Influence of the Perspectives on the Movement of One-Leg Lifting in an Interactive-Visual Virtual Environment: A Pilot Study
Influence of the Perspectives on the Movement of One-Leg Lifting in an Interactive-Visual Virtual Environment: A Pilot Study
Chien-Hua Huang 0 1
Chun Pei 1
Tien-Lung Sun 0 1
0 Department of Industrial Engineering & Management, Yuan Ze University , Chung-Li, Taiwan , 2 Feng Yuan Hospital, Ministry of Health and Welfare, Taichung City, Taiwan, 3 Department of Geronic Technology and Service Management, Nan Kai University of Technology , Nantou County , Taiwan
1 Editor: Mikhail A. Lebedev, Duke University , UNITED STATES
Numerous studies have confirmed the feasibility of active video games for clinical rehabilitation. To maximize training effectiveness, a personal program is necessary; however, little evidence is available to guide individualized game design for rehabilitation. This study assessed the perspectives and kinematic and temporal parameters of a participant's postural control in an interactive-visual virtual environment.
Data Availability Statement: All relevant data are
within the paper.
Funding: The author(s) received no specific
funding for this work.
Competing Interests: The authors have declared
that no competing interests exist.
Twenty-four healthy participants performed one-leg standing by leg lifting when a posture
frame appeared either in a first- or third-person perspective of a virtual environment. A foot
force plate was used to detect the displacement of the center of pressure. A three-way
mixed factor design was applied, where the perspective was the between-participant factor,
and the leg-lifting times (0.7 and 2.7 seconds) and leg-lifting angles (30Êand 90Ê) were the
within-participant factors. The reaction time, accuracy of the movement, and ability to shift
weight were the dependent variables.
Regarding the reaction time and accuracy of the movement, there were no significant main effects of the perspective, leg-lifting time, or angle. For the ability to shift weight, however, both the perspective and time exerted significant main effects, F(1,22) = 6.429 and F(1,22) = 13.978, respectively.
Participants could shift their weight more effectively in the third-person perspective of the virtual environment. The results can serve as a reference for future designs of interactivevisual virtual environment as applied to rehabilitation.
Standing on a single leg by lifting the opposite leg is often used in clinical tests of balance,
including the one-leg-standing (OLS) test [
], the Berg Balance Scale [
], and choice stepping
reaction time [
]. Shorter OLS times or stepping reaction times indicate an increased risk of
]. Therefore, one-leg lifting training is crucial for balance.
Stepping exercises are common in traditional rehabilitation training because this type of
functional practice is considered beneficial for ambulation. For example, a therapist might
teach patients to step over obstacles that they might encounter when walking outside. A
sidewalk curb is approximately 20 cm high; however, because the height may vary, patients should
practice lifting their leg higher than this curb height. In addition to height, patients must walk
within a fixed time, as when crossing a street at a traffic light. The optimal training model must
entail real-life contingencies, which are difficult to recreate. Combining virtual reality (VR)
and motion capture technology can solve this problem by providing additional sensory
feedback and different custom-made environments. Numerous studies have confirmed the
feasibility and effectiveness of active video games, such as the functional recovery of upper and lower
extremities for post-stroke patients [
]. Although many therapists understood the
advantages of active video games, they reported difficulty finding time to learn how to use them [
]. Thus, a simple, quick platform must be developed.
