In vivo measurements of muscle specific tension in adults and children


  • Thomas D. O’Brien,

    1. Institute for Biomedical Research into Human Movement and Health (IRM), Manchester Metropolitan University, John Dalton Tower, Chester Street, Manchester M1 5GD, UK
    Search for more papers by this author
  • Neil D. Reeves,

    1. Institute for Biomedical Research into Human Movement and Health (IRM), Manchester Metropolitan University, John Dalton Tower, Chester Street, Manchester M1 5GD, UK
    Search for more papers by this author
  • Vasilios Baltzopoulos,

    1. Institute for Biomedical Research into Human Movement and Health (IRM), Manchester Metropolitan University, John Dalton Tower, Chester Street, Manchester M1 5GD, UK
    Search for more papers by this author
  • David A. Jones,

    1. Institute for Biomedical Research into Human Movement and Health (IRM), Manchester Metropolitan University, John Dalton Tower, Chester Street, Manchester M1 5GD, UK
    2. School of Sport and Exercise Sciences, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK
    Search for more papers by this author
  • Constantinos N. Maganaris

    1. Institute for Biomedical Research into Human Movement and Health (IRM), Manchester Metropolitan University, John Dalton Tower, Chester Street, Manchester M1 5GD, UK
    Search for more papers by this author

Corresponding author T. D. O’Brien: Department of Sport, Health and Exercise Science, University of Hull, Cottingham Road, Kingston upon Hull HU6 7RX, UK.  Email:


To better understand the effects of pubertal maturation on the contractile properties of skeletal muscle in vivo, the present study investigated whether there are any differences in the specific tension of the quadriceps muscle in 20 adults and 20 prepubertal children of both sexes. Specific tension was calculated as the ratio between the quadriceps tendon force and the sum of the physiological cross-sectional area (PCSA) multiplied by the cosine of the angle of pennation of each head within the quadriceps muscle. The maximal quadriceps tendon force was calculated from the knee extension maximal voluntary contraction (MVC) by accounting for EMG-based estimates of antagonist co-activation, incomplete quadriceps activation using the interpolation twitch technique and magnetic resonance imaging (MRI)-based measurements of the patellar tendon moment arm. The PCSA was calculated as the muscle volume, measured from MRI scans, divided by optimal fascicle length, measured from ultrasound images during MVC at the estimated angle of peak quadriceps muscle force. It was found that the quadriceps tendon force and PCSA of men (11.4 kN, 214 cm2) were significantly greater than those of the women (8.7 kN, 152 cm2; P < 0.01). Both adult groups had greater values than the children (P < 0.01) but there were no differences between boys (5.2 kN, 99 cm2) and girls (6.1 kN, 102 cm2). Agonist activation was greater in men and women than in girls (P < 0.05), and antagonist activation was greater in men than in boys (P < 0.05). Moment arm length was greater in men than in boys or girls and greater in women than in boys (P < 0.05). The angle of pennation did not differ between the groups in any of the quadriceps heads. The specific tension was similar (P > 0.05) between groups: men, 55 ± 11 N cm−2; women, 57.3 ± 13 N cm−2; boys, 54 ± 14 N cm−2; and girls, 59.8 ± 15 N cm−2. These findings indicate that the increased muscle strength with maturation is not due to an increase in the specific tension of muscle; instead, it can be attributed to increases in muscle size, moment arm length and voluntary activation level.

It is self evident that, on average, adults are stronger than children (McComas et al. 1973; Parker et al. 1990; Kanehisa et al. 1995a,b; Round et al. 1999). The increase in strength has been suggested to be greater than the increase in muscle size (Kanehisa et al. 1995a,b); hence, it could be assumed that the force per unit area of muscle, known also as muscle specific tension, increases with puberty. This notion draws support from some reports of an increased proportion of type II fibres in adults (Lexell et al. 1992; Sjöstrom et al. 1992). Since type II fibres are reported to be intrinsically stronger than type I fibres in some studies (Kanda & Hashizume, 1992; Stienen et al. 1996), it is possible that adults may have a greater specific tension than children. The suggestion of an increasing specific tension with maturation also arises from studies that have quantified muscle size from anthropometric measurements or the anatomical cross-sectional area from a single site (Kanehisa et al. 1995a,b). These approaches assume that muscle size is proportional to anatomical dimensions and that measures from single areas in the muscle are representative of the whole muscle. Furthermore, they fail to account for any changes in pennation of the muscle fibres that will affect the apparent specific tension of the muscle, and therefore provide a poor estimate of the force-producing potential.

