Elsevier

Speech Communication

Volume 49, Issue 1, January 2007, Pages 49-58
Speech Communication

An MRI analysis of the extrinsic tongue muscles during vowel production

https://doi.org/10.1016/j.specom.2006.09.004Get rights and content

Abstract

Functions of the extrinsic tongue muscles in vowel production were examined by measurements of muscle length and tongue tissue deformation using MRI (magnetic resonance imaging). Results from the analysis of Japanese vowel data suggested: (1) Contraction and relaxation of the three subdivisions of the genioglossus (GG) play a dominant role in forming tongue shapes for vowels. (2) The extra-lingual part of the styloglossus (SG), which was previously thought to cause a high-back tongue position by pulling its insertion point in the tongue, was found to be nearly constant across all vowels both in length and orientation. (3) The tongue shape for back vowels is mainly achieved by internal deformation of the tongue tissue, and the medial tissue of the tongue showed lateral expansion in front vowels, and medial compression in back vowels.

Introduction

Speech is a human-specific capacity for communication that relies on the faculty of the human vocal organs, and its production is due in large part to the activities of the human tongue to control the acoustic resonance of the vocal tract. Despite their importance in human speech communication, the physiological mechanisms of the tongue muscles are poorly understood, or only assumed by rudimentary information based on gross anatomy and muscle action potentials. This study aims to re-evaluate the previously reported functions of the extrinsic tongue muscles through measurements of muscle length and tongue tissue deformation using magnetic resonance imaging (MRI) data obtained during Japanese vowel production. The functions of the tongue muscles in vowel production have been discussed based mainly on anatomical knowledge (Miyawaki, 1973) and electromyographic (EMG) data (Baer et al., 1988). Anatomical literature describes muscle functions as relying on the geometry of the muscles and the expected effects of their shortening (Gray et al., 1858/1989). Such descriptions regard only the muscles’ contraction effects and do not explore the mechanism of the three-dimensional deformation of the tongue tissue. Many studies have noted that tongue deformation is three-dimensional, as seen in surface reconstruction (Stone and Lundberg, 1996, Engwall, 2003) and computational modeling (Kakita et al., 1985, Wilhelms-Tricarico, 1995). EMG studies have examined the activities of individual muscles during speech production (Hirose, 1971, Baer et al., 1988), and analyses based on such data have proposed a schematic view of the functional relationship among the extrinsic tongue muscles (Honda, 1996, Maeda and Honda, 1994). However, it is generally difficult to speculate about the three-dimensional nature of tongue deformation mechanisms only from muscle action potentials.

Geometrical information regarding the muscle bundles is of primary importance for estimating their functions. Miyawaki, 1973, Takemoto, 2001 examined tongue muscle morphology based on cadaver specimens. Miyawaki (1973) sketched tongue muscle structure using tongue tissue slices in the sagittal, coronal, and transverse planes. The muscle fibers of the tongue were manually traced on each slice to depict their locations. Takemoto (2001) performed semi-microscopic analysis of the extrinsic and intrinsic muscles of the tongue and revealed that the tongue muscle tissue can be divided into the inner layer with laminar fiber assembly and the outer layer with the extrinsic muscle bundles. These anatomical studies give us reliable knowledge of the tongue muscles for inferring their functions.

EMG data also provide useful information for estimating muscle functions. Kakita et al. (1985) used the EMG data from (Baer et al., 1988) to compute the three-dimensional tongue shape employing a finite element method (FEM) model of the tongue. They further interpreted the function of the styloglossus muscle in back vowels: the extra-lingual and intra-lingual bundles of this muscle together generate a combined force that raises the tongue dorsum. Maeda and Honda (1994), also using the same EMG data, proposed an antagonistic scheme among the tongue muscles: vowels are controlled by two orthogonally arranged pairs of antagonistic muscles.

The previous studies noted above lack sufficient evidence to determine the functions of the extrinsic muscles, since anatomical descriptions alone offer little insight into three-dimensional deformation mechanisms, and EMG data alone fail to explain the quantitative contributions of individual muscles to vowel production. Recently, a few MRI methods were introduced to visualize tongue tissue deformation, such as “cine-MRI” (Masaki et al., 1999) and “tagging cine-MRI” (Stone et al., 2001). In the latter, a mesh pattern is labeled onto the tissue image by adding “tag” pulses to decrease MR signals along parallel lines or a grid, and the deformation of the mesh pattern during a short speech sequence is monitored to infer the changes of tongue muscle geometry. Although this method provides a possible means to investigate tongue muscle functions, image resolution is yet insufficient to identify the precise geometry of each muscle.

The present study employs high-resolution static MRI techniques to investigate the functions of the extrinsic tongue muscles in vowel production. Image analyses for muscle shortening and tongue deformation were carried out on the basis of the following assumptions: (1) The static muscle length reflects an equilibrium state of tongue muscle forces and their effects on tongue deformation. (2) The internal deformation of the tongue tissue helps interpret the contractile effects of deep tongue muscles. (3) The function of an individual muscle can be inferred by exploring the three-dimensional muscle geometry for each vowel.

