Full text of DOI:10.1093/gerona/56.8.M489

Journal of Gerontology: MEDICAL SCIENCES
Copyright 2001 by The Gerontological Society of America
2
001, Vol. 56A, No. 8, M489–M496
The Interacting Effects of Cognitive Demand and
Recovery of Postural Stability in Balance-Impaired
Elderly Persons
1
1
2
Sandra G. Brauer, Marjorie Woollacott, and Anne Shumway-Cook
1Department of Exercise and Movement Science, University of Oregon, Eugene.
2Department of Rehabilitation Medicine, University of Washington, Seattle.
Background. Although postural recovery is attentionally demanding in healthy elderly persons, an inability to re-
cover balance due to competition for attentional resources between the postural system and a second task could contrib-
ute to falls in older adults with poor balance. This study examined the attentional demands of balance recovery from a
mild postural disturbance in balance-impaired elderly persons. A second purpose of this research was to determine the
effect of performing a cognitive task on the recovery of balance in balance-impaired elderly persons.
Methods. Fifteen healthy older adults and 13 older adults with clinical balance impairment were exposed to balance
disturbances by means of sudden movement of a platform on which they stood. A dual-task paradigm where postural re-
covery served as the primary task and verbal reaction time to auditory tones served as the secondary task was used to as-
sess attentional demand. To determine the effect of the cognitive task on postural recovery, kinetic, kinematic, and
neuromuscular measures of a feet-in-place response were investigated.
Results. Balance recovery using a feet-in-place response was attentionally demanding in both groups of older adults
and was more demanding in balance-impaired than in healthy elderly persons. With the concurrent performance of a
cognitive task, balance-impaired elderly persons took longer to stabilize their center of pressure and regain balance than
in a single task, while healthy elderly persons showed no change between conditions. In addition, only balance-impaired
elderly individuals had a greater center-of-pressure resultant velocity during recovery in a dual-task compared with a
single-task situation.
Conclusions. The ability to recover balance using a feet-in-place response was more attentionally demanding in balance-
impaired than in healthy elderly persons. The recovery of balance was also slower and less efficient in balance-impaired
elderly persons when simultaneously performing a cognitive task, whereas the ability of healthy elderly individuals to
recover was not influenced by concurrent task demands. This suggests that dual-task performance may contribute to
postural instability and falls in balance-impaired elderly individuals.
EALTHY older adults show a marked reduction in the
ability to perform a postural and a cognitive task si-
mands of postural recovery in balance-impaired versus
healthy older adults. We expected that in a dual-task situa-
tion, balance-impaired older adults would demonstrate a
greater reduction in cognitive task performance compared
with healthy older adults, suggesting that attentional de-
mands associated with recovery of stability are greatest in
balance-impaired older adults.
Dual-task paradigms have also been used to examine the
effects of a secondary task on the efficiency of postural con-
trol. In healthy older adults, the ability to recover stability
following an external perturbation is affected by the simul-
taneous performance of a secondary task. Rankin and col-
leagues (10) compared neuromuscular response characteris-
tics associated with postural recovery using a moving
platform paradigm in single- versus dual-task conditions.
They found that in the dual-task situation, healthy older
adults exhibited a reduced magnitude in the gastrocnemius
response, the primary muscle used to regain stability. This
could make recovery of balance without taking a step less
effective and could explain the greater prevalence of step
responses by older adults in a dual-task situation (9). Al-
though recovery of postural stability in dual-task conditions
H
multaneously compared with young adults. This has been
demonstrated as a reduction in the performance of the cog-
nitive task, specifically, an increase in reaction time (1–3).
In a number of studies using this paradigm, a decrement in
the performance of the postural task has also been found.
Greater postural instability when performing dual tasks has
been reported in stance (4–6) and during obstacle avoidance
in gait (7).
Dual-task paradigms have been used to examine the rela-
tive attentional demands associated with different types of
postural tasks. Changes in the secondary task are used to in-
fer task-dependent changes in attentional demands. Using
this approach, researchers have shown that attentional de-
mands vary as a function of task complexity (2–8), age (1),
and balance abilities (8). Brown and colleagues (9) found
that recovery of postural stability was attentionally demand-
ing and that attentional demands increased with age. The
attentional demands associated with postural recovery in
balance-impaired older adults are not known. Thus, one
purpose of this study was to compare the attentional de-
M489

M490
BRAUER ET AL.
has been studied in healthy older adults, its relevance to ex-
plaining falls in balance-impaired elderly persons is un-
known.
cal therapist conducted the clinical examination, which in-
cluded the Berg Balance Scale (15), the Timed Up and Go
Test (17), and the Dual-Task Timed Up and Go Test (13).
