Relationships Between Vocal Stand Craniocervical Posture InvesMagnetic Resonance Imaging

*Nicola A. Miller, Jennifer S. Gregory, Scott I. K. Sempland *Fiona J. Gilbert, *yzxScotland, United Kingdom

Summary: Objectives. Traditional voice research focusestheir direct/indirect attachments (skull, cervical spine, and steinvestigate vocal structures within this wider context and asse

desiathinblesbetw

cal sas afrom(lowsho

ocalionaay w

ostu

INTROSpeectimelysystemlengednisms.couldeffectiof magful toosee thein threreductof MRTrad

vocal

alate;ds tond/orfacialthat

shipsvocalsture,irwayvocalcouldments

sure-upperng tol sur-disci-X-ray

is quicker and cheaper than MRI but lacks its superior soft tis-rations limit therticularly whereired.17 Recently,y dominated by

Address correspondence and reprint requests to Nicola A. Miller, Aberdeen Biomedicalsue definition. Additionally, ethical consideusefulness of cephalometry in research, pamatched controls or repeated images are requthe potential of MRI in fields traditionall

Imaging Centre, Lilian Sutton Building, Foresterhill, Aberdeen, Aberdeenshire AB252ZD, United Kingdom E-mail: n.a.miller@abdn.ac.ukJournal of Voice, Vol. 26, No. 1, pp. 102-1090892-1997/$36.00 2012 The Voice Foundationdoi:10.1016/j.jvoice.2010.10.016radiography, is an established method whereby soft tissuedimensions can be related to bony landmarks.16 Accordingly,the variables of interest extend beyond the vocal tract and artic-ulators to other regions within the head and neck. Cephalometry

Accepted for publication October 26, 2010.From the *Aberdeen Biomedical Imaging Centre, University of Aberdeen, Aberdeen,

Scotland, United Kingdom; yBone and Musculoskeletal Programme, Institute of MedicalSciences, University of Aberdeen, Aberdeen, Scotland, United Kingdom; zClinicalResearch Imaging Centre, Queens Medical Research Institute, University of Edinburgh,Edinburgh, Scotland, United Kingdom; and the xDepartment of Music, University ofAberdeen, Aberdeen, Scotland, United Kingdom.DUCTIONh is a complex behavior requiring rapid, precise, andcoordination of numerous components within the vocal.1 The largely hidden nature of this system has chal-progress toward understanding its underlying mecha-

2 Such an understanding is important because thislead to improved vocal performance in health and moreve interventions and therapies in disease. The applicationnetic resonance imaging (MRI) in 1987 as a safe and use-l in speech research meant that for the first time we couldsoft tissue outline of the entire vocal tract (glottis to lips)e dimensions.3 Since then, advances in technology andion in image acquisition time have secured the positionI as a leading imaging tool in voice research.itionally, voice research using MRI has focused on thetract and its appendages, such as the piriform fossae; the

articulators, particularly the lips, jaws, tongue and soft pand the voice source, the larynx.4,5 However, this tenignore the fact that all these structures have direct aindirect structural and functional links to the cranioskeleton, cervical spine, and sternum. It followsadjustments to these structural and functional relationcan lead to quantitative changes in variables describingstructures and airway dimensions. Alterations of head pofor example, are strongly correlated with changes of adimensions.612 We hypothesized that consideringstructures within the context of these wider relationshipslead to a better understanding of the coordinated adjustthat underpin activity within the vocal system.Established protocols are not yet available for these mea

ments in MRI.8,13,14 However, the dimensions of theairway (glottis to nasopharynx) and factors contributithese are also of interest in orthodontics,15 maxillofaciagery,11 and obstructive sleep apnea research.8 In theseplines, lateral cephalometry, traditionally usingduction studies.Study Design/Method. Using a cross-sectional studyadults (five males and five females) while at rest and bre17 craniocervical, craniocaudal, and anteroposterior variacervical posture, and airway dimensions. Relationshipscoefficient.Results. We found widespread correlations relating vo(r > 0.6). Increasing airway size (hyocervical distance) wthe hyoid, larynx, epiglottis tip and uvula tip, and of C3ciated with a shorter and higher soft palate, and a greateryngeal tube opening, narrower airway at the uvula tip andbase.Conclusion. Finding widespread correlations relating vconfirms the potential of this approach to uncover functimportance of considering vocal structures and the airwto be missed.Key Words: MRICephalometryVocal tractSpeechPructures, the Airway,tigated Using

e, Richard M. Aspden, Peter J. Stollery,

on the vocal tract, articulators, and larynx. By ignoringrnum) important information may be missed. We aim toss the validity of this approach for subsequent voice pro-

gn, we obtained midsagittal MR images from 10 healthyg quietly. With reference points based on cephalometry,were chosen to describe craniofacial morphology, cranio-een variables were sought using Pearsons correlation

tructures to the craniofacial skeleton and cervical spinessociated with greater distances from the cranial base ofthe menton. Awider velopharyngeal opening was asso-er) craniocervical angle was associated with a wider lar-rter distances of the hyoid and uvula tip from the cranial

structures to the craniofacial skeleton and cervical spinel activity during voice production and demonstrates theithin this wider context if important information is not

re.

coil (Philips Healthcare). Deformable foam wedges were used adopted a hyperextended neck position and was unable to

nial base, craniofacial skeleton, and cervical spine (Table 4).

Nicola A. Miller, et al Relationships Between Vocal Structures, the Airway, and Craniocervical Posture 103duction studies. The FOVextended from just above the pituitarygland to the sternal notch, taking in the width of the whole headand neck. Each individual was imaged with a 20 second acqui-sition while breathing quietly. The MRI slice closest to the mid-sagittal planewas chosen for analysis (identified by the presenceof the pituitary fossa, the tip of the odontoid process, the outlineof the trachea and spinal cord, and the spinal processes).

Image analysisImages were converted from Digital Imaging and Communi-cations in Medicine to Bitmap format using ImageJ (U.S.to make the volunteer comfortable and to restrain the head posi-tion. Earplugs and headphones helped attenuate the scannernoise and allowed two-way communication. Volunteers were re-quired to adopt a relaxed posture in the MRI scanner and wereinstructed to look straight ahead while holding the lips and teethtogether and to rest the tongue dorsum comfortably against thehard palate. Parasagittal imageswere obtained using a turbo spinecho pulse sequencewith the following parameters: field of view(FOV) 3403 340 mm; a 7683 768 matrix; repetition time of4,106 milliseconds; echo time of 100 milliseconds; 6 slices4.0 mm thick with a gap of 1.0 mm centered on the midsagittalplane. The number of slices was dictated by the need to optimizeimage resolution within the time constraint of a single breathhold thus allowing for comparisons with subsequent vocal pro-duction where the aim is to relate changes in vocal structures totheir wider anatomical links within the head and neck. The aimsof this pilot study are to measure craniocervical, craniocaudal,and anteroposterior dimensions based on reference points usedin cephalometry and to assess the potential of this approach touncover functional changes during subsequent experimentsinvestigating voice production.

METHODS

RecruitmentTwelve healthy volunteers were recruited to the study. All butone (who had a tonsillectomy and adenoidectomy as a child)had no history of speech or hearing pathology. Exclusion crite-ria included a history of claustrophobia, an inability to maintaina closed mouth position within the 20-seconds time frame nec-essary for image acquisition during this and subsequent parts ofthe study and the presence of contraindications to MRI, such aspacemakers and metallic orthodontic appliances. Approvalfrom Grampian Research Ethics Committee (now North ofScotland Research Ethics Service) was obtained, and all sub-jects gave written informed consent.

ProcedureUsing a 3.0T Achieva MRI system (Philips Healthcare, Best,The Netherlands), volunteers were imaged in a supine positionwith the head placed in a Sense-Neurovascular array-16 elementcephalometry has resulted in calls to validate and standardizeMRI protocols.13,14,18

