The jaw muscles are the motors of the masticatory apparatus. Given their role in producing forces and doing work during chewing, the musculoskeletal morphology of the masticatory apparatus plays a fundamental role in jaw-muscle and masticatory performance. Numerous studies have examined where the jaw muscles are positioned on the skulls of primates and how muscle location impacts masticatory leverage (Biegert, 1963; Zingeser, 1973; Hylander, 1975, 1979; Ravosa, 1990; Anapol and Lee, 1994; Spencer, 1999; Williams et al., 2002; Vinyard et al., 2003; Wright, 2005; Eng et al., 2009). The architectural configuration of the jaw muscles is a second fundamental component of muscle form that impacts performance. Over the past few years, there has been a resurgent interest in the architecture of primate jaw muscles (Taylor and Vinyard, 2004, 2009; Anapol et al., 2008; Perry and Wall, 2008; Taylor et al., 2009). These analyses have focused primarily on how variation in jaw-muscle architecture relates to body size, feeding behavior, and diet in primates.
In this study, the scope of investigation into primate jaw-muscle architecture is expanded by considering the relationship between jaw-muscle fiber architecture and jaw-muscle recruitment patterns during chewing. This relationship is examined for a handful of primate species for which we currently have both architectural and jaw-muscle electromyographical (EMG) data during chewing. Because this study is a preliminary analysis based on a small sample size, a simplistic hypothesis is tested: that relative variation in jaw-muscle architectures and EMG parameters is correlated across primates. It is predicted that primates recruiting a greater percentage of their jaw muscles during chewing will exhibit jaw-muscle architectures geared for relatively increased force production. Significant correlations among architectural and EMG variables would imply functional relationships that exist across primates. Alternatively, the absence of significant correlations would not preclude functional relationships, but rather indicate clade and/or species-specific relationships that vary throughout the order.
Jaw-Muscle Architecture: Linking Form and Function
Fiber architecture is a key determinant of a muscle's ability to generate force and excursion during contraction (Gans and Bock, 1965; Gans, 1982; Powell et al., 1984) and therefore plays an important role in muscle performance. Fiber architecture describes the internal arrangement of fibers relative to the force-generating axis(es) of a muscle. Individual fibers may run parallel to a force-generating axis or they may be pinnate and aligned at an angle to this axis. Holding nonarchitectural factors constant, muscles with predominantly pinnate fibers tend to produce relatively greater force. Alternatively, muscles composed primarily of parallel fibers tend to yield relatively greater muscle excursion.
The functional contrast between parallel- and pinnate-fibered muscles results in an architectural tradeoff between force production and excursion along this continuum of muscle morphologies. Parallel muscle fibers tend to be relatively longer and contain more sarcomeres in series. Given a constant stretch per individual sarcomere, the greater number of serially-arranged sarcomeres increases the distance that a muscle can shorten (or lengthen) (Williams and Goldspink, 1978). Following on this physiological relationship, fiber length is proportional to muscle excursion and by extension contraction velocity.
Pinnate fibers tend to be relatively shorter, but their orientation at an angle to the force-generating axis allows more fibers to be packed next to each other in a muscle. The maximum force a muscle can generate during isometric contraction is related to the number of fibers, their diameter, and their alignment relative to the force-generating axis. The angular orientation of the fiber to the force-generating axis results in some loss of force along the axis, but at small pinnation angles (e.g., <20° angle) this loss is smaller than the gain from the increased packing of muscle fibers per unit area (Gans, 1982). The increased number of fibers per unit area explains why pinnate-fibered muscles tend to produce greater force compared with parallel-fibered muscles of similar volume. The physiologic cross-sectional area (PCSA) of a muscle represents the cross-sectional areas of all fibers within a muscle and therefore is proportional to the maximum force a muscle can generate (Powell et al., 1984).
