Original Article

Crying Abnormalities in Congenital Hypothyroidism:

Preliminary Spectrographic Study
Daniela Lenti Boero, MD, PhD; Giovanna Weber, MD; Maria Cristina Vigone, MD; Carlo Lenti, MD, PhD

ABSTRACT

The aim of this preliminary study was to evaluate the acoustic patterns of the cries of hypothyroid newborns at the time
of diagnosis and after the beginning of therapy. Cries were recorded at the nursery of the San Raffaele Hospital, Milan,
Italy from 12 full-term subjects (three boys and nine girls) affected by congenital hypothyroidism. Results show that
untreated hypothyroid infants at first recording had fewer voiceless and partially voiced cries than normal controls. The
percent distribution of this pattern did not change at the second recording after the onset of substitutive therapy. Also,
untreated hypothyroid infants had many more cry units showing a vibrato contour than did controls, and this pattern did
not change after the onset of treatment. Starting, maximum, minimum, and end frequencies measured on the fundamen-
tal were significantly lower in the hypothyroid sample. Four hypothyroid subjects recorded before therapy and within
4 weeks after therapy onset significantly augmented their fundamental frequency parameters; however, in 25% of the
sample, sound parameters remained unaltered after 3 or more weeks of treatment. To our knowledge the present prelim-
inary study is the first one performed on follow-up of hypothyroid newborns and indicates that both central and peripheral
damage might influence the pattern of crying in untreated hypothyroid infants. (J Child Neurol 2000;15:603-608).

Among the typical early signs of congenital hypothyroidism
is a husky, hoarse, low-pitched cry well known to clini-
cians. Although sophisticated investigations on infant cry
in different pathologic states have been performed in the past
40 years, 1>2 only two studies have been done on the cries of
hypothyroid infants, before the introduction of routine
screening for hypothyroidism. These lacked a control after
the onset of substitutive therapy.3,1 It is now acknowledged
that between 15% and 40% of infants treated with substitu-
tive therapy from birth might have minimal brain dysfunc-
tion and have problems ranging from hearing impairment
to behavioral disturbances and difficulties in practical rea-
Received April 27, 1999. Received revised Feb 8, 2000. Accepted for publi-
cation Feb 9, 2000.
From the Facoltà di Scienze della Formazione, University of Urbino (Dr
Lenti Boero), Urbino, the Pediatric Department and Childhood and
Adolescence Endocrinology Centre, San Raffaele Hospital, Milan (Drs Weber
and Vigone), the Institute of Neurologic and Psychiatric Sciences of Childhood
and Adolescence, University of Milan (Dr Lenti), and the Laboratory of
Bioacoustics and Comparative Analysis of Behavior and Cognitive Functions,
University of Milan (Drs Lenti Boero and Lenti), Milan, Italy.
Presented in part at the 8th International Child Neurology Association
Congress, Ljubljana, Slovenia, Sept 13-17, 1998.
Address correspondence to Prof. Carlo Lenti, Istituto di Scienze Neurologiche
e Psichiatriche dell’Infanzia e dell’Adolescenza, Facoltà di Medicina e
Chirurgia, Università di Milano, Via G.F. Besta 1, 20100 Milan, Italy. Tel:
390-281844702; fax: 390-289120107; e-mail: dlenti@digibank.it.
soning, perceptomotor and visuomotor discrimination, spa-
tiomotor skills, sensorineural hearing, language compre-
hension, fine motor skills, and extrinsic motor eye
movement; they also show delays in neurophysiologic mat-
uration, have a mean global development quotient (Griffiths
Developmental Test) at age 36 months significantly lower
than that of normal children, and have a higher probability
of being in full-time special education classes. 115 This led
to the suggestion that the consequences of in utero thyroid
deficiencies could not be entirely averted with neonatal
hormone replacement. For this reason, it would be of great
relevance to test this hypothesis in order to detect early those
infants who are less responsive to substitutive therapy by
means of subtle behavioral screening. Because infant cry has,
as do all other human vocalizations, both central and periph-
eral components, 4,16,17 cry analysis is a powerful method to
detect subtle behavioral modification in infants at risk, 2, 18-21
The aim of this preliminary study was to evaluate the
acoustic patterns of the cries of hypothyroid newborns at
the time of diagnosis and after the beginning of therapy in
order to detect possible subtle changes in cry emissions.