Regarding game design, two considerations are paramount: perspective and parameters. The
two perspectives used in games are first-person (1PP) and third-person (3PP). In the 1PP
perspective, the player sees through the eyes of the character or object that he or she controls,
whereas in the 3PP, the player sees an avatar onscreen [
]. Prasad and Shiffrar [
] reported that
people can identify their own movement more easily in 3PP than in 1PP, despite having little
real-world experience seeing themselves in 3PP. In rehabilitation, therapists typically place a
mirror in front of patients to facilitate learning movements. The patients can observe their
movements concurrently and determine how they can move more easily [
]. Therefore, 3PP may be
more helpful for rehabilitation training. Additionally, certain parameters must be set. Koritnik
et al. [
] conducted a preliminary investigation to assess the feasibility of a virtual mirror system
for lower-extremity training. In that study, a healthy participant saw his or her own figure
superimposed on that of a virtual instructor, and attempted to match the instructor’s cadences and
knee-joint lifting heights. Ustinova et al. [
] created an avatar in a virtual environment (VE) to
analyze body movement and performance at different arm-reaching angles. Kinematics data were
collected, and the results showed that reach distance was greatest at angles between 45°and 77.5°,
without a corresponding loss of balance. Thus, manipulating kinematic or temporal parameters
in VR may provide useful information for game design or rehabilitation programming.
Patients recovering from stroke often experience difficulty when applying weight to their
hemiplegic leg. Traditionally, to facilitate shifting weight, patients are first asked to place their sound
foot on a small step in front of them. The height of the step is increased, and patients are asked
2 / 11
to lift their sound foot higher as their balance improves [
]. As mentioned, patients are trained
to step over a sidewalk curb. Based on the information processing model [
], they look at the
curb first (sensory input), determine how distant and how high the curb is (perceptual stage),
decide whether to step over the curb (decision making stage), shift their weight to one side
(programming stage), and then lift the other leg to step over the curb (response output). In
assessing movement performance, time and error are typically recorded. Time (speed) and
error (accuracy) are frequently the criteria most indicative of performance success.
Reaction time (RT) is a measurement of the time required for a person to perceive, process,
and initiate a movement in response to a stimulus. Movement time (MT) is the time between
the beginning of movement in response to a stimulus signal and the completion of the
response. Although RT entails no movement component, it entails the time prior to the
initiation of a movement in response to a stimulus. The combination of RT and MT is referred to as
response time (i.e., RT + MT = response time).
On the basis of the described theory, we designed a system that combines VR technology
with motion capture devices to evaluate and train a participant’s balance control.
In our previous work [
], we designed a three-dimensional posture frame (PF) and a human
avatar in a VE. The participants who wore motion capture devices interactively manipulated
an avatar to follow the fixed-shaped PF. As the PF approached the avatar in the scene, the
participants had to raise one leg to conform to the shape of the PF. The width and traveling time
of the PF were set as different types. The width denoted the distance between the PF boundary
and the avatar. The traveling time denoted the time of the PF that approached the avatar
following its appearance. Our preliminary results showed that the width and running times of the
PF design affected the participants’ RT. The RT of two participants was approximately 2
seconds with a wide PF and 5 seconds of traveling time.
We then designed two PFs: static and dynamic [
]. In contrast to the static PF, the
leg-raising angle and time can be manipulated in a dynamic PF. The participants were asked to lift one
leg for 1 or 2 seconds at 45°or 90°. We also used a force plate to detect the center of pressure
(COP) change to evaluate balance control. Four performance metrics were evaluated: mean
distance-anterior posterior (MDIST-AP), mean distance-medial lateral (MDIST-ML), sway
area (AREASW), and total excursions (TOTEX). Twenty-three healthy young individuals were
investigated. The results indicated that for the static PF, only MDIST-AP was affected by the
frame travel time (p = 0.000), whereas the other metrics were not related to the offset of the
frame or the frame travel time. For the dynamic PF, an impact of the leg-raising time on
MDIST-ML (p = 0.003) and TOTEX (p = 0.000) was observed.
In addition to the leg-lifting angles and times, we varied the perspectives in the present
study. We hypothesized that the participant would demonstrate improved movement
performance when an avatar was present in 3PP. Similar to traditional stepping exercises, small steps
(smaller leg-lifting angles) and longer leg-lifting times should make leg lifting easier. Thus, we
also hypothesized that the participant would exhibit improved movement performance at a
smaller lifting angle and for a longer lifting time. The present study explored the influence of
different perspectives on the movement performance in an interactive-visual VE.
A motion capture system (Animazoo-Gypsy Gyro 18, Synertial, UK) was used to capture the
participants’ movements. The system comprised 18 inertial rotational gyroscope sensors placed
3 / 11
Fig 1. Experimental virtual environment and the parameter description of measured values. Tp: The
time when the patients prepared for leg-lifting; Ts: The time when the patients actually started leg-lifting.
on the limbs and sampled them at 120 Hz. Data were delivered to a computer running motion
analysis software (SIMI) and instant animation software (Motion Builder 7.5). A virtual
immersive environment designed using 3D graphics software (CINEMA 4D) was developed,
and a PF with or without a real-time avatar was then shown on a projection screen (Fig 1).
A foot force plate (AMTI Advanced Mechanical Technology, Inc., Watertown, MA, USA)
under the participant’s feet detected the COP at a sampling frequency of 1000 Hz. The platform
was connected to the computer to collect COP data, which included the COP displacement
The participants stood 2.5 m in front of the projection screen and saw a dynamic PF in the VE
moving toward them. When the PF appeared after 2 seconds, the participants were asked to
perform one-leg lifting (right side) without making contact with the PF. In the 3PP group, the
participants saw a real-time avatar and directed the avatar through the PF. However, in the
1PP group, the participants saw only a PF and were asked to follow the PF independently (Fig
1). The PF was set for two lengths of time (0.7 and 2.7 seconds) and at two angles of knee
flexion (30°and 90°), totaling four experimental conditions. Each participant performed
one-leglifting movement under four experimental conditions, the order of which was determined
using 4 × 4 balanced Latin squares to avoid the learning effect (S1 Table). After three practice
trials, the mean values of the three test trials were recorded for analysis. A film of the
experiment can be seen here: http://youtu.be/zMxvWcnZrZo.
The outcome measured was movement performance. The outcome parameters were the
reaction time (ΔT1), accuracy of the movement (ΔT2), and ability to shift weight (Δd). The
operational definitions are as follows (Fig 1):
Tp: The time at which the participant prepared to lift the leg (i.e., COP displacement was
detected, but there was no actual leg-lifting movement) after the PF had begun to move,
representing the time at which the anticipatory postural reaction of the participant was observed.
Ts: The time at which the participant actually started lifting his or her leg.
ΔT1: = Ts−Tp, which represents the participant’s reaction time.
ΔT2: = |Ts –2|, which represents the error between the point when the participant actually
started leg lifting (Ts) and the point when the PF had appeared (2 seconds into the
4 / 11
experiment). Whether the original value is negative or positive, the absolute value represents
the accuracy of the movement. The closer the absolute value is to zero, the smaller the error
is. However, a negative original value (Ts –2) indicates that the participant started to lift his
or her leg before the appearance of the PF, meaning that the participant used anticipatory
postural adjustments (APAs) to complete the movement. By contrast, a positive value
indicates that the participant started to lift the leg after the appearance of the PF, meaning that
the participant used compensatory postural adjustments (CPAs). Thus, we also considered
the sign of (Ts− 2) in addition to its magnitude.
Δd: The COP X-Y displacement from the start to finish of leg lifting represents the ability to
shift weight. If the participant transferred his or her weight completely to the left leg when
lifting the right leg, the value of the COP displacement would be at a maximum.
A mixed factorial design (2 × 2 × 2) was applied, where the perspective was the
between-participant factor, and the leg-lifting times (0.7 and 2.7 seconds) and leg-lifting angles (30°and 90°)
were the within-participant factors.
The purpose of this study was to explore the influence of different parameters on the
movement performance in an interactive-visual VE. We developed the following three hypotheses
H1: The perspective factor (3PP vs. 1PP) influences posture performance in terms of RT (ΔT1),
accuracy of the movement (ΔT2), and ability to shift weight (Δd). We expect the 3PP group
to have shorter ΔT1, shorter ΔT2, and larger Δd than those of the 1PP group.
