acoustic comparison of vowel sounds produced before and after orthognathic surgery for mandibular...
TRANSCRIPT
J6
Arsvttst
U
D
T
T
H
F
Oral Maxillofac Surg4:910-916, 2006
Acoustic Comparison of Vowel SoundsProduced Before and After Orthognathic
Surgery for Mandibular AdvancementMatti Niemi, DDS,* Juha-Pertti Laaksonen, MA,†
Timo Peltomäki, DDS, PhD,‡ Jukka Kurimo, DDS,§
Olli Aaltonen, PhD,� and Risto-Pekka Happonen, DDS, PhD¶
Purpose: The effects of orthognathic surgery on the phonetic quality of speech were studied byanalyzing the main acoustic features of vowel sounds.
Patients and Methods: Five men with dentofacial deformities undergoing surgical operation forcorrection of malocclusion were enrolled in the study. The speech material consisted of 8 vowels insentence context. Every utterance was repeated 10 times in 3 different sessions: before the operation, 6weeks after the operation, and 30 weeks after the operation. The acoustic features (F1, F2, F0, duration)of vowels were measured and analyzed.
Results: At the group level, no significant acoustic changes were found between the 3 differentsessions in any parameter measured (all F values �1). The results show that the operation had individualand variable effects on vowel quality, ranging from slightly affected to completely unaffected. The 2lowest vocal-tract resonances changed in frequency for 2 of the subjects, and 1 subject had short-termchanges returning to the presurgical level. Significant changes of F0 were observed for 1 subject, and 3of the subjects had short-term changes. No significant changes were found for duration. One subject hadno significant changes in any parameter measured.
Conclusions: No long-lasting perceptually significant changes were identified in vowel productionin patients undergoing a variety of orthognathic procedures. The facial skeleton (consisting of palate,maxilla, mandible, dentition, nasal cavity, etc) imposes direct limits on the morphology of theresonating vocal tract cavities, and is therefore of immediate relevance to both speech articulationand acoustics.© 2006 American Association of Oral and Maxillofacial Surgeons
J Oral Maxillofac Surg 64:910-916, 2006pa
oompp
v
T
D
L
p
©
0
ny organic change to the vocal apparatus (eg, as theesult of a surgical operation) may have implications forpeech production if it alters the morphology of theocal organs and the resonating cavities or the consis-ency and mechanical properties of the vocal apparatusissues. However, a number of studies have demon-trated that speakers can compensate the structural al-erations in the vocal tract, at least to some degree, and
*PhD Student, Department of Oral and Maxillofacial Surgery,
niversity of Turku, Turku, Finland; Resident, Department of Oral
iseases, Tampere University Hospital, Tampere, Finland.
†PhD Student, Department of Phonetics, University of Turku,
urku, Finland.
‡Senior Lecturer, Department of Orthodontics, University of
urku, Turku, Finland.
§Orthodontist, Department of Oral Diseases, Tampere University
ospital, Tampere, Finland.
�Professor, Department of Phonetics, University of Turku, Turku,
inland.d
910
roduce acceptable speech, for example, while wearingn “artificial palate” or using a “bite block.”1-3
A number of studies concerning the effects ofrthognathic surgery on speech have been carriedut by using such research methods as straighteasures of articulation using electropalatogra-hy,4 evaluation based on patients’ individual ex-eriences using questionnaires,5 evaluation carried
¶Professor, Department of Oral and Maxillofacial Surgery, Uni-
ersity of Turku; Chief Physician, Department of Oral Diseases,
urku University Central Hospital, Turku, Finland.
Address correspondence and reprint requests to Dr Happonen:
epartment of Oral and Maxillofacial Surgery, University of Turku,
emminkäisenkatu 2, FIN20520 Turku, Finland; e-mail: risto-
2006 American Association of Oral and Maxillofacial Surgeons
278-2391/06/6406-0007$32.00/0
oi:10.1016/j.joms.2006.02.009
oda
sphap
dbsdtiscgt
tvoapcm
S
af
shrmmtaUT
rsdmbtopaomtmp
bTcsl
N xillofa
NIEMI ET AL 911
ut by professionals of speech and hearing disor-ers based on phonetic transcription,6 and bycoustic analysis of speech.7
Most of the studies have shown positive effects onpeech (eg, improvement in articulation),8,9 but op-osite findings have also been shown.6,10 It may be,owever, that different measurement procedures andgreat variability of surgical operations can also ex-lain these conflicting results.Consonants have primarily been observed to be
istorted.11 Distortion of sibilants (eg, /s/ and /z/) haseen reported frequently in speakers with malocclu-ion.12-14 Vowel sounds have been studied, but sel-om compared to consonantal sounds. In addition,hese studies have focused usually on single vowelsnstead of systematic examining of the whole vowelystem. For example, Bowers et al15 found significanthange in F2 of /i/ after surgery returning to presur-ical level with time interpreted as evidence of “ac-ive articulatory accommodation.”