The appropriate measure of muscle size to describe the contractile force-producing potential is the physiological cross-sectional area (PCSA), which is the sum of cross-sectional areas of all the muscle fibres and can be calculated as the ratio of muscle volume to muscle fascicle length. To date, the only study that has calculated the specific tension of in vivo childrens’ muscle using the PCSA found that the specific tension of the lateral gastrocnemius muscle was 21% higher in boys than men (Morse et al. 2008), the opposite of what might be expected. However, as acknowledged and discussed by these authors, the study still made several assumptions in the calculation of muscle force. First, it was assumed that the pattern of the moment arm–joint angle relationship is the same in children and adults, although this assumption is now partly supported by recent evidence in the knee extensors (O’Brien et al. 2009a). Second, the proportional contribution of the lateral gastrocnemius PCSA to total plantarflexor PCSA, and the joint angle at which optimal fascicle length occurs was assumed to be the same in both groups. Additionally, it was assumed that the voluntary activation was 100% in adults and children and that antagonist muscles were co-activated to the same level in the two groups.

In situ studies of mouse muscle in which muscle force and muscle size were quantified directly, without having to make the assumption necessary for in vivo human muscle, reported specific tension to increase in the first weeks following birth (Gokhin et al. 2008), but to be similar in young and adult animals (Brooks & Faulkner, 1988). However, the methodological limitations associated with the estimation of muscle size and force in the existing studies on children do not allow a firm conclusion as to whether specific tension of human muscle changes with maturation and growth. This is an important issue because accurate estimations of specific tension are essential for realistic modelling of several musculoskeletal functions, e.g. simulation of force–length and moment–angle relationships (Hoy et al. 1990; Van den Bogert et al. 1998; Maganaris, 2004). At present, the specific tension values used in adult musculoskeletal models are also used in child models, and the lack of valid comparable specific tension data between the two populations prevents the development of appropriate musculoskeletal models for comparison. In the absence of such information, the purpose of the present study was to address the question of whether there are differences in the specific tension of adults and children by quantifying muscle force and size and circumventing most of the methodological limitations of previous in vivo studies.



Forty participants, consisting of 10 men, 10 women, 10 boys and 10 girls, volunteered to participate in this study (ages, heights and weights are presented in Table 1). The physical maturity of the children, who were all aged 8–10 years, was not directly assessed, but at this age both boys and girls will all have been prepubertal (Tanner, 1962). The magnetic resonance imaging (MRI) scans of the knee used to measure moment arm length (described below) showed that the growth plate of the femoral condyle was completely unfused in all the children and completely fused in the adults. The adult participants recruited were all sedentary in their daily lives, thus negating any effect that training status might have on the measurements. The study complied with the Declaration of Helsinki and was approved by the ethics committee of the Institute for Biomedical Research into Human Movement and Health of Manchester Metropolitan University. Written informed consent was obtained from all participants and the parents or guardians of the children prior to testing.

Table 1.  Participant characteristics
GroupAge (years)Height (m)Mass (kg)
Mean s.d. Mean s.d. Mean s.d.
Men28.23.61.8 0.0878.814

Experimental testing schedule

Data were collected during two separate testing sessions. In the first session, the moment arm of the patellar tendon (PTMA) and the volume of the quadriceps muscle were measured and, as recommended for the testing of children, the participants were familiarized with the dynamometer testing procedures (De Ste Croix et al. 2003). The second testing session was completed between 2 days and 2 weeks later and was used to identify the knee angle of peak quadriceps tendon force (Fquad) in the dominant leg and to make architectural measurements of the muscle fascicles.

Moment arm

The PTMA was measured from sagittal-plane magnetic resonance images (MRI) taken at each knee angle with a 0.2 T MRI scanner (E−scan, Esaote Biomedica, Genoa, Italy). Participants lay at rest while images of the knee joint of the dominant leg were taken using a spin-echo T1 HF sequence with a slice thickness of 6 mm, interslice gap of 0.6 mm and the parameters time to repetition (TR)/echo time (TE)/number of excitations (NEX), 420/18/1; field of view (FOV), 160 × 160 mm2; and Matrix, 256 × 256.

