Article Text

other Versions

Effects of Olympic-style taekwondo kicks on an instrumented head-form and resultant injury measures
  1. Gabriel P Fife1,2,
  2. David M O'Sullivan3,4,
  3. Willy Pieter5,
  4. David P Cook6,
  5. Thomas W Kaminski7
  1. 1Department of Physical Education, Yonsei University, Seoul, Republic of Korea
  2. 2Department of Physical Education, Dong-A University, Busan, Republic of Korea
  3. 3Department of Physical Education, Chung-Ang University, Anseong, Republic of Korea
  4. 4Department of Physical Education, Seoul National University, Seoul, Republic of Korea
  5. 5Department of Taekwondo, Keimyung University, Daegu, Republic of Korea
  6. 6Department of Academy of Sport, London South Bank University, London, UK
  7. 7Department of Health, Nutrition and Exercise Sciences, University of Delaware, Newark, Delaware, USA
  1. Correspondence to Professor David M O'Sullivan, Department of Physical Education, College of Sport Science, Chung-Ang University, 72-1 Naeri Daedeokmyeon, Gyeonggido Anseong 456-756, Republic of Korea; tkd4{at}


Objective The objective of this study was to assess the effect of taekwondo kicks and peak foot velocity (FVEL) on resultant head linear acceleration (RLA), head injury criterion (HIC15) and head velocity (HVEL).

Methods Each subject (n=12) randomly performed five repetitions of the turning kick (TK), clench axe kick (CA), front leg axe kick, jump back kick (JB) and jump spinning hook kick (JH) at the average standing head height for competitors in their weight division. A Hybrid II Crash Test Dummy head was fitted with a protective taekwondo helmet and instrumented with a triaxial accelerometer and fixed to a height-adjustable frame. Resultant head linear acceleration, HVEL, FVEL data were captured and processed using Qualysis Track Manager.

Results The TK (130.11±51.67 g) produced a higher RLA than the CA (54.95±20.08 g, p<0.001, d=1.84) and a higher HIC15 than the JH (672.74±540.89 vs 300.19±144.35, p<0.001, ES=0.97). There was no difference in HVEL of the TK (4.73±1.67 m/s) and that of the JB (4.43±0.78 m/s; p=0.977, ES<0.01).

Conclusions The TK is of concern because it is the most common technique and cause of concussion in taekwondo. Future studies should aim to understand rotational accelerations of the head.

  • Concussion
  • Head injuries
  • Martial Arts

Statistics from


Sport-related concussions have been described as a ‘silent epidemic’ with 1.5–3.8 million cases in the USA annually.1 ,2 The Olympic sport of taekwondo with over 80 million participants worldwide,3 is not exempt from this dilemma with concussion incidence ranging from 5.5 to 50.2 per 1000 athlete-exposures,4 ,5 which may be attributed to differences in methodology (eg, reported to medical personnel versus video analysis followed by direct athlete examination). Furthermore, this rate is up to four times greater than in other collision sports, such as American football.6

The injurious results of forces to the head and the biomechanics of sport-related concussion are well documented.7–10 and have in part led to an international initiative to improve the care of athletes.11 Research in this area has focused on American football,9 ice hockey,7 soccer and other full-contact sports.2 These concussions have been associated with the velocity of impact and the mass of the impacting object.12 As such, with end-point velocities of the taekwondo turning kick (TK) ranging from 13 to 17 m/s4 ,13 ,14 coupled with the effective mass of the foot, it is not surprising that impact forces of greater than 200 g have been estimated.4 However, although such values would indicate risk of significant brain injury, there still remains a dearth of research specific to concussion mechanisms in taekwondo.

Consequently, the purpose of this study was to investigate the head impact characteristics resulting from the most common kicks reported to cause concussion in taekwondo.



Twelve male taekwondo athletes (22.5±3.5 years, 176.9±7.2 cm, 70.8±8.6 kg) were recruited for this study. The subjects competed at a minimum of A-class international level across all ranges of taekwondo weight categories. Preparticipation screening consisted of completing the Physical Activity Readiness Questionnaire (PAR-Q), along with a demographic and injury history questionnaire to ensure the athletes’ physical readiness to participate. In addition, an institutional review board-approved informed consent form was completed. Testing occurred at a biomechanics laboratory and subjects were orally briefed on the testing procedures prior to data collection.

