Journal of Sport Rehabilitation, 2011, 20, 442-456

2011 Human Kinetics, Inc.

Reliability of Force-Platform
Measures of Postural Sway
and Expertise-Related Differences
Zohreh Meshkati, Mehdi Namazizadeh,
Mahyar Salavati, and Masood Mazaheri
Context: Although reliability is a population-specific property, few studies have
investigated the measurement error associated with force-platform parameters in
athletic populations. Objective: To investigate the skill-related differences between
athletes and nonathletes in reliability of center-of-pressure (COP) summary measures under eyes-open (EO) and eyes-closed (EC) conditions. Design: Testretest
reliability study. Setting: COP was recorded during double-leg quiet standing
on a Kistler force platform before and after a fatiguing treadmill exercise, with
EO and EC. Participants: 31 male participants including 15 athletes practiced in
karate and 16 nonathletes. Main Outcome Measures: Standard deviation (SD) of
amplitude, phase-plane portrait, SD of velocity, mean total velocity, and area were
calculated from 30-s COP data. Intraclass correlation coefficient (ICC), standard
error of measurement, and coefficient of variation (CV) were used as estimates
of reliability and precision. Results: Higher ICCs were found for COP measures
in the athlete (compared with the nonathlete) group, postfatigued (compared with
prefatigued) condition, and EC (compared with EO) tests. CVs smaller than 15%
were obtained for most of the COP measures. SD of velocity in the anteroposterior direction showed the highest reliability in most conditions. Conclusions:
Tests with EC and to a lesser extent tests performed in the athlete group and in
the postfatigued condition showed better reliability.
Keywords: center of pressure, athlete, fatigue, vision

Sport participation has been shown to enhance humans postural capability.1,2

Athletes demonstrate less postural sway during quiet standing than nonathletes,
which demonstrates a healthier system of postural control.1,2 To quantify postural
sway as an index of balance performance, different parameters have been derived
from center-of-pressure (COP) data recorded from force platforms.3 Decision
making about the selection of an appropriate COP parameter relies on criteria such
Meshkati is with the Science and Research Branch, Islamic Azad University, Tehran, Iran. Namazizadeh is with the Khorasgan Branch, Islamic Azad University, Isfahan, Iran. Salavati is with the Dept of
Physical Therapy, University of Social Welfare and Rehabilitation, Tehran, Iran. Mazaheri is with the
Musculoskeletal Research Center, Isfahan University of Medical Sciences, Isfahan, Iran.
442

Force-Platform Measures of Postural Sway 443

as reliability. To evaluate changes of balance over time or discriminate standing

balance between participants, it is necessary, although not sufficient, to use reliable
COP parameters.4
Despite the frequent use of COP parameters in the literature related to sport,1,2
few studies5,6 have specifically addressed the measurement error associated with
these parameters in athletes. This may be because of burdens of extra time, expense,
and human resources imposed on subjects and researchers in reliability studies.7
Ageberg et al5 assessed intersession reliability of balance variables derived from a
force platform in single-leg stance in recreationally active individuals. In another
study, Harringe et al6 investigated the testretest reliability of COP measures in
a group of young female gymnasts with different levels of back and leg pain. To
date, most of the studies on reliability of COP parameters have used young and
elderly individuals as their participants (Table 1).4,818 The generalizability of the
results of nonathletes to the athletic population may be questionable because of
differences in postural control according to level of expertise.1,2
Apart from sport-participation level, fatigue is a factor that can potentially
influence neuromuscular functioning19,20 including balance control.2123 Postural
sway during double-leg stance24 or single-leg stance5,25 has been adversely affected
by strenuous treadmill24 or cycling5 exercise inducing generalized fatigue or by
lower extremity exercise25 inducing localized fatigue. This may lead to increased
risk of injury as demonstrated by the occurrence of injuries at the end of a sport
activity, when an athlete is fatigued.19 Although the ability to sustain muscle force
under fatigued conditions has a substantial effect on sport performance, most studies
have investigated the measurement properties of postural-control variables under
nonfatigued conditions. To our knowledge, no studies evaluating the reliability of
postural measures under fatigued conditions have been reported.
The aim of the current study was to estimate the testretest reliability of COP
measures in young athletes and nonathletes, tested before and after fatiguing exercise, under eyes-open (EO) and eyes-closed (EC) conditions.

Methods
Design
A convenient sample of 31 individuals with different levels of sports expertise was
assessed at 2 sessions with 48 to 72 hours between to determine the testretest
reliability of force-platform measures of postural sway in different conditions of
fatigue and visual feedback.

Participants
In the athlete group, 15 male karateists (age 21.47 3.31 y, weight 64.20 10.15
kg, height 1.74 0.07 m) participated in the study. They all competed at least at
regional level (4 regional level, 1 national level, 10 international level) and had at
least a black belt. Their average duration of karate participation was 9.67 5.02
years. In the nonathlete group, 16 male subjects (age 21.25 1.95 y, weight 66.44
10.89 kg, height 1.75 0.07 m) agreed to participate in the study. They engaged
in physical activity less than once a week, with irregular intervals between sessions.

