Original Paper

Folia Phoniatr Logop 2002;54:247–257

DOI: 10.1159/000065197

Spectrographic Measures of the

Speech of Young Children with
Cleft Lip and Cleft Palate
Carmen Casal a Catalina Domı́nguez b Ana Fernández c
Ramón Sarget b Eugenio Martı́nez-Celdrán c Joan Sentı́s-Vilalta d
Cosme Gay-Escoda e
a Department of Orthodontics and Pediatric Dentistry, b Craniofacial Unit,
Hospital Sant Joan de Déu, c Laboratory of Phonetics, Faculty of Arts,
d Faculty of Medicine, and e Oral and Maxillofacial Surgery, Faculty of Dentistry,

University of Barcelona, Spain

Key Words u, e, o]: first formant, second formant, dura-

Acoustic analysis W Bilateral cleft lip and
palate W Cleft lip W Cleft palate W Palatoplasty W
Spectrographic measures W Speech
development W Unilateral cleft lip and
palate W Vocalization
Abstract
Twenty-two consecutive children with re-
paired cleft lip and/or palate [isolated cleft lip
(CL) 6, isolated cleft palate (CP) 7, unilateral
cleft lip and palate (UCLP) 7, and bilateral
cleft lip and palate 2] with a mean age of 27
months underwent spectrographic measures
of tape-recorded speech (DSP Sona-Graph
digital unit). Controls were 22 age- and sex-
matched noncleft children. Data analyzed in-
cluded (1) the Spanish vocalic variables [a, i,
tion, and context; (2) obstruent variables [p, t,
k]: burst, voice onset time, and duration, and
(3) nasal variables [m]: first formant, second
formant, and duration. Statistically signifi-
cant differences were observed between the
CL group and the control group in the first
formant of [e] and in the increase of the fre-
quency of the [t] burst. Comparison between
UCLP and controls showed differences in the
second formant of [a], in the first formant of
[o], and in the second formant of [o]. These
results suggest a small but significant in-
fluence of either the cleft lip or its repair on
lip rounding for [o] and [u]. In addition,
tongue position differences were most likely
responsible for the differences seen with [a]
and [e]. Spectrographic differences in the
current patients did not contribute to mean-

