This paper asked whether children fitted with bilateral hearing aids (BHA) develop normal perception of binaural cues which are the basis of spatial hearing. Data from children with BHA (n = 26, age = 12.6 ± 2.84 years) were compared to data from a control group (n = 12, age = 12.36 ± 2.83 years). Stimuli were 250 Hz click-trains of 36 ms and a 40 ms consonant-vowel /da/ at 1 Hz presented through ER3A insert-earphones unilaterally or bilaterally. Bilateral stimuli were presented at different interaural level difference (ILD) and interaural timing difference (ITD) conditions. Participants indicated whether the sound came from the left or right side (lateralization) or whether one sound or two could be heard (binaural fusion). BHA children lateralized ILDs similarly to the control group but had impaired lateralization of ITDs. Longer response times relative to controls suggest that lateralization of ITDs was challenging for children with BHA. Most, but not all, of the BHA group were able to fuse click and speech sounds similarly to controls. Those unable to fuse showed particularly poor ITD lateralization. Results suggest that ITD perception is abnormal in children using BHAs, suggesting persistent effects of hearing loss that are not remediated by present clinical rehabilitation protocols.
The objective of the present study was to assess the binaural hearing abilities of children with hearing loss whose access to sounds during development was provided through bilateral hearing aids. Because each of the two hearing aids functions independently, access to binaural cues is distorted, potentially compromising binaural hearing during important developmental periods.
It is clear that binaural hearing is essential to auditory development. Studies in children with unilateral hearing loss have consistently shown that, even if there is a normal hearing ear, sound localization (Bess et al., 1986; Murphy et al., 2011) and language development (Fischer and Lieu, 2014; Klee and Davis-Dansky, 1986; Lieu et al., 2010) are significantly impaired. Outcomes are similar in children with bilateral hearing loss who use only one cochlear implant to hear; this group lacks access to interaural level and timing cues and thus struggles with sound localization (Litovsky et al., 2006; Grieco-Calub et al., 2008; Van Deun et al., 2010: Johansson et al., 2019). Children spend much of their days in complex listening environments where adept sound localization helps them interact and develop. While unilateral hearing gives children access to most speech sounds, it has been well established that bilateral hearing, whether from two hearing ears or through the aid of hearing prostheses/devices (e.g., cochlear implants and hearing aids), with minimal delay is most beneficial to the development of binaural hearing in children (Litovsky et al., 2006; Polonenko et al., 2017; van Hoesel and Tyler, 2003).
Given the importance of binaural hearing in development, provision of the most appropriate auditory prosthesis in each ear as soon as possible has been recommended for children with hearing loss (Gordon et al., 2015). Nonetheless, bilateral auditory prostheses may not be providing access to accurate binaural cues. For many children with permanent bilateral hearing loss, current treatment focuses on fitting a hearing aid in each ear individually. There are clear guidelines for monaural hearing aid fitting (Seewald et al., 2005; Bagatto et al., 2010), but little attention is paid in this process to ensuring that consistent and accurate binaural cues are being provided by the two devices.
Both interaural level differences (ILDs) and interaural timing differences (ITDs), respectively, can be distorted by independent processing in bilateral hearing aids. The extent of the distortion to ILDs and ITDs depends on the frequency and azimuth position of the sound stimulus and on specific hearing aid settings (Osman et al., 2018) with particular effects of dynamic compression (Cubick et al., 2018; Hassager et al., 2017; Keidser et al., 2006), frequency compression (Brown et al., 2016), and adaptive directional microphones (Keidser et al., 2006; Van den Bogaert et al., 2011; Van den Bogaert et al., 2006) that work independently in each device. Synchronization between hearing aids has been developed to maintain one directional adaptation for both microphones and reduce distortions in interaural cues but with only minimal benefits (Ibrahim et al., 2012; Keidser et al., 2006). Deleterious effects of bilateral hearing aid use have been shown in behavioral tasks including binaural fusion (perception of one sound image), sound localization, and spatial release from masking (Cubick et al., 2018; Noble and Byrne, 1990; Wiggins and Seeber, 2012, 2011). These functional abilities are further compromised by hearing loss (Noble and Byrne, 1990; Van den Bogaert et al., 2006) and reverberant conditions (Hassager et al., 2017). Increased hearing aid experience can lessen these challenges in adults who had the benefit of developing binaural systems with normal hearing prior to hearing loss (Drennan et al., 2005).
