Content uploaded by Jing Mu
Author content
All content in this area was uploaded by Jing Mu on Dec 16, 2024
Content may be subject to copyright.
1
1Department of Biomedical Engineering, The University of Melbourne, Melbourne, Australia
2Virtetic Pty Ltd, Melbo urne, Austr alia
3Department of Mechanical Engineering, The University of Melbourne, Melbourne, Australia
4Otto Bock Healthcare Products GmbH, Vienna, Austria
5The Alfred Hospital, Melbourne, Australia
6Royal Melbourne Hospital, Melbourne, Australia
7Department of Surgery, St Vincent’s Hospital, The University of Melbourne, Melbourne, Australia
Feasibility Study of a Game-Based Virtual Reality
Intervention for Functional Prosthesis Use Training:
A Pre-Clinical Assessment
Jing Mu1,2, Ricardo Garcia-Rosas2,3, Raphael Maria Mayer2,3, Daniel Meise4, Jim Lavranos5, Mark Graf6,
Ying Tan3, Denny Oetomo3, and Peter Choong7
Abstract
Introduction: This paper presents a pre-clinical feasibility study to investigate the efficacy of a game–based
virtual reality (VR) intervention in the functional training of upper-limb prosthesis use.
Materials and Methods: The study compared the skill of two able-bodied groups (intervention and control,
5 participants in each group) using a surface electromyography (sEMG) controlled self-experience prosthetic
hand and wrist through the Box and Blocks Test (BBT). The Intervention Group followed the game-based
VR intervention for four weeks and underwent real-life BBT assessments before and after the intervention
using the self-experience prosthesis. The Control Group performed the real-life BBT assessments four weeks
apart with no VR intervention in between. As a follow-on study, the Control Group then underwent the VR
intervention for further validation. The VR intervention consisted of two games designed to train the control
and use of an sEMG-controlled prosthetic hand and an assessment (VR BBT). The VR intervention protocol
included three 30-minute sessions per week over a period of four weeks.
Results: A significantly larger improvement was shown in the BBT scores from the intervention group (mean
6.375) compared to the control group (1.5). The follow-on study further confirmed the result.
Conclusion: The study demonstrated promising outcomes in the use of a game-based virtual reality
intervention for training functional prosthesis use, with participants showing significant improvements in
prosthesis use skills measured with real-life BBT. This motivates further development of VR interventions
for prosthesis use training and their validation through clinical trials.
Keywords
Prosthesis, functional training, virtual reality, feasibility study, skill transfer
I. INTRODUCTION
Training and practicing the use of a prosthetic hand are key to the effective usage and the adoption
of a prosthesis. Amputees need to practice the use of their prosthesis in real-world context to effectively use
it in daily life. However, practice with a prosthesis is only available to people several months post-
2
amputation. This is due to the need to wait for the amputation wound to heal, for the muscle mass of the
stump to stabilise, and for a socket and the prosthesis to be fabricated and made available1.
During this waiting period, people have few opportunities to make use of the affected limb and the
remaining muscles in the stump. Existing options are traditional rehabilitation or occupational therapy
practices, such as repeated motion, mirror therapy2, graded motor imagery3, and the use of myoelectric
sensors to practice muscle activations in the stump4. The key challenges with these existing practices are the
repetitiveness of the exercises (limited user engagement) and the lack of relevance to daily life prosthesis
use (disconnected function). None of these allow the user to practice the 3-dimensional (3D) movements
required to use a prosthesis in a real setting, the interactions between the prosthesis and the environment,
and do not highlight the limitations of prosthetic devices.
Virtual reality (VR) offers the possibility to address this key shortcoming, by allowing a virtual
prosthesis to be used in a 3D environment and to interact with virtual objects. Game-based interventions and
VR have shown significant promise in the world of rehabilitation due to the engaging nature of the
interventions and their potential relevance to daily life use5-8. In the world of prosthetics, VR has mostly
been used for research purposes9-13; and most recently, it has gathered interest as a potential intervention for
prosthesis use training14-16 and phantom limb pain management17,18. However, the research that has been
conducted on VR training for upper limb prostheses thus far has focused on training muscle activations and
the use of myoelectric sensors19, as opposed to functional prosthesis use training for activities of daily living.