H2: The leg-lifting time factor influences posture performance in terms of RT (ΔT1), accuracy
of the movement (ΔT2), and ability to shift weight (Δd). We expect a longer lifting time of
the PF to cause the participants to have shorter ΔT1, shorter ΔT2, and larger Δd.
H3: The leg-lifting angle factor influences posture performance in terms of RT (ΔT1), accuracy
of the movement (ΔT2), and ability to shift weight (Δd). We expect a smaller lifting angle of
the PF to cause the participants to have shorter ΔT1, shorter ΔT2, and larger Δd.
Twenty-eight volunteers were recruited from a university. The Berg Balance Scale was used to
screen for movement disorders. Those with a history of central nervous system or lower limb
IV: independent variables; DV: dependent variables; ΔT1: reaction time; ΔT2: accuracy of the movement; Δd: CoP X-Y displacement of leg-lifting; 1PP:
®rstperson perspective; 3PP: third-person perspective.
5 / 11
muscle injury, and one volunteer unable to fully participate for other reasons, were excluded. A
total of 24 healthy participants (10 female and 14 male) were enrolled, with a mean age of 23.8
years (range 22–25 years). The participants were randomly divided into two groups (3PP
group vs. 1PP group). The study was approved by the Institutional Review Board of National
Taiwan University Hospital, and written informed consent was received from all the
participants. A sample size estimate of 11 participants per group (22 total) was calculated using
Gpower 3.1 (Franz Faul, University of Kiel, Germany) to provide 0.84 power to detect a medium
effect size of 0.4 [
], and through a repeated measures ANOVA, within-between interaction
with alpha = 0.05.
Data were calculated in statistical means and standard deviations from three trials for each
variable and each participant. First, one-way ANOVA on each of the four experimental conditions
was performed to determine whether the learning effect was avoided. Second, three-way
ANOVA (two within and one between factor) with repeated measures on leg-lifting times and
angles was performed to assess the interaction of the perspective, leg-lifting time, and angle,
and their main effects on RT, accuracy of the movement, and ability to shift weight [
analyses were performed using SPSS 15.0 from PASW (SPSS, Inc., 2009, Chicago, IL, USA,
http://www.spss.com/), and the significance level was set at p < 0.05.
In the one-way ANOVA, the order did not cause any significant effects (S2 Table), indicating
that the counterbalancing was successful. Fig 2 presents the distribution of each value for the
reaction time (ΔT1), accuracy of the movement (ΔT2), and COP X-Y displacement for leg
lifting (Δd) for both groups (1PP vs. 3PP). In the three-way ANOVA, no significant two- or
three-way interactions were noted (Table 2). Table 2 shows the reaction time (ΔT1) and
accuracy of the movement (ΔT2); there were no significant main effects of the avatar, time, or angle.
Fig 2. Distributions of the reaction time, the accuracy of the movement, and CoP X-Y displacement of
leg-lifting for both groups in different leg-lifting angles and speed. ΔT1: Reaction time; ΔT2: Accuracy of
the movement; Δd: CoP X-Y displacement of leg-lifting. Blue dot: 3PP group; red plot: 1PP group.
6 / 11
SS: Sum of squares; df: degrees of freedom; MS: Mean Square; F: F-value; p: p-value; ES: effect size; power: power of the F test
1PP: ®rst-person perspective; 3PP: third-person perspective
*p < 0.05
For the ability to shift weight (Δd), the perspective had a significant main effect, F(1,22) =
6.429, p = 0.019. The 3PP group had a larger average displacement than that of the 1PP group.
There was also a significant main effect of time, F(1,22) = 13.978, p = 0.001. In the 1PP group, a
longer leg-lifting time corresponded to a larger COP X-Y displacement (Fig 3). The
displacement was 70.5% larger for the 2.7-second time compared with the 0.7-second time. All values
were lower than those of the 3PP group.
The main goal of the current study was to explore the influence of the perspective on the
movement performance in an interactive-visual VE. We also sought to determine which parameters
of the PF could influence postural control. Our results show that the perspective in VR training
and the leg-lifting time of the PF can affect the ability to shift weight during a one-leg-lifting
7 / 11
Fig 3. CoP X-Y displacement of leg-lifting (Δd) in different leg-lifting times for both groups. Blue line:
3PP group; red line: 1PP group.
task, but not the participants’ RT or the accuracy of their movement. Our results can be applied
in clinical settings, such as balance or ambulation training.
Regarding this perspective, Salamin et al. [
] reported similar results. They investigated the
different effects of training participants in a ball-catching task between 3PP and 1PP
perspectives. The study analyzed response times and error rates during the training. In 3PP training,
the trend in error rates was similar to that in the normal daily life perspective, but not in 1PP
training. For the response times, however, no significant differences were found between the
perspectives. Regarding sports science, Covaci et al. [
] compared the motor performance
between the two perspectives in a basketball-throwing task. Their results showed that the ball
speed after 3PP training was more similar to that after real-life training than 1PP training.
Overall, the findings of these studies and our own suggest that the 3PP is more effective than
the 1PP in VR training for transferring motor skills to real practice. For developing the ability
to shift weight for a patient recovering from stroke, a VR training program in the 3PP appears
to be more effective for rehabilitation than one in the 1PP.
The results for the RT and accuracy of the movement were contrary to our hypothesis.
Many factors affected the RT [
]: stimulus types, intensity, anticipation, and distractions,
among many others. However, neither perspective nor the leg-lifting time or angle was
associated with the RT (ΔT1) in our study. Some factors that we did not control for, such as
psychological factors, may have influenced our results. For instance, children with motor impairment
typically depend excessively on visual feedback in controlling their movement [
may extend their reaction time. Thus, the RT may not have shown an association with different
perspectives because of the healthy sample in our study.
Moreover, no associations were observed between our independent variables and the
accuracy of the movement (ΔT2). However, we examined the data (Fig 2) and identified a
compelling trend: the value is negative in 3PP, whereas it is positive in 1PP. A negative mean value
8 / 11
indicates that the participants started to lift their leg before the 2-second mark when the PF
began to move; by contrast, a positive mean value indicates that the participants started to lift
their leg after the 2-second mark. Thus, in 3PP, the participants started to perform the
leg-lifting movement slightly before the motion of the PF, whereas in 1PP, the participants did so
immediately after the motion of the PF. On the basis of the results, we infer that the
participants used APAs in 3PP but used CPAs in 1PP. Therefore, the perspective may affect the
accuracy of a movement’s timing. Further research is necessary to clarify this point.
Another parameter that we manipulated is the leg-lifting time. The participants
demonstrated greater ability to shift weight with a longer leg-lifting time, which was consistent with
our hypothesis. Furthermore, Fig 3 shows that the participants exhibited greater ability to shift
weight in 3PP. To achieve training goals, the initial training program can incorporate longer
leg-lifting times. In other words, the treatment program for each patient must be specific and
can be adjusted according to the progress of the medical condition.
Galna et al. [
] developed a Kinect-based game to enhance balance control for people with
Parkinson disease. The game implemented multidirectional reaching and stepping tasks with
12 levels of difficulty, and an avatar was shown to the user. Most participants reported that
they felt safe and enjoyed the games, but that some tasks were difficult. Thus, to practically
apply active video games in clinical rehabilitation, designing the virtual environment in a
personalized manner must be considered. In our study, the design of the posture frame and an
avatar in 3PP can feasibly be personalized, but some parameters may have to be corrected.
As previously mentioned, we created a VE for application in clinical rehabilitation. Guo
et al. [
] investigated people with or without mobility impairment and found that they
respond differently to VE. In addition to objective gait and physiological response, they also
investigated subjective presence. Thus, future research is required for different groups, such as
older adults or patients with stroke; both subjective and objective measurement must also be
We hypothesized that, in an interactive-visual VE, the presence of an avatar as a visual aid and
different parameters (i.e., times and angles) of one-leg lifting would influence the postural
control of the participants. The participants demonstrated greater ability to shift weight when an
avatar was present in the VE. The experimental results can serve as a reference for future
designs of interactive VE for rehabilitation.
S1 Table. Sequence of four experimental conditions in 4×4 Balanced Latin squares. n:
numbers of subjects, A: 0.7second × 30°; B: 0.7 second × 90°; C: 2.7second × 30°; D: 2.7 second × 90°.
S2 Table. Repeated measure ANOVA table of four orders for reaction time (ΔT1), accuracy
of the movement (ΔT2), and COP X-Y displacement for leg lifting (Δd). SS: Sum of squares;
df: degrees of freedom; MS: Mean Square; F: F-value; p: p-value; ES: effect size; power: the
power of the F test. 1PP: first-person perspective; 3PP: third-person perspective.
9 / 11
Data curation: CHH.
Formal analysis: CHH.
Funding acquisition: TLS.
Project administration: TLS.
Supervision: TLS CP.
Writing – original draft: CHH TLS.
Writing – review & editing: CHH TLS.
10 / 11
1. Michikawa T , Nishiwaki Y , Takebayashi T , Toyama Y . One-leg standing test for elderly populations . Journal of orthopaedic science: official journal of the Japanese Orthopaedic Association . 2009 ; 14 ( 5 ): 675 ± 85 . Epub 2009/10/06. doi: 10 .1007/s00776-009-1371-6 PMID: 19802686 .
2. Berg KO , Wood-Dauphinee SL , Williams JI , Maki B . Measuring balance in the elderly: validation of an instrument . Canadian journal of public health = Revue canadienne de sante publique . 1992 ; 83 Suppl 2 : S7 ± 11 . Epub 1992/07/01. PMID: 1468055 .
3. Lord SR , Fitzpatrick RC . Choice stepping reaction time: a composite measure of falls risk in older people . The journals of gerontology Series A, Biological sciences and medical sciences . 2001 ; 56 ( 10 ): M627± 32 . Epub 2001/10/05. PMID: 11584035 .
4. Levin MF , Knaut LA , Magdalon EC , Subramanian S. Virtual reality environments to enhance upper limb functional recovery in patients with hemiparesis . Studies in health technology and informatics . 2009 ; 145 : 94 ± 108 . Epub 2009/07/14. PMID: 19592789 .
5. Lohse KR , Hilderman CG , Cheung KL , Tatla S , Van der Loos HF. Virtual reality therapy for adults poststroke: a systematic review and meta-analysis exploring virtual environments and commercial games in therapy . PloS one . 2014 ; 9 ( 3 ): e93318 . Epub 2014 /04/01. doi: 10 .1371/journal.pone.0093318 PMID: 24681826; PubMed Central PMCID : PMC3969329 .
6. Glegg SMN , Holsti L , Velikonja D , Ansley B , Brum C , Sartor D . Factors influencing therapists' adoption of virtual reality for brain injury rehabilitation . Cyberpsychology, Behavior, and Social Networking . 2013 ; 16 ( 5 ): 385 ± 401 . doi: 10 .1089/cyber. 2013 .1506
7. Levac DE , Miller PA . Integrating virtual reality video games into practice: Clinicians' experiences . Physiotherapy Theory and Practice . 2013 ; 29 ( 7 ): 504 ± 12 . doi: 10 .3109/09593985. 2012 .762078 PMID: 23362843
8. Novak J . Game Development Essentials: An Introduction: Cengage Learning ; 2011 .
9. Prasad S , Shiffrar M. Viewpoint and the Recognition of People From Their Movements . J Exp Psychol Hum Percept Perform . 2009 ; 35 ( 1 ): 39 ± 49 . doi: 10 .1037/a0012728 PMID: 19170469
10. Shumway-Cook A , Woollacott MH . Motor Control: Theory and Practical Applications: Lippincott Williams & Wilkins; 2001 .
