Videofluoroscopic Investigation of Body Position on Articulatory Positioning Purpose To quantitatively examine the effects of body position on the positioning of the epiglottis, tongue, and velum at rest and during speech. Method Videofluoroscopic data were obtained from 12 healthy adults in the supine and upright positions at rest and during speech while the participants produced 12 VCV ... Research Article
Free
Research Article  |   August 01, 2014
Videofluoroscopic Investigation of Body Position on Articulatory Positioning
 
Author Affiliations & Notes
  • Youkyung Bae
    New Mexico State University, Las Cruces
  • Jamie L. Perry
    East Carolina University, Greenville, NC
  • David P. Kuehn
    University of Illinois at Urbana–Champaign
  • Disclosure: The authors have declared that no competing interests existed at the time of publication.
    Disclosure: The authors have declared that no competing interests existed at the time of publication.×
  • Correspondence to Youkyung Bae: youkyungsong@gmail.com
  • Editor: Jody Kreiman
    Editor: Jody Kreiman×
  • Associate Editor: Tim Bressman
    Associate Editor: Tim Bressman×
Article Information
Swallowing, Dysphagia & Feeding Disorders / Speech, Voice & Prosody / Speech / Research Articles
Research Article   |   August 01, 2014
Videofluoroscopic Investigation of Body Position on Articulatory Positioning
Journal of Speech, Language, and Hearing Research, August 2014, Vol. 57, 1135-1147. doi:10.1044/2013_JSLHR-S-12-0235
History: Received July 25, 2012 , Revised February 22, 2013 , Accepted August 14, 2013
 
Journal of Speech, Language, and Hearing Research, August 2014, Vol. 57, 1135-1147. doi:10.1044/2013_JSLHR-S-12-0235
History: Received July 25, 2012; Revised February 22, 2013; Accepted August 14, 2013
Web of Science® Times Cited: 3

Purpose To quantitatively examine the effects of body position on the positioning of the epiglottis, tongue, and velum at rest and during speech.

Method Videofluoroscopic data were obtained from 12 healthy adults in the supine and upright positions at rest and during speech while the participants produced 12 VCV sequences. The effects of body position, target sounds, and adjacent sounds on structural positioning and vowel formant structure were investigated.

Results Velar retropositioning in the supine position was the most consistent pattern observed at rest. During speech, all structures, with varying degrees of adjustment, appeared to work against the gravitational pull, resulting in no significant narrowing in the oro- and nasopharyngeal regions while in the supine position. Minimal differences in the formant data between the body positions were also observed. Overall, structural positioning was significantly dependent on the target and adjacent sounds regardless of body position.

Conclusions The present study demonstrated that structural positioning in response to gravity varied across individuals based on the type of activities being performed. With varying degrees of positional adjustment across different structures, fairly consistent articulatory positioning in the anterior–posterior dimension was maintained in different body positions during speech.

Speech production involves complex interactions between the environment and central neural representations of the speech system to produce dynamic movement patterns. Speech structures, with no exception, are subject to the constant force of gravity as an external load. Depending on body position, the orientation of the speech structures changes relative to gravitational force. Under most circumstances, speech production occurs in an upright position, and the majority of speech-related phenomena have been discussed based on the assumption of the speech apparatus being in an upright position. Several studies have investigated how the speech mechanism operates in different orientations relative to gravitational force using various techniques including X-ray, ultrasound, magnetic resonance imaging (MRI), electromagnetometry, and electromyography ( Engwall, 2003; Kitamura et al., 2005; Moon & Canady, 1995; Perry, 2011; Shiller, Ostry, & Gribble, 1999; Stone et al., 2007; Tiede, Masaki, & Vatikiotis-Bateson, 2000). Speech studies using MRI are most commonly obtained in the supine position. It is necessary to determine to what extent speech and anatomical data obtained in the supine position can be interpreted and generalized to upright speech functions.
Although individual variation exists in the literature, researchers generally agree that repositioning of the speech structures in response to different body positions causes minimal changes in the acoustic output ( Engwall, 2006; Kitamura et al., 2005; Stone et al., 2007; Tiede et al., 2000). The least amount of adjustment observed in the major constriction points, regardless of changes in body position, appears to physiologically contribute to preserving the acoustic targets ( Engwall, 2003, 2006; Stone et al., 2007; Tiede et al., 2000). Unlike the lingual gestures that have been the focus in the vast majority of studies, the extent and direction of adjustment made by other internally located structures are not well understood and are mainly confined to descriptive analyses. Thus, the present study aimed to examine the effects of body position on the displacement of various structures in the oro- and nasopharynx in a quantitative manner.
Stone et al. (2007)  used ultrasound to investigate tongue behavior in different body positions during speech activities. A significant variety of speaker-dependent compensatory maneuvers was reported in the global tongue and pharyngeal contours, whereas the range of tongue motion and vowel formant structure remained unchanged between the two body positions. The investigators concluded that individuals vary their responses to different body positions in an effort to preserve acoustic targets and maintain a sufficient airway. Similar findings were reported by Tiede et al. (2000)  using an X-ray microbeam system. The postural effects were minimal in the major constriction points for both consonant and vowel productions.
Engwall (2003, 2006)  examined midsagittal vocal tract tracings and area functions obtained from MRI on a single subject between the supine and upright positions. Noticeable changes were observed, particularly in the regions where no major constriction was involved (i.e., the spaces that would not be acoustically critical for the targeted phoneme). Kitamura et al. (2005)  compared midsagittal contours of the oropharyngeal structures between the supine and upright positions with MRI data acquired using an open-type scanner. The results indicated a greater degree of tongue retraction in the supine position for back vowels compared to high vowels. The points of major constriction for different vowels, however, appeared to remain constant regardless of the body position according to the majority of the midsagittal tracing data (see Figure 3 in Kitamura et al., 2005). Thus, the least variability observed at the points of constriction, as a means to preserve the acoustic output, is in agreement with previous literature ( Engwall, 2003, 2006; Stone et al., 2007; Tiede et al., 2000).
Most of the aforementioned studies provided articulatory data in conjunction with acoustic data. It is well agreed that vowels can be distinguished from one another based on the pattern constructed by the first two resonant frequencies (formants), where the first formant (F1) and the second formant (F2) are commonly described in relation to the size of the pharyngeal cavity or tongue height and the size of the oral cavity or tongue advancement, respectively ( Raphael, Borden, & Harris, 2007). Upon changes in body position, the effective direction of the gravitational pull on the articulators would change. Findings from previous literature are inconclusive whether changes in body position would result in changes in the shape of the vocal tract and possibly affect the acoustic output. It appears that it is articulatory variability that is more commonly observed, which may or may not be accompanied by significant acoustic changes.
Relatively limited information is available regarding the effects of body position on other articulators. Shiller et al. (1999)  showed minimal effects of body position on jaw movements. Despite a limited degree of compensatory adjustment observed for jaw movements in response to gravity, the acoustic output (i.e., formant frequencies) indicated variation between the different body positions. As the investigators noted, more active compensatory movements would be anticipated if intelligibility of speech is at risk, which partially explains this minimal level of compensatory adjustment observed in jaw movement.
Unlike the jaw, velar movement would likely have a more direct impact on speech intelligibility. Variations in velar positioning would occur to the extent that targeted nasality or non-nasality can be preserved, given that the role of the velopharyngeal port for speech is to maintain an appropriate level of oral–nasal balance by adjusting the port size. In addition to the degree of oral–nasal coupling, the shape of the oral cavity may be modified by the positioning of the velum, resulting in subsequent changes in the resonant characteristics ( Baker, Mielke, & Archangeli, 2008). Moon and Canady (1995)  reported reduced average peak levels of the levator veli palatini (levator) muscle activity in the supine position compared to those in the upright position using electromyographic data. The reduced activity level in the supine position was explained by gravity exerting force in the same direction as the force produced by the levator muscle, thus requiring less muscular effort in the supine position. The actual positioning of the velum would presumably be affected by the different activity level of the levator muscle. Using upright MRI, Perry (2011)  reported minimal differences between the supine and upright positions in velar thickness, velar length, levator muscle contraction, and nasopharyngeal dimensions. Velar height during high vowel production was the only measure that demonstrated significant differences between the two body positions. Results suggested that individual maneuvers in positioning the velopharyngeal structures appeared to offset the effect of different body positions. The pattern of velar displacement between the two body positions was also found to vary depending on the type of activity performed ( Perry, Bae, & Kuehn, 2012). Specifically, the velum demonstrated a compensatory response to gravity by its significantly posterior positioning in the upright position during swallowing; on the other hand, the velum conformed to the pull of gravity by its posterior positioning in the supine position at rest.
Accordingly, the effects of body position may vary across different structures—in particular, what type of activity is performed and to what extent individual structures are involved in the targeted activity may affect the positioning in response to gravity. For breathing purposes, it could be hypothesized that all oro- and nasopharyngeal structures would yield to gravitational pull as long as a patent airway opening is maintained. For speech purposes, variation across different structures in response to gravity could be anticipated; specifically, the degree of response would possibly be affected by the target and adjacent sounds, which determine to what extent different structures are involved in the sound production. Therefore, the present study quantitatively examined the effects of body position on structural displacement at different levels along the oro- and nasopharynx. More specifically, the positioning of the epiglottis, tongue, and velum was investigated at rest and during speech using videofluoroscopic imaging procedures. In addition, the effects of the target and adjacent sounds on structural positioning in the supine and upright positions were examined. Vowel formant analyses supplemented the structural positioning information in the two different body positions.
Method
Participants
A total of 12 Caucasian adults, six males and six females, between 19 and 27 years of age who speak American English as their first language participated in the study. Participants reported no history of speech, language, swallowing, or hearing disorders. Individuals with a history of a tonsillectomy, adenoidectomy, or any oropharyngeal structural or functional abnormalities were excluded from the study. The study was approved by the institutional review boards of the University of Illinois at Urbana-Champaign and New Mexico State University.
Speech Sample
The speech sample consisted of four consonants /s, n, f, m/ embedded into three vowel contexts /a, i, u/ as VCV sequences: /asa, isi, usu, ana, ini, unu, afa, ifi, ufu, ama, imi, umu/. The three corner vowels were included to sample the range of vocal tract shape and potentially different mechanical influences on velar and lingual positioning ( Kuehn, 1976). In order to maximize the contrast in the velopharyngeal status (open vs. closed), the voiceless consonants, which require greater velopharyngeal closure force than the voiced counterparts, were chosen ( Kuehn & Moon, 1998). In addition, the fricative consonants, with more sustainable articulatory posture than the stop consonants, were selected. The four consonants can be divided into two groups based on the place of articulation. The lingual consonants /s, n/ require active tongue movement, specifically the tongue tip approximating or contacting the alveolar ridge, whereas the labial consonants /f, m/ do not encumber the tongue with any particular position. Thus, the tongue presumably is more at liberty in positioning for the latter consonants, and the structural responses to gravity might be more evident for these consonants.
To keep the radiation dosage at a minimal level, only one production of each bisyllabic utterance was spoken by each participant. The participants were instructed to repeat each speech sample after the experimenter's production. The word list was rehearsed before recording commenced. During the actual recording, two native speakers of American English, trained in phonetics, listened to each of the participants' productions to ensure that the productions were accurate realizations of the intended utterance. Any unacceptable productions were repeated. The speech stimuli were presented in random order across participants.
Data Collection and Analyses
To assist with head positioning, a procedure similar to that described by Showfety, Vig, and Matteson (1983)  was used. Specifically, while each participant was standing and looking forward, a small transparent box containing water-diluted barium was affixed to the participant's head such that a line (Line 1) at the level of the barium was parallel to the floor. A second line (Line 2), perpendicular to Line 1 on the box, was used subsequently to position the head in the supine bodily position.
Immediately prior to the videofluoroscopic recordings, each participant received 1½ cc of barium transnasally to delineate soft tissue structures, particularly the velum, pharyngeal walls, and tongue in the sagittal view. Half of the female and male participants were randomly assigned to begin in the upright position; the others began in the supine position. A radiopaque metric ruler was inserted into each participant's mouth at the beginning of each recording, in both the supine and upright positions, for calibration purposes.
Instrumentation. Images of the participants' oropharyngeal structures were collected using lateral-view videofluoroscopy. A microphone and VCR were used to record the speech samples onto a super VHS master-quality videocassette, and the collected data were digitized using MicroVideo DC 30 software. Adobe After Effects CS4 was used to generate image sequences, and frame-by-frame analyses were conducted using Amira 5 two-dimensional length tools. Two of the investigators reviewed the entire image sequences and selected the images that were considered best representations of the resting state and the individual target sounds based on mutual agreement. During speech activities, each selected image depicted the peak of the articulatory gesture required for the target sound, such as maximally raised/lowered gesture of the tongue or closed/opened gesture of the lips, with a steady-state configuration of the targeted structure.
Structural variables. Four linear-distance variables were measured based on the images representing a resting state and each individual sound of the speech samples. In Figure 1, the four structural variables were measured relative to two reference lines on a videofluoroscopic image in the upright position. The first reference line connected the anterior-most edges of the second and third cervical vertebrae, which will hereafter be referred to as Reference Line 1 (RL1). Three ventral pharyngeal cavity landmarks (velar knee, tongue, and epiglottis) were selected and measured using a horizontal line positioned perpendicular to RL1. The second reference line (RL2) connected the posterior tip of the spinous process of the first cervical vertebra and the superior tip of the anterior tuberculum sellae. The superior-to-inferior velar position was measured relative to RL2. It should be noted that this line was arbitrarily established based on the identifiable bony structures, unlike the reference lines employed in previous cinefluorographic studies for the measurement of velar movement (e.g., Moll & Daniloff, 1971; Moll & Shriner, 1967).
Figure 1.

