Learning curve and motor retention of a video game in young and older adults

The purpose of the study was to compare the learning curve and motor retention of the Dance Dance Revolution (DDR) video game in healthy adults. Twenty young (M= 23.9 ± 2.8 yr.) and 18 older adults (M= 60.7 ± 5.9 yr.), were randomly assigned to two experimental conditions: a) DDR 7 trials and b) DDR 14 trials. Participants danced the same song six sessions, followed by a detraining period of eight days. Then participants returned to the laboratory and danced again in order to detect a motor retention effect. A three-way ANOVA revealed mean score interactions (sessions x groups x trials; p = 0.017). Follow-up analyses revealed differences between young and older participants in both trial sessions (p < 0.05). Compared to young adults, older adults showed a learning curve of four sessions when performing 14 trials per session. After eight days of detraining only older participants in the DDR14 condition reduced motor performance. In conclusion, older subjects can learn the DDR game after playing four sessions; however, those dancing the same song 14 times were more likely to reduce their scores after eight days of detraining. Younger participants scored high regardless of the number of trials and sessions with little variability.


Introduction
A recent report (The Entertainment Software Association, 2013), revealed that the videogame industry produced in 2011, approximately $24.75 billion, and these figures are likely to increase in the years to come.Indeed, videogames are becoming an integral part of our daily lives.Both, children and adults, are widely engaged in videogame playing for several hours during a day (Richards, McGee, Williams, Welch, & Hancox, 2010;Wethington, Sherry, Park, Blanck, & Fulton, 2013).
Scientific studies regarding the acute and chronic effects of playing videogames on different variables have been done before (Moncada-Jiménez & Chacón-Araya, 2012).In general, the findings are contradictory for psychological variables.Some meta-analysis and other reviews suggest a correlation between excess time video gaming on negative social and psychological aspects such as isolation and aggressive behavior (Anderson, 2004;Anderson & Bushman, 2001;Anderson et al., 2010;Cummings & Vandewater, 2007;Richards et al., 2010;Sherry, 2001); while other research suggests a positive association with motor learning, adoption of healthy habits (including increased energy expenditure), motor re-training and resilience (Baranowski et al., 2011;Thompson, Baranowski, & Buday, 2010;White, Schofield, & Kilding, 2011).
The Nintendo Wii console and the Dance Dance Revolution (DDR) videogame (Konami Corporation, Japan) are among the most studied exergaming products (Gao & Podlog, 2012;Murphy et al., 2009).This game is a more sophisticated form of an earlier arcade game.The purpose of the game is to use the feet to push arrows drawn in a mat according to a stimuli (i.e., music) shown on a video display, usually a television screen.It is expected that the user move the body according to a specific rhythm or choreography displayed in the video game to score high in the game.
To the best of our knowledge, the manner exergames are learned has not been recently described in the literature.For some authors (Singer & Berrocal, 1986), processes of adaptation to a movement arise depending on the activity and the situation, and when internal control is achieved then external variables no longer distract the performer.It has been proposed that states of motivation, attention, concentration and decision making are improved when a motor skill is mastered (Schmidt & Lee, 2005;Singer & Berrocal, 1986).It follows that a learning curve may allow determining whether or not a skill has been mastered (Song, 2009;Vendituoli, 2008).A learning curve function describes the degree of success attained during a period of instruction (i.e., score over time) (Singer, 1982).When mastery is achieved it is expected a gradual increase in performance (e.g., speed, accuracy) accompanied by a reduction in variability (Adi-Japha, Karni, Parnes, Loewenschuss, & Vakil, 2008).
Learning curves are also used as a feedback tool for performers since motor learning is unique for a particular subject and a particular skill (Magill, 1993;Schmidt & Lee, 2005;Singer, 1982).Skill performance might be impaired or lost due to a lack of practice following training (i.e., detraining) or inappropriate learning and practice, which might be related to motor retention (Magill, 1993;Schmidt, 1975;Singer, 1982).Motor retention has been defined as the ability to maintain high standards of physical performance (e.g., speed, coordination), effectiveness and consistency in a motor task (Bertollo, Berchicci, Carraro, Comani, & Robazza, 2010;Hynes-Dusel, 2002).Performance in a motor task is impaired when practice is discontinued and this reduction tends to appear earlier in high demanding tasks as opposed to low demanding tasks (Bertollo et al., 2010).
In spite of a growing body of evidence showing the usefulness of video games for improving health-related variables, there is still scarce information regarding how long does it take for a person to learn how to play a video game; in other words, the motor learning and motor control as it relates to exergaming.Therefore, the purpose of the study was to determine the learning curve of the DDR videogame in a sample of young and older adults, as well as to determine if any retention occurred after a detraining period.

Participants
Thirty-eight volunteers were recruited from a university campus and a service program for older adults.Potential participants were naive to the DDR 2 Hottest Party ® (Nintendo Wii) videogame or had little or no experience in exergaming.

