18

Motor Control, 2013, 17, 18-33 © 2013 Human Kinetics, Inc.

Official Journal of ISMCwww.MC-Journal.comORIGINAL RESEARCH

Miura, Kudo, Ohtsuki, and Nakazawa are with the Department of Life Sciences, Graduate School of Arts and Sciences, the University of Tokyo, Tokyo, Japan. Miura is also with the Japan Society for the Promotion of Science, Tokyo, Japan. Kanehisa is with National Institute of Fitness and Sports in Kanoya, Kanoya, Kagoshima, Japan.

Relationship Between Muscle Cocontraction and Proficiency in

Whole-Body Sensorimotor Synchronization: A Comparison Study of

Street Dancers and Nondancers

Akito Miura, Kazutoshi Kudo, Tatsuyuki Ohtsuki, Hiroaki Kanehisa, and Kimitaka Nakazawa

In this study, we investigated muscle cocontraction during a street dance movement task to provide evidence that the level of muscle cocontraction is associated with degree of proficiency in whole-body sensorimotor synchronization movement. Skilled street dancers and nondancers were required to synchronize a knee-bending movement in a standing position to a metronome beat. The dancer group showed significantly smaller variability of temporal deviation (defined as the peak knee-flexion time minus beat onset time), and lower level of muscle cocontraction analyzed by electromyographic data of the agonist and antagonist muscles of the upper and lower leg than did the nondancer group. In addition, multiple regres-sion analyses revealed that the group effect significantly predicted the level of muscle cocontraction. These results show that the level of muscle cocontraction in the lower limbs during whole-body sensorimotor synchronization movement is associated with the degree of proficiency of the movement.

Keywords: sensorimotor synchronization, cocontraction, human motor learning, whole-body movement, electromyography, dance

Electromyographic (EMG) analysis is used to investigate proficiency in motor skills. EMG activity in skilled movements can be quantified in terms of timing (Chapman, Vicenzino, Blanch, & Hodges, 2008a; Girard, Micallef, & Millet, 2005; Hirashima, Kadota, Sakurai, Kudo, & Ohtsuki, 2002; Sakurai & Ohtsuki, 2000), duration (Chapman, et al., 2008a), or amount (Ertan, Kentel, Tümer, & Korkusuz, 2003; Kudo & Ohtsuki, 1998; Takaishi, Yamamoto, Ono, Ito, & Moritani, 1998). In addition, measures of cocontraction have been used to quantify the coordination of agonist and antagonist muscle pairs; previous studies have reported decreased

Muscle Cocontraction in Dancers and Nondancers 19

levels of agonist/antagonist cocontraction to associate with increased proficiency or motor learning (Fujii, Kudo, Shinya, Ohtsuki, & Oda, 2009; Gribble, Mullin, Cothros, & Mattar, 2003; Osu et al., 2002; Thoroughman & Shadmehr, 1999). In discrete movements such as goal-directed arm tasks, practice has been found to lead to decreased cocontraction of arm flexor and extensor muscles (Gribble, et al., 2003; Osu, et al., 2002; Thoroughman & Shadmehr, 1999). In continuous rhythmic movements such as a rapid finger-oscillating task, Heuer (2007) reported a decreased level of muscle cocontraction in the skilled (i.e., dominant) hand compared with the unskilled (i.e., nondominant) hand as measured by the relative difference signal (RDS). RDS is calculated from normalized EMG signals from flexor and extensor muscles, such that the distribution of RDS values reflects the balance between cocontraction and reciprocal contraction. This analysis has been helpful in assessing the level of muscle cocontraction in continuous rhythmic movement; Fujii et al. (2009) used RDS to reveal that expert drummers showed less muscle cocontraction during rapid tapping movement than did novices.

The decreased level of muscle cocontraction associated with motor learning has been investigated mainly in finger and arm movements. However, the functional, neuroanatomical, and physiological differences between the finger/arm and other body parts mean that these observations cannot be generalized to other parts of the body. For example, several differences have been found between the upper and lower limbs; the lower limb muscles have higher innervation ratios (muscle fibers per motor unit) than the upper limb ones (Galea, De Bruin, Cavasin, & McComas, 1991), and it has been suggested that corticospinal projections to lower limb moto-neurons are weaker than those to the upper limb ones (Brouwer & Ashby, 1990). Moreover, the H-reflex recovery curve, defined as the time course of recovery of H-reflex depression after activity, differs between the upper and lower limbs, sug-gesting a difference in synaptic efficacy (Rossi-Durand, Jones, Adams, & Bawa, 1999). In spite of these differences, less cocontraction of the lower limb muscles during cycling movement in the more skilled was observed (Candotti et al., 2009; Chapman, Vicenzino, Blanch, & Hodges, 2008b). Asaka, Wang, Fukushima, & Latash (2008) reported decreased cocontraction of lower limb muscles following motor learning in a standing load-release task on an unstable surface. In this task, the activity of the lower limb muscles related with postural control, a different functional role than that of lower limb muscles during cycling movement or upper limb muscles. These findings suggest that the neural mechanism underlying the decrease in the level of muscle cocontraction may be common to human motor learning, leading us to hypothesize that it may also be observed in lower limb muscles during whole-body coordinated movement in a standing position. To clarify the relationship between muscle cocontraction and proficiency in whole-body coordinated movement, we conducted a comparative study between experts in street dance and novices, using a whole-body sensorimotor synchronization task similar to that typically performed by street dancers.

We adopted a street dance movement as an example of whole-body rhythmic movement in a standing position. Street dance is a style of dancing to music that was developed by African-Americans (Shichirui, 1999). The music is characterized by strong beats, typically at a rate of at least 100 beats per minute. Street dance involves a basic movement in which a dancer bends the knees to the beat while in a standing position. Because this movement involves dynamic stance postural

20 Miura, et al.

control, the level of muscle cocontraction in the legs can be investigated while they function in a different role from that of the arms, that is, postural control. Fransson, Gomez, Patel, & Johansson (2007) reported greater EMG activity in the tibialis anterior and gastrocnemius muscles of subjects standing on an unstable foam surface than those standing on a firm one. Similarly, Anderson and Behm (2005) observed increased EMG activity in the soleus and trunk muscles in subjects per-forming a squat movement on an unstable surface. Muscle cocontraction has also been revealed to enhance spatial accuracy in arm movement (Gribble, et al., 2003). Considering these findings, it may be assumed that if the postural movement used in the current study becomes unstable, the level of muscle cocontraction in the leg muscles will increase. Asaka et al. (2008) have reported that practice in a standing load-releasing task on an unstable surface led to decreased cocontraction of the leg muscles, which demonstrates that decreased muscle cocontraction can be caused by motor learning in a whole-body task requiring postural control. Therefore, we hypothesized that skilled street dancers would show a lesser level of leg muscle cocontraction (as measured with RDS) during a coordinated whole-body movement when compared with novices.

Another topic of this study was the relationship between movement frequency and the decreased level of muscle cocontraction associated with increased profi-ciency. Thus far, research on the decreased level of muscle cocontraction associ-ated with increased proficiency in rhythmic movement has been conducted only at maximum movement frequencies (Fujii, et al., 2009; Heuer, 2007). At maximum movement frequency, it may be assumed that movement speed is quite large, and therefore, greater recruitment of fast-twitch fibers are elicited. As movement fre-quency decreases, contribution of slow-twitch fiber to force production is thought to increase. Because slow-twitch fibers take more time to relax than fast-twitch ones (Bellemare, Woods, Johansson, & Bigland-Ritchie, 1983) at slow movement speed muscle cocontraction level is assumed to be higher than that at high movement speed. Hence, it is hypothesized that the decreased level of cocontraction associated with increased proficiency at slow movement speed would not be observed. We changed movement speed by setting wide range of movement frequency specified metronome beat. This study, therefore, fills a secondary purpose by investigating the relationship between muscle cocontraction level and movement frequency during whole-body rhythmic movement in both skilled street dancers and novice controls.

MethodsParticipants

Eight skilled street dancers and 9 control participants took part in this experiment. The street dancers (male, age 24.3 ± 4.4 years, mean ± SD) had 7.3 ± 4.3 years of dancing experience and included 6 winners of celebrated national or international street dance competitions. The novice controls (male, age 24.0 ± 1.3 years) had no experience in street dance. Hereafter, we refer to the dancer group simply as “Dancers” and the control group as “Nondancers.” All participants were free of musculoskeletal disorders and were in excellent physical condition. They gave writ-ten informed consent to participate in the experiment. This study was approved by the Ethics Committee of the Graduate School of Arts and Sciences, the University of Tokyo.

Muscle Cocontraction in Dancers and Nondancers 21

Experimental Task

The experimental task was a basic street dance movement, consisting of bouncing up and down repeatedly by bending at the knees in synchronization to a metronome beat (Figure 1). The participants were instructed to flex their knees on the beat, and to keep a 1:1 relationship between knee movement and beat. Tested frequencies ranged from 40 to 180 beats per minute (bpm) at 20 bpm intervals. The participants were also instructed not to move the neck or parts of the body other than the hip, knee, and ankle joints.

Design and Procedure

The participants performed a total of 8 trials, each of which using a different fre-quency. The trial duration was about 30 s, 24 s of which were analyzed. The trial order was randomized. Immediately before each trial, all participants watched a short instructional video in which a street dancer performed the trial movement. The participants were allowed to practice the movement for a few minutes during and after watching the video, before beginning the trial. Between trials, the participants had sufficient rest breaks to exclude the influence of fatigue. The participants were instructed to try to adjust their movements to the beat while in motion, even if they could not precisely match the beat.

Apparatus and Data Collection

EMG activity was recorded with bipolar Ag/AgCl disposable electrodes (10 mm in diameter) placed 3 cm apart between the center of the electrodes on the tibialis anterior (TA), soleus (SOL), medial gastrocnemius (MG), vastus lateralis (VL), rectus femoris (RF), and biceps femoris (BF) muscles of the right lower limb. Before the electrodes were attached, the skin was shaved, treated with alcohol, and then rubbed with fine sandpaper to reduce the interelectrode resistance. EMG signals were sampled at 1,000 Hz with an MP100 recording system (Biopac Systems, Tokyo, Japan) and stored in a personal computer using wave recording/analyzing software (AcqKnowledge 3.7.3 for Windows; Biopac Systems).

Figure 1 — Experimental movement task. Participants were instructed to flex their knees rhythmically according to a metronome beat.