The Trail-Making Test (parts A and B) were performed to
indicate overall attentional ability (18), and the Mini-Mental
State Examination (MMSE) (19) was performed to describe
general cognitive ability. Questionnaires determining medi-
cal history and exercise frequency were completed.
It has been suggested that an inability to produce an ap-
propriate postural response due to the competition for atten-
tional resources between the postural system and the cogni-
tive task contributes to falls in older adults with poor
balance (11). Older adults with clinical balance impairments
either stop (12) or take a longer time to complete a gait task
when performed with a secondary task (13). The inability to
walk and talk simultaneously was associated with a greater
risk of falls in the succeeding 6-month period (14). Further
evidence that competition for attention may play a role in
instability and falls in balance-impaired elderly persons was
reported by Shumway-Cook and colleagues (11). They
found that the addition of a simple cognitive task could in-
crease center-of-pressure (COP) motion during quiet stance
in older adults with a history of falls. A subsequent study
examining the effects of sensory context on attentional de-
mands of postural control found that as sensory conditions
became more difficult, balance-impaired older adults who
had been able to maintain stability in a single-task context,
lost balance in a dual-task context (8).
However, many falls in older adults occur as a result of
slips or trips, suggesting the need for studies related to re-
covery of postural stability in balance-impaired older adults.
Thus, a second purpose of this study was to determine the
effects of performing a simultaneous cognitive task on the
ability of balance-impaired older adults to recover postural
stability following a platform translation. We hypothesized
that, due to competition for attentional resources, balance-
impaired older adults would demonstrate a decrement in
postural recovery.
In the experimental session, subjects were asked to main-
tain stability in response to external perturbations in the
backward direction under single- versus dual-task condi-
tions. In the dual-task condition, subjects responded vocally
to an auditory choice reaction-time stimulus. Subjects com-
pleted 15 trials of the balance task only, 3 trials of the cogni-
tive task only, and 15 trials of the two tasks together, in ran-
dom order. Five trials were performed with the platform
moving forward to minimize preparation for the movement.
The 30 perturbations (nonlinear ramp-to-parabola wave-
forms) consisted of 3 each at the velocities of 10, 20, 30, 40,
and 50 cm per second. Only the 10-cm-per-second perturba-
2
tions (peak acceleration 0.13 m/s ) were analyzed in this
study, as balance-impaired older adults began taking steps
to recover at the faster perturbation velocities. Subjects
were instructed to try to keep their balance by keeping their
feet in place and avoid taking a step, but also were told to
respond as quickly and accurately as possible in the cogni-
tive task.
The cognitive task consisted of a verbal response (“high”
or “low”) to two frequencies of tone (500 Hz and 2 kHz)
generated by a mixer (Institute of Neuroscience at the Uni-
versity of Oregon) presented for 250 milliseconds in series
of 10 through a unilateral headphone transducer. The re-
sponse was recorded into a microphone attached to the
headset. Immediately after each response, the investigator
manually triggered the next tone generation. In the dual-task
condition, the platform movement was synchronized with
the production of a tone, such that it occurred at the same
time (ꢀ40 ms to account for variability in movement me-
chanics) as the presentation of a particular tone in the se-
quence. The tone in the sequence that triggered platform
motion was randomized between trials to eliminate any pre-
diction of plate movement.
METHODOLOGY
Subjects
Twenty-seven community-dwelling adults aged over 65
years (15 healthy, 72.1 ꢀ 7 years; 12 balance-impaired,
7
9.2 ꢀ 7 years) volunteered for the study. Exclusion criteria
included major sensory impairments and any neurological
or musculoskeletal diagnosis that could account for postural
instability. Informed consent was obtained from all sub-
jects.
Instrumentation
Subjects were classified as balance-impaired if they
scored ꢁ50 out of 56 on the Berg Balance Scale (15) and
reported a history of postural imbalance. Conversely, sub-
jects were categorized as healthy if they scored ꢂ51 out of
In the perturbation conditions, subjects stood with one
foot on each of two electronically synchronized force plat-
forms (Institute of Neuroscience at the University of Ore-
gon). Foot position was standardized (feet parallel, 10 cm
apart). The subjects wore a harness attached to an overhead
trolley, and an assistant remained by the side of the subject
to prevent a fall. A four-camera 120-Hz Peak Performance
motion analysis system (Peak Performance Technologies,
Inc., Englewood, CO) was used to construct a four-segment,
two-dimensional model of body motion (foot, shank, thigh,
and head/arms/trunk). Markers were placed on the fifth
metatarsal head, lateral malleolus, lateral tibio-fibular joint
line, greater trochanter, and superior aspect of the acromio-
clavicular joint. This was used in conjunction with known
anthropometric data to calculate center of mass (COM)
(20).
