G Model JSAMS-1157; No. of Pages 5
ARTICLE IN PRESS Journal of Science and Medicine in Sport xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Journal of Science and Medicine in Sport journal homepage: www.elsevier.com/locate/jsams
Original research
Physiological and psychophysiological responses to an exer-game training protocol Shaw Bronner a,∗ , Russell Pinsker b , Rutika Naik c , J. Adam Noah d a
ADAM Center, Dept. of Physical Therapy, Movement and Rehabilitation Sciences, Northeastern University, USA ADAM Center, Health Sciences, Long Island University, USA ADAM Center, Division of Sports Sciences, Long Island University, USA d Brain Function Laboratory, Department of Psychiatry, Yale University School of Medicine, USA b c
a r t i c l e
i n f o
Article history: Received 28 October 2014 Received in revised form 18 January 2015 Accepted 5 March 2015 Available online xxx Keywords: Energy expenditure Exercise METs Engagement Dance Video games
a b s t r a c t Objectives: Exer-games and virtual reality offer alternative opportunities to provide neuro-rehabilitation and exercise that are fun. Our goal was to determine how effective they are in achieving motor learning goals and fitness benefits as players gain experience. Design: We employed a repeated measures design to determine changes in physical exertion and engagement with training. Methods: Fourteen healthy adults trained on the XBOX Kinect video game Dance Central using a skillbased protocol to examine changes in energy expenditure (EE), heart rate (HR), METs, limb movement, game proficiency, and player engagement in initial, post-training, and transfer-testing of a full-body dance exer-game. Data were analyzed using repeated measures analysis of variance, p < 0.05. Results: Both EE, HR, and METs increased from initial (EE 4.89 ± 1.35, HR 103 ± 18, METs 4.25 ± 0.72) to post-training (EE 5.92 ± 1.25, HR 110 ± 15, METs 5.05 ± 0.75) and were greatest during transfertesting (EE 6.34 ± 1.35, HR 115 ± 17, METs 5.42 ± 0.88, p ≤ 0.001). Proficiency, measured by game scores, also increased from initial to post-training and transfer-testing (p ≤ 0.002). Limb movement and player engagement remained unchanged. Conclusions: It is important to understand whether player physiological and psychophysiological responses change with continued game-play. Although Dance Central involves whole-body movement, physical exertion remained at moderate levels after training. As exer-game and virtual reality systems move from their initial novelty, research about how players react to continued involvement with a game can guide game developers to maintain a freshness through game progression that preserves the participant’s attentional focus, minimizes attrition and maintains a prescribed level of energy exertion. © 2015 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
1. Introduction Exercise video games (exer-games) and virtual reality (VR) based systems are increasingly being used for physical activity and rehabilitation goals in the treatment of different neuropathologies such as traumatic brain injury,1 stroke,2 spinal cord injury,3 and Parkinson’s disease.4 High intensity task-specific exercise promotes neural reorganization, neuroplasticity, and improvements in function,5 as well as improvements in health-related physical fitness and cognitive function.6 In fitness training or neurologic rehabilitation, exer-games and VR training offer the potential to
∗ Corresponding author. E-mail address: [email protected] (S. Bronner).
deliver these gains in a cost effective manner, in socialized group activities, tele-rehabilitation, and on-line interactive exer-gaming.7 To achieve exercise and rehabilitation goals, improved performance requires intensive skilled practice, focused attention, progressive challenges, immediate feedback, and knowledge of results.8 To reach these goals, exercise adherence is critical, requiring that the exer-game or VR is engaging and players are motivated to return. Additionally, exer-games and VR must record data on the level of play and performance in order to ascertain whether the participant attains specific goals. If the outcome is improved physical fitness, do the players perform at a high enough intensity with ongoing progressive challenges as they improve? If the goal is rehabilitation-related, is there accurate tracking and feedback of performance related to functional outcomes? Research to date on exer-games or VR systems for physical fitness and neuro-rehabilitation has primarily focused on proof of
http://dx.doi.org/10.1016/j.jsams.2015.03.003 1440-2440/© 2015 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Bronner S, et al. Physiological and psychophysiological responses to an exer-game training protocol. J Sci Med Sport (2015), http://dx.doi.org/10.1016/j.jsams.2015.03.003
G Model JSAMS-1157; No. of Pages 5
ARTICLE IN PRESS S. Bronner et al. / Journal of Science and Medicine in Sport xxx (2015) xxx–xxx
2
concept in small samples. In the case of exer-games for the goal of physical fitness, the literature does not reflect energy expenditure (EE) of well-trained advanced players. Rather, participants are briefly ‘familiarized’ prior to determining EE during gameplay. The majority of research reports that participants exert 3–5 metabolic equivalents (METs) (light to moderate activity) when they play exer-games such as Wii Fit, Wii Sports, Dance Dance Revolution, Kinect Dance Central, and Eye Toy activities.9–11 Others have demonstrated moderate to vigorous exercise intensity (4.2–7.1 METs).9,12,13 Therefore, exer-games have potential to stimulate adequate conditioning of the body, but changes with skill development are unstudied. Games that use sensors that measure full body motion or require activities that use the large muscles of the lower extremities result in greater physical activity.14 The incorporation of whole body movement may also result in greater player engagement while playing exer-games through immersion.15 Player engagement is important because a positive experience makes it more likely that a player will return to play again. Engagement is also important for any training program (exercise or rehabilitation), as attrition rates tend to be high.16 We focused on two aspects of game-play, including both physiological and psychophysiological measures relevant to any type of exercise. To study the effect of exer-game training on these variables, we selected the interactive dance game, Dance Central for the Xbox 360, as it uses the Kinect sensor to process full body motion capture and determine performance scores. Dance Central offers a number of choreographed dance styles and music, potentially appealing to a range of players. We hypothesized that compared to an initial baseline trial following a “familiarization” period, player training would result in the following: (1) increased physical exertion; (2) increased upper and lower extremity movement; (3) improved game proficiency; (4) demonstration of motor learning by transfer of proficiency; and (5) greater player engagement.
2. Methods Fourteen healthy participants (seven women, seven men, ages 26.6 ± 9.5 years, height 172 ± 11 cm, mass 69 ± 19 kg), recruited by flier at a local university, volunteered for this study. A prior power analysis determined for a repeated measures design, with one group, three measurements, a medium effect size (f = 0.50) and alpha = 0.05, a sample size of 12 was required. All individuals were free of recent musculoskeletal injury, visual or auditory impairment, color blindness, neurological disease, or heart condition, based on their answers to a questionnaire for participation inclusion or exclusion. Each participant had minimal or no experience playing Dance Central on the Kinect prior to participation. During training and recording sessions, participants wore workout clothing. Each participant gave informed written consent and the Institutional Review Board for human ethics granted ethical approval for this study. Participants trained on Dance Central (Harmonix Music Systems, Inc., Cambridge, MA) for the Xbox 360 gaming console and Kinect motion sensor accessory (Microsoft Corp., Redmond, Washington, U.S.). The gaming console was attached to a 50-in video monitor. Participants played in a 1.8 m × 3.6 m square, approximately 2 m from the Kinect motion sensor and video monitor. During Dance Central game-play, players stand in front of the monitor and follow an avatar dancing on the screen. The gaming console includes a motion sensor, skeletal estimation sensor, with facial and voice recognition that captures the player’s movement and compares it to those of the z-score underlying the avatar movements.