This study evaluates the effects of the various or-hognathic surgery procedures on phonetic quality ofowel sounds by analyzing the main acoustic featuresf these sounds. It is assumed that the mandibulardvancement has effects on formant frequencies, es-ecially on second formant frequencies (F2s), be-ause it results in the altered, more anterior place-ent of the tongue.15
ubjects and Methods
SUBJECTS
Five men ranging in age from 31 to 42 years (meange, 36 years; Table 1) were enrolled consecutively
Table 1. SUMMARY OF THE TREATMENT AND SKELETAL
SubjectAge(Yrs)
DiagnosisICD-10 Treatment
Changein ANB O
1 42 K07.13G47.3
BSSObox-osteotomyextr 11, 21, 22
4°
2 39 K07.13G47.3
BSSO 2° 8
3 31 K07.13K07.23
BSSO 3.5° 5
4 34 K07.13 BSSO 2.5° 8.5 34 K07.13
G47.3BSSO
Le Fort I3° 6
Abbreviation: BSSO, bilateral sagittal split osteotomy.*Not possible to define exactly because of extraction of 3 front†Initially open bite, overbite increased at surgery.
iemi et al. Orthognatic Surgery and Vowel Acoustics. J Oral Ma
or orthognathic surgery. All of them were native m
peakers of Finnish and had no history of speech orearing disorders or difficulties. Only men were en-olled to enhance the accuracy of formant value esti-ation. Men have lower F0s than women, the har-onics defining a formant are closer, and its value is
herefore easier to estimate. The study protocol wasccepted by the ethical committees of the Turkuniversity Central Hospital, Turku, Finland and theampere University Hospital, Tampere, Finland.
SURGICAL PROCEDURES
All the subjects had been diagnosed to have a ret-ognathic mandible (Table 1). Three of them wereuffering from obstructive sleep apnea and 1 had aeep bite. Mandibular advancement (range, 6 to 10m) was carried out for all the patients by means of
ilateral sagittal split ramus osteotomy (BSSO). Addi-ionally, BSSO was combined with mandibular boxperation in 1 and with Le Fort I osteotomy in anotheratient. All the operations were done under generalnesthesia, and the patients were hospitalized post-peratively for 1 to 3 days. No disturbances of theandibular nerve were disclosed in any patients after
he surgery. Pre- and postoperative orthodontic treat-ent was an integral part of the treatment of eachatient.
CEPHALOMETRIC ANALYSIS
Lateral cephalograms were obtained immediatelyefore the surgery and 6 to 9 months postoperatively.he cephalograms were traced and conventionalephalometric points and lines were used to study thekeletal and dental changes. SNA, SNB, and particu-arly the difference angle ANB were used to examine
NGES IN THE PATIENTS STUDIED
ction in Increase inLower Face
Height Skeletal ChangesOverbite
* 8 mm Mandibular advancement: 7 mm
6 mm 5 mm Mandibular advancement: rightside 6 mm, left side 7 mm
5 mm 8 mm Mandibular advancement: 6 mm
2.5 mm 2 mm Mandibular advancement: 6 mm3 mm† 3 mm Mandibular advancement: right
side 10 mm, left side 9 mmMaxillary advancement: 6 mm
c Surg 2006.
CHA
Redu
verjet
*
mm
mm
5 mmmm
teeth.
axillary and mandibular changes. Lower anterior
fAlctc
odadCkTV/sttnptwoaerp
SdaTfitdthgcmottcaTifyca
vddpflRd5opm
iatrtvurwsseqd
R
asgfdvstF.2sobvmatlfr
912 ORTHOGNATIC SURGERY AND VOWEL ACOUSTICS
ace height was measured (in mm) from the pointNS (anterior nasal spine) to point menton at the
ower border of the mandibular symphysis. Thehange in the dimension was used, and the change inhe overjet and overbite were also measured on theephalograms.
RECORDINGS OF THE MATERIAL
The recordings of the speech samples were carriedut in a sound-isolated room with a high-quality con-enser microphone (C 1000 S; AKG, Vienna, Austria)t a distance of about 20 cm from the mouth and aigital audio tape recorder (TASCAM DA-P1; Teacorp, Tokyo, Japan) with a sample frequency of 48Hz on DAT tapes (TDK DA-R60STEA; TDK Corp,okyo, Japan). The recording level of input was 40 dBU. The speech material consisted of 8 vowels (/i/,y/, /e/, /ø/, /æ/, /a/, /o/, /u/) produced in the sameentence context “sano sana tVkin vielä kerran” (“sayhe word tVkin once more”) by speakers. The sen-ence context was used to produce the effects oformal connected speech. Every utterance was re-eated 10 times with the speech rate and intonationypical for subjects before the operation (session 1) asell as 6 weeks (session 2) and 30 weeks after theperation (session 3). In all, 1,200 vowel sounds werenalyzed (5 speaker � 3 recording sessions � 8 vow-ls � 10 list readings). The speech samples wereecorded at the Department of Oral Diseases, Tam-ere University Hospital.
ACOUSTIC ANALYSIS
Speech samples were analyzed on Computerizedpeech Lab (CSL, Model 4300B; software version 5.X)eveloped by Kay Elemetrics Corp (Lincoln Park, NJ)t the Department of Phonetics, University of Turku.he main acoustic features of vowels, frequencies ofrst formant (F1) and second formant (F2), fundamen-al frequency (F0) and duration, were analyzed. Theuration was measured by using a combination ofime-domain waveform (ie, oscillogram) displays andigh-resolution gray-scale digital broadband spectro-rams (frequency range display, 0 to 4,000 Hz). Theursors were manually placed to designate measure-ent points. The initial cursor was placed at the onset
f the vowel sound and the final cursor was posi-ioned at the end of the vowel sound, and the dura-ion value displayed was read. The formant frequen-ies were obtained using linear predictive codingnalysis (LPC; frequency range display, 0 to 4,000 Hz).he autocorrelation technique was used. The numer-
cal values of the formant frequencies were obtainedrom the “frequency/bandwidth” table. The LPC anal-sis was carried out by placing the cursor on theenter of the “steady-state” part of the vowel. F0 was
nalyzed from the pitch contours by choosing mean .alue from the “pitch results statistics” table (F0 rangeisplay, 70 to 350 Hz). In the spectrograms the stan-ard Blackman window weighting and 0.8 pre-em-hasis factor was used. In the LPCs the pre-emphasis
actor was 0.9 and LPC coefficient was 12; the frameength and the frame advance were 20 microseconds.eliability of estimated formant values was checked inoubtful cases when the formant was too weak using12-point fast Fourier tracking analysis (20-millisec-nd Hammning window). The cursor was placed atoints of high amplitude corresponding to the for-ants, and the frequency value displayed was read.
STATISTICAL ANALYSES
Statistical analyses were used to evaluate the signif-cance of differences in the acoustic correlates ofrticulatory parameters measured before, 6 weeks af-er, and 30 weeks after the surgical operation. Theepeated measures analysis of variance (ANOVA) withime (before vs after 6 weeks vs after 30 weeks) andowel type (8 levels) as within participant factors wassed. In individual level comparison the non-paramet-ic Wilcoxon matched pairs test was used (the levelas set at 0.01 to avoid type I error). Perceptual
ignificance was evaluated by comparing the mea-ured acoustic differences to just noticeable differ-nce (JND) values for formant and fundamental fre-uencies (as presented by Flanagan16), and foruration (as presented by Lehiste17).
esults
The results show that the operation had individualnd variable effects on vowel quality, ranging fromlightly affected to completely unaffected. At theroup level, there were no significant changes inormant frequencies, fundamental frequencies, or inurations of the vowels between the 3 sessions (all Falues �1). However, the effect of vowel type wasignificant for all the dependent variables suggestinghe intrinsic acoustic differences between vowels [F1:(7,28) 60,063, P � .0001; F2: F(7,28) 279.486, P �0001; F0: F(7,28) 6,738, P � .0001; DUR: F(7,28)1.652, P � .0001]. The vowels produced by theubjects are plotted on an acoustic plane as a functionf F1 and F2 (Fig 1). The normal vowels are markedy diamonds (�) connected by dashed lines, theowels produced 6 weeks after the operation arearked by triangles (‘) connected by square dots,
nd the vowels produced 30 weeks after the opera-ion are marked by circles (●) connected by solidines. Statistically significant differences detected inormant frequencies, fundamental frequency, and du-ation for subjects studied are listed below.
