Introduction: Hearing loss is a major global public health issue that negatively impacts quality of life, communication, cognition, social participation, and mental health. The cochlear implant (CI) is the most efficacious treatment for severe-to-profound sensorineural hearing loss. However, variability in outcomes remains high among CI users. Our previous research demonstrated that the existing subjective methodology of CI programming does not consistently produce optimal stimulation for speech perception, thereby limiting the potential for CI users to derive the maximum device benefit to achieve their peak potential. We demonstrated the benefit of utilising the objective method of measuring auditory-evoked cortical responses to speech stimuli as a reliable tool to guide and verify CI programming and, in turn, significantly improve speech perception performance. The present study was designed to investigate the impact of patient- and device-specific factors on the application of acoustically-evoked cortical auditory-evoked potential (aCAEP) measures as an objective clinical tool to verify CI mapping in adult CI users with bilateral deafness (BD). Methods: aCAEP responses were elicited using binaural peripheral auditory stimulation for four speech tokens (/m/, /g/, /t/, and /s/) and recorded by HEARLab™ software in adult BD CI users. Participants were classified into groups according to subjective or objective CI mapping procedures to elicit present aCAEP responses to all four speech tokens. The impact of patient- and device-specific factors on the presence of aCAEP responses and speech perception was investigated between participant groups. Results: Participants were categorised based on the presence or absence of the P1-N1-P2 aCAEP response to speech tokens. Out of the total cohort of adult CI users (n = 132), 63 participants demonstrated present responses pre-optimisation, 37 participants exhibited present responses post-optimisation, and the remaining 32 participants either showed an absent response for at least one speech token post-optimisation or did not accept the optimised CI map adjustments. Overall, no significant correlation was shown between patient and device-specific factors and the presence of aCAEP responses or speech perception scores. Conclusion: This study reinforces that aCAEP measures offer an objective, non-invasive approach to verify CI mapping, irrespective of patient or device factors. These findings further our understanding of the importance of personalised CI rehabilitation through CI mapping to minimise the degree of speech perception variation post-CI and allow all CI users to achieve maximum device benefit.

Hearing loss is a major global public health issue that negatively impacts communication, social participation, cognition, psychological well-being, and quality of life [Punch et al., 2019; Harithasan et al., 2020]. The cochlear implant (CI) is a neural prosthesis used in the treatment of severe-to-profound hearing loss for individuals who receive limited or no benefit from conventional acoustic amplification devices [Buchman et al., 2020]. The CI electrode is inserted within the cochlea to bypass the non-functional structures by direct electrical stimulation of the auditory nerve [Naples and Ruckenstein, 2020]. The electrical impulses are processed along the auditory neural pathway and interpreted at the auditory cortex as sound [Naples and Ruckenstein, 2020].

CI mapping establishes the electrical dynamic range of each electrode to determine the minimum (T level) and maximum comfortable level (MCL) of electrical current necessary to stimulate the auditory nerve and generate sound detection without causing discomfort [Shapiro and Bradham, 2012; Vaerenberg et al., 2014]. It is crucial to avoid under- or over-stimulation. Under-stimulation reduces CI performance, which in turn may reduce user motivation [Martins and Goffi-Gomez, 2021], while discomfort as a result of over-stimulation may lead to the rejection of the CI entirely, yielding poor rehabilitation outcomes [Távora-Vieira and Marino, 2019]. Current programming procedures rely on patient-subjective psychoacoustic responses to ensure that the dynamic range of electrical stimulation simultaneously maintains good sound quality, loudness, and user comfort. However, obtaining reliable subjective responses is not practicable for all CI users and may pose a risk of under- or over-stimulation in patients with complex needs. This is often the case when programming devices for infants, young children, pre-lingually deafened individuals, or difficult-to-test patients who cannot reliably report loudness sensation [Vargas et al., 2013].

To address these limitations, several objective measures have been proposed to establish the dynamic range in CI users, including the electrically evoked stapedius reflex threshold and electrically evoked compound action potential [Abbas and Brown, 2015; de Vos et al., 2018; Kosaner et al., 2018]. Our own previous work has demonstrated the potential benefit of non-invasive recording acoustically-evoked cortical auditory-evoked potential (aCAEP) responses as an objective tool to verify subjective CI mapping in adult CI users [Távora-Vieira et al., 2018, 2022a, 2022b]. CAEP responses are assessed non-invasively using electroencephalography to record the synchronised activation of neurons in the auditory cortex in response to acoustic stimuli [Hillyard and Picton, 1978; Martin and Boothroyd, 1999]. The subsequent P1-N1-P2 complex waveform is characterised by positive and negative peaks labelled according to polarity (“P” for positive and “N” for negative deflections) corresponding to the site of generation [a detailed summary is provided in Lightfoot, 2016]. In adults, P1 is primarily generated in the primary auditory cortex, whereas N1 and P2 have multiple and more widely distributed generators in the primary and secondary auditory cortices, with typical latencies of approximately 50 ms, 100 ms, and 200 ms, respectively, post-stimulus onset [Lightfoot, 2016]. Given the cortical origin of CAEPs, these responses are theoretically more representative of conscious sound perception than responses recorded from peripheral structures such as electrically evoked stapedius reflex threshold or electrically evoked compound action potentials, or from sub-cortical structures such as the auditory brainstem response or auditory middle-latency responses [Kim, 2015]. As such, CAEP measures provide an ideal surrogate for subjective reporting of sound perception.