Children are likely to be at greater risk for poor binaural hearing than adults with late onset hearing loss because they rely on bilateral hearing devices (hearing aids and cochlear implants) during sensitive periods in bilateral auditory development (Gordon and Kral, 2019). Poor perception of fundamental binaural cues in children using bilateral cochlear implants (Easwar et al., 2017a; Easwar et al., 2018; Steel et al., 2015) likely reflects deterioration of binaural processing at the level of the cortex, as measured by deficits in responses to binaural cues in the primary auditory cortex of deaf white cats provided with bilateral CIs (Tillein et al., 2010) and the inability of the bilateral devices to promote normal binaural development (Litovsky and Gordon, 2016). Indeed, children who received bilateral cochlear implants in the same surgery at young ages demonstrate higher than normal errors in sound localization tasks (Van Deun et al., 2010; Grieco-Calub and Litovsky, 2010; Killan et al., 2018), impaired perception of interaural cues (Steel et al., 2015; Grantham et al., 2008; Gordon et al., 2014), and poor cortical processing of binaural level or timing cues (Easwar et al., 2018; Easwar et al., 2017a,b).
Children using bilateral hearing aids could have advantages for developing binaural hearing relative to their peers using two cochlear implants. First, bilateral hearing aid users are less likely to experience mismatches in cochlear place of stimulation relative to bilateral cochlear implant users which impair binaural hearing (van Hoesel and Tyler, 2003; Hu and Dietz, 2015, Bernstein et al., 2018; Fitzgerald et al., 2015) and fusion (the ability to hear bilateral input as one fused image rather than separate sounds coming from each ear) (Kan et al., 2015; Steel et al., 2015). Second, with sufficient residual hearing, some children using hearing aids could have better access to the fine temporal structure of sounds than can be provided by cochlear implants.
In the present study, we tested the hypothesis that binaural hearing deficits in children using bilateral hearing aids are revealed as impairments in both sound lateralization and in binaural fusion.
Thirty-two children (14 girls, 18 boys) with bilateral hearing loss were recruited as part of study No. 1000002954, which was approved by the Research Ethics Board at the Hospital for Sick Children and adheres to the Tri-Council Policy on the Ethical Conduct for Research Involving Humans. The most recent audiograms relative to the time of testing are shown in Fig. 1.
The protocol for hearing aid fitting at our centre is to fit to desired sensation level (DSL) targets with real ear to coupler corrections. Six children with hearing loss were excluded from analyses due to incomplete testing or unreliable responses, leaving 26 children (12 girls, 14 boys) in the experimental group (mean age = 12.6 ± 2.84 years). These children had received their hearing aids at 3.82 ± 3.23 years of age. At testing, they had used bilateral hearing aids for 9.12 ± 3.77 years and were 12.6 ± 2.84 years of age. A control group of 32 children (18 girls, 14 boys) who reported that they had normal hearing were recruited (mean age of 12.9 ± 3.21 years). Audiometric pure tone thresholds were available in 12 children in the control group (seven girls, five boys, aged 12.36 ± 2.83 years at the time of testing) and analyses were restricted to these children. There was no significant difference in age between the hearing aid and control groups [t(18.9) = 0.25, p = 0.81]. As shown in Fig. 2, the pure tone average (PTA) of hearing thresholds at 500, 1000, and 2000 Hz were ≤20 dB hearing level (HL) in the control group. All participants had normal or corrected-to-normal vision and had no reported developmental deficits.