Four linear-distance variables in relation to two reference lines on a videofluoroscopic image in the upright position: Reference Line 1 (RL1) = a line connecting the anterior-most edges of the second and third cervical vertebrae; and Reference Line 2 (RL2) = a line connecting the posterior tip of the spinous process of the first cervical vertebra and the superior tip of the anterior tuberculum sellae. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to RL1; Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to RL2.

 Four linear-distance variables in relation to two reference lines on a videofluoroscopic image in the upright position: Reference Line 1 (RL1) = a line connecting the anterior-most edges of the second and third cervical vertebrae; and Reference Line 2 (RL2) = a line connecting the posterior tip of the spinous process of the first cervical vertebra and the superior tip of the anterior tuberculum sellae. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to RL1; Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to RL2.
Figure 1.

Four linear-distance variables in relation to two reference lines on a videofluoroscopic image in the upright position: Reference Line 1 (RL1) = a line connecting the anterior-most edges of the second and third cervical vertebrae; and Reference Line 2 (RL2) = a line connecting the posterior tip of the spinous process of the first cervical vertebra and the superior tip of the anterior tuberculum sellae. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to RL1; Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to RL2.

×
The first variable was the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to RL1 (referred to as Epiglottis-PPW). This variable described the positioning of the epiglottis at the inferior border of the oropharyx. The second variable, referred to as Tongue-PPW, was the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1. This variable allowed monitoring lingual displacement in the anterior–posterior dimension at a constant point along the oropharyngeal space. The other two variables pertained to velar positioning at the nasopharyngeal level. The third variable, referred to as Velum-RL1, was the perpendicular distance from the velar knee to RL1. The velar knee was identified as a point on the nasal surface of the velum at which the curve turned inferiorly with a larger curvature compared to the rest of the contour. For cases when the velum had a relatively flat nasal surface contour, especially during nasal sounds, the midpoint in the nasal surface of the velum served as a reference point. Thus, the Velum-RL1 variable described the positioning of the velum in the anterior–posterior dimension. The last variable, referred to as Velum-RL2, was the perpendicular distance from the velar knee to RL2. Given that the posterior and superior movements of the velum coincide with each other, the line drawn between the velum and RL2 was regarded as a measure of the velar position in the superior–inferior dimension.
All structural variables measured in the supine and upright positions for individual speech tokens were converted into ratio values of supine to upright to supplement descriptive statistics, which would minimize the influence of different vocal tract sizes across participants. The ratio would approximate one as the difference between the supine and upright positions for a given variable decreases. On the other hand, the ratio would deviate from one as the difference between the two body positions increases.
Monitoring head movement. Head movement (i.e., flexion and extension) was visually monitored during the recording by observing the barium within the transparent box, ensuring that the level of the barium coincided with Line 1 (parallel to the floor) in the upright position and Line 2 (perpendicular to Line 1) in the supine position. The level of the barium was easily observable on the videofluoroscopic images as well as during direct visualization.
The degree of head movement was additionally assessed with the collected videofluoroscopic data at a resting state by comparing the angles created by the intersection between RL1 and RL2 (see Figure 1). The difference between the angle constructed by the intersection of RL1 and RL2 in the supine and upright positions ranged from 0.0° to 15.5°, with a mean angle difference of 5.70° ( SD = 4.07°), which indicates minimal effects on upper airway dimensions ( Jan, Marshall, & Douglas, 1994).
Acoustic analyses. Audio data were extracted from the video files at the sampling rate of 48 kHz using VLC media player software ( VideoLAN, 2012). Praat ( Boersma & Weenink, 2010) was used to obtain F1 and F2 based on the mid one-third segment of the vowel duration. The automated formant analysis was performed based on the linear predictive coding analysis using the Burg algorithm, with the analysis window length of .025 s. Before acquiring the average F1 and F2 values for a vowel, the formant contours were closely inspected. When the overall formant values appeared to be faulty, the ceiling of the formant frequency range (i.e., maximum formant frequency) value was adjusted. Linear interpolation was used to replace outlier formant values. Unlike the robust F1–F2 structure that distinguishes vowels from one another, no such acoustic parameter that can characterize and differentiate nasal and fricative consonants is available. Therefore, acoustic analysis was confined to formant analysis for the vowels only in the present study.
Reliability
Inter- and intra-investigator reliability tests were performed on the four structural variables measured in the supine and upright positions for 12 randomly selected speech samples of the 12 participants. Approximately 10% of the data were used for the interinvestigator reliability test. A Pearson product–moment correlation coefficient of .94 was observed between the measurements made by the two raters. The cumulative percent frequencies for the measurement differences (mm) between the two raters showed that 80% of the measurements were agreed within 2 mm (see Table 1). The largest differences observed between the two raters were 3.89 mm ( M = 0.78 mm), 3.25 mm ( M = 1.06 mm), 4.60 mm ( M = 1.06 mm), and 4.55 mm ( M = 0.94 mm) for the Epiglottis-PPW, Tongue-PPW, Velum-RL1, and Velum-RL2 variables, respectively.
Table 1. Interinvestigator reliability: Cumulative percent frequencies (%) of the differences between the two raters in measurement (mm) for the four structural variables.
Interinvestigator reliability: Cumulative percent frequencies (%) of the differences between the two raters in measurement (mm) for the four structural variables.×
Difference between measurements Epiglottis-PPW Tongue-PPW Velum-RL1 Velum-RL2
≤ 0.5 42.0 27.8 25.0 42.0
≤ 1.0 73.9 59.7 62.5 59.4
≤ 1.5 85.5 73.6 80.6 76.8
≤ 2.0 91.3 87.5 86.2 86.9
≤ 2.5 94.2 90.3 90.4 92.7
≤ 3.0 94.2 95.9 96.0 97.0
≤ 3.5 98.5 100.0 97.3 98.4
≤ 4.0 100.0 98.7 98.4
≤ 4.5 98.7 98.4
≤ 5.0 100.0 100.0
Note. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to Reference Line 1 (RL1); Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to Reference Line 2 (RL2).
Note. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to Reference Line 1 (RL1); Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to Reference Line 2 (RL2).×
Table 1. Interinvestigator reliability: Cumulative percent frequencies (%) of the differences between the two raters in measurement (mm) for the four structural variables.
Interinvestigator reliability: Cumulative percent frequencies (%) of the differences between the two raters in measurement (mm) for the four structural variables.×
Difference between measurements Epiglottis-PPW Tongue-PPW Velum-RL1 Velum-RL2
≤ 0.5 42.0 27.8 25.0 42.0
≤ 1.0 73.9 59.7 62.5 59.4
≤ 1.5 85.5 73.6 80.6 76.8
≤ 2.0 91.3 87.5 86.2 86.9
≤ 2.5 94.2 90.3 90.4 92.7
≤ 3.0 94.2 95.9 96.0 97.0
≤ 3.5 98.5 100.0 97.3 98.4
≤ 4.0 100.0 98.7 98.4
≤ 4.5 98.7 98.4
≤ 5.0 100.0 100.0
Note. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to Reference Line 1 (RL1); Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to Reference Line 2 (RL2).
Note. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to Reference Line 1 (RL1); Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to Reference Line 2 (RL2).×
×
An overall correlation of .96 was observed for intra-investigator reliability using the Pearson product–moment correlation coefficient. Table 2 summarizes the cumulative percent frequencies in the test–retest measurement differences (mm), where 80% of the measurements were agreed within 2 mm. The largest differences in the test–retest measurements were 4.90 mm ( M = 0.94 mm), 4.95 mm ( M = 0.98 mm), 3.48 mm ( M = 0.76 mm), and 3.08 mm ( M = 1.05 mm) for the Epiglottis-PPW, Tongue-PPW, Velum-RL1, and Velum-RL2 variables, respectively.
Table 2. Intra-investigator reliability: Cumulative percent frequencies (%) of the differences between measurements (mm) made at two different time points for the four structural variables.
Intra-investigator reliability: Cumulative percent frequencies (%) of the differences between measurements (mm) made at two different time points for the four structural variables.×
Difference between measurements Epiglottis-PPW Tongue-PPW Velum-RL1 Velum-RL2
≤ 0.5 49.3 34.7 27.8 47.2
≤ 1.0 63.8 65.3 58.4 68.0
≤ 1.5 76.8 80.6 75.1 88.8
≤ 2.0 85.5 88.9 86.2 94.4
≤ 2.5 88.4 90.3 95.9 98.6
≤ 3.0 92.7 95.9 98.7 98.6
≤ 3.5 95.6 97.3 100.0 100.0
≤ 4.0 95.6 97.3
≤ 4.5 98.5 98.7
≤ 5.0 100.0 100.0
Table 2. Intra-investigator reliability: Cumulative percent frequencies (%) of the differences between measurements (mm) made at two different time points for the four structural variables.
Intra-investigator reliability: Cumulative percent frequencies (%) of the differences between measurements (mm) made at two different time points for the four structural variables.×
Difference between measurements Epiglottis-PPW Tongue-PPW Velum-RL1 Velum-RL2
≤ 0.5 49.3 34.7 27.8 47.2
≤ 1.0 63.8 65.3 58.4 68.0
≤ 1.5 76.8 80.6 75.1 88.8
≤ 2.0 85.5 88.9 86.2 94.4
≤ 2.5 88.4 90.3 95.9 98.6
≤ 3.0 92.7 95.9 98.7 98.6
≤ 3.5 95.6 97.3 100.0 100.0
≤ 4.0 95.6 97.3
≤ 4.5 98.5 98.7
≤ 5.0 100.0 100.0
×
Statistical Treatment
A series of participants' paired-samples t tests were used to examine the differences between the supine and upright positions at rest and during speech, separately, using SAS Version 9.2. Data collected during speech activities were divided into two sets including vowel and consonant productions. Each data set was analyzed using a series of three-way linear mixed-model analyses with repeated measures, in which posture (supine, upright), vowel (/a, i, u/), and consonant (/f, s, m, n/) served as within-subject fixed factors, and participants served as a random factor. Due to the unbalanced sample sizes, the Tukey–Kramer multiple comparison procedure was used to test all pair-wise contrasts for the statistically significant fixed factor effects. The formant data were supplemented by analyzing F1 and F2 using the same three-way mixed-model analyses. A significance level of α < .05 was used for all statistical tests.
Results
Overall Position Effects
The epiglottis of Participant 8 was not depicted in the videofluoroscopic images, and thus, the means of the supine and upright positions for the Epiglottis-PPW variable were based on 11 participants rather than 12. Table 3 provides descriptive data for the structural variables measured in millimeters at rest and during speech activities in the supine and upright positions. The ratio of supine to upright was computed across individual tokens, and summary statistics are also provided in Table 3. Results from the series of paired-samples t tests indicated no statistically significant differences between the two body positions for the structural variables, except for Velum-RL1. That is, the structural variables, including Epiglottis-PPW, Tongue-PPW, and Velum-RL2, measured in the supine position were not significantly different from those measured in the upright position. Regarding Velum-RL1, the distance from the velar knee to RL1 was significantly shorter while in the supine position than in the upright position, t(11) = 4.89, p < .05, at rest, which was consistently observed for 11 of the 12 participants. In such, the velum was positioned more posteriorly in the supine position than the upright position at rest. The mean ratio (0.84) of supine to upright across the participants also confirmed a more posterior velar position at rest in the supine position compared to the upright position.
Table 3. Summary data of the structural variables measured at rest and during speech activities: Group means (mm) in the supine and upright positions, mean ratios of supine to upright computed across individual tokens, and standard deviations (in parentheses).
Summary data of the structural variables measured at rest and during speech activities: Group means (mm) in the supine and upright positions, mean ratios of supine to upright computed across individual tokens, and standard deviations (in parentheses).×
Variable At rest During speech
Supine Upright Ratio Supine Upright Ratio
Epiglottis-PPW ‚5.95 (1.97) ‚6.06 (3.18) 0.99 (0.35) 10.49 (2.32) 10.33 (2.32) 1.02 (0.11)
Tongue-PPW 10.64 (3.18) 10.99 (3.63) 1.03 (0.31) 14.45 (2.34) 14.09 (2.39) 1.04 (0.13)
Velum-RL1 15.32 (2.32) 18.41 (2.34) 0.84 (0.11) ‚8.09 (1.95) ‚8.30 (2.82) 1.03 (0.27)
Velum-RL2 30.43 (3.20) 29.41 (3.31) 1.04 (0.10) 24.57 (3.18) 22.75 (2.80) 1.08 (0.11)
Table 3. Summary data of the structural variables measured at rest and during speech activities: Group means (mm) in the supine and upright positions, mean ratios of supine to upright computed across individual tokens, and standard deviations (in parentheses).
Summary data of the structural variables measured at rest and during speech activities: Group means (mm) in the supine and upright positions, mean ratios of supine to upright computed across individual tokens, and standard deviations (in parentheses).×
Variable At rest During speech
Supine Upright Ratio Supine Upright Ratio
Epiglottis-PPW ‚5.95 (1.97) ‚6.06 (3.18) 0.99 (0.35) 10.49 (2.32) 10.33 (2.32) 1.02 (0.11)
Tongue-PPW 10.64 (3.18) 10.99 (3.63) 1.03 (0.31) 14.45 (2.34) 14.09 (2.39) 1.04 (0.13)
Velum-RL1 15.32 (2.32) 18.41 (2.34) 0.84 (0.11) ‚8.09 (1.95) ‚8.30 (2.82) 1.03 (0.27)
Velum-RL2 30.43 (3.20) 29.41 (3.31) 1.04 (0.10) 24.57 (3.18) 22.75 (2.80) 1.08 (0.11)
×
The effect of body position during speech was also found to be negligible. A statistically significant difference between the supine and upright positions was found only for Velum-RL2, t(11) = −2.569, p < .05. Specifically, the distance from the velar knee to RL2 was significantly longer while in the supine position compared to the upright position. The mean ratio of supine to upright was 1.08, which also confirmed a more inferiorly positioned velum in the supine than upright position during speech activities. For the rest of the structural variables, no statistically significant differences were found between the supine and upright positions during speech.
Supine–Upright Comparisons During Speech Activities
The effects of the target and adjacent sounds on the positioning of the epiglottis, tongue, and velum in the supine and upright positions were examined. For this purpose, the data for the structural variables during vowel productions and consonant productions were analyzed separately. Along with the structural positioning information, the formant analyses for the vowel production in the supine and upright positions were provided.
Supine–upright comparisons during vowel productions. The first set of analyses involved the data of structural positioning and formant structure during vowel productions. Each mixed model was composed of posture (supine, upright), target vowel (/a, i, u/), and adjacent consonant (/f, s, m, n/) as within-subject fixed factors and participants as a random factor. The summary data of the structural variables in the supine and upright positions are displayed across different vowels in Figure 2.
Figure 2.

