interval 1 tions above the group mean and was

SPSS Inc, 233 S Wacker Dr, Chicago, IL (F4.36; df1,35; P.044), with excluded from subsequent analysis.
60606. A visit time group interaction

1278 f Physical Therapy Volume 87 Number 10 October 2007

 Balance Training in Older Adults

was revealed (F3.13; df5,165; A

P.018). 4.0

COP Velocity Change (cm/s)

3.5 * Baseline
Analysis of group differences for
each time interval across baseline 3.0 Postintervention
and postintervention visits indicated 2.5
a difference in the extent of desta- 2.0
bilization during time interval 1

1.5
(F4.90; df1,34; P.034) post- 1.0
intervention, with lower mean
change scores in the exercise group 0.5
*
(1.120.58 cm/s) than in the educa- 0.0
tion group (1.710.85 cm/s) -0.5
(Fig. 2B). Separate analysis of the ex-

-1.0
ercise group revealed a visit time 1 2 3 4 5 6
interaction (F5.76; df5,80; Time Interval
P.001) (Fig. 3B). No improvements
in the time to stabilize were noted.
Further analysis of change scores be- B
4.0
COP Velocity Change (cm/s)

tween baseline and postintervention 3.5 Baseline

as a function of each time interval
revealed a difference during time in-
terval 1 (P.002). This finding con-
firmed that there was less destabili-
zation during the 5 seconds
immediately following vibration as a
result of the exercise intervention.
Analysis of the education group re-
vealed a nonsignificant visit time
interaction (F1.27; df5,85;
P.30).
The inclusion of the 8-week
follow-up change scores for the ex-
ercise group revealed a visit time
interaction (F2.93; df10,140;
P.012), with higher change scores
at the 8-week follow-up (2.05
1.47 cm/s) than postintervention
(1.140.61 cm/s) during time inter-
val 1 (P.023). These results further
support the fact that improvements
in the ability to stabilize after the
exercise intervention were not
maintained.
The response accuracy and speed of
performance of the secondary task
are shown in Table 2. No differences
were identified between groups, nor
was there a group visit interaction.
3.0 * Postintervention
2.5
2.0
1.5
1.0
0.5
*
0.0
-0.5
-1.0
1 2 3 4 5 6
Time Interval
Figure 3.
Change in center-of-pressure (COP) outcomes following vibration at baseline and
postintervention in exercise group (XSD). (A) COP velocity change scores following
vibration perturbation. (B) COP velocity change scores following vibration perturbation
and while performing a secondary task. *Significant difference between baseline and
postintervention (P.006). Significant difference between baseline time intervals
(P.025). Significant difference between postintervention time intervals (P.001).
Clinical Measures postintervention and at follow-up
Means and standard deviations for were maintained. Specific details re-
clinical measures are shown in Table garding changes in the mean and
3. In terms of the FAB Scale, the standard deviation of each FAB Scale
Friedman test revealed differences item are shown in Table 4. Post hoc
between visits in the exercise group analysis revealed differences be-
(216, P.001). Post hoc analysis tween baseline and postintervention
revealed differences between base- for items 6 (one-leg stance), 7 (stand-
line and postintervention total ing on foam), and 9 (walking with
scores (P.001) and between base- head turns) and differences between
line and 8-week follow-up total baseline and the 8-week follow-up
scores (P.001), suggesting that the for item 9 (Bonferroni-adjusted P
improvements in functional balance value of .025).

October 2007 Volume 87 Number 10 Physical Therapy f 1279

Balance Training in Older Adults

Table 2.
Accuracy and Speed of Backward Counting by 3 During Condition 4a

Parameter XSD for the Following Group:

Baseline Postintervention 8-wk
Follow-up
Exercise Education Exercise Education Exercise
Accuracy (no. of correct responses/total no. of 0.940.093 0.960.035 0.950.06 0.970.05 0.940.06
responses)
Speed (no. of responses/s) 0.410.13 0.440.16 0.470.16 0.470.15 0.430.19
a
Condition 4quiet standing, eyes closed, vibration, secondary task.