1. Bilateral stimuli
Responses to two different stimuli presented bilaterally were at 50% of the dynamic range (dB) and bilaterally balanced: (1) clicks presented at 250 Hz in a train of 36 ms delivered at 1 Hz, and (if the child was able to complete further testing), and (2) a 40 ms consonant-vowel (cv) /da/ construction synthesized by Skoe and Kraus (2010), also presented at 1 Hz. Both stimuli were presented through ER-3A insert earphones in both groups. The dynamic range for each of the two stimuli was defined in each ear from levels of maximum comfortable loudness (intensity increased in 2 dB step sizes to 120 dB or any behavioural indication of discomfort, whichever came first) to the threshold (determined using a bracketing procedure (Arlinger, 1979; American Speech-Language-Hearing Association, 1979). Data shown in Fig. 2 indicates a significant correlation between the audiometric PTA and the threshold levels for both the click and /da/ stimuli in both ears. The bilateral balance was determined by first presenting bilateral stimuli at 50% of the dynamic range (between minimum and maximum presentation intensities) in both ears. Levels were then increased or decreased in step sizes of 2 dB in the ear with the higher hearing threshold and the child was asked to indicate which side of the head they heard the sound coming from (left or right). Levels were considered balanced when the proportion of responses to the right decreased to 0.5 ± 0.2 across trials.
Interaural level differences were generated by increasing the intensity of the sound in one ear by the same dB as it decreased in the other. Six ILD conditions were presented: ±1.5, ±3, ±4.5 dB, where + indicates weighting of the level to the right ear and – to the left. Level balanced stimuli (ILD = 0) were also presented at six conditions of interaural timing difference: ±0.2, ±0.4, ±1 ms, where – indicates left leading and + indicates right-leading. Balanced stimuli (ILD = 0) were also presented simultaneously (ITD = 0).
B. Tests of binaural hearing
Perceived lateralization of sound was measured in all children in response to 13 conditions of bilateral stimuli (six ILDs, six ITDs, and one at ILD and ITD = 0). Children indicated whether they heard the bilateral input as coming from the left or the right side of their head by clicking on the respective buttons of a wireless mouse. Large blue “L” and red “R” stickers were placed on the two buttons of the mouse to avoid ambiguity. Stimuli were presented in a randomized block design; each block contained all bilateral conditions as well as two unilateral conditions presented at the same level (input to the left ear only or to the right ear only). Click train stimuli were always presented first. Stimuli were presented from 1 s after the previous response until the following response. Response decisions and reaction times were recorded for each stimulus presentation. Participants completed six blocks for each stimulus (click train and speech sound). Lateralization responses from participants were considered unreliable and were excluded if the proportion of accurate responses to unilateral conditions was <0.8 across trials (2 unilateral presentations × 6 blocks = 12 trials).
A measure of binaural fusion was also completed. Children indicated on the wireless computer mouse whether they heard one sound or two disjointed sounds. The mouse cover was changed to display one dot or two dots on the appropriate mouse buttons. Stimuli were presented in a randomized block design; each block contained the six ITD conditions presented in the bilateral lateralization task (±0.2, ±0.4, ±1 ms) as well as two control conditions (ITD = ±16 ms) (Babkoff and Sutton, 1966; Furst et al., 1985; Steel et al., 2015). Participants completed six blocks of ten (60 total trials) for each stimulus (click train and speech sound). Click train stimuli were always presented first. Responses and reaction times were recorded for each stimulus presentation. Fusion responses from participants were considered unreliable and were excluded if the proportion of accurate responses (not fused) to the two large ITD conditions was <0.8 across trials.
Of the 32 children recruited into the experimental group, complete and reliable data were collected in 26 children for at least one task. Of the 12 children in the control group whose data were included for analyses, all 12 provided complete and reliable data for at least one task.
1. Data analyses
The proportion of right-sided responses were calculated for each bilateral stimulus condition presented in the lateralization task. The proportion of “one” responses were calculated for all bilateral stimuli presented in the fusion task. Binary logistic regressions were calculated for these proportions across ILD and ITD separately for each stimulus in each child. For the fusion task, slopes at x = 0 were calculated separately for left and right leading ITDs as it was possible participants might fuse sounds from one side better than the other. Linear regression models were used to compare demographic predictors (age, age at hearing loss, stimulus thresholds, group) to ILD and ITD sensitivity for each stimulus. The Breusch-Pagan test was used to test for heteroscedasticity. All statistical analyses and plots were done in R (R Core Team, 2018) using the R-Studio IDE (RStudio Team, 2015).
