Gait & Posture 36 (2012) 172176

Contents lists available at SciVerse ScienceDirect

Gait & Posture

journal homepage: www.elsevier.com/locate/gaitpost

Inuence of real and virtual heights on standing balance

Taylor W. Cleworth, Brian C. Horslen, Mark G. Carpenter *
Neural Control of Posture and Movement Laboratory, School of Kinesiology, The University of British Columbia, Vancouver, BC, Canada

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 30 July 2011
Received in revised form 6 February 2012
Accepted 15 February 2012

Fear and anxiety induced by threatening scenarios, such as standing on elevated surfaces, have been
shown to inuence postural control in young adults. There is also a need to understand how postural
threat inuences postural control in populations with balance decits and risk of falls. However, safety
and feasibility issues limit opportunities to place such populations in physically threatening scenarios.
Virtual reality (VR) has successfully been used to simulate threatening environments, although it is
unclear whether the same postural changes can be elicited by changes in virtual and real threat
conditions. Therefore, the purpose of this study was to compare the effects of real and virtual heights on
changes to standing postural control, electrodermal activity (EDA) and psycho-social state. Seventeen
subjects stood at low and high heights in both real and virtual environments matched in scale and visual
detail. A repeated measures ANOVA revealed increases with height, independent of visual environment,
in EDA, anxiety, fear, and center of pressure (COP) frequency, and decreases with height in perceived
stability, balance condence and COP amplitude. Interaction effects were seen for fear and COP mean
position; where real elicited larger changes with height than VR. This study demonstrates the utility of
VR, as simulated heights resulted in changes to postural, autonomic and psycho-social measures similar
to those seen at real heights. As a result, VR may be a useful tool for studying threat related changes in
postural control in populations at risk of falls, and to screen and rehabilitate balance decits associated
with fear and anxiety.
2012 Elsevier B.V. All rights reserved.

Keywords:
Virtual reality
Fear
Anxiety
Posture

1. Introduction
Fear of falling is prevalent in a number of populations, such as
the elderly and individuals with Parkinsons disease, and has been
associated with balance decits and increased risk of falls [15]. As
a result, recent efforts have been made to examine how fear and
other related factors such as anxiety, arousal and balance
condence may directly inuence balance in such populations.
In young healthy adults during quiet stance, signicant changes in
frequency, amplitude and mean position of center of pressure
(COP) displacements occur when fear and anxiety are induced by
standing at the edge of an elevated surface. While there is also a
clear need to investigate how fear and anxiety contribute to
balance decits associated with age and disease, signicant safety
and feasibility issues are raised if individuals with known fear and/
or balance problems are required to stand, physically, at high
surface heights.
One possible solution to this problem is to use virtual reality
(VR) to recreate the experience of standing on elevated surfaces.

* Corresponding author at: School of Kinesiology, The University of British

Columbia, 6108 Thunderbird Blvd, Vancouver, BC V6T 1Z4, Canada.
Tel.: +1 604 822 8614.
E-mail address: mark.carpenter@ubc.ca (M.G. Carpenter).
0966-6362/$ see front matter 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.gaitpost.2012.02.010

Past research with acrophobic individuals (those with a fear of

heights) has shown that VR is benecial in improving attitudes
toward heights while reducing height induced anxiety and
avoidance [68]. In addition, standing at simulated heights has
been associated with changes in physiological arousal, anxiety, and
perceptions of height that are similar to those recorded under real
height conditions in healthy adults [914]. However, the
effectiveness of using simulated heights to induce threat-related
changes to postural control is not known. Simeonov et al. [9]
compared the effects of real and virtual experiences of standing at
different heights, in combination with different support surfaces
and visual target locations, on postural control. Post hoc analyses
for 3-way interactions observed between visual environment,
height and surface compliance for AP sway and sway velocity,
revealed no signicant differences between real and virtual
environments within each height condition. However, the question
of whether height induced changes in posture depend on visual
environment was not specically tested in this study. Therefore,
there is still a need to investigate how changes in height affect
standing balance in real and virtual environments in order to
justify VRs use in studying the relationship between fear, anxiety
and postural control.
The purpose of this study was to examine how changes in
height within real and virtual environments can inuence standing

T.W. Cleworth et al. / Gait & Posture 36 (2012) 172176

postural control, autonomic function and self-reported perceptions of fear, anxiety, perceived stability and balance condence in
healthy young adults. We hypothesized that the effects of height
on COP, psycho-social and physiological measures would not be
different between real and virtual environments.