Game scores are based on movement and timing accuracy of the individual compared to the z-scores. To determine metabolic expenditure via indirect calorimetry, participants were outfitted with a portable gas analysis system (K4b2 Cosmed Inc., Rome, Italy) comprised of a small metabolic analyzer, battery pack, and face-mask and a chest strap heart rate (HR) monitor (Nike, Beaverton, OR). The facemask was attached to a turbine flow-meter allowing for real-time collection of VO2 values. The K4b2 unit (∼925 g) was strapped to the participant using the manufacturer’s harness. Prior to each testing session, the K4b2 was calibrated according to the manufacturer’s instructions. The K4b2 has been well validated for measurement of metabolic expenditure.17 During testing, VO2 , VCO2 , and HR were monitored continuously and recorded using the K4b2 breath-by-breath analysis. The beginning and end of each game-song were marked using the K4b2 marker feature. To determine limb movement, we collected kinematic data using a 10-camera motion capture system (Vicon, Oxford, UK), sampled at 120 Hz. Participants were outfitted with a full-body marker set using the Vicon Plug-In Gait model. To determine changes in player proficiency, game scores were recorded during data collection. To determine whether player engagement altered with greater experience with the exer-game, all participants were asked to fill out a game-play questionnaire at initial and post training data collection. The questionnaire, based on the Game Engagement Questionnaire,18 included 19 questions using a 7-point Likert scale. Elements included questions related to immersion, flow, fun, and exercise usefulness. We used a skill-based protocol in which participants trained until they attained certain designated skill levels for testing. The game was initially played at the easiest level (familiarization period) and progressed to medium and then hard levels, each with increased complexity of choreographed movements. Dance Central uses a star scoring system; from 0 stars to a perfect score of 6 stars. Players were required to achieve a minimum of 3 stars to advance to the next level of difficulty. Participants trained twice per week for 30–40 min per training session. Total time spent training by each participant was approximately 4.5 h (261.79 ± 34.34 min, range of training time from 262 to 342 min over 7 to 9 sessions). Prior to testing, participants were introduced to Dance Central with instruction in the operations of the gaming console and game menu. This included how to log in, select game-songs, unlock the game song’s next level of difficulty, and interpret their scores. Players further familiarized themselves with Dance Central by playing several game-songs, with instructions to perform the choreography as accurately as possible to maximize their game score. Metabolic and kinematic data collection occurred twice: (1) following the first session of training and (2) upon qualification for post-testing. At initial testing (Initial-E), metabolic data was collected on the easy level for three designated game-songs (‘Poker Face’, 120 beats min−1 ; ‘I Know You Want Me’, 128 beats min−1 ; and ‘Move Ya Body’ 124 beats min−1 ). These songs were designated at a difficulty level of 1, 1, and 4, respectively, according to the games ranking system. Kinematic data were also collected on easy for the game-song ‘Poker Face’. To qualify for post-training testing, each individual had to obtain 4 stars at the hard level on the designated game-songs. All participants met the skill-based criterion for final testing at all levels. Post-training testing (Post-E and Post-H) consisted of playing the same game-songs on the easy and hard levels. Also during post-testing, to demonstrate generalizability of their learning, players were tested while playing three new game-songs on the hard level (New-H) (‘Funkytown’, 106 beats min−1 ; ‘Galang’05 , 101 beats min−1 ; and ‘Crank that Soulja Boy’, 140 beats min−1 ). The new game-songs were designated at a difficulty level of 2, 2, and 3, respectively, which was sufficient to demonstrate a transfer of
Please cite this article in press as: Bronner S, et al. Physiological and psychophysiological responses to an exer-game training protocol. J Sci Med Sport (2015), http://dx.doi.org/10.1016/j.jsams.2015.03.003
G Model JSAMS-1157; No. of Pages 5
ARTICLE IN PRESS S. Bronner et al. / Journal of Science and Medicine in Sport xxx (2015) xxx–xxx
knowledge when players were playing these songs for the first times. There were no breaks between songs other than the time it took to move through the menu to the next song in order to simulate realistic game-play. Game scores were recorded during data collection. Players answered the game-play questionnaire following the initial and post-testing sessions. Metabolic data recorded from K4b2 was transferred to Microsoft Excel for analysis and expressed in terms of METs to allow comparison with pervious exer-game research. Mean ± SD HR beats min−1 , EE kcal min−1 , and METs per game-song were calculated for each participant. The mean of the three songs was calculated for each of the tests (e.g. Initial-E, Post-E, Post-H, and New-H). Kinematic data was processed and filtered with an FIR filter using the Vicon Nexus software. Sagittal plane peak angular displacements of the hip, knee, shoulder, and elbow were determined in a customized program in Labview 7.0 (National Instruments Corp., Austin, TX). To determine game proficiency for each player, game scores for each song during Initial-E, Post-E, Post-H, and New-H tests were entered into Microsoft Excel. Game-play questionnaire data was entered into Microsoft Excel and mean ± SD were calculated for the each player. For statistical analysis, separate repeated measures ANOVA were used to compare time [Initial-E, Post-E, Post-H, New-H] for physiological and psychophysiological variables (HR, EE, METs, game-scores, game-play questionnaires) in SPSS (21.0, IBM Corp, Armonk, NY). All tests were conducted with an ˛ of p < 0.05. 3. Results Repeated measures comparisons of HR found significant differences between tests (p = 0.016) (Table 1). In pairwise comparisons, New-H HR exceeded Post-H (p < 0.001), both exceeded Post-E (p < 0.001), New-H also exceeded Initial-E (p = 0.014). EE also differed between tests (p = 0.011). In pairwise comparisons, New-H and Post-H EE exceeded Initial-E (p ≤ 0.013). Similarly, METs differed across tests (p < 0.001). Pairwise comparisons found New-H and Post-H exceeded Post-E and Initial-E (p ≤ 0.003). Sagittal plane mean peak angular displacement of hip, knee, shoulder and elbow were examined for the song Poker Face. Hip angular displacements were 30.08 ± 6.69◦ and 28.82 ± 4.28◦ and knee angular displacements were 44.12 ± 17.57◦ and 40.26 ± 16.52◦ during Initial-E and Post-H testing. Shoulder angular displacements were 52.94 ± 35.20◦ and 37.43 ± 11.29◦ and elbow angular displacements were 120.65 ± 10.87◦ and 119.67 ± 5.49◦ during Initial-E and Post-H testing. Repeated measures comparisons did not yield significant differences for any joint (p > 0.05). Game scores differed across testing (Fig. 1) (p < 0.001). Pairwise comparisons found Initial-E and Post-E scores were lower than Post-H (p < 0.001) and New-H (p ≤ 0.003) scores. Post-H scores exceeded New-H scores (p = 0.012). There were no differences between Initial-E and Post-E scores.
3
Fig. 1. Mean (SD) game scores for Initial testing: Initial-E; and Post testing: Post-E, Post-H, and New-H gameplay. Repeated measures comparisons [F(1,13) = 53.617, p < 0.001]. Pair-wise comparisons: New-H and Post-H v. Initial-E and Post-E scores, p ≤ 0.003; New-H v. Post-H, p = 0.012. Abbreviations: Initial-E = Initial testing on easy level; New-H = Post training testing on new songs on hard level; Post-E = Post training testing on easy level; Post-H = Post training testing on hard level.