The means of the four structural variables (Epiglottis-PPW, Tongue-PPW, Velum-RL1, and Velum-RL2) measured in the supine (S) and upright (U) positions during vowel productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.

 The means of the four structural variables (Epiglottis-PPW, Tongue-PPW, Velum-RL1, and Velum-RL2) measured in the supine (S) and upright (U) positions during vowel productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.
Figure 2.

The means of the four structural variables (Epiglottis-PPW, Tongue-PPW, Velum-RL1, and Velum-RL2) measured in the supine (S) and upright (U) positions during vowel productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.

×
A significant main effect for posture was found only for Velum-RL2, F(1, 11) = 81.01, p < .05. The mean for Velum-RL2 was greater than the means for the other variables, indicating that the velum was more inferiorly positioned while in the supine position (24.44 mm) than in the upright position (22.57 mm) during vowel productions. For the remaining structural variables, no significant main effects were found for posture. That is, structural positioning was not significantly influenced by different body positions. The two-way and three-way interactions between posture and the other factors were not statistically significant.
The target vowels and the adjacent consonants showed statistically significant effects on structural positioning regardless of body position. With regard to the Epiglottis-PPW variable, a significant main effect was found for vowel, F(2, 20) = 436.79, p < .05, where the Epiglottis-PPW mean increased in the order of vowel /a/ (4.99 mm), /u/ (12.24 mm), and /i/ (13.82 mm). Subsequent pair-wise comparisons showed that all three target vowels were significantly different from one another ( p < .05). The main effect for the consonant factor was not statistically significant. In addition, none of the two-way or three-way interactions was statistically significant.
Similar results were found for the Tongue-PPW variable, in which the main effect for vowel was statistically significant, F(2, 22) = 505.07, p < .05. Although increasing in the order of vowel /a/ (8.30 mm), /u/ (14.99 mm), and /i/ (18.75 mm), the Tongue-PPW means for the three vowels were significantly different from one another ( p < .05). Neither the main effect for the consonant factor nor the interactions were statistically significant.
For Velum-RL1, the main effects for both vowel, F(2, 22) = 23.19, p < .05, and consonant, F(3, 33) = 19.81, p < .05, were statistically significant during vowel productions. Results showed that the Velum-RL1 mean increased in the order of vowel /u/ (7.37 mm), /i/ (7.97 mm), and /a/ (8.84 mm). The Velum-RL1 means for the three vowels were significantly different from one another ( p < .05). Regarding the significant main effect for the adjacent consonants on Velum-RL1, the mean values increased in the order of consonant /f/ (7.30 mm), /s/ (7.48 mm), /m/ (8.67 mm), and /n/ (8.80 mm). Results from the pair-wise comparisons showed that the following pairs of adjacent consonants, comparing the oral and nasal consonants, were found to be significantly different ( p < .05): /f/ and /m/, /s/ and /m/, /f/ and /n/, and /s/ and /n/. That is, the target vowels next to the nasal consonants (/m, n/) were produced with the significantly greater Velum-RL1 mean than those next to the oral consonants (/f, s/). No significant two-way or three-way interactions were found.
Results showed that there also were statistically significant main effects found for both vowel, F(2, 22) = 40.67, p < .05, and consonant, F(3, 33) = 24.45, p < .05, in Velum-RL2. Regarding the vowel effect, the Velum-RL2 mean increased in the order of vowel /u/ (22.65 mm), /i/ (23.05 mm), and /a/ (24.81 mm), where the mean for /a/ was significantly greater than those for the other two vowels ( p < .05). The effect of the adjacent consonants on Velum-RL2 was also found to be statistically significant. The Velum-RL2 mean varied depending on the adjacent consonants, which increased in the order of /f/ (22.56 mm), /s/ (22.67 mm), /n/ (24.39 mm), and /m/ (24.40 mm). Results from the pair-wise comparisons showed that only those pairs comparing the oral and nasal consonants were significantly different ( p < .05): /f/ and /m/, /s/ and /m/, /f/ and /n/, and /s/ and /n/. Similar to Velum-RL1, the target vowels adjacent to the nasal consonants (/m, n/) had significantly greater Velum-RL2 means than those adjacent to the oral consonants (/f, s/). None of the interactions were statistically significant.
With regard to the formant data, no significant main effects were found for posture in the acoustic measures of F1, F(1, 9) = 4.65, p = .06, and F2, F(1, 9) = 0.02, p = .89, although there were some numerical differences between the supine and upright positions. Acoustic data confirmed that both F1, F(2, 18) = 1,387.51, p < .05, and F2, F(2, 18) = 2,073.59, p < .05, significantly varied across different target vowels. In addition, the type of adjacent consonants was found to have a significant effect on F2, F(3, 27) = 8.14, p < .05. Table 4 summarizes descriptive formant data, including F1 and F2 expressed in Hz across different target vowels.
Table 4. Summary data of the means and standard deviations (in parentheses) in the supine and upright positions for the first formant (F1) and second formant (F2) in Hz across different target vowels.
Summary data of the means and standard deviations (in parentheses) in the supine and upright positions for the first formant (F1) and second formant (F2) in Hz across different target vowels.×
Position Vowel /a/ Vowel /i/ Vowel /u/
Supine Upright Supine Upright Supine Upright
F1 913.64 (161.27) 887.71 (161.76) 371.18 (51.20) 359.89 (51.57) 427.40 (72.82) 403.03 (63.54)
F2 1481.70 (193.91) 1492.15 (227.81) 2553.21 (315.63) 2537.59 (294.05) 1360.18 (243.46) 1345.51 (247.37)
Table 4. Summary data of the means and standard deviations (in parentheses) in the supine and upright positions for the first formant (F1) and second formant (F2) in Hz across different target vowels.
Summary data of the means and standard deviations (in parentheses) in the supine and upright positions for the first formant (F1) and second formant (F2) in Hz across different target vowels.×
Position Vowel /a/ Vowel /i/ Vowel /u/
Supine Upright Supine Upright Supine Upright
F1 913.64 (161.27) 887.71 (161.76) 371.18 (51.20) 359.89 (51.57) 427.40 (72.82) 403.03 (63.54)
F2 1481.70 (193.91) 1492.15 (227.81) 2553.21 (315.63) 2537.59 (294.05) 1360.18 (243.46) 1345.51 (247.37)
×
Supine–upright comparisons during consonant productions. The second set of analyses involved the data of structural positioning during consonant productions. Each mixed model was composed of posture (supine, upright), target consonant (/f, s, m, n/), and adjacent vowel (/a, i, u/) as within-subject fixed factors and participants as a random factor. The descriptive data of the structural variables during consonant productions are displayed in Figure 3. Similar to the results for vowel productions, a significant main effect for posture was found in Velum-RL2, F(1, 11) = 44.15, p < .05. The means for Velum-RL2 were significantly greater than the means for the other variables, indicating that the velum was positioned more inferiorly while in the supine (24.88 mm) than upright (23.10 mm) position during consonant productions. For the rest of the structural variables, no statistically significant main effects were found for posture. That is, Epiglottis-PPW, Tongue-PPW, and Velum-RL1 measured in the supine position were not significantly different from those measured in the upright position. The two-way and three-way interactions between posture and the other factors were not statistically significant.
Figure 3.

The means of the four structural variables measured in the supine and upright positions during consonant productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.

 The means of the four structural variables measured in the supine and upright positions during consonant productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.
Figure 3.

The means of the four structural variables measured in the supine and upright positions during consonant productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.

×
Results also showed that the main effects for the target consonant and the adjacent vowel factors were statistically significant for all four structural variables. In other words, structural positioning during consonant productions was significantly influenced by the target consonants and adjacent vowel contexts. In brief, there were significant main effects found for both vowel, F(2, 20) = 220.89, p < .05, and consonant, F(3, 30) = 29.57, p < .05, factors in Epiglottis-PPW. The interaction between vowel and consonant factors was also statistically significant, F(6, 60) = 2.95, p < .05. The means of the consonant and vowel combinations are plotted in Figure 4, which indicates that the effect of target consonants on Epiglottis-PPW changed depending on the type of adjacent vowels.
Figure 4.

The means of target consonant by adjacent vowel combinations for Epiglottis-PPW during consonant productions.

 The means of target consonant by adjacent vowel combinations for Epiglottis-PPW during consonant productions.
Figure 4.

The means of target consonant by adjacent vowel combinations for Epiglottis-PPW during consonant productions.

×
Similar results were found for Tongue-PPW, in which statistically significant main effects were found for both vowel, F(2, 22) = 272.31, p < .05, and consonant, F(3, 33) = 33.67, p < .05, as well as the interaction between the two, F(6, 66) = 4.34, p < .05. The means of the consonant and vowel combinations are plotted in Figure 5, which indicates that the effect of target consonants on Tongue-PPW was influenced by the type of adjacent vowels.
Figure 5.

The means of target consonant by adjacent vowel combinations for Tongue-PPW during consonant productions.

 The means of target consonant by adjacent vowel combinations for Tongue-PPW during consonant productions.
Figure 5.

The means of target consonant by adjacent vowel combinations for Tongue-PPW during consonant productions.