5
6 on the Berg Balance Scale and did not report a history of
imbalance. The Berg Balance Scale is a 14-item scale (each
item scored 0–5) that assesses the ability to perform func-
tional balance tasks. It has demonstrated high inter- and in-
trarater reliability (r ꢃ .98 and .99, respectively) and inter-
nal consistency (Cronbach’s ꢄ ꢃ .96) in older adults. The
maximum score is 56, with a score of ꢂ51 associated with a
low risk of falls (16).
Protocol
All subjects underwent a neurological examination, a
clinical examination, and an experimental session. A physi-

COGNITIVE DEMAND AND POSTURAL RECOVERY
M491
Table 1. Between-Group Differences in Characteristics
Table 2. Between-Group Differences in Clinical Balance Measures
Balance-impaired
Healthy
Older Adults
n ꢃ 14
Balance-impaired
Older Adults
Healthy Older Adults
Older Adults
Characteristic
n ꢃ 13
Performance Measures
n ꢃ 15
n ꢃ 13
Berg Balance Scale†*
Range
54.9 ꢀ 1.5
51–56
46.7 ꢀ 3.0
42–50
Age, y* (SD)
Range
Gender, % women
No. of falls in past year, %
0
72.1 (6.7)
64–86
67
79.4 (6.5)
68–92
77
Timed Up and Go (TUG), s*
Motor Dual-Task TUG, s*
Cognitive Dual-Task TUG, s*
Functional reach distance, cm*
Lateral reach distance, cm*
Single-leg stance time, s*
6.8 ꢀ 1.2
7.8 ꢀ 1.4
8.1 ꢀ 1.6
31.1 ꢀ 5.3
15.4 ꢀ 4.3
20.0 ꢀ 11.6
11.2 ꢀ 2.5
13.1 ꢀ 3.8
16.2 ꢀ 3.4
20.7 ꢀ 4.6
10.2 ꢀ 3.5
2.7 ꢀ 3.0
79
11
0
46
54
69
46
ꢂ1
Report imbalance outdoors, %**
Use aid outdoors, %**
No. of comorbidities, %**
0
Note: Values are mean ꢀ SD.
0
2
–1
–3
28
36
36
0
38
62
*
Significant difference (p ꢇ .05) between groups by one-way analysis of
variance.
†Maximum score of 56.
ꢂ4
No. times exercise/wk, %**
0
3
–2
–4
21
50
77
15
across groups, tasks, and trials. Several measures indicative
of stability following platform perturbation were calculated
with MATLAB. These included the time for the COP and
COM to return to a stable velocity (within 3 SDs of the pre-
motion velocity for 25 milliseconds); the peak x, y, and re-
sultant velocity of the COP over 2.5 seconds; and the maxi-
mum range of COP motion over 2.5 seconds.
A repeated-measures ANOVA was also performed to de-
termine whether the addition of a secondary task resulted in
a significant change in the EMG recovery response. Group,
task, and Group ꢅ Task interactions were investigated.
Univariate analyses were performed if a significant multi-
variate result was found. The onset time of the EMG signal
(ꢈ3 SDs from baseline for 50 ms) was determined by com-
puter algorithm (22) and checked visually. To determine the
magnitude of muscle activity, the root mean squared error
was calculated from the onset of activity for 100 milliseconds
and normalized to a 200-millisecond baseline. A ratio of
shank muscle activity (gastrocnemius: tibialis anterior activ-
ity) was calculated to provide an index of distal coactivation.
ꢂ5
29
8
MMSE* (SD)
29.5 (0.7);
range, 28–30
15.8 (5.4)
29.6 (10.7)
28.5 (1.6);
range, 25–30
25.7 (8.8)
72.6 (60.9)
Trail Making A Test, s* (SD)
Trail Making B Test, s* (SD)
Note: MMSE ꢃ Mini-Mental State Examination.
Significant difference (p ꢇ .05) between groups by one-way analysis of
variance.
