PSYCHOLOGICAL SCIENCE
Research Article
220 Copyright © 2003 American Psychological Society VOL. 14, NO. 3, MAY 2003
EFFECTS OF EXPERIENCE ON FETAL VOICE RECOGNITION
Barbara S. Kisilevsky,
1
Sylvia M.J. Hains,
1
Kang Lee,
1
Xing Xie,
2
Hefeng Huang,
2
Hai Hui Ye,
2
Ke Zhang,
2
and Zengping Wang
2
1
Queen’s University, Kingston, Ontario, Canada, and
2
Women’s Hospital, Zhejiang University School of Medicine,
Hangzhou, People’s Republic of China
Abstract—The ability of human fetuses to recognize their own mother’s
voice was examined. Sixty term fetuses were assigned to one of two
conditions during which they were exposed to a tape recording of their
mother or a female stranger reading a passage. Voice stimuli were de-
livered through a loudspeaker held approximately 10 cm above the
maternal abdomen and played at an average of 95 dB SPL. Each con-
dition consisted of three 2-min periods: no stimulus, voice (mother or
stranger), and no stimulus. Fetal heart rate increased in response to
the mother’s voice and decreased in response to the stranger’s; both
responses were sustained for 4 min. The finding of differential behav-
ior in response to a familiar versus a novel voice provides evidence
that experience influences fetal voice processing. It supports an epige-
netic model of speech perception, presuming an interaction between
genetic expression of neural development and species-specific experi-
ence.
Scientists have been fascinated with the perceptual abilities of the
fetus at least since the late 1800s (see review by Kisilevsky & Low,
1998). However, systematic, sustained investigations began only in the
early 1980s following the development and ready availability of so-
phisticated ultrasound and fetal physiological monitoring technology.
It is now known that from about 30 weeks gestational age (GA), fe-
tuses begin to react to auditory stimulation. They show reliable cardiac
and body movement responses to a brief 2- to 3-s, high-pass-filtered
white noise at 110 dB SPL (estimated uterine attenuation varies with
increasing frequencies from 0 to about 35 dB; e.g., Querleu, Renard,
Versyp, Paris-Delrue, & Crepin, 1988) during high-heart-rate-variabil-
ity behavioral states (2F–4F; Kisilevsky, Pang, & Hains, 2000). By
term (37 to 42 weeks GA), stimulus intensity (Kisilevsky, Muir, &
Low, 1989) and frequency (Lecanuet, Granier-Deferre, & Busnel,
1988) modulate the threshold and magnitude of the response.
Fetuses also show the ability to perceive and discriminate seg-
mented speech sounds. For example, Groome et al. (1999) reported
that by 36 to 40 weeks GA, fetuses respond to lower intensity, 83- to
95-dB SPL speech stimuli by displaying heart rate decelerations to
vowel sounds (/i/ and /â/). Near-term fetuses can also discriminate the
reversal of pairs of consonant-vowel sounds, /babi/ to /biba/ or /biba/ to
/babi/ (Lecanuet et al., 1987), and a change in the gender of a speaker,
male to female or female to male (Lecanuet, Granier-Deferre, Jacquet,
Capponi, & Ledru, 1993). Preterm fetuses (26 to 34 weeks GA) are re-
ported (Zimmer et al., 1993) to discriminate vowel sounds (/ee/ and
/ah/), albeit at higher stimulus intensity levels, 100 to 110 dB. Al-
though these studies demonstrate that fetuses can discriminate sounds
and voices, the findings can be most parsimoniously accounted for on
the basis of the acoustic properties of the signals used in the studies.
They leave unanswered the question of whether fetal speech discrimi-
nation is based on genetic predisposition or influenced by in utero ex-
perience with voices.
Evidence of the influence of in utero experience on fetal ability to
recognize familiar voices and differentiate between familiar and unfa-
miliar ones is unclear. To date, few studies have examined whether
fetuses are capable of recognizing familiar voices or speech patterns.
Hepper, Scott, and Shahidullah (1993) failed to show discrimination
of the mother’s and a novel female’s voice, but did show discrimina-
tion between the mother speaking and a tape recording of her voice.
Similarly, Lecanuet, Manera, and Jacquet (2002) found that the mother
speaking elicited a fetal heart rate (FHR) acceleration in both the 2F
state (i.e., frequent and periodic body movements, REM, heart rate ac-
celerations during movements, irregular respirations) and the 1F state
(i.e., quiescent except for occasional startle, REM absent, heart rate
stable with little variability, regular respirations). The tape-recorded
voice of the mother or a stranger played via a loudspeaker elicited de-
celerations in both states. Again, although the response differed de-
pending on whether the mother was speaking or a tape recording was
played, no differential responding to mother’s and stranger’s voice
was reported. In one other study, DeCasper, Lecanuet, Busnel, Granier-
Deferre, and Maugeais (1994) found that fetuses could discriminate a
tape recording of a familiar versus a novel rhyme read by a female
stranger when the familiar rhyme had been repeated by their mother
every day from 33 to 37 weeks GA. However, the fetuses failed to re-
spond to an unfamiliar rhyme, which would seem to contradict the
findings, noted earlier, that fetuses reliably respond to speech. The
findings from these studies leave unanswered the question of whether
fetuses are capable of remembering and recognizing their mother’s
voice and differentiating between familiar and unfamiliar voices.
There is some evidence from newborn studies to indicate that fe-
tuses may be able to recognize and remember human voices. Using a
nonnutritive-sucking paradigm, researchers have demonstrated new-
born preference for the mother’s voice over that of a stranger (De-
Casper & Fifer, 1980), a story read by the mother during the last 6
weeks of pregnancy over a novel story (DeCasper & Spence, 1986),
and the native language over a foreign language (Moon, Cooper, &
Fifer, 1993). The timing of these preferences, occurring shortly after
birth, leads to the hypothesis that fetuses are indeed able to remember
and recognize human voices that they are exposed to in utero (the in-
utero-speech-experience hypothesis). Support for this hypothesis would
not only provide evidence of the early onset of speech perception and
memory, but would also have implications for theories of the early de-
velopment of social and cognitive abilities (e.g., language acquisition
and attachment).
The present study addressed this hypothesis. Specifically, we used
both FHR and body movement measures to index fetal ability to dis-
criminate between maternal and female strangers’ voices. Given that it
takes time to perceive the distinct qualities of a voice that distinguish
Address correspondence to B.S. Kisilevsky, 90 Barrie St., Kingston, ON
K7L 3N6, Canada; e-mail: kisilevb@post.queensu.ca.