KEY WORDS: Isometric, isokinetic, dynamometry, martial art, reaction.

Participants
With local ethical approval and informed consent 13 practitioners of hard-style martial arts (9 male: 4 female, mean (± SEM) age 23.7 (± 3.1) years and 12 control participants (8 male: 4 female, mean age 22.2 (± 0.6) years were recruited for this study. The martial arts group was composed of 8 who practiced Tae Kwon Do, 3 Shaolin Nam Pai Chauan Kung-Fu and 2 Wu Shu Kwan Kung-Fu. Eleven of the martial artists had black belts and 2 were senior brown belts in training for their black belt. All were in regular training. Participants' height and weight were measured; the dynamometer normalised the torque measurements to the participants' body weight.

Study protocol

Strength testing
Parameters of trunk and lower limb flexion and extension strength were recorded under isometric and isokinetic conditions on a Cybex Norm Isokinetic Testing System (Henley Healthcare, USA).

Leg strength
Each participant was seated comfortably and the knee isolated as much as possible using a strap across the thigh of the leg being tested and a four-strap seat belt to prevent use of thorax and abdominal muscles. The equipment was arranged so that the dynamometer was aligned with centre of rotation of the knee joint being tested. The shin pad was strapped as distally on to the tibia making sure dorsiflexion of the ankle was not restricted. The protocol was repeated for the opposite leg and the order in which the legs were tested was randomised. Participants performed isokinetic concentric knee flexion and extension at speeds of 30, 90 and 210°·s^-1 with a trial and five repetitions at each speed with the maximal value being used in the subsequent analysis. This was followed by a maximal isometric contraction in flexion followed by extension at 45° of flexion. Participants were encouraged to be relaxed and a level of verbal encouragement constant for all participants was given throughout testing to prevent "psyching-up", known to affect muscle power (Tod et al., 2003).

Trunk strength
The lower limbs were stabilised by tibial and thigh pads. A belt secured the pelvis to limit the involvement of the hip flexor muscles during testing. Range of motion was recorded from -10° of hyperextension to 80° flexion as recorded through the Cybex system, which represented the limits of range of the system rather than the ranges of the individuals.

Participants performed isokinetic concentric trunk flexion and extension at speeds of 30 and 90°·s^-1 with five repetitions at each speed. This was followed by a maximal isometric contraction at +10° flexion, and then at -10° extension.

Measurement of reaction times
All participants performed trials to assess simple reaction time (SRT), choice reaction time (CRT) and their speed of movement once they had reacted in both the SRT (simple movement time; SMT) and CRT (choice movement time; CMT) tasks. Movement times were measured between release of an initiation button and depression of a stop button 25 cm away. Both hands were used in random order to perform the reaction time test by releasing a button in response to a tone. An in-house computer program running on an IBM compatible PC was used to measure reaction and movement times (for full protocol see Davey et al., 2001).

Statistical analyses
Body weight adjusted peak torques under the various conditions were investigated for differences between the left and right legs and also between the martial artists and the controls using the Student's t-test. Reaction and movement times were compared between the dominant hand and the non-dominant hand within the groups and also between the groups using the Student's t-test. Results were considered statistically significant when p < 0.05.

Strength
Leg strength
There were no differences (for either flexion or extension) between the dominant and non-dominant legs for either group under isometric testing or at any of the speeds under isokinetic testing. The data were therefore pooled for both legs. The martial artists had a higher mean body weight adjusted peak isometric torque than the controls only in extension (Figure 1a). Under isokinetic testing the martial artists had a higher torque than the controls in flexion and extension at all speeds (Figure 1b).

The hamstrings/quadriceps torque ratios were not significantly different between the martial artists the controls at any of the isokinetic speeds (mean [± SEM] ratio over all speeds; martial artists 72.67 ± 2.19%, controls 73.33 ± 1.86%).

Trunk strength
There were no significant differences between the martial artists and the controls under isometric testing or at any of the speeds under isokinetic testing for either flexion or extension. However, martial artists did consistently have slightly higher (not significantly) torque for all testing conditions (Figure 2a and Figure 2b).

The flexor/extensor torque ratios were not significantly different between the martial artists the controls at either of the isokinetic speeds (mean [± SEM] ratio over all speeds; martial artists 89.93 ± 6.61%, controls 95.84 ± 3.76%). Concurrent with previous results on isokinetic testing (Chan and Maffulli, 1996; Lord et al., 1992) it can be observed that as the speed of movement increases the peak torque decreases (Figure 1b and Figure 2b).

Reaction times
There were no differences in RT or MT in the simple or choice tests between the dominant and non-dominant hands for either group. The data were therefore pooled for both hands. The RTs for the simple and choice tests were not different between the martial artists and the controls (simple RT for controls = 222.18 ± 6.61ms, martial artists = 210.91 ± 3.75 ms; choice RT for controls = 343.10 ± 15.69ms, martial artists = 312.33 ± 8.77ms). However, the MT was significantly faster in the martial artists for both tests and resulted in a faster total RT for the martial artists (simple MT for controls = 165.79 ± 4.58ms, martial artists = 128.07 ± 3.90ms; choice MT for controls = 197.82 ± 5.23ms, martial artists = 174.29 ± 8.54ms; see Figure 3).

