J Am Acad Audiol 6 : 113-123 (1995)

Development of Hearing. Part III. Postnatal Development James E. Peck*

Abstract

As humans, we hear the way we do because of at least three major forces . The first is phy-logeny, the evolutionary changes in the auditory system since its beginnings . The second is embryology, the formation of the system in each individual . Finally, there is the interaction between the biologically determined mechanism we are born with and the environment. This series of articles reviews each aspect in turn, so that we may have a fuller appreciation of how it is we come to hear the way we do . Part I examined how the auditory system evolved, and Part II outlined the sequence of ear formation in prenatal life . In this concluding article, we examine the development of auditory perception from when we first hear (in utero) through 12 months of age, which is when speech usually appears. Humans seem to come "pro-grammed" to perform all the processes necessary to learn the language of their environment . Interaction with the environment hones our abilities to process sound and thereby optimizes language acquisition and production .

Key Words: Auditory development, auditory discrimination, auditory localization, auditory perception, experience, feature detectors

T

his article is the concluding segment to this three-part series on the development of hearing. In the first part, the evolution

of the hearing apparatus was traced from the ori-gins of life to the large-brained humans (Peck, 1994a) . The second outlined the embryonic devel-opment of the various parts of the auditory mech-anism (Peck, 1994b) . This third part focuses on the development of hearing from the earliest auditory experience to the onset of speech .

Earliest Auditory Experience . This article necessarily overlaps with the preceding one on embryology because, in all likelihood, we experience sound before we are born . Indeed, the fascinating question arises : when do we first hear? As we saw in Part II, all auditory mech-anisms seem ready at 6 to 7 months of gestation. What is the final event that allows us to hear? Is it mechanical, neural, chemical? Various struc-tures have at one time or another been assigned the role of "trigger" of onset of cochlear function. Closer to the truth seems to be a "simultaneous

'Communicative Disorders Laboratory, University of Mississippi Medical Center, Jackson, Mississippi

Reprint requests : James E. Peck, Communicative Dis-orders Laboratory, University of Mississippi Medical Center, 250 North State Street, Jackson, MS 39216

and synchronous maturation of many mechan-ical and neural properties" (Rubel, 1985, p. 57). Moreover, various factors influence one another for most of pregnancy (see Rubel, 1978, 1985). For example, certain central auditory ganglia, deprived of input from the peripheral receptors, will not develop fully (Rubel, 1985). In short, there is probably no one final cause.

Clinical observations suggest that some very premature infants born at 25 weeks react to sound. Support for these observations comes from high-resolution ultrasound imaging. Birn-holz and Benacerraf (1983) studied response to vibroacoustic stimuli in human fetuses by ultra-sound. Blink responses were first detectable at between 24 and 25 weeks and were consistent at 25 weeks . One may speculate, based on such observations and our knowledge of embryology, that fetuses are capable of experiencing sound by the seventh month of intrauterine life .

An equally intriguing question is what do we first hear? Since the baby's sound environment consists of fluid insulated by layers of tissues, outside sounds are less intense than inside ones . Sound levels inside the womb average 85 dB SPL (Walker et al, 1971). Regardless of origin, low fre-quencies travel more efficiently than high fre-quencies through such dense media, so that sounds in the 20- to 200-Hz range predominate

Journal of the American Academy of Audiology/Volume 6, Number 2, March 1995

(Walker et al, 1971). Even though external sounds are attenuated, specially above 800 Hz, nearby voices or loud sounds may reach the fetus (Armitage et al, 1980).

To study this matter, Armitage et al (1980) measured sound inside sheep . Hydrophones were placed inside intact amniotic sacs of preg-nant ewes . A loudspeaker was strapped to but did not touch the ewe's skin . Comparing exter-nal and internal levels showed attenuation rang-ing from 16 to 37 dB with peak attenuation occurring just below 1000 Hz, although gener-ally high frequencies were attenuated more than low frequencies. Further, the examiners could hear internal sounds, such as the mother's eat-ing, drinking, ruminating, breathing, and mus-cular movement. They could also hear, but not always understand, outside conversation . Sur-prisingly, heart beat and blood flow were not per-ceptible . Despite this information, one must keep in mind that the sound spectrum inside a uterus may be different from that which arrives at a fetus's inner ear due to altered transmission characteristics of an outer and middle ear filled with liquid (Rubel, 1985). What such signals "sound like" to the fetus may not be how they would sound to a person listening by air con-duction via earphones or a loudspeaker. Rather, fetal sound experience may be more akin to bone-conduction hearing.

In sum, the sounds we hear first are prob-ably those that are strongest during the entire pregnancy (i .e ., low-frequency components of internal sounds, principally the mother's vocal-izations and physiologic noise) . Fetal sound experience in general may help fine tune the developing auditory system (Fifer and Moon, 1988). More specifically, prenatal hearing of maternal voice may influence postnatal speech perception, an issue discussed below (Aslin et al, 1983 ; Fifer and Moon, 1988).

Postnatal Hearing. With birth comes a major shift from fluid- and tissue-conducted sound to air-conducted sound (somewhat anal-ogous to amphibians shifting from an aquatic to a terrestrial environment, as we saw in Part I [Peck, 1994a]). The infant is suddenly exposed to a very different sound spectrum . Also with birth, the organism is much more accessible for investigation, both by electrophysiologic and behavioral techniques .