11. Koritnik T , Bajd T , Munih M. Virtual environment for lower-extremities training . Gait & posture . 2008 ; 27 ( 2 ): 323 ± 30 . Epub 2007/06/29. doi: 10 .1016/j.gaitpost. 2007 . 04 .015 PMID: 17596945 .
12. Ustinova KI , Perkins J , Szostakowski L , Tamkei LS , Leonard WA . Effect of viewing angle on arm reaching while standing in a virtual environment: potential for virtual rehabilitation . Acta psychologica . 2010 ; 133 ( 2 ): 180 ± 90 . Epub 2009/12/22. doi: 10 .1016/j.actpsy. 2009 . 11 .006 PMID: 20021998 .
13. Davies PM , Br uÈhwiller DJ , Kesselring J , Gierig R . Steps to Follow: The Comprehensive Treatment of Patients with Hemiplegia: Springer Berlin Heidelberg; 2012 .
14. Edwards WH . Motor Learning and Control: From Theory to Practice: Cengage Learning; 2010 .
15. Lai CL , Shiuan-Yu T , Huang CH , Pei C , Chi WM , Liang-Ching H , et al. Fun and Accurate Static Balance Training to Enhance Fall Prevention Ability of Aged Adults: A Preliminary Study . Human Factors & Ergonomics in Manufacturing & Service Industries . 2013 ; 23 ( 6 ): 517 ± 27 .
16. Sun TL , Lee CH . An impact study of the design of exergaming parameters on body intensity from objective and gameplay-based player experience perspectives, based on balance training exergame . PloS one . 2013 ; 8 ( 7 ): e69471 . Epub 2013 /08/08. doi: 10 .1371/journal.pone.0069471 PMID: 23922716; PubMed Central PMCID : PMC3724862 .
17. Cohen J. Statistical Power Analysis for the Behavioral Sciences: L. Erlbaum Associates; 1988 .
18. Portney LG , Watkins MP . Foundations of Clinical Research: Applications to Practice: Prentice Hall Health; 2000 .
19. Salamin P , Tadi T , Blanke O , Vexo F , Thalmann D. Quantifying effects of exposure to the third and first-person perspectives in virtual-reality-based training . IEEE Trans Learn Technol . 2010 ; 3(3):272± 6 . doi: 10 .1109/tlt. 2010 .13
20. Covaci A , Olivier AH , Multon F , editors. Third person view and guidance for more natural motor behaviour in immersive basketball playing . 20th ACM Symposium on Virtual Reality Software and Technology , VRST 2014 ; 2014: Association for Computing Machinery .
21. Schmidt RA . Motor Learning & Performance: From Principles to Practice: Human Kinetics Books; 1991 .
22. van der Meulen JH , Denier van der Gon JJ , Gielen CC , Gooskens RH , Willemse J. Visuomotor performance of normal and clumsy children. I: Fast goal-directed arm-movements with and without visual feedback . Developmental medicine and child neurology . 1991 ; 33 ( 1 ): 40 ± 54 . Epub 1991/01/01. PMID: 1995409 .
23. Wilson PH , McKenzie BE . Information processing deficits associated with developmental coordination disorder: a meta-analysis of research findings . Journal of child psychology and psychiatry, and allied disciplines . 1998 ; 39 ( 6 ): 829 ± 40 . Epub 1998/10/03. PMID: 9758192 .
24. Galna B , Jackson D , Schofield G , McNaney R , Webster M , Barry G , et al. Retraining function in people with Parkinson's disease using the Microsoft kinect: game design and pilot testing . Journal of neuroengineering and rehabilitation . 2014 ; 11 ( 1 ): 60 .
25. Guo R , Samaraweera G , Quarles J , editors. The effects of VEs on mobility impaired users: Presence, gait, and physiological response . 19th ACM Symposium on Virtual Reality Software and Technology , VRST 2013 ; 2013 ; Singapore.