Abstract
Upper limb exercise is often neglected during post-stroke rehabilitation. Video games have been shown to be useful in providing environments in which patients can practise repetitive, functionally meaningful movements, and in inducing neuroplasticity. The design of video games is often focused upon a number of fundamental principles, such as reward, goals, challenge and the concept of meaningful play, and these same principles are important in the design of games for rehabilitation. Further to this, there have been several attempts for the strengthening of the relationship between commercial game design and rehabilitative game design, the former providing insight into factors that can increase motivation and engagement with the latter. In this article, we present an overview of various game design principles and the theoretical grounding behind their presence, in addition to attempts made to utilise these principles in the creation of upper limb stroke rehabilitation systems and the outcomes of their use. We also present research aiming to move the collaborative efforts of designers and therapists towards a model for the structured design of these games and the various steps taken concerning the theoretical classification and mapping of game design concepts with intended cognitive and motor outcomes.
Introduction
In the United Kingdom, stroke is the single biggest cause of disability 1 and there are around 150,000 first-time strokes every year. 2 Some 80–85% of survivors3,4 are affected by hemiparesis and only 5–34% eventually regain complete functionality of their upper extremity (UE). 5 Immediate hospital-based physiotherapy can capitalise on spontaneous recovery, 6 with 80% of recovery said to occur within the first 6 weeks. 7 However, hospital-based exercise programmes are costly, 8 resulting in shorter average inpatient stays: 23.7 days in 2008 to 19.5 days in 2010 in the UK, 9 for example. In addition to this, not only are patients spending less time in hospital but it would also seem that staff are stretched, with just 42% of physiotherapy, 16% of speech and language therapy and 84% of occupational therapy services meeting the stroke strategy staffing guidelines. 10 Furthermore, increasing lower limb movement is often a higher priority to enable discharge than increasing upper limb movement, meaning the hospital stay is often focused exclusively on improving mobility, posture, balance and gait. 11 In a 2014 study, only ∼22% of health care professionals (out of 264) and ∼17% of patients and carers (out of 123) agreed or strongly agreed that there is “good arm and hand therapy in hospital”. 12 Because of these problems, self-management approaches are an attractive prospect 13 if a patient’s care can be supported by home-based practice routines 14 focused on the previously omitted upper limb exercises. For the UE, rehabilitation is often focused on the improvement of performance in functional tasks through goal-directed practice and repetition.15,16 Repetitive movements associated with specific skills, if intensive and able to deliver feedback, can promote neural plasticity and motor recovery. 17 One method, constraint-induced movement therapy, encourages use of the more-affected limb by constraining the less-affected limb and is effective in facilitating both immediate and long-term functional improvements. 18 Another, bilateral training, requires the patient to perform mirrored movements with both limbs and can also be advantageous for motor retraining. 19 These repetitive exercises can, however, be uninteresting and tedious, reducing motivation for sustaining treatment 20 both at home and in hospital. Game-based therapy can be used to augment the aforementioned therapy techniques, 21 helping fulfil various principles of motor learning, 22 and is a practice which is constantly gaining considerable ground in terms of becoming a widely accepted approach in the area. This article is motivated by the belief that the principles of game design described herein, namely meaningful play, feedback, goals, rewards, challenge, difficulty, failure and flow, can be as integral to the creation of custom-made games for stroke rehabilitation as they are to the creation of commercial games. 23 Games can be a powerful therapeutic tool when integrated into rehabilitation, although the critical focus and resulting discourse typically omits the important and very salient link between these principles and rehabilitation. This may be particularly true in terms of the merit in both their use and the understanding of them.
Why games for rehabilitation?
