The widespread introduction of universal newborn hearing screening (UNHS) has resulted in a dramatic reduction to the age of diagnosis for children with congenital hearing impairment. Parents of children with hearing impairment can now seek intervention for their child within the first few months of life. Early cochlear implantation provides a child with congenital severe-to-profound hearing loss the opportunity to develop oral language skills at a rate comparable to their typically hearing peers and as a result, cochlear implant (CI) services are under increased pressure to make CI recommendations for infants as early as possible. Providing evidence-based and timely recommendations for implantation is a significant challenge for pediatric CI programs.
Auditory evoked potential (AEP) measures, such as the auditory brainstem response (ABR) and auditory steady state response (ASSR), are widely accepted and routinely used for identification and diagnosis of hearing loss in infants. The use of these measures to estimate the degree of hearing loss, for children who are too young or developmentally unable to participate in behavioral audiometry, is supported by a body of evidence demonstrating the relationship between these AEP measures and behavioral pure-tone hearing thresholds (HTLs) (1–3).
For decades the click-ABR has been used to assess the hearing status of infants and young children. Research has demonstrated that there is a strong correlation (≥0.9) between ABR thresholds and behavioral pure-tone HTLs (4,5). The click-ABR threshold is believed to most closely relate to high-frequency HTLs, although in some cases, the resulting ABR can correspond to the frequencies at which the hearing is best (5). Baldwin and Watkin (4) and Gorga et al. (5) concluded that ABR thresholds could be used to predict pure-tone behavioral thresholds but conceded that there were instances when the behavioral thresholds would be under and overestimated.
There are known limitations of the ABR for assessment of children with hearing loss in the severe-to-profound range. Maximum stimulus intensity for the ABR is limited to around 90 to 100 dBnHL. Beyond this level significant distortion occurs with most standard transducers. ABR has also been reported to overestimate the degree of hearing loss in those with severe-to-profound hearing loss (6,7). Therefore, absence of an ABR threshold at maximum presentation level does not imply absence of useful residual hearing.
It should be highlighted that the established correlations between ABR and pure-tone HTLs do not imply that they are predictive and there remains a great deal of variation regarding the accuracy of the ABR as a predictor of behavioral HTLs (8). The same could be said for the ASSR. Differences between ABR/ASSR and behavioral thresholds across most studies show standard deviations of approximately 10 dBHL; thus, for 1 in 20 infants, behavioral thresholds will be under/overestimated by 20 dBHL (9).
There seem to be some advantages to the ASSR over the ABR, particularly for the assessment of children with severe-to-profound hearing loss. The automated detection of a response based on Fourier and statistical analysis for the ASSR eliminates the subjective errors associated with waveform interpretation which may occur with the ABR (10). The ASSR also has the ability to stimulate at presentation levels as high as 120 dBHL. Rance and Briggs (11) reported strong positive correlations (0.81–0.93) between ASSR thresholds and subsequent behavioral HTLs for infants as young as 3 months and in a subsequent study, Rance and Rickards (3) demonstrated correlations between the ASSR and pure-tone HTLs at frequencies between 500 and 4 kHz to exceed 0.95. It is important to note that ASSR is not immune to misinterpretation, as current clinical use of ASSR measures routinely violates statistical assumptions and thus even significant ASSRs can sometimes be random noise (12).
Despite the shortcomings of the ABR and ASSR, the published strong relationships between these measures and behavioral HTLs support their use to estimate pure-tone HTLs for young children and form the basis of early hearing aid fitting. However, clinicians interpreting these measures should be aware of the variance in the data sets used to create normative data for both measures as the confidence intervals for estimating pure-tone HTLs from an ABR/ASSR can be quite large. Clinicians should also be mindful that most of the literature supporting the strong relationships between ABR/ASSR and behavioral HTLs has been conducted under controlled research conditions which are challenging to replicate in clinical practice. The relationship between ABR/ASSR assessed in the diagnostic clinical setting and behavioral HTLs typically assessed several months later (for confirmation of the degree of hearing loss) may be quite different to what is currently published and is the focus of the present study.
The benefits of early cochlear implantation, for children with congenital severe-to-profound hearing loss, have been well described over the past decade (13–17). UNHS has resulted in infants being referred for CI investigations as young as 4 weeks of age; however, many children do not receive their first CI until after 12 months of age. If the benefits of early implantation are to be accepted, then the consequent challenge for clinicians assessing infants for CI is to establish reliable HTLs upon which to make audiological recommendations for cochlear implantation.