2. Methods
18 young healthy adults (11 females) aged between 19 and
28 years volunteered to participate in this study. Participants were
clear of any known neurological or balance disorders and provided
written informed consent prior to the beginning of testing. The
UBC Clinical Research Ethics Board approved all experimental
procedures. One subject was removed from the experiment due to
a malfunction of the VR system that caused motion sickness.
2.1. Apparatus
Ground reaction forces were collected from a force platform
(#K00407, Bertec, USA) and used to calculate COP. The force
platform was placed at the edge (as seen in Fig. 1) of a
2.13 m  1.52 m hydraulic lift (Pentalift, Guelph, Canada) that
was situated at two heights, 0.8 m (low condition) and 3.2 m (high
condition) above ground. A table was placed in front of the
platform for the low condition in order to reduce threat and was
removed before the platform was raised for the high condition. The
low condition was presented rst in order to optimize the fear and
anxiety effects induced by height [15]. Throughout the experiment,
the participants were securely fastened to the ceiling in case of a
fall. The supporting rope did not provide any cutaneous sensation
or assistance unless a fall was to occur.

173

2.2. Procedures
All participants were subjected to the following two visual
environment conditions (real and virtual), the order of which was
counterbalanced across subjects.
2.2.1. Real environment protocol
Subjects rst performed a practice trial at the low height to
eliminate any rst trial effects in each visual environment [15], and
to familiarize the participant with the protocol.
Following the practice trial, subjects stood quietly for 120 s at
the low (0.8 m) height. After the low trial, subjects were seated and
raised to a height of 3.2 m. Subjects were then assisted to the edge
of the platform prior to beginning a second 120 s standing trial
(high). In all trials, the feet were positioned with the toes at the
edge of the platform and the most lateral borders of their feet
spaced at a distance equal to their foot length. Foot position was
marked prior to the experiment to ensure a constant foot position
across standing conditions. During each trial, participants stood
with their eyes open and focused on a marker placed at eye level on
the wall 3.84 m in front of them.
2.2.2. Virtual environment protocol
Subjects also performed three 120 s standing trials within the
virtual environment following the same order (practice, low, high)
as the real environment. However, in the virtual high condition, the
participant only perceived themselves to be raised to a simulated
height while physically remaining at the low height (Fig. 1). To
improve the illusion of being elevated to height, the table in front of
the platform was moved 5 cm away to create the sensation of an
edge (similar to the one experienced at the real height). Additional
auditory and haptic cues were used to increase sense of presence

Fig. 1. Diagram of experimental setup for virtual and real environments, comparing the visual scene (observed by the participant) in the low and high conditions. The model
located to the left of each visual scene illustrates the actual standing orientation of the participant during each condition.

174

T.W. Cleworth et al. / Gait & Posture 36 (2012) 172176

[12], and enhance the subjects virtual experience of being raised to

height. In particular, an audio recording of the sounds generated by
the motor during the actual elevation of the platform were
provided through headphones. The experimenter also lightly
shook the platform during simulated ascent, to mimic the platform
vibration experienced during the real elevation of the platform.

COP signal in the AP direction. The mean COP was removed prior
to calculation of root mean square (RMS) and mean power
frequency (MPF) of COP displacements in both the AP and ML
directions as measures of the amplitude and the frequency of
COP displacements, respectively.
2.4. Statistical analysis

2.2.3. VR setup
There are limitations to VR that must be controlled for in order
to: reduce the possibility of motion sickness; create a sense of
presence; and minimize postural changes due to VR. The virtual
environment was displayed using a piSight head mounted display
(HMD) placed on the head of the participant (Fig. 1), with a
resolution of 20 pixels/degree in approximately a 1508 horizontal
and 608 vertical eld of view. This relatively large eld of view
closely matches a real human eld of view creating a greater sense
of presence. The virtual scene was developed using Vizard
software (Worldviz, California, USA), and closely matched the
visual scene of the lab in both scale and detail (Fig. 1). Optotrak
(Northern Digital, Inc, Canada) cameras were used to determine
head position and updated the scene at a frequency of 250 Hz. This
minimized the time difference between actual head movement
and movement of scene (known as end-to-end latency). With such
small end-to-end latencies, that are essentially undetectable to
participants, presence should increase [11], and the risk of motion
sickness decrease [16].
Immersion protocols have been used to help further increase
sense of presence. Therefore, prior to testing, a brief immersion
period was used to orient the subject within the new virtual world.
This also allowed them to adapt to mechanical factors, such as the
weight of the HMD which may affect postural control. First, the
HMD was adjusted into an optimal position for displaying a single,
seamless image, then a black cloth was placed around the edges of
the HMD to ensure the subject could only see the virtual scene, and
therefore could not gain any visual inputs from the real world [17].
Prior to the practice trial subjects stood and performed a series of
object search and identication exercises. Shapes and words were
placed in the room while subjects were asked to locate, identify
and/or track these objects. The immersion period lasted no less
than ve minutes to ensure a sense of presence was accomplished.
2.3. Measurements
Two disposable surface Ag/AgCl electrodes were placed on the
thenar and hypothenar eminences on the non-dominant hand and
used to record electrodermal activity (EDA) with a sampling rate of
1 kHz (model 2502, CED, UK).
A series of questionnaires were projected on the wall in front of
the participant in both visual environments. Balance condence
questionnaires were completed prior to each standing trial (low
and high). State anxiety, perceived fear of falling and perceived
standing stability questionnaires were assessed immediately after
each standing trial (low and high) [18,19]. Subjects answered the
questionnaires verbally and had their responses recorded by an
experimenter. With the exception of two initial pilot subjects,
participants were also asked to provide a verbal numerical
estimate of the perceived height at which they stood following
the standing trial at 3.2 m. This estimate was then used to calculate
a ratio of overestimation, which is calculated by dividing estimated
height by actual height [20].
Ground reaction forces and moments were sampled at
100 Hz from the force plate and low pass ltered with a 5 Hz
dual pass Butterworth lter before calculating COP displacements in both the anteriorposterior (AP) and medial-lateral
(ML) directions. Mean position was calculated from the ltered

The effects of height (low and high) and visual environment

(real and VR) were examined using a 2  2 repeated measures
analysis of variance (ANOVA) for all dependent variables (RMS and
MPF of AP and ML COP, mean position of AP COP and all
physiological and psycho-social measures). Signicance was
determined using p  0.05. For signicant interactions, paired
samples t-tests were used to determine the signicant effects of
height within each visual environment. A paired samples t-test
was used to analyze the ratio of overestimation across the two
visual environments.

3. Results
3.1. Physiological and psycho-social measures
There were main effects of height and visual environment on
EDA, but no interaction effect. EDA was larger in the high
conditions compared to the low conditions (F(1,16) = 12.459,
p = 0.003), and was also larger in real environment compared to
virtual environment (F(1,16) = 5.086, p = .039) (Fig. 2B). Signicant main effects of height on psycho-social measures were
observed. When standing in the high condition compared to the
low condition, participants reported signicantly higher levels
of anxiety (F(1,16) = 35.635, p < 0.001), and lower levels of
stability (F(1,16) = 60.697, p < 0.001) and balance condence
(F(1,16) = 40.433, p < 0.001), independent of visual environment
(Fig. 2B). There was an interaction between height and visual
environment for fear of falling (F(1,16) = 4.914, p = 0.041). Post
hoc analyses revealed signicant effects of height in both visual
environments, but with greater increases in fear observed
between heights in the real (t(16) = 6.390, p < 0.001) compared
to the VR conditions (t(16) = 5.055, p < 0.001) (Fig. 2B). Ratios of
overestimation were also signicantly different across visual
environments (t(14) = 2.316, p = 0.036). Although height was
overestimated in both visual environments (ratio > 1), larger
overestimates of height were reported in the VR (1.55  0.14)
compared to the real (1.41  0.14) height environment.
3.2. COP measures
Repeated measures ANOVA revealed a signicant interaction
between height and visual environment for mean position of COP
in the AP direction (F(1,16) = 10.446, p = 0.005). Post hoc analysis
revealed a signicant shift in mean COP position away from the
edge at the high height compared to the low height in the real
environment (t(16) = 4.209, p = 0.001), with a similar trend
observed in the virtual environment (t(16) = 1.899, p = 0.076)
(Fig. 2A). There was a signicant main effect of height on both COP
MPF in the AP (F(1,16) = 7.735, p = 0.013) and ML (F(1,16) = 4.871,
p = 0.042) directions, where MPF was higher in the high height
compared to low height, independent of visual environment
(Fig. 2A). There was also a main effect of height on COP RMS in the
AP plane (F(1,16) = 5.092, p = 0.038), but not ML. AP-RMS was
lower in the high height compared to the low height, independent
of visual environment (Fig. 2A). There were no signicant
differences between visual environments on AP-MPF (p = 0.192)
or AP-RMS (p = 0.155).

T.W. Cleworth et al. / Gait & Posture 36 (2012) 172176

175

Fig. 2. A comparison between height conditions (low and high) for all (A) postural (COP mean position, RMS and MPF) in the AP (solid line) and ML (dashed line) directions
(mean  standard error mean), and (B) psycho-physiological measures (mean  standard error). The dark lines represent real conditions and the light lines represent VR conditions.
Note that in the mean position graph, the center of the force plate is located at 0 mm, and negative values indicate a mean position closer to the edge.