Section snippets

Functions of the extrinsic tongue muscles in the literature

This study focuses on the geometry and function of the three extrinsic tongue muscles. They have been described in the literature as follows. The genioglossus (GG) arises from the mental spine of the mandibular symphysis and spreads mid-sagittally in a fan shape in the medial part of the tongue. The posterior fibers (GGp) course to the tongue root, and the anterior fibers (GGa) project toward the dorsal surface of the tongue. The fibers of the GGp are thicker, stronger, and wider at the base of

Muscle length measurements

The lengths of the tongue muscles in mid-sagittal projection were measured on sagittal MRI data on the basis of distance scaling of makers and landmarks, and their variation across vowels was analyzed.

Results

The MRI data collected for the measurements of muscle lengths and tongue shape changes were found to have sufficient quality for image analysis. All of the four subjects produced the mid-vowel /ɯ/ regardless of the subjects’ dialects.

Discussion

In this study, the lengths of the extrinsic tongue muscles were measured and the three-dimensional deformation of the tongue tissue was visualized for the five Japanese vowels. In this section, the results obtained are compared with previous descriptions in the literature, and the roles of the extrinsic tongue muscles in vowel articulation are discussed.

Conclusion

The tongue muscles make the most significant contributions to the articulation of speech sound. Despite their critical functions supporting human communication, knowledge of how each muscle performs its role has mostly been speculative. While EMG and anatomical data have been the major source of information for estimating the physiological function of the tongue muscles, these data do not provide direct evidence of the causal relationships. This is because the tongue muscles blend with each

Acknowledgements

We are grateful to Dr. Shinobu Masaki, Mr. Yasuhiro Shimada, and Mr. Ichiro Fujimoto at the ATR Brain Activity Imaging Center for their help in conducting this study. We also thank Dr. Masaaki Honda at Waseda University for his effort in realizing the work as the P.I. of the CREST project. This work was conducted while the first author belonged to the CREST of JST (Japan Science and Technology). This research was also supported in part by the National Institute of Information and Communications

References (15)

There are more references available in the full text version of this article.

Cited by (33)

  • An acoustic-articulatory study of bilingual vowel production: Advanced tongue root vowels in Twi and tense/lax vowels in Ghanaian English

    2017, Journal of Phonetics
    Citation Excerpt :

    Our data show that all groups use tongue root advancement to produce the respective contrasts, with tongue height only playing a significant role in British English, and a lesser role amongst some high vowels in Twi. The finding that British English uses both tongue root and height is not surprising as the hydrostatic properties of the tongue mean that contraction of the posterior genioglossus (GGp) muscle advances the tongue root and also advances and raises the tongue body (Baer, Alfonso, & Honda, 1988; Takano & Honda, 2007). Tiede (1996) hypothesises that Akan languages do not show this tongue raising due to active contraction of the anterior genioglossus (GGa).

  • Human Tongue Biomechanical Modeling

    2017, Biomechanics of Living Organs: Hyperelastic Constitutive Laws for Finite Element Modeling
  • A study on coarticulatory resistance and aggressiveness for front lingual consonants and vowels using ultrasound

    2016, Journal of Phonetics
    Citation Excerpt :

    High and mid back rounded vowels involve mainly the activation of the styloglossus muscle, which pulls the tongue body upwards and backwards and, in addition, the posterior genioglossus for [u] and the pharyngeal constrictors for [o]. At least for [u] the entire genioglossus muscle could assist the tongue back raising by relaxation, which could account for the high degree of contextual sensitivity towards the tongue front when the vowel occurs next to palatal and dentoalveolar consonants (see Takano & Honda (2007) for Japanese). Low vowels, on the other hand, are produced by activating not only the hyoglossus which pulls the tongue backwards and downwards and the pharyngeal constrictors, but also the anterior genioglossus which acts to keep the tongue dorsum surface flat and in a relative front position more so when the low vowel is front ([a], also [æ]) than when it is back [ɑ].

  • Evaluating muscles underlying tongue base retraction in deglutition using muscular functional magnetic resonance imaging (mfMRI)

    2016, Magnetic Resonance Imaging
    Citation Excerpt :

    Electromyography, a technique involving the use of hook wire electrodes to measure muscle activity, has been used to verify the hyoglossus and styloglossus in mammalian models [4,5]. In humans, previous MRI studies have documented the deformation of the tongue muscles including the hyoglossus during speech via structural or cine-MRI [6,7]. However, no previous reports have verified the activity of the styloglossus and hyoglossus muscles in human swallowing.

  • Semi-automatic segmentation for 3D motion analysis of the tongue with dynamic MRI

    2014, Computerized Medical Imaging and Graphics
    Citation Excerpt :

    The first step in these analyses is an algorithm to accurately and repeatably extract the tongue (or structure of interest) for use in further analysis. Magnetic resonance imaging (MRI) is a widely used imaging modality for structural and functional analysis of the tongue and the vocal tract as it is nonionizing and shows excellent soft tissue contrast [3,7–13]. In particular, fast MR imaging with tagging [14,15] enables fast measurement and quantitative analysis of tongue motion during a specific speech or swallowing task [16,17].

View all citing articles on Scopus
1

Tel.: +81 774 95 1051.

View full text