This study is part of a larger MRI investigation of voice pro-National Institutes of Health, Bethesda, MD).19 Because theTo assist in their visual representation, uncorrelated variableswere omitted from the matrix and the remainder arranged, asfar as possible, to demonstrate relationships between them.Two distinct patterns of correlations were noted: two overlap-ping groups of correlations, where variables within each groupare correlated with each other (r 0.71) and correlations as-sociated with individual variables. The variables singled outfor further discussion are shown in bold italics.Three main groups of correlations were observed; those

based around the hyocervical distance (hy-c3), the velophar-yngeal opening (VPO) (u-ppw), and the craniocervical angle(evt/nsl). These are illustrated in Figure 2. Topographical corre-lations (where variables share reference points or lines) wereobserved between variables that reflect the overall size of thecraniofacial skeleton and airway and between craniocervicaltake part in subsequent experiments where subjects wererequired to produce voice comfortably and without unduestrain. We obtained a full data set for the remaining 10 subjects(five males, five females; age range 2047 with a median of 25years). Table 3 contains the descriptive statistics for the studypopulation (mean, standard deviation, and range).We observed widespread correlations between the cranio-

cervical, craniocaudal, and anteroposterior variables relatingthe larynx, hyoid, epiglottis, soft palate and airway to the cra-aim of subsequent experiments is to investigate the nature of ad-justments that occur during voice production, the choice of vari-ables was dictated by the need to capture adjustments secondaryto any changes in craniocervical posture, craniocaudal, or ante-roposterior dimensions. Software tools developed by the Uni-versity of Manchester, United Kingdom20 were used to markreference points as shown in Figure 1 and described in Table1. From these points, a program was written to automaticallymeasure the 17 variables listed in Table 2 and illustrated inFigure 1. Small case letters are used to identify landmarksused in MRI to distinguish them from those used in the cepha-lometric literature because they may not be directly compara-ble. In this study, C3 refers to the body of the third cervicalvertebra whereas c3 refers to its most anteroinferior point.

Statistical analysisStatistical analysis was performed using Sigmastat (Version 11;Systat Software, Inc., San Jose, CA). Descriptive statistics wereobtained and expressed as mean (standard deviation) and rela-tionships between variables were sought using Pearsons corre-lation coefficient. Students t test was used to detect possiblesignificant differences between the means of variables describ-ing hyoid position in males and females. For all tests, a P valueof 0.05 or less was taken to indicate statistical significance.

RESULTSTen of 12 subjects met the inclusion criteria for the study. Onesubject was unable to adopt the necessary tongue position withthe tongue dorsum resting against the hard palate. Anotherangles at the uppermost part of the cervical spine. The

c2

Journal of Voice, Vol. 26, No. 1, 2012104 ans pns cv2tg

cv2ip

optcvt

evt

u

n

s

nsl

ut

n

pns

s

A

B Chyocervical distance was strongly and positively correlatedwith the perpendicular distances from the nasion-sella line(NSL) of the hyoid (hy-nsl), larynx (l-nsl), uvula tip (ut-nsl),and epiglottis tip (et-nsl) and with the width of the oropharyn-geal airway at the uvula tip (pt-ppw-ut) (r 0.81, P < 0.01).Weaker positive correlations were observed for the anteroposte-rior distance between C3 and the menton (c3-me) (r 0.64,P < 0.05). Figure 2A shows that, on average, an increase in cra-niofacial dimensions is associated with an increase in airwaysize. Airway size at the epiglottis tip (pt-ppw-et) is correlatednot with these variables but with c3-me (r 0.69, P < 0.05).However, pt-ppw-ut and pt-ppw-et are positively correlatedwith each other (r 0.76, P < 0.05). The craniocervical anglesopt/nsl and cvt/nsl were very highly correlated (r 0.96,P < 0.001).Nontopographical correlations (where variables have no

common reference points or lines) and topographical correla-tions were observed between dimensions of the VPOand between variables associated with the craniocervicalangle evt/nsl. The narrowest part of the VPO, the minimal dis-tance separating the uvula from the posterior pharyngeal wall(u-ppw), was negatively correlated with ut-nsl (r0.72,P < 0.05) and the length of the soft palate (pns-ut) (r0.83,

cv4ip

cv6ip

utet

c4

c

hy

l

ster

FIGURE 1. Variables describing craniocervical, angular, craniocaudal,image. B. Craniocervical and angular variables. C. Craniocaudal variables.nsl

ppwu

DP < 0.01). Figure 2B shows that, on average, a wider VPO isassociated with a higher and shorter soft palate. Conversely,a narrowerVPO is associatedwith a lower and longer soft palate.The craniocervical angle evt/nsl, influenced by alignmentof C46, was correlated negatively with hy-nsl, ut-nsl, andpt-ppw-ut, and positively with the width of the laryngeal open-ing (ltw) (all correlations, r 0.63, P < 0.05). Figure 2C showswidening of evt/nsl to be associated, on average, with shorterperpendicular distances of the hyoid and soft palate tip fromthe cranial base, narrowing of the oropharyngeal airway at theuvula tip, and widening of the laryngeal tube opening. Con-versely, a reduction of evt/nsl was associated with greater dis-tances of the hyoid and soft palate tip from the cranial base,widening of the oropharyngeal airway at the uvula tip, and nar-rowing of the laryngeal tube opening. The craniocervical anglecvt/nsl, influenced by upper cervical alignment (C24), was alsopositively and nontopographically correlated with ltw.We found significant differences between themean values for

hy-c3 and hy-nsl for males (39.1 mm [2.0] and 115.2 mm [5.9],respectively) and females (34.2 mm [3.3] and 97.6 mm [8.2],respectively), for both groups (P 0.032 and 0.005, respec-tively) indicating that, on average, the hyoid occupied a moreposterosuperior position in females compared with males.

6

n(um)

me

hy

a

c3 pt

ltw

utet

and anteroposterior dimensions. A. Midsagittal magnetic resonance

D. Anteroposterior variables.

ior most aof sttis tal veintost iost iost iost athrothroteroteriealntt ovjoinithroost pior tior pint om

Nicola A. Miller, et al Relationships Between Vocal Structures, the Airway, and Craniocervical Posture 105TABLE 1.Bony and Soft Tissue Reference Points and Planes

Reference Point

a Superans The masp Angleet Epiglocv Cerviccv2tg The pocv2ip The mcv4ip The mcv6ip The mc3 The mcvt A lineevt A linehy The anl The anltw Laryngme Menton A poinnsl A lineopt A linepns The mpt Posterppw Posters Midpostern SternuDISCUSSIONIn this study, the superior soft tissue definition ofMRI was com-bined with bony reference points commonly used in cephalom-etry to investigate the vocal tract and related structures withinthe context of their direct and indirect structural attachmentsto the craniofacial skeleton, cervical spine, and sternum. Our re-sults demonstrate the validity of this method for quantitativeanalysis of vocal tract-related dimensions and demonstratethe potential of this approach to uncover functional correlationsbetween vocal tract-related structures and airway size in subse-quent studies investigating voice production.Cephalometry is generally performed in erect subjects.

Because moving from erect to supine can result in changes ofdimensions,21 the present study does not allow a direct compar-ison of all our results with equivalent cephalometric findings.However, unlike other variables, hy-c3 remains unchanged inthemove from an erect to a supine posture,21 and themeanvaluefor hy-c3 fell within the range reported for cephalometry.2224

Additionally, our demonstration that the hyoid occupiesa more posterosuperior position in females compared withmales is consistent with cephalometric reports.24,25 Together,these findings support earlier assertions that cephalometricvariables can be successfully adapted for use in quantitativeMRI investigations of vocal structures and the airway.The power of this approach to uncover functional correla-

tions during subsequent studies investigating voice productionis demonstrated by the discovery of a group of correlations

u Uvulaut Uvula tipDefinition

argin of arytenoid cartilagenterior point of maxilla at level of hard palateoft palateiprtebraat the superior extremity of the odontoid process of cv2nferoposterior point on the body of cv2nferoposterior point on the body of cv4nferoposterior point on the body of cv6nteroinferior point on the body of cv3ugh cv2tg and cv4ipugh cv4ip and cv6ipsuperior margin of outer cortex of hyoid boneor point of vocal foldstube widthhe most inferior point of bony chinerlying the nasionng n and s reflecting orientation of anterior cranial baseugh cv2tg and cv2iposterior point of the hard palateongueharyngeal wallf sella turcicaobserved in association with the lower craniocervical angleevt/nsl (Figure 2C). The effect of these correlated dimensionsis clearly illustrated in tracings of images obtained from sub-jects (both females) possessing the greatest and the smallestcraniocervical angle evt/nsl (Figure 3). Volunteer 5 had thegreatest angle evt/nsl and the shortest distance hy-nsl amongstall volunteers, whereas volunteer 4 had the smallest angle evt/nsl and the second greatest distance hy-nsl amongst all volun-teers but the greatest distance amongst female volunteers. Involunteer 4, evt/nsl is associated with straightening of the nor-mal cervical curvature, a greater hyoid-cranial base distance,widening of the oropharyngeal airway at the uvula tip (x), nar-rowing of the laryngeal tube opening (y), and a more posteriortongue position (Figure 3A). In marked contrast, evt/nsl involunteer 5 is associated with extension of the head, lordosisof the cervical spine, shortening of the hyoid-cranial base dis-tance, narrowing of the oropharyngeal airway at the uvulatip (x), and widening of the laryngeal tube opening (y)(Figure 3B).These findings are important because nontopographical asso-

ciations, such as those observed between evt/nsl and ltw, pointto the presence of underlying growth coordinating mechanisms,which contribute to the development of an individuals intrin-sic shape and facial appearance.26 Although coordinated pat-terns of growth-related changes associated with the uppercraniocervical angles (C24) have been observed previously,26

as far as we are aware this is the first study to demonstrate

TABLE 2.Craniocervical, Angular, Craniocaudal, and Anteroposterior Dim

Variables

Craniocervicalcvt/nsl The angle between cvt and nsevt/nsl The angle between evt and nsopt/nsl The angle between opt and n

Angularasp The angle between lines joini

Craniocaudalet-nsl The perpendicular distance ofhy-nsl The perpendicular distance ofl-nsl The perpendicular distance ofpns-ut The distance between posteristern-hy The distance from sternum tout-nsl The perpendicular distance of

Anteroposteriorc3-me The anteroinferior point of C3hy-c3 The anterosuperior point of hhy-me The anterosuperior point of hltw The superior margin of aryten

osteostewee

Journal of Voice, Vol. 26, No. 1, 2012106similar patterns of coordinated changes in association with thelower craniocervical angle (C46).

pt-ppw-et The posterior tongue to ppt-ppw-ut The posterior tongue to pu-ppw The minimal distance betIn 1941, Brodie27 showed that during normal growth of thecraniofacial skeleton, an increase in the size of one variable isaccompanied by a proportionate increase in the size of other(bony) variables. In Figure 2A, we show that, on average, the

TABLE 3.Descriptive Statistics: Mean (SD) Range

Variables Mean (SD) Range

opt/nsl 102.2 (7.0) 93115.6cvt/nsl 101.9 (7.0) 92.8115evt/nsl 103.3 (7.7) 90.9118.6asp 132.5 (5.8) 121.8139.3pns-ut 39.3 (5.7) 30.650.0u-ppw 4.1 (2.4) 0.07.0ut-nsl 71.5 (8.8) 59.884.8et-nsl 87.9 (9.1) 76.3101.0hy-nsl 106.4 (11.5) 89.7125.1l-nsl 129.0 (15.1) 111.2152hy-c3 36.7 (3.7) 30.641.2c3-me 82.8 (7.8) 75.794.7hy-me 47.2 (6.4) 38.357.6ltw 9.0 (2.6) 4.912.5stern-hy 110.1 (15.2) 81.4135.6pt-ppw-ut 8.6 (2.5) 5.7217.2pt-ppw-et 13.2 (2.8) 9.717.2

All linear measurements are in millimeter.larger the individual, the greater the dimensions relating tothe bony and soft tissue structures of the head and neck and

ensions

Definition

llsl

ng anspns and pns-ut

epiglottis tip from nasion-sella linehyoid bone from nasion-sella linelarynx from nasion-sella lineor nasal spine and uvula tipanterosuperior point of hyoid bodyuvula tip from nasion-sella line

to mentonyoid body to anteroinferior point of C3yoid body to mentonoid cartilage to base of epiglottis, perpendicular to airwayrior pharyngeal wall at epiglottis tiprior pharyngeal wall at uvula tipn uvula and posterior pharyngeal wallthe greater the size of the airway. However, correlationsobserved between the size of the VPO and the height and lengthof the soft palate (Figure 2B), and between the lower craniocer-vical angle, craniofacial morphology, and airway dimensions(Figure 2C), suggest that development of adult morphologydepends, in part, on the presence of coordinated underlyinggrowth-related mechanisms.Knowledge and awareness of underlying patterns of head and

neck development has important clinical and research implica-tions. For example, the factors that contribute to the immensevariability observed between subjects in voice research arestill not fully understood.28 In part, these may be because ofanatomical variations, such as differences in vocal tractlength.29 However, the results of this and other studies pointingto the presence of coordinated patterns of growth affecting headand neck development3032 suggest that improved knowledgeand awareness of underlying global patterns of head and neckdevelopment might lead to a better understanding of factorscontributing to normal and pathological variations ofstructural (and functional) relationships between bony andsoft tissues. Such knowledge could lead to improved vocalperformance in health, more effective interventions andtherapies for those with speech difficulties, and more realisticautomatic speech synthesis.Interpretation of our results is limited by the small sample

sizewith mixed ages, sexes, and varying subject heights, all fac-tors known to influence vocal tract dimensions. However, find-ing highly correlated patterns is indicative of fundamental

TABLE 4.Correlation Matrix for Craniocervical, Craniocaudal, and Anteroposterior Variables

Variables cvt/nsl evt/nsl hy-c3 hy-nsl l-nsl et-nsl ut-nsl pns-ut u-ppw pt-ppw-ut pt-ppw-et ltw c3-me hy-me

opt/nsl 0.96* 0.43 0.13 0.26 0.27 0.25 0.38 0.26 0.01 0.20 0.18 0.56 0.38 0.49cvt/nsl 1.00 0.59 0.3 0.43 0.4 0.38 0.54 0.30 0.04 0.33 0.2 0.71y 0.3 0.48evt/nsl 0.59 1.0 0.51 0.63y 0.46 0.55 0.67y 0.23 0.36 0.64y 0.32 0.74y 0.05 0.1hy-c3 0.3 0.51 1.00 0.87z 0.81z 0.83z 0.81z 0.6 0.44 0.81z 0.62 0.47 0.64y 0.27hy-nsl 0.43 0.63y 0.87z 1.00 0.95* 0.97* 0.74y 0.44 0.34 0.68y 0.37 0.51 0.34 0.02l-nsl 0.4 0.46 0.81z 0.95* 1.00 0.96* 0.71y 0.49 0.34 0.57 0.33 0.32 0.26 0.07et-nsl 0.38 0.55 0.83z 0.97* 0.96* 1.00 0.76y 0.58 0.47 0.56 0.31 0.38 0.26 0.09ut-nsl 0.54 0.67y 0.81z 0.74y 0.71y 0.76y 1.00 0.80z 0.72y 0.60 0.53 0.47 0.23 0.17pns-ut 0.30 0.23 0.6 0.44 0.49 0.58 0.80z 1.00 0.83z 0.2 0.44 0.10 0.19 0.14u-ppw 0.04 0.36 0.44 0.34 0.34 0.47 0.72y 0.83z 1.00 0.24 0.49 0.09 0.20 0.06pt-ppw-ut 0.33 0.64y 0.81z 0.68y 0.57 0.56 0.60 0.2 0.24 1.00 0.76y 0.48 0.6 0.35pt-ppw-et 0.2 0.32 0.62 0.37 0.33 0.31 0.53 0.44 0.49 0.76y 1.00 0.13 0.69y 0.49ltw 0.71y 0.74y 0.47 0.51 0.32 0.38 0.47 0.10 0.09 0.48 0.13 1.00 0.11 0.01c3-me 0.3 0.05 0.64y 0.34 0.26 0.26 0.23 0.19 0.20 0.6 0.69y 0.11 1.00 0.89*hy-me 0.48 0.1 0.27 0.02 0.07 0.09 0.17 0.14 0.06 0.35 0.49 0.01 0.89* 1.00Statistically significant correlations are shown in bold.

Nicola A. Miller, et al Relationships Between Vocal Structures, the Airway, and Craniocervical Posture 107relationships between these structures and requires furtherstudy. We suggest that the method used here offers greaterpotential for improved quantitative and qualitative informationconcerning structural (and functional) relationships betweenbony, airway, and soft tissue variables than is possible withthe use of cephalometry alone. Although MRI is more expen-sive than cephalometry, its lack of nonionizing radiationand superior soft tissue definition offer clear advantages forresearch purposes, particularly where matched controls ormultiple image acquisitions are required. Not all cephalometric