Regarding the PASE scores, no differ-
ences were identified between
groups or visits at baseline or postint-
ervention; significantly lower scores
were found in the exercise group at
the 8-week follow-up than postinter-
vention (P.003). Interestingly, a
Pearson correlation analysis between
PASE scores and time interval 1 for
conditions 3 and 4 revealed no rela-
tionships (r.16, P.13, and r
.06, P.57).
For the ABC Scale questionnaire, a
group visit interaction was identi-
fied (F4.27; df1,34; P.047),
with a lower balance confidence
score postintervention than at base-
line only for the education group
(F4.56; df1,18; P.047).
In terms of strength at the hip, knee,
and ankle, no main effect of group
(P.66) or visit (P.072) and no
group visit interaction (P.44)
were identified, thereby supporting
group equivalences at baseline and
postintervention.
Discussion
The results of the present study ap-
pear to support the original hypoth-
esis that the ability of older adults
to reintegrate proprioceptive inputs
is augmented following sensory-
specific training, and this effect is
not likely to be attributable to an
increase in lower-extremity strength
or activity level. Although a few stud-
ies reported training effects on pos-
tural stability when proprioceptive
input was reduced,28,29 an improved
ability of older adults to recognize
and use the restoration of accurate
proprioceptive information had not
previously been documented.
Although it may be argued that en-
hanced signals arising at the level of
the proprioceptive receptors may
account for the postural improve-
ments, we recently demonstrated
that this mechanism is not likely to
be the substrate for change.30 With
only 1 of 3 proprioceptive measures
indicating improvements with train-
ing in our previous work, it was dif-
ficult to ascribe a training effect at
the peripheral level. However, with-
out sufficient physiological evidence
from receptor isolation techniques,
such as microneurography,14 the
possibility of an increase in the dis-
charge of these receptors cannot be
discounted.
A more probable explanation for
these results is an increase during
the training intervention in the atten-
tion allocated to proprioceptive cues
(explicit learning), which eventually

Table 3.
Clinical Measures for Exercise and Education Groups at Baseline, Postintervention, and 8-Week Follow-up

Clinical Measurea XSD Score for the Following Group:

Baseline Postintervention 8-wk
Follow-up
Exercise Education Exercise Education Exercise
b,c b
FAB Scale 315 328 354 338 355c
PASE 102.445.0 99.144.0 127.357.6d 101.651.4 79.524.9d
ABC Scale 84.89.3 87.510.7e 85.79.5 79.124.2e 85.09.7
a
ABCActivities-specific Balance Confidence, FABFullerton Advanced Balance, PASEPhysical Activity Scale for the Elderly.
b
Significant difference between baseline and postintervention FAB Scale scores (P.001).
c
Significant difference between baseline and 8-week follow-up FAB Scale scores (P.02).
d
Significant difference between postintervention and 8-week follow-up PASE scores for exercise group (P.003).
e
Significant difference between baseline and postintervention ABC Scale scores for education group (P.047).

1280 f Physical Therapy Volume 87 Number 10 October 2007

 Balance Training in Older Adults

Table 4.
Mean (SD) and Friedman Mean Rank for Scores on the Fullerton Advanced Balance (FAB) Scale for the Exercise Group

FAB Scale Item Baseline Group Postintervention 8-Week Follow- P

Group up Group
XSD Mean XSD Mean XSD Mean
Rank Rank Rank
1. Stand with feet together, eyes closed 40 1.9 40 2.0 40 2.0 .39
2. Forward reach 41 1.9 40 2.1 41 1.9 .26
3. 360 turn 31 1.8 31 1.9 41 2.3 .05
4. Step up and over 40 1.9 40 2.0 40 2.0 .37
5. Tandem walk 31 1.8 41 2.3 31 1.9 .14
a a
6. One-leg stance 31 1.6 31 2.2 31 2.2 .02b
7. Stand on foam, eyes closed 31 1.7a 40 2.2a 40 2.0 .02b
8. 2-foot jump 31 1.8 31 2.1 31 2.1 .35
a,c a c
9. Walk with head turns 21 1.5 31 2.3 31 2.2 .01b
10. Reactive postural control 21 1.6 31 2.2 31 2.2 .12
Total score 315 1.2a,c 354 2.5a 355 2.3c .001b
a
Significant difference between baseline and postintervention scores (P.025).
b
Significant difference between testing sessions (P.05).
c
Significant difference between baseline and 8-week follow-up scores (P.025).

