sign in sign up
<<
Home , Fetus

The Fetus &&&&&&&&&&&&&& Evidence of Transnatal Auditory Learning

Christine M. Moon, PhD hurdles to surmount in asking newborn whether they William P. Fifer, PhD have learned about specific sounds before birth. In order for transnatal retention of auditory experience to be possible, three key conditions must be satisfied. The fetus must There is converging evidence for fetal retention of auditory experience into be able to hear, there must be sound to detect, and the fetus early postnatal life, but critical tests with appropriate controls are rare due must have the ability to learn. There is much evidence that the to methodological hurdles. Research has been conducted on newborn intrauterine environment is acoustically rich (see Gerhardt and response to naturally occurring stimuli such as heartbeats, intrauterine Abrams, this volume), that the fetus is capable of recordings, pre-and postnatal versions of the maternal voice, father's hearing (see Hall, this volume), and does respond to sound in voice, and unfamiliar voices. Postnatal experience cannot be ruled out as a the womb (see Abrams and Gerhardt, this volume). Bearing on possible explanation for many results. Only one critical prenatal exposure the learning question, there is growing evidence from non- experiment with postnatal testing has been carried out and published in a human animals and from that learning occurs prior to peer-reviewed scientific journal. Interpretation of acoustic and linguistic birth. Birds, e.g., learn about the calls of conspecifics while still information on intrauterine recordings suggests that the prosodic features in the egg, and the birds' own post-hatching vocalizations are of speech (pitch contours, rhythm, and ) are available to the fetus. affected by this experience (see Lickliter, this volume). In This is compatible with newborn responses and may contribute to addition to birds, non-human have provided evidence acquisition during the first year. There is no sound evidence that providing of prenatal learning. For example, elegant experiments with rat extra prenatal auditory stimulation benefits the developing , and there fetuses have demonstrated that it is possible to classically are potential risks. condition aversion to a specific odor prior to birth that affects Journal of Perinatology 2000; 20:S36±S43. the rat's behavior after birth.1 Empirical investigation into whether human learning can be LEARNING BY THE FETUS demonstrated during the fetal period has a long, if not replete, history. In 1932, Ray2 described a single subject classical Many, if not most, people believe that we can learn before we conditioning experiment. The CS was the clapper of an electric bell are born and that some of the things we have learned are gong, and it was placed so that it vibrated on the 's abdomen. retained after birth. It is not difficult to find someone who can It was paired with the UCS that was fashioned out of two wood relate a personal or family story about a baby who seemed, blocks made to snap together by a spring from a rat trap, thus shortly after birth, to respond in a special way to a familiar making a loud sound. Problems with timing of stimulus voice or sound. Examples of such accounts are the 1-day-old presentation and fetal response interval prevented Ray from who turns to look in the direction of a family member visiting reporting the data. In 1948, Spelt,3 using the same methodology in the hospital room for the first time after the birth, or a an experiment with 13 pregnant women, measured four types of newborn who seems to reliably become quiet when a familiar and concluded that classical conditioning of the piece of music is played. From this perspective, it is surprising human fetus had occurred, although incomplete information in the that although there is much converging evidence for human report renders independent interpretation of the results difficult. transnatal auditory learning, there are few carefully controlled More recently, Hepper4 referred to his replication of Spelt, but did not human studies on the topic. From the researcher's perspective, describe the experiment. If the literature on human fetal classical the paucity of published empirical studies in peer-reviewed conditioning is sparse, the literature on instrumental learning is journals that directly address transnatal learning is under- non-existent. standable. There are daunting methodological and practical Despite the paucity of research on fetal associative learning, several studies of fetal habituation learning do exist. Habituation Pacific Lutheran University(C.M.M.), Tacoma, WA 98447;Department of Developmental occurs when there is a reduction in fetal responding to a repeated Psychobiology(W.P.F.), New York State Psychiatric Institute;Columbia University, Box 40, 722 N. 168th St., New York, NY 10032. auditory stimulus, and the decrement cannot be attributed to sensory or response system fatigue.5,6 Demonstrations of habituation learning Address correspondence and reprint requests to Christine M. Moon, PhD, Department of Psychology, Pacific Lutheran University, Tacoma, WA 98447. are especially convincing when, after a decrease in responding to a Journal of Perinatology 2000; 20:S36 ± S43 # 2000 Nature America Inc. All rights reserved. 0743-8346/00 $15 S36 www.nature.com/jp Transnatal Auditory Learning Moon and Fifer