A recent Cochrane review assessed the effect of virtual reality (VR) treatment on the recovery of motor, gait, balance, cognitive functions and activities of daily living (ADL) in stroke patients, finding that, particularly for upper limb treatment, the VR approach yielded better functional outcomes than conventional therapy. 24 Playing games has been shown to have many positive behavioural and physiological effects, leading to meaningful improvements across cognitive, motor and affective measures. 21 Neurophysiologically, the increase in physical movement, opportunities for bilateral activity and augmentation of biofeedback can help to induce neuroplasticity.25,26 Influencing neuroplastic adaptations in the brain following a stroke through behavioural/motor experience is important in protecting remaining neurons and producing/strengthening neuronal connections.27,28 The full scope of the advantages of VR/game-based therapy has been discussed in previous research,23,29,30 though one of the chief advantages is the enhanced motivation such interventions can grant across an ever-growing demographic, 31 increasing the amount of practice and, consequently, the degree of performance improvement. 32 It has been suggested that to effect any kind of cortical change, at least 20 hours of training, 33 or 300–400 repetitions of any one movement 34 are needed, although neuroplastic changes in primate brains have been shown to be the result of as many as 12,000 repetitions over 11 days. 35 Therefore, dependence on existing therapies alone in order to promote neuroplastic changes might not always be practical, 36 as these figures are far above what research submits are achievable dosages in clinical practice: 4 hours of training per week 37 and just 30 repetitions of a single movement in a traditional rehabilitation session. 38
In upper limb rehabilitation discourse, both commercial and bespoke video games are attractive prospective tools. Their application has been explored since the arrival of ‘exergames’ in the 1980s 39 and the consequent and exponential creation of games without entertainment as the primary focus, often termed as ‘serious games’. 40 In games for rehabilitation, ecologically relevant functional movements can be masked within the context of the game and presented as steps to accomplish an objective.23,29 For example, reaching, grasping, pulling and pushing movements were masked in the context of Slingshot, Space Race and Balloon Pop games for the ‘virtual glove system’. 41 In this manner, games for rehabilitation may also function as ‘persuasive’ games, a concept coined by Ian Bogost, 42 which states that games can be strong rhetorical tools in conveying ideas and altering the user’s interpretation of the kinds of activities or movements they invoke. The combination of less expensive, but higher-fidelity, technology has been facilitating a rise in the user acceptance and clinical validity of games for serious purposes, such as stroke rehabilitation, 43 although it has been suggested that the rapid pace of innovation provides a ‘moving target’ for researchers, presenting additional challenges. 44 Motivation is a hugely important factor in rehabilitation, often linked to better therapeutic outcomes, but is also a complex and difficult concept to describe. 45 It can be thought of as a psychological property that encourages a patient to elicit and/or sustain goal-directed behaviour, 21 and a motivated patient is one willing to expend effort without needing undue encouragement or complaining about the rigours of treatment. 46 We are intrinsically motivated when we find interest, enjoyment and satisfaction in an activity, 47 and the playing of games is fundamentally an intrinsically motivated activity. 48 Lack of motivation and slow progress have been noted as impediments for poor adherence to home-based practice, 49 although VR programmes deployed at home have reported improved user satisfaction and a marginal increase in performance. 50
Commercial systems
Commercial games are attractive because of the overwhelming market-derived evidence that these games are effective at motivating individuals to use them, 21 and some may offer features suitable for stroke rehabilitation straight out of the box. 51 The Nintendo Wii and Sony PlayStation EyeToy have been two popular choices in the past, not least because of their low cost and the burgeoning research supporting their application. 25 Both systems have been used to play a number of commercial games, primarily EyeToy: Play (Sony) and Wii Sports (Nintendo), both collections of short, simple mini-games intended for children. Their simplicity reduces concern about cognitive overload 52 and attractive, bright graphics make them instantly relatable, although adult participants from one recent study did find the childlike aesthetics of Wii Sports unappealing. 