Strong relationships exist between muscle volume, joint power and whole-body external mechanical power in adults and children
Abstract
The present study investigated whether differences between adults and children in mechanical power during single-joint knee extension tasks and the complex multijoint task of jumping could be explained by differences in the quadriceps femoris muscle volume. Peak power was calculated during squat jumps, from the integral of the vertical force measured by a force plate, and during concentric knee extensions at 30, 90, 120, 180 and 240 deg·s−1, and muscle volume was measured from magnetic resonance images for 10 men, 10 women, 10 prepubertal boys and 10 prepubertal girls. Peak power during jumping and isokinetic knee extension was significantly higher in men than in women, and in both adult groups compared with children (P < 0.01), although there were no differences between boys and girls. When power was normalized to muscle volume, the intergroup differences ceased to exist for both tasks. Peak power correlated significantly with quadriceps volume (P < 0.01), with r2 values of 0.8, 0.86, 0.81, 0.78 and 0.81 from isokinetic knee extension at angular velocities of 30, 90, 120, 180 and 240 deg·s−1, respectively, and with an r2 value of 0.9 from squat jumps. These results indicate that the quadriceps femoris muscle volume accounts largely for the increase in power that occurs with maturation in the two genders not only in kinematically constrained knee extensions but also in multijoint tasks. Future studies should examine the role of other factors relating to the generation and transmission of contractile power, such as muscle architecture, tendon stiffness and external mechanical leverage.
It is well documented that performance of many high-intensity, short-duration anaerobic tasks, such as maximal running speed and jumping, improves throughout childhood and adolescence (for reviews see Malina & Bouchard, 1991; Van Praagh, 1998; Van Praagh & Doré, 2002). This improvement with age requires a considerable increase in muscle power (Van Praagh & Doré, 2002; De Ste Croix et al. 2003), since not only are the individuals running faster or jumping higher but they also have to move a greater body mass. This period of life is associated with the growth of skeletal muscle and, in boys, the additional development associated with the increase in circulating testosterone (Round et al. 1999). The increase in strength during this time is most probably associated with increases in muscle fibre diameter and muscle cross-sectional area. Furthermore, Jones & Round (2000) point out that increases in muscle length during growth will also contribute to power, since the power of a muscle is determined by its volume, as the product of the mean cross-sectional area and length.
Measures of muscle thickness or cross-sectional area of the muscle group of interest have failed to account for the differences in power between adults and children (Ferretti et al. 1994; Deighan et al. 2003a; Kanehisa et al. 2006), for the obvious reason that they do not take account of differences in muscle length. However, estimates of muscle volume using a number of simple approaches (Kanehisa et al. 1994; Kanehisa et al. 2003) have also failed to explain the changes in isokinetic peak power that occur with maturation. In contrast, estimates of the lean volume of the whole thigh were found to be useful for scaling muscle power during isokinetic single-joint tasks in adults and children (Barrett & Harrison, 2002).
Jumping represents a complex multijoint movement, from which the resulting external mechanical power produced as the feet push off the ground is frequently used as an indicator of the power-producing capacity of the knee extensors (Thorpe et al. 1998; Finni et al. 2000; Van Praagh & Doré, 2002). In accordance with the findings of Barrett & Harrison (2002), Temfemo et al. (2009) found that the mean power produced during a maximal vertical jump correlated significantly with lean thigh volume. However, this correlation was associated with a weak r2 value of only 0.33 for boys and 0.11 for girls (both sexes aged between 11 and 16 years), suggesting that only a small proportion of jumping performance was explained by lean thigh volume.
The finding in many previous studies that muscle size fails to account for the differences in single-joint and whole-body mechanical power in adults compared with children indicates that maturation-induced changes in other parameters that affect contractile velocity and joint moment may be important. Such factors include muscle architecture and fibre type, tendon stiffness, level of agonist and antagonist activation, mechanical advantage of muscles, and co-ordination of movement patterns. Before investigating the role of these factors, however, it is imperative to establish whether the increases in muscle power with maturation are largely explained by muscle size, by accurately measuring muscle volume rather than obtaining a gross estimate of it. Following this approach, it was hypothesized that, owing to the contributing role of the hip and ankle extensors and the co-ordination of all contracting muscles in the multijoint task of jumping, the relationship between the quadriceps femoris muscle volume and power throughout the transition from prepubertal childhood to adulthood would be weaker for jumping than for single-joint knee extensions.
