The Human Tongue Slows Down to Speak: Muscle Fibers of the Human Tongue


  • Ira Sanders,

    1. Alice and David Jurist Institute for Biomedical Research, Hackensack University Medical Center, Hackensack, New Jersey
    Search for more papers by this author
  • Liancai Mu,

    Corresponding author
    • Upper Airway Research Laboratory, Department of Research, Hackensack University Medical Center, Hackensack, New Jersey
    Search for more papers by this author
  • Asif Amirali,

    1. Alice and David Jurist Institute for Biomedical Research, Hackensack University Medical Center, Hackensack, New Jersey
    Search for more papers by this author
  • Hungxi Su,

    1. Upper Airway Research Laboratory, Department of Research, Hackensack University Medical Center, Hackensack, New Jersey
    Search for more papers by this author
  • Stanislaw Sobotka

    1. Upper Airway Research Laboratory, Department of Research, Hackensack University Medical Center, Hackensack, New Jersey
    2. Department of Neurosurgery, Mount Sinai School of Medicine, New York, New York
    Search for more papers by this author

Correspondence to: Liancai Mu, M.D., Ph.D., Upper Airway Research Laboratory, Department of Research, Hackensack University Medical Center, Hackensack, NJ 07601. E-mail:


Little is known about the specializations of human tongue muscles. In this study, myofibrillar adenosine triphosphatase (mATPase) histochemical staining was used to study the percentage and distribution of slow twitch muscle fibers (slow MFs) within tongue muscles of four neurologically normal human adults and specimens from a 2-year-old human, a newborn human, an adult with idiopathic Parkinson's disease (IPD), and a macaque monkey. The average percentage of slow MFs in adult and the 2-year-old muscle specimens was 54%, the IPD was 45%, while the neonatal human (32%) and macaque monkey (28%) had markedly fewer slow MFs. In contrast, the tongue muscles of the rat and cat have been reported to have no slow MFs. There was a marked spatial gradient in the distribution of slow MFs with the highest percentages found medially and posteriorly. Normal adult tongue muscles were found to have a variety of uniquely specialized features including MF-type grouping (usually found in neuromuscular disorders), large amounts of loose connective tissue, and short branching MFs. In summary, normal adult human tongue muscles have by far the highest proportion of slow MFs of any mammalian tongue studied to date. Moreover, adult human tongue muscles have multiple unique anatomic features. As the tongue shape changes that are seen during speech articulation are unique to humans, we hypothesize that the large proportion of slow MFs and the anatomical specializations observed in the adult human tongue have evolved to perform these movements. Anat Rec, 296:1615–1627, 2013. © 2013 Wiley Periodicals, Inc.

Tongue movements during human speech are among the most complex motor activities and appear to be unique among mammals. Speaking each vowel and consonant requires accuracy in both the spatial positioning of the tongue as well as the shape of its dorsal surface (Stone and Lundberg, 1996). These changes in tongue position and shape are rapidly performed yet precisely timed and coordinated with other articulators. The capability to perform these movements is related to well-documented evolutionary changes in the human central nervous system. These include a specialized speech motor control center in the cerebral cortex as well as direct cortical connections to the motoneurons controlling the tongue (Broca, 1861; Kuypers, 1958). Less understood is whether the human tongue muscles have specialized in order to execute these movements.

It has long been known that the human tongue has a unique curved shape with differences in the size and position of some muscles compared to other mammals (Abd-El-Malek, 1938a; Miyawaki, 1974); however, the significance of these differences is unknown. Recent studies have shown that the tongue muscles are composed of smaller neuromuscular compartments with separate nerve branches (Mu and Sanders, 1999, 2000, 2010; Sokoloff, 2000). This organization allows for the independent movement of highly localized parts of tongue muscles. The complex interweaving of human tongue muscles makes it technically difficult to reliably isolate and identify intrinsic tongue muscles for physiological or histological studies. As a result, there are few studies of any human intrinsic tongue muscle by electromyographic recording or by electrical stimulation. At present, the functions of most of the individual tongue muscles are presumed to be related largely to their anatomical arrangement as well as extrapolation of data obtained from experiments in animals (Abd-El-Malek, 1938b; Bennett and Hutchinson, 1946; Hellstrand, 1981; Gilliam and Goldberg, 1995; McClung and Goldberg, 1999).

Although the contractile properties of human tongue muscles are not known this information can be studied indirectly by studying MFs using histochemical stains. Specifically, MFs can be categorized into either slow (type I) or fast (type II) contracting types by their reaction to myofibrillar ATPase staining (Padykula, 1955; Barany, 1967). These two types of MFs often have different motor unit sizes, force production, contractile speeds, and resistance to fatigue. Fast fibers are primarily adapted for rapid contractions, generate much greater force, and most rely on glycolysis for their energy. In contrast, slow MFs are resistant to fatigue and generally involved in activities requiring precise control of low forces.

ATPase typing of tongue muscles in the mouse, rat and cat show that they are composed of 100% fast MFs (Prigozy et al., 1997; Hellstrand, 1980; Sokoloff, 2000). In contrast, the intrinsic muscles of the macaque tongue contain 25% slow MFs (DePaul, 1996). Further, the distribution of these MFs is not uniform. In the macaque, the anterior end of the tongue the muscles are composed of nearly 100% fast MFs, while the posterior end contains an equal number of fast and slow MFs. A gradient was also seen regarding MF diameters with small MFs in the anterior tongue and large diameter MFs posteriorly. A study sampling three human tongue muscles reported that 40% of the MFs were slow MFs. Moreover, there is a similar gradient of both slow MFs and MF diameters (Stahl, 2003).