444

Healthy young (10)

COP length, area

Faller (11) and nonfaller

(26) older adults

Healthy older adults (63)

Bauer et al18

ICC

ICC, LOA

Generalizability coefficient,
%SEM
ICC, LOA, SEM

Generalizability
coefficient
ICC, SEM

ICC, TEM, CV

For 1 30-s trial: ICC = .22 (area), .29 (rangeAP), .44

(rangeML), .39 (RMSaAP), .35 (RMSaML), .73 (velocityAP), .87
(velocityML)
For rigid surface-eyes open: ICC = .43 (rangeAP), .71
(rangeML), .12 (velocityAP), .29 (velocityML), .49 (area)
For 3 30s trials with eyes open: G = .50 (SDAP), .49 (SDML),
.83 (velocity), .49 (area)
For between-days reliability in young adults: .71 (RMS), .77
(velocityAP), .79 (velocityML), .72 (area)
For eyes open: G = .34 (positionAP), .43 (positionML), .40
(SDAP), .45 (SDML), .55 (rangeAP), .48 (rangeML), .44 (velocityAP), .46 (velocityML), .40 (area)
For 3 trials of nonnormalized measures: ICC = .88 (range), .84
(mean velocity), .80 (Maxv), .94 (area)
For between-days reliability of single task trials with eyes
open in nonfallers: ICC = .43 (MaxaAP), .75 (MaxaML), .31
(RMSAP), .73 (RMSML), .84 (mean velocity), .62 (area)
For narrow-stance trials with eyes open: ICC = .89 (total
length), .91 (lengthAP), .84 (lengthML), .88 (area)

According to different conditions and different measures, CV

range = .13.23

For between-days reliability: ICC = .58

Results
For 2-leg stance: R = .11 (positionAP), .30 (positionML)
R = .81 (SDML), .86 (SDAP), .84 (velocity)

COP, center of pressure; AP, anteroposterior; ML, mediolateral; SD, standard deviation; RMSa, root mean square of amplitude; Maxa, maximal amplitude; Maxv, maximal velocity; ICC, intraclass correlation coefficient; TEM, technical error of measurement; CV, coefficient of variation; SEM, standard error of measurement; LOA, limit of agreement.

Range, mean velocity,

Maxv, area
Maxa, RMSa, mean
velocity, area

Pinsault and
Vuillerme16
Swanenburg
et al17

Santos et al4

Lin et al15

Doyle et al14

Healthy young adults (30) Range, peak velocity,

area
Healthy young adults (15) SD COP, mean velocity,
area
Healthy young (16) and
RMS COP, mean velocolder (16) adults
ity, area
Healthy young adults (12) Mean position, SD COP,
range, mean velocity, area

Doyle et al13

CV

Statistics
Linear regression
Reliability coefficient
ICC

Area, range, RMSa, mean ICC

velocity

Healthy older adults (7)

Older adults with differMean velocity

ent levels of disequilibrium (36)
Healthy young adults (29) Mean velocity, Maxv,
sway index

Population (N)
Measures
Healthy young adults (28) Mean position
Healthy young adults (25) SD COP, mean velocity

Lafond et al12

Rogind et al11

Study
Goldie et al8
Le Clair and
Riach9
Benvenuti et
al10

Table 1 Summary of Selected Articles Related to Reliability of COP Measures

Force-Platform Measures of Postural Sway 445

Potential participants were excluded if they reported a history of uncorrected vision

impairment, vestibular problems (eg, vertigo), cardiovascular and respiratory disorders, diabetes, systemic inflammatory diseases, lower limb injury within last year,
or alcohol or drug abuse. Participants were required to perform no vigorous exercise
in the 48 hours before testing. The study was approved by the ethics committee of
the Science and Research Branch, Islamic Azad University. Informed consent was
obtained from all participants.

Procedures
A Kistler 9281C force platform (400 600 mm) based on piezoelectrical measurement of ground-reaction force in the anteroposterior (AP), mediolateral (ML), and
vertical planes was used to collect COP data. Data were sampled at 100 Hz,26,27
stored on a Pentium-based PC, and then exported to Excel to calculate COP
parameters. COP signals were filtered with a sixth-order Butterworth zero-phase
low-pass filter at 10 Hz.26,27
Subjects stood barefoot with their feet together and arms hanging at their
sides. Postural sway was assessed under 2 different conditions of visual feedback:
EO and EC. During the EO condition, participants were asked to focus their gaze
on a stationary target at approximately eye level located 2 m in front of them, and
during EC, they wore a blindfold to eliminate visual feedback. Participants performed 3 trials of quiet standing with EO and 3 trials with EC. Each trial lasted 35
seconds, followed by a rest period of approximately 25 seconds. The order of trials
was randomized. After that, subjects were asked to engage in a fatiguing protocol.
Immediately after the end of the protocol, each subject repeated 6 trials with EO and
EC, identical to those in the nonfatigued condition. All participants were invited to
come for a second visit 48 to 72 hours later, when the same protocol with the same
condition order was repeated. Participants were evaluated by the same rater, at the
same time of day, and in the same environment on the second occasion.
To take the adaptation period into account, the first 5 seconds of the 35- second
COP time series were not included in data analysis. Trial lengths of 15 to 30 seconds with 1 to 3 trials of data collection have been commonly used in research.14
Parameters calculated from 30-second COP data were standard deviation (SD)
of amplitude in the AP and ML directions; phase-plane portrait in the APML
(combined), AP, and ML directions; SD of velocity in the AP and ML directions;
mean total velocity; and area (95% confidence ellipse).28 Phase-plane portrait, a
less often used parameter, provides insight into static and dynamic aspects of postural control by including both amplitude and velocity in the analysis.26,28,29 The
reason to choose these parameters among numerous measures was because of their
common use in research, which allows the comparison of results across different
studies, and because each parameter represents a unique dimension of position (eg,
SD of amplitude), velocity (eg, mean total velocity), and both position and velocity
(ie, phase-plane portrait) of COP. In addition, some parameters such as minimum,
maximum, and peak-to-peak amplitude are not recommended for use because they
use 1 or 2 data points among all the data points recorded in a trial, which can cause
great variance between subjects and trials.3