Engagement scores from the game-play questionnaire decreased from initial-testing to post-training, from 89.14 ± 7.82 to 88.43 ± 19.07, but were not significant (p > 0.05). We also analyzed specific scores related to immersion and flow. Immersion decreased with training (p < 0.01); there was no change in flow. 4. Discussion The results of this study indicate that with training, player physical exertion and game proficiency increased, confirming two of the four hypotheses tested in this study. However, limb movement did not change and players were not more engaged post-training, rejecting these hypotheses. Our METs results confirm that Dance Central is a moderate physical activity regardless of training experience [Moderate activity = 3–6 METs,19 Dance Central = 4.27–5.42 METs]. These moderate findings confirm MET levels reported previously for advanced experienced adults.12 The length of a song and intricacy of choreographed steps may alter the physical challenge compared to other dance games or sport games available on Xbox Kinect.12,13 Although increased song difficulty did not produce vigorous levels of physical activity, we measured increased METs during advanced game-play (e.g. Post-H and New-H tests). Increased skill in performing a prescribed task is generally associated with increased efficiency; but there were no MET differences between Initial-E and retention (Post-E) testing. Playing the same game-songs at more advanced levels involved more complex choreography at the same tempo, which resulted in greater physical activity. In order to progress in their training, players were required to increase accuracy to unlock the new levels of the game. Dance Central provided real-time feedback so that players can adjust their movements for correction. Increased proficiency was reflected in
Table 1 Heart rate, energy expenditure, and METs. Variable
Initial-E
Post-E
Post-H
New-H
F, p
HR a beats min−1 EEb kcal min−1 METSc
103.17 ± 17.86
98.77 ± 13.14
109.57 ± 15.39
115.15 ± 17.11
F(13,1) = 12.696, p = 0.003
4.89 ± 1.35
5.54 ± 3.13
5.92 ± 1.25
6.34 ± 1.35
F(13,1) = 8.877, p = 0.011
4.25 ± 0.72
4.10 ± 1.00
5.05 ± 0.75
5.42 ± 0.88
F(13,1) = 24.296, p < 0.001
Abbreviations: Initial-E = Initial testing on easy level; New-H = Post training testing on new songs on hard level; Post-E = Post training testing on easy level; Post-H = Post training testing on hard level. a HR Pairwise comparisons: Post-H and New-H v. Post-E, p ≤ 0.001; New-H v. Post-H, p < 0.001; New-H v. Initial-E, p = 0.014. b EE Pairwise comparisons: Post-H and New-H v. Initial-E, p ≤ 0.013. c METS Pairwise comparisons: Post-H and New-H v. Initial-E, Post-E, p ≤ 0.003; New-H v. Post-H, p ≤ 0.001.
Please cite this article in press as: Bronner S, et al. Physiological and psychophysiological responses to an exer-game training protocol. J Sci Med Sport (2015), http://dx.doi.org/10.1016/j.jsams.2015.03.003
G Model JSAMS-1157; No. of Pages 5 4
ARTICLE IN PRESS S. Bronner et al. / Journal of Science and Medicine in Sport xxx (2015) xxx–xxx
increased game scores in advanced game-play of Post-H and NewH; however there were no differences between Initial-E and Post-E. Generalizability of skills, defined as a component of successful motor learning, was demonstrated in increased game scores during performance of the New-H songs compared to Post-H. Players, when confronted with songs they had no previous experience playing, were able to perform at an advanced level. Players learned a skill requiring quick recognition and mimicry of choreography that could be generalized when they were presented with this new choreography, suggesting a transfer of learning. This suggests that Kinect exer-games have promising applications for rehabilitation. Although players demonstrated improved proficiency and increased physical exertion, they did not increase their limb movement as hypothesized. We anticipated that players would demonstrate a freezing-freeing phenomenon in moving from novice to advanced play.20,21 However, movement may have been constrained by song tempo and spatial constraints of the Kinect motion sensor. These results indicate that although participants did not move more in later testing, they did not significantly reduce their motions either. Xbox Kinect, therefore, can be argued to have developed a sensor that does not allow players to minimize their movements. Players were unable ‘game’ the system to improve their performance as in other dance game systems like Nintendo Wii. In an investigation into player motivation and experience during gameplay, Pasch et al.22 interviewed motion-based video gamers and reported they had two motivations: to relax or to achieve. They observed that when experienced players played to achieve on Nintendo Wii games, they reduced their movements to the minimum necessary to win. When they played to relax, movements resembled the realistic movement of the sport (e.g. tennis, boxing, etc.). If the controller or game platform guides what movements a player selects to score well, the lack of change in limb movement in this study suggests that the Xbox Kinect has the potential to optimize greater physical activity using this paradigm. Similar evidence of the controller guiding player movement was reported in a comparison of three different platforms with similar dance songs.12 Physical exertion, while playing the same game-song and dance exer-game, dropped in advanced players when playing with the Wii controller (4.2 METs) compared to the Kinect Xbox (6.8 METs).12 However, another comparison of dance games on three different platforms found no difference in energy expenditure in players who were briefly ‘familiarized’ with the games.23 This example points out again that brief familiarity may not accurately reflect an exer-game’s capability to promote physical activity. Expertise should be carefully defined to distinguish experienced players from novices. In our case, expertise was defined by achieving the accuracy necessary to move to more advanced levels. Researchers have defined expertise as players who have previously played the games for an accumulated time exceeding 30-min;24 however, this does not delineate any type of skill. Other studies defined experienced players as individuals who regularly played on the game’s advanced level.12,25 In these studies, energy expenditure levels when playing whole body exer-games, were greater in experienced players compared to inexperienced players, defined as playing on beginner modes. Full body and realistic movements appear to contribute to enhancement of the exer-game experience. Engagement scores were increased when players played Guitar Hero on a guitar compared to a pad.15 Similarly, engagement questionnaire scores increased when participants played dance exer-games on the Kinect compared to playing with the Wii Remote controller.12,23 While most would agree that flow and immersion are critical to positive gaming experiences,26–28 there is little research on how training affects these constructs. In this study, training did not result in change in engagement scores. One study investigated the
relationship between immersion and gameplay experience in the game Beowulf, across a continuum of gamer experience.29 They reported that immersion is more positive when the gamer has less experience. When we deconstructed the engagement scores in this study, players reported decreased immersion with training. A recent study found that players who exercised on Dance Central 2 for four weeks rated game enjoyment highly.30 However, the research was focused on enhancing the game with story content to improve long term adherence, so it did not focus on skill development or elements of flow traditionally used in video game research. One limitation to this study was that all participants were required to train onsite and we were unable to control for social interaction factors. Due to the time commitment required for participants to train and a lack of multiple Xbox 360 and Kinect systems, we did not control whether they trained alone or with a partner. Social interaction can be a powerful exercise motivator and may ultimately encourage further training. Playing games in a group has previously been shown to produce higher energy expenditure in Kinect Reflex Ridge and Wii Sport Boxing exer-games.31 All of the participants had previous experience playing Dance Dance Revolution and four had previous formal dance training. This may be a factor in the limited amount of training time required to meet the skill-based criteria. We did not ascertain the fitness levels of the participants prior to enrollment. However, the focus of this study was a repeated measures design, with each participant serving as their own control. It is unlikely that the amount of training resulted in a conditioning effect on physiological variables. Rather, we attribute these physiologic changes to skill development.
5. Conclusion The current investigation is, to our knowledge, the first to examine a Kinect game in adults using a skill-based paradigm. Players trained until they were able to attain the necessary skills to progress in the testing level. Findings indicate that cutting corners did not occur due to the full body tracking and thus argues that Dance Central and other Kinect games may serve as a good model for interactive rehabilitation programs. Increased skill at the game did not produce increased engagement and resulted in reports of decreased immersion. This is concerning because one of the tenants of skilled motor learning is focused attention. As exer-game and VR systems move from their initial novelty, information about how players react to continued involvement with the game over time can guide game developers to maintain game progression freshness that preserves each participant’s attentional focus. The initial attraction to exer-games, compared to exercise and rehabilitation repetition, is that they are fun,32 but ongoing engagement using innovative and challenging progression may be necessary to keep players returning to achieve long-term goals.
6. Practical implications • Although Dance Central involves whole-body movement, physical exertion remained at moderate levels after training to advanced levels. • There was no change in engagement. • Immersion decreased with training.
Acknowledgements The authors thank the players for their time spent on this project and the students in our lab who assisted us in this project. This research was supported in part by grants from the Robert Wood
Please cite this article in press as: Bronner S, et al. Physiological and psychophysiological responses to an exer-game training protocol. J Sci Med Sport (2015), http://dx.doi.org/10.1016/j.jsams.2015.03.003
G Model JSAMS-1157; No. of Pages 5
ARTICLE IN PRESS S. Bronner et al. / Journal of Science and Medicine in Sport xxx (2015) xxx–xxx
Johnson Foundation Health Games Research (Grant No. #66729) and Long Island University Brooklyn Health and Wellness Institute. References 1. Betker AL, Desai A, Nett C et al. Game-based exercises for dynamic short-sitting balance rehabilitation of people with chronic spinal cord and traumatic brain injuries. Phys Ther 2007; 87:1389–1398. 2. Deutsch JE. Using virtual reality to improve walking post-stroke: translation to individuals with diabetes. J Diabetes Sci Technol 2011; 5:309–314. 3. Gaffurini P, Bissolotti L, Calza S et al. Energy metabolism during activitypromoting video games practice in subjects with spinal cord injury: evidences for health promotion. Eur J Phys Rehabil Med 2013; 49:23–29. 4. Mirelman A, Maidan I, Deutsch JE. Virtual reality and motor imagery: promising tools for assessment and therapy in Parkinson’s disease. Mov Disord 2013; 28:1597–1608 (official journal of the Movement Disorder Society). 5. Ungerleider LG, Doyon J, Karni A. Imaging brain plasticity during motor skill learning. Neurobiol Learn Mem 2002; 78:553–564. 6. Kramer AF, Erickson KI. Capitalizing on cortical plasticity: influence of physical activity on cognition and brain function. Trends Cogn Sci 2007; 11:342–348. 7. Lange B, Flynn SM, Rizzo AA. Game-based telerehabilitation. Eur J Phys Rehabil Med 2009; 45:143–151. 8. Shumway-Cook A, Woollacott MH. Motor Control Translating Research into Practice, 3rd ed. Philadelphia, PA, Lippicott Williams & Wilkins, 2007. 9. Bailey BW, McInnis K. Energy cost of exergaming: a comparison of the energy cost of 6 forms of exergaming. Arch Pediatr Adolesc Med 2011; 165:597–602. 10. Graves LE, Ridgers ND, Williams K et al. The physiological cost and enjoyment of Wii Fit in adolescents, young adults, and older adults. J Phys Act Health 2010; 7:393–401. 11. Lyons EJ, Tate DF, Ward DS et al. Energy expenditure and enjoyment during video game play: differences by game type. Med Sci Sports Exerc 2011; 43:1987–1993. 12. Bronner S, Pinsker R, Noah JA. Energy cost and game flow of 5 exer-games in trained players. Am J Health Behav 2013; 37:369–380. 13. Mellecker RR, McManus AM. Active video games and physical activity recommendations: a comparison of the Gamercize Stepper, XBOX Kinect and XaviX J-Mat. J Sci Med Sport 2014; 17:288–292. 14. Whitehead A, Johnston H, Nixon N et al. Exergame effectiveness: what the numbers can tell us, In: Proceedings of the fifth ACM SIGGRAPH symposium on video games. New York, NY, Association for Computing Machinery (ACM) Press, 2010. p. 55–62. 15. Bianchi-Berthouze N, Kim WW, Patel D. Does body movement engage you more in digital game play? and why? In: Paiva ACR, Prada R, Picard RW, editors. Affective computing and intelligent interaction, second international conference, ACII, vol. 4738. Portugal, Lisbon, 2007, p. 102–113.
5
16. Piva SR, Gil AB, Almeida GJ et al. A balance exercise program appears to improve function for patients with total knee arthroplasty: a randomized clinical trial. Phys Ther 2010; 90:880–894. 17. Husswirth C, Bigard AX, Le Chevalier JM. The Cosmed K4 telemetry system as an accurate device for oxygen uptake measurements during exercise. Int J Sports Med 1997; 18:449–453. 18. Brockmyer JH, Fox CM, Curtiss KA et al. The development of the game engagement questionnaire: a measure of engagement in video game-playing. J Exp Social Psychol 2009; 45:624–634. 19. World Health Organization. Global recommendations on physical activity for health, Geneva, Switzerland, World Health Organization, 2010. 20. Bernstein N. The Coordination and Regulation of Movement, NY, Pergamon Press, 1967. 21. Vereijken B, van Emmerik REA, Whiting HT et al. Free(z)ing degrees of freedom in skill acquisition. J Mot Behav 1992; 24:133–142. 22. Pasch M, Bianchi-Berthouze N, van Dijk B et al. Movement-based sports video games: investigating motivation and gaming experience. Entertain Comput 2009; 1:49–61. 23. Thin AG, Brown C, Meenan P. User experiences while playing dance-based exergames and the Influence of different body motion sensing technologies. Int J Comput Games Technol 2013; 2013:1–7. 24. O’Donovan C, Hussey J. Active video games as a form of exercise and the effect of gaming experience: a preliminary study in healthy young adults. Physiother 2012; 98:205–210. 25. Sell K, Lillie T, Taylor J. Energy expenditure during physically interactive video game playing in male college students with different playing experience. J Am Coll Health 2008; 56:505–511. 26. Lin HH, Wang YS, Chou CH. Hedonic and utilitarian motivations for physical game systems use behavior. Int J Hum-Comput Interact 2012; 28: 445–455. 27. Thin A, Hansen L, McEachen D. Flow experience and mood states while playing body movement-controlled video games. Games Cult 2011; 6:414–428. 28. Jennett C, Cox AL, Cairns P et al. Measuring and defining the experience of immersion in games. Int J Hum-Comput Stud 2008; 66:641–661. 29. Örtqvist D, Liljedahl M. Immersion and gameplay experience: a contingency framework. Int J Comput Games Technol 2010:1–11. 30. Charbonneau E, Buchanan S, Bott J et al. Dance enhanced: investigating how earning content through exertion impacts dance game enjoyment, Paper presented at: ninth annual conference on the foundations of digital games, 2014. 31. O’Donovan C, Hirsch E, Holohan E et al. Energy expended playing Xbox Kinect and Wii games: a preliminary study comparing single and multiplayer modes. Physiotherapy 2012; 98:224–229. 32. Leninger LJ, Coles MG, Gilbert JN. Comparing enjoyment and perceived exertion between equivalent bouts of physically interactive video gaming and treadmill walking. Health Finess J 2010; 3:12–18.