×
The main effects for vowel, F(2, 22) = 3.88, p < .05, and consonant, F(3, 33) = 50.57, p < .05, were found to be statistically significant in Velum-RL1. The means of Velum-RL1 during consonant productions increased in the order of /f/, /s/, /m/, and /n/ (see Figure 3), where those pairs comparing the oral and nasal consonants were significantly different ( p < .05): /f/ and /m/, /s/ and /m/, /f/ and /n/, and /s/ and /n/. The Velum-RL1 mean during consonant productions significantly varied across different vowels adjacent to the target consonants, in which the mean was significantly smaller in the /i/ vowel context (8.13 mm) than the /u/ (8.39 mm) or /a/ (8.90 mm) vowel contexts. None of the interactions were statistically significant in Velum-RL1.
The main effects for vowel, F(2, 22) = 9.88, p < .05, and consonant, F(3, 33) = 77.23, p < .05, were also statistically significant in Velum-RL2. The means of Velum-RL2 during consonant productions increased in the order of /s/, /f/, /n/, and /m/ (see Figure 3). Subsequent pair-wise comparisons showed that those pairs comparing the oral and nasal consonants were statistically significant ( p < .05): /f/ and /m/, /s/ and /m/, /f/ and /n/, and /s/ and /n/. The adjacent vowels significantly influenced the Velum-RL2 measure; the measure for the /a/ (24.83 mm) vowel context was significantly greater than those for the /i/ (23.55 mm) and /u/ (23.59 mm) vowel contexts. No significant interactions were found in Velum-RL2.
Discussion
The purpose of this study was to provide quantified information on the possible positional changes of the pharyngeal structures in response to gravity at rest and during speech. The positioning of the structural variables along the pharyngeal region, including the epiglottis, tongue, and velum, was examined using videofluoroscopic imaging procedures. Results showed that only the positioning of the velum was significantly affected by different body positions.
Effects of Body Position at Rest
While maintaining a normal patent airway, differences between the supine and upright position in structural positioning were negligible along the pharyngeal cavity, with the exception of Velum-RL1. Participants demonstrated no remarkable or consistent patterns of oropharyngeal narrowing, represented as the distance from the tongue and epiglottis to the PPW, in the anterior–posterior dimension. Minimal differences between the supine and upright positions may suggest a compensatory response of pharyngeal structures to counterbalance the gravitational pull while in the supine position. This compensatory response would presumably maintain a patent airway for normal breathing. The positioning of the epiglottis was observed to be contingent on the positioning of the tongue in 10 out of 11 participants at rest, which is not surprising given that the two structures are mechanically bound to each other. Although previous research ( Pae et al., 1994; Sutthiprapaporn et al., 2008) reported narrowing in the oropharyngeal area, especially observed in the tongue, epiglottis, and cross-sectional oropharyngeal area, following the positional change from upright to supine, inconsistent responses seen in the present study suggest that substantial individual variability may exist in the positioning of the oro- and nasopharyngeal structures upon changes in body position.
Among the four structural variables, the distance between the velar knee and RL1 nearby the PPW was significantly shorter in the supine than upright position at rest. This posterior positioning of the velum in the supine position is consistent with findings from previous studies ( Engwall, 2003; Kitamura et al., 2005; Perry, 2011; Perry et al., 2012). Although the velum gave way to the gravitational pull during rest breathing, it did so only to the extent that a minimal level of opening necessary for normal breathing could be preserved. This suggests a considerable muscle activity in order to keep the velum from being completely retrodisplaced in the supine position. The best candidate for this activity (i.e., pulling the velum anteriorly and inferiorly) would be the palatoglossus muscle. Moon and Canady (1995)  reported variations of the palatoglossus muscle activity in response to gravitational pull; although not statistically significant, the palatoglossus muscle was found to have greater peak activity levels in the supine than upright position for seven out of the 12 participants in the study. In addition to an active palatoglossus muscle contraction keeping the velopharyngeal port open, there might also be a passive component by virtue of the large investment of elastic fibers in the velum, anterior faucial pillars, and between palatoglossus muscle fascicles ( Kuehn & Azzam, 1978). It is likely that both active and passive components were involved in keeping the velopharyngeal airway patent for the participants in this study, who were relatively young adults (i.e., between 19 and 27 years of age). However, for elderly individuals, the active muscle force might be reduced, and the passive force of elastic tissue might be less resilient, which may lead to a more posteriorly positioned velum in the supine position during sleep, resulting in snoring, especially in older individuals ( Kuehn & Azzam, 1978).
The consistent pattern across the participants observed during rest breathing for Velum-RL1 was not observed for Velum-RL2. Although both variables measured velar positioning, the first pertains to velar positioning in the anterior–posterior dimension and the latter to velar positioning in the superior–inferior dimension (i.e., velar height). This discordant information regarding velar positioning can be viewed as (a) velar height represented as the Velum-RL2 variable may not be as critical as velopharyngeal port opening in the anterior–posterior dimension represented as Velum-RL1 in maintaining a normal airway, or (b) the variability found across participants in velar height may account for individual maneuvers to maintain a certain level of nasopharyngeal space necessary for normal breathing. Three-dimensional imaging would provide better information on the volumetric changes of the oro- and nasopharyngeal regions in response to different body positions.
Effects of Body Position During Speech Activities
Results from the study showed that the overall effects of body position on structural positioning varied across different structures during speech activities. No significant narrowing in the oro- and nasopharyngeal structures, especially in the anterior–posterior dimension, was observed in the supine position. Although not statistically significant, a consistent pattern of structural positioning in response to gravity was observed in which all of the structures appeared to work against the gravitational pull as a compensatory maneuver in the supine position during speech activities. Such a pattern, in fact, was effectively detected by the supplementary measure of the supine-to-upright ratios greater than one (see Table 4). This velar (Velum-RL1) response pattern observed during speech is contrary to that observed during rest breathing. This task-dependent response in velar positioning is consistent with the findings from Perry et al. (2012), which reported varying patterns of velar displacement between supine and upright positions depending on the type of activity performed (swallowing vs. resting). This finding can also be viewed in relation to the different activity level of the levator muscle in different body positions. Moon and Canady (1995)  reported a lower peak activity level of the levator muscle in the supine position compared to the upright position among the majority of participants. This lower activity level is presumably due to gravity working in the same direction as the force produced by the levator muscle in the supine position. Given that the resultant effect of the levator muscle contraction is velar elevation, a reduced level of this muscle activity in the supine position would likely result in a longer length of the muscle ( Perry, 2011) and a more inferiorly positioned velum, as observed in the present study. Velum-RL2 was the only variable that showed a statistically significant difference between the supine and upright positions during speech overall, where the velum was found to be significantly inferiorly positioned while in the supine rather than the upright position.
It was expected that structural positioning in the supine and upright positions would be influenced by the target and adjacent sounds that determine to what extent different structures are involved in sound production. Data from the study showed that structural positioning during speech significantly varied depending on the target and adjacent sounds. However, the type of sounds did not seem to affect the extent of structural adjustment in different body positions, as none of the interaction terms between posture and vowel or consonant were statistically significant. This finding is in agreement with Stone et al. (2007), in which the type of different phonemes was found to have little effect on the global tongue contours in two different body positions.
Results showed that Velum-RL2, pertaining to velar height, was the only variable that had significant differences between the supine and upright positions during both vowel and consonant productions. The positioning of the velum was significantly inferior while in the supine position, suggesting that the velum worked against the direction of gravity, than in the upright position during both vowel and consonant productions. For the rest of the structural and acoustic variables, no significant effects of body position were found. That is, the positioning of the epiglottis, tongue, and velum measured in the anterior–posterior dimension while in the supine position was not significantly different from that measured in the anterior–posterior dimension while in the upright position. Acoustic data also showed negligible effects of body position on the formant frequencies. Therefore, it may be safe to conclude that the positioning of different oro- and nasopharyngeal structures is adjusted to offset the gravitational pull in the supine position during speech, although the degree of adjustment varies across different structures. More importantly, minimal differences between the supine and upright positions in the structural and acoustic data strongly suggest that individuals make proper adjustments for different forces (i.e., gravity) acting on speech structures to maintain fairly consistent articulatory positioning and, presumably, a relatively constant vocal tract shape so as not to compromise an accurate acoustic realization of each speech sound.