The PTMA was quantified with respect to the tibio-femoral contact point (TFCP) using a MatLab-based script that has been described in detail previously (Tsaopoulos et al. 2006). Briefly, the TFCP position was calculated as the mid-point on a sagittal plane line between the medial and lateral condyle contact points with the tibial plateau. The PTMA was then measured as the distance from the TFCP perpendicular to the line of action of the patellar tendon.

Muscle volume

The volumes of the vastus lateralis (VL), vastus intermedialis (VI), vastus medialis (VM) and rectus femoris (RF) muscles of the participants’ dominant leg were measured from axial-plane MRIs. Participants lay supine in the MRI scanner with their knee fully extended. Images were taken 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, 3:57 min; TR/TE/NEX, 40/16/1; FOV, 180 × 180 mm2; and Matrix, 256 × 256. From these images, the cross-sectional areas of individual muscles were measured in scans at 18.9 mm intervals, using OsiriX dicom viewer (v.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 distance between each measured slice (1.89 cm).

Measurement of joint moment

Participants were seated and securely strapped on the seat of an isokinetic dynamometer (Cybex NORM, New York, NY, USA) with the centre of rotation of the dynamometer's lever aligned with the lateral femoral condyle of the participant's dominant leg during a submaximal contraction. The back of the chair was adjusted to set the hip angle at 85 deg and to allow the knee to hang freely just over the edge of the chair. A standardized warm-up, consisting of several submaximal and two maximal efforts across a range of joint angles, was completed before the test trials were performed.

The test trials consisted of two isometric maximal voluntary knee extensions (MVCs) at every 5 deg of knee flexion from 55 to 90 deg (where 0 deg is full extension), performed in a randomized order and with ∼2 min rest between trails. To aid with motivation and effort, verbal encouragement was provided for all trials and participants were able to see a real-time graph of the moment produced on a computer screen in front of them (Baltzopoulos & Kellis, 1998).

During MVC, compression of the soft tissue and of the padding on the seat can result in the knee joint angle changing from that originally set by the dynamometer system (Arampatzis et al. 2004). Consequently, the knee angle during contraction was measured with an electronic goniometer (Biometrics Ltd, Newport, UK) placed on the lateral side across the knee joint. The proximal and distal ends of the goniometer were taped securely, equidistantly either side of the tibio-femoral cleft, to the thigh in line with the femur and to the shank in line with the tibia. Zero degree was set with the participant at full knee extension.

Antagonist co-activation

The level of antagonistic co-activation during the knee extension MVC was estimated from the moment–electromyogram (EMG) relationship of the knee flexor muscles. At each joint angle, participants also performed isometric knee flexions at 10, 20, 50, 70 and 100% of MVC, during which the surface EMG from the biceps femoris muscle was recorded through two Ag–AgCl percutaneous electrodes 10 mm in diameter (Ambus A/s, Ballerup, Denmark), placed 20 mm apart at the distal third of the muscle belly and smoothed using the root mean square over 50 ms. The EMG of the biceps femoris during the knee extensions was then compared with the knee flexion moment–EMG relationship to determine the corresponding joint moment that this muscle would produce as an antagonist (Kellis & Baltzopoulos, 1997; Reeves et al. 2003, 2004).

Voluntary activation level

The level of voluntary activation was quantified from superimposed and resting doublets using the interpolated twitch technique (ITT; eqn (1); Rutherford et al. 1986; Behm et al. 1996; Newman et al. 2003; Bampouras et al. 2006). The ITT is susceptible to errors resulting from changes in series elastic stiffness between joint angles, which makes it an unsuitable means of quantifying activation level across a range of joint angles (Bampouras et al. 2006). However, recent experiments in adults and children indicate that the level of knee extensor muscle activation does not differ across knee joint angles and that the ITT outcome at 90 deg knee flexion is a realistic estimate of voluntary activation for the quadriceps muscle across all knee joint angles (O’Brien et al. 2009b). Therefore, resting and superimposed doublets were evoked at 90 deg of knee flexion by percutaneus muscle magnetic stimulation (pulse duration, 1 ms; rise and fall durations, 1 μs; interstimulus gap, 10 ms; and a magnetic field of 2.0 T, produced by a 70 mm diameter ‘double coil’; MagStim, Bi-Stim2, Whitland, UK). Magnetic stimulation has previously been shown to have a high test–retest reliability and to yield estimates of voluntary activation level comparable to electrical stimulation (O’Brien et al. 2008). The superimposed doublet was applied once the MVC had reached a steady value, with the resting doublet being applied approximately 4 s later. The percentage activation was calculated as:

display math(1)

Maximal quadriceps muscle force

The MVC moment produced by the quadriceps (QuadMVC) was estimated as the sum of the measured joint moment (JointMVC) and the calculated antagonist moment.

To correct for any deficit in voluntary activation and thus to estimate the moment that would be produced by the quadriceps if they were fully activated (Quad100%) the QuadMVC moment and ITT were used in eqn (2), assuming a linear relationship between moment and voluntary activation level:

display math(2)

The calculated Quad100% was then used to calculate peak patellar tendon force (Fpt) during a fully activated effort by dividing Quad100% by the PTMA at the knee angle measured by the goniometer during that trial.

Finally, the Fquad was calculated from the Fpt using the ratios provided by Buff et al. (1988), which account for the fact that the patella is not a frictionless pulley for transmitting force from the quadriceps to the tibia.

Muscle architecture

Specific tension reflects the intrinsic force-generating potential of a muscle at optimal contractile length. In the present study, optimal length was considered to occur at the knee angle of peak Fquad. During the experimental protocol, the angle of peak Fquad was identified to make the measurements of muscle architecture within the same testing session. This approach assumed that agonist and antagonist activation were consistent across joint angles, as has been shown in adults and children (O’Brien et al. 2009b).

At the joint angle that corresponded to the greatest estimated Fquad, additional MVCs were performed during which ultrasound recordings (21 Hz, 104 mm probe length, MyLab 70, Esaote Biomedica, Genoa, Italy) were made of the VL, VI, VM and RF muscles. The recordings were taken from the mid-belly of each muscle, which, in the case of VL, VM and RF, was at 39, 22 and 56% of the distance from the proximal edge of the patella to the anterior superior iliac spine, respectively. The VI was considered to be separated into two portions of equal size and independent architectural properties: the lateral VI and the anterior VI (Blazevich et al. 2006). The mid-bellies were considered to be at 39% for the lateral VI and 56% for the anterior VI. These relative distances were chosen because they have previously been reported to equate to the mid-belly of each muscle in adults (Blazevich et al. 2006), and pilot testing confirmed this was also the case in children.

The ultrasound frame for each muscle that corresponded to the maximum Fquad was identified and used for analysis of muscle architecture. The fascicle length, taking into account any curvature, was measured from the point of insertion on the superficial and deep aponeuroses. The long ultrasound probe allowed the imaging of entire fascicles, meaning that no estimation or extrapolation was required for measures of fascicle length. The angle of pennation was measured at the deep fascicular insertion into the aponeurosis. All architectural measurements were made using ImageJ (version 1.38x, NIH, USA) and repeated using three different fascicles in each muscle, the mean of the three being used for subsequent analysis. For the VI muscle, the mean of the fascicle architecture from the anterior and lateral portion was used to describe the architecture of the whole muscle.

Calculation of specific tension

Specific tension was calculated as the ratio of fascicle force to the PCSA of the muscle, as follows:

display math(3)

where θ is the angle of pennation and therefore tendon force × cos θ−1 is equal to the fascicle force, and PCSA is equal to the ratio between muscle volume and optimal fascicle length.

By incorporating the size and architecture measurements of all four quadriceps heads in the calculation of a specific tension for the entire quadriceps muscle, the errors associated with distributing the tendon force between the separate heads was minimized. The PCSA × cos θ of the entire quadriceps was therefore calculated as the sum of the PCSA × cos θ of each quadriceps head. The Fquad was then divided by the quadriceps PCSA × cos θ to determine the specific tension of the entire quadriceps muscle.