This paper presents a pre-clinical feasibility study for a VR intervention for functional prosthesis
use training co-designed with clinicians and users. The proposed VR intervention allows users to practice
the functional use of a prosthesis, i.e., interacting with objects in ways relevant to daily tasks in 3D space
using a gamified approach. The study seeks to contribute to the question: does VR training translate to
improvements in real-life functions with a prosthesis? We have hypothesised that users’ real-life prosthesis
3
use skills could be improved through game-based VR interventions. The outcomes of this study will serve
as the foundation of game-based VR intervention for functional prosthesis use training, guide future
clinical studies, and facilitate purposeful developments in VR interventions to complement existing
solutions.
II. METHODOLOGY
A. Participants
Ten able-bodied participants, 4 females and 6 males, were recruited in the study (aged 25 to 35 years,
median age 27) through online advertisements.
None of the participants had previous experience using a self-experience prosthetic hand.
Participants had varying experience with VR, from none to casual users. Participants were randomly
assigned to the Control and Intervention Groups, with participant gender balanced (2 females, 3 males each).
This study was approved by the University of Melbourne Human Research Ethics Committee, project
number 23319. Written informed consent was obtained from all participants in the study before participation.
B. Hardware Setup
The hardware used in the study (shown in Figure 1a-b) includes a self-experience prosthesis, a
standalone VR system, and a surface electromyography (sEMG) system to supply control commands to both
the self-experience prosthesis and the VR system. And a Box and Blocks Test (BBT)20, a functional test to
assess the unilateral upper limb motor skills.
1) Self-Experience Prosthesis: For this paper, a self-experience prosthesis is a prosthesis
attached to the limb of an able-bodied participant, allowing them to experience using a prosthesis. The self-
experience prosthesis consists of an Xlimb hand prosthesis21 configured to allow pinch-grasp (open, close)
4
and a wrist allowing for rotation (pronation, supination). Straps are used to mount the self-experience
prosthesis to the user’s forearm as shown in Figure 1a. A left-hand and a right-hand self-experience
prosthesis were used in the experiment. Participants used the one matches to their dominant hand throughout
the study.
2) VR System: Meta Quest 2 (Meta Platforms, Inc., USA) was used in this study to deliver the
VR game-based intervention. A Meta Quest 2 controller was used for tracking the forearm of the user in the
virtual environment as shown in Figure 1b.
3) Surface Electromyography (sEMG): The MyoCuff sEMG pattern recognition system
(Ottobock SE & Co. KGaA, Germany) was used to collect muscle activations in the forearm for prosthesis
control in both real life and VR. Four gestures were used in the experiment: hand open, hand close, wrist
flexion, and wrist extension (Figure 1 c-f).
A Bluetooth dongle (Virtetic Pty Ltd) is mounted onto the Myo Cuff to send classification output
to the self-experience prosthesis and VR system (Figure 1b).
C. Virtual Reality Intervention
The VR experience simulates a prosthetic hand and its interactions with the virtual environment.
SEMG patterns are mapped to functions in the virtual prosthesis, such as hand open/close and wrist
rotations.
The virtual reality intervention consists of two games and an assessment. The goal of the games is
to help people practice skills required for prosthesis control and functional use of a prosthesis in daily life.
The goal of the assessment is to help people assess their skills through a standardised test in VR.
5
1) Game 1: Paintball Town: In this game, participants exercise their skills in interacting with
the environment with proportionally controlled actions. Participants were tasked with shooting paint balls
at targets of two different colours that were placed around the environment at different heights and distances
(Figure 2a). Participants regulate the strength of shooting the paint balls with the “hand close” pattern and
switch the paint ball colour to match the target colour with the “pronation” and “supination” patterns. Game
score was calculated based on number of targets hit successfully and the difficulty of the targets.
2) Game 2: Pick & Cook: In this game, participants exercise their skills in pick-and-place tasks
with assisted proportionally controlled prosthesis in the context of BBQ (Figure 2b). The ingredients were
placed around the environment at different heights. The BBQ grill and plates were place at waist height.