Methods
Participants
Forty participants, consisting of 10 men, 10 women, 10 boys and 10 girls, volunteered to participate in this study, and their ages, heights and body masses are presented in Table 1. The maturation state of the children was not directly assessed, but because the age range examined was 8–10 years, the children were considered to be prepubertal based on the observations of Tanner (1962). The adult participants recruited were sedentary in their daily lives and the children were not taking part in any organized sport or physical activity outside of school to negate any training effect on the measurements. The study complied with the Declaration of Helsinki and was approved by the ethics committee for the Institute for Biomedical Research into Human Movement and Health of Manchester Metropolitan University, and written informed consent was obtained from all participants and the parents or guardians of the children prior to testing.
Age (years) | Height (cm) | Mass (kg) | |
---|---|---|---|
Men | 28.2 ± 3.6 | 170.6 ± 8 | 88.1 ± 14 |
Women | 27.4 ± 4.2 | 167.3 ± 7.3 | 64.0 ± 9.4 |
Boys | 8.9 ± 0.7 | 138.3 ± 9.2 | 35.6 ± 9.5 |
Girls | 9.2 ± 0.8 | 139 ± 5.6 | 39.3 ± 7.9 |
- Data are reported as the means ±s.d.
Experimental measurements
Data were collected during two separate testing sessions. In the first session, muscle volume and squat jump performance were measured, and the participants were familiarized with the isokinetic testing procedures, as recommended particularly for the testing of children (De Ste Croix et al. 2003). The second testing session was completed between 2 days and 2 weeks later and was used to determine the power–velocity relationship of the knee extensor muscles of the dominant leg on the isokinetic dynamometer.
The volumes of the vastus lateralis, vastus intermedialis, vastus medialis and rectus femoris muscles of the participants’ dominant leg were measured using magnetic resonance imaging (MRI). No physical activity was performed for 2 days prior to MRI, and the participants lay at rest in the scanning position for ∼15 min before they were scanned to allow for fluid shifts to occur (Berg et al. 1993). Participants lay supine in a 0.2 T MRI scanner (E-scan, Esaote Biomedica, Genoa, Italy) with their knee fully extended. Images were obtained along the entire thigh length using a Turbo 3-D T1 sequence with the following parameters: slice thickness, 6.3 mm; interslice gap, 0 mm; acquisition time, 3 min 57 s; time to repetition (TR)/echo time (TE)/number of excitations (NEX), 40/16/1; field of view (FOV), 180 × 180 mm2; Matrix, 256 × 256 pixels. From these images, the cross-sectional area of individual muscles (see Fig. 1) was measured in axial-plane scans with 18.9 mm interval, using OsiriX dicom viewer (version 2.2.1, Osirix Foundation, Geneva, Switzerland). The volume of each muscle head was calculated as the sum of the cross-sectional area in all scans multiplied by the thickness between each measured slice (1.89 cm). The total quadriceps femoris muscle volume was then calculated as the sum of the volumes of all four individual muscle heads.
The peak external mechanical power produced during a squat jump was recorded to provide a measure of the power-producing capability of each group in a complex, whole-body task that requires the co-ordinated activation of several muscle groups, whilst remaining highly dependent upon the capacity of the quadriceps femoris muscle (Thorpe et al. 1998; Finni et al. 2000). The squat jumps were performed on a Leonardo jump platform (Novotec Medical, Pforzheim, Germany), and the jumping external mechanical power was calculated as the product of the ground reaction force and the velocity of movement of the body calculated from the force integral. Although this measurement approach is limited to the measured ground reaction force, it is a practical means of determining a close representation of the real external mechanical power, providing that the body remains rigid during the push-off phase to minimize transfer of kinetic energy between segments. To this effect, the jumps were performed with the hands on the hips throughout the push-off phase. Following several practice and warm-up attempts, the jump was performed from a squat position, with the knee and hip joints flexed to a depth of the individual participant's choosing. This approach was chosen over controlling the depth of squat, because pilot testing showed that doing so could negatively affect the jumping performance of children, an observation that is in line with previous reports of an altered optimal jumping technique with changes in musculo-skeletal properties (Hunter & Marshall, 2002). The participants were asked to jump as high as possible three times, and the jump producing the greatest peak power was considered for further analysis.