446 Meshkati et al

Fatiguing Protocol
Generalized fatigue was induced by having the participants perform exercise on
a treadmill. Subjects wore a heart-rate monitor during the entire duration of the
exercise. Subjective effort level was assessed during exercise with Borgs scale
for rating of perceived exertion.30 Borgs scale correlates with several physiological variables including heart rate and blood lactate concentration.31 A significant
increase in blood lactate concentration has been observed after reaching a workload
of 50% to 60% of maximal oxygen consumption or 60% of maximal heart rate
(HRmax).24 According to our pilot study in a group of 10 individuals, heart rate
remained well above the anaerobic threshold (ie, 60% HRmax) during the 6-minute
period of postural assessment after the fatiguing protocol, when it reached 85%
HRmax at the end of fatiguing exercise. HRmax was predicted from the following
equation:5 HRmax (beats/min) = 220 age (y).
A fatigue protocol developed by Nardone et al24 was used in the current study.
Subjects were asked to walk on the treadmill for 2 minutes at 3 km/h at 0% grade
as a warm-up, followed by increasing grade to 7% for 3 minutes while the speed
remained constant. Thereafter, the grade was increased to 14% and the speed was
increased to 4 km/h. Afterward, the grade remained constant and the speed was
increased with steps of 0.5 km/h every 5 minutes. The protocol continued until the
subjects reached a heart rate of 85% HRmax or perceived the exercise as very hard
on the Borg scale. The exercise was stopped at reaching 85% HRmax in all participants. At the end of protocol, all individuals perceived the exercise as very hard.

Statistical Analyses
For age, weight, and height, comparisons between the 2 groups were made with
independent t tests. An alpha of .05 was considered significant for all statistical
analyses.
The mean of 3 trials of the COP parameters in each condition was used for
statistical analysis to determine reliability measures.
Reliability was assessed using the 2-way random model of intraclass correlation coefficient (ICC2,3) described by Shrout and Fleiss.32,33 For each ICC, the 95%
confidence interval was calculated to take the sampling distribution into account.
Munros classification for reliability coefficients was used to describe the degree
of reliability: .00 to .25little if any correlation, .26 to .49low correlation, .50
to .69moderate correlation, .70 to .89high correlation, and .90 to 1.00very
high correlation.34 A key problem with the ICC is its sensitivity to between-subjects
variability. In the presence of high between-subjects variability, a large ICC can be
obtained even when measurement precision is low.35 Therefore, use of the standard
error of measurement (SEM) and the coefficient of variation (CV), as 2 statistics
for measuring precision, in conjunction with the ICC has been recommended in
literature. The mean-square-error term derived from the analysis-of-variance table
provides an estimate of random error.36 The square root of this mean-square-error
term is a standard deviation that represents the within-subject variation from test to
retest session. The SEM was calculated as the square root of the mean-square-error
term. The SEM is useful for computing the minimal detectable change (MDC), or
change that could be considered really different between 2 measurements. MDC

Force-Platform Measures of Postural Sway 447

was defined as the 95% confidence interval of the SEM of the COP measure (
1.96 SEM).36 In addition, CV was determined for comparison of precision between
stability indexes (SD/mean 100).35

Results
There was no significant difference for age (P = .83), weight (P = .56), or height
(P = .58) between the 2 groups. Table 2 shows the mean and SD of COP measures
by expertise, fatigue, and vision levels, and Tables 3AC show ICCs and their 95%
confidence intervals, SEMs, MDCs, and CVs.
Higher ICC values were found for 77% (28/36) of COP measures in the athlete
group than in the nonathlete group. ICC values .70 were obtained for 58% (21/36)
and 44% (16/36) of COP measures in the athlete and nonathlete group, respectively.
The SD of AP velocity was the most reliable parameter in both athletes (ICC range:
.73.89) and nonathletes (ICC range: .62.79).
Compared with those performed in the prefatigued condition, 63% (34/54) of
COP measures showed better reliability for tests performed in the fatigued condition.
ICC values were .70 for 55% (30/54) of COP measures in the fatigued condition
and for 48% (26/54) in the prefatigued condition. The SD of AP velocity was the
most reliable parameter in both prefatigued (ICC range: .62.89) and fatigued (ICC
range: .74.88) conditions.
Tests with EC were more reliable than tests with EO. All but 3 of the COP
measures showed higher ICC values for EC tests than EO tests. In the EC and
EO tests, respectively, 85% (46/54) and 19% (10/54) of COP measures reached a
high level of reliability (ICC .70). The SD of AP velocity was the most reliable
parameter in EO tests (ICC range: .73.80), and mean total velocity was most reliable in EC tests (ICC range: .72.89).
In general, 86% (31/36), 56% (30/54), and 88% (48/54) of COP measures
had smaller CVs in the nonathlete (relative to athlete) group, prefatigue (relative
to postfatigue) condition, and EC (relative to EO) tests, respectively. All COP
measures yielded CV values smaller than 15% in all visual and fatigued conditions
for athletes, nonathletes, and total participants, with the exception of area and SD
of AP amplitude.

Discussion
Postural capabilities in each type of sport are specific to the conditions in which
the athlete has been trained.1 Karate is no exception. Incorporation of straight-line
powerful blocking, kicking, and punching techniques discriminates karate from
other sports in terms of balance abilities.37 To compare balance between individuals
who practice karate and those who train in other sports, the first step is to identify
reliable and valid postural measures. The current study was designed to investigate
the reliability of force-platform measures intended to evaluate balance in athletes
and nonathletes under 2 different testing conditions (EO and EC).
ICC values varied according to expertise level (athlete vs nonathlete) and
test condition (EO vs EC and prefatigue vs postfatigue). The largest difference in
reliability was between EO and EC conditions. This finding is in agreement with

448

Mean total
velocity
PP (APML)
Area (95%
ellipse)

PP

SDVEL

SDAMP

ML

PP

SDVEL

SDAMP

AP

Eyes Open

0.48
(0.12)
1.07
(0.20)
1.18
(0.19)
1.18
(0.20)
1.57
(0.25)
4.17
(1.41)

0.51
(0.18)
0.87
(0.16)
1.02
(1.20)

Test

0.44
(0.10)
1.16
(0.26)
1.25
(0.26)
1.23
(0.25)
1.62
(0.31)
3.94
(1.59)