© 2002 S. Karger AG, Basel Prof. Cosme Gay-Escoda

ABC 1021–7762/02/0545–0247$18.50/0 Department of Oral and Maxillofacial Surgery, Faculty of Dentistry
Fax + 41 61 306 12 34 University of Barcelona, Campus de Bellvitge
E-Mail karger@karger.ch Accessible online at: E–08907 L’Hospitalet de Llobregat, Barcelona (Spain)
www.karger.com www.karger.com/journals/fpl Tel. +34 93 4024274, Fax +34 93 4024212, E-Mail cgay@bell.ub.es
130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by:
ingful differences in speech sound develop-
ment. Individualized care (orthodontics, sur-
gery, speech therapy) in children with cleft lip
and/or palate attended at specialized cranio-
facial units contributes to normalization of
speech development.
Copyright © 2002 S. Karger AG, Basel
Introduction
Today, children with cleft deformity of the
lip and/or palate are treated through com-
bined surgical, orthodontic, psychologic, and
speech therapy. Difficulties in speech devel-
opment have extensively been documented in
children with clefts [1, 2]. Such problems may
result in poor self-esteem and in impairment
of psychosocial development of children [3,
4]. Bjornsson and Agustsdottir [5] analyzed
the social outcome in individuals with cleft
palate (CP) and found a higher percentage of
unmarried subjects and a higher incidence of
poor self-esteem among patients with clefts,
although education level was similar to peers
without cleft. It is important to develop a
methodology of examination that allows the
assessment of craniofacial growth, facial har-
mony, and speech development of the af-
fected child.
One of the most commonly used methods
to assess speech features of children with cleft
lip (CL) and/or CP is analysis of phonetic
transcription by speech-language pathologists
[6–8]. Spectrography, the acoustic analysis of
speech in terms of its spectral composition,
shows acoustic cues relevant to articulatory
dimensions. It is a valuable supplement to
phonetic transcription allowing some quanti-
tative comparisons across subjects. The re-
sults of spectrographic measures will provide
further information about differences be-
tween children with and without clefts. Spec-
trography can also be useful in the assessment
248 Folia Phoniatr Logop 2002;54:247–257
of early results of treatment and in the follow-
up of patients with craniofacial anomalies, or
as an outcome measure in cohort studies.
Although a systematic description of
pathological sound phenomena based on
spectrographic analyses with parallels be-
tween acoustic characteristics and diagnoses
have been initiated by Hungarian authors in
the 1970s [9], there are only a few studies of
spectrographic analysis in CP patients. Mi-
chelsson et al. [10] compared the cries of 13
CP neonates with the crying of 75 normal
babies of the same age and found that of the
17 phonetical attributes included in the study,
only vibrato and the ‘tonal pit’ (a sudden fall
of pitch followed by a sudden rise to about the
previous fundamental frequency) occurred
significantly more often in CP infants. Forner
[11] examined five consonant-vowel-conso-
nant nonsense syllables formed using /p/, /t/,
/k/, /tf/, and /s/ in 15 children with congenital
clefts aged between 5 and 6 years with varying
degrees of hypernasality and 15 controls. The
stop gap of the plosive and affricate conso-
nants was the most deviant individual acous-
tic segment. Haapanen and Iivonen [12]
showed a more normal quality of the vowel [i]
in 4 patients with modified Honig (MH) velo-
pharyngeal flap than in 4 patients with the
Sanvenero-Rosselli (SR) velopharyngeal flap
as compared with controls. This is explained
by the lower position of the base of the flap in
the SR technique, so that the oropharynx is
smaller in SR flap patients than in MH ones.
Kataoka et al. [13] studied patients with lip
and/or CP between 4 and 29 years of age
(mean 15 years) for evaluating hypernasality.
Nasal vowels were studied by using an acous-
tic approach and a spectrographic method
and found a high degree of correlation be-
tween both procedures. Bressmann et al. [14,
15] developed a computer-assisted method
for apparative analysis of voice quality and
found that the incidence of voice disorders
Casal et al.
130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by:
(6.5%) in a sample of 154 German-speaking
cleft lip and palate patients was only slightly
higher than in the normal population.
To further contribute to the knowledge of
spectrographic analysis in CP patients, a case-
control study was conducted to evaluate the
acoustic characteristics of children with CL
and/or CP from 10 months to 5 years of age as
compared to a control group of healthy non-
cleft children. Spectrographic measures ob-
tained for 22 children with CL and/or CP who
underwent early surgical repair were com-
pared with 22 controls who demonstrated
normal speech and language development.
Patients and Methods
Patients
This study is based on a series of measurements
performed on the first 22 patients with CL and/or CP
who attended the Craniofacial Unit of Hospital de
Sant Joan de Déu in Barcelona between June 1986 and
January 1991. This pediatric teaching hospital serves a
reference population of 2 million inhabitants who have
the lowest birth rate of Spain and Europe.
The patients with clefts were 14 boys and 8 girls
with a mean age of 27 months distributed into four
groups: 6 patients with isolated CL (mean age 31.5
months), 7 patients with isolated CP (mean age 20.8
months), 7 patients with unilateral cleft lip and palate
(UCLP, mean age 33.4 months), and 2 patients with
complete cleft lip and palate (BCLP, mean age, 23
months). The surgical protocol followed the recom-
mendations of Malek and Psaume [16] and included a
push-back repair of the soft palate and uvula (staphylo-
plasty), different types of Z-plasty repair (cheilorra-
phy), and two-flap (mucocutaneous flap of anterior
palate and mucous flap of vomer) palatoplasty. Sta-
phyloplasty was performed at a mean age of 4.7
months, cheilorraphy at a mean age of 5.8 months, and
two-flap palatoplasty at a mean age of 6.8 months. All
of the children with lip involvement underwent repair
of the underlying orbicularis oris musculature as part
of the surgical procedures. All operations were carried
out by the same surgeon. All patients underwent maxil-
lary correction (within the first days and the 2nd week
of life) with placement of a passive maxillary appliance
that was left in place for 6–8 weeks. In addition they
received psychologic support and speech therapy to
Speech of Children with Clefts of the
Lip and Palate
improve production and coordination of speech ges-
tures. In relation to speech characteristics, none of the
children exhibited glottal stops or hypernasality. The
incidence of recurrent airway infection and otitis me-
dia with effusion among the present population was
36%. Ear infection was managed with insertion of a
tympanostomy tube. None of the children had hearing
loss when the speech samples were taken.
Controls
The control group consisted of age- and sex-
matched children without cleft lip or palate who visit-
ed the outpatient clinic of the Department of Otorhi-
nolaryngology of our institution. To be included in the
control group, the children were required not to dem-
onstrate facial congenital malformations, gross cranio-
facial dysmorphology, growth retardation, psychic
and/or motor disabilities, adverse oral habits, and den-
tal malocclusion. Children with chronic rhinitis or oth-
er conditions possibly affecting nasality were excluded.
Pairing of children was carried out in the sampling
phase of the study. A constant matching age:sex ratio
1:1 was established. Mean ages of the controls were 32
months for the 6 children matched to those in the CL
group, 20.8 months for the 7 children matched to those
in the CP group, 34.7 months for the 7 children
matched to those in the UCLP group, and 23.5 months
for the 2 children matched to those in the BCLP group.
There were no statistically significant differences in
the mean ages of cases and controls included in each of
the four subgroups (Wilcoxon test).
Spectrographic Measurements
The voices of the 44 children were tape-recorded
with a Sony TCM-77V tape recorder whose small size
enabled it to be put out of the child’s visual field. Sony
UX Chrome Extrauniaxial audio tapes were used. The
speech therapist stimulated spontaneous, nonrepeti-
tive speech in order to obtain the speech sounds that
had to be analyzed. For each subject, five vowels and
four consonants [p, t, k, m] in the middle position were
chosen for analysis (i.e., vowels in CVC syllables and
consonants in VCV syllables). When a vowel or conso-
nant in the middle position was absent, an initial or
final vowel or an initial consonant was selected. With
regard to vowels, the following variables were ana-
lyzed: first and second formant frequency steady states
(taken at the midpoint of the steady state) and the
durations of the steady states. Three variables from the
Spanish obstruents [p, t, k] were analyzed: (1) burst fre-
quency, (2) voice onset time (VOT), and (3) duration
of oral closure for the consonant as measured by the
silent interval on the spectrogram or that filled in with
Folia Phoniatr Logop 2002;54:247–257 249
130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by:
 Table 1. Comparison of spectrographic measurements (Hz) between patients with isolated
CL and controls