Apparatus: anthropometric test dummy

To simulate head impact in taekwondo, the target consisted of a Hybrid II 50th Percentile Crash Test Dummy (Hybrid II) head (mass=5.1 kg) and neck (fitted with a taekwondo head guard) secured to an aluminum support frame with locations for peg bolts to be fastened at predetermined increments (figure 1), which allowed for adjustment of the Hybrid II head and neck to comply with average weight category standing heights of male Olympic taekwondo participants from the Athens 2004 Olympic Games.15 The Hybrid II head is an aluminum substructure covered by a deformable skin layer. The neck is made of a rubber cylinder that allows movement in all planes of motion (flexion/extension, right and left lateral flexion, left and right rotation) when the head is kicked.

Figure 1

Peg bolt height adjustment (right) and height adjustable Hybrid II (left).

Testing procedures

Subjects wore lightweight athletic clothing during testing and performed individualised warm-ups followed by a series of light kicking techniques for specific preparation. The athletes were then asked to begin a familiarisation period with the Hybrid II by performing the high TK, front leg jump axe kick, clench axe kick, jump back kick, and the jump spinning hook kick at the head to provide an opportunity to adjust their kicking speed and accuracy. These kicks were reported to be the most common causes of diagnosed concussion in taekwondo.5 Following the familiarisation and a brief rest period, subjects were asked to perform five repetitions of each kick aimed at the Hybrid II head. All participants were asked to kick as if they were competing, while wearing protective foot pads that are commonly used in competition. The athletes personally indicated that after the familiarisation period they felt as if they were ‘engaged in a typical training session like when kicking a bag’. The subjects also stated that although kicking the dummy was similar to a typical training session, it was ‘not the same as the person-to-person response experienced during real competition’. This feedback highlights the importance for future research groups to make efforts to collect live competition data, which might be absent from the constraints of a laboratory setting. The response by one athlete, who experienced no discomfort when kicking, may be best reflected in our personal concerns that they might actually damage the head-form due to such strong kicks. The years of impact conditioning employed by these competitors might explain their ability to execute kicks to the Hybrid II head with no discomfort.

Data acquisition

The Hybrid II head form was instrumented with a 500 g triaxial accelerometer (PCB Piezotronics 356A66, Depew, New York, USA) mounted to the head centre of gravity to obtain resultant linear accelerations (RLA). The accelerometer was firmly fixed inside the Hybrid II head on a 4.0 cm×4.0 cm aluminium plate secured to the head base by four-socket head cap screws. Furthermore, a plastic mounting base (manufacturer provided) that allows for the sensor to be mechanically grounded was glued to the aluminium mounting plate to ensure no movement of the accelerometer occurred during each trial (figure 2).

Figure 2

Triaxial accelerometer mounted to the base of head (right) on a 4.0 cm×4.0 cm aluminum plate secured to the head base by four socket head cap screws (left).

The accelerometer was interfaced via a three-channel, battery-powered integrated circuit piezoelectric sensor signal conditioner (PCB Piezotronics, Depew, New York, USA) and connected to a desktop computer for analysis. Acceleration data were captured using QTM (Qualisys Track Manager, Sweden) and processed in accordance with SAE J211-1 channel frequency class 1000.16

Three-dimensional kinematic data were collected using an eight-camera motion analysis system (OQUS 3-series, Qualisys, Sweden) sampling at 500 Hz to observe Hybrid II head as well as peak foot velocities.17 ,18 Data were collected from retro-reflective markers placed on the Hybrid II chin, head apex and base of the head to quantify peak head velocity (HVEL) at impact (figure 3) and the lateral malleolus and base of the fifth metatarsal of the kicking foot to quantify peak foot velocity (figure 4). The analogue and motion capture data were synchronised at 500 Hz by QTM.

Figure 3

Reflective apex, occiput (right) and chin markers (left). The Hybrid II Dummy neck is made of rubber and allows motion in all planes. Arrows indicate location of flexible rubberneck.

Figure 4

Reflective foot markers placed at the lateral malleolus and base of the fifth metatarsal to track peak foot velocity.

Data analysis

To assess concussion potential, the head injury criterion (HIC) was calculated over a 15-ms time window.16 ,19 Kinematic data were smoothed by a second-order Butterworth filter with a cut-off frequency of 20 Hz.13 Distributional characteristics were verified before statistical analysis and coefficients for skewness and kurtosis were calculated. In cases of skewness and/or kurtosis, the data were rank ordered. A one-way analysis of variance was used to determine differences between kicks, while the Tukey HSD test was employed for pairwise comparisons. The contribution of height, body mass, foot and head velocity to RLA by kick, as well as height, body mass and foot velocity as predictors of HVEL by kick was assessed by multiple regression analysis. The level of significance was set to 0.05.