53 In one study, statistically significant improvements in Fugl-Meyer Assessment and Motricity Index scores were noted after just six 30-minute sessions using Wii Sports, 54 while one of only a few randomised controlled trials in this area, which focused on the use of Wii Sports and Cooking Mama in comparison to recreational therapy involving table-top games, noted improved Wolf Motor Function Test scores for the Wii group at 4 weeks post-intervention. 55 User acceptance and functional improvement were also noted in several other smaller-scale studies.56–59 While the Wii requires precise aiming and controlled motions, video-capture systems such as the PlayStation EyeToy can illicit gross, ongoing movements,60,61 whilst allowing for more naturalistic, unencumbered interaction. 62 A randomised controlled trial analysed the effect of the EyeToy on 20 stroke patients over 4 weeks. 63 In addition to conventional therapy, one half received 30 minutes of treatment with the EyeToy, while the other half watched the games for the same duration. Significant improvements were noted in the experimental group via the Functional Independence Measure and the games were also deemed motivating and enjoyable, a feeling expressed in another study featuring the EyeToy: Play games Wishy Washy and Kung Foo and the IREX, a rehabilitation-specific video-capture gaming system. 64
The more powerful Microsoft Kinect grants highly accurate 65 full-body motion tracking and gesture/scene recognition. 66 Due to the abilities of this device, much work with it has been in the creation of intuitive applications focusing on technical possibilities. 67 One trial did evaluate the games Kinect Sports (Rare) and Kinect Adventures! (Good Science Studio), aesthetically similar to the EyeToy and Wii games, with 14 chronic stroke patients 68 and found Kinect therapy helped improve Functional Independence Measure scores, plus shoulder and elbow flexion and extension. Sophisticated, modern technology like this, such as a sub-millimetre accurate 69 optoelectronic device called the Leap Motion Controller and especially touchscreens, are leading to ‘Natural Interaction’, a type of human–computer interaction which allows for faster, more accurate and more intuitive communication with game systems. 70 A popular commercial game which has been adapted for stroke patient use with both the Kinect and the Leap Motion is Fruit Ninja (Halfbrick Studios), which recently reached 1 billion downloads. 71 The game is traditionally played on touchscreen devices with the player swiping their finger across the screen, chopping up fruit as it is launched across. When used with the Kinect, Fugl-Meyer Assessment and Wolf Motor Function Test scores were significantly increased 3 weeks after training and at a 12-week follow-up, while functional magnetic resonance imaging (fMRI) results showed contralateral primary sensorimotor cortex activity, demonstrating functional neuroplasticity. 72 Fruit Ninja played with the Leap Motion yielded a high correlation between game scores and Fugl-Meyer Assessment/Box & Blocks Test scores in one trial 73 and improvements in hand abilities and grasp force in another. 74 Also, with the Leap Motion, a group recently created three virtual rehabilitation tasks: a virtual rendition of the Nine Hole Peg Test, Cotton Balls and Stacking Blocks. 75 All three were well received by the eight therapists that tested them, highlighting the possible usefulness of modern devices such as these in stroke rehabilitation.
The combined results of studies applying commercial games to rehabilitation are promising, not just in terms of their cost, safety and ability to elicit intense arm movements, but also regarding their positive reception and the consequent enjoyment.76,77 Because the game designers try to maximise the likelihood that these games will be appealing and played for a large number of hours, and, consequently, commercially successful, 21 they are designed according to a number of well-researched principles. However, owing to the omission of medical considerations during development, and despite accommodative design for a diverse target audience, commercial game systems typically lack the levels of adaption needed to fit the heterogeneous nature of stroke-related dysfunctions,43,78 making play, or even control of input devices, difficult and frustrating. 20 One possible solution is to re-purpose commercial games with more appropriate input devices, as has been attempted with Fruit Ninja73,74 and Tetris. 79 A much more credible and prevalent solution, however, is the design of custom-made games and game systems, built to better accommodate patients’ physical needs while concomitantly providing games which are designed by the same principles. This could be key in creating patients as engaged with rehabilitative games as ‘hardcore gamers’ are with the latest commercial games.
Game design concepts and their use in rehabilitation
Examples of how game design principles exist in various custom and commercial games/game systems.