Variable n Cleft sample Controls p value

mean SD mean SD

First formant [a] 5 688.00 221.63 880.00 203.96

Second formant [a] 5 1,664.00 216.52 1,912.00 315.47
Duration [a] 5 0.12 0.02 0.14 0.03
First formant [i] 6 533.33 140.10 593.33 160.83
Second formant [i] 6 3,100.00 426.15 2,966.67 452.14
Duration [i] 6 0.15 0.06 0.13 0.07
First formant [u] 4 420.00 174.36 370.00 88.69
Second formant [u] 3 986.67 325.78 893.33 23.09
Duration [u] 4 0.14 0.08 0.15 0.06
First formant [e] 6 620.00 21.91 706.67 78.66 0.043
Second formant [e] 6 2,753.33 740.77 2,306.67 640.33
Duration [e] 6 0.15 0.04 0.20 0.10
First formant [o] 6 673.33 89.14 653.33 187.05
Second formant [o] 6 1,413.33 294.66 1,486.67 184.90
Duration [o] 6 0.09 0.05 0.17 0.08
Burst [p] 6 747.33 103.70 646.00 282.66
VOT [p] 6 0.03 0.03 0.02 0.01
Duration [p] 5 0.14 0.05 0.15 0.07
Burst [t] 5 3,320.00 663.93 2,504.00 261.69 0.043
VOT [t] 5 0.07 0.11 0.04 0.04
Duration [t] 4 0.18 0.04 0.14 0.03
Burst [k] 6 1,626.67 814.14 1,393.33 863.27
VOT [k] 6 0.03 0.01 0.03 0.02
Duration [k] 6 0.22 0.08 0.17 0.07
First formant [m] 6 386.67 100.13 333.33 54.65
Second formant [m] 4 1,380.00 545.04 1,510.00 434.97
Duration [m] 6 0.11 0.06 0.17 0.10

prevoicing before burst release. The final duration Spectrographic Analysis

measure for [p, t, k] was made from the start of the oral
closure to the start of the following vowel or the end of
the VOT. One nasal sound [m] was also measured. The
first and second formant frequencies and the duration
of their steady states was measured from [m], particu-
larly the constant, low-frequency energy band, which is
a characteristic feature of the bilabial nasal consonant
[17].
Spectrographic data were recorded in a specifically
designed diagram. Spectrographic analysis of voices
was carried out at the Laboratory of Phonetics in the
University of Barcelona using a DSP Sona-Graph (Kay
Elemetrics model 5500) digital unit. Two investiga-
tors, a staff member of the laboratory of phonetics (A.
F.) and a speech therapist (C. D.) who were always
present at the time of spectrographic measures ana-