There were significant differences in RLA (p=0.005, η2=0.054), HIC (p=0.002, η2=0.059), head velocity (p=0.002, η2=0.059) and foot velocity (p=0.004, η2=0.055) among kicks (see table 1). A graphical example of one HIC value produced by one athlete is presented in figure 5.

Table 1

Descriptive statistics (95% CI) of resultant linear acceleration (RLA), head injury criterion (HIC), head velocity (HVEL) and foot velocity (FVEL) at impact for each kick

Figure 5

A graphical representation of the time frame of which the head injury criterion (HIC) is determined from one trial of the turning kick. Because the HIC is determined from the impulse of the peak acceleration, differences in the impulse of the accelerations may explain differences in HIC magnitudes (672.74+40.89, 95% CI 329.1 to 1016.4 and 462.95+556.72 95% CI 109.23 to 816.67) when two kicks produce similar peak head velocities (4.73±1.67 m/s and 4.43±0.78 m/s, p=0.977, ES<0.01).

Tables 2 (RLA) and 3 (HIC) depict the probability and effect size matrices of the post hoc analyses. Height, weight, head and foot velocities predicted 71.43% (SEE=11.99) of the variance in RLA following the front axe kick (p=0.044, f2=2.50; table 2).

Table 2

Probability and effect size matrix of the post hoc analysis of resultant linear acceleration (g)

Table 3

Probability and effect size matrix of the post hoc analysis of head injury criterion


The findings from the present study indicate that the taekwondo TK has the greatest potential for concussive impacts owing to its foot and head kinematics as well as associated RLA and HIC values. Mean head acceleration (130 g) and HIC (672) exceed those reported for American football and boxing.19 ,20 Guskiewicz et al9 provided an insight into a large number of live sport impacts (total=104 714) with 13 of them concussive (average 102.8 g, range 60.51–168.71 g), which were recorded over five competitive seasons of American football. In taekwondo, the most common technique that was reported to cause concussion is the TK (130.11±51.67 g (95% CI 97.28 to 162.94 g), which falls within the range of impacts recorded by Guskiewicz et al.9 Research on Olympic boxing provided descriptions of energy transfer to the head from common punches.8 ,19 However, the current investigation did not include this variable, as reliable anthropometric measurements were not readily available. Future research may consider this variable as an explanation of injury potential relative to headgear safety testing.21

Pellman et al20 re-enacted concussive impacts using Hybrid III dummies, while Guskiewicz et al9 employed a newer technology of accelerometer-embedded helmets (ie, the Helmet Instrumented Technology System, Simbex LLC, Lebanon, New Hampshire, USA), which may more accurately reflect the magnitude of concussive impacts measured during live games. In Olympic boxers, peak translational head acceleration for the hook punch was 71.2 g.19 Not only was the TK head acceleration nearly double that observed from the boxing hook punch, but two other kicks (ie, spinning hook kick and jump back kick) also elicited greater RLAs. It must be understood that the complete picture of the biomechanics of sport-related head injury is multifaceted, that is, rotational acceleration, location of impact,22 etc need to be considered to understand concussion and severe head trauma.

The importance of this is illustrated when compared with an HIC of 1000 that was reported as life-threatening.23 In addition, it should be noted that all kicks in the present study can be classified, based on biomechanical measures, as potentially concussive. Some reports, however, indicate life-threatening potential24 and at times, fatal25 ,26 outcomes. Recently, a medical case study24 reported the axe kick to cause a massive brain haemorrhage, although it was found to be one of the lower-magnitude techniques examined in our study. Attention must be brought to the TK as it is arguably the most frequently used technique in taekwondo27 and the opportunities for such an impact to occur are increased (see the HIC ranges in table 1).

