Game-Based Rehabilitation for Myoelectric Prosthesis Control

Background: A high number of upper extremity myoelectric prosthesis users abandon their 26 devices due to difficulties in prosthesis control and lack of motivation to train in absence of a 27 physiotherapist. Virtual training systems, in the form of video games, provide patients with an 28 entertaining and intuitive method for improved muscle coordination and overall better 29 control. Complementary to the established rehabilitation protocols, it is highly beneficial for 30 this virtual training process to start even before receiving the final prosthesis and to be 31 continued at home for as long as needed. Objectives: The aim of Results: Results show a significant improvement in fine accuracy electrode activation (p<0.01), 45 electrode separation (p=0.02) and endurance control (p<0.01) from Pre Gaming EMG 46 assessments to the Follow Up measurement. The deviation around the EMG goal value 47 diminished and the opposing electrode was activated less. Participants had most fun playing 48 the games when collecting items and facing a challenging gameplay 49 Conclusion: Most upper limb amputees use a 2-channel myoelectric prosthesis control. This 50 study has shown that this control can be effectively trained by employing a video game based 51 rehabilitation protocol. 52 53


57
The initial control of a myoelectric prosthesis can be a frustrating experience especially after 58 the already traumatic event of losing a limb. Due to the non-intuitive interface, which handles 59 a complex mechatronic system, the cognitive demand for controlling the prosthesis is high 60 and further delays the actual use of the device in everyday life [1] [2]. At least 50% of upper 61 extremity amputees report problems with the prosthesis control and functionality [3] [4], 62 which can be attributed to the need of receiving more training in handling the prosthesis [5] 63 [6]. By providing more training opportunities, the user can fully benefit from the technical 64 functions of the prosthesis. 65 In order to prepare the residual muscles as well as induce the specific brain plasticity, having 66 an access to a prosthesis itself is not necessarily needed. Regaining muscle strength and 67 coordination is a cognitively exhausting and repetitive process during which the proper 68 execution of movements is re-established using surface electromyographic (EMG) feedback 69 [2] [7] [8]. Through physiotherapy patients are presented with a variety of tasks promoting 70 the development of coping strategies for dealing with the activities of daily living and as such 71 introducing the embodiment of the prosthetic system itself. To effectively control their 72 prosthesis, patients need to learn how to properly contract their muscles. Especially the 73 strength of activation and isolation of a single muscle are important parameters [1] [9] [10]. 74 Even though the standard rehabilitation program offers direct functional benefits, its main 75 shortcoming is the lack of motivation for patients to pursue it without the involvement of a 76 therapist throughout the lengthy process. To add to the loss of functionality, patients may 77 suffer from post-traumatic depression, further decreasing motivation for rehabilitation [11]. 78 Transferring traditional EMG rehabilitation protocols to a virtual setting and incorporating 79 video games into the training process can potentially increase the patient's engagement and 80 perseverance [12]. It additionally provides medical professionals with quantitative data of the 81 one's performance. 82 Many studies report that the progress achieved during rehabilitation based on a playful 83 concept is faster and superior to conservative physiotherapeutic exercises [9] [13] [14] [15]. 84 These rehabilitation games are especially popular in older adults [16] and when treating 85 patients affected by stroke [17] [18] and Parkinson's disease [19] [20]. Various research groups 86 have addressed adding virtual games to an otherwise rather dull routine in the area of upper 87 limb amputee rehabilitation. There is, however, a difference between virtual or augmented 88 reality environments [9] [21] and using commercially available video games during therapeutic 89 interventions. The latter provides a higher accessibility and allows patients to easily set up the 90 games at home. This way, games can be chosen that are proven to motivate the players, also 91 over a longer period of time [15]. An example of such commercially available video games that 92 have been interfaced using EMG signals is "Guitar Hero" [14]. This game based on rhythm and 93 speed requires a fast reaction from the player and also an immediate transmission of the 94 processed EMG signals to the gaming system. Similar to this, a rehabilitation concept of stroke 95 patients uses modified version of the WiiMote control is used for rehab purposes of upper 96 limb amputees in which EMG signals are matched to the keys of the WiiMote [22]. However, 97 controls are only limited to two motions. Other groups chose a game similar to the arcade 98 classic "Pong" during which the muscle activity is mapped into a paddle motion that is used to 99 hit a ball into the opponent's court [23]. Although those approaches can be motivating, the 100 necessary actions are not very intuitive and are hardly transferable to the handling of a 101 prosthesis [24]. 102 This study presents an interface between a computer and a commercially available surface 103 EMG electrode system (Ottobock Healthcare GmbH, 13E200), which is commonly used for 104 controlling prostheses, to evaluate the short term effects on controllability after a video game 105 based rehabilitation protocol. 106 Here, compared to previous studies, participants are prompted to not only conduct repetitive 107 agonist and antagonist muscle activation, but also to train and exert sustained contractions 108 over a short period of time, perform precisely timed contractions and elicit simultaneous 109 contraction of both muscles or muscle groups -similarly to how patients would control a real 110 prosthesis in order to interact with their environment.   of the recorded EMG signal following the embedded filtering and rectification. 138 Participants were invited two times and had three test sessions in total: One Pre Gaming and 139 Post Gaming measurement, both conducted on the same day, and a Follow Up measurement 140 to evaluate short term retention rate two days later. 141 Initially, participants were instructed to perform three basic EMG assessments -the 142 provisional maximum voluntary contraction level, accuracy of electrode control and muscle 143 endurance. After a short break, participants were presented with three computer games in 144 randomized order. After each EMG controlled game, they were asked to complete a short user 145 evaluation survey considering the gaming experience. After the third and final game, a 146 modified questionnaire aiming at intrinsic motivation (IMI) was completed.