Please cite this article in press as: Bronner S, et al. Physiological and psychophysiological responses to an exer-game training protocol. J Sci Med Sport (2015), http://dx.doi.org/10.1016/j.jsams.2015.03.003
ARTICLE IN PRESS Journal of Science and Medicine in Sport xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Journal of Science and Medicine in Sport journal homepage: www.elsevier.com/locate/jsams
Original research
Physiological and psychophysiological responses to an exer-game training protocol Shaw Bronner a,∗ , Russell Pinsker b , Rutika Naik c , J. Adam Noah d a
ADAM Center, Dept. of Physical Therapy, Movement and Rehabilitation Sciences, Northeastern University, USA ADAM Center, Health Sciences, Long Island University, USA ADAM Center, Division of Sports Sciences, Long Island University, USA d Brain Function Laboratory, Department of Psychiatry, Yale University School of Medicine, USA b c
a r t i c l e
i n f o
Article history: Received 28 October 2014 Received in revised form 18 January 2015 Accepted 5 March 2015 Available online xxx Keywords: Energy expenditure Exercise METs Engagement Dance Video games
a b s t r a c t Objectives: Exer-games and virtual reality offer alternative opportunities to provide neuro-rehabilitation and exercise that are fun. Our goal was to determine how effective they are in achieving motor learning goals and fitness benefits as players gain experience. Design: We employed a repeated measures design to determine changes in physical exertion and engagement with training. Methods: Fourteen healthy adults trained on the XBOX Kinect video game Dance Central using a skillbased protocol to examine changes in energy expenditure (EE), heart rate (HR), METs, limb movement, game proficiency, and player engagement in initial, post-training, and transfer-testing of a full-body dance exer-game. Data were analyzed using repeated measures analysis of variance, p < 0.05. Results: Both EE, HR, and METs increased from initial (EE 4.89 ± 1.35, HR 103 ± 18, METs 4.25 ± 0.72) to post-training (EE 5.92 ± 1.25, HR 110 ± 15, METs 5.05 ± 0.75) and were greatest during transfertesting (EE 6.34 ± 1.35, HR 115 ± 17, METs 5.42 ± 0.88, p ≤ 0.001). Proficiency, measured by game scores, also increased from initial to post-training and transfer-testing (p ≤ 0.002). Limb movement and player engagement remained unchanged. Conclusions: It is important to understand whether player physiological and psychophysiological responses change with continued game-play. Although Dance Central involves whole-body movement, physical exertion remained at moderate levels after training. As exer-game and virtual reality systems move from their initial novelty, research about how players react to continued involvement with a game can guide game developers to maintain a freshness through game progression that preserves the participant’s attentional focus, minimizes attrition and maintains a prescribed level of energy exertion. © 2015 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
1. Introduction Exercise video games (exer-games) and virtual reality (VR) based systems are increasingly being used for physical activity and rehabilitation goals in the treatment of different neuropathologies such as traumatic brain injury,1 stroke,2 spinal cord injury,3 and Parkinson’s disease.4 High intensity task-specific exercise promotes neural reorganization, neuroplasticity, and improvements in function,5 as well as improvements in health-related physical fitness and cognitive function.6 In fitness training or neurologic rehabilitation, exer-games and VR training offer the potential to
∗ Corresponding author. E-mail address: [email protected] (S. Bronner).
deliver these gains in a cost effective manner, in socialized group activities, tele-rehabilitation, and on-line interactive exer-gaming.7 To achieve exercise and rehabilitation goals, improved performance requires intensive skilled practice, focused attention, progressive challenges, immediate feedback, and knowledge of results.8 To reach these goals, exercise adherence is critical, requiring that the exer-game or VR is engaging and players are motivated to return. Additionally, exer-games and VR must record data on the level of play and performance in order to ascertain whether the participant attains specific goals. If the outcome is improved physical fitness, do the players perform at a high enough intensity with ongoing progressive challenges as they improve? If the goal is rehabilitation-related, is there accurate tracking and feedback of performance related to functional outcomes? Research to date on exer-games or VR systems for physical fitness and neuro-rehabilitation has primarily focused on proof of
http://dx.doi.org/10.1016/j.jsams.2015.03.003 1440-2440/© 2015 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Bronner S, et al. Physiological and psychophysiological responses to an exer-game training protocol. J Sci Med Sport (2015), http://dx.doi.org/10.1016/j.jsams.2015.03.003
G Model JSAMS-1157; No. of Pages 5
ARTICLE IN PRESS S. Bronner et al. / Journal of Science and Medicine in Sport xxx (2015) xxx–xxx
2
concept in small samples. In the case of exer-games for the goal of physical fitness, the literature does not reflect energy expenditure (EE) of well-trained advanced players. Rather, participants are briefly ‘familiarized’ prior to determining EE during gameplay. The majority of research reports that participants exert 3–5 metabolic equivalents (METs) (light to moderate activity) when they play exer-games such as Wii Fit, Wii Sports, Dance Dance Revolution, Kinect Dance Central, and Eye Toy activities.9–11 Others have demonstrated moderate to vigorous exercise intensity (4.2–7.1 METs).9,12,13 Therefore, exer-games have potential to stimulate adequate conditioning of the body, but changes with skill development are unstudied. Games that use sensors that measure full body motion or require activities that use the large muscles of the lower extremities result in greater physical activity.14 The incorporation of whole body movement may also result in greater player engagement while playing exer-games through immersion.15 Player engagement is important because a positive experience makes it more likely that a player will return to play again. Engagement is also important for any training program (exercise or rehabilitation), as attrition rates tend to be high.16 We focused on two aspects of game-play, including both physiological and psychophysiological measures relevant to any type of exercise. To study the effect of exer-game training on these variables, we selected the interactive dance game, Dance Central for the Xbox 360, as it uses the Kinect sensor to process full body motion capture and determine performance scores. Dance Central offers a number of choreographed dance styles and music, potentially appealing to a range of players. We hypothesized that compared to an initial baseline trial following a “familiarization” period, player training would result in the following: (1) increased physical exertion; (2) increased upper and lower extremity movement; (3) improved game proficiency; (4) demonstration of motor learning by transfer of proficiency; and (5) greater player engagement.
2. Methods Fourteen healthy participants (seven women, seven men, ages 26.6 ± 9.5 years, height 172 ± 11 cm, mass 69 ± 19 kg), recruited by flier at a local university, volunteered for this study. A prior power analysis determined for a repeated measures design, with one group, three measurements, a medium effect size (f = 0.50) and alpha = 0.05, a sample size of 12 was required. All individuals were free of recent musculoskeletal injury, visual or auditory impairment, color blindness, neurological disease, or heart condition, based on their answers to a questionnaire for participation inclusion or exclusion. Each participant had minimal or no experience playing Dance Central on the Kinect prior to participation. During training and recording sessions, participants wore workout clothing. Each participant gave informed written consent and the Institutional Review Board for human ethics granted ethical approval for this study. Participants trained on Dance Central (Harmonix Music Systems, Inc., Cambridge, MA) for the Xbox 360 gaming console and Kinect motion sensor accessory (Microsoft Corp., Redmond, Washington, U.S.). The gaming console was attached to a 50-in video monitor. Participants played in a 1.8 m × 3.6 m square, approximately 2 m from the Kinect motion sensor and video monitor. During Dance Central game-play, players stand in front of the monitor and follow an avatar dancing on the screen. The gaming console includes a motion sensor, skeletal estimation sensor, with facial and voice recognition that captures the player’s movement and compares it to those of the z-score underlying the avatar movements.