Many factors may contribute to the age at which a child receives a cochlear implant. Dettman et al. (18) recently described the barriers to early implantation and identified a large amount of variation on the clinical pathway for infants undergoing assessment as contributing to older age at implantation for some children. Clinical practices at most tertiary institutions providing CI services for young children require the establishment of reliable behavioral audiological results before recommending a CI. In the Dettman et al. (18) study sample, the number of weeks between the first and last behavioral hearing assessment ranged from 1 to 24 weeks. This implies that the process of establishing behavioral hearing levels is contributing to a delay in age at CI for many children.
As clinical practice in pediatric CI programs shifts toward making a CI recommendation as early as possible, the role of AEP testing in guiding CI candidacy decisions has become topical. Paediatric CI programs within Australia are performing CI surgery as young as 4 months of age (14) and reports from Europe suggest that children are receiving cochlear implants as young as 2 months of age (19). Recommendations for these very young children are being made on the basis of AEP findings. The efficacy of using AEP results obtained in the diagnostic clinical setting to determine CI candidacy is currently untested.
An evidence-based protocol for evaluating infants’ suitability for CI within the first 6 months of life needs to be established. Developing such a protocol has the potential to improve access to CI for infants and their families, reduce the number of clinical attendance appointments and result in earlier CI. This in turn has the potential to improve long-term speech and language outcomes for these children with severe-to-profound hearing impairment.
This study aimed to challenge the requirement for behavioral audiological assessment as part of the standard CI assessment battery for children with congenital hearing loss by investigating the relationship between diagnostic ABR and ASSR results and subsequent behavioral audiometry for a cohort of children referred for CI. It aimed to determine if there was a clear pattern of objective audiology results which warranted a recommendation for CI, without the requirement of behavioral testing, with the goal of reducing the age at implantation for children with congenital severe-to-profound hearing loss.
In the state of Victoria, Australia, where this study was conducted a state-wide diagnostic audiology protocol has been established. Infants typically have their hearing screened within the first 1 to 2 days of life using an automated ABR (aABR) screen. Infants who do not receive a “pass” result (aABR >40 dB) on their hearing screen are referred for comprehensive diagnostic audiology assessment at one of the 19 approved sites across the state. Each of the approved sites must adhere to an agreed diagnostic audiology assessment protocol that includes separate ear high-frequency tympanometry, bone conduction ABR, air conduction click-ABR and ASSR or tone-burst ABR (TB-ABR) at 500, 1k, 2k, and 4 kHz. The equipment and exact parameters used may vary from site to site and providing details of each piece of equipment was beyond the scope of this study. All children were tested under natural sleep for diagnosis or confirmation of hearing loss. Referral to the CI service was made if the estimated hearing loss is severe or worse in both ears. The study institution is the sole provider of pediatric CIs in the state (current population 6 million).
All children received hearing aids before or during their evaluation for a CI. Hearing aids were fitted based on the threshold estimates from the ABR/ASSR by a national service provider, Australian Hearing (20).
Results for click-ABR, ASSR at 500, 1k, 2k, and 4 kHz and tympanometry, provided to the CI service by the diagnostic hearing centers, were collected from the patient file. For the ABR, the physiologic ABR threshold was taken as the lowest stimulus level at which a wave V response could be visually detected in the response. For the ASSR, the threshold was determined automatically using Fourier analysis of the EEG with a statistical criterion of p < 0.05.
Unaided behavioral audiology was conducted on site at the study institution by experienced pediatric audiologists. Testing was performed using standard visual reinforcement or play audiometry techniques in a sound treated booth. Behavioral testing commenced at the earliest opportunity from 6 months of age. Children were assessed at regular intervals until two reliable sets of behavioral HTLs were obtained.
A retrospective review of 123 pediatric patient files for children referred to the CI service before 3 years of age over a 3-year period (2012–2014) was undertaken. Results for click-ABR, ASSR, and behavioral audiology and tympanometry were collected and relationships were investigated for 64 children who met the inclusion criteria. No children were assessed using TB-ABR. Data were excluded for 59 children due to the presence of auditory neuropathy (AN) findings, middle ear pathology at the time of testing, if no response (NR) was obtained on ASSR at 85 dB and testing was not conducted at a higher intensity level (equipment test protocol limited output to 85 dB or testing discontinued because of parental distress), behavioral testing was judged to be unreliable due to poor response state of the child or inability to successfully condition the child to the test technique and/or the family did not proceed with behavioral testing.