22 Miura, et al.

All trials were videotaped with a high-speed camera (HAS-220, Ditect, Tokyo, Japan) in the sagittal plane. The EMG signals and the video images were synchro-nized by the common electronic pulse which triggered the EMG and the image recordings. Digital images were recorded at an acquisition rate of 200 Hz for 24 s, and all of these images were included in the analysis. Reflective markers were attached to the right side of the acromion, the greater trochanter, rotation axis of the knee joint, lateral malleolus and first metatarsophalangeal joint. Marker coordi-nates were digitized using motion analysis software (DIPP-Motion Pro, Ditect). To eliminate start-up effects, we began recording after the participant had performed about 10 movement cycles.

Data Analysis

Out of 136 trials (8 trials × 17 participants), 6 were excluded before analysis. Three were excluded due to large EMG noise and 3 due to participants not maintaining a 1:1 relationship between the movement and the beat.

Hip, knee, and ankle angular displacements were computed from the coor-dinates of the reflective markers. For each joint, full extension is of about 180 degrees, with a smaller angle indicating flexion. The angular displacement data were smoothed using a low-pass filter (cut-off frequency: 6 Hz) and were dif-ferentiated numerically to obtain angular velocities. Angular displacement range was determined by averaging peak to peak (positive to negative and negative to positive) amplitudes within a cycle. The negative and positive peaks of angular velocity within a cycle were also averaged.

Temporal deviation (the peak knee-flexion time minus beat onset time) was used to evaluate the accuracy of the movement.

We also evaluated the extent of muscle cocontraction based on RDS (Heuer, 2007). Initially, the direct-current component of each EMG signal was removed and the resulting signals were full-wave rectified, smoothed by means of a moving average of 9 ms per window width that is corresponding to low-pass filter: cut-off frequency about 50 Hz (Letchford, Sandri, Levitan, & Mehta, 1992), and scaled to a mean of 1. For each time point, the relative difference (E − F)/(E + F) was computed, where E and F denote the scaled extensor and flexor signals respectively. Values near 0 indicate dominant cocontraction, whereas values approaching +1 or −1 indicate a dominant reciprocal contraction. The distribution of RDS values therefore reflects the balance between cocontraction and reciprocal contraction. A small variance in RDS values indicates that cocontraction of antagonistic muscles is dominant, while a large variance indicates that reciprocal contraction is dominant. We computed this value for each of the agonist-antagonist pairs, i.e., SOL–TA, MG–TA, VL–BF, and RF–BF.

Statistical Analysis

First, any missing data were replaced with the mean value of all participants at that movement frequency. Then, a 2-way analysis of variance (ANOVA) with 1 within-subject factor (beat rate; 40, 60, 80, 100, 120, 140, 160, or 180 bpm) and 1 between-subject factor (group; Dancer vs. Nondancer) was performed on (1) the

Muscle Cocontraction in Dancers and Nondancers 23

mean and sample SD of the temporal deviation, (2) the angular displacement range, (3) the angular velocity peaks (positive and negative), and (4) the RDS variance for each muscle pair. In the ANOVA, when Mauchly’s test of sphericity revealed heterogeneity of covariance, the more conservative Greenhouse–Geisser test was performed. Statistical significance was set at p < .05.

To determine which variable best predicted the level of muscle cocontraction, we conducted forward-backward stepwise multiple regression analyses, in which independent variables were entered into and removed from the discriminant func-tion one at a time on the basis of their discriminating power (Hair, Black, Babin, & Anderson, 2009). We set RDS variance as the dependent variable and group (as dummy variable), beat rate, and kinematic variables (angular displacement range and negative and positive peaks of angular velocity) as independent variables, and used these to build a multiple regression model for each of the agonist-antagonist pairs. We used data obtained from each participant for the regression analysis. To avoid multicollinearity, variance inflation factor (VIF) was calculated. Because the VIFs larger than 10 indicates serious problem of multicollinearity (Montgomery, Peck, & Vining, 2007), independent variables that showed VIFs larger than 10 were removed and the regression analysis was performed again.

Results

Performance Evaluation

Figure 2 shows the mean and sample SD of the temporal deviation as a function of beat rate. Interactions were not significant either for the mean of temporal deviation (F7, 105 = .477, p > .84) or the sample SD of the temporal deviation (F7, 105 = .642, p > .72). The mean of the temporal deviation did not differ significantly between groups (F1, 15 = 3.953, p > .06), and the sample SD of the temporal deviation was significantly different between groups (F1, 15 = 5.851, p < .05), which indicates that Dancers performed the knee-bending movement with greater temporal precision than Nondancers regardless of movement frequency. Significant main effects of beat rate were found on the mean of the temporal deviation (F3.372, 50.587 = 49.993, p < .001) and on the sample SD of the temporal deviation (F2.628, 39.425 = 16.287, p < .001).

Angular Displacement Range

Figure 3 shows the range of angular displacement of each joint as a function of beat rate. As beat rate increased, the range of angular displacement got small at all joints, and Dancers exhibited significantly greater motion of knee and ankle joints at all beat rates than did Nondancers. There was no significant interaction for the hip joint (F7, 105 = .400, p > .90), for the knee joint (F7, 105 = 1.568, p > .15), and for the ankle joint (F7, 105 = 1.564, p > .15). Significant main effects of group were found in the knee joint (F1, 15 = 5.830, p < .05), in the ankle joint (F1, 15 = 4.633, p < .05), and not in the hip joint (F1, 15 = .926, p > .35). Significant main effects of beat rate were revealed in the hip joint (F2.015, 30.232 = 55.959, p < .001), the knee joint (F2.110,

31.644 = 173.337, p < .001), and the ankle joint (F1.736, 26.043 = 146.968, p < .001).

24 Miura, et al.

Angular Velocity Peaks

Figure 4 shows mean positive and negative peaks of angular velocity in each joint as functions of beat rate, and they depict characteristic curves. For example peaks of positive and negative angular velocity in the knee joint were not at 180 bpm but at around 100 bpm. Dancers displayed significantly greater absolute angular velocity mainly in knee joint than did Nondancers over wide range of beat rates. For positive peaks of angular velocity, significant interaction effects were revealed in the hip joint (F7, 105 = 2.915, p < .01), in the knee joint (F7, 105 = 4.842, p < .001), and in the ankle joint (F7, 105 = 4.480, p < .001). This indicates that the effect of beat rate differed between Dancers and Nondancers.

For negative peaks of angular velocity, no interaction was found in the hip joint (F7, 105 = .411, p > .89), the knee joint (F7, 105 = .423, p > .88), and the ankle joint (F7, 105 = .558, p > .78). Significant main effect of group was found in the knee joint (F7, 105 = 5.684, p < .05), not in the hip joint (F7, 105 = 1.973, p > .18) and ankle joint (F7, 105 = 4.187, p > .05), and significant main effects of beat rate in the hip joint (F2.921, 43.809 = 3.299, p < .05), in the knee joint (F3.032, 45.474 = 10.606, p < .001), and in the ankle joint (F1.808, 27.120 = 7.959, p < .01) were found.

Figure 2 — (A) Mean temporal deviation (the peak knee-flexion time minus beat onset time). (B) Sample SD of temporal deviation. Vertical bars represent between-subjects standard errors.

Figure 3 — Ranges of angular displacement for the hip, knee, and ankle joints. Vertical bars represent between-subjects standard errors.

Muscle Cocontraction in Dancers and Nondancers 25

RDS Variance

Figure 5 illustrates examples of EMG activity and RDS in upper leg muscles during the movement task at 120 bpm for a typical participant from each group. The RF activated around the point of peak knee-flexion in both participants (Figure 5, third row). However, whereas the BF cocontracted with the RF in the Nondancers, the 2 muscles showed reciprocal contraction in the Dancer (Figure 5, fourth row). Hence, the RDS of Dancers tends to jump between values close to ± 1, whereas that of Nondancers tends to cluster around zero (Figure 5, fifth row).

Figure 6 depicts RDS variance for each of the agonist-antagonist muscle pairs as functions of beat rate. Nondancers exhibited significantly greater cocontraction of antagonist muscle pairs in all agonist-antagonist pairs than did Dancers. As for RF–BF pair, a significant interaction effect was found (F7, 105 = 3.591, p < .01), indicating that the effect of beat rate differed between Dancers and Nondancers. No interaction was found in the SOL–TA pair (F7, 105 = .571, p > .77), in the MG–TA pair (F7, 105 = .679, p > .69), and in the VL–BF pair (F7, 105 = 1.835, p > .08). Significant main effects of group were found in the SOL–TA pair (F1, 15 = 8.358, p < .05), in the MG–TA pair (F1, 15 = 5.530, p < .05), and in the VL–BF pair (F1, 15 = 10.738, p < .01). Significant main effects of beat rate were found in the MG–TA pair (F3.876, 58.137 = 3.950, p < .01), in the VL–BF pair (F2.603, 39.046 = 37.817, p < .001) and not in the SOL–TA pair (F3.106, 46.589 = 1.867, p > .14).

Multiple Regression Analysis

A separate regression analysis was run for each agonist-antagonist muscle pair (SOL–TA, MG–TA, VL–BF, or RF–BF), with RDS variance as the dependent vari-able, and beat rate, group, angular displacement range, and positive and negative

Figure 4 — Angular velocity peaks for the hip, knee, and ankle joints. The graphs in the top row depict positive peaks and the bottom row, negative peaks. Vertical bars represent between-subject standard errors.

26

Figure 5 — Examples of kinematics, processed EMG signals, and relative difference signals (RDS). The top row illustrates stick pictures representing stages of movement. The second row shows graphs of angular knee displacement. The next 2 rows show processed EMG activity in the RF (third row) and BF (fourth row). The bottom row shows the RDS.

Figure 6 — RDS variances for each agonist-antagonist muscle pair. Vertical bars represent between-subject standard errors.

Muscle Cocontraction in Dancers and Nondancers 27

peaks of angular velocity as independent variables. For example, for those agonist-antagonist pairs that include biarticular muscles, angular displacement range and positive and negative peaks of angular velocity in the 2 joints involved were used as the independent variables. Table 1 shows the summary of the results. The level of muscle cocontraction in all agonist-antagonist muscle pairs can be attributed to group effect, independent of other variables, because group was found to be a significant predictor of RDS variance for the all muscle pairs: the SOL–TA pair (standardized β = .328, p < .001), the MG–TA pair (standardized β = .252, p < .01), the VL–BF pair (standardized β = .238, p < .001), and the RF–BF pair (standard-ized β = .349, p < .001). Beat rate was also found to be a significant predictor of RDS variance for the MG–TA pair (standardized β = .202, p < .01), the VL–BF pair (standardized β = .702, p < .001), and the RF–BF pair (standardized β = .652, p < .001). A negative peak of angular velocity in ankle joint was a significant pre-dictor of RDS variance for the SOL–TA pair (standardized β = −.173, p < .05). A positive peak of angular velocity in the ankle joint and negative peak of angular velocity in the knee joint were found to be significant predictors of RDS variance for the MG–TA pair (standardized β = −.300, p < .05 and standardized β = −.532, p < .01, respectively).

