Neuroscience Letters 459 (2009) 6973

Contents lists available at ScienceDirect

Neuroscience Letters
journal homepage: www.elsevier.com/locate/neulet

Tapping performance and underlying wrist muscle activity of non-drummers,

drummers, and the worlds fastest drummer
Shinya Fujii a,b,c, , Kazutoshi Kudo b , Tatsuyuki Ohtsuki b , Shingo Oda a
a
b
c

Laboratory of Human Motor Control, Graduate School of Human and Environmental Studies, Kyoto University, Yoshida-nihonmatsu-cho, Sakyo-ku, Kyoto 606-8501, Japan
Department of Life Sciences, Graduate School of Arts and Sciences, University of Tokyo, Komaba 3-8-1, Japan
Japan Society for the Promotion of Science, Chiyoda-ku, Tokyo 102-8472, Japan

a r t i c l e

i n f o

Article history:
Received 23 March 2009
Received in revised form 25 April 2009
Accepted 27 April 2009
Keywords:
Tapping
Rate limit
Tremor
Electromyography
Practice
Drummer

a b s t r a c t
Studies of rapid unimanual tapping have assumed that the human rate limit for voluntary rhythmic
movement is 57 Hz, which corresponds to an inter-tap interval (ITI) of 150200 ms. In fact, the winner
of a recent contest to nd the worlds fastest drummer (WFD) can perform such movements using a
handheld drumstick at 10 Hz, which corresponds to an ITI of 100 ms. Because the contest measured only
the number of taps by the WFD, we examined the stability of the ITI and the underlying wrist muscle
activity of the WFD. By comparing the performance and wrist muscle activity of the WFD with those
of two control groups (non-drummers (NDs) and ordinary skilled drummers (ODs)), we found that the
WFD had a relatively stable ITI and more pronounced reciprocal wrist muscle activity during the 10-Hz
performance. Our result indicates that very fast, stable tapping performance can be achieved by keeping
the wrist joint compliant rather than stiff.
2009 Elsevier Ireland Ltd. All rights reserved.

To assess manual motor function, studies of human motor control have used a simple motor task called the rapid tapping task,
in which participants are asked to tap as fast as possible for several dozen seconds using nger [24], wrist [3,11], elbow, and/or
shoulder movement [13,29]. These studies have found that the
rate limit at which the motor effectors of the nger and wrist
can move voluntarily is 57 Hz, which corresponds to an intertap interval (ITI) of 150200 ms (see also the review [26]). Several
studies have investigated the effect of practice on the rate limit of
rapid tapping by training participants for several weeks or months
[14,15,25] or comparing the performances of musical instrument
players (i.e., pianists, string players, and drummers) with those of
non-players [2,8,9,12]. These studies reported that the rate limit
improved slightly following practice, but that the tapping frequency
was still within the 57 Hz limit.
In fact, the winner of a recent international contest to nd
the worlds fastest drummer (WFD) can perform such movements
using a handheld drumstick at 10 Hz, which corresponds to an ITI
of 100 ms. Specically, the WFD performed a record 1247 taps in
60 s with alternate left- and right-handed drumming movements,
which approximately equals 10 Hz movement with each hand over

Corresponding author at: Laboratory of Human Motor Control, Graduate School

of Human and Environmental Studies, Kyoto University, Yoshida-nihonmatsu-cho,
Sakyo-ku, Kyoto 606-8501, Japan. Tel.: +81 75 753 6876; fax: +81 75 753 6876.
E-mail address: shinya-fujii@h01.media.kyoto-u.ac.jp (S. Fujii).
0304-3940/$ see front matter 2009 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.neulet.2009.04.055

