Early vocalization of preterm infants with extremely low birth weight (ELBW), part I: from birth to expansion stage

Clinical Linguistics & Phonetics, April 2012; 26(4): 330 344 2012 Informa UK Ltd ISSN: 0269-9206 print / ISSN 1464-5076 online DOI: 10.3109/02699206.2011.636499 Early vocalization of preterm infants with extremely low birth weight (ELBW), part I: from birth to expansion HELENA TÖRÖLÄ 1,2, MATTI LEHTIHALMES 1, HANNA HEIKKINEN 3, PÄIVI OLSÉN 2, & ANNELI YLIHERVA 1 1 Faculty of Humanities, Logopedics, University of Oulu, Oulu, Finland, 2 Department of Paediatrics and Adolescence, Oulu University Hospital, Oulu, Finland, and 3 Department of Mathematical Sciences/IT Administration Services, University of Oulu, Oulu, Finland (Received 7 February 2011; revised 6 October 2011; accepted 16 October 2011) Abstract The vocalization of preterm infants with extremely low birth weight (ELBW) up to the expansion was systematically described and compared with those of healthy full-term infants. The sample consisted of 18 preterm ELBW infants and the control group of 11 full-term infants. The follow-up was performed intensively using video-recordings. The vocalization of the preterm and full-term infants was analyzed quantitatively according to the categorical s created by Oller. A descriptive analysis of all the vocalizations produced by the infants was performed. The preterm infants entered the primitive articulation later than the full-term infants and failed to produce more skills during that. According to this sample, there was no difference in entering the expansion, but the preterm infants failed to produce more skills than the full-term infants. The number of vocalization acts varied differently by age between the groups. Keywords: preverbal vocalization, prematurity, extremely low birth weight, infant Introduction Infant vocalization has been studied since the 1940s and in the beginning the studies on preverbal vocalization concentrated on full-term infants. There are few follow-up studies on preterm infants' early vocalization (Beckwith, Sigman, Cohen, & Parmelee, 1977; Van Beek, Hopkins, & Binne Hoeksma, 1994). In a study by Beckwith et al. (1977), preterm infants' early vocalizations started to increase between 3 and 8 of age, while in full-term infants this happened earlier, by the age of 3. Nevertheless, both infant groups produced approximately the same number of vocalizations. Van Beek et al. (1994) studied the development of communicative behaviors in healthy full-term and preterm infants, whether small or appropriate for gestational age from the age of 6 weeks up to the Correspondence: Helena Törölä, MA, Speech and Language Therapist, Department of Paediatrics and Adolescence, Oulu University Hospital, PO Box 23, Children and Women, FI-90029 OYS, Finland; Faculty of Humanities, Logopedics, University of Oulu, PO Box 1000, FI-90014, Oulu, Finland. Tel: +358405148385. Fax: +35883155858. E-mail: helena.torola@oulu.fi, helena.torola@dnainternet.net

Early vocalization of preterm infants with ELBW 331 age of 18 weeks of corrected age. In producing speech-like vocalizations, infants born between 32 and 34 weeks of gestation vocalized less than preterm infants born before 32 weeks of gestation and full-term infants. In general, typically developing infants start to produce a rich variety of vocalizations soon after birth. The basic sensory capacities to discriminate speech sounds already exist before that, even prenatally (Draganova et al., 2005). The human fetus is exposed to the native language prenatally, and by the last trimester of pregnancy the fetus becomes sensitive to melody contour in both language and music. Consequently, the newborn infant reflects the prosody of the ambient language in the cry melody (Mampe, Friederici, Christoph, & Wermke, 2009). According to behavioral studies, the newborn infant is able to discriminate between the voice of the mother and a stranger (DeCasper & Fifer, 1980), and according to recent work using near-infrared spectroscopy, the human brain seems to become sensitive to human voice between 4 and 7 of age, being able to discriminate between various emotional expressions (Grossmann, Oberecker, Koch, & Friederici, 2010). Furthermore, the infant experiences a continuous change of the structures of the vocal tract, which modifies vocalizations and supports future development (Kent & Miolo, 1995). The developmental changes are functional, histological and biomechanical, continuing up to the age of 18 years (Kent, 1999). There are different models concerning the preverbal vocalization of infants from birth up to the first meaningful word. Researchers emphasize that the maturation of vocalization follows a specific order in typically developing children (Iivonen, 1991; Oller, Eilers, Neal, & Schwartz, 1999; Oller, Wieman, Doyle, & Ross, 1976; Roug, Landberg, & Lundberg, 1989; Stark, Bernstein, & Demorest, 1993). Stage theories have been presented in many studies (Elbers, 1982; Koopmans-van Beinum & van der Stelt, 1979, 1986; Nathani, Ertmer, & Stark, 2006; Oller, 1980, 2000; Stark et al., 1993). The three models that are often cited when studying infant vocalization are presented in Table I. Originally, Oller et al. (1976) concentrated on describing speech-like utterances in preverbal vocalizations; the basic element of vocalization was a syllable with a consonantal element. Later on, Oller (1980) described infant vocalizations as a gradual, developmental process from arbitrary sounds to spoken language. At that time Oller introduced the term meta-phonology to refer to the general parameters of language. Later Oller (2000) revised the s as phonation, primitive articulation, expansion, canonical, and integrative. In addition, Oller revised the idea of infra-phonology, which includes the development from the first vocalizations through well-formed syllables up to meaningful utterances. Oller considers the wellformed syllables in canonical babbling as a border between pre-canonical protophones and later speech-like productions. The term protophones, unique to humans, refers to the utterance types that appear to be precursors to speech, from quasi-vowels to canonical babbling. However, non-protophones, such as vegetative sounds, cry and laughter, are precursors to speech, as well (Oller, 2000). A totally different perspective of preverbal vocalization than that of Oller's was presented by Koopmans-van Beinum and van der Stelt (1979, 1986). The development of early sound production was studied in two infants from birth up to the age of 1 year and five s were summarized in the development of preverbal vocalization. The researchers took into account both the phonation and the articulation movements during one respiratory phase and described them, arguing that the development of respiratory skills affects the quality of voice and articulation (see Table I). Stark, Rose, and Benson (1978) described in their categorization the preverbal vocalizations from early cry and discomfort, vegetative, laugh, and comfort sounds up to speech of vocal play. Later on Stark (1979) defined the five s of preverbal vocalization: (1) reflexive crying and vegetative sounds, (2) cooing and laughter,

