Women's voice pitch lowers after pregnancy

https://doi.org/10.1016/j.evolhumbehav.2018.04.002Get rights and content

Abstract

Women's voice pitch (the perceptual correlate of fundamental frequency, F0) varies across the menstrual cycle and lowers after menopause, and may represent a putative signal of women's fertility and reproductive age. Yet, despite dramatic changes in women's sex hormone levels and bodies during and after pregnancy, previous between-subject and case studies have not found systematic changes in F0 due to pregnancy. Here, we tracked within-individual variation in 20 mothers' voices during their first pregnancy, as well as up to 5 years before conception and 5 years postpartum. Voice recordings from 20 age-matched nulliparous women were measured as a control. Linear Mixed Models indicated that F0 mean, range and variation changed significantly following pregnancy in mothers, controlling for age at time of recording, whereas we did not observe any F0 changes across corresponding timeframes in our sample of nulliparous controls. Mothers' voices became significantly lower-pitched and more monotonous during the first year postpartum compared to during pregnancy or before. These F0 parameters did not decrease within-individuals over a 5-year period prior to conception above and beyond the effects of ageing. Although voice pitch decreased following pregnancy, mothers' F0 parameters reverted after the first year postpartum, approaching pre-pregnancy levels. Our results demonstrate that pregnancy has a transient and perceptually salient masculinizing effect on women's voices.

Introduction

Nonverbal properties of the human voice, particularly fundamental frequency (F0, perceived as voice pitch), correlate with a range of physical traits and behaviours, and predict individual differences in reproductive and social success (Pisanski and Bryant, 2018; Puts et al., 2016). Fundamental frequency is more sexually dimorphic in humans than in any other great ape (Puts et al., 2016). This sex difference emerges following a pubertal surge in the androgen levels of adolescent boys that masculinizes the larynx, lengthening and thickening the male vocal folds and resulting in a voice pitch that is around 75% lower in adult men than women (Titze, 1989). Fundamental frequency also varies considerably within the sexes. Numerous studies, focusing mostly on men, have linked individual differences in men's F0 to testosterone levels, facial masculinity, social dominance, and number of sex partners (Pisanski and Bryant, 2018 for review). There is therefore strong consensus that low F0 has been sexually selected in men to communicate threat potential and mate quality (Puts et al., 2016), and constitutes a secondary sex characteristic (Feinberg, 2008).

Women's voices, while comparatively less studied, also appear to communicate reproductive potential and fertility, and predict men's mate preferences. Men show preferences for indices of femininity and youth in women's traits, including high mean F0 in women's voices, presumably because these traits suggest a woman is of reproductive age and potentially fecund (Pisanski and Feinberg, 2018; Puts, Jones, and DeBruine, 2012 for reviews). Vocal attractiveness ratings are highest for women in their 20's and comparatively low for pre-pubertal and post-menopausal women (Röder, Fink, and Jones, 2013; Wheatley et al., 2014). Indeed, cross-sectional and longitudinal studies show that women's F0 either decreases gradually or stays relatively constant (barring cyclic fluctuations, see below) throughout their reproductive years, but consistently report a 10–35 Hz decrease in women's F0 following menopause (Abitbol, Abitbol, and Abitbol, 1999; Amir and Biron-Shental, 2004). This drop in F0 may be caused by menopausal decreases in estrogens and increases in androgens resulting in swelling (edema) that alters the thickness and mass of the vocal folds (Derman, 1995; Hirano, Kurita, and Nakashima, 1983). Hormone therapies provide further evidence that testosterone permanently decreases women's F0 (Baker, 1999), whereas estrogen injections partially counter postmenopausal decreases in F0 (Lindholm, Vilkman, Raudaskoski, Suvanto-Luukkonen, and Kauppila, 1997). Taken together, there is good evidence that a woman's voice pitch can indicate whether she is of reproductive age.

Studies have also shown intra-individual fluctuations in women's F0 across the menstrual cycle, suggesting that F0 may indicate cycle-to-cycle changes in fertility. Acoustic analyses reveal increases in women's F0 just prior to ovulation (Fischer et al., 2011) or at peak fertility (Bryant and Haselton, 2009), and playback studies show covariation between menstrual fluctuations in estrogen and progesterone levels and listeners' assessments of women's voice quality (Çelik et al., 2013). Although hormonal contraceptive usage, which stabilizes the cyclical variation in women's hormone levels, does not appear to predict differences in women's mean F0, lower levels of vocal perturbation among women taking monophasic birth control pills have been attributed to a stable hormonal balance leading to increased regularity in vocal fold vibration (Amir, Biron-Shental, and Shabtai, 2006).

Indeed, the vocal folds appear particularly sensitive to sex hormone levels which are known to fluctuate throughout the life cycle. Sex hormones may act on neuromotor control of the larynx (Higgins and Saxman, 1989) or act directly on the vocal fold mucosa that contain specific receptors for androgens, estrogens and progesterone (Newman, Butler, Hammond, and Gray, 2000).

Section snippets

Does pregnancy affect women's voice pitch?

Women's bodies change dramatically during and immediately following pregnancy. Estrogen, progesterone, testosterone and cortisol levels increase during pregnancy, followed by dramatic decreases following parturition (O'Leary, Boyne, Flett, Beilby, and James, 1991; Tan and Tan, 2013). Changes in anatomy and physiology during pregnancy include Reinke's edema (vocal fold swelling), decreased lung capacity, and altered respiration patterns (Speroff and Fritz, 2005; Tan and Tan, 2013). Many women

Present study

Here, we acoustically analysed archival recordings of 20 women experiencing their first pregnancy (Table S1a). We measured changes in F0 parameters before pregnancy (5 years), during pregnancy, and after pregnancy (5 years), and compared this to voice measures taken from 20 age-matched nulliparous controls (Table S1b) during corresponding timeframes. This allowed us to directly test whether mothers' F0 values decreased during or after pregnancy relative to pre-conception, and by how much, and

Between-sample comparisons of F0

Linear Mixed Models (LMMs) fit by maximum-likelihood estimation, including participant identity as a subject variable with random intercept, sample (mothers, nulliparous controls) as a fixed factor, and woman's age at the time of voice recording as a random covariate, showed no significant differences in any F0 parameters between mothers and controls collapsing across phases (all F < 1.3, all P > 0.271). This result confirmed that the two samples of age-matched women could effectually be

Discussion

The results of this longitudinal study show that women's average voice pitch, pitch range and pitch variability decrease after giving birth to their first child. These voice changes endure for at least one year postpartum, later returning to near pre-pregnancy levels controlling for ageing. Previous between-subject and case studies have focused on voice changes during pregnancy (Cassiraga et al., 2012; Hamdan et al., 2009; Hancock and Gross, 2015; Lã and Sundberg, 2012; Saltürk et al., 2016;

Subjects

We analysed voice recordings collected from 40 adult women, including 20 women who had given birth to their first child, and 20 nulliparous controls who had never given birth. We first searched online for female public figures who had given birth to their first child no less than four years ago. This search produced a list of twenty-nine women. We then excluded women (n = 9) for whom we could not find 10 years' worth of frequent (i.e., at least bi-monthly), standardized, high-quality voice

Ethics

The study was reviewed and approved by the Sciences and Technology Cross-Schools Research Ethics Committee (C-REC) of the University of Sussex (ER-REBY-2,8). The audio samples used for acoustic analysis were extracted from publicly or institutionally available video files.

Open practices

The dataset supporting this article is openly available on the Sussex Research Online (SRO) repository (http://sro.sussex.ac.uk/id/eprint/74811).

Authors' contributions

K.P. and D.R. designed the investigation. K.P. and K.B. collected the data and performed acoustic analysis. K.P. and D.R. conducted statistical analysis. K.P. drafted the manuscript and created all figures. The manuscript was reviewed, edited and approved by all authors, who agree to be accountable for the work.

Competing interests

The authors report no competing interests.

Funding

This project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No. H2020-MSCA-IF-2014-655859 to K.P.

Acknowledgements

We thank Francis Steen and Tim Groeling for curating and providing access to the UCLA Digital Civic Learning Search Engine, and Richard E. Lucas for statistical advice.

References (52)

  • M.B. Higgins et al.

    Variations in vocal frequency perturbation across the menstrual cycle

    Journal of Voice

    (1989)
  • F.M.B. et al.

    Pregnancy and the singing voice: Reports from a case study

    Journal of Voice

    (2012)
  • P. Lindholm et al.

    The effect of postmenopause and postmenopausal HRT on measured voice values and vocal symptoms

    Maturitas

    (1997)
  • S.-R. Newman et al.

    Preliminary report on hormone receptors in the human vocal fold

    Journal of Voice

    (2000)
  • K. Pisanski et al.

    Voice modulation: A window into the origins of human vocal control?

    Trends in Cognitive Sciences

    (2016)
  • J. Raine et al.

    Tennis grunts communicate acoustic cues to sex and contest outcome

    Animal Behaviour

    (2017)
  • Z. Saltürk et al.

    Objective and subjective aspects of voice in pregnancy

    Journal of Voice

    (2016)
  • E.K. Tan et al.

    Alterations in physiology and anatomy during pregnancy

    Best Practice & Research. Clinical Obstetrics & Gynaecology

    (2013)
  • J.R. Wheatley et al.

    Women's faces and voices are cues to reproductive potential in industrial and forager societies

    Evolution and Human Behavior

    (2014)
  • A. Abramson et al.

    Estrogen receptors in the human larynx: Clinical study of the singing voice

  • E.L. Alford et al.

    Care of the professional voice: A management guide for singers, actors, and professional voice users

    Otolaryngology and Head and Neck Surgery

    (2000)
  • O. Amir et al.

    The impact of hormonal fluctuations on female vocal folds

    Current Opinion in Otolaryngology & Head and Neck Surgery

    (2004)
  • O. Amir et al.

    Birth control pills and nonprofessional voice: Acoustic analyses

    Journal of Speech, Language, and Hearing Research

    (2006)
  • P. Boersma et al.

    Praat: Doing phonetics by computer v 6.0.21

  • F.S. Brodnitz

    Hormones and the human voice

    Bulletin of the New York Academy of Medicine

    (1971)
  • G.A. Bryant et al.

    Vocal cues of ovulation in human females

    Biology Letters

    (2009)
  • Cited by (0)

    View full text