Functional assessment during physical rehabilitation exercises using serious games

Functional assessment during physical rehabilitation exercises using serious games

Thesis presented by Bruno BONNECHERE

with a view to obtaining the PhD Degree in Doctor in Biomedical and Pharmaceutical Sciences (ULB - “Docteur en Sciences Biomédicales et Pharmaceutiques”) and in Doctor in Engineering Sciences (VUB) Academic year 2018-2019

Supervisors: Professor Serge VAN SINT JAN (Université libre de Bruxelles)

Laboratory of Anatomy, Biomechanics and Organogenesis

and Professor Bart JANSEN (Vrije Universiteit Brussel)

Department of Electronics and Informatics - ETRO

J U R Y M E M B E R S

• Prof. C. Melot (chairman), Université Libre de Bruxelles

• Prof. R. Pintelon (vice-chairman), Vrije Universiteit Brussel

• Prof. J. Deutsch, Rutgers University, The state university of New Jersey

• Prof. P. Feys, UHasselt

• Prof. B. Dachy, Université Libre de Bruxelles

• Prof. E. Swinnen, Vrije Universiteit Brussel

• Prof. J. Cornelis, Vrije Universiteit Brussel

• Prof. S. Van Sint Jan (promotor), Université Libre de Bruxelles

• Prof B. Jansen (promotor), Vrije Universiteit Brussel

i

Education is not something you can finish.

- Isaac Asimov

iii

A C K N O W L E D G E M E N T

Firstly, I would like to thank my supervisors Prof. Serge Van Sint Jan and Bart Jansen for the continuous support of my Ph.D study and related research.

Besides my supervisors, I would like to thank the rest of my thesis committee: Prof. Véronique Feipel and David Gall, for their insightful comments and encouragement.

This thesis would also not have been possible without the hard work of two of my colleagues: Prof. Victor Sholukha, his extensive mathematical skills, knowledge and work dedication were particu- larly useful and of course my friend Lubos Omelina for his enor- mous work on the development of the rehabilitation platform and the serious games, obviously without him none of this would have been possible. But of course there is not only work in life and so I also appreciated the many discussions we had with him and Katka while enjoying good Belgian beers (or Coke!).

My sincere thanks also goes to Prof. Stéphane Louryan, director of the laboratory of Anatomy, which has always given me a great deal of autonomy in the way I organize my time, otherwise the research presented in this thesis would have been much more difficult to orga- nize.

I thank my fellow labmates from ULB and VUB in for the stimulat- ing discussions during lunch or seminars.

Some people did not participate directly in this thesis but still con- tributed to its successful completion. A special thank you to Sonia for her constant good humour in and out of the lab as well as Olivier whose passion for Research and Science is still an inspiration to me!

Last but not the least, I would like to thank my parents and my sister for supporting me throughout writing this thesis and my life in general as well as my friends.

v

A B S T R A C T

Lack of motivation has been highlighted as a major cause of adverse patients outcomes during the rehabilitation process.

Thanks to the evolution in the way video games are controlled, from totally passive to active situation, clinicians have had the idea of integrating video games into rehabilitation to increase adherence to treatments, especially during at-home exercises.

Specially developed games, the Serious Games (SG), must be created to fulfill the needs and specificities of clinicians and patients.

As the rehabilitation exercises are performed in the SG, it is easy to record the movements performed by the patients with the gaming hardware.

The aim of this thesis was to determine if the gaming hardware, the Kinect™sensor and the Wii Balance Board™, can be used to: i) perform functional evaluation of the patients, ii) monitor motions performed by the patients during the rehabilitation exercises and iii) assess the severity of the disease based on the performance during the SG.

In order to validate this new approach several validation processes were done: from the strict controlled environment gait laboratory to the clinics.

The different developments and validation phases are presented in this thesis.

vii

C O N T E N T S

i general introduction 1

1 background of the research project 3

1.1 Context . . . . 3

1.2 Aims and outline of the project . . . . 5

1.2.1 Aims and objectives . . . . 5

1.2.2 Outline . . . . 6

2 the rehabilitation medicine 7 2.1 Definition and principles . . . . 7

2.2 A highly multidisciplinary work . . . . 8

2.3 The neuromusculoskeletal system . . . . 9

2.3.1 The central nervous system (CNS) . . . . 9

2.3.2 The peripheral nervous system . . . . 10

2.3.3 The musculoskeletal system . . . . 10

2.4 Treatment and exercises . . . . 11

2.5 Adherence to treatment and motivational issues . . . . 12

2.6 Functional evaluation and physical rehabilitation . . . 14

2.6.1 Quality of measurement tools . . . . 14

2.6.2 Classical approach . . . . 15

2.6.3 New trends in functional evaluation . . . . 17

2.6.4 Summary and problems to solve . . . . 21

3 the serious games 23 3.1 The games in rehabilitation . . . . 23

3.2 Definition of Serious Games . . . . 25

3.3 History . . . . 26

3.4 Principle of action . . . . 27

3.5 Field of applications . . . . 28

4 the use of commercial video games in rehabili - tation: a systematic review 29 4.1 Introduction . . . . 31

4.2 Methods . . . . 31

4.2.1 Search strategy . . . . 31

4.2.2 Inclusion criteria . . . . 32

4.2.3 Data extraction . . . . 33

4.3 Results . . . . 33

4.4 Discussion . . . . 35

4.5 Conclusion . . . . 36

5 cost- effective ( gaming) motion and balance de - vices for functional assessment: need or hype? 37 5.1 Introduction . . . . 39

5.2 High demand from the clinical field and public health services . . . . 40

ix

ity of morphological measurements using the

kinect ™ sensor: comparison with standard stereopho-

togrammetry. 53

6.1 Introduction . . . . 55

6.2 Methods . . . . 56

6.2.1 Participants . . . . 56

6.2.2 Materials . . . . 57

6.2.3 Data collection . . . . 57

6.2.4 Data processing and statistics . . . . 58

6.3 Results . . . . 61

6.3.1 Correlation between manually determined height and MMC values . . . . 61

6.3.2 Ratio between segments’ length and body height 64 6.3.3 Morphological data comparison . . . . 66

6.4 Discussion . . . . 66

6.5 Conclusion . . . . 69

7 validity and reproducibility of the kinect™ within functional assessment activities: comparison with standard stereophotogrammetry 73 7.1 Introduction . . . . 75

7.2 Methods . . . . 76

7.2.1 Participants . . . . 76

7.2.2 Material and data . . . . 76

7.2.3 Data collection . . . . 77

7.2.4 Data processing and statistical analysis . . . . . 77

7.3 Results . . . . 80

7.3.1 MMC-MBS discreptancies . . . . 80

7.3.2 Test-Retest reliability . . . . 84

7.4 Discussion . . . . 84

8 interchangeability of the wii balance board ™ for bipedal balance assessment 89 8.1 Introduction . . . . 91