Our previous work has demonstrated that the amplitude and latency of the P1-N1-P2 complex are consistent between high-performing CI users with bilateral and single-sided deafness with normal hearing individuals [Távora-Vieira et al., 2018, 2022a, 2022b]. Furthermore, the presence of the P1-N1-P2 complex response is associated with higher speech perception scores in CI users [Távora-Vieira et al., 2022a, 2022b]. These studies validated the clinical use of aCAEP measures as an objective CI verification tool by demonstrating a significant improvement in speech perception scores in CI users who underwent aCAEP-guided CI map adjustments based on the presence of the P1-N1-P2 complex [Távora-Vieira et al., 2022a, 2022b]. However, while these findings reveal the benefit of objectively verifying CI mapping using aCAEP measures to improve speech perception outcomes, there is limited knowledge or understanding of the impact of variables that may limit the use of this method across all patient populations.

The present study was designed to further our understanding of the impact of patient- and device-specific factors on the use of aCAEP measurements as an objective tool to guide and verify CI mapping in adult CI users. The study aimed to identify potential predictive variables that may limit the use of aCAEP measures by investigating the correlation between aCAEP responses post-CI and the factors linked to suboptimal CI outcomes and inter-patient variability. Specifically, patient-specific factors were examined including age and sex, the nature, aetiology, and duration of hearing loss, age at CI intervention, side of implantation, CI experience, as well as device-specific factors, namely, electrode array, magnet, and audio processor type. Furthering our understanding of the factors that may impact objective CI verification is crucial to advance the personalised approach to verifying and optimising CI mapping in order to enhance the speech perception of the user. This can further facilitate all CI users to maximise their hearing outcome potential with their devices.

Ethics

This study was designed and conducted in accordance with the Declaration of Helsinki, and ethics approval was obtained from the South Metropolitan Area Health Service, Human Research Ethics Committee (Reference Number: 3258). Written informed consent was obtained from all individuals who met the inclusion criteria and agreed to participate.

Subjects

Adult (≥18–90 years of age) CI recipients with acquired post-lingual (>2 years of age) bilateral deafness (BD) were recruited from the Audiology department at the Fiona Stanley Hospital, Murdoch, Western Australia. The inclusion criteria were a severe or worse sensorineural hearing loss, defined as a pure-tone average at 0.5, 1, 2, and 4 kHz of ≥70 dB HL, in both the implanted ear (ipsilateral) and non-implanted ear (contralateral). All participants were oral communicators, with a required receptive language ability to understand the instructions given for testing. All participants presented with normal cognitive function and psychological status, which was confirmed by patient medical records.

aCAEP Response Recording

The HEARLab™ system (Frye Electronics, Inc., Beaverton, OR, USA) was used to generate the auditory stimuli and record aCAEP responses. Participant aCAEP responses were recorded for the three speech tokens /m/, /g/, and /t/. The HEARLab™ system also records responses for the speech token /s/. However, in our prior work, it was observed that the 50 ms duration of the /s/ stimulus can create an overlapping artefact that can interfere with aCAEP detection [Távora-Vieira et al., 2022a]. In light of this, we omitted analysis of the /s/ speech and classified participants according to the presence or absence of the P1-N1-P2 complex in response to only the /m/, /g/, and /t/ tokens.

Stimuli were presented in an anechoic free-field environment at 55 dB SPL using a loudspeaker placed at 0° azimuth angle, 1 m from the seated participant. Participant alertness during the recording was supervised by the testing clinician. The aCAEP responses were non-invasively recorded using a three-electrode montage, consisting of a non-inverting, inverting, and ground electrode, placed on the vertex (Cz), mastoid contralateral to the CI, and forehead, respectively. Electrode impedance was maintained below 5 kΩ and the residual noise level below 3.2 μV. Spectrum analysis verified the signal duration (ms) and frequency range (Hz) for each speech token.

HEARLab™ applies an automatic statistical criterion (Hotelling’s T2 statistical test) to determine the presence or absence of the aCAEP response by calculating the statistical significance of the difference between the signal and noise. A p value <0.05 is regarded as a detected response. The response was recorded from 200 ms before stimulus onset (baseline) to 600 ms post-stimulus onset. The minimum number of responses for each speech token was set at 200 epochs. Data reliability was confirmed by reproducibility of the recording. To validate the accurate identification of present and absent aCAEP responses by the HEARLab™ software, two experienced audiologists independently visually inspected the waveform morphology of the P1, N1, and P2 components. The aCAEP responses were deemed present only when both audiologists unanimously agreed on the presence of the P1-N1-P2 complex.

aCAEP Response Group Classification

The aCAEP response group classification procedure followed the protocol described by Távora-Vieira et al. [2022a]. Participants attended a standard follow-up appointment as per the post-CI protocol and CI-recipient rehabilitation program. The aCAEP protocol was performed for all participants using their preferred CI program, confirmed by the testing clinician as per device data logging information, where available.