A. Children with bilateral hearing loss have normal sensitivity to changing interaural level cues
Results of the lateralization task for the ILD conditions in each child, modeled by logistic regression, are plotted in Fig. 3(A). The mean [±1 standard error (SE)] proportion of right-sided responses are also shown. Responses to click-trains [Fig. 3(A), left panel] and /da/ stimuli [Fig. 3(A), right panel] reveal that children in both groups heard ILDs that were weighted from left ear to right as moving from the left to the right side of their head. Mixed model linear analyses indicated that children using bilateral hearing aids perceived ILD changes similarly to the controls for both stimuli [clicks: F(1,33) = 0.89, p = 0.35; /da/: F(1,14) = 1.23, p = 0.29]. Responses to the right at ILD = 0 were not significantly different between groups and not significantly different from 0.50 [t(13.5) = −1.22, p = 0.24 between groups, 95% confidence interval 0.46–0.57], confirming balanced bilateral input at this condition. The sensitivity to ILDs, as measured by the regression slopes, ranged from 0.17 to 3.33 (proportion of right response/dB) in both groups as shown in Fig. 3(B). Mean (±1 SE) reaction times across ILD conditions are shown in Fig. 3(C), revealing that responses by children with hearing aids were only slightly delayed relative to children without hearing aids for both stimuli. No significant differences in reaction time were found between groups for either stimulus [clicks: t(13.3) = 0.98, p = 0.34; /da/: t(28.3)= −1.61, p = 0.13].
B. Children with bilateral hearing loss have impaired perception of interaural timing cues
Results of the lateralization task for the ITD conditions in each child, modeled by logistic regression, are plotted in Fig. 4(A). The mean (±1 SE) proportion of right responses are also shown. Regression slopes, plotted in Fig. 4(B), were more variable in the hearing aid group than in the control participants and were significantly shallower [clicks: F(1,32) = 13.11, p = 0.001; /da/: F(1,14) = 21.51, p = 0.0004], reflecting poorer sensitivity to ITDs. As shown in Fig. 4(C), the children with hearing loss also took significantly longer [t(37.1) = 2.31, p = 0.03] to decide the side of the head to which ITD clicks were lateralized compared with normal hearing participants. Reaction time delays were also observed when ITDs were applied to the /da/ speech stimuli, but these differences did not reach statistical significance [t(13.3) = −1.60, p = 0.13].
C. Perception of interaural timing differences declines with increasing hearing loss
For each of the two stimuli, regression slopes for ILDs and ITDs were assessed for changes with age at testing, degree of hearing loss (defined by the threshold to each stimulus), and asymmetry in hearing (defined by the difference in thresholds between the ears to each stimulus). Figure 5(A) plots the ILD slopes against the stimulus threshold. There was no significant change in ITD slope with the increasing hearing threshold as shown by the dashed lines (all p > 0.05). Figure 5(B) plots ITD slopes against the stimulus threshold. As shown by the solid lines, ITD slope significantly decreased with the increasing hearing threshold for all conditions (left and right for both stimuli). In support, ITD slope significantly decreased with the pure tone hearing thresholds at 250, 500, 1000, 2000, 4000, and 8000 Hz (R2 ≥ 0.29, p < 0.001).
Figure 6 plots ILD and ITD slopes against hearing asymmetry, calculated as the difference in stimulus thresholds between the better and worse hearing ears. As shown, there are no significant effects of hearing asymmetry on ILD slopes, whereas ITD slopes significantly decrease with increasing hearing asymmetry.