The present study demonstrated that target and adjacent sounds have significant effects on determining structural positioning regardless of body position. The epiglottis and tongue showed similar patterns during vowel and consonant productions. With regard to vowel production, the mean values of Tongue-PPW and Epiglottis-PPW were highly dependent on the target vowels, with little effect from the adjacent consonants. The oropharyngeal spaces at the level of the tongue and epiglottis were found to be greater for the high-front vowel /i/ than the back vowels /u/ and /a/. Although not statistically significant, the difference between the supine and upright positions for Tongue-PPW was slightly greater (>1 mm) while producing /i/ than /a/ or /u/. This overshoot-like maneuver that was allowed for the high-front vowel but not for the back vowels may partly support previous findings ( Engwall, 2003, 2006; Tiede et al., 2000) in which speech structures were found to have the least variability in their positioning upon gravitational changes for acoustically sensitive targets. The vowel-dependent positioning of the tongue can further be interpreted in relation to standing wave theory ( Raphael, Borden, & Harris, 2007). In this regard, the vowel /i/ has a low F1 but is characterized by an F2 that is much higher than that for all other vowels ( Peterson & Barney, 1954). According to standing wave theory, the best way to achieve this acoustic effect is to have an expansion in the oropharyngeal region at which a standing wave volume velocity is at a maximum corresponding to F2. The vowels /a/ and /u/, on the other hand, have a lower F2 and benefit from a more constricted oropharyngeal space compared to that for /i/. The reason is that a constriction at a volume velocity maximum tends to lower the corresponding resonant frequency. Presumably, the constraints imposed on oropharyngeal volume are more stringent for proper realization of F2 for /a/ and /u/ compared to /i/. This is probably true because no other vowel is produced with a higher F2 than /i/ so that overshoot in the lower tongue region in the supine position is allowable for this vowel. With the external load of gravity effective in an unusual anterior–posterior dimension in the supine position, more precise lingual adjustment might also have been required. Nonetheless, the formant data from the present study showed minimal effect of body position on F2, which most likely pertains to tongue advancement information. Rather, the effect of body position appeared to be greater in F1 than F2, although not statistically significant ( p = .06). The increased F1 values and presumably lowered tongue position in the supine position relative to the upright position can be viewed as another compensatory maneuver to offset the gravitational pull. Thus, the information on the positioning and the formant structure related to the tongue in the supine and upright positions can be viewed in agreement with previous literature, which demonstrated varying degrees of lingual positioning adjustments working toward preserving proper acoustic output ( Stone et al., 2007; Tiede, Masaki, & Vatikiotis-Bateson, 2000).
With regard to lingual displacement during consonant productions, it was anticipated that, upon changes of body position, Tongue-PPW would show more variability during nonlingual consonant productions than during lingual consonant productions. Results, however, showed that there was no statistically significant interaction between posture and the target consonants, and thus, the hypothesis was not supported in the present study. The limited amount of lingual displacement in the supine position during nonlingual consonant productions can be partly explained by the coarticulatory effects from the adjacent vowels, which were also stringently controlled. In addition, it is possible that, along with the positioning of the tongue at the point of constriction, the shaping of the tongue in the oropharyngeal region may be strictly controlled in order to properly modulate the airflow for consonant production.
The velum-related variables clearly demonstrated the effects of vowel height (i.e., high vs. low) and the status of the velopharyngeal port (i.e., open vs. closed) required for the target or adjacent consonants on velar positioning. It was not surprising that the velum was positioned more anteriorly and inferiorly when it was targeting or surrounded by the low vowel /a/ and the nasal consonants /m, n/ compared to the high vowels /i, u/ and the oral consonants /f, s/. Supposedly, the formant data would be affected by the coupling between the nasal and oral cavities as well as lingual gestures. In the present study, however, the adjacent consonant context (oral vs. nasal) was found to have little effect on F1. That is, no remarkable distinction between oral and nasalized vowels was observed in F1, despite the significant effect of the adjacent consonant context (oral vs. nasal) on velar positioning. The discrepancy between structural positioning and acoustic data may be partly due to small sample size in the present study. In addition, given that this study was based on the midsagittal information only, possible articulatory maneuvers recruited in different dimensions are largely unknown, which needs further investigation.
In order to keep the radiation dosage low, only one token of each speech sample was collected from each participant. This raised the challenge of ensuring that the one single token of each speech sample was representative of the participant's general production in terms of structural positioning. Such a challenge, however, should resolve by using a tool that allows image acquisitions repeatedly with ease while visualizing the dynamically moving structures. MRI has become useful in speech research because it offers free choice of the image planes in all three dimensions with excellent soft tissue contrast and is free from ionizing radiation, making it easily repeatable. Bae, Kuehn, Conway, and Sutton (2011)  demonstrated the feasibility of using real-time MRI in conjunction with simultaneous audio recordings to bridge information between the lingual and velar articulatory gestures with their acoustic features in the supine position. Most of the current upright or open MRI scanners may not have the capacity to accommodate real-time imaging pulse sequences due to their use of low magnetic field strength. However, the potential of using real-time MRI with repeated image acquisitions in different body positions remains promising. Findings from the present study also support the use of supine imaging (such as traditional MRI) for anatomic and acoustic investigations as structural positioning in the anterior–posterior dimension along the pharyngeal cavity was not affected by different body positions.
Study Limitations
There are several limitations to be noted in this study. Although videofluoroscopy allows visualization of articulatory gestures and swallowing function, the spatial resolution offered by the technique may not be sufficient to detect the smallest changes in articulatory positioning induced by different body positions. Sia, Carvajal, Carnaby-Mann, and Crary (2012)  pointed out that “radial distortion causes an image to be increasingly stretched toward the periphery” (p. 192) of the field of view. To offset this error as much as possible in the current study, care was taken to confine measurements largely to the center of the field of view. Along with the issue of spatial resolution, the rate of 30 frames per second introduces motion artifacts such as blurring of images when imaging articulatory structures that are constantly in motion. Both factors might have contributed to the inter- and intra-investigator reliability test results, which indicated that approximately 80% of measurements were agreed within 2 mm.
As noted, information on velar positioning in response to different body positions is largely lacking. This can be partly attributed to the internally located velum that is not easily accessible through various laboratory techniques. For example, although the use of ultrasound has provided detailed information on lingual gestures, it is not suitable for visualization of velar movement. Attaching a point-tracking device on the velum is also challenging, thus limiting the use of electromagnetic articulography or an X-ray microbeam system. Videofluoroscopy has been commonly employed for visualizing velopharyngeal function, which allows quantitative measurements on the structures of interest. Nevertheless, it should be acknowledged that the image quality of videofluoroscopy may not allow the level of accuracy and precision necessary to detect small differences induced by changing body position.
In addition, the speech samples employed in the present study were nonsense syllable sequences. The use of nonsense syllables may induce variations in the speed and accuracy of speech production (e.g., Munson, 2001). Little is known about how articulators would make adjustments in response to varying degrees of articulatory demands in an unusual body position, and thus further research is needed.
Conclusion
The present study provided quantitative information regarding the effects of body position on structural displacement at rest and during speech activities. Findings from the present study confirmed that structural positioning is influenced by different body positions; however, structural positioning in response to gravity varied across individuals based on the type of activities performed. Moreover, the degree of adjustment in response to gravity also varied across different structures within a given individual. Nonetheless, the overall adjustments were made in order to maintain fairly consistent articulatory positioning in the anterior–posterior dimension so that, presumably, a relatively constant vocal tract shape and a proper acoustic realization of each speech sound could be preserved. It is still unclear what underlying factors would determine the extent of positional adjustment across different structures upon changes in body position. Thus, further investigation is warranted to systematically account for the gravity effects on the articulatory structures in multiple dimensions and their acoustic output while other factors such as mechanical constraints and coarticulatory effects are taken into account.