Effects of joint angle and voluntary activation

There are two main potential sources of error in the measured fascicle architecture. First, the knee angle at which muscle architecture was measured may not correspond to optimal sarcomere length. Second, the ultrasound recordings were made of a muscle that was voluntarily activated, not fully activated. This was because the large size of the magnetic stimulation coil prevented the placement of the ultrasound probe over the necessary region of the muscles. Since it is known that both joint angle and contraction intensity alter the muscle architecture and the resulting estimate of PCSA, these errors may compromise the validity of the calculated specific tension. To quantify the extent of this problem, further trials were performed by two of the men, one woman, two boys and two girls. During these additional trials, muscle architecture was recorded during MVC at a knee angle 10 deg more extended than where Fquad occurred and also during contraction at only 90% of MVC. These additional architecture measurements were then used to calculate specific tension and compared to the best estimates of specific tension.

Statistical analysis

An analysis of variance (ANOVA) with Bonferroni corrected post hoc tests was performed to identify any differences between the groups in the following parameters: JointMVC, antagonist moment, voluntary activation level, Quad100%, PTMA length, Fpt, Fquad, quadriceps volume, quadriceps PCSA and specific tension. The volume, fascicle length, angle of pennation and PCSA of each head were analysed through additional ANOVAs to identify any differences in the architecture of each head between all of the groups. Significance was accepted at P≤ 0.05.

To estimate the extent of the errors resulting from the limitations of fascicle architecture measurements, the standard error was calculated between estimates of specific tension calculated with the correct muscle architecture and architecture from an altered knee angle and from a contraction at 90% of MVC.

All data are reported as means ±s.d., unless otherwise stated.


The peak Fquad occurred at a dynamometer set knee angle of ∼75 deg for men, women and girls, and ∼70 deg for boys; the standard deviations were between 6 and 9 deg for all groups. This equated to goniometer measured knee angles of 61.2 ± 4.8, 68.1 ± 8.9, 57.7 ± 9.4 and 62.1 ± 9.4 deg for men, women, boys and girls, respectively.

The ANOVAs examining JointMVC, antagonist moment, voluntary activation level, Quad100%, moment arm, Fpt and Fquad revealed that there were significant differences between the groups for all measurements (P= 0.014 for antagonist activation; for all others P < 0.01). The means ±s.d. of each measurement for each group are presented in Table 2, along with the results of the post hoc tests, which showed that adults were stronger than children, measured as a JointMVC and as the Fquad.

Table 2.  The mean ±s.d. values of MVC joint moment, antagonist co-activation moment, level of voluntary activation and moment arm at the angle of peak quadriceps force
 Joint MVC (N m)Antagonist moment (N m)Antagonist activation level (%)Quad100% (N m)Moment arm (cm) F pt (N) F quad (N)
  1. Significant differences between men (M), women (W), boys (B) and girls (G), *P < 0.01, †P < 0.05. Abbreviations: Quad100%, fully activated quadriceps moment; Fpt, patellar tendon force; Fquad, quadriceps tendon force.

Men271.79 ± 79.616.13 ± 9.586.66 ± 9.3328.85 ± 81.74.09 ± 0.48037 ± 163611439 ± 1938 
Women177.45 ± 60.113.68 ± 7.686.61 ± 6.6214.74 ± 63.63.73 ± 0.35757 ± 16008695 ± 2586
Boys 77.58 ± 17.17.08 ± 4  75.12 ± 12.8110.67 ± 33.43.01 ± 0.63699 ± 853 5234 ± 1248
Girls 90.96 ± 28.27.76 ± 5  67.96 ± 11.6142.69 ± 48.73.30 ± 0.44278 ± 12646117 ± 1852
Post hoc significant differences M > W, B, G*M > B†M > G*M > W, G, B*M > B, GM > B, G*M > B, G*
 W > B, G* W > G†W > B*, G†W > BW > B, G*W > B, G*

The ANOVAs examining the muscle architecture of each quadriceps head showed significant differences (P < 0.01) in muscle volume, fascicle length and PCSA between children and adults. However, the post hoc tests of fascicle length revealed that there were no differences in the fascicle length of the RF and no differences in the angle of pennation in any head of the quadriceps with respect to age or sex. The volume, fascicle length, angle of pennation and PCSA for each group are presented in Table 3 along with the results of the post hoc tests. The relative volume of each head to total quadriceps volume was similar in all groups: VL ∼35, VI ∼25–30, VM ∼20–25 and RF ∼15%. These were also the relative sizes of the PCSA of each head in all the groups.