Jaw-Muscle Electromyography: Behavioral Records of Muscle Function
When muscles contract they generate a bioelectrical signal that can be recorded, quantified, and compared across muscles, individuals, and species (Gans and Gorniak, 1980; Basmajian and De Luca, 1985; Loeb and Gans, 1986; Hylander and Johnson, 1993). EMG data from the jaw muscles of primates have been essential in assessing variation in jaw-muscle function and testing hypotheses relating morphological variation to masticatory function across the order (Hylander and Johnson, 1985, 1994; Hylander et al., 1987, 2000, 2004, 2005; Ross and Hylander, 2000; Vinyard et al., 2005, 2006, 2008; Wall et al., 2006, 2008). Based on these efforts, along with those focused on human evolution and biomedical applications, primates are the best sampled group among mammals for EMG data during chewing. With this broader sampling of species, one can begin to explore relationships among motor recruitment patterns and jaw-muscle morphology throughout primates (e.g., Vinyard et al., 2008).
Researchers have primarily used jaw-muscle EMGs collected from primates to study the timing and relative recruitment of these muscles during chewing. The timing of jaw-muscle activity demonstrates when a muscle experiences peak activity, how long a muscle is active and the relative order of activation among muscles involved in a given behavior. Describing the pattern of activation among jaw muscles has significantly advanced our understanding of primate mastication and its impact on jaw morphology (Hylander et al., 1987, 2000, 2005; Hylander and Johnson, 1994; Vinyard et al., 2005, 2006, 2007, 2008). The relative levels of recruitment on the chewing (or working) and nonchewing (or balancing) sides of the jaw have also been essential in developing our current understanding of primate masticatory functional morphology. In primates, the assessment of relative muscle recruitment has focused primarily on comparing relative levels of activation between working- and balancing-side muscle pairs during chewing of different food items. This study will focus on this working-side/balancing-side (W/B) ratio measuring relative recruitment in assessing the relationship between jaw-muscle architecture and jaw-muscle activity across primates.
MATERIALS AND METHODS
Muscle architectural parameters were measured for 11 primate species and Belanger's treeshrew (Tupaia belangeri) (Table 1). Data for Callithrix jacchus and Cebus apella are taken from previously published work, whereas new data are provided for the remaining species. Data for Lemur catta and Otolemur garnettii (temporalis weight and fiber length only) are from Perry and Wall (2008). All individuals used for fiber architecture analysis were dental adults lacking obvious craniofacial pathologies. Most were captive. Specimens were fixed in 10% buffered formalin or ethanol before analysis. Published EMG data were taken for 11 primate species and supplemented with unpublished data for C. apella (Table 1). EMG data were collected from subadults or adults lacking craniofacial pathologies. In some cases, data were collected from two species of the same genus (e.g., Macaca) because of animal availability. Rather than introduce marked discrepancies in phylogenetic distance at the interspecific level, these closely related species were combined into “species groups” for analysis (see Table 1). Eight species groups were analyzed.
In some cases, we collected data from two species of the same genus (e.g., Macaca) because of animal availability. Rather than introduce marked discrepancies in phylogenetic distance at the interspecific level, we combined these closely related species into “species groups” for analysis.
Primate taxonomy adapted from Fleagle (1999); taxonomic groupings selected based on the exclusion of tarsiers from our sample. Belanger's treeshrews (order Scandentia) are placed in Tupaiidae following Olson et al. (2005) and references therein.
Only EMG data are available for this species.
Perry and Wall (2008) could not calculate pinnation angle for O. garnettii specimens because of cadaver preservation. Given the difference in methods, we excluded their estimate of PCSA from this dataset.
Perry and Wall (2008) did not adjust data to a given sarcomere length, but used only specimens fixed in a near centric occlusion jaw posture.
Marmoset species means differ from estimates published in Vinyard et al. (2007) by the removal of the single tamarin from the callitrichid sample and the addition of a small amount of new data.
Only muscle architecture data are available for this species.