D. Many children with bilateral hearing aids hear bilateral input as one sound (fused)
Responses from the fusion task were modeled by logistic regression for each participant. Analyses for ITDs leading to the left (negative values) were assessed separately from ITDs leading from the right (positive values) as in previous work (Steel et al., 2015). The resulting regression lines for each participant are plotted in Fig. 7(A). The intercept of the lines ranges widely from a proportion of “one sound” responses from 0 to 1.0, reflecting that some children judged the bilateral input to be fused (as one sound) more consistently than others. Slopes were not significantly different from 0 (mean −0.005, 95% confidence interval from −0.017 to 0.0068), indicating no change in binaural fusion judgments at different ITDs. The proportion of “one sound” responses was therefore averaged across the different ITD conditions for each participant for further analyses. Bilateral input across ITD conditions was consistently perceived as one fused sound for children with normal hearing [clicks: mean (±1 SE) = 0.97 ± 0.02 and /da/: mean (±1 SE) = 0.90 ± 0.05]. Figure 7(B) shows reaction times for both /da/ and click stimuli during the fusion task; mean responses were not significantly different in the group with hearing loss relative to control participants for both stimuli [/da/: t(9.79) = −0.77, p = 0.46. clicks: t(28.9) = 1.82, p = 0.08]. Reaction times were also assessed separately for the four children with poor fusion of the bilateral input, and these were not significantly different from the remaining children with better fusion [t(5.34) = 0.27, p = 0.79].
Although many of the hearing aid users perceived one fused click across ITDs (13/17, 76%), four of the children (4/17, 24%) consistently responded that they heard two separate sounds [proportion of “one” responses to clicks: mean (±1 SE) = 0.77 ± 0.09 and /da/: mean (±1 SE) = 0.97 ± 0.01] across ITDs, reflecting poor binaural fusion. In Fig. 8, the proportion of fused (“one sound”) responses to ITD stimuli were plotted against the lateralization slopes of ITD sensitivity for the 25 children who completed both tests. Lateralization slopes were significantly more shallow in the group with poor fusion than in the other children [F(1,30) = 6.66, p = 0.02].
This study aimed to investigate whether children with hearing loss who wear bilateral hearing aids have impaired binaural hearing. Bilateral hearing aid users were found to have normal perception of ILDs but abnormal perception of ITDs relative to a control group of similar aged children. Sensitivity to ITD changes was reduced in bilateral hearing aid users and they took longer than the control group to respond during these tasks. ITD sensitivity decreased across both groups as the hearing thresholds to the stimuli worsened and hearing became more asymmetric between the ears. A small group of bilateral hearing aid users struggled to fuse bilateral input in both the ITD lateralization and fusion tasks.
As seen in Fig. 3, bilateral hearing aid users were able to differentiate even small changes in ILDs in a lateralization task and were not significantly less accurate than normal hearing participants. Response times, shown in Fig. 3(C), also indicate that the bilateral hearing aid group did not take significantly longer than controls to choose the side of the head on which they heard these bilateral stimuli. As reviewed by Litovsky and Gordon (2016), children using bilateral cochlear implants rely heavily on ILDs for spatial hearing, but these cues are not sufficient for them to develop normal sound localization accuracy. Moreover, despite the ability to identify changes in ILDs, cortical responses to these cues reveal abnormalities in central ILD processing (Easwar et al., 2018). Thus, children who wear bilateral cochlear implants may be using unique strategies to detect and use ILDs for spatial hearing. The lateralization of ILDs with accuracy and response times similar to the control group in the present cohort of bilateral hearing aid users suggests that these children have some advantages for developing ILD sensitivity relative to their peers using bilateral cochlear implants. The present cohort using bilateral hearing aids had better hearing during development (albeit still abnormal) than children using bilateral cochlear implants and also had better access to acoustic sound bilaterally.
