Architecture and functional ecology of the human gastrocnemius muscle‐tendon unit

Abstract

The gastrocnemius muscle‐tendon unit (MTU) is central to human locomotion. Structural variation in the human gastrocnemius MTU is predicted to affect the efficiency of locomotion, a concept most often explored in the context of performance activities. For example, stiffness of the Achilles tendon varies among individuals with different histories of competitive running. Such a finding highlights the functional variation of individuals and raises the possibility of similar variation between populations, perhaps in response to specific ecological or environmental demands. Researchers often assume minimal variation in human populations, or that industrialized populations represent the human species as well as any other. Yet rainforest hunter‐gatherers, which often express the human pygmy phenotype, contradict such assumptions. Indeed, the human pygmy phenotype is a potential model system for exploring the range of ecomorphological variation in the architecture of human hindlimb muscles, a concept we review here.

Introduction

The gastrocnemius muscle‐tendon unit (MTU) is essential to human locomotion. It supplies the main propulsive force of walking and running, and is therefore a central focus of diverse disciplines, including clinical biomechanics, exercise physiology, and human evolutionary biology. Structural variation in the human gastrocnemius MTU (Roy & Edgerton, 1992; Wilson & Lichtwark, 2011) is predicted to affect locomotor efficiency, a concept most often explored in the context of athletic performance. For example, the stiffness of the Achilles tendon is known to vary among individuals with different histories of competitive running. The stiffer tendons of sprinters (Arampatzis et al. 2007) are predicted to economize the high metabolic costs of sprinting and confer a competitive advantage (Lichtwark & Wilson, 2008). Such a finding highlights the functional variation that exists between individuals and raises the possibility of similar variation between populations, perhaps in response to specific ecological or environmental demands.

This concept is termed ecomorphology (Wainwright, 1991) and it is central to studies of primate evolution. Accordingly, a rich literature exists on the ecomorphology of nonhuman primates, including the architecture of the gastrocnemius MTU (review: Hanna & Schmitt, 2011). The architecture of skeletal muscle refers to the number and orientation of muscle fibers within a muscle (Gans, 1982; Lieber & Fridén, 2000). Architectural variables include muscle mass and length, fiber length, pennation angle, and sarcomere length. From these variables, several summary parameters can be calculated, such as the ratio of muscle fiber length to muscle length and the physiological cross‐sectional area (PCSA). Muscle fiber length is proportional to the distance over which the fiber can shorten (maximum muscle excursion), whereas PCSA is directly proportional to the maximum tetanic tension that can be generated by a muscle (maximum muscle force). PCSA is therefore the best predictor of muscle functional capacity (Lieber & Ward, 2011) and is calculated as:

$$\begin{array}{cc}\hfill \text{PCSA}({\text{cm}}^2)=& [\text{Muscle mass (g)}\times \text{cosine}\mathrm{\theta }]/\hfill \\ & [\mathrm{\rho }({\text{g/cm}}^3)\times \text{Fiber length (cm)}]\hfill \end{array}$$ (1)

where θ is the surface pennation angle and ρ is muscle density (1.056 g cm^–1 for fixed mammalian muscle) (Lieber, 2009). In theory, the PCSA represents the sum of the cross‐sectional areas of all the muscle fibers within the muscle.

The PCSAs of ape hindlimb muscles demonstrate ecomorphological variation (Thorpe et al. 1999; Payne et al. 2006a,b; Myatt et al. 2011; Holowka & O'Neill, 2013) and it is tempting to use the PCSAs of chimpanzees and humans to parameterize models of hominin bipedal efficiency (Wang et al. 2004; Pontzer et al. 2009). For example, Pontzer et al. (2009) used kinematic and energetic data from humans and chimpanzees (moving quadrupedally and bipedally) to model the energetic costs of bipedalism. Results of the model suggest that hindlimb length, posture (effective mechanical advantage) and muscle fascicle length contribute nearly equally to the different costs of walking for humans and chimpanzees. The authors then used the model to estimate the energetic cost of walking for an early, ape‐like hominin (AL 288‐1; Australopithecus afarensis). The cost of walking for AL 288‐1, across a range of hypothetical postures from crouched to fully extended, was below that of quadrupedal apes, but above that of modern humans. Such a finding is important because it supports the hypothesis that natural selection favored increasingly efficient bipedalism during the course of hominin evolution.