repeated stimulus, the fetus begins responding anew when a that they are capable of retaining information about specific auditory discriminately different stimulus is presented. Recovery of both stimuli for at least 24 hours,23 including differentiating two words motor7,8 and cardiac9,10 responding to a change in auditory spoken by the same female speaker.24 These latter results bear stimulation has been reported for late-term fetuses (see Lecanuet strongly on transnatal learning because it is unlikely that the and Schaal11 for a review). demonstrated capacity for neonates to retain auditory information Repeated experience over longer periods of time than those over a longer period had its onset only once the has been born. afforded by the laboratory habituation or classical conditioning In addition to learning in habituation procedures, newborns have setting may result in learning by the fetus. This has been described as also shown that they can be classically conditioned, although it has exposure learning12 and may result in a preference formed by been difficult to do so.25 Results with instrumental conditioning are familiarity.13 ± 15 In one study of response to the mother's voice, which much more prevalent. For example, infants less than 4 days old can is presumably a familiar sound, 36-week-old fetuses did not learn in less than an hour that headturning26 or sucking on a tube 27 demonstrate a difference in numbers of movements (visualized via is rewarded with a sweet taste. Non-nutritive sucking has proven to ultrasound images) to a tape recording of the mother's versus a be a relatively malleable behavior in instrumental conditioning stranger woman's voice played through speakers placed directly on procedures with newborns. Butterfield and Siperstein28 described the mother's abdomen. But in a separate condition, fetuses of the changes in sucking by 24- to 48-hour-old infants as a result of same did respond with fewer movements when the hearing music that was either contingent or non-contingent on mother was actually speaking versus when presented with a tape sucking. DeCasper et al. have developed an instrumental recording of her voice.16 This may mean that 36-week-old fetuses conditioning procedure using non-nutritive sucking to assess cannot discriminate between the mother's and another female's newborn infant perceptual and learning capacities. In this procedure, voice, but can discriminate characteristics of the maternal voice that 1- to 4-day-old infants learn to activate and maintain different result from different means of transmission. Alternatively, the sounds presented through headphones by altering the duration of methods used in the study may not have been sufficiently sensitive to their pauses between sucking bursts (interburst intervals) and/or detect fetal ability to discriminate between the externally presented the duration of their sucking bursts.13,29 ± 34 In a variant of this non- voices. A more recent study using fetal rate suggests that it may nutritive sucking preference procedure, 2-day-olds can learn within have been the latter. Pregnant women recited one of two different 15 minutes to initiate sucking bursts during auditory signals for the children's rhymes daily between weeks 33 and 37 of their availability of preferred sounds.29,32,34 ± 36 The parameters of neonatal . Fetal heart rate was monitored in week 37 during learning have been the focus of some recent studies using non- presentation of a tape recording of each rhyme separately. Heart rates nutritive sucking and auditory stimulation with investigators finding declined significantly during the rhyme that had been recited by the evidence for cognitive processes in newborns beyond simple mother, but not during the other rhyme.17 association learning.37,38 As with habituation learning, it seems If the results from all fetal learning studies are taken together, it unlikely that such capacities emerge only during the first day after appears that there is some evidence for human prenatal learning birth. although more research is certainly warranted. Another potential source of information about fetal learning is the neonate. It seems unlikely that birth marks an abrupt change in the perinate's ability NATURALLY OCCURRING SOUNDS to learn. Indeed, learning by the premature infant supports this, If we accept that fetuses can learn about sounds that they hear in although it is difficult to study the learning capacities of these infants utero, then it becomes reasonable to ask whether there is any due to their fragility 18 plus their range of medical conditions and indication that newborns respond differentially to sounds that state lability. Full-term neonates have revealed themselves to be occurred during the prenatal period either naturally or through adept learners and it is likely that much of this capacity was already deliberate exposure. One candidate natural prenatal sound is the present at the end of prior to birth. Within the first days rhythm of a heartbeat. It has long been surmised that heartbeats are after birth, newborns demonstrate habituation learning. For sounds available in the womb39,40 and indeed, maternal cardiovas- example, between 1 and 3 days after birth, infants demonstrate cular sounds have been present in recordings made in utero in habituation and recovery of localized headturning to rattle sounds 19 humans during labor after the rupture of amniotic membranes,41 and to spoken words.20 Infants 4 to 5 days of age suck rapidly on a although it is not clear that they actually are a prominent form of non-nutritive nipple when sucking produces a recording of speech stimulation for the fetus.11,42,43 Salk39,40 reported that playing recordings sounds for them to hear. Sucking gradually habituates but returns to of a heartbeat sound for groups of newborns in a hospital nursery high rates when a new speech sound is presented.21,22 Habituation resulted in greater gain and reduced crying time. Other experiments with newborns typically take place over a period of researchers have also found that intrauterine sounds calm minutes and demonstrate that infants retain stimulus characteristics neonates.44,45 Based partially on this research, commercial products long enough to compare them with a new exemplar. There have using heartbeat sounds have been devised for pacifying young been habituation experiments with neonates, however, that suggest infants.44 Other research has, however, called the soothing effect of

Journal of Perinatology 2000; 20:S36 ± S43 S37 Moon and Fifer Transnatal Auditory Learning