Meaningful play and feedback
‘Meaningful play’ emerges from the relationship between player action and system outcome, apparent to the player through feedback. 81 This interaction between the player and procedures of the game system is what primarily shapes the player’s experience, 42 and the goal of successful game design is the creation of meaningful play. 82 Playing a game means making choices and taking actions, with every action taken resulting in a change affecting the overall game system. 83 The principles of an ‘action plan’ to perform a specific movement 84 and the anticipation of feedback and outcomes as a result may also be critical to motor learning. 85 For play to become meaningful and for the continued incentive to keep playing, the player must be able to perceive not only the immediate result of their action, but also how the outcome of the action was integrated into the larger context of the game. 82 Several examples have cited this concept as an important aspect in the creation of games for stroke rehabilitation.86,87
The concept of feedback is paramount in creating and maintaining meaningful play, since feedback is how a player perceives the results of their actions. 88 Without a quantifiable advantage/disadvantage for performing an action successfully/unsuccessfully, the experience will be less engaging. 89 It is important in rehabilitation also, as increased levels of feedback, whether graded (degree) or absolute (correct/incorrect), 90 can lead to increased training effects and enhance motor learning. 91 Positive feedback (correct performance) can enhance motivation, self-efficacy and compliance, while negative feedback (incorrect performance) can facilitate skill improvement.92,93 Lack of negative feedback can lead to frustration, such as in the case of the Ammi Interactive Rehabilitation Touch games Memory and Puzzle, 94 where players were left confused as to why their actions were not yielding positive in-game outcomes. As well as showing the players the immediate results of their actions, 95 such as trace and aura effects in the tangible user interface system Elements, 96 progress identifiers such as an on-screen score can assist in demonstrating progress towards objectives. 97 A person’s own sensory-perceptual information, available as a result of their actions, is known as intrinsic feedback92,98 and can be mediated through vision (text, icons, scores), audition (sound effects, dialogue), or touch (haptic feedback, pressure, vibration). Advanced touch feedback was demonstrated with a collection of custom Kinect games played with a haptic glove, 99 which activated various palm- and finger-situated vibration motors to produce real-time vibration patterns based on the virtual interaction. During gameplay in rehabilitation, intrinsic feedback is often supplemented by extrinsic or augmented feedback, which can include verbal encouragement, points, charts, or a video replay. This can prove beneficial if cognitive impairments prevent appropriate transmission of intrinsic elements98,100 and may also enhance control, motivation, retention and transfer, if delivered at the patient’s own pace. 101 Indeed, the use of extrinsic feedback is advocated in much of the literature, categorised as either Knowledge of Performance (KP) or Knowledge of Results (KR). KP refers to the movement characteristics which lead to the performance outcome, while KR refers to the outcome of performing a skill or achieving a goal. 93 There is evidence to suggest that the use of KP during repetitive movement practice results in better motor outcomes than the use of KR, as the latter can discourage variety in learning strategies (e.g. active problem-solving activities) and can lead to feedback dependency. 92 Furthermore, feedback that is prescriptive rather than descriptive can make use of the brain’s declarative learning process, resulting in factual knowledge that can be recalled from one’s long-term memory. 102 A system that garnered the use of both KP and KR was the Gertner Tele-Motion Rehab system, 103 which provided feedback about the quality of upper limb motion and compensatory movements (KP), including both descriptive online data and prescriptive remote contact with a clinician, as well as ongoing feedback on the game progress and scores (KR).
Goals and rewards
Goals in games can range from short-term, typically quick navigational beacons designed to provide a constant source of satisfaction, 82 to long-term, more high-level, overriding objectives. 104 Low-level goals are usually bolstered by others which may take many play sessions and improvement in skill to achieve. 105 Circus Challenge (Limbs Alive), for instance, incorporates 100 separate upper limb actions into multiple difficulty levels over 10 game scenarios, including knife throwing and juggling. 106 While the short-term goal may be to achieve a 3-star rating on each of the current difficulties for each mini-game, the long-term goal would be to achieve 3-star ratings on every difficulty level for every mini-game, thereby achieving maximum scores for all 100 movements. The accumulation of these goals is a powerful means in order to get the player invested and involved 81 and can inspire participants to perform tasks for personal gain or satisfaction via incentives and rewards,107,108 as well as helping cultivate meaningful play by allowing the significance of an action to be understood in terms of progress towards winning the game. 82 Goals and clear instructions have been found to increase motivation, 46 while a lack of goals and instructions can result in confusion and frustration. 