The hypothesis of this study is that the slow MFs colocalize with other specializations and are related, either directly or indirectly, with speech capability. In this study, all tongue muscles from four neurologically normal adult humans (Normals) were examined by ATPase staining. In addition, a variety of specimens (Experimentals) were studied for a comparative, developmental, and pathological comparison to the Normals. Specifically, Experimental specimens consisted of a nonhuman primate (Macaque), a newborn human (Neonate), a 2-year-old human (Infant), and an adult human with idiopathic Parkinson's disease (IPD). These Experimental specimens were partly chosen for their different speech capabilities. The differences between Normal and Experimental specimens as well as between the various Experimental specimens were expected to provide insight on tongue structure–function relationships.


Tissue Source

All specimens were obtained from autopsies performed at Mount Sinai Medical Center. As the research involved only autopsy specimens, it was exempt from consent requirements. The human tongues (four normal adults, a 2-years-old, a neonatal, and an adult with IPD) were obtained from autopsies without known neuromuscular disorders, except for the case with IPD. The Normals were from a 56 year old Caucasian female; 52 year old Hispanic female; 60 year old African-American male; and the fourth had no information. The Experimentals were from a 2 year old African-American male, a 3 day old Oriental male, and the IPD was from a 57 year old Caucasian with documented speech and swallowing impairment. All specimens were obtained within 24 hr of death; this postmortem interval does not hamper reliable histochemical analysis of autopsied tissues (Eriksson et al., 1980; Eriksson and Thornell, 1983). A monkey tongue was obtained from an animal being euthanized by other investigators immediately after termination of their study.

Tissue Preparation

The length, width, and wet weight of each adult human specimen was recorded. Then, each tongue was bisected in the midline to divide the tongue into equal and symmetrical halves. Only one side was used for histochemistry, the other side was used for other experimental protocols. Previous studies by Stone and Lundberg (1996) have demonstrated significant changes in the human tongue shape during phonation (Fig. 1A). These findings suggest that tongue muscles in the blade, body, and base can act independently during phonation and speech articulation. In this study, regional differences in fiber-type distribution within the adult human tongue were examined. To this end, each hemitongue was thus divided into three blocks: the blade, body, and base (Fig. 1B). The blade consisted of a ∼2.0-cm segment that extends from the tip of the tongue to the frenulum. The body was a ∼3.0-cm segment that extended from the frenulum to the sulcus terminalis, the line of circumvallate papillae. The base was the remaining ∼2.5-cm segment of tongue. The blocks were frozen by immersion in isopentane cooled to −65°C by dry ice. The tissue blocks were mounted for cryosection and were trimmed until the middle of the block, and then serial frontal sections (Fig. 1C) were cut (10 µm thick) on a cryostat (Reichert-Jung 1800; Mannheim, Germany) at −25°C. The sections were stored at −80°C until staining procedures were performed.

Figure 1.

(A) Reconstructed views of the human tongue shape during phonation of two different vowels. Note the markedly different shape of the tongue. (From an ultrasound study of tongue shape during speech by Stone and Lundberg with permission, J Acoust Soci Am 99:3728–3737, 1996). (B) An adult human tongue. The division between base, body, and blade are marked by black lines. A white broken line down the midline shows how the tongue was bisected. White rectangles show general area where sections were obtained for analysis. (C) Section from anterior tongue body stained with hematoxylin and eosin to show muscle structure. (D) Frontal section from the body of a human tongue 55 mm posterior to the tip of the tongue. At this level all muscles can be seen as separate entities. The outlines of the muscles do not include areas where muscles overlap as this would make the image too complex. For example the genioglossus (GG) passes through the transverse/vertical (T/V) and superior longitudinal (SL) muscles. The internal structure of the inferior longitudinal (IL) muscle is not compact but heterogeneous and poorly demarcated. Lateral to the IL lies the hyoglossus (HG) and styloglossus (SG). These two muscles are thin planar muscles that overlap through most of their course along the lateral aspect of the tongue. Above the previously described muscles is the T/V muscle group. At the superior aspect of the tongue, the SL is seen clearly. It is well demarcated and thick near the midline. The vertical-striated appearance of the SL reflects the many thin fascicles of the SL between which the terminal muscle fibers of the GG and vertical muscle pass to insert into mucosa.

Myofibrillar ATPase Staining

Tongue sections were preincubated at pH 4.2 or 4.3 and then stained for ATPase (Guth and Samaha, 1969, 1970). At this pH, slow MFs (type I) stained dark while the fast MFs (type II) stained light, with a few MFs staining gray. Control sections were done for each specimen at alkaline preincubation at pH 10.6, and these demonstrated reversal of the acidic staining pattern. Further subcategories of type II fibers can be found in most muscles by staining at other pH. However, early attempts at staining human tongue muscle between pH of 4.4 and 4.8 always resulted in uninterpretable results with many intermediate-staining fibers. Therefore, further subcategorization of tongue MFs was not performed.