In the instance that NR was obtained for the ABR or ASSR at ≥95 dB, 5 dB was added to maximum level tested to enable inclusion in statistical analyses, e.g., NR at 110 dB for ASSR, given value of 115 dB for statistical analysis. For behavioral HTLs, if NR was obtained at maximum presentation level, 5 dB was added to the maximum presentation level on the audiometer, e.g., NR at 120 dBHL, given value of 125 dB for statistical analysis.
For statistical comparison ASSR thresholds were compared to the corresponding pure-tone HTL. For click-ABR, ABR threshold was compared to the child's best HTL level between 1 and 4 kHz. The relationships were tested using Minitab statistical software and the Pearson correlation coefficient with a linear regression model.
A total of 123 children were referred to the CI program at the study institution between 2012 and 2014. Of the children referred, 64 met the above inclusion criteria. All children underwent diagnostic testing using ABR/ASSR prior to being referred to the CI program. Mean age at referral for CI assessment was 0.51 years (SD 0.63 yr, range 0.04–2.51 yr).
A summary of hearing loss etiologies and significant comorbidities for the 64 children included in the study can be found in Table 1.
For the 64 children there were 129 occasions of service for ABR/ASSR testing. Not all measures (ABR and ASSR at 500, 1k, 2k, and 4 kHz) were able to be completed on the one occasion so multiple appointments were required for most children. For an individual child the number of appointments for ABR/ASSR testing ranged from one to five. On 31 (24%) occasions ABR and/or ASSR testing was not completed at stimulation levels ≥95 dB, for example results reported as NR on ASSR at 85 dBHL. This occurred when the equipment test protocol limited the output to ≤95 dB, when testing was discontinued because the baby became unsettled or due to parental distress. These results were excluded from further analysis leaving 98 sets of ABR/ASSR data available for analysis. The number of appointments for behavioral audiometry ranged from two to seven. Not all frequencies were able to be obtained for both ears for all children.
Mean age at first ABR/ASSR was 0.19 years (SD 0.33 yr, range 0.03–2.11 yr). Mean age at first behavioral audiometry session was 0.84 years (SD 0.65 yr and range 0.39–3.17). Mean timeframe between first ABR/ASSR and first behavioral audiometry session was 0.64 years (SD 0.56 yr, range 0.24–3.04).
Click-ABR thresholds for the 64 children ranged from 20 dBnHL to NR at ≥95 dBnHL. The corresponding best behavioral HTL from 1 to 4 kHz ranged from 20 dBHL to NR at 120 dBHL. There was a significant correlation (r = 0.658, p < 0.001) between the click-ABR threshold and the child's best behavioral HTL from 1 to 4 kHz. The relationship is illustrated in Figure 1. The 95% confidence interval and 95% prediction interval are also shown in Figure 1. The 95% prediction interval was 60 dB wide which suggested that for any given click-ABR threshold the best behavioral HTL from 1 to 4 kHz can be predicted within a 60 dB range with 95% confidence. The data also showed that for an NR click-ABR threshold at 95 dBnHL (indicated in Fig. 1 at 100 dBnHL) the corresponding best behavioral HTL ranged from 70 dBHL to no measurable hearing at 115 dBHL (indicated in Fig. 1 at 120 dB).
ASSR thresholds for the children ranged from 30 dBHL to NR at ≥110 dB at 500 Hz, 1 kHz and 4 kHz and ranged from 35 dB to NR at ≥110 dBHL for 2 kHz. The corresponding behavioral HTLs ranged from 25 dBHL to 115 dBHL at 500 Hz, from 25 dBHL to NR ≥120 dBHL at 1 kHz, from 35 dBHL to NR ≥120 dBHL at 2 kHz, and from 20 dBHL to NR ≥120 dBHL at 4 kHz. ASSR and behavioral HTL were significantly correlated for 500 Hz (r = 0.661, p < 0.001), 1 kHz (r = 0.522, p < 0.001), 2 kHz (r = 0.409, p < 0.001), and 4 kHz (r = 0.400, p < 0.001). The relationship, including 95% prediction and confidence intervals for 1 kHz, is shown in Figure 2 and for 4 kHz in Figure 3. For 1 kHz, the 95% prediction interval was 60 dB wide and for 4 kHz it was 85 dB wide. For a NR ≥110 dBHL on ASSR at 1 kHz the corresponding behavioral HTLs ranged from 80 to 120 dBHL. For a NR ≥110 dBHL on ASSR at 4 kHz the corresponding behavioral HTLs ranged from 75 to 120 dBHL.
The ability to use AEP measures to predict the degree of hearing loss for infants is fundamental to the early provision of hearing devices and intervention for children with congenital hearing loss. The present study aimed to investigate the validity of using such measures to guide early CI candidacy decisions.
Findings of this study supported previous research demonstrating significant correlations between click-ABR and ASSR and subsequently obtained behavioral pure-tone HTLs for infants with hearing loss. However, for both measures the predictive value was poor raising concern over the use of such measures to identify children who would benefit from a CI.
In the present study the correlation between click-ABR threshold and best behavioral HTL between 1 and 4 kHz was r = 0.66 (p < 0.001). This is consistent with previous reports demonstrating correlations between 0.55 and 0.94 (5,21). Despite the significant correlation the predictive value of the relationship was poor with the prediction interval being 60 dB wide. There was a >20 dB difference between the ABR and best HTL for 17 (24%) out of the 72 data comparisons. This is comparable to the findings of Baldwin and Watkin (4) who found that 28% of their participants had pure-tone averages at 2 to 4 kHz that differed by 20 dB from their neonatal ABR. For 11 comparisons, from 4 children, ABR underestimated the best HTL from 1 to 4 kHz by >20 dB. This can be accounted for by the presence of progressive loss (serial decline in hearing levels across multiple objective and/or behavioral hearing tests) for these four children. The more concerning finding, in the context of the present study, is for the six children where the ABR overestimated the best pure-tone HTL from 1 to 4 kHz by >20 dB. For one child the ABR threshold was reported as 55 dBnHL and corresponding HTL was found to be 30 dBHL. Further investigation revealed that ABR was not assessed below 55 dB so there remains the possibility that the true ABR threshold was below this level. For another child the ABR threshold was reported as 80 dBnHL and the corresponding HTL was 55 dBHL. This child had a history of middle ear pathology and there was possibility of an undiagnosed conductive component at the time of ABR. It may also be possible that the ABR threshold has been misinterpreted as raw traces where not viewed as part of this study. Four children with absent ABR at ≥95 dBnHL were found to have measurable acoustic hearing between 1 and 4 kHz, with best behavioral HTLs ranging from 70 to 80 dBHL. One child was born extremely prematurely (at 28 wk), one had cochleae-vestibular dysplasia and the remaining two had an etiology of Connexin 26 mutation with normal birth and developmental histories. It is known that the ABR may be temporarily affected by the baby's neurological status and interpretation of the ABR can be difficult in children with delayed central nervous system maturation. Baldwin and Watkin (4) came to a similar conclusion suggesting that the ABR was likely to overestimate hearing loss in infants who suffered perinatal illness or were born very/extremely prematurely. For the child with cochleae-vestibular dysplasia, it is possible that this anatomical abnormality may have affected the interpretation of the ABR. There is no clear explanation for the variation between ABR and behavioral results for the children with Connexin 26 mutation. The overestimation of behavioral HTLs by the click-ABR found in the present study is not a unique finding. Marttila and Karikoski (6) reported 65.9% of children with NR on click-ABR had hearing with pure-tone averages (2–4 kHz) ranging from 65 dBHL to 120 dBHL. It can be concluded that the absence of the ABR is suggestive of a significant hearing loss but does not imply nonviable residual hearing in all cases.
The present study found significant positive correlations between ASSR thresholds at 500, 1k, 2k, and 4 kHz and subsequent pure-tone behavioral HTLs at the same octave frequencies. The correlations ranged from r = 0.661 at 500 Hz to r = 0.400 at 4 kHz. These correlations are poorer than those reported by Rance and Rickards (3) on a similar aged cohort of children. Notable differences between the studies were the carefully controlled conditions for the collection of ASSR results, exclusion of children with progressive loss, and the removal of data where the ASSR or behavioral threshold was absent at maximum presentation levels in the Rance and Rickards (3) study. Despite the significant correlation between ASSR and behavioral hearing levels, the predictive value of the ASSR in the present study was limited. For 1 kHz the 95% prediction interval was 60 dB wide and for 4 kHz it was 85 dB wide.
For ASSR at 1 and 4 kHz, 38 (30%) of the 125 comparisons were found to be >20 dB from the estimated HTL. For 31 comparisons, ASSR underestimated the corresponding behavioral HTL and for seven comparisons, ASSR overestimated the corresponding behavioral HTL. Fourteen children were the source of the 31 comparisons where ASSR underestimated HTL. Nine of these children were identified as having a progressive hearing loss. For the remaining five children aetiology was Connexin 26 mutation (1) and unknown (4). Progressive loss cannot be excluded for these five children because there was insufficient objective and/or behavioral test sessions to determine if there was a serial decline in hearing over time. There were two instances from two children where the ASSR threshold at 1 kHz overestimated the behavioral HTL by >20 dB. In the first instance the 1 kHz ASSR threshold was reported as 70 dBHL and the subsequent behavioral HTL was found to be 35 dBHL. There is no obvious explanation for this discrepancy. The child's aetiology was genetic and there were no notable comorbidities or anatomical abnormalities. The step-sizes used to assess the ASSR were not available, so it can only be speculated as to how accurate the ASSR threshold was. In the second instance the 1 kHz ASSR threshold was reported as 105 dB and the subsequent behavioral HTL was found to be 75 dBHL. This child had cochleae-vestibular dysplasia and the ABR also overestimated the behavioral hearing levels by >20 dB. It is possible that the test conditions for this child were not ideal at the time of ABR/ASSR resulting in this overestimation. There were five instances from three children where the ASSR threshold at 4 kHz overestimated the estimated HTL by >20 dB. In one instance the ASSR threshold was 95 dB and the behavioral threshold was found to be 55 dBHL. The aetiology for this child was Connexin 26 mutation and there were no known comorbidities or anatomical abnormalities reported. There is no clear explanation for this discrepancy in thresholds other than the possibility that the test conditions for the ASSR were not ideal. ASSR threshold was reported as NR at 100 dB (indicated in Fig. 3 as 105 dB) and the corresponding behavioral threshold assessed on two separate occasions was found to be 65 dB HL and 75 dB for one child. This child had an aetiology of CMV and had spent 11 days in the special care nursery. If treated with antiviral therapy, it is possible to prevent or lesson the severity of hearing loss for infants with congenital CMV. It is unknown if this child received antiviral therapy but if they did, it is a possible explanation for the discrepancy between ASSR and behavioral hearing levels. The final child had two ASSR results at 4KHz, one where the threshold was reported as 110 dB and the second was NR at 110 (indicated in Fig. 3 as 115 dB). The corresponding behavioral HTL was found to be 75 dBHL. The aetiology for this child was Connexin26 and there were no known comorbidities or anatomical abnormalities reported. There is no clear explanation for the discrepancy between ASSR and behavioral thresholds in this instance. Less than ideal test conditions for the ASSR seem unlikely given the overestimated ASSR threshold was obtained on two occasions.
It is important to highlight that ABR/ASSR results used for the analysis in this study were not obtained at the study institution and no effort was made to control the procedure(s) and parameters used to obtain the diagnostic results for ABR/ASSR. Results were accepted on face value as interpreted by the diagnostic audiology center. This is a notable limitation of this study and raises the possibility of inaccurate/unreliable ABR/ASSR results. However, in the reality of clinical work such errors will occur, so it is imperative to check for consistency among objective, behavioral, and functional assessments conducted and investigate thoroughly when inconsistency occurs.
Across the comparisons detailed above (click-ABR and ASSR at 1 and 4 kHz) there were 10 instances where the AEP measure suggested profound hearing loss and the behavioral HTL was found to be in the severe hearing loss range. These 10 observations were obtained from 6 children out of the total group of 64. In the context of making cochlear implant recommendations, this finding is of concern and confirms that thresholds from clinically obtained click-ABR or ASSR alone do no provide sufficient accuracy to inform a cochlear implant recommendation.
Reducing the age children with severe-to-profound hearing loss receive a CI remains a priority for the study institution. An alternate evaluation pathway, with reduced reliability on the behavioral audiogram, is needed in order to make CI recommendations as early as possible. It will be necessary to explore alternative methods for evaluating infants for cochlear implantation that are not based primarily on the ABR/ASSR. An evaluation protocol that incorporates a range of assessments which informs the degree of hearing loss assesses audibility using hearing aids and measures functional benefit from the hearing aids within the first 6 to 12 months of life is desirable and will be the focus of a future project at the study institution.
This study indicates that the reality of clinical practice in pediatric audiology leaves considerable uncertainty in determining accurate hearing thresholds in infants under 6 months of age. It is clear that AEP assessments, if carried out carefully and interpreted conservatively, can provide estimates of hearing thresholds on the day of testing, but there is considerable variance between these estimates and behavioral thresholds obtained at a later date. Even in this cohort where there has been an attempt to exclude infants with middle ear disease and AN findings, the AEP tests provide threshold estimates that have an uncertainty of 25 to 30 dB. When making judgements about cochlear implantation for these children, there remains a need to consider all available evidence including objective, behavioral, and observational measures and ensure that there is consistency. Reliance on AEP testing alone could lead to inappropriate decisions in relation to cochlear implantation.
1. Rance G, Dowell RC, Rickards FW, et al. Steady-state evoked potential and behavioral hearing thresholds in a group of children
with absent click-evoked auditory brain stem response. Ear Hear
2. Stapells DR. Threshold estimation by the tone-evoked auditory brainstem response: A literature meta-analysis. J Speech Lang Pathol Audiol
3. Rance G, Rickards FW. Prediction of hearing threshold in infants
using auditory steady-state evoked potentials. J Am Acad Audiol
4. Baldwin M, Watkin P. Predicting the degree of hearing loss
using click auditory brainstem response in babies referred from newborn hearing screening. Ear Hear
5. Gorga M, Johnson T, Kaminski J, et al. Using a combination of click- and tone burst-evoked auditory brain stem response measurements to estimate pure-tone thresholds. Ear Hear
6. Marttila T, Karikoski J. Comparison between audiometric and ABR
thresholds in children
. Contradictory findings. Eur Arch Otorhinolaryngol
7. Sininger Y. The use of auditory brainstem response in screening for hearing loss
and audiometric threshold prediction. Auditory Evoked Potentials: Basic Principles and Clinical Application
2007; Baltimore, MD: Lippincott Williams & Wilkins, 254–274.
8. Hang AX, Roush PA, Teagle HF, et al. Is “no response” on diagnostic auditory brainstem response testing an indication for cochlear implantation in children
? Ear Hear
9. Picton TW, Dimitrijevic A, Perez-Abalo M-C, et al. Estimating audiometric thresholds using auditory steady-state responses. J Am Acad Audiol
10. Rance G, Rickards FW, Cohen LT, et al. The automated prediction of hearing thresholds in sleeping subjects using auditory steady-state evoked potentials. Ear Hear
11. Rance G, Briggs RJ. Assessment of hearing in infants
with moderate to profound impairment: The Melbourne experience with auditory steady-state evoked potential testing. Ann Otol Rhinol Laryngol
12. John MS, Purcell DW. Rance G. Introduction to technical principles of auditory steady-state response testing. The Auditory Steady-State Response. Generation, Recording, and Clinical Application
. San Diego, CA: Plural; 2008. 11–53.
13. Ching TY, Dillon H, Button L, et al. Age at intervention for permanent hearing loss
and 5-year language outcomes. Pediatrics
14. Dettman S, Dowell R, Choo D, et al. Long-term communication outcomes for children
receiving cochlear implants younger than 12 months: A multicenter study. Otol Neurotol
15. Dettman S, Pinder D, Briggs R, et al. Communication development in children
who receive the cochlear implant
younger than 12 months: Risks versus benefits. Ear Hear
16. Leigh J, Dettman S, Dowell R. Evidence-based guidelines for recommending cochlear implantation for young children
: Audiological criteria and optimizing age at implantation. Int J Audiol
17. Leigh J, Dettman S, Dowell R, et al. Communication development in children
who receive a cochlear implant
by 12 months of age. Otol Neurotol
18. Dettman S, Choo D, Dowell R. Barriers to early cochlear implantation. Int J Audiol
19. Colletti L, Mandalà M, Colletti V. Cochlear implants in children
younger than 6 months. Otolaryngol Head Neck Surg
20. King AM. The national protocol for paediatric amplification in Australia. Int J Audiol
21. Allyn and Bacon, Stapells D, Picton T, Durieux-Smith A. Jacobson JT. Electrophysiologic measures of frequency-specific auditory function. Principles and Applications in Auditory Evoked Potentials