Methods for quantifying jaw-muscle architecture follow recently published protocols in Taylor et al. (2009) and Taylor and Vinyard (2009). We briefly summarize methods here, although more detailed explanations can be found in these publications. Whole superficial masseter and temporalis muscles were dissected from the skull and weighed to the nearest 0.01 g (Table 2). After weighing, muscles were sectioned along their lengths into ∼1.5-cm-thick segments. Segments were oriented in cross section to visualize muscle fibers (Fig. 1). Sampling sites were identified at each end of the cross section, and up to six adjacent fibers were measured. For each fiber, we measured fiber length between the proximal and distal myotendinous junctions (Lf) and the perpendicular distance (a) from the distal end of the fiber to the central myotendinous junction. The pinnation angle (θ) was estimated as the arcsine of a/NLf (Anapol and Barry, 1996) (Fig. 1). Pinnation angles were computed for each fiber and then averaged across all fibers for a muscle.
Table 2. Species means and (standard deviations) for jaw-muscle (a) architectural and (b) EMG variables
N = sample size. Additionally, not all cadavers possessed both masseter and temporalis muscles.
Condyle-M1 lengths and jaw-muscle architectural variables are not necessarily from the same individuals.
The first number represents masseter sample size based on new data published here, while the second number indicates the sample size for temporalis weight and fiber length from Perry and Wall (2008).
Fiber lengths were normalized to a standardized sarcomere length, given that measured fiber lengths can vary based on fixation at different degrees of gape (see Felder et al., 2005). Muscle segments were chemically digested in 30% HNO3 and saline solution (Loeb and Gans, 1986) and 5–10 fiber bundles were dissected from each muscle segment. In situ sarcomere lengths were measured from each fiber bundle using laser diffraction (Lieber et al., 1984). We normalized to resting fiber length (NLf) by dividing a standard sarcomere length of 2.41 μm by the in situ sarcomere length and multiplying by raw fiber length. The standardized value of 2.41 μm is based on optimal sarcomere length in macaque limb muscle (Walker and Schrodt, 1974).
The following variables were computed based on these data:
1Mean fiber lengths (NLf) for the superficial masseter and temporalis estimated as the average of all fibers sampled from a respective muscle.
2PCSA, computed as: is the specific density of muscle (Mendez and Keys, 1960) (Table 2).
Jaw-Muscle Electromyography Data
Electromyographic data recording follows methods outlined in Hylander et al. (2000) and Vinyard et al. (2005). As the vast majority of these data are taken from published sources, the EMG data collection and manipulation methods are not described. Electrode construction, implantation, EMG recording, and data manipulation for Cebus apella followed the methods used in Hylander et al. (2000) and Vinyard et al. (2005).
Following data collection, filtered digital EMG data were rectified and integrated using a root-mean-square (rms) algorithm (with a 42 ms time constant) to create a single EMG waveform (Hylander and Johnson, 1993; see Fig. 2 in Hylander et al., 2000). The peak activity level of the rms-EMG waveform was recorded for each muscle during a chewing cycle. Because EMG amplitude can vary for several reasons (including signal quality and other non-biological factors) (e.g., Basmajian and De Luca, 1985; Loeb and Gans, 1986), comparisons among electrodes typically involve computing relative amplitudes. To scale electrode amplitudes, the largest peak rms value across all cycles in a muscle (i.e., per electrode) in an experiment was assigned a peak value of 1.0, and all remaining peaks in other cycles for that electrode are linearly rescaled to range between 0 and 1 (Hylander et al., 2000). These measures represent relative estimates of muscle recruitment during chewing (Hylander and Johnson, 1993; Hylander et al., 2000). Each chewing cycle was designated as a left- or right-side cycle allowing the determination of working-side (W), or chewing, and balancing-side (B), or nonchewing, muscles for each cycle. Ratios of working- to balancing-side scaled activity (W/B ratios) were calculated to estimate the amount of relative activity in a muscle pair during chewing. For W/B ratios, a value of 1.0 indicates equal levels of relative recruitment between a working- and balancing-side muscle pair, whereas values >1.0 suggest increased amounts of relative recruitment in the working-side muscle. We compare jaw-muscle architecture to W/B ratios for the superficial masseter, as well as the anterior and posterior temporalis across primates (Table 2; Fig. 2).