Table 1 Results of Multiple Regression Analyses for Predicting RDS Variance

Dependent variable R2 Explanatory variable

Standardized regression

coefficient βVariance of RDS in SOL-TA pair

.166 Group .328 ***

Negative peak of angular velocity in ankle joint

−.173 *

Variance of RDS in MG-TA pair

.262 Group .252 **

Beat rate .202 **

Positive peak of angular velocity in ankle joint

−.300 *

Negative peak of angular velocity in knee joint

−.532 **

Variance of RDS in VL-BF pair

.549 Group .238 ***

Beat rate .702 ***

Variance of RDS in RF-BF pair

.547 Group .349 ***

Beat rate .652 ***

Note. RDS: relative difference signal. * p < .05, ** p < .01, *** p < .001.

28 Miura, et al.

DiscussionThis study investigated the relationship between muscle cocontraction and pro-ficiency in whole-body coordinated movement in a standing position. Dancers and Nondancers performed a basic street dance movement, that is, a whole-body task consisting of bending the knees to different frequencies of auditory stimuli. Performance evaluation revealed that Dancers executed the task movement with greater temporal accuracy than did Nondancers. In addition, RDS analysis of the leg muscles revealed less cocontraction of antagonist muscle pairs of both the upper and lower leg in Dancers than in Nondancers. Multiple regression analyses indicated that the group significantly predicted the level of muscle cocontraction in all agonist-antagonist pairs. These results suggest that, in a rhythmic whole-body movement, the level of cocontraction is related to the degree of proficiency.

As an evaluation of performance, we calculated temporal deviation (the peak knee-flexion time minus beat onset time [Figure 2]). Over a wide range of fre-quencies, the peak knee-flexion time occurred after the beat onset in each cycle, with the mean temporal deviation not differing significantly between Dancers and Nondancers. As we instructed the participants to flex their knees on the beat, both groups successfully performed the task movement by matching knee flexion duration to the beat. However, the SD of temporal deviation differed significantly between Dancers and Nondancers, meaning that Dancers performed the task movement more precisely than did Nondancers over a wide range of frequencies. This result is consistent with our previous study, in which Dancers showed less variability of relative phase between a beat and a knee-bending movement than did Nondancers (Miura, Kudo, Ohtsuki, & Kanehisa, 2011).

Dancers showed greater temporal precision in the knee-bending movement, and they showed a lower level of muscle cocontraction than did Nondancers (Figure 6). Multiple regression analyses for each agonist-antagonist pair showed that group significantly predicted RDS variance (Table 1). These results indicate that the level of muscle cocontraction in each agonist-antagonist pair can be attrib-uted to group, independent of other variables such as angular displacement range and angular velocity, suggesting that the decreased level of muscle cocontraction is associated with increased proficiency. Muscle cocontraction has frequently been observed in unstable conditions. For example, wrist muscle cocontraction increases when unstable mechanical loads are attached to the wrist (De Serres & Milner, 1991; Milner, 2002; Milner, Cloutier, Leger, & Franklin, 1995). In addition, muscle cocontraction has revealed to decrease with motor learning or associate with the degree of proficiency (Fujii, et al., 2009; Gribble, et al., 2003; Osu, et al., 2002; Thoroughman & Shadmehr, 1999). Osu et al. (2002) reported that levels of muscle cocontraction around the shoulder and elbow joints in a reaching movement decreased with motor learning, and suggested that feedforward control acquired through practice allowed for a less stiff joint motion. That is, feedforward control acquired by practice is thought to compensate for unpredictable dynamic forces and therefore lead to decreased cocontraction (Heuer, 2007; Osu, et al., 2002). Feedforward control is thought to be achieved by internal models that can predict sensory consequences from copies of efferent motor commands and calculate the motor commands necessary to balance them (Kawato, 1999), or by anticipating synchronization arising from time delays in appropriately coupled dynamical sys-

Muscle Cocontraction in Dancers and Nondancers 29

tems (Stepp & Turvey, 2010). In the current study, when Nondancers are required to perform a movement task without the benefit of feedforward control, muscle cocontraction can be viewed as a strategy to deal with nonmuscular forces (e.g., interaction, inertial, and gravitational forces) encountered during the movement, although muscle cocontraction is energetically inefficient because it counteracts torque generated by agonist muscle activity as Thoroughman and Shadmehr (1999) called it a “wasted contraction.”