The stained muscle sections were examined under a Zeiss photomicroscope (Axiophot-2; Carl Zeiss, Goettingen, Germany) and photographed using a digital camera (Spot-32; Diagnostic Instruments, Keene, NH) that was attached to the photomicroscope and connected to a personal computer with image analysis software. All slow type I (dark staining) and fast type II (light staining) fibers were manually identified and marked using a computerized image analysis system (SigmaScan; Jandel Scientific, San Rafael, CA). Counts were made from all intrinsic and extrinsic muscles in all three segments (blade, body, and base) of the tongue. Relative proportions of slow and fast twitch fiber types in a given segment for each muscle were computed. In addition, as the superior longitudinal (SL) muscle spans the entire width of the tongue, it was separated into a medial and a lateral half which were counted separately.


The critical step in analysis was the proper identification of the different muscles. The complex interweaving of the human tongue muscles makes the identification of individual muscles extremely difficult in histological sections. Prior publications on the internal anatomy of the human tongue (Abd-El Malek 1938a, Miyawaki, 1974) have limited value when examining actual histological sections. In earlier work, we used data from the Visible Human Project to study the complex anatomy of human tongue muscles (Fig. 1D) and constructed a three-dimensional model to aid in identifying each muscle (Figs. 2 and 3; Sanders and Mu, 2013).

Figure 2.

Three-dimensional model of human tongue demonstrating intrinsic (A–F) and extrinsic (G–I) muscles. (A) The entire 3-dimensional model showing the tongue and mandible. (B) The mandible is removed to show the tongue by itself. (C) The surface of the tongue is made translucent so that the location of the intrinsic muscles can be appreciated relative to the surface. (D) The surface is made transparent to show the intrinsic muscles more clearly. Note that the intrinsic muscles are compactly layered; on the top and bottom are the superior (SL) and inferior (IL) longitudinal muscles with the transverse and vertical muscles (T/V) in between. (E) The SL is made translucent to show the other intrinsic muscles. The T/V muscles overlap throughout the tongue and are rendered as a single object in this model. (F) The model is rotated to the inferior view and the translucent mandible is replaced. (G) Superior view of the extrinsic muscles, surface is translucent. Note that the hyoglossus (HG) and styloglossus (SG) muscles originate posterior to the tongue and course anteriorly along the lateral aspect of the tongue. In contrast, the genioglossus (GG) originates anteriorly from the inner surface of the mandible and courses posteriorly at various angles. (H) Superior view, surface is removed. (I) Frontal view of extrinsic muscles. The HG and SG form the lateral aspect of the tongue, while the GG forms the center of the tongue. Note that the muscles are all rendered as separate objects for the sake of clarity, but in reality, there is extensive overlap. The most important example is the GG that becomes the vertical muscle in the medial third of the tongue. Therefore, the GG actually passes between sheets of the transverse muscle and then between SL fascicles to insert into the dorsal mucosa of the tongue.

Counts of slow and fast MFs were converted into percentages and the variability of percentages was expressed by the standard deviation. Comparison of percentages was made between the blade and body, body and base, and blade and base, and these were evaluated by student's t-test, using P < 0.05 as the level of significance.

Experimental Group

The internal muscle anatomy of the Experimental specimens as seen in histologic sections was substantially different from normal adults, therefore it was not possible to identify all individual muscles from sections of the tongue base and blade. Therefore, the only area that allowed proper identification of individual muscles in all experimental specimens was the tongue body, specifically its medial aspect. Even then, the extrinsic muscles and the intrinsic inferior longitudinal (IL) muscle could not be reliably identified in all specimens and these were not analyzed. Therefore, for purposes of comparison, only counts were made from three intrinsic muscles from the medial aspect of the sections from the tongue body, the SL, vertical (VM), and transverse muscle (TM). Therefore, the reported total muscle fiber counts were obtained from these three intrinsic muscles. To allow a valid comparison between the experimental group and the Normals, the four normal adult human specimens were recounted using the same protocol.


A review of basic tongue anatomy is necessary before describing specific findings for each muscle. Tongue muscles of most mammals are divided into two main groups: the extrinsic and the intrinsic (Sonntag, 1925). The intrinsic muscles originate and insert within the tongue and have no bony attachments. These muscles are presumed to largely function as muscular hydrostats, altering the shape of the tongue. In contrast, the extrinsic muscles have one attachment to the mandible (GG), styloid process (SG) or hyoid bone (HG) while the other end inserts within the tongue. These muscles act more to change the position of the tongue within the oral cavity. However, the parts of the extrinsic muscles that enter into the tongue may be considered intrinsic muscles which also make significant contributions to shaping of the tongue. Individual muscles are shown and described in Figs. 2 and 3.

Figure 3.