8.2 Methods . . . . 92

8.2.1 Participants . . . . 92

8.2.2 Measurement setup . . . . 92

8.2.3 Procedure . . . . 92

8.2.4 Data Processing . . . . 93

8.2.5 Statistics . . . . 94

contents xi

8.3 Results . . . . 96

8.4 Discussion . . . . 97

iii validation through serious games 99 9 validation of the balance board ™ for clinical evaluation of balance during serious gaming re- habilitation exercises 101 9.1 Introduction . . . . 103

9.2 Material and Method . . . . 104

9.2.1 Participants . . . . 104

9.2.2 Procedures . . . . 104

9.2.3 Data processing . . . . 105

9.2.4 Statistical analysis . . . . 106

9.3 Results . . . . 110

9.4 Discussion . . . . 110

10 3d analysis of upper limbs motion during reha - bilitation exercises using the kinect ^tm sensor: development, laboratory validation and clini- cal application 117 10.1 Introduction . . . . 119

10.2 Methods . . . . 122

10.3 Laboratory validation . . . . 129

10.3.1 Participants . . . . 129

10.3.2 Material . . . . 129

10.3.3 The Serious Games . . . . 130

10.3.4 Data processing and statistics . . . . 132

10.3.5 Results of the laboratory validation . . . . 132

10.4 Validation in Clinical Environment . . . . 134

10.4.1 Participants . . . . 134

10.4.2 Material . . . . 134

10.4.3 Data Processing and Statistics . . . . 134

10.4.4 Results of the Validation in the Clinical Envi- ronment . . . . 135

10.5 Discussion . . . . 138

10.6 Conclusion . . . . 141

iv clinical validation 143 11 suitability of functional evaluation embedded in serious game rehabilitation exercises to as- sess motor development through ages. 145 11.1 Introduction . . . . 147

11.2 Materials and methods . . . . 147

11.3 Results . . . . 149

11.4 Discussion . . . . 149 12 automated functional upper limb evaluation of

patients with friedreich’s ataxia using serious

games 155

12.4 Discussion . . . . 167

12.5 Conclusion . . . . 170

v general discussion 171 13 the end of active video games and the conse - quences for rehabilitation 173 14 discussion 177 14.1 Summary of the results and other studies . . . . 177

14.1.1 Laboratory validation . . . . 177

14.1.2 Validation through Serious Games . . . . 179

14.1.3 Clinical validation . . . . 181

14.2 Clinical efficacy . . . . 182

14.3 Technology’s sustainability . . . . 184

14.4 Simple analysis or 3D evaluation? . . . . 186

14.5 Ethical issues . . . . 188

14.6 What about the brain? . . . . 189

14.7 Limitations . . . . 193

14.8 Future developments . . . . 194

14.8.1 Integration of SG in the conventional rehabilita- tion . . . . 194

14.8.2 Is it limited to rehabilitation? . . . . 199

14.8.3 Virtual reality . . . . 202

14.8.4 Robotics . . . . 202

14.8.5 A solution for emerging countries? . . . . 203

14.8.6 Pose detection . . . . 205

14.8.7 Balance evaluation . . . . 207

14.8.8 Tele-monitoring . . . . 208

14.8.9 Machine learning & Data Mining . . . . 208

15 conclusion 211

bibliography 217

Appendix 247

L I S T O F F I G U R E S

Figure 3.1 Success criteria for rehabilitation exercises . . . 23

Figure 3.2 The four pillars of learning . . . . 25

Figure 4.1 Flow diagram of the study selection. CP: Cere- bral Palsy, PD: Parkinson’s Disease . . . . 32

Figure 4.2 Number of studies per year . . . . 33

Figure 4.3 Patients divided by pathology . . . . 34

Figure 4.4 Device use in the selected studies . . . . 34

Figure 5.1 Example of software developed for functional assessment . . . . 42

Figure 6.1 Skeleton model used for MMC and MBS com- parison . . . . 59

Figure 6.2 Body segment lengths expressed as a ratio of the body height . . . . 60

Figure 6.3 Distribution of expected errors after regression 62 Figure 6.4 Results of Session 1 and 2 for MMC and MBS with regression line . . . . 67

Figure 7.1 Illustrations of the four motions performed in the study . . . . 78

Figure 7.2 Stick figure model obtained from the Kinect™ sensor . . . . 80

Figure 7.3 Bland and Altman plots for MMC and MBS comparison . . . . 85

Figure 8.1 Differences between FP and WPP for the 72 trials 95 Figure 9.1 Illustrations of the serious gaming used in the study . . . . 105

Figure 9.2 Scatter plots and regression lines for the DOT and Area obtained during the games. . . . 111

Figure 9.3 Bland & Altman plots for the DOT and Area.X axes are the means of the two systems in de- grees; Y axes are the mean of the differences in degrees. Red lines (middle one) represent the mean difference between the devices. Blue lines (extremities) indicate upper and lower ag- greement (1.96 SD). . . . 112

Figure 9.4 Scatter plots, correlation lines and equations for the speed related parameters . . . . 113

Figure 10.1 Joint center estimation from the Kinect (red cir- cle), reconstructed PiG like data (transparent 34 circles) and nineteen local coordinate sys- tem origins (indicated by numbers) . . . . 121

xiii

Figure 11.2 Evolution of motor function through ages . . . 153 Figure 12.1 Screenshot of the game . . . . 161 Figure 12.2 Mean and 95% confidence interval (CI) of the

time for healthy subjects (blue) [from [200]]

and patient results (black). Black dots repre- sent individual results of the patients. . . . 164 Figure 12.3 Mean and 95% CI of the dexterity index for

healthy subjects (blue) [from [200]] and patient results (black). Black dots (right side) and grey squares (left side) represent individual results of the patients. Since no statistically significant difference were found between right and left side mean of the two sides was used for the fitting. . . . 164 Figure 12.4 Results of the time and dexterity index expressed

in percentage of the values of healthy subjects according to the duration of the disease. Linear fitting with 95% CI is presented with R² . . . . 168 Figure 12.5 Scatter plots and linear regression for the rela-

tion between time and dexterity according to age. . . . 169 Figure 14.1 Comparison between Kinect V1 and V2 . . . . 178 Figure 14.2 Influence of the position of the body related to

the WBB on the results . . . . 180 Figure 14.3 The paradox between the speed of technology

development and the validation process . . . . 186 Figure 14.4 The communication between the different ac-

tors of rehabilitation . . . . 187 Figure 14.5 Cognitive evaluation using mobile games . . . 193 Figure 14.6 The three main points to discuss during the

development of a serious game . . . . 196 Figure 14.7 Evolution of the motivation during the treatment198 Figure 14.8 Examples of balance between conventional ex-

ercises and SG. . . . 199

Figure 14.9 Evidence-Based Practice . . . . 201

Figure 14.10 Inequalities in access to professional health care 203

L I S T O F TA B L E S

Table 2.1 Patients’ habits during rehabilitation exercises 13 Table 2.2 Characteristics of the different tests currently

used to assess stroke patients . . . . 19 Table 5.1 Studies on WBB validation . . . . 44 Table 5.2 Studies on Kinect sensor validation. . . . 46 Table 6.1 Correlation between measured and MMC seg-

ment length . . . . 63 Table 6.2 Correlation between segment lengths determined

from MMC and MBS measurements . . . . 65 Table 6.3 Results of Intraclass Coefficient Correlation (ICC),