Acknowledgment
We would like to thank Rebecca Ray Ebert and Joseph Barkmeier for their involvement in collecting the videofluoroscopic data.
References
Bae, Y. Kuehn, D. P. Conway, C. A. Sutton, B. P. ( 2011). Real-time magnetic resonance imaging of velopharyngeal activities with simultaneous speech recordings. Cleft Palate–Craniofacial Journal, 48, 695– 707. [Article]
Bae, Y. Kuehn, D. P. Conway, C. A. Sutton, B. P. ( 2011). Real-time magnetic resonance imaging of velopharyngeal activities with simultaneous speech recordings. Cleft Palate–Craniofacial Journal, 48, 695– 707. [Article] ×
Baker, A. Mielke, J. Archangeli, D. ( 2008). More velar than /g/: Consonant coarticulation as a cause of diphthongization. In Chang, C. B. Haynie, H. J. (Eds.), Proceedings of the 26th West Coast Conference on Formal Linguistics (pp. 60– 68). Somerville, MA: Cascadilla Proceedings Project.
Baker, A. Mielke, J. Archangeli, D. ( 2008). More velar than /g/: Consonant coarticulation as a cause of diphthongization. In Chang, C. B. Haynie, H. J. (Eds.), Proceedings of the 26th West Coast Conference on Formal Linguistics (pp. 60– 68). Somerville, MA: Cascadilla Proceedings Project.×
Boersma, P. Weenink, D. ( 2010). Praat: Doing phonetics by computer (Version 5.1.29) [Computer software]. Retrieved from www.praat.org/
Boersma, P. Weenink, D. ( 2010). Praat: Doing phonetics by computer (Version 5.1.29) [Computer software]. Retrieved from www.praat.org/×
Engwall, O. ( 2003). A revisit to the application of MRI to the analysis of speech production: Testing our assumptions. In Palethorpe, S. Tabain, M. (Eds.), Proceedings of the 6th International Seminar on Speech Production (pp. 43– 48). Sydney, Australia: Macquarie University.
Engwall, O. ( 2003). A revisit to the application of MRI to the analysis of speech production: Testing our assumptions. In Palethorpe, S. Tabain, M. (Eds.), Proceedings of the 6th International Seminar on Speech Production (pp. 43– 48). Sydney, Australia: Macquarie University.×
Engwall, O. ( 2006). Assessing MRI measurements: Effects of sustenation, gravitation and coarticulation. In Harrington, J. Tabain, M. (Eds.), Speech production: Models, phonetic processes and techniques (pp. 300– 314). New York, NY: Psychology Press.
Engwall, O. ( 2006). Assessing MRI measurements: Effects of sustenation, gravitation and coarticulation. In Harrington, J. Tabain, M. (Eds.), Speech production: Models, phonetic processes and techniques (pp. 300– 314). New York, NY: Psychology Press.×
Jan, M. A. Marshall, I. Douglas, N. J. ( 1994). Effect of posture on upper airway dimensions in normal human. American Journal of Respiratory and Critical Care Medicine, 149, 145– 148. [Article]
Jan, M. A. Marshall, I. Douglas, N. J. ( 1994). Effect of posture on upper airway dimensions in normal human. American Journal of Respiratory and Critical Care Medicine, 149, 145– 148. [Article] ×
Kitamura, T. Takemoto, H. Honda, K. Shimada, Y. Fujimoto, I. Syakudo, Y. … Senda, M. ( 2005). Difference in vocal tract shape between upright and supine postures: Observations by an open-type MRI scanner. Acoustical Science and Technology, 26, 465– 468. [Article]
Kitamura, T. Takemoto, H. Honda, K. Shimada, Y. Fujimoto, I. Syakudo, Y. … Senda, M. ( 2005). Difference in vocal tract shape between upright and supine postures: Observations by an open-type MRI scanner. Acoustical Science and Technology, 26, 465– 468. [Article] ×
Kuehn, D. P. ( 1976). A cineradiographic investigation of velar movement variables in two normals. Cleft Palate Journal, 13, 88– 103.
Kuehn, D. P. ( 1976). A cineradiographic investigation of velar movement variables in two normals. Cleft Palate Journal, 13, 88– 103.×
Kuehn, D. P. Azzam, N. A. ( 1978). Anatomical characteristics of palatoglossus and the anterior faucial pillar. Cleft Palate Journal, 15, 349– 359.
Kuehn, D. P. Azzam, N. A. ( 1978). Anatomical characteristics of palatoglossus and the anterior faucial pillar. Cleft Palate Journal, 15, 349– 359.×
Kuehn, D. P. Moon, J. B. ( 1998). Velopharyngeal closure force and levator veli palatini activation levels in varying phonetic contexts. Journal of Speech, Language, and Hearing Research, 41, 51– 62. [Article]
Kuehn, D. P. Moon, J. B. ( 1998). Velopharyngeal closure force and levator veli palatini activation levels in varying phonetic contexts. Journal of Speech, Language, and Hearing Research, 41, 51– 62. [Article] ×
Moll, K. L. Daniloff, R. G. ( 1971). Investigation of the timing of velar movements during speech. The Journal of the Acoustical Society of America, 50, 678– 684. [Article]
Moll, K. L. Daniloff, R. G. ( 1971). Investigation of the timing of velar movements during speech. The Journal of the Acoustical Society of America, 50, 678– 684. [Article] ×
Moll, K. L. Shriner, T. H. ( 1967). Preliminary investigation of a new concept of velar activity during speech. Cleft Palate Journal, 4, 58– 69.
Moll, K. L. Shriner, T. H. ( 1967). Preliminary investigation of a new concept of velar activity during speech. Cleft Palate Journal, 4, 58– 69.×
Moon, J. B. Canady, J. W. ( 1995). Effects of gravity on velopharyngeal muscle activity during speech. Cleft Palate–Craniofacial Journal, 32, 371– 375. [Article]
Moon, J. B. Canady, J. W. ( 1995). Effects of gravity on velopharyngeal muscle activity during speech. Cleft Palate–Craniofacial Journal, 32, 371– 375. [Article] ×
Munson, B. ( 2001). Phonological pattern frequency and speech production in adults and children. Journal of Speech, Language, and Hearing Research, 44, 778– 792. [Article]
Munson, B. ( 2001). Phonological pattern frequency and speech production in adults and children. Journal of Speech, Language, and Hearing Research, 44, 778– 792. [Article] ×
Pae, E. K. Lowe, A. A. Sasaki, K. Price, C. Tsuchiya, M. Fleetham, J. A. ( 1994). A cephalometric and electromyographic study of upper airway structures in the upright and supine positions. American Journal of Orthodontics and Dentofacial Orthopedics, 106, 52– 59. [Article]
Pae, E. K. Lowe, A. A. Sasaki, K. Price, C. Tsuchiya, M. Fleetham, J. A. ( 1994). A cephalometric and electromyographic study of upper airway structures in the upright and supine positions. American Journal of Orthodontics and Dentofacial Orthopedics, 106, 52– 59. [Article] ×
Perry, J. L. ( 2011). Variations in velopharyngeal structures between upright and supine positions using upright magnetic resonance imaging. Cleft Palate–Craniofacial Journal, 48, 123– 133. [Article]
Perry, J. L. ( 2011). Variations in velopharyngeal structures between upright and supine positions using upright magnetic resonance imaging. Cleft Palate–Craniofacial Journal, 48, 123– 133. [Article] ×
Perry, J. L. Bae, Y. Kuehn, D. P. ( 2012). Effect of posture on deglutitive biomechanics in healthy individuals. Dysphagia, 27, 70– 80. [Article]
Perry, J. L. Bae, Y. Kuehn, D. P. ( 2012). Effect of posture on deglutitive biomechanics in healthy individuals. Dysphagia, 27, 70– 80. [Article] ×
Peterson, G. E. Barney, H. L. ( 1954). Control methods used in a study of the identification of vowels. The Journal of the Acoustical Society of America, 24, 175– 184. [Article]
Peterson, G. E. Barney, H. L. ( 1954). Control methods used in a study of the identification of vowels. The Journal of the Acoustical Society of America, 24, 175– 184. [Article] ×
Raphael, L. J. Borden, G. J. Harris, K. S. ( 2007). Speech science primer: Physiology, acoustics, and perception of speech ( 5th ed.). Baltimore, MD: Lippincott Williams & Wilkins.
Raphael, L. J. Borden, G. J. Harris, K. S. ( 2007). Speech science primer: Physiology, acoustics, and perception of speech ( 5th ed.). Baltimore, MD: Lippincott Williams & Wilkins.×
Shiller, D. M. Ostry, D. J. Gribble, P. L. ( 1999). Effects of gravitational load on jaw movements in speech. The Journal of Neuroscience, 19, 9073– 9080.
Shiller, D. M. Ostry, D. J. Gribble, P. L. ( 1999). Effects of gravitational load on jaw movements in speech. The Journal of Neuroscience, 19, 9073– 9080.×
Showfety, K. J. Vig, P. S. Matteson, S. ( 1983). A simple method for taking natural-head-position cephalograms. American Journal of Orthodontics, 83, 495– 500.
Showfety, K. J. Vig, P. S. Matteson, S. ( 1983). A simple method for taking natural-head-position cephalograms. American Journal of Orthodontics, 83, 495– 500.×
Sia, I. Carvajal, P. Carnaby-Mann, G. D. Crary, M. A. ( 2012). Measurement of hyoid and laryngeal displacement in videofluoroscopic swallowing studies: Variability, reliability, and measurement error. Dysphagia, 27, 192– 197. [Article]
Sia, I. Carvajal, P. Carnaby-Mann, G. D. Crary, M. A. ( 2012). Measurement of hyoid and laryngeal displacement in videofluoroscopic swallowing studies: Variability, reliability, and measurement error. Dysphagia, 27, 192– 197. [Article] ×
Stone, M. Stock, G. Bunin, K. Kumar, K. Epstein, M. Kambhamettu, C. … Prince, J. ( 2007). Comparison of speech production in upright and supine position. The Journal of the Acoustical Society of America, 122, 532– 541. [Article]
Stone, M. Stock, G. Bunin, K. Kumar, K. Epstein, M. Kambhamettu, C. … Prince, J. ( 2007). Comparison of speech production in upright and supine position. The Journal of the Acoustical Society of America, 122, 532– 541. [Article] ×
Sutthiprapaporn, P. Tanimoto, K. Ohtsuka, M. Nagasaki, T. Lida, Y. Katsumata, A. ( 2008). Positional changes of oropharyngeal structures due to gravity in the upright and supine positions. Dentomaxillofacial Radiology, 37, 130– 135. [Article]
Sutthiprapaporn, P. Tanimoto, K. Ohtsuka, M. Nagasaki, T. Lida, Y. Katsumata, A. ( 2008). Positional changes of oropharyngeal structures due to gravity in the upright and supine positions. Dentomaxillofacial Radiology, 37, 130– 135. [Article] ×
Tiede, M. K. Masaki, S. Vatikiotis-Bateson, E. 2000). Contrasts in speech articulation observed in sitting and supine conditions. In Hoole, P. (Ed.), Proceedings of the 5th Seminar on Speech Production: Models and data (pp. 25– 28). Munich, Germany: Ludwig-Maximilians–Universität München.
Tiede, M. K. Masaki, S. Vatikiotis-Bateson, E. 2000). Contrasts in speech articulation observed in sitting and supine conditions. In Hoole, P. (Ed.), Proceedings of the 5th Seminar on Speech Production: Models and data (pp. 25– 28). Munich, Germany: Ludwig-Maximilians–Universität München.×
VideoLAN. ( 2012). VLC Media Player (Version 2.0.5) [Computer software]. Retrieved from www.videolan.org/
VideoLAN. ( 2012). VLC Media Player (Version 2.0.5) [Computer software]. Retrieved from www.videolan.org/×
Figure 1.

