Presented at the Annual Convention of the American Speech-Language-Hearing Association. Atlanta, GA, November 15–17.

It is now widely accepted that hormones affect vocal function in women. These vocal changes have been reported after the menopause known as menopausal vocal syndrome, but to also during the menstrual cycle (Abitbol et al, 1998). The reported vocal changes have been attributed to hormonal induced physiological and anatomical changes in the vocal folds such as alterations in the vocal fold mass and the glottal membrane resulting in lowering of voice pitch and sometimes vocal roughness. Whiteside et al (2006) has studied the changes in the acoustic characteristic of speech during the menstrual cycle in women aged 20 to 25 years old and compared to age-matched men. Specifically, they collected their data during days 2 to 5 of the menstrual cycle, when the hormones are low, and during days 18 to 25, when the hormones are high. They reported that the menstrual cycle was related to the voice onset time (VOT), which was reduced for voiced consonants but increased for voiceless ones.
Another period in a woman's life that the hormonal balance is changing is during pregnancy. During this period, some hormones are increasing dramatically over the forty-week period and some new ones appear. Specifically, the Human Chorionic Gonadotrophin (HCG) acts mainly in the first months, while Progesterone and Estrogens reach their peak in the last months of the pregnancy. Furthermore, changes in the woman's body such as the increase of weight in the abdominal region result in reduced lung function and subsequently in vocal function. Abdul-Latif Hmadan (2007) in his study of the vocal function of women before and after giving birth described that pregnant women in the last months of pregnancy exhibit symptoms of vocal fatigue and a reduction of maximum phonation time. After giving birth there is an increase in the fundamental frequency and the maximum phonation time.
In this study we aim to examine the vocal changes, observed in the first months of pregnancy as a result of the hormonal changes, and compare them to age-matched non-pregnant women. We have chosen this period as there are not observed major changes in a woman's body size yet which might affect vocal function as in the later months of pregnancy.

Method
Subjects: Sixty two women were recruited from two different cities of Greece and were allocated into two groups. The first group included 31 pregnant women (mean age=29.2 years, range=21–35 years) in the first 13 weeks of pregnancy (mean=9.3 weeks, range=6–13 weeks). The control group consists of 31 non-pregnant women (mean age=28.6 years, range=21–35 years). Subjects with a history of gynecological or hormonal related health problems or other health problems which might affect vocal performance were excluded. All patients gave informed consent to participate in the study.
Assessment: The assessment protocol includes acoustic measures and rate of speech. The tasks used were: sustained phonation, reading a list of words, reading texts, and a one-minute monologue describing the cookie theft picture of BDAE. The word list used for reading consisted of 40 words. Ten words were chosen to contain each of the five Greek vowels /a, i, o, u, e/ between plosive sounds in open and closed initial and final syllables, stressed and unstressed. The remaining 30 words began with the six plosive consonant sounds /p, b, t, d, k, g/ followed by the five vowels, to measure voice onset time.
Data analysis: For the analysis of the recorded data we used the Praat program. For the statistical analysis the t test was used.

Results
The analyses showed differences between pregnant and non-pregnant women in vocal function but not in speech rate. Specifically, pregnant women had significantly higher fundamental frequency during sustained vowel production (t=2.569, p=.013), but significantly shorter duration in the sustained production of vowel /a/ (t=−2.702, p=0.010), voiceless consonant /s/ (t=−2.340, p=.025), and voiced consonant /z/ (t=−2.508, p=.016). Significantly reduced pitch was also observed in some vowels in stressed syllables during word reading (/i/: t=−2.362, p=.021; /o/: t=−2.226, p=.30). No significant differences between the two groups were observed in the pitch of the other vowels during word reading. Pregnant women produced voiced consonants /b/ (t=−2.169, p=.034) and /d/ (t=−2.637, p=.011) with a significantly shorter VOT. Regarding speech parameters, the two groups did not differ significantly in their mean pitch and speech rate (syllables per minute) during speaking (monologue) or passage reading.

Conclusion
Our results suggest that vocal changes but not speech changes are present in the first months of pregnancy. These were demonstrated as increased pitch during sustained vowel phonation, which might be related to increased tension in the vocal cords as a compensatory mechanism to reduced maximum phonation time. However, as the changes in the body size that might cause reduced lung capacity are not evident at this stage of pregnancy, our findings may be attributed to the hormonal changes taking place at the early stages of pregnancy. More specifically, the observed differences may be related to increased levels of HCG hormone, which reaches its peak at the early stages of pregnancy, rather than to progesterone or estrogens. Further research is needed to confirm this, studying parameters of voice and speech in relation to respiratory function and hormonal levels.