Game scores are based on movement and timing accuracy of the individual compared to the z-scores. To determine metabolic expenditure via indirect calorimetry, participants were outfitted with a portable gas analysis system (K4b2 Cosmed Inc., Rome, Italy) comprised of a small metabolic analyzer, battery pack, and face-mask and a chest strap heart rate (HR) monitor (Nike, Beaverton, OR). The facemask was attached to a turbine flow-meter allowing for real-time collection of VO2 values. The K4b2 unit (∼925 g) was strapped to the participant using the manufacturer’s harness. Prior to each testing session, the K4b2 was calibrated according to the manufacturer’s instructions. The K4b2 has been well validated for measurement of metabolic expenditure.17 During testing, VO2 , VCO2 , and HR were monitored continuously and recorded using the K4b2 breath-by-breath analysis. The beginning and end of each game-song were marked using the K4b2 marker feature. To determine limb movement, we collected kinematic data using a 10-camera motion capture system (Vicon, Oxford, UK), sampled at 120 Hz. Participants were outfitted with a full-body marker set using the Vicon Plug-In Gait model. To determine changes in player proficiency, game scores were recorded during data collection. To determine whether player engagement altered with greater experience with the exer-game, all participants were asked to fill out a game-play questionnaire at initial and post training data collection. The questionnaire, based on the Game Engagement Questionnaire,18 included 19 questions using a 7-point Likert scale. Elements included questions related to immersion, flow, fun, and exercise usefulness. We used a skill-based protocol in which participants trained until they attained certain designated skill levels for testing. The game was initially played at the easiest level (familiarization period) and progressed to medium and then hard levels, each with increased complexity of choreographed movements. Dance Central uses a star scoring system; from 0 stars to a perfect score of 6 stars. Players were required to achieve a minimum of 3 stars to advance to the next level of difficulty. Participants trained twice per week for 30–40 min per training session. Total time spent training by each participant was approximately 4.5 h (261.79 ± 34.34 min, range of training time from 262 to 342 min over 7 to 9 sessions). Prior to testing, participants were introduced to Dance Central with instruction in the operations of the gaming console and game menu. This included how to log in, select game-songs, unlock the game song’s next level of difficulty, and interpret their scores. Players further familiarized themselves with Dance Central by playing several game-songs, with instructions to perform the choreography as accurately as possible to maximize their game score. Metabolic and kinematic data collection occurred twice: (1) following the first session of training and (2) upon qualification for post-testing. At initial testing (Initial-E), metabolic data was collected on the easy level for three designated game-songs (‘Poker Face’, 120 beats min−1 ; ‘I Know You Want Me’, 128 beats min−1 ; and ‘Move Ya Body’ 124 beats min−1 ). These songs were designated at a difficulty level of 1, 1, and 4, respectively, according to the games ranking system. Kinematic data were also collected on easy for the game-song ‘Poker Face’. To qualify for post-training testing, each individual had to obtain 4 stars at the hard level on the designated game-songs. All participants met the skill-based criterion for final testing at all levels. Post-training testing (Post-E and Post-H) consisted of playing the same game-songs on the easy and hard levels. Also during post-testing, to demonstrate generalizability of their learning, players were tested while playing three new game-songs on the hard level (New-H) (‘Funkytown’, 106 beats min−1 ; ‘Galang’05 , 101 beats min−1 ; and ‘Crank that Soulja Boy’, 140 beats min−1 ). The new game-songs were designated at a difficulty level of 2, 2, and 3, respectively, which was sufficient to demonstrate a transfer of
Please cite this article in press as: Bronner S, et al. Physiological and psychophysiological responses to an exer-game training protocol. J Sci Med Sport (2015), http://dx.doi.org/10.1016/j.jsams.2015.03.003
G Model JSAMS-1157; No. of Pages 5
ARTICLE IN PRESS S. Bronner et al. / Journal of Science and Medicine in Sport xxx (2015) xxx–xxx
knowledge when players were playing these songs for the first times. There were no breaks between songs other than the time it took to move through the menu to the next song in order to simulate realistic game-play. Game scores were recorded during data collection. Players answered the game-play questionnaire following the initial and post-testing sessions. Metabolic data recorded from K4b2 was transferred to Microsoft Excel for analysis and expressed in terms of METs to allow comparison with pervious exer-game research. Mean ± SD HR beats min−1 , EE kcal min−1 , and METs per game-song were calculated for each participant. The mean of the three songs was calculated for each of the tests (e.g. Initial-E, Post-E, Post-H, and New-H). Kinematic data was processed and filtered with an FIR filter using the Vicon Nexus software. Sagittal plane peak angular displacements of the hip, knee, shoulder, and elbow were determined in a customized program in Labview 7.0 (National Instruments Corp., Austin, TX). To determine game proficiency for each player, game scores for each song during Initial-E, Post-E, Post-H, and New-H tests were entered into Microsoft Excel. Game-play questionnaire data was entered into Microsoft Excel and mean ± SD were calculated for the each player. For statistical analysis, separate repeated measures ANOVA were used to compare time [Initial-E, Post-E, Post-H, New-H] for physiological and psychophysiological variables (HR, EE, METs, game-scores, game-play questionnaires) in SPSS (21.0, IBM Corp, Armonk, NY). All tests were conducted with an ˛ of p < 0.05. 3. Results Repeated measures comparisons of HR found significant differences between tests (p = 0.016) (Table 1). In pairwise comparisons, New-H HR exceeded Post-H (p < 0.001), both exceeded Post-E (p < 0.001), New-H also exceeded Initial-E (p = 0.014). EE also differed between tests (p = 0.011). In pairwise comparisons, New-H and Post-H EE exceeded Initial-E (p ≤ 0.013). Similarly, METs differed across tests (p < 0.001). Pairwise comparisons found New-H and Post-H exceeded Post-E and Initial-E (p ≤ 0.003). Sagittal plane mean peak angular displacement of hip, knee, shoulder and elbow were examined for the song Poker Face. Hip angular displacements were 30.08 ± 6.69◦ and 28.82 ± 4.28◦ and knee angular displacements were 44.12 ± 17.57◦ and 40.26 ± 16.52◦ during Initial-E and Post-H testing. Shoulder angular displacements were 52.94 ± 35.20◦ and 37.43 ± 11.29◦ and elbow angular displacements were 120.65 ± 10.87◦ and 119.67 ± 5.49◦ during Initial-E and Post-H testing. Repeated measures comparisons did not yield significant differences for any joint (p > 0.05). Game scores differed across testing (Fig. 1) (p < 0.001). Pairwise comparisons found Initial-E and Post-E scores were lower than Post-H (p < 0.001) and New-H (p ≤ 0.003) scores. Post-H scores exceeded New-H scores (p = 0.012). There were no differences between Initial-E and Post-E scores.
3
Fig. 1. Mean (SD) game scores for Initial testing: Initial-E; and Post testing: Post-E, Post-H, and New-H gameplay. Repeated measures comparisons [F(1,13) = 53.617, p < 0.001]. Pair-wise comparisons: New-H and Post-H v. Initial-E and Post-E scores, p ≤ 0.003; New-H v. Post-H, p = 0.012. Abbreviations: Initial-E = Initial testing on easy level; New-H = Post training testing on new songs on hard level; Post-E = Post training testing on easy level; Post-H = Post training testing on hard level.