led to a less attentionally demanding
recovery of postural stability (im-
plicit learning).31 Improvements in
postural control in the exercise
group without evidence of reduc-
tions in the accuracy or speed of the
counting task support this theory. As
the accuracy of peripheral input de-
clines with age, attentional resources
become more focused on the control
of posture.1 Thus, the introduction
of a sufficiently challenging second-
ary task or postural condition often
results in reduced task performance
or instability.6,32 Because the specific
instructions provided to participants
in the present study were to main-
tain focus on the secondary task, ev-
idence that stability was increased
suggests an implicit learning effect.
Our previous finding that a velocity
discrimination test was the only pro-
prioceptive outcome to improve
with exercise also suggests en-
hanced central mechanisms.30 This
test required subjects to identify the
faster of 2 presented velocities until
the smallest velocity difference was
identified correctly. Thus, a greater
extent of cognitive resources was
necessary for this test than for other
proprioceptive measures, meaning
that the possibility of improved at-
tention cannot be ruled out. Besides
the possible influence of attention in
recognizing and selecting proprio-
ceptive information, it also has been
suggested that attention is involved
in sensory integration under condi-
tions of sensory conflict.33
One surprising outcome was the de-
crease in COP velocity change scores
(ie, reduced destabilization) during
secondary task performance relative
to the results obtained in the no-
secondary-task condition (Fig. 2). Al-
though the difference was not signif-
icant, it was evident in both the
exercise and the education groups.
In contrast, previous studies6,33,34
demonstrated a destabilizing rather
than a stabilizing effect with the ad-
dition of a secondary task. These
conflicting results may be reconciled
by a recent study demonstrating that
postural stability improved or de-
clined relative to baseline perfor-
mance depending on the cognitive
demands of the secondary task.32
Perhaps the task of counting back-
ward by 3, used in the present study,
did not represent a sufficiently chal-
lenging cognitive task to tax atten-
tional resources effectively. Never-
theless, the finding that the extent of
destabilization was reduced postin-
tervention with or without a second-
ary task suggests that either a shift in
attention or increased attentional ca-
pacity is possible. A follow-up train-
ing study involving a more cogni-
tively demanding secondary task
under conditions of sensory conflict
may bring further clarity to this
discussion.
Several authors3,4,35 have proposed
that the explanation for impaired
postural responses in older adults
lies in age-related changes in central
integration mechanisms. During the
exercise intervention in the present
study, sensory inputs were manipu-
lated by altering the support surface