In the current study, Dancers are thought to have acquired feedforward control to compensate for such forces by long-term practice, considering following previ-ous studies revealing experienced dancers to have a style-specific ability to control movements. Previous research on postural control in classical ballet dancers has revealed enhanced postural stability acquired through specific dance training (Crotts, Thompson, Nahom, Ryan, & Newton, 1996; Gerbino, Griffin, & Zurakowski, 2007; Golomer, Cremieux, Dupui, Isableu, & Ohlmann, 1999; Golomer & Dupui, 2000; Golomer, Dupui, & Monod, 1997a, 1997b; Golomer, Dupui, Séréni, & Monod, 1999). Skilled classical ballet dancers have been found to use feedforward control to preserve verticality of the head-trunk axis when laterally raising one extended leg (Mouchnino, Aurenty, Massion, & Pedotti, 1992). In street dance, skilled dancers have been reported to exhibit less variability of the relative phase between a beat and a knee-bending movement than nondancers do (Miura, et al., 2011). The previous research and the present result that Dancers showed more accurate performance than Nondancers did (Figure 2B) suggest that Dancers relied on feedforward control more than Nondancers did.

Previous neuroimaging research observed a reduction in brain activities during motor learning process, reflecting an improvement in neural efficiency (Gobel, Parrish, & Reber, 2011; Reithler, van Mier, & Goebel, 2010). It was reported that expert shooters showed a reorganized cortical network that is characterized by less cortico-cortical communication than novices (Deeny, Haufler, Saffer, & Hatfield, 2009; Deeny, Hillman, Janelle, & Hatfield, 2003). In addition, trained athletes also exhibited less cortical activity than novices (see Nakata, Yoshie, Miura, & Kudo, 2010, for review). These findings, along with a previous study that reviewed the changes of the neural correlates that are associated with musical practice (Münte, Altenmüller, & Jäncke, 2002), suggest that the plastic changes in Dancers’ brain that are associated with long-term practice would enable feedforward control in whole-body rhythmic movements requiring dynamic postural control, thus lead-ing to decreased levels of muscle cocontraction, which needs to be examined in future research.

Previous research on the decreased level of muscle cocontraction associated with motor learning has mainly focused on finger/arm movements. Our results, along with those of Asaka et al. (2008), suggest that decreased muscle cocontrac-tion can occur in body parts other than the finger/arm despite many functional, neuroanatomical, and physiological differences between the arm and other body parts (see introduction). Moreover, in the current study, the motor task consisted of a basic movement that is the basis for almost all street dance forms. Hence, it is noteworthy that we observed a decreased level of muscle cocontraction in this skill closely associated with street dance performance. This suggests that the level of muscle cocontraction could be an important parameter in street dance, and possibly in other dance or sports performances. Impression of supple and pliable

30 Miura, et al.

performance of skilled dancers could be due to a lower level of muscle cocontrac-tion. On the contrary, the stiffened movements of beginners in street dance could be partly due to a greater level of cocontraction.

Previous research has shown that a decreased level of muscle cocontraction is associated with increased proficiency in continuous rhythmic movements such as a finger-oscillating movement (Heuer, 2007) or wrist-drumming movement (Fujii, et al., 2009) only at maximum movement frequency. Therefore, we attempted to confirm that the decreased level of muscle cocontraction associated with an increased degree of proficiency would occur at a lower movement frequency, at which muscle cocontraction was likely to occur originally because of greater recruitment of slow-twitch muscles that need longer time to relax (Bellemare, et al., 1983). Our results showed that angular velocity in each joint was widely changed by changing movement frequency (Figure 4), indicating that proportion of fast-twitch and slow-twitch fibers for force production was different over range of movement frequencies. Still, there were significant and reliable differences in the level of muscle cocontraction between Dancers and Nondancers over a wide range of movement frequencies (Figure 6). This suggests that the decreased level of muscle cocontraction associated with motor learning may also occur at lower movement frequencies, that is, regardless of recruited muscle fiber type.

ConclusionThis study revealed a relationship between degree of movement proficiency and level of muscle cocontraction in the lower limbs during a standing rhythmic whole-body movement, thus providing evidence that the neural mechanism underlying the decrease in the level of muscle cocontraction is common to human motor learn-ing. Moreover, we have demonstrated the level of muscle cocontraction to be an important parameter in whole-body sensorimotor synchronization.

Acknowledgments

This study was partly supported by a Grant-in-Aid for Scientific Research (B) (No.21300215) from Japan Society for the Promotion of Science (JSPS) awarded to K. Kudo, a Grant-in-Aid for Scientific Research (B) (No.19300216) from JSPS awarded to T. Ohtsuki, and a Grant-in-Aid for JSPS Fellows (No.23·9480) awarded to A. Miura.

ReferencesAnderson, K., & Behm, D.G. (2005). Trunk muscle activity increases with unstable squat

movements. Applied Physiology, Nutrition, and Metabolism, 30(1), 33–45. PubMedAsaka, T., Wang, Y., Fukushima, J., & Latash, M.L. (2008). Learning effects on muscle modes

and multi-mode postural synergies. Experimental Brain Research, 184(3), 323–338. PubMed doi:10.1007/s00221-007-1101-2

Bellemare, F., Woods, J.J., Johansson, R., & Bigland-Ritchie, B. (1983). Motor-unit dis-charge rates in maximal voluntary contractions of three human muscles. Journal of Neurophysiology, 50(6), 1380-1392. PubMed

Muscle Cocontraction in Dancers and Nondancers 31

Brouwer, B., & Ashby, P. (1990). Corticospinal projections to upper and lower limb spinal motoneurons in man. Electroencephalography and Clinical Neurophysiology, 76(6), 509–519. PubMed doi:10.1016/0013-4694(90)90002-2