Table 3.  The mean ±s.d. volume, fascicle length (L), angle of pennation (θ) and physiological cross-sectional area (PCSA) for the vastus lateralis, vastus intermedialis, vastus medialis and rectus femoris muscles
MuscleGroupVolume (cm3) L (mm)θ (deg)PCSA (cm2)
  1. Significant differences between men (M), women (W), boys (B) and girls (G), *P < 0.01, †P < 0.05.

Vastus lateralisMen 691.22 ± 147.994.55 ± 15.415.41 ± 4.3  74.04 ± 17.04
Women 455.94 ± 108.394.39 ± 11.114.07 ± 3.5 48.74 ± 12.5
Boys236.13 ± 42.376.65 ± 10.615.97 ± 2.331.43 ± 7.4
Girls268.57 ± 50.381.97 ± 10.713.96 ± 3.732.94 ± 6.0
Post hoc differencesM > W, B, G*M > B†M > W, B, G*
 W > B, G*W > B†W > B, G*
Vastus medialisMen 523.18 ± 133.895.99 ± 15.525.45 ± 7.6  55.40 ± 16.12
Women350.73 ± 87.884.97 ± 9.3 26.85 ± 7.141.35 ± 9.9
Boys155.46 ± 29.972.75 ± 7.9 23.33 ± 4.821.71 ± 5.4
Girls182.79 ± 40.981.15 ± 6.2 25.84 ± 6.122.40 ± 3.8
Post hoc differencesM > W, B, G*M > B, G*M > W, B, G*
 W > B, G*W > B, G*W > B, G*
Vastus intermedialisMen 557.58 ± 143.195.33 ± 11.213.61 ± 3.4  59.28 ± 17.87
Women373.69 ± 81.086.62 ± 7.6 11.83 ± 2.043.37 ± 9.8
Boys200.81 ± 47.664.78 ± 6.8 11.88 ± 1.630.99 ± 6.7
Girls231.98 ± 42.870.49 ± 7.2 13.52 ± 1.833.11 ± 6.7
Post hoc differencesM > W, B, G*M > B*, G†M > W, B, G*
 W > B, G*W > B, G*
Rectus femorisMen280.71 ± 66.167.72 ± 16.5 29.46 ± 10.2  43.06 ± 11.88
Women178.81 ± 32.864.17 ± 16.8 25.22 ± 12.029.30 ± 8.8
Boys116.17 ± 23.958.84 ± 15.120.85 ± 4.420.46 ± 4.8
Girls110.45 ± 23.656.89 ± 8.5 24.12 ± 8.619.77 ± 5.3
Post hoc differencesM > W, B, G*M > W, B, G*
 W > B, G*W > B, G*

As with the size of each individual head, the volume, PCSA and PCSA × cos θ of the entire quadriceps was greater in men compared with women and in adults compared with children (P < 0.01; Table 4). The magnitude of these differences was proportional to the differences in Fquad, which led to similar estimates of specific tension in all groups (Table 4).

Table 4.  The volume, PCSA and specific tension of the entire quadriceps muscle
GroupVolume (cm3)PCSA (cm2)Specific tension (N cm−2)
  1. Significant differences between men (M), women (W), boys (B) and girls (G), *P < 0.01.

Men2052.7 ± 453.1231.8 ± 55.555.02 ± 11 
Women1359.2 ± 267.8162.8 ± 29.8 57.3 ± 12.6
Boys 708.6 ± 136.3104.6 ± 20 54.06 ± 14.2
Girls793.8 ± 145 108.2 ± 18.559.77 ± 15.3
Post hoc significant differencesM > W, B, G*M > W, B, G*
W > B, G*W > B, G*

For half of the participants, muscle architecture was measured at the angle of peak Fquad. For those whose muscle architecture was measured at the incorrect knee angle, the difference was typically only 5 deg, although in two of the boys the difference was 15 deg, with the number of participants in more extended or more flexed knee angles being roughly equal. For the seven participants who were retested to determine the effect of these errors, the specific tension was 54.74 ± 11.8 N cm−2. The specific tension when using the muscle architecture from MVC at a more extended knee angle was 52.8 ± 11.5 N cm−2, with a standard error from the specific tension at the correct knee angle of 2.9 N cm−2. When using muscle architecture from a contraction at 90% of MVC, the specific tension was 52.7 ± 10.8 N cm−2, with a standard error from the specific tension calculated with architecture at 100% of MVC of 1.4 N cm−2.


The present study quantified the specific tension of the quadriceps femoris muscle in adults and children of both sexes. There were no differences in the specific tension between groups, indicating that the increased knee extensor muscle strength with maturation is not due to a change in muscle quality. Musculoskeletal modelling applications for which the quadriceps specific tension is a necessary input parameter, for example, simulations of the quadriceps femoris force–length and moment–angle properties (Hoy et al. 1990; Van den Bogert et al. 1998; Maganaris, 2004), should treat this parameter as a constant, independent of maturation status and sex.

The measured knee extension JointMVC moment is dependent upon the specific tension, the PCSA and the level of voluntary activation of the quadriceps muscle, the moment arm through which the quadriceps muscle force acts and the level of co-activation of the antagonist knee flexors. Since the present study has shown that specific tension is similar between groups, the different JointMVC moments must result from differences in the remaining parameters listed above. Table 5 shows that accounting for each of these parameters increases the QuadMVC, Quad100% and Fpt compared with the JointMVC of women and boys relative to men, and of girls relative to women. It can be seen that the level of antagonist co-activation has very little effect in reducing the differences in JointMVC. This was because antagonist co-activation moment relative to the QuadMVC moment was very similar in all groups (∼6–9%). The voluntary activation levels of men and women did not differ. Therefore, the sex difference in JointMVC moment postpuberty was mainly due to differences in the PTMA and the PCSA, with the former explaining ∼20% of the difference, meaning that the PCSA explains the remaining ∼80%.

Table 5.  The JointMVC, QuadMVC, Quad100% moments and Fpt of women and boys relative to men, and of girls relative to women
ComparisonJointMVC (%)QuadMVC (%)Quad100% (%) F pt (%)
  1. Abbreviations: QuadMVC, voluntary quadriceps moment; Quad100%, fully activated quadriceps moment; Fpt, patellar tendon force; JointMVC, measured joint moment.

Boys:Men28.529.433.7 46.02

In the boys and girls, voluntary activation level was lower than in the men and women, respectively, and this explained 7% of the difference in JointMVC moment of the males and 31% of the difference in females. The greater increase of JointMVC moment in females compared with males as a result of increased voluntary activation level is in accordance with previous observations that girls display lower motivation and effort than boys during maximal strength tasks (Faust, 1977). Although the differences here were not significant (65 versus 78%, respectively), in combination with the smaller difference in the JointMVC moment between girls and women compared with boys and men (Table 5), voluntary activation level explained a much greater proportion of the increase in JointMVC moment in females compared with males. The PTMA in the adults was also larger and explained 17 and 16% of the difference in males and females, respectively. This means that, in males, 75% of the increased JointMVC moment with pubertal maturation is due to the increased PCSA of male muscle. In females, the increased PCSA of women compared with girls explained only 50% of the increase. The greater growth of PCSA in males compared with females is most probably due to the increased levels of testosterone in the males, since the increase in testosterone is known to result in significant increases in muscle mass (Bhasin et al. 1996) and coincides with the occurrence of the strength differences between the sexes (Round et al. 1999). Fibre type distribution was not quantified in the present study, but our findings indicate that either fibre type distribution was not different between the groups studied, or that if it was different this had no significant effect on the specific tension at a whole-muscle level in vivo.

The similar specific tension values in adults and children found here are in line with previous reports in isolated mouse soleus and extensor digitorum longus muscles (Brooks & Faulkner, 1988) and in contrast with the previous in vivo findings of Morse et al. (2008). The disagreement with the findings of Morse et al. (2008) indicates that either there is a muscle specific effect, which is difficult to explain, or that at least one of the assumptions relating to the joint angle of optimal fascicle length, tendon force distribution or voluntary activation level made by Morse et al. were not valid.