Dimensionless variables were created to compare relative differences among primates that differ in absolute body size. Jaw-muscle W/B ratios are dimensionless, and no further manipulation of published data was performed. Masseter and temporalis architectural variables were divided by the distance from the condyle to M1. Condyle-M1 length estimates the load arm during chewing and provides a mechanically-relevant standard for comparison of jaw-muscle variables (Vinyard, 2008). To make dimensionless shape ratios for muscle weight (g) and PCSA (cm2), the cube root of weight and the square root of PCSA were calculated before dividing by condyle-M1 length.
Jaw-muscle architecture and EMG variables were compared using product–moment correlations (a priori α = 0.05) to test our hypothesis that jaw-muscle morphology is related to jaw-muscle activity across primates. Because our small samples limit statistical power, we also indicate trends tending toward significance where 0.05 < P < 0.10.
Masseter architectural variables are not correlated with the superficial masseter W/B ratio or the average W/B ratio across this primate sample [Table 3a; Fig. 3]. Based on the lack of association, the hypothesized primate-wide association is not apparent among masseter architectural variables and W/B ratios. However, closer examination of Fig. 3 suggests that after excluding platyrrhines the remaining primates in the sample may share a relationship between masseter EMG and both relative masseter weight and relative fiber length.
Table 3. Product–moment correlations among jaw-muscle architecture and EMG variables for the (a) superficial masseter and (b) temporalis
Product–moment correlations are followed by the P-value in parentheses.
Bolded correlations are significant at α = 0.05. Italicized correlations indicate P-values < 0.1 suggesting trends at the small sample sizes.
In contrast to the masseter data, relative temporalis architectural variables exhibit multiple significant inverse relationships with temporalis W/B ratios [Table 3b; Fig. 4]. Relative temporalis muscle weight shares a significant negative correlation (r = −0.75) with the posterior temporalis W/B ratio (Fig. 4a) and a negative correlation with the average of the W/B ratios (r = −0.66) that approaches significance (Table 3b). Relative fiber length is not related to either W/B ratio, although relative fiber length and the posterior temporalis W/B ratio approach a meaningful association. Finally, relative temporalis PCSA is strongly negatively correlated with the anterior temporalis W/B ratio (r = −0.82) and the average W/B ratio for superficial masseter and temporalis muscles (r = −0.85). The negative correlation between relative PCSA and the posterior temporalis (r = −0.71) does not reach statistical significance, but trends similarly. The strong negative association suggests that across primates relative increases in balancing-side temporalis recruitment are associated with relatively greater force-producing ability in the temporalis muscles.
Results for the temporalis muscle support the hypothesized relationship linking jaw-muscle architecture and activity during chewing across primates. The masseter, however, does not demonstrate the predicted primate-wide association among masseter architectural variables and W/B ratios. The mosaic pattern of results is not surprising, given that previous analyses of jaw-muscle architecture and EMGs have demonstrated clade-specific shifts in muscle morphology (Taylor and Vinyard, 2004, 2009; Taylor et al., 2009) and activity patterns (Hylander et al., 2000, 2004, 2005; Vinyard et al., 2007, 2008). Recently, Crompton et al. have demonstrated significant variation in jaw-muscle recruitment patterns in marsupial herbivores, both among marsupials and compared to placental mammals, suggesting similar patterns of variation and mosaicism likely exist outside of primates (Crompton et al., 2008a, b).
The contrasting patterns of relationships between the temporalis and masseter support previous work, suggesting these two muscles have evolved as distinct functional units in primates (Hylander et al., 2000, 2005; Vinyard et al., 2007, 2008). Based on the results of this study, this interpretation can be extended to include the evolution of jaw-muscle architecture and its relationships with recruitment patterns throughout the order. Beyond a certain level of evolutionary independence between the temporalis and masseter, the potentially distinct evolutionary histories of the two muscles may indicate that various primate groups exhibit differing levels and patterns of musculoskeletal integration in their masticatory apparatus (Ackermann and Cheverud, 2000; Vinyard, 2007). This interpretation requires further testing and should not be extended to broadly characterize the cranium based on a general conclusion of similarity in cranial integration in papionins (Cheverud, 1989), platyrrhines (Marriog and Cheverud, 2001), and hominoids (Ackermann, 2002).
Masseter Jaw-Muscle Architecture and EMG
The absence of a primate-wide relationship between masseter architecture and relative recruitment levels follows previous arguments that the masseter has experienced divergent evolution across primate clades, particularly in its function as a balancing-side muscle (Hylander et al., 2000; Ravosa et al., 2000; Vinyard et al., 2007, 2008). Based on its orientation in the skull, the masseter likely plays a relatively large role in generating side-to-side, or transverse, jaw movements during chewing and transversely oriented forces during the power stroke of mastication (Hylander, 1979; Ravosa et al., 2000). This observation is particularly evident when compared with the more vertically-aligned anterior temporalis of primates (Cachel, 1979; Ross, 1995). One of the trends observed in primate dental evolution highlights a shift toward a more transversely-oriented power stroke in anthropoids (Hiiemae and Kay, 1973; Kay and Hiiemae, 1974). Given its anatomical orientation, the masseter may exhibit relatively greater functional variation as an indicator of evolutionary change related to these shifts in dental form and function during mastication across primates. Increased functional variation in the masseter may also be related to potential functional tradeoffs with the medial pterygoid muscle. Possible functional redundancy between these muscles could provide both with greater degrees of evolutionary freedom. These non-mutually exclusive ideas largely remain conjecture until we can directly measure occlusal force directions across primates as well as increase the data on medial pterygoid architecture and EMG patterns to relate occlusal forces to muscle-activity patterns and jaw-muscle architecture.
The strongest evidence for the argument of clade-specific evolution of the superficial masseter in primates focuses on the derived firing pattern observed in platyrrhines. Previous work in owl monkeys (Hylander et al., 2000), callitrichids (Vinyard et al., 2007), and capuchins (Vinyard et al., unpublished data) documents a superficial masseter activity pattern where the working- and balancing-side muscles exhibit peak activity nearly simultaneously during the chewing cycle. This timing pattern differs from other primates and treeshrews, where the balancing-side superficial masseter tends to peak earlier than the working-side masseter (Hylander et al., 2000; Vinyard et al., 2007). Although largely speculatory without additional data on muscle mechanics during chewing, the EMG data suggest that the platyrrhine masseter may have diverged in its functional role during chewing compared with other primates. By recruiting the balancing-side muscle later in the chewing cycle, platyrrhines may be incorporating their balancing-side masseter later in the power stroke to augment occlusal forces and/or movements, emphasizing this role over other functions such as closing the jaw. The observation that both relative masseter weight and fiber length may exhibit different relationships with the superficial masseter W/B ratio in platyrrhines versus other primates suggests that future investigations should compare jaw-muscle architecture in platyrrhine and nonplatyrrhine primates. Finally, it is worth noting that the pattern of association between relative masseter PCSA and the superficial masseter W/B ratio does not differ markedly between platyrrhines and other primates compared with relative fiber length and masseter weight. The relatively smaller difference in PCSA may indicate a functional convergence and suggest the possibility of multiple architectural solutions to meet similar functional demands during chewing in primates.