0.53
(0.15)
0.86
(0.17)
1.02
(0.21)

Retest

Prefatigue

0.51
(0.12)
1.20
(0.16)
1.31
(0.17)
1.31
(1.18)
1.74
(0.25)
4.81
(2.25)

0.56
(0.21)
0.96
(0.16)
1.13
(0.22)

Test

0.49
(0.12)
1.17
(0.20)
1.28
(0.20)
1.30
(0.20)
1.71
(0.26)
4.50
(1.64)

0.55
(0.13)
0.94
(0.24)
1.13
(0.20)

Retest

Postfatigue

Nonathletes (n = 16)

0.45
(0.14)
1.34
(0.31)
1.13
(0.23)
1.42
(0.30)
1.83
(0.36)
3.96
(2.51)

0.48
(0.19)
1.00
(0.23)
1.13
(0.23)

Test

0.49
(0.15)
1.19
(0.26)
1.05
(0.25)
1.26
(0.26)
1.67
(0.32)
3.82
(2.28)

0.49
(0.23)
0.90
(0.22)
1.05
(0.25)

Retest

Prefatigue

0.48
(0.13)
1.45
(0.36)
1.53
(0.35)
1.57
(0.40)
2.02
(0.52)
4.87
(2.96)

0.56
(0.25)
1.15
(0.38)
1.30
(0.42)

Test

0.50
(0.16)
1.37
(0.34)
1.47
(0.33)
1.50
(0.36)
1.96
(0.45)
5.22
(3.00)

0.59
(0.21)
1.11
(0.32)
1.27
(0.34)

Retest

Postfatigue

Athletes (n = 15)

0.47
(0.13)
1.20
(0.29)
1.30
(0.27)
1.30
(0.28)
1.69
(0.33)
4.06
(1.98)

0.50
(0.18)
0.93
(0.21)
1.08
(0.22)

Test

0.44
(0.12)
1.17
(0.26)
1.26
(0.26)
1.25
(0.25)
1.64
(0.31)
3.88
(1.92)

0.51
(0.19)
0.88
(0.20)
1.04
(0.22)

Retest

Prefatigue

0.49
(0.13)
1.32
(0.29)
1.42
(0.29)
1.44
(0.33)
1.88
(0.42)
4.84
(2.57)

0.56
(0.23)
1.05
(0.30)
1.22
(0.33)

Test

(continued)

0.50
(0.14)
1.27
(0.29)
1.37
(0.28)
1.40
(0.30)
1.83
(0.38)
4.85
(2.38)

0.57
(0.17)
1.03
(0.29)
1.20
(0.28)

Retest

Postfatigue

Total (N = 31)

Table 2 COP Measures in Different Visual and Fatigue Conditions for Athletes, Nonathletes, and Total Participants,
Mean (SD)

 449

Mean total
velocity
PP (APML)
Area (95%
ellipse)

0.69
(0.15)
1.75
(0.42)
1.89
(0.41)
1.86
(0.34)
2.41
(0.43)
7.29
(3.22)

0.57
(0.18)
1.35
(0.22)
1.48
(0.22)

0.65
(0.21)
1.86
(0.43)
1.97
(0.45)
1.94
(0.37)
2.49
(0.49)
7.52
(4.47)

0.60
(0.18)
1.38
(0.23)
1.52
(0.24)

Retest

0.70
(0.20)
1.83
(0.37)
1.98
(0.39)
1.94
(0.33)
2.52
(0.45)
9.15
(5.14)

0.67
(0.23)
1.38
(0.26)
1.55
(0.30)

Test

0.68
(0.20)
1.86
(0.44)
1.99
(0.45)
1.98
(0.40)
2.54
(0.52)
8.12
(4.14)

0.66
(0.19)
1.42
(0.28)
1.58
(0.31)

Retest

0.74
(0.27)
1.87
(0.51)
2.02
(0.54)
1.98
(0.54)
2.56
(0.68)
8.29
(5.54)

0.59
(0.25)
1.42
(0.42)
1.55
(0.45)

Test

0.71
(0.27)
1.87
(0.57)
2.02
(0.60)
1.96
(0.60)
2.50
(0.73)
7.40
(4.17)

0.53
(0.16)
1.36
(0.44)
1.47
(0.45)

Retest

0.73
(0.24)
2.04
(0.66)
2.17
(0.69)
2.14
(0.69)
2.76
(0.91)
9.04
(5.71)

0.67
(0.27)
1.54
(0.58)
1.69
(0.62)

Test

0.66
(0.20)
1.89
(0.59)
2.01
(0.60)
2.00
(0.65)
2.58
(0.84)
8.27
(6.35)

0.64
(0.36)
1.43
(0.57)
1.59
(0.65)

Retest

Postfatigue

Athletes (n = 15)
Prefatigue

0.71
(0.22)
1.81
(0.46)
1.95
(0.48)
1.92
(0.44)
2.48
(0.56)
7.77
(4.44)

0.58
(0.21)
1.38
(0.33)
1.51
(0.34)

Test

0.68
(0.23)
1.86
(0.50)
1.99
(0.52)
1.95
(0.49)
2.50
(0.61)
7.46
(4.26)

0.57
(0.17)
1.37
(0.34)
1.50
(0.35)

Retest

0.71
(0.22)
1.93
(0.53)
2.07
(0.56)
2.04
(0.54)
2.64
(0.71)
9.10
(5.33)

0.67
(0.25)
1.46
(0.44)
1.61
(0.48)

Test

0.67
(0.20)
1.87
(0.51)
2.00
(0.52)
1.99
(0.52)
2.56
(0.68)
8.19
(5.23)

0.65
(0.28)
1.43
(0.44)
1.58
(0.49)

Retest

Postfatigue

Total (N = 31)
Prefatigue

COP, center of pressure; AP, anteroposterior; AMP, amplitude; VEL, velocity; PP, phase plane; ML, mediolateral. Units of COP measures are as follows: cm (SD of amplitude);
cm/s (SD of velocity/mean total velocity); cm2 (area). Phase plane is in an arbitrary unit.