Standard Error of Measurement (SEM), Root Mean Square Error (RMSE)and Mean of Ab- solute Difference (MAD) for MMC and MBS between Session 1 and 2. Correlations between manual measurements (Pearson r) . . . . 71 Table 6.4 Correlation between segment lengths determined

from MMC and MBS measurements before and after regression . . . . 72 Table 7.1 Mean results (std) expressed in degrees for MMC

and MBS . . . . 82 Table 7.2 Test-retest reliability analysis for both the MMC

and MBS system . . . . 87 Table 8.1 Intraclass correlation coefficient (ICC) for the

different studied parameters for the four dif- ferent devices, lower and upper bound and the confidence interval are indicated between the brackets. p-value are the results of Friedman tests . . . . 96 Table 8.2 Contingency table of the number of observa-

tions that are out of the confidence interval.

Values into bracket are the ratio between out- side values and number of observations. . . . . 97 Table 8.3 Spearman coefficient correlation between FP and

WBB for the four different devices. . . . 98 Table 9.1 Mean (std) results for the Flight Simulator . . 107 Table 9.2 Mean (std) results for the Wipe Out . . . . 108 Table 9.3 Mean (std) results for both games . . . . 109 Table 10.1 Relative Coordinate Systems topology . . . . . 122 Table 10.2 List of parameters evaluated for trajectory anal-

ysis . . . . 123

xv

Table 11.2 Coefficient of determination (R ² ) for the differ- ent fitting methods . . . . 152 Table 12.1 Characteristics of the patients included in the

study . . . . 160 Table 12.2 Mean (std) results for patients and control. . . 163 Table 12.3 Pearson’s correlation coefficients (R) between

scores obtained from the SG and the clinical evaluation. . . . 166 Table 14.1 Correlation (Spearman’s ⇢) between score com-

puted from CoP displacement (WBB) and the different clinical scales used to assess balance and posture. n = 9 . . . . 181 Table 14.2 Mean (std) TCMS before and after interven-

tions for the three categories and the total scores.182 Table 14.3 Results of the survey about SG acceptance in

Morocco . . . . 205

A C R O N Y M S

AL Anatomical Landmark

AVG Active Video Games

BA Bland & Altman plot

CMC Coefficient of Multiple Correlations

CP Cerebral Palsy

CVME Coefficient of Variation of the Method Error

CoP Center of Pressure

FP Force Plate

FRDA Friedreich ataxia

ICC intraclass Correlation Coefficient

ISB International Society of Biomechanics

Kinect Microsoft Kinect™sensor

LCS Local Coordinate System

LOA Limits of Agreement

MBS Marker Based System

MDC Minimal Detectable Change

MMC Markerless Motion Capture

PR Physical Rehabilitation

RCP Reproducibility Coefficient

RCT Randomized Clinical Trials

RMSE Root Mean Square Error

RoM Range of Motion

SEM Standard Error or Measurement

SG Serious Games

VG Video Games

VR Virtual Reality

WBB Nintendo Wii Balance Board™

WHO Word Health Organization

xvii

Part I

G E N E R A L I N T R O D U C T I O N

Because of the highly multidisciplinary aspects of this the- sis and the wide audience it must reach (e.g. doctors, phys- iotherapists and engineers), many aspects must be addressed and defined in this introduction. As the technical aspects are discussed in the experimental parts of this thesis, we will focus here more on the aspects of physical medicine and rehabilitation and discuss the technological implica- tions and functional evaluation in the discussion. The con- cept of serious games is defined with the current limita- tions and the different fields of applications.

Results of a systematic review about the use of commer- cial video games in physical rehabilitation are also pre- sented to give an overview of the type of pathologies can be addressed by this new technology.

Finally, a paper presenting the needs of the clinic but also

the specificities that the devices should satisfy are pre-

sented at the end of this section in order to introduce the

different chapters on the experimental validation of play

equipment in clinics.

1

B A C K G R O U N D O F T H E R E S E A R C H P R O J E C T

Imagination is more important than knowledge.

Knowledge is limited.

Imagination encircles the world.

- Albert Einstein

1.1 context

Commercial video games have significantly evolved over the last decade.

Today computer performance and play experience allow new perspec- tives for rehabilitation. Thanks to new gaming controllers (Nintendo Wii Fit, Microsoft Xbox Kinect, etc.) video game playing has changed from a passive (i.e., the player is seated on a sofa) into an active expe- rience: players have to move in order to interact with games.

Clinicians are now prospecting the new potential use of these games in rehabilitation mainly through testing available commercial games with patients suffering from various pathologies, such as in the neuro- logical field. Physical rehabilitation must be based on active exercises, and these new gaming strategies allow for this. Currently there are three main limitations with the use of commercial video games in re- habilitation which are not personalized to specific patient problems.

First of all, these games are not adapted for rehabilitation (e.g. not based on physical rehabilitation exercises, based on speed, too com- plex visual background) since they are developed for fun and enter- tainment purposes only.

Secondly, motion accuracy requested from the player to achieve the game goals is low while most therapists will aim to improve the patient’s joint control and coordination in more complex ways.

Thirdly, there is currently no possibility to record the motion per- formed by the patients during exercises. However, collecting this in- formation could be important : (i) to enable direct feedback to the patient and eventually correct the motion if they are not performed in the right way and (ii) to provide information to therapists in case of telerehabilitation exercises when the patient is performing exercises at home without the clinicians’ supervision.

3

the Microsoft Xbox Kinect™(Kinect) in 2007 and 2010 respectively numerous studies have been done in order to evaluate if these devices can be used to perform simple biomechanical evaluation (Kinect) and balance or posture assessment (WBB) but very few work has been performed on their evaluation during SG rehabilitation exercises.

It appears thus that specially-developed SGs coupled to the WBB and the Kinect sensor could be used to follow patients’ evolution dur- ing rehabilitation exercises. However, before being used in clinics to assess patients, the data collected during the games and parameters extracted from it must be validated.