Hand closing and opening were proportionally controlled by the “hand close” and “hand open” patterns,
respectively. The grasping of ingredients was assisted by the object attaching to the prosthetic hand when
70% of the hand closure was achieved, and vice-versa. Wrist pronation and supination were proportionally
controlled by the “wrist flexion” and “wrist extension” patterns, respectively. Score was calculated based
on the number of dishes finished and the number of ingredients correctly cooked to perfection.
Assessment: VR Box and Blocks Test: In this challenge, participants were tasked with completing
a standard BBT in VR (Figure 2c). This assessment requires participants to move one block at a time from
one side of the box to the other. The score of the assessment is the number of blocks moved in one minute.
The box was place at waist height. Hand closing and opening were proportionally controlled by the “hand
close” and “hand open” patterns, respectively. No grasping assistance was provided for this challenge,
meaning that the blocks responded to physics-based grasping. Wrist pronation and supination were
proportionally controlled by the “wrist flexion” and “wrist extension” patterns.
D. Experimental Protocol
6
The entire study consists of two parts, the Experiment Period and the Validation Period as shown
in Figure 3. Both Intervention Group and Control Group participated in the Experiment Period, but only
the Control Group was asked to also complete the Validation Period to validate the random assignment of
participants. Each period is four weeks long. Hence, the study length was four weeks for the Intervention
Group and eight weeks for the Control Group.
All participants engaged in this study with their dominant hand as the side receiving training and
being assessed. Participants do not have access to either the VR intervention or the self-experience
prosthesis at all other times during their engagement in this project.
1) Experiment Period: During the Experiment Period, the Intervention Group was asked to
attend 3 sessions per week on 3 separate days (12 sessions in total). The Control Group was only asked to
attend 2 sessions 4 weeks apart.
At the start of each session, sEMG was calibrated with 4 different patterns (hand open, hand close,
wrist flexion, wrist extension) in 3 arm positions (arm hanging, 90 deg elbow bend, arm straight forward).
Participants were assessed with real-life BBT with the self-experience prosthesis in the first and
the last day (the only two sessions for Control Group) in the Experiment Period. The standard 60-second
BBT was performed 2 times in each real-life assessment.
The Intervention Group also underwent the VR intervention in all sessions. In each session, the
VR intervention followed the sequence as shown in Table I, with Game A and B being randomised
between Paintball Town or Pick & Cook for each participant.
2) Validation Period: During the Validation Period, the Control Group was asked to complete
3 sessions per week on 3 separate days for 4 weeks, same as what the Intervention Group went through in
the Experiment Period as described in Table I.
E. Data Collection and Analysis
7
The scores of the real-life BBTs were recorded for the real-life assessments. In VR, the scores for
the games and assessment in timed challenges (see Table I) were collected.
The participants’ skills were measured by the real-life BBT. The average scores from the two BBT
repeats in each real-life assessment were used in the analysis. A Jarque-Bera test was first performed (with
Matlab command ‘jbtest’) to test if the data fits in a normal distribution. The data from both groups in the
Experiment Period was first compared to investigate the effectiveness of the VR intervention. Then the data
from Control Group in both the Experiment Period and the Validation Period was compared to validate the
random assignment of participants and further demonstrate the effectiveness of the VR intervention. BBT
scores from both real-life assessments and VR BBT were also compared. And participants’ performance
progression in the two games and the VR BBT assessment was also studied.
III. RESULTS
A. Effectiveness of the VR Intervention
The effectiveness of the VR intervention is measured by the improvements in participants’ scores in
the real-life BBT in the Experiment Period. Figure 4a-c shows the changes and improvements in participants’
scores in both groups before (Day 1) and after (Day 12) the four-week Experiment Period. Figure 4a-b show
that the Intervention Group has a larger improvement in the BBT score compared to the Control Group, as
can be seen from the slope of the dark red lines. From Figure 4b we can see that there is one participant
scoring 1 and 1 in both before and after tests, and is considered as an outlier in the Intervention Group.