250 Folia Phoniatr Logop 2002;54:247–257 Casal et al.

130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by:
lyzed the voices of all children. An analysis of agree-
ment for spectrographic measurements of the first for-
mant [a] in a subset of 164 recordings showed a mean
(SD) frequency of 923.25 (136.10) for the first examin-
er and 900.35 (183.90) for the second examiner.
Statistical Analysis
The SPSS computer program was used for the anal-
ysis of data. Means, SDs and range of the 27 speech
variables of speech development were calculated for
each of the four subgroups. The Wilcoxon matched-
pairs sign rank test was used to determine the statisti-
cal significance of group differences. The chi-square
(¯2) test was used for the analysis of the five nonnumer-
ical variables of the spectrographic speech study. A p
value of ! 0.05 was considered statistically significant.
Data are expressed as mean (SD).
Results
A comparison of spectrographic measure-
ments between CL patients and controls (ta-
ble 1) showed statistically significant differ-
ences for two variables. The first formant of
vowel [e] showed a frequency decrease
[620.00 (21.91) versus 706.67 (78.66) Hz, p !
0.043]. The other statistically significant vari-
able was the frequency of the [t] burst, which
was higher among CL patients [3,320.00
(663.93) versus 2,504.00 (261.69) Hz, p !
0.043]. In order to rule out potential interfer-
ences, we also analyzed the vowel following
the consonant [t] (data not shown), as its burst
presents some variability depending on the
type of vowel that followed it. Thus, if conso-
nant [t] is followed by a front vowel, its burst
is more acute (e.g., higher frequency empha-
sis) and if it is followed by a back vowel, its
burst is more grave (e.g., lower frequency em-
phasis). When vowels in the CVC syllable
were analyzed, it was found that vowels [a]
and [o] had been acquired in 100% of the
cases, vowel [u] in 90%, vowel [e] in 83.3%,
and vowel [i] in 66.6%.
Speech of Children with Clefts of the
Lip and Palate
Spectrographic measurements in the CP
group and controls (table 2) showed no statis-
tically significant differences. Most of the
vowels analyzed appeared between two con-
sonants. Thus, vowel [e] was in a middle word
location in 81.8% of cases, vowel [a] in 80%,
vowel [u] in 75%, vowel [o] in 70%, and vowel
[u] in 58%.
When spectrographic measurements be-
tween UCLP patients and controls were com-
pared (table 3), there were three statistically
significant variables. The second formant of
[a] showed a frequency increase [1,912.33
(466.76) versus 1,493.33 (195.41) Hz, p !
0.046], although such an increase did not have
a large phonetic repercussion. The first for-
mant of [o] showed an increase [720.00
(101.98) versus 616.00 (96.33) Hz, p ! 0.043]
as did the second formant of [o]: [1,656.00
(267.73) versus 1,360.00 (248.18) Hz, p !
0.043]. The vowel [a] was in a middle word
position in 76.9% of cases, vowel [i] in 81.8%,
[u] in 80%, [e] in 90.9%, and [o] in 70%.
Spectrographic measures of 2 children
with BCLP and controls are shown in table 4.
The BCLP children had not acquired the vow-
el [u], but this abnormality was transitory
because they were studied at an early age. At
this time, they also produced consonants [p, t]
as [k]. Although data were recorded for only 2
children, there were differences for first for-
mant of [a], for the first formant of [o], and for
the second formant of [i] (probably related to
back tongue position) as compared with con-
trols.
Discussion
In the current study, spectrographic analy-
sis of CL children showed statistically signifi-
cant differences in only the first formant of
vowel [e]. It seems that there is not proper
lowering of the tongue as the first formant
Folia Phoniatr Logop 2002;54:247–257 251
130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by:
 Table 2. Comparison of spectrographic measurements (Hz) between patients with isolated
CP and controls