MATERIALS AND METHODS
and Data Collection
Subjects
Twelve full-termsubjects (three boys and nine girls) affected by
congenital hypothyroidism diagnosed by routine screenings were

603
Downloaded from jcn.sagepub.com at GEORGIAN COURT UNIV on March 18, 2015
604
sent to the nursery of the San Raffaele Hospital, Milan, Italy and
recruited for this study. All had uneventful gestation and perina-
tal histories(mean Apgar scores, 9 at 1 minute). At arrival, mean
(~ SD) hormonal blood profile was thyroid-stimulating hormone,
524 mU/L (± 354); free T3, 3.18 ng/L (± 1.42); and free T4, 0.72 ng/L
(± 0.77). Immediately before the onset of substitutive therapy
(L-thyroxine) cries were induced by manipulation stimuli during
routine neurologic examination done by an expert clinician (C.L.)
and recorded on Sony digital audio tape (DT-90) with DAT sound
recorders (Sony TCD D7 and TASCAM DAPL) and a Sony unidi-
rectional microphone positioned 2 to 5 cm from the mouth of the
crying baby. Mean time of recording was 15 (± 3) days from birth
(range, 12-18). Eight subjects (two boys and six girls) were recorded
a second time after about 1 month (mean, 55 ± 16 days from birth)
during a follow-up visit, when blood levels of thyroid-stimulating
hormone, free T3, free T4, T3, and T4, were within the normal range.
A sample of 16 infants (seven boys and nine girls) aged 3 to 4 days
recruited in the same nursery served as controls.
Sound Analysis
Recorded cries were sampled by means of an Audiomedia II card
mounted on a Macintosh CI and Sound Designer II (Digidesign Inc)
software. Sampling rate was 44.100 Hz and the sample size was 16
bits (216 value resolution). In order to have homogeneous sample
along the time axis, we subdivided each cry into subsamples of four
to six wails each; the first subsample was collected at the begin-
ning of the cry, the last subsample at the end of the cry, and the oth-
ers along the time axis at homogeneous intervals calculated by
dividing the time remaining.
Sonograms were produced with Canary version 1.2 software22
and a Powermac 7600 with a high-resolution 17&dquo;-screen (Trinitron
Applevision).
Figure 1. Melodic contour of an untreated hypothyroidic infant aged
19 days. Sonogram was drawn with the following parameters: fast
Fourier transform size: 8192 points; frame length: 2048 points; filter
bandwidth: 85.49 Hz; window function: Hanning; clipping level: -100;
amplitude: logarithmic; overlap: 93.75%; grid resolution set at 2.902
ms and 5.383 Hz. In lower window waveform is shown. First cry unit
in upper window has nearly 50% vibrato contour; second and third
cry unit have normal rising contour; fourth unit shows partial phona-
tion with the first half showing evident fundamental frequency; the
second half does not.
Downloaded from jcn.sagepub.com at GEORGIAN COURT UNIV on March 18, 2015
Sound Parameters
Human sound production is a complex, multifaceted phenomenon
and parameters are chosen by researchers according to the aim of
the investigation.23 In the present study we chose both quantitative
and qualitative parameters that differentially contribute to the
understanding of the mechanisms underlying infant cry phonation.
Qualitative
Voicing. Cries in normal infants can be voiced, in which case the
sonographic trace is well evident; voiceless, in which case the
sonographic trace is not evident at all; or partially voiced, in which
case the fundamental frequency is only partially evident.24 The lat-
ter two situations can happen in two opposite conditions: too
much acoustic energy (air) from the lungs drives the vocal folds
apart so they cannot vibrate and produce sound&dquo; or too little
energy makes the fundamental’s trace feeble or not detectable. In
the latter case the fundamental frequency appears if the signal is
amplified with the proper algorithm provided by Canary 1.2. This
gives an important clue for differential diagnosis and for an indi-
rect assay of the acoustic energy of the cry, which could be lack-
ing in the presence of a peripheral motor deficit.