The martial artists in this study showed greater torque in isokinetic and isometric testing of the legs and a slightly (but not significantly) higher trunk torque. Our results support and extend previous studies on quadriceps strength in practitioners of various forms of martial arts (Heller et al., 1998; Jacobson et al., 1997; Lan et al., 1998; 2000). No previous study, however has investigated the effects of training in the hard style martial arts on knee flexor strength. The significantly higher strength of training in the hard style martial arts on knee flexor strength. The significantly higher strength of the knee flexors reported here suggests that these are also developed in martial arts training. Training in the softer style Tai Chi martial art has been shown to increase the strength of the knee extensors and flexors (Lan et al., 1998; 2000), but Tai Chi places emphasis on different qualities in its practitioners to those in the harder styles studied here. Indeed, the increases in strength found by Lan et al., 1998 after training were less than the differences found between the martial artists and the controls in this study. This could be due to the age differences between the participants in their study and ours or could indicate that training for the harder style martial arts has a more substantial effect on the strength of the knee flexors and extensors.

The results of isometric testing of the legs support a previous study (Douris et al., 2004) showing that martial artists practising a mixed soft and hard style were stronger in extension. Since the form of martial arts practised by the participants in the present study require the participants to adopt low stances with their legs flexed for long periods of time (using predominantly quadriceps muscles), this may account for the increased strength.

The hamstring/quadriceps ratios is a commonly used indicator of the balance of agonist and antagonist muscles (Aagaard et al., 1995; Seto et al., 1988; St Clair et al., 2000) and may be a predictor of injury. Although the martial artists here had higher torques this ratio was no different from the control group. Therefore, although participation in the martial arts is associated with higher risk of injury (Zetaruk et al., 2005) muscle imbalance may not be a key factor.

The strength of the trunk muscles is enhanced by physical training (Andersson et al., 1988; Chan et al., 1996; Peltonen et al., 1998; Williams and Singh, 1997) and in martial arts these muscles are required to stabilise the body during kicking and punching, and maximise the effectiveness of force transfer onto the target. However, although the trunk flexors and extensors did tend to have larger torque values under both isokinetic and isometric testing, these differences did not reach significance. On further analysis of the isokinetic peak torque, there was a greater difference between the two groups in trunk extension than in flexion. Trunk extension movements are necessary to avoid kicks or punches to the face level, and certain kicking techniques require extension of the trunk.

There was no significant difference in the trunk flexion to extension ratio. However, the martial artists did tend to have lower ratios at both speeds, most likely due to stronger trunk extensors. While the relevance of trunk muscle strength with respect to low back pain has been widely studied, results are equivocal. Back extensor strength has been reported to be low in low back pain sufferers (Iwai et al., 2004; Kankaanpaa et al., 1998) and exercises that strengthen the extensors have been shown to be beneficial for low back pain (Carpenter and Nelson, 1999). However, previous research has not found significant correlations between trunk strength and back pain (Balague et al., 1993; Shirado and Kaneda, 1992). None of the martial artists in the present study reported any back pain.

It has been reported that physically fit participants show faster simple reaction times than their less fit counter parts (Arito and Oguri, 1990; Brisswalter et al., 1997; Hascelik et al., 1989; Malathi and Parulkar, 1989) and sportspeople have superior sports-specific perceptual skills compared to novices (Kioumourtzoglou et al., 1998; Mori et al., 2002). In the present study, we were able to divide the more commonly measured total reaction time into the reaction time (RT) and the movement time (MT) components. Results indicated that movement time scores largely explained faster reaction times for martial artists. Therefore, we conclude that difference in the total reaction time is solely due to the martial artists being able to move their limb faster than the controls and not due to any superiority in simple or choice reaction or pre-motor time itself. This is in-keeping with previous studies showing highly skilled professional cricketers have been shown to take as long as novice players to pick up the ball flight information from a film of bowlers (McLeod, 1987) and martial arts training improves the speed of movements (Lee et al., 1999).

This study has revealed (rather surprisingly) that the trunk strength is not significantly increased in practitioners of hard-style marital arts. Given that this physically demanding sport is associated large and random movements of the trunk, it would be wise for training regimes to increase focus on trunk stability exercises to increase strength of the abdominal and back muscles. This may lead to an increase in performance and possibly lessen the likelihood of injury or incidence of low back pain; this however, would require further research.

This study reveals that practitioners of hard-style martial arts produce higher torques in their knee flexors and extensors under isokinetic testing and only in their extensors under isometric testing. Furthermore, martial artists can produce slightly (but not significantly) higher torques in the trunk muscles. In addition, they have faster reaction times; however, this effect is largely explained by a faster movement time.

We would like to thank all participants who took part in this study. In addition we would like to Imperial College London for providing the funds for this undergraduate research project.

• Martial artists undertaking hard-style martial arts have greater strength in their knee flexor and extensor muscles as tested under isokinetic testing. Under isometric testing conditions they have stronger knee extensors only.
• The trunk musculature is generally higher under both conditions of testing in the martial artists, although not significantly.
• The total reaction times of the martial artists to an auditory stimulus were significantly faster than the control participants. When analysed further it was revealed that the decrease in reaction time was due to the movement time component of the total reaction time.
• The training involved for the practice of the hard-style martial arts increases the strength of muscles involved in kicking. This increased strength is not seen in the trunk muscles. Furthermore, martial artists have a faster response time; the cause of which appears to be only the faster movement time.