Variable n Cleft sample Controls p value

mean SD mean SD

First formant [a] 7 880.00 273.25 754.29 210.94

Second formant [a] 7 1,754.29 331.81 1,554.29 421.06
Duration [a] 7 0.15 0.06 0.12 0.02
First formant [i] 3 693.33 266.33 506.67 100.66
Second formant [i] 3 3,333.33 100.66 2,053.33 981.70
Duration [i] 3 0.14 0.08 0.16 0.06
First formant [u] 2 380.00 28.28 540.00 197.99
Second formant [u] 1 840.00 – 1,320.00 –
Duration [u] 2 0.16 – 0.12 0.04
First formant [e] 3 746.67 61.10 813.33 140.48
Second formant [e] 3 2,053.33 789.26 2,106.67 625.25
Duration [e] 3 0.13 0.00 0.14 0.02
First formant [o] 3 626.67 61.10 660.00 34.64
Secondformant [o] 3 1,413.33 122.20 1,360.00 40.00
Duration [o] 3 0.20 0.06 0.17 0.09
Burst [p] 4 780.00 124.37 2,440.00 3,440.16
VOT [p] 4 0.02 0.01 0.02 0.01
Duration [p] 6 0.16 0.05 0.12 0.03
Burst [t] 4 3,310.00 1,339.60 2,410.00 591.83
VOT [t] 4 0.08 0.06 0.08 0.13
Duration [t] 4 0.16 0.08 0.14 0.08
Burst [k] 1 2,106.00 – 2,080.00 –
VOT [k] 1 0.02 * 0.02 *
Duration [k] 1 0.20 * 0.11 *
First formant [m] 4 390.00 38.30 520.00 242.21
Second formant [m] 4 1,320.00 353.27 1,330.00 264.07
Duration [m] 4 0.10 0.03 0.11 0.03

* Variable missing.

depends on the constriction of the pharynx
[18]. In addition, production of [e] appears to
be affected by differences in mouth opening,
which may also have been the result of lip sur-
gery or inherently small mouth openings be-
cause of differences in orbicularis oris. We
have not found differences for [i] and [u] per-
haps because these vowels require small
mouth openings, which apparently can be
handled by these children. Any statistically
significant difference in these vowels in the
production of which lip action is essential
would alert us of an inadequate surgical tech-
nique or an inadequate accommodation to the

252 Folia Phoniatr Logop 2002;54:247–257 Casal et al.

130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by:
 Table 3. Comparison of spectrographic measurements (Hz) between patients with isolated
UCLP and controls

Variable n Cleft sample Controls p value

mean SD mean SD

First formant [a] 6 850.00 302.46 913.33 127.54

Second formant [a] 6 1,913.33 466.76 1,493.33 195.41 0.046
Duration [a] 5 0.15 0.06 0.20 0.14
First formant [i] 5 528.00 76.94 488.00 177.54
Second formant [i] 4 3,420.00 189.03 2,920.00 340.98
Duration [i] 5 0.14 0.05 0.17 0.15
First formant [u] 5 496.00 161.49 488.00 153.36
Second formant [u] 3 1,253.33 61.10 800.00 40.00
Duration [u] 5 0.19 0.07 0.17 0.09
First formant [e] 5 544.00 92.09 696.00 72.66
Secondformant [e] 5 2,912.00 320.50 2,616.00 394.56
Duration [e] 5 0.14 0.05 0.16 0.10
First formant [o] 5 720.00 101.98 616.00 96.33 0.043
Second formant [o] 5 1,656.00 267.73 1,360.00 248.19 0.043
Duration [o] 5 0.12 0.02 0.13 0.06
Burst [p] 5 752.00 232.21 816.00 171.11
VOT [p] 5 0.18 0.25 0.03 0.02
Duration [p] 5 0.15 0.06 0.13 0.03
Burst [t] 4 2,700.00 360.00 2,460.00 333.87
VOT [t] 3 0.05 0.05 0.03 0.01
Duration [t] 4 0.16 0.07 0.15 0.07
Burst [k] 5 2,088.00 818.49 3,064.00 1,453.71
VOT [k] 5 0.08 0.09 0.02 0.01
Duration [k] 5 0.14 0.04 0.16 0.07
First formant [m] 3 333.33 61.10 386.67 184.75
Second formant [m] 3 1,026.7 201.33 1,280.00 138.56
Duration [m] 3 0.09 0.01 0.10 0.05

mechanism. Spectrographic analysis of CP ond [o] formants probably relates to an ab-

children showed no statistically significant
differences as compared with controls. Fur-
ther, in contrast to UCLP children who had
an increase in the formants of vowel [o], CL
children showed no abnormalities. On the
other hand, the increase in the first and sec-
sence of lip rounding, resulting in an increase
in the formant frequencies of this vowel. In
addition, spectrographic analysis of UCLP
children showed a statistically significant
higher frequency of the first formant of the
vowel [a], which had little phonetic impor-