Melodic Contours. The contour shape of voiced cries in nor-
mal infants can have different morphologies; they can rise or rise
and fall; the latest pattern can be repeated twice. They can have
either a flat contour, a falling contour, or a falling and rising con-
tour, which also can be repeated twice. They can show a vibrato
contour when the fundamental frequency shows a continuous
saw-like line (Figure 1). The latter is prominent in preterm infants25
and in other pathologic states such as neonatal asphyxia.’1 A
vibrato pattern consists of fluctuations in the fundamental fre-
quency that do not show any stabilization toward some constant
value. Sources of fluctuation are found at several stages of the vocal
production chain: neurologic, biomechanics, aerodynamic, and
acoustic
Quantitative
Quantitative analysis was carried out only on the fundamental fre-
quency of voiced cries.
Time Parameters (in ms). Length of wail, silent interval to
the next wail, and time for reaching the maximum frequency were
recorded. A previous report24 demonstrated that those parameters
are highly correlated and can be considered a &dquo;time factor&dquo; because
of the respiratory cycle of the infant.
Frequency Parameters (in Hz). Starting, end, maximum,
and minimum frequency, correlated as the &dquo;frequency factor,&dquo;
were recorded.24 The pitch of the fundamental depends on the
length of the vocal folds at rest, the contractile stress imparted to
them, the imbalance of intrinsic agonistic and antagonistic laryn-
geal muscles, and the pressure of the air flow from the lungs.&dquo; In
addition, we measured the dynamic gamma (difference from max-
imum to minimum frequency in each wall), which is indicative of
the acoustic competence of the infant.
Peak Frequency (in Hz). This was measured for each cry unit.
The frequency with the highest acoustic energy is an indicator of
the spectral energy distribution of the sound. It is a supraglottal
factor result of the of the sound in the
as a resonance pharyngeal
and buccal chamber, controlled by the higher corticothalamic cen-
ter.’6 In normal infants this is generally located on the second or

third harmonics, but never on the fundamental. A very interesting
indicator of the acoustic structure of the cry is the ratio between
the peak frequency of the whole cry and the peak of the fundamental
frequency.
Data Analysis and Statistics
Voicing characteristics and melodic contours do not seem
affected by weight differences or by a short maturation span24; for
thisreason the cries of all of the hypothyroid patients were com-
pared with those recorded from the controls. Because time and fre-
quency parameters change rapidly during ontogeny,25 only the
untreated hypothyroid infants recorded within 15 days from birth
(n 7) were compared with the normal sample.
=
To compare different percentages of cry types we used GLIM
version 3.77,~~ which allows modeling on the distribution more
suitable to the data. In analyzing the differences for the frequency
parameters we employed multivariate statistics because
sider them as related random variables29 taken on a single
the fundamental of each single cry unit. Those statistical analyses
were performed using SPSS.
RESULTS
Weight in pathologic subjects at birth was significantly
higher than in the control subjects (F1,15 = 6.700, P .021;
mean ± SD 3727 g ± 369 and 3523 g ± 319, respectively,
= Untreated Patients Versus Normal Controls
for hypothyroid and normal subjects). Mean weight at the
time of the first recording for hypothyroid subjects was
3707 g ± 304, slightly but not significantly lower than at birth.
In the subsample of infants whose second recording was
done before 45 days of age, the mean weight had risen to
3548.75 g ± 361.16, which was not significantly augmented
(t-value3 = 1.58, two-tailed P .213).
=
Voicing
A total of 152 cries from the 12 hypothyroid subjects before
substitutive therapy were compared for voicing against a
total of 208 cries from the normal sample. Untreated hypothy-
roid infants at first recording had fewer voiceless and par-
tially voiced cries than normal control (GLIM 3.77, data
modeled on a binomial distribution, X2 10.835, DF 1,
= =
P < .001; Figure 2). The percent distribution of this pattern
did not change at the second recording after the onset of sub-
stitutive therapy (X2 0.365, DF 1, P not significant). In
= =
addition, as much as 30% of the cries needed amplification,
while none of the cries of normal infants did.
Melodic Contours
Untreated hypothyroid infants at the first recording hadI
many more cry units showing a vibrato contour than did nor-
mal controls (X2 12.66, DF 1, P < .001); this pattern didl
= =
not change after the onset of treatment (X2 0.608, DF 1.,
= =
P not significant; Figure 3). A lesser percentage of flat cons
tours was found to a nearly significant degree in the~
untreated hypothyroid sample in comparison with the nor
mal controls (Xz 3.377, DF 1, 0.05 < P < .10); this patterri
= =
did not change after treatment (X2 0.419, DF l, P not sig
= =
Downloaded from jcn.sagepub.com at GEORGIAN COURT UNIV on March 18, 2015
605
to be
Figure 2. Percentage amount of voiceless and partially voiced cries
in untreated hypothyroid infants before therapy was lower than in nor-
mal controls (X2 = 10.835, DF 1, P< .001). After onset of substitutive
=
therapy in pathologic subjects this pattern did not change (XZ 0.365,
=
DF 1, P not significant).
=
we con-
object:
nificant). No significant differences were found in the per-
centage of rising and falling contours (Xz 0.277, DF 1, P
= =
not significant) or in falling and rising contours (X~ 1.352, =
DF l, P not significant) between the untreated sample and
=
the controls. This pattern did not change after therapy.
=
Time and Frequency Parameters
Because weight affects acoustic parameters, and hypothy-
roid infants were heaver than controls, in this analysis we
kept as controls only those infants whose weight was more
than 3 kg, thus reducing the factor weight difference to
nonsignificance (SPSS, one-way analysis of variance
[ANOVA], F1,14 0.761, P = .398). In the hypothyroid infants,
=
the values of the starting frequency did not depend on thy-
roid-stimulating hormone blood level (SPSS, ANOVA,
Fl187 0.887, P .347, not significant) and were positively
= =
related to weight at the time of recording, as well as the min-
imum frequency (SPSS, regression analysis, F1460 8.2288, =
P 0.0043 and F1458 = 25.5636, P .000, respectively, for start-
= =
ing and minimum frequency); length of wails did not depend
on weight at the time of recording (SPSS, regression analy-
sis, F1459 = 0.0001, P .9981). In the control sample the
=
starting frequency was not positively related to weight of
recording, although the statistics approached significance
(SPSS, regression analysis, F1,93 4.5373, P .08).
= =
The mean and standard deviations of the parameters
analyzed in phonated cries recorded from the untreated
sample and normal controls are displayed in Table 1. There
was no difference in time parameters in the two samples
(SPSS, ANOVA, multivariate Pillais trace, F ,3, 14,3 = 1.052,
.
P .371). But starting, maximum, minimum, and end fre-
=
quencies were significantly lower in the hypothyroid sam-
ple (SPSS, ANOVA, multivariate Pillais trace, F,t.:298 12.15, =
-
P .000; univariate F test: Fu01 = 28.16 for the starting fre-
=
quency, 19.313 for the end, 17.69 for the maximum, and
-
45.39 for the minimum frequency, P .000). =
There was no difference in the dynamic gamma within
-
each wail in the two samples (SPSS, one-way ANOVA,
606 J
Figure 3. Percentage of vibrato contour in untreated hypothyroid
infants before therapy was higher than in normal controls (Xz 12.66,
DF = 1, P< .001 After onset of substitutive therapy in pathologic sub-
jects this pattern did not change (Xz 0.608, DF = 1, P not significant).
=
No significant differences were found in percentage of rising-falling
and falling-rising contours between untreated patients and controls.