4. Discussion
Whether standing at a physical height or a simulated height of
3.2 m, signicant decreases in perceptions of stability and balance
condence, and signicant increases in perceptions of anxiety
were observed independent of visual environment. While fear also
signicantly increased with height in both visual environments,
the effect of height was larger in the real compared to the virtual
environment. This is in line with previous work that has shown
that VR can produce physiological indicators of anxiety and
arousal, such as skin conductance, heart rate and skin temperature
changes when subjects perceive themselves to be standing at a
simulated height [1012].

The observation that subjects overestimated height in the real

high condition (ratio of 1.41) is consistent with previous reports of
subjects overestimating heights when looking down from an
elevated height [9,20,21]. The fact that subjects also overestimated
height in the virtual high condition supports the notion that
subjects had a strong sense of presence and perceived themselves
to be standing on, and looking down from, a high height in the
virtual environment.
Not only were there similar changes in physiological and
psycho-social measures found between heights across visual
environments, there were also similar changes to standing
postural control. Specically, the virtual environment was
successful at reproducing the effects of standing at a physical

176

T.W. Cleworth et al. / Gait & Posture 36 (2012) 172176

height on postural control, resulting in increases in frequency and

decreases in amplitude of AP-COP displacements. These changes
are congruent with the postural changes observed in previous
studies in which subjects stood on elevated surfaces [15,22,23].
The change in COP mean position between heights was not the
same across visual environments, as the real environment induced
a larger lean away from the edge than the virtual. The direct
relationship between fear and lean is not fully understood; as it
seems to be inuenced by other factors such as focus of attention
[24] and different sources of anxiety, arousal and fear [2527].
Therefore, further investigation is needed in order to distinguish
why participants lean further away from the edge in a real
compared to a virtual height.
It has been reported previously that posture is more unstable in
a virtual environment [16,17,28], even when system latency, eld
of view and HMD weight have been controlled for [16,28]. The
mechanisms behind this destabilizing effect are unknown, but
have been suggested to come from technical limitations related to
sense of presence and sensory conicts. Therefore, the absence of
an observed effect of VR on postural control during quiet stance in
the current study suggests that we have optimized our end-to-end
latencies, realism (creating our virtual environment as a duplicate
to the real environment), immersion period, eld of view and/or
sense of presence. By optimizing these mechanical aspects and
limiting sensory conict, the effects of VR on postural control were
minimized. Likewise, no adverse effects of standing in a virtual
environment were observed for measures of physiological arousal
or psycho-social state. Anxiety, balance condence and perceived
stability were unchanged between visual environments, while EDA
was signicantly decreased in the VR compared to real environment. Similar observations of decreased EDA when standing in a
VR environment have been reported previously [9].
The advantage of VR is that it can be used to place subjects in a
variety of environments that they may otherwise avoid due to fear or
safety restrictions. As a result fear related changes in postural control
can be studied under a wide range of threatening contexts (such as
standing at a high height, walking on ice or riding a moving bus)
within the safe connes of a lab or clinic. In young healthy adults,
virtual heights induced increases in physiological and psycho-social
measures of anxiety and fear, altered perceptions of height and
changes in postural control. These results suggest that VR may be a
useful tool for studying how threat inuences postural control in
populations at risk of falls, and to screen and rehabilitate balance
decits associated with fear and anxiety.
Acknowledgments
This project was supported by the Natural Sciences and
Engineering Research Council of Canada and Canadian Foundation
of Innovation grants.
Conict of interest statement
None of the authors have any conicts of interest with this
study.