the 60 s period [21]. Because the contest measured only the number
of taps within a certain length of time, the stability of the ITI and
underlying muscle activity of the WFD remain unknown.
Previous studies of human motor learning showed that motor
performance becomes stable during the acquisition of a skill,
whereas the level of muscle co-contraction gradually decreases
[10,23]. In our recent study examining the wrist muscle activity
of the preferred right hand during rapid unimanual drumming
[8], we also showed that drummers had a lower level of muscle
co-contraction (i.e., clearer reciprocal contraction of antagonistic
muscle groups) and a more stable ITI compared to non-drummers.
Based on these ndings, we hypothesized that the WFD would show
pronounced reciprocal activity of the wrist muscles and have a stable ITI.
Another possibility exists, however, concerning the muscle
activity of the WFD. In studies of physiological tremor (PT) (i.e., very
small involuntary oscillating movements, such as those observed
while maintaining a posture), it has been argued that the human
nervous system can produce rapid oscillatory movements of a
body part, including 10 Hz movement [6,19]. PT is believed to
originate from multiple factors, including the oscillatory activity of the central nervous system, motor unit ring properties,
mechanical resonances, and reex loop resonance [6,19]. Although
the amplitude of the PT is usually too small (e.g., 0.097 mm in
the index nger of young adults [5]) to use for rapid drumming
movement, several studies have shown that it can be increased
(but becomes more variable) by increasing limb stiffness through

70

S. Fujii et al. / Neuroscience Letters 459 (2009) 6973

voluntary co-contraction of antagonistic muscle groups [4,20]. If the

WFD employs a PT through increased limb stiffness, we hypothesized that the WFD would show increased muscle co-contraction
and a variable ITI.
This study tested whether the WFD shows (i) pronounced reciprocal muscle activity and a stable ITI or (ii) increased muscle
co-contraction and a variable ITI. To do so, we examined the tapping
rate, variability of the ITI, and wrist muscle activity of the WFD and
compared them with those of two control groups: non-drummers
(NDs) and ordinary skilled drummers (ODs).
The WFD is a 44-year-old man who began drum training at 3
years of age and has 41 years of drumming experience. His handedness score, assessed using the Edinburgh Inventory [22], was 90 (i.e.,
strongly right-handed). He is a professional drummer who plays all
styles of music and won an international contest to nd the worlds
fastest drummer in 2005. All the NDs (n = 23, 17 men and 6 women)
and ODs (n = 44, 37 men and 7 women) were right-handed. The
handedness scores assessed using the Edinburgh Inventory were
89.6 18.7 (mean standard deviation) and 89.1 16.7 for the NDs
and ODs, respectively. The mean ages of the NDs and ODs were
25.3 7.2 and 30.1 10.5 years old, respectively. Participants in the
ND group had no drum training experience, whereas the OD group
included professional drummers and students from a special music
school. The ODs began drum training at 16.7 4.9 years of age and
had 12.8 11.4 years of experience. Typically, the ODs played rock,
popular, or jazz music.
In accordance with the Declaration of Helsinki, the participants were given a clear explanation of the experimental procedure
and provided written informed consent before participating in the
study. The experimental procedure was approved by the Ethics
Committee of the Graduate School of Arts and Sciences of the University of Tokyo.
The drum throne was adjusted to provide a comfortable height
and position for each participant. The participants sat on it while
holding a drumstick in one hand with the palm down. The participants were asked to use the drumstick to hit a drum practice pad
(6 in. in diameter, VIC-PAD6S, Pearl), under which a strain gauge
force transducer (LUR-A-1KNSAI, Kyowa) was mounted, using wrist
exion/extension movements. The participants were asked to tap
the center of the drum pad to ensure recording of the force data. The
participants were instructed to tap with the drumstick, using each
hand separately (i.e., unimanual approach) as fast as possible for
12 s. The left- and right-hand tasks were repeated three times each.
The order of the left- and right-hand tasks was randomized. Before
the experimental measurement, the participants were allowed to
practice until they felt accustomed to performing the task.
The electromyogram (EMG) was recorded for 8 NDs (5 men and
3 women, 28.5 10.7 years old) and 10 ODs (9 men and 1 woman,
29.3 9.4 years old, who began drum training at 15.2 4.6 years
and had 14.1 10.8 years of drum experience) from among the control groups. As in our previous study [8], surface EMG electrodes
(AgAgCl, 8 mm diameter) were attached to the exor carpi ulnaris
and extensor carpi radialis. We focused on wrist muscle activity
because the drummers reported that they mainly use these muscles
for high frequency movements in their routine drum playing. We
also conrmed that during the experiment, all of the participants
mainly used the wrist joint to perform the drumming task.
The signals from the force transducer and EMG were amplied
(SA-100D; TEAC), converted from analog to digital at a frequency
of 1000 Hz, and recorded on a computer. The data from 12 s of
drumming were truncated to eliminate start-up effects and the nal
slowing period by omitting data collected during the rst 1.5 s and
last 0.5 s, thereby analyzing only the data collected during the middle 10 s. The signal of the force transducer (rst rows in Fig. 1A)
enabled the detection of tap time [9]. The interval between two
sequential taps was dened as the ITI. We calculated the mean tap-