*
**Significant difference (p ꢇ .05) between groups by Mann-Whitney U test.
Surface bipolar electrodes (1 mm ꢅ 10 mm; DE-02, Del-
sys, Inc, Boston, MA) were placed bilaterally on the agonist
muscle, gastrocnemius, and the antagonist muscle, tibialis
anterior. All data were collected at 900 Hz by AMLABII
(
AMLAB International, Australia). Electromyographic (EMG)
data were band-pass filtered from 20 to 500 Hz and further
analysis was performed in MATLAB (The MathWorks,
Inc., Natick, MA).
Data Analysis
To determine whether there was a difference in cognitive
task performance between the single and dual tasks, and
whether this varied with balance impairment, a repeated-
measures analysis of variance (ANOVA) was performed on
the first two reaction times following plate motion. Main ef-
fects of the group (healthy vs balance-impaired), task (sin-
gle vs dual), reaction times (one vs two) and trials (1, 2, and
RESULTS
Demographic information and results from the clinical
examination are reported in Tables 1 and 2. Balance-
impaired older adults were significantly older, performed
less well on clinical balance tests, reported imbalance and
used an aid outdoors more frequently, exercised less fre-
quently, reported more comorbidities, demonstrated poorer
attention ability (high scores on the Trail-Making Test), and
had lower overall cognitive functioning (MMSE) than the
healthy older adults.
3
) were investigated, along with the associated interactions.
Difference scores (dual-task ꢆ single-task reaction time)
were calculated and an ANOVA was performed to deter-
mine whether there was a difference in the attentional de-
mand of recovery between healthy and balance-impaired
older adults. The cognitive task response accuracy (proba-
bility of making an incorrect response) was determined
across groups using a generalized estimating equation, or
GEE, due to the discrete nature of the data and the repeated-
measures design (21). The significance level was set at p ꢇ
Because there was a significant difference in age between
groups of older adults, a repeated-measures analysis of co-
variance with age as a covariate was also performed. No dif-
ferences in significance between the analyses were found;
thus, age did not influence the between-group results re-
ported for healthy versus balanced-impaired subjects. The
data presented in this study include only feet-in-place re-
sponses.
.
05 for all analyses.
To determine whether the addition of a secondary task re-
sulted in a significant change in the postural recovery, a re-
peated-measures ANOVA was performed. Individual analy-
ses determined differences in COP or COM measures
The Attentional Demands of Postural Recovery
The first question in this study asked if attentional de-
mands associated with recovery of stability were greater in

M492
BRAUER ET AL.
Figure 2. Verbal reaction time for the first two time periods fol-
lowing plate onset in healthy and balance-impaired older adults in
single- and dual-task conditions.
When the first two postperturbation reaction times were
investigated individually, a difference between the healthy
and balance-impaired older adults was found. For both re-
sponses of the balance-impaired subjects, the dual-task re-
action time was longer than with the single task (Figure 2).
In contrast, the initial response of the healthy older adults
was not affected by task, whereas the second response was
longer in the dual-task situation. This suggests that in
healthy older adults, the initial response to instability may
be less attentionally demanding than the later portions of the
postural response. In contrast, in balance-impaired older
adults, the whole balance response appears to be attention-
ally demanding.
The inaccuracy measures revealed a significant task
F(1,488) ꢃ 34.31, p ꢇ .001] and group effect [F(1,488) ꢃ
[
Figure 1. Difference in reaction times between dual and single
tasks (dual ꢆ single) of the first reaction time following plate motion.
A, mean data; B, individual data.
1
2.16, p ꢇ .001], but no Group ꢅ Task interaction. All sub-
jects had significantly more mistakes with an added balance
task than when performing the cognitive task alone. The
probability of making an incorrect response increased sig-
nificantly from single to dual tasks, and from healthy to bal-
ance-impaired older adults.
balance-impaired older adults compared with healthy older
adults. To determine this, differences in cognitive task per-
formance between single (prior to platform perturbation)
and dual conditions (immediately after platform perturba-
tion) were investigated across groups. There was a task ef-
fect [F(1,26) ꢃ 6.84, p ꢃ .011], a group effect [F(1,26) ꢃ
Influence of a Cognitive Task on Postural Recovery
The second question of this study addressed the effect of
a secondary cognitive task on the ability to recover balance
in older adults with and without balance impairments. To
address this question, characteristics of the muscle re-
sponse, COP, and COM were analyzed across tasks for each
group.
7
.84, p ꢃ .017], but no Group ꢅ Task interaction in reac-
tion time. Reaction time was longer in the dual task than the
single task for all subjects, suggesting that balance recovery
is attentionally demanding for older adults.