Table I. The preverbal s of development and age range according to Oller, Koopmans-van Beinum and van der Stelt, and Nathani, Ertmer, and Stark. 332 H. Törölä et al. Oller (1980, 2000) 1980 2000 Koopmans-van Beinum and van der Stelt (1979, 1986) Nathani et al. (2006) Stages Age range Stages Age range Stages Age range Stages Age range 1. Phonation 0 1 month 1. Phonation 2. Goo 2 3 2. Primitive articulation 3. Expansion 4 6 3. Expansion 4. Canonical 7 10 4. Canonical 5. Variegated babbling 11 12 5. Integrative 0 2 1. Uninterrupted phonations without any articulatory movement 1 4 2. Interrupted phonation within one breath unit without any articulatory movement 3 8 3. One articulatory movement per breath unit combined with continuous or interrupted phonation 5 10 4. Strikingly large decrease of articulatory movements and increase of uninterrupted phonation without articulation; all kinds of variations in the phonatory domain concerning intonation, duration, and intensity 5. Reduplicated articulatory movements during one respiration cycle combined with continuous or interrupted phonation 0 1.5 1.5 2.5 2.5 5 5 6.5 6.5 10 1. Reflexive level 0 2 2. Control of phonation 1 4 3. Expansion 3 8 4. Basic canonical syllables level 5. Advanced forms 5 10 9 18

Early vocalization of preterm infants with ELBW 333 (3) vocal play, (4) reduplicated babbling and (5) non-reduplicated babbling and expressive jargon. Stark also argued that infants use the same vocal segments in cry, vegetative and comfort sounds as they later use in various contexts. The Stark Assessment of Early Vocal Development-Revised by Nathani et al. (2006) is based on Stark's earlier work; it includes five s, which are quite similar to those of Oller (Table I). In spite of differences between models, in all of them, the first syllable production is distinguished from the earlier developmental s. The names and the ages-of-emergence of the different s seem to vary across models. In this study, we focused on the first three s of vocalization, namely phonation, primitive articulation and expansion s that precede syllable production. All researchers distinguish the start of syllable production from the earlier developmental s. Canonical babbling starts at approximately the age of 7. Our aim to describe the early vocalization s is based on the need in clinical follow-up of preterm infants to know what is typical for these infants and whom to keep under intensive observation. For this purpose, we collected every vocalization act of the infants and described them in addition to those described by Oller (2000). We wanted to analyze the early vocalizations of preterm infants from birth up to the expansion and to compare the development with that of healthy full-term infants. In addition, we wanted to describe as literally as possible every vocalization act the infants produced in order to find out whether there are differences in the numbers, types and diversity of the vocal productions between the preterm and full-term infants within the s. Method Study group The study group included 18 preterm infants, born with birth weights less than 1000 g, except for one with birth weight of 1020 g. There were 7 girls and 11 boys with normal hearing. The mean birth weight was 880 g (range 670 1020 g) and the mean gestational age 27 weeks (range 23 30 weeks). More health information on the preterm infants is presented in Table II. All possible consecutive preterm births at Oulu University Hospital in Northern Finland from October 2003 until April 2005 were included, in the case that the parents had consented to participate and the family was geographically attainable and Finnish speaking. The exclusion criteria were congenital or chromosomal anomalies and hearing loss. Only one of the infants in the final study sample had a mild diplegia, which was diagnosed later on; others did not have any motor problems. One infant had severe visual impairment and was included because he had no other developmental problems. The preterm infants were most often firstborn children in their families. Control group The control group of 11 full-term infants, 4 boys and 7 girls, born at Oulu University Hospital, was collected at the maternity ward from August to November 2003. All the healthy full-term infants with normal hearing, whose parents volunteered and whose families were geographically attainable, were included in the study and their preverbal vocalizations were followed from birth up to the expansion. The families were Finnish speaking, except one family, where the father was Spanish speaking. In that case, the Finnish-speaking mother stayed at home taking care of the infant throughout the data collection.

334 H. Törölä et al. Table II. Background information of the preterm group. Code Birth weight Gest. age at birth SGA RDS BPD Asphyxia PVL IVH IIIº P1 925 g 27 P2 925 g 30 x x x P3 910 g 26 x x P4 905 g 29 x x P5 980 g 30 x P6 680 g 25 x x P7 980 g 28 x x P8 680 g 24 x x x P9 790 g 26 x x P10 975 g 27 x P11 810 g 27 x x P12 890 g 27 x P13 1020 g 30 x x P14 932 g 26 x x P15 670 g 23 x x x P16 975 g 25 x P17 760 g 25 x x P18 990 g 27 x x Notes: All the full-term infants were healthy newborns. x, diagnose exists; SGA, small for gestational age; RDS, respiratory distress syndrome; BPD, bronchopulmonary dysplasia; PVL, periventricular leucomalasia; IVH, intraventricular hemorrhage, n = 18. Mothers in both the preterm and full-term groups had similar levels of education, i.e. a college or university degree. The occurrence of otitis media episodes was controlled consulting the mother at every home visit during the study, and the number of these episodes was similarly distributed in both groups. Data collection Video-recordings were used to collect data on early vocalizations. First, a pilot study with one infant was performed to test the whole procedure. This child was not included in the sample. The preterm infants were video-recorded for the first time at the gestational age of 37 weeks at the hospital. The second recording took place at 40 weeks of gestational age, and after that at home every 2 weeks up to the corrected age of 3, and then once a month up to the end of the expansion. The first video-recording of the full-term infants took place at the hospital before discharge. The infants were followed at home every second week up to the age of 3 and after that once a month up to the end of data collection. The total number of recordings was 329. The median number of recordings was 12 for the preterm infants and 11 for the full-term infants. The vocalization data were collected when the infant was alert and was actively interacting with the mother or a nurse. The video-recording was stopped when the infant became tired or hungry and had vocalized enough for an assessment to be performed. The alertness of the infants became longer with increasing age and therefore also the length of the video-recordings increased. During the entire data collection period, the durations of the video-recordings varied from 7 up to 26 min in the preterm infants and from 5 up to 25 min in the full-term infants. The median duration of a video-recording session was 11 min during the first four video-recordings and 17.5 min for the rest of the recordings in the preterm infants. The median duration was 10 min during the first four video-

Early vocalization of preterm infants with ELBW 335 recordings and 21.5 min for the rest of the recordings in the full-term infants, respectively. Usually, the mother and sometimes the father participated in the sessions. All vocalizations a child produced during each video session were transcribed and analyzed. Only those vocalization episodes were excluded when the infant had food, a utensil or a toy in the mouth or was crying simultaneously. Vegetative sounds, such as hiccups, burps and coughs were also excluded, because of their reflexive nature. However, cry was included, but not laughter. In several cases, laughter was not heard during the video-recordings, although the parents reported that it had appeared. Notes on each video-recorded session were recorded in a diary. The video-recordings were stored in Transana (Fassnacht & Woods, 2007, version 2.20), a program for transcription and qualitative analysis of large volumes of digitized video and audio data. It provides tools for building a database and file manipulation as well as for organizing and storing data. Analysis The analysis of early vocalization was based on Oller's (2000) model, which offers a method to analyze this type of data and is widely used in the analyzes of preverbal vocalization. We had to build subgroups of vocalizations heard during the various s in order to describe the preverbal development of preterm infants in detail and to compare it with that of the fullterm infants. An individual skill was accepted when it first appeared. The infant was defined to have reached a certain developmental of early vocalization if at least 20% of the vocalization acts belonged to it (Nathani et al., 2006). In some infants that 20% consisted of only one vocalization type, in others it may have consisted of five different vocalization types belonging to the. The time lag between the first appearance of a vocalization and the time to reach the 20% criterion was reported in both groups. Because the vocalizations in the phonation are reflexive by nature, the calculation was started since the first skills from the primitive articulation were heard. The number of utterances produced during each video-recording was calculated in order to define whether there was a quantitative difference in vocalization between the preterm and full-term infants. The ages of emergence of vocalization skills are presented as medians and the ages that describe the skills of the preterm infants are all given in corrected ages. An experienced speech and language therapist (SLT) analyzed 10% of the randomly chosen vocalizations of the infants. The inter-rater agreement in analyzing the vocalizations was 97% in the preterm and 100% in the full-term infants. In defining individual vocalization acts, the inter-rater agreement between the first author and the SLT was 96% in the preterm and 97% in the full-term infants. Statistical analysis Because the sample size was quite small and distributions of most of the variables were skewed, the Mann Whitney U test was used in order to compare the ages of entering s and the number of skills the infants failed to produce at each between preterm and full-term infants. Video-recording durations varied by infant and by recording time, thus the vocalization frequency was constituted by dividing the number of utterances by the duration of that recording. Vocalization frequency was observed until the age of 9, when most of the infants had passed the expansion. A mixed model (McCulloch & Searle, 2001) was used to assess the age-dependent difference between the vocalization of preterm and full-term infants. In addition to age, its second and third powers were included in the model as fixed explanatory variables to allow a cubic trend to be included

336 H. Törölä et al. in the vocalization by age, and the infant was treated as a random explanatory variable. Residual variance was allowed to vary by age. Statistical analyzes were performed, using the IBM SPSS statistical software (version 19.0, SPSS Inc., Chicago, IL). Ethical aspects The research protocol was approved by the Ethical Committee of The Northern Ostrobothnia Hospital District. A written, informed consent was obtained from the parents. Results The preterm infants were at the phonation from the beginning of the study, 37 weeks of gestational age, up to the corrected age of 2.5, and the full-term infants from birth up to the age of 2 (Table III), when they entered the primitive articulation. Both groups entered the expansion at the age of 5. In general, in preterm infants, the variation in the development of early vocalization was more extensive compared with fullterm infants. I Phonation At the phonation, in addition to vegetative sounds, cry and normal phonation, we separated three different phenomena of quasi-resonant nuclei, where the cavity friction of the vocal tract is more or less compressed, namely creaking, groaning and fussing. Creaking means a strong, sometimes even metallic vocalization that is connected with slow wriggling and stretching movements of the body, which may possibly precede crying. Groaning sounds are less compressed, but are also connected with comprehensive body movements. Groaning precedes, for instance, feeding and crying. The friction noise, which is a voiceless [h]-like resonance caused by the air passing through a narrowed part of the vocal tract (see Pike, 1943, p. 71), is defined in the present study as fussing. When fussing first appeared, it was connected with the movement of the head or body or flailing of the extremities. Later, during the following s, it could be produced without body movements as with the rest of the phonation- skills. Examples of all the various vocalization manners can be found in the supplementary material. All infants produced vocalizations from the first recording (Table III). The median ages of and age ranges in reaching the vocalization skills, as well as the appearance of certain skills and the number of infants reaching the various skills at the phonation, are also presented in Table III. Although cry was not heard until 2 weeks of corrected age in some preterm infants, and 1 month in some full-term infants, the parents reported that the preterm infants had expressed cry before the gestational age of 37 weeks and the full-term infants since birth. II Primitive articulation Cooing and joyful normal phonation were separated because not every infant produced the velar closure, heard as gargling-like vocalizations, in connection with joyful phonation and there seemed to be a time difference in their appearance (Table III). Normal phonation appears already at the phonation, but the tone is different at this, being delighted and more controlled. Vowel- and diphthong-like vocalizations as well as distress were extracted.

Table III. Early vocalization of the preterm and the full-term infants. Preterm infants (n = 18) Full-term infants (n = 11) I Phonation Md Age range Number of infants Md Age range Number of infants p-value a reaching skill reaching skill Entering the 37 wk ga 37 wk ga Birth Birth Cry 37 wk ga 37 wk ga 2 wk ca 18 Birth Birth 1 mo 11 Normal phonation 37 wk ga 37 wk ga 2 mo ca 18 Birth Birth 1 mo 11 Creaking 37 wk ga 37 40 wk ga 18 Birth Birth 2 wk 10 Groaning 37 wk ga 37 37 wk ga 18 Birth Birth 2 wk 11 Fussing 1.5 mo ca 1.5 2 mo ca 18 1.5 mo 1.5 2 mo 11 II Primitive articulation Entering the 2.5 mo ca 1.5 3 mo ca 2 mo 1.5 2.5 mo 0.019 Joyful phonation 2.5 2 6 17 2 1.5 3 10 Diphthong-like phonation 4 3 6 9 3 1.5 6 11 Vowel-like phonation 3.5 2.5 6 16 2.5 1.5 5 11 Ingressive phonation 7 2 13 8 3 1.5 8 9 Distress 3 1.5 5 18 2 2 3 11 Cooing 3 2 6 16 3 1.5 4 11 Fussing + normal phonation 2.5 1.5 4 18 2.5 1.5 3 11 III Expansion Entering the 5 mo ca 4 6 moca 5mo 4 6 mo 0.740 Interruption of phonation 5 5 7 18 5 4 7 11 Blowing 6.5 6 10 10 9 6 11 7 Raspberries 10 5 14 14 5 5 9 11 Marginal babbling 5 47 18 6 4 7 11 Exploring the extremities of phonation 5 4 7 18 5 4 6 11 Exploring the extremities of articulation 10.5 6 15 10 8 6 12 7 organs Play with phonation manners 7 5 14 14 9 5 11 10 Notes: Ca, corrected age; ga, gestational age; Md, median; mo, month; n, number; wk, week. a Mann Whitney U test was used to compare the groups in entering the s and presented as p-value. Early vocalization of preterm infants with ELBW 337

338 H. Törölä et al. At the primitive articulation the infants combined fussing with normal phonation. Ingressive phonation was found already at the primitive articulation. The preterm infants entered the primitive articulation at the median age of 2.5, while the full-term infants did so at the median age of 2 (U = 47.500, p = 0.019), the difference being statistically significant. The preterm infants reached all milestones except cooing on average 2 4 weeks later than the full-term infants (Table III). Among the primitive articulation skills found in the groups was diphthong-like phonation. The eight diphthong-like phonations found in both the preterm and the full-term infants were the following: [au], [ey], [ou], [yi], [æi], [æu], [æy],[øy]. The appearance of the individual primitive articulation skills in the preterm infants varied from the median age of 2.5 up to 7, and the differences between the infants were extensive. The fullterm infants reached all the skills at this during a period of 2, although there were individual differences between the infants. In three cases, developing skills from the primitive articulation were heard 1 month before the preterm infant entered the, while every full-term infant entered the with no previously heard skills from the. At the primitive articulation, only 4 of the 18 preterm infants produced all the individual vocalization skills typical for this, while 9 of the 11 full-term infants did so (Table III). The development in reaching the different skills showed more variation in the preterm than in the full-term infants. All but three preterm infants failed to produce at least one developmental milestone, whereas only three fullterm infants failed to produce one. The mean of skills the infants failed to produce at the primitive articulation was 1.27 and 0.27 for the preterm and full-term infants, respectively (U = 37.000, p = 0.004). The difference was statistically significant (Table IV). Table IV. Number of early vocalization skills that the preterm and full-term infants failed to produce from birth up to the expansion Number of missing skills Preterm infants (n = 18) Full-term infants (n = 11) M Range M Range p-value a I Phonation 0 0 0.09 0 1 0.707 II Primitive articulation 1.27 0 3 0.27 0 1 0.004 III Expansion 1.38 0 3 0.72 0 3 0.076 Note: M, Mean. a Mann Whitney U test was used to compare the groups. III Expansion At the expansion, interruption of phonation was analyzed as a separate vocalization (Table III). Interruption of phonation has an essential role in consonant production. In addition, play with phonation manners was observed, which means that the infants produced utterances moving rapidly, e.g. from oral to nasal sounds or vice versa, such as [ngn], or arbitrarily combined many kinds of speech sounds, such as [he:stn] or [mby:dn]. Glides from stops to the respective voiced consonants [kŋ] and fricatives to a combination of voiced and nasal consonants [hæbm] were also typical of this type of early vocalizations. Marginal babbling includes single syllables or arbitrary syllable repetitions that lack the rhythm of the canonical syllable chains, whereas exploring the extremities of articulation organs include skills

Early vocalization of preterm infants with ELBW 339 such as smacking and cracking. The terminology exploring the extremities of voice refers to variations in characteristics such as volume, pitch, intonation and stress. The median age of entering the expansion was 5 for both groups of infants. During the expansion the age variation between the first and the last skill was more extensive than at the primitive articulation : medians were within a 5.5 month range for the preterm infants and within a 4-month range for the full-term infants (Table III). The preterm infants reached most of the skills on average at the same age or later than the full-term infants. Two preterm infants expressed occasional vocalizations characteristic of the expansion 1 month before producing 20% of skills belonging to that and entering the. Instead, six full-term infants showed signs of developing skills characteristic of the expansion before entering that : three infants 2 and three infants 1 month before entering the. At the expansion, half of the preterm infants did not produce various vocalizations to the extent of the full-term infants. At the expansion the mean of the skills that the infants failed to produce was 1.33 for the preterm infants and 0.7 for the full-term infants. The quantity of vocalizations The number of vocalizations per minute varied by age (Figure 1). For both groups the vocalization trend over age first increased, then decreased and finally increased again. These Figure 1. Group-specific estimated curves for vocalization over age.

340 H. Törölä et al. Table V. Estimated effects of the group variable (pre or full-term), age variables and their interactions on vocalization by mixed model. Estim. coefficient Standard error p-value Intercept 0.435 0.103 <0.001 Preterm versus full-term 0.893 0.117 <0.001 Age 0.352 0.124 0.005 Age (pre versus full) 0.520 0.140 <0.001 Age 2 0.063 0.037 0.092 Age 2 (pre versus full) 0.126 0.043 0.004 Age 3 0.003 0.003 0.241 Age 3 (pre versus full) 0.009 0.003 0.010 Note: The full-term group was treated as a reference category. changes in trend were placed differently between the preterm and full-term infants. In addition, during the phonation the preterm infants vocalized more than the full-term ones. Hence, the differences in vocalization between the preterm and full-term infants were assessed using a mixed model with a cubic trend by age. The estimated parameters were different for the groups (Table V) inducing quite different estimated curves for the groups. In practice this means that first, during the phonation, the preterm infants vocalized more than the full-term infants. During primitive articulation the preterm infants decreased and the full-term infants increased their vocalization. By the end of the expansion, the vocalization of preterm infants dramatically declined, while that of the full-term infants increased. Not all the infants were equally active in their vocal productions during the video-recordings. There were two preterm infants who did not vocalize at all during one video-recording session and 11 preterm infants produced fewer than six utterances during 1 2 video-recordings. One full-term infant was silent during two recordings and 10 full-term infants produced fewer than six utterances during 1 5 video-recordings. Discussion The preterm infants entered the primitive articulation later than the full-term infants. No other differences were found in the development of early vocalization between the preterm and the full-term infants when analyzed, using Oller's (2000) categorization. However, there was wider age variation among the preterm infants in attaining the various skills included in each of vocalization (see Table III) than among the full-term infants. Especially, in the primitive articulation and expansion s they developed fewer skills compared with the full-term infants, which has not been reported in earlier studies. The number of vocalization acts per time unit varied differently by age in the groups. The results of this study are in line with previous studies concerning the early vocalization of preterm infants (Nathani et al., 2006; Oller, 2000). In the present study, even though there was no great variation when the preterm and the full-term infants reached various vocalization s, the differences between individuals were notable in both groups, especially in the preterm group. All infants failed to demonstrate skills whether they had diagnoses or not. The number of diagnoses the infant had got during the neonatal period did not seem to be connected to the number of skills the infant failed to produce. There were preterm infants with, for example, three different diagnoses and who failed to produce only one

Early vocalization of preterm infants with ELBW 341 vocalization type and, on the other hand, infants with no diagnoses at all or only RDS who failed to produce four or five vocalization types. In summary, it could be said that both in preterm and full-term groups, there was a lot of variation in preverbal vocalization and there was a lot of variation among the preterm infants as well as among the full-term infants. The preterm infants were most often firstborn children in their family while the full-term infants were born later. Although it has been shown in several studies that firstborn children, in general, score higher in communicative skills than later born children (Berglund, Eriksson, & Westerlund, 2005), in the present study that was not observed. It may be that the extremely low birth weight (ELBW) preterm infants have a greater risk of speech and language problems because of their difficulties during the neonatal period more than because of environmental factors. The previous studies concerning the development of early vocalization of preterm infants have concentrated on the number of vocalization acts (Beckwith et al., 1977; Van Beek et al., 1994). Beckwith et al. (1977) observed an increase in the vocalization of preterm infants between 3 and 8, and in full-term infants at the age of 3, which are similar findings to those of the present study. The preterm infants vocalized more than the full-term infants since the beginning of data collection, and the nature of the vocalization was reflexive during the phonation (see Nathani et al., 2006; Oller, 1980). The fullterm infants, who were in the reflexive period in the present study, vocalized and moved less during this phase of development. The lesser degree of primitive vocalization at the phonation could reflect the maturity of the infant. It has been reported that preterm infants are not able to differentiate actions from each other, as seen, for example, in gross motor functions (Lüchinger, Hadders-Algra, Van Kan, & de Vries, 2008) and feeding (Gewolb, Vice, Schweitzer-Kenney, Taciak, & Bosma, 2001; Tsu-Hsin, Ching-Fan, Yu-Wei, & Ching-Lin, 2002). In the present study, the preterm infants moved and vocalized simultaneously, while full-term infants were able to vocalize lying peacefully, an observation which could be interpreted as a sign of immaturity in the preterm infants. When entering the primitive articulation, which is the start of volitional vocalization, the preterm infants produced less vocalization, while the full-term infants did the opposite (Figure 1). At the end of the primitive articulation, the preterm infants vocalized more than the full-term infants, and towards the end of the expansion, the preterm group were again vocalizing less than the full-term infants, who started to vocalize more frequently. In the present study, it seemed that the transitions were difficult for the preterm infants during every. Pridham, Steward, Thoyre, Brown, and Brown (2007) found the same phenomenon in preterm infants who usually had difficulties in transitioning from one food texture to another, that is, from one oral motor pattern to another. There were infants who vocalized minimally during some video-recording sessions. For example, in the preterm infants, the video-recording sessions with minimal vocalization occurred in half of the cases (7/14) after they had entered the primitive articulation, while 20/27 of the full-term infants' vocalization was minimal during the phonation. This difference may reflect the maturity level, discussed above. However, there were altogether 329 video-recording sessions, thus the number of silent sessions was not extensive. One of the main differences between the preterm and full-term infants in the present study was that the preterm infants failed to produce more skills (1 6) at the primitive articulation and the expansion s than did the full-term infants (1 3). These skills did not appear during the data collection at all. On the other hand, many of the infants failed to produce skills only during a certain, which seemed to happen to most of the infants in both groups (17/18 and 7/11). The data suggests that the preterm infants also failed to produce more skills than the full-term infants at the expansion, although

342 H. Törölä et al. the difference was not statistically significant. One of the skills that most (10/18) preterm infants failed to produce was exploring the extremities of articulation organs, which includes various oral motor functions. The oral motor patterns have been reported to be a problem for preterm infants (Arvedson, Clark, Lazarus, Schooling, & Frymark, 2010). In the present study, it is not possible to define which skills are precursors for later speech capacity and what later effects on infant speech development are associated with the inability to produce certain vocalization skills. Previous studies have not discussed skills that infants fail to produce. In this study that has been reported because those skills may have diagnostic importance. Not all of the vocalization types of the phonation were heard from the beginning (Table III), even though the preterm infants produced vocalizations from the beginning of the follow-up, at the gestational age of 37 weeks. However, all vocalization types at the phonation appeared before entering the primitive articulation. This may be due to the low alertness of the preterm infants. It has been reported that preterm infants, in general, present more passive attention behavior than full-term infants (Van de Weijer-Bergsma, Wijnros, & Jongmans, 2008). The preterm and full-term infants reached skills characteristic of the primitive articulation and expansion s at different ages. Differences between the groups in reaching individual skills were, however, not calculated owing to small sample sizes and many skills that especially the preterm infants failed to produce. For example, ingressive phonation was a late skill, which may reflect respiration problems (RDS, BPD) that almost every (16/18) preterm infant had suffered from. The same reason might explain the failure to produce diphthong-like phonations and blowing, since the duration of these sounds is long and demands good breathing. Half of the preterm infants did not produce diphthong-like phonations at all, whereas both vowel- and diphthong-like phonation occurred in all the full-term infants. There are 18 diphthongs in Finnish: [yi], [øi], [æi], [ui], [oi], [αi], [æy], [αu], [yø], [øy], [uo], [ou], [ie], [ei], [eu], [iu], [ey], [iy] (Suomi, Toivanen, & Ylitalo, 2008). In addition, the diphthong-like phonation [æu] occurred frequently in the present sample, which does not correspond to any Finnish diphthong. Interruption of phonation is the basis for consonant production and marginal babbling for syllable production (Oller, 1980). Infants explore the extremities of phonation by experimenting with loudness, pitch and intonation and all of these vocalizations are important steps toward speech-like phonation. The appearance of these preverbal milestones seemed to show a tendency to concentrate around the median age of 5 in both groups. In the present study, all vocalizations were analyzed to find the possible differences between preterm and full-term infants. In general, non-speech-like vocalizations are not considered precursors to speech (Oller, 2000). For example, creaking, groaning and fussing at the phonation are neither fixed signals nor vegetative sounds and they appeared in both preterm and full-term infants. This is why they could be paralleled with normal phonation and thus be considered precursors to speech. However, due to the small sample size, it was not possible to define which amount and quality of vocalization could be essential in later speech development. Few of the preterm and full-term infants performed later developmental skills before entering this, and the preterm infants performed these skills even more seldom than the full-term infants. On the other hand, the preterm infants showed signs of developing skills from both the primitive articulation and expansion s, while the full-term infants did so only from the expansion. Based on the present study, it seemed that the infants in both groups started to rapidly produce skills at a certain following the appearance of

Early vocalization of preterm infants with ELBW 343 the first skills, which seemed to be a universal phenomenon for both the preterm and the fullterm infants. In summary of the present study, it could be said that preterm infants produced more vocalizations at the phonation and the full-term infants especially produced more vocalizations during the primitive articulation and expansion s. In addition, the preterm infants failed to produce half as many various individual skills at each of vocalization than the full-term infants. The present study provides valuable information on the development of early vocalization of preterm infants with ELBW. The intensive follow-up provides information on how the early vocalization of the ELBW infants develops as well as the potential differences between preterm infants with ELBW and full-term infants. This study also provides a basis for following preterm children's pre-speech development and finding possible features, which could predict the risk of later delay in speech development that is frequently found, especially in preterm children with ELBW (Foster-Cohen, Edgin, Champion, & Woodward, 2007). The differences between the groups seem to be explained by the difference in maturity, oral motor and breathing problems. Acknowledgments We wish to thank Professor Emeritus Dr Antti Iivonen (the University of Helsinki) for his many valuable notes on preverbal development. Special thanks to physiotherapist Merja I. Luukinen for assisting with the data collection and speech-language pathologist Eija Meriläinen for double-checking the vocalizations. Declaration of Interest: This work was supported by Rinnekoti Research Centre, the Alma and K. A. Snellman Foundation, Oulu, Finland, the Graduate School of Culture and Interaction, and the Faculty of Humanities, University of Oulu. References Arvedson, J., Clark, H., Lazarus, C., Schooling, T., & Frymark, T. (2010). Evidence-based systematic review: Effect of oral motor interventions on feeding and swallowing in preterm infants. American Journal of Speech-Language Pathology, 19, 321 340. Beckwith, L., Sigman, M., Cohen, S. E., & Parmelee, A. H. (1977). Vocal output in preterm infants. Developmental Psychobiology, 10, 543 554. Berglund, E., Eriksson, M., & Westerlund, M. (2005). Communicative skills in relation to gender, birth order, childcare and socioeconomic status in 18-month-old children. Scandinavian Journal of Psychology, 46, 485 491. DeCasper, A. J., & Fifer, W. P. (1980). Of human bonding: Newborns prefer their mothers' voices. Science, 280, 1174 1176. Draganova, R., Eswaran, H., Murphy, P., Huotilainen, M., Lowery, C., & Preissl, H. (2005). Sound frequency change detection in fetuses and newborns, a magnetoencephalographic study. NeuroImage, 28, 354 361. Elbers, L. (1982). Operating principles in repetitive babbling: A cognitive continuity approach. Cognition, 12, 45 63. Fassnacht, C., & Woods, D. (2007). Transana (Version 2.20) [Computer Software]. Madison, WI: School of Education at the University of Wisconsin-Madison. Foster-Cohen, S., Edgin, J. O., Champion, P. R., & Woodward, L. J. (2007). Early delayed language development in very preterm infants: Evidence from the MacArthur-Bates CDI. Journal of Child Language, 34, 655 675. Gewolb, I. H., Vice, F. L., Schweitzer-Kenney, E. L., Taciak, V. L., & Bosma, J. F. (2001). Developmental patterns of rhythmic suck and swallow in preterm infants. Developmental Medicine & Chils Neurology, 43, 22 27. Grossmann, T., Oberecker, R., Koch, S. P., & Friederici, A. D. (2010). The developmental origins of voice processing in the human brain. Neuron, 65, 852 858. Iivonen, A. (1991). The fonological-phonetic development of the child: 2 the development of the utterance types during the preverbal period. Lapsen fonologis-foneettinen kehitys: 2 ilmaisutyyppien kehitys esikielellisellä kaudella. Suomen logopedis-foniatrinen aikakauslehti, 1, 15 25.

344 H. Törölä et al. Kent, R. D. (1999). Motor control: Neurophysiology and functional development. In A. J. Caruso & E. A. Strand (Eds.), Clinical management of motor speech disorders in children (pp. 29 71). New York, NY: Thieme. Kent, R. D., & Miolo, G. (1995). Phonetic abilities in the first year of life. In P. Fletcher & B. Mac Whinney (Eds.), The handbook of child language (pp. 303 334). Oxford: Blackwell. Koopmans-van Beinum, F., & van der Stelt, J. M. (1979). Early s in infant speech development. Proceedings of the Institute of Phonetic Sciences, 5, 30 43. Koopmans-van Beinum, F., & van der Stelt, J. M. (1986). Early s in the development of speech movements. In Lindblom B. & Zetterström R. (Eds.), Precursors of early speech: Proceedings of an international symposium held at the Wenner-Gren Center, Stockholm, September 19 22, 1984 (p. 37 50). Basingstoke: Macmillan. Lüchinger, A. B., Hadders-Algra, M., Van Kan, C. M., & de Vries, J. I. P. (2008). Fetal onset of general movements. Pediatric Research, 63, 191 195. Mampe, B., Friederici, A. D., Christoph, A., & Wermke, K. (2009). Newborns' cry melody is shaped by their native language. Current Biology, 19, 1994 1997. McCulloch, C. E., & Searle, S. R. (2001). Generalized, linear, and mixed models (p. 156 157). Canada: John Wiley & Sons, Inc. Nathani, S., Ertmer, D. J., & Stark, R. E. (2006). Assessing vocal development in infants and toddlers. Clinical Linguistics and Phonetics, 20, 351 369. Oller, D. K. (1980). The emergence of the sounds of speech in infancy. In G. H. Yeni-Komshian, J. F. Kavanagh, & C. A. Ferguson (Eds.), Child phonology, volume 1, production (pp. 93 112). New York, NY: Academic Press. Oller, D. K. (2000). The emergence of speech capacity. Mahwah, NJ: Lawrence Erlbaum Associates, Publishers. Oller, D. K., Eilers, R. E., Neal, A. R., & Schwartz, H., K. (1999). Precursors to speech in infancy: The prediction of speech and language disorders. Journal of Communication Disorders, 32, 223 245. Oller, D. K., Wieman, L. A., Doyle, W. J., & Ross, C. (1976). Infant babbling and speech. Journal of Child Language, 3, 1 11. Pike, K. L. (1943). Phonetics. A critical analysis of phonetic theory and a technic for the practical description of sounds. Ann Arbor, MI: The University of Michigan Press, 11, printing. Pridham, K., Steward, D., Thoyre, S., Brown, R., & Brown, L. (2007). Feeding skill performance in premature infants during the first year. Early Human Development, 83, 293 305. Roug, L., Landberg, I., & Lundberg, L.-J. (1989). Phonetic development in early infancy: A study of four Swedish children during the first eighteen of life. Journal of Child Language, 16, 19 40. Stark, R. (1979). Prespeech segmental feature description. In P. Fletcher & M. Garman (Eds.), Language acquisition. Studies in first language development (pp. 15 32). Cambridge: Cambridge University Press. Stark, R. E., Bernstein, L. E., & Demorest, M. E. (1993). Vocal communication in the first 18 of life. Journal of Speech and Hearing Research, 36, 548 558. Stark, R. E., Rose, S. N., & Benson, P. J. (1978). Classification of infant vocalization. British Journal of Disorders of Communication, 13, 41 47. Suomi, K., Toivanen, J., & Ylitalo, R. (2008). Finnish sound structure. Phonetics, phonology, phonotactics and prosody. Oulu: University of Oulu. Tsu-Hsin, H., Ching-Fan, S., Yu-Wei, H., & Ching-Lin, H. (2002). Psychometric characteristics of the Neonatal Oral-Motor Assessment Scale in healthy preterm infants. Developmental Medicine & Child Neurology, 49, 915 919. Van Beek, Y., Hopkins, B., & Binne Hoeksma, J. (1994). Development of communicative behaviors in preterm infants: The effects of birthweight status and gestational age. Infant Behavior and Development, 17, 107 117. Van de Weijer-Bergsma, E., Wijnroks, L., & Jongmans, M. J. (2008). Attention development in infants and preschool children born preterm: A review. Infant Behavior & Development, 31, 333 351.

Copyright of Clinical Linguistics & Phonetics is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.