While a nearly significant lesser percentage of flat contours was
found in untreated hypothyroid patients in comparison with normal
controls (X2 3.377, DF 1, 0.05 < P< .10), this pattern did not change
= =
after treatment in pathologic subjects (XZ 0.419, DF 1, P not sig-
= =
nificant).
Fl,2ol 1.187, P .277). The peak frequency was higher in
= =
the untreated hypothyroid infants than in the controls (mean
± SD, 4.527 -!- 3.784 and 2.765 -!-- 1.045, respectively, for the
hypothyroid infants and the control; SPSS, one-way ANOVA,
of the whole cry to the peak of the fundamental frequency
~1202 ~ 11.505, P .001) and the ratio of the peak frequency
= and F140 3.306, P .077 not significant, for peak frequency
was significantly different in the pathologic group (SPSS,
one-way ANOVA, F1,201 156,539, P = .000).
=
Cry Parameters Before and After
Treatment in the Hypothyroid Sample
The time interval between the first and second recording in
our sample varied from 21 to 61 days. To make it more
homogeneous, we analyzed the five infants whose second
recording was done within 4 weeks of the first one (mean
± SD, 24.2 ± 2.49; range, 21 to 26 days) separately from the
remaining three, who had a wide range of interval times (38,
55, and 61 days). One of the five infants in the first subgroup
did not significantly vary his time and fundamental fre-
Table 9. Time and Frequency Parameters of Phonated Cries of Untreated Hypothyroid Infants and Controls
. ~
*Calculated with SPSS, ANOVA multivariate Pillais trace.
’Calculated with SPSS, one-way ANOVA.
Downloaded from jcn.sagepub.com at GEORGIAN COURT UNIV on March 18, 2015
quency parameters after 26 days (Table 2). The other sub-
jects did not change their time parameters, significantly
augmented their fundamental frequency parameters (SPSS,
ANOVA, multivariate Pillais trace, F4,123 6.714, P .000),
= =
and showed individual responses for the peak frequency that
were significantly lower in only one infant (subject 5). The
dynamic gamma was significantly augmented in that child
and in one other (subject 7). Means ± SD were as follows:
2745.37 ± 1449.65 for peak frequency, 1534.53 ± 980.9 for
dynamic gamma, 91.25 ± 48.46 for before treatment, and
150.07 ± 43.27 for after treatment for subject 5 and 209.87
± 71.78 for before treatment and 284.23 ± 86.11 for after
treatment for the dynamic gamma for subject 7.
=
Among the other subjects, one did not show any change
in any parameter after 38 days from the first recording
(SPSS, ANOVA, multivariate Pillais trace, F336 1.253, =
P .305, not significant for the time and
=
F435 1.774, P = .156,
=
not significant for the frequency parameters on the funda-
mental ; and ANOVA, factorial: F138 = 1.1, P .301, not sig- =
nificant for the gamma of the fundamental frequency and
F~, 38= 3.68, P .063, not significant for the peak frequency).
=
The other two infants significantly lowered their funda-
mental frequency parameters; one of them also significantly
changed his time parameters, his dynamic gamma, and his
peak frequency (SPSS, ANOVA, multivariate Pillais trace,
F~,37 =
3.095, P .027 for the frequency and
=
0.308, F3,28 =
P .819, not significant, for the times; ANOVA, factorial:
=
F140 = 0.514, P .478, not significant, for the dynamic gamma
=
= =
of the first infant, and F4,13 6.864, P .003 for the fre-
= =
quency and F328 4.985, P .015 for the time; ANOVA, fac-
= =
torial : F1,16 = 4.764, P .044 for the dynamic gamma and
=
F1,16 = 9.117, P .008 for peak frequency of the second infant).
=
DISCUSSION
Infant vocalizations are complex acoustic structures depend-
ing on integrated central and peripheral factors: the air
flow from the lungs, which provides the acoustic energy for
crying, the dimensions of the phonatory apparatus at rest,
the dynamic stress of the vocal folds during vocalization, and
the hierarchic neurocontrol systems: the brain stem, mid-
~ . ~~ ~ .~ ~~ ~.

