Human infant crying has been researched as a non-invasive tool for assessing neurophysiological states at an early developmental stage. Little is known about the acoustic features of spontaneous cries in preterm infants, although their pain-induced cries are at a higher fundamental frequency (F0) before term-equivalent age. In this study, we investigated the effects of gestational age, body size at recording and intrauterine growth retardation (IUGR) on the F0 of spontaneous cries in healthy preterm and full-term infants at term-equivalent age. We found that shorter gestational age was significantly associated with higher F0, although neither smaller body size at recording nor IUGR was related to increased F0 in preterm infants. These findings suggest that the increased F0 of spontaneous cries is not caused by their smaller body size, but instead might be caused by more complicated neurophysiological states owing to their different intrauterine and extrauterine experiences.
For several decades, acoustic features of infant crying have been studied as a possible non-invasive tool for assessing neurophysiological states [1–5]. Previous studies have indicated that an abnormally high frequency (F0) (e.g. mean F0 > 600 Hz) of infant cries is associated with medical conditions, including chromosomal, endocrine, metabolic and neurological disturbances at an early developmental stage . Preterm birth is also a factor in higher F0 cries during early infancy [1–3]. Pain-induced cries in preterm infants have been reported to be higher in F0 before term-equivalent age compared with those of full-term newborns [2,3], although such differences disappeared around term-equivalent age [2,4]. A higher vocal F0 is generally related to smaller body size, especially shorter vocal folds ; therefore, it is possible that the higher F0 of preterm infants simply reflects premature body development.
Vocal F0 also depends on a complex interaction between laryngeal and respiratory controls [1,6]. Specifically, vagal inputs from the right nucleus ambiguus of the medulla are assumed to have inhibitory effects on laryngeal muscle contraction and tightening of the vocal folds . Therefore, diminished vagal activity might cause laryngeal muscle contraction and tightening of the vocal folds, resulting in a higher F0 . As preterm infants exhibit reduced vagal activity even at term-equivalent age , the F0 of their cries might be affected by the altered vagal activity as well as smaller body size.
However, the effects of body size and other factors related to neurophysiological states (e.g. intrauterine growth retardation (IUGR)) on the F0 of cries in preterm infants have not been investigated [2–4]. In addition, although spontaneous cries (those unaffected by external acute stress) have a higher internal consistency than pain-induced cries in full-term neonates , to our knowledge, no studies have assessed the F0 of spontaneous cries in preterm infants. In this study, we performed acoustic analysis of the F0 of spontaneous cries before feeding in both healthy preterm infants at term-equivalent ages and full-term newborns. We investigated the effects of gestational age, body size at recording and IUGR on F0 to assess the relationship between preterm birth and the F0 of spontaneous cries at term-equivalent ages.
2. Material and methods
Forty-four healthy preterm infants (gestational age less than 37 weeks, weight at birth less than 2500 g) and 20 full-term (FT) newborn infants (gestational age more than or equal to 37 weeks, weight at birth more than or equal to 2500 g) were enrolled in this study. Preterm infants were assigned to two subgroups according to gestational age: very preterm (VP) infants (gestational age less than 32 weeks, n = 22) and moderate-to-late preterm (MLP) infants (gestational age between 32 and less than 37 weeks, n = 22). They were also grouped according to intrauterine growth: small for gestational age preterm (SGAP) infants (n = 19), and adequate for gestational age preterm (AGAP) infants (n = 25) (see the electronic supplementary material).
(b) Cry recording and acoustic analysis
Cry recordings were performed at term-equivalent age (postmenstrual age between 37 and less than 42 weeks) at Kyoto University Hospital. The spontaneous cries of each infant in an open crib less than 30 min before feeding were recorded for 60 s using a wave recorder (EDIROL R-09; Roland Corp., Los Angeles, CA, USA) at a 44.1 kHz sampling rate and 16 bit-quantization. Cry utterances containing broad regions of environmental noise were excluded from acoustic analyses to avoid artefacts when determining F0. F0 measurements were performed using PRAAT v. 5.2.35 . A total of 2321 cry utterances were extracted manually from the PRAAT spectrogram of all cry recordings for the acoustic analysis. The cry utterances were down-sampled to 22.05 kHz and low-pass filtered at 10 kHz to eliminate outliers and artefacts; then the minimum, mean and maximum F0 were determined using the PRAAT autocorrelation algorithm. The F0 values were averaged for each infant (see the electronic supplementary material for more information).
One-way ANOVA of the gestational group revealed significant group differences in the minimum (F2,61 = 7.87, p<0.001, η2 = 0.10), mean (F2,61 = 9.17, p < 0.001, η2 = 0.23) and maximum F0 (F2,61 = 13.32, p < 0.00001, η2 = 0.30) (table 1). Post hoc testing (Bonferroni) revealed that the minimum (p < 0.0001) and mean F0 (p = 0.04) were significantly different between the VP and MLP groups. In addition, the minimum (p = 0.03), mean (p < 0.001) and maximum F0 (p < 0.0001) were significantly different between the VP and FT groups, whereas the maximum F0 (p = 0.01) was significantly different between the MLP and FT groups. We also examined the differences between the SGAP and AGAP groups using the two-tailed Student's t-test; however, there were no differences between the groups for any of the F0 variables (table 2).
Pearson's and Spearman's correlation analyses for all participants (n = 64) revealed that gestational age was significantly correlated with the minimum (rS = −0.32, p < 0.01; figure 1a), mean (rS = −0.38, p < 0.01; figure 1b) and maximum F0 (rS = −0.48, p < 0.0001; figure 1c). However, infant weight at recording was marginally correlated with minimum F0 (rS = 0.24, p = 0.06), but not mean (rS = 0.11, p = 0.39) or maximum F0 (rS = –0.01, p = 0.96). Other body size variables (height, head and chest circumference) at recording were not significantly correlated with any F0 variable (see the electronic supplementary material).
For multiple regression analyses, we used gestational age, postmenstrual age and weight at recording to avoid the collinearity of predictors due to strong correlations among the variables related to preterm birth or body size at recording. All F0 variables could be predicted by gestational age rather than postmenstrual age or weight at recording (table 3).
This is the first study to determine the relationships between preterm birth and the F0 of spontaneous cries in human infants at term-equivalent age. We found that shorter gestational age was significantly associated with higher F0 of spontaneous cries; however, neither smaller body size at recording nor IUGR was related to increased F0 of spontaneous cries in preterm infants.
Preterm infants are usually smaller in body size than full-term infants at term-equivalent age. Therefore, the higher F0 of spontaneous cries may reflect the smaller body size at recording in preterm infants. However, we did not find any negative effects of body size at recording on any F0. As the relationship between body size and vocal folds is not strong, especially in the same age and sex class within species , it is possible that the body size measurements used in this study did not reflect individual differences in vocal fold size. By contrast, we found a weak positive effect of weight at recording on minimum F0. Vocal F0 is related to subglottic pressure from the lungs as well as the size of the vocal folds . Therefore, higher subglottic pressure might increase the minimum F0 in infants with increased body weight at recording. As the MLP group had lower body weight at recording than the FT and VP groups (see the electronic supplementary material), the relatively low minimum F0 in the MLP group might have been due to their lower body weight at recording. However, our data at least suggest that the increased F0 of spontaneous cries in preterm infants is not due to their smaller body size at recording.
Nevertheless, it remains unclear why preterm birth is associated with an increased F0 of spontaneous cries at term-equivalent age. Our data did not provide a direct answer to this question. However, one possibility is that the increased F0 might reflect the reduced vagal activity in preterm infants. Vagal input has an inhibitory effect on laryngeal contraction and vocal fold tightening; therefore, diminished vagal activity is assumed to cause increased vocal fold tension and higher F0 . As preterm infants at term-equivalent age exhibit lower resting vagal activity than full-term newborns , their reduced vagal activity might cause higher tension in the vocal folds, resulting in an increased F0 of spontaneous cries. To clarify this relationship, further studies involving direct investigation of vagal activity by measuring respiratory sinus arrhythmia are required.
An additional important point is that we included preterm infants who exhibited IUGR because no studies have assessed F0 of cries in SGA infants. As some preterm infants could exhibit IUGR, which can negatively affect neurobehavioural maturation , it is possible that this negative maturation might be related to the increased F0 of spontaneous cries in preterm infants. However, we were unable to identify the effects of IUGR on the F0 in preterm infants, suggesting that the increased F0 in preterm infants does not reflect a neurophysiological vulnerability specific to IUGR. Moreover, it is important to note that our reported F0 value in preterm infants did not necessarily deviate from the normal F0 range (i.e. 200–600 Hz) reported previously . This might have been because we included low-risk preterm infants without severe complications. Nevertheless, we should investigate the effects of preterm birth in a study with stronger homogenization including full-term SGA infants.
It is also possible that the increased F0 of spontaneous cries in preterm infants is due to the longer postnatal period. This study did not include postnatal age in the major analyses because of a strong negative correlation between gestational and postnatal age. Nevertheless, each postnatal and postmenstrual age was positively correlated with every F0, even after controlling for gestational age (see the electronic supplementary material). In mammals, the production of a cry involves cortical regions including the anterior cingulate gyrus, as well as reflexive central pattern generators in the periaqueductal grey and nucleus retroambiguus [12,13]. Recent study of full-term infants shows that cry melodies (variation in F0 over time) become increasingly complex during early interaction with the environment in the first months of life , suggesting the early contribution of cortical control to the F0. However, it remains unclear in preterm infants at term-equivalent age; therefore, additional longitudinal studies will help to assess the effects of the postnatal and maturational period on the F0 in preterm infants.
In conclusion, this study revealed that preterm birth was associated with an increased F0 of spontaneous cries at term-equivalent age, regardless of the smaller body size at recording or IUGR. Hence, the increased F0 in preterm infants may have been caused not by smaller body size at recording, but rather by more complicated neurophysiological states due to different intrauterine and extrauterine experiences.
The research reported here was supported by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science and the Ministry of Education Culture, Sports, Science and Technology (24119005, 24300103 to M.M-Y.; 14J06302 to Y.S.) and ERATO Okanoya Emotional Information Project, Japan Science and Technology Agency.
We thank M. Shibata, M. Inagawa and the hospital staff for their invaluable help in data collection, Y. Nonaka and N. Naoi for technical support, and Y. Maehara and M. Imafuku for helpful comments on earlier drafts of this paper.
- Received April 27, 2014.
- Accepted July 24, 2014.
© 2014 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited.