Towards Optimized Virtual Reality Interventions for Improving Balance and Muscle Strength in Healthy Elderly

Summary

Independence at older age is compromised on a large scale by fall-related injuries. About 40% of those aged 65 and older, fall at least once per year. The high incidence of falls is attributable to risk factors such as age-related decreases in postural control and muscle strength. However, balance training programs can improve postural control and muscle strength and thereby reduce the risk of falls. Challenging weight shifts and muscle activation have shown to be important components of training programs that aim to reduce fall risk. Virtual Reality (VR) balance training may have advantages over regular exercise training in older adults. However, results so far are conflicting, potentially due to the lack of challenge imposed by weight shifts and muscular engagement in those VR training applications. The main aim of this research project was to design and test new VR training games for balance and muscle training in the elderly that optimally exploit age- specific movement requirements to challenge balance and muscle activation, as well as user requirements to increase intrinsic motivation.

In chapter one we conducted a systematic review to study the effectiveness of different VR-training interventions in improving balance in healthy elderly, and to investigate whether the balance challenge of VR-training has been assessed. A computer aided search of the databases PubMed, Web of Science and Cinahl was performed until August 2015, to identify RCTs that studied the effectiveness of VR-balance-training in healthy elderly. The systematic search resulted in 41 articles, of which 26 were RCTs. Most common outcome measures were the timed up and go,Berg balance scale and a battery of force plate measures. The results and effect sizes of different studies showed large variability but mean change scores rarely exceeded minimal detectable change. Studies analyzing the balance challenge imposed by the training were missing. An improved understanding of the balance challenge in different types of games is needed to adequately select and prescribe them in the context of balance training programs and fall prevention tools.

To investigate the challenge imposed on balance in VR balance games, the study in chapter two assessed to which extent two similar skiing games induce challenging weight shifts, as reflected in center of mass (COM) movements relative to participants’ functional limits of stability. The functional limits of stability (FLOS) represent the limits to which individuals can move their COM without the need to take a step. Thirty young and thirty elderly participants performed two skiing games, one on the Wii Balance board (Wiiski), which uses a force plate, and one with the Kinect sensor (Kinski), which performs motion tracking. The effect of the games, age groups and the progression over trials on COM displacement were tested with Generalized Estimated Equations. The results show that in all directions with anterior and medio-lateral, but not with a posterior component, subjects showed significantly larger maximal COM displacements during the Kinski game than during the Wiiski game. Furthermore, it was shown that during the skiing games young subjects quickly learned that they did not have to move their COM much in order to be successful in the game, whereas elderly did not show such a quick decrease in COM displacements. Moreover, elderly moved more in posterior directions whereas young subjects moved more anteriorly. These findings might be related to the game incentives, since posterior displacements slowed the game down, and anterior displacements speeded the game up. For the older participants, the game speed was more likely to be experienced as too fast than too slow. As a result, elderly moved their COM posteriorly to slow down the game. By appropriately adapting the settings of the games to the skills of the participants, such a quick decline in COM displacements might be prevented. These results emphasize the importance of assessing the movement challenge in games used for balance training. Similar games impose different challenges due to the controllers and their gain settings. In conjunction with game mechanics, controllers and their gain settings determine the movements that are elicited by the games.

Muscle weakness is another important risk factor for falls in elderly. In chapter three, the intensity and duration of muscle activity in Virtual Reality balance games was assessed. To this end, thirty young and thirty healthy elderly subjects played seven different VR balance games. Muscle activity of the m.Vastus Lateralis, m.Vastus Medialis, m.Soleus and m.Gluteus Medius was obtained using surface EMG. The processed EMG signals were divided in 200ms blocks, after which each block was categorized by its average normalized EMG activity i.e. >80%, 60-80%, 40-60% or <40% of maximum voluntary contraction (MVC). We calculated the total number of blocks in each category to score intensity, as well as the maximal number of Consecutive Blocks (MCB) >40% MVC, to identify prolonged muscle activity. We found that muscle activity during these VR training games was mostly below 40% MVC and prolonged activation was lacking. Although the activation levels during these VR games were low in general, faster movements could potentially provide a strength-training stimulus when longer activity bouts with more repetitions are included.

To be able to adjust VR training as to optimally benefit from improved motivation in VR training, we need to evaluate which games and underlying game mechanics are considered motivating by older adults. In chapter four, we studied 30 elderly who played eight different VR-training games, and afterwards filled out the Intrinsic Motivation Inventory (IMI). Differences in intrinsic motivation between games were analyzed using Friedman’s ranked ANOVAs. In addition, depth interviews were conducted according to the laddering technique, to unveil the underlying game mechanics that players preferred. Overall, IMI scores were relatively high for all games, indicating that these VR games might be effective for increasing intrinsic motivation. Wii yoga and Kinect Adventures were the best scoring games on all IMI subscales. Both games provided regular positive feedback. An important game mechanic was Variation, which showed a strong link to important values such as: Stay Focused, Improve Fitness and Health and Independency. Furthermore, the game mechanics Visual Feedback and Positive Feedback, which lead to an increased Drive to Perform, were perceived valuable. Seemingly contradicting, but both important attributes such as Speed versus Slow Movements, emphasize the importance of designing VR training that adapts to the skill level of the player. We have shown that games with different game mechanics can induce high intrinsic motivation. When designing or selecting VR balance training games for elderly, these game mechanics should be incorporated to optimize a positive user experience and increase intrinsic motivation.

Finally, novel VR balance games that are controlled with off-the-shelf hardware were developed based on the findings in previous chapters and recommendations for conventional training to prevent falls in healthy elderly. In chapter five the challenge in our novel VR balance games was evaluated. More specifically, we studied to which extent these games elicited challenging weight shifts and muscle activity, by evaluating muscle activity blocks and COM displacements relative to the participants’ functional limits of stability. Furthermore, the potential motivational pull of the VR training was evaluated by administering the intrinsic motivation inventory. Sixteen healthy elderly were recruited to play the novel games and two reference games that were found to be the most challenging ones for inducing muscle activity, or weight shifts in previous studies. The results show that we succeeded in creating motivating balance games that successfully challenged participants to elicit challenging weight shifts, by setting the game parameters to the functional limits of stability of each player. Our novel VR balance games induced COM displacements that resulted in medians of around 80% of FLOS or higher for all directions. Furthermore, the COM displacements in our novel games were larger for each direction compared to the reference game, although for Slingshot the left direction only reached significance at the third trial. Additionally, we improved the elicitation of consecutive muscle activity, by introducing long bouts of exercises, but it seems hard to elicit high intensity muscle activity through unloaded VR training.
We conclude that affordable hardware can be effectively used to create challenging and enjoyable VR training programs. VR training programs that are optimized to elicit challenging weight shifts and muscle activity should be further studied in longitudinal interventions. These longitudinal interventions should uncover the effects of optimized VR training on balance, muscle performance and eventually the reduction of fall risk in healthy elderly.

However, further research is needed to study the effect of optimized VR balance games that are developed and thoroughly tested to meet the requirements for effective training, of muscle strength, balance performance and finally a reduction of fall risk.

Overall, this thesis contributes to a better understanding of how different VR training concepts can be used to optimally exploit age-specific movement requirements to challenge balance and muscle activation, as well as user requirements to increase intrinsic motivation.