45 Goal-directed movements, either in-game or out, are essential for stroke rehabilitation,11,15,16 particularly ‘concrete’ goals, which focus upon physical interactions with objects or persons. 100 The virtual re-imagination of real-world activities, such as supermarket shopping 109 and posting an envelope 110 have been incorporated into game-based rehabilitation. In addition, hammering a nail and pouring a glass into a carafe are central gameplay aspects of the VRRS system from Italy, which yielded increased Fugl-Meyer Assessment and Functional Independence Measure scores in a 376-patient study. 111
Game rewards have been categorised by various researchers. Hallford et al. 112 claimed rewards to be either of glory, sustenance, access or facility, while Bateman 113 further categorised them as either currency, rank, mechanical, narrative, emotional, new toys, new places, completeness, or victory. The type of reward attained from the achievement of a goal, while different across game genres, is even more different with regards to what is desirable for different players. There is no way of elucidating an ideal reward for any one player, whether healthy or recovering from stroke, as this relates to the player’s ‘player type’, 114 the reward’s form/characteristics and how the player wishes to utilise it. 115 However, this mentioned, all gameplay is motivated by rewarding experiences. 21 While rewards in games traditionally come from level progression or goal attainment, 116 games for rehabilitation should reward engagement and effort over success. 39 The intrinsic/extrinsic motivation to receive particular personal, competitive or material rewards has also been discussed in research. 117 One of the main systems of reward in rehabilitative games is a difficulty level/star system. The Elinor gaming system, 118 for instance, allowed players to earn performance-contingent stars for each level before moving on. If the player failed a level five consecutive times, they were returned to the former level to ensure that reaching the limit of their ability did not prevent ongoing achievement. Recently, research has given insight into the neuroscience of reward and motivation. Activity in the nucleus accumbens, the area of the brain associated with the pursuit of rewards and pleasure, 119 has been found to scale linearly to the probability of receiving a reward, 120 as well as the release of dopamine, associated with reward-based learning, feelings of pleasure/enjoyment and motivation to perform specific behaviours.21,89 The properties of the dopaminergic reward system that underlies gameplay, however, may change as a function of disease or injury. 21 Psychological and physiological rewards such as these have been found to result from decision-making, 119 uncertainty 121 and exploration, 122 all factors inherent in video games.
Challenge and difficulty
Typically, games are designed to be playable by individuals with vastly different levels of motivation, experience, skill, and different play styles and strategic approaches. 123 Individuals left with heterogeneous cognitive/motor dysfunctions, however, require an even wider-ranging scope of difficulty. 124 For this reason, many games for rehabilitation, such as the Kinect-based Jewel Thief 125 and the Rutgers Arm 126 include pre-game calibration tasks to assess range of movement and then form the proceeding game parameters on these measurements. Despite precautions, failure is still a prominent concern and the risk may be more sizeable in rehabilitative games. Some games prevent explicit failure, such as a recent mirror-training game 127 which skipped over failed hand posture tasks to avoid detrimental feelings of inefficacy. However, it should be noted that failing can grant opportunities for satisfaction and proof of improvement when obstacles are overcome, 128 if handled in a positive way. Failing positively can increase excitement and eagerness to retry a task, 129 and even frustration can contribute to a pleasurable and cathartic experience. 123
Game difficulty should begin low to match levels of ability and familiarity, 21 gradually increasing alongside player skill,104,130 although this cannot be at a gradient ideal for everyone. Methods such as selecting easy, medium or hard play, or using increasingly difficult levels, are both prevalent forms of adaptable or predetermined difficulty, while in-game difficulty adjustment is adaptive or dynamic difficulty. 131 From a motor learning perspective, an optimal level of challenge and ‘desirable difficulties’ are important,85,132 and the aim of in-game difficulty adjustment is to achieve a harmonious balance between challenge and skill. This is required to achieve a state of flow, 133 the psychological description of energised focus, full immersion, involvement and enjoyment. 