heartbeat sounds into question. There have been failures to replicate maternal voice has been compared to stranger female voices. The this.46,47 There have also been many experiments with sounds other study of subjects with the least postnatal experience was carried out than heartbeats that show that a neonatal arousal response to sound with infants within 2 hours after birth, during which time efforts depends on many variables including broad characteristics of the were made to keep vocal stimulation at a minimum. The researchers stimulus, the infant's initial state of arousal, and the experimenter's do report, however, that some words were spoken by the on choice of response.47,48 Although evidence for soothing infants with postnatal contact with the infants although they were three or four heartbeats is ambiguous, there are more convincing results on the words at most. Mothers and four other women attending the birth salience of heartbeat sounds for newborns in a non-nutritive sucking spoke the infant's name. The neonates were videotaped for later blind preference experiment that showed that 66-hour-old infants altered coding of infant head, eye, hand, and sucking movements. Infants interburst intervals when the consequence was hearing a recording of reacted with more movement when the mother was speaking intrauterine heartbeat sounds. These results demonstrated that compared with the strangers.51 Although researchers attempted to heartbeat sounds acted as a reinforcer for the infants.14 The control for voice quality, it is not possible to rule out the likelihood experiment did not compare responding to heartbeat sounds with that the voices of women who had just given birth to these infants other rhythmic sounds. It is, thus, not clear whether it was the were different in some systematic way compared to the other voices. presumably familiar prenatal cardiovascular sound or the In a study using recordings of the mothers' and stranger females' characteristic of rhythmicity that was reinforcing for the newborns. voices, DeCasper and Fifer13 used the non-nutritive sucking Demonstration of infant responsiveness to rhythmic sounds in preference procedure to compare responses by infants less than 3 days general would not, however, rule out the effectiveness of prenatal after birth to familiar and unfamiliar female voices. Infants activated exposure to heartbeats because experience in utero could generalize the recordings of their own mothers' voices proportionately more to other rhythmic sounds. frequently than the recordings of the stranger females. DeCasper and In addition to the cardiovascular sounds present on recordings Fifer13 replicated this finding in the same study using a new group of made in utero, the maternal voice has been noted to be a prominent neonates and a variant of the sucking procedure. The maternal voice sound, more intense than other voices. It is, however, not more preference was again replicated in a different non-nutritive sucking intelligible than other voices. All voices on the intrauterine recordings experiment with 2-day-olds. 50 In yet a different study comparing sound muffled, largely due to the attenuation of frequencies above newborn movement during recordings of the maternal voice versus a 500 Hz resulting from the low-pass filtering properties of the stranger female voice, infants 27 to 95 hours old made fewer head, intrauterine environment (see Gerhardt and Abrams, this volume). arm, and leg movements when the recording was the mother's For the perinate experiencing the sound of the maternal voice, this voice.16 Interestingly, the study included a comparison of responding suggests two things: (1) the mother's voice is salient; and (2) her to the mother versus stranger female with the voices talking in both voice sounds different after birth than it did before birth. If -directed (AD) and infant-directed (ID) speech. ID speech familiarity results in preference for salient sounds, then the more has been shown to differ in several ways compared to AD speech.52,53 familiar prenatal version of the mother's voice should be preferred ID speech would normally not be a maternal speech pattern during the early postnatal period. Three different experiments have available prenatally and would thus be novel for the newborn.54 used a strategy of utilizing a simulated version of the maternal Results showed that there was an interaction between voice and prenatal voice in preference procedures with newborns. Simulated speech pattern such that infants made significantly fewer movements prenatal versions were created by low-pass filtering recordings of the when the voice was maternal AD speech compared to the three other mother's voice. In two experiments, the simulated prenatal maternal types (maternal ID and non-maternal ID and AD) of speech voice was compared to an unaltered recording which served as a samples.16 The interaction of voice and speech type would seem to postnatal sample for the infant. Results of these prenatal versus support prenatal familiarization with maternal AD speech. It could postnatal experiments showed that newborns detect the equivalence be, however, that even though mothers do use ID speech when of the prenatal and postnatal recordings,32 and they do prefer a low- speaking to their newborns,52 the relative number of ID speech pass filtered recording of the maternal voice compared to an samples is small compared to AD speech during the early postnatal unfiltered recording of her voice.49 Another strategy has been to use period. It is, thus, possible that the infants could have been the preference procedure to compare responding to stimulated responding to maternal AD speech on the basis of postnatal, rather versions of the mother's prenatal voice versus a stranger female's than prenatal, experience. prenatal voice. In these cases also, the mother's prenatal voice was Father's voice has received some attention by researchers; preferred.34,49,50 This preference for the prenatal version of the mother's although newborns discriminate between the paternal and a stranger voice strongly suggests that newborns can learn about sounds male voice,31,55 in one study of six infants, the neonates did not available in the womb and that early postnatal responding is demonstrate a preference.31 The difference in newborn responding to influenced by this experience. parental voices has been interpreted as evidence for prenatal auditory The salience of the mother's voice to infants shortly after birth has learning and is the result of the relative prominence of the mother's been the focus of several experiments in which responding to the voice in utero.31 Interestingly, a recent study of older infants also

S38 Journal of Perinatology 2000; 20:S36 ± S43 Transnatal Auditory Learning Moon and Fifer

showed no preference for the paternal voice versus a stranger male the two melodies.61 In a similar non-nutritive sucking preference voice although the 4-month-olds in the study could discriminate study, the mothers recorded two lullabies and played the recording of between male voices.56 Perhaps prenatal exposure has lasting effects one of them daily for 10 minutes during the last 8 weeks of on infant response to the maternal voice, but there are also . Sixteen infants were tested with the two maternal lullaby important differences in the newborn's postnatal experience with recordings, and there was a significant preference for the familiar each that could begin to affect infants as early as the first day lullaby.62 In the third dissertation study of prenatal exposure to a after birth. lullaby, a different methodology was employed. The mothers of 15 In addition to demonstrating preferences for listening to infants began singing or playing a recording of themselves singing particular voices, newborns have shown that they respond one of several lullabies 7 weeks prior to their expected delivery dates differentially to . Within the first 4 days after birth, infants as well as for 4 weeks after delivery. A different group of mothers sang discriminate the language that their mother speaks compared to a or played a lullaby only after birth for 4 weeks, and a third group of foreign language,57 and 2-day-olds have demonstrated a preference mothers was not instructed to sing to their infants either pre-or for the maternal language compared to a foreign language.58 The postnatally. At 4 weeks of age, infants were exposed to a recording of rhythmic structure of the two comparison languages appears to be the maternal lullaby while crying, and infant state of arousal was important in the neonates' ability to classify utterances.59 coded to determine whether prenatal exposure had a differential effect on soothing. Results showed that infants who received prenatal and postnatal lullaby exposure spent more time calm when presented DELIBERATE EXPOSURE TO SOUND with the lullaby tape than either of the control groups.63 This could Given the prominence of the mother's voice on intrauterine tape be evidence that prenatal experience directly influences postnatal recordings41,60 and the robust newborn response to the maternal voice behavior or it could be an effect of the total duration of exposure in compared to other voices, it seems very likely that prenatal learning which case prenatal and postnatal experience would add together. It has influenced postnatal responding. However, in the experiments cannot be ruled out, however, that infants' response to the prenatal with voices and languages described above, it is not possible to rule lullaby was due to more sensitive parenting that resulted from a out the effect of postnatal experience. Remarkably, there is only one longer period of parent involvement for those who started singing to published experiment that is a direct test of prenatal learning about their infants before they were born. speech. DeCasper and Spence32 asked women to read a particular 3- In a quite different study of the effect of prenatal music exposure, minute speech passage out loud two times a day during the last 6 infants 2 to 4 days of age were monitored while listening to a weeks of pregnancy. After these infants were born, 16 of them were recording of the theme song of a television soap opera that their participants in the non-nutritive sucking preference procedure, and mothers reported watching daily during pregnancy (Learning their activation of recordings of familiar or novel speech passage was group). Their responses were compared to those of a control group compared. For nine infants, the voice reading the two stories on the of infants whose mothers did not watch the television program. recording was the mother, and for seven infants it was a stranger Results showed that compared to the Control group, the Learning female who read the two stories. Results showed that infants activated group experienced a decrease in heart rate, decrease in number of the tape recording of the prenatal story relatively more frequently movements, and change in behavioral state to a more alert state compared to the novel story, independent of who was reading the while listening to the music.64 Since none of the infants had heard stories on the tapes. For a group of control infants who had not been the soap opera theme song after birth, these results suggest that the previously exposed to the speech passages, there was no systematic fetus can learn about incidental ex utero environmental sounds as preference. This is one experiment in which postnatal experience well as the mother's speaking and singing voice, and that prenatal with the auditory stimulus can be ruled out as an explanation for learning affects postnatal behavior. Intriguingly, these results and the newborn response to a prenatal event. results from deliberate prenatal exposure to music seem to be at odds Infants have also been shown to be responsive to music to which with the newborn's failure to respond to the paternal voice. Father's they have been exposed prenatally in three different unpublished voice would seem to occur more frequently in the fetus' auditory dissertation studies. In one study of 13 neonates, infants were tested environs compared to the music selections. An explanation for the with the non-nutritive sucking preference procedure for response to a preference for familiar music, but not for the paternal voice, must be familiar and an unfamiliar melody. Their mothers had repeated the based on something other than frequency of exposure.56 melody of ``Mary Had a Little Lamb'' using the syllable ``la'' instead Taken altogether, the studies of perinatal learning plus the of words for about 5 min/d during the 14 days prior to expected experiments on fetal and newborn responses to familiar prenatal delivery of their infants. Results showed that despite the brief sounds warrant two conclusions: (1) there is sufficient evidence to exposure period, the 4-day-olds who had been exposed to the support the claim that transnatal learning occurs; and (2) we know melody prenatally activated it more frequently relative to a novel very little about transnatal learning. Clearly, the conditions under melody. A control group of 4-day-olds who had not been which transnatal learning occurs are an area of fertile ground for deliberately exposed to the melody showed no differential response to future research.