Muscles of the human tongue. (A) The superior longitudinal (SL) muscle spans the length of the tongue just beneath the mucosa of its superior surface, it is believed to shorten the tongue and curve the tip superiorly. The SL is the only unpaired muscle of the tongue. (B) The inferior longitudinal (IL) muscles span the length of the tongue just above the mucosa of its inferior surface, they are believed to shorten the tongue and curve the tip inferiorly. (C) The transverse and vertical (T/V) muscle group. Transverse muscles connect the medial septum to the lateral aspect of the tongue; they are believed to narrow the tongue. The vertical muscles connect the inferior surface to the superior surface; they are believed to flatten the tongue. (D) The hyoglossus (HG) muscles originate from the hyoid bone and also insert into the inferior-lateral margins of the tongue, they are believed to cause tongue retraction and depression. (E) The styloglossus (SG) muscles originate from the styloid processes and also insert along the inferior-lateral margins of the tongue, they are believed to cause tongue retraction and elevation. (F) The genioglossus (GG) muscles are midline muscles which originate from the posterior surface of the mandible. The GG muscles are believed to cause tongue protrusion with a secondary effect of drawing the midline of the tongue inferiorly, such as seen in the tongue on the right in Fig. 1A. As mentioned in the legend for Fig. 2, the muscles are all rendered as separate objects for the sake of clarity, but in reality there is extensive overlap. One example is that the GG muscle becomes the vertical muscle in the medial third of the tongue. Therefore, the GG muscle actually passes between sheets of the transverse muscle and then between SL fascicles to insert into the dorsal surface of the tongue. The other example is that the T/V muscles are rendered as a single object.

Morphological Measurements

For the four normal adult human tongue specimens, the average wet weight was measured to be 70 g (range: 50.8–93.4). The average length was 9.0 cm (range: 7.5–10.5). The average width was 2.9 cm (range: 2.5–3.0) for the anterior tongue, 6.4 cm (range: 6.0–6.6) for the midtongue, and 5.0 cm (range: 4.0–6.0) for the posterior tongue. The depth was measured using the ATPase stained sections (n = 5). The average tongue depths in the medial and lateral aspects of the tongue were 14/11 mm for the blade, 17/12 mm for the body, and 17/11 mm for the base.

ATPase of Normal Adult Tongue Intrinsic Muscles

A total of 29,611 MFs were counted in the four normal adult tongue specimens. Histochemically identified slow and fast MFs and their distribution in the adult tongue muscles are shown in Fig. 4. Overall, the tongue body (57% slow MFs) and the tongue base (58% slow MFs) were slower than the tongue blade (46% slow MFs) (Table 1; P < 0.01).

Figure 4.

ATPase of individual human tongue muscles. This figure contains examples of adult human tongue muscles stained for ATPase (pH 4.2). The figure in the center of the lower row is a photograph of a 2-mm frontal section through the body a human tongue. Unlike histological sections, in which individual tongue muscles are difficult to identify, the location of each muscle can be seen in this thick section. Counterclockwise from the right side: The genioglossus (GG) muscle is composed of a horizontal part, which is the muscle cut in cross section at the bottom of the black circle and an oblique part, which is the muscle cut along its length at the top of the circle. The ATPase section is from the horizontal part of the GG which has the highest percentage of slow MFs for the whole muscle. The inferior longitudinal (IL) muscle is the most difficult intrinsic muscle to identify in fontal sections. The IL had the highest percentage of slow MFs of all tongue muscles. It also was notable for not having an anterior to posterior gradient, but did show fiber-type grouping. The transverse and vertical (T/V) muscles are both displayed in same photograph. This section was cut in the sagittal plane such that the center column of MFs cut in cross section are the transverse, while the MFs cut lengthwise on both sides of the transverse are vertical MFs. The transverse had the second highest percentage of slow MFs, while the vertical muscle had by far the lowest percentage. Medial and lateral fascicles of the superior longitudinal muscle (SL) are displayed next. There were large differences between different parts of the tongue. MF type grouping is seen in both photographs of the SL. The stylglossus (SG) and hyoglossus (HG) are shown together as they had similar histological appearance and similar percentages of slow MFs.

Table 1. Percentage of slow muscle fibers in adult human tongue muscles

The Superior Longitudinal Muscle

The SL is the only unpaired muscle of the tongue. Its gross appearance is that of a sheet covering the dorsal surface of the tongue just beneath the mucosa. In frontal sections, it can be seen that the SL is actually composed of dozens of fascicles. The SL is present and easily identifiable from frontal sections from any part of the tongue. Because of this, and its importance in shaping the dorsal surface of the tongue, the anatomy and histology of this muscle was examined in greater detail than the others. The thickness of the SL was measured using the ATPase stained sections. On average, the thickness of the medial SL was 1.5 mm for the blade, 3.0 mm for the body, and 4.0 mm for the base. In contrast, the thickness of the lateral SL was ∼1.0, 2.0, and 2.8 mm, respectively.

The distribution patterns of the fiber types in the SL are shown in Fig. 5. There was a gradual anterior to posterior increase in the proportion of slow (type I) MF (blade, 45% type I; body, 58% type I; base, 63% type I) (Table 1). However, this anterior-to-posterior gradient was limited to the medial half of the SL (blade 45%, body 69%, base 74%; P < 0.01) as there was no significant difference between these regions in the lateral SL. Further, a medial-lateral comparison showed that the medial SL in tongue body and base had a higher percentage of slow MFs (63%) than the lateral SL (48%) (P < 0.001). In contrast, there were no significant differences between the medial and lateral portions of SL in the tongue blade.

Figure 5.