134 Flow is important in the realms of rehabilitation, as an activity will be deemed rewarding if an individual can use their sensory and physical potential in a novel/challenging way. 135 The theory has been applied extensively to games too, particularly with GameFlow, 95 a contemporary manifestation, outlining core elements of player enjoyment. Autonomous adjustment of difficulty contingent on current performance and success, via heuristic observations for example, can keep the player in a state of flow by increasing the challenge when the game appears too easy and decreasing it when it appears too hard, 136 helping to avoid anxiety or boredom. 137 Furthermore, maintaining the balance and extending the time between a player losing or winning is greatly advantageous for maximising potential for functional improvements. Intelligent control of difficulty was achieved on the Rehabilitation Gaming System (RGS) by the Personalised Training Module 138 and the Intelligent Game Engine for Rehabilitation, used with a Kinect Fruit Catcher game, 139 to name two examples. Autonomous control was achieved in earnest by the Adaptive Mixed Reality Rehabilitation System’s Adaptive Therapy Sequence, which guided the patient through an entire rehabilitation session, depending upon their performance and kinematic improvements. 140
Custom systems
Many custom game systems have used game design principles extensively to influence their development, although very few have cited these in detail. A system that did highlight a particular affinity for design, according to these principles, was the RehabMaster™ (D-Gate, Seoul, Korea), a gesture-based system played with upper limb and trunk movements while seated opposite a large monitor. 87 Four games were designed that not only used goal-directed movements inspired by the Fugl-Meyer Assessment, Action Research Arm Test and Motricity Index, but were also bolstered by knowledge of meaningful play and an optimal level of challenge to try and inspire maximum enjoyment, immersion and intrinsic motivation. Three user groups were involved in designing the RehabMaster – stroke patients, occupational therapists and physiatrists – showing that, far from being supplementary or a novelty, the development and use of game systems such as this may represent the successful collaboration of industry, design, research and health care. 23 The considerations for all stakeholders associated with a patient’s care within a multidisciplinary approach is important, as focus on the patient alone could compromise the success of the technology. 39 In the games Underwater Fire, Goalkeeper and Bughunter, the speed, size and trajectory of the objects was controllable and adaptable depending on current levels of success or failure, similar to the game Spheroids on the aforementioned RGS. 138 Unlike some systems, however, an occupational therapist sat next to the monitor, having access to real-time performance data and controlling the difficulty. They could also provide extrinsic feedback and physical assistance. A training feature and the fourth game Rollercoaster both tasked the patient with mimicking the positions and movements of an on-screen avatar as accurately as possible, 141 promoting bilateral movement and mirror training. The use of the RehabMaster was found to result in functional improvements in both chronic and acute/sub-acute stroke patients, with positive reports of skill development, improved immersion and motivation, when compared with conventional therapy. 87 In addition, improvements in health-related quality of life, particularly role limitation due to emotional problems and depression, were noted in another trial. 142 A usability test yielded results indicating improved attention and a ‘flow experience’ for the patients, characterised by control, attentional focus, intrinsic interest and curiosity, which may have minimised the drop-out rate. 87 The highly realistic and immersive environment may have helped facilitate a sense of presence, 143 which may have a positive correlation with task performance. 144
The Message-in-a-Bottle (MIB) game 145 was played on a large projector screen using a HapticMaster (Moog FCS) robotic device. Research undertaken which inspired the design of the MIB system alluded to the importance of realism (realistic setting, sound and graphics), customisation, rapid absorption rate, winning and losing and a variety of control options, 146 while objectives requiring higher-order thinking skills and creative problem solving 147 were also deemed significant. In MIB, the player is presented with highly realistic views of a tropical island and tasked with using the haptic device to pick up virtual bottles that float towards him/her, the size, weight and speed of which can be altered according to the individual’s ability. After this initial physical task, the message in the bottle presents one of an available 300 questions, split over nine categories, before asking the user to place the bottle in one of two virtual baskets, according to which answer he/she thinks is correct. The combination of both motor and cognitive challenges is perhaps the most important feature of the MIB. Later levels require more difficult motor tasks, and eight motor difficulty levels are stepped through according to the continual observation of task performance and biomechanics. As well as attempting to focus the resultant game on principles of good game design, the system was created with the objective of increasing levels of intrinsic motivation. An Intrinsic Motivation Inventory questionnaire demonstrated the system’s success in this area, motivating players to play for more beneficial amounts of time.