Journal of Perinatology 2000; 20:S36 ± S43 S39 Moon and Fifer Transnatal Auditory Learning

TEACHING THE FETUS intended and produce an enhanced loss of neurons, e.g., through neuronal excitotoxicity.73 Some practitioners of have not waited The Prenatal University, Project Eastman, and Project Prelearn for basic research to proceed on transnatal learning. Programs of have had enhancement of cognitive development as goals, but other deliberate instruction of the fetus have emerged in the last 20 interventions have had other goals such as enhancement of parenting years, and products and practices for the purpose of teaching the practices74 and relaxation of infants, children, and .75 The unborn are available in many cities in the U.S. and over the accounts of prenatal teaching programs are fascinating and they Internet.65,66 Many of these programs have been described in point the way to possible future research, but they must be interpreted articles in the journal, Pre- and Perinatal Psychology. The with skepticism. The published studies of prenatal stimulation each longest-running practice of regular stimulation of the fetus seems contains methodological flaws that render claims of change due to to be that of Rene Van de Carr, MD, of Hayward, CA, who has the interventions questionable. The accounts provide data on single directed the Prenatal University program since 1979. In this cases or small numbers of subjects and the studies are rarely curriculum, the mother and the father are taught to stimulate the conducted with adequate controls. To be convincing, these studies fetus 5 min/d by doing such activities as repeating bisyllables should include comparison groups of subjects such as fetuses who aloud, patting the mother's abdomen in the ``kick game,'' and receive a different kind of prenatal stimulation, different amounts of singing and reciting rhymes to the fetus. Although the central prenatal stimulation, and no postnatal stimulation that may affect focus of the practice is enhancement of parent±infant the variable of interest. Testing for effects should be blind (see communication, developmental gains by infant participants have Philbin, this volume, for a description, a checklist, and examples of been described. These include comparative group data on early good research methodology). communication and speech, single cases of early speech (e.g., In addition to the methodological issues mentioned above, the words spoken at 4 months),67 and a case of apparent specific research on purposeful prenatal stimulation must be questioned for memory by a of a prenatal event.68 use of extrauterine stimuli that may be too intense for the developing Another program, the Eastman Project, was initiated to auditory system or may interfere with normal fetal sleep/wake cycles enhance music and speech development through prenatal (see Gerhardt and Abrams, this volume, for a discussion of research exposure to music via earphones placed directly on the skin of the on the effect of intense noise on auditory development). It would be maternal abdomen. Preliminary results included ``early develop- considered cruel to deliberately and repeatedly subject a sleeping ment of highly organized and remarkably articulate speech''.69 newborn infant to loud noise near his/her ears. Yet, for the fetus, There does not appear to be a published account of the final what is likely to be the equivalent of this practice has been claimed results of the project. by some researchers to have benefits. In a third program of fetal instruction, Project Prelearn, the auditory stimuli were variants of the maternal heartbeat and were designed to present auditory patterns in 2-hour daily ``lessons'' that became progressively more complex over time. A central premise of In total, there appears to be sufficient evidence that prenatal auditory the program was that pattern detection by the fetus provides experience affects postnatal responding to the sound of voices stimulation to the developing brain that results in prevention of talking. This conclusion carries important implications for the apoptotic neurodegeneration or programmed neuron loss, which is investigation of the development of speech perception and language otherwise normal during the perinatal period.70,71 The retention of acquisition. The most obvious implication is that the process does neurons was proposed to support increased cognitive ability for not begin at birth, but before birth. Certainly, there is a major individuals in the program relative to normal children. Results of increase in the amount of speech information available to the infant prenatal stimulation of 12 participants included larger-than- after birth, and much learning occurs in the early postnatal period. normal weight, height, and cranial circumference at birth; cooing, But characteristics of the prenatal environment may render the babbling, and relaxed body posture on delivery; and advanced sound of voices and the speech signal particularly salient to the receptive and expressive .72 The idea that newborn.15 The mother's voice, e.g., is not only a prominent acoustic extrauterine stimulation will enhance infant development by event in utero by virtue of its amplitude60,76 ± 78 and how often it occurs, preventing normal neuron degeneration is unsubstantiated but the sound of the voice is accompanied by stimulation in other elsewhere in either the human or animal literature. It is neither sensory modalities. There may be some vestibular stimulation of the clear that this type of extra stimulation will prevent apoptosis nor is fetus by movement of the maternal diaphragm and muscles when it known whether this would result in cognitive benefits if this she speaks, and this would presumably co-vary with loudness. There normal process were prevented from occurring. And contrary to the is a direct relationship between the maternal body and speculation that the experimental subjects may have more neurons presence versus absence of voice. When the mother is lying in bed, than other infants is the possibility that the extrauterine stimuli there will be a greatly reduced likelihood of hearing her voice. The may act as stressors which could have results opposite those sound of her voice is likely to be accompanied by maternal upright

S40 Journal of Perinatology 2000; 20:S36 ± S43 Transnatal Auditory Learning Moon and Fifer

body position and large motions, e.g., of walking. Even when seated, languages.85 It appears that rhythm is a particularly important the mother's torso is likely to be in frequent motion due to change of prosodic cue for newborns.21,86 Prosody has gained attention in body position and movement of legs and arms. Thus, for the fetus, language acquisition theory in recent years as a potentially the mother's voice would seem to be available as a multimodal important means for older infants to discern communicative event. This differs from the sounds of other voices and may enhance intent of the speaker87 and in the notion of ``prosodic the salience of the maternal voice, but also may also cause the sound bootstrapping'' or using intonation, rhythm, or stress cues in of the human voice, in general, to be an object of the infant's speech to unpack the underlying syntactic organization of attention. utterances88,89 (see Jusczyk82 for an expanded discussion). It is Although this is speculative, another way in which the maternal not unreasonable to suppose that prenatal exposure to the voice could retain its prominence in early postnatal life is through its prosodic features of speech plays a role in a cascade of perinatal constancy relative to that of other sensory experiences. True, developmental events that contribute to eventual language the difference between the muffled, low-pass filtered sound of the acquisition. intrauterine version of the voice and the extrauterine version is not Given that transnatal learning about speech and voices may play insignificant, but it may be small compared to the magnitude of a role in the infant's acquisition of language, is it advisable to stimulus change in other senses like vision and the cutaneous, provide extra exposure to speech during gestation as recommended somesthetic, olfactory, and gustatory senses (see Lecanuet and by proponents of such programs as the Prenatal University? We Schaal 11 for a review of sensory development). Hearing the maternal remain very skeptical of claims that deliberate prenatal teaching/ voice may, thus, provide a sensory bridge into postnatal life that is learning episodes can enhance postnatal cognitive functioning, but not available from other prenatal sensations. we are in agreement with those who think of talking and singing to The salience of the maternal voice is likely to generalize to an unborn child as the beginning of a caring relationship.63,74,75,90 attention to other voices, and there is evidence from a variety of There is sufficient sound scientific evidence to conclude that the research paradigms that newborns6,54,79 and older infants80,81 attend to sound of a pregnant woman's speaking and singing voice is speech even when the voices are not familiar. For example, newborns available to her unborn child. In fact, her voice is available whenever change sucking patterns when the consequence is hearing a stranger she speaks Ð not only when she is talking to the fetus. Moreover, it female voice versus silence.50 Prenatal experience may be an appears very likely that much of the acoustic environs of mother is important factor in the early postnatal bias to listen to speech (see also available to the fetus in addition to the ongoing internal Jusczyk82 for a discussion). physiological sounds that have been recorded in utero (see Abrams In addition to a general bias for speech, prenatal experience may and Gerhardt, this volume). Given that this is the case, increasing predispose newborns to attend to some components of the speech fetal stimulation beyond what is provided by the ambient sound level signal and to ignore others.15,82 Infants appear to be particularly is not warranted. It may be detrimental not only to auditory responsive to the prosodic elements of speech which include pitch development (Gerhardt and Abrams, this volume) but also contours, rhythm, and stress. These characteristics of speech are potentially to the development of behavioral state regulation, clearly represented in intrauterine tape recordings, whereas the attention, and other cognitive processes about which much is individual sound segments of speech or phonemes are not well unknown. We highly recommend against using devices such as represented76,77 (see Gerhardt and Abrams, this volume). In the miniature audio speakers or megaphones to present sound to the laboratory, newborns less than 4 days of age have proven themselves fetus. Those who promote these types of practices based on selective, to be very capable of responding to segmental components of speech unfounded, and illogical application of scientific data on early by, e.g., discriminating between two vowels,35,83,84 by discriminating two learning and brain changes are taking risks with the development of consonants in brief consonant±vowel syllables,21 and by discrimi- the unborn child. nating syllables that differ in order of consonants.35 Although neonates demonstrate competence at the segmental level, they recognize and prefer to listen to a filtered recording of the mother's voice which has References attenuated the frequencies that carry much of the information about 1. Robinson, SR, Smotherman WP. Habituation and classical conditioning in the rat fetus: opioid involvements. In: Lecanuet J-P, Krasnegor NA, Fifer WP, segments, leaving only prosodic information.33,34,49 Furthermore, research with whispered voices supports the notion that the prosodic Smotherman WP, editors. Fetal Development: A Psychobiological Perspective. components are salient to newborns. In whispered speech, voicing is New York: Erlbaum; 1995. pp. 295±314. 2. Ray WS. A preliminary report on a study of fetal conditioning. Child Dev absent and thus much of pitch contour information. Infants less than 1932;3:175±177. 3 days of age can discriminate between unfamiliar whispered voices, 3. Spelt DK. The conditioning of the human fetus in utero. J Exp Psychol but they do not alter sucking patterns in order to activate a recording 1948;38:338±346. of a whispered voice versus silence.34 4. Hepper PG. Memory in utero? Dev Med Child Neurol 1997;39:343±346. Newborns are capable of using prosodic information to 5. Hepper PG. Fetal habituation: another Pandora's box? Dev Med Child Neurol recognize the maternal voice34,49 and to discriminate between 1997;39:274±278.

Journal of Perinatology 2000; 20:S36 ± S43 S41 Moon and Fifer Transnatal Auditory Learning

6. Zelazo PR, Weiss MJS, Tarquinio N. Habituation and recovery of neonatal 29. DeCasper AJ, Carstens AA. Contingencies of stimulation: effects on learning orienting to auditory stimuli. In: Weiss MJS, Zelazo PR, editors. Newborn and emotion in neonates. Infant Behav Dev 1981;4:19±35. Attention: Biological Constraints and the Effect of Experience. Norwood, NJ: 30. DeCasper AJ, Fifer WP. Of human bonding: newborns prefer their mothers' Ablex; 1991. pp. 120±141. voices. Science 1980;208:1174±1176. 7. Shahidullah S, Hepper PG. The developmental origins of fetal 31. DeCasper AJ, Prescott PA. Human newborns' perception of male voices: responsiveness to an acoustic stimulus. J Reprod Infant Psychol preference, discrimination, and reinforcing value. Dev Psychobiol 1993;11(3):135±142.8. 1984;17(5):481±491. 8. Shahidullah S, Hepper PG. Frequency discrimination by the fetus. Early Hum 32. DeCasper AJ, Spence MJ. Prenatal maternal speech influences newborns' Dev 1994;36:13±26. perception of speech sounds. Infant Behav Dev 1986;9:133±150. 9. Lecanuet J-P, Granier-Deferre C, Busnel M-C. Differential fetal auditory 33. Spence MJ, DeCasper AJ. Prenatal experience with low-frequency maternal reactiveness as a function of stimulus characteristic and state. Semin voice sounds influences neonatal perception of maternal voice samples. Perinatol 1989;13(5):421±429. Infant Behav Dev 1987;16:133±142. 10. Lecanuet J-P, Granier-Deferre C, Jacquet A-Y, Capponi I, Ledru L. Prenatal 34. Spence MJ, Freeman MS. Newborn infants prefer the maternal low-pass discrimination of a male and a female voice uttering the same sentence. filtered voice, but not the maternal whispered voice. Infant Behav Dev Early Dev Parent 1993;2(4):217±228. 1996;19(2):199±212. 11. Lecanuet J-P, Schaal B. Fetal sensory competencies. Eur J Obstet Gynecol 35. Moon C, Bever TG, Fifer WP. Canonical and non-canonical syllable Reprod Biol 1996;68:1±23. discrimination by 2-day-old infants. J Child Lang 1992;19(1):1±17. 12. Hepper PG, Shahidullah S. The beginnings of mind Ð evidence from the 36. Moon C, Fifer WP. Syllables as signals for 2-day-old infants. Infant Behav behaviour of the fetus. J Reprod Infant Psychol 1994;12:143±154. Dev 1990;13:377±390. 13. DeCasper AJ, Fifer WP. Of human bonding: newborns prefer their mothers' 37. DeCasper AJ, Spence MJ. Auditory-mediated behavior during the perinatal voices. Science 1980;208:1174±1176. period: a cognitive view. In: Weiss MJS, Zelazo PR, editors. Newborn 14. DeCasper AJ, Sigafoos AD. The intrauterine heartbeat: a potent reinforcer for Attention: Biological Constraints and the Influence of Experience. Norwood, newborns. Infant Behav Dev 1983;6:19±25. NJ: Ablex; 1991. pp. 142±176. 15. Fifer WP, Moon C. Auditory experience in the fetus. In: Smotherman WP, 38. Bower TGR. Contingencies, logic and learning. Behav Anal 1997;20:141± Robinson SR, editors. Behavior of the Fetus. Caldwell, NJ: Telford; 1988. 148. pp. 175±188. 39. Salk, L. The role of the heartbeat in the relations between mother and infant. 16. Hepper PG, Scott D, Shahidullah S. Newborn and fetal response to maternal Sci Am 1973;228:24±29. voice. J Reprod Infant Psychol 1993;11:147±155. 40. Salk, L. Mothers' heartbeat as an imprinting stimulus. Trans N Y Acad Sci 17. DeCasper AJ, Lecanuet J-P, Busnel M-C, Granier-Deferre C, Maugeais R. 1962;24:753±763. Fetal reactions to recurrent maternal speech. Infant Behav Dev 41. Querleu D, Renard X, Boutteville C, Crepin G. Hearing by the human fetus? 1994;17:159±164. Semin Perinatol 1989;13(5):409±420. 18. Thoman EB, Ingersoll EW. Learning in premature infants. Dev Psychol 42. Gerhardt KJ. Characteristics of the fetal sheep sound environment. Semin 1993;29(4):692±700. Perinatol 1989;13(5):362±370. 19. Zelazo PR, Brody L, Chaika H. Neonatal habituation and dishabituation of 43. Damstra-Wijmenga SMI. The memory of the newborn baby. Midwives Chron headturning to rattle sounds. Infant Behav Dev 1984;7:311±321. Nurs Notes 1991;104(1238):66±68 (March). 20. Zelazo PR, Weiss MJ, Randolph M, Swain IU, Moore DS. The effects of delay 44. Murooka H. Lullaby from the womb. (Audio recording.) Hollywood, CA: on neonatal retention of habituated headturning. Infant Behav Dev Capitol Records. 1974. 1987;10:417±434. 45. Rosner BS, Doherty NE. The response of neonates to intra-uterine sounds. 21. Bertoncini J, Bijeljac-Babic R, Blumstein SE, Mehler J. Discrimination in Dev Med Child Neurol 1979;21:723±729. neonates of very short CVs. J Acoust Soc Am 1987;82(1):31±37. 46. Tulloch JD, Brown BS, Jacobs HL, Prugh DG, Greene WA. Normal heartbeat 22. Bertoncini J, Bijeljac-Babic R, Kennedy LJ, Jusczyk PW, Mehler J. An sounds and the behavior of human infants: a replication study. Psychosom investigation of young infants' perceptual representations of speech sounds. J Med 1964;26:661±670. Exp Psychol Gen 1988;117(1):21±33. 47. Detterman DK. The effect of heartbeat sound on neonatal crying. Infant 23. Keen R, Chase H, Graham FK. Twenty-four-hour retention by neonates of Behav Dev 1978;1:36±48. an habituated heart rate response. Psychon Sci 1965;2:265±266. 48. Gerber SE. Stimulus, response and state variables in the testing of neonates. 24. Swain IU, Zelazo PR, Clifton RK. Newborn infants' memory for speech Ear Hear 1985;6(1):15±19. sounds retained over 24 hours. Dev Psychol 1993;29(2):312±323. 49. Fifer WP, Moon C. The effects of fetal experience with sound. In: Lecanuet J- 25. Blass EM, Ganchrow JR, Steiner JE. Classical conditioning in newborn P, Krasnegor NA, Fifer WP, Smotherman WP, editors. Fetal Development: A humans 2±48 hours of age. Infant Behav Dev 1984;7:223±235. Psychobiological Perspective. New York: Erlbaum; 1995. pp. 351±366. 26. Siqueland ER, Lipsitt LP. Conditioned headturning in human newborns. J 50. Fifer, WP, Moon C. Psychobiology of newborn auditory preferences. Semin Exp Child Psychol 1966;3:356±376. Perinatol 1989;13(5):430±433. 27. Lipsitt LP, Kaye H, Bosack TN. Enhancement of neonatal sucking through 51. Querleu D, Lefebvre C, Titran M, Renard X, Morillion M, Crepin G. Reactivite reinforcement. J Exp Child Psychol 1966;4:163±168. du nouveau-ne de moins de deux heures de vie a la voix maternelle. J 28. Butterfield EC, Siperstein GN. Influence of contingent auditory stimulation Gynecol Obstet Biol Reprod 1984;13:125±134. upon non-nutritional suckle. In: Bosma J, editor. Oral Sensation and 52. Fernald A, Simon T. Expanded intonation contours in mothers' speech to Perception: The Mouth of the Infant. Springfield, IL: Charles C. Thomas; 1972. newborns. Dev Psychol 1984;20(1):104±113.

S42 Journal of Perinatology 2000; 20:S36 ± S43 Transnatal Auditory Learning Moon and Fifer

53. Fernald A, Taeschner T, Dunn J, Papousek M, de Boysson-Bardies B, Fukui I. 72. Logan, B. Infant outcomes of a prenatal stimulation pilot study. Pre- A cross-language study of prosodic modifications in mothers' and fathers' Perinat Psychol 1991;6(1):7±31. speech to preverbal infants. J Child Lang 1989; 16: 477±501. 73. Anand, KJ, Scalzo FM. Can adverse neonatal experiences alter brain 54. Cooper RP, Aslin RN. Preference for infant-directed speech during the first development and subsequent behavior? Biol Neonate 2000;77(2):69± month after birth. Child Dev 1990;61:1584±1595. 82. 55. Ockleford EM, Vince MA, Layton C, Reader MR. Responses of neonates to 74. Rosada Montemurro RN. Singing lullabies to unborn children: experiences ' and others' voices. Early Hum Dev 1988;18:27±36. in Village Vilamarxant, Spain. Pre-Perinat Psychol 1996;11(1):9±16. 56. Ward CD, Cooper RP. A lack of evidence in 4-month-old human infants for 75. Olds C. A sound start in life. Pre-Perinat Psychol 1986;1(1):82±85. paternal voice preference. Dev Psychobiol 1999;35:49±59. 76. Querleu D, Renard X, Versyp F, Paris-Delrue L, Crepin G. Fetal hearing. Eur 57. Mehler J, Jusczyk P, Lambertz G, Halsted N, Bertoncini J, Amiel-Tison C. A J Obstet Gynecol Reprod Biol 1988;29:191±212. precursor of language acquisition in young infants. Cognition 1988;29:143± 77. Griffiths SK, Brown WS, Gerhardt KJ, Abrams RM, Morris RJ. The perception 178. of speech sounds recorded within the of a pregnant sheep. J Acoust Soc 58. Moon, C, Cooper RP, Fifer WP. Two-day-olds prefer their native language. Am 1994;96:2055±2063. Infant Behav Dev 1993;16(4):495±500. 78. Lecanuet J-P, Granier-Deferre C. Speech stimuli in the fetal environment. In: 59. Nazzi T, Bertoncini J, Mehler J. Language discrimination by newborns: de Boysson-Bardies B, de Schonon S, Jusczyk P, MacNeilage P, Morton J, toward an understanding of the role of rhythm. J Exp Psychol Hum Percept editors. Developmental Neurocognition: Speech and Face Processing in the Perform 1998;24(3):1±11. First Year of Life. Dordrecht: Kluwer Academic; 1993. 60. Richards DS, Frentzen B, Gerhardt KJ, McCann ME, Abrams RM. Sound levels 79. Butterfield EC, Cairns GF. Whether infants perceive linguistically is uncertain, in the human uterus. Obstet Gynecol 1992;80(2):186±190. and if they did its practical importance would be equivocal. In: Scheifelbush 61. Panneton RK. Prenatal Experience with Melodies: Effects on Postnatal RL, Lloyd LL, editors. Language Perspectives: Acquisition, Retardation and Auditory Preference in Human Newborns. Greensboro, NC: Univ. of NC Intervention. Baltimore: University Park Press; 1974. Greensboro; 1985. Thesis. 80. Colombo J, Bundy RS. A method for the measurement of infant auditory 62. Satt BJ. An Investigation Into the Acoustical Induction of Intrauterine selectivity. Infant Behav Dev 1981;4:219±223. Learning. Los Angeles, CA: Calif. Sch. Prof. Psychol.; 1984. Thesis. 81. Glenn SM, Cunningham CC, Joyce PF. A study of auditory preferences in 63. Polverini-Rey RA. Intrauterine Musical Learning: The Soothing Effect on non-handicapped infants and infants with Down's syndrome. Child Dev Newborns of a Lullaby Learned Prenatally. Los Angeles, CA: Calif. Sch. Prof. 1981;52:1303±1307. Psychol.; 1992. Thesis. 82. Jusczyk PW. The Discovery of Spoken Language. Cambridge: MIT; 1997. 64. Hepper PG. An examination of fetal learning before and after birth. Ir J 83. Clarkson MG, Berg WK. Cardiac deceleration and vowel discrimination in Psychol 1991;12(2):95±107. newborns: crucial parameters of acoustic stimuli. Child Dev 1983;54:162± 65. Trilogy Teaching Tapes. Writeside Publishing. Available at URL: http:/ / 171. www.willamette.net/~writeside/trilogy.html (cited December 1, 1999). 84. Weir CG, Lamb J. Differential sensitivity of newborns to synthetic vowel 66. Fetal Phone. Chrischild, Inc. Available at URL: http:/ /www.fetalphone.com stimuli. Br J Audiol 1990;24:419±430. (cited December 1, 1999). 85. Nazzi T, Floccia C, Beroncini J. Discrimination of pitch contours by neonates. 67. Van de Carr R, Lehrer M. Enhancing early speech, parental bonding and Infant Behav Dev 1998;21(4):779±784. infant physical development using prenatal intervention in standard obstetric 86. Sansavini A, Bertoncini J, Giovanelli G. Newborns discriminate the rhythm of practice. Pre-Perinat Psychol 1986;1(1):20±30. multisyllabic stressed words. Dev Psychol 1997;33(1):3±11. 68. Van de Carr R, Lehrer M. Prenatal University: commitment to fetal±family 87. Fernald A. Intonation and communicative intent in mothers' speech to bonding and the strengthening of the family unit as an educational infants: is the melody the message? Child Dev 1989;60:1497±1510. institution. Pre-Perinat Psychol 1988;3(2):87±102. 88. Gleitman L, Wanner E. The state of the state of the art. In: Wanner E, 69. Shetler DJ. The inquiry into prenatal musical experience: a report of the Gleitman L, editors. Language Acquisition, The State of the Art. New York: Eastman project 1980±1987. Pre-Perinat Psychol 1989;3(3):171±189. Cambridge; 1982. 70. Logan B. Teaching the unborn: precept and practice. Pre-Perinat Psychol 89. Morgan JL, Demuth K. Signal to Syntax: Bootstrapping From Speech to 1987;2(1):9±24. Grammar in Early Acquisition. Mahwah, NJ: Erlbaum; 1996. 71. Ikonomidou C, Bosch F, Miksa M, Bittigau P, Vockler J, Dikranian K, et al. 90. Lind J. Music and the small human being. Acta Paediatr Scand Blockade of NMDA receptors and apoptotic neurodegeneration in the 1980;69:131±136. developing brain. Science 1999;283(1):70±73.

Journal of Perinatology 2000; 20:S36 ± S43 S43