PP

SDVEL

SDAMP

ML

PP

SDVEL

SDAMP

AP

Test

Postfatigue

Nonathletes (n = 16)

Prefatigue

(continued)

Eyes Closed

Table 2

450

0.15
0.10
0.15
0.09
0.18
0.17
0.15
0.21
1.53

0.11
0.14
0.16
0.11
0.23
0.25
0.21
0.27
2.22

.48 (.04 to .79)

.73 (.29 to .91)
.57 (.13 to .83)

.66 (.23 to .87)

.52 (.06 to .81)
.55 (.09 to .82)
.62 (.13 to .86)
.57 (.12 to .83)
.61 (.15 to .85)

.69 (.31 to .88)

.89 (.71 to .96)
.86 (.65 to .95)

.83 (.56 to .94)

.83 (.56 to .94)
.82 (.55 to .94)
.87 (.66 to .96)
.86 (.63 to .95)
.79 (.50 to .93)

0.22
0.46
0.49
0.41
0.54
4.35

0.22
0.28
0.32

0.18
0.36
0.34
0.30
0.41
2.99

0.30
0.21
0.30

Prefatigue
SEM
MDC

11.56%
11.09%
10.97%
9.00%
9.92%
16.62%

11.17%
8.51%
8.50%

14.98%
13.68%
12.04%
10.85%
11.62%
22.57%

21.49%
9.71%
12.33%

CV

.71 (.35 to .89)

.83 (.56 to .94)
.82 (.54 to .93)
.89 (.69 to .96)
.84 (.59 to .94)
.74 (.39 to .90)

.72 (.35 to .90)

.88 (.68 to .96)
.85 (.62 to .95)

.51 (.01 to .80)

.62 (.19 to .85)
.60 (.14 to .84)
.74 (.39 to .90)
.67 (.27 to .88)
.56 (.08 to .83)

.59 (.13 to .84)

.80 (.51 to .93)
.77 (.44 to .92)

ICC (95% CI)

0.11
0.25
0.26
0.22
0.34
3.13

0.17
0.19
0.24

0.10
0.21
0.22
0.19
0.28
2.00

0.15
0.16
0.19

0.22
0.49
0.52
0.42
0.67
6.13

0.33
0.38
0.48

0.21
0.42
0.42
0.38
0.55
3.91

0.30
0.31
0.37

Postfatigue
SEM
MDC

9.49%
11.43%
10.64%
9.38%
10.23%
25.45%

20.28%
10.64%
11.28%

13.86%
13.43%
12.90%
11.09%
12.24%
29.22%

22.95%
9.44%
12.63%

CV

COP, center of pressure; AP, anteroposterior; AMP, amplitude; VEL, velocity; PP, phase plane; ML, mediolateral. ICC: intraclass correlation coefficient; CI: confidence
interval; SEM: standard error of measurement; MDC: minimal detectable change; CV: coefficient of variation. Units of COP measures are as follows: cm (SD of amplitude); cm/s (SD of velocity/mean total velocity); cm2 (area). Phase plane is in an arbitrary unit. All correlations above .43 are significant at P < .05.

Eyes Open
AP
SDAMP
SDVEL
PP
ML
SDAMP
SDVEL
PP
Mean total velocity
PP (APML)
Area (95% ellipse)
Eyes Closed
AP
SDAMP
SDVEL
PP
ML
SDAMP
SDVEL
PP
Mean total velocity
PP (APML)
Area (95% ellipse)

ICC (95% CI)

Table 3A TestRetest Reliability of COP Measures in Different Visual and Fatigue Conditions for Athletes (n = 15)

 451

0.13
0.09
0.12
0.07
0.20
0.18
0.17
0.20
0.92
0.10
0.14
0.14
0.09
0.22
0.23
0.19
0.25
2.19

.39 (.13 to .74)

.73 (.38 to .90)
.68 (.28 to .87)

.60 (.19 to .84)

.28 (.20 to .67)
.36 (.13 to .72)
.45 (.04 to .77)
.50 (.02 to .79)
.64 (.22 to .86)

.71 (.36 to .89)

.62 (.19 to .85)
.63 (.21 to .85)

.73 (.39 to .89)

.73 (.39 to .89)
.72 (.37 to .89)
.72 (.37 to .89)
.71 (.35 to .89)
.70 (.32 to .88)

0.19
0.42
0.45
0.37
0.49
4.29

0.20
0.28
0.28

0.14
0.39
0.36
0.33
0.40
1.80

0.26
0.18
0.23

Prefatigue
SEM
MDC

12.01%
8.57%
8.02%
7.68%
7.97%
17.81%

11.50%
8.07%
8.36%

9.85%
11.89%
9.63%
9.14%
8.55%
18.33%

17.03%
6.76%
8.21%

CV

.74 (.40 to .90)

.90 (.74 to .96)
.88 (.68 to .96)
.84 (.61 to .94)
.83 (.58 to .94)
.64 (.24 to .86)

.63 (.21 to .86)

.79 (.50 to .92)
.74 (.41 to .90)

.66 (.26 to .87)

.34 (.18 to .71)
.45 (.05 to .77)
.45 (.07 to .77)
.46 (.05 to .77)
.51 (.03 to .80)

.46 (.05 to .77)

.79 (.50 to .92)
.54 (.05 to .81)

ICC (95% CI)

0.10
0.13
0.15
0.14
0.20
2.81

0.13
0.12
0.16

0.07
0.14
0.14
0.14
0.19
1.39

0.13
0.12
0.14

0.21
0.26
0.30
0.28
0.40
5.50

0.26
0.24
0.31

0.14
0.28
0.28
0.28
0.37
2.72

0.26
0.24
0.28

Postfatigue
SEM
MDC

10.63%
4.62%
5.35%
5.20%
5.39%
20.08%

12.86%
6.84%
6.22%

9.60%
11.96%
8.37%
8.46%
8.73%
23.58%

20.05%
11.96%
9.70%

CV

COP, center of pressure; AP, anteroposterior; AMP, amplitude; VEL, velocity; PP, phase plane; ML, mediolateral. ICC: intraclass correlation coefficient; CI: confidence
interval; SEM: standard error of measurement; MDC: minimal detectable change; CV: coefficient of variation. Units of COP measures are as follows: cm (SD of amplitude); cm/s (SD of velocity/mean total velocity); cm2 (area). Phase plane is in an arbitrary unit. All correlations above .43 are significant at P < .05.

