Braz J Otorhinolaryngol.

2010;76(6):723-8.
ORIGINAL ARTICLE BJORL .org

Auditory steady state response in pediatric audiology

Ana Emilia Linares 1, Orozimbo Alves Costa Filho 2, Maria Angelina Nardi de Souza Martinez 3

Keywords: Abstract
hearing,
electrophysiology,
hearing loss.
T he main issue regarding pediatric audiology diagnosis is determining procedures to configure
reliable results which can be used to predict frequency-specific hearing thresholds.

Aim: To investigate the correlation between auditory steady-state response (ASSR) with other tests
in children with sensorineural hearing loss.

Methods: Prospective cross-sectional contemporary cohort study. Twenty-three children (ages 1 to

7; mean, 3 years old) were submitted to ASSR, behavioral audiometry, click audiometry brain stem
response (ABR), tone burst ABR, and predicting hearing level from the acoustic reflex.

Results: the correlation between behavioral thresholds and ASSR was (0.70- 0.93), for the ABR
tone burst it was (0.73 -0.93), for the ABR click it was (0.83-0.89) only at 2k and 4 kHz. The match
between the ASSR and the hearing threshold prediction rule was considered moderate.

Conclusion: there was a significant correlation between the ASSR and audiometry, as well as
between ABR click (2k and 4 kHz) and for the ABR tone burst. The acoustic reflex can be used to
add information to diagnosis in children.

1
PhD-FMUSP, Speech and Hearing therapy - Associação dos Pais e Amigos dos Deficientes Auditivos de Sorocaba.
2
Full Professor - University of São Paulo at Bauru.
3
PhD; Professor - PUC-São Paulo and Clinical Director - Associação dos Pais e Amigos dos Deficientes Auditivos de Sorocaba.
Paper submitted to the BJORL-SGP (Publishing Management System – Brazilian Journal of Otorhinolaryngology) on August 4, 2009;
and accepted on August 22, 2010. cod. 6543

Brazilian Journal of Otorhinolaryngology 76 (6) November/December 2010

http://www.bjorl.org / e-mail: revista@aborlccf.org.br
723
 INTRODUCTION vibrator. The audiometry was carried out in a broad
acoustic room. In order to do the BAEP with click, tone
The attempt to early detect and identify hearing burst and ASSBA we used the SmartEP - Intelligent Hea-
impairment is associated with the need for proper inter- ring Systems (Auditory Evoked Potencials System):
vention, providing the child with conditions to develop The procedures done were: otoscopy and tym-
speech, language, hearing and also the social, psyche and panometry as acoustic admittance (Ya), with a 226 Hz
educational aspects. probe frequency. In order to investigate the contralateral
With the Neonatal Auditory Screening programs acoustic reflexes we employed 0.5k; 1k; 2k; 4kHz stimuli
(NAS), together with the Auditory Health programs, we saw and broad band noise (white noise), recorded with the
the possibility of early diagnosis and treatment. However, conventional probe at 226 Hz dBHL.
the diagnostic process can only be considered complete In order to study the behavioral auditory threshold
when one specifically identifies the type, degree and con- prediction based on the acoustic reflex we used the
figuration of this population’s hearing loss1,2,3. auditory threshold prediction rule4. All the thresholds
The Brainstem Auditory Evoked Potential (BAEP) obtained in dBHL were converted to dBSPL, with the aim
with the click and tone burst stimuli, otoacoustic emissions, of employing them in the auditory threshold prediction
Immittance values and Visual Reinforcement Audiometry rule. The Immittance metering device was physiologically
(VRA) and, today, the most promising method for hearing calibrated as per suggested by the author5.
assessment: auditory steady state brainstem audiometry In the Visual Reinforcement Audiometry we used
(ASSBA), are all part of the set of tests (electrophysiolo- in-the-ear phones6,7.
gical, electroacoustic and behavioral) to which the child Ludic tonal audiometry was carried out after two
is submitted for audiological investigation. years of age with in-the-ear phones; it was necessary to
The electrophysiological measurement is the most condition the child in order to perform the ludic tasks
employed tool to identify and characterize the hearing loss (fitting), each time he/she perceived the sound stimulus.
in the population of infants and children without cognitive The BAEP (Click) was initially carried out in the
and motor conditions to undergo VRA, or those who do intensity of 80 dBnHL with 20 dB increments to study the
not provide reliable responses during the test. response threshold. The ground and positive electrodes
The main issue surrounding the pediatric audiologi- were positioned on the forehead and the negative on the
cal diagnosis is to establish the procedures which provide mastoid or on the zygomatic arc (near the ears), and the
reliable and objective results, which can be used in the impedance was < 5kohm. The window used to study the
prediction of hearing thresholds by specific frequency. latencies was 20ms, the stimulus presentation rate was
Thus, these thresholds are applicable in the prescription 49.1/s; and 2000 responses were collected. The stimulus
of the technological characteristics of amplification, foste- was supplied by the EAR 3A insertion phone. The pola-
ring the auditory and language development of children. rity used was alternated. We used 30Hz high pass and
This study aimed at correlating the ASSBA with the 1500Hz low pass filters. The amplifier’s rejection level
click and tone burst BAEP results, tonal audiometry (VRA), was of 10± 25uV8.
as well as correlating the degree of hearing loss indicated in For the BAEP (Tone Burst) the initial intensity was
the ASSBA as the one suggested by the hearing threshold 80 dBnHL, with 20dB increments or decrements used
prediction rule based on the acoustic reflex in children to study the threshold. The electrodes were positioned
with different degrees of sensorineural hearing loss. as were the click ASSBA. The window for latency study
was 25ms, and the stimulus presentation rate was 37/s;
MATERIALS AND METHODS and 2000 responses were collected. The stimulus was
supplied by the EAR 3A insertion phones. The polarity
The present study (research protocol # 0492/07)
used was rarefaction. We used filters for 1k, 2k and 4kHz
was submitted to and approved by the Ethics in Research
(100-3,000Hz) and for Tone burst of 500Hz (30-3000H)z9,10.
Committee of Project analysis.
For ASSBA, the placement of the electrodes and the
The sample was made up of 23 children with bi-
earphones were the same as those for BAEP. The acoustic
lateral sensorineural hearing loss, aged between 1 and 7
signal provided was made up by carrier frequencies of
years, of both genders.
500, 1k, 2k e 4kHz, respectively modulated in the am-
Inclusion criteria included bilateral sensorineural
plitudes of 75, 85, 93, 101Hz on the left and 79, 87, 95,
hearing loss, of consistent and reliable responses upon
103Hz on the right. The parameters used were a maximum
audiometry (VRA or ludic), as well as lack of neurological
of 400 responses analyzed at every 20 collections. The
dysfunction and proper middle ear function.
filters used were 70Hz and 110Hz.
The equipment used was the AZ7 Middle Ear
The initial intensity was 60dBSPL in the multifre-
Analyzer from - Interacoustics, Interacoustic (AC33) au-
quency dichotic mode, and the increase or decrease in
diometer with in-the-ear phone, supra-aural and bone

Brazilian Journal of Otorhinolaryngology 76 (6) November/December 2010

http://www.bjorl.org / e-mail: revista@aborlccf.org.br
724
intensity depended on having a response. When no res- 1, and values near zero indicate no correlation between
ponse was recorded, the ears were monaurally stimulated the variables.
and the frequencies were presented separately. The agreement between the degree of hearing loss
The data was submitted to the Fast Fourrier Trans- obtained from the ASSBA with the one obtained from the
form analysis and angular analysis at every 20 collections, threshold prediction rule based on the acoustic reflex
and p<0.05 was used as level of significance. We con- was assessed by means of the Kappa weighed statistical
sidered valid the frequency peaks corresponding to the analysis12. The values of this coefficient can vary between
modulation frequencies which were statistically higher -1 and 1. Values equal to or lower than 0.4 indicate poor
than the noise level - we used the statistical method of agreement; between 0.41 and 0.6 moderate agreement;
the instrument, as depicted on Chart 1, below. between 0.61 and 0.80 strong agreement; and from 0.81
to 1 indicates almost perfect agreement.
Chart 1. Values of the relations which were analyzed using the
equipment’s statistical method. RESULTS

Signal/noise ratio > 6.13 dB The complete protocol was applied to 23 children 12
Side-bins signal/noise ratio > 6.13 dB;
(52%) aged between 1 and 3 years, 9 (39%) were between
4 and 5 years and 2 children (9%) were between 6 and 7
Signal amplitude > 0.0125 µV
years. Of these, 15 (65%) were females and 8 (35%) were
Noise amplitude < 0.05 µV. males. Values for the descriptive statistical values for age
are in years and are depicted on Table 1.
The maximum intensity supplied was 117dBSPL. We notice that most of the children have the same
During the procedure, if there were no responses recor- level of hearing loss in both ears. Only 1 child (4.4%) had
ded in the steady state register, we considered it a lack profound hearing loss in the right ear and severe in the
of response and the test was ended. left; and 2 (8.7%) had severe hearing loss in the right ear
and profound in the left, as depicted on Table 2.
Statistics
In the study of the correlation between the mini- Table 1. Values from descriptive statistics for the age (years)
mum response level at the ASSBA and the responses in
Standard
the remaining tests, we calculated the Pearson’s correla- N Mean
Deviation
Minimum Median Maximum
tion coefficient11. This coefficient varies between -1 and 23 3,4 1,6 1 3 7

Table 2. Distribution of joint, marginal percentages and frequencies of the degree of hearing loss on the right and left ears

Left ear
Right ear Mild Moderate Severe Profound Total
Mild 1 (4,4%) 1 (4,4%)
Moderate 5 (21,7%) 5 (21,7%)
Severe 3 (13,0%) 2 (8,7%) 5 (21,7%)
Profound 1 (4,4%) 11(47,8%) 12(52,2%)
Total 1 (4,4%) 5 (21,7%) 4 (17,4%) 13(56,5%) 23(100%)

Table 3. Values from the Pearson’s correlation coefficient (R) between ASSBA, Audiometry, click and tone burst BAEP per frequency and ear.

Frequency (kHz) ASSBA and Audiometry ASSBA and click BAEP ASSBA and BAEP Tone Burst
Right Left Right Left Right Left
R R R R R R
0.5 0,87 0,83 0,23 0,43 0,90 0,93
1 0,93 0,82 0,37 0,57 0,92 0,84
2 0,74 0,70 0,87 0,87 0,73 0,80
4 0,85 0,81 0,89 0,83 0,80 0,76

Brazilian Journal of Otorhinolaryngology 76 (6) November/December 2010

http://www.bjorl.org / e-mail: revista@aborlccf.org.br
725
 The results depicted on Table 3 (p-value) show that The results obtained showed that there was a signi-
there is a significant correlation between audiometry and ficant correlation between the ASSBA and the click BAEP;
the ASSBA in the four frequencies considered, both in the however, only for the 2k and 4kHz frequencies.
right and the left ears. The same happened in the analysis We also observed a strong correlation in the litera-
of the correlation between the tone burst BAEP and the ture between the ASSBA threshold in 2kHz and the click
ASSBA. We notice that there was a significant correlation BAEP (0.96) and between the 2k and 4kHz mean with the
between the click BAEP and the ASSBA only in the 2k click (0.97)22. Recent studies have shown a good correlation
and 4kHz frequencies. between ASSBA and click BAEP (0.63 - 0.70), being better
There was a moderate agreement between the he- for the 1kHz frequency (0.70)23.
aring loss degree classifications and the ASSBA and the Click BAEP had already been correlated with beha-
rule of predicting the degree of hearing loss based on the vioral assessment and results have shown correlations with
acoustic reflex - because there was an error in the degree the 2k and 4kHz frequencies. The click BAEP is usually
of hearing loss (Tables 4 and 5). the first measure applied in the audiological evaluation of
children when the behavioral auditory responses are not
Table 4. Frequencies and percentages of the hearing loss level successfully obtained. The click is not capable enough of
based on the threshold estimation rule from the acoustic reflex and estimating the threshold by specific frequency within the
the ASSBA - Right ear. 500 to 4kHz spectrum23.
The click BAEP enables one to estimate the hea-
PEAEE ring loss degree, for a proposal of initial intervention, not
Mild or providing details - as information per specific frequency.
Severe Profound Total
Moderate One of the main limitations of the click BAEP is the lack
Severe 5 (21,7%) 2 (8,7%) 4 (17,4%) 11 (47,8%) of a specific frequency. BAEP is highly dependent on
Profound 12 (52,2%) 12 (52,2%) neural synchrony. ASSBA overcomes some of the BAEP
limitations, since it is a response evoked by a pure tone
Total 5 (21,7%) 2 (8,7%) 16 (69,6%) 23 (100%)
modulated in frequency and amplitude. It is common for
Kappa = 0.45; standard deviation = 0.09 the examiner to decide on the presence or absence of
BAEP based on wave morphology under strong intensity15.
Table 5. Frequencies and percentages of the hearing loss level The discrepancy between click BAEP and ASSBA
based on the threshold estimation rule from the acoustic reflex and found in the present study was clear in the cases of hearing
the ASSBA - Left ear. loss with descending configuration, since the click BAEP
was absent and the ASSBA showed a preservation of the
PEAEE auditory threshold for the frequencies of 500 and 1kHz;
Mild or
Severe Profound Total
then, we can consider that the click BAEP underestimated
Moderate the hearing in children15. The prescription of amplification
Mild or
1 (4,3%) 1 (4,3%)
technological characteristics, referral for evaluation in a
Moderate cochlear implant program and hearing rehabilitation are
Severe 5 (21,7%) 2 (8,7%) 2 (8,7%) 9 (39,1%) all based on the responses obtained.
Profound 13 (56,5%) 13 (56,5%) The values of the correlation coefficients between
Total 6 (26,1%) 2 (8,7%) 15 (65,2%) 23 (100%) Tone Burst BAEP and ASSBA varied between 0.73 and 0.93
for the frequencies of 0.5k and 4kHz. The results are in
Kappa = 0.60; standard deviation = 0.10 agreement with the ones already presented in the literature,
since studies have shown a 0.86 correlation between the
two tests for the frequency of 0.5kHz23 as well as for 0.25k
DISCUSSION and 0.5kHz we obtained the values of 0.9 and 0.7917. One
In the analysis of the correlation between ASSBA recent study indicated that the correlation coefficient values
and audiometry, seen on Table 3, we can notice that there were: 0.77; 0.60; 0.66 and 0.50 between 0.5k and 4kHz22.
was a significant correlation on the four frequencies con- Although the literature challenges the difficulty in
sidered to the right and the left. Other authors found a reproducing the 500Hz frequency of the Tone Burst BAEP,
significant correlation for these procedures and suggested in the present paper, the correlation with the AASSBA
the use of ASSBA in clinical practice. A proper agreement was 0.90 on the right side and 0.93 on the left side. Some
between ASSBA and behavioral audiometry reinforces authors described the morphology of the waves found
the possibility of its application in children who did not as poor and which in many cases only wave V could be
undergo VRA, favoring the diagnosis and the hearing detected. Each change to the wave shape makes it difficult
rehabilitation as early as possible13-21. to identify and interpret the response15,24.

Brazilian Journal of Otorhinolaryngology 76 (6) November/December 2010

http://www.bjorl.org / e-mail: revista@aborlccf.org.br
726
 Moreover, they reported that the wave shape for REFERENCES
the low frequency stimulus tends to be less differentiated
1. Mencher G.; Davis A.; DeVoe S. Beresford D; Bamford, J. Universal
and more difficult to identify its click. The output limitation neonatal hearing screening: past, present, and future. Am J Audiol.
for low frequencies can then limit its use in the clinical 2001; (10): 3-12.
practice as a protocol. They suggested that the ASSBA is 2. Martinez, MANS. Tratado de Fonoaudiologia Ed. Roca São Paulo,
an alternative because it used the continuous stimulus22. 2004.
3. Joint Committee on Infant Hearing - Year 2007 Position Statement:
There was a moderate agreement between the principles and guidelines for early detection and intervention pro-
ASSBA classification and the hearing threshold estimate grams. Pediatrics. 2007; 120(4):898-921.
done based on the acoustic reflex. 4. Jerger JF, Hayes D, Anthony L. Effect of age on prediction of hearing
On Tables 4 and 5 one can notice that 5 children level from the acoustic reflex. Arch Otolaryngol. 1978; (194): 393-4.
5. Jerger JF, Burney P, Mauldin L, Crump B. Predicting hearing loss
(21.7%) had mild right and left side hearing loss according from the acoustic reflex. J Speech Hear Disord. 1974; (39):11-22.
to the ASSBA, nonetheless, the rule for estimating the 6. Gravel JS. Audiologic assessment for the fitting of hearing instruments:
auditory threshold based on the acoustic reflex suggested big challenges from tiny ears. In Seewald RC. A sound foundation
that these children had severe hearing loss on the right through early amplification: proceedings of an international confer-
ence. Edited by Richard C Seewald, National center of audiology,
and left. In the cases of profound hearing loss, 4 children London Ontario Canada, 2000.
(17.4%) on the right and 2 (8.7%) on the left were classified 7. Versollato MC. Relações entre o desenvolvimento sensório motor,
as having severe loss. características individuais e desempenho na audiometria de reforço
Although we have errors with the rule, it is worth visual em crianças com cinco a nove meses de idade. [Dissertação].
São Paulo: Pontifícia Universidade Católica; 2005.
mentioning that in no case the rule suggested normal 8. Elliott C, Lightfoot G, Parker D, Stapells D, Stevens J, Sutton G, et al.
hearing, since all the children assessed had sensorineural Automated auditory brainstem response information an guidelines
hearing loss. The most severe error may happen when the for screening hearing in babies. www.nhsp.info/workbook.shtml.
rule estimates the hearing to be normal in cases of severe 2002.
9. Stapells DR. Current status of the auditory steady-state responses for
hearing loss. A moderate error happens when the rule estimating an infant´s audiogram. In: SEEWALD, R. C.; BANFORD,
suggests normal hearing and there is moderate hearing J. (Eds). A sound foundation through early amplification. Chicago:
loss; or when there is a severe hearing loss and the rule Phonak; 2005.p.43-59.
suggests moderate hearing loss. The errors in the hearing 10. Gorga MP, Johnson TA, Kaminski JR, Beauchaine KL, Garner CA,
Neely ST. Using a Combination of Click- and Tone Burst-Evoked
loss degree estimates have already been discussed in other Auditory Brain Stem Response Measurements to Estimate Pure-Tone
studies5,25-28. Thresholds. Ear Hear. 2006;(27): 60-74.
As we compare the values of pure tones in the study 11. Fisher L D, van Belle G. Biostatistics. John Wiley & Sons, New York,
of acoustic reflexes with broad band noise, it was possi- 1993.
12. Sim J, Wright CC. The Kappa Statistic in Reliability Studies: Use,
ble to notice that in 100% of cases the broad band noise Interpretation and Sample Size Requirements. Phys Ther. 2005; 85(
threshold was either high or absent, suggesting hearing 3): 257 - 68.
loss. When hearing is normal, according to summation, the 13. Lins OG. Audiometria fisiológica tonal utilizando respostas de estado
broad band noise threshold is less intense when compared estável auditivas do tronco cerebral. [Doutorado em Medicina]. São
Paulo: Escola Paulista de Medicina, Universidade Federal de São
to pure tone4. Paulo, 2002.
The analysis of the relation between the threshol- 14. Perez-Abalo MC; SavioG; Torres. Steady state responses to multiple
ds for pure tone and for broad band noise (white noise) amplitude-modulated tones: an optimized method to test frequency-
indicated the difference in the auditory pathway behavior specific thresholds in hearing-impaired children and normal-hearing
subjects. Ear Hear. 2001; 22 (3): 200-11.
after the stimulus and this analysis could add information 15. Cone-Wesson B, Parker J, Swiderski N, Rickards F. The auditory
to the children audiological diagnoses; however, the au- steady state response: full-term and premature neonates. J Am Acad
ditory threshold estimation rule based on the reflex could Audiol. 2002; 13 (5): 260-9.
not have been used alone4. 16. Rance G; Rickards F. Prediction of hearing threshold in infants using
auditory steady-state evoked potentials. J Am Acad Audiol. 2002; (13):
236-45.
CONCLUSÃO 17. Stueve MP, O´rourke C. Estimation of hearing loss in children:
comparison of auditory steady-state response, auditory brainstem
ASSBA was significantly correlated with the au- response, and behavioral test methods. Am J Audiol. 2003; 12(2):
diometry (VRA and ludic), with the click BAEP for 2k 25-136.
18. Swanepoell W, Hugo R, Rood R. Auditory steady-state responses
and 4kHz and with the tone burst in all the frequencies, for children with severe to profound hearing loss. Arch Otolaryngol
bilaterally. Head Neck Surg. 2004; 130(5):531-5.
The rule used to estimate the hearing loss can help 19. Luts H, Desloovere C, Wouters J. Clinical application of dichotic mul-
and add information used to confirm the hearing loss, but tiple- stimulus auditory steady-state responses in high risk newborn
and young children. Audiol Neurootol. 2006; 11(1):24-37.
it must not be used alone. 20. Duarte JL, Alvarenga KF, Costa Filho OA, Lins OG. A resposta auditiva
de estado estável na avaliação auditiva: aplicação clínica. Pró Fono.
2008; 20(2): 105-10.

Brazilian Journal of Otorhinolaryngology 76 (6) November/December 2010

http://www.bjorl.org / e-mail: revista@aborlccf.org.br
727
21. Rodrigues GRI. Potenciais evocados auditivos de estado estável em 24. Kaf W A, Sabo AD, Durrant JD. R. Reliability of electric response audiometry
crianças com perda auditiva neurossensorial. [Dissertação]. São Paulo: using 80Hz auditory steady-state responses. Int J Audiol. 2006; (45):477-
Pontifícia Universidade Católica, 2009. 86.
22. Werff KRV; Brown CJ; Gienapp BA; Clay KMS. Comparison of auditory 25. Tsappis A. Prediction of auditory sensitivity. Arch Otolaryngol. 1977;
steady-state response and auditory brainstem response thresholds in (103):322-5.
children. J Am Acad Audiol. 2002; (13): 227-35. 26. Keith RW. An evaluation of predicting hearing loss from the acoustic reflex.
23. Gorga MP, Worthington DW, Reiland JK, Beauchaine KL, Goldgar DE. Some Arch Otolaryngol. 1977; (103): 419-24.
comparisons betweem auditory brainstem response thresholds, latencies, 27. Van Wagoner RS; Goodwine SE. Clinical impressions of acoustic reflex
and the tone pure audiogram. Ear Hear. 1985; 6(2): 105-12. measures in adult population. Arch Otolaryngol. 1977; (103): 582-4.
28. Hall JW. Predicting hearing level from the acoustic reflex: a comparison
of three methods. Arch Otolaryngol. 1978;(104):601-6.

Brazilian Journal of Otorhinolaryngology 76 (6) November/December 2010

http://www.bjorl.org / e-mail: revista@aborlccf.org.br
728