167
The assessment of precision control test evaluates the participant's fine EMG control accuracy. 168 The range of this test is adapted based on the outcome of the MVC test. Within the range of 169 10-90% MVC, the participant had to perform three trials per electrode of reaching a mark 170 using the EMG signal. For each electrode, the participant is asked to reach 30 randomly 171 preselected activation levels, and sustain them for 300ms each. The required level of 172 activation is indicated by a triangular mark on the EMG bar (see Figure  measure was the EMG signal deviation from the desired sine curve given as correlation r² [26]. 203

204
Three different open-source games were used in this study: a racing game, a dexterity game 205 and a rhythm based game. Each game features its respective control method (see Figure 3).

206
The general input mechanism was to substitute keyboard events with EMG activation. Since immersion into the games. The questionnaire is adapted to the study by changing the words 245 "working" and "doing" to "playing". It includes statements such as "I found the games very 246 interesting" and "I felt tense while playing". The statements are rated on a 7-point Likert rating 247 scale ranging from 1 ("no, not at all") to 7 ("yes, definitely"). This test was used to relatively set the maximum contraction limit for the following EMG tests. 263 It is given as average of the RMS electrode activation. MVC as measured 3 times, before 264 playing the games, directly after and before the Follow Up measurement. 265  Participants had to rank the 28 statements from 1 to 7, where 1 represents "I do not agree" 286 and 7 represents "I agree". The statements belong to one of 5 categories and the ranking is 287 averaged. An independent samples Mann Whitney U test was performed to describe the data. 288 All categories except "pressure" have a high desirable rank.

332
Participants showed an improvement in muscle endurance control (see Figure 6). This is true 333 for all but one participant, whose EMG activation expresses an offset and is shifted by a half Participants were asked to answer five Questions (Q1 -Q5) for each game (see Figure 8).

355
According to the user evaluation survey the most favorite game (derived from the score of Q1 356 and Q2) was the racing game Super Tux Cart followed by the rhythm game Step Mania 5. 357 According to Q3 and Q4 participants preferred to control the games with EMG signals only 358 and to perform different contraction lengths as well as co-contractions. In terms of motivation 359 (Q5), the game for dexterity ranked far behind the racing and rhythm game which were almost 360 equally well received. Additionally, participants had to rate the EMG assessments after each session. They were 372 asked how important they thought of the importance of the EMG assessment increased until 373 the Follow Up the EMG assessments were, and to rate the fun they had while doing them. As 374 can be seen in Figure 9, rating measurement (however, not significantly), while the participants 375 enjoyed them significantly less (p=0.002). Interestingly, there is a slight rise in rating the fun 376 factor after the Follow Up session again. The motivation aspects of training gamification are clear and probably the main advantage 398 compared to the conventional techniques. Also, it is reasonable to assume that certain 399 improvement of the EMG control could be observed by sole application of the listed EMG tests. 400 However, the prolonged exposure to them would certainly lead to the loss of interest, which 401 is sure to be maintained by the appealing context of a video game [15]. The transferability of the obtained results to the amputee population might be questioned, 405 since this study has been conducted strictly on healthy participants and this is yet to be further 406 explored. However, based on the outcomes reported in other myocontrol based studies [34] 407 [35] it is reasonable to expect that the patient group will perform similarly. 408 Since this study was a short term intervention, it can be seen as a proof of concept. Further 409 research will incorporate a long term evaluation of videogame-based interventions, as well as 410 additional exploration of the advanced control mechanisms such as those based on machine 411 learning approaches [36] [37] [38].