Figure 4c directly compares the improvements taken as the difference between each participant’s before and
after scores. Since the data comes from two independent groups, a two-sample t-test was performed to test
if the two data are likely to come from the same distribution. The result from a two-sample t-test (with
Matlab command ‘ttest2’) showed statistically significant difference between the two groups (with outlier
8
removed) at 5% level, with the average improvements in the Control Group and Intervention Group being
1.5 and 6.375, respectively (four-fold increase with VR intervention). The results show that the VR
intervention significantly increased participants’ improvements in functional prosthesis use skills.
B. Control Group Validation
Figure 4d compares the performances of the Control Group in the Experiment Period (Day 1 and
Day 12) where there was no intervention between the two real-life BBTs and in the Validation Period (Day
13 and Day 24) where they underwent VR intervention between the two real-life BBTs. In Figure 4d, a clear
difference in the slope of change in score can be seen in the Experiment Period (Days 1-12, no VR
intervention) and validation period (Days 13-24, with VR intervention) of the study. The Validation Period
with VR interventions displays a larger increase in performance compared to the Experiment Period where
there were no VR interventions. Figure 4e plots the improvements in the two performances of the Control
Group in box plots, labelled as “No VR” for the performance in the Experiment Period where the Control
Group underwent no intervention, and “VR” for the Validation Period where the Control Group underwent
VR intervention. Since the data is from the same group obtained under different conditions, a paired-sample
t-test was performed to test if the difference between the two data is likely to form a distribution with zero
mean. Paired-sample t-test (with Matlab command ‘ttest’) showed a significant difference in the
improvements between the two periods at 5% significance level. The results confirm that the VR
intervention significantly increases users’ improvements in functional skills.
C. Comparison Between Real-Life (RL) and VR BBT Results
Comparison is done between the Real-Life (RL) and Virtual-Reality (VR) BBT results. The real-
life (RL) and VR BBT results were extracted from both groups when both RL and VR assessments were
conducted, i.e., on Day 1 and 12 for the Intervention Group and Day 13 and 24 for the Control Group. The
9
two attempts in RL BBT were averaged, and the VR BBT conducted right after the RL test was used for
comparison. The relationship between the RL and VR BBT scores are plotted in Figure 5a. Linear
regression was performed with the 20 data points and the resulting linear fitted line is shown in red (solid
line) along with the 95% confidence bounds (dashed red lines). The equation of the linear fit and the R2
value are shown in the text box. The results show a positive correlation between the real-life and VR BBT
scores with a R2 value of 0.50.
D. Performance Progression in VR
Figure 5b-d shows the average scores (blue lines) participants achieved in VR games on the 12
days they engaged in VR. For each day, each participant’s score was obtained by averaging the scores
from the two-timed attempts. Standard error of the mean was also calculated with the participants’ scores
on each day (shaded areas). In all three games and challenge, increasing trends can be seen as the
intervention progresses and the average scores on day 12 at least doubled the scores on day 1.
IV. DISCUSSION
A. Effectiveness of the VR Intervention in Functional Prosthesis Use Training
As can be seen in Figure 4a-c, the Intervention Group showed a significantly larger improvement in
BBT scores compared to the Control Group in the experiment stage, which suggests that the exercised VR
interventions can effectively help the users in achieving better functional prosthesis use outcomes.
The effectiveness of the VR intervention is further demonstrated by the Control Group when
undergoing the VR intervention in the Validation Period (Figure 4d-e). The improvements in the BBT scores
in the Validation Period is significantly higher than the change in the Experiment Period.
10
These results suggest that the VR intervention may help users in improving their skills in operating
myoelectric prosthetic hands. The results from the Validation Period in the Control Group also confirms
the capability of the Control Group and validates that there is no bias in the random assignment of
participants.
These results agree with other studies on the feasibility of using virtual reality for training the
control of a prosthetic device6,7,10,22, suggesting that VR may be an effective method to provide pre-
prosthetic training to amputees.
B. VR BBT
The comparison between the BBT scores obtained from real-life (with self-experience prosthesis and
test equipment) and VR (virtual equipment and environment) settings, as shown in Figure 5a, showed a
strong positive correlation. The linear regression fitted line has a slope smaller than 1, which indicates that
the VR BBT might be a more sensitive measure of the person’s skill level than the real-life BBT. The R2
value shows that the VR BBT scores predict 50% of the variations in real-life BBT scores.
These observations suggest that VR BBT could be used in assessing an individual's capability
especially when it is difficult to set up a physical test environment. VR BBT may also provide a higher
resolution in the assessment. However, improvements could be made to the simulation to increase the
reliability of the VR BBT in predicting real-life BBT scores.
C. User Learning During VR Intervention
The progression of users’ learning in VR can be observed from Figure 5b-d. Notably, all three games
and challenges exhibit a clear upward trend in performance. Initially, users demonstrate rapid learning, as
evidenced by the early session scores in the games, which then stabilise in later sessions (Figures 5cd).
Additionally, the VR BBT scores show a consistent and gradual increase throughout the intervention (Figure
11
5b). These findings highlight the active learning that occurs during the VR intervention and provide evidence
that the skills acquired in VR games directly translate to improved prosthesis use.
D. Future Work
With the positive results in mapping VR to real-life BBT scores, the VR platform presented in this
work has demonstrated the feasibility in replicating standardised functional assessments. For future work,
other standardised functional assessments such as the Jebsen-Taylor Hand Function Test23 can be added to
the VR platform. More outcome measures, such as user satisfaction on the VR-based trainings and the ease
of use of the technology, should also be included.
V. CONCLUSION
This paper presented a pre-clinical feasibility study investigating the effect of VR interventions for
functional prosthesis use training on users’ skill improvement. The results demonstrated that functional
VR interventions could effectively help users in improving their skills in using myoelectric prosthesis. In
addition, it was also shown that VR-based BBT could be an alternative skill assessment method when a
physical prosthesis or BBT is not available.
12
REFERENCES
1.Braza DW, Martin JNY. Chapter 119 - Upper Limb Amputations. Frontera WR, Silver JK, Rizzo TD,
eds. ScienceDirect. Published online January 1, 2020:651-657.
https://www.sciencedirect.com/science/article/abs/pii/B978032354947900119X
2.Ezendam D, Bongers R, Jannink M. Systematic review of the effectiveness of mirror therapy in upper
extremity function. Disability & Rehabilitation. Published online 2009:1-15.
doi:https://doi.org/10.1080/09638280902887768
3.Priganc VW, Stralka SW. Graded Motor Imagery. Journal of Hand Therapy. 2011;24(2):164-169.
doi:https://doi.org/10.1016/j.jht.2010.11.002
4.Ottobock. Myo plus Pattern Recognition in Practice Therapy and Training Brochure for Fast and Easy
Use Information for Qualified Personnel.; 2020. Accessed October 20, 2023.
https://shop.ottobock.us/store/medias/2945709MyoPlusTherapyandTrainingBrochure.pdf?context=bWFzd
GVyfHJvb3R8ODY4MzEyfGFwcGxpY2F0aW9uL3BkZnxoMzcvaDBmLzg4OTI2NjQzMTU5MzQucG
RmfDdiODY3NjcxOGMyZDAwMmExYWVkZTgxZWVmMjI0NjQ2OGE3MjU0NWEzMjM5ZTE4Zm
Y5OGQyN2JjYzcyMmM0YTY
5.Hashim NA, Razak NAA, Osman NAA. Comparison of Conventional and Virtual Reality Box and
Blocks Tests in Upper Limb Amputees: A Case-Control Study. IEEE Access. 2021;9:76983-76990.
doi:https://doi.org/10.1109/access.2021.3072988
6.Hargrove L, Miller L, Turner K, Kuiken T. Control within a virtual environment is correlated to
functional outcomes when using a physical prosthesis. Journal of NeuroEngineering and Rehabilitation.