Yet the estimated body size of AL 288‐1 is quite small (stature: 105.4–109.0 cm; mass: ~27 kg), as derived from regression formulae in a skeletal sample of Aka (pygmy) foragers and bonobos, Pan paniscus (Jungers, 1988, 1991; McHenry, 1992). The small body size of AL 288‐1 raises questions about the suitability of using architectural data from the hindlimb muscles of large‐bodied, industrialized humans to model the bipedal efficiency of early hominins. Another question concerns the sources of these data, which can be collected in vivo using non‐invasive imaging techniques, such as magnetic resonance imaging or ultrasonography (e.g. Fukunaga et al. 1992; Narici et al. 1996; Maganaris et al. 1998; Bolsterlee et al. 2015), or from cadavers, the gold standard (review: Yamaguchi et al. 1990).

The two most widely used datasets are based on cadaveric measures from five individuals of unknown size, sex and age (Wickiewicz et al. 1983; Friederich & Brand, 1990). Another cadaveric study focused solely on the gastrocnemius, but again, the size, sex and age of the eight individuals were unreported (Huijing, 1985). These small sample sizes limit the value of confidence intervals for generating accurate predictions of muscle force or excursion. They also frustrate any appreciation for scaling effects on variables such as muscle fascicle length, an important parameter in models of hominin bipedal efficiency (Wang et al. 2004; Pontzer et al. 2009).

More recently, Ward et al. (2009) reported cadaveric measures from the hindlimbs of 21 humans in San Diego County, California, USA (mean age ± 1 SD, 83 ± 9 years; male : female ratio, 9 : 12; height, 168.4 ± 9.3 cm; weight, 82.7 ± 15.3 kg). This dataset is a signal and commendable achievement; however, it is also limited to members of an industrialized population. Some evolutionary psychologists would describe this population as WEIRD (Western, Educated, Industrialized, Rich, and Democratic), an epithet designed to call attention to the problem of sample bias in human evolutionary studies (Henrich et al. 2010). The advanced age of the population is also problematic, as aging has been shown to significantly affect skeletal muscle architecture, leading to a reduction in the gastrocnemius muscle fascicle length, pennation angle, and volume (Narici et al. 2003).

Researchers – often implicitly – assume minimal variation in human populations, or that industrialized populations represent the human species as well as any other. Yet recent studies have revealed unexpected variation in the architecture and mechanical properties of the gastrocnemius MTU. For example, differences exist between men and women (Chow et al. 2000), sprinters and endurance runners (Arampatzis et al. 2007), and even women with distinct histories of wearing high‐heeled shoes (Csapo et al. 2010). Perhaps more surprising are the differences between populations. For example, Venkataraman et al. (2013) focused on the gastrocnemius MTU of small‐bodied (pygmy) populations, and suggested that habitual tree‐climbing can have an effect on the relative length of muscle fiber bundles. Given the small body size and inferred climbing behavior of AL 288‐1, the hindlimb muscles of human pygmy populations could be more appropriate than those of industrialized peoples for informing theoretical models of hominin walking efficiency, a concept we explore here.