For subject 1, F1 decreased clearly (W � 2,065; P �
0001) in the comparison of session 1 and session 2 (6
wt53æviidcpo
�iws�iHHmIv
F(tCvc
N xillofa
NIEMI ET AL 913
eeks after operation). The decreases were percep-ually significant in high vowels /i, y, u/ (range, 19 to6 Hz; mean, 37 Hz), in mid vowels /e, ø, o/ (range,5 to 104 Hz; mean, 69 Hz), in front vowels /i, y, e, ø,/ (range, 22 to 44 Hz; mean, 29 Hz), and in backowels /u, o, a/ (range, 27 to 51 Hz; mean, 34 Hz). F2ncreased (W � �1,278; P � .0018). However, thesencreases were perceptually insignificant. In addition,uration decreased (W � 1,333; P � .0009), but thehanges were perceptually insignificant. In the com-arison of session 2 and session 3 (30 weeks after the
900
800
700
600
500
400
300
200
/y/
/æ/
/i/
500100020003000
Subject 1
F2 (Hz)
/e/
/o/
/a/
/ø/
/u/
900
800
700
600
500
400
300
200
/y/
/æ/
/i/
500100020003000
Subject 3
F2 (Hz)
/e/ /o/
/a/
/ø/
/u/
900
800
700
600
500
400
300
200
/y/
/æ/
/i/
500100020003000
Subject 5
F2 (Hz)
/e/
/o/
/a/
/ø/
/u/
IGURE 1. Eight vowels produced by the speakers plotted on an acoF2) of the figures is roughly equivalent to the tongue advancement. The mouth cavity. Accordingly, the higher the frequency value oforrespondingly, the higher the frequency value of F1, the more openowels are marked by diamonds (�) connected by dashed lines, theonnected by square dots, and the vowels produced 30 weeks after
iemi et al. Orthognatic Surgery and Vowel Acoustics. J Oral Ma
peration), no significant changes were observed (P d
.01) for formant frequencies, whereas durationncreased (W � �1,692; P � .0001). These increases
ere perceptually insignificant. In the comparison ofession 1 and session 3, F1 decreased (W � 1,913; P
.0001). The decreases were perceptually significantn high vowels /i, y, u/ (range, 17 to 36 Hz; mean, 24z), in low vowels /æ, a/ (range, 34 to 80; mean, 53z), and in back vowels /u, o, a/ (range, 27 to 65 Hz;ean, 43 Hz). F2 increased (W � �1,561; P � .0001).
ncreases were perceptually significant only in backowels /u, o, a/ (range, 52 to 84 Hz; mean, 56 Hz). For
900
800
700
600
500
400
300
200
/y/
/æ/
/i/
500100020000
Subject 2
F 1(Hz)
F2 (Hz)
/e//o/
/a/
/ø/
/u/
900
800
700
600
500
400
300
200
/y/
/æ/
/i/
500100020000
Subject 4
F 1(Hz)
F2 (Hz)
/e//o/
/a/
/ø/
/u/
ane as a function of F1 and F2 frequencies. The horizontal dimensioncal dimension (F1) corresponds closely to the height of the tongue inmore fronted is usually the position of the tongue in the mouth.
mouth cavity because of the lower position of the tongue. The normalproduced 6 weeks after the operation are marked by triangles (‘)
eration are marked by circles (●) connected by solid lines.
c Surg 2006.
F 1(Hz)
300
F 1(Hz)
300
F 1(Hz)
ustic plhe vertiF2, theis thevowelsthe op
uration, no significant changes were observed (P �
.a
�tacmdssm3cddw5
co.(sip1(/dw6t�atsb
co.ciaaeimP(s�hHm
9sdc
iascp(ctchmTcFc
D
omdavdmna
eecpfobbas2ceibiacap
914 ORTHOGNATIC SURGERY AND VOWEL ACOUSTICS
01). For F0, no significant changes were observed inny comparison between sessions (P � .01).