2018;15(S1). doi:https://doi.org/10.1186/s12984-018-0402-y
7.Hashim NA, Abd Razak NA, Gholizadeh H, Abu Osman NA. Video Game–Based Rehabilitation
Approach for Individuals Who Have Undergone Upper Limb Amputation: Case-Control Study. JMIR
Serious Games. 2021;9(1):e17017. doi:https://doi.org/10.2196/17017
8.Nissler C, Nowak M, Connan M, et al. VITA—an everyday virtual reality setup for prosthetics and
upper-limb rehabilitation. Journal of Neural Engineering. 2019;16(2):026039.
doi:https://doi.org/10.1088/1741-2552/aaf35f
13
9.Michalski SC, Szpak A, Saredakis D, Ross TJ, Billinghurst M, Loetscher T. Getting your game on:
Using virtual reality to improve real table tennis skills. Capio CM, ed. PLOS ONE. 2019;14(9):e0222351.
doi:https://doi.org/10.1371/journal.pone.0222351
10.Hunt CL, Sun Y, Wang S, et al. Limb loading enhances skill transfer between augmented and physical
reality tasks during limb loss rehabilitation. Journal of NeuroEngineering and Rehabilitation. 2023;20(1).
doi:https://doi.org/10.1186/s12984-023-01136-5
11.Dhawan D, Barlow M, Lakshika E. Prosthetic Rehabilitation Training in Virtual Reality. In: 2019 IEEE
7th International Conference on Serious Games and Applications for Health (SeGAH). IEEE; 2019:1-8.
doi:https://doi.org/10.1109/SeGAH.2019.8882455
12.Kisiel-Sajewicz K, Marusiak J, Rojas-Martínez M, et al. High-density surface electromyography maps
after computer-aided training in individual with congenital transverse deficiency: a case study. BMC
Musculoskeletal Disorders. 2020;21(1). doi:https://doi.org/10.1186/s12891-020-03694-4
13.Garcia-Rosas R, Tan Y, Oetomo D, Manzie C, Choong P. Personalized Online Adaptation of
Kinematic Synergies for Human-Prosthesis Interfaces. IEEE Transactions on Cybernetics.
2021;51(2):1070-1084. doi:https://doi.org/10.1109/tcyb.2019.2920376
14.Boschmann A, Dosen S, Werner A, Raies A, Farina D. A novel immersive augmented reality system
for prosthesis training and assessment. ieeexplore.ieee.org. Published 2016. Accessed October 20, 2023.
https://ieeexplore.ieee.org/abstract/document/7455889
15.Woodward RB, Hargrove LJ. Adapting myoelectric control in real-time using a virtual environment.
Journal of NeuroEngineering and Rehabilitation. 2019;16(1). doi:https://doi.org/10.1186/s12984-019-
0480-5
16.Sharma A, Niu W, Hunt CL, Levay G, Kaliki R, Thakor NV. Augmented Reality Prosthesis Training
Setup for Motor Skill Enhancement. arXiv.org. Published March 5, 2019. https://arxiv.org/abs/1903.01968
17.Ambron E, Buxbaum LJ, Miller A, Stoll H, Kuchenbecker KJ, Coslett HB. Virtual Reality Treatment
Displaying the Missing Leg Improves Phantom Limb Pain: A Small Clinical Trial. Neurorehabilitation and
Neural Repair. 2021;35(12):1100-1111. doi:https://doi.org/10.1177/15459683211054164
14
18.Tong X, Wang X, Cai Y, et al. “I Dreamed of My Hands and Arms Moving Again”: A Case Series
Investigating the Effect of Immersive Virtual Reality on Phantom Limb Pain Alleviation. Frontiers in
Neurology. 2020;11. doi:https://doi.org/10.3389/fneur.2020.00876
19.Boschmann A, Neuhaus D, Vogt S, Kaltschmidt C, Platzner M, Dosen S. Immersive augmented reality
system for the training of pattern classification control with a myoelectric prosthesis. Journal of
NeuroEngineering and Rehabilitation. 2021;18(1). doi:https://doi.org/10.1186/s12984-021-00822-6
20.S. R. A. Lab. Box and Block Test. Shirley Ryan AbilityLab. Published April 26, 2012.
https://www.sralab.org/rehabilitation-measures/box-and-block-test
21.Mohammadi A, Lavranos J, Zhou H, et al. A practical 3D-printed soft robotic prosthetic hand with
multi-articulating capabilities. PLOS ONE. 2020;15(5):e0232766.
doi:https://doi.org/10.1371/journal.pone.0232766
22.Nissler C, Nowak M, Connan M, et al. VITA—an everyday virtual reality setup for prosthetics and
upper-limb rehabilitation. Journal of Neural Engineering. 2019;16(2):026039.
doi:https://doi.org/10.1088/1741-2552/aaf35f
23. Jebsen R, Taylor N, Trieschmann R, Trotter M, Howard L. An objective and standardized test of hand
function. Arch Phys Med Rehabil. 1969; 50:311–319.