The human pygmy phenotype

The prevailing definition of the human pygmy phenotype is based on a mean male stature of < 155 cm (Perry & Dominy, 2009). It has evolved three or more times independently in Central Africa, Southeast Asia, and South America (Migliano et al. 2013; Perry et al. 2014), and almost always in association with tropical rainforest habitats. The convergent origins of pygmy size have led anthropologists to hypothesize that small body size confers direct or indirect fitness benefits in response to the ecological challenges of hunting and gathering in tropical rainforests (review: Perry & Dominy, 2009). Such challenges include: (i) limited food and/or nutritional resources (O'Dea, 1993; Shea & Bailey, 1996; López Herráez et al. 2009); (ii) warm, humid conditions (Cavalli‐Sforza, 1986b); (iii) dense forest undergrowth (Diamond, 1991); and (iv) high adult mortality rates (Migliano et al. 2007).

Some researchers have focused on the gastrocnemius as an instructive line of evidence in support of the hypothesis that small body size is a heat‐minimizing, energy‐economizing adaptation to rainforest habitats. For example, Cavalli‐Sforza (1986a) tabulated the anthropometric data collected by Gusinde (1948) and observed that:

Pygmies (inhabiting the Ituri Forest, Democratic Republic of Congo) have especially thin calves, legs, and arms. Particularly for calves, which are easier to measure, I have found data confirming this superficial impression. Calf circumference is among the smallest pygmy measures when expressed as a proportion of the total height (Cavalli‐Sforza, 1986b, pp. 398–399).

Cavalli‐Sforza's argument was twofold. First, given that body heat during exercise is proportional to body mass, it follows that natural selection favored absolutely smaller bodies in warm, humid habitats, i.e. a lower body mass mitigates the risk of foraging‐induced overheating. Secondly, as skeletal muscle is metabolically costly, natural selection should favor relatively small muscles in calorie‐limited environments. This latter premise is compelling; however, Hiernaux (1977) reported comparable height–weight relationships among African peoples in a wide range of habitats, providing no evidence of systematic muscle reduction in pygmy populations, including the Mbuti of the Ituri Forest. Instead, the pygmy gastrocnemius MTU appears to be particularly sensitive to ecomorphological variation.

If calf circumference is proportional to the cross‐sectional area (CSA) of the gastrocnemius, the gastrocnemius CSAs are predicted to be relatively low in populations that express the pygmy phenotype. To test this prediction – and its corollary, that gastrocnemius PCSAs (and hence, force capacities) are also relatively low – it is plausible to use portable ultrasound equipment (Fig. 1A) to measure architectural variables in vivo (Fig. 1B). This method has the practical advantage of enabling research on remote study populations. For example, Venkataraman et al. (2013) used field‐portable ultrasound equipment to estimate the lengths of muscle fiber bundles in the gastrocnemius of two African populations with different subsistence ecologies, the Bakiga and Batwa.

Figure 1.

Field‐portable ultrasonography. (A) MicroMaxx ultrasound system outfitted with an L52e transducer (SonoSite, Bothell, WA, USA). (B) Posterior view of the lower leg and transverse view of the lateral head of the gastrocnemius (GAS; red shade) and soleus (SOL) muscles. This view enables estimates of cross‐sectional area (CSA). (C) Variation in the CSA of the lateral head of the gastrocnemius as a function of adult stature for the Bakiga (farmers) and Batwa (pygmies) of southwestern Uganda.

Bakiga and Batwa

The two populations live in the vicinity of Bwindi Impenetrable National Park, southwest Uganda. The Bakiga are a Bantu‐speaking agricultural population (Edel, 1996), whereas the Batwa hunted and gathered in Bwindi until it was declared a national park in 1991. (Note: a convention in the ethnographic literature is to omit the Bantu prefix Ba‐, and therefore use the ethnonym Twa. A growing consensus, however, favors Batwa as it is the preferred nomenclature of the Batwa themselves). The Bakiga and Batwa have coexisted for at least 500 years (Ngologoza, 1998) and genetic admixture is common (Perry et al. 2014). Mean statures of self‐identified Batwa (66 men, 152.9 cm; 103 women, 145.7 cm) are less than those of the Bakiga (20 men, 165.4 cm; 41 women, 155.1 cm) (Perry et al. 2014) and comparable to Batwa who live in the vicinity of Lake Kivu: years 1929–30: 17 men, 152.9 ± 6.0 cm; eight women, 140.2 ± 6.5 cm (Schebesta & Lebzelter, 1933); 1934: 101 men, 153.0 cm; 84 women, 144.2 cm (Gusinde, 1949); 1970s: 15 men, 153.3 ± 6.4 cm (Ghesquiere & Karvonen, 1981). Collectively, these populations of Batwa are termed lacustrine (Schadeberg, 1999) or Great Lakes Batwa (Lewis, 2000), and they meet the definition of the pygmy phenotype.