Extensive work has been undertaken in this field over the last decade by James Burke and his colleagues at the University of Ulster, Northern Ireland. Game design principles, particularly the ideas behind challenge, handling failure and meaningful play, were extensively explored and integrated into a number of custom-made games.51,148 Sensor-based games included Fishing, Catch the Orange and Whack-a-Mouse,148 use of which contributed to improved functional measurements with eight stroke patients, after 10 sessions and also at a 6-week follow-up. 149 The latter, a ‘Whack-a-Mole’ style game, adapted both the length of time mice would appear out of the holes and the locus of them upon performance data and level progression. A dog was also added in later levels which needed to be avoided, adding elements of visual discrimination and problem solving. An increase in speed/difficulty was also a feature of the vision-based game Rabbit Chase 150 and deemed an important factor for overall enjoyment.148 For Rabbit Chase, Arrow Attack and Bubble Trouble, an initial calibration task tested the reach ability of the patient in eight directions. 51 These games were designed to include specific features which could combat certain afflictions; for example, time limits and direct feedback were used to combat coordination/timing problems. 151 Brick’a’Break and Shelf Stack were a pair of augmented reality games developed to satisfy the ability to customise the tasks and handle failure positively/conservatively. 86 For Shelf Stack, the player was encouraged to simply increase their score over play sessions, so they essentially could not fail. In Brick’a’Break, there was no time limit (unlike the popular game Breakout on which it was based) and only a short pause when the ball was missed (without any further penalties) in order to avoid discouragement from playing.
Towards a model for structured design
Proposed features of importance when designing game systems for stroke rehabilitation according to published survey papers, taxonomies and frameworks.
Prominent features of a system for ‘elderly entertainment’ included such well-known factors as motivational feedback, the creation of new learning experiences and social activity. A simple objective/interface, like the toned down colours and sounds of the Elinor system 118 to avoid cognitive overload and fatigue, was deemed as important and, additionally, presenting a sensitivity to decreased sensory acuity and slower responses, particularly important for individuals who are either elderly or suffering from cognitive impairments. An extreme example of this would be the vision-based system from Hadassah University Hospital, Israel, which replaced the impaired limb with a virtual representation and accentuated the user’s limited movements into amplified full-range movements, thus promoting recovery. 156 Appropriateness of genre was also deemed an important factor, and many systems have used reimagined versions of popular game concepts such as Pong, 157 Tennis, 118 Breakout, 86 Guitar Hero 158 and Snake.118,127 The most popular games out of the 15 included on the Elinor system were proven to be based on well-established games. 118 Genre was also an element of another taxonomy by Rego et al., 159 which was recommended for the analytical comparison and determination of serious games for rehabilitation, but could also be applied to the design too.
The last element of ‘elderly entertainment’ was appropriate cognitive challenge, an element mentioned by other research in this area,160,161 particularly in relation to the learning outcomes achievable from engaging cognitive mechanisms. 22 Cognitive elements in games for stroke rehabilitation are included to either break the monotony of repetitive movements, 145 to evoke timed responses, hand-to-eye co-ordination 161 or, finally, to inspire short-term visual and memory challenges, such as the Rutgers Arm Card Island game. 162 Furthermore, integrating varied problem solving may be a powerful pedagogical tool, helping also develop high-order thinking skills. 163 Cognitive challenge was one of three key attributes outlined by Alankus et al., 160 in addition to social context and motion type, as a prerequisite to the design of nine stroke rehabilitation games, incorporating differentiating input methods, cognitive challenges and competitive/collaborative elements. Motivation was outlined as an important factor, and the use of both therapist-determined difficulty and automatic difficulty adjustment were advocated, as well as the inclusion of non-player characters and storylines. Several of these findings have alluded to the concept of socialisation152,159 and, indeed, the ability to play these games with or against other players seems to be beneficial for motivation, engagement, learning and may be central to player enjoyment. 164 Sociality was also an aspect of a recent and exhaustive investigation into the advantages of design principles in physical therapy. 21
During the creation of the 3D VR Activity Station at the University of Göteborg, 165 Goude et al. 151 mapped specific and common post-stroke motor impairments against various game design principles and features. 108 Focusing on the techniques used to combat certain impairment, precise game features were postulated to be of benefit, if included, to the correlative dysfunction. For example, rhythm-based actions could be used to combat timing problems associated with discoordination, and precise manoeuvring and dexterity-based actions could be used to diminish scaling/proprioception-related problems. Only a subset of the taxonomy was published, but the development of the 15 games which came pre-loaded on the 3D VR Activity Station were influenced by this research. This work was also cited in the creation of several vision-based serious games at the University of Ulster,51,148,150 and Burke’s work in this area (previously mentioned) presents a very functional example of how existing research into game principles can be utilised in the creation of custom-made games. Resulting from the extensive work at the University of Ulster, a vast framework called the GAMER framework was recently presented by Charles and McDonough. 161 GAMER categorises components relating to the player, primarily their therapeutic goals in terms of transferability to ADL, coordination, extremity focus, cognition and movements, and components relating to the interface, such as the input, output, deployment and ability to data log and the audio and visual consistency in the design of the user interface itself. Components relating to the game itself are the most exhaustive, comprising elements such as varied, easy to learn/play game content, concepts of meaningful play comprised of feedback, goals and targets and play rewards and, finally, challenge, comprised of appropriate difficulty, failure and competition.