15
Figure 1: Experimental hardware set-up. (a) The self-experience prosthesis consists of an Xlimb hand
prosthesis (H), a wrist rotator (W), straps (S) to attach it to the user’s forearm, and an sEMG (E). (b) The
VR System uses a VR Headset (V) and its motion tracking controller (C), a Bluetooth dongle (D), and an
sEMG sensor (E). Figure c-f shows the four gestures (c) hand open, (d) hand close, (e) wrist flexion, and (f)
wrist extension which are trained in 3 arm positions before each session: arm hanging, 90 deg elbow bend,
and arm straight forward.
(a) (b)
(c) (d) (e) (f)
16
Figure 2: Games for the Study (a) 1: Paintball Town (b) 2: Pick & Cook and Challenge (c) VR BBT
Figure 3: Timeline of intervention showing the Experiment Period and the validation period.
(a) (b) (c)
Experiment Period
Real-Life Box & Blocks
Control Group
Day 1 Day 2-11 Day 12
Real-Life Box & Blocks No Intervention
VR Intervention
Intervention Group
Validation Period
Day 13 Day 14-23 Day 24
VR Intervention
Real-Life Box & Blocks
VR Intervention
Real-Life Box & Blocks
Real-Life Box & Blocks
VR Intervention
VR Intervention
Real-Life Box & Blocks
VR Intervention
17
Figure 4: Performance comparison between Control Group and Intervention Group in the Experiment Period
(a-c). Gray dashed lines plot the performance change in Box and Blocks Test score for each participant
before (Day 1) and after (Day 12) the four-week period; dark red lines show average; light blue shaded areas
show one standard deviation. Red lines in box plots mark median; boxes show 25-75 percentiles; whiskers
mark the minimum and maximum values; cyan dots label the improvement of each participant. Outlier at 0
in Intervention Group is labelled with red dashed line in (b) and red plus in (c). ∗† labels statistical
significance at 5% when the outlier is removed. Performance of Control Group with no VR intervention and
with VR intervention (d-e). Dashed lines plot the performance change in Box and Blocks Test score for each
participant before and after the four-week period both without VR intervention and with VR intervention,
participants are shown in different colours; dark red lines show average; light blue shaded areas show one
standard deviation. Red lines in box plots mark median; boxes show 25-75 percentiles; whiskers mark the
minimum and maximum values; cyan dots label the improvement of each participant. ∗ labels statistical
significance at 5%.
Control Intervention
0
2
4
6
8
10
12
Improvements
∗†
No VR VR
0
2
4
6
8
10
12
Improvements
∗
(a) (b) (c)
(d) (e)
18
Figure 5: (a) Scores achieved by the participants in Box and Blocks Test both in VR and using real-life
setup. The 20 data points are collected from both before and after VR intervention in both groups (only
validation period is considered in Control Group as there were no VR component in the Experiment Period
for this group). Solid red line shows the linear regression fitted line. Dashed red lines mark 95% confidence
bounds. The Average participant performance in VR: (b) VR BBT, (c) Paintball Town, (d) Pick & Cook.
Blue lines show the average score achieved by participants from both groups in different VR games. Shaded
areas show standard error of the mean.
0 2 4 6 8 10 12 14 16 18
Virtual Reality
0
2
4
6
8
10
12
14
16
18
Real Life
Data
Fit
Confidence bounds
y = 0.6094*x + 2.4486
R2 = 0.50
(a) (b)
(c) (d)
19
TABLE I: VR intervention protocol for each session.
Mode
Activity
Duration
Timed challenge
VR BBT
1 minute
Timed challenge
Game A
3 minutes
Timed challenge
Game B
3 minutes
Rest
1 minute
Practice
Game A
5 minutes
Practice
Game B
5 minutes
Practice
VR BBT
5 minutes
Rest
1 minute
Timed challenge
Game A
3 minutes
Timed challenge
Game B
3 minutes
Timed challenge
VR BBT
1 minute