Candotti, C.T., Loss, J.F., Bagatini, D., Soares, D.P., da Rocha, E.K., de Oliveira, A.R.,, & Guimarães, A.C. (2009). Cocontraction and economy of triathletes and cyclists at dif-ferent cadences during cycling motion. Journal of Electromyography and Kinesiology, 19(5), 915–921. PubMed doi:10.1016/j.jelekin.2008.04.008

Chapman, A.R., Vicenzino, B., Blanch, P., & Hodges, P.W. (2008a). Is running less skilled in triathletes than runners matched for running training history? Medicine and Science in Sports and Exercise, 40(3), 557. PubMed doi:10.1249/MSS.0b013e31815e727a

Chapman, A.R., Vicenzino, B., Blanch, P., & Hodges, P.W. (2008b). Patterns of leg muscle recruitment vary between novice and highly trained cyclists. Journal of Electromyog-raphy and Kinesiology, 18(3), 359–371. PubMed doi:10.1016/j.jelekin.2005.12.007

Crotts, D., Thompson, B., Nahom, M., Ryan, S., & Newton, R.A. (1996). Balance abilities of professional dancers on select balance tests. The Journal of Orthopaedic and Sports Physical Therapy, 23(1), 12-17. PubMed

Deeny, S.P., Haufler, A.J., Saffer, M., & Hatfield, B.D. (2009). Electroencephalographic coherence during visuomotor performance: A comparison of cortico-cortical commu-nication in experts and novices. Journal of Motor Behavior, 41(2), 106–116. PubMed doi:10.3200/JMBR.41.2.106-116

Deeny, S.P., Hillman, C.H., Janelle, C.M., & Hatfield, B.D. (2003). Cortico-cortical com-munication and superior performance in skilled marksmen: an EEG coherence analysis. Journal of Sport & Exercise Psychology, 25(2), 188–204.

De Serres, S.J., & Milner, T.E. (1991). Wrist muscle activation patterns and stiffness associ-ated with stable and unstable mechanical loads. Experimental Brain Research, 86(2), 451-458.

Ertan, H., Kentel, B., Tümer, S., & Korkusuz, F. (2003). Activation patterns in forearm muscles during archery shooting. Human Movement Science, 22(1), 37–45. PubMed doi:10.1016/S0167-9457(02)00176-8

Fransson, P., Gomez, S., Patel, M., & Johansson, L. (2007). Changes in multi-segmented body movements and EMG activity while standing on firm and foam support surfaces. European Journal of Applied Physiology, 101(1), 81–89. PubMed doi:10.1007/s00421-007-0476-x

Fujii, S., Kudo, K., Shinya, M., Ohtsuki, T., & Oda, S. (2009). Wrist muscle activity during rapid unimanual tapping with a drumstick in drummers and nondrummers. Motor Control, 13(3), 237-250. PubMed

Galea, V., De Bruin, H., Cavasin, R., & McComas, A.J. (1991). The numbers and relative sizes of motor units estimated by computer. Muscle & Nerve, 14(11), 1123–1130. PubMed doi:10.1002/mus.880141114

Gerbino, P.G., Griffin, E.D., & Zurakowski, D. (2007). Comparison of standing balance between female collegiate dancers and soccer players. Gait & Posture, 26(4), 501-507. PubMed doi:10.1016/j.gaitpost.2006.11.205

Girard, O., Micallef, J.P., & Millet, G.P. (2005). Lower-limb activity during the power serve in tennis: effects of performance level. Medicine and Science in Sports and Exercise, 37(6), 1021-1029. PubMed

Gobel, E.W., Parrish, T.B., & Reber, P.J. (2011). Neural correlates of skill acquisition: Decreased cortical activity during a serial interception sequence learning task. Neuro-Image 58(4), 1150-21157. PubMed

Golomer, E., Cremieux, J., Dupui, P., Isableu, B., & Ohlmann, T. (1999). Visual contribu-tion to self-induced body sway frequencies and visual perception of male professional dancers. Neuroscience Letters, 267(3), 189–192. PubMed doi:10.1016/S0304-3940(99)00356-0

32 Miura, et al.

Golomer, E., & Dupui, P. (2000). Spectral analysis of adult dancers’ sways: Sex and interac-tion vision-proprioception. International Journal of Neuroscience, 105(1-4), 15–26. PubMed doi:10.3109/00207450009003262

Golomer, E., Dupui, P., & Monod, H. (1997a). The effects of maturation on self-induced dynamic body sway frequencies of girls performing acrobatics or classical dance. European Journal of Applied Physiology and Occupational Physiology, 76(2), 140–144. PubMed doi:10.1007/s004210050226

Golomer, E., Dupui, P., & Monod, H. (1997b). Sex-linked differences in equilibrium reac-tions among adolescents performing complex sensorimotor tasks. Journal of Physiology, Paris, 91(2), 49–55. PubMed doi:10.1016/S0928-4257(97)88937-1

Golomer, E., Dupui, P., Séréni, P., & Monod, H. (1999). The contribution of vision in dynamic spontaneous sways of male classical dancers according to student or pro-fessional level. Journal of Physiology, Paris, 93(3), 233–237. PubMed doi:10.1016/S0928-4257(99)80156-9

Gribble, P.L., Mullin, L.I., Cothros, N., & Mattar, A. (2003). Role of cocontraction in arm movement accuracy. Journal of Neurophysiology, 89(5), 2396–2405. PubMed doi:10.1152/jn.01020.2002

Hair, J.F., Black, W.C., Babin, B.J., & Anderson, R.E. (2009). Multivariate Data Analysis (7th ed.). Upper Saddle River, NJ: Prentice Hall.

Heuer, H. (2007). Control of the dominant and nondominant hand: exploitation and taming of nonmuscular forces. Experimental Brain Research, 178(3), 363–373. PubMed doi:10.1007/s00221-006-0747-5

Hirashima, M., Kadota, H., Sakurai, S., Kudo, K., & Ohtsuki, T. (2002). Sequential muscle activity and its functional role in the upper extremity and trunk during overarm throwing. Journal of Sports Sciences, 20(4), 301–310. PubMed doi:10.1080/026404102753576071

Kawato, M. (1999). Internal models for motor control and trajectory planning. Current Opinion in Neurobiology, 9(6), 718–727. PubMed doi:10.1016/S0959-4388(99)00028-8

Kudo, K., & Ohtsuki, T. (1998). Functional modification of agonist-antagonist electromyo-graphic activity for rapid movement inhibition. Experimental Brain Research, 122(1), 23–30. PubMed doi:10.1007/s002210050487

Letchford, C.W., Sandri, P., Levitan, M.L., & Mehta, K.C. (1992). Frequency response requirements for fluctuating wind pressure measurements. Journal of Wind Engineering and Industrial Aerodynamics, 40(3), 263–276. doi:10.1016/0167-6105(92)90379-O

Milner, T.E. (2002). Adaptation to destabilizing dynamics by means of muscle cocontrac-tion. Experimental Brain Research, 143(4), 406–416. PubMed doi:10.1007/s00221-002-1001-4

Milner, T.E., Cloutier, C., Leger, A.B., & Franklin, D.W. (1995). Inability to activate muscles maximally during cocontraction and the effect on joint stiffness. Experimental Brain Research, 107(2), 293–305. PubMed doi:10.1007/BF00230049

Miura, A., Kudo, K., Ohtsuki, T., & Kanehisa, H. (2011). Coordination modes in sen-sorimotor synchronization of whole-body movement: A study of street dancers and non-dancers. Human Movement Science, 30(6), 1260–1271. PubMed doi:10.1016/j.humov.2010.08.006

Montgomery, D.C., Peck, E.A., & Vining, G.G. (2007). Introduction to linear regression analysis (4th ed.). Hoboken, NJ: Wiley.

Mouchnino, L., Aurenty, R., Massion, J., & Pedotti, A. (1992). Coordination between equi-librium and head-trunk orientation during leg movement: a new strategy build up by training. Journal of Neurophysiology, 67(6), 1587-1598. PubMed

Münte, T.F., Altenmüller, E., & Jäncke, L. (2002). The musician’s brain as a model of neu-roplasticity. Nature Reviews. Neuroscience, 3(6), 473–478. PubMed

Nakata, H., Yoshie, M., Miura, A., & Kudo, K. (2010). Characteristics of the athletes’ brain: Evidence from neurophysiology and neuroimaging, Brain Research Reviews 62(2), 197-211.

Muscle Cocontraction in Dancers and Nondancers 33

Osu, R., Franklin, D.W., Kato, H., Gomi, H., Domen, K., Yoshioka, T., & Kawato, M. (2002). Short- and long-term changes in joint cocontraction associated with motor learning as revealed from surface EMG. Journal of Neurophysiology, 88(2), 991–1004. PubMed

Reithler, J., van Mier, H.I., & Goebel, R. (2010). Continuous motor sequence learning: Cor-tical efficiency gains accompanied by striatal functional reorganization. NeuroImage, 52(1), 263–276. PubMed doi:10.1016/j.neuroimage.2010.03.073

Rossi-Durand, C., Jones, K.E., Adams, S., & Bawa, P. (1999). Comparison of the depres-sion of H-reflexes following previous activation in upper and lower limb muscles in human subjects. Experimental Brain Research, 126(1), 117–127. PubMed doi:10.1007/s002210050721

Sakurai, S., & Ohtsuki, T. (2000). Muscle activity and accuracy of performance of the smash stroke in badminton with reference to skill and practice. Journal of Sports Sciences, 18(11), 901–914. PubMed doi:10.1080/026404100750017832

Shichirui, S. (1999). Kokujinrizumukannohimitsu. Tokyo: Ikuhousha. [The secret of the sense of rhythm in African-American].

Stepp, N., & Turvey, M.T. (2010). On strong anticipation. Cognitive Systems Research, 11(2), 148–164. PubMed doi:10.1016/j.cogsys.2009.03.003

Takaishi, T., Yamamoto, T., Ono, T., Ito, T., & Moritani, T. (1998). Neuromuscular, meta-bolic, and kinetic adaptations for skilled pedaling performance in cyclists. Medicine and Science in Sports and Exercise, 30(3), 442-449. PubMed doi:10.1097/00005768-199803000-00016

Thoroughman, K.A., & Shadmehr, R. (1999). Electromyographic correlates of learning an internal model of reaching movements. The Journal of Neuroscience, 19(19), 8573-8588. PubMed