Measurement instruments
The DDR videogame was played from a Nintendo Wii console and four control mats of the same brand (Nintendo of America, Inc., Redmond, WA) placed in front a 36" television screen (Sony Bravia, Japan).Participants wore a heart rate monitor Polar, model T-61 (Polar Group, Oulu, Finland).
Data were collected from the scoring system of the DDR display; however, the original scores were multiplied and transformed into new scores by using the beginner mode of the DDR game.The original scores were multiplied as follows: a) Perfect x 7; b) Great x 6 pts., c) Good x 5 pts., d) Almost x 4 pts., e) Boo x 3 pts., f) OK x 2 pts., and g) NG x 1 pt.
The highest achievable score was 322 pts., defined as the number of correct arrow steps shown for the song (46 arrows) multiplied by the maximal score in the scale (i.e., 7 pts.).The lowest achievable score was 48 pts., as defined by lowest number of arrow steps (16 arrows) multiplied by the score of 3 pts.allowed by the software.Notice that the scores OK and NG are not taken into consideration by the videogame software even when a participant stands still in front of the video display.

Procedures
Younger and older participants were randomly assigned to two groups based on the number of trials to be performed in the DDR.Participants in both age groups were randomly assigned to perform either seven (DDR7) or fourteen (DDR14) trials of the same song during six sessions (days) within a three week period (Bertollo et al., 2010).All participants were required to perform the study in the Human Movement Sciences Research Center under controlled environmental conditions and a predefined distance of the mat from the television screen.Participants read and signed an informed consent to participate in the study in accordance to the standards set forth by the Ethics Scientific Committee of the University of Costa Rica.
Participants were required to visit the laboratory to read and sign an informed consent form.During the same visit, participants were instructed on how to play the videogame by an oral explanation of one of the researchers, who read the same document to all participants to avoid confusion and ensuring for homogenous instructions.Physical exertion was monitored at the end of each trial by placing a heart rate monitor to each participant.
The experimental sessions required participants to dance the same song on each of the six exercise sessions.A pilot study on a different group of people with similar characteristics to the participants of this study revealed that the song Black or White (Jackson, 1991) and the Beginner level was the easiest combination of a song and a difficulty level for participants to play.Performance scores and heart rate values were recorded at the end of each trial.Once the experimental protocol was finished, a detraining period of eight days was implemented, followed again by a single exercise session.Participants were instructed on not playing any videogames during eight days.Then, participants were appointed to perform one last time the DDR videogame.The score in this last exercise trial was called the «retention» effect.

Statistical analysis
Statistical analysis was performed with the IBM SPSS Statistics for Windows version 20 (IBM Software Group, Chicago, IL, USA).Values are presented as mean and standard deviation (M ± SD).Statistical significance was set a priori at P < 0.05.Inferential statistics included mixed three-way (groups x experimental treatments x sessions) ANOVA for the score variable and appropriate follow-up analyses.Coefficient of variation ([M/SD] x 100) was also used to describe learning curves (Vincent & Weir, 2012).

Results
Young (n = 20) and older adults (n = 18) participated in the study.Physical characteristics are described in table 1. Mean scores and heart rate response to combined sessions is presented in table 2 for younger and older participants.
Three-way ANOVA showed triple (p = 0.017) and double significant interactions (groups x sessions) (p = 0.001), and significant main effects for sessions (p = 0.001) and groups (p = 0.001) in DDR scores.Two-way ANOVA comparing the two age groups and the two trial regimens (i.e., DDR7 vs. DDR14) revealed significant interaction (p < 0.05), and main effects for trials (p < 0.05), and sessions (p < 0.05) in DDR scores.One-way ANOVA revealed significant differences in the within-session DDR mean scores in the experimental group DDR7 (p < 0.05).Tukey's HSD post hoc analysis indicated that significant mean differences were obtained in session 6 vs. sessions 1, 2, 3; session 5 vs. sessions 1 and 2; and between session 1 vs. sessions 3, 4 and 7. One-way ANOVA also revealed significant differences in the withinsession DDR mean scores in the experimental group DDR14 (p < 0.05).Tukey's HSD post hoc analysis indicated that significant mean differences were obtained in sessions 4, 5, 6 vs. sessions 1, 2, 7; sessions 3, 7, 2 vs. session 1; and between session 2 to 6 in both groups.
Among younger participants performing in the DDR7 and DDR14 groups, only the main effect session was statistically significant (p < 0.05) for the DDR scores.One-way ANOVA failed to reveal significant mean differences in DDR scores in the within-sessions factor.
Two-way ANOVA revealed a significant interaction (p < 0.05) and main effect group (p d» 0.05) and sessions (p < 0.05) in DDR scores among younger and older participants performing DDR7.One-way ANOVA revealed significant mean differences among all sessions in both age groups (p < 0.05).
Two-way ANOVA revealed a significant interaction (p < 0.05) and main effect group (p < 0.05) and sessions (p < 0.05) in DDR scores among younger and older participants performing in the DDR14 groups.One-way ANOVA revealed significant mean differences among all Note: DDR = Dance Dance Revolution; DDR7 = Dance Dance Revolution 7 trials; DDR14 = Dance Dance Revolut ion 14 trials; HR = Heart rate * Resting heart rate before session start (pre).** Mean session's heart rate (post) Mean score and heart r ate resp onses in trial grou ps.

Figure 1 .Figure 2 .
Figure 1.Learning curves following s even or fourteen DDR video game trials in young adults (n= 20) (panel A) and older adults (n = 18) (panelB).Significant differences were observ ed between young and older participants (p < 0.05).Asterisk represent s significant differences between trials (p < 0.05).Differen t letters represent significant differences between sess ions (p < 0.05).