Conclusion
This article has demonstrated how game theory can and has been applied to the creation of games for upper limb stroke rehabilitation. Low motivation and engagement with physical therapy are very real and prominent concerns, 166 but encouraging evidence has shown that games can be a powerful therapeutic tool when intelligently integrated into rehabilitation and can help combat these problems.23,24,29,77 In addition to improving compliance, games can also fulfil principles of motor learning 22 and induce/modulate neuroplastic adaptations in the brain. 26 Both commercial and custom-made game systems have been re-purposed/developed for use in upper limb stroke rehabilitation, although where one type seems to excel, the other suffers and vice versa. This article includes a brief and recent history of commercial interventions in this area, although their efficacy, accessibility and cost-effectiveness must be offset against the adaptive and sometimes improvisational nature of their application. Their detriments seem to support the case for custom-made interventions.
The creation of a video game is a complex and multifaceted process, and the gold standard for their modern design is the result of over half a century of painstaking research and exhaustive development. As a result of an extensive literature survey, a number of core factors of game design were distilled and discussed in detail: meaningful play, feedback, goals, rewards, challenge, difficulty, failure and flow. These factors could be as integral to the creation of custom-made games for stroke rehabilitation as they are to the creation of commercial games for the general public, particularly if these aspects, which scientifically contribute to motivation, engagement and commercial success, are combined with technologically advanced control systems that compensate for motor impairment. The increasing success of the games industry is a testament to the ability of such design factors to maximise engagement and create meaningful, addictive experiences for players, although research on these factors in the area of stroke rehabilitation is relatively limited. As a result, few custom-made games cite these principles in their body of research, but those which do have been presented and discussed in this article.
The so far limited research elucidating these game design concepts and advocating their utilisation in stroke rehabilitation has also been presented, and this evidence in particular helps draw a salient and inextricable link between the two fields. Further work is needed, however, to cement the two fields together and further highlight the importance of the relationship between the motor/cognitive outcomes achievable from the design principles, to the most important aspect of the entire area: the stroke survivor. Further work is also needed to empirically demonstrate the significant advantages of using video games in therapy. While they appear to have much promise, evidence associated with the application of commercial games is typically considered low quality due to small participant numbers, inconsistent and poorly reported results,24,77 plus an erroneous focus on time spent in therapy over the number of movement repetitions. 21 In addition, evidence associated with the application of custom-made games seems to suffer too, from data too small to draw strong conclusions but encouraging enough to support future exponential development.30,76
Footnotes
Declaration of conflicting interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Ian Swain is employed part-time by Odstock Medical Limited who part fund this project, in addition to his appointment at BU. He also holds a number of stocks in Odstock Medical. Dr Mecheraoui is full-time at OML.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was conducted within a PhD entitled Enhanced rehabilitation of the upper limb following stroke by an adaptive Virtual Reality and FES approach, co-funded by Bournemouth University and Odstock Medical Limited.
Authorship statement
Nathan Barrett is a PhD student at Bournemouth University, part-funded also by Odstock Medical Limited where he is based. He was responsible for the research, first manuscript draught, manuscript review and editing. The supervisory team consists of Ian Swain, Clinical Director, PhD, at Odstock Medical Limited and part-time Professor at Bournemouth University (SciTech), Dr Christos Gatzidis, Principal Academic at Bournemouth University, and Dr Choukri Mecheraoui, a Design Engineer with Odstock Medical Limited.