October 2007 Volume 87 Number 10 Physical Therapy f 1281

Balance Training in Older Adults
or by reducing the sensory redun-
dancy of the visual and vestibular
systems; these manipulations forced
participants to effectively reweigh
remaining inputs within the CNS.17
The direct beneficial consequences
of these tasks were reflected in the
ability of the participants to regain
stability, likely by taking advantage
of the restored proprioceptive infor-
mation and integrating it with vestib-
ular inputs and other sensorimotor
cues. Evidence of similarly enhanced
central integration following sensory
training has been found in studies
demonstrating improved stability
during the manipulation of proprio-
ceptive, vestibular, or visual systems
or all of these by use of the Sensory
Organization Test (SOT).28,36 Al-
though the use of a sway reference
standing surface during the SOT is
considered to be a proprioceptive
perturbation, the muscle spindles
serving this system cannot be tar-
geted as precisely as with vibration.
These studies also were limited be-
cause of the use of a cross-sectional
design including seasoned tai chi
practitioners36 and the use of the
SOT for both training and testing
procedures,28 which may have re-
sulted in a learning effect.37
The functional significance of the re-
sults of the present research was ev-
ident because of improvements in
the FAB Scale scores in the exercise
group. These results, demonstrating
responsiveness to training, further
support the validity of FAB Scale
scores. Interestingly, the items dem-
onstrating improvements across vis-
its were items 6 (one-leg stance), 7
(standing on foam), and 9 (walking
with head turns), each of which
comprises an element of sensory in-
tegration. However, even though
subjects showed improvements in
and maintenance of FAB Scale scores
at the 8-week follow-up, the im-
provements did not translate to the
maintenance of enhanced postural
stability following vibratory pertur-
1282 f Physical Therapy Volume 87
bation. This lack of an effect suggests
that there are context-dependent dif-
ferences following a targeted train-
ing intervention and thereby sup-
ports the need for ongoing sensory
training. In turn, compensatory sen-
sory mechanisms may be selected
more efficiently under conditions of
sensory deprivation38 or restoration.
The decrease in balance confidence
in the education group postinterven-
tion may be explained by discussions
centered on effective means of
reducing fall risk. An increased
awareness of these topics may have
underscored the apprehension expe-
rienced during functional balance
tasks until changes could be imple-
mented. Two recent studies examin-
ing the effectiveness of falls preven-
tion education reported similar
findings, with almost half of the par-
ticipants demonstrating increased
fear of falling39 and a 28% increase in
1 or more falls40 at follow-up.
Seasonal variations in PASE scores
may account for the reduction in
scores at the 8-week follow-up in the
exercise group.26 Follow-up testing
occurred during both the winter and
the summer months, when either
snow and ice or high heat and hu-
midity may have forced participants
indoors. Arguably, the reduction in
activity level may explain the lack of
retention in the ability to reintegrate
proprioceptive information effec-
tively. However, without a signifi-
cant correlation between PASE
scores and COP velocity in the first 5
seconds following vibration, this the-
ory is not substantiated.
Conclusion
The results of the present study sug-
gested that sensory-specific exercise
had a training effect on propriocep-
tive reintegration. However, 2 limita-
tions should be mentioned. The first
limitation is that 8-week follow-up
scores were not obtained for the ed-
ucation group. Therefore, although
Number 10
improvements in the FAB Scale
scores were maintained in the exer-
cise group, it remains unclear
whether the control group also ex-
perienced changes over the 8-week
time period. The second limitation is
that the participants were older
adults who were healthy rather than
older adults with balance impair-
ments, who may have benefited to a
greater extent. Future research may
include a group of older adults with
declining balance to assess the effect
of training on central sensory reinte-
gration. Such an investigation also
may include a kinematic and kinetic
analysis of the effect of vibration on
dynamic stability tasks. A combina-
tion of findings from these studies
and those from the present study
may lead to more efficient balance
exercise interventions and, ulti-
mately, to a reduction in fall risk in
older adults.
Both authors provided concept/idea/re-
search design, writing, and data analysis. Dr
Westlake provided data collection, project
management, fund procurement, and sub-
jects. Dr Culham provided facilities/equip-
ment and consultation (including review of
manuscript before submission).
This study was approved by Queens Univer-
sity Health Science and Affiliated Teaching
Hospitals Research Ethics Board.
Data from this study were presented at the
International Congress of the World Confed-
eration for Physical Therapy; June 2 6, 2007;
Vancouver, British Columbia, Canada.
Dr Westlake was supported by a Canadian
Institutes of Health Research Fellowship.
This article was received September 7, 2006,
and was accepted May 8, 2007.
DOI: 10.2522/ptj.20060263
References
1 Kristinsdottir EK, Fransson PA, Magnusson
M. Changes in postural control in healthy
elderly subjects are related to vibration
sensation, vision and vestibular asymme-
try. Acta Otolaryngol. 2001;121:700 706.
2 Hay L, Bard C, Fleury M, Teasdale N. Avail-
ability of visual and proprioceptive affer-
ent messages and postural control in el-
derly adults. Exp Brain Res. 1996;
108:129 139.
October 2007