References
[1] Adkin AL, Frank JS, Jog MS. Fear of falling and postural control in Parkinsons
disease. Movement Disord 2003;18:496502.
[2] Bloem BR, Grimbergen YA, Cramer M, Willemsen M, Zwinderman AH. Prospective assessment of falls in Parkinsons disease. J Neurol 2001;248:9508.
[3] Hindmarsh JJ, Estes EH. Falls in older persons: causes and interventions. Arch
Intern Med 1989;149:221722.
[4] Li F, Fisher KJ, Harmer P, McAuley E, Wilson NL. Fear of falling in elderly
persons: association with falls, functional ability, and quality of life. J Gerontol
B Psychol Sci Soc Sci 2003;58B:28390.
[5] Maki BE, Holliday PJ, Topper AK. Fear of falling and postural performance in the
elderly. J Gerontol 1991;46(4):M12331.
[6] Emmelkamp PM, Bruynzeel M, Drost L, van der Mast CA. Virtual reality
treatment in acrophobia: a comparison with exposure in vivo. Cyberpsychol
Behav 2001;4:3359.
[7] Hodges LF, Kooper R, Meyer TC, Rothbaum BO, Opdyke D, de Graaff JJ, Williford
JS, North MM. Virtual environments for treating the fear of heights. Computer
1995;28(7):2734.
[8] Rothbaum BO, Hodges LF, Kooper R, Opdyke D, Williford JS, North M. Effectiveness of computer generated (virtual reality) graded exposure in the
treatment of acrophobia. Am J Psychiat 1995;152:6268.
[9] Simeonov PI, Hsiao H, Dotson BW, Ammons DE. Height effects in real and
virtual environments. Hum Factors 2005;47(2):4308.
[10] Meehan M, Insko B, Whitton M, Brooks FP. Physiological measures of presence
in stressful virtual environments. In: Proceedings of ACM SIGGRAPH, vol. 21;
2002. p. 64552.
[11] Meehan M, Razzaque S, Whitton MC, Brooks FP. Effect of latency on presence in
stressful virtual environments. In: Proceedings of the IEEE Virtual Reality, vol.
2226; 2003. p. 1418.
[12] Meehan M, Razzaque S, Insko B, Whitton M, Brooks FP. Review of four studies
on the use of physiological reaction as a measure of presence in stressful
virtual environments. Appl Psychophysiol Biofeedback 2005;30(3):23958.
[13] Wilhelm FH, Pfaltz MC, Gross JJ, Mauss IB, Kim SI, Wiederhold BK. Mechanisms of
virtual reality exposure therapy: the role of the behavioural activation and
behavioural inhibition systems. Appl Psychophysiol Biofeedback 2005;30:27184.
[14] Zimmons P, Panter A. The inuence of rendering quality on presence and task
performance in a virtual environment. In: Proceeding of the IEEE Virtual
Reality, vol. 2226. 2003. p. 2934.
[15] Adkin AL, Frank JS, Carpenter MG, Peysar GW. Postural control is scaled to level
of postural threat. Gait Posture 2000;12(2):8793.
[16] Akizuki H, Uno A, Arai K, Morioka S, Ohyama S, Nishiike S, Tamura K, Takeda N.
Effects of immersion in virtual reality on postural control. Neurosci Lett
2005;379(1):236.
[17] Horlings CG, Carpenter MG, Kung UM, Honegger F, Wiederhold B, Allum JH.
Inuence of virtual reality on postural stability during movements of quiet
stance. Neurosci Lett 2009;451(3):22731.
[18] Adkin AL, Frank JS, Carpenter MG, Peysar GW. Fear of falling modies anticipatory postural control. Exp Brain Res 2002;143:16070.
[19] Carpenter MG, Frank JS, Adkin AL, Paton A, Allum JHJ. Inuence of postural
reactions to multi-directional surface rotations. J Neurophysiol 2004;92:325565.
[20] Clerkin E, Cody M, Stefanucci J, Proftt D, Teachman B. Imagery and fear
inuence height perception. J Anxiety Disord 2009;23:3816.
[21] Teachman B, Stefanucci J, Clerkin E, Cody M, Proftt D. A new mode of fear
expression: perceptual bias in height fear. Emotion 2008;8:296301.
[22] Carpenter MG, Frank JS, Silcher CP, Peysar GW. The inuence of postural threat
on the control of upright stance. Exp Brain Res 2001;138(2):2108.
[23] Davis JR, Campbell AD, Adkin AL, Carpenter MG. The relationship between fear
of falling and human postural control. Gait Posture 2009;29(2):2759.
[24] Huffman JL, Horslen BC, Carpenter MG, Adkin AL. Does increased postural
threat lead to more conscious control of posture? Gait Posture
2009;30(4):52832.
[25] Geh CLM, Beauchamp MR, Crocker PRE, Carpenter MG. Assessed distressed:
white-coat effects on clinical balance performance. J Psychosom Res
2011;70(1):4551.
[26] Horslen BC, Carpenter MG. Arousal, valence and their relative effects on
postural control. Exp Brain Res 2011;215:2734.
[27] Maki BE, Holliday PJ, Topper AK. A prospective study of postural balance and
risk of falling in an ambulatory and independent elderly population. J Gerontol
1994;49(2):M7284.
[28] Kelly JW, Riecke B, Loomis JM, Beall AC. Visual control of posture in real and
virtual environments. Percept Psychophys 2008;70:15865.