Temporalis Jaw-Muscle Architecture and EMG
Unlike the masseter, the temporalis demonstrates a significant association between relative muscle architecture and W/B ratios across primates (Table 3b; Fig. 4). The association between relative PCSA and temporalis W/B ratios suggests that as primates increase their relative recruitment of the balancing-side temporalis, they can generate relatively greater temporalis muscle force compared with species exhibiting lower levels of balancing-side activity. The strength of the relationship between relative temporalis PCSA and anterior temporalis W/B ratios supports a hypothesis that anterior temporalis form and activity are functionally constrained across primates. As one of the primary vertically-oriented jaw muscles recruited late in the power stroke as a balancing-side muscle (i.e., Triplet II), it is inferred that the relatively conserved functional role involves producing vertically-oriented occlusal forces during chewing (Hylander et al., 2005; Vinyard et al., 2007, 2008). Additional support for the functional constraint hypothesis is provided by the consistency in the timing of activation across primates (Hylander et al., 2005) and the relatively low variance among species for anterior temporalis EMG activity patterns compared with other jaw muscles (Vinyard et al., 2008). These results extend the hypothesis focused on conserved EMG patterns to include temporalis muscle architecture as part of this integrated complex. We are careful to point out that a functional constraint linked to being a primary vertically oriented muscle does not translate into behavioral constraints, given that the temporalis can play a significant role in multiple different behaviors across primates (Hylander and Johnson, 1985; Ross and Hylander, 2000). Additionally, functional constraint does not prohibit the temporalis from exhibiting marked levels of behavioral modulation within species during these different behaviors (Blanksma and van Eijden, 1990, 1995; Blanksma et al., 1997; Vinyard et al., 2008) nor does it preclude evolutionary change across species. The functional constraint hypothesis would argue, however, that this evolutionary change was dictated by selective factors focused on vertical force production during behaviors involving the masticatory apparatus.
We also observed that relatively larger temporalis weights are correlated with increased levels of balancing-side muscle recruitment (i.e., lower W/B ratios) (Table 3b, Fig. 4a). It is possible that allometric changes in temporalis weight underlie this inverse correlation across primates as this shape variable does not adjust for size-correlated changes in shape (Jungers, 1995). Relative temporalis weight, however, is not correlated with condyle-M1 length (P > 0.05). Furthermore, previous allometric studies suggest that primate jaw-muscle weights scale close to isometry relative to body mass (Cachel, 1984; Perry and Wall, 2008; Ross et al., 2008). The functional consequences of this association suggest that the relationship between relatively increased temporalis PCSA and balancing-side recruitment of the temporalis muscle in primates is associated at least in part with relatively larger temporalis weight. Previously, Hylander et al. (2005) noted that suborder differences in temporalis W/B ratios are related to suborder differences in temporalis weight (e.g., Cachel, 1979). The positive association between relative temporalis weight and W/B ratios supports and extends this conclusion to suggest that anthropoids and strepsirrhines may inhabit the two ends of a continuum relating temporalis weight and recruitment patterns.
Despite the small sample size, our analysis demonstrates that jaw-muscle architecture is an integrated component of the primate masticatory system. Jaw-muscle morphology likely has evolved with changes in muscle-activity patterns, suggesting a link between jaw-muscle form and physiological performance during chewing. Temporalis architecture was strongly correlated with anterior temporalis W/B ratios across primates, suggesting a potential functional linkage relating muscle architecture to vertical force production during chewing. Alternatively, masseter architecture was not correlated with superficial masseter W/B ratios during chewing, highlighting the possibility of evolutionary changes related to diverging functions during mastication in this muscle across primate clades. The observed differences in these relationships between the masseter and temporalis highlight the potential independent evolution of jaw-muscle form and function in primates. The mosaic relationship across muscles underscores the morphological and functional complexity of the masticatory apparatus in primates.
The authors thank Qian Wang, Valerie DeLeon, Timothy Smith, and Jason Organ for the invitation to participate in this special issue of the Anatomical Record. The authors thank Carel van Schaik, Anthropological Institute and Museum (Zürich), Richard W. Thorington and Linda Gordon (USNM), Elizabeth Curran (NEPRC), Amanda Trainor (WPRC), and Susan Gibson (SMBRR) for supplying the cadaveric specimens used in the analysis of muscle fiber architecture. They also thank A. Doherty, N. Robl, C. Thompson, B. Armfield, and A. Mork for assisting in unpublished EMG data collection from Cebus apella, and C. Eng, E. Han, and K. Jones for assisting in jaw-muscle fiber architecture data collection. The veterinary staff at NEOUCOM provided animal care for the capuchins.