The difference between the dual- and single-task reaction
time was compared between groups to determine if the atten-
tional demand was greater for balance-impaired older adults.
Balance-impaired older adults demonstrated a greater differ-
ence score [F(1,26) ꢃ 6.56, p ꢃ .017], and therefore a longer
reaction time in a dual task, for the first postperturbation re-
action time. The dual-task reaction time was at least 50 mil-
liseconds longer than the single task in 75% of balance-im-
paired older adults, but only in 26% of healthy older adults.
This suggests that attentional demands were greater with
balance impairment (Figure 1).
COP parameters.—Investigation of the COP variables
(time to stabilization, peak velocity, and range of COP mo-
tion) revealed a main effect of group [F(1,26) ꢃ 6.34, p ꢇ
.019] and task [F(1,26) ꢃ 4.82, p ꢃ .043], but no Group ꢅ
Task interaction [F(1,26) ꢃ 2.13, p ꢃ .158]. The main ef-
fects were then investigated for each variable.
As shown in Figure 3, in balance-impaired older adults,
the time to stabilization (COP return to preperturbation ve-
locity levels) in the dual-task condition was significantly
longer (p ꢃ .008) in comparison with the single-task condi-

COGNITIVE DEMAND AND POSTURAL RECOVERY
M493
a single healthy and a balance-impaired subject is shown in
Figure 4.
The peak COP velocity reached during recovery was then
investigated in order to infer the efficiency of the neuromus-
cular response in slowing the forward momentum of the
body induced by the perturbation. Balance-impaired older
adults had a greater COP resultant velocity during recovery
in a dual-task situation, compared with a single-task situa-
tion (Figure 5; p ꢃ .022). This suggests that, compared with
healthy older adults, balance-impaired older adults were less
efficient in recovering stability in the dual-task condition.
The peak range of the COP excursion during recovery
was then investigated as a measure of efficiency of recov-
ery. No task differences were found for the balance-
impaired older adults. In contrast, in the anterio-posterior
direction, the healthy older adults demonstrated a smaller
COP excursion in the dual-task compared with the single-
task condition (p ꢃ .016). This suggests that in the dual-task
condition, healthy older adults may constrain COP excur-
sion more than in the single-task condition. In the medio-
lateral direction, no task effect was found.
Figure 3. Mean (ꢀSEM) values of the time taken for the center-of-
pressure (COP) and center-of-mass (COM) velocity to return to a
preperturbation level in healthy older adults and balance-impaired
older adults in single and dual tasks. ST ꢃ single task; DT ꢃ dual
task. *p ꢇ .05.
COM excursion.—COM measures are often considered
a better global measure of postural stability than COP. Al-
though COP reflects the sum of muscle forces acting against
the surface, the effect of these forces is reflected in COM
movements. Investigation of the COM variables revealed a
group effect [F(1,26) ꢃ 10.61, p ꢃ .003], but no task effect
tion. In contrast, there was no task difference in the time to
stabilization in the healthy older adults. The balance-
impaired took longer than the healthy older adults to stabi-
lize in both tasks. An example of the time to stabilization of
(
p ꢃ .187). There was a trend for the balance-impaired
older adults to take a longer time to stabilize their COM in a
dual-task situation in comparison with the single task (Fig-
ure 3).
Muscle responses.—The repeated-measures ANOVA of
the EMG onset and magnitude variables for the gastrocne-
mius muscle revealed a significant group effect [F(1,26) ꢃ
7
.11, p ꢃ .013]; thus, interactions and univariate effects
were investigated. Because there were no differences in
Figure 5. Peak resultant center-of-pressure (COP) velocity
(ꢀSEM) from plate onset for 2.5 s in single and dual tasks for healthy
and balance-impaired older adults.
Figure 4. Time for the center of pressure to stabilize in a healthy
and a balance-impaired older adult in a dual task.

M494
BRAUER ET AL.
EMG onset or magnitude between the right and left sides,
data were combined.
older adults had a greater COP resultant velocity during re-
covery in a dual-task situation, compared with a single-task
situation, suggesting that compared with healthy older
adults, they were less efficient in recovering stability in the
dual-task condition.
There was also a trend for the balance-impaired older
adults to take a longer time for their COM to stabilize to a
preperturbation level in a dual-task compared with a single-
task condition, with COM changes being close to those of
the COP. These smaller changes in COM between the two
conditions may be due to the fact that the COP must actu-
ally move out beyond the COM in recapturing balance after
a perturbation. Thus, these movements tend to be slightly
larger than COM changes when recovering balance.