Speech of Children with Clefts of the Folia Phoniatr Logop 2002;54:247–257 253
Lip and Palate
130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by:
 Table 4. Comparison of spectrographic measurements (Hz) between patients with isolated
BCLP and controls

Variable n Cleft sample Controls p value

mean SD mean SD

First formant [a] 2 660.00 141.42 820.00 367.70

Second formant [a] 2 1,860.00 593.97 1,480.00 56.57
Duration [a] 2 0.24 0.08 0.13 0.01
First formant [i] 2 580.00 141.42 520.00 226.27
Second formant [i] 2 2,700.00 84.85 3,480.00 339.41
Duration [i] 2 0.14 0.05 0.18 0.08
First formant [u] * * *
Second formant [u] * * *
Duration [u] * * *
First formant [e] 2 620.00 197.99 640.00 169.71
Second formant [e] 2 1,980.00 424.26 2,420.00 84.85
Duration [e] 2 0.24 0.04 0.22 0.01
First formant [o] 1 800.00 – 480.00 –
Second formant [o] 1 1,560.00 – 1,520.00 –
Duration [o] 1 0.09 – 0.13 –
Burst [p] * * *
VOT [p] * * *
Duration [p] * * *
Burst [t] * * *
VOT [t] * * *
Duration [t] * * *
Burst [k] 1 2,040.00 – 2,600 –
VOT [k] 1 0.03 – 0.02 –
Duration [k] 1 0.16 – 0.14 –
First formant [m] 2 560.00 56.57 460.00 197.99
Second formant [m] 2 1,380.00 141.42 2,440.00 226.27
Duration [m] 2 0.13 0.03 0.17 0.05

* Variable missing.

tance because vowel quality does not depend
on the frequency of only one of its formants,
but on its overall spectral shape. Impaired lip
rounding seems a possible reason for higher
formants in UCLP patients as compared with
controls, although an anterior tongue position
could also be responsible for differences in the
second resonance. Thus, some effects from
surgery or the inherent differences in the orbi-
cularis oris may result in small but significant
differences in vowel formant frequencies.
Spectrographic analysis of the speech of 2
BCLP children showed that the most charac-
teristic and distinctive feature of these pa-