Eyes Open
AP
SDAMP
SDVEL
PP
ML
SDAMP
SDVEL
PP
Mean total velocity
PP (APML)
Area (95% ellipse)
Eyes Closed
AP
SDAMP
SDVEL
PP
ML
SDAMP
SDVEL
PP
Mean total velocity
PP (APML)
Area (95% ellipse)

ICC (95% CI)

Table 3B TestRetest Reliability of COP Measures in Different Visual and Fatigue Conditions for Nonathletes
(n = 16)

452

0.14
0.10
0.14
0.08
0.20
0.19
0.18
0.21
1.23
0.11
0.14
0.15
0.10
0.22
0.24
0.20
0.26
2.20

.44 (.11 to .69)

.73 (.51 to .86)
.61 (.34 to .79)

.63 (.36 to .80)

.43 (.09 to .68)
.49 (.17 to .71)
.56 (.27 to .76)
.55 (.25 to .75)
.61 (.33 to .79)

.69 (.44 to .84)

.82 (.66 to .91)
.80 (.63 to .90)

.79 (.61 to .89)

.78 (.60 to .89)
.78 (.59 to .89)
.82 (.66 to .91)
.80 (.63 to .90)
.75 (.54 to .87)

0.20
0.44
0.46
0.39
0.51
4.32

0.21
0.28
0.30

0.15
0.39
0.38
0.35
0.42
2.41

0.28
0.20
0.27

Prefatigue
SEM
MDC

11.67%
9.76%
9.46%
8.31%
8.92%
17.23%

11.26%
8.29%
8.45%

12.24%
12.74%
10.75%
9.99%
10.07%
20.37%

19.08%
8.21%
10.21%

CV

.72 (.49 to .85)

.84 (.71 to .92)
.83 (.68 to .91)
.87 (.75 to .94)
.83 (.68 to .92)
.69 (.46 to .84)

.69 (.45 to .84)

.86 (.73 to .93)
.82 (.68 to .91)

.56 (.25 to .76)

.62 (.34 to .79)
.61 (.34 to .79)
.71 (.49 to .85)
.66 (.40 to .82)
.54 (.23 to .75)

.54 (.22 to .75)

.74 (.53 to .87)
.73 (.51 to .86)

ICC (95% CI)

0.11
0.20
0.22
0.19
0.28
2.92

0.15
0.17
0.20

0.09
0.18
0.16
0.17
0.23
1.70

0.14
0.15
0.16

0.21
0.40
0.43
0.37
0.55
5.72

0.29
0.33
0.40

0.18
0.35
0.32
0.33
0.46
3.33

0.27
0.30
0.32

Postfatigue
SEM
MDC

10.11%
7.89%
7.90%
7.21%
7.71%
22.68%

16.51%
8.62%
8.77%

11.40%
11.36%
10.65%
9.74%
10.41%
26.30%

21.60%
10.69%
11.13%

CV

COP, center of pressure; AP, anteroposterior; AMP, amplitude; VEL, velocity; PP, phase plane; ML, mediolateral. ICC: intraclass correlation coefficient; CI: confidence interval;
SEM: standard error of measurement; MDC: minimal detectable change; CV: coefficient of variation. Units of COP measures are as follows: cm (SD of amplitude); cm/s (SD
of velocity/mean total velocity); cm2 (area). Phase plane is in an arbitrary unit. All correlations above .43 are significant at P < .05.

Eyes Open
AP
SDAMP
SDVEL
PP
ML
SDAMP
SDVEL
PP
Mean total velocity
PP (APML)
Area (95% ellipse)
Eyes Closed
AP
SDAMP
SDVEL
PP
ML
SDAMP
SDVEL
PP
Mean total velocity
PP (APML)
Area (95% ellipse)

ICC (95% CI)

Table 3C TestRetest Reliability of COP Measures in Different Visual and Fatigue Conditions for All Subjects (N = 15)

