Sound spectrographic analysis of pain cry in preterm infants

Eurly Human Development. Elsevier

8 ( 1983) 14 I- 149

141

Sound spectrographic analysis of pain cry in preterm infants K. Michelson,

A-L. JkvenpS

and A. Rime

Children’s Hospital, University of Helsrnkr, Helsinki, Accepted

for publication

20 February

Finland

1983

Summary Sound spectrographic cry analysis was performed on 302 cries of 48 preterm infants born at 30-37 gestational weeks. The cries were recorded during the first week of life and thereafter weekly until the infants were discharged. The control series comprised 54 cries from 27 fullterm healthy infants. The results showed that the cries of the smallest prematures compared with the controls were shorter, more high-pitched, and included bi-phonation and glide more often. The cry characteristics changed with increasing conceptual age and the older the child the more the cry pattern resembled that of the fullterm. The cries of the preterm infants when they had reached 38 conceptual weeks were similar to those of newly born fullterm infants. The results indicate that the gestational age should be taken into consideration in cry analysis. cry; sound

spectrography;

preterm

infant

Introduction Sound spectrographic studies on the cries of preterm infants were reported by Michelsson in 1971 [9]. She noted that the cry of prematures was different from the crying of healthy fullterm infants. The cries correlated with the maturity of the infants; the pitch was higher, the cries were shorter and bi-phonation and glide were more common in younger individuals. Furthermore, it was found that the fullterm infants small for gestational age cried similarly to those of normal birthweight. This

Address for correspondence; Helsinki 29. Finland.

037%3782/83/$03.00

Katarina

Michelsson,

M.D., Children’s

0 1983 Elsevier Science Publishers

B.V.

Hospital,

Stenbtickstreet

11, SF-00290

142

indicated that the cry characteristics were more dependent on the gestational age and the maturation of the child than on the birthweight. In 1981, Michelsson et al. [12] made a blind study of cries of 200 infants admitted to the newborn ward at the Children’s Hospital, University of Helsinki. They included both symptomless and sick preterm infants. The results of this study were comparable with those of Michelsson [9]. Both studies showed that the cries of sick preterm infants were more different from cries of healthy fullterm infants than the cries of symptomless prematures were. In 1982, Thoden et al. [17] reported on the sound spectrographic cry analysis of 69 prematures. The infants were divided into three groups according to the conceptual age at which the cry was recorded (31-33, 34-37 and 2 38 weeks). The cries were collected when the infants were newly born and then weekly to term. The first, second and third cry signal after the pain stimulus, in all 285 cries, were analysed. The results indicated that there was a clear change in the cry characteristics of the pain cry with increasing conceptual age; the duration of the cry signals lengthened and the pitch of the signals decreased. When the preterm infants had reached the age of 38 conceptual weeks, the cry characteristics were almost identical to the pain cry of fullterm, newly born infants. In the present study, the cry characteristics of preterm infants were analysed weekly in order to find out if the cry characteristics change with increasing gestational age and if the preterm infants’ cries reach the pattern which has been found in the fullterm.

Material Subjects The material consisted of 48 preterm infants born at 30-37 weeks of gestation. The control series comprised 27 fullterm infants born after uneventful pregnancies and deliveries. The cries of the preterm infants and the control series were partly the same as in the series by Thoden et al. [ 171 and Thodtn and Koivisto [ 161. The gestational age was estimated from the last menstrual period and clinically by the scoring system by Dubowitz et al. [2]. When these two estimations differed, the one by Dubowitz et al. was used. The birthweight of the premature infants was appropriate for the gestational age [l], and the infants did not have IRDS, septic infections, severe hypoglycemia, hyperbilirubinemia, or neurological symptoms. The fullterm control infants were all healthy and had normal birthweights. Cry analysis The infants were awake at the moment of the induction of the cry. The cries were elicited by a pinch of the infant’s ear. The cries of the 48 prematures were collected during the first week of life and weekly thereafter until discharge from the hospital. Two cry signals were analysed from each recording. The cries were recorded on 15 1 occasions totalling 302 cries.

143

The cries of the control infants were recorded at one and five days of age. The first cry signal after the pain stimulus was analysed from the recordings. Methods The cries were recorded on a Uher 4000 L tape recorder or a Sony TC 55 cassette recorder. The cries were visualised as sonograms using a Sona-Graph 7029A (Kay Elemetrics USA) provided with a narrow band filter. The sonograms were analysed according to the principles given by Sirvio and Michelsson [ 151. The following cry characteristics were analysed from the sound spectrograms: the duration of the cries, the occurrence of shift (an abrupt change in the fundamental frequency), the maximum pitch both including and excluding the shift, the minimum pitch, the melody type and the occurrence of bi-phonation (two simultaneous sounds), glide (a rapid up- or downward change of the fundamental), vibrato, and glottal roll. Results Table 1 shows the maximum and minimum pitch of the cry signals at different conceptual ages. The cries of the smallest prematures were the most high-pitched, TABLE

I

Mean duration

and mean maximum

Conceptual weeks

Cries

Duration

(n)

(s)

Preterm 30-31 32 33 34 35 36 37 38 39 40 41-42

11 22 34 40 52 44 33 30 13 12 11

1.4kO.2 ** 2.6& 1.7 **+ 3.Ok2.0 ** 3.9k2.3 4.2 k 2.5 4.3 + 1.6 4.3* 1.5 4.0 * 2.2 3.8*2.1 3.8* 2.3 3.5* 1.2

Controls 38-40

54

4.3_+ 1.9

and minimum

pitch of the cries in prematures

Maximum

pitch

and controls Minimum

pitch

(Hz) excl. shift

incl. shift

(Hz)

(Hz)

1220+_640 *** 990*590*** 920?510*** 880_+520*** 970+590*** 790*400*** 790+430*** 690 i 360 610+ 150 600+ 90 630+ 170

1230+_620 1250+860 1040+560 1040_+640 12001f690 970 * 550 1060+750 710+_390 690 k 360 750 i 350 960 k 620

600 + 300

Values are means+ S.D. The significance levels were determined variables.* P c 0.05; ** P i 0.01; *** P < 0.001.

*** *** *** *** *** ***

830 _+570 with the Wilcoxon

490+_370 * 450+260*** 390+280*** 370_+250 400 i 290 280+ 130 340+250 330*310 33oi I70 330* 130 36Ok 170

33oi

90

test for continuous

144

HZ 1300 I

1000

30-31

32

33

34

Fig. 1 The mean maximum

35

36

37

pitch excluding

36

39

40 41-42

9.~.

39

40 41-42

g.w.

shift.

HZ 1300 -

1000 -

500 -

30-31

32

33

34

Fig. 2. The mean maximum

30-31

32

33

34

Fig. 3. The mean minimum

35

36

37

pitch including

35 pitch.

36

37

36 shift.

36

39

40 41-42

g.w.

II

20

54

Controls

0

2

*

***

* ** *

Glide

type Glottal

11

84

90 86 73 83 76 81 79 82 84 73 73

F/RF

Melody

levels were determined

6

2

5 3 3 0 5 9 7 8 0 18 9 6 14 7 0 7 8 18 9

0 0

Flat

10

R/FR

type a

flat. and no type. The significance

28

0

14 21 20 15 27 27 33 31 17 9

10

roll

14 24 25 17 27 21 9 8 0 0

Vibrato

roll and the melody

20 23 15 13 23 9 15 0 0 8 9

glottal

10 14 15 13 8 20 ** 18 * 3 8 8 0

Bi-phonation

glide, vibrato,

a The melody types are: F = falling, RF = rising-falling, R = rising, FR = falling-rising, X*-test with Yates’s correction.* P < 0.05; ** P < 0.01; *** P < 0.001.

20 18 24 20 25 20 30 13 15 17 2-l

I1

Shift

22 34 40 52 44 33 30 13 12 11

(n)

Cries

(Sk;)of shift, bi-phonation,

Preterm 30-31 32 33 34 35 36 31 38 39 40 41-42

Conceptual weeks

The occurrence

TABLE

with the

8

9 15 8 10 7 12 4 0 9 0

0

No

146

set 5432l-

30-31

32

33

34

35

36

37

38

39

40 41-42

g.w.

Fig. 4. The mean duration.

and significant differences from those of the fullterm healthy controls were observed up to 37 weeks conceptual age. The minimum pitch was significantly higher only at the lowest conceptual ages. Figs. 1, 2 and 3 show the mean pitch according to conceptual age in both the maximum and minimum pitch of the cries of the prematures. The duration of the cries was significantly shorter in the lowest conceptual age groups (Table I, Fig. 4). From 34 weeks onwards no significant differences were observed in the cry duration between the groups. No significant differences were seen in the melody type according to conceptual age (Table II). The majority of the cries of both the premature and the fullterm infants had a falling or rising-falling melody type. Shifts occurred in 13-30% of the cries of the prematures (Table II). Shifts were seen in 20% of the control infants. The occurrence of shift did not differ significantly with the degree of prematurity. There was a decrease in the occurrence of both bi-phonation and glide with increasing conceptual age (Table II). In the fullterm healthy controls the occurrence of bi-phonation was 2% and glide 0%. Bi-phonation and glide occurred in 8-23% of the prematures from 31 to 37 conceptual weeks. From 38 weeks onwards bi-phonation was seen in O-8% and glide in O-9%. Bi-phonation occurred significantly more often at 36-37 conceptual weeks and glide at 30-37 conceptual weeks when compared to controls. No significant differences were seen in the occurrence of vibrato and glottal roll according to conceptual age.

Discussion All infants with any symptoms or signs of disease in the newborn period were excluded from the study. The preterm infants were born at 30-37 weeks of gestation. Thus the sonograms analysed at 31-37 conceptual weeks were from cries of both newly born and 1-6-week-old infants. This type of grouping was considered justified

147

by previous results which had shown that the infants at 40 conceptual weeks cry similarly regardless of whether they had grown intra- or extrauterinely [9]. From 38 weeks onwards the prematurely born were at least one week old at the time of the cry analysis. The present study confirms earlier findings of an increased prevalence of more high-pitched cries in prematures of low gestational age [9,17]. However, even if the cries of the smallest prematures were usually high-pitched they were also sometimes more low-pitched and thus resembled the cries of fullterm infants. The differences in the pitch of the fundamental frequency in cries of prematures of low gestational age have also been noted previously [9,17]. Whether this is due to maturational factors or reflects the presence of some pathological condition, such as cerebral haemorrhage, is not known. At the time of the cry recordings ultrasound and CT-scanning were not available. Lester and Zeskind [6] observed that the cry of infants born small for gestational age was different from that of the normal newborn. Their speculation was that the pain cry characteristics may be sensitive to central nervous system stress, and are related to the risk status of the neonate, and thus analysis of infant cry features could be useful in assessment of healthy infants who might be at risk. The shifts in pain induced cries usually have had a higher pitch than the main fundamental frequency. The maximum pitch including shift is thus dependent on how common shifts are: the more shifts, the more high-pitched cries. The first cry response to a pain stimulus in healthy newborn infants is often a high-pitched shift part, mainly of very short duration, and the signal continues with a more low-pitched normal phonation. In the cries of the prematurely born the high-pitched parts were longer and occurred both together with the main fundamental frequency and as the sole cry response. In cries of infants with asphyxia [9] and meningitis [lo] the phonations have been still more unstable, with alternating highand low-pitched parts. The sudden shifts in pitch can, according to Golub (51, be due to decreased vagal tone, which allows the laryngeal muscles to contract. Lester and Zeskind [7] assume that lack of inhibition or decreased vagal tone is probably responsible for the increases and imbalance of pitch that are common in groups of infants at risk. Lester and Zeskind thought that the organization of crying in early infancy is a mirror that reflects the organization of the nervous system. Their studies indicated that prenatal risk factors affected the cry in a way that was related to the behavioural organization of the infant [6]. The cry seems to be an indicator of the maturity of the neonate as the cry characteristics change with increasing age. When the premature infant had reached the age of 38 conceptual weeks the cries were similar to those of newly born fullterm babies. In a previous study [9] we noted that infants born at 32 weeks gestational age cried when they had reached the age at which they really should have been born similarly to a fullterm newly born baby. Thus the development of the cry characteristics seems to change identically, regardless of whether the infant had lived an extraor intrauterine life. This finding is in accordance with the studies by Nijhuis et al. [ 131 who noted that the time course of development of behavioural states in utero is

148

similar to that observed by Prechtl et al. [ 141 in low-risk premature neonates. The pain cry characteristics of healthy fullterm newborn infants do not change substantially during the first days or months of life. Wasz-Hiickert et al. [18] and Thoden and Koivisto [16] studied cries of infants of different ages up to the age of six months. Neither study revealed any meaningful differences in the cry characteristics with increasing age. The cries of sick fullterm and premature infants who have diseases affecting the central nervous system have been more high-pitched than the cries of healthy infants of corresponding age [9-12,181. The changes in cry characteristics of sick infants were not only the pitch of the fundamental frequency. There was also a change in the melody type, and an increased occurrence of bi-phonation and glide. The results of the present study showed that all these cry changes were also found in cries of symptomless prematures, and they were more common the more the infant was premature. Because of the differences in the cry characteristics between the fullterm and premature infants, the gestational age should be taken into consideration in cry analysis. The high-pitched cries of small prematures may have a negative effect on the mother-infant interaction. Frodi et al. [4] pointed out that cries of premature infants elicited more negative emotions than did the term infant cry. Cries of high-risk infants were rated in a study by Zeskind and Lester [19] as sick, discomforting, aversive, and distressing. Lester and Zeskind [7] suggested that the responses of the adults to the high-risk infant cries may be unique to infants with high-pitched cries, for Freudenberg et al. [3] found that cries of infants with Down’s syndrome, which were low-pitched and flat [8], were not found to be as unpleasant as normal infant cries. Early attention to the high-pitched cry patterns in small prematures could benefit the infant-parent relationship.

References 1 Backstrom, L. and Kauppinen, M.A. (1968): The fetal weight growth. (In Finnish). Finn. Med. J., 23, 1553-1557. 2 Dubowitz, L.M.S., Dubowitz, V. and Goldberg, C. (1970): Clinical assessment of gestational age in the newborn infant. J. Pediat., 77, I-10. 3 Freudenberg, R.P., Driscoll, J.W. and Stern, G.S. (1978): Reactions of adult humans to cries of normal and abnormal infants. Infant Behav. Dev., 1, 224-227. 4 Frodi, A.M., Lamb, M.E., Leavitt, L.A., Donovan, W.L., Neff, C. and Sherry, D. (1978): Fathers’ and mothers’ responses to the faces and cries of normal and premature infants. Dev. Psychol., 14, 490-498. 5 Golub, H.L. and Corwin, M.J. (1982): Infants cry: A clue to diagnosis. Pediatrics, 69, 197-201. 6 Lester, B.M. and Zeskind, P.S. (1978): Brazelton scale and physical size correlates of neonatal cry features. Infant Behav. Dev., 1, 393-402. 7 Lester, B.M. and Zeskind, P.S. (1982): A biobehavioral perspective on crying in early infancy. In: Theory and Research in Behavioral Pediatrics, Vol. 1, pp. 133- 180. Editors: Fitzgerald, Lester, Yogman. Plenum Publishing Co., New York, NY. 8 Lind, J., Vuorenkoski, V., Rosberg, G., Partanen, T.J. and Wasz-Hiickert, 0. (1970): Spectrographic analysis of vocal response to pain stimuli in infants with Down’s syndrome. Dev. Med. Child Neurol., 12, 478-486.

149 9 Michelsson, K. (1971): Cry analyses of symptomless low birth weight neonates and of asphyxiated newborn infants. Acta Paediatr. Stand., Supplement 216. 10 Michelsson, K., Sir&, P. and Wasz-Hiickert, 0. (1977): Sound spectrographic cry analysis of infants with bacterial meningitis. Dev. Med. Child. Neurol., 19, 309-315. 11 Michelsson. K., Sin%, P. and Wasz-Hiickert, 0. (1977): Pain cry in full-term asphyxiated newborn infants correlated with late findings. Acta Paediatr. Stand., 66, 61 I-616. 12 Michelsson. K., Raes, J., Thoden, C-J. and Wasz-Hiickert, 0. (1982): Sound spectrographic cry analysis in neonatal diagnostics. An evaluative study. J. Phonet., 10, 79-88. 13 Nijhuis, J.G., Prechtl, H.F.R., Martin, C.B., Jr. and Bots, R.S.G.M. (1982): Are there behavioral states in the human fetus? Early Hum. Dev., 6, 177-195. 14 Prechtl, H.F.R., Fargel, J.W., Weinmann, H.M. and Bakker, H.H. (1979): Posture, motility and respiration in low-risk preterm infants. Dev. Med. Child Neurol., 21, 3-27. 15 Sirvio, P. and Michelsson, K. (1976): Sound spectrographic cry analysis of normal and abnormal newborn infants. Fol. Phoniatr., 28, 161-173. 16 Thoden, C-J. and Koivisto, M. (1980): Acoustic analysis of normal pain cry. In: Infant Communication: Cry and Early Speech, pp. 124-151. Editors: T. & J. Murry. College-Hill Press, Houston, Texas. 17 Thoden, C-J., Jarvenpll, A-L. and Michelsson, K. (1982): Sound spectrographic pain cry in prematures. In: Infant Cry: Theoretical and Research Prospectives. Editor: B. Lester. Plenum Publishing Co., New York, NY, in press. 18 Wasz-Hockert, 0.. Lind, J., Vuorenkoski, V., Partanen, T.J. and Valanne, E. (1968): The Infant Cry. A Spectrographic and Auditory Analysis. Clinics in Developmental Medicine No. 29. Spastics International Medical Publications in Association with William Heinemann Medical Books Ltd. London. 19 Zeskind, P.S. and Lester, B.M. (1978): Acoustic features and auditory perceptions of the cries of newborns with prenatal and perinatal complications. Child Dev.. 49, 580-589.