Table 2. Sound Parameters of
Note convergent difference after treatment for frequency factor for all
frequency.
ns = not significant.
brain, and limbic circuit, respectively, are responsible for the
mechanics, configuration, and initiation of the cry. 16,17,26 In
the present study we chose parameters that could shed
light both on the causation of the well-known acoustic
specificity of the cry of the hypothyroid infant, and on the
possibility of monitoring fine details of therapy effects, as
is made possible by the use of sound spectrography.30
We found that the differences between hypothyroid
and normal infants were straightforward regarding voic-
ing ; fewer disphonated wails were found in the pathologic
sample and many cries needed amplification. This points to
a lack of acoustic energy from the lungs, which is a deficit
peripheral factor. Regarding the melodic contour, our find-
ings were in contrast with those of other authors, who
reported a higher percentage of flat melodic contour in
hypothyroid infants; we did not find this. A possible expla-
nation for this discrepancy could be the different cry stim-
ulus (pain) or the age at recording: Michelsson and Sirvio’s
sample was composed of infants aged from 10 days to
4 months.4
As in other physiopathologic states, such as prematu-
rity25 or brain damage due to neonatal asphyxia,21 we found
greater vibrato contour in the hypothyroid sample. Because
such a severe deviation from a steady normal melodic pat-
tern could be caused at different levels of the sound pro-
duction chain, 17 it is impossible to disentangle the reciprocal
roles of central and peripheral factors in vibrato causation
in this case, as both levels might indeed influence this find-
ing. In accordance with the findings of Michelsson and
Sirvio,4 starting, end, maximum, and minimum frequency
were significantly lower in the patient group before therapy
than in the normal controls. This could be because of relax-
ation of the vocal folds and overall phonatory apparatus, as
well as a lack of peripheral energy from the lungs.
An unexpected finding was that the peak frequency
was significantly higher in the pathologic than in the con-
trol sample. To conserve the harmonic structure of the cry,
the peak frequency should have been lower because the
pathologic infants had a lower fundamental frequency. The
fact that it was not so suggests that in the hypothyroid cry
the normal harmonic structure is not conserved, as is also
pointed out by the significant difference between the ratio
of the peak frequency to the whole cry and the peak of the
fundamental frequency. This discrepancy in harmonic struc-
ture might be the element responsible for the perceived
sharpness of the hypothyroid cry. An altered harmonic
Downloaded from jcn.sagepub.com at GEORGIAN COURT UNIV on March 18, 2015
607
Hypothyroid Subjects Recorded Before andAfterTherapy
but one infant, and the individual responses for the fundamental frequency range and the peak
structure is also found in brain-damaged asphyxiated
infants.21 Because the peak frequency is a supraglottal fac-
tor, depending on central control over the buccal and pha-
ryngeal resonating chamber, those alterations might point
to central damage.
The only comparable parameters between the patho-
logic and the control infants were the time factor, which
reflects the basic life-sustaining respiratory cycle, and the
dynamic gamma, which suggests that pathologic infants
are able to use the same acoustic extension as normal
infants, though on a lower register.
Because both blood and clinical parameters were nor-
malized after the onset of treatment in all of the infants
recorded for the second time, an unexpected finding was that
both the voicing pattern and the percentage of vibrato con-
tour were not changed, thus suggesting at least a retarda-
tion of the effect of the hormones on cry efficiency and a
discrepancy between the spectrographic and clinical bio-
chemical findings.
The fundamental frequency parameters should have
changed at least because of development, if not because of
the therapy. This happened homogeneously in four of five
infants recorded within 4 weeks of the onset of therapy. Indi-
vidual responsiveness to therapy and differential suffering
of in utero thyroid deficiency could have been responsible
for the fact that in two infants (25%), the parameter of the
fundamental frequency remained unaltered, thus suggest-
ing the utility of longer-term monitoring of cry characteris-
tics. Individual differences were clear for two other
parameters: the dynamic gamma and the peak frequency,
which changed in only some of the infants.