Force-Platform Measures of Postural Sway 453

previous studies that reported higher reliability of tests with EC than with EO. In a
study to establish the reliability of some commonly used COP measures in a healthy
population, Doyle et al14 found generalizability coefficients (a measure equivalent
to ICC) of .50, .49, .83, and .49 for the SD of AP amplitude, SD of ML amplitude,
mean total velocity, and area, respectively, in the EO condition and .66, .66, .84 and
.68 in the EC condition, based on 3 trials of 30 seconds duration. Bauer et al,18 in a
study to investigate the intrasession reliability of COP measures in a large sample of
healthy elderly individuals, found higher ICCs for tests with EC (.71, .95, .93, and
.95 for area, length, ML sway, and AP sway, respectively) than those with EO (.88,
.89, .84, and .91) in a condition of narrow bilateral stance. Generally these results
support our finding regarding higher reliability of COP measures in the EC condition. Based on these data, it is recommended that researchers employ tests under
blindfolded conditions rather than using only tests with EO to measure postural
sway in athletes. Quiet stance in the absence of visual feedback might be perceived
as a new and challenging postural condition by most participants.18 This might lead
to adoption of variable postural strategies (as indicated by the larger SDs of COP
measures in EC than in the EO condition in Table 2), resulting in higher ICCs. As
indicated before, the ICC is a ratio of between-subjects variance to total variance;
hence, the greater the differences between individuals, the greater the ICC.35
The preceding rationale may also be a likely explanation for skill- and fatiguerelated differences in the reliability of COP measures. The higher ICC values
obtained for athletes and for fatigued conditions may have resulted from a higher
variation of measures among subjects in these skill and fatigue levels (as shown in
Table 2). Contrary to reliability, the measurement precision (as indexed by SEM
and CV) was higher for COP measures in the nonathlete group and prefatigue.
The reverse patterns of reliability and precision support the notion that the ICC
and SEM or CV are not appropriate surrogates for each other.35,36 They provide
different information regarding reliability. Although ICCs can be used for diagnostic purposesthe ability of a measure to discriminate between individuals (ie,
between-subjects designs)the SEM is useful for evaluative purposesto detect
changes over time after an intervention (ie, within-subject designs).4
Depending on the COP variable, our results can be compared with other studies. Salavati et al28 investigated between-days reliability of COP measures similar
to those reported in the current study in a group of people with musculoskeletal
disorders including low back pain, anterior cruciate ligament injury, and functional ankle instability. Mean total velocity (ICCs of .84 and .91 for EO and EC,
respectively), phase-plane portrait (ICCs of .71 and .82), and phase-plane portrait
in the ML direction (ICCs of .79 and .69) were the best parameters with respect
to reliability. Differences in studied populations may be an explanation for the
disparity in results. Santos et al4 obtained ICC values of .40 (root mean square
[RMS], equivalent to SD of amplitude, in the AP direction), .45 (RMS in the ML
direction), .53 (mean velocity), and .40 (area) in the EO condition and .46 (RMS
in the AP direction), .43 (RMS in the ML direction), .44 (mean velocity), and .43
(area) in the EC condition in a sample of 12 young healthy people. The reliability
coefficients of COP measures in the current study were higher, maybe because of
the larger sample size.
Although some researchers have provided evidence for the positive effect of
exercise training on postural stability,38,39 they have not determined whether the

454 Meshkati et al

change in postural performance is real or a result of measurement error. The MDC

can be used to determine the real change in an individuals performance.36 The
difference between 2 measurements should be beyond the MDC to be regarded
as a real change in performance. For example, for an athlete assessed in EO and
the nonfatigued condition with an initial mean velocity of 1.30 m/s, changes in
subsequent values after sport training would be considered real if they were outside
the range of 1 to 1.60 m/s (Table 3). A measure with a lower MDC detects smaller
changes in an individuals performance.
A few notes are warranted with regard to this study. In addition to biological
variability, the lack of reliability of COP measures in some conditions (eg, EO) may
be a result of insufficient number or duration of the trials employed.14 Although
increasing the number or duration of trials would improve the reliability of these
measures,14 the burdens of time, cost, and fatigue imposed on the athlete and examiner must be taken into account. Furthermore, it is important to note that because
sex differences have been found in COP parameters,40 the results of the current
study may be more generalizable to men, who constituted all of our participants.
Although a relationship between base of support and the ability of a measure to
assess specific aspects of postural control in athletes has been demonstrated in the
literature,41 we also recommend evaluating the reliability of COP measures in other
support conditions such as unilateral stance. Moreover, because of the relationship
between sport type and postural performance, the results might vary in athletes in
other sports not considered in the current study.

Conclusions
In conclusion, tests with EC and to a lesser extent tests performed by athletes and
in a fatigued condition showed better reliability. The SD of AP velocity showed the
highest reliability in most conditions. These results may be useful for those whose
major concern is to select reliable COP measures for studying postural control in
people with different levels of expertise. It would be necessary to assess the reliability of COP measures in another gender and other sport activities in future studies.
Acknowledgments
The study was supported by the Science and Research Branch, Islamic Azad University,
Tehran, Iran.

References
1. Perrin P, Deviterne D, Hugel F, Perrot C. Judo, better than dance, develops sensorimotor
adaptabilities involved in balance control. Gait Posture. 2002;15:187194.
2. Golomer E, Crmieux J, Dupui P, Isableu B, Ohlmann T. Visual contribution to selfinduced body sway frequencies and visual perception of male professional dancers.
Neurosci Lett. 1999;267:189192.
3. Palmieri RM, Ingersoll CD, Stone MB, Krause BA. Center-of-pressure parameters
used in the assessment of postural control. J Sport Rehabil. 2002;11:5166.
4. Santos BR, Delisle A, Larivire C, Plamondon A, Imbeau D. Reliability of centre of
pressure summary measures of postural steadiness in healthy young adults. Gait Posture.
2008;27:408415.