Engagement scores from the game-play questionnaire decreased from initial-testing to post-training, from 89.14 ± 7.82 to 88.43 ± 19.07, but were not significant (p > 0.05). We also analyzed specific scores related to immersion and flow. Immersion decreased with training (p < 0.01); there was no change in flow. 4. Discussion The results of this study indicate that with training, player physical exertion and game proficiency increased, confirming two of the four hypotheses tested in this study. However, limb movement did not change and players were not more engaged post-training, rejecting these hypotheses. Our METs results confirm that Dance Central is a moderate physical activity regardless of training experience [Moderate activity = 3–6 METs,19 Dance Central = 4.27–5.42 METs]. These moderate findings confirm MET levels reported previously for advanced experienced adults.12 The length of a song and intricacy of choreographed steps may alter the physical challenge compared to other dance games or sport games available on Xbox Kinect.12,13 Although increased song difficulty did not produce vigorous levels of physical activity, we measured increased METs during advanced game-play (e.g. Post-H and New-H tests). Increased skill in performing a prescribed task is generally associated with increased efficiency; but there were no MET differences between Initial-E and retention (Post-E) testing. Playing the same game-songs at more advanced levels involved more complex choreography at the same tempo, which resulted in greater physical activity. In order to progress in their training, players were required to increase accuracy to unlock the new levels of the game. Dance Central provided real-time feedback so that players can adjust their movements for correction. Increased proficiency was reflected in
Table 1 Heart rate, energy expenditure, and METs. Variable
Initial-E
Post-E
Post-H
New-H
F, p
HR a beats min−1 EEb kcal min−1 METSc
103.17 ± 17.86
98.77 ± 13.14
109.57 ± 15.39
115.15 ± 17.11
F(13,1) = 12.696, p = 0.003
4.89 ± 1.35
5.54 ± 3.13
5.92 ± 1.25
6.34 ± 1.35
F(13,1) = 8.877, p = 0.011
4.25 ± 0.72
4.10 ± 1.00
5.05 ± 0.75
5.42 ± 0.88
F(13,1) = 24.296, p < 0.001
Abbreviations: Initial-E = Initial testing on easy level; New-H = Post training testing on new songs on hard level; Post-E = Post training testing on easy level; Post-H = Post training testing on hard level. a HR Pairwise comparisons: Post-H and New-H v. Post-E, p ≤ 0.001; New-H v. Post-H, p < 0.001; New-H v. Initial-E, p = 0.014. b EE Pairwise comparisons: Post-H and New-H v. Initial-E, p ≤ 0.013. c METS Pairwise comparisons: Post-H and New-H v. Initial-E, Post-E, p ≤ 0.003; New-H v. Post-H, p ≤ 0.001.
Please cite this article in press as: Bronner S, et al. Physiological and psychophysiological responses to an exer-game training protocol. J Sci Med Sport (2015), http://dx.doi.org/10.1016/j.jsams.2015.03.003
G Model JSAMS-1157; No. of Pages 5 4
ARTICLE IN PRESS S. Bronner et al. / Journal of Science and Medicine in Sport xxx (2015) xxx–xxx
increased game scores in advanced game-play of Post-H and NewH; however there were no differences between Initial-E and Post-E. Generalizability of skills, defined as a component of successful motor learning, was demonstrated in increased game scores during performance of the New-H songs compared to Post-H. Players, when confronted with songs they had no previous experience playing, were able to perform at an advanced level. Players learned a skill requiring quick recognition and mimicry of choreography that could be generalized when they were presented with this new choreography, suggesting a transfer of learning. This suggests that Kinect exer-games have promising applications for rehabilitation. Although players demonstrated improved proficiency and increased physical exertion, they did not increase their limb movement as hypothesized. We anticipated that players would demonstrate a freezing-freeing phenomenon in moving from novice to advanced play.20,21 However, movement may have been constrained by song tempo and spatial constraints of the Kinect motion sensor. These results indicate that although participants did not move more in later testing, they did not significantly reduce their motions either. Xbox Kinect, therefore, can be argued to have developed a sensor that does not allow players to minimize their movements. Players were unable ‘game’ the system to improve their performance as in other dance game systems like Nintendo Wii. In an investigation into player motivation and experience during gameplay, Pasch et al.22 interviewed motion-based video gamers and reported they had two motivations: to relax or to achieve. They observed that when experienced players played to achieve on Nintendo Wii games, they reduced their movements to the minimum necessary to win. When they played to relax, movements resembled the realistic movement of the sport (e.g. tennis, boxing, etc.). If the controller or game platform guides what movements a player selects to score well, the lack of change in limb movement in this study suggests that the Xbox Kinect has the potential to optimize greater physical activity using this paradigm. Similar evidence of the controller guiding player movement was reported in a comparison of three different platforms with similar dance songs.12 Physical exertion, while playing the same game-song and dance exer-game, dropped in advanced players when playing with the Wii controller (4.2 METs) compared to the Kinect Xbox (6.8 METs).12 However, another comparison of dance games on three different platforms found no difference in energy expenditure in players who were briefly ‘familiarized’ with the games.23 This example points out again that brief familiarity may not accurately reflect an exer-game’s capability to promote physical activity. Expertise should be carefully defined to distinguish experienced players from novices. In our case, expertise was defined by achieving the accuracy necessary to move to more advanced levels. Researchers have defined expertise as players who have previously played the games for an accumulated time exceeding 30-min;24 however, this does not delineate any type of skill. Other studies defined experienced players as individuals who regularly played on the game’s advanced level.12,25 In these studies, energy expenditure levels when playing whole body exer-games, were greater in experienced players compared to inexperienced players, defined as playing on beginner modes. Full body and realistic movements appear to contribute to enhancement of the exer-game experience. Engagement scores were increased when players played Guitar Hero on a guitar compared to a pad.15 Similarly, engagement questionnaire scores increased when participants played dance exer-games on the Kinect compared to playing with the Wii Remote controller.12,23 While most would agree that flow and immersion are critical to positive gaming experiences,26–28 there is little research on how training affects these constructs. In this study, training did not result in change in engagement scores. One study investigated the
relationship between immersion and gameplay experience in the game Beowulf, across a continuum of gamer experience.29 They reported that immersion is more positive when the gamer has less experience. When we deconstructed the engagement scores in this study, players reported decreased immersion with training. A recent study found that players who exercised on Dance Central 2 for four weeks rated game enjoyment highly.30 However, the research was focused on enhancing the game with story content to improve long term adherence, so it did not focus on skill development or elements of flow traditionally used in video game research. One limitation to this study was that all participants were required to train onsite and we were unable to control for social interaction factors. Due to the time commitment required for participants to train and a lack of multiple Xbox 360 and Kinect systems, we did not control whether they trained alone or with a partner. Social interaction can be a powerful exercise motivator and may ultimately encourage further training. Playing games in a group has previously been shown to produce higher energy expenditure in Kinect Reflex Ridge and Wii Sport Boxing exer-games.31 All of the participants had previous experience playing Dance Dance Revolution and four had previous formal dance training. This may be a factor in the limited amount of training time required to meet the skill-based criteria. We did not ascertain the fitness levels of the participants prior to enrollment. However, the focus of this study was a repeated measures design, with each participant serving as their own control. It is unlikely that the amount of training resulted in a conditioning effect on physiological variables. Rather, we attribute these physiologic changes to skill development.