254 Folia Phoniatr Logop 2002;54:247–257 Casal et al.

130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by:
tients was their inability to attain normal
labialization as shown by the lack of acquisi-
tion of vowel [u] and a higher frequency of the
first formant of [o]. The BCLP patients stud-
ied had not acquired the vowel [u], but they
did produce the vowel [o]. In addition, these
patients did evidence correct auditory dis-
crimination of [p], [t], and [k], but they did
not correctly produce [p] and [t]. Both sounds
were produced and phonetically evaluated as
[k].
Some characteristics of the Spanish lan-
guage, however, deserve to be mentioned. In
Spanish, vowel arrangement by phonetic or-
der (from front to back) is [i, e, a, o, u]. In
relation to vowel duration, the wider the
mouth opening is, the longer the duration.
The most prolonged vowel is [a] followed by
[e, o, i, u]. In addition, the duration of stressed
vowels with tonic accent is also more pro-
longed. Vowel position also alters vowel dura-
tion. The most neutral vowel is that between
two consonants. Initial vowels have different
durations and are less homogeneous; final
vowels have lower intensity and longer dura-
tion. On the other hand, the consonant burst
occurs as a release of air pressure accumulated
within the closed oral cavity. In the case of the
voiceless [p], the air is retained within the
mouth with total occlusion of the lips, the
vocal folds are at rest, and a moment of
silence is produced. The frequency of the
burst energy increases in voiceless consonants
[t] and [k]. However, [k] bursts are higher in
frequency relative to [p], but typically lower in
frequency relative to [t]. Data on bursts indi-
cate the absence of nasal escape of air during
consonant production since escape of air
through the nostrils creates insufficient in-
traoral pressure to generate a burst. VOT
measures the time from the oral release of the
constriction until the start of vocal fold vibra-
tion. For voiceless stops such as [p, t, k],
VOTs are positive indicating a lag in the onset
Speech of Children with Clefts of the
Lip and Palate
of vocal fold vibration. It should be noted that
Spanish voiceless plosives are not aspirated,
so that vibration may start very close to the
release of the burst. On the other hand, the
VOT of voiced consonants [b, d, g] is negative
as vocal fold vibration does not coincide with
the beginning of the vowel and in Spanish, is
made before the burst for Spanish voiced con-
sonants.
Acquisition of the vowels was similar in
the children with clefts and the control groups
in terms of order. The vowel [a] was present in
most subjects of both groups, and the vowel
[u] was the most infrequently represented in
the study population, indicating a trend to-
wards neutralization in cleft patients. Chap-
man [2] reported that vowel acquisition was
slower among children with clefts compared
to children without clefts. On the other hand,
vowel quality in children with clefts was also
similar to that of control children. A balance
between mouth opening and position of the
lip and tongue as well as the empty space
created in the intraoral cavity was observed,
which may be induced by placement of early
maxillary orthopedic palate. Similar effects
have been observed in patients with clefts
after maxillary expansion, thus improving lin-
guopalatal contact [19]. Patients with clefts
have been reported to present less mobility of
the tongue that occupies a low, posterior posi-
tion. Yamashita and Michi [20] evaluated
compensatory articulation by electropalato-
graphic and spectrographic methods and con-
firmed the abnormal position of the tongue in
patients with clefts, inducing the subsequent
compensatory articulation.
In the present study, we evaluated not only
the features of the five vowels, but also effects
due to word position. The selection of vowels
in CVC syllables ensured a lower variability
in duration of production of such phones [20,
21]. Almost 80% of children produced the
five vowels in middle placement. Acquisition
Folia Phoniatr Logop 2002;54:247–257 255
130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by:
of voiceless consonants [p, t, k] and the nasal
consonant [m] was similar in both groups (38
and 34% of children, respectively). Conso-
nants [p, m] were acquired before [t, k] in both
groups. The pattern of consonant acquisition
was normal in both groups. However, some
authors observed delay in the acquisition of
consonants such as [t, d] [2, 7]. Such delay is
expected when oronasal communication re-
mains patent.
Spectrographic analysis is a method that
can supplement the phonetic transcription of
speech commonly used at present for phonet-
ic analysis of children with CP. The study of
normal patterns of speech development dur-
ing early childhood [23, 24] allowed the iden-
tification of the most significant deviations
characterizing children with cleft of the lip
and/or palate undergoing early surgical cor-
rection. However, statistically significant
spectrographic differences between cleft (CL,
UCLP) and noncleft children did not appear
important for speech development in these
patients and seem to reinforce the value of
early surgical repair and speech therapy. Giv-
en the small sample size and the length of fol-
low-up, these short-term results cannot be
considered definitive.
From the spectrographic standpoint, chil-
dren with BCLP presented the most severe
speech limitations. Taking into account that
these patients also underwent early surgical
repair, these results suggest that the severity of
the initial malformation is critical for consid-
eration of prognosis. Many speech studies
could not separate effects of timing vs. severi-
ty of the cleft [1, 25]. The present BCLP were
more severe than the other participants,
showing delays in vowel production and other
consonant features.
The possibility of normalization of speech
development at early ages in children with CL
and/or CP is very important and confirms the
need for specialized craniofacial units in or-
der to provide individualized care and follow-
up for patients with congenital orofacial de-
fects.
Acknowledgments
We are indebted to Marta Pulido, MD, for editing
the manuscript and editorial assistance.