Four linear-distance variables in relation to two reference lines on a videofluoroscopic image in the upright position: Reference Line 1 (RL1) = a line connecting the anterior-most edges of the second and third cervical vertebrae; and Reference Line 2 (RL2) = a line connecting the posterior tip of the spinous process of the first cervical vertebra and the superior tip of the anterior tuberculum sellae. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to RL1; Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to RL2.

 Four linear-distance variables in relation to two reference lines on a videofluoroscopic image in the upright position: Reference Line 1 (RL1) = a line connecting the anterior-most edges of the second and third cervical vertebrae; and Reference Line 2 (RL2) = a line connecting the posterior tip of the spinous process of the first cervical vertebra and the superior tip of the anterior tuberculum sellae. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to RL1; Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to RL2.
Figure 1.

Four linear-distance variables in relation to two reference lines on a videofluoroscopic image in the upright position: Reference Line 1 (RL1) = a line connecting the anterior-most edges of the second and third cervical vertebrae; and Reference Line 2 (RL2) = a line connecting the posterior tip of the spinous process of the first cervical vertebra and the superior tip of the anterior tuberculum sellae. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to RL1; Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to RL2.

×
Figure 2.

The means of the four structural variables (Epiglottis-PPW, Tongue-PPW, Velum-RL1, and Velum-RL2) measured in the supine (S) and upright (U) positions during vowel productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.

 The means of the four structural variables (Epiglottis-PPW, Tongue-PPW, Velum-RL1, and Velum-RL2) measured in the supine (S) and upright (U) positions during vowel productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.
Figure 2.

The means of the four structural variables (Epiglottis-PPW, Tongue-PPW, Velum-RL1, and Velum-RL2) measured in the supine (S) and upright (U) positions during vowel productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.

×
Figure 3.

The means of the four structural variables measured in the supine and upright positions during consonant productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.

 The means of the four structural variables measured in the supine and upright positions during consonant productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.
Figure 3.

The means of the four structural variables measured in the supine and upright positions during consonant productions. A statistically significant difference (* p < .05) between supine and upright was found for Velum-RL2 only.

×
Figure 4.

The means of target consonant by adjacent vowel combinations for Epiglottis-PPW during consonant productions.

 The means of target consonant by adjacent vowel combinations for Epiglottis-PPW during consonant productions.
Figure 4.

The means of target consonant by adjacent vowel combinations for Epiglottis-PPW during consonant productions.

×
Figure 5.

The means of target consonant by adjacent vowel combinations for Tongue-PPW during consonant productions.

 The means of target consonant by adjacent vowel combinations for Tongue-PPW during consonant productions.
Figure 5.

The means of target consonant by adjacent vowel combinations for Tongue-PPW during consonant productions.

×
Table 1. Interinvestigator reliability: Cumulative percent frequencies (%) of the differences between the two raters in measurement (mm) for the four structural variables.
Interinvestigator reliability: Cumulative percent frequencies (%) of the differences between the two raters in measurement (mm) for the four structural variables.×
Difference between measurements Epiglottis-PPW Tongue-PPW Velum-RL1 Velum-RL2
≤ 0.5 42.0 27.8 25.0 42.0
≤ 1.0 73.9 59.7 62.5 59.4
≤ 1.5 85.5 73.6 80.6 76.8
≤ 2.0 91.3 87.5 86.2 86.9
≤ 2.5 94.2 90.3 90.4 92.7
≤ 3.0 94.2 95.9 96.0 97.0
≤ 3.5 98.5 100.0 97.3 98.4
≤ 4.0 100.0 98.7 98.4
≤ 4.5 98.7 98.4
≤ 5.0 100.0 100.0
Note. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to Reference Line 1 (RL1); Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to Reference Line 2 (RL2).
Note. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to Reference Line 1 (RL1); Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to Reference Line 2 (RL2).×
Table 1. Interinvestigator reliability: Cumulative percent frequencies (%) of the differences between the two raters in measurement (mm) for the four structural variables.
Interinvestigator reliability: Cumulative percent frequencies (%) of the differences between the two raters in measurement (mm) for the four structural variables.×
Difference between measurements Epiglottis-PPW Tongue-PPW Velum-RL1 Velum-RL2
≤ 0.5 42.0 27.8 25.0 42.0
≤ 1.0 73.9 59.7 62.5 59.4
≤ 1.5 85.5 73.6 80.6 76.8
≤ 2.0 91.3 87.5 86.2 86.9
≤ 2.5 94.2 90.3 90.4 92.7
≤ 3.0 94.2 95.9 96.0 97.0
≤ 3.5 98.5 100.0 97.3 98.4
≤ 4.0 100.0 98.7 98.4
≤ 4.5 98.7 98.4
≤ 5.0 100.0 100.0
Note. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to Reference Line 1 (RL1); Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to Reference Line 2 (RL2).
Note. Epiglottis-PPW = the distance between the tip of the epiglottis and the posterior pharyngeal wall (PPW) along the line perpendicular to Reference Line 1 (RL1); Tongue-PPW = the distance between the tongue and the PPW at the level of the anterior edge of the second cervical vertebra along the line perpendicular to RL1; Velum-RL1 = the perpendicular distance from the velar knee to RL1; and Velum-RL2 = the perpendicular distance from the velar knee to Reference Line 2 (RL2).×
×
Table 2. Intra-investigator reliability: Cumulative percent frequencies (%) of the differences between measurements (mm) made at two different time points for the four structural variables.
Intra-investigator reliability: Cumulative percent frequencies (%) of the differences between measurements (mm) made at two different time points for the four structural variables.×
Difference between measurements Epiglottis-PPW Tongue-PPW Velum-RL1 Velum-RL2
≤ 0.5 49.3 34.7 27.8 47.2
≤ 1.0 63.8 65.3 58.4 68.0
≤ 1.5 76.8 80.6 75.1 88.8
≤ 2.0 85.5 88.9 86.2 94.4
≤ 2.5 88.4 90.3 95.9 98.6
≤ 3.0 92.7 95.9 98.7 98.6
≤ 3.5 95.6 97.3 100.0 100.0
≤ 4.0 95.6 97.3
≤ 4.5 98.5 98.7
≤ 5.0 100.0 100.0
Table 2. Intra-investigator reliability: Cumulative percent frequencies (%) of the differences between measurements (mm) made at two different time points for the four structural variables.
Intra-investigator reliability: Cumulative percent frequencies (%) of the differences between measurements (mm) made at two different time points for the four structural variables.×
Difference between measurements Epiglottis-PPW Tongue-PPW Velum-RL1 Velum-RL2
≤ 0.5 49.3 34.7 27.8 47.2
≤ 1.0 63.8 65.3 58.4 68.0
≤ 1.5 76.8 80.6 75.1 88.8
≤ 2.0 85.5 88.9 86.2 94.4
≤ 2.5 88.4 90.3 95.9 98.6
≤ 3.0 92.7 95.9 98.7 98.6
≤ 3.5 95.6 97.3 100.0 100.0
≤ 4.0 95.6 97.3
≤ 4.5 98.5 98.7
≤ 5.0 100.0 100.0
×
Table 3. Summary data of the structural variables measured at rest and during speech activities: Group means (mm) in the supine and upright positions, mean ratios of supine to upright computed across individual tokens, and standard deviations (in parentheses).
Summary data of the structural variables measured at rest and during speech activities: Group means (mm) in the supine and upright positions, mean ratios of supine to upright computed across individual tokens, and standard deviations (in parentheses).×
Variable At rest During speech
Supine Upright Ratio Supine Upright Ratio
Epiglottis-PPW ‚5.95 (1.97) ‚6.06 (3.18) 0.99 (0.35) 10.49 (2.32) 10.33 (2.32) 1.02 (0.11)
Tongue-PPW 10.64 (3.18) 10.99 (3.63) 1.03 (0.31) 14.45 (2.34) 14.09 (2.39) 1.04 (0.13)
Velum-RL1 15.32 (2.32) 18.41 (2.34) 0.84 (0.11) ‚8.09 (1.95) ‚8.30 (2.82) 1.03 (0.27)
Velum-RL2 30.43 (3.20) 29.41 (3.31) 1.04 (0.10) 24.57 (3.18) 22.75 (2.80) 1.08 (0.11)
Table 3. Summary data of the structural variables measured at rest and during speech activities: Group means (mm) in the supine and upright positions, mean ratios of supine to upright computed across individual tokens, and standard deviations (in parentheses).
Summary data of the structural variables measured at rest and during speech activities: Group means (mm) in the supine and upright positions, mean ratios of supine to upright computed across individual tokens, and standard deviations (in parentheses).×
Variable At rest During speech
Supine Upright Ratio Supine Upright Ratio
Epiglottis-PPW ‚5.95 (1.97) ‚6.06 (3.18) 0.99 (0.35) 10.49 (2.32) 10.33 (2.32) 1.02 (0.11)
Tongue-PPW 10.64 (3.18) 10.99 (3.63) 1.03 (0.31) 14.45 (2.34) 14.09 (2.39) 1.04 (0.13)
Velum-RL1 15.32 (2.32) 18.41 (2.34) 0.84 (0.11) ‚8.09 (1.95) ‚8.30 (2.82) 1.03 (0.27)
Velum-RL2 30.43 (3.20) 29.41 (3.31) 1.04 (0.10) 24.57 (3.18) 22.75 (2.80) 1.08 (0.11)
×
Table 4. Summary data of the means and standard deviations (in parentheses) in the supine and upright positions for the first formant (F1) and second formant (F2) in Hz across different target vowels.
Summary data of the means and standard deviations (in parentheses) in the supine and upright positions for the first formant (F1) and second formant (F2) in Hz across different target vowels.×
Position Vowel /a/ Vowel /i/ Vowel /u/
Supine Upright Supine Upright Supine Upright
F1 913.64 (161.27) 887.71 (161.76) 371.18 (51.20) 359.89 (51.57) 427.40 (72.82) 403.03 (63.54)
F2 1481.70 (193.91) 1492.15 (227.81) 2553.21 (315.63) 2537.59 (294.05) 1360.18 (243.46) 1345.51 (247.37)
Table 4. Summary data of the means and standard deviations (in parentheses) in the supine and upright positions for the first formant (F1) and second formant (F2) in Hz across different target vowels.
Summary data of the means and standard deviations (in parentheses) in the supine and upright positions for the first formant (F1) and second formant (F2) in Hz across different target vowels.×
Position Vowel /a/ Vowel /i/ Vowel /u/
Supine Upright Supine Upright Supine Upright
F1 913.64 (161.27) 887.71 (161.76) 371.18 (51.20) 359.89 (51.57) 427.40 (72.82) 403.03 (63.54)
F2 1481.70 (193.91) 1492.15 (227.81) 2553.21 (315.63) 2537.59 (294.05) 1360.18 (243.46) 1345.51 (247.37)
×