5. Conclusion The current investigation is, to our knowledge, the first to examine a Kinect game in adults using a skill-based paradigm. Players trained until they were able to attain the necessary skills to progress in the testing level. Findings indicate that cutting corners did not occur due to the full body tracking and thus argues that Dance Central and other Kinect games may serve as a good model for interactive rehabilitation programs. Increased skill at the game did not produce increased engagement and resulted in reports of decreased immersion. This is concerning because one of the tenants of skilled motor learning is focused attention. As exer-game and VR systems move from their initial novelty, information about how players react to continued involvement with the game over time can guide game developers to maintain game progression freshness that preserves each participant’s attentional focus. The initial attraction to exer-games, compared to exercise and rehabilitation repetition, is that they are fun,32 but ongoing engagement using innovative and challenging progression may be necessary to keep players returning to achieve long-term goals.
6. Practical implications • Although Dance Central involves whole-body movement, physical exertion remained at moderate levels after training to advanced levels. • There was no change in engagement. • Immersion decreased with training.
Acknowledgements The authors thank the players for their time spent on this project and the students in our lab who assisted us in this project. This research was supported in part by grants from the Robert Wood
Please cite this article in press as: Bronner S, et al. Physiological and psychophysiological responses to an exer-game training protocol. J Sci Med Sport (2015), http://dx.doi.org/10.1016/j.jsams.2015.03.003
G Model JSAMS-1157; No. of Pages 5
ARTICLE IN PRESS S. Bronner et al. / Journal of Science and Medicine in Sport xxx (2015) xxx–xxx
Johnson Foundation Health Games Research (Grant No. #66729) and Long Island University Brooklyn Health and Wellness Institute. References 1. Betker AL, Desai A, Nett C et al. Game-based exercises for dynamic short-sitting balance rehabilitation of people with chronic spinal cord and traumatic brain injuries. Phys Ther 2007; 87:1389–1398. 2. Deutsch JE. Using virtual reality to improve walking post-stroke: translation to individuals with diabetes. J Diabetes Sci Technol 2011; 5:309–314. 3. Gaffurini P, Bissolotti L, Calza S et al. Energy metabolism during activitypromoting video games practice in subjects with spinal cord injury: evidences for health promotion. Eur J Phys Rehabil Med 2013; 49:23–29. 4. Mirelman A, Maidan I, Deutsch JE. Virtual reality and motor imagery: promising tools for assessment and therapy in Parkinson’s disease. Mov Disord 2013; 28:1597–1608 (official journal of the Movement Disorder Society). 5. Ungerleider LG, Doyon J, Karni A. Imaging brain plasticity during motor skill learning. Neurobiol Learn Mem 2002; 78:553–564. 6. Kramer AF, Erickson KI. Capitalizing on cortical plasticity: influence of physical activity on cognition and brain function. Trends Cogn Sci 2007; 11:342–348. 7. Lange B, Flynn SM, Rizzo AA. Game-based telerehabilitation. Eur J Phys Rehabil Med 2009; 45:143–151. 8. Shumway-Cook A, Woollacott MH. Motor Control Translating Research into Practice, 3rd ed. Philadelphia, PA, Lippicott Williams & Wilkins, 2007. 9. Bailey BW, McInnis K. Energy cost of exergaming: a comparison of the energy cost of 6 forms of exergaming. Arch Pediatr Adolesc Med 2011; 165:597–602. 10. Graves LE, Ridgers ND, Williams K et al. The physiological cost and enjoyment of Wii Fit in adolescents, young adults, and older adults. J Phys Act Health 2010; 7:393–401. 11. Lyons EJ, Tate DF, Ward DS et al. Energy expenditure and enjoyment during video game play: differences by game type. Med Sci Sports Exerc 2011; 43:1987–1993. 12. Bronner S, Pinsker R, Noah JA. Energy cost and game flow of 5 exer-games in trained players. Am J Health Behav 2013; 37:369–380. 13. Mellecker RR, McManus AM. Active video games and physical activity recommendations: a comparison of the Gamercize Stepper, XBOX Kinect and XaviX J-Mat. J Sci Med Sport 2014; 17:288–292. 14. Whitehead A, Johnston H, Nixon N et al. Exergame effectiveness: what the numbers can tell us, In: Proceedings of the fifth ACM SIGGRAPH symposium on video games. New York, NY, Association for Computing Machinery (ACM) Press, 2010. p. 55–62. 15. Bianchi-Berthouze N, Kim WW, Patel D. Does body movement engage you more in digital game play? and why? In: Paiva ACR, Prada R, Picard RW, editors. Affective computing and intelligent interaction, second international conference, ACII, vol. 4738. Portugal, Lisbon, 2007, p. 102–113.
5
16. Piva SR, Gil AB, Almeida GJ et al. A balance exercise program appears to improve function for patients with total knee arthroplasty: a randomized clinical trial. Phys Ther 2010; 90:880–894. 17. Husswirth C, Bigard AX, Le Chevalier JM. The Cosmed K4 telemetry system as an accurate device for oxygen uptake measurements during exercise. Int J Sports Med 1997; 18:449–453. 18. Brockmyer JH, Fox CM, Curtiss KA et al. The development of the game engagement questionnaire: a measure of engagement in video game-playing. J Exp Social Psychol 2009; 45:624–634. 19. World Health Organization. Global recommendations on physical activity for health, Geneva, Switzerland, World Health Organization, 2010. 20. Bernstein N. The Coordination and Regulation of Movement, NY, Pergamon Press, 1967. 21. Vereijken B, van Emmerik REA, Whiting HT et al. Free(z)ing degrees of freedom in skill acquisition. J Mot Behav 1992; 24:133–142. 22. Pasch M, Bianchi-Berthouze N, van Dijk B et al. Movement-based sports video games: investigating motivation and gaming experience. Entertain Comput 2009; 1:49–61. 23. Thin AG, Brown C, Meenan P. User experiences while playing dance-based exergames and the Influence of different body motion sensing technologies. Int J Comput Games Technol 2013; 2013:1–7. 24. O’Donovan C, Hussey J. Active video games as a form of exercise and the effect of gaming experience: a preliminary study in healthy young adults. Physiother 2012; 98:205–210. 25. Sell K, Lillie T, Taylor J. Energy expenditure during physically interactive video game playing in male college students with different playing experience. J Am Coll Health 2008; 56:505–511. 26. Lin HH, Wang YS, Chou CH. Hedonic and utilitarian motivations for physical game systems use behavior. Int J Hum-Comput Interact 2012; 28: 445–455. 27. Thin A, Hansen L, McEachen D. Flow experience and mood states while playing body movement-controlled video games. Games Cult 2011; 6:414–428. 28. Jennett C, Cox AL, Cairns P et al. Measuring and defining the experience of immersion in games. Int J Hum-Comput Stud 2008; 66:641–661. 29. Örtqvist D, Liljedahl M. Immersion and gameplay experience: a contingency framework. Int J Comput Games Technol 2010:1–11. 30. Charbonneau E, Buchanan S, Bott J et al. Dance enhanced: investigating how earning content through exertion impacts dance game enjoyment, Paper presented at: ninth annual conference on the foundations of digital games, 2014. 31. O’Donovan C, Hirsch E, Holohan E et al. Energy expended playing Xbox Kinect and Wii games: a preliminary study comparing single and multiplayer modes. Physiotherapy 2012; 98:224–229. 32. Leninger LJ, Coles MG, Gilbert JN. Comparing enjoyment and perceived exertion between equivalent bouts of physically interactive video gaming and treadmill walking. Health Finess J 2010; 3:12–18.
Please cite this article in press as: Bronner S, et al. Physiological and psychophysiological responses to an exer-game training protocol. J Sci Med Sport (2015), http://dx.doi.org/10.1016/j.jsams.2015.03.003