3 Teasdale N, Simoneau M. Attentional de-
mands for postural control: the effects of
aging and sensory reintegration. Gait Pos-
ture. 2001;14:203210.
4 Quoniam C, Hay L, Roll JP, Harlay F. Age
effects on reflex and postural responses to
propriomuscular inputs generated by ten-
don vibration. J Gerontol A Biol Sci Med
Sci. 1995;50:B155B165.
5 Camicioli R, Panzer VP, Kaye J. Balance in
the healthy elderly: posturography and
clinical assessment. Arch Neurol. 1997;54:
976 981.
6 Shumway-Cook A, Woollacott M. Atten-
tional demands and postural control: the
effect of sensory context. J Gerontol A
Biol Sci Med Sci. 2000;55:M10 M16.
7 Woollacott M, Shumway-Cook A. Atten-
tion and the control of posture and gait: a
review of an emerging area of research.
Gait Posture. 2002;16:114.
8 Rankin JK, Woollacott MH, Shumway-
Cook A, Brown LA. Cognitive influence on
postural stability: a neuromuscular analy-
sis in young and older adults. J Gerontol A
Biol Sci Med Sci. 2000;55:M112M119.
9 Peterka RJ. Sensorimotor integration in hu-
man postural control. J Neurophysiol.
2002;88:10971118.
10 Peterka RJ, Loughlin PJ. Dynamic regula-
tion of sensorimotor integration in human
postural control. J Neurophysiol. 2004;91:
410 423.
11 Xu D, Hong Y, Li J, Chan K. Effect of tai chi
exercise on proprioception of ankle and
knee joints in old people. Br J Sports Med.
2004;38:50 54.
12 Beard DJ, Dodd CA, Trundle HR, Simpson
AH. Proprioception enhancement for an-
terior cruciate ligament deficiency: a pro-
spective randomised trial of two physio-
therapy regimes. J Bone Joint Surg Br.
1994;76:654 659.
13 Gauchard GC, Gangloff P, Jeandel C,
Perrin PP. Influence of regular propriocep-
tive and bioenergetic physical activities on
balance control in elderly women.
J Gerontol A Biol Sci Med Sci. 2003;58:
M846 M850.
14 Burke D, Hagbarth KE, Lofstedt L, Wallin
BG. The responses of human muscle spin-
dle endings to vibration of non-
contracting muscles. J Physiol. 1976;261:
673 693.
15 Vuillerme N, Teasdale N, Nougier V. The
effect of expertise in gymnastics on pro-
prioceptive sensory integration in human
subjects. Neurosci Lett. 2001;311:7376.
October 2007
16 Richardson JK, Ashton-Miller JA. Periph-
eral neuropathy: an often-overlooked
cause of falls in the elderly. Postgrad Med.
1996;99:161172.
17 Rose DJ. FallProof! A Comprehensive Bal-
ance and Mobility Training Program.
Windsor, Ontario, Canada: Human Kinet-
ics; 2003.
18 Maki BE, Holliday PJ, Fernie GR. Aging and
postural control: a comparison of
spontaneous- and induced-sway balance
tests. J Am Geriatr Soc. 1990;38:19.
19 Lafond D, Corriveau H, Hebert R, Prince F.
Intrasession reliability of center of pres-
sure measures of postural steadiness in
healthy elderly people. Arch Phys Med Re-
habil. 2004;85:896 901.
20 Raymakers JA, Samson MM, Verhaar HJ.
The assessment of body sway and the
choice of the stability parameter(s). Gait
Posture. 2005;21:48 58.
21 Uimonen S, Sorri M, Laitakari K, Jamsa T. A
comparison of three vibrators in static pos-
turography: the effect of vibration ampli-
tude on body sway. Med Eng Phys. 1996;
18:405 409.
22 Ribot-Ciscar E, Rossi-Durand C, Roll JP.
Muscle spindle activity following muscle
