Music in Pregnancy

by Julia Priest, M.S., M.Mus., for Perspectives, the journal of the Early Childhood Music and Movement Association, in press Spring 2010,


Parents often ask music and movement teachers whether class is beneficial for their newborns and young infants. Some parents tell us they plan to wait “until the baby can really get something out of it.” What can we tell them to support our intuitive conviction that they should come to class “early and often”? This article will report on four recent studies that should help early childhood music and movement specialists speak with evidence-based clarity. For the sake of simplicity, we will honor the scientific convention of calling a human being before birth a fetus, reserving the terms infant and baby for the neo-natal period.


The twenty-first century studies described in this article were designed to be more convincing than some of their predecessors by virtue of being prospective (the researchers ran the experiment on people instead of asking people to reconstruct and narrate past events from memory), randomized (substantially similar subjects were randomly assigned to the control or experimental groups, or else a single subject experienced the control stimulus and the research stimulus in random order), and double-blind (there was a sham stimulus; the statisticians didn’t know which data they were analyzing). We will close with a synopsis of a foundational twentieth-century study.


“Foetal response to voice and music.” Australian and New Zealand Journal of Obstetrics and Gynecology 2005; 45: 414-417, Al-Qahtani, Noura H.


Despite the drumbeat of the heart, rush of blood, and gurgle of digestion, the womb is quiet enough that the mother’s voice and other nearby sounds emerge clearly. Dr. Al-Qahtani reminds us that scientists were observing fetal response to external sounds as early as the 1920’s. However, measuring the responses is always confounded by the fact that fetuses move unpredictably and even their heart rates vary unpredictably. Statisticians have developed a panoply of arithmetical tests to separate out spontaneous variations in the data from variations that might be a response to the research stimuli.


Ten women whose pregnancies were at least 37 weeks advanced participated in the study. Dr. Al-Qahtani played each fetus both instrumental music and vocal sounds. She chose Spanish guitar music because, she said, it comprised a large range of frequencies and dynamics. The vocal stimulus was a female voice reciting nursery rhymes.


Each fetus was observed and measured for ten minutes while in the quietest state of sleep, as evidenced by still eyes and body and a characteristic heart pattern. After this baseline recording, a single earphone, placed against the mother’s belly, played either music, voice, or sham (tape recorder playing silence) for 15 seconds each. After a 3.5-minute silence, another of the stimuli was played, and this sequence was repeated nine times, with music, voice, or sham occurring in random order.


When all the data were collected and analyzed, it turned out that fetal movements and heart rates during the sham stimuli didn’t differ significantly from the random fetal events during baseline or silent times. There was a real reaction, however, during both spoken voice and guitar music: fetal heart rate accelerated, on average, to a degree that would be very hard to explain by chance. There were no measurable differences between responses to music and responses to speech.


One potential problem with this study is that, to avoid causing maternal anxiety, the mothers did not have earplugs, nor did they listen to alternate music. We know that fetuses do react to their mothers’ emotions; so, in this case, it is possible that mothers somehow transmitted information about each stimulus from their nervous system to their fetuses.


Dr. Qahtani points out that some other similar studies have found decelerations of heart rate rather than accelerations. She posits that different kinds of musical stimuli may explain contradictory findings. For scientists, studying music can be mysterious because it is such a complex stimulus. It can vary, as we know, by pitch, volume, duration, timbre, tempo, and so many other dimensions. For musicians and music teachers, the same mystery presents a world of opportunity—to soothe people or to energize them, or to communicate subtle and deep emotions. Fetuses show incipient responses to the complexity of music. Fetuses are so attuned to vocal quality and other characteristics of speech that they can discriminate, in utero, between a stranger’s voice and that of their own mother. If fetuses are so sensitive to timbre, then how much more so might newborns be?


“Maturation of fetal responses to music.” Developmental Science 7:5 (2004), pp. 550-559, Kisilevsky, B.S., Hains, S.M.J., Jacquet, A.-Y., Granier-Deferre, C., and Lecanuet, J.P.


In the 1980’s a psychologist and amateur musician named Dr. J.-P. Lecanuet began shifting the focus of his Paris lab from perception in cats to auditory perception in the human fetus. He became particularly interested in tempo perception. At the same time, independently, a researcher named Barbara Kisilevsky in Ontario was also beginning to study fetal auditory perception after some years studying newborns. In the 1990’s, when Dr. Kisilevsky spent a sabbatical year at Dr. Lecanuet’s laboratory, the two strands of research merged.


Early childhood music specialists often hear mothers assert “I listened to Baby Mozart throughout my pregnancy,” with a satisfied or defensive tone, as if insisting that they took their prenatal vitamins and never touched wine. The idea that music magically, effortlessly benefits development is widespread, but—no matter how much parents wish it--the support for it is “mostly anecdotal and is perhaps reinforced by a plethora of commercial audio-recordings (e.g. music, heart sounds) and devices purported to enrich the fetal environment and increase infant IQ.”


In order to test such claims, Drs. Kisilevksy and Lecanuet needed to begin by finding out the bare facts of how fetuses respond to music: do they 'tell us' (by changing their heart rate or moving) that they hear what we play to them? Do they change their response as they mature? Does music excite or calm them? Can they detect different tempi? The newly-formed team chose to study one of the musical genres that infants hear most frequently: lullabies. A graduate student in Dr. Kisilevsky’s lab was already using Brahms’ Lullaby to soothe babies in the NICU after painful heelsticks, so it made sense to continue working with that piece. Dr. Lecanuet prevailed on a professional pianist friend to record a five-minute track of the Brahms at 69 bpm, the traditional tempo, and another track at 118 bpm, for contrast.


At the time, prior research had already verified that external music is indeed heard clearly enough to be recognizable in utero; that the auditory system is substantially working after 33 weeks gestational age; and that immediately before birth, babies can detect certain gross auditory differences such as male/female voice, familiar/unfamiliar voice, and low/high piano pitch.


While playing the lullaby recording to the fetuses, the scientists attempted to separate out mothers’ responses from fetal responses by masking the mother’s hearing with earphones playing dissimilar music (either country or guitar music). They then measured fetal heart rate and body movements, especially in the thirty seconds before and after the music started, and again the thirty seconds before and after the music stopped.


During data analysis, they separated the fetuses into groups by gestational age (GA). A group of half a dozen fetuses were between 28 and 32 weeks GA when studied, which means that their auditory systems were not fully developed. Another half-dozen fetuses were 33-34 weeks GA, the age at which, previous research has suggested, the auditory system is fully developed. Over a dozen fetuses were at 35-36 weeks GA when they participated in this study. Another dozen were called “term fetuses” because, at 37 or more weeks, they would no longer have been considered premature had they been born. Indeed, the only fetuses included in the data analysis were those who subsequently turned out to be healthy term babies.


When the researchers turned the music on, all age groups showed changes of heart rate. Statistical testing strongly suggested that these changes could not have occurred by chance. In the youngest fetuses, heart rates generally accelerated at the onset of music if it was played loudly: this was thought to indicate arousal. If the same music was played more softly, heart rates generally decelerated at the onset of music, which was thought to indicate attention. The deceleration effect became more common in the older fetuses. The “term” fetuses showed heart rate deceleration no matter how loud or soft the music was played. In other words, the most mature fetuses may have been paying some primitive form of attention to the music.


The Canadian team used only the traditional-tempo recording, but the Paris team played the lullaby twice for each fetus, in random order: for half the fetuses, they played it first at an ordinary tempo, 69 beats per minute and then again much faster, at 118 beats per minute. The other half of the group of fetuses heard the fast version before the a tempo version. During the course of the five-minute music, term fetuses showed an increase in heart rate to the faster tempo, but no change with the normal-tempo music.


Understanding this study depends on getting comfortable with the technical terms “arousal” and “attention.” The youngest fetuses may have been exhibiting a startle response in reaction to the onset of music in their environment. The more mature fetuses, because they may have been able pay some primitive form of attention to the music, could also respond differentially to faster and slower tempo. If tempo and timbre are both salient features of music for fetuses, we are forced to ask how many other features of music fetuses can also experience with some degree of clarity, such as melodic contour or rhythmic pattern. We will also want to know whether they will remember, as newborns, the music they heard while in utero.


“Fetal learning: a prospective randomized controlled study.” Ultrasound Obstetrics and Gynecology 2002; 20: 431-438. James, D.K., Spencer, C.J., and Stepsis, B.W.


David James had long had a research interest in fetal neurobehavioural development. In the 1990s and there was a lot of discussion both in the lay press and in the scientific literature about whether fetuses could learn; Dr. James designed a study to address this question. Some people studied the spoken word (both maternal and non-maternal), but he decided to study music because he is an amateur musician and his mother was a professional musician. In their quest for a music stimulus that was fairly rhythmic and with a significant bass part, his team started with Carl Orff's Carmina Burana. Parents who recognized the track from the soundtrack of the 1976 film “The Omen” soon expressed anxiety about the music’s effects on their fetus, so the researchers switched to another very rhythmic track with a “wide range of tone: Glenn Miller's Little Brown Jug—dated, but not controversial.”


Twenty pregnant women were studied 92 hours prior to their elective date of delivery. Each pregnant woman wore headphones on her abdomen for five consecutive hours. None of the women had been previously exposed to the Little Brown Jug during pregnancy, so their fetuses could have formed no associations, whether positive or negative, with it. For one initial silent hour, researchers took a baseline recording of heart rate and body movements for all the fetuses. Thereafter, the experimental group was differentiated from the control group by having a continuously looping recording of the Glenn Miller track played through the headphones to the fetus.


Mothers in the control group continued to wear silent headphones on their bellies. This “sham stimulus” was an attempt to keep the mothers unaware of whether they were in the experimental group or the control group, to make this study nearly a “double blind” study.


For the fetuses in the experimental condition, music was played through the headphones on their mothers’ bellies the entire time. During the first and fourth hours of this period, fetal heart rate and movements were recorded for all the fetuses, both control and experimental. Strikingly, during the fourth hour of hearing music, fetuses who heard music showed significant differences in their heart-rate and movements compared to fetuses in the control group. They showed more state transitions (awake to asleep, etc.) and spent more time awake.


Three to five days after birth, all babies were studied again in a quiet room, thirty minutes after feeding. The babies’ heart rates were measured through special neonatal electrodes attached to an electrocardiogram machine. A trained observer took notes of the infants’ body, limb, and eye movements. In this neonatal situation, thirty minutes of “baseline” (no music) were followed by an hour of looped Little Brown Jug music, played through a headphone 2-3 feet from the baby’s head. All of the newborns, including those in the control group, now heard this music. All the neonates showed significant differences in their heart rate and movements while hearing music compared to the silent condition. They showed more state transitions (awake to asleep, etc.) and spent more time awake.


There were, however, also marked differences between those babies who had heard four hours of music in utero and those who had not. The music group changed sleep-wake state more efficiently and spent more time awake than the non-music group.


The limitations of this study include the fact that mothers in the experimental group may have become aware of the music vibrating on their abdomens, and their knowledge or feelings about this may have had an effect on fetal behavior. Another problem is that fetuses and newborns change spontaneously and unpredictably: even the control group exhibited significant changes during their time of listening to silence.


The authors believe their study demonstrates that fetal exposure to music results in the development of altered behavior in the fetus before birth and in the newborn period, and consider this to be evidence of a kind of learning, albeit the most primitive kind. Music and movement specialists should feel slightly more comfortable than ever before in assuring parents that their babies’ remarkable abilities include some ability to learn music.


“An examination of fetal learning before and after birth.” The Irish Journal of Psychology 1991; 12: 95-107. Hepper, P.G.


One of the most elegant studies of learning a piece of music in utero was carried out three decades ago by British psychologist Peter Hepper. He recruited a group of mothers who watched one particular television show regularly during pregnancy and compared them to mothers who did not watch that show. Fetuses whose mothers did not watch the program exhibited no body movement changes when researchers played the program’s theme song. Likewise, even in mothers who watched the show regularly, younger fetuses (29-30 weeks gestational age) displayed no particular reaction to the theme song of that television show. But fetuses who were 36-37 weeks along, whose mothers watched the show regularly, changed the pattern of their body movements when that theme song was played. This suggested that somewhere between 30 and 36 weeks gestation, fetuses developed the ability to recognize familiar music. When Dr. Hepper tested these same fetuses as newborns, again they showed by changing their behavior that they had formed memory traces of the music. Memory for the theme song disappeared by three weeks after birth if the mother did not continue to watch the program.


In conclusion, we can tell parents:

  • music is clearly audible in utero and, after 33 weeks gestional age, fetal ears are mature enough to hear it;
  • as they mature, fetuses begin to distinguish between music and other rhythmic noises such as the mother’s heartbeat;
  • fetuses can discriminate between their mother’s voice and that of a female stranger;
  • term fetuses can detect differences such as male/female voice and low/high piano pitch;
  • term fetuses seem to give music some form of attention;
  • between 30-36 weeks, fetuses begin to form some kind of primitive memory traces that persist into the newborn period.
  • to maintain the memory trace, newborns need to hear music repeatedly.


Rather than asking whether their baby is old enough to benefit from music class, perhaps parents should be asking us whether to start music class when their fetus is at 33 weeks gestational age, or wait until 37 weeks, just prior to the estimated due date. The field of fetal auditory perception is (forgive the pun) in its infancy, but earliest reports tend towards supporting Kodaly’s vision that music education should begin “nine months before the birth of the mother”!


The body of research on all types of prenatal cognitive abilities is small. The sense of hearing, being the fetus’s most well-developed distal sense, is naturally the easiest to study and will provide the most natural opportunity for scientists to determine more about the way that thinking itself develops in the unborn human.