Force-Platform Measures of Postural Sway 455

5. Ageberg E, Roberts D, Holmstrm E, Fridn T. Balance in single-limb stance in healthy

subjects-reliability of testing procedure and the effect of short-duration sub-maximal
cycling. BMC Musculoskelet Disord. 2003;4:14.
6. Harringe ML, Halvorsen K, Renstrm P, Werner S. Postural control measured as the
center of pressure excursion in young female gymnasts with low back pain or lower
extremity injury. Gait Posture. 2008;28:3845.
7. Perkins DO, Wyatt RJ, Bartko JJ. Penny-wise and pound-foolish: the impact of
measurement error on sample size requirements in clinical trials. Biol Psychiatry.
2000;47:762766.
8. Goldie PA, Bach TM, Evans OM. Force platform measures for evaluating postural
control: reliability and validity. Arch Phys Med Rehabil. 1989;70:510517.
9. Le Clair K, Riach C. Postural stability measures: what to measure and for how long.
Clin Biomech (Bristol, Avon). 1996;11:176178.
10. Benvenuti F, Mecacci R, Gineprari I, et al. Kinematic characteristics of standing
disequilibrium: reliability and validity of a posturographic protocol. Arch Phys Med
Rehabil. 1999;80:278287.
11. Rogind H, Simonsen H, Era P, Bliddal H. Comparison of Kistler 9861A force platform
and Chattecx Balance System for measurement of postural sway: correlation and test
retest reliability. Scand J Med Sci Sports. 2003;13:106114.
12. Lafond D, Corriveau H, Hbert R, Prince F. Intrasession reliability of center of pressure measures of postural steadiness in healthy elderly people. Arch Phys Med Rehabil.
2004;85:896901.
13. Doyle TL, Newton RU, Burnett AF. Reliability of traditional and fractal dimension
measures of quiet stance center of pressure in young, healthy people. Arch Phys Med
Rehabil. 2005;86:20342040.
14. Doyle RJ, Hsiao-Wecksler ET, Ragan BG, Rosengren KS. Generalizability of center
of pressure measures of quiet standing. Gait Posture. 2007;25:166171.
15. Lin D, Seol H, Nussbaum MA, Madigan ML. Reliability of COP-based postural sway
measures and age-related differences. Gait Posture. 2008;28:337342.
16. Pinsault N, Vuillerme N. Testretest reliability of centre of foot pressure measures to
assess postural control during unperturbed stance. Med Eng Phys. 2009;31:276286.
17. Swanenburg J, de Bruin ED, Favero K, Uebelhart D, Mulder T. The reliability of
postural balance measures in single and dual tasking in elderly fallers and non-fallers.
BMC Musculoskelet Disord. 2008;9:162.
18. Bauer C, Grger I, Rupprecht R, Gassmann KG. Intrasession reliability of force
platform parameters in community-dwelling older adults. Arch Phys Med Rehabil.
2008;89:19771982.
19. Gutierrez GM, Jackson ND, Dorr KA, Margiotta SE, Kaminski TW. Effect of fatigue
on neuromuscular function at the ankle. J Sport Rehabil. 2007;16:295306.
20. Sandrey MA, Kent TE. The effects of eversion fatigue on frontal plane joint position
sense in the ankle. J Sport Rehabil. 2008;17:257268.
21. Yaggie J, Armstrong WJ. Effects of lower extremity fatigue on indices of balance. J
Sport Rehabil. 2004;13:312322.
22. Harkins KM, Mattacola CG, Uhl TL, Malone TR, McCrory JL. Effects of 2 ankle fatigue
models on the duration of postural stability dysfunction. J Athl Train. 2005;40:191194.
23. Gribble PA, Robinson RH, Hertel J, Denegar CR. The effects of gender and fatigue on
dynamic postural control. J Sport Rehabil. 2009;18:240257.
24. Nardone A, Tarantola J, Giordano A, Schieppati M. Fatigue effects on body balance.
Electroencephalogr Clin Neurophysiol. 1997;105:309320.
25. Yaggie JA, McGregor SJ. Effects of isokinetic ankle fatigue on the maintenance of
balance and postural limits. Arch Phys Med Rehabil. 2002;83:224228.
26. Salavati M, Mazaheri M, Negahban H, et al. Effect of dual-tasking on postural control
in subjects with nonspecific low back pain. Spine. 2009;34:14151421.

456 Meshkati et al

27. Riley MA, Baker AA, Schmit JM, Weaver E. Effects of visual and auditory shortterm
memory tasks on the spatiotemporal dynamics and variability of postural sway. J Mot
Behav. 2005;37:311324.
28. Salavati M, Hadian MR, Mazaheri M, et al. Testretest reliability of center of pressure
measures of postural stability during quiet standing in a group with musculoskeletal
disorders consisting of low back pain, anterior cruciate ligament injury and functional
ankle instability. Gait Posture. 2009;29:460464.
29. Riley PO, Benda BJ, Gill-Body KM, Krebs DE. Phase plane analysis of stability in
quiet standing. J Rehabil Res Dev. 1995;32:227235.
30. Borg GAV. Psychophysical bases of perceived exertion. Med Sci Sports Exerc.
1982;14:377381.
31. Borg G, Hassmn P, Lagerstrm M. Perceived exertion related to heart rate and blood
lactate during arm and leg exercise. Eur J Appl Physiol Occup Physiol. 1987;56:679
685.
32. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol
Bull. 1979;86:420428.
33. Denegar CR, Ball DW. Assessing reliability and precision of measurement: an introduction to intraclass correlation and standard error of measurement. J Sport Rehabil.
1993;2:3542.
34. Domholdt E. Rehabilitation Research: Principles and Applications. 3rd ed. Philadelphia:
Elsevier Saunders; 2005.
35. Atkinson G, Nevill AM. Statistical methods for assessing measurement error (reliability)
in variables relevant to sports medicine. Sports Med. 1998;26:217238.
36. Weir JP. Quantifying testretest reliability using the intraclass correlation coefficient
and the SEM. J Strength Cond Res. 2005;19:231240.
37. Douris P, Chinan A, Gomez M, Aw A, Steffens D, Weiss S. Fitness levels of middle
aged martial art practitioners. Br J Sports Med. 2004;38:143147.
38. Hue OA, Seynnes O, Ledrole D, Colson SS, Bernard PL. Effects of a physical activity
program on postural stability in older people. Aging Clin Exp Res. 2004;16:356362.
39. Wolf SL, Barnhart HX, Ellison GL, Coogler CE. The effect of Tai Chi Quan and
computerized balance training on postural stability in older subjects. Atlanta FICSIT
Group. Frailty and Injuries: Cooperative Studies on Intervention Techniques. Phys
Ther. 1997;77:371381.
40. Nolan L, Grigorenko A, Thorstensson A. Balance control: sex and age differences in
9- to 16-year-olds. Dev Med Child Neurol. 2005;47:449454.
41. Asseman FB, Caron O, Cremieux J. Are there specific conditions for which expertise
in gymnastics could have an effect on postural control and performance? Gait Posture.
2008;27:7681.