ping rate and within-trial standard deviation (S.D.) of the ITI over
10 s.
Unfortunately, we could not obtain right-hand EMG data for
the WFD due to difculties with the apparatus during the limited time available. Therefore, only the left-hand EMG data for
the WFD were analyzed and compared to those of the NDs and
ODs. As in our previous study [8], the left-hand EMG time series
were full-wave rectied and smoothed using a moving average
of 9 ms per window width (see the second and third rows in
Fig. 1A). To compare the wrist exor and extensor muscle activities, cross-correlations between the moving-averaged EMG signals
of the exor and extensor muscles were calculated [17]. The signal
from the wrist extensors was shifted with respect to the wrist exors within the range of 0 ms to the duration of the mean ITI for each
participant. We calculated the highest value of the cross-correlation
coefcient and the time lag at the moment for each participant.
The higher the value of the cross-correlation coefcient was, the
more pronounced was the reciprocal activity of the wrist exor and
extensor muscles.
The values of each dependent variable were averaged over the
trials for each participant. For the mean tapping rate and the S.D.
of the ITI, the asymmetry score was calculated using the following
equation:
asymmetry score =

RL
100,
R+L

(1)

where R and L denote the performances of the right and left

hands, respectively. The mean tapping rates of the NDs and ODs
were subjected to two-way repeated-measures analysis of variance
(ANOVA) because the data met the prerequisites for conducting
parametric analyses (the KolmogorovSmirnov test for a normal
distribution and the Levene test for homogeneity of variances).
The performing hand (left or right) was the within-participant
variable, and drumming experience (non-drummer or drummer)
was the between-groups variable. When a signicant interaction
was observed, between-group differences were analyzed using
the unpaired t-test, and within-group differences were analyzed
using paired t-tests. Regarding the S.D. of the ITI for the NDs
and ODs, the between-group differences were analyzed using the
MannWhitney U-test, and within-group differences were analyzed using the Wilcoxon matched-pairs signed-ranks test. The
asymmetry scores of the mean tapping rate and S.D. of the ITI for the
NDs and ODs were subjected to the MannWhitney U-test. The peak
cross-correlation coefcient and the time lag for the NDs and ODs
were subjected to unpaired t-tests. Statistical tests were deemed
signicant at an level of 0.05. When conducting multiple comparisons, the level was adjusted using the Bonferroni procedure
to compensate for the increased probability of nding signicant
results.
The mean tapping rates for the left and right hands of the NDs
were 6.2 0.8 and 7.0 0.9 Hz, respectively, and those of the ODs
were 6.4 0.6 and 6.8 0.6 Hz, respectively (see Fig. 2A). The twoway ANOVA for the mean tapping rate of the NDs and ODs indicated
a signicant interaction (F1,65 = 5.09, p < 0.05). No signicant NDOD
difference was observed for either the left (t65 = 1.31, ns) or right
(t65 = 0.76, ns) hands. The mean tapping rate of the left hand was
lower than that of the right hand for both the NDs (t22 = 4.26,
p < 0.01) and ODs (t43 = 4.98, p < 0.01). However, the asymmetry
score of the mean tapping rate for the ODs (2.86 3.86) was signicantly smaller than that for the NDs (5.67 6.25; z = 2.51, p < 0.05),
indicating that the ODs showed less tapping speed asymmetry compared to the NDs. The WFD had a mean tapping rate of 10.2 and
10.0 Hz for the left and right hand, respectively (see Fig. 2A). The
mean tapping rate for the left and right hands of the WFD was the
highest among the participants. The asymmetry score of the WFD
was 1.15 (i.e., an almost symmetrical tapping speed).

S. Fujii et al. / Neuroscience Letters 459 (2009) 6973

71

Fig. 1. Typical examples of the tap force and EMG signals in a non-drummer (left), an ordinary drummer (middle), and the worlds fastest drummer (right). (A) Typical signals
for 500 ms. The vertical dashed lines indicate the tap time detected from the signal of the tap force (rst rows). The second and third rows of graphs show the rectied and
smoothed EMG signals of the wrist exors and extensors, respectively. (B) Ensemble average of the smoothed EMG signals with reference to the tap time. Gray areas indicate
the within-participant standard deviation of muscle activity.

The S.D. of the ITI for the ODs was much smaller than that for
the NDs for both the left (z = 5.94, p < 0.001) and right (z = 5.27,
p < 0.001) hands (see Fig. 2B). The S.D. of the ITI of the right hand
was signicantly smaller than that of the left hand for both the NDs
(z = 3.07, p < 0.01) and the ODs (z = 3.89, p < 0.001); however, the
ODs had signicantly lower asymmetry scores than did the NDs
(13.16 19.04 vs. 25.13 24.79, z = 2.42, p < 0.05). The left- and
right-hand S.D. of the ITI for the WFD was 9.5 and 6.6 ms, respectively. The value of the WFD was similar to that of the ODs. The
asymmetry score of the WFD was 17.70. When we calculated the
coefcient of variation (CV) of the ITI, the statistical results were
consistent with the S.D. of the ITI. The left- and right-hand CV of
the ITI for the NDs was 24.0 12.0 and 18.0 12.7%, respectively,
those for the ODs were 7.0 5.7 and 5.1 3.2%, respectively, and
those for the WFD were 9.6 and 6.6%, respectively.
Fig. 1B shows the typical ensemble average of the moving average signals in reference to the tapping point in a ND, an OD, and the
WFD. The ensemble averages demonstrate that the WFD had relatively less muscle co-contraction with smaller within-participant
variability compared to the ND and OD. The time lag between
the wrist exors and extensors of the NDs, ODs, and WFD were
67.6 13.7, 71.5 9.1, and 39.0 ms, respectively (see Fig. 3A). No signicant difference was found in the time lag between the NDs and
ODs (t16 = 0.72, ns). The time lag of the WFD was the shortest
observed among the participants. The peak value of the crosscorrelation coefcient of the NDs, ODs, and the WFD was 0.47 0.10,
0.51 0.16, and 0.70, respectively (see Fig. 3B). No signicant difference was found in the cross-correlation coefcient between the

NDs and ODs (t16 = 0.62, ns). The cross-correlation coefcient of

the WFD was the highest among the participants, showing that the
reciprocal muscle activity was most pronounced in the WFD.
We investigated whether the WFD showed (i) pronounced reciprocal muscle activity and a stable ITI or (ii) increased muscle
co-contraction and a variable ITI. Our results demonstrated that the
WFD showed pronounced reciprocal activity of the wrist muscles
and a stable ITI, supporting our rst hypothesis. The results indicate
that the WFD kept his wrist joint compliant, rather than stiff, while
drumming at 10 Hz.
We postulate two functional reasons why the WFD showed
pronounced reciprocal muscle activation rather than increased
muscle co-contraction. First, a lower level of muscle co-contraction
or higher coherence between antagonistic muscle groups would
be physiologically more efcient and prevent muscle fatigue
[16,23,27]. Second, using the PT through increased muscle cocontraction inevitably results in more variable movements [20].
This would not be suitable for drum playing, as producing a regular
rhythm by exactly timed, repetitive tapping movements is arguably
an important element in musical performance.
Nevertheless, it is still possible that the WFD utilizes the neural
substrate of the PT, given that Carignan et al. [4] recently showed
that the PT amplitude can be modulated voluntarily without using
muscle co-contraction. They showed that the PT amplitude of the
postural index nger was successfully modulated not by using the
co-contraction of nger muscles, but through strategies such as
concentration, relaxation, visualization, and control of breathing.
From these observations, they indicated that cortical inuence is

72

S. Fujii et al. / Neuroscience Letters 459 (2009) 6973

Fig. 2. Tapping performance with the left (white bars) and right (black bars) hand by
non-drummers (NDs), ordinary drummers (ODs), and the worlds fastest drummer
(WFD). The error bars indicate the standard deviation between participants (n = 23
for the NDs and n = 44 for the ODs). (A) Mean tapping rate and (B) within-participant
standard deviation (S.D.) of the inter-tap interval (ITI).