The muscle power–velocity relationships for the knee extensors were determined using an isokinetic dynamometer (Cybex NORM, New York, USA). Participants were seated and securely strapped on the dynamometer chair with the hip angle set at 85 deg (where 0 deg is full extension), and the centre of rotation of the dynamometer lever arm was aligned with the lateral femoral condyle during a submaximal contraction. Following a standardized warm-up, participants performed single maximal concentric extensions of their dominant knee from a stationary flexed start position; single interrupted cycles have previously been shown to elicit greater peak moments than continuous flexion–extension cycles in children, owing to the neural and co-ordination demands of reversing the action (Docherty & Gaul, 1991). Extensions were performed across the participants’ full range of motion at angular velocities of 30, 90, 120, 180 and 240 deg·s−1. Although angular velocities of up to only 180 deg·s−1 have previously been reported as reliable for the testing of children (Burnie & Brodie, 1986; Deighan et al. 2003b), pilot testing indicated that children could still reliably generate peak powers at 240 deg·s−1; a trial at this velocity was therefore included to provide the greatest range across which differences between adults and children could be assessed. Trials at each angular velocity were repeated twice and in a randomized order, with a rest period of ∼2 min between trials. To aid with motivation and effort, verbal encouragement was provided in all trials and participants were able to see a real-time graph of the moment they were producing, as recommended for the testing of children (Baltzopoulos & Kellis, 1998). The resulting joint moments from the isovelocity phase were gravity corrected and multiplied by the angular velocity (in radians) to calculate power. The peak power across the two trials at each angular velocity was identified and used for all further analyses.
Statistical analysis
The peak power produced during the isokinetic trials and jumps was normalized to the volume of the quadriceps femoris muscle (Powervol). Jumping external mechanical power and Powervol were analysed using a multivariate general linear model (GLM) between groups. Two separate repeated measures 5 × 4 (angular velocity × group) GLMs were used to test for differences in the isokinetic muscle power and Powervol with joint angle and between the groups. A Pearson's correlation was performed between muscle volume and peak power for each angular velocity and the squat jump for all subjects combined to quantify the strength of the relationship between muscle power and muscle volume for each set of testing conditions. Preliminary statistical analysis indicated that for the isokinetic trials at 90, 180 and 240 deg·s−1 the data were heteroscedastic (P < 0.05). For these tests, natural logarithmic transformation was completed before the correlation analysis was performed. Significance was accepted at P≤ 0.05. Data are reported as the means ±s.e.m., unless otherwise stated.
Results
The participants in each group were well matched, with relatively small variance of body mass, height and age (Table 1). Unsurprisingly, the adult men were taller and heavier than the women, who were in turn taller and heavier than the children. There were no notable differences between the boys and girls.
Despite being heavier, the adults jumped higher than the children. The men jumped significantly higher (∼40 cm) than the women (∼30 cm) who, in turn, jumped higher than the boys and girls (∼25 cm). The height jumped is clearly a function of the power generated by the muscles and the body mass they have to accelerate. Both jump height and power are shown in Table 2.
Jump height (cm) | Jump power (W) | Jump height/power/body mass (cm W−1 kg−1) | Powervol (W cm−3) | |
---|---|---|---|---|
Men | 38.2 ± 1.8* | 3451 ± 217* | 0.87 ± 0.03 | 1.7 ± 0.09 |
Women | 31.1 ± 2† | 2308 ± 152† | 0.86 ± 0.03 | 1.74 ± 0.08 |
Boys | 24.7 ± 0.9 | 1102 ± 81 | 0.79 ± 0.02 | 1.55 ± 0.03 |
Girls | 24.8 ± 0.7 | 1235 ± 79 | 0.79 ± 0.03 | 1.64 ± 0.04 |
- Data are reported as the means ±s.e.m.*P < 0.01, significantly different from women, boys and girls; and †P < 0.01, significantly different from men, boys and girls.
We next explored to what extent these differences in power could be explained by differences in muscle volume. Typical MR images are shown in Fig. 1, with the boundaries of each muscle marked. Total quadriceps femoris muscle volume was greater in men than in women, who in turn had a greater muscle volume than children, with no differences between boys and girls (Table 3). The proportionate contribution of each individual muscle head to the total quadriceps femoris muscle volume was approximately the same for all the groups (Table 3).
