FormalPara Core Messages

  • 48 boys and 47 girls were randomly chosen and analyzed from 3rd to 12th school classes in a stratified study at an elementary school and high school. A basic test of capability of reproducing tones and rhythms had been made routinely at the entrance to the school.

  • High-speed video examples were presented with 4.000 frames per second and 256 × 256 pixels in pubertal, prepubertal, and postpubertal pupils.

  • Voice Range Profiles were made with the standardization proposal of the Union of European Phoniatricians.

  • Fundamental frequency (F0) and semitone range during reading of a standard text with a conversational voice were based on 2.000 cycles of electroglottographic measurements.

  • Hormone measures at the Danish Statens Serum Institut included serum testosterone, dehydroepiandrosterone, androstenedione, estradiol, estrone, and estrone sulfate plus sex hormone-binding globulin. The five pubertal Tanner stages were measured by our pediatrician.

  • Statistical programs at the Danish Statens Serum Institut, based on logarithms of geometrical results when relevant, were used for the one-way multivariate analysis and predictive calculations.

3.1 Test Persons

48 boys and 47 girls took part in the stratified study. 4–5 pupils came randomly from 3 to 12 school classes, ages 8 to 19 years. All of them had passed a test at the entrance to the school, which included reproducing a rhythm by clapping, repeating tones by singing, and singing a given song (Figs. 3.1 and 3.2) as it is done by others [1]. The tests were also used for determining the type of voice (the voice category) because hormonal changes were thought eventually to be related. The test was necessary to define a standard of the study for comparison with other studies but was later considered without specific influence on the general results. Video analysis showed no neoplasms or other anatomic abnormalities of the larynx for all of them. The Tanner stage of puberty was measured by our pediatrician for the age of each pupil [2, 3].

Fig. 3.1
A musical notation chart lists 4 rows segmented into 24 parts titled 2, school class. The notation chart also lists two additional rows titled 7, school class.

Reproduction of tones

Full size image
Fig. 3.2
A musical notation chart lists segments labeled 4 slash 4, 6 slash 4, 6 slash 8, and 2 slash 4 under 2, school class, segment labeled 6 slash 8 under only 4 to 7 class, and segments labeled 5 slash 4 and 7 slash 4 under only 7 class.

Rhythm test

Full size image

3.2 Method of Investigation

3.2.1 High-Speed Videos

High-speed video equipment is continuously improved, and the frames per second and pixels increase. The equipment used for the HSV examples was developed by Richard Wolf GmbH, Knittlingen, Germany (Endocam 5562), which features a high-speed camera mounted on a rigid scope. The equipment allowed for recordings of 4.000 frames per second (fps), with 256 × 256 pixels. An intonation of /a/ for 2 s, with the rigid scope of 90 degrees was mostly used, and videos were stored in the setup for software analysis. The software included analysis hereof as kymograms [4]. The kymograms varied in a large scale with the position for the analysis on the vocal folds, and calculations were without usable information. An analysis was therefore abandoned. Examples of kymograms in the middle of the membranous vocal folds are given together with presentation of at least one whole cycle of vocal fold movements for the test persons.

The vocal folds were described for surface changes: shiny, matte, or less shiny. Regularity and thickness were looked for as well as changes in the edges of the glottis, glottal gaps, and mucus in the various stages of childhood. Firstly, the pubertal period was of interest as defined in pediatrics. Since the normal variations in the puberty period were interesting, high-speed video examples were stored of the beginning, the middle, and the end of the pubertal periods in girls and boys. Secondly, there were differences in the vocal fold changes during the postpubertal period where the pubertal findings continued. Thirdly, in the prepubertal stages, as could be expected due to the adrenarche, some changes in the vocal folds were already found [3].

3.2.2 Voice Range Profile Measurement

Measurements of the Voice Range Profile were performed in accordance with the standardization proposal from the Union of European Phoniatricians, as also used by Dienerowitz et al. [5,6,7]. The method used is described and discussed in the introduction.

There is no standard for the measurement setup, an overview of used setups is given by Rychel et al. [8], and the various setups found in the literature in their study are presented in Fig. 2.5. The results of measurement seem to be stable if the distance from a calibrated sound pressure level meter is 30 cm. The main discussion is whether dB(A) or dB(C) is preferable. Traditionally, dB(A) has been used, which is also the case in this study. As referred to in Sect. 2.2, a study has been made of the measurements showing a difference for the lower tones measured with dB(A) and dB(C) [9]. The difference is found stable in several measurements. It does not influence the comparisons with the other measurements (e.g., hormones and pediatric stage measurements).

The Voice Range Profile areas were recorded in tones × dB(A) on a diatonic scale, calculated for statistical evaluation. The total tonal range was given traditionally in semitones on a chromatic scale, to maintain a logarithmic scale on both the abscissa and the ordinate and to avoid misunderstandings and errors in the statistics.

A normal school room was used for the investigation with a background noise lower than 40 dB(A)–50 dB(A). This gives a realistic working situation. The author of this book, a medical ear-nose-throat specialist and concert singer, took care of the setup including a 30 cm distance to the calibrated Brüel & Kjaer frequency and sound pressure level phonetograph 8301. The first attempt was used after instruction. But no training or warming up was made. The test person intonated from the lowest to highest intensity on the given tones starting with the lowest measurable tones. All the data were stored in the computerized Voice Range Profile apparatus, our phonetograph 8301, as adjusted with Brüel & Kjaer frequency and intensity [9,10,11]. The lowest and highest physiological tones where there is mostly only one intensity were measured. We have given the results of the singing tone range as evaluated by the test persons in some figures.

3.2.3 Measurement of the Fundamental Frequency

The fundamental frequency during reading of a standard text of the speaking voice (F0) with conversational voices was registered by electroglottography and computed over 2.000 cycles. The standard text was “the Northwind and the sun” translated phonetically into Danish. For the youngsters, the reading of the standard text was asked to be performed with a conversational voice (not like actors). From the mean value of the frequency in Hz, the fundamental frequency was worked out. The tonal range during continuous speech was averaged to give the tonal range of the speaking voice with the tonal chromatic scale in semitones. As specified in Sect. 2.3.2, the measurement took place in a standard school classroom. The background noise was measured under 40 dB(A), as the given text was being read out. The speech studio electrolaryngograph from Laryngograph Ltd., UK, was used with reference to an earlier firm model 830 electroglottograph (FJ electronics) [12].

3.2.4 Puberty Stages and Hormonal Status Analysis

Body size and weight, testicle volume (for boys), stage of pubic hair development, and (for girls) stage of breast development were determined for each of the test persons by the pediatrician in our team [13, 14].

The hormonal analysis was based on those parameters which to our knowledge change during puberty, with advice from Statens Serum Institut. Children from the age of 8 were included in the investigation, as the adrenarche (increased prepubertal function of the adrenal glands at this age) can possibly provide information that will help in understanding puberty also of the vocal folds. The following values were analyzed with standard procedures of the Danish Statens Serum Institut: serum testosterone (free and total, there is a close relationship between the two values), dehydroepiandrosterone-sulfate (DHEAS), androstenedione and the transport globulin for testosterone, and sex hormone-binding globulin (SHBG). For the girls, the program of investigation also included the following parameters: estradiol (E2) produced mainly in the ovaries, estrone (E1) produced also in the adrenal glands and fat tissue, and estrone sulfate (E1SO4).

The buildup and working period of androgens and estrogens are complex, and the same is true of the possible interactions between the individual hormones (Fig. 3.3). All the androgen-regulating hormones in the hypothalamus in the investigation could have been included in this study. It is to be hoped that our work can provide the starting point for detailed hormonal brain research in the future.

Fig. 3.3
A flowchart is as follows. Cholesterol, pregnenolone, D H E A, progesterone, androstenedione, and estrone E 1. Androstenedione interconnects with testosterone and estrone E 1 interconnects with estradiol E 2.

Downstream conversion of cholesterol into androgens and estrogens

Full size image

3.2.5 Statistical Analysis

Measurement results do not have any evidence-based scientific value before they have been subjected to statistical analysis to determine the significance of their relations. All statistic calculations were made at the Danish Statens Serum Institut.

There was the question of whether linear or logarithmic relationships should be used to find correlations and predictions. The logarithmic criteria that were used, based on geometric cross sections, are considerably stricter than the linear ones.

A one-way multivariate analysis was performed, using the fundamental frequency of the speaking voice (F0) as a classifier to determine whether there were differences between children and groups. F0 was used to find predicting values. For all variables, we determined the correlation coefficients to be able to calculate the relationships between them and their dependency on age by using the partial correlation coefficients. Further group analyses were also made for pre- and postpubertal voice categories.