Contrary to our hypothesis that range of the COP excur-
sion during recovery would be higher during the secondary
task for balance-impaired older adults, we found no signifi-
cant change in excursion in the balance-impaired group. This
supports previous work by Stelmach and colleagues (6), who
also found that the time to restabilize the COP was influ-
enced more by the addition of a cognitive task than the total
range of motion when investigating age-related changes in
balance recovery.
Interestingly, in the dual-task condition, COP excursion
used to recover from the postural threat was smaller than in
single-task conditions for healthy elderly individuals. This
unexpected result suggests that dual-task conditions may re-
sult in tighter constraints on postural control in healthy pop-
ulations, as a strategy to ensure stability. For example, an
earlier study (9) showed that, in dual-task conditions,
healthy elderly persons stepped when the COM was closer
to the center of the base of support than in single-task condi-
tions, suggesting they used a more conservative balance
strategy (stepping when COM was well within the base of
support) with a second task. This is consistent with research
by Andersson and colleagues (23), who reported that under
quiet-stance conditions, patients with peripheral vestibular
disorders and poor balance reduce sway amplitude when
performing a secondary mental task. Thus, certain popula-
tions may use a strategy to constrain their motion to better
cope with the additional demand.
This increased constraint on the COP displacement in
healthy older adults for the dual-task condition may be cor-
related with their significantly greater co-contraction of
muscles at the ankle joint in this condition. This strategy of
co-contracting agonist and antagonist muscles at a joint in-
creases joint stiffness and has been observed in both less
skilled performers and in older adults when constraining
movement during postural tasks (24).
Performance of a secondary task did not affect the onset
latency or magnitude of the primary neuromuscular re-
sponse responsible for recovery. This is unlike the results of
Rankin and colleagues (10), who reported a reduced pos-
tural muscle response magnitude in older adults when per-
forming a dual task. This difference could be due to meth-
odological issues, in that the postural or the cognitive task
may not have been as attentionally demanding in the current
study as in previous studies. Rankin and colleagues (10)
used a math task that could be more demanding than the au-
ditory reaction time task used in this study. In addition, the
Investigation of gastrocnemius muscle onset times re-
vealed a significant group effect [F(1,26) ꢃ 8.32, p ꢃ .008],
but no significant task, or Group ꢅ Task interaction. Gas-
trocnemius onset was slower in the balance-impaired com-
pared with the healthy older adults; however, there was no
difference in onset times between tasks. Thus, performance
of a secondary task did not affect the onset latency of the
neuromuscular response.
The magnitude of postural muscle activity from the onset
of activity for 100 milliseconds revealed a significant group
effect [F(1,26) ꢃ 5.40 p ꢃ .029], but no task effect or
Group ꢅ Task interaction. Although balance-impaired older
adults demonstrated lower response magnitudes compared
with the healthy older adults, there was no effect of task on
muscle response amplitude.
Coactivation of agonist and antagonist muscles increases
stiffness at a joint and is a strategy, albeit an inefficient one,
used to increase stability. There was a significant difference
in coactivation levels (gastrocnemius/tibialis anterior) be-
tween groups in the single (p ꢃ .043) and dual (p ꢃ .026)
tasks. Although balance-impaired older adults demonstrated
greater co-contraction than healthy older adults, their coac-
tivation was equivalent in both conditions. In contrast, in
healthy older adults, there was a significant task effect on
coactivation, with a greater coactivation level in the dual
task than in the single task.
DISCUSSION
The Attentional Demand of the Feet-in-Place Response
One purpose of this study was to compare the attentional
demands of postural recovery in balance-impaired versus
healthy older adults. We expected that in a dual-task situa-
tion, balance-impaired older adults would demonstrate a
greater reduction in cognitive task performance compared
with healthy older adults, suggesting that attentional de-
mands associated with recovery of stability are greatest in
balance-impaired older adults. Supporting our hypothesis,
recovery of stability was found to be attentionally demand-
ing in both groups of older adults, but more attentionally de-
manding in balance-impaired older adults.
A second purpose of this study was to determine the ef-
fects of performing a simultaneous cognitive task on the
ability of balance-impaired older adults to recover postural
stability following a platform translation. We hypothesized
that due to competition for attentional resources, balance-
impaired older adults would demonstrate a decrement in
postural recovery in the dual-task conditions.