Climbing

Hunter‐gatherers worldwide climb to harrowing heights. Honey most strongly motivates climbing behavior, but hunting and fruit and seed collection are also important activities (Demps et al. 2012; review: Kraft et al. 2014). Tree climbing for honey is a frequent and important practice of hunting and gathering populations living near the Albertine Lakes of East Africa (Ichikawa, 1981; Bailey, 1991). The two main styles of unassisted climbing, first described by Skeat & Blagden (1906) and then by Schebesta (1928), involve (i) applying the plantar surface of the feet to a tree or vine and ‘walking’ upward with the legs and arms advancing alternately, or (ii) ‘pulsing’ up a tree by ‘hugging’ it with the arms and gripping the trunk on either side with the soles of the inverted feet. We have observed the former style among Batwa men (Fig. 2A) and measured extreme dorsiflexion (> 45°) at the tibiotalar joint (Venkataraman et al. 2013). We also used portable ultrasound to estimate the lengths of muscle fiber bundles in the gastrocnemius (Fig. 2B) and found a positive association with tree‐climbing behaviors at the population level (Fig. 2C; Venkataraman et al. 2013). This result suggests that habitual dorsiflexion of the ankle during vertical climbing favors a more excursive joint and, thus, proportionally longer muscle fibers.

Figure 2.

Ecomorphology of the gastrocnemius MTU. (A) A Batwa male ascends a tree in the pursuit of honey, demonstrating extreme ankle dorsiflexion (photograph by George H. Perry, reproduced with permission). (B) Lateral view of the leg and sagittal view of the lateral head of the gastrocnemius (GAS; red shade) and soleus muscle (SOL), separated by the deep aponeurosis. Arrows indicate individual muscle fiber bundles. This view enables estimates of pennation angle (θ) and muscle fiber bundle lengths. (C) Boxplots of muscle fiber bundle lengths, normalized for total muscle length, of the lateral and medial heads of the gastrocnemius (GAS) among Bakiga (n = 9) and Batwa men (n = 29).

Among industrial populations, the plantar flexors, i.e. the medial and lateral heads of the gastrocnemius and the soleus, have large PCSAs, high pennation angles, and short fibers (low fiber length to muscle length ratio) (Lieber, 2009). Such architecture suggests that the primary role of these muscles is to generate high forces over small excursions. However, if greater excursion is required at the ankle – in the case of tree climbing – then the muscle fibers will lengthen. Any increase in muscle fiber length will produce lower PCSAs (see Eq. (1)) and CSAs (Fig. 1C; Table S1), which could account for Cavalli‐Sforza's (1986b) observations of relatively small calf circumferences among pygmy peoples in the Ituri Forest, Democratic Republic of Congo.