being exerted on the PT without muscle co-contraction. In addition, recent brain imaging studies have revealed changes in the
activity of the cortical motor areas associated with tapping practice [14,15] and with asymmetry of tapping skill [18]. In view of
these studies, it might be possible that the WFD achieved 10-Hz
rhythmic movement by utilizing the neural substrate of the PT
through cortical modulation that was not accompanied by wrist
muscle co-contraction.
Is the achievement of 10-Hz performance and the pronounced
reciprocal muscle activity of the WFD related to the early age at
which he began drumming (i.e., 3 years old) or to his 41 years of
drumming experience? To address this question, we examined the
correlation of the performance and myoelectric variables with the
age of initiation of drumming and years of drum training experience among the drummers (i.e., ODs and WFD). For the years of
drum training experience, Spearmans correlation coefcient found
no signicant correlation with any variable (|r| = 0.28, ns), indicating that the 41 years of drumming practice was not directly related
to the extraordinary performance of the WFD. Because the WFD
stated that he had specically practiced to improve his tapping rate
to win the contest and to achieve impressive drumming performance, we suggest that not the mere number of years of experience,
but the amount of deliberate practice (i.e., concentrated practice
especially designed to improve specic aspects of an individuals
performance) [7] was related to his 10-Hz performance.
Regarding the age of commencement of drum training, signicant correlations were found between age of onset and the
left-hand mean tapping rate (r = 0.31, p < 0.05) and the S.D. of the
ITI (r = 0.34, p < 0.05). That is, the earlier participants started drum
training, the more rapid and stable the performance of the nonpreferred left hand was. Therefore, the early initiation of drum
training by the WFD might be related to his non-preferred lefthand performance. This supports previous neuroanatomical studies
suggesting that the early hand skill training of musicians interacts
with the cortical organization of hand motor dominance [1,28]. In
the future, an investigation of the brain structure or function of the
WFD will provide further insight into the neural substrates that
allow voluntary control of 10-Hz rhythmic movements.
Acknowledgments
We thank Mr. H. Tsunoda and the Wild Music School (Tokyo,
Japan) for their support in running this experiment. This study was
supported by a grant for the fellows of the Japan Society for the
Promotion of Science (JSPS) awarded to S. Fujii, a Grant-in-Aid for
Scientic Research (Nos. 17500416 and 21300215) from the JSPS
awarded to K. Kudo, and a Grant-in-Aid for Scientic Research (B)
(No. 19300216) awarded to T. Ohtsuki.
References

Fig. 3. Myoelectric variables of non-drummers (NDs, white bars), ordinary drummers (ODs, gray bars), and the worlds fastest drummer (WFD, black bars). The error
bars indicate the standard deviation (S.D.) (n = 8 for the NDs, n = 10 for the ODs).
(A) Time lag between the wrist exors and extensors at the peak cross-correlation
coefcient and (B) peak value of the cross-correlation coefcient.

[1] K. Amunts, G. Schlaug, L. Jancke, H. Steinmetz, A. Schleicher, A. Dabringhaus, K.

Zilles, Motor cortex and hand motor skills: structural compliance in the human
brain, Hum. Brain Mapp. 5 (1997) 206215.
[2] T. Aoki, S. Furuya, H. Kinoshita, Finger-tapping ability in male and female
pianists and nonmusician controls, Motor Control 9 (2005) 2339.
[3] T. Aoki, H. Kinoshita, Temporal and force characteristics of fast double-nger,
single-nger and hand tapping, Ergonomics 44 (2001) 13681383.
[4] B. Carignan, J.F. Daneault, C. Duval, The amplitude of physiological tremor can
be voluntarily modulated, Exp. Brain Res. (2009).
[5] C. Duval, J. Jones, Assessment of the amplitude of oscillations associated with
high-frequency components of physiological tremor: impact of loading and
signal differentiation, Exp. Brain Res. 163 (2005) 261266.
[6] R. Elble, W.C. Koller, Tremor, John Hopkins University Press, Baltimore, 1990.
[7] K.A. Ericsson, A.C. Lehmann, Expert and exceptional performance: evidence of
maximal adaptation to task constraints, Annu. Rev. Psychol. 47 (1996) 273305.
[8] S. Fujii, K. Kudo, T. Ohtsuki, S. Oda, Wrist muscle activity during rapid unimanual
tapping with a drumstick in drummers and non-drummers, Motor Control,
133, 2009 (July issue).