* Level of significance P < 0.001.y Level of significance P < 0.05.z Level of significance P < 0.01.measurements can be adapted for MRI as some bony land-marks, such as the gonion (a point at the bisection of the inferiorand posterior borders of the mandible), are not present in the

nsl

hyc3

ppwu

A B

FIGURE 2. Significant correlations associated with A. hy-c3, B. u-ppw, anative correlations by a dashed line.midsagittal plane. In some instances, adaptations permit theuse of an equivalent landmark. For example, a carefully placedpoint overlying the nasion, rather than the nasion itself, allowsuse of a line equivalent to the NSL. More work is necessarybefore the reliability of this method can be established. How-ever, previous work has shown MRI results to be highly repro-ducible and stable over time.13 Only one observer annotated theimages but findings of low intrainvestigator variability com-pared with interinvestigator variability have led to the sugges-tion that preference should be given to such intrainvestigator

evaluation when comparing a series of MR images or for studypurposes.13 Future studies that include the use of a positionalMRI scanner, which permits imaging in erect and supine

nsl nsln

s

evt

c4

c6

C

d C. evt/nsl. Positive correlations are indicated by a solid line and neg-

BB. l

Journal of Voice, Vol. 26, No. 1, 2012108subjects, could further understanding of structural and func-tional relationships between bony and soft tissues.

CONCLUSIONSWe aimed to investigate vocal structures within the context oftheir direct and indirect structural links to the craniofacial skel-eton, cervical spine, and sternum by combining MRIs superiorsoft tissue definition with bony reference points used in cepha-lometry. Observations of widespread patterns of correlationslinking craniocervical, craniocaudal, and anteroposteriordimensions support earlier work pointing to the dependenceof structural relationships on underlying growth coordinatingmechanisms. Our results demonstrate the potential of thisapproach to offer valid qualitative and quantitative information

evt/nsl

cv4

cv6

hy

utx

y

A

FIGURE 3. Comparison between volunteers with the A. smallest andand y indicates airway size at laryngeal tube opening).in subsequent experiments investigating voice productionand, further, that important information may be missed ifvocal structures are considered without taking into accounttheir wider structural relationships within the head and neck.Recognition of and accounting for these wider associationshas important clinical and research implications and couldpotentially lead to a better understanding of mechanisms under-lying structural and functional coordinated activity within thisregion. In turn, this could lead to improved interventions andtherapies across disciplines.

AcknowledgmentsWe thank the magnetic resonance imaging radiographers forhelp with imaging, the volunteers for freely donating theirtime, and the University of Aberdeen for sponsoring this study.

REFERENCES1. Kent RD. The uniqueness of speech among motor systems. Clin Linguist

Phon. 2004;18:495505.

2. Munhall K. Functional imaging during speech production. Acta Psychol

(Amst). 2001;107:95117.

3. Baer T, Gore JC, Boyce S, Nye PW. Application of MRI to the analysis of

speech production. Magn Reson Imaging. 1987;5:17.4. Honda K, Takemoto H, Kitamura T, Fujita S, Takano S. Exploring human

speech production mechanisms by MRI. IEICE Trans Inf Syst. 2004;87:

10501058.

5. Ventura SR. Application of MRI and biomedical engineering in speech

production study. Comput Methods Biomech Biomed Engin. 2009;12:

671681.

6. Hellsing E. Changes in the pharyngeal airway in relation to extension of the

head. Eur J Orthod. 1989;11:359365.

7. Pae EK, Lowe AA, Sasaki K, Price C, Tsuchiya M, Fleetham JA. A ceph-

alometric and electromyographic study of upper airway structures in the

upright and supine positions. Am J Orthod Dentofacial Orthop. 1994;

106:5259.

8. Solow B, Skov S, Ovesen J, Norup PW, Wildschidtz G. Airway dimen-

sions and head posture in obstructive sleep apnoea. Eur J Orthod. 1996;

18:571579.

9. Muto T, Takeda S, Kanazawa M, Yamazaki A, Fujiwara Y, Mizoguchi I.

nslevt/nsl

ut

hycv4

cv6

x

y

argest craniocervical angle evt/nsl (x indicates airway size at uvula tipThe effect of head posture on the pharyngeal airway space (PAS). Int

J Oral Maxillofac Surg. 2002;31:579583.

10. Muto T, Yamazaki A, Takeda S, Kawakami J, Tsuji Y, Shibata T,

Mizoguchi I. Relationship between the pharyngeal airway space and cra-

niofacial morphology, taking into account head posture. Int J Oral Maxil-

lofac Surg. 2006;35:132136.