It follows that longer muscle fibers will correspond with proportionally shorter Achilles tendons. The Achilles tendon, when measured as a proportion of tibia length, is reported to be 65% (Frey, 1913), a figure that is often used in models of hominin bipedal efficiency. A practical use of ultrasonography is the visualization of the gastrocnemius muscle‐tendon junction in vivo, which enables estimates of Achilles tendon length (Fig. 3A; Supporting Information Table S1). For example, Rosso et al. (2012) reported mean (± 1 SD) Achilles tendon lengths as a proportion of tibia length (49 ± 5%) in a sample of 52 Europeans. Here we report comparable means (± 1 SD) as a proportion of fibula length in a sample of 32 Bakiga (57.2 ± 0.04%) and 78 Batwa (54.2 ± 0.05%; Fig. 3B; Table S1), a difference that reached statistical significance (Welch's t = 3.5, F = 11.96, P < 0.001). The Achilles tendon plays an important role in the storage and release of elastic energy, but the functional significance for the Batwa of a marginally shorter Achilles tendon during walking and running is uncertain (Sellers et al. 2010).

Figure 3.

Surface landmarks used for anthropometry. (A) Lateral and medial views of the left leg. Manual palpation was used to locate the head of fibula (HF) and the most prominent aspect of the lateral (LM) and medial malleolus (MM). Ultrasound was used to determine the muscle‐tendon junction (MTJ) of the gastrocnemius and Achilles tendon. From these landmarks, we estimated the lengths of the fibula and Achilles tendon. The moment arm of the ankle was calculated as the mean horizontal distance between the midpoint of the lateral and medial malleoli and the Achilles tendon. This view also enables estimates of total foot length. (B) Achilles tendon length as a proportion of fibula length (mean ± 1 SD) for the Bakiga (57.2 ± 0.04%; n = 32) in comparison with the Batwa (54.2 ± 0.05%; n = 78). *This difference reached statistical significance (Welch's t = 3.5, F = 11.96, P < 0.001).

Walking and running

Rainforest understories are complex and challenging environments for humans. Vegetation obstructs and encumbers movement and the ground itself is often sodden and fraught with tortuous roots, creating a slippery, uneven surface. Such impediments could explain the ‘springy’ (Schebesta, 1928) or ‘high‐stepping’ gait (Evans, 1937) that pygmy peoples employ to negotiate dense rainforest habitats swiftly (Putnam, 1948). The kinematics of high‐stepping include increased hip and knee flexion to provide greater foot clearance and a corresponding reduction in stride length, which is energetically costly due to increased muscle activations (Voloshina et al. 2013). Such a distinctive gait is perhaps expected to result in energy‐economizing, ecomorphological variation of the hindlimb.

A variable relevant to walking and running and also easily measured under field conditions is the moment arm of the Achilles tendon (Fig. 3A). By definition, the moment arm of the Achilles tendon is the shortest distance from the line of action of the Achilles tendon to the center of rotation of the ankle. In addition to skeletal muscle architecture, moment arms can be tailored to a particular performance requirement. While the moment arm affects the maximum torque at a given joint, the ratio of muscle fiber length‐to‐moment arm provides additional information about the relative muscle‐joint influence. Muscles with longer fibers have a longer functional range than muscles with shorter fibers, but this does not necessarily mean that muscles with longer fibers are associated with joints with larger ranges of motion or that muscles with shorter fibers are associated with joints with smaller ranges of motion. The amount of muscle length change that occurs as a joint rotates strongly depends on the muscle moment arm (Lieber, 2009), i.e. a muscle will change length much less for a given change in joint angle when attached with a short moment arm, as compared with a long moment arm. Thus, a muscle that appears to be adapted for force production due to a large PCSA and short fiber lengths may actually produce a larger joint excursion if placed in a position with a small moment arm.

Viscoelastic energy storage in the Achilles tendon is most sensitive to the moment arm: the smaller the moment arm, the more energy is stored in the tendon (Scholz et al. 2008). As the Achilles tendon plays an important role in the storage and release of elastic energy during running (Alexander, 1991), it follows that a smaller moment arm will be associated with better running economy. Endurance running, particularly during persistence hunting, is a potentially important aspect of hunting and gathering ecology (Liebenberg, 2006). Such persistence hunting is unknown among the hunter‐gatherers of Central Africa, who tend to ambush or capture prey by means of communal bow‐ and net‐hunting (Bahuchet, 2014). In some cases, however, hunters sprint to subdue moribund or netted prey, and spear‐hunts of elephants can entail brief pursuits on trails (Turnbull, 1965). Spear‐hunts of swift prey such as bushpigs and duikers (Yasuoka, 2006) probably require off‐trail sprints.