References

1 O’Gara MM, Logemann JA: Pho-
netic analyses of the speech develop-
ment of babies with cleft palate.
Cleft Palate J 1988;25:122–134.
2 Chapman KL: Vocalizations of tod-
dlers with cleft lip and palate. Cleft
Palate Craniofac J 1991;28:172–
178.
3 Tobiasen JM: Social judgements of
facial deformity. Cleft Palate J 1987;
24:323–327.
4 Richman L: Facial and speech rela-
tionships to behavior of children
with clefts across three age levels.
Cleft Palate Craniofac J 1997;34:
390–395.
5 Bjornsson A, Agustsdottir S: A psy-
chosocial study of Icelandic individ-
uals with cleft lip or cleft lip and pal-
ate. Cleft Palate J 1987;24:152–
157.
6 Bradford PW, Culton GL: Parents’
perceptual preferences between
compensatory articulation and nasal
escape of air in children with cleft
palate. Cleft Palate J 1987;24:299–
303.
7 Harding A, Campbell RC: A com-
parison of the speech results after
early and delayed hard palate clo-
sure: A preliminary report. Br J
Plast Surg 1989;42:187–192.
8 Asher-Mcdade C, Roberts C, Shaw
WC, Gallager C: Development of a
method for rating nasolabial appear-
ance in patients with clefts of the lip
and palate. Cleft Palate Craniofac J
1991;28:385–390.
9 Safrán A, Szende T: Esophageal
speech as a compensatory process.
Folia Phoniatr 1973;25:347–364.
10 Michelsson K, Sirvio P, Koivisto M,
Sovijarvi A, Wasz-Hockert O: Spec-
trographic analysis of pain cry in
neonates with cleft palate. Biol Neo-
nate 1975;26:353–358.

256 Folia Phoniatr Logop 2002;54:247–257 Casal et al.

130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by:
11 Forner LL: Speech segment dura-
tions produced by five and six year
old speakers with and without cleft
palates. Cleft Palate J 1983;20:185–
198.
12 Haapanen ML, Iivonen A: Sound
spectra in cleft palate patients with a
Sanvenero-Rosselli and modified
Honig secondary velopharyngeal
flap. Folia Phoniatr 1992;44:291–
296.
13 Kataoka R, Michi KI, Okabe K,
Miura T, Yosjida H: Spectral prop-
erties and quantitative evaluation of
hypernasality in vowels. Cleft Palate
Craniofac J 1996;33:43–50.
14 Bressmann T, Merk M, Sader R,
Ziegler W, Horch HH: Computer-
assisted acoustic speech analysis in
therapy follow-up of patients with
cleft lip-jaw-palate. Biomed Tech
(Berl) 1997;42(suppl):93–94.
Speech of Children with Clefts of the
Lip and Palate
15 Bressmann T, Sader R, Merk M,
Ziegler W, Busch R, Zeilhofer HF,
Horch HH: Perceptive and instru-
mental examination of voice quality
in patients with lip-jaw-palate clefts.
Laryngorhinootologie 1998;77:700–
708.
16 Malek R, Psaume J: Nouvelle con-
ception de la chronologie et de la
technique chirurgicale du traite-
ment des fentes labio-palatines. Ann
Chir Plast Esthet 1983;28:237–247.
17 Kent RD, Read CH: The Acoustic
Analysis of Speech. San Diego, Sin-
gular Publishing Group, 1992.
18 Jakobson R, Halle M: Fundamentos
del lenguaje. Madrid, Ayuso, 1956.
19 Ohkiba T, Hanada K: Adaptative
functional changes in the swallowing
pattern of the tongue following ex-
pansion of the maxillary dental arch
in subjects with and without cleft
palate. Cleft Palate J 1989;26:21–
30.
Folia Phoniatr Logop 2002;54:247–257
20 Yamashita Y, Michi K: Misarticula-
tion caused by abnormal lingual-pa-
latal contact in patients with cleft
palate with adequate velopharyn-
geal function. Cleft Palate Craniofac
J 1991;28:360–366.
21 Martı́nez-Celdrán E: Análisis espec-
trográfico de los sonidos del habla.
Barcelona, Ariel Practicum, 1998.
22 Navarro Tomás T: Manual de Pro-
nunciación Española. Madrid,
CSIC, 1918.
23 Mitchell P, Kent RD: Phonetic vari-
ation in multisyllable babbling. J
Child Lang 1990;17:247–265.
24 Oller DK, Wiedman LA, Doyle WJ,
Ross C: Infant babbling and speech.
J Child Lang 1975;3:1–11.
25 Dorf DS, Curtin JW: Early cleft pal-
ate repair and speech outcome. Plast
Reconstr Surg 1982;70:74–81.
257
130.239.20.133 - 5/11/2018 6:48:12 PM
Umea University Library
Downloaded by: