Archives of Disease in Childhood 1994; 71: F8 1-F87

Development of fetal hearing

Peter G Hepper, B Sara Shahidullah

AbstractPrevious research has revealed that thehuman fetus responds to sound, but todate there has been little systematic inves-tigation of the development of fetal hear-ing. The development of fetal behaviouralresponsiveness to pure tone auditorystimuli (100 Hz, 250 Hz, 500 Hz, 1000 Hz,and 3000 Hz) was examined from 19 to 35weeks of gestational age. Stimuli werepresented by a loudspeaker placed on thematernal abdomen and the fetus'sresponse, a movement, recorded by ultra-sound. The fetus responded first to the 500Hz tone, where the first response wasobserved at 19 weeks of gestational age.The range of frequencies responded toexpanded first downwards to lower fre-quencies, 100 Hz and 250 Hz, and thenupwards to higher frequencies, 1000 Hzand 3000 Hz. At 27 weeks of gestationalage, 96% of fetuses responded to the 250Hz and 500 Hz tones but none respondedto the 1000 Hz and 3000 Hz tones.Responsiveness to 1000 Hz and 3000 Hztones was observed in all fetuses at 33 and35 weeks of gestational age, respectively.For all frequencies there was a largedecrease (20-30 dB) in the intensitylevel required to elicit a response as thefetus matured. The observed pattern ofbehavioural responsiveness reflectsunderlying maturation of the auditorysystem. The sensitivity of the fetus tosounds in the low frequency range maypromote language acquisition and resultin increased suscepdbility to auditorysystem damage arising from exposure tointense low frequency sounds.(Arch Dis Child 1994; 71: F81-F87)

Queen's University ofBelfast, BelfastBT7 1NN, NorthernIreland, School ofPsychologyP G Hepper

Department ofObstetrics andGynaecologyB S Shahidullah

Correspondence to:Professor Hepper.Accepted 24 May 1994

There is now little doubt that the ability torespond to sound commences in the prenatalperiod,' 2 however, there have been few sys-tematic investigations of hearing in the fetusand thus the development of fetal auditoryabilities remains largely unknown.3 Fetal audi-tory system functioning may only be studiedindirectly by the observation of a behaviouralresponse to a sound stimulus.4 Techniques ofauditory evoked brainstem responses orevoked otoacoustic emissions, which maydirectly assess auditory function, cannot beapplied during the prenatal period. As such,investigations of fetal hearing provide directevidence of auditory thresholds for behaviouralresponsiveness and only indirect, by inferepce,evidence of absolute auditory thresholds (thisissue will be returned to later). The followingpaper examines the very early development ofbehavioural responsiveness to auditory tones,

that is, in the prenatal period, and may provideinformation on the functioning and maturationof the auditory system. Two aspects offetal auditory responsiveness are examined:the range of frequencies responded to andthe intensity required to elicit a response('sensitivity') at different frequencies.The adult human has the ability to hear

frequencies ranging from approximately 20 Hzto over 20 000 Hz, although the ability to hearfrequencies at the top end of this scaledecreases with age. Some authors have arguedthat the newborn hears in only a very restrictedportion of this range, for example, 500Hz-l000 Hz,5 suggesting a considerableexpansion in the range of frequenciesresponded to after birth.

Observation of the fetus's response to soundstimuli do not support this as responses havebeen elicited by a wide range of frequencies.Responses to vibroacoustic stimuli have beenobserved to frequencies ranging from 83 Hz6to 3000 Hz,7 while studies using acousticstimulation have reported responses from250 Hz8 to 5000 Hz.9 Most of these studieshave been performed in the final weeks of ges-tation and illustrate that the fetus, albeittowards the end of gestation, responds to awide range of frequencies. The range offrequency responsiveness at earlier gestations,nor how this develops, is unknown.

Early studies of thresholds for behaviouralresponsiveness to auditory tones in newbornsrevealed these to be much higher than those ofadults10-13 and this has been confirmed bymore recent studies.14-16 Auditory brainstemresponse thresholds have been reported to behigher in newborns than in adults17-20 andevoked otoacoustic emissions appear to growin the first few days postpartum.21 Thisevidence indicates that newborn auditorythresholds are higher than those of adults.

Examination of the sensitivity of the adultear to different frequencies indicates that peaksensitivity occurs around 4 kHz,22 due in partto the resonance of the external ear canal.Attempts have been made using pure tonestimuli to assess neonatal threshold levels atdifferent frequencies. A number of studieshave found enhanced sensitivity to low fre-quency signals (100 Hz-1000 Hz) comparedwith higher frequencies'9 23 24 particularly inthe region of 200 Hz-500 Hz.23 25 In contrastsome have found no differences in sensitivityacross frequencies,26 27 greater sensitivity tohigher frequencies (3000 Hz compared with500 Hz),"I or greatest sensitivity to frequenciesaround 1000 Hz compared with 500 Hz and4000 Hz.28 In general the evidence suggeststhat newborn infants are most sensitive tofrequencies in the lower part of the adultfrequency range, especially below 1000 Hz,

F81

on Novem

ber 12, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.71.2.F

81 on 1 Septem

ber 1994. Dow

nloaded from

Hepper, Shahidullah

and their threshold levels are higher than thoseof adults. The sensitivity of the fetal auditorysystem to different frequencies is unknown.The following experiment examined the

prenatal development of behavioural respon-siveness to different pure tone stimuli. Fetusesfrom 19 to 35 weeks of gestational age werestudied and their response to different frequen-cies, presented at increasing levels of intensitydocumented. Given that hearing first developsin the lower frequency range of the adult hear-ing range the following study used stimuliranging from 100 Hz to 3000 Hz, that is, thelow frequency range of adult auditory abilities.

Subjects and methodsSUBJECTSFour hundred and fifty fetuses, from singletonuncomplicated pregnancies, were studied.Subjects were obtained from the population ofmothers attending the antenatal clinic at RoyalMaternity Hospital, Belfast. All mothers gaveconsent to take part in the study. Fetuses werestudied at one of nine gestational ages: 19, 21,23, 25, 27, 29, 31, 33, or 35 weeks of gestationalage. Fifty fetuses were studied at each age. Allfetuses who took part in the study had anuncomplicated delivery and Apgar scores ofgreater than eight at one and five minutes.Furthermore, all subjects were screened forhearing problems at 8-12 months of age usinghealth visitors' screening assessment and allwere found to possess normal hearing at thistime.

APPARATUSThe fetus was observed on real time ultrasoundscan using an ATL Ultramark 4plus ultrasoundmachine with a 5 MHz scanhead. The stimuliwere presented to the fetus via a speaker (AKGcapsule 160, Z-61A, Z-60A) placed on themother's abdomen. A synthesised functiongenerator (Wavetek, Model 23), connected tothe speaker, produced the stimuli used in thisexperiment. Five different pure tone sine wavefrequencies were used: 100 Hz, 250 Hz, 500Hz, 1000 Hz, and 3000 Hz. The intensity of thestimulus, dB(A), was measured 1 cm from thespeaker face. At present there is no suitable scalefor measuring intensity levels with regard to thefetus' and in order to provide consistencywith previous studies examining fetal auditoryabilities the dB(A) scale was used.

All ultrasound observations were recordedon videotape along with a visual indication ofwhen the sound stimulus was presented and atimer to indicate time from onset of stimuluspresentation. These tapes were used for laterindependent review by observers blind to theparticular conditions (frequency, intensity)under study.

PROCEDUREThe mother was scanned while lying on acouch in a semirecumbent position with a headtilt of 45 degrees. A standard longitudinalsection of the fetus was used to observe the

fetus so that the fetal head, upper body, andarms could be visualised. The position of thehead of the fetus was determined and thespeaker placed on the abdomen directly overthe head of the fetus.

For any particular frequency, stimuli werepresented to the fetus in 5 dB(A) steps ofincreasing intensity starting at 60 dB(A) andending at 120 dB(A). Stimuli were presentedfor 2-5 seconds with an interstimulus interval of7*5 seconds. The fetus was considered to haveresponded to the stimulus if it moved (head,arms, or upper body) either during the 2-5seconds of stimulus presentation or within theimmediate 2-5 seconds after cessation of thestimulus. (Such observations of fetal responseshave a high interobserver reliability.29) Stimuluspresentation continued until either the fetusexhibited a response on two consecutivestimulus presentations or the 120 dB(A)stimulus was presented. The intensity level atwhich the fetus exhibited its first response, if itresponded on two consecutive occasions, wasrecorded. If the fetus did not respond a score of120 dB(A) was recorded but this was not usedin any subsequent analysis. Fetuses were testedat one, and only one, gestational age but at thatage were presented with all five frequencies andthe dB(A) level at which each fetus respondedwas noted. Across all fetuses tested, and withineach gestational age, the order in which the fivedifferent frequencies were presented wascompletely randomised. The first stimulusfrequency was presented after a period of 120seconds of inactivity. Presentation of eachsubsequent series of stimulus frequencies com-menced five minutes after the end of the previ-ous series if the fetus had been inactive for thepreceding 120 seconds. If the fetus was activeduring the final two minutes of this five minuteperiod presentation of the next series wasdelayed until a period of 120 seconds ofinactivity had passed.

ResultsTHE NUMBER OF FETUSES RESPONDINGThe first response was observed at 19 weeks ofgestational age by a single fetus who respondedto the 500 Hz tone. The number of fetusesresponding to each frequency increased withgestational age (see fig 1). Fetuses respondedfirst to the low frequency tones 500 Hz, 250 Hz,and 100 Hz, with 96% responding by 27 weeksof gestational age to the 500 Hz and 250 Hztones. At this age none responded to the 1000Hz or 3000 Hz tone. Fetuses first responded tothe 1000 Hz and 3000 Hz tones at 29 and 31weeks of gestational age respectively. One hun-dred per cent of the fetuses responded to the1000 Hz tone at 33 weeks of gestational age andto the 3000 Hz tone at 35 weeks of gestation.

INTENSITY LEVEL REQUIRED TO ELICIT ARESPONSEFor those who responded to each frequencythe mean intensity level at which the responsewas elicited at each gestational age wasobtained (see figs 2 and 3). For all frequencies

F82

on Novem

ber 12, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.71.2.F

81 on 1 Septem

ber 1994. Dow

nloaded from

Development offetal heaing

Frequency (Hz)A* 100 Hz 0- 1000 Hz

-°- 250 Hz * 3000 Hz*- 500 Hz

Weeks of gestation-4- 21 - 290- - 23 -o--- 31-*- 25 A 33

o----- 27 o 35120 n

m

115

110

105

100

95

90

85Qn

0.01

19 21 23 25 27 29 31 33 35Gestational age (weeks)

0.1 0.25 0-5 1Frequency (kHz)

Figure 3 Mean intensity level required to elicit a responseat each frequency for each gestational age.

Figureeach fre

thereintensgestatwere

especias gerequirall fre

DiscuTheresporthat istoryquencalthouportio

120

115

110

105

100m

95

90

85

80

Figure 2responseexaminey=152-y=202-

1 Percentage of individuals (n=50) responding to gestation progresses the fetus exhibits aSquency at each gestational age. response to a wider range of frequencies.

Furthermore, the intensity level required tois a significant decrease (p<0-01) in the elicit a response decreases for all frequencies asity level required to elicit a response as gestation increases, this may suggest that fetalion progressed. Lowest intensity levels hearing becomes more sensitive during gesta-required for the low frequency tones, tion (fig 3). The lines for each gestational ageially those of 500 Hz and 250 Hz. Overall represent the mean intensity level required to!station advanced the intensity level elicit a response for the fetus and may be con-ed to elicit a response decreased across sidered as equal loudness contours. Thesequencies. results support previous studies examining the

postnatal development of hearing which haveindicated that hearing first develops in the low-

ission middle frequency portion of the adult auditoryresults indicate that the fetus first range and that thresholds decrease withads to frequencies of 500 Hz and 250 Hz, maturity.30363, in the lower portion of the adult audi- The results reported above may be used torange. Responsiveness to higher fre- make inferences regarding the development of:ies 1000 Hz and 3000 Hz develops later auditory functioning in the fetus. However,igh these are still in the low-middle this must proceed with caution. The maternal)n of the adult hearing range. Thus as abdomen attenuates sound, and the level of

attenuation is dependent on frequency.2 Thusthe externally measured sound levels do not

Frequency (Hz) necessarily correspond to intensity levels inside*- 100 Hz 0 000 Hz of the womb. While this factor would not

0 500 Hz change the development of responsivenesswithin a single frequency it may alter the rela-tionship between frequencies and have impli-cations for the onset of responsiveness toparticular frequencies.

In order to assess the effects of attenuation

by the maternal abdomen, attenuation factors

reported by Querleu et al2 were taken into

account and the following values subtractedfrom the observed mean intensity levels

required to elicit a response: 2 dB(A) at 100Hz and 250 Hz, 14 dB(A) at 500 Hz, 20 dB(A)at 1000 Hz, and 24 dB(A) at 3000 Hz. The

results of this subtraction (representing the

mean intrauterine sound intensity required toelicit a response) are presented in fig 4. This

ffi figure may be more representative of the sensi-19 21 23 25 27 29 31 33 35 tivity of the ear to auditory stimuli as it

Gestational age (weeks) more closely represents sound intensity as

2 Mean (SE) intensity level required to elicit a experienced by the fetus in utero and is henceat each gestational age for each of the frequencies the intensity level which elicits the behavioural

38-19 x; 500Hzy=15150-1x97x;51000Hz response (although see later). The results of73-334 x; 3000 Hzy=232-97-3-98 x. this manipulation support observations from

U,a)

L-LI10-

100

90

80

70

60

50

40

30

20

10

o

3 10OVuI

F83

on Novem

ber 12, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.71.2.F

81 on 1 Septem

ber 1994. Dow

nloaded from

Hepper, Shahidullah

FrequencyA-10Hz-2dB o- 1000 Hz - 20 dB0- 250 Hz - 2 dB - --3000 Hz -24 dB- 500 Hz- 14dB

115 r

m

110

105

100

95

90

85

80

75

70

6519 21 23 25 27 29 31 33 35

Gestational age (weeks)Figure 4 Mean intrauterine intensity level required toelicit a response at each gestational age for each frequencydetermined after accountingfor the possible atenuationprovided by the maternal abdomen.

data using intensity levels measured ex uteroand indicate that the fetus is most responsive tothe low frequency, 500 Hz, tone and thatintensity levels required to elicit a responsedecrease as gestation progresses. They differ inthe fact that the high frequency tones, 1000 Hzand 3000 Hz, now have lower thresholdsrequired to elicit a response than the 250 Hzand 100 Hz tones, although the first responseto these higher frequencies occurs later duringgestation than that to the lower frequencies.Whether accounting for the attenuation of

the mother's abdomen produces results whichrepresent the true intensity levels required toelicit a behavioural response, and indirectlyinformation on the sensitivity of the ear, todifferent frequencies must be considered care-fully. A number of different attenuation factorshave been reported by various authors,3740those used here probably represent the bestestimates at the present time, however, theyare only estimates. These values were obtainedduring the final weeks of pregnancy, afterrupture of membranes and may not be repre-sentative of attenuation with membranes intactor at earlier gestations. The attenuation factorsrepresent only the effect of the maternalabdomen. There are a number of structuraldifferences between the fetal ear and that of theadult which may further differentially affect itssensitivity to different frequencies.' Thus theseresults must be treated with care.

One difference between the results usingintensity measured ex utero and in uterorelates to the development of responsiveness todifferent frequencies. The fetus was observedto respond last to the frequencies of 1000 Hzand 3000 Hz. However, if the abdomenattenuates high frequencies more than lowfrequencies, the 1000 Hz tone, for example,presented at 120 dB(A) may in fact be only100 dB(A) in utero which may not be intenseenough to elicit a response at earlier gestations.Thus the delayed responsiveness observed to

the high frequency tones may be due to lowerintrauterine intensity levels rather than any dif-ference in the onset of auditory responsivenessto these frequencies.One means to examine this would be to

increase the intensity levels of the 1000 Hz and3000 Hz frequencies by 20 dB and 24 dB,respectively, to counter the attenuation offeredby the maternal abdomen. Ethical concernsregarding the possible harmful effects of highintensity sounds' on the hearing of the fetusrule out this possibility. However, two observa-tions from the present study suggest thatpossible intensity differences do not accountfor the difference in onset of responsivenessexhibited to low frequency and high frequencytones.

First, given the reported attenuation factorsof 14 dB at 500 Hz and 20 dB at 1000 Hz2 itwould be expected that there would be a 6 dBdifference in the womb for these frequencieswhen of equal intensity as measured outside ofthe womb. Assuming that frequency respon-siveness to the 500 Hz and 1000 Hz tonesdevelops at the same time it would be expectedthat there should be a difference of 6 dB in theresponsiveness to these tones. However, this isnot the case, differences of 19-1 dB, 13 dB,12-5 dB, and 5 dB in responsiveness at 29, 31,33, and 35 weeks of gestational age wereobserved. Only at 35 weeks is the difference inintensity level required to elicit a responseapproximately 6 dB, at earlier ages the differ-ence is 2-3 times greater. These results suggestthat intensity differences due to differentialattenuation of high and low frequencies do notsolely explain the earlier behavioural respon-siveness of the fetus to the 500 Hz tone overthe higher frequency tones.

Second, if the response of the fetus to differ-ent frequencies differed only because of theintensity differences it would be expected thatthe increase in responsiveness with gestationwould be equal for all frequencies. This is notthe case. Although all frequencies exhibit ahighly significant decrease in intensity levelrequired to elicit a response, the slopes of thedecline are different for different frequencies.Observation of fig 2 indicates that the 1000 Hzand 3000 Hz tones show a steeper decline inintensity level required to elicit a response thanthe 100 Hz, 250 Hz, and 500 Hz tones. For thetwo higher frequencies the decline in intensitylevel required to elicit a response for the 3000Hz tone appears slightly faster than for the1000 Hz tone. The decline in intensity levelrequired to elicit a response for the three lowerfrequencies appears visually similar. In order todetermine whether these visual observationswere correct a statistical analysis4l was per-formed to compare the slopes of the decline inintensity level required to elicit a response foreach frequency. For this the difference inobserved threshold for each fetus thatresponded to each pair of frequencies (forexample, 100 Hz-250 Hz, 100 Hz-500 Hz andso on) was obtained and regressed against ges-tational age. The resulting slopes were thencompared against zero using a t statistic.41The results of these statistical tests (table)

F84

on Novem

ber 12, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.71.2.F

81 on 1 Septem

ber 1994. Dow

nloaded from

Development offetal heaning

Table of results obtainedfrom comparing the slopes of linesproduced by auditoty threshold data for each frequency

FrequencyFrequency 100 250 500 1000 3000

100 - t=0-37 t=110 t=9-40 t=8-58df=264 df=265 df= 131 df= 100NS NS p<0-0001 p<0-0001

250 - - t=1-42 t=4-69 t=5-34df=288 df=131 df=100NS p<0-0001 p<0-01

500 - - - t=6-83 t=7-31df=131 df=100p<0-0001 p<O-OOOl

1000 - - - - t=3-32df=100p<0-Ol

3000 - - - - -

confirmed observations made from the graph.The decline in intensity level required to elicita response for the three lower frequencies (100Hz, 250 Hz, and 500 Hz) was not significantlydifferent from one another but for all three fre-quencies was significantly different from thedecline in intensity level required to elicit a

response for the two higher frequencies (1000Hz and 3000 Hz). Furthermore, there was a

difference in the slopes of the decline inintensity level required to elicit a response forthe two higher frequencies (1000 Hz and3000 Hz). These results suggest that intensitydifferences are not solely responsible for thedifference in the onset of responsivenessobserved to different frequencies.

In summary, although the maternalabdomen will differentially attenuate differentfrequencies, the developmental differencesobserved in responsiveness to these frequenciesmay represent true differences in their matura-tion. The range of frequencies responded towidens as the fetus ages. The intensity levelrequired to elicit a response decreases withadvancing gestation. When sound intensity ismeasured ex utero the fetus is most responsiveto the 500 Hz tone and below. When usingintensity assessed in utero, the fetus may bemore sensitive to 500 Hz tones and above. Theresults indicate that fetal response to soundcommences in the lower parts of the adultfrequency range and expands its frequencyrange and increases in sensitivity as the fetusmatures.

IMPLICATIONS FOR THE DEVELOPMENT OF THEFETAL AUDITORY SYSTEMAs mentioned previously examination of fetalauditory abilities may only proceed indirectlyat present, via the observations of behaviouralresponsiveness to sound. After birth auditorythreshold levels obtained by observations ofbehavioural responding are higher than thoseobtained by more direct means, for example,auditory evoked brainstem responses.'8 Thusthe findings obtained here cannot be used toassess absolute thresholds. However, theobservations of behavioural responding mayprovide information regarding the develop-ment of auditory function.4The responses of fetuses of different gesta-

tional ages to the acoustic stimuli used in thisexperiment indicate two aspects of auditory

fumction which develop from 19-35 weeks ofgestation. First, the fetus begins hearing in arestricted range of the adult's frequency range,around 500 Hz and this range expands duringthe course of development. Second, the initialresponses elicited by a particular frequencyrequire an acoustic stimulus of greater inten-sity than later during gestation, indicating thatfetal hearing becomes more sensitive duringdevelopment. The intensity levels of stimulirequired to elicit a response at 35 weeks of ageis 20-30 dB less than when the fetus first startsto respond. Observations of premature infantsreveal a 20 dB decrease in the intensity levelrequired to elicit an auditory evoked brainstemresponse from 28-34 weeks' gestational age toterm.42

This development of auditory responsive-ness suggests a parallel development in theauditory system. The cochlea commencesdevelopment around 10-12 weeks of age and ithas been estimated that the mechanical func-tioning of the cochlea, except for the mostbasal part, is mature by 30-35 weeks of gesta-tional age.43 The developmental changesobserved here thus reflect underlying changesin the auditory system occurring between20-35 weeks of gestation.One important development around this

time is the innervation and completion of theouter hair cells which are responsible for turn-ing the basilar membrane into an active trans-ducer of acoustic stimuli rather than simply apassive one. The possible role of this for thedevelopment of the hearing frequency range,or the increase in sensitivity, is unknown.While it is undoubtedly true that the

periphery sets the limits on what can beprocessed, the development of fetal hearingmust also be affected by the development ofthe auditory nervous system and this toomay influence the increased responsivenessobserved in this experiment. For example, thedevelopment of the ventral cochlear nucleusaccelerates between 18-33 weeks of gestationalage44 and this may play a part in the develop-ment of auditory responsiveness.The changes in behavioural responsiveness

observed in this study correlate with changesobserved in auditory brain stem responsesfrom premature infants of similar ages. Fromthe initial reliable recordings of the auditoryevoked brainstem response in prematureinfants at 28 weeks' gestational age, waveamplitude increases, wave latency decreasesand interwave interval decreases with advanc-ing age, until approximately 2-4 years of agewhen 'mature' values are recorded.45 Of par-ticular interest is the observation of a rapidreduction in interwave response latencybetween 28-36 weeks of gestation, a reductionwhich slows over the next 1-2 years. Theunderlying mediation of this, argued to be theresult ofmaturation ofthe brainstem leading todecreased brainstem conduction times,possibly due to greater synaptic efficiency, mayplay a part in the observed changes in fetalresponsiveness.The finding of initial low frequency respon-

siveness corresponds to other behavioural and

F85

on Novem

ber 12, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.71.2.F

81 on 1 Septem

ber 1994. Dow

nloaded from

Hepper, Shahidullah

physiological findings which indicate that thedeveloping auditory system responds first tolow frequency sounds.46 48 The frequenciesused in this experiment, with the exception ofthe 3000 Hz tone, would in adults be encodedby the volley principle, that is, the firing rate ofthe auditory nerves would match that of thestimulus. The 3000 Hz tone falls in the regionwhere both the volley and place principlewould operate and thus both mechanisms mayaccount for encoding this frequency. However,whether, at this early developmental stage, theplace and volley principle work at the samefrequencies as observed in adults is unknown.

Given that the frequencies used are, inadults, encoded in the firing pattern of theauditory nerves, the developments in respon-siveness may be a result of the development ofthe auditory nerves and their firing rate. Ofparticular interest is the apparent differencebetween the 100 Hz, 250 Hz, 500 Hz tonesand 1000 Hz and 3000 Hz tones in the gesta-tional age at which the fetus first responds andtheir different rates of sensitivity change. Thefiring rate of auditory nerves is about 1000impulses per second, set by the refractory rateof the nerves. Thus tones below this may beencoded in the firing pattern of a single nerve,but above this nerve fibres have to act together,phase locked firing to encode the stimulus.41 Itmay be that the difference observed herebetween tones below 1000 Hz (100 Hz, 250Hz, 500 Hz) and tones 1000 Hz and above isdue to the development of the volley principleand phase locked firing enabling groups ofnerves to act together.

There are thus a number of maturationaldevelopments of the fetus's auditory systemwhich may account for the increased range andsensitivity of hearing observed in these experi-ments, although the actual mechanism remainsunknown at this stage.

FUNCTIONALIMPLICATIONSThe behavioural response of the fetus developsfirst to frequencies around 500 Hz and lowerand the fetus is more responsive to these fre-quencies than to high frequencies. This willhave implications for the fetus's response toauditory stimuli generated in the externalenvironment. The sensitivity of the fetus to lowfrequency sounds means that the fetus will beexposed to sounds of speech and language.The fundamental frequency of the humanvoice is around 225 Hz for females and 128 Hzfor males and thus the voice will form a salientauditory stimulus for the fetus. It is known thatthe fetus recognises its mother's voice fromprenatal exposure5>52 and is able to discrimi-nate between different speech sounds inutero.53-55 This exposure to speech may beimportant for the development of language3and for recognition of the mother and develop-ment of attachment.56 The frequencies towhich the fetus first responds are those whichare relatively unattenuated by the maternalabdomen and thus will perhaps be 'loudest' inthe womb, increasing their saliency.The increased responsiveness of the fetus to

low frequency sounds is also of importancewhen considering the possible harmful effectsof sounds on hearing.1 For a broad band soundof equal intensity ex utero the low frequencieswill pass relatively unattenuated into thewomb. As these sounds correspond to the fre-quency range in which the fetus first respondsto sound, low frequencies may be potentiallymore harmful to the developing ear of thefetus than high frequency sound.I Care shouldbe taken to avoid intense exposure to lowfrequency sounds.We thank the Staff of Royal Maternity Hospital, Belfast andProfessors K Brown (psychology) and W Thompson(obstetrics) for their help and cooperation. PGH acknowledgesthe support of the Wellcome Trust and the Medical ResearchCouncil.

Hepper PG, Shahidullah S. Noise and the fetus. Sudbury:HSE Books, 1994.

2 Querleu D, Renard X, Versyp F, Paris-Delrue L, Crepin G.Fetal hearing. Eur J Obstet Gynecol Reprod Biol 1988; 29:191-212.

3 Hepper PG. Fetal psychology: an embryonic science. In:Nijhuis JG, ed. Fetal behaviour: developmental and pen-natal aspects. Oxford: Oxford University Press, 1992:129-56.

4 Hepper PG, Shahidullah S. The development of fetal hear-ing. A review. Fetal and Maternal Medicine Review (inpress).

5 Rubel EW. Auditory system development. In: Gottlieb G,Krasnegor NA, eds. Measurement of audition and vision inthe first year ofpostnatal life: a methodological overview. NewJersey: Ablex, 1985: 53-90.

6 MadisonLS, Adubato SA, Madison JK, et al. Fetal responsedecrement: true habituation? Developmental andBehavioral Pediatrics 1986; 7: 14-20.

7 Johansson B, Wendenberg E, Westin B. Measurement ofthe tone response by the human fetus. Acta Otolaryngol(Stockh) 1964; 57: 188.

8 Hepper PG, Shahidullah S. Habituation in normal andDown's syndrome fetuses. QJf Exp Psychol [B] 1992; 44:305-17.

9 Lecanuet JP, Granier-Deferre C, Busnel M. Fetal cardiacand motor responses to octave-band noises as a functionof central frequency, intensity and heart rate variability.Early Hum Dev 1988; 18: 81-93.

10 Wedenberg E. Auditory tests on new-born infants. ActaOtolaryngol (Stockh) 1956; 46: 446-61.

11 Schulman CA. Heart rate audiometry. Part 1. An evaluationof heart rate response to auditory stimuli in newbornhearing screening. Neuropddiatrie 1973; 4: 362-74.

12 Steinschneider A, Uipton EL, Richmond JB. Auditorysensitivity in the infant: effect of intensity on cardiac andmotor responsivity. Child Dev 1966; 37: 233-52.

13 Eisele WA, Berry RC, Shriner TH. Infant sucking responsepatterns as a conjugate function of changes in the soundpressure level of auditory stimuli. J Speech Hear Res 1975;18: 296-307.

14 Berg KM. A comparison of thresholds for'1/3-octave filteredclicks and noise bursts in infants and adults. Perception andPsychophysics 1993; 54: 365-9.

15 Olsho LW, Koch EG, Carter EA, Halpin CF, Spetner NB.Pure-tone sensitivity of human infants.Jf Acoust Soc Am1988; 84: 1316-24.

16 Werner LA, Folsom RC, Mancl LR. The relationshipbetween auditory brainstem response and behavioralthresholds in normal hearing infants and adults. Hear Res1993; 68: 131-41.

17 Adelman C, Levi H, Under N, Sohmer H. Neonatalauditory brain-stem response threshold and latency: 1hour to 5 months. Electroencephalogr Clin Neurophysiol1990; 77: 77-80.

18 EggermontJJ. Development of auditory evoked potentials.Acta Otolaryngol (Stockh) 1992; 112: 197-200.

19 Foxe JJ, Stapells DR. Normal infant and adult auditorybrainstem responses to bone conducted tones. Audiology1993; 32: 95-109.

20 Stuart A, Yang EY, Stenstrom R, Reindorp AG. Auditorybrainstem response thresholds to air and bone conductedclicks in neonates and adults. Am J Otol 1993; 14:176-82.

21 Kok MR, van Zanten GA, Brocaar MP. Growth of evokedotoacoustic emissions during the first days postpartum. Apreliminary report. Audiology 1992; 31: 140-9.

22 Glattke TJ. Anatomy and physiology of the auditory system.In: Rose DE, ed. Audiological assessment. New Jersey:Prentice Hall, 1978: 22-51.

23 Hutt SJ, HuttC, Lenard HG, Bernuth H, MuntjewerffWJ.Auditory responsiveness in the human neonate. Nature1968; 218: 888-90.

24Lenard HG, Bernuth H, Hutt SJ. Acoustic evokedresponses in newborn infants: the influence of pitch andcomplexity of the stimulus. Clinical Neurophysiology 1969;27: 121-7.

25 Eisenberg RB.The development of hearing in man: anassessment of current status.J'ournal of the AmericanSpeech and HearingAssociation 1970; 12: 119-2 1.

F86

on Novem

ber 12, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.71.2.F

81 on 1 Septem

ber 1994. Dow

nloaded from

Development offetal hearing

26 Crowell DH, Jones RH, Nakagawa JK, Kapuniai LE. Heartrate responses of human newborns to modulated puretones. Proceedings of the Royal Society ofMedicine 1971; 64:472-4.

27 Taguchi K, Picton TW, Orpin JA, Goodman WS. Evokedresponse audiometry in newborn infants. ActaOtolaryngology 1969; 252: 5-17.

28 Werner LA, Gillenwater JM. Pure tone sensitivity of 2- to 5-week-old infants. Infant Development and Behavior 1990;13: 355-75.

29 Shahidullah S. Hearing in the fetus. Belfast: Queen'sUniversity, 1993. (MD thesis.)

30 Brown PE, Grinnell AD, Harrison JB. The development ofhearing in the pallid bat, Anthrozous pallidus. CompPhysiol [A] 1978; 126: 169-82.

31 Crowley DE, Hepp-Raymond M-C. The development ofcochlear function in the ear of the infant rat. Journal ofComparative and Physiological Psychology 1966; 62:427-32.

32 Ehret G. Development of absolute auditory thresholds inthe house mouse (Mus musculus). Journal of the AmericanAudiological Society 1976; 1: 179-84.

33 Ehret G. Postnatal development in the acoustic system ofthe house mouse in the light of developing maskedthresholds. JAcoust Soc Am 1977; 62: 143-8.

34 Ehret G, Romand R. Postnatal development of absoluteauditory thresholds in kittens. Journal of Comparative andPhysiological Psychology 1981; 95: 304-11.

35 Finck A, Schneck CD, Hartman AF. Development ofcochlear function in the neonate Mongolian gerbil(Meriones unguiculatus). Journal of Comparative andPhysiological Psychology 1972; 78: 375-80.

36 Romand R. Maturations des potentiels cochleaires dans laperiode perinatale chez le chat et chez le cobaye. Jf Physiol(Paris) 1971; 63: 763-82.

37 Bench J. Sound transmission to the human foetus throughthe maternal abdomen wall. Journal of Genetic Psychology1968; 113: 85-7.

38 Richards DS, Frentzen B, Gerhardt KG, McCann M,Abrams R. Sound levels in the human fetus. ObstetGynecol 1992; 80: 186-90.

39 Szmeja Z, Slomka Z, Sikorsku K, Sowinski H. The risk ofhearing impairment in children from mothers exposed tonoise during pregnancy. Int J7 Pediatr Otorhinolaryngol1979; 1: 221-9.

40 Walker D, Grimwade J, Wood C. Intrauterine noise: a com-ponent of the fetal environment. Am Jf Obstet Gynecol1971; 109: 91-5.

41 McNemar Q. Psychological statistics. New York: Wiley,1962.

42 Lary S, Briassoulis G, de Vries L, Dubowitz LM, DubowitzV. Hearing response threshold in preterm and terminfants by auditory brainstem response. Jf Pediatr 1985;107: 593-9.

43 Pujol R, Lavigne-Rebillard M, Uziel A. Development of thehuman cochlea. Acta Otolaryngol (Stockh) 1991; 482(suppl): 7-12.

44 Nara T, Goto N, Nakae Y, Okada A. Morphometric devel-opment of the human auditory system: ventral cochlearnucleus. Early Hum Devl 1993; 32: 93-102.

45 Eggermont nI, Salamy A. Maturational time course for theABR in preterm and full-term infants. Hear Res 1988; 33:35-48.

46 Dolan DF, Teas DC, Walton JP. Postnatal development ofphysiological responses in auditory nerve fibres. AcoustSoc Am 1985; 78: 544-54.

47 Saunders JC, Dolgin KG, Lowry LD. The maturation offrequency selectivity in C57BIJ6J mice studied withauditory evoked response tuning curves. Brain Res 1980;187: 69-79.

48 Teas DC, Klein AJ, Kramer SJ. Cochlear responses toacoustic transients. An interpretation of whole nerveaction potentials. JfAcoust SocAm 1982; 34: 1438-89.

49 Rose JE, Brugge JF, Anderson DJ, Hind JE. Phase lockedresponse to low frequency tones in single auditory nervefibres of the squirrel monkey. Neurophysiol 1967; 30:769-93.

50 DeCasper AJ, Fifer WP. Of human bonding: newbornsprefer their mothers' voices. Science 1980; 208: 1174-6.

51 DeCasper AJ, Spence MJ. Prenatal maternal speechinfluences newborns' perception of speech sound. InfantBehavior and Development 1986; 9: 133-50.

52 Hepper PG, Scott DS, Shahidullah S. Newborn and fetalresponse to maternal voice. Journal of Reproductive andInfant Psychology 1993; 11: 147-54.

53 Lecanuet J-P, Granier-Deferre C, Jacquet AY, Capponi I,Ledru L. Prenatal discrimination of a male and femalevoice uttering the same sentence. Early Development andParenting 1993; 2: 217-28.

54 Lecanuet J-P, Granier-Deferre C, DeCasper AJ, MaugeaisR, Andrieu AL, Busnel MC. Perception et discriminationfoetales de stimuli langagiers, mise en evidence a partirde la reactivite cardiaque, resultats preliminaires.Compte-Rendus de l'Academie des Sciences, Paris 1987; III:161-4.

55 Shahidullah S, Hepper PG. Frequency discrimination bythe fetus. Early Hum Dev 1994; 36: 13-36.

56 Fifer WP. Attachment in the neonate: the ontogeny of maternalvoice preference. Greensboro: University of NorthCarolina, 1981. (Doctoral dissertation.)

F87

on Novem

ber 12, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.71.2.F

81 on 1 Septem

ber 1994. Dow

nloaded from