Evidence of impaired perception of interaural timing cues (ITDs) in the present cohort suggests continued impairment in the immature/developing binaural hearing system. If so, the advantages of better hearing and access to bilateral acoustic input alone have not been sufficient to counter these binaural impairments. As shown in Figs. 4(A) and 4(B), children using bilateral hearing aids showed decreased regression slopes for both stimuli compared to normal hearing controls when asked to lateralize bilateral stimuli with different ITDs. Moreover, they appear to struggle with this task, as shown in Fig. 4(C), by longer response times compared to controls for the click stimulus. Impaired access to these binaural cues could have functional effects as shown by impaired sound localization in children with hearing loss both with and without the use of hearing aids (Meuret et al., 2017; Johnstone et al., 2018; Johansson et al., 2019). To our knowledge, this is the first evidence of impaired ITD processing in children using bilateral hearing aids and, importantly, indicates that concerns about binaural processing in the immature/developing auditory system are not limited or particular to the use of bilateral pulsatile electrical stimulation in children delivered by cochlear implants (Litovsky and Gordon, 2016). Rather, deficits in binaural hearing likely occur as a result of hearing impairment in early life and are not completely reversed by even the early provision of bilateral hearing devices.
Impairments to binaural/spatial hearing appear to be of particular concern when they occur during development. ITD processing in adult bilateral cochlear implant users is poorer when the deafness was acquired in childhood rather than in adulthood (Litovsky and Godar, 2010). Adults with late-onset hearing loss had the advantage of binaural hearing during childhood and retain sensitivity to binaural cues (Spencer et al., 2016). Errors in sound localization impairments in adults with late-onset deafness thus likely arise mainly from distortions to binaural cues created by the device itself (as reviewed in Sec. I). Impairments of early-onset deafness include disruptions to the neural coding of binaural cues. Responses of cortical neurons to bilateral cochlear implant input and to ITDs, in particular, have been shown in congenitally deaf white cats (Tillein et al., 2010). Congenital deafness also affects subcortical processing of binaural cues with distinctly fewer neurons in the inferior colliculus remaining sensitive to ITDs compared to those in acutely deafened animals (Hancock et al., 2010). Similar to deafness-induced changes in the developing cortex, those neurons in the midbrain which did respond to cochlear implant evoked ITDs in the congenitally deaf cats were more broadly tuned and therefore less suitable for processing natural ITDs. It is important to note that the animals in these studies were completely deafened and thus more comparable to children with congenital deafness who receive cochlear implants than the present cohort of children whose residual hearing varied considerably. The benefits of residual hearing on binaural hearing during childhood are further supported by findings presented in Fig. 5; ITD sensitivity increased with better thresholds to the same stimuli. Thus, better bilateral hearing in early life is essential to developing binaural hearing.
It is becoming clear that binaural hearing development will also be impacted by asymmetric hearing in development (Gordon et al., 2013; Gordon et al., 2015; Litovsky et al., 2018; van Wieringen et al., 2018). Data shown in Fig. 6 further confirms this as ITD sensitivity in the lateralization task that significantly decreased with increasing differences in hearing thresholds between the ears. The concern is that an aural preference develops which disrupts the balance of inputs from each ear needed by the auditory system to detect interaural differences. Electrophysiological responses from children receiving bilateral cochlear implants sequentially or simultaneously are consistent with results from unilaterally deaf white cats, demonstrating the development of an aural preference syndrome as the delay to bilateral input lengthens (Kral et al., 2013; Gordon et al., 2013). The period of asymmetry in early life allows pathways from the stronger ear to become strengthened while pathways from the deprived ear are weakened (Kral et al., 2012), thus further deteriorating expected bilateral input to the auditory brainstem and cortices (Tillein et al., 2010). The effects are long-lasting, as measured in children with several years of bilateral hearing device use (Litovsky and Gordon, 2016). Indeed, even children who were experienced bilateral cochlear implant users have trouble perceiving changes in ITD (Gordon et al., 2014). Children listening with a hearing aid in one ear and a cochlear implant in the other (bimodal) find it particularly challenging to detect changes in ITD in a sound lateralization task (Polonenko et al., 2018; Zaleski-King et al., 2018) likely due to large mismatches in place and timing of stimulation between the ears (Polonenko et al., 2015).
