Spectrographic analysis of cries prom children with maladie du cri du chat

V. VUORENKOSKI, J. LIND, T. J. PARTANEN, J. LEJEUNE, J. LAFOURCADE and O. WASZ-HÖCKERT

Ann. Paediat. Fenn. 1966:12, 174-180.


Résumé :

Abstract : Spontaneous cries of 8 infants with the chromosomal syndrome Maladie du cri du chat have been analyzed with the sound spectrographic method and compared with the results obtained from 75 hunger and 60 pain cries of 105 normal and healthy infants. It has been shown that the spontaneous cry produced by these cases is typical and specific, and different from the normal material.

Sommaire

The major symptom which has attracted interest to the syndrome Maladie du cri du chat is indicated by its name. This symptom is the very peculiar type of cry, reminiscent of the cry of a cat, occurring specially during early infancy. The cases reported in the last years (1, 2, 3, 4, 6, 7, 10, 11, 12) have, except for the typical cry, mental and physical retardation as main symptoms. Other symptoms are: microcephaly, moan face, hypertelorism, epicanthus, non-mongoloid slant of palpebral fissures, laryngeal abnormalities, low-set ears and abnormal dermatoglyphics.

The first cases of Maladie du cri du chat have been reported in 1963 by Lejeune et al. (6). They explained this disease as a deletion of the short arm of a chromosome in the 4-5 group, which has been designated as number 5. Today, 61 cases with similar clinical features, chromosome findings and cat-like cry have been reported from several countries, and a suggestion has been made that as the clinical ramifications of the syndrome become more widely recognized, many more cases will be diagnozed (10).

When Lejeune et al. reported their series consisting of 8 cases at the XI International Congress of Pediatrics in Tokyo 195 (5), a suggestion was made to perform a cooperative study to analyse spectrographically the cry produced by infants with this disease. The intention was to determine the specificity of the spontaneous crying of these cases which could serve as diagnostic-tool.

Haut

Material.

Recordings of spontaneous crying of 8 cases with the syndrome Maladie du cri du chat have. been included in this study. The clinical findings and the chromosome analyses are reported elsewhere (5). A total number of 4-1 different, cry signals of these cases have been analyzed. The number of the cries analyzed and the age of each case on the recording day are presented in Table 1.

The normal reference material consisting of 75 spontaneous (hunger) and 60 pain signals from 105 infants, analyzed and published in previous papers (8, 14), have been used. For the division of this material according to age, see Table 2.

Additionally, same recordings have been made from newborn and young cats.

The cry material was collected on magnetic cape and recorded on a sound spectrograph (Sona-Graph 661 A, made by Kay Electric N. J.) as reported elsewhere (8, 9, 14). The spontaneous cries of both the case and normal material were obtained before the meal; thus these cries heve also been called hunger cries. As the emotional situation of the baby with the syndrome Maladie du cri du chat cannot be compared without consideration of that of the normal infants, a second normal cry type has been used for comparison, namely pain. The pain cry recordings have been made in connection with BCG or PDT vaccination or after pinching the infant's arm (8).

The following variables were measured from sound spectrograms and included in the material:

1. Length, sec.

2. Minimum pitch, cps

3. General pitch, cps

4. Maximum pitch, cps

5. General form of melody

6. Occurence of subharmonic break

7. Occurence of shift

8. Occurence of vocal fry

9. Occurence of glottal plosives

10. Occurence of nasality

11. Occurence of tenseness

12. Occurence of inspiration-expiration phona-tion

13. Continuity

For détails about these variables and technique of measurement, see Wasz-Höckert et al. (14). The

data obtained from these measurements were statistically analysed with computers (Computations were carried out with programs and computers of the Medical Computing Center of the College of Medicine at the University of Cincinnati, Cincinnati, Ohio). In order to find significant differences between cri du chat and normal reference material, two different analyses were performed: the first was concerned with discrimination between cri du chat and normal pain signal, the second between cri du chat and normal hunger signal. A method (13) was used that searched for the best discriminating sets of the variables studied. The screening of the variables took place by the criterion of the joint predictability of the dependent, variable (cri du chat - normal) provided by the different sets of independent variables. The predictability figures are sample estimates of correct classifications, multiplied by 100.

Table 1. - Cri du chat material: number of spontaneous cry analyses and age of different cases
CaseNumber of AnalysesAge of Case, months
1410
284
31042
447
5613
6438
7442
8430
Total44
Table 2. - Normal material: division according to cry situation and age
Age Group
Situation1-3 Months4-5 Months6-7 MonthsTotal
Hunger25252575
Pain20202060

Haut

Results

The typical spontaneous cri du chat signal is characterized by the following acoustical values in comparison with normal spontaneous (hunger) and pain cry signals (see Table3) :

The average length of the cri du chat signal (2.6 sec.) is practically equal to that of the normal pain cry, but the s.d. value is greater in the former (1.6). The average length of the normal hunger cry is shorter (1.2 sec.).

All the pitches are very high (general pitch 860 cps, maximum pitch 1.030 cps, minimum pitch 750 cps) and there is a remarkable difference with the normal material - both normal pain and hunger signals have lower values. The shift does not occur in cri du chat or normal hunger cry signals, while it is very common in normal pain cries (.83). The subharmonic break is also typical for normal pain cry (.62), it occurs seldom in cri du chat (.09) and normal hunger cry signals (.03).

Flat melody form is most common in cri du chat signals (.36); in our normal cry material only birth and pleasure signals have a corresponding amount of flat melody form (14). The next common melody form in cri du chat signals is rising (.23); only .02 of normal pain signals are of this form, and none of the hunger signals. One of the cri du chat signals has a very unusual falling-rising melody form. - The most common melody form in normal hunger signals is rising-falling (.69), in pain signals falling (.58).

The cri du chat signals are voiced (.84), tense (.95) and oral (1.00) and seldom joined to glottal plosives (.20).

Two phenomena occuring in some cri du chat signals are not found in normal cry material used in this study:

1) Dichotomous signal: the phonation stops for a short interval (0.4-0.6 sec.) and continues without any glottal attack (.23).

2) Inspiration-expiration signal: one part of the signal is phonated during the inspiration, the other during the expiration (18).

In f igures 1-6 there are examples of cri du chat signals, while fig. 7 presents a cry of a real cat. Normal pain signals are seen in figures 8 and 9, and normal hunger signals in figures 10-12 (In tape recording of the cry samples presented in figs. 1-2 will be supplied by request.

The signals in figures 1, 3, 4 and 5 are of flat form, signal 2 is falling-rising, signals 6, 7 and 8 rising, signal 9 falIing and signals 10-11 rising-falling. In signal 12 there are only glottal plosives. Signals 2, 3 and 4 are dichotomous, and signal 2 is at the same time with inspiration-expiration phonation. Signals 9 and 11 are half-voiced and the former one contains also a shift. Subharmonic break can be seen in figures 1 and 10.

Comparing the cri du chat signals with normal ones it can easily be seen, that the former ones are all higher in pitch.

From the small material of cries of real cats analyzed (q cries from 5 cats) several similarities can be found with signals of cri du chat cases, specially when taping account of younger cats (1-3 months). The cries of young cats have an equal pitch (600-1.200 cps) and melody form (rising or flat). It is evident that these are shorter than the signals of cri du chat cases, but a short cri du chat signal can shoow an amazing identity with the cry of a young cat (see Figs. 6 and 7).

Discrimination between cri du chat and normal signals: In terms of predictability, number of observations and number of variable subcategory combinations, best discrimination was obtained by using one single variable, namely minimum pitch. Table 4 illustrates the relationship. If minimum pitch is equal to or higher than 500 cps, the signal is very probably a cri du chat; otherwise it should be classified as a normal pain signal. This rule results in correct classifications in 97.1 per cent of the cases, as estimated from the information provided by the sample. No other single variable reached this level of predictability. Using more than one independent variable simultaneously, the predictability was higher than 97.1 per cent in some combinations. However, the improvement was slight, and, considering the number of observations and the number of variable subcategory combinations involved, it was nonexistent.

The other single predictors that resulted in predictability of more than 80 per cent are lined in table 5 along with their respective decision rules and predictability percentages.

In the next best discriminator to general pitch, the predictability dropped more than 6 per cent, and, consequently, the probability of misclassification rose to over 26 per cent.

The best discriminator between cri du chat and hunger signal was again a single variable, and again minimum pitch (Table b).

Its predictability was 91.5 %, the classification being normal hunger signal, if minimum pitch is not measurable or less than 500 cps; otherwise cri du chat. No combination of independent variables improved the predictability from this level. There was another variable that provided a predictability above 80 per cent, namely, general pitch. The decision rule for it is :

Not measurable or less than 600 cps normal : hunger;

Equal or more than 600 cps : cri du chat;

resulting in 87.2 per cent predictability.

Table 3. - Means, medians and percentages of different variables in the groups of normal hunger and pain signals and spontaneous signals from children with maladie du cri du chat
AttributeNormal MaterialCri du chat
Pain 60 CriesHunger 75 Cries8 Babies 44 Cries
Length, sec. 2.71.22.6
Stand, déviation 1.10.61.7
Général pitch, cps530500860
Stand, déviation 8070165
Maximum pitch, cps 6805201 030
Stand, déviation 8080200
Minimum pitch, cps 380420750
Stand, déviation 7070165
Occurrence of shift, % 83
Pitch of shift, cps 910
Occurrence of subharm break, % 6239
Rising-falling, % 10695
Falling, % 58316
Rising, % 223
Flat, % 336
Falling-rising, % 2
Interrupted, % 272818
Occurrence of nasality, % 25
No glottal plosives, % 7231
Glottal plosives, % 285420
Oniy glottal plosives, % 15
Occurrence of vocal fry, % 7520
Occurrence of tenseness, % 874295
Dichotomous signal, % 23
Expiration-inspir signals, % 18
Voiceless signals, % 89
Half-voiced signals, % 202316
Voiced signais, % 726884
Table 4. - Distribution of minimum pitch in normal pain and cri du chat signals
Minimum PitchNormal painCri du chatTotal
Not measurable33
< 500 cps55156
> 499 cps24345
Total6044104
Table 5. - Classifications based on vocal fry and general pitch.
VarianteDecision rulePredicatbility
Vocal fryPresence : normal pain84.4
Absence : cri du chat
General pitch < 600 cps : normal pain80.4
> 599 cps : cri du chat
Not meas. : cri du chat
Table 6. - Distribution of minimum pitch in normal hunger and cri du chat signals
Minimum PitchNormal hungerCri du chatTotal
Not measurable1111
< 500 cps54155
> 499 cps104353
Total7544119


Fig. 1 - Cri du chat signal - case 1.