The aspects of postural control most influenced by the
performance of a second task were the time for the COM
and COP to stabilize. The time for balance-impaired older
adults to stabilize COP in the dual task was significantly
longer in comparison with the single task, while healthy
older adults showed no change between tasks. The peak
COP velocity reached during recovery was then investi-
gated in order to infer the efficiency of the neuromuscular
response in slowing the forward momentum of the body in-
duced by the perturbation. Again, only balance-impaired

COGNITIVE DEMAND AND POSTURAL RECOVERY
M495
perturbations investigated in this study were of a lower ve-
locity than in Rankin’s study, to accommodate the poorer
balance abilities of the balance-impaired elderly subjects.
This study affirms and extends results from previous re-
search demonstrating the deleterious effects of performing a
secondary cognitive task on postural control in balance-
impaired older adults (8,11–14). This reduction in balance
ability could result from several factors. First, it could be
due to a reduction in total attentional capacity, which lim-
ited the amount of attention the subjects could direct to the
task of postural recovery. General capacity theory suggests
there is a finite amount of processing space available in the
brain to perform tasks (25). If there was a reduction in over-
all capacity in older adults with balance impairments that
did not permit sufficient attention to be directed to both
tasks, then reduced abilities in both tasks could be expected.
We found that the balance-impaired older group showed
significantly lower scores on the Trail-Making Tests, sug-
gesting that they had reduced attentional abilities.
Alternatively, the postural response used by balance-
impaired older adults may have a greater attentional de-
mand than the response made by the nonimpaired adults.
Older adults with a deterioration in one or more systems re-
quired to maintain balance may need to allocate a greater
proportion of attention to postural control to achieve the
same level of stability as a nonimpaired adult. Although the
balance-impaired older adults in this study demonstrated a
clinical decrement in balance ability, we can only surmise
that these changes required a greater allocation of attention
to balance. Similarly, we have evidence that the balance-
impaired older adults demonstrated changes in their pos-
tural response (longer time to recover) that could require the
allocation of additional attention to recover balance. Thus,
the inability of balance-impaired older adults to maintain
high levels of performance on both tasks when performed
simultaneously could be due to both a reduction in atten-
tional capacity and a change in the allocation of attention to
the tasks. To determine whether attentional misallocation is
a contributor to the inability to perform dual tasks, studies
where attention is specifically directed to each task are re-
quired.
to their reduced cognitive ability, in addition to their bal-
ance impairment.
Clinical Applications
It is known that many factors contributing to balance con-
trol show deterioration in older adults and place them at risk
of falling. We have found evidence of decrements in the
postural stability of balance-impaired older adults when
performing tasks requiring both cognitive processing and
control of balance. Clinical applications include the devel-
opment of intervention strategies for balance-impaired older
adults in which postural tasks are first practiced alone, then
simultaneously, with a secondary cognitive task in order to
improve the ability of older adults to balance under these at-
tentionally more challenging conditions.
Acknowledgments
This study was supported by National Institutes of Health Grant AG-
05317 to M.H. Woollacott and A. Shumway-Cook. We gratefully acknowl-
edge the contribution of Denise Gravelle for assistance in programming and
data collection. Dave Brumbley of the University of Oregon Institute of
Neuroscience technical support group is also acknowledged for the design
of the moveable forceplate system. In addition, statistical advice from
Robin High was greatly appreciated.
Address correspondence to Dr. Sandra Brauer, Department of Physio-
therapy, University of Queensland, St. Lucia, QLD, 4072, Australia.
E-mail: s.brauer@shrs.uq.edu.au
References
1. Lajoie Y, Teasdale N, Bard C, Fleury M. Attentional demands for
static and dynamic equilibrium. Exp Brain Res. 1993;97:139–144.
2
. Teasdale N, Bard C, LaRue J, Fleury M. On the cognitive penetrability
of postural control. Exp Aging Res. 1993;19:1–13.
3
. Eichhorn J, Orner J, Rickard K, Craik R. Ageing effects on dual task
methodology using walking and verbal reaction time. Issues Aging.
1
998;21:8–12.
4
. Maki BE, Whitelaw RS. Influence of experience, expectation and
arousal on posture control strategy and performance. In: Woollacott
M, Horak F, eds. Posture & Gait: Control Mechanisms. Portland, OR:
University of Oregon Books;1992:123–126.
5. Maylor EA, Wing AM. Age differences in postural stability are increased
by cognitive demands. J Gerontol Psych Sci. 1996;51B:P143–P154.
6
. Stelmach GE, Zalaznik HN, Lowe D. The influence of ageing and at-
tentional demands on recovery from postural instability. Aging. 1990;
2
:155–161.