Photomicrographs of ATPase stained cross sections (pH 4.2) from the left human superior longitudinal (SL) muscle. Sections were taken from blade, body, and base, from left to right, respectively. Top row is the medial side and bottom row is the lateral side. Note that the slow MFs (black) progressively increased in the medial SL (top column) from the anterior to the posterior regions. In contrast, slow MFs were relatively more evenly distributed in the lateral SL. Also, note the variation in MF diameter from blade to base. Finally, fiber-type grouping is seen in all parts of the SL. An extreme example of fiber-type grouping is seen in the medial tongue base where the arrow points to a muscle fascicle that is almost entirely composed of slow MFs, yet surrounded largely by the lighter staining fast MFs. Bar = 100 µm.

The arrangement of fast and slow MFs in the SL was also notable for fiber-type grouping (Fig. 5). Specifically, the slow MFs were found to be concentrated together in clusters. In normal limb muscle, the usual distribution pattern is that MFs of different types are arranged in a random mosaic pattern. Specifically, the number of slow MFs surrounding any given slow MFs was no greater than the overall percentage of slow MFs in that muscle. When fiber type grouping is found in limb muscle it usually occurs as a result of neurological disease or injury causing denervation of MFs, followed by reinnervation of surrounding MFs by surviving motor neurons (Karpati and Engel, 1968). In this study, it is notable that in contrast to Normals, the fiber-type grouping was absent or much decreased in all Experimental specimens, including the IPD specimen. Another unusual attribute of the SL was that its MFs were highly variable in diameter and length, with a large percentage of its MFs only a few millimeters in length. Moreover, a high percentage of these MFs split into two MFs that either terminated or joined to other MFs (Fig. 6).

Figure 6.

Muscle fibers of the superior longitudinal (SL) muscle. Twelve 15-μm serial sections spanning 2 mm from a fascicle of the SL muscle (sequence numbered 1–12). The upper ovals in each photo encircle a group of three MFs that split into four MFs (photo 2) then five MFs (photo 4), before recombining or terminating back to three MFs (photo 11). Arrows point to MFs that will split in the next frame or have arisen de novo. The lower ovals mark another group of MFs that go through similar changes. These examples are representative of the whole SL.

The SL appeared to have three distinct regions with different percentages of slow MFs, MF size, degree of fiber-type grouping and amount of loose connective tissue. The tongue blade had the least amount of slow MFs (45%), had the smallest diameter MFs, the least fiber-type grouping, and the loosest connective tissue. The lateral parts of the tongue body and base had comparatively greater number of slow MFs (50%), and had an intermediate diameter MFs, fiber-type grouping, and connective tissue content. Finally, the medial side of the body and base had the highest number of slow MFs (72%), the largest diameter MFs, the most fiber-type grouping, and the least connective tissue. Another unusual attribute of the SL was that its MFs were also highly variable in length, with a large percentage of its MFs only a few millimeters in length. Moreover a high percentage of these MFs split into two MFs that either terminated or joined to other MFs (Fig. 6).

The Inferior Longitudinal Muscle

The IL originates from the tongue base and passes anteriorly to merge with other tongue muscles and terminate at the tongue tip. The IL is the most difficult intrinsic muscle to identify as it often is divided into two or more fascicles that are equal in size and thus can be difficult to distinguish from the surrounding fascicles of the GG, HG, and SG.

The IL was found to be the slowest muscle (61%) of the tongue, with no statistically significant difference in the percentage of slow MFs in the blade (59%), body (63%), or base (62%) (Table 1; P > 0.05). Therefore, the IL does not exhibit the anterior-to-posterior gradient seen in the SL and other tongue muscles.

The Transverse Muscle

The TM is oriented horizontally in frontal sections. It originates from the median septum in tongue midline through the length of the tongue and passes horizontally towards lateral tongue. The TM is composed of many separate thin sheets of fascicles that alternate with sheets of the VM. There was no difference in the proportion of slow MFs in the blade (59%), body (60%), or base (60%) (Table 1; P > 0.05). Therefore, no anterior-to-posterior gradient of slow MFs was seen in this muscle.

The Vertical Muscle

The VM is oriented vertically in frontal section. It is composed of separate sheets of fascicles that alternate with the TM. It is notable that in the medial third of the tongue the VM appears to be a continuation of genioglossus MFs. However, the VM has its own motor endplates and is innervated by different terminal nerve branches. In the lateral two thirds of the tongue the VM originates from connective tissue overlying the IL and courses superiorly to pass between fascicles of the SL to insert into the connective tissue of the dorsal surface of the tongue. Contraction of the VM flattens the tongue. The VM was the fastest muscle in the tongue with only 41% slow MFs (Table 1; P < 0.001).

ATPase of Normal Adult Tongue Extrinsic Muscles

The styoglossus (SG) and hyoglossus (HG) muscles contained 48% and 54% slow MFs, respectively, while the genioglossus (GG) muscle contained 56% slow MFs (Table 1). There was no significant difference in the percentage of slow MFs between the tongue extrinsic muscles (P > 0.05). However, regional differences in the fiber-type distribution were present in each of the adult tongue extrinsic muscles (Table 1). On average, the three extrinsic tongue muscles (GG, HG and SG) in the blade contained less slow type I fibers (42%) than in the body (58%) and base (58%) (P < 0.01). The SG was notable for having the greatest regional variation in slow MF percentage: blade (35%); body (61%); base (48%); as well as a significantly lower overall proportion of slow MF (48%) (P < 0.05). The reason for this is unknown; however, it may be related to the SG's role in rapidly raising the tongue base during swallowing.


Figure 7 shows the distribution of the slow MFs in the ATPase-stained SL muscle sections from a normal human adult, a patient with IPD, a human newborn, and a macaque monkey. Figure 8 shows the differences in the percentage of slow MFs in the intrinsic tongue muscles between Normals and Experimentals. Specifically, intrinsic tongue muscles in the normal human adult and 2-year-old had markedly more slow MFs (52% and 54%, respectively) than those in the subject with IPD (45%), neonate (32%), and monkey (28%; P < 0.05).

Figure 7.

Slow muscle fibers in adult and experimental tongues. Photomicrographs of ATPase stained cross sections (pH 4.2) of the superior longitudinal (SL) muscles from a normal human adults (A), a patient with idiopathic Parkinson's disease (B), human newborn (C), and macaque monkey (D). The adult sample has markedly more slow MFs (black) than the other samples. 100×.

Figure 8.

A bar graph showing the differences in the proportions of slow muscle fibers (MFs) between the normal human adult and experimental samples. Note that normal human adult and human 2-year-old samples contained more slow MFs as compared to the other samples.

For purposes of comparing the Normal and Experimental specimen, a separate protocol was implemented for analysis. This was done due to the limitation in identifying and counting the other muscles in the Experimental specimen. It is important to note that the counts reported here are total counts of only the medial aspect of three of the four intrinsic muscles and none of the extrinsic muscles. This allowed a direct comparison to be made between Normal and Experimental specimen. This is reflected by differences in total slow MF counts in the intrinsic muscles of normal adults using this protocol (52%) compared to the comprehensive counting of the same muscles reported above (54%). In addition, as these Experimental specimens are extremely rare only a single specimen was obtained and counted in each category. Therefore, a statistical comparison between Normal and Experimental specimen is not possible.


The most notable finding in this study was that 54% of adult human tongue MFs are slow MFs, which is substantially more than prior sampling studies (40%) (Stahl et al, 2003), more than double that reported for the macaque monkey (26%) (DePaul and Abbs, 1996), and quite different from nonprimate mammals whose tongue muscles are 100% fast twitch. Slow MFs are associated with fine control of movement and suggest that the human tongue has become specialized for this purpose. Other specialized features of the adult tongue muscles not present or much decreased in Experimentals was the presence of fiber-type grouping of MFs, variability in MF size, splitting and interconnecting of MFs and large amounts of loose connective tissue. Fiber-type grouping is especially notable as it is considered pathological in limb muscles. The greatest difference between human and nonhuman tongue movements is shape change that occurs during speech. Therefore, a reasonable and parsimonious explanation is that these specializations are related to the tongue shape changes seen during speech articulation.

In Normals among the intrinsic tongue muscles, the IL, TM, and SL muscles contained significantly more slow MFs (61%, 60%, and 55%, respectively) than the VM (41%). The extrinsic muscles SG (48%), HG (54%), and GG (56%) were similar to most of the intrinsic muscles.

In the SL muscle, the slow MFs were arranged in a spatial gradient with higher percentages found in the posterior tongue compared to the anterior tongue and the medial aspect of the tongue compared to the lateral aspect. A similar gradient was found with MF size and diameter. Notably, slow and fast MFs were of similar sizes in different regions of the tongue. This is different from limb muscles where slow MFs are usually smaller than fast MFs. The SL was examined more closely than other muscles as it was most easily identified in sections from any part of the tongue. The SL appeared to have three distinct regions with different percentages of slow MFs, MF size, degree of fiber-type grouping, and amount of loose connective tissue. The tongue blade had the least slow MFs (46%), smallest MFs, the least fiber-type grouping and the loosest connective tissue. The lateral side of the body and base had comparatively greater number of slow MFs (50%), intermediate MF size, fiber-type grouping, and connective tissue content. Finally, the medial side of the body and base had the highest number of slow MFs (72%), the largest MFs, the most fiber-type grouping, and the least connective tissue.

The reason for these regional differences is not known, however, the small MFs and the large amount of loose connective tissue appear to correlate with the degree to which these areas can change their shapes. For example, the tongue blade is the most mobile part of the tongue, it can elongate, shorten, and ventroflex and dorsiflex more than 90°. In contrast, very little shape change occurs in the tongue base. The difference between these areas during performance of a speech task can be seen in Fig. 1A. Therefore, the small MFs and the large amount of connective tissue seen in the tongue blade may reflect the ability of this area to deform extensively. Surprisingly, the tongue blade had the least amount of slow MFs, as it would appear to perform very delicate tasks.

Part of the explanation for the spatial distribution of slow MFs may lay in the peculiar muscular hydrostat biomechanics of the tongue (Kier and Smith, 1985, 1989). Without any bony skeleton, the muscles themselves must provide the scaffolding upon which they can mechanically interact. Movement of the lateral aspect of the tongue requires support from the medial aspect. Similarly, movement of more distal segments of the tongue depends on support from the tongue base. Recently, electromyographic studies in humans documented the supportive role that the genioglossus muscle plays during tongue protrusion (Pittman and Bailey, 2009).This support function probably requires the tonic contraction that is performed by slow MFs.

Tongue Function in Nonprimate Mammals

The MF types within tongue muscles of the rat and cat have been studied (Hellstrand, 1980). In these animals, the intrinsic tongue muscles are entirely composed of fast MFs, and this is reflected in their tongue movements. Electrical stimulation of the lateral division of the XII nerve in the rat causes tongue retrusion in 12 msec, one of the fastest contraction times of any mammalian muscle (Gilliam and Goldberg, 1995). Similarly, stimulation of the styloglossus, genioglossus, and longitudinal muscles individually all show contraction times between 12 and 14 msec (McClung and Goldberg, 1999). Cat tongue muscles have comparable contraction speeds; stimulation on the surface of the cat tongue causes dimpling with a contraction time of 22 msec (Hellstrand, 1981). The data in nonhuman mammals suggest that the primary attribute of the intrinsic muscles is speed. This may be related to the use of the tongue for rapid extraoral activities such as drinking fluids by lapping. During lapping, the tongue is rapidly protruded from the oral cavity, then dorsiflexed and retruded (Thexton and McGarrick, 1988).

In comparison to nonhuman mammals, the shape changes seen during speech do not require fast movement of the tongue over long distances but are largely shape changes caused by internal rearrangement of muscle groups relative to each other. Specifically, the tongue coarticulates with itself; the blade of the tongue may be in position to articulate one sound while the tongue base moves to anticipate the next sound. The tongue movements during coarticulation overlap both spatially and temporally so the need for rigid support conflicts with internal movement.

Tongue Function in Nonhuman Primates

Anatomical and physiological evidence suggests suggest that the monkey has evolved increased central nervous control over tongue function. Primates have direct cortical projections to the hypoglossal nucleus, suggesting increased volitional control of the tongue. Within the tongue, the intrinsic muscles contain significant numbers of slow MFs and muscle spindles (Cooper, 1953), both of which are absent from the tongues of nonprimates. Within the tongue of the monkey, the VM and TM have the highest percentages of slow MFs as well as the greatest numbers of muscle spindles. Contraction of both the TM and the VM thins and elongates the tongue, and the equal levels of slow MFs and muscle spindles suggest that elongation is an important and finely controlled tongue action in the monkey.

The exact purpose of these specializations is unknown. Monkeys do not have the ability to speak in any way comparable to humans. However, certain species have been shown to use vocalization extensively in social interactions. These vocalizations include primitive versions of consonants and fricatives, sounds that require tongue participation (Richman, 1987).

During mastication and swallowing, subtle differences have been noted which differentiate macaques from nonprimates (German et al., 1992). During the chewing cycle, nonprimates show a strong linkage between their jaw and tongue position with swallowing occurring at a specific time in the cycle. In contrast, the tongue movements of the macaque are much more independent and swallowing can occur at any time during the cycle. In addition, while chewing macaques lengthen and shorten different parts of their midtongue independently. These results suggest that slow MFs in macaques may be related more to increased control of mastication and swallowing then vocalization.

Specializations for Human Tongue Shaping During Speech

Tongue movements during speech involve the articulation of either vowels or consonants, and these require movements of the tongue body as well as changes in tongue shape. During vowels, the larynx is phonating into an open airway and the tongue is shaped to produce a local area of narrowing. The part of the tongue that is maximally displaced is called the maxD (Stone and Lundberg, 1996). According to the acoustical theory of speech, these localized constrictions “tune” the pharynx so as to produce the specific harmonics of each vowel (Fant, 1960). In contrast, when the tongue articulates consonants it forms a complete or nearly complete constriction of the vocal tract. The shape of the tongue is similar for vowels and consonants when their maxD is in the same area of the vocal tract, with the consonants requiring slightly farther displacement.

The biomechanics of tongue speech movement differs from other tongue movements as well as limb movements in that the tongue is not exerting significant external force but rather it is reshaping itself. In limb movements, muscles frequently show activity during the beginning of a movement that disappears as the movement nears completion. During speech, the opposite occurs; electromyography along the tongue's midline surface shows that background activity is relatively low in the tongue and that the activity rises until maximum activity is seen at the time when the tongue reaches its maxD (MacNeilage and Sholes, 1964). One interpretation is that the tongue “congeals” into its articulatory position and then relaxes to allow itself to rearrange for the next position.

The active shaping of the tongue dorsum is more important in humans than for other mammals. During the production of anterior consonants and vowels such as “ma” the dorsal surface of the tongue is curved along the sagittal plane (anterior to posterior) so that the tongue tip and blade are higher than the remainder of the tongue (dorsiflexion). In addition, the tongue assumes a curvature in the frontal plane (side to side) so that the lateral edges of the tongue blade are in contact with the hard palate. The exact amount of curvature of the tongue and contact with the hard palate varies for different vowels and consonants but the basic pattern is always present. Dorsiflexion of the tongue is the main action of the SL, and the TMs largely control side-to-side curvature, and it would be expected that these muscles would contain some indication of their specialized roles.

The SL appeared to be the most specialized of the human tongue muscles. Recent reports have documented the presence of multiple “en grappe” motor endplates on SL MFs (Slaughter et al., 2005; Mu and Sanders, 2010). This morphology suggests the presence of the extremely specialized slow tonic MF. Slow tonic MFs do not contract with a single twitch like skeletal MFs but have prolonged tonic contraction that can be finely controlled. The SL also has unusual features such as MF-type grouping, and small interlacing MFs. The SL has approximately double the number of muscle spindles than the TM while the IL and VMs have significantly less (Kubota et al., 1975). The high concentration of muscle spindles in the human SL is notable considering that nonhuman primates have a different pattern of muscle spindle distribution, with the SL containing fewer muscle spindles than the VM and TM (Bowman, 1968).

Although the percentage of slow MFs in the human SL was similar to that of the other intrinsic muscles, this reflects the percentage for the entire muscle. In examining the SL closer, it was seen that it appeared to be separable into superior and inferior layers, with the superior having much higher percentages of slow MF than the inferior. Separate quantitative counts of the two levels were not practical as there is no clear demarcation between them. However, the superior layer had the highest percentages of slow MFs in the tongue, in some places reaching 100% (Fig. 5). The possibility exists that there is a functional division in the SL with the superior layer primarily active during speech and the inferior layer during swallowing and other tongue activities.

Another notable difference in the tongue muscles of the adult human is the specialization of the TM relative to the VM. In monkeys, the TM and VM are similar in size, percentages of slow MFs, and number of muscle spindles. In contrast the human TM is much larger then the VM, has higher percentages of slow MFs (58%:41%) and has more muscle spindles (79:15). Similar to the SL the TM has an ill-defined layered structure with the highest percentages of slow MFs found superiorly. Finally, another specialized aspect of the TM that it shares with the SL is the presence of multiple en grappe motor endplates (Mu and Sanders, 2010) suggesting the presence of MFs with slow tonic morphology. Interestingly, the VM does not have this specialization. The increased specialization of the TM may be related to its control of the convex shaping of the tongue superior surface as seen in Fig. 1A and the curling of the sides of the tongue, two motions prominent during tongue articulation.

Developmental Specimens

At birth the human has a different repertoire of tongue movements than the adult; suckling and drinking are the two major activities. Vocalization in humans is limited to crying which does not require significant tongue articulation. Because of the different activities of the neonatal tongue, it would be expected that it show a different MF pattern than the adult. Overall, the percentage of slow MFs is much decreased and in at least one muscle, the SL, the relative distribution of the slow MFs is markedly different. A notable aspect of the amount and distribution of the slow MFs in the neonatal tongue are their similarity to that found in the macaque specimen. Although speculative, it is possible that this similarity is a case of ontogeny recapitulating phylogeny, as the developing human briefly passes through the activity patterns of its distant primate ancestors.

The 2-year-old specimen was remarkably different from the neonate and similar to that of the adult, both in total amounts of slow MFs and in their muscle distribution. Muscle type grouping as well as other specializations was also seen in this specimen. This is consistent with what is known about the morphology and function of the upper airway in 2-year olds. Although human are born with a flat tongue and upper airway, it soon begins morphing into the adult curved morphology. Much of this occurs during the second half of the first year of life; this period is marked by the onset of speech, albeit a primitive version known as babbling. By the second year of life, most normal 2-year olds can articulate most of the speech sounds used by adults.

Idiopathic Parkinson's Disease (IPD)

IPD is a progressive neurodegenerative disease. The disease is primarily a motor disorder that causes bradykinesia and tremor. The upper airway is affected early in the disease and is manifest by dysarthria and dysphagia. Limb muscles show a loss of fast MFs that is consistent with the bradykinesia. In this study a selective loss of the slow MFs was found, the opposite pattern than found in the limbs. Moreover, the histology of the IPD specimen appeared to be more similar to that of a limb muscle than the normal adult tongues. Specifically, MFs of both types were compactly arranged with little connective tissue between MFs or fascicles. In addition, the MFs were of relatively uniform size and formed a random mosaic pattern rather than the normal specimen's variable sized MFs that fiber-type grouping.

The significance of these findings is unclear but the overall impression is that the IPD specimen had lost much of the specialized features that made normal adult tongues distinctive. As IPD is a central nervous system disorder it appears that much of the specialization of the normal adult tongue depends on the activity patterns and/or trophic factors that originate centrally. This raises an interesting question of whether tongue muscle differentiation from cortically driven language results in muscle differentiation, rather than muscle differentiation being necessary to express cortically driven language.


The data from this study have shown that the adult human tongue has distinctive specializations. These specializations are at many levels; gross anatomy, muscle fascicular structure, MF size and connections, and MF contractile characteristics. Many of these specializations appear to be present only in humans who speak, including the 2-year old, and are absent in those who only vocalize without speaking, such as the human neonate and the macaque monkey. The purpose of these specializations is hypothesized to be the precise control of localized deformations in the tongue, the shape changes that contribute to speech articulation. However, other interpretations are possible and more work is needed to clarify the special role of the human tongue in speech. It should be pointed out that this study has limitations as indicated by a small sample size and limited muscle samples for the Experimental specimens. Therefore, further comparative studies with a large sample size are desirable for fully documenting function-related structural and histochemical specializations of the human tongue muscles.


The authors thank the anonymous reviewers for their constructive comments on the manuscript. Appreciation is extended to the Department of Pathology of Mount Sinai Medical Center in New York for providing human specimens for this study.