11. Muto T, Yamazaki A, Takeda S. A cephalometric evaluation of the pharyn-

geal airway space in patients with mandibular retrognathia and prognathia,

and normal subjects. Int J Oral Maxillofac Surg. 2008;37:228231.

12. Anegawa E, TsuyamaH, Kusukawa J. Lateral cephalometric analysis of the

pharyngeal airway space affected by head posture. Int J Oral Maxillofac

Surg. 2008;37:805809.

13. Stuck BA, Kopke J, Maurer JT, Verse T, Kuciak G, Duber C, Hormann K.

Evaluating the upper airway with standardized magnetic resonance imag-

ing. Laryngoscope. 2002;112:552558.

14. Stuck BA, Maurer JT. Airway evaluation in obstructive sleep apnea. Sleep

Med Rev. 2008;12:411436.

15. Haralabakis NB, Toutountzakis NM, Yiagtzis SC. The hyoid bone position

in adult individuals with open bite and normal occlusion. Eur J Orthod.

1993;15:265271.

16. Broadbent BH. A new x-ray technique and its application to orthodontia.

Angle Orthod. 1931;1:4566.

17. Tangugsorn V, Skatvedt O, Krogstad O, Lyberg T. Obstructive sleep

apnoea: a cephalometric study. Part I. Cervico-craniofacial skeletal mor-

phology. Eur J Orthod. 1995;17:4556.

18. Leggett B, Murray M, Buzzell J, Leggitt V, Kim JS. Validation of a non-

ionizing radiation (MRI) technique of cephalometric analysis. Poster,

International Association for Dental Research Conference March 811,

2006, Orlando, FL.

19. Rasband WS. ImageJ. Bethesda, MD: U.S. National Institutes of Health;

19972009. Available at: http://rsb.info.nih.gov/ij/. Accessed August 26,

2008.

20. Available at: http://personalpages.manchester.ac.uk/staff/timothy.f.cootes/

software/am_tools_doc/index.html. Accessed August 22, 2008.

21. Sutthiprapaporn P, Tanimoto K, Ohtsuka M, Nagasaki T, Iida Y,

Katsumata A. Positional changes of oropharyngeal structures due to gravity

in the upright and supine positions. Dentomaxillofac Radiol. 2008;37:

130135.

22. Muto T, Kanazawa M. Positional change of the hyoid bone at maximal

mouth opening. Oral Surg Oral Med Oral Pathol. 1994;77:451455.

23. Hoekema A, Hovinga B, Stegenga B, De Bont LG. Craniofacial morphol-

ogy and obstructive sleep apnoea: a cephalometric analysis. J Oral Rehabil.

2003;30:690696.

24. Sahin Saglam AM, Uydas NE. Relationship between head posture and

hyoid position in adult females and males. J Craniomaxillofac Surg.

2006;34:8592.

25. Marsan G. Head posture and hyoid bone position in adult Turkish Class III

females and males. World J Orthod. 2008;9:391398.

26. SolowB, Tallgren A. Head posture and craniofacial morphology.Am J Phys

Anthropol. 1976;44:417435.

27. Brodie AG. On the growth pattern of the human head. From the third month

to the eighth year of life. Am J Anat. 1941;68:209262.

28. Moore RK. Spoken language processing: piecing together the puzzle.

Speech Commun. 2007;49:418435.

29. FitchWT, Giedd J. Morphology and development of the human vocal tract:

a study using magnetic resonance imaging. J Acoust Soc Am. 1999;

106(3 Pt 1):15111522.

30. Houston WJB. Mandibular growth rotations-their mechanisms and impor-

tance. Eur J Orthod. 1988;10:369373.

31. Solow B, Sandham A. Cranio-cervical posture: a factor in the develop-

ment and function of the dentofacial structures. Eur J Orthod. 2002;24:

447456.

32. Sonnesen L, Pedersen CE, Kjaer I. Cervical column morphology related to

head posture, cranial base angle, and condylar malformation. Eur J Orthod.

2007;29:398403.

Nicola A. Miller, et al Relationships Between Vocal Structures, the Airway, and Craniocervical Posture 109

Relationships Between Vocal Structures, the Airway, and Craniocervical Posture Investigated Using Magnetic Resonance ImagingIntroductionMethodsRecruitmentProcedureImage analysisStatistical analysis

ResultsDiscussionConclusionsAcknowledgmentsReferences