The specific tensions reported here (∼55 N cm−2) are in line with those estimated in optimized musculoskeletal models (Cholewicki et al. 1995; Pain & Forrester, 2009), but notably higher than the typical values of ∼30 N cm−2 reported in previous in vivo studies of the knee extensor muscles (Narici et al. 1992; Chow et al. 1999; Reeves et al. 2004). However, Chow et al. (1999) acknowledged that they had underestimated specific tension because of inaccurate estimates of fascicle pennation, failure to account for antagonist activation and the use of mean muscle force across a number of trials rather than maximal muscle force. Narici et al. (1992) and Reeves et al. (2004) did not account for the force that is lost in the transmission of quadriceps force through the patella to the tibia (Buff et al. 1988), and used Fpt to calculate specific tension rather than Fquad. At the mean knee angle of 65 deg, this loss of force is equal to a third of Fquad. If the specific tension calculated in the present study had been calculated from the Fpt, rather than Fquad, this would have resulted in estimates of ∼37 N cm−2, which are much closer to those previously reported. Additionally, in the present study Fpt was calculated from PTMA lengths measured at rest, which are smaller than those during contraction (Tsaopoulos et al. 2007). Had PTMA lengths during contraction been used, according to the findings of Tsaopoulos et al. (2007), the specific tension in the men would have been ∼34 N cm−2. Moreover, the use of the tibio-femoral contact point to quantify the PTMA length may have also contributed to some of the differences between specific tension estimates of the present and previous studies. This method produces PTMA values smaller than would be obtained from the instant centre of rotation method (Tsaopoulos et al. 2009), which will result in an overestimation of the tendon force and the accompanying specific tension. However, measuring PTMA length from the instant centre of rotation is a complex and time-consuming process that is not always practicable. The remaining difference is likely to be a result of the incorporation of muscle architecture from all quadriceps heads, rather than just the VL, resulting in slightly smaller PCSA estimates. The specific tension values obtained in the present study are also much higher than those reported by Maganaris et al. (2001; 15 N cm−2). Such a large difference is most probably due to the fact that specific tension in that study was underestimated because muscle forces were not maximal, since the muscles studied were activated by percutaneous electrical stimulation, which is unlikely to recruit the whole muscle.

A number of limitations in this study must be considered. Muscle architecture was measured at MVC and in some instances not at optimal joint angle, but the potential error introduced was small (∼4%) and not likely to have any significant impact on the validity of the specific tension estimates. Cross-talk between the quadriceps and hamstrings EMG would affect the estimated level of antagonistic joint moment, but it has been shown that the magnitude of cross-talk between knee extensors and flexors is generally small (Koh & Grabiner, 1992), hence unlikely to have had a significant effect for the EMG measurements in the adults. However, because the children have smaller muscles and the same EMG electrodes were used for both children and adults, the possibility of cross-talk is relatively greater in children. It should also be considered that the calculation of specific tension in the present study relies on a number of necessary assumptions relating to contractile force generation and transmission. It was assumed that muscle volume consists entirely of contractile material, that optimal quadriceps muscle length corresponds to the optimal sarcomere length in each of the four individual quadriceps muscle heads and that sarcomere length within each head is homogeneous. It was also assumed that the aponeurosis and the tendon of the quadriceps muscle are in line and that only the quadriceps muscle contributes to the force in the quadriceps tendon, i.e. no force transmission occurs from other muscles that act across the knee joint (e.g. sartorius). Finally, it should be acknowledged that the exact stage of biological maturation for the boys and girls studied was not directly assessed, and the application of the present findings to studies with children at various stages of maturation must be done with care.

In conclusion, adults were significantly stronger than children, and men were significantly stronger than women. However, when antagonist co-activation and agonist voluntary activation levels and the tendon moment arm length were accounted for, the specific tension of the quadriceps muscle was similar in all groups. The increased size of adult muscle compared with the children explained approximately 75 and 50% of the externally measured strength differences in males and females, respectively. Compared with women, the increased PCSA in men explained 80% of these strength differences. These findings indicate that the strength gains as a result of maturation are not due to an improved ‘muscle quality’.