tendon vibration in man. Neurosci Lett.
1998;258:147150.
23 Rose DJ, Lucchese N, Wiersma LD. Devel-
opment of a multidimensional balance
scale for use with functionally indepen-
dent older adults. Arch Phys Med Rehabil.
2006;87:1478 1485.
24 Powell LE, Myers AM. The Activities-
specific Balance Confidence (ABC) Scale.
J Gerontol A Biol Sci Med Sci. 1995;50A:
M28 M34.
25 Myers AM, Powell LE, Maki BE, et al. Psy-
chological indicators of balance confi-
dence: relationship to actual and per-
ceived abilities. J Gerontol A Biol Sci Med
Sci. 1996;51:M37M43.
26 Washburn RA, Smith KW, Jette AM, Janney
CA. The physical activity scale for the el-
derly (PASE): development and evaluation.
J Clin Epidemiol. 1993;46:153162.
27 Washburn RA, McAuley E, Katula J, et al.
The physical activity scale for the elderly
(PASE): evidence for validity. J Clin Epide-
miol. 1999;52:643 651.
28 Hu MH, Woollacott MH. Multisensory
training of standing balance in older
adults, I: postural stability and one-leg
stance balance. J Gerontol. 1994;49:M52
M61.
Volume 87
Balance Training in Older Adults
29 Tsang WW, Hui-Chan CW. Effect of 4- and
8-week intensive Tai Chi training on bal-
ance control in the elderly. Med Sci Sports
Exerc. 2004;36:648 657.
30 Westlake KP, Wu Y, Culham EG. Sensory-
specific balance training in older adults:
effect on position, movement, and veloc-
ity sense at the ankle. Phys Ther. 2007;87:
560 568.
31 Orrell AJ, Eves FF, Masters RS. Implicit mo-
tor learning of a balancing task. Gait Pos-
ture. 2006;23:9 16.
32 Huxhold O, Li SC, Schmiedek F, Linden-
berger U. Dual-tasking postural control:
aging and the effects of cognitive demand
in conjunction with focus of attention.
Brain Res Bull. 2006;69:294 305.
33 Redfern MS, Jennings JR, Martin C, Furman
JM. Attention influences sensory integra-
tion for postural control in older adults.
Gait Posture. 2001;14:211216.
34 Condron JE, Hill KD. Reliability and valid-
ity of a dual-task force platform assessment
of balance performance: effect of age, bal-
ance impairment, and cognitive task. J Am
Geriatr Soc. 2002;50:157162.
35 Stelmach GE, Teasdale N, Di Fabio RP,
Phillips J. Age related decline in postural
control mechanisms. Int J Aging Hum
Dev. 1989;29:205223.
36 Tsang WW, Wong VS, Fu SN, Hui-Chan
CW. Tai Chi improves standing balance
control under reduced or conflicting sen-
sory conditions. Arch Phys Med Rehabil.
2004;85:129 137.
37 Grindstaff TL, Christiano KE, Broos AM,
et al. Assessment of a practice effect in
serial sensory organization testing scores
of healthy adults. Percept Mot Skills. 2006;
102:379 386.
38 Bayouk JF, Boucher JP, Leroux A. Balance
training following stroke: effects of task-
oriented exercises with and without al-
tered sensory input. Int J Rehabil Res.
2006;29:5159.
39 Rucker D, Rowe BH, Johnson JA, et al. Ed-
ucational intervention to reduce falls and
fear of falling in patients after fragility frac-
ture: results of a controlled pilot study.
Prev Med. 2006;42:316 319.
40 Chang JT, Morton SC, Rubenstein LZ, et al.
Interventions for the prevention of falls in
older adults: systematic review and meta-
analysis of randomised clinical trials. BMJ.
2004;328:680.
Number 10 Physical Therapy f 1283