In our dataset (see Table S1), the mean (± 1 SD) moment arms of the Batwa (3.86 ± 0.46 cm) and Bakiga (3.94 ± 0.40 cm) are indistinguishable (F = 0.98, P = 0.32). Accordingly, the relative length of the moment arm, when expressed as a proportion of the gastrocnemius MTU, is longer among the Batwa (10.9 ± 1.3% vs. 10.0 ± 0.9%; F = 12.7, P < 0.001), although the functional significance of this negligible difference is uncertain. The moment arm of six European men of unreported stature was ~4.75 cm for 0° dorsiflexion at rest (Maganaris, 2004). This value is nearly 1 cm greater than that of the Batwa, but it remains to be determined how the moment arm differs as a proportion of the gastrocnemius MTU.

To investigate locomotor efficiency in pygmies and non‐pygmies, Minetti et al. (1994), recruited 13 Baka pygmies and seven Europeans to walk and run on a treadmill under field conditions in Cameroon, while recording simultaneous metabolic measurements and two‐dimensional motion analysis. From these measures, energy expenditure and mechanical external and internal work were calculated. Among the Baka, the metabolic energy cost, stride frequency, and internal mechanical work were higher than among the Europeans during walking at all speeds (P < 0.05). The major determinant of the higher cost of walking for the Baka was an increased stride frequency, which led to an increase in mechanical internal work. Most differences between the Baka and Europeans during walking tended to disappear when speed was normalized to the Froude number, a dimensionless value representing the ratio of centripetal force to gravitational force during walking:

$$\text{Froude number}={[\text{velocity (m/s)}]}^2/[\text{leg length (m)}\times \text{acceleration due to gravity}({\text{m/s}}^2)]$$ (2)

However, this was not the case for running, suggesting that the Baka differ in some aspects of the mechanics of locomotion. During running, the Baka's metabolic energy cost tended to be lower and the stride frequency and internal mechanical work were higher (NS), even though the external mechanical work was similar to that of the Europeans. For this study population, the inverse relation between energetic cost of locomotion and body mass (Taylor et al. 1970, Heglund & Taylor, 1988) was consistent during walking but not running. To explain the exceptional running efficiency of the Baka, Minetti et al. (1994) speculated that they might have proportionally shorter or stiffer Achilles tendons. The former hypothesis is supported by recent findings among the Batwa (Fig. 3C), whereas the latter hypothesis is unlikely given the tandem prediction of low PCSAs and correspondingly low strengths of the gastrocnemius among pygmy peoples (see above), and the positive correlation between plantar flexor muscle strength and Achilles tendon stiffness (Muraoka et al. 2005).

Summary

There are multiple factors related to skeletal muscle architecture that affect locomotion, and these factors vary among human populations. The differences reviewed here have implications for theoretical models of hominin bipedal efficiency (Wang et al. 2004; Pontzer et al. 2009). If early hominin ecology included arboreal behaviors, which seems highly likely (White et al. 2009; Crompton et al. 2010; Venkataraman et al. 2013), then models based on the PCSAs and Achilles tendon lengths of industrialized peoples are bound to overestimate the maximum potential force capacities (and hence, energetic costs) of hominin bipedalism. This supposition highlights the need for greater diversity in skeletal muscle sampling, especially in relation to the functional ecology of the human gastrocnemius MTU.

Supporting information

Table S1. Anthropometric data, including architectural properties of the gastrocnemius muscle‐tendon unit, collected in 2009 from 92 Batwa and 94 Bakiga in the vicinity of Bwindi Impenetrable National Park, Uganda.