Fig. 2 - Cri du chat signal - case 2.


Fig. 3 - Cri du chat signal - case 3.


Fig. 4 - Cri du chat signal - case 4.


Fig. 5 - Cri du chat signal - case 5.


Fig. 6 - Cri du chat signal - case 6.


Fig. 7 - Cry of a young cat.


Fig. 8 - Pain signal of a normal infant.


Fig. 9 - Pain signal of a normal infant.


Fig. 10 - Hunger signal of a normal infant.


Fig. 11 - Hunger signal of a normal infant.


Fig. 12 - Hunger signal of a normal infant.

Haut

Discussion

The validity of the results of the present study is lowered by the fact that there were only 8 cases with cri du chat disease in the sample, the total number of 44 cri du chat signals having been obtained by multiple recordings of the same infants. In other words, the total variability in the present case sample is partly interindividual, partly intraindividual.

However, it has been shown that the spontaneous cry produced by infants with the chromosome disease known as Maladie du cri du chat is typical and specific, mostly for its relatively constant high pitch and voiced, tense phonation. The dichotomous and inspiration-expiration cry types found in this material are of special interest.

The cries analyzed in this study are different from the normal material and no resemblance has been found with other pathological cry material studied by the group. On the other hand, it was interesting to note similarities between the cri du chat signals and cries of young cats. This observation shows that the name given to this syndrome is pertinent.


Haut

References

1. BÖÖK, J. A., ATKINS, L., and SANTESSON, B.: Some new data on autosomal aberrations in man. Path. et Biol. 1963:11:1159.

2. GROUCHY, J. DE, ARTHUIS, M., SALMON, C,. LAMY, M. and THIEFFRY, S. : Le syndrome du cri du chat : une nouvelle observation. Ann. Génét. 1964:7:13.

3. KAJII, T., HOMMA, T., OIKAWA, K., FURUYAMA, M. and KAWARAZAKI, T.: Cri du chat syndrome. Arch. Dis. Childh. 1966:41:97.

4. LEJEUNE, J., LAFOURCADE, J., BERGER, R., and TURPIN, R.: Ségrégation familiale d'une translocation 5-13 déterminant une monosomie of une trisomie partielles du bras court du chromosome 5 : Maladie du 'cri du chat' et sa 'reciproque'. C.R. Acad. Sci. (Paris) 1964: 258:5767.

5. LEJEUNE, J., LAFOURCADE, J., BERGER, R,, and TURPIN, R.: La maladie du cri du chat. Proc. Int. Gong. Ped. Tokyo 1965.

6. LEJEUNE, J., LAFOURCADE, J., BERGER, R., VIALETTE, J., BOESWILLWALD, M, SERINGE, P., and TURPIN, R.: Trois cas de délétion partielle du bras court d'un chromosome 5. C.R. Acad. Sci. (Paris) 1963:257:3098.

7. LEJEUNE, J., LAFOURCADE, J., DEGROUCHY, J.; and BERGER, R.: Délétion partielle du bras court du chromosome 5. Individualisation d'un nouvel état morbide. Sem. Hôp. Paris 1964:40:1069.

8. LIND, J., WASZ-HÖCKERT, O., VUORENKOSK. V., PARTANEN, T. J., THEORELL, K., and VALANNE, E.: Focal response to painful stimuli in newborn and young infant. Ann. Paediat. Fenn. 1966:12:55.

9. LIND, J,, WASZ-HÖCKERT, O., VUORENKOSKI, V., and VALANNE, E.: The vocalisation of a newborn, brain-damaged child. Ann. Paediat. Fenn. 1965:11:32.

10. MACINTYRE, M. N., STAPLES, W. L, LaPOLLA, J., and HEMPEL, J. M.: The 'cat cry' syndrome. Amer. J. Dis. Child. 1964:108:538.

11. MCCRACKEN, J. S., and GORDON, R. R: 'Cri du chat.' syndrome. A new clinical and cytogenetic entity. Lancet 1965:1:23.

12. PUNNETT, H. H., CARPENTER, G. G., and DIGEORGE, A. M.: Deletion of short arm of chromosome 5. Lancet 1964:2:588.

13. STERLING, T. D., and POLLACK, S. V.: Computers and the life sciences. Columbia University Press, New-York 1965, p. 173.

14. WASZ-HÖCKERT, O., LIND, J., VUORENKOSKI, V., PARTAKEN, T. J., and VALANNE, E.: The spectrographic and auditive identification of the cry of the normal newborn and young infant,. Developm. :Med. and Child. Neur. In print.