Findings of the present study suggest that access to acoustic input in both ears through bilateral hearing aids does not protect the immature binaural auditory system from the effects of asymmetric hearing. This means that current fitting protocols through present hearing aid technology used by the present cohort have not adequately met the needs of these children to develop binaural hearing. It is already clear that bilateral hearing aids result in distortions to both ILDs and ITDs (Osman et al., 2018; Brown et al., 2016) due to mismatches in gain and other automatic processing strategies (Musa-Shufani et al., 2006; Ou et al., 2015; Wiggins and Seeber, 2012; Van den Bogaert et al., 2006). Present guidelines focus on monaural targets for hearing aids which leaves these mismatches unaddressed. Future studies are needed to determine which mismatches are most detrimental to binaural hearing and how more effective binaural settings might be achieved. The same concerns have been raised in bilateral cochlear implant users. Present cochlear implant programming software concentrates on setting threshold and comfort levels for each cochlear implant separately. This fitting strategy permits mismatches in interaural level, timing, and place of stimulation. These create additional asymmetries between the ears that amplify the negative consequences on binaural hearing (Kan et al., 2013; Fitzgerald et al., 2015; Hu and Dietz, 2015).
Asymmetries in bilateral input could affect the integration of bilateral input, compromising the ability to hear sounds presented to both ears as one fused image. This was tested in the present study by presenting bilateral input with different ITDs. As shown in Fig. 7(A), many (13/17, 76%) of the bilateral hearing aid participants judged a high proportion of these stimuli (both click trains and short speech stimuli) as one fused sound. Moreover, their proportion of “one” responses were not significantly different from the participants in the control group and, as shown in Fig. 7(B), their response times in this task were similar to the control group. Thus, children with bilateral hearing aids did not demonstrate the same impairments in binaural fusion previously shown by reduced proportion of “one” responses and increased response times in children with bilateral cochlear implants (Steel et al., 2015) and impaired binaural fusion in adults with bilateral cochlear implants (Fitzgerald et al., 2015). This suggests that bilateral electrical stimulation through cochlear implants is particularly difficult to integrate. Mismatches in place of stimulation generated by cochlear implants may play an overstated role in impairing binaural fusion explaining why the children who use bilateral hearing aids did not suffer the same deficits for the stimuli tested. Subtle differences in pitch between the ears might yet exist in children with hearing aids and could have actually broadened the pitch range which was binaurally fused as shown in adults with hearing loss (Reiss et al., 2017).
Although binaural fusion was similar to controls across the group of bilateral hearing aid users, data shown in Fig. 7 demonstrate that 4 of the 17 (24%) participants with hearing aids who completed this task were consistently unable to fuse ITDs. Analyses of this small group revealed no significant demographic differences from the remaining 13 bilateral hearing aid users (age at test, onset of deafness, and etiology of hearing loss). As shown in Fig. 8, ITD lateralization was consistently poor in these four children. The reverse, however, does not hold true, with several participants in the bilateral hearing aid cohort exhibiting relatively poor lateralization while maintaining normal fusion. Thus, binaural fusion appears to be required for ITD lateralization, as also shown in adult bilateral cochlear implant users (Kan and Litovsky, 2015) but does not guarantee normal ITD sensitivity.
V. SUMMARY AND CONCLUSIONS
Binaural hearing provides the foundation for spatial hearing, which is essential for children who spend much of their day in complex listening environments. Data from the present study demonstrates that children who wear bilateral hearing aids can integrate bilateral input into one fused image and retain sensitivity to ILDs. On the other hand, changes in ITDs remain difficult for these children to accurately lateralize; both accuracy and response times were significantly impaired relative to children in the control group. The ITD impairments increased with poorer hearing and increased asymmetry in bilateral hearing, confirming that access to ITDs can be disrupted by childhood hearing loss and may not be entirely restored with present bilateral hearing aid fitting and technology. Future efforts should concentrate on providing earlier and better access to accurate and consistent interaural timing cues through bilateral hearing aids in children with hearing loss. Future studies should also assess the contribution of fluctuating hearing loss due to such issues as middle ear infection on binaural hearing development in children with and without permanent hearing loss.
The authors would like to first and foremost acknowledge the time and effort of our participants who participated in this study. We would also like to thank Carmen McKnight for her work creating the stimuli and response systems and Melissa Polonenko for her help with statistics. Last, we wish to acknowledge the SickKids foundation which provided funding that helped make this research possible.