The unaltered level of the ratio of the peak frequency
of the whole cry to the peak of the fundamental frequency
in the follow-up sample recorded within 28 days of the first
recording indicates that the finer motor skills needed for the
control of the supraglottal factors might need more time to
normalize as is evidenced by the fact that two of the three
infants recorded after more than 28 days changed their
parameters.
CONCLUSIONS
To knowledge the present preliminary study is the first
our
one performed on follow-up of hypothyroid newborns and
indicates that both central and peripheral damage might
influence the pattern of crying in untreated hypothyroid
608
infants and that individual differences in responsiveness
might be present. In addition, at the biosocial and psycho-
logic level it should be underscored that an unaltered har-
monic structure might negatively influence the establishment
of a normal attachment in the mother or other caregivers. 18,31
This also suggests that, beyond medical treatment, hypothy-
roid infants and their parents also require some kind of
psychologic support. In conclusion, though preliminary,
our study pinpoints the great clinical usefulness of sound
spectrography investigation in pediatric neurology. Future
research will investigate the pattern of cry ontogeny in
hypothyroid and normal infants during the first 3 months of
life in order to exactly compare their progressive changes
and to monitor the effects of medical treatment in the patho-
logic group.
Acknowledgments
This study was performed at the San Raffaele Hospital, Milan, Italy and was sup-
ported in 1995 by a grant from the Ministry for University and Scientific Research
for the project &dquo;Acoustic cues for individuality, motivational states, and ontogeny
of communication: A study on human normal and pathologic infants in a com-
parative perspective,&dquo; by funds awarded by the University of Milan to Carlo
Lenti and by the Pierfranco and Luisa Mariani Foundation. We wish to thank the
nursing staff of the Department of Neonatology of the Ospedale San Raffaele, Milan
for precious and patient collaboration, and an anonymous referee for useful sug-
gestions on a previous version of this paper.
References
1. Wasz-Hockert O, Michelsson K, Lind J: Twenty-five years of Scan-
dinavian cry research, in Lester BM, Boukydis CFZ (eds): Infant
Crying: Theoretical and Research Perspectives. New York,
Plenum Press, 1985, pp 83-104.
2. Lester BM, Corwin M, Golub H: Early detection of the infant at
risk through cry analysis, in Newman J (ed): The Physiological
Control of Mammalian Vocalization. New York, Ablex, 1989, pp
99-118.
3. Vuorenkoski L, Vuorenkoski V, Anttolainen I: Cry analysis in con-
genital hypothyroidism: An aid to diagnosis and clinical evalua-
. Acta Paediatr Scand Suppl
tion 1973;236:27-28.
4. Michelsson K, Sirvio P: Cry analysis in congenital hypothyroidism.
Folia Phoniatr 1976;28:40-47.
5. Glorieux J, La Vecchio FA: Psychological development in con-
genital hypothyroidism, in Dussault JH, Walker P (eds): Congen-
ital Hypothyroidism. New York, Marcell Dekker, 1983, pp 410-430.
6. Ilicki A, Larsson A: Psychomotor development of children with
congenital hypothyroidism diagnosed by neonatal screening. Acta
Paediatr Scand 1988;77:142-147.
7. Letarte J, Garagorri JM: Congenital hypothyroidism laboratory and
clinical investigations, in Ducharme JR, Guyda HJ (eds): Pediatric
Endocrinology. New York, Raven Press Ltd, 1989 pp 449-471.
8. Aronson R, Ehrlich RM, Bailey JD, Rovet JF: Growth in children
with congenital hypothyroidism detected by neonatal screening.
J Pediatr 1990;116:33-37.
9. Porterfield SP, Hendrich CE: The role of thyroid hormones in
prenatal and neonatal neurological development&mdash;Current per-
spectives. Endocrinol Rev 1993;14:94-106.
10. Kooistra L, Laane C, Vulsma T, et al: Motor and cognitive devel-
opment in children with congenital hypothyroidism: A long-term
Downloaded from jcn.sagepub.com at GEORGIAN COURT UNIV on March 18, 2015
evaluation of the effects of neonatal treatment. JPediatr 1994;124:
903-909.