S. Fujii et al. / Neuroscience Letters 459 (2009) 6973

[9] S. Fujii, S. Oda, Tapping speed asymmetry in drummers for single-hand tapping
with a stick, Percept. Mot. Skills 103 (2006) 265272.
[10] P.L. Gribble, L.I. Mullin, N. Cothros, A. Mattar, Role of cocontraction in arm
movement accuracy, J. Neurophysiol. 89 (2003) 23962405.
[11] J. Hermsdorfer, C. Marquardt, S. Wack, N. Mai, Comparative analysis of diadochokinetic movements, J. Electromyogr. Kinesiol. 9 (1999) 283295.
[12] L. Jancke, G. Schlaug, H. Steinmetz, Hand skill asymmetry in professional musicians, Brain Cogn. 34 (1997) 424432.
[13] D. Kimura, W. Davidson, Right arm superiority for tapping with distal and
proximal joints, J. Hum. Mov. Stud. 1 (1975) 199202.
[14] S. Koeneke, K. Lutz, M. Esslen, L. Jancke, How nger tapping practice enhances
efciency of motor control, Neuroreport 17 (2006) 15651569.
[15] S. Koeneke, K. Lutz, U. Herwig, U. Ziemann, L. Jancke, Extensive training of
elementary nger tapping movements changes the pattern of motor cortex
excitability, Exp. Brain Res. 174 (2006) 199209.
[16] B.S. Lay, W.A. Sparrow, K.M. Hughes, N.J. ODwyer, Practice effects on coordination and control, metabolic energy expenditure, and muscle activation, Hum.
Mov. Sci. 21 (2002) 807830.
[17] L. Li, G.E. Caldwell, Coefcient of cross correlation and the time domain correspondence, J. Electromyogr. Kinesiol. 9 (1999) 385389.
[18] K. Lutz, S. Koeneke, T. Wustenberg, L. Jancke, Asymmetry of cortical activation
during maximum and convenient tapping speed, Neurosci. Lett. 373 (2005)
6166.
[19] J.H. McAuley, C.D. Marsden, Physiological and pathological tremors and rhythmic central motor control, Brain 123 (Pt 8) (2000) 15451567.

73

[20] S. Morrison, K.M. Newell, Limb stiffness and postural tremor in the arm, Motor
Control 4 (2000) 293315.
[21] Ofcial Site of WFD Worlds Fastest Drummer Extreme Sport Organization:
Speed Drumming Contest World Finals, McAfee Enterprises Inc., Retrieved April
25, 2009, from http://www.extremesportdrumming.com/.
[22] R.C. Oldeld, The assessment and analysis of handedness: the Edinburgh inventory, Neuropsychologia 9 (1971) 97113.
[23] R. Osu, D.W. Franklin, H. Kato, H. Gomi, K. Domen, T. Yoshioka, M.
Kawato, Short- and long-term changes in joint co-contraction associated with
motor learning as revealed from surface EMG, J. Neurophysiol. 88 (2002)
9911004.
[24] M. Peters, Why the preferred hand taps more quickly than the nonpreferred hand: three experiments on handedness, Can. J. Psychol. 34 (1980)
6271.
[25] M. Peters, Handedness: effect of prolonged practice on between hand performance differences, Neuropsychologia 19 (1981) 587590.
[26] B.H. Repp, Sensorimotor synchronization: a review of the tapping literature,
Psychon. Bull. Rev. 12 (2005) 969992.
[27] S. Sakurai, T. Ohtsuki, Muscle activity and accuracy of performance of the smash
stroke in badminton with reference to skill and practice, J. Sports Sci. 18 (2000)
901914.
[28] G. Schlaug, The brain of musicians. A model for functional and structural adaptation, Ann. N.Y. Acad. Sci. 930 (2001) 281299.
[29] J.I. Todor, P.M. Kyprie, H.L. Price, Lateral asymmetries in arm, wrist and nger
movements, Cortex 18 (1982) 515523.