Such kind of evaluation, done during the rehabilitation, has many advantages;

1. it is done in the natural environment of the patient (it is known that patients are not exhibiting the same performance when they are wearing underwear in a gait laboratory)

2. serious games seem to divert patients’ attention, and patients are therefore less focused on pain and tasks to be reached; conse- quently, patients can reach larger motions amplitudes and show increased motivation

3. time is saved

4. financial benefices are achieved (the devices are affordable and

since the evaluation is done within the therapy session there is

no dual pricing)

1.2 aims and outline of the project 5

1.2 aims and outline of the project 1.2.1 Aims and objectives

The overall aim of the project is to validate the use of serious games to perform functional evaluation during rehabilitation exercises.

The specific objectives of the doctoral project are

1. To determine the field of application and the current limitations of the use of SG in rehabilitation

2. To validate the use of commercial gaming hardware in the con- text of functional evaluation

3. To integrate and validate the recordings of motion performed by the patients during the rehabilitation exercises

4. To create relevant scores based on the motions performed dur- ing SG

5. To validate the scores in clinics

In order to answer these various questions, several studies have been conducted, results of those studies have been published in the literature (a total of 23 papers have been published in international peer-reviewed journals).

Among these studies we selected and presented ten chapters in

this thesis based on papers to illustrate the path from simple com-

mercial games up to the use of serious games to perform functional

assessment with patients. All of these studies have been published in

international peer-reviewed journals.

Laboratory validation

• Determination of the precision and accuracy of morphological measurements using the Kinect™

sensor: comparison with standard stereopho- togrammetry. (Chapter 6)

• Validity and reliability of the Kinect™ within functional assessment activities: comparison with standard stereophotogrammetry. (Chapter 7)

• Interchangeability of the Wii Balance Board™ for bipedal balance assessment. (Chapter 8)

Validation through Serious Games

• Validation of the Balance Board™ for clinical evaluation of balance during serious gaming rehabilitation exercises. (Chapter 9)

• 3D analysis of upper limbs motion during reha- bilitation exercises using the Kinect ^TM sensor:

development, laboratory validation and clinical application. (Chapter 10)

Clinical validation

• Suitability of functional evaluation embedded in serious game rehabilitation exercises to assess motor development through ages (Chapter 11)

• Functional upper limb evaluation of patients with Friedreich’s Ataxia using serious games (Chapter 12)

Discussion • The end of active video games and the conse-

quences for rehabilitation (Chapter 13)

2

T H E R E H A B I L I TAT I O N M E D I C I N E

Health is like money, we never had a true idea of its value until we lose it.

- Josh Billings

2.1 definition and principles

Before discussing the use of SG in rehabilitation, it seems appropriate to define and obtain some precise notions about rehabilitation.

The definition and description of the rehabilitation presented here are presented more in details in the book Serious Games in Physical Rehabilitation: from theory to practice [1] written by the candidate.

Rehabilitation is a key health strategy to address disability [2]. Re- habilitation is a branch of physical medicine. Rehabilitation, in the area of health, can be defined as the ability to rehabilitate patients in his environment: ’the aim of rehabilitation is to restore or return a person to a state of optimal functioning in interaction with his/her environment’

[3]. Another interesting definition of rehabilitation is the one from the World Health Organization ’a set of measures that assist individuals who experience, or likely to experience, disability to achieve and maintain optimal functioning in interaction with their environments’ [4].

Two key elements of rehabilitation are included in this definition.

The first one is the importance of trying to restore optimal func- tioning of the patient. Of course, unfortunately, in some cases it is not possible to restore ’normal’ mobility or activity; therefore, reha- bilitation is mainly focusing on the function and autonomy during activities of daily living.

The second important aspect is the interaction with the environ- ment: interaction with the surrounding objects (to move, to eat, to wash) to be as independent as possible but also to interact with peers.

The social aspect is one of the key, and unfortunately often neglected, aspects of the rehabilitation and, more generally, of the overall man- agement and integration of disabled people in the society at large.

7

Expertise from various medical and paramedical points of view is needed to have a holistic approach of a particular patient suffering from a specific pathology living in his own environment.

It is important to note that two patients presenting the same pathol- ogy will not systematically receive the same treatment because lots of parameters have to be taken into consideration [5]. The Interna- tional Classification of Functioning, Disability and Health (ICF) from the World Health Organization (WHO) has been created to underline the importance that personal and environmental factors could and should play in rehabilitation ¹ .

The ICF is divided into four main categories: Body functions, Ac- tivities and participation, Environmental factors and Body structures.

The thesis focuses on physical rehabilitation and hence mainly on body function as described in the ICF but also involved the other main categories.

The title of this thesis contains the word ’Functional assessment’ this term is very broad and must be specified here.

According to ICF the body function (part b) is separated into eight parts: Mental functions - Sensory function and pain - Voice and speech function - Function of cardiovascular, hematological, immunological and respiratory systems - Functions of the digestive, metabolic and endocrine systems, Genitourinary and reproductive functions - Neu- romusculoskeletal and movement related functions - Function of the skins and related structures.

In this thesis we will work within the framework of Neuromuscu- loskeletal and movement related functions part of the ICF, more precisely at the level of movement function (b750-b789). We can further specify the area of interest of this thesis up to the level of Control of voluntary movements functions (b760) which includes control of simple (b7600) and complex (b7601) voluntary movements, the coordination (b7602) and the supportive functions (b7603).

Note that to perform a motion correctly, especially if it is coupled with video games, other functions such as cognition (e.g. attention functions [b140]) and vision (seeing functions [b210]) are also needed.

1 http://www.who.int/classifications/icf/en/

2.3 the neuromusculoskeletal system 9

2.3 the neuromusculoskeletal system

The overall aim of rehabilitation is to restore or return a person to a state of optimal functioning. For achieving this goal, different ob- jectives are targeted such as posture, balance, strength, coordination, endurance, and dexterity.

For a healthy subject, a lot of daily activities are done automatically, without even thinking about it, despite this ’automation’ even a rela- tively simple motion such as taking a glass and drinking it requires optimal functioning and the synchronization of several anatomical systems.

It is interesting to have a simplified overview of these various anatom- ical systems, and how they interact with each other, to understand the pathologies, and therefore the different strategies of rehabilita- tion. Here is a non-exhaustive list of the principle components of the nervous and musculoskeletal system involved in voluntary contrac- tion with the main function of each component, and a few examples of pathologies that can affect each level.

2.3.1 The central nervous system (CNS)

• The motor cortex : Located in the frontal lobe, it is involved in the programming (mainly in the premotor cortex that is located in front of the motor cortex), the execution and the control of voluntary movement. Lesions (e.g., stroke, traumatic brain in- jury, tumors) in the motor cortex induce paralysis. The motor cortex is composed of different components: the primary motor cortex, the premotor cortex, the supplementary motor cortex and the posterior parietal cortex.

• The cerebellum : Involved in the coordination, precision and ac- curate timing of the motion. The cerebellum is important in the acquisition and learning of new skills and motricity. It allows for supervised learning of a new task by transmitting the error be- tween the expected motion and the performed one to the motor cortex: it works as a feedforward learning. Therefore, a lesion in the cerebellum induces posture and balance disorders, difficul- ties with motor learning and fine movement controls (dexterity).

The cerebellum has also an important cognitive function mainly

for language and attention.

system and the peripheral nerves. It relays informations from brain to muscles through the efferent or descending pathways (pyramidal and extrapyramidal tracts) and from peripheral (mus- cles, joints, glands, organs) to brain through the afferent or as- cending pathways (dorsal column medial lemniscus system, an- terolateral system, spinocerebellar tracts). Patients with spinal cord injuries could present various symptomatologies depend- ing on the severity and localisation of the lesions: from muscle weakness and loss of sensitivity within limited regions (i.e. her- niated disc) to full body paralysis (i.e. laceration of the spinal cord in the cervical spine due to trauma).

2.3.2 The peripheral nervous system

• The nerves : They relay between the spinal cord and the muscles, joints, ligaments, bones. Nerves can be stretched, compressed, cutted inducing muscles’ weakness or paralysis (efferent), loss of sensitivity and proprioception (afferent).

2.3.3 The musculoskeletal system

• The muscles : Muscles’ contraction is a complex mechanism

involving neuromuscular junction (release of neurotransmitter

[mainly acetylcholine]) and the slide of myo-filaments (actin

and myosin) to produce changes in muscles’ lengths and finally

producing motion. The muscle spindles also provide informa-

tion to the CNS relative to the length and the stretch of the

muscle allowing for proprioception. In rehabilitation, it is im-

portant to restore or preserve the proprioception to maintain

optimal functioning of the musculoskeletal system (fine motor

function and balance). The muscles are attached to bones by

tendons (also containing proprioceptive fibers).

2.4 treatment and exercises 11

• The joints : Movement can be defined as the change of position of one bone relative to another one. Bones are linked together through joints; these joints are reinforced by ligaments. Liga- ments provide lots of proprioceptive and nociceptive informa- tion (avoiding luxation of the joint).

2.4 treatment and exercises

Depending on the level of the lesions and the severity of the disease (e.g., moderate balance problem due to lack of vascularization of the cerebellum, quadriplegia in case of complete spinal cord injury), the objectives and techniques of rehabilitation are totally different.

The general aim of the rehabilitation is to increase muscle coordina- tion; reduce muscle weakness and/or muscle tone; improve balance;

improve gait and motion independence; limit bone deformation and bone deviation.

There is one constant in every rehabilitation program: the more the patient is performing the exercises, the more efficient the program will be [6]. Therefore, the rehabilitation does not only take place dur- ing the supervised sessions with the clinicians, patients also need to perform some exercises alone (at home, in the room, etc.). This is due to the fact that in order to create efficient new connections in the brain, the plasticity, high number of repetitions are needed to create and strengthen new nerve connections.

To be efficient these exercises must be performed and most impor- tant they must be well achieved. If the exercises are not well per- formed (e.g. due to the inability of the patient to perform alone due to the disability or if the patient forgets how to perform correctly the exercises), it can lead to some perverse adverse effects such as increasing compensatory movements or leading to vicious posture.

Therefore, it is important to have a regular feedback between clin- icians and patients to ensure that the patients are performing the exercises and that they are doing them correctly.

Depending on the pathology, the age of the patient, the infrastruc-

ture, etc., the number (and the durations) of the session vary between

several sessions a day (in specialized centers in case of severe disabil-

ity) to one weekly session.

For some pathologies, rehabilitation schemes are advised on a, very, frequent basis. Therefore the motivation of the patients is the main key to as successful rehabilitation treatment [8].

The main challenge is therefore to keep patients motivated enough despite the feeling he/she could have of ’inefficacy’, ’lack of progress’,

’tiredness’ etc.

Such problems are even more present with teenagers during the puberty identity crisis. Even if progress is slow, or when patients feel that there is no progress at all, it is important to pursue the rehabil- itation program. If not, physical condition could deteriorate and the benefits of the previous sessions can be lost.

In order to collect data on patient compliance related to rehabilita- tion recommendations by clinicians, we realized a survey on patients’

habits during rehabilitation exercises. 340 patients (319 in Europe and USA, 21 in Morocco) participated in this survey [9]. Results are pre- sented in Table 2.1 (page 13). The results are similar for both prescrip- tion and adherence in developed and emerging countries: 86% of clin- icians prescribed home-exercises but only 33% of those exercises are really performed by the patients.

Based on this survey, there is clearly a need to develop solutions to:

• Show how to realize the rehabilitation exercises

• Correct them (feedback)

• Remind patients when and what exercises they should do

• Motivate the patients

2.5 adherence to treatment and motivational issues 13

Table 2.1: Patients’ habits during rehabilitation exercises

Question Answer Morocco Europe

and US

Prescription? Yes 89% 86%

No 11% 14%

Adherence?

Totally 42% 29%

Partially 42% 54%

No 16% 17%

Causes

Due to lack of time 14% 28%

I forgot to do the exercises 28% 24%

I felt there was no evolution (use-

less) 0% 10%

Exercises are too boring 0% 25%

I could not remember how to do

the exercises myself at home 43% 19%

Motivation

Reminder (smartphone, email, cal-

endar) with instructions 19% 18%

Reminder (smartphone, email, cal- endar) with video instructions on how to perform the exercises

6% 23%

Application showing how to per-

form exercises with live feedback 0% 30%

Motivating and fun rehabilitation

exercises in computer games 6% 20%

Other (specify)

63%

(sheets of paper with instruc- tion)

9%

Most of the time functional evaluation is performed by physiothera- pists (general function, gait, trunk) or occupational therapists (mainly for the upper limbs) not by medical doctors.

There are two types of functional evaluation, although the distinc- tion between the two is not always clear: the qualitative ( ⇡ subjective) and the quantitative ( ⇡objective) evaluations as explained below.

2.6.1 Quality of measurement tools

Before discussing the use and limitations of the devices and scales used to perform functional evaluation several terms need to be de- fined.

A measure can be affected by bias: systematic bias when the error is always the same (e.g. problem of calibration of the device) or ran- dom bias when the direction of the error changes over time (when performing measures on humans, plenty of variables present varia- tion during the day [e.g. height, inflammatory levels, fatigue, etc.] or between days).

It is therefore important to know the performance of the devices used to perform clinical measurement in order to interpret them care- fully.

The validity of a test is the degree of closeness of measurements (i.e. how close is the value compared to the real value). Several indi- cators are used to assess the validity of a device compared to a gold- standard. The most common one is the method developed by Bland &

Altman [10] for continuous data. They defined a Limit of Agreement

(LOA) between two method as LOA = mean(differences) ± 1.96 ⇤

std(differences) [at ↵ = 5%]. In order to be used in clinics this LOA

must be smaller than the limits of acceptability defined by experts

and clinicians.