Finally, changes in balance ability in quiet stance have
been associated with increased anxiety (4). Performing dual
tasks may be a more stressful situation, and this could lead
to a reduction in balance ability. Older adults with an admit-
ted balance problem may have found the experimental situ-
ation more stressful than did the nonimpaired adults. How-
ever, the perturbations studied here were very small, and all
subjects were able to maintain balance with no foot motion
and minimal upper-limb response.
7. Chen HC, Schultz AB, Ashton-Miller JA, Giordani B, Alexander NB,
Guire KE. Stepping over obstacles: dividing attention impairs perfor-
mance of old more than young. J Gerontol Med Sci. 1996;51A:M116–
M122.
8
. Shumway-Cook A, Woollacott M. Atentional demands and postural
control: the effect of sensory context. J Gerontol Med Sci. 2000;55A:
M10–M16.
9
. Brown LA, Shumway-Cook A, Woollacott MH. Attentional demands
and postural recovery: the effects of aging. J Gerontol Med Sci. 1999;
5
4A:M165–M171.
1
0. Rankin J, Woollacott MH, Shumway-Cook A, Brown L. Cognitive in-
fluence on postural stability: a neuromuscular analysis in young and
older adults. J Gerontol Med Sci. 2000;55A:M112–M119.
1. Shumway-Cook A, Woollacott M, Kerns KA, Baldwin M. The effects
of two types of cognitive tasks on postural stability in older adults with
and without a history of falls. J Gerontol Med Sci. 1997;52A:M232–
M240.
1
Limitations
Balance-impaired older adults were less healthy, reported
more comorbidities, and exercised less frequently. Thus,
differences we found may be due to differences in health
and exercise status between the groups. In addition, they
demonstrated poorer attention ability and lower overall cog-
nitive functioning. The greater demand postural recovery
posed for the balance-impaired older adults could be related
12. Lundin-Olsson L, Nyberg L, Gustafson Y. Attention, frailty and falls:
the effect of a manual task on basic mobility. J Am Geriatr Soc. 1998;
4
6:758–761.
1
3. Shumway-Cook A, Brauer S, Woollacott M. Predicting the probability
for falls in community-dwelling older adults using the timed up and go
test. Phys Ther. 2000;80:869–903.

M496
BRAUER ET AL.
1
1
4. Lundin-Olsson L, Nyberg L, Gustafson Y. “Stops walking when talk-
22. Hodges PW, Bui BH. A comparison of computer based methods for
the determination of onset of muscle contraction using electromyogra-
phy. Electroenceph Clin Neurophysiol. 1996;101:511–519.
23. Andersson G, Yardley L, Luxon L. A dual-task study of interference
between mental activity and control of balance. Am J Otol. 1998;19:
632–637.
24. Manchester D, Woollacott M, Zederbauer-Hylton N, Marin O. Visual,
vestibular and somatosensory contributions to balance control in the
older adult. J Gerontol Med Sci. 1989;44:M118–M127.
ing” as a predictor of falls in elderly people. Lancet. 1997;349:617.
5. Berg K, Wood-Dauphinee S, Williams JI, Gayton D. Measuring bal-
ance in the elderly: preliminary development of an instrument.
Physiother Can. 1989;41:304–311.
6. Shumway-Cook A, Baldwin M, Polissar NL, Gruber W. Predicting the
probability for falls in community-dwelling older adults. Phys Ther.
1
1
997;77:812–819.
1
7. Podsiadlo D, Richardson S. The timed Up and Go: a test of basic functional
mobility for frail elderly persons. J Am Geriatr Soc. 1991;39:142–148.
8. Reitan R. Validity of the Trail Making Test as an indicator of organic
brain damage. Percept Motor Skills. 1958;8:271–276.
25. Kahneman D. Attention and effort. In: Prentice-Hall Series in Experi-
mental Psychology. Englewood Cliffs, NJ. Prentice-Hall; 1973.
1
1
9. Folstein MF, Folstein SE, McHugh PR. “Mini-Mental State”: a practi-
cal method for grading the cognitive state of patients for the clinician.
J Psychiatr Res. 1975;12:189–198.
2
0. Winter DA. Biomechanics and Motor Control of Human Movement.
2
nd ed. New York: Wiley & Sons; 1990.
Received May 24, 2000
2
1. Liang K, Zeger S. Longitudinal data analysis using generalized linear
Accepted June 1, 2000
models. Biometrika. 1986;73:13–22.
Decision Editor: John E. Morley, MB, BCh