Total volume (cm3) | VL (%) | VM (%) | VI (%) | RF (%) | |
---|---|---|---|---|---|
Men | 2053 ± 453 | 34 ± 2 | 25 ± 1 | 27 ± 3 | 14 ± 1 |
Women | 1359 ± 268 | 33 ± 4 | 26 ± 3 | 28 ± 3 | 13 ± 2 |
Boys | 709 ± 117 | 34 ± 2 | 22 ± 1 | 28 ± 3 | 16 ± 1 |
Girls | 754 ± 140 | 34 ± 2 | 23 ± 2 | 29 ± 2 | 14 ± 2 |
- Data are reported as the means ±s.d.
The relationships between joint angular velocity and peak muscle power and Powervol are presented in Fig. 2. The GLM revealed significant differences in power between the men, women and children at all angular velocities, although there were no differences between the boys and girls (Fig. 2A). There were no differences between groups when the muscle power was normalized to muscle volume (P > 0.05; Fig. 2B). For all groups, power and Powervol increased significantly as a function of velocity (P≤ 0.01). There was also a significant interaction between power and velocity (P < 0.01; Fig. 2A), with men achieving much greater increases in peak power with increasing angular velocity than the other groups. However, once power was normalized to muscle volume there were no differences in the slope of the Powervol–velocity relationship between groups.
The correlation between muscle power and muscle volume for all the subjects combined produced high r2 values for all angular velocities (r2= 0.8, 0.86, 0.81, 0.78 and 0.81 for angular velocities of 30, 90, 120, 180 and 240 deg·s−1, respectively, P < 0.01; Fig. 3A–E).
Peak external mechanical power during the squat jumps differed between groups (P < 0.01; Table 2). The post hoc tests revealed that for power in absolute terms the differences lay between men, women and children, but there was no difference between boys and girls. The Pearson's correlation between peak external mechanical power and muscle volume for all the participants combined revealed a very high r2 value (r2= 0.9, P < 0.01; Fig. 3F) and, accordingly, no differences were seen between the groups in external mechanical Powervol (Table 2).
Discussion
The ability to jump improves with the transition from childhood to maturity and depends on a number of factors, including muscle power and the body mass which must be displaced. Based on the strong correlations found, we have shown that, like knee extension contractions, the differences in power-producing capability during jumping between boys, girls, men and women can be explained largely in terms of differences in the quadriceps femoris muscle volume.
It has been suggested that children may not be able to optimally co-ordinate their movement in multijoint actions (Van Praagh & Doré, 2002), leading to reduced external mechanical power in tasks such as jumping. Moreover, in contrast to knee extension, jumping is not only dependent on the quadriceps muscle but also on the hip and ankle joint extensors. Consequently, the differences in the relationship between the quadriceps femoris muscle volume and power between the two different tasks would be further amplified. In line with this concept are the findings of previous studies showing that measures of lean thigh volume accounted for the differences in the knee extension joint power–velocity relationship between adults and children (Barrett & Harrison, 2002), but not for the differences in external mechanical power while jumping (Temfemo et al. 2009). Accordingly, we hypothesized that muscle volume would not correlate with external whole-body mechanical power as strongly as with single-joint mechanical power. However, in contrast to the above reports and our hypothesis, we found that the correlation between the quadriceps femoris muscle volume and peak external mechanical power during squat jumps was in fact slightly stronger than between volume and muscle power during isokinetic knee extensions. One possible explanation for this surprising finding is that all the muscles contributing to jumping have proportionally similar volumes between groups, and therefore a high correlation can be obtained between power and the volume of the quadriceps femoris muscle alone. If this is the case, then the precise measurement of volume in a single representative muscle to describe the combined power of all contracting muscles is paramount. The errors resulting from inaccurate estimates of muscle volume, and the inclusion of non-relevant tissue (bone and hamstrings muscle), may explain why previous studies have failed to account for the changes in external mechanical power with maturation (Temfemo et al. 2009). One other possibility is that the relative kinematic freedom of the squat jumps compared to the isokinetic trials allows the contracting muscles to operate closer to their specific optimal shortening velocity in both adults and children. Since it is peak power at optimal velocity that would be best accounted for by volume, this might explain the high correlation between external jumping mechanical power and muscle volume. In addition, although jumping is a complex, multijoint movement, which requires high levels of co-ordination, it forms part of many recreational activities frequently performed by children and adults, which may result in them having developed efficient movement patterns. In contrast, the dynamometer task, albeit a simple joint action with which the participants have been familiarized before testing, represents an unusual movement; therefore, there may have been some decrement in activation or increase in antagonist coactivation. However, since the level of antagonist muscle activation has been shown to be similar in adults and children across angular velocities up to 180 deg·s−1 (Bassa et al. 2005), it seems unlikely that antagonist activation may play a role in altering the power–velocity relationship of adults or children.