For analysis, participants were divided into three groups based upon their aCAEP responses. In the first group, aCAEP responses were elicited for each speech token with the subjectively programmed CI map. This was regarded as optimal for speech detection, and no further CI map adjustments or testing was conducted. These participants were classified as subjectively optimised [Távora-Vieira et al., 2022a].

Participants with absent aCAEP responses for one or more speech tokens underwent CI map adjustments. This involved modifying the MCL for the electrode(s) corresponding to the frequency range of the absent speech token(s). As per fast Fourier transform spectrum analysis, the frequency ranges for the tokens were as follows: /m/: 200–500 Hz, /g/: 800–1,600 Hz, and /t/: 3,000–8,000 Hz.

Following CI map adjustments, the aCAEP protocol was repeated. The participants that exhibited present aCAEP responses for each speech token following CI map adjustments were deemed objectively optimised [Távora-Vieira et al., 2022a]. Participants that maintained absent aCAEP responses for one or more speech token(s), despite CI map adjustments, were classified as non-optimised [Távora-Vieira et al., 2022a].

Speech Perception Testing

The ability to perceive speech in quiet was measured using the Consonant-Nucleus-Consonant (CNC) word test [Peterson and Lehiste, 1962]. The CNC word test was performed in an anechoic free-field environment using a loudspeaker placed at 0° azimuth angle, 1 m from the seated participant. Baseline CNC word test scores were obtained preceding the aCAEP protocol for all participants. The CNC word test was repeated post-aCAEP and scores were recorded for group 2 and group 3 participants following CI map adjustments. To control for learning effects, the CNC word list was changed for each test condition.

Statistical Analysis

All statistical analyses were performed in the R programming language using RStudio. Normality tests were conducted using the Shapiro-Wilk test, on the basis of which non-parametric data analysis was conducted. The Kruskal-Wallis sum rank test was used to assess the statistical significance of differences between pre- and post-optimisation CNC scores, both within and between aCAEP response groups. The Kruskal-Wallis sum rank test was also used to test the statistical significance of differences between aCAEP response groups and patient pre- and post-operative factors, including participant age at testing, hearing loss onset and implantation, duration of hearing loss, and CI experience. The χ2 test of independence was used to test the significance of association between the aCAEP responses and patient demographic factors including sex, cause, and onset nature of hearing loss. The χ2 test of independence was also used to test the significance of association between aCAEP responses and device factors, including manufacturer, type of implant, electrode, and audio processor as well as the side of implantation. The Mann-Whitney-Wilcoxon test was performed as the non-parametric post-hoc test. To control for the risk of type 1 error, we implemented the Bonferroni correction to adjust the significance level for multiple comparisons. Specifically, we divided the original alpha level of 0.05 by the number of statistical tests conducted, resulting in an adjusted alpha level of 0.0167. Therefore, we considered p values less than or equal to 0.0167 as statistically significant.

Participants

A total of 132 adult (82 males; 50 females) CI users with BD participated in this study. Their demographic, aetiological, and CI-related data are summarised in Table 1. All participants were implanted with MED-EL electrode arrays: standard, FORM24, FLEX24, FLEXSoft, or FLEX28, and received one of the following speech processors: OPUS 2, RONDO, RONDO 2, SONNET, or SONNET 2 (MED-EL, Innsbruck, Austria). MCLs were measured using the subjective CI mapping procedure prior to participation in this study. Self-reported evidence of full-time CI usage during waking hours was cross-validated with device data logging information extracted from the audio processor, when available.

Table 1.

Summary of participant demographics, hearing loss characteristics, and device parameters

Total cohort
Total, n 132 
 Male 82 
 Female 50 
Nature of hearing loss, n 
 Birth 11 
 Sudden 53 
 Gradual 68 
Cause of hearing loss, n 
 Congenital 
 Genetic 
 Idiopathic sudden 14 
 Menière’s disease 12 
 Middle ear pathology 13 
 Noise-induced 14 
 Otosclerosis 
 Presbycusis 19 
 Trauma 
 Vascular 
 Virus 
 Unknown 29 
Implanted ear, n 
 Right 55 
 Left 77 
Average in years (±SD) 
 Age at test 65.02±16.2 
 Age at hearing loss onset 45.22±24.01 
 Age at implantation 64.45±15.98 
 Hearing loss duration 21.15±17.91 
 CI experience 1.9±2.52 
Implant, n 
 CONCERTO 23 
 SONATA 
 SYNCHRONY 100 
 SYNCHRONY 2 
Electrode, n 
 FLEX24 
 FLEX26 
 FLEX28 108 
 FLEXSoft 
 FORM19 
 FORM24 
 Standard 
Audio processor, n 
 OPUS 2 11 
 RONDO 
 RONDO 2 14 
 RONDO 3 
 SONNET 38 
 SONNET 2 60 
Total cohort
Total, n 132 
 Male 82 
 Female 50 
Nature of hearing loss, n 
 Birth 11 
 Sudden 53 
 Gradual 68 
Cause of hearing loss, n 
 Congenital 
 Genetic 
 Idiopathic sudden 14 
 Menière’s disease 12 
 Middle ear pathology 13 
 Noise-induced 14 
 Otosclerosis 
 Presbycusis 19 
 Trauma 
 Vascular 
 Virus 
 Unknown 29 
Implanted ear, n 
 Right 55 
 Left 77 
Average in years (±SD) 
 Age at test 65.02±16.2 
 Age at hearing loss onset 45.22±24.01 
 Age at implantation 64.45±15.98 
 Hearing loss duration 21.15±17.91 
 CI experience 1.9±2.52 
Implant, n 
 CONCERTO 23 
 SONATA 
 SYNCHRONY 100 
 SYNCHRONY 2 
Electrode, n 
 FLEX24 
 FLEX26 
 FLEX28 108 
 FLEXSoft 
 FORM19 
 FORM24 
 Standard 
Audio processor, n 
 OPUS 2 11 
 RONDO 
 RONDO 2 14 
 RONDO 3 
 SONNET 38 
 SONNET 2 60 

Age and duration summary data are represented in years (± standard deviation).

aCAEP Response Groups

All participants were grouped according to the presence or absence of the P1-N1-P2 aCAEP response to the three speech tokens, /m/, /g/, and /t/. For 63 participants (48%), aCAEP responses were elicited for all three speech tokens with the subjectively programmed map, classified as subjectively optimised. For 37 participants (28%), aCAEP responses were elicited for all three speech tokens after aCAEP-guided map optimisation, classified as objectively optimised (Fig. 1). In the remaining 32 participants (24%), aCAEPs could either not be elicited for at least one speech token by aCAEP-guided optimisation, or the participant did not accept the optimised CI map adjustments. These participants were classified as objectively non-optimised.

Fig. 1.

Comparison of acoustic cortical auditory-evoked potential (aCAEP) responses amplitude (µV) and latency (ms) to speech tokens (/m/, /g/, and /t/) pre- and post-cochlear implant (CI) aCAEP-guided map optimisation adjustments. Averaged responses at 55 dB SPL were recorded using the HEARLab™ system.

Fig. 1.

Comparison of acoustic cortical auditory-evoked potential (aCAEP) responses amplitude (µV) and latency (ms) to speech tokens (/m/, /g/, and /t/) pre- and post-cochlear implant (CI) aCAEP-guided map optimisation adjustments. Averaged responses at 55 dB SPL were recorded using the HEARLab™ system.

Close modal

Speech Perception

Speech perception scores were collected using the CNC speech test pre- and immediately post-aCAEP guided MCL optimisation were obtained during the same appointment from the objectively optimised and non-optimised groups. For the subjectively optimised group, their responses are classified as pre-optimisation. Data are shown in Figure 2. A comparison between groups pre-optimisation showed that the subjectively optimised group achieved a significantly higher average CNC score (mean 68.56% ± SD 14.96%) than the objectively optimised group (53.54% ± 16.30%; W = 259, p = 0.000405) and the non-optimised group (50.82% ± 13.09%) (W = 142, p value = 0.000151).

Fig. 2.

Boxplot depicts group speech perception scores (%) assessed using the Consonant-Nucleus-Consonant (CNC) word test in quiet. The box represents the interquartile range of speech perception scores (%), with the central line indicating the median. The whiskers extend to show the full range of the data, except for outliers which are shown as individual data points. Pre-optimisation scores are presented for the subjectively optimised group (n = 63), that did not undergo the MCL adjustment optimisation procedure. Pre- and post-optimisation speech perception scores are depicted for the objectively optimised (n = 37) and non-optimised (n = 32) participant groups that underwent MCL adjustments. A p value ≤0.0167 was considered statistically significant (*).

Fig. 2.

Boxplot depicts group speech perception scores (%) assessed using the Consonant-Nucleus-Consonant (CNC) word test in quiet. The box represents the interquartile range of speech perception scores (%), with the central line indicating the median. The whiskers extend to show the full range of the data, except for outliers which are shown as individual data points. Pre-optimisation scores are presented for the subjectively optimised group (n = 63), that did not undergo the MCL adjustment optimisation procedure. Pre- and post-optimisation speech perception scores are depicted for the objectively optimised (n = 37) and non-optimised (n = 32) participant groups that underwent MCL adjustments. A p value ≤0.0167 was considered statistically significant (*).

Close modal

No significant differences were observed between the pre-optimisation CNC scores of the subjectively optimised group (68.56% ± 14.96%) and the post-optimisation CNC scores of the objectively optimised group (68.63% ± 14.37%) (W = 729, p value = 0.574) or the post-optimisation CNC scores of the non-optimised group (57.25% ± 20.68%) (W = 125.5, p value = 0.179). No significant difference was observed between the average CNC scores of the objectively optimised and non-optimised groups, either at pre-optimisation (W = 178.5, p = 0.508) or post-optimisation (W = 104, p = 0.267).

The objectively optimised group demonstrated a significant increase in the average CNC score from pre-optimisation (53.54% ± 16.30%) to post-optimisation (68.63% ± 14.37%; W = 43, p = 0.002). Although the non-optimised group showed an increase in average CNC score from pre-optimisation (50.82% ± 13.09%) to post-optimisation (57.25% ± 20.68%), this increase was not statistically significant (W = 13, p = 0.547).

Pre- and Post-Operative Factor Analysis between aCAEP Groups

For the purpose of analysis, CI experience was defined as the time elapsed between the date of CI activation and the date that aCAEP testing was performed. Aetiologies of hearing loss were subdivided into classes of similar nature, namely, middle ear pathology, Ménière’s disease, sudden sensorineural hearing loss (SSNHL), progressive, and other. Hearing loss caused by middle ear pathology, surgery, and otosclerosis was included in the middle ear pathology subset. Trauma-induced, viral, and idiopathic causes of SSNHL were included in the SSNHLsubset. Age-related presbycusis, vascular, genetic, and noise-induced hearing losses were included in the progressivesubset. Hearing loss present at birth from unknown cause was included in the othersubset.

No significant differences were observed between the aCAEP response groups with regard to patient pre- and post-operative factors including patient age at the time of testing, hearing loss onset and CI intervention, the nature, cause, and duration of hearing loss, and CI experience and use status (Tables 2, 3). Similarly, no significant differences were observed between the aCAEP response groups with regard to device factors, including implant, electrode, or audio processor type (Table 3).

Table 2.

Statistical analyses using Kruskal-Wallis rank sum test to determine the statistical differences between aCAEP groups and patient-specific factors

FactorGroupTestStatisticdfp value
subjectively optimisedobjectively optimisednon-optimised
Age at test n 63 37 32 Kruskal-Wallis 0.1473 0.929 
Mean (±SD) 64.63 (±17.12) 64.89 (±15.51) 65.91 (±15.57)     
Median 70 67 68.5     
Age at CI n 63 37 32  0.44335 0.8012 
 Mean (±SD) 63.28 (±63.28) 64.85 (±64.85) 66.27 (±15.34)     
 Median 69.73 65.51 69.36     
Age at HL n 63 37 32  2.1761 0.3369 
 Mean (±SD) 43.05 (±22.56) 46.74 (±24.5) 47.74 (±26.5)     
 Median 50.11 53.58 55.88     
Duration of HL n 58 35 27  0.43497 0.8045 
 Mean (±SD) 21.98 (±18.09) 19.15 (±16.35) 21.97 (±19.83)     
 Median 20 18 13     
CI experience n 63 37 32  4.0351 0.133 
Mean (±SD) 2.41 (±2.87) 1.39 (±2.02) 1.47 (±2.15)     
Median 0.69 0.38 0.42     
FactorGroupTestStatisticdfp value
subjectively optimisedobjectively optimisednon-optimised
Age at test n 63 37 32 Kruskal-Wallis 0.1473 0.929 
Mean (±SD) 64.63 (±17.12) 64.89 (±15.51) 65.91 (±15.57)     
Median 70 67 68.5     
Age at CI n 63 37 32  0.44335 0.8012 
 Mean (±SD) 63.28 (±63.28) 64.85 (±64.85) 66.27 (±15.34)     
 Median 69.73 65.51 69.36     
Age at HL n 63 37 32  2.1761 0.3369 
 Mean (±SD) 43.05 (±22.56) 46.74 (±24.5) 47.74 (±26.5)     
 Median 50.11 53.58 55.88     
Duration of HL n 58 35 27  0.43497 0.8045 
 Mean (±SD) 21.98 (±18.09) 19.15 (±16.35) 21.97 (±19.83)     
 Median 20 18 13     
CI experience n 63 37 32  4.0351 0.133 
Mean (±SD) 2.41 (±2.87) 1.39 (±2.02) 1.47 (±2.15)     
Median 0.69 0.38 0.42     

A p value ≤0.0167 was considered statistically significant.

Table 3.

Statistical analyses using Pearson’s χ2 test to determine the statistical significance of differences between aCAEP groups and patient-specific factors

FactorGroup (n)TestStatisticdfp value
Subjectively optimisedObjectively optimisedNon-optimised
Sex 
 Male 39 22 21 Pearson’s χ2 test 0.27961 0.8695 
 Female 24 15 11     
Implanted ear 
 Right 34 23 20  0.9456 0.6233 
 Left 29 14 12     
Onset nature 
 Birth  0.5578 0.9676 
 Sudden 26 15 12    
 Gradual 31 19 18     
Cause of HL 
 MD  5.593 0.6927 
 MEP 11    
 Progressive 19 11 10    
 SSNHL 10    
 Other 15 11     
Implant 
 CONCERTO 16  10.487 0.1056 
 SONATA    
 SYNCHRONY 43 29 28    
 SYNCHRONY 2     
Electrode 
 FLEX24  16.313 12 0.1773 
 FLEX26    
 FLEX28 53 27 28    
 FLEXSoft    
 FORM19    
 FORM24     
 Standard     
Audio processor 
 OPUS 2  5.441 0.245 
 RONDO    
 RONDO 2    
 RONDO 3    
 SONNET 23    
 SONNET 2 20 21 19    
FactorGroup (n)TestStatisticdfp value
Subjectively optimisedObjectively optimisedNon-optimised
Sex 
 Male 39 22 21 Pearson’s χ2 test 0.27961 0.8695 
 Female 24 15 11     
Implanted ear 
 Right 34 23 20  0.9456 0.6233 
 Left 29 14 12     
Onset nature 
 Birth  0.5578 0.9676 
 Sudden 26 15 12    
 Gradual 31 19 18     
Cause of HL 
 MD  5.593 0.6927 
 MEP 11    
 Progressive 19 11 10    
 SSNHL 10    
 Other 15 11     
Implant 
 CONCERTO 16  10.487 0.1056 
 SONATA    
 SYNCHRONY 43 29 28    
 SYNCHRONY 2     
Electrode 
 FLEX24  16.313 12 0.1773 
 FLEX26    
 FLEX28 53 27 28    
 FLEXSoft    
 FORM19    
 FORM24     
 Standard     
Audio processor 
 OPUS 2  5.441 0.245 
 RONDO    
 RONDO 2    
 RONDO 3    
 SONNET 23    
 SONNET 2 20 21 19    

A p value ≤0.0167 was considered statistically significant.

This study aimed to assess patient- and device-specific factors that may affect the clinical use of aCAEP measures to objectively verify optimal CI mapping for speech perception. The results revealed that participants who exhibited aCAEP responses for all speech tokens after subjective mapping had significantly higher (p = 0.0001) speech perception scores than those who did not produce an aCAEP response for at least one token. This is consistent with previous research on CI users with BD [Távora-Vieira et al., 2022a] and single-sided deafness [Távora-Vieira et al., 2022b], supporting the notion that subjective loudness perception measures alone may not be adequate to achieve optimal speech perception outcomes for all CI users [Távora-Vieira et al., 2018].

When aCAEP-guided MCL adjustment was applied for users who did not produce an aCAEP response for at least one token, their scores immediately and significantly improved, reaching the same level as those of the subjectively optimised group. This immediate improvement suggests that enhanced speech perception can be achieved by providing sound access at cortical level, even without the recognised effects of patient learning and adaptation to CI map adjustments [Dornhoffer et al., 2022; Ma et al., 2022]. Nonetheless, it is likely that further prolonged CI usage may lead to even greater enhancement in speech perception.

In a subset of participants (32 individuals, 24% of the cohort), it was not possible to fully optimise the map via aCAEP-guided MCL adjustment, either because aCAEPs could not be elicited for all three speech tokens, or because the user did not accept the resultant map settings. This group also showed improved CNC speech scores, but the improvement was not statistically significant. Although this suggests partial CI map optimisation may improve speech perception outcomes, future investigations are necessary to identify the underlying reasons for why aCAEPs cannot be elicited in some users (other than those who rejected the new map). The underlying cause may be neurological, but whether this issue is peripheral or central in origin remains to be investigated.

The study found that 52% (n = 69) of experienced CI participants did not elicit an aCAEP response to at least one speech token, a lower proportion than that reported by Távora et al. [2022a], where 61% of their sample had absent aCAEP responses prior to CI map adjustments. This difference may be attributed to the variation in study design. The previous work used four speech tokens (/m/, /g/, /t/, and /s/). It was found that the /s/ token presented a challenge for detection, as its 50 ms stimulus duration can create an overlapping artefact with P1 [Távora-Vieira et al., 2022a]. Consequently, we excluded /s/ from the present study. Since in the previous study, /s/ was the token most likely to fail to elicit an aCAEP in CI users, its exclusion from the present study likely accounts for the lower proportion of users with one or more absent aCAEP responses.

In this study, the three groups did not differ significantly in terms of patient demographics (age and sex), pre-CI factors (age at hearing loss onset, duration, nature, and aetiology of hearing loss), post-CI factors (age at implantation, side of implant, duration of CI experience), and device factors (type of implant, electrode, and audio processor). As such, it can be concluded that none of these factors inhibit the ability to use aCAEP-guided MCL adjustment as an objective technique to verify CI mapping. This was unexpected, as some of these factors have previously been identified as being associated with poor CI outcomes. Duration of deafness and age at implantation, have been extensively investigated and are widely recognised as crucial determinants of CI outcomes [Blamey et al., 1996, 2013; Simon et al., 2020; Zhao et al., 2020]. However, the present investigation found no correlation between aCAEP response groups and patient age at testing or at implantation. This suggests that aCAEP can be utilised irrespective of age to ensure that older patients also receive optimised CI maps optimal for speech perception. It is well established that the duration of severe-to-profound deafness prior to implantation has a significant impact on CI outcomes, with longer deafness duration prior to implantation leading to poorer CI outcomes [Bernhard et al., 2021; Goudey et al., 2021]. However, the present study observed no correlation between aCAEP response groups and duration of hearing loss pre-CI. It is crucial to note that variability in the collection of self-reported symptomatic hearing loss data on onset and duration may limit the interpretation of these findings, as accurate reporting can be challenging, particularly in individuals with gradual hearing loss who cannot precisely define the onset of their hearing loss [Zhao et al., 2020].

The present investigation also explored the impact of CI experience on aCAEP responses. Previous research has shown that longer periods of CI use and rehabilitation facilitate adaptation and cortical response development over time [Glennon et al., 2020]. However, the present results revealed no correlation between CI experience and aCAEP groups, which is consistent with the findings of Távora-Vieira et al. [2022a], suggesting that objective CI mapping facilitates the elicitation of auditory cortex-level responses, even in the early stages of CI use.

Prior research has shown that right-ear implantation is associated with superior outcomes post-CI in post-lingually deafened adults [Liang et al., 2020; Goudey et al., 2021]. As such, the current study aimed to investigate the impact of implantation side on the use of aCAEP measures. In this study, no association was observed between the side of implantation and the ability to elicit aCAEP responses. In contrast, previous research has examined asymmetries in auditory-evoked potentials in response to monaural stimulation and found that the right auditory cortex may be more involved in processing monaurally presented tone and noise stimuli than the left auditory cortex [Hine and Debener, 2007].

We hypothesised that the nature (sudden or progressive) or aetiology of hearing loss could inhibit the ability to elicit aCAEP responses. However, our results did not support this hypothesis, rather the present findings indicate no correlation between the nature and cause of hearing loss and aCAEP response groups.

Previous studies have indicated that evoked potential recordings may be contaminated by CI stimulation artefacts [Gilley et al., 2006; Alemi et al., 2021]. No correlation was observed between the aCAEP response group and device factors, including the type of implant, magnet, or audio processor. This may be due to the limited sample consisting solely of one device manufacturer, and thus results may differ if a wider range of manufacturers were included.

Limitations

The current investigation utilised an automated system to record and determine the presence or absence of CAEP responses. This was done based on statistical detection, that is, the degree to which a response can be distinguished from noise. As such, we assessed the presence or absence of the CAEP waveform as a whole, and did not examine discrete waveform characteristics such as the amplitudes and latencies of the P1, N1, and P2 components. These characteristics warrant further investigation in future research.

The current results demonstrate that aCAEP measurements provide an objective, non-invasive, and personalised approach to verify and optimise CI mapping to enhance speech perception scores. Patient demographics (age and sex), pre-CI factors (age at hearing loss onset, duration, nature, and aetiology of hearing loss), post-CI factors (age at implantation, side of implant, duration of CI experience), and device factors (type of implant, electrode, and audio processor) do not significantly influence the ability to elicit aCAEP responses.

The authors thank Patrick Connolly who helped with English clarity and accuracy on a version of this manuscript.

Ethics approval was obtained from the South Metropolitan Health Ethics Committee (reference number: 3258). Participants have given their written informed consent to participate in this study.

The authors report no competing interests.

Dayse Távora-Vieira holds a research fellowship grant from the Rayne Medical Research Foundation. This project did not receive any other specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Caris Bogdanov and Dayse Távora-Vieira: design, data collection, drafting, interpretation, and final approval. Helmy Mulders and Helen Goulios: interpretation and final approval.

All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author.

1.
Abbas PJ, Brown CJ. Assessment of responses to cochlear implant stimulation at different levels of the auditory pathway. Hearing Res. 2015;322:67–76.
2.
Alemi R, Nozaradan S, Lehmann A. Free-Field Cortical Steady-State Evoked Potentials in Cochlear Implant Users. Brain Topography. 2021;34(5):664–80.
3.
Bernhard N, Gauger U, Romo Ventura E, Uecker FC, Olze H, Knopke S, et al. Duration of deafness impacts auditory performance after cochlear implantation: A meta-analysis. Laryngoscope Investig Otolaryngol. 2021;6(2):291–301.
4.
Blamey P, Arndt P, Bergeron F, Bredberg G, Brimacombe J, Facer G, et al. Factors affecting auditory performance of postlinguistically deaf adults using cochlear implants. Audiol Neurootol. 1996;1(5):293–306.
5.
Blamey P, Artieres F, Başkent D, Bergeron F, Beynon A, Burke E, et al. Factors affecting auditory performance of postlinguistically deaf adults using cochlear implants: an update with 2251 patients. Audiol Neurootol. 2013;18(1):36–47.
6.
Buchman CA, Gifford RH, Haynes DS, Lenarz T, O’Donoghue G, Adunka O, et al. Unilateral cochlear implants for severe, profound, or moderate sloping to profound bilateral sensorineural hearing loss: a systematic review and consensus statements. JAMA Otolaryngol Head Neck Surg. 2020;146(10):942–53.
7.
de Vos JJ, Biesheuvel JD, Briaire JJ, Boot PS, van Gendt MJ, Dekkers OM, et al. Use of electrically evoked compound action potentials for cochlear implant fitting: a systematic review. Ear Hear. 2018;39(3):401–11.
8.
Dornhoffer JR, Reddy P, Ma C, Schvartz-Leyzac KC, Dubno JR, McRackan TR. Use of auditory training and its influence on early cochlear implant outcomes in adults. Otol Neurotol. 2022;43(2):e165–73.
9.
Gilley PM, Sharma A, Dorman M, Finley CC, Panch AS, Martin K. Minimization of cochlear implant stimulus artifact in cortical auditory evoked potentials. Clin Neurophysiol. 2006;117(8):1772–82.
10.
Glennon E, Svirsky MA, Froemke RC. Auditory cortical plasticity in cochlear implant users. Curr Opin Neurobiol. 2020;60:108–14.
11.
Goudey B, Plant K, Kiral I, Jimeno-Yepes A, Swan A, Gambhir M, et al. A multicenter analysis of factors associated with hearing outcome for 2,735 adults with cochlear implants. Trends Hear. 2021;25:23312165211037525.
12.
Harithasan D, Mukari SZMS, Ishak WS, Shahar S, Yeong WL. The impact of sensory impairment on cognitive performance, quality of life, depression, and loneliness in older adults. Int J Geriatr Psychiatry. 2020;35(4):358–64.
13.
Hillyard SA, Picton TW. On and off components in the auditory evoked potential. Percept Psychophys. 1978;24(5):391–8.
14.
Hine J, Debener S. Late auditory evoked potentials asymmetry revisited. Clin Neurophysiol. 2007;118(6):1274–85.
15.
Kim J-R. Acoustic change complex: clinical implications. J Audiol Otol. 2015;19(3):120–4.
16.
Kosaner J, Van Dun B, Yigit O, Gultekin M, Bayguzina S. Clinically recorded cortical auditory evoked potentials from paediatric cochlear implant users fitted with electrically elicited stapedius reflex thresholds. Int J Pediatr Otorhinolaryngol. 2018;108:100–12.
17.
Liang C, Wenstrup LH, Samy RN, Xiang J, Zhang F. The effect of side of implantation on the cortical processing of frequency changes in adult cochlear implant users. Front Neurosci. 2020;14:368.
18.
Lightfoot G. Summary of the N1-P2 cortical auditory evoked potential to estimate the auditory threshold in adults. Semin Hear. 2016;37:1–8.
19.
Ma C, Fried J, Nguyen SA, Schvartz-Leyzac KC, Camposeo EL, Meyer TA, et al. Longitudinal speech recognition changes after cochlear implant: systematic review and meta-analysis. Laryngoscope. 2023;133(5):1014–24.
20.
Martin BA, Boothroyd A. Cortical, auditory, event-related potentials in response to periodic and aperiodic stimuli with the same spectral envelope. Ear Hear. 1999;20(1):33–44.
21.
Martins KVC, Goffi-Gomez MVS. The influence of stimulation levels on auditory thresholds and speech recognition in adult cochlear implant users. Cochlear Implants Int. 2021;22(1):42–8.
22.
Naples JG, Ruckenstein MJ. Cochlear implant. Otolaryngol Clin North Am. 2020;53(1):87–102.
23.
Peterson GE, Lehiste I. Revised CNC lists for auditory tests. J Speech Hear Disord. 1962;27(1):62–70.
24.
Punch JL, Hitt R, Smith SW. Hearing loss and quality of life. J Commun Disord. 2019;78:33–45.
25.
Shapiro WH, Bradham TS. Cochlear implant programming. Otolaryngol Clin North Am. 2012;45(1):111–27.
26.
Simon M, Campbell E, Genest F, MacLean MW, Champoux F, Lepore F. The impact of early deafness on brain plasticity: a systematic review of the white and gray matter changes. Front Neurosci. 2020;14:206.
27.
Távora-Vieira D, Marino R. Re-training the deaf ear: auditory training for adult cochlear implant users with singlesided deafness. Cochlear Implants Int. 2019;20(5):231–6.
28.
Távora-Vieira D, Wedekind A, Ffoulkes E, Voola M, Marino R. Cortical auditory evoked potential in cochlear implant users: an objective method to improve speech perception. PLoS One. 2022a;17(10):e0274643.
29.
Távora-Vieira D, Wedekind A, Marino R, Purdy SC, Rajan GP. Using aided cortical assessment as an objective tool to evaluate cochlear implant fitting in users with single-sided deafness. PLoS One. 2018;13(2):e0193081.
30.
Távora-Vieira D, Wedekind A, Voola M. Single-sided deafness: using cortical auditory evoked potential to improve cochlear implant fitting. Otol Neurotol. 2022b;43(9):e976–83.
31.
Vaerenberg B, De Ceulaer G, Szlávik Z, Mancini P, Buechner A, Govaerts PJ. Cochlear implant programming: a global survey on the state of the art. Sci World J. 2014;2014:646590.
32.
Vargas JL, Sainz M, Roldan C, Alvarez I, de la Torre A. Analysis of electrical thresholds and maximum comfortable levels in cochlear implant patients. Auris Nasus Larynx. 2013;40(3):260–5.
33.
Zhao EE, Dornhoffer JR, Loftus C, Nguyen SA, Meyer TA, Dubno JR, et al. Association of patient-related factors with adult cochlear implant speech recognition outcomes: a meta-analysis. JAMA Otolaryngol Head Neck Surg. 2020;146(7):613–20.