11. Simons WF, Fuggle PW, Grant DB, Smith I: Intellectual develop-
ment at 10 years inearly treated congenital hypothyroidism. Arch
Dis Child 1994;71:232-234.
12. Tillotson SL, Fuggle PW, Smith I, Grant DB: Relation between bio-
chemical severity and intelligence in early treated congenital
hypothyroidism. BMJ 1994;309:440-445.
13. Leentjens IFJ, Kappers EJ: Persistent cognitive defects after cor-
rected hypothyroidism. Psychopathology 1995;28:235-237.
14. Rovet J, Walker W, Bliss B, et al: Long-term sequelae of hearing
impairment in congenital hypothyroidism. J Pediatr 1996;128:
776-783.
15. Siragusa V, Terenghi A, Rondanini GF, et al: Congenital hypothy-
roidism: Auxological retrospective study during the first six years
of age. J Endocrinol Invest 1996;19:224-229.
16. Liebermann P: The physiology of cry and speech in relation to lin-
guistic behavior, in Lester BM, Boukydis CFZ (eds): Infant Cry-
ing: Theoretical and Research Perspectives. New York, Plenum
Press, 1985, pp 29-58.
17. Titze IR: Principles of voice Production. New Jersey, Prentice Hall,
1994.
18. Lenti Boero D: From crying to speech: A review of literature in a
comparative perspective. Ric Psicol 1997;4:33-79.
19. Mende W, Wermke K, Schindler S, et al: Variability of the cry
melody and the melody spectrum as indicators for certain CNS
disorders. Early Child Dev Care 1990;65:95-109.
20. Lenti Boero D, Lenti C, Volpe C, Bianchi C: A preliminary analy-
sis of the cries in preterm and full newborn babies, in Cosmi EV,
Di Renzo GC (eds): 2nd World Congress of Perinatal Medicine.
Proc comm and posters. Bologna, Monduzzi Editore, 1993, pp
523-527.
21. Lenti C, Lenti Boero D, Volpe C, et al: II neonato con sofferenza
cerebrale: Correlazioni clinico sonografiche e di neuroimmagini.
Riv Neuroradiol 1994;7:276-277.
22. Charif RA, Mitchell S, Clark CW: Canary 1.1 User’s Manual.
Ithaca, NY, Cornell Laboratory of Ornithology, 1993.
23. Green JA, Gustafson GE: Interrelations among the acoustic fea-
tures of cries: How many features do we need? Early Child Dev
Care 1990;65:31-44.
24. Lenti Boero D, Lenti C, Volpe C, et al: Newborns crying in differ-
ent situational contexts: Discrete or graded signals? Percept Mot
Skills 1998;86:1123-1140.
25. Sebastio A: Analisi sonospettrografica longitudinale del pianto nel
neonato prematuro. MD thesis, University of Milan, 1998.
26. Lester BM, Boukydis CFZ: No language but a cry, in Papoushek
H, Jurgens U, Papoushek M (eds): Nonverbal Vocal Communi-
cation. Comparative and Developmental Approaches. Cam-
bridge, Cambridge University Press, 1992, pp 145-173.
27. Aitkin M, Anderson D, Francis B, Hinde J: Statistical Modelling
in GLIM. Oxford, Oxford University Press, 1989.
28. Lenti Boero D: Scent-deposition behaviour in alpine marmots
(Marmota marmota L.): Its role in territorial defense and social
communication. Ethology 1995;100:26-38.
29. Manly BFJ: Multivariate Statistical Methods. A Primer. Lon-
don, Chapman & Hall, 1986.
30. Lester BM: Developmental outcome prediction from acoustic cry
analysis in term and preterm infants. Pediatrics 1987;80:529-534.
31. Frodi A: When empathy fails: Aversive infant crying and child
abuse, in Lester BM, Boukydis CFZ (eds.): Infant Crying: Theo-
retical and Research Perspectives. New York, Plenum Press,
1985, pp 263-278.