Sensitivity. Perhaps the most basic audi-tory function is the ability to detect acoustic energy ; how well it is done is described as sen-

sitivity. Among the electrophysiologic techniques to study sensitivity in infants, the most often uti-lized has been auditory brainstem response (ABR) audiometry. ABR results have shown that babies' thresholds for clicks are about 17 dB higher than those of adults (Hecox, 1975 ; Schul-man-Galambos and Galambos, 1979) . Also, the latency of the infant's brainstem response decreases from before birth until the second year of life . This decreasing latency reflects a maturing auditory system, attributed to the process of myelination (Starr et al, 1977) .

It must be kept in mind, however, that the ABR is a neural response and that ABR testing is a measure of synchronous neural discharge . A synchronous neural response is not to be equated with hearing or with the ability of an organism to make meaningful interpretations about its acoustic environment (Rubel, 1978), a capacity without which a child cannot acquire auditory language .

Even more productive than electrophysio-logic means have been behavioral techniques for learning about what and how babies hear. Merely through observation, one sees obvious reflexive responses (usually "startle") to intense sounds in term babies and in many "preemies." Regardless of level, broad-band sounds are more likely than narrow-band sound to elicit a response from infants. Speech, a broad-band sound, has been shown to be more response elic-iting than a narrow band or discrete frequency, perhaps due to listening experience in utero (Eisenberg, 1976), as mentioned above in the dis-cussion of what we first hear .

Responses to intermediate sound levels are subtle or absent and are highly influenced by physiologic states of hunger and fatigue. While it is true that one can observe behavioral changes to very low-level signals, even in newborns, these changes have traditionally been considered unreliable .

Recently, however, the observer-based psy-choacoustic procedure (OPP) was developed for behavioral study of auditory ability of very young infants down to 3 months of age (Olsho et al, 1987). Under the OPP paradigm, no particular response is specified (e.g ., head turn). Rather, any behavioral change is accepted (eye widening, activity suppression, sucking changes) . Sounds are presented to the infant via loudspeakers . To control for observer bias, the observer wears earphones and hears all trials, whether gen-uine or no-signal control trials . Thus, the observer is ignorant of when signals are actually presented to the baby. The observer decides

114

Postnatal Hearing Development/Peck

whether a given behavior occurred as a result of a signal . If the observer is correct, visual rein-forcement is supplied to the baby for whatever the behavior was. Such reinforcement also helps the observer to become a more accurate judge of infant responses . Using OPP, infants in the 2-to 5-week age range showed thresholds at 500 and 4000 Hz of about 55 dB and 60 dB SPL, respectively (Werner and Gillenwater, 1990).

Visual reinforcement audiometry (VRA) has been tremendously successful with most babies above about 5 months of age. Using VRA, Moore and Wilson (1978) showed that infants 6- and 12-months of age have quite low sensitivity thresh-olds . They plotted sound field thresholds of 15 to 20 dB SPL at 500, 1000, and 4000 Hz for infants of 6 to 12 months of age.

While infant sensitivity is very good, it still is not quite as good as that of adults . Trehub et al (1980) compared infants and adults (using VRA) for six octave-band noises from 200 to 10,000 Hz in sound field. In general, 6-month-old infants' thresholds were 20 to 25 dB poorer than those of adults . By 12 months, the differ-ence was about 10 dB less .

The gap between child and adult sensitiv-ity continues to close with age but at different rates for different frequencies (see review in Trehub et al, 1988, 1989) . The most rapid improvement in sensitivity occurs in the region of 10,000 Hz, where adult levels are attained as early as 4 or 5 years of age. The next frequency region to reach maximal detectability is 20,000 Hz, which occurs by ages 6 to 8 years. Indeed, 7-year-old children have more sensitive hearing at 20,000 Hz than do adults; around age 10 years, children begin to "lose" hearing until adult levels are reached. For stimuli in the 2000-to 4000-Hz range, maximal sensitivity is achieved by 8 years, and finally the 400- to 1000-Hz range at 10 years of age.

The bases for these changes are not clear. Speculation has included mechanical changes in the outer, middle, or inner ear. For example, the smaller pinna and canal of children, as com-pared to adults, might enhance high-frequency detection. Also, earlier research suggested that children had wider critical bands than adults (Schneider and Trehub, 1981), but recent evi-dence shows that critical band width does not change with age (Schneider et al, 1990).

Localization . In addition to detecting soft sound, infants can localize sound. While all sound processing may have survival value, local-ization is probably one of the more basic auditory

abilities . Virtually all organisms that hear can localize sound in space (Erulkar, 1972). This harkens back to the first paper in this set of three, which referred to the survival value of hearing in vertebrates (Peck, 1994a). Hearing is a distance sense; thus, the ability to detect a sound is even more helpful when the organism can tell where the sound is coming from (consider hunting or escape). A great deal has been learned about human infants' abilities to resolve audi-tory space.

As with most auditory abilities, resolution of auditory space improves with age. Improve-ment in localization accuracy among infants occurs first in the horizontal plane and then in the vertical plane. Newborns make head turns to the right or left of midline in an orienting response to sound (Clifton et al, 1981). The new-born places sounds in a long narrow band in front and just above horizontal with almost no verti-cal dimension. This unidimensional region "dis-appears" around 2 months and "reappears" in two dimensions at about age 41/z months with a wider horizontal and some vertical aspect . Also, infant response becomes brisker. For example, head-turn latency is 7 sec in newborns and decreases to about 2 sec by 5 months of age (Morrongiello and Clifton, 1984). These changes in localization are interpreted as a shift from reflexive behavior (subcortical) at birth to vol-untary/controlled behavior (cortical) after the neonatal period (see review in Morrongiello and Gotowiec, 1990).

At 8 weeks of age, the smallest angle of change that can be detected (minimum audible angle or MAA) is about 27 degrees. Performance improves systematically, so that by 6 months, infants' MAA is about 14 degrees (Morrongiello et al, 1990). After 6 months of age, improve-ment in localization is slower but still system-atic . The MAA is 8 degrees at 12 months and only 4 degrees at 18 months (Morrongiello, 1988b) . (The difference is not due to limitation of head control; infants at 6 months can visually locate a light to within about 5 degrees [Morrongiello and Rocca, 1987a]).

Throughout life, our sense of auditory space is most keen directly in front. In other words, we can detect smaller right-left distinctions in front than changes in sound location within the hemi-field on the right or left sides. In infancy, the MAA for sounds referenced to 60 degrees to the side is about 5 degrees poorer than for noting right-left shifts directly in front (Morrongiello and Rocca, 1990). As with other skills, localiz-ing within hemifields improves with age. At 6

Journal of the American Academy of Audiology/Volume 6, Number 2, March 1995

months of age, the MAA is 19 degrees, while at 12 months, it is 15 degrees (Morrongiello and Rocca, 1990).

One of the two main cues used in horizon-tal localization is time of arrival (the other being relative intensity at the two ears). When a sound is on one side, stimulation of the near ear pre-cedes, by a fraction of a second, stimulation of the far ear. Auditory systems give precedence to the first ear stimulated and identify the sound as coming from the near-ear side . The effect is greatest in the low frequencies. Even if the first sound is followed by reverberations of this wave-form, the auditory system regards only the first wavefront in deciding about sound location . An auditory illusion of sidedness, known as the precedence effect, can be created by producing a sound through two separated loudspeakers with very small differences in time of onset. A listener experiences the sound as coming from the side of the first loudspeaker.

The illusion of location from the precedence effect is strongly age related and can be exploited for studying infants' processing of time cues for localizing . If a single-source sound occurs on a particular side, neonates can orient to that sound as well as 5-month-old infants. However, if a sound is presented through two separated loud-speakers with a 7-msec delay at one loudspeaker, neonates fail to orient, while 5-month-old babies can orient to the lead side (Clifton, 1985). By 6 months, infants consistently turn toward the side of the leading loudspeaker (Clifton et al, 1984). In turn, preschool children can localize smaller delays than infants, and adults can localize still smaller delays than preschoolers . Investigators speculate that improving prece-dence perception is related to progressing myeli-nation of the cortex (see literature review in Morrongiello and Gotowiec, 1990).

As mentioned above, infants resolve audi-tory space more accurately in the horizontal than in the vertical plane. For example, at 6 and 12 months of age, infants have a larger angle of error by 1 or 2 degrees for vertical versus hor-izontal tasks (Morrongiello and Rocca, 1987b) . This implies that there are different cues for sound location . The explanation seems to involve high-frequency reception and pinna shape. The sense of verticality is strongly influ-enced by high frequencies, and the pinnae strongly influence high-frequency reception (Morrongiello and Gotowiec, 1990). Perhaps the changing size and shape of the child's pin-nae help resolve the high-frequency cues of everyday sounds .

In short, at first, the infant can respond to gross laterality (right-left) changes ; later, the infant is increasingly accurate in placing sound in space (Morrongiello and Rocca, 1987a; Mor-rongiello, 1988a) .

Localization is not purely auditory; it sub-serves visual searching and finding (Morrongiello and Gotowiec, 1990) . After 2 months, infants begin to integrate auditory and visual space (e.g ., the mother's voice and her facial move-ments coinciding in a unified space) (Muir, 1985). Also, Muir (1985) observed that a blind infant was less likely than a sighted child to localize at 0 to 6 months of age. It seems, then, that local-ization is an integrated, bimodal function with audition supporting vision in locating the source of a sound.

The phenomenon of localization holds rel-evance for the clinical audiologist in assessing hearing of very young children. Generally, such children do not tolerate earphones well, so their hearing is assessed in a sound field. The goal is to determine whether or not there is at least one normally hearing ear, since having a hearing loss in both ears will impair commu-nication development. However, it is now clear that a long-standing hearing loss in even just one ear can have serious consequences, prompt-ing the need to identify unilateral hearing deficits (Bess and Tharpe, 1986 ; Oyler et al, 1988). Testing an infant's ability to localize sound can be helpful in suggesting the presence or absence of a unilateral hearing loss (Mor-rongiello, 1989 ; Auslander et al, 1991) . See Figure 1 for examples of localization behav-iors one might see in informal clinical assess-ment of infants.

Recognition. It seems a universal belief that very young babies recognize and prefer their own mother's voice. Data support this popular belief. Neonates no older than 3 days have been taught that if they produced a par-ticular sucking pattern, they could listen to their own mother's voice. Even when the crite-rion to hear one's mother's voice was reversed, the neonates still "worked" to hear their mother's voice rather than the voices of the other mothers (DeCasper and Fifer, 1980). More-over, newborns are able to distinguish between their mother speaking in her native tongue ver-sus speaking in an unfamiliar language, even when the signals have been low-pass filtered (Mehler et al, 1988). Perhaps babies are sensi-tized prenatally to respond to their mother's voice (DeCasper and Fifer, 1980).

116

Postnatal Hearing Development/Peck

Newborn period to 4 too. Normal infant is aroused from sleep by sound signals of 90 dB (SPL) in a noisy environment, 50-70 dB (SPL) in quiet.

4to7mo. Baby turns head directly toward the side of a signal 40-50 d0 (SPL) but cannot find it above or below.

7 to 9 too. Baby directly locates

a sound source of 30-40 dB (SPL) to the side and

indirectly below.

A

9 to 13 ma . Baby directly locates a sound source of 25-35 dB (SPL) to the side and below.

16to21mo. j n) ~)/r O~ 21to24mo. Toddler localizes Child locates directly sound / ) directly a signals of sound signal of 25-30 d0 (SPL) 25 dB (SPL) on the side, below ~t at all angles. and above.

Figure 1 Illustrations of typical development of sound localization in infants. Left panel shows infant responses from newborn to 9 months . Right panel shows infant responses from 9 to 24 months . (Adapted with permission from North-ern and Downs, 1991 .)

Not only do babies recognize their own mother's voice, but they also recognize and pre-fer certain kinds of vocalizations. Interestingly, the vocal patterns they prefer are the sorts grown-ups use in speaking to babies (Aslin et al, 1983; Fernald, 1992). Indeed, infant responses probably reinforce our adult behavior.

Universally, when adults speak to infants, they use slow tempo, high pitch, and marked intonation changes. Adults seem to have an almost irresistible urge to address infants in this specific way. This speech style is called "motherese," even though it is used by males and females, parents and nonparents . Adults spon-taneously use "motherese" because its features are highly salient auditory stimuli for recruit-ing infants' attention (Fernald, 1984, 1992).

It is clear that young babies can discern the acoustic parameters that typify infant-directed speech . Spring and Dale (1977) demonstrated that young infants in the 1- to 4-month age range can detect acoustic cues that serve to mark intonation (e .g ., changes in fundamental frequency, intensity, and duration of voice) . The basis for these phenomena is undeter-mined, but speculation includes intrauterine

exposure to such sounds and the attention-get-ting property of variable sounds . Apparently, infants arrive ready to be responsive to those acoustic characteristics of the "motherese" sig-nal (Fernald, 1992) .

"Motherese" is not merely a social interac-tion to delight adults and babies alike. Rather, it plays an important role in caregiver-infant bonding. From the outset, mothers use their voice to arouse their babies, to attract and main-tain their attention, and to soothe and warn them . In short, a mother's vocalizations com-municate affect, that is, her feelings and mood . Later in infancy, as a result of experience, the vocalizations take on additional meaning as "signs" of events around them. Finally, the infant attributes specific significance to speech and arrives at linguistic meaning. Concomitantly, the infant uses sound initially to convey affect and ultimately to express lexical meaning through words (Fernald, 1984).

Summarizing this section, there is a mutual interaction at work here . On the one hand, adults spontaneously speak to infants in a highly stylized way. On the other, infants prefer those vocal patterns.

Journal of the American Academy of Audiology/Volume 6, Number 2, March 1995

Why should humans engage in such elabo-rate and ritualistic vocal-auditory behavior? Taking a highly Darwinian approach, Fernald (1992) argues persuasively that infant-directed speech is biologically determined through evo-lution by natural selection . The basis, she claims, is the lengthening period of dependency in pri-mates. Humans are the most neotenous of all, born the least developed and the most help-less . Prolonged immaturity among hominids favored certain parenting behaviors. Among them was adopting vocal behaviors that were suited to the offsprings' limited auditory per-ceptual systems, thereby promoting attention and bonding.

Effective, early mother-infant communi-cation had adaptive advantages for hominids extending far past infancy, reasons Fernald (1992) . The quality of attachment in infancy impacts on social competence in childhood and on emotional relationships and parenting in adulthood. Good nurturing produces a socially competent member of the group and an effec-tive parent . The nurtured becomes the nurturer. In keeping with natural selection, such behav-iors improve the chances of survival of both the individual and of the group and thereby tend to be passed along to subsequent members (Fernald, 1992).

Discrimination. It has been known for 30 years that not only can neonates respond to sound, but that they can also discriminate among different sounds (Papousek, 1961; Siqueland and Lipsitt, 1966). We saw above that neonates can discriminate between their mother's voice and the voices of other mothers . In a classic study of speech sound discrimination, Eimas et al (1971) showed that infants aged 1 to 4 months could distinguish between voiced and voiceless sounds of b/p and d/t . Infants were presented syn-thetic syllables, which varied in voice onset time (VOT), the main cue for voiced-voiceless dis-tinction . The high amplitude sucking paradigm was'used, based on the phenomenon that when infants are presented a novel stimulus, their sucking rate increases suddenly. Those infants listening to syllables with the same VOT did not alter sucking rate, whereas those infants pre-sented with syllables of differing VDTs consis-tently accelerated sucking rates.

Since the early investigations, a considerable body of research has emerged showing that infants in the first year of life have the ability to make virtually all of the auditory discrimi-nations necessary to learn language (see Kuhl,

1987). They have also shown that infants can dis-criminate other aspects of speech, such as time and pitch changes. For example, Olsho (1984) compared difference limens for frequency (DLF) of babies 5 to 8 months of age with those of adults . Tones were octave points 250 to 8000 Hz . For the infants, she conditioned head-turn responses to frequency changes. For both groups, relative frequency discrimination limens (change in frequency/frequency) at 2000 Hz and above were similar at about 1 percent (i .e ., detected a change of as few as 20 cycles per second in a 2000-Hz signal) . For frequencies below 2000 Hz, infants' DLF was about 1 to 3 percent larger than those of adults . For adults, the DLFs at 250, 500, and 1000 Hz were about 2 percent, 1.5 per-cent, and 0.4 percent, respectively. In contrast, infants' DLFs were approximately 3.5 percent, 2 .5 percent, and 1 percent for the same tones. As other examples of discrimination ability, 6-month-old infants can detect small temporal changes on the order of 20 msec (Morrongiello and Trehub, 1987), and 9- to 11-month-old infants can detect changes in melodic contour in different contexts (Trhub et al, 1987).

Being able to detect time and pitch changes (i .e ., suprasegmental features) enhances speech discrimination . Karzon (1985) studied the abil-ity of 1- to 4-month-old infants to discriminate between three-syllable utterances contrasted by one phoneme. Vocal stress was varied by intensity, duration, and fundamental frequency. Karzon (1985) found increased ability to dis-criminate when the target syllable was exag-gerated in the highly intonated manner in which speech is directed to infants. Perhaps prenatal listening to low-frequency sound heightened their response to prosody.

Categorization. Languages have cate-gories of speech sounds called phonemes . The ability to perceive speech sounds as belonging to categories denotes the ability to detect a dif-ference between stimuli, even if varying in only one dimension (see Kuhl, 1987). An example is noting the difference in VOT that distinguishes /p/ from /b/. Numerous studies since the early 1970s demonstrate that infants group speech sounds into phonetic categories in adult-like fashion (see Kuhl, 1987).

However, categorizing speech sounds re-quires not only noting differences but also sim-ilarities. This means that the listener must treat sounds that are discriminatively different as equivalent, a phenomenon known as equiva-lence classification . Equivalence classification

118

Postnatal Hearing Development/Peck

involves ignoring irrelevant differences and rec-ognizing those cues that make sounds alike. Infants show remarkable ability to make dif-ferent-but-similar judgments under varying con-ditions. For example, they can recognize that a speech sound belongs to a particular phonemic category, although it appears in different contexts (such as initial and final word positions), in dif-ferent intonation contours, when spoken by per-sons of different ages, gender, or dialect, or when uttered at different rates (see reviews by Aslin et al, 1983 and Kuhl, 1987). "By 6 months of age, infants appear to be ̀natural' categorizers" (Kuhl, 1987, p . 351) .

Innate Abilities and Special Mechanisms . With infants exhibiting such extensive speech "processing" abilities, investigators, like Eimas et al (1971), hypothesized that human infants were born with special auditory processors, which he dubbed feature detectors. In other words, it was felt that babies came ready to make linguistically relevant auditory discrimi-nations even before they had been exposed to them . It was as if babies' auditory systems had a "speech mode" for listening to speech . Two related hypotheses surfaced that this speech mode was species specific-only humans had it-and that it was speech specific (i.e., the abil-ity evolved specifically to perceive speech). How-ever, research in the 1970s called into question the speech-specific idea that humans were pro-grammed to process speech sounds . Studies showed that adults made categorical distinc-tions between nonspeech sounds much as they did with speech sounds (Cutting and Rosner, 1974; Pisoni, 1977). (Here, the term nonspeech sounds means sounds that mimic some property or properties of speech without being identified as speech .) Later studies involving infants found that they, also, could make fine discriminations between nonspeech sounds (Jusczyk et al, 1977, 1980; Clarkson and Clifton, 1985).

By the late 1970s, the species-specific notion that we humans are the only species capable of such exacting auditory behavior had also been seriously weakened . Reports appeared of vari-ous mammals making many categorical speech sound distinctions (Kuhl, 1978). With both the speech-specific and species-specific notions in doubt, Kuhl (1986) saw no need to invoke "spe-cial mechanisms" to explain categorical per-ception or speech sound discrimination . On the contrary, she hypothesized that as speech evolved, certain "phonetic" sounds were selec-ted for production because they possessed

characteristics that the auditory system could detect easily. Thus, Kuhl argued, infant speech perception was based on certain properties of the mammalian auditory system, which the human baby draws upon in processing speech .

Recent evidence from Kuhl herself suggests that, indeed, there may be a degree of species-specific speech processing ability after all . Kuhl (1991) demonstrated that some speech sounds are more typical of a given speech sound cate-gory than others . In other words, there is an ideal sound for each category, or a "prototype ." This typicality strongly influences speech perception . Kuhl found, for both human adults and 6-month-old infants, that the closer a sound is to the pro-totype, the easier it is to recognize it as a mem-ber of that category. However, in testing mon-keys, Kuhl found that they did not show a dif-ference in discriminating speech sounds based on prototype or nonprototype sounds . She con-cluded that human adults and infants have an auditory image of what constitutes the center of a speech sound category.

How humans come to have this internal auditory map of categories is uncertain. Kuhl (1991) suggests two possibilities . First, infants may be biologically endowed with these images of prototypes . Alternatively, experience may have a strong determining force, even by 6 months of age. Finally, it could be that there is a mix of these two factors. Whichever the case, the critical point here is that there seems to be at least some species-specific ability to process speech. In the larger context of this series of papers, the relevance of the preceding para-graphs is that our remarkable perceptual abil-ities, however special or general, facilitate the acquisition of language .

Experience. As impressive as our inborn perceptual capabilities may be, they are not enough for postnatal development of all auditory skills and speech production . Experience seems critical in refining and maintaining speech per-ception. As an obvious example, profound con-genital deafness, without intervention, imposes a nearly insurmountable difficulty in acquiring aural-oral communication. The biologic systems underlying auditory perception cannot fulfill their roles unless sound is at least perceptible . As we have seen, experience begins prenatally. Indeed, experience is critical for proper neuro-genesis; it finetunes our perceptual-motor net-works to those characteristics of human voice that are available to the fetus (Werker and Tees, 1992). Prenatal experience shapes postnatal

Journal of the American Academy of Audiology/Volume 6, Number 2, March 1995

sound preference and enhances the ability to process sound in general and speech in partic-ular. In other terms, speech perception has an endogenous basis with an experiential overlay (Werker and Tees, 1992).

Despite general agreement that interaction with our environment is important in commu-nication development, the way and degree to which it is important are not at all clear. Views run the range of extremes from experience being absolutely crucial to being completely unneces-sary for each aspect of development. Perhaps the role of environment is multidimensional and involves several parallel processes at work simul-taneously (Aslin et al, 1983 ; Werker, 1989). Fig-ure 2 shows five hypothetical roles of experience on development (see Werker, 1989).

In maturation, the first process shown, the environment has no role at all in development. Accordingly, certain perceptual skills emerge "on their own" regardless of experience . In a second process, perceptual skills are completely ready at birth, but experience is necessary to maintain the perceptual skills . Some would offer the above case of deafness as an example of the maintenance/loss process. It implies that skills are fully developed but will not be maintained unless sounds of language are experienced by the individual . Next, experience facilitates percep-tual development. Facilitation does not alter the final level of ability but rather it speeds the rate of reaching that level. Fourth, experience attunes the perceptual mechanisms to become sharper in their auditory discriminations. Finally, experience induces the skills . In the induction hypothesis, skills are latent but unde-veloped, and their appearance is wholly depen-dent upon environmental exposure . This process is similar to imprinting . A well-known example of imprinting is ducklings developing a strong and instant recognition of the first sound they hear-usually their mother's call-to the exclu-sion of later-heard animal calls.

In at least one aspect, experience can seem, paradoxically, to diminish rather than aid audi-tory skills . Werker (1989) and her colleagues tested persons from infancy to adulthood in their ability to discriminate sound contrasts that occur in their language and sound con-trasts that occur in other languages. Not unex-pectedly, they found that adults can distinguish between phones used in their language but not between phones used in an unfamiliar language . In contrast, infants can distinguish between phones in a language they have never heard. It seems that infants are born with a universal

Maturation

Maintenance/Loss

Facilitation

Attunement

Induction

Onset of Experience Figure 2 An illustration of five hypothesized roles of postnatal experience on the development of perceptual skills . (Adapted with permission from Werker, 1989 .)

phonetic sensitivity, which fades between ages 6 and 12 months .

These initial findings had been interpreted as an example of maintenance/loss in experience . However, later investigations showed that, given considerable training, adults could regain some ability to make these distinctions (Werker, 1989 ; Werker and Tees, 1992) . In other words, their capacity to make those distinctions is not lost entirely.

Werker (1989) suggests that decreased pho-netic sensitivity at 10 to 12 months of age is not a decline in auditory skill but rather a language-based reorganization of phone categories . She sees a parallel in oral language develop-ment . Babies seem to be able to make all sounds of any language, but as soon as spoken lan-guage emerges, vocal productions narrow to approximate the sounds of that language . In other words, phonetic sensitivity gives way to language sensitivity. Werker (1989) considers these changes as perceptual and productive streamlining, by limiting perception and pro-duction to categories useful for communicating

120

Postnatal Hearing Development/Peck

in the language of one's environment and ignor-ing the rest .

Data on the overall role of experience in auditory development are limited. Part of the problem is the familiar "nature versus nurture" controversy. Resolving it in humans is difficult because it is not ethically possible to perform the studies, which would isolate the critical vari-ables. For the foreseeable future, we may need to be content with the notion that experience is crucial in the development of audition and com-munication, but to what extent is conjectural .

SUMAL4,RY

T his series of three articles has given an overview of the development of hearing.

The papers addressed in turn evolution, embry-ology, and postnatal development of hearing. While much has been learned, "the mechanisms underlying auditory development are largely unknown" (Aslin et al, 1983, p. 669) . Here is a brief synopsis of what is currently believed about auditory development. Our auditory system has evolved over a span of nearly 500 million years . The human ear's internal fluid-filled labyrinth has its origin in an equilibrial organ in fish . This early internal balance organ was adapted for hearing separately by fish, amphibia, and rep-tiles. Amphibia adapted their auditory organ for air-conduction hearing on land by coupling it to the outside world with a "middle ear" mech-anism. The middle ear plan, a large membrane mechanically coupled to the inner ear, became part of the genetic heritage of later vertebrates. Early reptiles evolved, perhaps in more than one line, an auditory labyrinth from an equi-librial one. The auditory organ of at least two rep-tilian species served as the basis for the audi-tory organ of later mammals and birds. Mam-mals developed two distinct types of auditory sensory cells, the outer and inner hair cells. Eventually, through primate and hominid devel-opment, humans came to possess an exquisite, large brain with which to process the neural codes from the peripheral receptor system .

In the case of individual humans, the embryo rapidly develops an intricate auditory system according to genetic instructions inherited over the ages . The inner ear has its beginnings at the end of the third gestational week, when cells in tissue from the outer layer of the embryo fold upon themselves, forming a fluid-filled cyst . By 10 weeks, the cyst has transformed into a cochlea of adult shape and by 5 months it attains adult size . Between 3 and 5 months, some of the cells

lining the cyst specialize to become a mature organ of Corti. The combined sensory and neural systems are ready to respond to sound early in the seventh month. The middle ear originates in the fourth week as outpouchings of the embry-onic pharynx. Ossicles form during months 2 through 4, and the middle ear space is open (but fluid filled) near the end of 7 months . Finally, the ear canal and outer ear originate in the fourth and fifth weeks, respectively. Tissue from six tiny bulges blend to form the pinna into its familiar shape by about 5 months . The ear canal deepens and widens and terminates in a thin tympanic membrane in the seventh month.

By birth, we humans possess biologic sys-tems, some ready and others soon to be ready, to process complex sounds in a linguistically relevant manner. Our biology is influenced by exposure to sound, particularly the sounds and content of language . A superb central nervous system detects, discriminates, recognizes, cate-gorizes, sequences, and stores sounds and, in a way surpassing that in all other species, extracts meaning from those sounds . Ultimately, near 12 months of age, we lay claim to perhaps our most human characteristic and begin to speak.

Acknowledgment . I am grateful to Drs. David Lip-scomb, Noel D. Matkin, and Robert Oyler for their edi-torial help in preparing this series of papers . Thanks to Drs. Barbara A. Morrongiello, Sandra E. Trehub, and Janet F. Werker for their copious and generous sugges-tions for this third paper. Particular appreciation is given to L. Judson Farmer, friend and colleague, who sug-gested I undertake this project and urged its completion .

REFERENCES

Armitage SE, Baldwin BA, Vince M . (1980) . The fetal sound environment of sheep . Science 208:1173-1174.

Aslin RN, Pisoni DB, Jusczyk PW. (1983). Auditory devel-opment and speech perception in infancy. In: Mussen PH, ed . Handbook of Child Psychology. 4th ed . Vol II . Haith MM, Campos JJ, eds. Infancy and Developmental Psychobiology. New York : John Wiley, 573-687.

Auslander MC, Lewis DE, Schulte L, Stelmachowicz PG . (1991) . Localization ability in infants with simulated uni-lateral hearing loss . Ear Hear 12:371-376 .

Bess FH, Tharpe AM. (1986) . Case history data on uni-laterally hearing-impaired children . Ear Hear 7:14-26 .

Birnholz JC, Benacerraf BR. (1983) . The development of human fetal hearing. Science 222:516-518 .

Clarkson MG, Clifton RK. (1985) . Infant pitch percep-tion : evidence for responding to pitch categories and the missing fundamental. J Acoust Soc Am 77:1521-1528 .

Clifton RK . (1985) . The precedence effect: its implica-tions for developmental questions. In : Trehub SE,

Journal of the American Academy of Audiology/Volume 6, Number 2, March 1995

Schneider BS, eds. Auditory Development in Infancy. New York : Plenum, 85-99.

Clifton RK, Morrongiello BA, Kulig JW, Dowd JM. (1981) . Newborns' orientation toward sound: possible implica-tions for cortical development. Child Deu 52 :833-838 .

Clifton RK, Morrongiello BA, Dowd JM . (1984). A devel-opmental look at an auditory illusion : the precedence effect . Deu Psychobiol 17:519-536 .

Cutting JE, Rosner BS. (1974) . Categories and bound-aries in speech and music. Percept Psychophys 16:564-570 .

Kuhl P (1991) . Human adults and human infants show a "perceptual magnet effect" for the prototypes of speech categories, monkeys do not . Percept Psychophys 50(2):93-107.

Mehler J, Jusczyk PW Lambertz G, Halsted N, Bertoncini J, Amiel-Tison C. (1988). A precursor of language acqui-sition in young infants. Cognition 29:143-178 .

Moore JM, Wilson WR. (1978) . Visual reinforcement audiometry (VRA) with infants. In : Gerber SE, Mencher GT, eds. Early Diagnosis of Hearing Loss. New York : Grune & Stratton, 177-212.

DeCasper AJ, Fifer W. (1980) . Of human bonding: new-borns prefer their mothers' voices . Science 208:1174-1176 .

Eimas PD, Siqueland PD, Jusczyk PW Vigorito J. (1971) . Speech perception in infants. Science 171:303-306 .

Eisenberg R. (1976) . Auditory Competence in Early Life . Baltimore: University Park Press.

Erulkar SD . (1972). Comparative aspects of spatial local-ization of sounds . Physiol Rev 52:237-360 .

Fernald A. (1984) . The perceptual and affective salience of mothers' speech to infants. In : Feagans L, Garvey C, Golinkoff R, eds. The Origins and Growth of Communi-cation . Norwood, NJ : Ablex, 5-29 .

Fernald A. (1992) . Human maternal vocalisations to infants as biologically relevant signals: an evolutionary perspective . In : Barkow JH, Cosmides L, Tooby J, eds. The Adapted Mind: Evolutionary Psychology and the Generation of Culture. New York : Oxford University Press, 391-428.

Fifer WP, Moon C. (1988) . Auditory experience in the fetus. In : Smotherman W Robertson S, eds. Behavior of the Fetus. West Caldwell, NJ : Telford Press, 175-188.

Hecox K. (1975) . Electrophysiological correlates of human auditory development. In : Cohen LB, Salapatek P, eds. Infant Perception : From Sensation to Cognition: Vol 2. New York: Academic Press, 151-191.

Jusczyk PW, Pisoni DB, Walley AC, Murray J. (1980) . Discrimination of relative onset time to two-component tones by infants. J Acoust Soc Am 67:262-270 .

Jusczyk PW, Rosner BS, Cutting JE, Foard F, Smith LB . (1977). Categorical perception of non-speech sounds by two-month-old infants. Percept Psychophys 21 :50-54 .

Karzon RG . (1985) . Discrimination of polysyllabic sequences by one- to four-month-old infants. J Exp Child Psychol 39:326-342 .

Kuhl PK. (1978) . Predispositions for the perception of speech-sound categories . A species specific phenomenon? In : Minifie FD, Lloyd LL, eds . Communicative and Cognitive Abilities-Early Behavioral Assessment . Baltimore: University Park Press, 229-255.

Kuhl PK . (1986) . Theoretical contributions of tests on animals to the special-mechanisms debate in speech . Exp Biol 45:233-265 .

Kuhl P. (1987) . Perception of speech and sound in early infancy. In : Salapatek P, Cohen LB, eds. Handbook of Infant Perception . Vol. 2. Orlando: Academic Press, 275-382.

Morrongiello BA. (1988a). Infants' localization of sound along two spatial dimensions : horizontal and vertical axes . Inf Behav Dev 11:127-143 .

Morrongiello BA . (1988b). Infants' localization of sounds along the horizontal axis ; estimates of minimum audi-ble angle. Deu Psychol 24:8-13.

Morrongiello BA . (1989) . Infants' monaural localization of sounds ; effects of unilateral infection. JAcoust Soc Am 86:597-602 .

Morrongiello BA, Clifton RK. (1984) . Effects of sound frequency on behavioral and cardiac orienting in new-born and five-month-old infants. J Exp Child Psychol 38:429-446 .

Morrongiello BA, Fenwick KD, Chance G. (1990) . Sound localization acuity in very young infants: an observer-based testing procedure. Deu Psychol 26:75-84 .

Morrongiello BA, Gotowiec A. (1990) . Recent advances in the behavioral study of infant audition : the develop-ment of sound localization skills . J Speech Lang Pathol Audiol 14:51-63 .

Morrongiello BA, Rocca PT. (1987a). Infants' localization of sounds in the horizontal plane: effects of auditory and visual cues . Child Dev 59:918-927 .

Morrongiello BA, Rocca PT. (1987b). Infants' localization of sounds in the vertical plane: estimates of minimum audible angle. J Exp Child Psychol 43:181-193 .

Morrongiello BA, Rocca PT. (1990) . Infants' localization of sounds within hemifields : estimates of minimum audi-ble angle. Child Dev 61 :1258-1270 .

Morrongiello BA, Trehub SE . (1987) . Age-related changes in auditory temporal perception . J Exp Child Psychol 44:413-426 .

Muir DW. (1985) . The development of infants' auditory spatial sensitivity. In : Trehub SE, Schneider BS, eds. Auditory Development in Infancy. New York : Plenum, 51-83.

Northern J, Downs M. (1991) . Hearing in Children. 4th ed. Baltimore: Williams & Wilkins.

Olsho LW (1984) . Infant frequency discrimination . Inf Behav Dec 7:27-35 .

Olsho LW, Koch EG, Halpin CR, Carter EA . (1987) . An observer-based psychoacoustic procedure for use with young infants. Deu Psychol 23:627-640 .

Oyler RF, OylerAL, Matkin ND. (1988). Unilateral hear-ing loss : demographics and educational impact . Lang Speech Hear Serv Sch 19:201-210 .

122

Postnatal Hearing Development/Peck

Papousek H. (1961) . Conditioned head rotation reflexes in infants in the first month of life . Acta Paediatr 50:565-576 .

Spring DR, Dale PS . (1977) . Discrimination of linguistic stress in early infancy. J Speech Hear Res 20:224-232 .

Peck JE . (1994a) . Development of hearing . Part I . Phylogeny. J Am Acad Audiol 5 :291-299 .

Peck JE . (1994b). Development of hearing. Part II . Embryology. J Am Acad Audiol 5:359-365 .

Pisoni DB . (1977) . Identification and discrimination of the relative onset of two component tones: implica-tions for voicing perception in stops. J Acoust Soc Am 61:1352-1361 .

Rubel ER . (1978) . Ontogeny of structure and function in the vertebrate auditory system . In : Johnson M, ed . Hand-book of Sensory Physiology 9. Development of Sensory Systems. New York : Springer, 135-257.

Rubel EW. (1985) . Auditory system development. In : Gottlieb G, Krasnegor NA, eds. Measurement ofAudition and Vision in the First Year of Postnatal Life : AMethodo-logical Overview. Norwood, NJ : Ablex, 53-90.

Schneider BA, Morrongiello BA, Trehub SE . (1990) . Size of critical band in infants, children, and adults . J Exp Psychol: [Hum Percept] 16:642-652 .

Schneider BA, Trehub SE . (1981) . Behavioral assessment of basic auditory abilities . In : Trehub SE, Schneider B, eds. Auditory Development in Infancy. New York: Plenum, 101-114.

Schulman-Galambos C, Galambos R. (1979) . Brain stem evoked response audiometry in newborn hearing screen-ing. Arch Otolaryngol 105 :86-90 .

Siqueland ER, Lipsitt LP. (1966) . Conditioned head turn-ing in human newborns . J Exp Child Psychol 3:356-376 .

Starr A, Amlie RN, Martin WH, Sanders S. (1977) . Development of auditory function in newborn infants revealed by auditory brainstem potentials . Pediatrics 60:831-839 .

Trehub SE, Schneider BA, Endman M. (1980) . Develop-mental changes in infants' sensitivity to octave-band noises . J Exp Child Psychol 29:282-293 .

Trehub SE, Schneider BA, Morrongiello BA, Thorpe LA. (1988) . Auditory sensitivity in school-age children . JExp Child Psychol 46:273-285 .

Trehub SE, Schneider BA, Morrongiello BA, Thorpe LA . (1989) . Developmental changes in high-frequency sensi-tivity. Audiology 28:241-249 .

Trehub SE, Thorpe LA, Morrongiello BA . (1987) . Organizational processes in infants' perception of audi-tory patterns . Child Dev 58:741-749 .

Walker D, Grimwade J, Wood C . (1971). Intrauterine noise: a component of the fetal environment. Am J Obstet Gynecol109 :91-95 .

Werker JF. (1989) . Becoming a native listener. Am Scient 77:54-59 .

Werker JF, Tees RC . (1992) . The organization and reor-ganization of human speech perception . In : WM Cowan, ed . Annu Rev Neurosci 15 :377-402.

Werner LA, Gillenwater JM . (1990) . Pure-tone sensitiv-ity of 2- to 5-week-old infants. Inf Behav Dev 13:355-375 .