For subject 2, both F1 and F2 increased (F1: W �1,553; P � .0003, F2: W � �1,694; P � .0001) in
he comparison of session 1 and session 2 (6 weeksfter operation). However, these alterations were per-eptually insignificant for both parameters. Further-ore, significant changes were noticed neither for
uration nor for F0 (P � .01). In the comparison ofession 2 and session 3 (30 weeks after operation), noignificant changes were noticed in any parametereasured. In the comparison of session 1 and session
, F2 increased (W � �2,323; P � .0001). The in-reases were perceptually insignificant. In addition,uration increased (W � �1,239; P � .0030) and F0ecreased (W � 1,440; P � .0003). The alterationsere perceptually significant for F0 only (range, 2 toHz; mean, 3 Hz).For subject 3, the only significant change, in the
omparison of session 1 and session 2 (6 weeks afterperation), was increasing of F0 (W � �1,593; P �
0001), which was also perceptually significantrange, 3 to 8 Hz; mean, 5 Hz). In the comparison ofession 2 and session 3 (30 weeks after operation), F1ncreased (W � �1,584; P � .0002). Changes wereerceptually significant in mid vowels /e, ø, o/ (range,9 to 33 Hz; mean, 22 Hz), in low vowels /æ, a/range, 25 to 68 Hz; mean, 45 Hz) and in back vowelsu, o, a/ (range, 21 to 45 Hz; mean, 30 Hz). F0ecreased (W � 1,962; P � .0001). These decreasesere perceptually significant (range, 3 to 9 Hz; mean,Hz). In the comparison of session 1 and session 3,
he only significant change was increasing of F1 (W �1,109; P � .0091). The alterations were perceptu-
lly significant merely in low vowels /æ, a/ (range, 36o 108 Hz; mean, 72 Hz). For F2 and duration, noignificant changes were observed in any comparisonetween sessions (P � .01).For subject 4, the only significant change, in the
omparison of session 1 and session 2 (6 weeks afterperation), was decreasing of F0 (W � 2,030; P �
0001). The changes were also perceptually signifi-ant (range, 5 to 14 Hz; mean, 9 Hz). In the compar-son of session 2 and session 3 (30 weeks after oper-tion), F2 decreased (W � �1,363; P � .0009). Thelterations were perceptually significant in high vow-ls /i, y, u/ (range, 83 to 191 Hz; mean, 127 Hz), andn front vowels /i, y, e, ø, æ/ (range, 92 to 200 Hz;
ean, 133 Hz). In addition, F0 increased (W � 1,617;� .0001). Increases were perceptually significant
range, 4 to 11 Hz; mean, 7 Hz). In the comparison ofession 1 and session 3, F2 decreased (W � 1,740; P
.0001). Changes were perceptually significant inigh vowels /i, y, u/ (range, 82 to 162 Hz; mean, 108z), in mid vowels /e, ø, o/ (range, 74 to 136 Hz;
ean, 90 Hz), and in front vowels /i, y, e, ø, æ/ (range, f1 to 180 Hz; mean, 120 Hz). In this comparison, noignificant changes were observed for F0. For F1 anduration, no significant alterations were found in anyomparison between sessions (P � .01).For subject 5, no significant alterations were found
n the comparison of session 1 and session 2 (6 weeksfter operation). In the comparison of session 2 andession 3 (30 weeks after operation), duration de-reased (W � 2,345; P � .0001). These changes wereerceptually insignificant. In addition, F0 increasedW � �1,360; P � .0009). Increases were also per-eptually significant (range, 3 to 9 Hz; mean, 6 Hz). Inhe comparison of session 1 and session 3, F2 in-reased (W � �1,373; P � .0012). The alterations,owever, were perceptually insignificant. Further-ore, duration decreased (W � 2,608; P � .0001).hese changes were perceptually insignificant. In thisomparison, no significant changes were observed for0. For F1, no significant changes were found in anyomparison between sessions (P � .01).
iscussion
According to the results of the acoustic analysis, therthognathic surgery procedures carried out effectsainly on the frequencies of F1 and F2, but not on the
uration of vowel sounds. Acoustically and perceptu-lly significant changes were highly individual andariable. These changes applied to single vowels, sel-om to the whole vowel system. It must be kept inind, however, that only the effects of surgery, butot preoperative orthodontic treatment, were evalu-ted.
The 2 lowest formants and the total duration of vow-ls did not change systematically for any subject. How-ver, there were some permanent, perceptually signifi-ant changes confined to different vowel groups in 2ersons. For subject 1, F1 decreased, except in mid
ront vowels /e, ø/, and F2 increased in back vowels /u,, a/; and for subject 4, F2 decreased, except in lowack vowel /a/. In addition, F1 decreased for subject 3,ut only in low vowels /æ, a/. Note that most of theselterations came up only 30 weeks after operation (ie, inession 3) but not at the earlier recordings (ie, in session). These changes developed over time. Alternatively, itan also be hypothesized that the operation itself has noffects of vocal tract resonances, but the changes arenduced by subjects’ individual articulatory acts. It mighte that speakers’ awareness about the vocal tract mod-
fication caused by operation necessitates compensatoryrticulation, even when it would not be necessary toorrect speech. The speakers modify their existingcoustic-articulatory mapping or develop a new map-ing. In addition, short-term alterations were found for
ormant frequencies up to 6 weeks postoperatively,
wl
Aenv
cwvstatsrtmotgttaaimcuctctn
witaavm
oisput
mbani
vdbra
stsmrriwpmostooctdm
R
1
1
1
1
1
NIEMI ET AL 915
hich returned completely or partly to the presurgicalevel at 30 weeks after the operation.
Most of the effects detected for F0 were short-term.s mentioned previously, no perceptually significantffects were found for duration indicating that orthog-athic surgery has no effects on temporal acts ofowel production.The acoustic analysis did not show systematic, per-
eptually significant changes indicating that speakersere able to completely or partially compensate the
ocal-tract modification using different motor controltrategies.2,3 The results suggest that vowel produc-ion tasks might be specified primarily in terms ofcoustic output serving as guides for articulatory ges-ures.18,19 The auditory feedback is used by thepeech motor control system to modify an underlyingepresentation mapping vocal gestures to their acous-ic consequences. Auditory feedback provides infor-ation that the speech motor system can use to adjust
ngoing articulation, and over time, these compensa-ory modifications are learned so that the new speechestures can be reproduced in the situation, wherehe vocal tract is altered. There is evidence that audi-ory feedback is used for both online compensationnd long-term adaptation.20 Kent et al21 suggest thatlso proprioceptive and tactile information as well asntraoral pressure are likely essential as auditory infor-
ation for speech production indicating that theompensatory acts could have been the results of these of all feedback modalities in combination. Be-ause vowels are produced without any contacts be-ween articulators (ie, between tongue and palate), aonclusion can be drawn that the importance of tac-ile information is only minor or even entirely insig-ificant in vowel production.22
Individual differences in the magnitude of effectsere observed suggesting that speakers might differ
n their strategy to compensate structural modifica-ions of their vocal tract. In other words, speakers’bilities to compensate vary. These differences canlso have other origins such as different shapes ofocal tract, the type of surgery carried out or differentanners for producing vowels normally.In this study only acoustic analyses were carried
ut, meaning that the interpretations made concern-ng the actual movements of articulators are onlyuppositions based on the acoustic theory of speechroduction.23 It is highly possible that speakers usennatural articulatory configurations for compensa-ions of altered vocal tract.15
Results indicate that the operation has effects pri-arily on the filter characteristics of the vocal tract
ut not on the functioning of the voice source mech-nisms. Orthognathic surgery seems to have only mi-or potential to change vowel quality permanently
ndicating that even large anatomical changes in the
ocal tract have no significant effects on speech pro-uction. These results support the views presentedy Stevens24 that there might be stable articulatoryegions where large physical movements produce rel-tively small acoustic changes.
Patients undergoing different types of orthognathicurgery were enrolled in the present study. In addi-ion to mandibular advancement carried out for all theubjects, Le Fort I osteotomy was done for one andandibular box osteotomy for another patient. The
esults demonstrate that orthognathic surgeries car-ied out had only minor effects on vowel productionn the study subjects with normal speech qualities. It
ould be meaningful to focus the future research onersons with acoustically noticeable speech abnor-alities. To obtain accurate knowledge on the effects
f comprehensive treatment of malocclusions andkeletal disturbances on speech it is necessary to dohe first recordings before the onset of preoperativerthodontic treatment. The increasing sophisticationf measurement techniques within speech science,oupled with a heightened awareness of the impor-ance of knowledge concerning the effects of organiceformities on speech, should open opportunities forore rigorous research in this area.
eferences1. Hamlet S, Cullison BL, Stone ML: Physiological control of sibi-
lant duration: Insights afforded by speech compensation todental prosthesis. J Acoust Soc Am 65:1276, 1979
2. Lindblom B, Lubker J, Gay T: Formant frequencies of somefixed-mandible vowels and a model of speech motor program-ming by predictive simulation. J Phon 7:147, 1979
3. McFarland DH, Baum SR: Incompletely compensation to artic-ulatory perturbation. J Acoust Soc Am 97:1865, 1995
4. Wakumoto M, Isaacson KG, Friel S, et al: Preliminary study ofarticulatory reorganization of fricative consonants followingosteotomy. Folia Phoniatr Logop 48:275, 1996
5. Ruscello DM, Tekieli ME, Jakomis T, et al: The effects oforthognathic surgery on speech production. Am J Orthod 89:237, 1986
6. Dalston RM, Vig PS: Effects of orthognathic surgery on speech:A prospective study. Am J Orthod 86:291, 1984
7. Lee ASY, Whitehill TL, Ciocca V, et al: Acoustic and perceptualanalysis of the sibilant sound /s/ before and after orthognathicsurgery. J Oral Maxillofac Surg 60:364, 2002
8. Glass L, Knapp J, Bloomer H: Speech and lingual behaviorbefore and after mandibular osteotomy. J Oral Surg 35:104,1977
9. Witzel MA, Ross R, Munro I: Articulation before and after facialosteotomy. J Maxillofac Surg 8:195, 1980
0. Farrell CD, Kent JN: Evaluation of the surgical stability of 20cases of inverted-L and C osteotomies. J Oral Surg 35:239, 1977
1. Gardner AF: Dental, oral, and general causes of speech pathol-ogy. Oral Surg Oral Med Oral Pathol 2:742, 1949
2. Ruscello DM, Tekieli ME, Van Sickels JE: Speech productionbefore and after orthognathic surgery: A review. Oral Surg59:10, 1985
3. Vallino LD: Speech, velopharyngeal function, and hearing be-fore and after orthognathic surgery. J Oral Maxillofac Surg48:1274, 1990
4. Yamamoto T, Imai T, Umeda K: Acoustic characteristics of the
fricatives in the surgical class III patients, in Morimoto T,Matsuya T, Takada K (eds): Brain and Oral Functions: Oral
1
1
1
1
1
2
2
2
2
916 ORTHOGNATIC SURGERY AND VOWEL ACOUSTICS
Motor Function and Dysfunction. Amsterdam, The Netherlands,Elsevier, 1995
5. Bowers J, Tobey EA, Shaye R: An acoustic-speech study ofpatients who received orthognathic surgery. Am J Orthod 88:373, 1985
6. Flanagan JL: Speech Analysis Synthesis and Perception (ed 2).Berlin, Germany, Springer-Verlag, 1972, pp 280-281
7. Lehiste I: Suprasegmentals. Cambridge, MA, MIT Press, 1970,p 13
8. Hawkins S: Re-evaluating assumptions about speech percep-tion: Interactive and integrative theories, in Pickett JM (ed):The Acoustics of Speech Communication: Fundamentals,Speech Perception Theory, and Technology. Boston, Allyn and
Bacon, 1999 29. Perkell JS, Guenther FH, Lane H, et al: A theory of speechmotor control and supporting data from speakers with nor-mal hearing and with profound hearing loss. J Phon 28:233,2000
0. Houde JF, Jordan MI: Sensorimotor adaptation in speech pro-duction. Science 279:1213, 1998
1. Kent RD, Martin RE, Sufit RL: Oral sensation: A review andclinical prospective, in Winitz H (ed): Human Communicationand Its Disorders. Norwood, New York, Ablex, 1990
2. Hardcastle WJ: Physiology of Speech Production. New York,Academic Press, 1981
3. Fant G: Acoustic Theory of Speech Production. (ed 2). TheHague, Mouton & Co NV, 1970
4. Stevens K: On the quantal nature of speech. J Phon 17:3, 1989