Peak muscle power during the isokinetic testing was only measured from within the isovelocity phase of the knee extension. With the exception of one girl at 240 deg·s−1, all participants produced their peak power within the isovelocity phase for all angular velocities examined. For this one trial, the muscle power peaked prior to the isovelocity phase and decreased throughout the range of movement. For all other participants and trials, the peak powers produced during the isokinetic knee extensions were in line with those previously reported for adults and children over the same range of angular velocities (Burnie & Brodie, 1986; Seger & Thorstensson, 1994; Paasuke et al. 2001; Barrett & Harrison, 2002; Deighan et al. 2003a,b). Peak power values during the squat jumps are also comparable to those previously reported for similar groups (Ferretti et al. 1994). Data on knee extensor muscle volume in children could not be found to provide comparisons with the present data, but the adult muscle volume data in the present study are in line with those previously reported (Friederich & Brand, 1990; Tracy et al. 2003).
Children produced a similar muscle and external mechanical peak power regardless of sex, while men produced significantly higher peak powers than women. This finding supports previous reports of differences in the way power-producing capability develops in each sex during puberty, with males experiencing greater increases than females (Gilliam et al. 1979; Seger & Thorstensson, 1994, 2000; Deighan et al. 2003a). It is apparent from the present findings that these sex differences are a result of the greater muscle volume gains from prepubertal age to adulthood in males. This is clearly illustrated in Fig. 3, where it can be seen that the muscle volumes of boys and girls are rather similar, whilst the volumes and associated powers are notably greater in men than in women, with very little overlap between the two groups. The discrepancy in muscle size difference between adults and children with sex is unlikely to reflect any differences in physical activity histories between men and women. It is more likely to reflect inter-gender muscle size gain differences with maturation due to the increased testosterone levels in males following the onset of puberty, which is known to be associated with significant gains in muscle mass (Bhasin et al. 1996) and has been shown to coincide with the development of differences in isometric strength between the sexes (Parker et al. 1990; Round et al. 1999).
Differences were found in the slope of the muscle power–velocity relationships between groups, with men specifically showing greater increases in peak power with increasing angular velocity than the other groups. This finding is in accordance with some previous reports (Kanehisa et al. 1994, 2003; Ramos et al. 1998; Paasuke et al. 2001; Barrett & Harrison, 2002), but it contrasts others (Seger & Thorstensson, 1994, 2000). However, those studies (Seger & Thorstensson, 1994, 2000) only used velocities of up to 180 deg·s−1, which may not have been sufficiently high for the differences between the groups to become apparent. The greater peak muscle power produced at high contraction velocities by adults compared with children may be a result of the muscles of adults having longer muscle fibres (i.e. a greater number of sarcomeres in series), which would allow a greater maximal shortening velocity (Gans & de Vree, 1987). If this is the case, then theoretically, for the same external angular velocity, the sarcomere shortening velocity would be lower in an adult muscle with longer fibres compared with that in a child's muscle with shorter fibres, meaning that the fibres can generate a greater force. Differences in the rate of torque development, as a consequence of neural maturation (Paasuke et al. 2001), changes in fibre type (Lexell et al. 1992; Sjöström et al. 1992) and tendon stiffness (Vogel, 1980), may allow adults to achieve peak power earlier in a movement, and possibly closer to the knee angle that corresponds to optimal fibre length.
Maturation-induced changes in the parameters listed above will also affect the external mechanical power during jumping. Potential differences in the mechanical advantage of muscles (ratio of muscle moment arm and ground reaction force moment arm) between children and adults should also be considered during jumping. This is because the moment arm length of the ground reaction force depends not only on anatomical dimensions but also on the technique employed (Biewener, 2005), which may change with maturation. The contribution of these factors to the changes in mechanical power with growth should be investigated in future studies. However, it is important to consider that any correlation between these factors and muscle volume will also be reflected in their correlation with power.