I am a full-time academic researcher and consultant, currently working as a freelancer for several highly regarded universities and institutions in the UK, various European countries, the Middle East and South Asia (such as the Open University and the School of Oriental and African Studies, University of London). I recently completed my ESRC-funded doctorate from the Institute of Education (University of London). The topic for my doctorate was the use of singing with children in the classroom for facilitating enhanced learning, behaviour and well-being. During my doctorate, I worked for several research projects at the Institute. Prior to commencing my doctorate, I underwent intensive research training and completed a Master's degree in Social and Educational Research. I also hold a Master's degree in Arts and a Bachelor of Science degree in Psychology. The research projects that I have been involved in have been in the fields of: education, psychology, biology, research methods, statistics, sociology, culture, child development, music and other art forms. My recent research work has been published in international peer-reviewed journals (such as the British Journal of Music Education) and presented in international conferences (such as the International Conference on Special Education Needs and Inclusion).

Perceptual connections between pre-pubertal children’s speaking and singing behaviours

Introduction
Traditionally, it has been believed that children possess ‘a speaking voice’ used for generating speaking behaviour and ‘a singing voice’ used for generating singing behaviour (Rinta, 2005; Rinta and Welch, 2008). On the contrary, the voice-scientific view advocates the idea that ‘speaking’ and ‘singing voices’ are an integrated entity (Rinta and Welch, op.cit.; Sundberg, 2001). Nevertheless, only a limited amount of empirical research has been conducted in regard to investigating such claims.

The aim of the current study was to systematically investigate whether pre-pubertal children’s speaking and singing behaviours are connected perceptually as evidenced through children’s vocal functioning and voice quality. Firstly, theoretical connections between children’s speaking and singing behaviours are discussed. Secondly, perceptual voice assessment as a method to investigate connections between children’s speaking and singing behaviours is demonstrated. Thirdly, the methodological framework used in the study and the empirical findings are presented. Fourthly, the findings from the study are discussed in relation to existing literature.

The Child Voice
Children speaking and singing behaviours, vocal functioning and voice quality have remained an under-researched area (McAllister, 1997; Rinta and Welch, 2008, Sederholm, 1996; White, 2001). Not much is known about the ‘normality’ of the child voice nor its functioning in speaking and singing behaviours (McAllister, op.cit.; Sederholm, op.cit.). The field has primarily been undermined on the basis of the belief that children will grow out of their speech and voice distortions (Rinta, 2005). Nevertheless, such lack of knowledge complicates professional speech and voice therapy practice and the teaching of singing to children.

Furthermore, we have generally been consigned to the notion that children’s speaking and singing behaviours are two independent entities that rely on different voices as their main instruments (Kersner and Wright, 2002; Rubin et al., 2003; Welch, 2005). Particularly in Western countries, children are educated to view speaking and singing behaviours as distinct entities from one another (David, 1995; Welch, op.cit.). Such perception of speaking and singing behaviours has shaped professional speech and voice therapy practice so that speaking behaviour is primarily focussed on in therapeutic settings (Rinta, 2005). For instance, a greater deal of importance has been placed on children’s speaking behaviour in professional practice and in prior research due to the notion that speaking behaviour is of prime importance in treating children’s speech and voice distortions (Hunt and Slater, 2003; Mathieson and Greene, 2003; Rinta, op.cit.). Yet, it may be that children’s singing behaviour is causing the children’s vocal output to be distorted (Bunch, 1997; Rubin et al., op.cit.). Thus, speech and voice therapy practice could greatly benefit from studies investigating children’s singing behaviour.

The voice-scientific perspective
According to the voice-scientific perspective, our speaking and singing behaviours rely on the same voice as their main instrument (Rinta and Welch, 2008; Sundberg, 1996; Thurman and Welch, 2000). This voice is generated by specific physiological mechanisms (including the throat, the vocal folds and the larynx) that, consequently, underlie both speaking and singing behaviours (Rinta and Welch, op.cit.; Sundeberg, op.cit.; 2001). These particular physiological elements in the voice mechanism adapt according to the vocal behaviour in generation (Bunch, 1997; Reid, 2001; Rubin et al., 2003). For instance, the vocal folds are used differently between speaking and singing behaviours, mainly due to the fact that singing behaviour takes place in a higher pitch-range in comparison to speaking behaviour (Bunch, op.cit.).

It can be claimed, then, that the same physiological mechanisms and the same voice underlie speaking and singing behaviours, with the perceptual differences between the two vocal behaviours mainly being of functional nature (Rinta and Welch, 2008; Rubin et al., 2003). In other words, since speaking and singing behaviours are different in nature, they exploit the physiological mechanisms and the voice somewhat differently (Bunch, 1997). Their different natures, subsequently, facilitate the recorded perceptual differences between the two vocal behaviours (Sundberg, 2001). Therefore, the perceptual differences recorded between the two vocal behaviours manifest due to their differing natures and functions rather than the behaviours relying on different physiological mechanisms and different voices as their main instruments (Rinta and Welch, op.cit.).

The Evolutionary Perspective
The voice mechanism has been the main instrument for all vocalisation throughout centuries (Newman, 1993). Originally, all vocalisation (including speaking and singing behaviours) developed for the purpose of survival and communication (Newman, op.cit.; Thiessen, 1996). Consequently, there were no clear distinctions between different types of vocal behaviours, with all such behaviours being used for the same purpose and indicating that all vocal functioning shared a common root (Thiessen, op.cit.). It was only later on that speaking and singing behaviours began to be perceived as two distinct types of vocal behaviours (Sloboda, 2000; Welch, 2005).

As a result of the distinct perceptions of speaking and singing behaviours, our vocal functioning became limited in its expression capability (Bruyninck et al., 1994; Thiessen, 1996). By the age of six, children will have learned to distinguish speaking behaviour from singing behaviour in cultures that advocate the distinctions between the two vocal behaviours (such as Western culture) (Fourcine, 2005; Welch, 2005). Based on their cultural conditioning, children learn to relate to speaking and singing behaviours as independent entities (Laurence, 2001; Welch, op.cit.). For example, in African cultures, speaking and singing behaviours are viewed as a continuum, whilst in Western cultures, speaking and singing behaviours are perceived as distinct from one another (Brown et al., 2004).

However, despite the recorded cultural differences in regard to the concepts of speaking and singing behaviours and the concepts of ‘speaking’ and ‘singing voices’, the perceptual connections and differences between the two vocal behaviours have not been exposed to empirical research (Rinta, 2005). Therefore, the current study aimed to investigate potential perceptual connections between pre-pubertal children’s speaking and singing behaviours.

Perceptual voice assessment
In professional speech and voice therapy practice, most often therapists use a perceptual voice assessment protocol when assessing the child voice (Carding et al., 2000; Hirano, 1987; Yamaguchi et al., 2003). Such a protocol measures the child client’s general voice quality and vocal functioning primarily in speaking behaviour (Carding et al., op.cit.; Yamaguchi et al., op.cit.). More specifically, the standardized formally-established perceptual voice assessment protocols have focussed on child clients’ speaking behaviour and speech (Rinta, 2005; Yamaguchi et al., 2003). None of the assessment protocols have been designed for the purpose of assessing children’s singing behaviour (Carding et al, 2000; Rinta, op.cit.).

The majority of the perceptual voice assessment protocols have been formulated for assessing adult clients (Wilson, 1987). Nevertheless, the child voice differs significantly from the adult voice (Hunt and Slater, 2003; Mathieson and Greene, 2003). The physiologically immature child voice mechanism results in perceptually different vocal characteristics in the child voice in comparison to the adult voice (Welch, 1994; White, 1998). Therefore, children should be compared to one another rather than to adults in order to obtain a reliable assessment outcome.

Due to the demonstrated lack of formally-established protocols that could be used reliably in perceptual voice assessment with children in regard to their speaking and singing behaviours, a protocol that considered both vocal behaviours and that was specially designed for children was formulated for the purposes of the current study (see Appendices for the full protocol). The protocol was designed on the basis of the existing standardised formally-established perceptual voice assessment protocols (Carding et al., 2000; McAllister, 1997; Sederholm, 1996; Yamaguchi et al., 2003). When formulating the new protocol, elements from established singing competency assessment protocols were taken into account in order for the protocol to be an appropriate tool for assessing the products from both vocal behaviours (Mang, 2001).

The specially designed protocol consisted of separate sections for both speaking and singing behaviours. The section for each vocal behaviour included 13 identical vocal elements (i.e. hoarseness, breathiness, hyperfunction, hypofunction, gratings, roughness, voice breaks, unstable pitch/ voice quality, hard glottal attacks, vocal fry, audible inhalation, hypernasality and hyponasality) (see Appendices for the full protocol). In addition, sections for vocal pitch and register were included. Continuous lines were adopted to the rating-scale since such form of a scale was less likely to result in a loss of informative data (McAllister, 1997; Sederholm, 1996).

On the protocol, the left end of the line (i.e. rating 1) represented healthy and normal voice quality and the right end (i.e. rating 7) represented unhealthy and abnormal voice quality. The protocol was piloted with 30 children and three professional judges prior to conducting the main study and it was found to be a reliable and valid tool in assessing children’s vocal functioning and voice quality in their speaking and singing behaviours.

Aims of the study
The research question for the study was:

Are pre-pubertal children’s speaking and singing behaviours connected, as evidenced through perceptual connections between the functioning and the quality of the two vocal behaviours?

Methods

Participants
Purposive sampling was used as a method to recruit participants for the study. The requirement for the participants was to be pre-pubertal. The schools, in which the participant children were located, were chosen on the basis of prior professional contact.

76 -year-old children performed as the participants. 36 of the participants were female and 26 were male. Half of the children (N=38) were from a primary school located in west London and the other half (N=38) were from a primary school located in southern Finland. Both of the schools were mainstream primary schools.

The participant children were from four different classes. Two of the classes were from the British school and two were from the Finnish school. The UK classes were from Year 5 (i.e. children between the ages of 9 and 10). The first Finnish class was from Year 4 (i.e. children between the ages of 9 and 10). The second Finnish class was from Year 2 (i.e. children between the ages of 7 and 8). The last class was included in the study in order to investigate the proposed research question with a wider age-range (i.e. three years).

Ethical issues
Ethical guidelines of the British Educational Research Association (BERA) were followed throughout the study. Parental and teacher approval was requested prior to commencing any data collection. Initially, the Head Teachers of the schools and the classroom teachers were contacted. Once approval had been granted from the schools for conducting the study, a letter was sent to the parents of the potential participants in order to request approval for their child to participate in the study.

Voice tasks
Speaking and singing tasks were carried out with the participant children. The tasks were formulated on the basis of an extensive literature review, concerned with voice tasks that professional speech and voice therapists use when assessing a child client’s overall vocal functioning and voice quality (McAllister, 1997; Sederholm, 1996). The most common speaking tasks were adopted to the current study.

The three speaking tasks were: reading a pre-selected text passage, speaking spontaneously once provoked by three pictures and sustaining vowels (/a/, /e/, /i/).Additional singing tasks were selected on the basis that they had been used for gathering singing data in previous studies (White, 2001). The two singing tasks included in the study were: vocalising pitch-glides and singing a simple, familiar song.

Procedure
Each child was taken to a quiet room, one at a time. The child was instructed to sit by a table. The researcher adjusted a small microphone on the child’s shirt at a distance of 10 centimetres from the child’s mouth. The researcher explained the experimental procedure to the child.

The tasks were administered to the children in a pre-set order: a) reading a pre-selected text passage; b) speaking spontaneously once being provoked by three pictures; c) sustaining vowels (/a/e/i/); d) vocalising pitch-glides; and e) singing a simple pre-selected song. A high-quality mp3-player was used as a recorder. The procedure was approximately fifteen minutes long.

Data analyses
Three independent judges rated the vocal functioning and the voice quality of each participant child in the child’s speaking and singing behaviours with the use of the specially-designed perceptual voice assessment protocol. The researcher herself and two professional speech therapists performed as the judges. The judges were provided with a copy of the children’s voice recordings, the voice assessment protocol and the instructions as to how to perform the assessment task. Each judge performed the voice assessment task at their workplace with the use of a high-quality mp3-player and a pair of headphones.

The protocol consisted of continuous lines that were subsequently divided into seven separate categories when analysing the voice data. Thus, each parameter was rated on a seven-point continuous scale. Individual children were provided with separate scores for each of the 13 parameters in speech and in singing. Based on the ratings for the individual voice parameters, a meaned rating for overall voice quality was calculated for each child. Separate overall voice quality ratings were calculated for speaking behaviour and for singing behaviour. Based on such assessment, the researcher was subsequently able to compare the ratings generated for speech and those generated for singing.

Once the researcher had received the completed rating sheets from the three judges, the inter-judge reliability of the ratings was calculated. The inter-judge reliability was high (Pearson correlation, r=0.89, p<0.05). The finding indicates that the protocol was a reliable assessment tool with children’s and that the outcomes from the assessment were not dependent on the person carrying out the task.

Results
The mean score of the ratings provided for the 13 individual voice parameters was used as a representative for the general impression of the child’s vocal output for each participant. Prior to using the meaned ratings in comparing perceptual voice quality in speech to that in singing, other central tendency measures were calculated in order to investigate whether the meaned ratings were appropriate in fulfilling their proposed purpose. Mode, median and standard deviations were calculated for each voice parameter in speech and those in singing.

The finding was that the correlation between each measure in speech and the same measure in singing was statistically significant (mode: r=0.587; p<0.05; median: r=0.851; p<0.05; standard deviation: r=0.617; p<0.05). Since the findings were the same for all the central tendency measures, the mean ratings were regarded a suitable measure for representing the general impression of the children’s overall voice quality and vocal functioning.

The meaned ratings were compared between speaking and singing behaviours for the individual children. Parametric correlation was conducted in order to investigate the relationship between the ratings in speaking behaviour and those in singing behaviour. The correlation between the overall voice quality ratings in speech and those in singing was statistically significant (r=0.625, p<0.05). The finding indicates that the overall quality of the participant children’s voices was perceptually similar in the two vocal behaviours. The scatterplot below illustrates the finding (see Figure 1).

In addition, to comparing the overall voice quality ratings between the two vocal behaviours, parametric correlations were calculated between the 13 individual voice parameters in speech and those in singing for all the participants. Paired-sample t-tests were carried out for each parameter. The test was not significant for any of the parameters (p> 0.05, n.s.), indicating that the voice parameters were perceptually similar between the two vocal behaviours. The bar-chart below demonstrates the finding that the quality of the participant children’s voices was similar between the two vocal behaviours (see Figure 2).

Figure 2: Bar chart for overall voice quality ratings in speech and those in singing for the participant children

Descriptive statistics further suggested that the participant children’s general voice quality in speech was similar to their general voice quality in singing. The descriptive statistics demonstrated that there was minimal difference between voice quality ratings in speech and those in singing. The mean rating in speech for the whole group was 1.95 compared with 1.69 in singing. The standard deviation was slightly greater for speech than for singing (0.59 versus 0.82). The range of the ratings varied by 2.90 (1.00-2.90) and by 2.20 for singing (1.00-2.20). The descriptive statistics suggest that the participant children’s voice quality in their speaking behaviour was similar to their voice quality in their singing behaviour (see Table 1).

Participants	Speech			Singing
	Mean	Standard Deviation	Range	Mean	Standard Deviation	Range
Whole group (N=76)	1.95	0.823	2.90	1.69	0.589	2.20
Table 1: Descriptive statistics for overall voice quality ratings in speaking and in singing

Discussion
It has been claimed that our speaking and singing behaviours are connected through their main instrument - the voice (Rinta and Welch, 2008; Sundberg, 2001; Welch, 2005). The voice is, in turn, generated by specific physiological mechanisms (Bunch, 1997; Rubin et al., 2003; Welch, op.cit.). The current study supported such claims by advocating the idea that children’s vocal functioning and voice quality in their speaking behaviour is similar to those in their singing behaviour. The findings suggest that, perceptually, the same voice is used by pre-pubertal children for generating their speaking and singing behaviours.

Consequently, the current study indicated that it is appropriate to talk about children’s speaking and singing behaviours rather than their ‘speaking’ and ‘singing voices’. The functional differences between the two vocal behaviours are likely to be the primary reasons for any evident differences recorded between the two vocal behaviours (Bunch, 1997; Sundberg, 2001). In other words, the same voice is used for generating both vocal behaviours, with the different natures of the two vocal behaviours facilitating the recorded differences. For example, singing requires a child to follow melodic lines that most often take place in a vocal pitch-range that is considerably higher from that in speaking. Although the main focus of the study was not on investigating the exact physiological mechanisms underlying children’s vocal behaviours, the findings from the current study suggested that the same physiological elements are likely to underlie children’s speaking and singing behaviours since the same voice that derives from such physiological source was the main instrument used for both vocal behaviours. The study, therefore, suggests that children’s speaking and singing behaviours are connected through their main instruments that can be used in a variety of ways according to the desired vocal behaviour.

Moreover, it has been claimed that all vocal behaviour share a common root (Newman, 1993; Thiessen, 1996), with children’s cultural surroundings shaping their vocal behaviours to resemble either speaking or singing behaviour (Rinta and Welch, 2008; Welch, 2005). For instance, in Western cultures, speaking and singing behaviours are generally regarded as two distinct behaviours, whilst in African cultures, the two vocal behaviours are regarded as inter-connected (Welch, op.cit.). Nevertheless, the current study suggested that the vocal functioning and voice quality of children in Western culture were similar between the two vocal behaviours. The findings suggest that, despite the Western differentiation between speaking and singing behaviours or ‘speaking’ and ‘singing voices’, the same voice was used by children in generating both vocal behaviours from the perceptual perspective. It seems, therefore, that traditional concepts that operate in western cultures may educate children in perceiving their speaking and singing behaviours, as well as their ‘speaking’ and ‘singing voices’, as two separate entities when such entities are, in fact, inter-connected.

It should be noted that the participant children in the current study were pre-pubertal. It may be that the vocal functioning and voice quality of children in their puberty exhibit distinctively different vocal output in comparison to their younger peers. Physiological changes taking place when the children enter puberty are likely to be reflected in their vocal products (Hunt and Slater, 2003; Mathieson and Greene, 2003). It may be that greater differences are recorded between vocal output from speaking behaviour and that from singing behaviour with older children. Furthermore, older children will have been exposed to cultural conditioning for a longer period of time than their younger peers, potentially further facilitating perceptual differences between the two vocal behaviours when, in fact, the behaviours are integrated (Rinta and Welch, 2008; Welch, 2005). In addition, external matter (such as air pollutants) or internal matter (such as a child’s psychological state) may significantly contribute to a child’s perception of his/ her speaking and singing behaviours as different from one another (Hunt and Slater, op.cit.; Mathieson and Greene, op.cit.; Rinta and Welch, 2008). For instance, when a child holds negative associations with singing activities, such negative perceptions may deteriorate the child’s vocal functioning and voice quality in his/ her singing behaviour but not in his/ her speaking behaviour.

Since the same voice is used by children for generating their speaking and singing behaviours, the ‘normal’ and the ‘abnormal’ vocal characteristics in their speaking and singing behaviours are likely to be similar between the two vocal behaviours. The current study provided evidence for such characteristics, with slight differences being recorded between specific voice parameters (such as unstable pitch and audible inhalation being stronger in singing than in speaking behaviour) from one vocal behaviour to another. A greater amount of voice data on the ‘normal’ and ‘abnormal’ characteristics of children’s vocal functioning and voice quality could be gathered with the use of the new specially-designed perceptual voice assessment protocol.

The perceptual connections between the participant children’s speaking and singing behaviours were evident in the current study. The specially-designed perceptual voice assessment protocol used in the study was found to be a reliable and valid instrument in gathering data on pre-pubertal children’s vocal functioning and voice quality in their speaking and singing behaviours. The protocol enabled the researcher to formulate a comprehensive vocal profile for the children in reference to both vocal behaviours. The protocol was found to be a reliable instrument to be exploited in perceptual assessment with pre-pubertal children, in particular due to the fact that it took both speaking and singing behaviours into consideration and was specifically designed for assessing children’s vocal functioning and voice quality.

Despite the fact that the protocol was found to be a reliable instrument in perceptual voice assessment with children, the protocol could be exposed to further investigation with a larger number of children and judges from different cultural origins. Acoustic analyses could also potentially be used in such a study in order to compare voice data through the two different assessment approaches and in order to investigate whether the outcomes from the two different assessment approaches differed from one another. In such a study a grater number of individual voice parameters could also be looked at and specific relationships between individual voice parameters in both vocal behaviours could be explored.

A medically-oriented study could be conducted in order to investigate the underlying physiological differences behind voice production in children’s speaking and singing behaviours. Such a study could further highlight the fact that the same physiological mechanisms underlie children’s speaking and singing behaviours. In addition, a further study focussing on investigating specific voice quality characteristics in children’s speaking and singing behaviours across cultures could highlight more distinct cultural differences between children of a variety of cultural origins. Such a study could also highlight the fact whether there are cultural differences as to how segregated or integrated children’s speaking and singing behaviours are.

Conclusion
The study suggests that pre-pubertal children’s speaking and singing behaviours are connected through their vocal functioning and voice quality. Perceptually, pre-pubertal children appear to possess one voice that is used as the main instrument for generating their speaking and singing behaviours. Therefore, it seems to be appropriate to talk about children’s speaking and singing behaviours rather than their ‘speaking’ and ‘singing voices’. Therefore, teachers and therapists should be educated on considering children’s speaking and singing behaviours as integrated rather than separate entities.

References
Brown, S., Martinez, M.J., Hodges, D.A., Fox, P.T. and Parsons, L.M. (2004). The song system of the human brain. Cognitive Brain Research, March, 40-56

Bruyninck, M., Harmegnies, B., Listerri, J. and Poch-Olive, D. (1994). Language-induced voice quality variability in bilinguals. Journal of Phonetics, 22, 19-37

Bunch, M. (1997). Dynamics of the singing voice. New York: Springer

Carding, P., Carlson, E., Epstein, R., Mathieson, L. and Shewell, C. (2000). Formal perceptual evaluation of voice quality in the United Kingdom. Logopedics Phoniatrics Vocology, 25, 133-138

David, M. (1995).The new voice pedagogy. London: Scarecrow Press

Fourcine, A.. (2005). Hearing and singing. In: J. L.Chapman (ed.). (227-239).

Singing and teaching. A holistic approach to classical voice. London: Plural Publishing House Inc.

Hirano, M. (1989). Objective evaluation of the human voice: clinical aspects. Folia Phoniatrica et Logopedica, 41, 89-144

Hunt, J. and Slater, A. (2003). Working with children’s voice disorders. Oxon, UK: Speechmark Publishing Ltd.

Kersner, M. and Wright, J. (2002). How to manage communication problems in young children. London: David Fulton

Laurence, F. (2001).Children’s singing. (pp. 221-225). In: J. Potter (ed.) Cambridge Companion to Singing. Cambridge: Cambridge University Press

Mang, E. (2001). A cross-language comparison of pre-school children’s fundamental frequency in speech and song production. Research Studies in Music Education, 16, 4-14 Mathieson, L. and Greene, I. (2003). The voice & its disorders. London: Whurr Publishers

McAllister, A. (1997). Acoustic, perceptual and physiological studies of ten-year-old children’s voices. Dissertation from the Department of Logaoedics and Phoniatrics, Karoliska Institute Huddinge University Hospital and from the Department of Speech, Music and Hearing, Royal Institute of Technology (KTH), Stockholm, Sweden

McMullen, E. and Saffran, J.R. (2004). Music and language: a developmental comparison. Music Perception; 21 (3), 289-311

Newman, P. (1993). The singing cure - an introduction to voice movement therapy. Massachusets: Shambala Publications, Inc.

Papousek, H. (1996). Musicality in infancy research: biological and cultural origins of early musicality. (pp. 37-55.) In: I. Deliege and J. Sloboda (eds.). Musical beginnings. Oxford: Oxford University Press

Reid L. (2001). Vocal mechanics. (pp.1-23).In: a. Bybee and J. Ford (eds.). The modern singing master - essays in the honour of Cornelius L. Reid. London: Scarecrow Press

Rinta, T. (2005). Inclusion of singing in perceptual voice assessment: practice and perceptions of speech and voice therapy professionals. Unpublished MRes-dissertation, Institute of Education, University of London

Rinta, T. and Welch, G.F. (2008). Should singing be included in speech and voice therapy settings with pre-pubertal children suffering from speech and voice disorders?. Journal of Voice, 17(2), 100-112

Rubin, J., Sataloff, R. and Korovin, G. (2003). Diagnosis and treatment of voice disorders - second edition. New York: Thomson Delmar Learning

Ruzza, B., Rocca F., Boero, D.L. and Lenti, C. (2002). Investigating the musical qualities of early infant sounds. (pp. 527-529.)In G. Avanzini, C. Faienza , D. Minciacchi, L. Lopez and M. Majno (eds). The neurosciences and music. New York: New York Academy of Sciences

Sederholm, E. (1996). Hoarseness in ten-year-old children, perceptual characteristics, prevalence and etiology. Thesis from the Department of Logaoedics and Phoniatrics Karoliska Institute, Huddinge University Hospital and the Department of Speech, Music and Hearing Royal Institute of Technology (KTH), Stockholm, Sweden

Sloboda, J. (2000). The musical mind - the cognitive psychology of music. Oxford: Oxford University Press

Sundberg, J. (1996). The human voice. (pp. 1096-1104.). In: R. Greger and U. Windhorst (eds.) Comprehensive human physiology. Berlin: Springer

Sundberg, J. (2001). How vocal sounds are made. (pp. 204-247.). In: J. Potter (ed.). Cambridge companion to singing. Cambridge: Cambridge University Press

Stacy, R.., Brittain, K. and Kerr, S. (2002). Singing for health: an exploration of the issues. Health Education, 102(4): 156-162

Thiessen, D.D. (1996). Bitter sweet destiny – the stormy evolution of human behaviour. New Jersey, USA: Transaction Publishers

Thurman, L. and Welch, G.F. (2000) (eds.). Bodymind and voice - foundations of voice education. Iowa: National Centre for Voice and Speech

Welch, G.F. (1998). Early childhood musical development. Research Studies in Music Education, 11

Welch, G.F. (2005). Singing as communication. (pp. 239-258.). In: D. Miell , R. MacDonald and D.J. Hargreaves (eds.). Musical communication.: Oxford: Oxford University Press

Wilson, D.K. (1987). Voice Problems of Children – Third Edition. Wlliams & Wilkins: London

White, P. (1998). Formant frequency analysis of children's spoken and sung vowels Using sweeping fundamental frequency production. Logopedics, Phoniatrics, and Vocology, 1(2), 43-52

White, P. (2001). The child voice. Stockholm: KTH Voice Research Centre

Yamaguchi, H., Shrivastav, R. and Andrews, M.L. (2003). A comparison of voice quality ratings made by Japanese and American listeners using the GRBAS scale. Folia Phoniatrica et Logopedica, 55, 147-157