2.6 functional evaluation and physical rehabilitation 15

The reproducibility (i.e., the degree to which repeated measurements perform under the same conditions will show the same results). The distinction must be made between measurements performed by the same observer (intra rater) or between different observers (inter rater).

The most used method to assess reproducibility is the Intraclass Cor- relation Coefficient (ICC). Based on the 95% confidence interval of the ICC estimate, values less than 0.5, between 0.5 and 0.75, between 0.75 and 0.9, and greater than 0.90 are indicative of poor, moderate, good, and excellent reliability, respectively [11].

Two other relevant clinical parameters can be processed: the Stan- dard Error of Measurement (SEM = std(firsttest) ⇤ p

1 - ICC) is a reliability measure to assess response stability [12] and the Minimal Detectable Change (MDC = 1.96 ⇤ SEM ⇤ p

2) that estimates the small- est amount of change that can be detected by a measure that corre- sponds to a noticeable change in clinics [13]. In clinics the distinction must be made between the MDC (intrinsic characteristics of the mea- surement tool) and the Minimal Clinically Important Difference or minimally important change [14].

2.6.2 Classical approach

The qualitative evaluation

To perform qualitative evaluation clinicians are using validated scales and scores.

The evaluation can be done directly by the clinicians: the patient is requested to perform several tasks and the evaluator rates it accord- ing to some given scale. Typically the scores for each task is ’O’ when the patients is not able to do the task, ’1’ if he can do it with help or partially and ’2’ if he can do it independently. The advantages of this kind of evaluation are that it is relatively quick, easy to perform and does not required expensive materials. On the other hand those scales are poorly sensitive and can not detect small modifications of the status of the patients. For example, daily clinics evaluation of bal- ance is performed using scales such as the qualitative Berg Balance Scale [15]. Despite the fact that these scales have been validated for various neurological conditions [16], they are not sensitive enough to detect small clinically relevant changes [15]. Another point is that the results are dependent on the observer ² .

2 Due to the large amount of available, and validated, tests and scales we can not present here all the characteristic of the tests. The characteristics of the most used scales are summarized in the Rehabilitation Measures Database ( https://www.

sralab.org/rehabilitation-measures )

the patients), there is a tendency to exaggerate the results [19].

Due to the limitations of the qualitative evaluation explained here above, there is a need to have more sensitive data: the quantitative evaluation.

The quantitative evaluation

People have been thinking about how they walk since the earliest times. Aristotle (384-322 BCE) can be attributed with the earliest recorded comments regarding the manner in which humans walk. It was not until the renaissance that further progress was made through the ex- periments and theorising of Giovanni Borelli (1608-1679). Although several scientists wrote about walking through the enlightenment pe- riod it was the brothers Willhelm (1804-1891) and Eduard (1806-1871) Weber, working in Leipzig who made the next major contribution based on very simple measurements [20].

Both Jules Etienne Marey (1830- 1904), working in France, and Ead- weard Muybridge (1830-1904), working in the USA, made significant advances in measurement technology thanks to the invention of pho- tography. First studies on motion analysis were performed, first on animals then on humans, based on pictures during the 1880s.

Nowadays, technology has developed and there are plenty of dif- ferent kind of devices available to perform motion analysis. Those de- vices can be separated according to their complexity (attempt to clas- sify from the simplest to the most complex): photogrammetric assess- ment [21], goniometers [22], accelerometers and gyroscopes, wearable inertial sensors [23], markerless motion capture (MMC) system [24]

and stereophotogrammetry devices (Marker Based System - MBS) [25, 26, 27, 28].

In this thesis we are going to analyze the results of MMC compared to gold-standards for 3D motion analysis, the MBS.

The advantages, disadvantages and limitations of these two kind

of devices are going to be discussed in Part ii.

2.6 functional evaluation and physical rehabilitation 17

2.6.3 New trends in functional evaluation

Over the last years there have been major advances in the develop- ment and use of new technologies, not only the (serious) games re- lated ones. The most important point is, probably, not the exponential curve of the technological development but the speed at which this technology is becoming widely available for the public. Smartphones, connected watches, motion tracking systems and other smart clothes have turned the people into a connected body, true sensors in their own right. While these technologies are widely used by the public (e.g., Endomondo ³ to track and monitor running and walking pro- gresses using the gyroscopes of the smartphone, Lost It! ⁴ for weight management using a scanner, connected balance and activity bands, Peak ⁵ for cognitive assessment and training using smartphone, etc.) they are only used in limited cases in clinics and a fortiori even less during clinical trials.

In this part, we will present the use of new devices to perform quantitative evaluation and follow-up of the patients during the re- habilitation, why this development is needed and its current limita- tions. We will discuss the case of stroke because stroke is the second most common cause of death (11.8% of all deaths worldwide), after ischemic heart disease (14.8% of all deaths, and the third most com- mon cause of disability after neurological conditions and ischemic heart disease [29]. It is estimated that 33 million new cases of stroke occur each year [30]. Among the new cases about 10% of the patients will recover completely, 25% with mild deficits, 40% with moderate to severe deficits, 10% of these patients must be placed in specialized centers because of the functional impact and 15% die shortly after stroke [30].

Current state of the art

There are two types of stroke: ischemic and hemorrhagic. Ischemic strokes are the most common (about 80% of cases) [31]. Patients with stroke usually are affected by sensory and motor dysfunctions and by cognitive impairment.

3 https://www.endomondo.com

4 https://www.loseit.com

5 http://www.peak.net

successful recovery is highly variable across measures and cut-off points for defining successful outcomes vary [33].

The Fugl-Meyer scale was developed as the first quantitative evalu- ative instrument for measuring sensorimotor stroke recovery. Limita- tions of the motor domain include a ceiling effect, omission of some potentially relevant items, and weighting of the arm more than the leg [34].

Several scales are used to assess the global function such as the NIH Stroke Scale [35], the Barthel Index of Activities of Daily Living [36] and the Modified Rankin Outcome Scale [37].

Concerning the balance and mobility impairment the most used metrics are the Berg Balance Scale (BBS) and the 2 or 6-Minute Walk Test (2MWT-6MWT) [38].

Fatigue Severity Scale (FSS) is the most widely used measure to assess directly the fatigue and weakness [39] but other tests such as the 2MWT or the 6MWT are also used as an indirect indicator [38].

Except the walking tests that are objective and quantitative mea- surements, all the other tests are scores and scales. There are two potential problems related to the use of scales: it is qualitative evalua- tion depending, to a certain extent, on the subjectivity of the observer and the time required to perform those evaluations. Therefore, it is difficult, in daily clinics, to have a regular follow-up and evaluation of the patient. Such monitoring is however important for stroke patients.

The characteristics of the different tests are presented in Table 2.2.

It is striking to note that even though most of the tests present good to excellent results for test-retest and interrater reliability, the values of the MDC are fairly high, representing 5 to 10% of the maximal value of the test.

Another very frequent problem found in research is the poor trans- fer between results of randomized clinical trials (RCT) and the clinic.

Some treatments (e.g. drugs, medical devices, etc.) show good results

in the controlled research environment (i.e. efficacy) but only limited

results in the real clinical situation (i.e. effectiveness) [40].

2.6 functional evaluation and physical rehabilitation 19

Table 2.2: Characteristics of the different tests currently used to assess stroke patients

Domain Test Intra Inter SEM MDC

General function

Fugl-Meyer (mo- tor function) (/100)

0.99 0.98 9.4 5.2 NIH Stroke Scale

(/42) 0.97 0.95 NE NE

Barthel Index

(/100) 0.94 0.93 1.45 4

Modified Rankin Outcome Scale (/6)

0.95 0.95 NE NE

Balance & Mobility

BBS (/56) 0.98 0.97 1.7 4.6

2MWT (m) 0.97 12 16

6MWT (m) 0.98 0.94 19 36

Fatigue FSS (/63) 0.75 NE 0.7 2

NE: Not established, from the Rehabilitation Measures Database.

In this context, new solutions must be developed to perform reg- ular and quantified evaluation of the patient in his natural environ- ment to get more precise results and to increase the ecological valid- ity.

The evolution of the technology

Technological advances and the evolution of computers have trans-

formed the whole world, including the healthcare sector. The reduc-

tion of the cost and the miniaturization of the new technology is prob-

ably even more impressive than the increase of speed and power of

calculation. Current smartphones can perform three times more oper-

ations (Floating Operations Per Seconds) than supercomputers from

15 years ago. Almost every patient has therefore a supercomputer in

his pocket. This power of calculation allows the development of ma-

chine learning methods. Machine learning relies on the construction

of a model that accounts for a given collection of data and can be

used to predict further data; the evolution of one particular patient

compared to a database of patients presenting the same conditions for

example. In order to ’feed’ the model a large amount of data should

be collected and stored in the database [41].

to assess range of motion [45].

There are two potential issues related to the use of real world data compared to data obtained during current clinical trials: the quality and the quantity of the data.

The quality of the data

The biggest challenge related to the use of real world data is to en- sure the quality of the data, i.e., the so-called data curation. Com- pared to the data usually collected during clinical trials (e.g., biologi- cal samples, scales, clinical measurement, etc.) the conditions of mea- surement cannot be controlled and therefore it is difficult to apply strict and rigorous protocols.

Nowadays data are collected by the clinicians or investigators ac- cording to guidelines, tomorrow’s data will be collected by the pa- tients themselves since the patients need to become actor of their treatments and not spectator anymore. Patient’s empowerment is cur- rently limited to treatment [46] but is this is more than likely that it will also be the case in clinical trials [47]. Patients could only partic- ipate actively in the treatment, or the evaluation and assessment, if they believe in the technology and accept to use it for health-related purposes [48]. For the ’Y’ generation (generation marked by the use and familiarity with IT technologies) also called generation 2.0 in ref- erence to the technological field, this is probably not an obstacle, for the elderly subjects this is definitely a challenge.

The lack of standardization related to real word data could however be counterbalanced by the fact that the measurements are performed on a very regular basis. Therefore, a large amount of data is collected and it is relatively easy to assess the quality of them (i.e. precision) using statistical methods.

Opposed to what is generally well admitted the reproducibility of

results and measurements performed during well designed and con-

trolled clinical trials is relatively low, a recent survey published in

Nature concluded that 75% of the researchers failed to reproduce re-

sults of published studies [49]! There is thus a demand from the field

to have more confirmation before publishing results [50].

2.6 functional evaluation and physical rehabilitation 21

The quantity of the data

The storage of the large amounts of data that can be generated is, currently, no longer a problem thanks to the development of storage facilities (i.e. cloud storage). On the other side the interpretation and the extraction of relevant information from such amount of data re- mains a true challenge.

There is a shifting paradigm in the way of processing the data from the classical statistical analysis: the researcher emits a hypothe- sis, designs a protocol and performs the statistics to test this hypoth- esis to the data mining and other unsupervised methods. In these approaches, the data are processed automatically, without assump- tions. The different algorithms try to classify and differentiate the data, based on these results the researcher - or the clinician - must determine if the proposed results have a clinical sense.

2.6.4 Summary and problems to solve

The complexity of the rehabilitation process and the different anatom- ical systems involved in motor controls was discussed in this chapter.

We then discussed the notions of functional evaluation and the cur- rent limitations of both the qualitative and quantitative approaches as well of the new trends of functional evaluation allowed by the de- velopment of technology. we have seen that there is a need to develop new solutions to perform functional assessment on a more regular basis and to be able to detect small changes in patients’ conditions.

The objective of this thesis is to determine if it is possible to per-

form functional evaluation of patients when they perform rehabilita-

tion exercises using serious games. So now we will focus on serious

games before presenting the different steps of the development and

validation of this innovative solution.

3

T H E S E R I O U S G A M E S

The best way to predict the future is to invent it.

- Alan Kay

3.1 the games in rehabilitation

The use of traditional games in rehabilitation has been investigated for decades by psychologists and physiotherapists.

Several theories have been developed to explain social and cogni- tive development through games and play (e.g., Piaget in 1962 [51], Gottman in 1986 [52]).

Piaget’s theory explained that make-believe play provides player the opportunity to reproduce real-life conflict and ameliorate nega- tive feelings.

Gottman emphasized that play constitutes an emotionally signifi- cant context through which pain loss and anxiety can be enacted.

These two theories could, at least partially, explain the success of integrating games in rehabilitation. Some might think that games are only used in pediatric rehabilitation, it is not the case. Games can be used to increase the quality of rehabilitation for various patients.

Figure 3.1: Success criteria for rehabilitation exercises

23

they are focused on the games, patients can perform more repetitions before getting bored when they are playing.

The last point is the selectivity: is the patient performing the right motion? Is there no compensatory or coupled motions? Since the games are goal-oriented, patients are somehow guided to perform the right motion.

Another way of presenting the positive effects of the games on learning is presented in Figure 3.2. The authors presented the four pillars of learning and how the games can be used to reinforce the training [53]. They discussed the use of video games for learning but it can be easily transfered to rehabilitation.

The attention - orienting network is used to select the relevant infor- mation. Games can promote strategies that help to select relevant in- formation (modelling/examples, modality/use of the audio channel for verbal explanations to guide the exercises, feedback, integration of relevant information in virtual tools).

The active learning is the fact of being active during the learning.

Games promote interactivity rather than being passive (text, audio or video to explain how to do the exercises).

Feedback during rehabilitation exercises is one of the most impor- tant aspects. Patients must evaluate the gap between the objective and the actual performance. This term is related to the proprioception (i.e.

how the brain perceives the body in space), proprioception is affected in various pathologies and one of the goals of the rehabilitation is to restore the proprioception [54]. Games promote the use of feedback which deals with the task completed, not with the self-esteem.

The last point is the consolidation: achieving long-term improve- ment. Games promote the repetition of interactions with important learning content inside the game.

Despite these at least potential positive aspects, very little informa-

tion, or no information at all, can be found in textbooks of rehabilita-

tion, on the use of games for children’s or adults’ rehabilitation (e.g.,

[55, 56]).

3.2 definition of serious games 25

Figure 3.2: The four pillars of learning, a new framework to enhance the learning effectiveness of serious games, from [53].

There is a certain contradiction here between the fact that during the games the patients focused more on the games than on the motor tasks and thus performed more repetitions of exercises and the fact that one of the four pillars of learning using game is to increase the attention on the task.

We will discuss latter on of the link between the motor rehabilita- tion and the cognition but it is clear that the games must be adapted according to the capacity and need of the patients, exactly as we adapted the conventional treatment to each patient. Depending on the pathology and severity of the disease, the clinician will choose to focus on motor exercises only or involving both motor and cognitive aspects.

3.2 definition of serious games

Before explaining goals and principles of this new approach it is im- portant to define the different terms that are used in research and literature.

The term Serious Games - defined as games designed with a pri- mary purpose other than pure entertainment (health care, rehabilita- tion, education, injuries’s prevention. . . ) - is very often used but for some people there is one inadequacy between those two terms, ac- cording to them if a game is too serious this is not a game anymore.

On the other hand some people find that all games are serious, or

at least have serious purposes, and therefore the serious games term

is not an oxymoron [57]. The serious aspect of the games should not

tation or in the health care sector.

The term Exergames (contraction of exercises and games) describes the use of video games in the rehabilitation but this term is less used compared to serious games in the literature [59].

The term Virtual Reality (VR) is also used but this term is much wider than only performing physical exercises during rehabilitation.

The aim of VR is to immerse patients in a virtual environment in or- der to (re)create some sensory inputs and/or putting patients in var- ious situations that will help him to perform those kinds of activities later on in real life [60].The interactive computer play is one subset of VR-based therapy wherein users can interact with virtual objects in a simulated game environment (usually on a two-dimensional screen, not being immersed in a 3D virtual environment) and receive real time feedback on their actions [61].

3.3 history

It is not obvious to determine when the first serious game was created.

According to some authors it could even be that purely entertainment home video games only appeared after the first digital serious games in the seventies [57].

The term serious games, as used today, dates to 2002. Serious games have been developed in various fields. Before 2002 the ’ancestors’ of serious games were mainly developed for education (66%), advertis- ing (11%), ecology (8%) and healthcare (5%). After 2002 there has been an important decrease in education (26%) an increase in adver- tising (31%). The proportion of healthcare application between 2002 and 2010 represents 8% of the 1265 published studies [57].

Two famous games popularized the use of serious games in the

healthcare domain and in health prevention. In 1992 the pharmaceu-

tical company Novo Nordisk™ and the video game company Raya

System™ developed ’Captain Novolin™’ a game developed to teach

children how to manage diabetes and insulin. The superhero is di-

abetic and the aim of the game is to control his glucose-level. The

efficacy of this kind of intervention has been tested in a clinical trial.

3.4 principle of action 27

Diabetic’s children were invited to play ’Packy & Marlon™’ a game similar to the previous one. Compared to a control group, children who played the games managed better their diabetes: the number of children who have to go to the hospital due to a glucose crisis decreased by 77% in the intervention group [62]. Currently this com- pany is a leading publisher of interactive entertainment products de- signed to promote children and adolescent’s health. Their products are endorsed by the American Academy of Pediatrics, Juvenile Dia- betes Foundation, and Asthma and Allergy Foundation of America.

The name has been changed into Health Hero Network™ and is now into the Bosh Healthcare™ group.

Another very popular game that had significant impact on health- related behavior is the ’Re-Mission’ game. This game was designed for children with cancer in order to teach them how to deal with cancer treatment (mainly chemotherapy) to maximize adherence to treatment. In this game patients must shoot cancer cells to fight the infections and manage clinical signs and adverse effects (e.g., consti- pation, nausea, etc.). A clinical trial has been done to compare the ef- fect of the Re-Mission game compared to commercial game (Indiana Jones and the Emperor’s Tomb™) in adolescents and young adults who were undergoing cancer therapy. Results showed that patients playing Re-Mission significantly improved treatment adherence, indi- cators of cancer-related self-efficacy and knowledge.

The findings support current efforts to develop effective video-game interventions for education and training in health care [63]. Currently a second version of the games is freely available on the Internet ¹ . 3.4 principle of action

Serious games could be used to assist patients and clinicians in the rehabilitation of various pathologies (see Chapter 4, page 29). A com- mon feature, regardless the approach or the targeted pathology, is that serious games are used, at least partially, to motivate patients to perform their exercises during the rehabilitation process [64].

A group of Canadian researchers has made a lot of work about the motivational aspect of SG for children [65, 66, 67, 68, 69, 70]. Their main findings are summarized below.

1. The key point is the variability of the games. Patients, and more typically children and teenagers, are quickly getting bored of the games. Thus it’s important to have different kinds of games.

Even in the same rehabilitation sessions games must change (for example two or three different games for a one hour session).

1 http://www.re-mission.net/

4. The last point is the flexibility of the games. In order to keep the game attractive patients must be allowed to modify some parameters of the games (speed, number of balls, number of other players, colors, sounds).

Although these studies have been performed with children, the same observations and conclusion can be drawn with adults [71].

3.5 field of applications

The use of serious games in rehabilitation is a relatively new topic, therefore it is currently difficult to define precisely the field of appli- cation.

The field of applications, the most studied group of patients and

the kind of devices used are presented in Chapter 4 (page 29).

4

T H E U S E O F C O M M E R C I A L V I D E O G A M E S I N R E H A B I L I TAT I O N : A S Y S T E M AT I C R E V I E W

International Journal of Rehabilitation Research. 2016; 39(4):

277-90

Bonnechère B. ^1,2,3 , Jansen B. ^2,3 , Omelina L. ^2,3,4 , Van Sint Jan S. ¹

1 Laboratory of Anatomy, Biomechanics and Organogenesis (LABO), Université Libre de Bruxelles, Brussels, Belgium

2 Department of Electronics and Informatics – ETRO, Vrije Universiteit Brussel, Pleinlaan 2, B-1050 Brussels, Belgium

3 iMinds, Dept. of Future Health, Gaston Crommenlaan 8 (box 102), B-9050 Ghent, Belgium

4 Institute of Computer Science and Mathematics, Slovak University of Technology, Bratislava, Slovakia

29