The TK is performed in a rapid proximal to distal sequential motion and serves to transfer momenta along the kinetic chain resulting in high end-point velocity at impact. However, the ideosyncracies of taekwondo lend themselves to various kicks being performed with similar early phase kinematics (akin to a baseball pitch) to prevent cue utilisation. As such, technique alteration may occur relatively late in the sequence, which may increase the potential of head trauma owing to the unpredictability of point of impact. The area of impact has been related to changes in HIC in American gridiron football and is often associated with the range of motion and stiffness of the neck structure.28


Although the current study provides insight into the possible head injury severity experienced by elite taekwondo athletes during full-contact competition, the ecological validity of the findings must be further investigated. The quasi-static nature of the Hybrid II head and neck complex enabled subjects to ‘load-up’ on each technique due to the fact that there were no time constraints of decision attributes as per true competition. In addition, the design of the Hybrid II structure does not permit realistic neuromuscular responses relating to the neck, particularly as there are no pretuning capabilities prior to impact, which may often be the case in real taekwondo bouts. Early studies on Hybrids II and III biofidelity reported nearly 3–5 times greater neck stiffness for the Hybrid III compared with the human neck, and poor response changes in velocity.17 ,29 ,30 However, the Hybrid II has also been observed to have minimised energy absorption capabilities as well as a less biofidelic head,31 which are attributed to a reduction in the approximation of head injury.32 An additional important consideration of the Hybrid II head/neck complex of inertia is needed to appropriately understand and apply the results of data from athletic head impact studies. Although the inertia of the head-form used in this study and others8 ,19 have not been reported, future research should consider measuring this variable to draw more comprehensive conclusions concerning sport head impacts. A more recently designed ATD (ie, THOR Advanced Crash Test Dummy) is currently used by the National Highway Traffic Safety Administration (USA) due to its significantly improved biofidelity compared to the earlier models designed in the 1960s and 1970s33; however, the current study used a more economical alternative. The issue of ATD biofidelity indicates the need for caution when applying the results from laboratory studies. A consideration of all limitations is important for a comprehensive understanding of the results and a move towards more ecologically valid research is recommended akin to the real-time impact data taken in American football.9 ,22

Caution is warranted, however, as far as the accuracy of the peak head velocities reported in this study is concerned, because our high-speed cameras were limited to a sampling rate of 500 Hz.34 Nevertheless, the novelty of these data may be helpful in providing a preliminary understanding of the injurious forces which cause head injury in this sport. Future research is encouraged to employ more customary instrumentation (eg, a more biofidelic ATD, higher sampling rate for head velocity data) and to investigate other biomechanical variables, such as rotational acceleration.


Previous biomechanical investigations on Olympic boxing and American gridiron football provide the groundwork for understanding injury mechanisms and threshold for sport-related concussion. Our results are intriguing because the high RLA and HIC produced by taekwondo athletes of all Olympic weight categories were greater than previously reported in boxing, a similar combative sport and some American gridiron football studies. These findings provide a biomechanical understanding of the nature of forces acting on the head during various kicks reported to cause concussion in taekwondo. Further investigations into other measures related to concussion and severe head injury in sport are also warranted.

With the high incidence of concussion reported in taekwondo, the results of the present study are an area for concern. This is heightened in response to the changes in scoring guidelines as specified by the WTF35 in October 2010, with four points allocated for a spinning kick to the head. Additionally, not only should medical personnel be aware of the severity of injuries possible in taekwondo but should also be up-to-date on recommendations and the standard of care brought forth by the International Conference on Concussion in Sport.11

What this study adds

  • The first known biomechanical study on head inury in taekwondo.

  • A broader picture of the concussion and head injury research field, especially regarding higher impact magnitudes not measured before in other combat sports.


This study was funded by a University of Delaware General University Research Grant and the National Athletic Trainers’ Association Paula Sammarone Turocy Postgraduate Grant. We would like to thank Humanetics ATD (Plymouth, Michigan, USA) for donating the Hybrid II device. A special thanks to Hyunju Lee for assistance with editing images in this paper.


View Abstract


  • Contributors GPF contributed to conception, design, collection and interpretation of data, drafting and revising the article critically for important intellectual content and final approval of the version to be published. DO contributed to conception, design, collection and interpretation of data, drafting and revising the article critically for important intellectual content and final approval of the version to be published. WP contributed to conception, design, analysis and interpretation of data, drafting and revising the article critically for important intellectual content and final approval of the version to be published. DPC contributed to conception, design, collection and interpretation of data, drafting and revising the article critically for important intellectual content and final approval of the version to be published. TWK contributed to conception, design and interpretation of data, drafting and revising the article critically for important intellectual content and final approval of the version to be published.

  • Funding National Athletic Trainers’ Association and University of Delaware GUR Grant.

  • Competing interests None.

  • Ethical approval University of Delaware Institutional Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.